A GRAMMAR OF THE HITTITE LANGUAGE

Part 1: Reference Grammar

00-HittiteGrammar.indb i 6/5/08 4:13:57 PM
LANGUAGES OF
THE ANCI ENT NEAR EAST
Editorial Board
Gozio Rtnro, Pennsylvania State University
Editor-in-Chief
Jxrs P. Aiir Brown University
Grr B. Gnoo The Oriental Institute, Univ. of Chicago
Jon Htrnrnonu Harvard University
Mrnru Knrnrnrx Friedrich-Schiller-Universitt Jena
Aforo Lornrro Universitt Basel
H. Cnro Mricnrnf University of California, Los Angeles
Profn Mrcniowsxr University of Michigan
P. Oxfon Sx+nv Harvard University
Mrcnri P. Sfnrcx Universitt Leipzig
1. A Grammar of the Hittite Language, by Harry A. Honer Jr. and H. Craig Melchert
Part 1: Reference Grammar
Part 2: Tutorial
00-HittiteGrammar.indb ii 6/5/08 4:13:58 PM
A Grammar of the
Hittite Language
Part 1
Reference Grammar
by
Hnn. A. Horrrn, Jn.
The University of Chicago
and
H. Cnro Mricnrnf
The University of California, Los Angeles
Winona Lake, Indiana
Ersrnnts
2008
00-HittiteGrammar.indb iii 6/5/08 4:13:58 PM
2008 by Harry A. Honer, Jr., and H. Craig Melchert
All rights reserved
Printed in the United States of America

www.eisenbrauns.com


































2 3 4 5 6 7 8 9
20 19 18 17 16 15 14 13 12 11 10 09
Library of Congress Cataloging-in-Publication Data
Honer, Harry A.
A grammar of the Hittite language / by Harry A. Honer Jr. and H. Craig Melchert.
p. cm. (Languages of the ancient near east ; 1)
Includes bibliographical references.
ISBN 978-1-57506-119-1 (hardcover : alk. paper)
1. Hittite languageGrammar. I. Melchert, H. Craig (Harold Craig), 1945
II. Title.
P945.H59 2008
491.998 dc22
2008018385
To three giants of Hittitology,
in whose shadow we stand
Hs Gtsfv Gtfrnnocx
Exxtri Lnocnr
Hrrnrcn Offr
00-HittiteGrammar.indb v 6/5/08 4:13:59 PM
00-HittiteGrammar.indb vi 6/5/08 4:14:00 PM
vii
CONTENTS
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
Abbreviations and Conventional Markings . . . . . . . . . . . . . . . . xvii
General Abbreviations xviii
Bibliographical Abbreviations xx
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
The Hittites and Their Language 1
Decipherment 1
The Text Corpus 2
Modern Resources for Study 3
Cuneiform Editions 3
Commentaries 4
Sign Lexicon 4
Grammars 5
Dictionaries 5
Text Catalogues and Name Collections 7
Chapter 1. Orthography and Phonology . . . . . . . . . . . . . . . . . . 9
The Cuneiform Writing System of the Hittites 9
Orthography 10
Writing Conventions 10
Transcriptional Conventions 14
Homophony 15
Polyphony 16
Multivalence 21
New Values 21
Logogram Pronunciation 22
Phonology 24
Individual Phonemes (Vowels and Consonants) 24
Vowels 25
Consonants 35
Phonotactics 48
Onomatopoeia 49
Accent 49
Chapter 2. Noun and Adjective Formation . . . . . . . . . . . . . . . . 51
Underived Stems 52
Derived Stems 53
Reduplicated Nouns and Adjectives 62
Compounded Nouns and Adjectives 63
00-HittiteGrammar.indb vii 6/5/08 4:14:00 PM
viii
Chapter 3. Noun and Adjective Inection . . . . . . . . . . . . . . . . . 64
Inectional Endings 69
The Basic Scheme 69
Common-Gender Nominative 70
Common-Gender Accusative 71
Neuter Nominative-Accusative 71
Ergative 72
Genitive 73
Dative-Locative 74
Vocative 74
Allative 76
Ablative 76
Instrumental 77
Stem Variation 78
Chapter 4. Noun and Adjective Declension . . . . . . . . . . . . . . . . 79
a-Stem Nouns 79
Common-Gender a-Stem Nouns 79
Neuter a-Stem Nouns 82
a-Stem Adjectives 83
i- and u-Stem Nouns 85
i-Stem Nouns 86
i-Mutation 86
Common-Gender i-Stem Nouns 87
Neuter i-Stem Nouns 90
ai-Stem Nouns 92
i-Stem Adjectives 94
u-Stem Nouns 97
Common-Gender u-Stem Nouns 98
Neuter u-Stem Nouns 101
au-Stem Nouns 101
u-Stem Adjectives 102
e-Stem Noun 104
Consonantal-Stem Nouns and Adjectives 105
-Stem Noun 105
l-Stem Nouns 105
al-Stem Nouns 106
elil-Stem Nouns 107
ul-Stem Nouns 107
n-Stem Nouns 108
Neuter n-Stem Nouns 108
Common-Gender n-Stem Nouns 111
r-Stem Nouns and Adjectives 114
s-Stem Nouns 117
t-Stem Nouns and Adjectives 120
Simple t-Stem Nouns 121
nt-Stem Nouns and Adjectives 123
rln-Stem Nouns 124
Contents
00-HittiteGrammar.indb viii 6/5/08 4:14:01 PM
ix
Non-Derived rln-Stem Nouns 124
Derived Nouns with Sux -tar 126
Derived Nouns with Sux -ear 128
Derived Nouns with Sux -war 129
Derived Nouns with Sux -mar 130
Verbal Substantives 130
Irregular Consonant-Stem Nouns 130
Chapter 5. Personal Pronouns . . . . . . . . . . . . . . . . . . . . . . 132
Distinctive Features of Pronominal Stems and Endings 132
Accented (Independent) Personal Pronouns 133
Enclitic Personal Pronouns 135
Chapter 6. Possessive Pronouns . . . . . . . . . . . . . . . . . . . . . 137
Chapter 7. Deixis: The Demonstratives . . . . . . . . . . . . . . . . . . 142
k- and ap- 143
ai, uni, ini 144
anni- 146
Adverbs Built to Demonstratives 147
Chapter 8. Relative and Indenite Pronouns . . . . . . . . . . . . . . . 149
The Interrogative and Relative Pronoun kui- 149
The Indenite Pronoun kuiki and the Distributive kuia 149
Other Stems with Partial Pronominal Inection 152
Chapter 9. Numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
Form and Declension of the Cardinals 153
One 154
Two to Ten 156
The Syntax of the Cardinals 158
Agreement in Case 158
Number Agreement with Non-Collectives 158
Counting Non-Decimal Sets 163
Word Order in Counting 165
Fractions 165
Ordinal Numbers 166
Ordinals in -t- 167
Multiplicatives 168
Distributives 170
Derivative Adjectives and Verbs 170
Numbers in Compounds 171
Compound Numbers 171
Chapter 10. Verb Formation . . . . . . . . . . . . . . . . . . . . . . . . 173
Reduplicated Roots 173
Verbal Suxes and Inxes 175
Overview 175
The Individual Suxes 175
Contents
00-HittiteGrammar.indb ix 6/5/08 4:14:01 PM
x
Chapter 11. Verb Inection . . . . . . . . . . . . . . . . . . . . . . . . 180
Types of Verbal Inection 180
Irregularities in the mi-Conjugation 183
Irregularities in the i-Conjugation 184
Medio-Passive Endings 184
Verbal Substantive, Participle, Innitive, and Supine 185
Chapter 12. Conjugation of mi-Verbs . . . . . . . . . . . . . . . . . . . 187
Consonantal Stems 187
Root Stems 187
Axed Stems 196
Vocalic Stems 198
Ablauting 198
Non-Ablauting 210
Suppletive Verb 211
Verbs with Mixed Stems 212
Chapter 13. Conjugation of i-Verbs . . . . . . . . . . . . . . . . . . . 214
Consonantal Stems 214
Vocalic Stems 218
a-Stems 218
Stems Originally in -Cw- 222
ai-Stems 222
Stems with Mixed Inection in -a- and -i- 224
Verbs with the Imperfective Sux -annali- 226
Mixture of mi- and i-Forms 227
Chapter14. Medio-Passive Conjugation . . . . . . . . . . . . . . . . . 230
Consonantal Stems 230
Vocalic Stems 231
Chronology of the Medio-Passive Endings 233
Medio-Passive Stem Formation 233
Chapter 15. Grammatical Agreement . . . . . . . . . . . . . . . . . . . 235
Types of Agreement 235
Lack of Agreement 239
In Gender 239
In Number 240
Chapter 16. Noun Cases . . . . . . . . . . . . . . . . . . . . . . . . . . 242
Generalities 242
Nominative 242
Vocative and Naming Construction 244
Other Forms of Direct Address 245
Accusative 246
Ergative 250
Genitive 250
Word Order in a Genitival Phrase 254
Dative-Locative 257
Indirect Object 257
Contents
00-HittiteGrammar.indb x 6/5/08 4:14:02 PM
xi
Indicating Possession 258
Dative of Disadvantage 258
Goal 259
Purpose or Result 260
Location 260
Temporal Uses 261
Units of Measure and Dimensions 262
Additive-Incremental 262
Allative 262
Ablative 264
Instrumental 268
Cases with Particular Verbs 270
Chapter 17. Adjectives . . . . . . . . . . . . . . . . . . . . . . . . . . 271
Comparison of Adjectives 273
Comparative Degree 273
Superlative Degree 275
Chapter 18. Pronouns . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
Independent Personal Pronouns 277
Enclitic Personal Pronouns 279
(Third-Person) Subject Clitics 280
Intransitive Verbs with Subject Clitics 280
Intransitive Verbs without Subject Clitics 281
Other Clauses without Referential Subjects 282
Special Cases 282
Demonstrative Pronouns 283
k-, ap-, and ai : Word Order 283
Declinable k . . . k as a Distributive Expression 285
Correlated k-Forms Having Dierent Cases 286
Other Deictic Elements 286
The Indenite Pronoun kuiki 286
tamai- 287
Chapter 19. Adverbs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
Local Adverbs 289
Temporal Adverbs 290
Adverbs of Manner 291
Denominative Adverbs 291
The Adverbial Sux -ili 292
Adverbial Circumlocutions 293
Chapter 20. Local Adverbs, Preverbs, and Postpositions . . . . . . . . . 294
Free-standing Adverbs 295
Preverbs 295
Postpositions 297
Chapter 21. Verb Voice . . . . . . . . . . . . . . . . . . . . . . . . . . 302
Active Voice 302
Middle Voice 303
Passive Voice 303
Contents
00-HittiteGrammar.indb xi 6/5/08 4:14:03 PM
xii
Chapter 22. Verb Tense . . . . . . . . . . . . . . . . . . . . . . . . . . 306
Present 306
Preterite 309
The Analytic Perfect Construction 310
Function 310
Form 310
Uniqueness within the Old Anatolian Indo-European Group 310
Examples 311
Stative Constructions with ar (k)- and - 311
Chapter 23. Verb Mood . . . . . . . . . . . . . . . . . . . . . . . . . . 313
Indicative 313
Imperative 313
Optative, Potential, and Irrealis 314
Optative 315
Potential and Irrealis 315
Chapter 24. Verb Aspect . . . . . . . . . . . . . . . . . . . . . . . . . 317
Imperfectives 317
Adverbial Markers 317
Sux Markers 318
Nuances of Imperfective Aspect 318
Choice of Sux 322
Sux Redundancy 323
ka (tta) and kma 323
The Serial Use of pai- to go and uwa- to come 324
Chapter 25. Non-Finite Verb Forms . . . . . . . . . . . . . . . . . . . 330
Verbal Substantive 330
Innitive 332
Supine 338
Participle 339
Chapter 26. Negation . . . . . . . . . . . . . . . . . . . . . . . . . . . 341
natta 341
Word Order with natta 341
Negative Rhetorical Questions (with Negatives of Assertion) 342
nawi not yet 343
Imperatival and Categorical Negative l 344
lman and nman 344
Categorical Negative l 345
Word Order with l 345
nelikku 345
Double Negatives 346
Carry-Over of Negative Force 346
Summary 346
Contents
00-HittiteGrammar.indb xii 6/5/08 4:14:03 PM
xiii
Chapter 27. Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
Intonation Marking 348
YesNo Questions 348
Direct Questions with Interrogative Pronouns, Adjectives,
and Adverbs 350
Questions Posing Alternatives 353
Indirect Questions 353
Chapter 28. Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . 354
The Particle -wa (r-) 354
The Particle -z (a) 357
Form 357
Reexive Function 357
Lexical Use of -z (a) 360
Use in Nominal Sentences and with the Verb To Be 362
The Local Particles -an, -apa, -ata, -kan, and -an 364
The Particle -kan 366
The Local Value 366
Accompanying the Dative of Disadvantage 371
The Terminative Value 372
With Verbs of Hostility 372
Aspect Independent of Verbs Lexical Meaning 372
Absence of Expected -kan 373
Summary 373
The Particle -an 374
Co-occurring with the Adverb er 374
Physical or Conceptual Movement toward an Object 375
The Particle -an 377
The Particle -apa 378
The Particle -ata 382
The Particle -pat 384
Anaphoric 385
Particularizing 386
Restrictive 387
Contrastive 388
Chapter 29. Conjunctions . . . . . . . . . . . . . . . . . . . . . . . . . 389
Clause-Initial Clause-Linking Conjunctions 390
Distribution 390
nu 390
u 392
ta 393
Enclitic Clause-Linking Conjunctions 395
-al-ma 395
-al-ya 399
Asyndeton 401
Form 401
Function 402
Contents
00-HittiteGrammar.indb xiii 6/5/08 4:14:04 PM
xiv
Disjunction 405
nau . . . na (u)ma 405
-(a) ku . . . -(a) ku 405
Chapter 30. Clauses . . . . . . . . . . . . . . . . . . . . . . . . . . . . 406
Word Order 406
Major Constituents 406
Deletion Processes 409
Sentential Clitics 410
Nominal and To Be Sentences 412
Dependent Clauses 414
Temporal Clauses 416
Causal Clauses 418
Concessive Clauses 419
Conditional Clauses 419
Simple Conditions 420
Contrary-to-Fact Conditions 422
Relative Clauses 423
Indirect Statements and Questions 426
Multiple Dependent Clauses 428
Chapter 31. Sumerian and Akkadian . . . . . . . . . . . . . . . . . . . 430
Sumerograms 430
Nouns, Adjectives, and Participles 430
Pronouns 430
Verbs 430
Akkadian Grammar 430
Orthography 431
Superscripting 432
Phonetic Complements 433
Phonology 433
Morphology 434
Nouns and Adjectives 434
Pronominal Suxes 437
Numbers 438
Verbs 438
Conjunctions 440
Calques 440
Prepositions 441
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 442
Contents
00-HittiteGrammar.indb xiv 6/5/08 4:14:04 PM
xv
PREFACE
This book originated in supplementary teaching materials developed by the authors
for classroom instruction. Several generations have learned Hittite with the help of the
outstanding Hethitisches Elementarbuch of Johannes Friedrich (second edition: 1960).
However, the passage of more than 40 years has inevitably rendered parts of Friedrichs
grammar outdated or incomplete. A number of recent works have tried to address cur-
rent instructional needs. Our own eorts to procure or produce teaching materials for the
classroom led us to conclude that the time had come for a more comprehensive reference
grammar of Hittite, along the lines of Wolfram von Sodens Grundri der akkadischen
Grammatik (latest posthumous edition in 1995), that could also serve as the basis for an
accompanying set of graded lessons for language learners.
No descriptive grammar can ever pretend to be truly denitive. We have done our
best to incorporate and synthesize the advances made since the work of Friedrich and to
make this grammar as broad and up-to-date in coverage as possible. We ask indulgence
in advance for the inevitable omissions and inadequacies. On points where there is no
consensus, we have not hesitated to make reasoned choices, while striving to acknowl-
edge dierent points of view. Our primary goal has been to describe the language sys-
tematically as it appears in the extant texts. We have referred to prehistoric factors only
where we feel that they help elucidate features of attested Hittite or are of broad interest.
We expressly disavow any intent of systematic coverage in this regard.
One of the most dramatic changes in Hittitology since 1960 has been our enhanced
ability to establish a relative chronology not only of texts (recognized since the earliest
days of the eld) but also of individual copies of those texts, often written many years
after the texts composition. We have sought to give full recognition to established nd-
ings in this area, but our initial intention of assigning Old, Middle, or New Hittite status
to all cited forms in the paradigms proved to be overly optimistic. Many issues about the
dating of texts and manuscripts remain unresolved. We have therefore limited ourselves
to marking consistently only examples assured as Old Hittite by their appearance in cop-
ies from the Old Hittite period (OS = Old Script). We have otherwise been selective in
making what we take to be valid generalizations about the date of various phenomena
in appropriate passages in the grammar. Further renements must be left for the future,
including in installments of the ongoing major lexica.
The tutorial is a series of graded lessons arranged in a typical fashion. Major morpho-
logical categories are introduced a few at a time, along with a limited but representative
sample of the lexicon. Each lesson has illustrative sentences suitable for practice in
translation. In order to avoid inventing more Hittite sentences than absolutely necessary,
we have insofar as possible used for the exercises Hittite sentences that actually occur in
00-HittiteGrammar.indb xv 6/5/08 4:14:05 PM
xvi xvi
the texts in either their original form or slightly adapted. We have keyed the tutorial to
the reference grammar and have provided extensive notes for the exercise sentences, es-
pecially on matters of syntax, but we have designed the tutorial primarily for classroom
use with an instructor who knows the language. Although some readers may be able to
use the tutorial for self-instruction, we cannot give assurance that such a method will
produce satisfactory results.
Our enormous overall debt to scholars past and present should be apparent through-
out. We are indebted to colleagues too numerous to mention for their prompt sending of
copies of published and unpublished works and responses to queries. We wish to thank
in particular Professors John A. Brinkman, Benjamin Fortson, Theo van den Hout, Jay
Jasano, Jared Klein, Norbert Oettinger, and Elisabeth Rieken for reading all or part
of an earlier draft of the grammar and oering innumerable helpful suggestions and
criticisms. The present version has been immeasurably improved due to their eorts.
Nevertheless, since we were not able to follow all their suggestions, they cannot be
accountable for whatever errors, omissions, or infelicities remain. For these we alone
are responsible. We would also like to thank Mr. Aaron Butts of Duke University for
working through an earlier draft of the tutorial and suggesting changes in it, and the
LANE series editor, Professor Gonzalo Rubio, for further helpful suggestions. Last but
not least, both authors are grateful to Winifred Honer for the hospitality, patience, and
unfailing good humor which she has shown to us during the long and sometimes trying
gestation period of this book and for standing together with us as loyal, long-suering
Chicago Cubs baseball fans!
The Authors
Christmas, 2005
Preface
00-HittiteGrammar.indb xvi 6/5/08 4:14:05 PM
xvii
ABBREVIATIONS AND CONVENTIONAL MARKINGS
In general, and unless otherwise noted in the remarks below, we follow the system of
abbreviation used in the Hittite Dictionary of the Oriental Institute of the University of
Chicago (CHD). For further details, see below under Bibliographical Abbreviations.
As we describe in more detail in the Introduction, the Hittite language can be re-
garded as developing in three stages: Old Hittite (ca. 16501450 n.c.), Middle Hittite
(ca. 14501350), and New Hittite (ca. 13501190). As in the CHD, we employ the sigla
OH, MH, and NH for Old, Middle, and New Hittite, and following a slash (l) indicate
the date of the copy with OS, MS, and NS for Old, Middle, and New Hittite Script.
Text datings follow the format of the CHD (e.g., OHlMS, etc., rather than the German
system of using ah and mh to denote the date of the individual copy). OS used by
itself implies OHlOS, NH by itself implies NS. But the datings themselves may dier
from those of earlier printed CHD volumes. Wherever possible, we seek to conform to
the datings of the copies now used in the on-line Konkordanz of the Mainz center for
Hittitological research.
In citing Hittite text references, joined pieces published in separate places are noted as
KUB 24.5 + KUB 9.13 obv. 31 or just as KUB 24.5+, which implies that there is a join
but does not specify its identity.
The slanted equal sign () in Hittite transcriptions indicates a boundary before a clitic
element. Since there is no convenient transliteration for the marker wedges that Hittite
scribes prexed to forms exhibiting foreign or unusual elements, we follow the CHD
practice of using a graphic representation of the one- or two-wedge variants: and .
Unlike verbs in the Semitic languages, Hittite verbs with third-person subjects are
indierent to the biological gender of their subjects. To avoid ugly renderings such as
helshelit . . . -s in our translations, when the actual gender of the subject is unknown,
we have arbitrarily used the masculine pronoun he.
In the paradigms the following conventions are followed.
Parentheses within forms may mark optional modes of writing the word: e.g., -az(a)
means that the ablative ending can be written either -Ca-za or -Ca-az; annaz(a) means
that the ablative of anna- mother can be written either an-na-az or an-na-za; wal(a)zi
indicates that this form can be written either wa-al-a-zi or wa-la-a-zi; pau(e)na
signals the existence of spellings pauena and pauna; a(u)wari(y)a signals the
possible spellings a-u-wa-ri-ya-a, a-u-wa-ri-a, a-wa-ri-ya-a, etc.; ui(u)wanza sig-
nals the existence of u-i-wa-an-za as well as u-i-u-wa-an-za. This applies also to
optional plene spellings (e.g., putezzi). The same convention is used with transliter-
ated forms: ar-re-(e-)ez-zi.
Parentheses may also occasionally mark a speech element lost through a regular
sound change but restored in our broad transcription to aid lexical andlor grammati-
00-HittiteGrammar.indb xvii 6/5/08 4:14:06 PM
abl. ablative
abbr. abbreviated, abbreviation
acc. accusative
act. active
adj. adjective
adv. adverb
Akk. Akkadian
all. allative
C consonant (in CV, CVC, VC)
CLuw. Cuneiform Luwian
col. column
coll. collective
cal identication: e.g., Old Hittite at-ta-a-a-an his father (acc.) will be transcribed
atta(n)an, la-a-am-ma-a-mi-it as lmm(n)mit, Old Hittite an-da-ma-pa, when the
form contains the clitic pronoun -mu, will be transcribed andam(u)apa (1.72, p. 32;
and 28.100, p. 378). Likewise we clarify instances of simplied spellings (1.10,
p. 12), where a nongeminate consonant represents two identical consonants straddling
a clitic boundary, by supplying one of the two identical consonants within parentheses
(e.g., i-ki-e-et will appear as iki()et, at-ta-a-mi-i their father as atta()mi ).
A parenthesized letter or syllable may also indicate alternative interpretations of a
single writing (e.g., kar (a)pzi means that kar-ap-zi could be interpreted as either lkarptsil
or lkaraptsil); annall(i)e indicates that an-na-al-li-e (or an-na-al-li-e-e) could also be
transliterated an-na-al-le-e (or an-na-al-le-e-e).
Bolded forms indicate that the form is attested in OS, but may or may not occur in
later periods. The bolding is OS-inclusive, not OS-exclusive.
Bolded forms with parenthesized letters (e.g., ui(u)wanza) require that the alter-
nate writings uiwanza and uiuwanza both occur in OS, not that one is OS and the
other from a post-OS copy. Instances of the latter scenario require a second recording in
the paradigm (e.g., nepii, nepi, not *nepi(i)). But as with all bolded forms, a writing
ui(u)wanza leaves open the possibility that one or both of the variant writings also
occurs in post-OS.
In rare cases (see 1.72, p. 32) a parenthesized letter indicates a phoneme lost through
a phonotactic change: uiwatarm(u)apa means that the form uiwatarmapa arose
through deletion of the u vowel before the suxed particle -apa. Similarly na-pa
(for n(u)apa) in 28.100 (p. 378) the conjunctions n(u), (u), t(a) in 29.1 (p. 389),
itamana(n)an his ear, tuzzi(n)man my army in 6.5 (p. 139).
Entire forms marked by parentheses indicate those occurring rarely in the texts.
Forms in the paradigms of mi-conjugation verbs marked by a following are i-
conjugation intrusions. Forms in the paradigms of i-conjugation verbs marked with
are mi-conjugation intrusions.
When translating ancient texts or individual ancient words, we use pairs of single
quotation marks (. . .). When quoting from a modern publication, however, we employ
pairs of double quotation marks (. . .).
General Abbreviations
xviii Abbreviations and Conventional Markings
00-HittiteGrammar.indb xviii 6/5/08 4:14:07 PM
com. common gender
conj. conjunction
d.-l. dative-loctive
dat. dative
DN divine name
dupl(s). duplicate(s)
ed. edition, edited (by)
e.g. for example
erg. ergative
Erg. Ergnzungsheft
esp. especially
etc. et cetera
ex(x). example(s)
f(f). and following
fem. feminine
fest. festival (text)
gen. genitive
Gilg. Gilgamesh Epic (CTH 341)
GN geographical name
Hatt. attuili
Hitt. Hittite
HLuw. Hieroglyphic Luwian
Hurr. Hurrian
ibid. in the same place
idem the same (author)
i.e. that is
IE Indo-European
imp. imperative
imperf. imperfective
impers. impersonal
indef. indenite
inf. innitive
ins. instrumental
interj. interjection
interrog. interrogative
intr. intransitive
iter. iterative
lit. literally
loc. locative
Luw. Luwian
m
marks the PN of a male person
masc. masculine
MH Middle Hittite
m.-p. medio-passive (voice)
MS Middle Hittite Script
Mur. Murili
Muw. Muwatalli
n. (foot)note
n(om).-a(cc). nominative-accusative
neut. neuter
NH New Hittite
no. number
nom. nominative
NS New Hittite Script
obj. object
obv. obverse
OH Old Hittite
OS Old Hittite Script
p(p). page(s)
part. participle
pass. passive
perf. perfect
pers. comm. personal communication
PIE Proto-Indo-European
pl. plural
pl. tantum only plural
PN personal name
poss. possessive
postpos. postposition
PrAn. Proto-Anatolian
pres. present (tense)
pret. preterite
prev. preverb(s)
pron. pronoun
rel. relative
ref(s). reference(s)
rev. reverse
rit. ritual
scil. namely
sg. singular
subst. substantive
Sum. Sumerian
sup. supine
upp. uppiluliuma
s.v. under the word
Tel. Telipinu
Tel. pr. Telipinu proclamation
tr. translation, translated (by)
trans. transitive
translit. transliteration
Tud. Tudaliya
V vowel (in CV, CVC, VC)
var(s). variant(s)
voc. vocative
vs. versus
xix Abbreviations and Conventional Markings
00-HittiteGrammar.indb xix 6/5/08 4:14:08 PM
x in transliteration indicates an illegible sign
x non-subscripted x stands for an indeterminate number
x subscripted following a sign value indicates a value not yet assigned a number in
ocial sign lists
within Sumerograms the multiplication sign precedes sign element inscribed
within another. See 1.14 (p. 15).
= equivalences in dupls., lexical texts, etc.
marks clitic boundaries
section (of this or other books)
* prexed to unattested forms
( . . . ) within a Hittite word encloses omissible part of word (see above, pp. xviixviii)
( . . . ) in translation encloses words not in the Hittite but needed in English
[ . . . ] encloses phonetic interpretations
1
[ . . . ] encloses material lost in text break
[ . . . ( . . . ) . . . ] ( . . . ) encloses material restored from a duplicate
< derives or develops from
> becomesldevelops into
. . . encloses material accidentally omitted by the scribe
( . . . ) encloses material omitted from main text but restored from a duplicate
. . . encloses material to be omitted
l . . . l encloses phonologicallphonemic interpretations
1
single-wedge marker
double-wedge marker
Bibliographical Abbreviations
In-text bibliographical citations are in the Author-Date form (Wilhelm 1992) except for a re-
stricted number of reference works (e.g., CTH, HE, HED, HW, HW
2
, HZL) which are so well-
known and commonly used that it seemed inadvisable to refer to them in the Author-Date style,
and others which, while not quite so standard as the above, we had occasion to refer to very
frequently: AHP and LH. Journal abbreviations occur only in the comprehensive bibliography
that closes this work.
AA Archologischer Anzeiger
AAA Annals of Archaeology and Anthropology
ABoT Ankara arkeoloji mzesinde bulunan Boazky tabletleri
AGI Archivio Glottologico Italiano
AfO Archiv fr Orientforschung
AHP Melchert 1994
AHw von Soden 196585
AIPHOS Annuaire de lInstitut de Philologie et dHistoire Orientales et Slaves
AM Annals of Murili, edited in Goetze 1933a
ANET Pritchard 1969
AoF Altorientalische Forschungen
1. It is generally impossible to capture phonetic detail for Hittite. We have therefore mostly limited our
interpretations of its sound system to matters of contrast, marked by l. . . l. We have resorted to phonetic
interpretations, given in [. . .], only where the distinction seemed especially salient. All such interpretations
are provisional.
xx Abbreviations and Conventional Markings
00-HittiteGrammar.indb xx 6/5/08 4:14:08 PM
ArAn Archivum Anatolicum
ArOr Archiv Orientaln
AuOr Aula Orientalis
Belleten Trk Tarih Kurumu Belleten
BiOr Bibliotheca Orientalis
Bo Inventory numbers of Boazky tablets excavated 19061912
Bo yearl . . . Inventory numbers of Boazky tablets excavated from 1968 to the present
BrTabl. The Bronze Tablet cited according to the edition of Otten 1988
BSL Bulletin de la Societe de Linguistique de Paris
BSOAS Bulletin of the School of Oriental and African Studies
CAD Gelb et al. 19562005
CDA Black, George, and Postgate 1999
CHD Gterbock, Honer, and van den Hout 1980
CLL Melchert 1993b
CoS Hallo and Younger 2003
CTH Laroche 1971
DLL Laroche 1959
D Deeds of uppiluliuma cited according to the edition of Gterbock 1956
EHS Kronasser 1966
FHG Laroche 1951a; 1952
FHL Durand and Laroche 1982
GAG von Soden 1952; latest revision von Soden 1995
att. Apology of attuili III cited according to the edition of Otten 1981
HE Friedrich 1960
HED Puhvel 1984
HFAC Beckman and Honer 1985
HKM Maat tablets cited by the cuneiform edition Alp 1991b
HS Historische Sprachforschung
HSCP Harvard Studies in Classical Philology
HT Hittite Texts in the Cuneiform Character in the British Museum
HW Friedrich 1952 (reprint Friedrich 1991)
HW
2
Friedrich, Kammenhuber, and Homann 1975
HZL Rster and Neu 1989; bare numbers refer to pages, numbers following # refer to
sign numbers
IBoT Istanbul arkeoloji mzelerinde bulunan Boazky tabletleri 14 Istanbul
1944, 1947, 1954, Ankara 1988
IF Indogermanische Forschungen
IM Istanbuler Mitteilungen
IncLing Incontri Linguistici
JAC Journal of Ancient Civilizations
JANER Journal of Ancient Near Eastern Religions
JANES Journal of the Ancient Near Eastern Society
JAOS Journal of the American Oriental Society
JBL Journal of Biblical Literature
JCS Journal of Cuneiform Studies
JIES Journal of Indo-European Studies
JNES Journal of Near Eastern Studies
KBo Keilschrifttexte aus Boghazki
KlF Kleinasiatische Forschungen
xxi Abbreviations and Conventional Markings
00-HittiteGrammar.indb xxi 6/5/08 4:14:09 PM
Konk. Konkordanz der hethitischen Texte at http:llwww.hethport.uni-wuerzburg.del
hetkonkl
KUB Keilschrifturkunden aus Boghazki
KuSa Wilhelm 1997
KuT Precedes inventory numbers of Kuakl tablets
KZ Zeitschrift fr Vergleichende Sprachforschung (Kuhns Zeitschrift)
Laws Hittite laws cited by in LH
Lg Language
LH Honer 1997d
LSU Riemschneider 1958
MDOG Mitteilungen der Deutschen Orientgesellschaft zu Berlin
MIO Mitteilungen des Instituts fr Orientforschung
MSpr Murilis Aphasia, ed. Lebrun 1985
MSS Mnchener Studien zur Sprachwissenschaft
N.A.B.U. Nouvelles assyriologiques breves et utilitaires
OA Oriens Antiquus
OLZ Orientalistische Literaturzeitung
Or NS Orientalia Nova Series
PP 14 First to Fourth Plague Prayers of Murili II, ed. Goetze 1930a
RA Revue dAssyriologie et dArcheologie orientale
RHA Revue hittite et asianique
RS Ras Shamra text, cited by inventory number
SBo 1 Gterbock 1940
SMEA Studi micenei ed egeo-anatolici
StBoT Studien zu den Boazky-Texten
SV 12 State treaties cited according to the editions by Friedrich 1926; 1930
THeth 11 Homann 1984
TIES Tocharian and Indo-European Studies
TPS Transactions of the Philological Society
Ullik. Ullikummi myth, ed. Gterbock 1952
VBoT Verstreute Boghazki-Texte (Goetze 1930c)
VO Vicino Oriente
WO Die Welt des Orients
WZKM Wiener Zeitschrift fr die Kunde des Morgenlandes
ZA Zeitschrift fr Assyriologie und verwandte Gebiete
ZDMG Zeitschrift der Deutschen Morgenlndischen Gesellschaft
xxii Abbreviations and Conventional Markings
00-HittiteGrammar.indb xxii 6/5/08 4:14:10 PM
1
INTRODUCTION
The Hittites and Their Language
0.1. The people we now call Hittites lived almost four thousand years ago in the
central highlands of what is today the Republic of Turkey.
1

0.2. The story of the rediscovery of the Hittite writing system, language, and civiliza-
tion has been told often. In its main lines it runs as follows. Stone blocks found in Syria
at the end of the 19th century with hieroglyphic inscriptions chiseled into them were
correctly connected to the people known from the Hebrew Bible and the Neo-Assyrian
annals as Hittites. At that time scholars could not know that the language of these
inscriptions was not Hittite proper but a closely related language now called Luwian.
Yet the assumption that these inscriptions related somehow to the ancient Hittites was
correct. Similar inscriptions on rock reliefs in central Anatolia led explorers and archae-
ologists to impressive ruins near the village of Boazky. Ocial excavations begun
there in 1906 under the direction of Hugo Winckler and Theodore Makridi revealed a
great city dating from the time of the New Kingdom pharaohs of Egypt and the Kassite
dynasty of Babylonia.
2
Several huge archives of clay tablets inscribed in a variety of
cuneiform writing very similar to the contemporary Amarna archives found in Egypt
were discovered. Although many tablets were composed in Akkadian and could be read
immediately, conrming the excavators suspicion that they had found the capital of
atti, the vast majority were written in the native language of the Hittites.
Decipherment
0.3. Two tablets in this native language had been found decades earlier in the Amarna
archives, representing correspondence between the Egyptian pharaoh and the king of a
land called Arzawa (later revealed to be located in southwestern Anatolia). A Norwe-
gian scholar, J. A. Knudtzon, claimed the two Arzawa letters were written in a previ-
ously unknown Indo-European language (1902). His claim came under heavy criticism
from specialists in Indo-European languages. In the second volume of the edition of the
Amarna tablets O. Weber maintained thataccording to a letter sent to Weber Knudt-
zon had eventually lost condence in his own discovery,
3
leading to the inaccurate
1. For coverage of the Hittites, their culture and historywritten for the general readersee any one of
the following: Bittel 1970; Klengel and Klengel 1975; Macqueen 1986; Gurney 1990; Honer 1994, 1997a;
Bryce 1998, 2002; de Martino 2003; Honer 2003a.
2. See Gterbock 1995b for the history.
3. Weber and Ebeling 1915: 1074.
00-HittiteGrammar.indb 1 6/5/08 4:14:10 PM
assertion by others that Knudtzon had retracted his claim to decipherment. Working
with a much larger corpus of well-preserved documents in the Arzawa language from
Boaz ky, a Czech Assyriologist named Bedich Hrozn demonstrated convincingly
that Knudtzons allegedly retracted theory was in fact correct and published the rst ad-
equate grammatical sketch of what became known henceforth as the Hittite language
(Hrozn 1915, 1917).
0.4. The language now called Hittite
4
was the principal administrative language
of the ancient kingdom of atti, attested in documents from the state archives in its
capital city, attua, and from a few other sites (see in detail 0.60.10). According
to the currently most widely accepted chronology these texts date from the 16th to
13th centuries n.c. Our limited written sources leave us almost wholly ignorant about
the status of Hittite as a spoken language in terms of place, time, and social classes or
population groups.
5
0.5. Hittite is a member of the Anatolian sub-branch of the widespread Indo- European
family that includes Sanskrit, Greek, Latin, and most of the modern languages of Eu-
rope. Other Indo-European languages of ancient Anatolia include Luwian, Palaic, Ly-
cian, Lydian, and Carian. Hittite shows the typical features of an older Indo-European
language: it is both synthetic, showing signicant use of derivational suxes to form
words, and inecting, marking the role of most words in a sentence by a system of end-
ings (word-nal suxes). The historical relationship of Hittite (more correctly of the
Anatolian sub-branch) to the rest of the Indo-European family is a matter of continuing
debate, but this issue lies beyond the purview of the present descriptive grammar.
The Text Corpus
0.6. Hittite cuneiform tablets and tablet fragments, most of them recovered from the
royal archives of the capital city, attua, near the modern town of Boazkale (gorge
castle), number well over 30,000. This town earlier bore the name Boazky (gorge
village), which, before the introduction of standardized writing of contemporary Turk-
ish, appeared in archeological and philological literature of the 20th century as Boghaz-
ky, Boghazki, Boghaz Keui, etc. Although the ocial name of the town today is
Boazkale (gorge castle), it is customary in scholarly literature to continue to spell the
name Boazky (sometimes spelled without the Turkish as Boghazky), and we will
do so in this grammar. The vast majority of the excavated tablets are conserved today in
Turkish museums in Ankara, Istanbul, Boazky, and orum. Other sizable collections
4. The alternative name Nesite is rarely used, although the Hittites own designation for their lan-
guage was nelaili, neumnili (in) the language of (the city of) Nea (also known as Kane, see Gterbock
1958). For orientation in the subject of the various names used in ancient and modern times for this people
and their language see Gterbock 1959.
5. For recent discussions of this problem see Wilhelm 2002b, Melchert 2005b, and van den Hout
2006.
0.4 Introduction 2
00-HittiteGrammar.indb 2 6/5/08 4:14:11 PM
3
are found in the Pergamon Museum in Berlin, the British Museum, and the Louvre,
with much smaller holdings in various academic institutions or museums in America,
England, Europe, and the Middle East, as well as in private holdings.
0.7. Most known Hittite cuneiform texts were found at sites in central Turkey (Bo-
azky, Alaca Hyk, Maat Hyk, Ortaky [orum], Kuakl, Kayalpnar). A much
smaller number were found in the tablet archives of ancient peoples of the Mediterra-
nean littoral (Syria and Egypt)
6
who had diplomatic relations with the Hittites.
0.8. Although throughout the nearly 100 years of discovery and publication a very
large corpus of Hittite texts and fragments has been published, there still remains a
substantial number of tablets in Turkish museums awaiting publication. And although it
seems unlikely that any additional large source of clay tablets will be found in attua,
beginning in 1990 a cache of more than three thousand tablets has been excavated at
Ortaky (ancient apinuwa) in the orum Province of Turkey (Sel 2002). The archive
of apinuwa dates from the Middle Hittite period (ca. 14001350). Once published, it
should shed valuable light not only on this relatively poorly understood period in the
history of the Hittites but also on a crucial period in the development of their language
between the Old and New Hittite periods. In 2005, during exacavations directed by
Mller-Karpe at Kayalpnar near Sivas, some Hittite, Hurrian, and Akkadian tablets
were found. Hand-copies of them are being prepared by Elisabeth Rieken. Although it
seems likely that in the coming decades additional small archives will be found at other
provincial centers of the Hittite heartland, the weight of textual evidence will continue
to be the large harvest of tablets from the capital city.
Modern Resources for Study
Cuneiform Editions
0.9. Of the excavated tablets, most of those from Boazky have been published as
facsimile editions (drawings) in the following publications: 60 volumes in the series
Keilschrifturkunden aus Boghazki (abbr. KUB), 45 volumes in Keilschrifttexte aus
Boghazki (abbr. KBo), 4 volumes in the Turkish series Istanbul arkeoloji mzelerinde
bulunan Boazky tabletleri (abbr. IBoT) and a handful of other volumes in smaller
series.
7
The task of publishing facsimile editions of Hittite cuneiform texts from the
Boazky excavations is in the hands of the Boazky Archive of the Academy of Sci-
ences of Mainz, Germany, whose former director was Heinrich Otten and whose current
director is Gernot Wilhelm. This center for research maintains a wide range of research
6. Syria: Alalakh, Ugarit, and Emar; Egypt: Amarna.
7. Ankara arkeoloji mzesinde bulunan Boazky tabletleri (ABoT), Hittite Fragments in American
Collections (HFAC), Fragments hittites de Geneve (FHG), Fragments hittites du Louvre (FHL), Hittite Texts
in the Cuneiform Character in the British Museum (HT), Kuakl-Sarissa (KuSa), Verstreute Boghazki-
texte (VBoT).
Introduction 0.9
00-HittiteGrammar.indb 3 6/5/08 4:14:12 PM
4
resources, some available only to scholars visiting the site at Mainz (such as the com-
prehensive lexical les), and some available online at the Hethitologie Portal Mainz,
including the online edition of Silvin Koaks Konkordanz der hethitischen Texte (see
further below, p. 7), digitized photos of some tablets, a collection of personal names, a
bibliography, and the beginnings of a set of digital editions of texts (Digitale Publika-
tion von Texten der Hethiter [DPTH]).
0.10. Approximately 116 tablets or large fragments from the now completed exca-
vations at Maat Hyk (ancient Tapikka) were published by Alp (1991a, 1991b). We
can therefore expect no additional tablets from Maat. A smaller number of tablets from
the ongoing excavations at Kuakl (ancient aria) directed by Andreas Mller-Karpe
have been published by Gernot Wilhelm (1995, 1997, 1998, 2002a). Scattered individual
tablets from Emar, Alalakh, Ugarit, and El-Amarna have been published in excavation
reports and journal articles.
Commentaries
0.11. Since the early days of Hittitology in the 20th century, Hittite compositions of
many textual genres have been reconstructed from the tablets and presented in translit-
eration, with critical notes, commentary, and (often) lexical indices. The earliest such
series was Boghazki Studien, edited by Ferdinand Sommer and published in Leipzig,
Germany, in which the following important editions appeared: Hrozn 1917; Sommer
1920, 1922; Weidner 1923; Sommer and Ehelolf 1924. A second important series, He-
thi tische Texte, also published in Leipzig and edited by F. Sommer, is a subdivision of
the more comprehensive series Mitteilungen der Vorderasiatisch-Aegyptischen Gesell-
schaft. Today there exist at least three signicant series of this type in Germany and
several in Italy.
8
The largest and best known of the German series is Studien zu den
Boazky-Texten (Wiesbaden: Harrassowitz; abbr. StBoT), currently edited by Gernot
Wilhelm.
Sign Lexicon
0.12. The authoritative sign lexicon for Hittite texts is Hethitisches Zeichenlexikon.
Inventar und Interpretation der Keilschriftzeichen aus den Boazky-Texten (Wies-
baden: Harrassowitz, 1989; abbr. HZL), compiled by two experts in the eld, Christel
Rster and Erich Neu. Rster assisted Heinrich Otten for many years in producing ex-
cellent hand-copies for the KBo series. Neu produced some of the standard guides for
the dating of Hittite cuneiform texts on the basis of paleography. The volume not only
contains the complete known repertory of signs but virtually all known variant forms
of the signs, arranged under each entry in roughly chronological order. Under each
sign entry are listed all known examples of its meaning as a logogram (see 1.5, p. 10;
8. A new series of text editions in Italy is the Series Hethaea (Italian University Press), which in turn is
a subseries of Studia Mediterranea, edited by Onofrio Carruba.
0.10 Introduction
00-HittiteGrammar.indb 4 6/5/08 4:14:13 PM
5
1.371.44, pp. 2224) and its use as a component of multisign logograms (Su mero-
grams and Akkadograms) and logographically written proper names that contain the
sign in question. Of particular value to beginners, at the back of HZL are alphabetized
lists of Sumerograms and Akkadograms with German and Turkish translations, logo-
graphically written proper names (divine names, personal names, geographical names),
tables of common CV, VC, and CVC signs, and of easily confused signs.
Grammars
0.13. For more than forty years, the best instructional grammar of Hittite has been
Johannes Friedrichs Hethitisches Elementarbuch (Heidelberg: Carl Winter, 1960; abbr.
HE).
9
Less satisfactory, but written in English, is the grammar by Held, Schmalstieg,
and Gertz (1987). Very brief surveys also exist, such as Kammenhuber 1969b, Lu-
raghi 1997a, Francia 2005, and Rieken 2005a. There is a grammatical sketch of the
Middle Hittite texts from Maat by Honer (forthcoming). From the comparative Indo-
Europeanist perspective an early inuential work was Sturtevant 1933, which appeared
in a revised and enlarged second edition as Sturtevant and Hahn 1951. Later signicant
systematic works from this perspective are Kronasser 1956, 1966, and Ivanov 1963.
Dictionaries
Hittite Language
0.14. The best concise coverage of the entire Hittite vocabulary is still Johannes
Friedrichs Hethitisches Wrterbuch (rst edition; Heidelberg: Carl Winter, 1952; abbr.
HW),
10
with its three supplements, reprinted posthumously under a single cover (Fried-
rich 1991). Although this work was last updated (in the third supplement) in 1966, it is
marked by a careful, cautious, and accurate approach and is a model of conciseness. It
provides a German translation of all words whose meanings were known to Friedrich, a
selection of inected forms, a brief bibliography of studies of the words meaning, and
sometimes a proposed etymology. A more recent Hittite-German word list that covers
the entire alphabet is Tischler 2001. But although this is more up-to-date than HW, it
lacks many useful features of Friedrichs earlier work, such as the inected forms, the
bibliographies, and the list of Hurrian vocabulary.
0.15. Two projects have been underway since the 1970s to produce complete dic-
tionaries of Hittite on the scale of Wolfram von Sodens Akkadisches Handwrterbuch
and the Assyrian Dictionary of the Oriental Institute of the University of Chicago. The
rst is the revised and augmented second edition of Johannes Friedrichs Hethitisches
Wrterbuch, begun under the direction of Annelies Kammenhuber of the University of
Munich, Germany, and continued now by Inge Homann. This dictionary (abbreviated
as HW
2
) began its coverage with A and has now partially completed the letter in
9. Reprint, third unaltered edition; Heidelberg: Winter, 1974.
10. To be distinguished from HW
2
, on which see 0.15.
Introduction 0.15
6
volume 3. The second project is The Hittite Dictionary of the Oriental Institute of the
University of Chicago. Work began in Chicago in 1974 under the joint direction of Hans
G. Gterbock and Harry A. Honer Jr. on the basis of lexical les collected over a period
of ten years (196575) by Honer. This database was augmented by CHD sta during
the following years. In order to avoid immediate overlap with the Munich project, the
CHD began its published coverage with L. To date, volumes covering words begin-
ning with L, M, N, P and the rst two fascicles of S have appeared. The Chicago team
has begun placing a partially modied version of the text of published volumes online
(the so-called eCHD), so that articles can be consulted over the internet. Both of these
projects attempt to include treatments of all known words appearing in published texts,
whether or not their meanings have been determined, and to produce transcriptions and
full translations of representative occurrences of these words. Such a full presentation
enables users with limited access to the original sources to appreciate and weigh the
evidence for determining the words meaning. The CHD also includes notations indicat-
ing the best estimate of the date of original composition and of the copy of many cited
sources. Using this documentation, a user can trace the chronological development of
various meanings and grammatical usages.
0.16. The important multivolume work by Jaan Puhvel, Hittite Etymological Dic-
tionary (7 vols. to date; Berlin: Mouton l de Gruyter, 1984; abbr. HED), is now more
than halfway through the alphabet (AN) and is useful for more than etymological
considerations. Puhvel conscientiously lists all inected forms of words included in
his corpus, gives translations of selected passages in which the inected forms occur,
and has useful semantic discussions. HED provides no dating of the forms and thus
cannot show diachronic development within the attested languages. A second ongoing
work, Johann Tischlers Hethitisches Etymologisches Glossar (14 Lieferungen to date;
Innsbruck: Institut fr Sprachwissenschaft der Universitt Innsbruck, 1977), focuses
primarily on etymology.
Luwian Language
0.17. The latest glossary of the cuneiform Luwian texts is Craig Melcherts Cunei-
form Luvian Lexicon (Chapel Hill, N.C.: Self-published, 1993; abbr. CLL), which in-
cludes all words attested in the corpus edited by Frank Starke (1985) and gives selected
coverage of Luwian words appearing in Hittite contexts. Melcherts lexicon may be
obtained in PDF format from the author. Some Hieroglyphic Luwian words are listed
and glossed in HW and its three supplements in the appendix Nachbarsprachen. But
the usefulness of glossaries and word-lists for Hieroglyphic Luwian published prior to
Hawkins, Morpurgo Davies, and Neumann 1974, which changed the readings of several
high-frequency signs and thus revocalized many words, is limited. Among these earlier
works is Meriggi 1962, which was once a standard tool. Even Laroche 1960, which in
many respects still remains a crucial tool, suers in this respect. The new corpus of Hi-
eroglyphic Luwian texts (Hawkins 2000) contains a very useful partial index of Luwian
0.16 Introduction
00-HittiteGrammar.indb 6 6/5/08 4:14:15 PM
7
words, as do the vocabulary lists at the back of three recent introductory grammars of
Luwian (Werner and Lscher 1991; Plchl 2003; Payne 2004), but a complete diction-
ary is still lacking.
Hurrian Language
0.18. An up-to-date Hurrian glossary is badly needed in view of the many newly
available texts. The latest complete glossary published is by Laroche (197879), which
must be supplemented by tabulations and lists in new publications, such as the Hurro-
Hittite bilingual Song of Release edited by Erich Neu (1996). Neus death in 1999
deprived us of his planned companion volume, which would have contained a complete
glossary of this extremely valuable text. Other valuable sources on grammar include
Andre-Salvini and Salvini 1999; Wilhelm 1992a, 1992b, 1992c; and Giorgieri 2000. For
vocabulary, see Neu 1996, Catsanicos 1996, and the list of Hurrian words in the back of
the introductory grammar by Ilse Wegner (2000).
Sumerograms and Akkadograms
0.19. An older listing of Sumerograms and Akkadograms in Hittite texts can be
found in HW 264315 and its three supplements. An up-to-date replacement is HZL
30469, in which each logogram is accompanied by the number of the sign in the rep-
ertoire under which it is booked, as well as a German and a Turkish translation. Their
results are also incorporated by Johann Tischler (2001) in his appendix of Sumerograms.
Not restricted to examples attested in Hittite texts are the complete dictionaires of Ak-
kadian, AHw and CAD, as well as the concise CDA.
Text Catalogues and Name Collections
Hittite Cuneiform Texts
0.20. No one can adequately keep up with Hittite textual evidence without a cata-
logue of text compositions. This is particularly so, because one not only has to identify
and locate all the known compositions but must also reconstruct texts from myriads of
joins and duplicates. In the 1950s, 1960s, and 1970s the leading authority in this eld
was the French Hittitologist Emmanuel Laroche, whose magnum opus was his catalogue
of Hittite texts, published originally in installments in the journal Revue hittite et asian-
ique and subsequently produced in a revised and enlarged second edition as Catalogue
des textes hittites (Paris: Klincksieck, 1971; abbr. CTH). Laroche later supplemented
this (Catalogue des textes hittites, premier supplement Revue hittite et asianique 30
[1972]: 94133). But more than thirty years have passed since the last update of this
work, and Laroches death precluded any further revision by him. Currently it is neces-
sary to supplement CTH with the online additions at the web site maintained by Billie
Collins (Hittite Home Page, currently at Emory University) and the online version of
S. Koaks Konkordanz der hethitischen Texte (currently at Universitt Wrzburg; here-
after abbreviated Konk.).
Introduction 0.20
00-HittiteGrammar.indb 7 6/5/08 4:14:16 PM
8
Toponyms
0.21. The rst signicant collection of toponyms in Hittite texts was published by
Hayri Ertem (1973). The great Tbingen Atlas of ancient Western Asia has produced
a series of valuable volumes cataloguing toponyms from the major text corpora and
time periods. The volume covering the Hittite empire is by Giuseppe del Monte and
Tischler (1978), with a supplement (1992). It contains not only the text references but
translations of the immediate context of the more signicant toponyms and a relatively
complete bibliography of studies in which a location for the toponym in question has
been proposed.
Personal Names
0.22. The rst comprehensive collection and study of the personal names of the Hit-
tite texts was by Laroche (1951b, 1955). A revised and much augmented second edition
appeared in 1966. Additions to this second edition were published by Gary Beckman
(1983a). The Hethitologie Portal Mainz web site now contains a Rpertoire onomas-
tique prepared by Marie-Claude Tremouille in 2002.
Divine Names
0.23. For many years, the only systematic and comprehensive collection of divine
names was Laroche 1947. For some reason, although he published supplements and,
eventually, revised editions of his collection of personal names, Laroche never at-
tempted to revise his collection of divine names. The new comprehensive collection by
Ben H. L. van Gessel (19982001) has the advantage of completeness and great detail.
Every attested occurrence is listed, together with bibliography on each deity. But un-
like Laroches work, van Gessels does not group the various deities according to their
ethnic provenience, nor is there much discussion of the deities whose names are cata-
logued. For this one must consult the recent comprehensive volumes on Hittite religion
by Volkert Haas (1994) and Maciej Popko (1995).
0.21 Introduction
00-HittiteGrammar.indb 8 6/5/08 4:14:18 PM
9
Chapter 1
ORTHOGRAPHY AND PHONOLOGY
The Cuneiform Writing System of the Hittites
1.1. The Hittite texts were written by professional scribes on clay tablets that were
impressed with a stylus and dried in the sun and, to a lesser extent, on metal and on
wax-covered wooden writing boards (referred to in the texts as gulzattar or, logographi-
cally,
GI
ti-i
5
[= Akk. lu]). Although it is unclear what script was used on the writ-
ing boards, none of which have survived, the script used on clay and metal tablets was
cuneiform.
1
The cuneiform (from the Latin word for wedge-shaped) system derives
ultimately from Southern Mesopotamia, present-day Iraq, where it was devised by the
Sumerians for writing their own language and adapted centuries later for writing Ak-
kadian, a Semitic language. In addition to cuneiform writing on clay tablets, the Hittites
occasionally utilized a hieroglyphic script.
2
During the earliest phase of their kingdoms
history, the Old Hittite period, ocials used this hieroglyphic script to inscribe stamp
and cylinder seals. Much later, during the so-called New Hittite (or Empire) period,
kings began to use this system for carving royal inscriptions on rock faces or stelae.
These royal inscriptions were composed in Luwian (Hittite luwili; for this designation
see 19.15, p. 292), a language closely related to Hittite.
3
1.2. Before the advent of the Old Assyrian merchant colonies at the beginning of
the second millennium n.c., there was no writing in Anatolia (Kammenhuber 1969b:
16162). Once the Old Assyrian writing system arrived, it was employed not only by the
Assyrian merchants but also on occasions by the local Anatolian rulers (Balkan 1957;
Kammenhuber 1969b: 162).
1.3. The exact time and the precise source from which the Hittites obtained the
cuneiform writing system and applied it to the recording of their own IE language is
unknown.
4
Since Assyrian trading colonies existed in central Asia Minor (Cappadocia)
1. For a surviving metal tablet see Neve 1987: 4058 and Otten 1988, with English translations by
Beckman 1999: 11424 and Honer 2000. A silver tablet sent from Hatti to Egypt, containing the text of a
treaty between the two countries, is described in the Egyptian translation of that tablet, which appears both
on the walls of the temple of the god Amon at Karnak and in the Ramesseum; see the English translation
by John A. Wilson (ANET 201).
2. On the origins of this script in Anatolia see Mora 1991 and Hawkins 2003.
3. On the Luwians and their language see now Melchert 2003d.
4. A good summary of the present state of our knowledge on this subject can be found in HZL 1516.
00-HittiteGrammar.indb 9 6/5/08 4:14:18 PM
10
as early as ca. 1950 n.c. and left behind written documents composed in cuneiform,
one might have expected that the Hittites obtained knowledge of the cuneiform writing
system from them. But even a cursory comparison of Old Assyrian and Old Hittite cu-
neiform writing reveals that (1) the shapes of the signs (palaeography), (2) the selection
of logograms (Sumerograms), and (3) the choice of signs for the expression of a given
syllable (orthography) are all quite dierent. For example, Old Assyrian uses the nr sign
for the syllable ti, while Hittite scribes used the fr or ur signs. It is therefore generally
assumed that attuili I (ca. 16501600),
5
during his military campaigns in North Syria,
captured scribes who were using a form of the late Old Babylonian syllabary, and these
captives formed the nucleus of the rst scribal academy at attua.
6

Orthography
Writing Conventions
1.4. Because we have no living speaker of the Hittite language, acoustic recording, or
transcription of Hittite words in an ancient contemporary alphabetic script, we have no
way of knowing the precise sounds of the language. We gain access to Hittite phonology
and morphology only through the lter of the conventions the ancient scribes employed
when they wrote on clay using the cuneiform syllabary.
7
1.5. The cuneiform syllabary from its earliest stages in Mesopotamia consisted of
congurations of one or more wedges comprising what are called signs. Signs on Hit-
tite tablets are written left to right, with spaces between words. Signs are functionally
distinguished as phonetic (or perhaps better syllabic) and logographic.
8
Logograms are
signs or combinations of signs that designate a particular word in the target language
(e.g., the noun king, the adjective large, or the verb to sit down). Logograms in Hit-
tite texts consist of words from the Sumerian and Akkadian languages; the former are
called Sumerograms, the latter Akkadograms. Sumerograms in Hittite texts (apart from
the information given by occasional Hittite phonetic complements; see 1.37, p. 22) usu-
ally fail to indicate the grammatical case of the noun or adjective
9
and the voice, tense,
5. Establishing an absolute chronology for the Hittite kings, and even for broad periods of their king-
dom, has proven both dicult and controversial. Since this is not a history textbook, we will use only ap-
proximate dates when it becomes necessary to identify a period of Hittite history.
6. But Hecker (1992 and 1996) argues that the Hittites borrowed an atypical form of cuneiform known
in the Old Assyrian milieu. Klinger (1998) also discusses the question of who taught the Hittites to write.
For the latest discussion of the problem see Rubio 2006.
7. For general treatments of the subject of writing systems in the ancient Near East see Hawkins 1979,
1986; Morpurgo Davies 1986.
8. Logograms can serve either to represent spoken words or as determinatives to classify semantically
an immediately following (much less commonly, a preceding) word; see 1.391.44 (pp. 2324).
9. Sumerian case markers (Thomsen 1984: 88109) were not employed by Hittite scribes. For example,
the Sumerogram itoi king (without added Hittite ending) can stand for subject, agent, direct or in-
direct object, or possessor, as can the adjective oi great. When a Sumerogram stands in a case other
1.4 Orthography and Phonology
00-HittiteGrammar.indb 10 6/5/08 4:14:19 PM
11
or person of the verb,
10
whereas Akkadograms usually indicate all of these.
11
Ak-
kadograms consist of one or more signs read with the normal phonetic values (it not,
+-ui father, +-w+-:i+ word). Sumerograms often consist of a single sign and not
infrequently are read with a value dierent from the signs syllabic value in Akkadian or
Hittite; for example, the sign , read x+ in Akkadian and ka in Hittite is read rrx in
both Babylonian and Hittite texts, when it is the Sumerogram meaning word. But Sum-
erograms also can consist of two or more signs ( utxt.xtts daughter, girl).
1.6. Each cuneiform sign has a syllabic value. The repertoire of phonetic signs con-
sists only of signs of the following structural types. Hittite scribes (unlike later Baby-
lonian and Assyrian scribes) never used a single sign to represent CVCV other than in
logograms.
Vowel V e.g., a, e, i, u
Consonant + Vowel CV e.g., ba, da, pe, ti, lu
Vowel + Consonant VC e.g., ab, e, il, ut
Consonant + Vowel + Consonant CVC e.g., bar, kap, kan, kir, arlur
1.7. When we transliterate, we write the most appropriate values of the individual
signs, connecting those belonging to a single word with hyphens (e.g., a-a-an-zi they
are). Most Hittite words in this grammar are not transliterated but presented in broad
transcription. When we use broad transcription, the hyphens are removed and adjacent
repetitions of identical vowels are simplied (e.g., a-a-an-zi > aanzi, na-at > nat,
but i-u > iu). Adjacent identical consonants are not simplied but remain geminate
(ap-pa-an-zi > appanzi). Stop-nal VC signs with voiced-voiceless possibilities (ulf,
nlr, olx, etc.) conform to the voice character of the following CV. For example,
u-ta-a > atta father, not *adta. If the VC sign in question is word nal, it is written
voiceless (e-a-u > eat he sat down). Neither transliteration nor broad transcription
pretend to reect the precise pronunciation of Hittite words, only an approximation.
1.8. What is called plene writing (see also 1.46, p. 25) occurs when a vowel already
represented in a CV or VC sign is redundantly expressed by an adjacent V sign, e.g.,
than subject or direct object, it is usually marked with an Akkadian preposition (see 31.37, p. 441) (e.g., s+
itoi of the king, +\+ itoi tolfor the king, ts:i
GI
ntn from the table). An exception is the use
of the sequence of noun plus its genitive complement xtn (s+) itoi land of the king, where the genitive
marker s+ is not obligatory.
10. For example, only the presence of a Hittite verbal ending attached as a phonetic complement to the
end of the Sumerogram urn seize can indicate whether the subject is I, you, he, she, or they.
11. Thus +-ui father is normally subject, +-u+ is direct object, +-ut is indirect object or possessor.
Similarly with hand: +-:i+ subject, +-:++ direct object, +-:t object of preposition (Akkadian geni-
tive case); see 31.20 (p. 436). See chapter 31 (pp. 430.) for a brief survey of Akkadian grammar. We say
normally because in Hittite contexts the Akkadian case forms are occasionally used erroneously. Further-
more, an Akkadogram occasionally lacks the case ending altogether, resembling the construct state of the
Akkadian noun (on which see 31.23, p. 437).
Orthography and Phonology 1.8
00-HittiteGrammar.indb 11 6/5/08 4:14:20 PM
12
word-initial as e-e-ta = ta or u-up-zi = pzi, word-nal as pa-ra-a = par, or word-
internal as ma-a-an = mn or i-da-a-lu-u = idlu. Vowels written plene are indicated
in broad transcription with macrons (, , , ). Much less common is a hyper-plene
writing, where a word begins with two a-signs: a-a-an warm, a-a-bi ritual pit. There
is one example of non-initial hyper-plene writing, -u--, found in the sg. nom.-acc. neut.
u-u- and sg. acc. com. u-u--un of the adjective uu- full, lled, alongside ordinary
plene spellings u-u and u-u-un (see the paradigm in 4.57, p. 104). The hyper-plene
spelling here probably points to a stem lsuwu-l with an unusual sequence l-uwu-l (see
Goetze 1954: 404 n. 13 and AHP 5455 and 115). The lwl lls a hiatus produced by
loss of a PIE laryngeal (see Watkins 1975: 378 and Oettinger 1976a: 39 n. 72). That the
hyper-plene spelling indicates a preserved hiatus lsu_u-l, as per Watkins and Oettinger,
cannot be entirely excluded (see 1.142, p. 48). A mere long vowel lsu:-l (Berman
1972b: 18889) would not account for the hyper-plene spelling (see also the discussion
of Weitenberg 1984: 13640). There are no known Hittite words consisting solely of a
hyper-plene vowel, only descriptions of sounds heard in nature, such as i-i, as a hunters
imitation of the repetitive cry of a bird or animal (24.11, p. 320).
1.9. Hittite scribes normally spelled single (non-geminate) intervocalic consonants
as (C)V-CV (te-pu few, a-p-ya here).
12
Rare exceptions to this rule are signicant
in that they often mark a clitic boundary. For example otu-un-a-ta (not *otu-u-na-
a-ta) and tut-un-a-ta (not *tut-u-na-a-ta) in KUB 30.10 obv. 15 signal a clitic
boundary between the accusative singular noun (otu-un ox, tut-un sheep) and the
following clitic local particle -ata.
1.10. Scribes write double (geminate) consonants intervocalically as (C)V-VC-CV
(na-at-ta not, a-ap-pa back, again) or CVC-CV (kat-ta down). However, along-
side regular spellings for geminate consonants we also nd shorthand (or simplied)
spellings. One type omits the VC sign:
GI
lu-ti-ya-az for lu-ut-ti-ya-az window.
13
For
shorthand spellings of geminates omitting the CV sign see 1.12.
1.11. Since Hittite (as well as Mesopotamian) cuneiform has no sign for a conso-
nant without a vowel, it is impossible to write initial or nal sequences of two or more
consonants or internal sequences of three or more consonants without using at least one
empty (i.e., unpronounced) vowel. As an example of an initial sequence, lsmen-l to
forfeit must be spelled a-me-en- or e-me-en-, and lspikusta-l pin must be spelled
12. The consonants in question are those underlined in this and the following paragraph.
13. Other examples of nongeminate writing of geminates are: nu-kn passim (see AHP 14); ar-mi-zi-i
KBo 13.86 rev. 2; a-a-nu-wa-an-za KBo 4.6 obv. 14; a-a-u-i-it KBo 11.1 rev. 18; a-u-u KBo 8.47 obv. 6;
a-da-na KUB 30.63 v 20; a-za-a-ta KUB 12.62 rev. 2 (second example); a-a-u-u KUB 15.34 iii 40;
a-ti-li KUB 7.3:16;
m
a-tu-i-urorn-tt+ KUB 26.68 i 5; []a-zi-wi
5
-a KBo 41.152:12; ki-e-ra-a-a-
an KBo 3.27 obv. 3; ki-a-an KUB 39.71 iv 16; pa-ur KBo 13.58 iii 17; a-ra-a-zi-ya-a-ta KUB 19.67 i
15; i-wa-ti KBo 39.76 + KBo 41.64 ii 13; tu-zi-u KBo 6.34 i 25, 27; tu-zi-ya-a KUB 2.1 ii 23; wa-i-ya-at
KUB 44.4 + KBo 13.241 rev. 2. On the nongeminate writings of postvocalic pres. sg. 3 verbal ending -zi
see Yoshida 1998.
1.9 Orthography and Phonology
00-HittiteGrammar.indb 12 6/5/08 4:14:21 PM
13
e-p-ik-ku-u-ta, i-p-ik-ku-u-ta- or a-p-ik-ku-u-ta-. Kuryowicz (1958), followed
by many others, cites ma-li-id-du- and mi-li-id-du- as evidence for an initial sequence
lml-l. Examples of a nal sequence are: kar-a cut! for lkarsl, wa-ar-a reap! for
lwarsl, wa-al-a hit! for lwalhl, a-an-a or a-an-a seek! for lsanhl, and ki-i-ar-
ta by the hand for lkissardl.
14
As an example of a medial triconsonantal sequence,
lharspawantsl must be spelled ar-a-pa-wa-an-za. As shown by the above examples,
Hittite regularly uses nal Ca signs to spell word-nal consonantal sequences. There are
two attested exceptions, where nal l-antsl is spelled -an-zi instead of the usual -an-za:
nammawara anti tuanzi And furthermore, it (i.e., the bee) is anti tuanza
(KUB 17.10 i 39; OHlMS), and mn a-me-e-kn-zi (for a-me-e-kn-za, itself a
late form of a-me-e-a-an-za, see 1.138, p. 48) ui-[ri] When it becom[es] spring
(KUB 38.26 rev. 19; cult inv., NH). For more on empty vowels in Hittite writing see
1.137 (p. 47), 6.4 with n. 10 (p. 138), and AHP 29. Internal sequences of three con-
sonants, the rst of which is lnl, sometimes left the lnl graphically unexpressed (so
Kimball 1999: 31516): li-ik-zi (llinktsil) he swears and li-ik-ta he swore.
1.12. The normal rules of cuneiform writing require that a VC sign either occur
word-nal or be followed by a CV(C) sign. But examples exist of writings of the type
VC-VC (see also above, 1.9). In one type, represented by the example wa-al-a-tal-zi,
it is assumed that the written vowel in a is not to be pronounced and represents a way
around the limitation of cuneiform in writing sequences of three consonants (lwalhtal
or lwalhtsil) (see 1.11). The same principle usually applies to spellings CVC-VC: e.g.,
pr-a-ta represents lparhtal. In some cases, however, evidence shows that a CVC-VC
spelling is merely shorthand for CV-CV-VC:
LU
i-ip-pr-a or u(-u)-up-pr-a along-
side u-up-pa-ra-a. A second type consists of cases where it appears that the scribe
has simplied a VC-CV sequence (with identical C) by omitting the CV sign. These
cases are not examples of scribal errors, but a system of scribal shorthand used mostly
15

in post-OH. A scribe could write uppi, uppin, uppa, or uppiya (for example) as
u-up-i, u-up-in, or u-up-ya-a, as well as u-p-i (VS 28.15 ii 15) ,
UZU
u-pa (KUB
25.32 iii 29, 30) , `u-p-in, or *u-p-ya-a. Both types would be abbreviated writings.
But while the latter forms give no hint that they are abbreviated, the former would be
marked as such by a special orthographic pattern and would be understood to be equiva-
lent to the writings u-up-p-i, u-up-p-in, and u-up-p-ya-a. This rule would ac-
count for the examples: ki-an thus, ma-a-a-an as, i-i-at-ti, u-up-i, u-up-ya-a,
m
P-ip-pa-ap-a, a-at-an-na-a (HFAC 12 4), a-a-u ashes (< a-) (KUB 39.14
i 13) , nepii kat-an (KUB 24.5 + KUB 9.13 obv. 31) ,
d
Ki-p-ik-a-du (KUB 51.87 +
KUB 20.19 iv 17) , na-ak-u-i-u (KUB 32.76:8) ,
d
Ya-ar-i, tar-kum-ya-an[-. . .],
NA
4
ku-
un-ku-nu-uz-in, a-ra-az-i, i-ya-an-er (KUB 33.115 iii 14) . When the geminate con-
sonant written in shorthand was a z, an additional rule applied: one should also read
the shorthand in these cases as including an i-vowel following the geminate z. rorn-az-
14. Also written ki-i-ar-at; 4.82 (p. 115).
15. me-e-ek-e-e is OH.
Orthography and Phonology 1.12
00-HittiteGrammar.indb 13 6/5/08 4:14:22 PM
14
ya-za (abl.), a-ra-az-ya-az, a-az-i-, ke(-e)-ez-ya (for kezziya), i-pa-an-tu-uz-a-ar
(KBo 22.198 i 1) (for ipantuzziaar); thus also the interesting pres. sg. 3 forms of
mi-verbs: -e-mi-ya-az-ya-kn (KBo 10.37 ii 26) (for wemiyazziyakan), [t]i-i-e-ez
(KBo 34.90:8) (for tiyzzi in dupl. KUB 15.34 i 43 ), u-u-it-ti-ya-az (KBo 13.194:7)
(for ittiyazzi), kap-pu-u-e-ez (KUB 58.105 ii 12) (for kappuezzi). Cases of this type
occur too frequently to be scribal omissions and too irregularly to be cases of syncope
(on syncope see 1.761.78, pp. 3233; 1.83, p. 35).
Transcriptional Conventions
1.13. Syllabically written Hittite, Luwian, and Hurrian words are always written in
lowercase italic letters (Hittite e-e-zi he is, Luwian zi-la-ti-ya in the future, Hurrian
al-la-ni the Lady), Akkadograms in upper case italic letters (+-w+-:i+ word), and
Sumerograms in upper case non-italic letters (itoi king). Signs forming a part of
a Hittite, Luwian, Hurrian, or Akkadian word are connected to each other and to an
adjacent Sumerogram with hyphens (at-ta-a father, +-ui-:+ my father, urorn-tt+
god, itoi-u king,
m
Tup-p-
d
rxtn Tuppi-Teup).
1.14. Signs forming a single Sumerian word (xi.nnnn silver) are usually con-
nected by a period. But there is no uniform system among Hittitologists regarding the
use of the period to connect component signs of Sumerian nouns or expressions. In
some cases one nds two-word transcriptions such as r srsxtn oerer (lit., lord of
the oering, Akk. bl niq), yet in other cases one-word transcriptions such as utxt.
itoi prince (lit., son of the king, Akk. mr arri) are used. The scribes themselves
do not seem to have left what is called word space between the component signs
in either case. An argument could therefore be made that, following the example of
the scribes, we should write r.srsxtn and utxt.itoithat is, always as a single
compound rather than two words. But this creates an awkward situation for the plurals
of several of these common compounds, where the plural marker occurs between the
rst and second components, creating ugly forms such as utxt.xr.itoi. In order to
avoid this, we shall transcribe utxt.itoi in the singular but utxt.xr itoi in the
plural. Although strictly speaking this is inconsistent, it is a reasonable accommodation
and allows us to continue with the standard transcription for Sumerograms found in
most of the existing tools in the eld of Hittitology (such as the HZL). There are also
dierent approaches for transcribing the word enemy (adjective and noun). Most treat
the i element as a determinative and transcribe
LU
xn, while the CHD transcribes it as
a compound i.xn. In this grammar we transcribe i.xn in the singular but i.xr
xn in the plural. In only a few cases a Sumerogram consisting of two or more signs
has a reading which is not the sum of the readings of its components: xr.tu is to be
read xrsin threshing oor, q.t.ut
8
is to be read sor cupbearer,
MUNUS
r.xr.ir
as
MUNUS
rsr. We also follow the CHD in writing xi.or gold (without prejudice as to
its probable Sumerian pronunciation) instead of otxr. Sumerian scribes created sign
complexes in which one sign was inscribed inside another. The reasons for this pro-
1.13 Orthography and Phonology
00-HittiteGrammar.indb 14 6/5/08 4:14:23 PM
15
cedure need not concern us here, but the Sumerian pronunciation of such groups was
usually not a sum of the components: x (mouth) with inscribed (water) was read
as the verb o to drink. Whenever the Sumerian pronunciation for a sign-complex
of this sort is known, the sign-complex is transcribed with that value (e.g., o). If
there is uncertainty about the pronunciation, the complex is written as x, with the
framinglcontainer sign rst, followed by the inscribedlcontained sign second, the latter
preceded by a multiplication sign. Most such complexes occurring in Hittite texts have
known pronunciations: x = o or o to drink, xru = x to eat, xrx
= n thunder, xtu = zt
9
tooth, xxr = rxr tongue. Only a few complexes of
this kind used in Hittite do not have known pronunciations, but at least one occurs rather
commonly: xt mouth. In the earliest period of Sumerian writing, the component
signs in words did not always have to be in order. By the time of the Hittite scribes,
this was no longer the case, but a few relics remained from the early period of writing
in Mesoopotamia. Thus, the Sumerogram for poor man,
LU
x.r.xx, represents the
old Sumerian writing of an early loan from Akkadian, *makaen (later Akk. muknu).
Sumerian scribes sometimes indicated plurality by simply repeating the sign of the plu-
ralized item: thus, xtn land, but xtn.xtn lands. Examples of this writing convention
are still found in Hittite texts, but since the Hittite scribes had lost the rationale for this
convention, they redundantly added a plural sux as well (e.g., xtn.xtn.xr). In some
cases, instead of transcribing the sign twice, scholars transcribe the pair with the same
roman letters but with a dierent identication number: srsxtn (representing a single
sign), in contrast with ssxtn (representing srsxtn+srsxtn).
Homophony
1.15. The cuneiform syllabary contains many signs of identical phonetic value (e.g.,
there are several signs each for the syllables a, i, e, u, ba, a, u, kan, etc.). Signs with
identical syllabic values are called homophones. In order to distinguish homophones
in transliteration, Assyriologists and Hittitologists mark them with accents or subscript
numbers. For example, unmarked ba in transliteration indicates the sign rst established
as having the lbal value, b (with acute accent) denotes the second, ba (with grave ac-
cent) the third, ba
4
the fourth, and subsequent values are all indicated with subscript
numbers. The accent mark always rests on the vowel or, in case the sign has two vowels,
on the rst (rnr, rso, uXn, nnn).
16
In reading and writing Hittite in translitera-
tion, it is very important to distinguish homophones. Some homophonous values of
signs in the Mesopotamian forms of the syllabary are not used for writing Hittite. For
example, the number one value of lpil is the sign orft (HZL #317), which in Hit-
tite texts, when it is not a logogram, is always to be read wa. The sign most commonly
16. This practice was initiated in Hittitology by HZL (1989) and has become the new standard. The
older practice, following an earlier Assyriological method, placed an acute accent on the second vowel
(sri) for the second value, a grave accent on second vowel (sriX) for the third, an acute accent on the rst
vowel (si) for the fourth, a grave accent on the rst vowel (sli) for the fth.
Orthography and Phonology 1.15
00-HittiteGrammar.indb 15 6/5/08 4:14:25 PM
16
used for the value lpil (or lbil) is the sign transliterated bi or p. Similarly, Hittite scribes
preferred the number two signs for lkanl, lparl, lpatl (also read lpitl), and ltuhl. But
in most cases the signs used are the number one variants. Very few homophonous
signs are used interchangeably in Hittite. In the case of the homophonous signs u and
, the latter is used almost exclusively in logograms (-si hislherlits) or proper names
(
m
Ka--). In the case of ur and r, the former is much more common than the latter.
In the following three pairs, the rst sign can also have the i-containing value given in
parentheses, while the second sign has only the e-containing value: e (i) and ; ze
(zi) and z; n (ni) and ne. In very late texts the syllabic value lel, which normally is
represented only by the sign r, can also be expressed by the sign xr (which then has
the transcriptional value e
15
, for which see 1.35, p. 21).
Polyphony
1.16. Some cuneiform signs have more than one phonetic value, that is, they are
polyphonous. Some CV type signs whose initial consonant is a stop can have either a
voiced or voiceless
17
interpretation: nt can be bu or pu.
18
Signs of the types VC and
CVC do not indicate whether the nal stop is voiced or voiceless (b or p, d or t, g or k).
For example, the sign n can be read ab or ap, ru as id or it, to as ug or uk. Moreover,
when writing Hittite, the scribes do not even use contrastively those CV signs with
initial stop that distinguish voicing in the Akkadian syllabary: a-ta-an-zi and a-da-an-zi
they eat, ta-ga-a-an and da-ga-a-an on the ground, ad-da-as and at-ta-a father
(1.841.86, pp. 3536). Nevertheless, when transcribing syllabically-written Hittite
words, Hittitologists normally transliterate the obstruent according to the value of the
cuneiform sign most favored by the tradition of Hittitologists. Usually the favored trans-
literation is that which uses the number one value ( pa, not b; du, not tu; ga, not ka).
Exceptions to this pattern are the preferred transliterations utilizing the voiceless stops
such as p or p (instead of bi), tn (instead of din or den), pr (instead of bar), pdlt
or pdlt (instead of be), tglk (instead of daglk). CV signs possessing a number-one
value of both voiced and voiceless nature, e.g., nt = bu or pu, are normally rendered
with the voiceless stop.
19
Hittitologists are divided as to how they transliterate signs
in the Sumero-Akkadian syllabary intended to express the Akkadian emphatic velar q
in combination with a vowel (primarily q, since lqil and lqul were rendered in the
17. We use the terminology voiced and voiceless in this book, although we are aware that for Hit-
tite other scholars prefer the terms fortis and lenis or tense and lax. See AHP 1321 and Luraghi
1997a: 3 with n. 1.
18. In Hittite, all CV signs in which the C is a stop (lbl, lpl, ldl, ltl, lgl, lkl) were probably pronounced
voiceless at the beginning of words (1.86, p. 36). In that sense we are not speaking of real polyphonic val-
ues in spellings such as ga-an-ki and ka-an-ki he hangs. Nor in the interest of phonetic realization do we
transliterate ga-an-ki as ka-an-ki (or ka-an-gi
5
).
19. The above rules are ideals. Most scholars have their own personal preferences or habits in transcrip-
tion. This is particularly notable in the case of proper names: some write
d
Ku-mar-bi, others
d
Ku-mar-p ;
some
d
Te-li-p-nu, others
d
Te-li-bi-nu.
1.16 Orthography and Phonology
00-HittiteGrammar.indb 16 6/5/08 4:14:27 PM
17
contemporary Akkadian syllabary with the xr and xt signs). Because Hittite possessed
no such consonant, many scholars prefer to transliterate q as ka
4
and transcribe
20
it
as k (e.g., HZL), while others (including the CHD) prefer to avoid the subscripted value
and rely on users to know that qa in transliterated Hittite does not represent an emphatic
velar. Because the signs Vg-gV, V-k-kV and Vq-qV all equally represent lVkVl in Hit-
tite, or something like this, and V-gV, V-kV and V-qV represent lVgVl, it is obvious
that the function of a transliteration of Hittite is not to show precise pronunciation but
to code the individual cuneiform signs in roman script, choosing where possible from
common alternative values those which most nearly approximate what we think was
the pronunciation. For this purpose, for example, it makes no dierence whether one
transliterates
NA
4
za-ap-za-q-ya KUB 51.43 i 9 as
NA
4
za-ap-za-qa-ya or
NA
4
za-ap-za-ka
4
-
ya, because both according to the rules of Hittite spelling (see 1.841.85, p. 35, and
1.7, p. 11) and from the writing of this word in alphabetic Ugaritic and Hebrew as spsg
we can tell that the consonant in question was lgl. Instead of employing in this case the
transcriptional value ga
5
for q, which is attested in Akkadian and listed in HZL but
rarely if ever used by Hittitologists, we would advocate the use of the unmarked value
qa. We have chosen to employ the CHD procedure in this grammar. Readers will see
the letter q used both in transliteration and transcription of Hittite, Luwian, and Hur-
rian words, and should not assume that this indicates the existence of emphatic velar
phonemes in those languages.
1.17. The stops which occur at the end of VC or CVC signs, when they occur word-
nal, are written as voiceless: u-u-da-ak promptly, e-ep seize!, e-et eat!
1.18. One sign in Hittite is used with both a CV value and a quite dierent CVC
value: the nr sign is usually read ri, but in certain environments, especially in writing
derived nouns in -(a)t(t)alla- (see 2.32, p. 57), it must be read -tal-: see -na-at-tal-
la-an-pt the merchant himself KBo 6.3 i 13 (OHlNS) with duplicate -na-at-ta-al-
la-an-pt KBo 6.2 i 6 (OS). The value tal of the nr sign is also used word-initial in the
verbs talliya- and tali-.
21
A few signs with primary CV values in Hittite are used in
either logograms or proper names with a diering CVC value: pa has the n+: value in
the geographical names n+:-:i-s+, n+:-:t, etc.; i has the value tt+ in Akkadograms.
1.19. Signs of the type CVC can have more than one reading. Seven signs each have
alternate CVC values that dier signicantly: pr can be read as ma, kal can be read
as dan, kur can be read as madlt, tar can also be read as a in Hittite words and as xi:
in Akkadograms such as +s-xi: I fell silent, kir can be read as pi or pa, arlur can
be read as mur (although the mur value is largely restricted to the royal name Murili
and the geographical name Amurru), and the nu ( pt) sign can be read as +t: in the
Akkadogram ++-+t: (Akk. mmtu curse).
20. On the distinction between transliteration and transcription see 1.7 (p. 11).
21. On the latter see Koak 1993 (reading tal-a-a-an-du).
Orthography and Phonology 1.19
00-HittiteGrammar.indb 17 6/5/08 4:14:28 PM
18
1.20. Sometimes, however, the consonants of a CVC sign are stable and the dier-
ence is only in the internal vowel. CiC signs routinely also have the value CeC (ptlpt,
erlir, kerlkir ). But other vowel dierences also exist: ar and ur are the same sign, as
are pt and plt, pr and pir
x
, and possibly kat and kit
9
and kar and kir
8
.
22
In the 13th
century the sign ur (B), usually read tn, acquired the value ltanl, which HZL (sign
#330) represents without an assigned number as tan
x
23
(see 1.35, p. 21).
1.21. Rarely scribes would attempt to disambiguate a CVC signs internal vowel by
adding a CV sign of unambiguous vowel immediately preceding it or a VC one follow-
ing it. In such cases we transliterate the sequence using a superscript as follows: e-
e
er
(i.e., eer) they were sitting.
24
1.22. In cases where the dierence is also in the consonants, scribes used the same
disambiguating methods, writing p-i-
ke
ker they used to give KUB 29.25:4 = Laws
119 copy aa (OS), where the superscripted kile sign guides the reader to select the ker
value of the following sign in preference to its pi value,
25
pr
ar
-na KUB 35.68:8 and
m
M
a
-u-i-lu-wa(-an) KUB 14.15 iv 37, 38 to distinguish pr and ma, and
m
Mi-id-
dan
an
-na-mu-u-wa-a KBo 4.12 obv. 22 to distinguish dan from kal.
26
1.23. But other examples of preposed CV or postposed VC signs seem to have no
disambiguating value, because the adjacent CVC signs had only one known syllabic
value in Hittite writing:
27
kar-
di
dim-mi-ya-az anger KUB 24.4 rev. 10 ,
NINDA
pur
u-ur
-
pu-ru-u KUB 9.17. i 23 ,
MUNUS.ME
kat
at
-re-e female musicians KBo 32.2 rev. 9, KBo
42.85 i 3,

p-ner
er
they drove KBo 4.2 ii 20, tuuear t
u
-a he uses tuuear.
Because in most of the cited cases the practice seems unnecessary, one wonders why it
was done. In cases where it occurs only once or twice, one could surmise scribal error.
22. For the proposal of kit
9
see Goetze 1927: 60 and Sturtevant and Hahn 1951: 13 n. 8. HZL #173 cau-
tiously writes this value (as well as kir
8
) within parentheses in view of Otten (1973: 23), who questioned the
use of kit
9
(especially in OH) in view of the observation by von Soden and Rllig (1991: 13) that the value
lkitl for the ou sign does not occur in Akkadian texts before the rst millenniun. Otten agreed with the
recommendation of Sommer and Falkenstein (1938: 69 n. 3) that the assured value lkatl be used regularly in
Hittite, and the ali variations thereby occasioned be explained in other ways within Hittite. The verb p-i-
xn they used to give KUB 38.3 i 17 (NH) might be read p-i-ker
8
to harmonize with the normal writings
p-i-ker and p-i-ke-er, but doing so might mask a rare and important variant form. See Neu 1989a on the
third-person plural preterite active ending -ar.
23. In transliteration a subscripted x (not the multiplication sign but the letter x representing the al-
gebraic symbol for unknown number) denotes a demonstrable phonetic value that has not yet received a
standardized numerical value in HZL or the Akkadian syllabaries.
24. nu araza ku e-
e
er And those who were sitting outside (. . . came in the gate-building) KUB
41.1 iv 9. The form was incorrectly explained as a late intransitive (sitzen) to the causative verb ae-l
ee- to seat by Jakob-Rost (1972: 5051, 74, followed by HED ElI 209 and HW
2
A 386b).
25. This example courtesy of T. van den Hout.
26. The last two examples courtesy of S. Koak.
27. The rst three examples were called to our attention by N. Oettinger.
1.20 Orthography and Phonology
00-HittiteGrammar.indb 18 6/5/08 4:14:29 PM
19
But in examples such as tuuear t
u
-a, which occurs fairly often, scribal error is
an improbable explanation.
28
1.24. Word-nal -m was lost in the case endings of Akkadian nouns and adjectives
during the historical period of Hittite (see 31.20, p. 436). But words were sometimes
still written conventionally with nal -CVm signs (e.g., -:i+, -:t+, -:++, -ti+, -tt+,
-t++, etc), and even -Vm (e.g.,
GI
x+-+\-\i-i+), although the -m was no longer pro-
nounced in concurrent Akkadian. Extrapolating from this practice, NH scribes felt free
to use CVm signs for lCVl in writing Hittite words, not only in word-nal position
(e-e-tum, read e-e-tu
4
for le:stul let be), but also in other cases (i-dam-ma-a-mi
for listamasmil I hear). See also use of the xx sign in ar-kam-mi- lyre (and ar-
kam-mi-ya-ya-la- lyre-player), which alternates with (a-)ar-ga-(a-)mi- but never with
`ar-ga-am-mi-. Spellings -CVm-mV- in NS are thus not always reliable indicators of
a genuine geminate l-mm-l. Even some -Vm-mV- spellings in NS are suspect. Note
NS du-um-me-e-ni and even tum-um-me-(e-)ni for consistent OS and MS tu-me-e-ni or
du-me-(e)-ni we take. It is very unlikely that the l-m-l in words such as this became
geminate from OH to NH. Such complications make it dicult to determine the reality
of geminate l-mm-l in a number of words.
1.25. Rarely one nds cases of a sign sequence CVC
1
-C
1
V (other than CVm) in
which C
1
is not to be understood as geminate.
29
This is the case with the fi sign in
wa-a-tl-lali-, alternating with wa-a-dltu-lali-, but never *wa-a-tuldu-ul-lali (see
4.66, p. 108). Other possible examples include naaratt- fear (written in sg. nom.
as na-a-a-ra-az, na-a-ar-az, and in a blended writing na-a-ar-ra-az, but never
written *na-a-a-ar-ra-az).
1.26. The vowel of each CV-type sign is usually specic. But some CV signs whose
vowel is i often have a second reading with e:

nr p pe
H
or gi ge
G
nr i e
&
xr ki ke
O
ir li le

r ni ne
J
nr ri re
28. t-u-a might have the same function as wa-al-a-ta (lwalhtal), to indicate ltuhhsal.
29. Reading cuneiform signs of the type CVC

as CV
x
recognizes a

scribal convention that derives from
the Sumerians, the inventors of cuneiform, where the nal consonant of nouns written CVC (e.g., otu ox)
is repeated when followed by suxes beginning with a vowel: otu-da (read gu
4
- da for lgudal) of the ox,
although this is not to be understood as gemination.
Orthography and Phonology 1.26
00-HittiteGrammar.indb 19 6/5/08 4:14:30 PM
20
1.27. For some of the above lCel sequences, a second sign with value restricted to
cosof + e is equally or even more common: ( nr) is more common than
e (nrlnr) for lhel; ne (/ r) is more common than n (r) for lnel. On the other
hand, mi and me are dierent signs, as are i and e, and ti and te.
1.28. Almost all i-containing VC signs can also have an e value (see also 1.63,
p. 29):

rn ib, ip eb, ep

ru id, it ed, et

ro ig, ik eg, ek
.
rx im em

rn ir er
B
rz iz ez
1.29. Exceptions: e and i are dierent signs, as are el and il, en and in. One VC
sign that can be read with any vowel is D n, read a, e, i, or u.
1.30. Aside from the n sign (a, e, i, or u), no CV or VC type sign used in Hit-
tite fails to distinguish between vowels other than e and i. The vowel categories a eli
u are always kept distinct in such signs. This is in contrast to contemporary Akkadian
usage, where the rr sign (* HZL #317), read syllabically only as wa in Hittite, has the
possible readings wa, we, wi, and wu.
1.31. The signs i and S (word-initial) and i, u, and S (intervocali-
cally), when followed by a vowel sign of a dierent quality, are often used to represent
lyl and lwl: OH i-an-zi lyantsil they dolmake, i-(e-)ez-zi lye:tsil he doeslmakes,
i--ga-an lyuganl yoke, ta-i-az-zi-la-an-ni ltayatsilannil in theft, u-wa(-a)-i(-e)-ez-
zi lsuwaye:tsil he shall look, i-da-a-la-u-e-e lida:lawe:sl evil (ones), a-ra-a-u-a
lara:wasl exempt, -ez-zi lwe:tsil he comes, na-a--i lna:wil not yet, a--i-ti-i-za
lsawidistsl weanling (animal), a-az-zi--i lhatsiwil ceremony, ritual.
1.32. Where there is doubt about the vocalic identity of a given ambiguous sign,
the scribes often added clues. p or p followed immediately by -e is almost always to
be interpreted as p-e, only rarely as p-e. lkarpye:tsil could be spelled kar-p-ez-zi but
usually as kar-p-e-ez-zi (with an extra e) to disambiguate the writing of the following
izlez sign. Similarly, we nd p-i-e-et-ta allotment. Even iear beer, which uses
the unambiguous r sign, is often further disambiguated by the writing i-i-e-e-ar, to
prevent a mistaken reading *ear. See the remarks on plene writing in 1.8 (p. 11),
1.46 (p. 25), and 4.19 (p. 87).
1.33. There appear to be a few instances of sequences spelled Ci-e with an un-
ambiguous Ci sign that must nevertheless be read as lCe:l mi-e-ni KUB 29.9 i 11, 15
1.27 Orthography and Phonology
00-HittiteGrammar.indb 20 6/5/08 4:14:31 PM
21
(sg. d.-l. of lme:na-l face), [t]i-e-pu KBo 25.23 rev. 7 (sg. nom.-acc. neut. of lte:bu-l
little). Nevertheless, we should be extremely cautious in appealing to this possibility
in questionable cases. For example, ti-i-et in the duplicate KBo 3.36 obv. 24 shows that
ti-e-et in KBo 3.34 ii 19 (OHlNS) is a form of tiye- (stepped [to Akaliya, saying]),
not of te- (said). What used to be considered a pronominal stem i-(i)-e- but now is
correctly recognized as the number one (see Goedegebuure 2006) should not be read
as lse:-l (see already Neu 1997: 147).
Multivalence
1.34. Many signs in the Hittite cuneiform syllabary are multivalent. That is, they
have logographic as well as syllabic (or phonetic) values. Examples are:
Sign
Name Sumerographic value(s) Akkadographic value Hittitographic value
water + in +-\+ to a in a-ri he arrives
urorn god or sky +\ in
LU
n+-z+-+\-\i city
administrator
an in ma-a-an when, if
ur ur judgment, legal case,
srirx well-being
ot in +u-ot slave and
ot-\t judgment, :t
4
in
ui-it-:t
4
my lady
di, de in p-di in place,
e-de-ez on that side
r r lord i\ in is-xi-i\ he
bowed
en in ku-en-zi he kills
o o milk ++-+o-o+-t+-:t districts ga in e-ga-a ice
o o eld, rxt (unit of
measure)
o+\, x+\ gn, kn in i-in-kn
plague
x rrx word, matter -x+ your (sg.) ka in ka-a-a this
xr or
6
dark or night +t in +t-t+-++ whatever mi in mi-li-it honey
fn fn, xtu to cut (o) xio l: in +s-xi: I fell
silent
tar in tar-a-zi he
conquers
New Values
1.35. In only a few cases the Hittite scribes appear to have introduced a new phonetic
value to an existing cuneiform sign. Because their word for wine (Sumerian orfr)
was wiyana, they gave to the orfr sign () the value lwil, which we transliter-
ate as wi
5
(see HZL #131). In texts written in the 13th century, the xr sign (^ or
@), which normally stands for the Sumerian plural marker must occasionally be read
as phonetic e
15
(see HZL #360A and 1.15, p. 16 here), and the sign ur (B), usually
read tn, acquired the value ltanl, represented without an assigned number as tan
x
(see
1.20, p. 18). The sign that has the logographic value zz wheat () and in most
other contemporary cuneiform systems the syllabic value has the syllabic value ta
Orthography and Phonology 1.35
00-HittiteGrammar.indb 21 6/5/08 4:14:32 PM
22
in Hittite (see HZL #241), as well as occasionally in Akkadian texts from El Amarna
and Ugarit (see Labat 1988: 155, sign #339). It was not therefore a Hittite innovation,
unless it spread from attua to Syria and Egypt.
1.36. Hittite scribes utilized ligatures of the rr (* wa, we, wi, wu) sign with the
vowel signs a, e, i, and u subscripted to render sounds in the Hurrian and Hattic lan-
guages that were nonexistent in their own language (probably labial fricatives such as
[f ] or [v]). These graphic innovations are conventionally transcribed with subscripted
vowels thus: = wa
a
, / we
e
, , wi
i
, : wu
u
, or . wu

(see HZL ##31826). This
contrasts with the customary writing of initial lwel and lwil as -e- and -i- (rarely u-
i-) and post-initial lwel as -u-e or --e- and lwil as -u-i or --i in native Hittite words.
The scribes also used spellings in which pV (or bV) alternated with wV to indicate the
fricative sounds in Hattic and Hurrian words.
Logogram Pronunciation
1.37. We assume that logograms in Hittite contexts were normally pronounced by
Hittite scribes with their Hittite equivalents, both when reading a tablet aloud and when
dictating. The Sumerogram itoi (king) was pronounced with the Hittite word for
king au, xtts.itoi-a queen as auara,
d
tft-u sun(god) as
d
Itanu,
and oi large, great as alli.
30
Words could be written with Sumerograms to which
the nal syllable (often containing the all-important inectional ending) was indicated as
a phonetic complement: itoi-u king (subject case [nominative], singular), itoi-
un (direct object case [accusative], singular), oi-i great (subject case [nominative],
singular), ui-zi (= Hitt. iyazi) he does (present tense third person singular), etc. Not to
be read in Hittite are apparent logograms in a few proper names, which are rebus writ-
ings, such as:
m.GI
orunt-i-urorn-tt+, to be read using Akkadian values as
m
attuili.
31

Some personal names of Hittite ocials are Akkadian in origin, such as
m
utxt.tu.20.xx
HKM 22:9, to be read with Akkadian values as
m
Mr-er HKM 31:20 son of day
twenty (i.e., boy born on the 20th day of a festival; see Alp 1991a: 7879).
1.38. When a single word or phrase contains Sumerograms, followed by both Akka-
dian and Hittite phonetic complements, the resulting form can be bewildering: urorn-
30. There are a very few cases where the spelling may suggest otherwise. For instance, +-ui-:+-an-na-
a-za for attaminaza att. i 9, and possibly also ui-ti-u-a-an HKM 52:25; 80 obv. 5, if this really
stands for ian. That Hittite scribes sometimes used the logograms own pronunciation in dictating
texts to copyists is clear from errors in hearing such as n.t for proper n. he died,
GI
o.z for cor-
rect
GI
ot.z throne, etc. For discussion of the broader subject of Fehlertypologie with examples see
Rster 1988. The auditory error (Hrfehler) is not mentioned in her article. It would be a subcategory of
her section 2 (Zeichenvertauschung), which however she describes only as eine Verwechslung hnlich
aussehender Keilschriftzeichen (p. 295, italics ours), although in her list of examples (pp. 299303) there
are cases clearly due to auditory (not visual) similarity: ui for ut, ui for ut
8
, or for xr, zt for z.
31.
GI
orunt = Akkadian au sta, stick. urorn-tt+ = Akkadian ili of a god. attu + i + ili = the
royal name attuili ! A parade example of a rebus writing (see HE 8c spielerische Schreibungen).
1.36 Orthography and Phonology
00-HittiteGrammar.indb 22 6/5/08 4:14:33 PM
23
tt+-i god uses Sumerian urorn god followed by the end of the Akkadian equiva-
lent noun ilum god (albeit in the genitive case, ilim), followed by the last syllable of
the Hittite noun iuni god (4.50, p. 100). Such hybrid writings are few in number,
even if those few words occur frequently.
1.39. Some Sumerograms prexed (much less commonly suxed) to nouns are de-
terminatives. They are class markers, identifying the noun in question (whether that
noun is written in Hittite, Sumerian, or Akkadian) as belonging to a particular semantic
class:
32
uto
ceramic or stone containers
or
trees or objects entirely or partially made of wood
ntn.so
mountain names
u
river names
xr
appended to place names, but also to the logogram tnt city in tnt.xr,
see 9.51 (p. 167)
xtn
country names
i
designations of male persons, excluding personal names
xt
reptiles
xtts
designations of female persons, including personal names
xtr
bird names; generally not prexed, but appended to the word it modies

4
minerals or objects of stone
sn
green plants; always appended
fo
textiles or garments
tnt
city names
tntut
objects entirely or partially made of copper
(Single vertical
wedge)
personal name of a male individual
1.40. In transliteration (but not in the cuneiform writing itself) these determinatives
are superscripted (e.g.,
URU
Nerik the city Nerik, ara
MUEN
eagle). Sometimes what
appears at rst to be a determinative is rather a logogram: xtn
URU
alpa land of (city)
Aleppo (not
KUR.URU
alpa), i )I+t messenger (literally man of the message) (not
LU
)I+t message belonging to the male class), i +isiot guard (lit., man of the
spear) (not
LU
+isiot spear belonging to the male class). In the last two cases the virtu-
ally consistent retention of the Akkadian genitive ending -t on the second noun betrays
its function in Akkadian as a genitive.
1.41. Proper names regularly bear determinatives. Thus, mountain names (and names
of mountain deities) have prexed ntn.so, river names (and names of river deities)
have prexed u, names of springs and spring deities have prexed fi (or r), city
32. A complete repertoire of these determinatives can be found in HZL.
Orthography and Phonology 1.41
00-HittiteGrammar.indb 23 6/5/08 4:14:35 PM
24
names have prexed tnt. Names of deities other than mountain, river, or spring dei-
ties have a prexed urorn (god) sign, which, however, is rendered as a superscripted
lowercase d:
d
Teub.
33
Male and female personal names have prexed signs that are
rendered by superscripted m and f : (mnemonic for malelmasculine and femalelfemi-
nine)
m
attuili,
f
Puduepa. The
m
sign is a single vertical wedge usually represent-
ing the numeral one. The
f
sign elsewhere represents the word xtts woman or the
syllabic value al.
34
1.42. A few determinatives are placed not in front of their nouns but after them:
xtr bird in ara
MUEN
eagle,
35
xr place in
URU
alpa
KI
Aleppo, sn appended
to the names of herbs or vegetables (e.g., aaittin
SAR
). See 31.10 (p. 432).
1.43. We follow the CHD system, where the Sumerian plural suxes xr, nr., and
uruir are superscripted as determinatives only when the noun that precedes them is Hit-
tite, Luwian, Hurrian, or Akkadian, but not when it is Sumerian, since then the Sumerian
plural marker is to be read as part of the Sumerogram. If we do not write a Boazky
Akkadogram ++-+o-o+-t+-:t watchposts (the plural of ++oo+t:i) as ++-+o-o+-t
+-:t

or the plural of English land as land
s
, then neither should we write a Sumerogram
xtn.xtn.nr. as xtn.xtn
I.A
(contra HE 6d). Sometimes the plural marker interrupts
a syllabic writing, such as al-ki
I.A
-u crops. For the unusual placement of determina-
tives on compound nouns see 2.60 (p. 63).
1.44. While the suxed plural markers (xr and nr.) are equally common in all
periods, the suxed plural marker uruir is rare after Old Hittite, and when it does occur,
it is redundantly combined with another plural marker (uruir.nr.).
36
While no regular
rule for the selection of one of these three Sumerian plural markers has been discovered,
there is a marked tendency to use xr with nouns denoting persons and nr. with other
objects. Prexed determinatives (as class markers) are less frequently used in OH than
in MH and NH.
Phonology
Individual Phonemes (Vowels and Consonants)
1.45. It is important to distinguish between the supposed ancient pronunciation (to
the extent that we can reconstruct it) and the conventional pronunciation employed by
33. So in the CHD. Many Hittitologists prefer an upper case D:
D
Teub. Occasionally, a noun denoting
an object, when it is to be thought of as deied, will also bear the urorn determinative: e.g.,
d
aa- (dei-
ed) hearth.
34. For this reason many other Hittitologists prefer the superscripted roman numeral one for the rst
(
I
attuili) and upper case SAL or MUNUS for the second (
MUNUS
Puduepa).
35. But also prexed:
MUEN
ara(n)-, especially in OS texts; see Neu 1983: 53 for examples.
36. This suggests that, early on, the meaning of uruir was lost to Hittite scribes, as was the xx in
tu.(numeral).xx (9.51, p. 167).
1.42 Orthography and Phonology
00-HittiteGrammar.indb 24 6/5/08 4:14:35 PM
25
modern scholars. The following description applies to the ancient pronunciation unless
explicitly stated otherwise. The cuneiform writing system has a limited ability to ex-
press the phonemes of the Hittite language.
Vowels
Plene Writing and Vowel Length
1.46. The so-called plene writing (e.g., pa-a-, ti-i-, u-u-, te-e-, etc.; see 1.8, p. 11)
oered a means of expressing vowel length, but we cannot be certain that plene writing
always indicates length (as opposed to stress) or that non-plene writing always denotes
short vowels. Many scholars assume the existence of both short and long vowels in
Hittite, but there is no consensus on their distribution or on the phonemic status of
long vowels (see Melchert 1992 and Kimball 1999, both with copious references to
other views). Melchert (1984b: 162) adduces three additional reasons for plene writings:
(1) to show e-coloring of the vowel with ambiguous Celi and eli C signs; (2) to mark
the position of the accented syllable (e.g., sg. nom. te-e-kn vs. gen. tk-na-a-a (4.71,
p. 110);
37
(3) in the case of all monosyllables except clause-initial conjunctions (nu,
ta), to avoid writing a word with only one sign (da-a take!, i-it go!). Other reasons
may include: (4) rarely to indicate interrogative intonation (27.2, p. 348); and (5) to
distinguish real vowels from mere graphs, as in pa-ra-a (lpral), where the rst a is
merely graphic, and
URU
Da-la-a-wa lTlawal (in later Greek letters ), and
URU
Pa-la-
a lPlal (see and ). Most of the Hittite words cited in this grammar
are cited in broad transcription, not in transliteration. In broad transcription, the plene
writings cited above are indicated with macrons: tkan, takn, d, t, par,
URU
Dalwa,
URU
Pal.
1.47. In the noun pairs uttar versus uttr and uidar versus uidr, the dierence
in vowel length marks a contrast between singular and plural (3.20, p. 71). The single
word nu-u-wa still, yet is dientiated from the combination of conjunction nu and
clitic -wa (nu-wa) consistently by the plene writing of the former. But other examples
of what appears to be the same word or form with longer and shorter spellingsfor
example, e-er and e-e-er above and pa-an-zi and pa-a-an-zi they goare not dif-
ferent words but dierent spellings of the same word. Such variant spellings in the
same document sometimes arose when a scribe who preferred the short writings copied
a document whose scribe preferred the long ones. In some cases, there may be a dia-
chronic dimension: e-e-er is the OS writing of what in later periods was written short
e-er (6.6, p. 140).
37. Although r can be read either ni or n, r is not read `e in Hittite-language texts. Thus, while the
i in OH dative-locative ut-ni-i vs. nominative-accusative ut-ne-e land might be for disambiguating the
vowel, there can be no such motivation for dative-locatives i-i-i mouth and a-a-i-i hearth. Some sort
of length or stress must be indicated in the latter cases.
Orthography and Phonology 1.47
00-HittiteGrammar.indb 25 6/5/08 4:14:36 PM
26
Inventory of vowels
1.48. Hittite seems to have distinguished four vowels (a, e, i, u), each of which could
be long or short. Despite the claim of Eichner (1980: 156) and Hart (1983: 12430),
there is no basis for assuming a Hittite vowel lol, spelled with the sign u (reecting
prehistoric diphthong `Vu), distinct from lul, spelled with the sign (reecting prehis-
toric `u). Both u and are used to spell reexes of `u and `Vu-diphthong (Melchert
1992: 18687; Kimball 1999: 7980).
38
See now Rieken 2005b for the possibility of a
secondary split of pre-Hittite *lul (from all sources) into phonetic [o] and [u], probably
with marginal phonemicization of lol.
Vowel Alternations
e and i
1.49. Despite the ambiguity of certain e- or i-containing cuneiform signs, the two
vowels were certainly distinct phonemes in Hittite (Otten and Souek 1969: 56; Mel-
chert 1984b; 1992).
1.50. Many words containing the vowels e or i show no uctuation over time:
Words with Stable e
1.51. Word-initial: e-e-zi he is (never *i-i-zi), e-ep-zi he seizes (never *i-ip-
zi), e-ed-mi I eat, e-u come! (never *i-u), e-ku-zi he drinks (never *i-ku-zi), e-
e(-a)-ri form, shape, image (never *i-a-ri), ega- ice (never *i-ga-), enant- tame
(animal), tri- food.
1.52. Word-internal: e(-e)-er above, over, te-ez-zi he speaks and te-et he spoke,
-e-te-et he built, ku-(e-)en-zi he kills or ku-(e-)en-ta he killed never alternate with
forms in i (e.g., *i-ir, `ti-iz-zi, `-i-te-et, `ku-(i-)in-zi, etc.).
39
1.53. Word-nal: ku(-i)-e which (pl. neut.), a-p-e those (pl. neut.), ut-ne-e and
ut-n-e land (nom.-acc. neut.) versus ut-ni-i (d.-l.) p-e in p-e ar(k)- to pre-
sent, and le-e let not. The contrast of the plural ke-e these with (neuter) singular ki-i
this is stable through OH, MH, and early NH, breaking down only in late NH.
38. While many words are spelled consistently with either u or , there are also many cases of alternate
spellings, even in the same manuscript. Examples include: u-ul-ni-ik-zi KBo 6.2 i 13 and 16 (OS), i-nu-
u-ri KBo 6.34 i 32 vs. i-nu--ri KUB 41.26 i 26, lu-u-ri-i KUB 13.5 iii 5; vs. lu--ri-i KUB 13.18 iii 6,
mu-u-ga-an-zi KBo 10.20 iii 43 vs. mu--ga-it KBo 3.7 i 13, pu-u-ul KBo 3.7 iv 10 etc. vs. pu--ul KBo
26.20 iii 24, pu-u-nu-u-ta KUB 36.35 i 8 vs. pu--nu-u-a-an-zi KBo 20.5 iii! 7, pu-ul-ti-i KUB 32.123
ii 18 and 40, tu-u-li-ya KUB 6.46 iii 51 vs. tu--li-ya KUB 21.1 iv 39, u-re-e-na-an-da KBo 11.10 iii 26 vs.
-re-e-na-an-ta KBo 11.72 iii 13.
39. There are extremely rare mixed spellings such as ku-i-en-zi, -i-e-e we KUB 30.36 ii 8. On ti-e-et
with dupl. ti-i-et (allegedly he said) see 1.33, p. 21. On ku-i-en-zi see Melchert 1984b: 78.
1.48 Orthography and Phonology
00-HittiteGrammar.indb 26 6/5/08 4:14:37 PM
27
Words with Stable i
1.54. Word-initial: iyatar proliferation, fertility, abundance and its cognate forms,
ilan- step(?), i-a-al-li spittle, i-i-

(the oblique stem of ai mouth), iiya- to
trace, track, spy out, ik(iya)- to anoint, iki back, inura- kneading trough, dough-
pan, i-it go! (in deliberate disambiguity from e-et eat!, 1.56), i-da-(a-)lu evil,
i-wa-ar (postposition) like, as, innar independently (written in- not *en-),
GI
in-ta-
lu-uz-zi shovel(?). OH spellings ya-an-zi they dolmake and ya-an-ni-i they set out
of what later is consistently written i-ya-an-zi and i-ya-an-ni-i suggest that these words
begin with y, not i. The consistent spellings i-tV- and i-kV- without alternate spellings
*V-tV- or *V-kV-, for words with etymological initial sequences `st- and `sk- argues
that these contain a real prothetic vowel, as in foreign words beginning with sequences
loaned into Turkish istasyon station (Oettinger 1979: 41617 against AHP 3133).
40

See further 1.137 (p. 47). The situation for `sp- is complicated. It is likely that consis-
tent i-pa-an-t night also shows a real prothetic i-, while aleli-p-ik-ku-u-ta- pin
has preserved lsp-l (see 1.11, p. 12). For the special case of i(-ip)-pa-an-d to libate
(also with consistent i-vocalism) beside i-pa-an-d , see Forssman (1994).
41

1.55. Word-nal: The sg. d.-l. ending on nouns in OH is always -i rather than -e. In
MH and NH, this is also almost always the case. For some exceptions in -e see 1.61.
1.56. There is contrast between e and i, as seen in e-et eat! and i-it go!, as well
as in clitic -e to himlherlit (dative) and the dative-locative of the clitic possessive -i
in hislherlits . . . (Otten and Souek 1969: 56). The last example is only found in OH;
see the next section.
Words with Fluctuation between e and i
1.57. Certain words changed e to i, or vice versa, within the attested history of Hit-
tite: (i-)i-a-i > e-e-a-i (s)he performs, i-ku-na- > e-ku-na- cold, cool, i-ni > e-ni
that (one),
42
i-ni-i-a-an > e-ni-i-a-an thus,
43
ke-e > ki-i these (pl. nom. com.
and pl. nom.-acc. neut., 7.3, p. 143), -e > -i to himlher. The present rst-singular i-
conjugation ending is often - in OS, but in post-OH always -i. The rst two examples
40. But Eichners reading i-i-ta-ma-a-a-du in KUB 6.46 i 39 (1992: 80) has not stood the test of
Singers collation (1996: 54 ad B i 38), which showed [x r]u.otn
4
.n +-\+
D
tft -t accidentally omit-
ted from the line following i-ta-ma-a-a-du inserted by the scribe between [u-u-ma-]an-te-i and i-ta-
ma-a-a-du, all of which the copyist Weber misunderstood and miscopied. There is, therefore, no example
of plene writing of this prothetic vowel in Hittite.
41. There is no proof for prothetic i- with initial *sm-. The equation of Palaic ameri- with Hittite
imeri(ya)- (op cited in AHP 155) is quite uncertain. Furthermore, even if the equation were valid, it would
not aect our understanding of the Hittite pronunciation, which must be determined exclusively on the basis
of inner-Hittite spellings.
42. There is still no certain example of either ini or eni in OS.
43. There is no certain example of enian in OS, whereas all certain examples are in NS. However,
inian occurs as late as in NH letters (KBo 18.86:29).
Orthography and Phonology 1.57
00-HittiteGrammar.indb 27 6/5/08 4:14:38 PM
28
may reect a phonetic change (Melchert 1984b: 15356; but see Rieken 1996: 29497),
and the last two instead reect a morphological replacement (-e > -i after the nominal
dative-locative ending -i; -e > -i after -mi, -i, -zi, etc.). The other changes are surely
real, but their conditioning is not fully explained. On the limited appearance of -i for
pl. nom. com. -e in NH see now Sideltsev 2002, especially p. 72.
1.58. Some words show a uctuation that cannot be explained as a development from
earlier to later phases of Hittite. The examples cited in the preceding paragraph show
that there were at least some real changes in the distribution of lel and lil from OH to
NH. The examples also raise the possibility, however, that NH copyists misunderstood
the patterns they found in older texts and introduced errors and even unreal creations.
Because much of our evidence consists of NS copies of texts of uncertain date and
compositional history, in the following cases (and many others) we cannot determine
the status of the variations in e- and i-spellings. One should always be prepared to nd
isolated examples of e for i or vice-versa. See 1.65, p. 30.
1.59. Word-initial ile vacillation: i-na- and e-na- dough, e- (medio-passive) to
sit down, die down has e-a-ri and i-a-ri it (a re) dies down, i-ga-at-ta-ru and e-ga-
at-ta-ru let it become cool,
GI
el-zi- and
GI
il-zi- scale of a balance, e-ne-ra-an, i-ni-
ra-a, and in-na-re-en eyebrow,
GI
e-ri-p- and
GI
i-ri(-im)-p- cedar (loanword from
Akkadian erinnu via Hurrian; 4.16, p. 86), i-a-a-ru and e-e-a-a-ru tear(s),
MU
il-lu-ya-an-k and
MU
el-li-ya-an-k serpent.
1.60. Word-internal ile vacillation: te-e-um-mi-u cups (acc.) and ti-i-um-mi-u
(both OS in Otten and Souek 1969: 13), p-e-ta and p-i-ta (s)he gave, ke-e-da-ni
and ki-i-da-ni
44
tolfor this, u-i-ya-at-tn and -e-ya-at-tn send! (pl.), -e-mi-ya-
and -i-mi-ya- to nd, p-(e-)ra-an and once p-i-ra-an before, i-ik-nu-. . . vs. e-
ek-nu-. . . robe, i-in-kn vs. -en-kn plague.
1.61. Word-nal ile vacillation: pres. sg. 2 mi-conjugation ending is usually -i, but
see ar-a-ni-e-e you envy, and wa-ar-i-a-at-te you help.
45
In substantives one
nds sporadic word-nal e for i in NH (or NS) in the sg. d.-l. of a-stem nouns (e-,
i-dam-ma-ne [4.2, p. 80], p-e-te, an-tu-u-e; Neu 1979a: 18788) and of the con-
sonantal stem adjective u-u-ma-an-te to all (see Honer forthcoming: 45) and of
u-stem adjectives (3.24, p. 74; 4.43, p. 98; a-a-a-u-e KUB 31.127 iv 1 , compare
also i-da-la-u-e KUB 29.1 iii 11 (OHlNS) and nti-u-e KUB 21.5 iii 66 , pa-an-ga-u-e
KUB 31.42 ii 22, te-pa-u-e-wa-mu KUB 33.106 ii 5 , Weitenberg 1984: 319 844).
46

44. All examples are late: KUB 6.46 iii 61 (Muw. II), KUB 7.1 + KBo 3.8 i 8, KUB 30.16 + KUB 39.1
i 12.
45. wa-ar-i-a-at-te for lwarrissa-l is another example of the system of scribal shorthand discussed in
1.12, p. 13.
46. In the sg. d.-l. of u-stem nouns and adjectives the stem vowel u tends (already in OH) to color
the d.-l. ending vowel i to e. This is so common as to resist interpretation as an anomaly. See Neu 1979a:
18788, Weitenberg 1984: 319. Weitenberg is cautious and claims that the formation is not prior to Muw. II.
1.58 Orthography and Phonology
00-HittiteGrammar.indb 28 6/5/08 4:14:39 PM
29
Also in the vocative of u-stem nouns and names: itoi-u-e (MS), itoi-e,
d
tft-e
(Weitenberg 1984: 314 833). For vocatives in -e, see 3.28 (p. 74). Also voc. at-ti-me
O my father for expected *at-ta-mi, i-tar-ni-u-um-me KUB 26.81 i 7 and iv 9 , vs.
i-tar-ni-um-mi, act. pres. sg. 1 verb n-me KUB 22.25 + KUB 50.55 rev. 12 ; and NS
dative-locatives like p-e-te and p-te for p-di. There is not yet any consensus on just
which of these orthographic variations represent real linguistic alternations and which
are merely unusual or nonstandard spellings. Under these circumstances, prudence
suggests that we should transliterate such cases according to the most frequent values
for the signs used, as done above. It would be premature to impose our individual judg-
ment by transliterating at-ti-m or p-ti
7
, implying that we know for certain that Hittite
scribes employed the respective signs with these values.

However, as in other cases,
such as the representation of glides (see 1.31, p. 20) or the interpretation of ambiguous
eli signs (see 1.64), we and other scholars do not hesitate to make judgments in broad
transcription. One should therefore be prepared to encounter normalized spellings
such as pti for p-e-te, to which we will always attach a warning that it is smoothed
over and indicate precisely what signs are used.
1.62. Similarly, there sometimes occur, even in carefully written texts, examples
of what are called broken writings (e.g., -me-in or -mi-en): te-e-u-um-me-in for
teummin (Otten and Souek 1969: 13), u-ul-lu-mi-en for ullumen KUB 23.21 obv.
29 , and examples such as ku-u-a-ni-e-i-it in Laws 157, copy p, where we must
understand kuaniit. Unless one thinks such writings actually reect hiatus (1.142,
p. 48), one has to assume that only one of the two indicated vowels is valid and make
a judgment as to which is intended. Once such a decision is made, some scholars indi-
cate the decision by using rarer transcriptional values: -m-in or -me-en
6
, while others
prefer to reserve the interpretive stage for broad transcription, transliterating -me-in but
transcribing the sequence as ullumen and always indicating to the reader the nature of
the spelling. In this grammar we follow the second procedure. In neither case should
one assume that Hittite scribes regularly used xr with the m value or r with the en
6

value.
47
1.63. Many cuneiform signs of the type vowel + consonant in which the vowel is
i may also be read e (see 1.28, p. 20). Thus, in words spelled with such signs, both
vowels must be attested in plene writing (e.g., e-ek- and *i-ik-, e-ep- and *i-ip-, e-et-
and i-it-, e-e- and i-i-; 1.8, p. 11 and 1.46, p. 25) in order for the words to count as
examples with uctuation. The following words which are represented as uctuating in
the bound transcriptions of Friedrich HW, are not yet attested with a plene i vocaliza-
tion: ekt- net (e-ek-ta-, never *i-ik-ta-), ekdu- (e-ek-du, but never *i-ik-du),
GI
eppiya-
But in view of three cases of OHlNS and two of MHlNS, we would hesitate to exclude its beginning in the
pre-NH period.
47. The same caution needs to be observed in the case of rare necessary values such as e
17
for xr
(e.g., in i-e
17
), which must not be invoked arbitrarily but only in restricted environments (pl. nom. com.
of nouns or adjectives) and in absolutely clear cases in late NH.
Orthography and Phonology 1.63
00-HittiteGrammar.indb 29 6/5/08 4:14:40 PM
30
(e-ep-p-ya-, never *i-ip-p-ya-), e-er- (in e-grade form e-er-te-ni of the verb ar- to
arrive), e-er-ma-an illness, disease, e-ez-za-an cha. The pret. pl. 3 verbal ending
-er, even when written -i-rn, must be read -i-er. When this ending is written with CVC
signs, it must be read -er (not ir), and -ker (not kir), just as nal ur in the pret. and
imp. pl. 2 must be read -tn, rather than -tin. For the legitimate syllabic values in Hittite
of the ur sign see 1.20 (p. 18).
1.64. For persons unaware of the complexity of the writing system it is also confus-
ing that many Hittitologists use the i value of ambivalent eli signs as the unmarked
valuei.e., the default readingwhenever there is no plene vowel to disambiguate,
e.g., p-ra-an (which could just as well be interpreted as p-ra-an; so Melchert 1984b
and CHD P s.v.). Of the major dictionaries in progress at present, both HW
2
and HED
use the so-called unmarked i value. The CHD, on the other hand, insists on an inter-
preted transliteration (*-et transliterated as -e-et, not -e-it). Many words whose writ-
ings always contain one of these signs and that never use plene writing are uncertain as
to their vocalization.
1.65. The cases of vocalic stability appear to refute the idea that e and i merged
completely in Hittite at any period (pace Kimball 1999: 7379). The picture, however,
is admittedly complex and to some extent confusing. We must allow for the possibility
that dierent phonological rules applied at dierent stages in the language and that texts
composed in one period but recopied many years later present an articial and unreliable
picture of the phonology.
eli and ai
1.66. Some of the alternations between eli and ai noted by HE 13 reect mor-
phological replacements in NH. For the replacement of kappuezzi and kappuet by
kappuwizzi and kappuwit see Oettinger 1979: 336. Both paita and peta he gave
are replacements of pai, and neut l niut turn yourself is replaced by naiut (see
Jasano 2003: 183). Quite dierent are the problematic examples of etymologically
well-founded e rarely written ai (HE 13b) or ae: a-ep-ta KBo 5.6 i 11 for e-ep-ta he
seized, ma-eq-qa-u KUB 26.1 iii 58 for me-eq-qa-u many, ap-p-e-ez-zi- along-
side [ap-p]a-iz-zi-an and [ap-p]a-iz-zi-u-a, and e(n)k-lai(n)k- (Neu 1974: 41 with
notes).
a and e
1.67. Most alternations of a and e in nouns and verbs reect so-called ablaut, for
which see 3.37 (p. 78).
1.68. There was a historical progression of thematic mi-conjugation verbs with forms
such as i-ez-zi and i-e(-ez)-zi in OH and early MH to i-ya-(az-)zi in late MH and NH
(Carruba, Souek, and Sternemann 1965). Forms such as i-ez-zi, ti-ez-zi, p-e-i-ez-zi,
etc., were not pronounced dierently from i-e-(ez-)zi, ti-e-ez-zi, p-e-i-e-ez-zi, etc. The
second set of spellings is merely a more explicit writing of the rst set, for the sign rz
1.64 Orthography and Phonology
00-HittiteGrammar.indb 30 6/5/08 4:14:41 PM
31
can be read as iz or ez. If members of the rst set really were pronounced litsil and ltitsil,
one should be able to nd *i-im-mi, *ti-im-mi (or *ti-i-mi) and *p-e-i-mi, `p-e-i-i,
`p-e-i-nu-un forms in the same OH paradigms.
48
The appearance of a nominative-
accusative plural com. i-stem ending -iya beside -i in New Hittite appears to show a
similar change (3.16, p. 71).
1.69. The variation in rst- and second-plural endings -wenil-teni and -wanil-tani is
the result of early dierences in the position of the stress: -wanil-tani were the outcomes
when the endings did not receive stress, while -wenil-teni were the stressed outcomes.
The latter endings then spread at the expense of the former.
49
Pairs like apteni and eptani
are particularly suggestive of a pattern resulting from dierent accent placement.
50
a and u
1.70. The clitic second-singular pronoun has two allomorphs, -tta and -ddu. The
latter is regularly selected when immediately followed by the particles -z(a) and -an
(Hrozny 1917: 12728; HE 40; Luraghi 1997a: 25; see 5.14, p. 135). The fact that
similar-sounding taza (and + reexive particle) remains unchanged shows that this
change was not phonological but morphophonemic (i.e., applying only in the case of
particular morphemes).
51
Combinatory Vowel Changes
Vowel Assimilation
1.71. Eichner has proposed that the form of the sg. d.-l. of pedan place, written pl-
di, is always to be read lpidil where the rst lil in place of expected lel is to be attributed
to the inuence of the following lil of the ending (briey 1973: 84 n. 5; in more detail,
1980: 14348, especially n. 65). Eichner would also compare mi-li-it honey < *melit
(compare Luwian mallit). The presence of p-e-di beside pl-di (even in the same ms.,
as in KBo 3.4 iii 28 and 33) suggests that the latter can and should be read as lpedil and
leaves the proposed rule in doubt.
48. Melchert 1984b: 5455. The pres. pl. 3 forms of the mi-conjugation, such as i-en-zi they do, make
and -en-zi they come for the more common i-ya-an-zi and -wa-an-zi, are probably pre-NH but are not
clearly older than the more common forms.
49. AHP 13738, crediting Warren Cowgill for the analysis. For the view that -wani and -tani are due to
Luwian inuence see Rosenkranz 1952: 1315; Carruba 1966a: 41; 1966b, 1968. Otten and Souek (1969:
78), however, showed from OH evidence that these endings are not due to Luwian inuence.
50. In the analysis cited in n. 49 Melchert assumed that the dierence in stress placement was morpho-
logically conditioned and carried no functional signicance. It is possible, however, that pairs like paiweni
and paiwani we will go (see 12.41, p. 209) reect diering sentential intonation, indicating modality:
interrogation, subjunctive ideas, etc., semantic features expressed in other IE languages through the sub-
junctive and optative moods, signicantly missing from Hittite (11.3, p. 180; 23.10, p. 314).
51. There are no regular alternations between i and u. For a few sporadic cases of uncertain status see
AHP 178.
Orthography and Phonology 1.71
00-HittiteGrammar.indb 31 6/5/08 4:14:42 PM
32
Elision
1.72. When the conjunctions nu and u are followed by a clitic beginning with a
vowel other than lul, the vowel of the conjunction is elided: nua > na, nuat > nat,
nue > ne, nuata > nata, etc. (Ungnad 1920). This rule also applies to the clitic
pronoun -mu- me: uiwatarm(u)apa (OHlMS), inganam(u)apa (OHlMS),
kinunam(u)apa (OHlMS), andam(u)apa (OHlNS), numm(u)ata (OHlMS); see
CHD -mu a, b 4; HE 38a; and in this grammar, 28.110 (p. 382). That this rule is
morpho phonemic is shown by the fact that it fails to occur in other u + a sequences:
pduan KUB 23.68 obv. 24 (MHlNS), kara KBo 4.14 ii 13 (NH), KUB 60.86 obv.
5 and karwat KUB 14.14 rev. 18 , or in takkua Law 27, takkuan Law 28, etc.
1.73. When the clitic conjunction -ma but is followed by another clitic beginning
with u, the a of -ma is elided: ta-a-i-mu-u-za (< timauza) KBo 20.32 ii 9, i-en-zi-
mu-u But they make them (< ienzimau) KBo 16.78 i 10, lu-uk-kat-ta-mu-u-kn
(< lukkattamaukan) KBo 27.165 rev. 15 (MS). nam'-mu-u(-za-kn) KUB 7.1 ii
11 is not an example of namm(a) + -u (see CHD LN 391).
1.74. The initial -a- of the local particles -apa and -ata is elided after the dative
singular clitic -e- or -i-: nu-u-e-pa < nueapa KUB 36.110 obv. 11 (OS), or
nu-u-i-i-ta < nuiata HKM 116:30 (MHlMS). This rule is morphophonemic, as
shown by its failure to apply in the case of the local particle -an and the non-geminating
clitic conjunction -a: nu-u-e-an Law 78 (OS) and a-p-e-a KUB 36.100 rev. 4 (OS).
It also does not occur in na-at-i-at (see 30.19, p. 411), where the same morpheme -i
is involved.
52

1.75. For the morphologically conditioned deletion of various stem-nal vowels be-
fore vowel-initial suxes see 2.2 (p. 51).
Syncope
1.76. Sequences -(i)ya- and -(u)wa- in many cases contract, respectively, to -i- and
-u-. Forms such as ti-in-ti-e (IBoT 1.36 ii 48 [MHlMS], for normal ti-ya-an-te-e), u-
u-la-li-it-ta-at (ibid. iii 39, for normal u-u-la-li-ya-at-ta-at), and u-u-un-ta-an (ibid.
ii 41, for u-wa-an-ta-an) show syncope, as do sro
5
-in well KUB 31.105:4 and pas-
sim (for *lazzin
53
< *lazziyan, sg. neut. of stem *lazziyant-), appezzin < appezziyan
subsequently, dapin (see dapian, 8.10, p. 152), terin (see teriyan), and umilin (sg.
neut. < *umiliyant-, 15.13, p. 239), and imperfectives of stems in -ye-l-ya- (ilalike-
< ilalye-, wemike- < wemye-, etc.). See further the occasional contraction of the NH pl.
nom.-acc. com. i-stem ending -ila to -i (see 4.19, p. 87). Possibly also in stem-nal
position in the imp. 2 uitti < uittiya- (Oettinger, pers. comm.).

Additional examples
include lelundi- to pour (see part. liluwant- poured), miyaunte- to be old,
52. Since na-at-i-at occurs only in text copies from late NH, it is possible that the elision rule seen in
the cases with -ata and -apa, neither of which survived in the living language of late NH, had ceased to
operate by the time this repeated pronoun construction was coined.
53. Perhaps attested in restored [la'-a]z-zi-in KUB 43.31 left col. 9 (OS) (van den Hout 1991: 198).
1.72 Orthography and Phonology
00-HittiteGrammar.indb 32 6/5/08 4:14:42 PM
33
miunta- to make old (see miya(u)wanda- to make old) all from *miyawant-
old (Melchert 1984b: 53, 58, 133), anunta roasted things from anuwant-. But
there are many examples where the contraction fails to occur, and neither the chrono-
logical distribution nor the precise conditioning for the syncope (in terms of accent or
word length) is clear. For a recent discussion of the problem see Rieken (2001), where
cases are also cited that may reect processes other than syncope: (1) possible examples
of ablaut (see 3.37, p. 78) such as antuwaa-lantua- human being, watul-lutul-
sin, or warnilurni burns; (2) multiple stems for the same word such as lau-i
and lauwa-i pours (see 13.1713.18, p. 222).
54
Some examples may show either
syncope or something else: uriwarant-lwariwarant- blazing (re),
NA
4
kunnan-l
NA
4
ku-
wannan- bead, pl. d.-l. eknuleknuwa robe, ulkiara-lwalkiara- skilled, etc.
In addition, it is possible that the genitive singular of personal names
m
Nunnu and
m
Taruu reects syncope instead of a genitive singular *-s ending. See also 1.79.
1.77. Syncope is also seen in OH nau . . . nauma either . . . or > post-OH nau
. . . nama (see 29.59, p. 405), and may be assumed for the protoform *aima > OH
ma (see below, 7.16, p. 146).
55

1.78. It is uncertain whether any alternations of ue~u or ui~u are also to be attributed
to syncope. One possible case is karli- for karu(w)ili-.
56
Ablaut (see 3.37, p. 78) is
more likely in OH u-u-wa-an-da-an and u-u-wa-an-da-an for later uiwant- liv-
ing, alive (see EHS 3738; Otten and Souek 1969: 57 for the forms). Here an original
zero-grade adjective has been renewed under the inuence of the related verb ueli-
to be alive. In examples such as kuera-lkura- eld and tuekka-ltukka- body, limb,
there is no good basis for choosing between syncope and ablaut. Even less clear is
the alternation in wilan-lulin- clay. Some apparent alternations reect rather dierent
stems. The pair uinu- and unu- cause to run are based on the respective weak stems
uy- and u(w)- of the base verb uwai- to run (see Melchert 1984b: 52).
57
54. This also seems the most probable explanation for memiyani vs. memini, sg. d.-l. of word (4.75,
p. 112). Text chronology shows that memiyani occurs later than memini, so that syncope seems unlikely.
55. There are no sure examples of apocope. One possible instance is the form na-at for na-at-ta not.
The form na-at in an attami
m
Kizzuwan na-at-tau-e-e--nu--ut and my father did not spare(?)
Kizzuwa KBo 3.28 ii 19 (OHlMS) in view of the many sign omissions in the immediate context is al-
most certainly a case of an accidental omission of the ta in na-at-ta (so Laroche 1973: 187, 189). nakku
nat [arnik]za KBo 6.2 iv 5455 (Laws 98 copy A, with var. in copy B nakku L arnikzi ) might be a
genuine example, if we interpret copy A as the NH scribe of copy B did: it is a loss; he shall not make
compensation KBo 6.2 iv 5455 (OS) (see CHD nakku, which takes this as two clauses). But one cannot
exclude the interpretation of Honer in LH 9596, 197, reading na-at as nat with -at referring to r-ri kuit
arkzi. The example [. . .] l nat aruwaizzi [. . .] does not bow KUB 44.63 + KUB 8.38 ii 2122) may show
genuine apocope as well. n[a-a]t-tu-wa-az KUB 43.31 left col. 7 (OS?) (which is a variant of i-it-wa-az
KUB 58.48 iv 3 [OHlNS]; CHD natta a 1 a) may either show syncope of expected na-at-ta-wa-az or be an
additional example of an old shorter variant of natta, nat not.
56. ka-ru--li-umaza urorn.xr-u l
d
rx-a zo-az aata KBo 32.13 ii 2627.
57. The stems uda- and wida(i)- simply belong to two dierent verbs meaning to bring. Only the sec-
ond (like peute- and uwate-) takes as its objects living beings capable of self-propulsion (see CHD P 352
Orthography and Phonology 1.78
00-HittiteGrammar.indb 33 6/5/08 4:14:43 PM
34
Diphthongization (Breaking)
1.79. In certain cases, words with original u come to be spelled with uwa: e.g.,
paur re appears as pauwar, kurke-
58
to cut as kuwarke-, etc. For further
examples see Rieken (2001: 37274), who plausibly suggests that these cases show a
genuine sound change whereby the u develops into a diphthong.
Vowel Insertion (Anaptyxis )
1.80. It is reasonably certain that some original sequences of consonants were bro-
ken up in pre-Hittite by secondarily inserted vowels (see English athlete pronounced
[lit]). However, the limitations of the cuneiform writing system in spelling conso-
nant sequences (1.11, p. 12) make it hard to distinguish with certainty genuine inserted
vowels from those that are merely graphic, and there is no consensus on this issue.
59

The discussion here is limited to those cases where it is likely that addition of suxes
to roots leads to genuine vowel insertion.
1.81. Insertion of ile. An i or e is inserted between a root ending in a labial or velar
stop or in -- and the sux -ke-: appileke-, arninkileke-, anileke-. The vowel is
consistently i in OH, but spellings with e begin in MH, and in NH one nds both i and e
with some frequency (Melchert 1984b: 108, 134). There is a similar insertion before the
pret. sg. 3 ending - (Oettinger 1993: 41): akki to akk- to die and iiyai to iiya-
to denounce. Likewise in pret. sg. 3 takkita and other forms of tak- to fashion. The
behavior of roots ending in a dental consonant is more complex. In MH we nd pret.
sg. 3 atte to att- to pierce, with insertion, as in the previous examples (Oettinger
1993). However, the usual treatment when -ke- is added to a root ending in a dental
stop is insertion of -i- after the [s]: azzikke- to elad- to eat (< lad-ske-l) and zikke- to
dai- to put (< ld-ske-l).
60
This development is also sometimes seen after roots ending
in other dental consonants: alongside tar-a-ke- (ltarske-l) to tar- to say we also nd
tar-i-ik-ke- with inserted -i-. Likewise for the addition to ann(a)- to dispute, to judge
we nd both annike- (with i-insertion before the s) and aikke- (< *ann-ske- with
i-insertion after the s and assimilation (see 1.120, p. 43). For possible anaptyxis of i in
the origin of the instrumental ending see 3.35, p. 77.
1.82. Insertion of u. Some verbs whose attested stem in Hittite ends in -a- (13.9,
p. 218) or -e- (12.23, p. 199) originally had variants without the nal vowel. When
these appear before suxes beginning with -w- (pl. 1 -wen(i); verbal noun -war ; inni-
in summary of peda- B).
58. Written kur-a-k.
59. A new systematic treatment of the problem is sorely needed. The analysis in AHP 2933 is both
incomplete and in some respects erroneous.
60. Against AHP 30 and with Oettinger (1979: 318) and others, the consistency of the spellings with
i argues that the vowel is real. In zikke-, the root appears in an archaic form without the -i- seen in the
inection of dai-. The rare alternate form zake- StBoT 25 #137 iii 15 must represent a genuine variant
treatment.
1.79 Orthography and Phonology
00-HittiteGrammar.indb 34 6/5/08 4:14:44 PM
35
tive -wanzi), a u is inserted before the -w-, and the new sequence -uw- is then dissimi-
lated to -um- (1.126, p. 44). E.g., `dweni > `duweni > tumni we take (OH) to d-,
`tarnwanzi > `tarnuwanzi > tarnu(m)manzi to release to tarna-. This insertion may
be viewed as anaptyxis, but see AHP 57 (with references to Eichner 1988 and Bernabe
1983) for other interpretations.
1.83. Insertion of a. There are two probable cases of syncope followed by anaptyxis
to resolve an internal sequence of three consonants of which the middle one is r. In the
rst example kutruwene witnesses by syncope became `kutrwene, which then re-
solved by anaptyxis to kutarwene (ku-tar--e-ni-e KUB 23.78b 9 ) (Oettinger 1982b:
164). In the second example etriyant- (well-)fed, fattened, robust by syncope became
`etryant-, which then resolved to etaryant- (e-tar-ya-an-t[a-an] KUB 12.63 obv. 16 )
(Melchert 1997b).
61
Consonants
Contrasts in Stops
1.84. Sturtevant (1933: 6667), following a suggestion by Mudge, rst argued that
there is a phonemic contrast in Hittite between single and double (geminate) stops in
intervocalic position.
62
Further research has fully conrmed this analysis. While there
are a few examples of inconsistent spellings (e.g., occasional i-ya-ta-ri for regular i-ya-
at-ta-ri ), most morphemes are written consistently with either single or double stops.
Patterned exceptions actually conrm the contrast (see AHP 14), while many others
may be explained as simplied spellings (see 1.9, p. 12, and 1.12, p. 13). One can
even cite a small number of semantically contrasting minimal or near-minimal pairs:
ap- that (with rare sg. acc. com. apn) vs. app- to seize (part. sg. nom.-acc. neut.
appn), padn of the feet vs. paddan dug (sg. nom.-acc. neut.), ekan cubit vs.
ekkan known (part. sg. nom.-acc. neut.).
1.85. In general Hittite intervocalic single stops reect PIE voiced and voiced aspi-
rate stops, while geminate stops continue PIE voiceless stops (Sturtevant 1933: 6667),
and explanations have been found for most apparent exceptions (AHP 16 and Kimball
1999: 26164, with references to op, Eichner, and others). There is no consensus as
to whether the phonetic contrast in attested Hittite is one of voicing or of some other
feature such as fortis vs. lenis or aspirated vs. unaspirated (see AHP 1618
with references to a variety of opinions). For the sake of simplicity we here describe
the contrast in stops as one of voicing, but we do not mean thereby to take a denitive
stance on this issue.
63
61. That the attested spelling variants both represent a sequence [trw] with non-syllabic [r] is highly
unlikely but cannot in principle be excluded.
62. Compare the similar conventions for writing Hurrian stops in syllabic cuneiform (Wegner 2000:
3940).
63. Whether or not the stops spelled as geminate in Hittite were voiceless, they seem to have been so
heard by Ugaritic scribes: see alphabetic writings of Hittite names such as u-up-p-lu-li-u-ma ( pllm).
Orthography and Phonology 1.85
00-HittiteGrammar.indb 35 6/5/08 4:14:45 PM
36
1.86. It is likely that the phonemic contrast in Hittite stops existed only word inter-
nally. It was neutralized in word-initial and word-nal position (see AHP 1821 and
Luraghi 1997a: 1.1). Since the device of marking the contrast by single vs. gemi-
nate spellings was only used consistently intervocalically, it is dicult to prove the
contrast for internal sequences of stop plus another consonant, but a handful of ex-
amples suggest that it probably applied there: see l-dr-l in appatriye- to commandeer
< apptar seizing spelled ap-pa-ta-ri-ez-zi and ap-pa-at-ri-ez-zi Law 76 (OS and NS,
respectively) vs. l-tr-l in atre(n)- crossroads, intersection < atta- to cut spelled
a-at-(ta)-re-e-n.
Inventory of Consonants
1.87. It is believed that Hittite possessed voiced and voiceless consonants of the fol-
lowing types: stops (b, p, d, t, g, k, g
w
, k
w
), aricate (ts), fricatives (, ), and sonorants
(m, n, l, r, y, w).
Stops
1.88. There are four series of stops in Hittitebilabial, dental, velar, and labiovelar
all of which can be voiceless or voiced when intervocalic. The following table of
Hittite stops is adapted from Luraghi 1997a: 1.2.1. As in Luraghis table, capitalized
voiceless stops represent cases where the voicing opposition is neutralized, and forms
that end in a hyphen represent stem forms. The # sign represents the word boundary.
lpl: a-ap-pa back, again, u-up-p- pure
lbl: a-pa-a-a that one
#lPl: p-e-ri-ya-a of the house, pa-ra-a forth, forwards
lPl#: e-ep seize (sg. imp.).
ltl: kat-ta or ka-at-ta
64
down(ward), at-ta-a or ad-da-a father, e-e-tu let him
be
ldl: i-di or e-di on this side, wa-a-tar water
#lTl: ta-a-ru wood, tree, tu-uk you (sg. acc. or dat.)
lTl#: ke-e-a-ri-it by the hand (instrumental case), e-et eat (sg. imp.), i-it go
(sg. imp.)
lkl: tu-e-eg-ga- body, lu-uk-ke-et he set re to, a-ag-ga-a-i I know
lgl. la-a-ki he bends, la-ga-(a-)an bent, inclined
#lKl: kat-ta down(ward), ke-e-ar- hand, ki-i-a-an thus, ku--a-ta bride-
price
lKl#: la-a-ak bend (sg. imp.), a-a-ak know, zi-(i-)ik you (sg. nom.), am-mu-uk
me (acc. or dat.), u-u-da-a-ak promptly
64. The usual spelling is kat-ta, but occurrences of ka-at-ta (KUB 20.4 vi 4, KUB 20.43:9, KUB 58.50
iii 7) and one of ka-at-ta-an (KUB 56.46 vi 21) show that the double writing of the dental is secure. HED
K 128 (rarely ka-at-) cites no example.
1.86 Orthography and Phonology
00-HittiteGrammar.indb 36 6/5/08 4:14:46 PM
37
lg
w
l: a-a-ku-wa eyes, a-ku-an-zi they drink
65
#lK
w
l: ku-i who, ku-en-zi he kills
lK
w
l#: e-ku drink!, -ku . . . -ku whether . . . or
1.89. In addition to signs indicating voiced and voiceless stops, the cuneiform syl-
labary possessed some that were intended to represent so-called emphatic stops of the
Semitic Akkadian language: the velar conventionally represented by the Roman letter
q and the dental represented by .
66
Because even the Babylonians and Assyrians often
wrote these consonants with the same signs they employed for the nonemphatic counter-
parts, we have only one sign in the Hittite syllabary whose number-one transcriptional
value contains an emphatic: q. On the question of how to transliterate q in Hittite
words see above, 1.16 (p. 16).
Aricates
1.90. The cuneiform signs conventionally transcribed as containing a z represent
a sound (or sounds) with three distinct sources in Hittite. (1) Some cases reect a se-
quence lt+sl: the sg. nom. of the common gender t-stem aniyatt- laniyat-sl is spelled
a-ni-ya-az vs. sg. gen. a-ni-ya-at-ta-a.
67
(2) In other cases it represents the aricate
ltsl resulting from the dissimilation of a dental stop before another dental stop (see
1.125, p. 44): imp. sg. 3 le:tstul let him eat is spelled e-ez(-za)-du. (3) In still others
it represents the sound resulting from prehistoric assibilation of `ty and `ti: sux -zziya-
< `-tyo- in a-an-te-ez-zi-ya- front, rst (see 2.53, p. 61) and pres. sg. 3 ending -zzi
< `-ti in e-ep-zi takes etc. The rst two sources suggest that z stands for a voiceless
dental aricate ltsl, and we follow most Hittitologists in adopting this value.
68
Nothing
precludes a palatal or palatalized aricate (see the sound of English church) for those
cases resulting from assibilation. However, there is no evidence for a phonemic contrast
between these instances and those reecting lt+sl.
1.91. The aricate ltsl written with z or zz may have had no voiced counterpart. It
occurs word initial, medial, and nal: za-a-a-i battle, -ez-zi he comes, a-ru-na-az
from the sea, ke-e-ez on this side, ne-ku-uz mur in the evening time. But see
Yoshida (1998), who claims that original single intervocalic -z- is ldzl and -zz- is ltsl.
65. On the absence of intervocalic -Vk-ku- for lk
w
l see the discussion in AHP 61.
66. Also patterning with q and in Semitic languages is the sibilant . But in the Hittite cuneiform syl-
labary, there are no signs for this that cannot as easily be read as z.
67. That z is representing an aricate ltsl here is conrmed by the fact that in those cases in clitic se-
quences when l-t+s-l becomes l-ss-l and the dental stop is lost, the z-containing signs are not used (see the
examples in 1.111, p. 41).
68. Even if the Ugaritic alphabetic place name lwsnd denotes the city name
URU
Lawa(z)zantiya (del
Monte and Tischler 1978: 23738), it would constitute evidence that Hittite z was not ltsl only if Ugaritic
had an available aricate ltsl phoneme. The uncertainty of interpreting Ugaritic as an aricate ltsl, as it is
in Hebrew, instead of an emphatic, as in Arabic, also makes this evidence uncompelling. The contrast lsl
vs. lzl in Ugaritic does, however, indicate that the -(z)z- in
URU
Lawa(z)zantiya was voiceless.
Orthography and Phonology 1.91
00-HittiteGrammar.indb 37 6/5/08 4:14:47 PM
38
For the possibility that some instances of z stand for a voiced fricative [z] see 1.136
(p. 47).
Fricatives
1.92. The precise value in Hittite of the sound represented by the cuneiform signs
containing cannot be determined at present. The cuneiform syllabary as adopted by
the Hittites had only two sets of CV and VC signs, used non-logographically,
69
contain-
ing sibilants: VlV and zVlVz. Having adopted the z-containing signs for an aricate
(1.90), Hittite scribes had no signs left to indicate a sibilant other than those with .
Therefore the fact that indicates a voiceless (alveo-)palatal sibilant (like English sh) in
Akkadian says nothing about the quality of the sound in Hittite. The voiceless sibilant of
PIE was surely a dental-alveolar lsl, but nothing assures us that it retained that value in
Hittite. All that one can declare with certainty is that Hittite had a single lsl phoneme.
It may have been pronounced as dental-alveolar [s], alveo-palatal [ ], or even palatal
[], or some combination of these, according to environment.
1.93. Evidence cited from the transliteration of names in other writing systems is
complex. Advocates of the lsl interpretation point to writings of Hittite royal names in
Egyptian Hieroglyphs with the sign representing a folded cloth (Gardiner 1957: 507,
sign S 29) and conventionally transcribed by Egyptologists as s+acute accent (on this,
see Edel 1948, 1973). However, the royal name uppiluliumawhich happens to be
the only Hittite royal name that is linguistically Hittite (rather than Luwian, Hurrian,
etc.: uppi- pure + luli- pond + ethnicon -um(a)n-) appears in alphabetic Ugaritic as
pllm. Hurrian names compounded with -arri and spelled with -containing signs at
Boazky appear in Ugaritic alphabetic texts as r. Tropper (2000: 10813) concludes
(p. 113): Es ist somit wahrscheinlich das ug. ll doch als nicht-arizierter, stimmloser
Interdental [ ] gesprochen wurde. Since Ugaritic cuneiform employed another sign for
lsl, it is highly unlikely that the was pronounced [s]. The pronunciation [s] or [ ] (sin
or shin) used by individual scholars today for the Hittite sibilant is entirely conven-
tional, and no inferences regarding the Hittite articulation should be drawn from it.
1.94. The cuneiform signs containing
70
(that is, a, a, i, u, al, ar, at, t,
etc.) represent a voiceless velar fricative [x] in Akkadian (like Arabic khalfah caliph,
Scottish loch, or Hebrew lkh l
e
kha get out of here!). Hittite () also probably stands
for a velar fricative, but other possibilities (such as a pharyngeal fricative) cannot be
excluded.
71
We use lhl in phonemic transcriptions merely for the sake of typographic
69. Obviously they did have [s] signs used exclusively for Sumerograms: s sinew, sr horn, st
esh, etc. These signs eventually comprised the third set of sibilant-containing signs in Akkadian (,
z, s).
70. For typographic economy some publications use h for and s for .
71. Ancient Ugaritic and ancient and modern Arabic use dierent written characters to designate the
two distinct phonemes realized as velar and pharyngeal fricatives (usually transliterated as and , respec-
tively). Since a few proper names written in Hittite with appear in alphabetic Ugaritic with the velar frica-
1.92 Orthography and Phonology
00-HittiteGrammar.indb 38 6/5/08 4:14:47 PM
39
simplicity and do not thereby intend any claim regarding its realization. The sequence
is pronounced as [s] or [ ] plus the separate strong fricative, not as the single sound
[ ] of English sh, German sch, or Turkish .
1.95. Both and show contrasting single and double writing intervocalically:
a-a-an-zi they are, a-a-a-an-zi they remain, e-a-an-zi they sit, and e-e-a-an-
zi they work, produce. For the l contrast, see the minimal pair a-a-a-an and
a-a-a-an blocked, clogged (part. sg. neut.) versus a-a-a-an (noun denoting a
kind of obligation), as well as a number of words that consistently write intervocalic :
e.g., lau(wa)i- to pour,
GI
zaurti-, the verb wa-lwe- to turn and its derivatives
waatar, waear, mur (muni, etc.) time, the second in the medio-passive
ending -()aari, -()aati, etc.
72
The l contrast is generally thought to reect a
tensellax or voicelesslvoiced distinction (Luraghi 1997a: 1.2.3). The laxlvoiced sound
of nongeminate is reected in the spelling of the Hittite personal and royal name
m
Tudaliya in Ugaritic alphabetic script as tdl (where the phoneme written is a laxl
voiced counterpart to , the contrast being [] to [x], the same as Arabic to ; see also
Hebrew tidal < `tudal in Genesis 14:1, 9), and in the place name written gt trnds
winepress of T.,
73
which as in the citylcountry name Taruntaa certainly contains
the name of the stormgod
d
Tarunta. If so, note that both cuneiform and t are reected
as voiced and d.
1.96. The examples of consistent contrasting and , on the other hand, are inter-
preted as consonantal length (i.e., gemination) (Luraghi 1997a: 1.2.3).
Sonorants
1.97. Hittite sonorants include two nasals lml and lnl, two liquids lll and lrl, and two
approximants or semivowels lyl and lwl (European scholars often write these as i _ and
u_). The phoneme lrl never occurs in intial position, nor lml in nal position. The simple
nasals and liquids contrast with geminates in intervocalic position: ku-na-an struckl
killed (part. sg. nom.-acc. neut.), ku-un-na-an righthand (adj. sg. acc. com.), a-ra-a
colleague, ar-ra-a anus, ma-a-la-i she agrees, ma-al-la-i she mills. Examples:
1.98. lml: la-a-ma-an name, a-me-e-i in spring, u-u-ma-an-te-e all (pl.
nom. com.), me-ma-al meal, groats.
1.99. lm:l: am-mu-uk me, kam-ma-ri in a swarmlcloud, nam-ma again, next,
dam-mi-i-a-a oppression, mi-im-ma-i refuses, wa-an-nu-um-mi-ya-a orphan.
tive , the velar fricative interpretation is slightly more likely. Also favoring the velar fricative interpretation
is the writing of
m
Tu-ud-a-li-ya as tdl (see 1.95). But given the mistakes in hearing phonemes from other
languages, this argument can only be suggestive, not conclusive.
72. There are, of course, occasional lapses: me-na-a-an-da (KBo 7.14 obv. 7, OS) for normal
me-na-a-a-an-da.
73. See also the place name trds, with nasal reduction.
Orthography and Phonology 1.99
00-HittiteGrammar.indb 39 6/5/08 4:14:49 PM
40
For the problem of apparent intervocalic geminate m in words using CVm-mV see
1.24, p. 19.
1.100. lnl: ne-e-wa-an new, da-me-da-ni to another, p-ya-na-iz-zi he rewards,
-wa-nu-un I came, pt-ti-nu-zi he runs o with, ga-p-nu-u threads, ki-nu-un
now, ku-na-an struck, killed, -na-at-ta-al-la-an merchant, u--ni-ik-zi injures,
-nu-wa-an-zi they decorate, pu-nu-u-a-an-zi they inquire.
1.101. ln:l: an-na-a mother, u-ul-la-an-na-az out of wantonness, t-a-an-na-i
he shall harvest (a grapevine), zi-in-na-i he nishes, a-ra-wa-an-ni-i free, an-ni-i-
ki-iz-zi he performs, an-nu-u mothers (acc.), in-na-ra-a independently, in-na-ra-
u-wa-an-za vigorous, in-na-ra-wa-tar vigor, zi-in-ni-it he nished, ti-in-nu-zi he
paralyzes, u-un-na-i he drives (here), u-un-na-i he lls.
1.102. lll: la-a-la-an tongue, ma-a-la-i he agrees, kar-ma-la-a-a-i he causes
permanent injury, da-a-la-i he leaves, a-a-li corral, pen, i-da-a-lu-u evil, i-i-li
in the courtyard, lu--li-ya in a pond, tu-li-ya to the assembly, wa-a-tu-la-a of
althe sin, me-ma-al meal.
1.103. ll:l: ma-al-la-i she mills, -na-at-ta-al-la-an merchant, i-kal-la-a-ri he
tears o, u-u-e-el-li-ya-az from the mud pit, il-lu-ya-an-ga-a serpent, e-i-il-
li-ya-a of purity (Hurrian loanword, 4.16, p. 86), ta-az-zi-il-li-i (a kind of priest),
u-ul-la-an-na-az out of wantonness, u-ul-la-an-zi they contest.
1.104. lrl: a-ra-a colleague, ke-e-a-ra-a hand, ka-ru- long ago, formerly,
e-a-ri he sits down, ku-u-ru-ri to a foreign (land), i-wa-a-ru dowry, inheritance
share, ta-ra-an-za promised,
GI
ma-a-ri-in spear, ut-tar word, thing, r above,
asdur brushwood, kurur hostility.
1.105. lr:l: ar-ra-a anus, a-ar-ra-an-zi they divide, -wa-ar-ra for help, pr-
ra-an-da across, mar-ri rashly, mar-mar-ri in a marsh(?), me-er-ra-a-an-ta-ru let
them disappear, mi-ir-mi-ir-ru-u drainage waters, foul waters (pl. acc.), u-ur-re-er
they came to the aid of, me-ur-ri
I.A
times.
1.106. lyl: i-ez-zi he doeslmakes, i-an-zi they go, ma-ya-an-za young adult, u-
wa-i-ez-zi he shall look, ta-i-ez-zi, ta-a-i-e-ez-zi and ta-ya-az-zi he steals, ta-a-i--ga-
a two-year-old, a-a-i--u (an animal), ta-ki-ya in another, da-a-i puts (< dai-).
1.107. lwl: wa-a-tar water, -ez-zi he comes, -e-te-ez-zi he builds, a-ra-a-u-
wa-a exempt, na--i not yet, ne-e-wa-an or ne-e-u-wa-an new, ne-e-u-[it] or n-
u-i-i[t]
74
with new, i-wa-ar like, i-wa-a-ru dowry, -wa-a-tar vision, sight, u-
wa-i-ez-zi he shall look, du-wa-a-an on one side, da-u-wa-an-za blind, a- see!,
a-az-zi- rite.
74. In preferring the reading n-u-i-i[t] to ni-u-i-i[t], the issue is similar to the case of pl-di, for which
see 1.71, p. 31.
1.100 Orthography and Phonology
00-HittiteGrammar.indb 40 6/5/08 4:14:50 PM
41
1.108. Rarely, an initial w followed by a vowel is written with or u preceding a wV
sign: -wa-ar-kn-ta-an (warkant-) fat, -wi
5
-te-ni-it with water, -wi
5
-ta-ar wa-
ters, u-wa-al-a-zi he strikes, etc. (see Melchert 1984b: 13, 2021).
1.109. As indicated above (1.97), there is a regular contrast intervocalically be-
tween single and geminate nasals and liquids, and, as expected, most words are stable
in showing either one or the other. However, in some instances, words that have a single
n, l, or r in the older language unexpectedly show geminate spellings in NH (NS): e.g.,
ku-en-ner for ku-e-ner they killed (see 12.6, p. 190), i-i--ul-la-a for i-i--la-a
of the treaty, kat-te-er-ra- for kat-te-ra- lower. This feature is particularly common in
verbal substantives with the ending -i: ar-ku-wa-ar-ri
I.A
pleas, minumarri
I.A

sooth-
ings, etc., but is by no means regular there (compare kururi
I.A

hostile). For further
examples see AHP 165. These late geminate spellings cannot be explained away in the
manner of those for m (see 1.24, p. 19). The question of their linguistic reality and
possible conditioning remains to be answered.
1.110. For the dissimilation of uw to um and wu to mu see 1.1261.127,
pp. 4445.
Assimilation of Consonants
1.111. t > This change, which occurs only when t and are divided by a clitic
boundary, seems to have been a phenomenon of OH, clearly reected in OS texts,
sometimes side-by-side with examples in which the lost dental has been restored. OS
rituals show instrumental ka-lu-u-lu-p-i-mi-t(a-a-ta) with their ngers or ka-lu-lu-
p-iz-mi-it with restored -t-. In OHlNS texts we nd further exx.: HE 36c claimed this
assimilation on the basis of kuit kuit p[eta] ta-a-e arnikzi whatever (kuit kuit, neut.)
(the rst man) gave (for the bride), (the eloper) shall give it (-at) to him as compensa-
tion KBo 6.3 ii 67 = Laws 28a (OHlNS) (HE 36 c: < taate und es ihm and
na-a-a-an KBo 5.2 iv 12 [< nu + -at + -an]); we see it also in (
d
Tarunta encoun-
tered eight sons of the gods on the road,) na-a-i inkantat and they bowed to him,
(saying: O Tarunta, our father! Where were you?) KUB 33.66 + KBo 40.333 iii 8
(OH?lMS) ; in the OH?lMS festival text from aria KuT 53 ii 1519 (ed. Wilhelm
2002a: 345) the neuter noun half (see 9.42, p. 165) in the phrase kuit x-++ arzi
what half he holds (in his right hand) is resumed na-a-a-an . . . it[anani] . . . di
he puts it (-at) on the altar; but restored in i 50: [na-a]t-a-an. Other exx. from OHlNS
texts:
GI
piniw[att]a kuit

ilamniit rta (15) nuwara(t)ta kardu nuwarat

URU
Agade
d
ts:+u-a (16) [
GI
ftxti].nr.-e iyandu Let him cut down the paini-tree
that stands for you(?) in his gate house, and let them make it into weapons of
d
ts:+u of
Akkad KBo 22.6 iv1416, see iv 21; ntn.so-ikuwatan . . . [tak']annikuwa(t)
an whether it (is) in the mountain . . . or it (is) in the [pl]ain KUB 43.60 i 57 , and
kad!-du-u-mi-it for kaddutmit with their talons KUB 43.60 i 17 .
Orthography and Phonology 1.111
00-HittiteGrammar.indb 41 6/5/08 4:14:51 PM
42
1.112. tn > nn atttar, gen. attanna (from *attatna), see 4.106 (p. 128).
Since most examples are found in nouns with nom.-acc. in -tar and oblique stem in
-ann- or in their transparent derivatives, the -tn- is sometimes trivially restored by anal-
ogy, whence exceptions like iyatna (gen. of iytar abundance) and takatniya- to
make level (< takatar, takann- (a) plain). There is no need or justication for attrib-
uting the -tn- in these examples to Luwian inuence. -tn- is also preserved in transparent
derivatives such as the causatives atnu- to dry (something) out to at- to become dry
and tit(ta)nu- to make stand up to tiya- to stand. Luwian inuence is clear in the case
of the loanword uitar (gen. uitna) wild animals and is likely in other examples such
as artar oense (d.-l. aratni, abl. aratnazi) and kattawatnalli- plainti. See HE
32 2 and Neumann 1964: 50 with literature, among others.
75
A loanword explanation
is possible but unprovable for examples such as mitnaz and wetna- (a wild animal).
76

One clear exception to the rule in a native Hittite word is utne land, country. For the
latest discussion on the derivation of this word and the reason for the nonassimilation,
see Oettinger 2000b: 18384.
1.113. The clitic -a that means also, and geminates a preceding consonant, reect-
ing a prehistoric assimilation of *-C+ha (see Luwian -a also, and). After vowels also,
and appears as -ya. This clitic must not be confused with the postconsonantal clitic -a
of OH that does not geminate a preceding consonant. See 29.24, 29.25 (p. 395).
1.114. Addition of -a and to a sg. nom. in l-antsl (spelled -an-za) leads to [-ants.sa],
spelled -an-za-(a-)a (HE 25b).
1.115. In OH and MH the clitic and appended to a logogram was attached to its last
consonant (i.e., on the case ending): itoi-a (= au + -a) and the king. In NH the
scribal habit changed: the case ending no longer occurred on the phonetic complement,
and the clitic took the form -ya: itoi-ya and the king (nom.).
1.116. When the clitic and occurs on ablatives, the endings -az-zi-ya and -ez-zi-
ya
77
are not to be interpreted as *. . .azya or *. . .ezya. The i is rather a genuine reex
in protected non-nal position of the earlier ablative ending *-ati > -azzi > -az (3.32,
p. 76): mandazziya KUB 26.1+ i 27 and
TUG
termazziya KUB 22.70 obv. 56 , UN-
azziya (i.e., antuazziya) ibid. 57. There are rare examples of this earlier, longer
ablative even in nal, unprotected position: kzzi KBo 4.2 iii 14 and kuzzi HKM 24:45
(MHlMS), KUB 15.34 iii 43 (MHlMS) , KUB 16.39 ii 13 , KBo 4.9 vi 6, 10 (for expected
kez and kuez) (Sommer and Ehelolf 1924: 29).
75. The aberrant hapax sg. gen. uitta with its geminate -tt- remains problematic, despite the eorts of
Carruba (1966a: 1819 n. 27). Luwian inuence is also possible for the unusual MH variant annetnit for
regular annenit (see AHP 27273 with reference to op for an apparent change sn- > -tn- in Luwian; and
compare the Southern American English pronunciation [rdnrt] for isnt it).
76. We disagree with the proposed restoration [. . . . . . u-]-e-et-na-a in KUB 1.16 ii 46 (HED H
353), which would assign the only occurrence to date of wetna- to uedar wildlife.
77. As in kez . . . kezziya.
1.112 Orthography and Phonology
00-HittiteGrammar.indb 42 6/5/08 4:14:52 PM
43
1.117. Some cases of z before or after a nasal may stand for a voiced fricative [z]
resulting from voicing assimilation of s (see 1.136, p. 47).
1.118. nm > mm At clitic boundaries n may assimilate to immediately following m,
producing a geminate consonant that is occasionally reduced to a single m:
78
ad-dam-
ma-an < attanman my father (sg. acc.) KBo 3.44:9; tu-ek-kam-ma-an < tuekkanman
my body KUB 30.10 obv. 14 , ln.xr-am-ma-an < ln.xr-anman of my servants
KBo 3.27 obv. 10; ma-a-am-ma-an < mnman KUB 30.10 obv. 22 ; not graphically
geminate a-a-a-me-et < aanmet my obligation for aan-services KBo 6.2
+ KBo 19.1 ii 19 (Laws 40, OS), la-a-am-ma-a-mi-it < lmanmit my name KUB
1.16 iii 13 (OHlNS).
1.119. n > The assimilation of n to at clitic boundaries is fairly common
already in OH (Otten and Souek 1969: 5758): e-er-ma-a-me-et (< rmanmet), er-
ma-a-ma-a-kn (< ermanmakan), an-na-a-ma-an-na (< annanmanna), a-tu-
galka-a-me-et, da-a-a-u-u-ma-a-ta < dunmata. The resulting geminate is
regularly spelled as such between vowels but is occasionally reduced to : an-na-a-a-
an, p-e-ra-a-i-it, [tar-l]i-pa-a-a-an, kar-di-mi-ya-at-ta-a-a-an, and na-a-a-an,
but
d
i--um-m[i-in] <
d
iunummin our deity KBo 3.22:39 (OS);
d
i--u[m-(mi-in)]
ibid. 41. Less usual, n > z: p-e-ra-az-mi-it (< pranmit) (OS). Beginning in MH, one
nds back-forms in which the n was secondarily restored: na-an-a-an, ma-a-an-i, ma-
a-a-an-a-ma-a, p-ra-an-i-it.
1.120. An assimilation n > is also seen in the -ke-sux forms of some roots end-
ing in -n-: ku(w)ake- to kuen- to kill. Some other verbs showing earlier assimilated
forms, such as imperfective stems aike- (< anna- to judge
79
) and tarike- and
tar(a)ke- (< tarna- to let; Otten 1969: 5 with n. 4), occasionally later developed more
regular, unassimilated forms: anneke- and tarneke-. The unassimilated sequence
n is rare in Hittite: a-an-i he wipes (and other forms of this verb an-), the noun
anaiwi- corpse(?).
1.121. An n can arise where it is etymologically unjustied by remote assimilation
to another nasal preceding or following in the word (Oettinger 1994; Kimball 1999:
31819): aannanza for aannaz from the family, tetenanza with thunder,
80

antenaz from the hole for attenaz, naarantan for naarattan fear (sg. acc. of
naaratt-), li-in-kn-ta KUB 13.35 i 9 for correct li-in-kat-ta he swore (HE 31b);
tanantezzi for tannattezzi or tannattezi it will be laid waste (Riemschneider 1970: 32).
See further 4.110 (p. 129).
78. Goetze 1930b: 56, with many exx.; HE 31, 36. But the claim for alkim pianzi (with dupl. alkin
pianzi) in Laws 112 has recently been rightly questioned by Otten (1990), who argues from a duplicate that
al-ki-im should be read n+t--t+, i.e., as an Akkadian genitive.
79. Kept distinct from ake-, the imperfective stem of a()- to give birth and aleke-, that of
a()-le- to open.
80. For an alternate account of the ablatives in -anza see Jasano 1973.
Orthography and Phonology 1.121
00-HittiteGrammar.indb 43 6/5/08 4:14:53 PM
44
1.122. mn > mmn > mm m is geminated before n (spelled -Vm-nV- or -Vm-mV-
nV-) and then the sequence -mmn- is often simplied to -mm- (see AHP 15253 with
refs.) and even written -m- (as in u-mu-ma-a unite! in KUB 29.1 ii 43 ; see 2.47,
p. 60): pl. d.-l. of
NINDA
aram(a)n- (a type of bread) is a-ra-am-na-a, a-ra-am-ma-na-
a, and a-ra-am-ma-a (see 4.69, p. 108); with the sux -uman- (2.45, p. 60) we
nd
URU
utummana the man of uta and tameumma- to make (of) another and
tameumme- to become (of) another from the adjective tameum(a)n- (belonging to)
another.
1.123. mn > nn does not exist in Hittite. The genitive ilanna in Laws 82 and 89
was derived by Friedrich (1959) from ilamna, genitive of ilammar portico, but by
Laroche (1957a: 1819) from unattested *ilatar. The former solution is unlikely, since
one would have to assume mn > mm and mn > nn in the same language during the same
time period.
1.124. Possible partial assimilations such as nt > nd are masked by the lack of a means
in Hittite cuneiform for showing the voiced-voiceless distinction in the obstruents when
they directly follow another consonant (-an-ta- and -an-da- are in free variation).
Dissimilatory Changes
1.125. dtltt > zt Hittite preserves a PIE rule by which the rst of two dental stops is
dissimilated to an aricate [ts].
81

It is by this change, e.g., that le:d-tul let him eat ap-
pears as [e:tstu], spelled e-ez-(za)-du or redundantly as e-ez-za-a-du (see 12.3 [p. 187]
for further examples). This aricate was also retained even in cases where the following
dental stop was itself assibilated to an aricate, producing an unusual geminate aricate.
Thus le:d-tsil he eats appears as [e:ts.tsi], spelled e(-ez)-za-(az-)zi.
1.126. uw > um A word-internal sequence -uw- is dissimilated to -um-,
82
as seen in
the rst-plural verbal endings -mani, -meni, and -men
83
and the verbal substantive end-
ing -mar and innitive ending -manzi (see 25.225.3, p. 330), which take the place of
regular -wani, -weni, -wen, -war, and -wanzi in the nu-causative verbs (HE 29a; Otten
and Souek 1969: 7879): warnu- + -weni > warnu(m)meni, in the verb au-lu- to see
in umeni we see and aumen we saw (OS), earnu- + -wi- > earnumi- make
blood-red (Melchert 1984b: 27), and possibly also in the bird name tarumaki- wood-
pecker(?) from taru- wood, tree + waki- (see wak- to bite) (see also below p. 63,
n. 28, 11.22 [p. 186], 12.4412.47 [pp. 210211]). This dissimilation also aects the
secondary sequence -uw- in forms like tarnum(m)eni we leave < tarn-uweni to tarn(a)-
to leave. For the source of -uw- in these cases see 1.82 (p. 34). Third-plural present
forms like arnuwanzi and participles like arnuwant- are not counter-examples, since
81. This rule is usually characterized as s-epenthesis, but as per Weriand Merlingen and others is more
accurately viewed as dissimilation (see Melchert 2003c: 154, with references).
82. Kammenhuber 1969b: 137; Melchert 1984b: 2227. Observed already by HE 29.
83. Already in OS in a--me-en.
1.122 Orthography and Phonology
00-HittiteGrammar.indb 44 6/5/08 4:14:54 PM
45
these represent arnu-+ -anzil-ant- without inherent -w-, and the rule does not apply to a
secondary -w- inserted to ll a hiatus in a sequence -uV-. The rule also does not apply
in secondary sequences of -uw- occurring at a clitic boundary, thus pnduwaz aandu
KBo 3.1+ ii 13 (ed. THeth 11:2829).
1.127. wu > mu idlu- bad forms its pl. nom. com. as idlawe, but its acc. pl.
com. as idlamu,
84
and the pl. acc. com. of newa- new is ne-mu-u. Similarly, in
texts from MH into the 13th century the pl. acc. of iu- deity, god is urorn.xr-
mu-u (= imu < iwu) (Neu 1974: 12122 with notes). Along with HE (29, 75,
169, 174) and Otten and Souek (1969: 57), we regard this as a true dissimilatory sound
change (Lautwechsel) against Neu (1974: 12122), who maintains that this writing is
a graphic substitute (Ersatzschreibung) and indicates a pronunciation lsiwusl, not
lsimusl. See other unusual pl. acc. com. forms such as zaimu dreams (< zaai-),
laemu errands, trips (< laa-?), allamu < alli- big (see p. 95, n. 76).
1.128. The opposite of remote assimilation, namely, remote dissimilation (Oettinger
1994: 31014), occurs in the following examples:
1. nm > lm lman name (< pre-OH `nman), lammar hour (< pre-
OH *nammar), perhaps also l let not (< pre-Hittite *n) arising from
combinations lman and lmu.
2. mn > ml irmanant- having sickness (2.24, p. 56) > irmalant- sick, ill.
3. ll > lr apalzil (kind of stew?) > apalzir.
4. rr > rn *urri > OH -ra-a-ni and post-OH wa-ra-a-ni it burns, and
wa-ra-a-nu let it burn from the verbal stem war- and the medio-passive
endings -ari and -aru.
Loss (Deletion) of Consonants
1.129. There are some examples of sporadic loss of stops in the middle of sequences
involving sonorants: e.g., kinu- (2) for lkistnu-l (usually written kitanu-) to extin-
guish, innuzi for inganuzi (lhinknutsil) makes to bow, amanzi for amangazi
(lhamanktsil)intertwines. For further examples, see AHP 16667. There are too many
examples for them all to be due to scribal errors.
1.130. There are genuine examples of sporadic loss of lsl before ltsl: a-ar-zi for
a-ar-a-zi, pa-ap-pr-zi for pa-ap-pr-a-zi, tk-ki-iz-zi for tk-ki-i-zi, ta-pu-uz-za for
ta-pu-u-za. For reasons against assuming this loss in pres. sg. 3 verb forms in -e(z)zi
(Riemschneider 1970: 65), see Watkins (1973), who showed that these belong to stative
verbs in -e- (10.11, p. 177).
1.131. There are a few instances that suggest a genuine if sporadic deletion of next
to stop: idlawatti beside idlawati you harm, and warput for expected `warput
Bathe! for which Neu (1968b: 191) compares i-ip-pa-an-du-un Bo 2738 iv 5 (now
84. Goetze 1927: 120 n. 4; Sommer and Falkenstein 1938: 76 n. 1; HE 29b; Melchert 1984b: 2223.
Orthography and Phonology 1.131
00-HittiteGrammar.indb 45 6/5/08 4:14:55 PM
46
KUB 58.79 iv 5 ) for regular i-ip-pa-an-da-a-u-un (lsipandhunl). For ar and geni-
tive ena see 4.102 (p. 126).
1.132. There was a weak articulation of intervocalic and preconsonantal r, causing
sporadic nonrepresentation in writing (HW 30b; EHS 6570; AHP 125). Intervocalic:
p-an for p-ra-an before,
TUG
kuear for
TUG
kurear (a garment), pu-u-ut for purut
mud. Preconsonantal: waggante for warkante fat ones, artati- for artarti- mush-
room(?). To these one might also add: naarnukewa[ran kuie] uritenukewaran
kuie [those who] were frightening [him], those who were terrifying him KUB 60.157
iii 78 , where one expects the third preterite plural ending -er. The loss in this case and
in a(r)tarti- may be dissimilatory, but the loss of r before w could also be compared with
nw > w (1.135). This might explain the use of the prop vowel i in kaniriwant- thirsty
(Elisabeth Rieken, pers. comm.). The appearance of the quotation particle as -wa- before
a consonant may be a regularization of the preconsonantal loss of r (see 28.4, p. 354).
1.133. Final r was lost prehistorically in nal position after an unaccented vowel
(Neu 1982; Melchert 1988b; Yoshida 1990: 10812). Most cases of preserved nal
-ar reect original syllabic *r. Old Hittite reects this contrast in plural paprta beside
singular paprtar impurity, myata alongside miytar prosperity, atrea along-
side atrear message, kuduwta slanders, false accusations, ullta disrespectful
acts; a-an-ne-e-a-e-et, [pr-t]a-a-u-wa-a-i-it-wa, a-at-ta-ta-mi-it. Compare also
plural
UZU
appea (also written
UZU
n.nr.-a) alongside singular
UZU
appear body
part. This occurs even when the nal r is protected by a sux: attatamit my wis-
dom (Sommer and Falkenstein 1938: 179). This loss probably also accounts for the
appearance of the quotation particle -war as -wa in nal position (see 28.4, p. 354) (cf.
Goetze and Pedersen 1934: 3032; Sommer and Falkenstein 1938: 66).
1.134. A prehistoric loss of intervocalic `y is still reected in the oblique forms
of i-stem adjectives such as sg. gen. alla (see Sommer 1932: 357 with lit., Melchert
1984b: 4445, and 4.37, p. 94) and in OH medio-passives such as ne(-e)-a turns (see
1.142, p. 48).
1.135. In certain circumstances n is lost before immediately following , k, t, w, or
z (Kimball 1999: 31619; see also Goetze 1930b: 56, with many exx.; HE 31 and
36). : a-a-a (imp.) for a-an-a seek; k: i-kn-ni KBo 3.36 i 9 for inkani, li-ku-
wa-an-ni KUB 9.31 i 42 for linkueni we will swear, ma-an-ni-ku-wa-an IBoT 1.36 i
17 (MHlMS) for manninkuwan near, pa-ga-u-wa-a KUB 24.13 iii 23 for pangawa
of the multitude,
LU.ME
a-ku-ni-e-e KUB 39.71 i 5 as plural of
LU
ankunni priest,
za-ki-la-tar KUB 42.91 iii 8 for zankilatar compensation; t: u-u-ma-da-az KUB 13.2
iii 35 (MHlNS) alongside mandaz, -e-e-a-ta HT 1:30 with dupl. -e-e-a-an-ta
KUB 9.31 i 37 they clothe themselves, i-u-uz-zi-ya-te-e KUB 9.31 i 39 with dupl.
i-u-uz-zi-ya-an-te-e HT 1 i 32 belted, i-ip-pa-ti KUB 6.46 i 60, 64 (NH) for sip-
panti he oers; also across word boundary: ki-i-a te-ez-zi thus he says (for kian
tezzi) (Kmmel 1967: 17); w: ma-a-wa < mnwa whenlif (+ quotation particle) KUB
1.132 Orthography and Phonology
00-HittiteGrammar.indb 46 6/5/08 4:14:55 PM
47
36.45:2 and passim; z: li-in-ga-zi < linkanzi they swear, a-an-a-zi < ananzi they
seek; t.xr-an-na-za < antuannanza (ergative of antutar) KUB 1.1 i 30 . Since
these spellings alternate with ones showing the n, it is possible that the nasalization was
preserved in the preceding a vowel and not entirely lost. We cannot decide this question
from the orthography.
85
In any case, the alternate spellings without n are too frequent
to be emended as scribal errors.
LU
alugatallattin < alugatallantin your messenger in
VBoT 1 obv. 19 may show the same loss or instead assimilation of -nt- to -tt- at a clitic
boundary. See for such assimilations 1.1181.119 (p. 43) and 6.5 (p. 140).
Epenthesis
1.136. A secondary sequence [ns], usually resulting from an earlier syllabic nasal
plus s, appears in Hittite as nz, contrasting with assimilation of original *lVnsVl to
lVssVl (AHP 121): *n

ss > *anss > anz us. Similarly, liquid (l or r) + alternates
with liquid + z in some cases:
d
Gule and (verb) gul- alongside noun gulzi-, nom. gul-
zatar IBoT 3.101 obv. 4, abl. gulzattanaz; as well as in maratarri- and marzatarri-,
d
Paurze and
d
Paure, piraanna and pirzaanna,

arzana and

arana, nazi
and nai (Carruba 1966a: 36; AHP 121), guripi- and gurzippant-. The status of the z
in such examples is not clear. It may stand for [ts], reecting epenthesis of t in sequences
of sonorant plus s (AHP 121), or it may represent a voiced fricative [z], showing spo-
radic assimilation of [s] to the preceding sonorant. The latter alternative is supported
by the alternation of and z in (a)mankurwant- bearded and z(a)mankur beard
(AHP 94), which appears to show sporadic assimilation of [sm-] to [zm-] (contra AHP
172). See Oettinger (1994: 322), however, for an explanation of z(a)mankur in terms
of epenthesis.
Other Alternations
1.137. The apparent word-internal alternation of and z in zaai- and zazai-
dream(HE 27 c, with literature) is to be interpreted as showing alternate graphic
representations of initial ltsh-l, i.e., a sequence of aricate [ts] + fricative [x] (Elisabeth
Rieken, pers. comm.). Likewise, initial zaklg- represents ltsk-l: (1) in the rare imperfec-
tive stem zake- to dai- to put (see p. 34, n. 60) without anaptyxis vs. regular zikke-;
(2) in zagarai anus (lit., feces-mouth). In the latter, original initial `sk- has for some
unknown reason become ltsk-l instead of regular ik- with prothesis (see igauwant-
covered with dung). The same development is seen in z(a)kkar beside regular akkar
dung (see AHP 121).
1.138. One occasionally nds k for and vice-versa when either is in contact with
a sibilant, an interchange that suggests a partial neutralization in this environment:
85. Possible support for the presence of a nasalized vowel may be found in the spelling ge-em-zu (KBo
5.3 i 24) for ge-en-zu lap; mercy, where an etymological m is extremely unlikely (see Weitenberg 1984:
165). The example u-in-pa-an (KUB 33.36 iii 4) for usual u-im-pa-an suggests the same reduction to a
nasalized vowel also before a labial stop.
Orthography and Phonology 1.138
00-HittiteGrammar.indb 47 6/5/08 4:14:56 PM
48
amekanza for ameanza spring, but also
UZU
iia- for
UZU
ikia- back KUB
33.112 iv 14 (HE 28a) ,
MUNUS.ME
azara- for
MUNUS.ME
azkara-
86
and iaru- for
ikaru- (Kronasser 1966: 98; Kmmel 1967: 109). More doubtful is the equivalence
of ta-a-a-at-ta-ri and `takattari (Kmmel 1967: 109).
1.139. The sequence nunu can appear haplologically as nu in: kitanun KUB 27.67
iii 11 for *kitanunun I extinguished, mernun KUB 13.35 i 28 for *mernunun I caused
to disappear. Such haplology is not, however, limited to nunu: kap-pu-u-wa-ar count-
ing HKM 21:6 (MHlMS) < `kappuwawar, arkuwar plea < *arkuwawar. On haplol-
ogy at Maat see Honer forthcoming.
1.140. Metathesis of and p is possible in the passage of Akkadian loan
LU
+-st-ri
into Hittite
LU
a-p-i- (4.16, p. 86) (Kmmel 1967: 9596), as well as AkkadianlHur-
rian and Hittite gurip-lgurzip-lgurpii- (Kmmel 1967: 1056). Metathesis of and r is
attested in the alternation aarali- and araali- (HED A 12829). Forms like OH e-uk-i
you drink alternating with post-OH e-ku-u-i (lek
w
sil) reect not metathesis but the
uncertainty of the scribes as to how to write a unitary labio-velar consonant [k
w
]. See
also tar-uk-zi for ltark
w
zil, elsewhere written tar-ku(-uz)-zi.
Phonotactics
1.141. Consonants. All voiceless obstruents, both glides (w and y), and all so-
norants except r occur word-initially (on the neutralization of voicing in word-initial
and word-nal stops see 1.86, p. 36). The aricate ltsl and the fricative lsl appear
freely in word-nal position; lhl occurs principally in the imperative second singular of
verbs with stems ending in . Word-nal lnl, lll, and lrl are common; there is no nal
lml. The glides lyl and lwl occur word-nally if one interprets diphthongs as consist-
ing of vowel plus glide. The nature of the cuneiform writing system makes it dicult
to determine the status of consonantal sequences. For a mere attempt to sketch some of
the possibilities see AHP 11014.
1.142. Vowels. Most Hittite vowels occur freely in word-initial and word-nal po-
sition. For possible restrictions see AHP 114. Hiatusi.e., a sequence of two vowels not
forming a diphthongis relatively rare in Hittite. The most certain common example
is the sequence [-e.a-] in cases such as ne-(e-)a(-ri) turns and xtn-e-a of the lands.
But there is a tendency for this hiatus to be lled with -y-, and one also frequently nds
forms such as ne-ya-ri and ut-ne-ya-a (see Melchert 1984b: 4647, with references).
Still another likely case consists of the sequences -a-(e-)e and -a-u in the pl. nom. and
acc. com. of i-stem adjectives with ablaut (3.37, p. 78), such as al-la-(e-)e and al-
la-u great. They are probably pronounced as [sal.la.e:s] and [sal.la.us], but absolute
proof is not possible. For other, more dubious cases see AHP 115. For the adjective
u-u(--u) full see 1.8, p. 12.
86. For this and some other examples as reecting a change kr > hr see now Oettinger 2000a.
1.139 Orthography and Phonology
00-HittiteGrammar.indb 48 6/5/08 4:14:57 PM
49
1.143. It is impossible to determine whether sequences of lil and lul followed with-
out written glide by a dierent vowel are always pronounced with an intervening glide
as [iyV] and [uwV] or are sometimes pronounced with hiatus as [iV] and [uV]. The
spelling i-e-a-ni-it alongside i-i-e-e-a-ni-it for with beer (ins.) may represent a
genuine alternate pronunciation [si.e:ssniT] alongside [siye:ssniT] or merely another
spelling for the latter. Likewise, kap-pu-ez-zi alongside kap-pu-u-(e)-ez-zi counts may
stand for [kapu.e:tsi] or [kapuwe:tsi], and pa-ri-an alongside pa-ri-ya-an for [pari.an]
or [pariyan].
87
The occasional insertion of a glide in vowel sequences at clitic bound-
ary (e.g., tk-ku-wa-at KBo 6.2 ii 35 [OS] for takkuat or a-a-a-ni-ya-a att. iv
85 [NH] for aania) also does not prove that all such sequences were pronounced
with an inserted glide.
Onomatopoeia
1.144. Onomatopoeia is dicult to identify, even in a living language. Nevertheless,
there are nouns for natural phenomena (storms), animals, or musical instruments that
appear to imitate the sounds made: ariari- thunderstorm, akuwakuwa- (name of a
small amphibian), aluwaaluwa- and kallikalli- (names of birds named for the repeated
call they make), uupal- (percussion instrument; see 2.19, p. 54), galgalturi- (per-
cussion instrument). These examples share the feature of reduplication (2.562.57,
p. 62), a universal characteristic of onomatopoetic words. A few of the reduplicated
verbal stems (10.310.5, pp. 173174) may therefore imitate the sounds they denote,
such as aar- to laugh, teta- to thunder, and tataiya- to whisper.
88
Compare
the archer-hunter who utters the repetitive cry ee ee of his prey (24.11, p. 320, and
1.8, p. 12).
Accent
1.145. Our knowledge of the Hittite accent is quite limited. Some scholars have
proposed that plene spellings (see 1.46, p. 25) directly mark the position of the accent
(Hart 1980; Carruba 1981). However, we subscribe to the view that plene spellings for
the most part directly indicate vowel length. Nevertheless, vowel length may indirectly
show the place of the accent in that prehistoric long vowels were preserved and short
vowels lengthened under the accent (see, e.g., Kimball 1983; Eichner 1980 15064).
Drawing inferences about the accent in this way faces several problems. First, the use
of plene spellings is far from consistent. Second, it is by no means assured that indica-
tion of vowel length is the only function of plene spellings (again see 1.46, p. 25).
89

Third, there is no guarantee that the accent in attested Hittite has consistently remained
87. To further complicate matters, in many instances we also cannot exclude that it is the high vowel that
is lacking: either i-e-ni-it or i-i-e-e-a-ni-it (or both) could represent spoken [sye:ssniT].
88. For an onomatopoetic verb to whistle in Hattic see Singer 2002a.
89. It at least seems clear that, when a word contains two plene spellings, one of the instances cannot
reect the accent: e.g., idlaw evil (pl. nom. com.). Such examples are far from rare.
Orthography and Phonology 1.145
00-HittiteGrammar.indb 49 6/5/08 4:14:58 PM
50
in the prehistoric position which led to the supposed lengthening eects. In sum, there
is a high degree of circularity in the reasoning just described. All statements about the
position of the accent in Hittite should be viewed with this caveat in mind.
1.146. There are a signicant number of examples in both nominal and verbal in-
ection where the position of the Hittite accent as reected by plene spellings matches
that reconstructed for Proto-Indo-European and where, consequently, it is widely as-
sumed that Hittite has preserved the inherited accentual pattern. For example, Hittite
sg. nom.-acc. tkan vs. sg. gen. takn, sg. d.-l. takn earth (see 4.71, p. 110) appears
to continue directly PIE `dhhm vs. `dhh- with mobile accent (for the PIE inec-
tion of earth see Schindler 1967). Likewise, zi is vs. ant- being < `h
1
sti vs.
`h
1
snt- (see Oettinger 1979: 8698). For the accompanying stem variation known as
ablaut see 3.37, p. 78). We will call attention to cases on which there is widespread
agreement, but readers should not infer that there is any consensus on the position of
the accent in general.
1.146 Orthography and Phonology
00-HittiteGrammar.indb 50 6/5/08 4:14:59 PM
51
Chapter 2
NOUN AND ADJECTIVE FORMATION
2.1. The Hittite lexicon includes the following parts of speech: nouns, adjectives,
pronouns, numbers, verbs, adverbs, postpositions, conjunctions, and particles.
1
Of
these, only nouns, adjectives, and verbs show productive patterns of stem formation.
2
2.2. Nominal and verbal stem formation in Hittite, as in other older Indo-European
languages, generally consists of addition of a derivational sux to a root or to a stem
consisting of a root plus one or more suxes: e.g., ark- to perish > ark-a- ruin,
destruction and > ark-nu- to cause to perish, destroy; ipant- to libate > ipant-
uzzi- libation > ipantuzzi-aar- libation vessel. However, Hittite also shows signi-
cant traces of the PIE phenomenon widely known as the Caland System (see Meier-
Brgger 2000: 27172, 2003: 28891). However it is to be explained historically, this
process appears descriptively as derivation by substitution: i.e., the (nal) sux of the
base is deleted with the addition of another sux. This happens most typically when
the base is an adjective, and Hittite shows many cases with stems in -u-, -i-, and -ant-.
3

Examples with -u-: park-u- high > parg-ati- and parg-atar height, also park-nu- to
make high and park-e- to become high, tall. Examples with -i-: pal-i- wide >
pal-ati-, pal-atar and pal-ear width, also pal-e- to become wide; parku-
i- pure, clear > parku-e- to become clear and parku-nu- to purify. Examples with
-ant-: appin-ant- rich > appin-a- to make rich and appin-e- to become rich;
papr-ant- impure, guilty > papr-e- to be proven guilty by ordeal. On the model of
the examples with -ant-, Hittite also extended this pattern to the sux -want- (see Mel-
chert 1984a: 80): miri-want- splendid, perfect was reinterpreted as miriw-ant- before
-ant- was deleted in the formation of miriw-a- to make splendid, perfect. For more
examples see the respective paragraphs on specic nominal and verbal derivational
suxes. Since the pattern with substitution was synchronically irregular for (pre-)Hit-
tite speakers, sux additionnot replacementbeing the regular pattern, one also
predictably nds examples of the same suxes with the regular process of addition:
atku- narrow > atku-e- to become narrow, uktri- long-lasting > uktri(y)-a-
1. Negations (such as not) and preverbs (such as up, down) are here subsumed under adverbs,
although in the body of this grammar the latter will often be referred to under the more explicit term
preverb.
2. For verbal stem formation see chapter 10.
3. Our goal here is to describe the situation within Hittite. We make no attempt to distinguish between
inherited examples and innovative examples modeled on the former.
00-HittiteGrammar.indb 51 6/5/08 4:14:59 PM
52
to make long-lasting nakk- important > nakki(y)-atar importance, dignity, mayant-
young > maya(n)d-atar youth and mayant-a- to make young.
2.3. Noun and adjective stems can be subdivided into derived and underived. In us-
ing the term underived, we intend no claim as to the words status in pre-Hittiteonly
that within the historical period of Hittite there existed no more basic form from which
the word in question was derived. We also include possible loanwords in this category,
since they too bear no Hittite derivational sux. What follows is a brief orientation to
the most important classes of underived stems and a survey of productive suxes. The
last comprehensive treatment of Hittite nominal stem formation is EHS. See also the
valuable discussions of selected stem classes in Weitenberg 1984 and Rieken 1999a.
Underived Stems
2.4. The largest class of underived stems ends in -a- (see EHS 16566). A few of
these continue PIE o-stems (Kammenhuber 1969c: 27980; Szemerenyi 1996: 18285;
Meier-Brgger 2000, 2003: F309, F311). The majority of adjectives in PIE were o-stems
(Meier-Brgger 2000: 201; 2003: 218), which are continued by Hittite a-stems.
2.5. Many common gender nouns denoting body parts are a-stem nouns: antaka-
loins(?) (see Melchert 2003a), arra- anus, elinera- eyebrow, arana- head
(also an rln-stem arar, arana), itamana- ear, kaka- tooth, kalulupa- digit,
keara- hand (also an r-stem), lala- tongue, laplipa- eyelid, paltana- shoulder,
pantua- bladder(?), pappaala- esophagus, pata- foot, patala- sole of the foot,
tetana- hair. Likely also belonging here are anaa- upper back(?) and titita- nose,
for which no gender-specic forms are yet attested.
2.6. Other common gender a-stem nouns with no apparent derivational sux are:
aka- gate, ira- border, boundary, laa- campaign, nata- reed; arrow, palzaa-
statue base, pedestal, nega- sister, negna- brother (logograms r and +ni),
dammea- oppression, injury, tea- sleep, etc.
2.7. Underived a-stem adjectives include: ndara- blue, arawa- free, exempt,
kunna- righthand, mara- false, deceitful, nwa- new and pittalwa- plain, simple.
For their inection see 4.10 (p. 83).
2.8. Assured loanwords among common gender a-stems are relatively few: for ex-
ample,

arzana- < Akkadian arzannu groats (Honer 1974),

LU.SANGA
kumra- < Old As-
syrian kumru (Otten 1990b) =
LU
otut
12
(Honer 1996a). Less certain are:
LU
atwaya- (a
cult functionary), perhaps < Hattic atwe (Soysal 2004: 17, 145, 279, 456),

kippa- reed
hut < Sumerian (r.)or.ru (see Taracha 2001), and ipa- (a secretion or infection of the
eye) < Akk. pu (see Fincke 2000).
2.9. Because in post-OH many nouns which were originally not i-stems took on i-
stem appearance in the nominative and accusative due to the so-called i-mutation (see
2.3 Noun and Adjective Formation
00-HittiteGrammar.indb 52 6/5/08 4:15:01 PM
53
4.17, p. 86), we limit our list of underived -i-stems here to those that show exclusively
i-stem inection.
2.10. Among nouns identiable in this manner are: (common gender) auli- artery(?),
sacricial animal, alki- grain, barley, urki- wheel, imeri- rein, kalmi- wooden
log, thunderbolt(?), kappni- caraway seed, cumin, karpi- anger, kurakki- pillar,
lli- pond, cistern, pri- lip, tuzzi- army, and zakki- latch; (neuter gender) elzi-
scales, a(a)ri- lung, diaphragm, li- corral, pen, uwai- stela, kenupi- pot-
tery, earthenware, kullupi- sickle, luzzi- forced labor, corvee.
2.11. Underived i-stem adjectives include: uelpi- fresh, young, kappi- small,
nakk- heavy; dicult, ukturi- lasting, durable.
2.12. The majority of assured loanwords appear in Hittite as i-stems. Common gen-
der examples: aganni- (a container) < Akk. agannu,
LU
apii- (incantation priest) < Akk.
ipu (with metathesis), kakkari- (a kind of bread loaf) < Akk. kakkaru, nura(n)ti-,
pomegranate < Akk. lurindu (`nurindu), ankunni- priest < Akk. ang < Sum.
sanga (with possible stem inuence from Hurrian as the intermediary of the loan?),
arui- censer < Hurr. arui, ai- (kind of wood) < Hattic ai. Neuter examples:
atupli- (a garment) < Akk. utuplu, azzizzi- ear < Akk. assu, tuppi- clay tablet
< Akk. uppu, keldi- well-being < Hurr. keldi,
GI
parnulli- (kind of tree and its wood)
< Hattic (Soysal 2004: 148, 300, 590).
2.13. Examples of u-stems underived within the historical period of Hittite are:
(common gender) au- king, eu- rain, pardu- sprout, shoot, bud, pupu- lover,
paramour, iu- god, taku- testicle(?), wappu- river bank, and wellu- meadow;
(neuter gender) u- goods, alenzu- duckweed, algae, attalu- door bolt, iaru-
tears, iwaru- dowry, inheritance share, kapanu- trunk (of tree), kapru- throat(?),
kapunu- (a unit of land measurement), genu- knee, genzu- lap, womb, mercy, ru-
plunder, booty, tru- tree, wood, and welku- grass.
2.14. Loanwords appearing in Hittite as u-stems are quite rare. Only one example is
at all likely, the common gender
NINDA
taparwau- (kind of bread) < Hattic.
Derived Stems
2.15. Derived nouns and adjectives consist of a root (whether noun, adjective, verb,
or adverb) to which one or more derivational suxes have been added. The resulting
combination constitutes the derived stem, to which the case endings are added (HE
4353; EHS 160343; Kammenhuber 1969b: 291300; Berman 1972b; and Meriggi
1980: 14). The following is a list of the principal suxes for forming derived stems
of nouns and adjectives and examples of the resulting forms. Luwian suxes that occur
in Hittite-language contexts are marked with the notation Luw.
Noun and Adjective Formation 2.15
00-HittiteGrammar.indb 53 6/5/08 4:15:01 PM
54
2.16. The sux -a- forms common gender action and result nouns from verbs (EHS
16667):
4
alwanzaa- sorcery < alwanza- to bewitch, hex, arga- ruin, destruc-
tion < ark- to perish, be destroyed, aa- grandchild (perhaps originally o-
spring) < a()- to give birth to, generate, kara- cut (of meat), sector (of territory)
< kar- to cut o, maniyaa- administration, district < maniya- to administer,
govern (in iter.), maninkuwaa- vicinity, neighborhood < maninkuwa- to ap-
proach, draw near,
NINDA
para- bread fragment < par- to break, fragment, arra-
part, share < arra- to divide.
5
2.17. The sux -ai(t-) (Luw.) (see 2.55, p. 62) forms neuter abstract nouns
from adjectives and nouns, expressing status: tuukantait- status of a crown prince
< tuukantali- crown prince, malaallait- practice of ritual magic < malaallali-
pertaining to ritual magic, nakkuait- status of a scapegoat < nakkuali- scapegoat,
tarpaait- status of a substitute < tarpaali- substitute, lapanallait- status of a
warden of the salt deposits < lapanallali- warden of the salt deposits, muwaddalait-
overpowering might, victoriousness < muwattallali- overpowering, victorious. It also
appears on a few terms for concrete objects: arandanait- (part of a wagon or chariot
wheel). Other nouns in -ai- in Hittite contexts may be of Hurrian derivation.
2.18. The sux -ai- productively forms common-gender action nouns from ver-
bal roots: urtai- curse < u(wa)rt- to swear, iamai- song < iamai- to
sing, itarninkai- illness < itarnink- to make ill, linkai- oath< link- to swear,
maniyaai- administrative district < maniya- to administer, agai- omen, sign
(see akiya- to give a sign), wagai- grain pest, biter < wag- to bite, watai- sin
< wata- to sin, zaai- ght < za- to strike. There are also a few common-
gender nouns formed to nominal bases: ullanzai- defeat (see ullant- defeated). Of
unclear derivation: kurkurai- verbal intimidation, threat, aklai- prerogative, right,
custom, rite, zaai- dream. Most neuter ai-stems are underived. For some examples
see 4.344.35 (pp. 9394).
6

2.19. The sux -al- forms neuter instrumental nouns from verbal bases. Some of
these are assuredly native Hittite: ardal- rock-cutting saw < ard- to saw, uupal-
(a percussion instrument) < up- to make a hollow sound (imitative, 1.144, p. 49),
iiyal- band < ii(ya)- to bind. Others are certainly loanwords from Luwian (see
2.55, p. 62): ariyal- basket < ariya- to lift, winal- sta, club. Some cases may be
either Hittite or Luwian in origin: e.g., tarmal- hammer. Some Luwian stems in -al- are
borrowed into Hittite as common gender stems with either a-stem or i-stem inection:
e.g., pappaalali- esophagus.
4. We dier from Kronasser (EHS 96), who interprets several these forms as containing a sux -a-.
5. Berman (1972b: 11516) adds further examples, including uncertain ones such as
LU
kta-.
6. On the problem of
(MUNUS)
azgarai female musicians to azgara- and iai tail beside a- see
Oettinger 1995: 21415. Also relevant is Rieken 2004a: 535, 540, who only cites Rle on azkara-. Also
see 3.21 (p. 72).
2.16 Noun and Adjective Formation
00-HittiteGrammar.indb 54 6/5/08 4:15:02 PM
55
2.20. The sux -ala- forms adjectives from nouns: genzuwala- merciful < genzu
mercy, liiyala- pertaining to the liver < lii- liver. Subsequent substantivization
of adjectives of this kind in the sense the one pertaining tolwho deals with . . . leads
to productive use of -ala- to form what are eectively agent nouns from other nouns:
LUlMUNUS
arkamiyala- player of the arkammi-instrument, arzanala- (a functionary) <

arzana- inn, auriyala- border guard < auri- border post, ariyala- one who oers
NINDA
ari-,
LUlMUNUS
uwaanala- worshiper of
d
uwaana, ipanduziyala- one who
oers libations < ipanduzzi libation, karimnala- temple servant < karimn- temple,
LU
erala- < er(a)- (a substance), `palala- (written i.xsxi-la-) guide, cara-
van leader < pala- road, journey, caravan,
7
tawalala- one who serves tawal-drink,
LU
waliyala- one who serves wali drink,
LU
zuppariyala- torch-bearer < zuppari-
torch. This pattern of forming agent nouns eventually leads to deverbal examples:
gangala- scale (for weighing) < gank- to weigh,
LU
tarwekala- dancer < imperfec-
tive form of tarwi- to dance,
LU
laiyala- traveler(?) < laiya- to travel. One such
adjective derives from an adverbial base: tuwala- distant < tuwa at a distance.
2.21. The suxes -alla- (Luw.) and -alli- (Luw.) form nouns from nouns (see 2.55,
p. 62). Nouns in -alla- or -alli- referring to persons are common gender (appaliyalla-
plotter, usurper < appali- trap, iru(i)talla- basket-carrier < iruit- basket), while
those referring to things are neuter (aranalli- crown < aran- head, puriyalli - lip
( pri-) cover,
8
akuwalli- eye (kuwa) cover, zeriyalli- cup-stand < zeri- cup).
Derived adjectives with this sux are: annalla- motherly, maternal; annalla li- for-
mer, previous,
9
attalla- fatherly, paternal.
2.22. The sux -an- forms neuter actionlresult nouns from verbal bases: enkan
death (< enk- to allot), naan fear (< na- to fear), mudan garbage (< mudi-
to remove).
2.23. The roles of the sux -ant-
10
are various and complex (HE 48; Melchert
2000: 5861). We may distinguish the following attested functions of -ant- as a deri-
vational sux in Hittite.
11
First, it is used to form participles (see 25.3925.45,
7. Since both stems pala- and `pali- (sg. nom. xsxi-i, derived verb xsxi-iya-) exist for the
base noun, both `palala- and `paliyala- (CHD P) are possible. The Hittite noun is formed on the pattern
of i.xsxi = Akk. a arrni(m) caravan leader (OAss, CAD 113; AHw 327a sub mng. 4).
8. The hapax puriyal- lip cover is either a scribal error for *puriyalli or a nonce creation based on
puriyalli- due to the supercial inuence of real nouns in -al. See parul- alongside more common parulli-
bread fragment due to nouns in -ul such as iiul.
9. See 4.10 (pp. 8384), with n. 40, and 4.17 (pp. 8687).
10. Descriptively, in Hittite we can speak of a single -ant- sux (hence, the sux). Historically, we
posit distinct origins for several of the functions. See, for example, Melchert 2000: 58 n. 20 on the dierent
PIE sources for the Hittite -ant- sux in its participial and possessive uses.
11. In drawing these distinctions we make no claim regarding the prehistoric source(s) of the -ant- suf-
x. We analyze forms of neuter noun stems ending in -anza and -ante not as derivational but as examples
of the ergative case. See 3.83.9 (pp. 6667) and 3.21 (p. 72).
Noun and Adjective Formation 2.23
00-HittiteGrammar.indb 55 6/5/08 4:15:03 PM
56
pp. 339340; on the -nt- participle in PIE see Szemerenyi 1996: 31718; Meier-Brgger
2000: 173; 2003: 185).
2.24. The sux -ant- also forms possessive denominal adjectives (HE 48 b 1):
12

perunant- rocky (< peruna- rock),
(NA
4
)
akuwant- (roads) covered with shells or
pebbles (
NA
4
aku-), natant- provided with a drinking tube (nata-),
TUG
kurenant- wear-
ing a (womans) headdress (
TUG
kurear), irmanant- having sickness (irman), ill. The
noun utneyant- population, inhabitants could also be a possessive denominal adjective
in origin (those) possessing the land (utne-). The mountain name
UR.SAG
Tarlipanta
may indicate a mountain on which the berries grew from which they made the dark red
tarlipa-drink.
2.25. The sux -ant- also has a delimiting or individualizing function (Melchert
2000: 5861, 6869). When it is added to names of the seasons, the reference is to a
particular instance (Goetze 1951: 46970): ameant- the (nextlfollowing) spring vs.
amea- spring-time (in general).
2.26. In a number of adjective and noun pairs, however, there is no longer any dis-
cernible dierence in meaning between the base stem and the extended stem in -ant-:
au- and au(w)ant- good, pleasant, ikuna- and ikunant- cold, irmala- and irma-
lant- ill, uppi- and uppi( y)ant- holy sacred, dapi- and dapiyant- entire, ankunni-
and ankunniyant- priest, gaena- and gaenant- in-law, ua- and uant- grand-
father. In the above cases, both short and extended stems of the adjectives and nouns
are still attested in Hittite. There are other examples in which only the extended stem
in -ant- has been preserved, but the shorter stem must be presumed as a base of other
derived forms: see 2.2 (p. 51).
2.27. A few Hittite stems in -(a)nt- reect the addition of a -t- to a stem in -(a)n-
(Oettinger 1982a; 2001b). Here belong such examples as ipant- night (see Avestan
xapan-) and kanint- thirst (see kanin-ant- thirsty with the possessive sux dis-
cussed in 2.24).
2.28. The sux -aa/i- forms denominal adjectives denoting appurtenance which
may themselves become nouns:
d
ilai- (demon) of the courtyard (ila-),
d
Wadulai-
(demon) concerned with sin (watul-),
d
Itamanaa- (deity) of the ear,
d
akuwaa-
(deity) of the eye,
URU
Taruntaa- (city) of (the god) Tarunta, etc.
2.29. The sux -(a)a- usually forms common gender actionlresult nouns from
verbs (EHS 167): unuwaa- adornment < unuwi- to adorn, arnamniyaa- re-
volt < arnamniya- to revolt, tariyaa- fatigue < ta(r)riya- to exert oneself,
12. See, correctly, Sturtevant and Hahn (1951: 119). Not all such cases reect participles of interven-
ing but un attested denominal verbs (against Goetze [1930b: 3435] and others). See on this point Oettinger
1981. Some examples naturally allow for both possibilities: annanuzziant- wearing a halter may be de-
rived directly from
KU
annanuzzi- halter or via an unattested verb *annanuzziya- to t with a halter, and
TUG
kurenant- wearing a womans headdress either from
TUG
kurear or *kureni- to t with a k .
2.24 Noun and Adjective Formation
00-HittiteGrammar.indb 56 6/5/08 4:15:04 PM
57
nuntariyaa- haste < nuntariya- to hasten, armuwalaa- moonlight < armuwali-
to shine (of the moon), kariyaa- compliance, pity < kariya- to comply, accede,
maliyaa- agreement, concurrence, approval (< CLuw mali-lmali- to think; see
Hitt. mali- to approve). The status of the apparent denominal example luliyaa-
swampland, moor < luli(ya)- pond and others with no attested base forms, such as
apparnuwaa- radiance, halo, karitaa- grass(?), and marruwaa- (a stone), is
unclear.
13
2.30. The sux -ati- forms common gender nouns to adjectival bases: dalugati-
length < daluki- long, pargati- height < parku- high, palati- width < pali-
wide. See also lu(m)pati- pain, grief without known base. The stem vowels of i- and
u-stem adjectives are deleted before this sux (see 2.2, p. 51).
2.31. The sux -att-
14
forms common gender actionlresult nouns from verbal and
adjectival bases: aniyatt- performance, equipment, tuiyatt- groaninglscreaming in
pain, kardimiyatt- anger, karuiliyatt- previous state, naaratt- fear, dukaratt-
amusement, entertainment. For archaic iwatt- day, without attested base, see p. 121,
n. 206.
2.32. The sux -(a)t(t)alla- (see 2.55, p. 62) forms common-gender agent nouns.
Denominal examples: aliyatalla- guard, watchman, sentinel < li sector, watch,
alugatalla- messenger < aluga- message,
LU
kinartalla- lyre-player < *kinar lyre
(< Semitic *kinnr, Hebr. kinnr),
MUNUS
kankatitalla- woman serving kankati-soup. De-
verbal examples: peran uyatalla- helper, iamatalla- singer < iamai- to sing,
palwatalla- crier or clapper < palwi-. Many are formed on the imperfective -kel
a- stem (12.3112.33, p. 204): iiyaikatalla- spy, informant, maniyaikatalla-
governor, adminstrator, memikattalla- eloquent, verbose, pikattalla- sparing
(one), ukikattalla- watchman, weekattalla- sentry, weurikatalla- strangler,
wiwikatalla- wailer. Of uncertain derivation: aranatalla- envier (see arane- to
envy 10.11, p. 177), ukmatalla- enchanter (see ukmai- spell),
LU
kuanatalla-,
kuaniyatalla- hired laborer, wage earner (see kuan wage).
2.33. The sux -tar forms neuter abstractlaction nouns from verbs, adjectives,
and nouns: (1) from verbs: aggtar death, akuwatar drinking, apptar seizing,
*armaatar conception, pregnancy, iyatar love, adtar eating, atarl
13. Since all clear examples are based on stems ending in -a-, one cannot tell whether the productive
sux was -a- or -aa-. Contra Starke (1979: 258), there is no solid evidence that all examples of the
productive deverbal sux -(a)a- are Luwianisms. Nor is there yet evidence for the sux even being pro-
ductive in either form of Luwian. In terms of attested Hittite, examples such as tea- dream are underived
(see Starke 1979: 24850).
14. See 4.91 (p. 120) and 4.94 (p. 121, paradigms). On this class of action noun see Berman 1972b:
15556 and Laroche 1975. For discussion of the semantic content of this sux see Rieken 1999a: 118. On
the problematic words maniyaiyatt-, watarnaaz, upparatt-, andatt-, and lamarandatt- see Rieken
1999a: 12025.
Noun and Adjective Formation 2.33
00-HittiteGrammar.indb 57 6/5/08 4:15:04 PM
58
aatar procreation, womb, clan, igatar heap of threshed grain(?), *maniyaatar
administration (< maniya- to administer), ulltar wantonness, disrespectful-
ness; (2) from adjectives: atugtar fright, idlawatar evil, marlatar foolishness,
nakkiytar importance, dignity, paltar width; (3) from nouns: anniyatar mother-
hood, *andayandatar status of a live-in husband, antutar humanity,
15
`gaina-
tar in-law-ship, pintar manliness, manhood, xtts-tar womanliness. The stem
vowels of some adjectives in -i- (atukali- frightening 4.39 [p. 97], pali- wide)
and -u- (parku- high) are deleted before this sux (see 2.2, p. 51).
2.34. The sux -ear forms neuter action nouns and abstracts from verbs and
adjectives: (deverbal) aear seated group, annear judgment, attear hole
(cut), atrear sending; message,
(TUG)
ikallear torn garment,
TUG
kurear wom-
ans headdress (< *cut piece of cloth < kuer- to cut), mukear prayer, pattear
hole (dug), iear beer (< ai- to press), taruppear total, tetear thunder-
ing, tiear orchard (*a planting of trees < dai- to put), uppear something sent
(uppa-), gift; (de-adjectival) alwanzear sorcery, palear width, paprear im-
purity, parkear height, uppear holiness, purity. The stem vowels of adjectives
in -i- and -u- (uppi- holy, pure, parku- high) are deleted before this sux (see 2.2,
p. 51). Also a few collectives from nouns: e.g., lalakuear swarm of ants. For the
inection see 4.108 (p. 128).
2.35. The sux -iya- forms adjectives from nouns and adverbs and probably forms
numeric adjectives: ipantiya- nocturnal < ipant- night, itarniya- middle < itarna
in the midst of, anduriya- internal, interior < *andur- (see andurza inside), arazeya-
outer, external, antiya- special, separate, nuntariya- swift, swiftly rising, *3-ya-
(*teriya-) threefold and `4-ya- (*mewiya-?) fourfold in 3-ya- to make threefold
and 4-ya- to make fourfold, iptamiya- sevenfold(?) (a beverage name based on
the numeral seven. For an alternate explanation of 4-ya- see 9.59 (p. 170).
2.36. The sux -ili- forms adjectives from various bases: karuili- previous, former
< kar formerly, annapili- empty < annapi (adverb of disputed meaning), taruili-
victorious, mighty. For the adverbial sux -ili see 19.15 (p. 292).
2.37. The sux -ima-
16
forms common-gender action nouns from verbs and ad-
jectives: ekunima- cold < elikuna- cold, kurkurima- intimidation < kurkuriya- to
intimidate, katkattima- trembling < katkattiya- to tremble, lalaima- agitation,
worry < lalaiya- to be agitated, lalukkima- brightness < lalukke- be(come)
bright, takupima- screaming < takupi- to scream, tetima- thundering < teta-
to thunder, tuima- groaning, agony, pain < tui- to groan, be in pain, wante-
15. The Hittite word may have originally meant status of a human being, but its use in historical Hittite
is always as a collective: mankind, humankind.
16. Initially analyzed by Laroche (1956). On the semantics and history of this sux see Oettinger
2001a. Sometimes written with geminate m.
2.34 Noun and Adjective Formation
00-HittiteGrammar.indb 58 6/5/08 4:15:05 PM
59
wantema- lightning, wawarkima- door-pivot, hinge, weritema- anxiety < werite- to
be anxious.
2.38. The sux -ri- forms common-gender and neuter nouns to verbal bases (some
not directly attested). Common: auri- observation post < au ( )- to see, look,
SG
kiri-
? < ki- to comb (wool). Neuter: tri- food < et- to eat. Indeterminate: *miri-
brilliance (< *mi- twinkle) in miriwant- brilliant and miriwtar brilliance.
2.39. A linguistic relic which, when added to nouns denoting human (or divine)
males, marks the corresponding female is the derivational sux -()ara-. This sux
distinguishes the pairs au king and auara queen, ia lord and iaara
lady, uppi holylconsecrated man and uppeara virgin (see HE 50a; Laroche
1966: 3026; EHS 109, 111, 124, 129; Kammenhuber 1969b: 18990, 198, 261, 269,
297; 1993 19698), and the Hittite words underlying ln male slave and orxr fe-
male slave.
17
It operates like the English -ess in pairs such as actor : actress, waiter :
waitress, and master : mistress. Compare Luwian nan-i- brother and nan-ar-i- sis-
ter (-ar-). On this element in onomastics see Laroche 1966: 3026. The source of this
sux is probably an old word for woman seen in Luwian *ar (i)- (base of the deriva-
tives arait- femininity, arul(i )- female, feminine). The Hittite word for woman,
however, reects the PIE root *g
w
en-, as does Luwian wanati- (see Starke 1980). On
the putative Hittite kuinna- woman see Neu 1990; Carruba 1991, 1993a, questioned by
Gterbock (1992; 1995c). Puhvel (HED K 3068) agreed with Gterbock in questioning
the philological basis of Neus example of kuinna- woman but agreed with Carrubas
identication of the adjective kuwana- and kuwaa- female and reconstructed the
Hittite noun woman as ku(w)an(a)-.
2.40. For -talla- see -(a)t(t)alla- in 2.32.
2.41. The sux -tara- forms agent nouns from verbal bases: wetara- herdsman
< we(iya)- to graze, akuttara- (an ocial) < eku-laku- to drink. See 4.82 (p. 115).
2.42. The sux -ul- forms neuter nouns from verbs (for their inection see 4.66,
p. 107): takul peace, friendship < tak- to join. The a of the verb stem is deleted
before this sux: watul sin, oense < wata- to miss the mark, sin, iiul obliga-
tion, treaty < iiya- to bind, immiul mixture < immiya- to mix, pariul crumb
< pariya- to crumble, and earul sieve < eariya- to sift.
18
There is one example
from an adjective: aul goodness, favor, well being < au- good, and two from nouns:
kazzarnul (a textile) < karza(n)- (weavers tool)
19
and paurul (an implement).
17. See the derived verb orxr-aare- to become a (female) slave for evidence that the noun under-
lying orxr woman slave is derived from that underlying ln man slave. See LH 43, 139, 18586, 263.
18. This word is not an i-stem (contra Neu apud Rieken 1999a: 470 n. 2317). The correct reading of
KBo 44.142 (183ln) rev. 5 is [. . . x
NINDA
wa-ge-]e-ar 10-li, not [. . . e-]e-ar-u-li.
19. On this derivation see Rle 1998.
Noun and Adjective Formation 2.42
00-HittiteGrammar.indb 59 6/5/08 4:15:05 PM
60
2.43. The sux -ula- forms common-gender nouns from various bases: aandula-
garrisoning, pittula- loop.
2.44. The sux -ulli-
20
forms neuter instrumental and result nouns from verbs:
itappulli- lid, stopper < itapp- to stop up, kariulli- body-veil < kariya- to cover,

DUG
ariulli- (a container), kukuulli- mortar bowl in which foodstus are crushed
< kuku- to crush, uttulli- tuft, something plucked < u(i)ttiya- to pull, pluck,
parulli- bread fragment < par(iya)- to break,
21
and pupulli- ruins (from an un-
clear base). The a of the verbal stem -(i)ya- is deleted before this sux.
2.45. The sux -uman- forms ethnica based upon place-names (Kammenhuber 1969b:
12021, 124, 215, 268, 271, 295; Oettinger 1982b; Melchert 1983: 46; Kammenhuber
1993: 19495, 66061, Oettinger 2003: 147); for its inection see 4.74 (p. 111). Ex-
amples: (sg. nom.)
m
uppiluliuma (man) of
URU
uppiluliya,
URU
utummana the (man)
of uta,
URU
attuumma,
22
(sg. gen.)
LU
luwiumna of a man of Luwiya, (sg. acc.)
URU
puruandumnan the man from Puruanda, (pl. nom.)
URU
udulumni (elders) of
udul,
URU
anunum(i)ne the men of Anunuwa,
URU
alampmene the men of alampa,
URU
ktapmene the men of Ktapa,
LU.ME
neumene men of Nea. Case uncertain
from lack of context:
URU
uumna man of Ua. For kaniumnili and neumnili in the
language of KanilNea (= Hittite) see 19.15 (p. 292); and note
URU
palaumnili in the
language of the land of Pala,
URU
itanumnili in the language of Itanuwa. Note too the
ethnic pronoun kuenzumna whatever his home KBo 1.35 iii 9 (Kammenhuber 1969b:
215, 271) built on the interrogative-relative pron. kui- who, what (EHS 102, 113, 194).
Note also arunuman- maritime from aruna- sea. The Luwian ethnicon is in -wann(i)-
(Kammenhuber 1969b: 153, 268, 295, 318): see
URU
ninuwawann(i)- of Nineveh.
2.46. Personal names were often based upon ethnica (see Laroche 1957b, 1966;
Honer 1998c). These names can use either the Hittite ethnicon -uman- or the old Hat-
tic -il + Hittite theme vowel -i-. The royal name uppiluliuma certainly contains the
toponym
URU
uppiluliya + the sux -uman-, and the following personal names may con-
tain that sux: uppiuman, Artumanna, Zardumanni, Piriyaauma, Nauma. The fol-
lowing contain -ili: attuili (from attua), Arinnili (from Arinna), Nerikaili (from
Nerik), Gagaili (from the Kaka people), upenaili (from upina), Katapaili
(from Katapa), Taurwaili (from Taurpa), anikkuili (from Ankuwa), imuili
(from emuwa), Karanuili (from Karana).
2.47. This same -uman- sux is preserved in factitive verbs: tameumma- to
make dierent (*tameuman- belonging to another < tamai-), and umuma- (for
`umumma-) to make (several objects) one (*umuman-), unify (see Rieken 2000b
and 9.62, p. 171). These factitive verbs are derived from the zero-grade form of the
20. Some examples may be collective plurals in -i of noun stems in -ul-.
21. Also attested is an l-stem parul-. CHD P 192 s.v.
22. See
URU
a-at-tu-[]um-ma-a KBo 7.14 rev. 3 (OS).
2.43 Noun and Adjective Formation
00-HittiteGrammar.indb 60 6/5/08 4:15:06 PM
61
sux *-umn- with assimilation to -umm- (see 1.122, p. 44), whose geminate is not
consistently written.
2.48. The sux -ur- forms neuter nouns from verbs: aniur- ritual < aniya- to per-
form, engur- gift < enk- to (pr)oer. Compare also with unclear bases: kudur- (a
body part), kurur- hostility; for the inection see 4.80 (p. 114). The a of the verbal
stem -(i)ya- is deleted before this sux.
2.49. The sux -uzzi- forms common-gender and neuter instrumental nouns from
verbs: kuruzzi- tool for cutting < kuer- to cut, ipanduzzi- vessel for libating < ipand-
to libate. The a of the verbal stem -(i)ya- is deleted before this sux (iuzzi- belt,
sash < iiya- to bind).
2.50. The sux -want-
23
forms possessive adjectives from nouns: amankurwant-
having a beard, dukarawant- having joy, pittuliyawant- having anxiety, kituwant-
having hunger, kanir(u)want- and kane(u)want- thirsty, innarawant- having
vigor, lalukkiwant- having brightness, miriwant- having brilliance, earwant-
having bloodstains, lupannawant- wearing a diadem, kurutawant- wearing a horned
crown(?), upigawant- wearing a veil, naduwant- (terrain) having reeds,
URU
Wiya na-
wanda (city) having vines. One also nds some deverbal examples: armauwant-
pregnant, kartimmiyawant- having anger, angry, naariyawant- having fear, afraid
(see on their formation Oettinger 1988). In view of the last two examples in -iya-, it is
possible that the forms tar (a)ganiyawant- covered with colostrum(?) (
GA
tar (a)gan,
see Honer 2005) and euganiyawant- covered with meconium(?) (*eugan-) are
built to denominative verbs *tar(a)ganiya- and *euganiya-.
2.51. The sux -(u)war forms the verbal substantive of most verbs (see 11.20,
p. 185).
2.52. The sux -zil- forms result nouns from verbs: arnikzil- compensation,
tayazzil- theft.
24
2.53. The sux -zzi(ya)- forms adjectives from local adverbs: antezzi(ya)- former,
rst, appezzi(ya)- latter, later, last, arazzi(ya)- higher, uppermost (< *-tyo- 1.90,
p. 37; paradigms in 4.10, p. 84, and 4.38, p. 96).
2.54. Nouns (frequently divine names) ending in -epa- or -zipa- (Laroche 1947:
6768) may contain a derivational sux or be true compound nouns containing a word
elipa- spirit(?) that is not directly attested.
25
For -(n)zipa- alongside -epa- see 1.136
23. See Goetze 1930b: 3435; Sturtevant and Hahn 1951: 119; HE 49d; Kammenhuber 1969b: 188
89, 294; Oettinger 1988; Sihler 1995: 346 (< the hysterokinetic sux -went- in PIE).
24. The pl. acc. [a]r-ni-ik-zi-lu-u KUB 46.42 iv 6, KUB 46.38 i 7, 10 (NH) indicates that at least
arnikzil- had common-gender forms as well as neuter.
25. The word pa-lipa- once suspected of being this word (see Laroche 1947: 6768 and AlHeth 84
n. 250) seems to denote a disease of the eyes or an impairment of vision.
Noun and Adjective Formation 2.54
00-HittiteGrammar.indb 61 6/5/08 4:15:07 PM
62
(p. 47). As a sux (or in a compound) it occurs in daganzipa- earth (< tekan, dagan
earth), taranzipa- platform,
d
antaepa- (< ant- forehead),
d
ilanzipa- (< ila-
court),
d
Ipanzaepa- (< ipant- night),
d
Miyatanzipa- (< miytar fruitfulness [rln-
stem]),
(d)
akaepa- (< aka- gate).
2.55. Some of the derivational suxes listed above are either borrowed from Lu-
wian or created within Hittite by modication of Luwian material. Direct borrowings:
(1) abstracts in -ait- (Starke 1990: 15376) and (2) participles in -m(m)i-. Modica-
tions: (3) nouns in -alla- and -alli- from the Luwian adjectival sux -allali- (Melchert
2005b) and (4) agent nouns in -(a)t(t)alla- from a reanalysis of the same Luwian sux
-allali- (Oettinger 1986a). The new sux -(a)t(t)alla- spreads at the expense of the na-
tive sux -ala-. In still other cases, a native Hittite sux is reinforced by borrowings
from Luwian: (5) nouns in -al-,
26
and (6) genitival adjectives in -aali-. In these last
two types, it is often dicult to decide whether a particular example involves borrow-
ing or not.
Reduplicated Nouns and Adjectives
2.56. Hittite and Luwian vocabulary contains a considerable number of reduplicated
nouns, verbs, and adjectives (van Brock 1964; EHS 11922; and Kammenhuber 1969b:
184, 17,3). For reduplicated verbal stems see 10.3 (p. 173). Reduplication in both
verbs and nouns is a feature of PIE (Beekes 1995: 171; Sihler 1995: 487 443), but the
process is also highly productive cross-linguistically. None of the examples below can
be shown to be inherited from PIE. Some are probably formed according to inherited
patterns, while others are wholly new.
2.57. Among the nouns and adjectives: uwaurti- throat, titita- nose, tetana-
hair, laplipa- eyelash (see Luwian lalpi-), gagali- tooth, pa(p)paalali- esopha-
gus, ia(i)- tail, duddumi- deaf; deaf person, memal- meal, alaltumar (i)- cor-
ner, ariari- stormy weather, akuwakuwa- (a small animal, possibly onomatopoetic
for the animals noise, but denied in HED A 26, citing lalakuea- ant), aluwaaluwa-
and kallikalli- (bird names, imitating their repeated call?), artarti- (a plant name),
wantewantema- lightning,
GI
karkaralli-, iggaigga-, and oi.orn
4
tariyattariyan.
Along with the many names for plants and animals, note the common occurrence of
reduplicated nouns for anatomical terms.
2.58. A special kind of reduplication is what in German is called Lallwrter (EHS
11719), words derived from baby talk (compare in English Mama, Daddy, bye bye
and terms for bodily discharges such as pee pee and poo poo). It is likely that in Hittite
such words as atta- father (Luwian ttali-), anna- mother, ua- grandfather, and
annaanna- grandmother belong to this class (see also pappa- father in the com-
26. For this type see Starke 1990: 300342 and Rieken 1999a: 44558, though Starke wrongly denies
and Rieken underestimates the number of native Hittite examples.
2.55 Noun and Adjective Formation
00-HittiteGrammar.indb 62 6/5/08 4:15:07 PM
63
pound pappanekne brothers having the same father). And since words for body parts
are among the rst acquired by young children, it is completely possible that several of
those listed in the preceding paragraph owe their form to baby talk.
Compounded Nouns and Adjectives
2.59. Compounded forms also exist (Gterbock 1955; HE 53 a; Honer 1966;
Kammenhuber 1969b: 18384; Neu 1986): dyuga- two-year-old (d two+ yuga-,
9.3, p. 153), appaiwatt- future (ppa after + iwatt- day), allakarttar negli-
gence (alla high + kart- heart + -tar), zagarai anus (zakar excrement + ai
mouth), pattarpali- (a bird name: broad of wing, pattar + pali-), annaneke sis-
ters having the same mother (anna- mother + nega- sister), pappanekne brothers
having the same father ( pappa father + negna- brother), tuzziyaear camplsettle-
ment of an army (tuzzi); tnt-riyaear settlement of a city
27
(second component
aear settlement), and possibly trumaki-
MUEN
woodpecker (tru- woodltree
+ wak(a)i- biter (analysis by Honer 1966),
28
with uw > um across word boundary
[1.126, p. 44]). The negation nman not want(ing) to (Honer 1982) seems to con-
sist of a negative element (n?) added to the optative man (see CHD sub both man and
nman; see 26.19, p. 344).
2.60. There are also forms that are better described as univerbations. The rst type
consists of nouns with a modifying genitive:
LU
maniyaiya ia-,
LU
mukena r-a,
LU.ME
azziwa ie,
NINDA
aramna li, etc. (Neu 1986, esp. his table on p. 116). The
seond type comprises nouns derived from a preverb + verb combination (HE 53 b):
(LU)
peran uyatalla- he who runs before, helper, par andandtar divine power,
providence, fate, luck,
LU
antiyant- son-in-law who moves in with his brides family
(< anda iya-), kattakurant- ( jug) cut o beneath, erkurant- (loaf) cut o above
(Honer 1966). The third type is formed by the combination attributive adjective + head
noun:
UZU
parkui atai (cut of meat called) pure bone,
UZU
dnati (cut of meat called)
second bone. That all of these combinations were treated as single words is shown by
the fact that, when they are marked by determinatives (1.391.44, pp. 2324), the
latter appear on the preceding genitive or preverb rather than on the head noun.
27. Friedrich (HW
1
[1952] 299), who was well aware of the syllabic reading appira- for tnt, never-
theless listed this word under the tnt Sumerogram because he was uncertain of its syllabic Hittite reading.
Kammenhuber (1954: 406) and Kronasser (EHS 291) considered it to stand for *appiriyaear and took
appiri as d.-l. of the normal a-stem appira- city. Neu (1974: 1067) and Oettinger (1976a: 46) consid-
ered it to be based not on a d.-l. but on an alternate stem appiri(ya)-. Although Neu and Oettinger may
be right about appiri(ya)- city, the interpretation of tnt-riyaear as *appiriyaear has now been
placed in doubt by KBo 27.12 iii 4, a duplicate to KBo 6.34 iii 29, which instead of tnt-ya-e-e-ar reads
[. . .-]x-u-ri-ya-e-e-ar! This suggests that underlying tnt in this word was a word [. . .-]uri(ya)-, not
*appiri(ya)-.
28. For wagai- with wakin (trumaki[n
MUEN
]) in the sg. acc., compare the derived action nouns
ullanzai- and zaai- with sg. acc. in ullanzin and zain. See 4.32 with n. 66 (p. 92).
Noun and Adjective Formation 2.60
00-HittiteGrammar.indb 63 6/5/08 4:15:09 PM
64
Chapter 3
NOUN AND ADJECTIVE INFLECTION
3.1. Hittite is an inecting language, marking the syntactic role of verbs and of
nominal categories (such as noun, adjective,
1
pronoun,
2
and some declinable cardinal
numbers) by word-nal suxes (endings). The inectional categories of nouns, ad-
jectives, and some pronouns are gender, case, and number.
3.2. Gender. Grammatical gender is evident (and pertinent) only in the nominative
and accusative cases. Hittite recognizes two grammatical gender classes, traditionally
called common and neuter, alternatively animate and inanimate. Characteristics of the
Hittite neuter or inanimate nouns are the same features that characterize neuter nouns
in other old IE languages: identical forms in the nominative and accusative cases, and a
zero ending in the singular (except in a-stems, where the ending is -(a)n).
3.3. Hittite nouns have no formal feminine gender (EHS 1067). It remains a matter
of debate whether this reects a prehistoric merger of inherited masculine and feminine
into a common (or animate) gender (so Kammenhuber 1969b: 253; Hararson 1994:
3239, and others) or an archaic system in which there was no feminine distinct from
the masculine (so, e.g., Neu 1969: 23741). Likewise controversial is the suggestion
that adjectives of the type dankui- dark, parkui- pure, are formal relics of the PIE
feminine that have lost their original connotation as feminines (Pedersen 1938: 3536;
see also Kronasser 1966: 107; Kammenhuber 1969b: 253; Oettinger 1987; and Starke
1990: 8587). In respect to the lack of a masculine-feminine contrast, the Hittite declen-
sion of pronouns and substantives resembles that of nouns such as Latin flis cat and
canis dog, which can be masculine or feminine according to the biological gender of
the animal referent (Sihler 1995: 244). See 2.39 (p. 59) for the sux used by Hittites
to designate females (-()ara-).
3.4. As in other IE languages, there is a broad correlation in Hittite between the
grammatical gender of a noun and the animacy of its referent: nouns denoting living
beingsdivine, human, or animalare usually common or animate. But as in other
systems, this correlation is an imperfect one, and numerous exceptions exist.
3
Nouns
1. Including the participles as a verbal adjective.
2. There is a special set of case endings for the pronouns, which show dierent forms for the gen., abl.,
ins., pl. nom. com., and neut. (see 5.4, p. 133).
3. Many common-gender nouns denote non-living things (aimpa- burden, aggala- furrow, alpa-
cloud, aruna- sea, peruna- rock, ega- ice, aluga- message) because they belong to the a-stem class,
and only common-gender a-stems are productive in Hittite (see 2.42.6, p. 52).
00-HittiteGrammar.indb 64 6/5/08 4:15:09 PM
65
referring to inanimate objects may be common or neuter gender. The determining fac-
tor is formal, not semantic: certain suxes or inectional types belong to a given gen-
der, regardless of the semantic eld to which the referent belongs (see 2.152.55,
pp. 5362, for these suxes).
4
For example, many body parts are common gender, not
because they are an inalienable part of the human being and therefore could be argued to
be inherently animate, but because they belong to the a-stem class, and only common
gender a-stems are productive in Hittite (see 2.42.8 [p. 52], 4.14.5 [pp. 7982]).
Hittite also preserves the inherited neuter u-stem genu- knee and the neuter root noun
kerlkard- heart, while lii- liver and ararlaran- head are also neuter (all these
belonging to well-attested neuter inectional types in Hittite). This formal factor over-
rides the general tendency of nouns referring to humans and animals to be common
gender. Of nouns referring to groups of humans, rnr.xr-t- troops is common gender,
because productive t-stems are common gender,
5
while antutar people, humanity,
atar family, aear assembly, appiriyaear population of a city, and
MUNUS.
ME
azkara(i)- (group of female musicians) are all neuter, because the suxes -tar and
-ear form neuter nouns, as does -a(i)- when forming collectives.
6
Of nouns referring
to groups of animals, uitar wild animals and lala(k)uear swarm of ants are neces-
sarily neuter because of their stems, although their referents are animate.
3.5. The essentially formal basis of grammatical gender described in 3.3 did not
prevent speakers from occasionally employing common gender to mark true animatiza-
tion, based on the undeniable widespread correlation of common gender with animacy
of the referent. For example, it can hardly be accidental that genu- knee, a neuter noun,
is inected as common gender only in the phrase the walking knee (sg. acc. iyantan
genun KBo 40.25 i 24)
7
, where it is being depicted as an active body part. Similarly,
in a ritual passage where a house is to be transformed into a ram, the neuter noun per
appears as a secondary common-gender sg. nom. parna in one version (KBo 10.45 iv
31), in contrast to the ergative parnanza of the duplicate (KUB 41.8 iv 30 ).
8
4. Nouns in -ait-, -al-, -tar, -ear, etc., are formally neuter or inanimate. Nouns in -ala-, -aa-, -att-,
-ima-, etc., are formally common or animate.
5. The noun for troops is unknown. One possibility might be tuzziyant-, as claimed by Tischler (2001:
222) and others. For although the sg. nom. rnr.xr-az in KBo 17.1 i 30 (OS) would be an unusual spelling
for l-antsl, it is not unparalleled (see 1.135, p. 46, and 3.21, p. 72): li-in-ki-ya-az KUB 30.34 iv 7 (MHl
NS) is an ergative (Laroche 1962: 26, 32), and
MUNUS.ME
azkaraiyaza l azkaranza are alternate spellings of
the ergative of azgarai- (see Honer 1998a: 3740 and 3.9 below [p. 67]).
6. The same principle applies to nouns referring to plants and to all kinds of manufactured items, such
as containers, buildings or parts of buildings, wooden and metal tools, and other semantic elds. For con-
venient lists see Tischler 2001 under the respective determinatives uto, r, or, tntut, and so on. Compare
the well-known example of the German diminutive suxes -chen and -lein. Since these form neuter nouns,
even diminutives referring to people are grammatically neuter: Mdchen girl, Frulein young woman.
7. See the scribal confusion about the same construction in i-ya-an-ni-ya-an (neut.) ge-nu-un (com.)
KUB 9.34 iii 37, ed. Hutter 1988: 3839, 8283.
8. This is no more remarkable or signicant for the overall system than the fact that grammatical gender
aects poetic imagery in modern IE languages (e.g., the moon is typically depicted as a female in Romance
Noun and Adjective Inection 3.5
00-HittiteGrammar.indb 65 6/5/08 4:15:10 PM
66
3.6. Since the gender opposition in Hittite contrasts only animate and inanimate,
there is no formal distinction between masculine and feminine gender, only a deriva-
tional sux -()ara- (2.39, p. 59) used to form nouns denoting human (or divine)
females from corresponding male-denoting nouns.
3.7. For two reasons it is necessary to insist on the unremarkable facts cited in 3.3
3.5 regarding the relationship of grammatical gender and semantic animacy in Hittite.
First, misconceptions about the distinction have prevented a proper appreciation of the
role of the ergative case (see 3.83.9 below). Second, Starke (1977: 12226) has in-
correctly claimed that in OH there was a grammatical contrast in the behavior of nouns
referring to persons and animals (Personenklasse) and those referring to inanimate ob-
jects (Sachklasse).
9
Specically: (1) case forms in -i of nouns referring to living beings
are allegedly attested in OH only in functions belonging to the dative (recipient, goal,
and one from whom something is removed), while case forms in -i of nouns referring to
inanimate objects occur only in the function of a locative, expressing location (goal for
these nouns being marked in OH only by the allative in -a, and place from which only
by the ablative); (2) nouns referring to living beings (Personenklasse) never occur in
OH in the allative, ablative, or instrumental. If the word attaz in KBo 20.31:1 (OS) is
a form of father, as the context suggests, it furnishes direct counterevidence for the
second claim.
10
For further evidence refuting both (1) and (2) see the examples cited
in the respective paragraphs on the case syntax of the dative-locative (16.65, p. 257;
16.68, p. 258; and 16.70, p. 259) and the instrumental (16.107, p. 269).
3.8. Split Ergativity. Case syntax in Hittite for the most part operates according
to an accusative system typical for an IE language: subjects of both transitive and in-
transitive verbs appear in the nominative, while direct objects of transitive verbs appear
in the accusative. However, as established by Laroche (1962), when a neuter noun func-
tions as the subject of a transitive verb, it obligatorily takes the form -anza (sg.) or -ante
(pl.) (see further 3.21, p. 72). He properly labeled this form an ergative, although
his own characterization of its behavior left ample room for doubt. Garrett (1990b) has
since demonstrated that Hittite nominal syntax shows precisely the features expected
of a language with a split ergative system.
11
Pronouns and common-gender nouns
follow the accusative pattern: as subjects they appear in the nominative (whether
languages, where the word is feminine gender, but as a male in German, where it is masculine gender).
9. Unfortunately, this analysis has been uncritically accepted by a number of scholars (e.g., Luraghi
1997a: 8 2.1.3, 13 2.1.5.910; and Tjerkstra 1999: 7 and 100 n. 22).
10. In KBo 20.31 and its duplicate KUB 57.69 (ed. Honer 1992a) the word attaz is followed im-
mediately by u-u-[a'-. . .], and the following context refers to utxt.xr sons, +ntx+ and negnan
brother, and x mother.
11. Split ergativity refers to a situation in which a language uses both accusative and ergative patterns
of case syntax side by side, the distribution of which can be conditioned by a variety of grammatical factors.
We stress that we insist here only on the validity of the split ergative system for attested Hittite. We take no
stand on its historical origin (for an alternative to the view presented by Garrett (1990b), compare Oettinger
2001b: 31112, with refs., and Josephson 2003a and 2004). Furthermore, the status of ergativity in PIE is
3.6 Noun and Adjective Inection
00-HittiteGrammar.indb 66 6/5/08 4:15:12 PM
67
the verb is transitive or intransitive) and as direct objects in the accusative. But neuter
nouns follow an ergative pattern: they appear in the absolutive case as the subjects
of intransitive verbs and objects of transitive verbs, but in the ergative as the subjects
of transitive verbs.
12
3.9. Rejection of an ergative case in Hittite has been based on two invalid objections
(for both see, e.g., Starke 1977: 182 n. 24):
13
(1) Hittite does not show the features of a
typical ergative language such as Hurrian; (2) -anzal-ante belong to a quasi-derivational
-ant-sux that semantically animatizes neuter nouns referring to inanimates (when
they are functioning in what is loosely labeled an active sense). The rst of these objec-
tions is irrelevant: since Hittite shows split ergativity, where the ergative pattern applies
only to a subset of the nominal system, we would not expect it to show the features of
a fully ergative language such as Hurrian. The second claim is simply false. As stressed
by Honer (1998a: 3740), the noun
MUNUS.ME
azkara(i)- (female musicians) is semanti-
cally animate, referring only to persons, but as a collective it happens to be grammati-
cally neuter. The fact that the word appears as
MUNUS.ME
azkaraiyaza l azkaranza just
when it is the subject of a transitive verb conrms that the ending -anza marks ergative
case. It is illogical to say that -anza here animatizes a noun that is already semantically
animate. Nor is the notion of being active the determining factor. Grammatically neu-
ter nouns referring to persons or things appear in the nominative-accusative (properly
absolutive) when they are the subjects of intransitive action verbs: aear ar tiyazi
the assembly stands up, wtar arazi water ows. These and other facts conrm that
the selection of the endings -anza and -ante is grammatically conditioned (by neuter
gender and subjecthood of transitive verbs) and has nothing to do with semantic ani-
macy. Ergative case is the only proper characterization for such a feature.
3.10. Just as formally dened grammatical gender occasionally is employed to mark
genuine animacy of a usually inanimate referent (see 3.5), so, unsurprisingly, the Hit-
tites in a few instances also used the available ergative marker for the same purpose,
extending its usage in such cases beyond its proper grammatical sphere. For example, in
a ritual context where a house is likened rst to a ram and then to a ewe, it appears in one
version (KUB 41.8 iv 30, 34 ) as parnanza as subject of both an intransitive and transi-
tive verb. That the ergative may occasionally function simultaneously as appositional
direct address without abandoning its ergative role, as in nu atta nepianza rorn-an
tarna O father heaven, turn them back! is also unremarkable.
3.11. Case. OH distinguishes as many as eight cases of the noun and adjective
in the singular and as many as six in the plural (Forrer 1922; HE 5456). They are:
an entirely separate issue, on which see, among others, Villar 1983; Luraghi 1987; Rumsey 1987; Comrie
1998; and Oettinger 2001b.
12. For the sake of simplicity, we elsewhere retain the traditional term nominative-accusative for the
case that should properly be termed absolutive.
13. Others who reject the notion of ergativity in Hittite include Benveniste 1962b; Kammenhuber 1985:
45255; Neu 1989b; Carruba 1992; and Marazzi 1996: 16162.
Noun and Adjective Inection 3.11
00-HittiteGrammar.indb 67 6/5/08 4:15:13 PM
68
(1) vocative, (2) nominative, (3) accusative, (4) genitive, (5) allative, (6) dative-locative,
(7) ablative, and (8) instrumental.
14
A ninth case, the ergative (3.83.9 and 3.21),
occurs only with neuter nouns. Dative and locative singular have already merged into
a common dative-locative form in the singular in OH. The plural in OH contrasts
nominative, accusative, genitive, and dative-locative, plus ablative and instrumental (but
the last two cases show no contrast between singular and plural). The allative and in-
strumental, productive in OH and MH, survive in NH only in isolated expressions. The
common gender nominative and accusative plural also merge in NH. Hittite nominal
cases function like those in other case languages: the nominative marks the subject, the
accusative the direct object, the genitive possession, the dative-locative the indirect ob-
ject or place inlto which, the ablative place from which, and the instrumental means or
accompaniment. The OH allative marks only place to which. For the uses of the cases
in detail see chapter 16.
3.12. Number. PIE substantives distinguished singular, dual, and plural (Szeme-
renyi 1996: 7.1.3). Animate nouns contrasted a count or distributive plural with
a collective plural (Eichner 1985). The Old Anatolian languages such as Hittite show
only indirect traces of the dual. These scattered examples are treated synchronically as
collective plurals.
15
Thus, in the declension of substantives and pronouns (for the lat-
ter, see chapters 58) Hittite distinguishes only singular and plural number. Inectional
endings of the noun, adjective, and clitic pronoun distinguish singular from plural in
OH in all cases but the instrumental (ending -(i)t) and ablative (ending -(a)z). The latter
two cases are indierent to number throughout the history of the language. In post-OH
the genitive is identical in singular and plural.
3.13. OH also preserves the PIE contrast between a count (or distributive) plural
and collective plural in common-gender nouns (Neu 1969; Eichner 1985; Neu 1992;
Melchert 2000). These collective forms have been previously regarded incorrectly as
neuters and their occurrence on substantives of common gender as evidence for hetero-
genericity in Hittite nouns (for heterogenericity see 15.14, p. 240). These examples do
not show alternation in gender but rather a contrast in number: alpa bank of clouds vs.
alpelalpu (individual) clouds, gul-a-a fate (of a person) vs.
d
GulelGulu the
Fates (goddesses). See Melchert 2000: 6264 for further instances. There was origi-
nally no such contrast for neuter nouns, which had only the collective plural (Eichner
1985). Hittite did develop several devices for providing a count plural for neuter nouns.
One such device was to use the only appropriate endings it hadnamely, those of the
14. Since neuter nouns do not distinguish nominative and accusative, their appearance in the ergative
case as subjects of transitive verbs still results in only an eight-way contrast within the paradigm of neuter
nouns.
15. For literature on the dual see Szemerenyi 1996: 7.13. For traces of the dual in Hittite and Luwian
see among others Rieken 1994: 5253, with references (citing da(n)ati- double-bone, imeri- reins and
elzi- scales). Of the terms for body parts occurring in pairs (eyes, ears, nostrils, etc.), only kuwa (two)
eyes has a possibility of ultimate derivation from an old dual.
3.12 Noun and Adjective Inection
00-HittiteGrammar.indb 68 6/5/08 4:15:15 PM
69
common-gender nouns: e.g., luttae and luttau windows for neuter luttai- window.
For the use of a sux -ant- to show a count plural of collectives see 9.26 (p. 159). The
contrast between collective and count plural appears to have been lost by NH, where one
nds only singular and plural, with the plural of neuter nouns marked by the endings of
the old collective. For the sake of simplicity, the collective endings are for the most part
referred to below as nominative-accusative plural neuter except in those cases where
they clearly belong to common-gender nouns.
Infectional Endings
The Basic Scheme
3.14. The following paradigm shows the basic scheme of noun and adjective endings
in Hittite. Extremely rare alternative endings have been omitted, as have all except the
most straightforward distributional restrictions. For these see the detailed paragraphs
which follow the paradigm.
16 17 18 19 20 21 22 23 24

Old Hittite New Hittite
Case
sg. pl. sg. pl.
nom. com. -, -
16
-e
17
- -u, -e, -a
18
acc. com. -n,
-an (cons. stems)
-u -n,
-an (cons. stems)
-u, -e, -a
18
n.-a. neut. -, -n (a-stems) -,
19
-a,
20, 21
-, -n -, -a,
21
-i
21, 22
erg. -anza -ante -anza -ante
gen. -a
23
-an, -a -a -a
d.-l. -i, -ya (i-stems),
24

- (cons. stems)
-a -i, -e, -ya (i-
stems), -ai
25
-a
16. See 3.15, 4.65, p. 107 (arnikzil), 4.82, p. 115 (some nouns in -ar).
17. Luraghi (1997a: 2.1.6) adds -a. But what is attested (as a hapax) is only -ia (see below, n. 31).
18. For the rules governing the choice see 3.16.
19. With and without lengthening of the vowel of the stem. See 3.20.
20. With consonantal stems.
21. Plural nom.-acc. in -a or -i belonging to certain common-gender forms are collectives.
22. The endings -a and -i are used with consonantal stems and -i with verbal substantives. See 3.20.
23. Also -s in nekuz (3.22). Some, including Luraghi (1997a: 2.1.6), claim sg. gen. in -an in OH. We
have found no unambiguous case in Old Script. See 3.23.
24. OS examples:
GI
u-lu-ka-an-ni-ya (ea) KBo 17.15 rev! 20 (OS), u-wa-i-ya KUB 28.75 iii 19
(OS), ta-ki-i-ya URU-ri in another city KBo 6.2 i 7 (OS), lu-li-ya in a vat KBo 6.2 i 56 (OS), a-ni-ya -
it-ti in the same year KBo 3.22:10 (OS), a-ni-ya i-wa-at on the same day KBo 3.22:60 (OS), lu-ut-ti-ya
KBo 8.74++ iii 20 (OS), KBo 17.74++ ii 5 and passim (OS?, see Konk.),
d
al-ki-ya ( pariya, in sequence
with all other DNs in d.-l.) KUB 41.10 rev. 6 (OS), KUB 28.75 iii 25 (OS).
Noun and Adjective Inection 3.14
00-HittiteGrammar.indb 69 6/10/08 7:37:54 AM
70
Old Hittite New Hittite
Case
sg. pl. sg. pl.
voc.
26
-e, -i, - -
all.
26
-a -a
27
abl. -az(a),
28
-za
29
-az(a)
ins. -it, -dlta -it, -dlta
30
Common-Gender Nominative
2526 27 28 29 30Table of nominative endings, end.
3.15. Singular common. With the exception of a few nouns with stems in -ar- (see
4.82, p. 115) and possibly a few with stems in -zil- (see 4.65, p. 107, for arnikzil),
all of which take a zero ending, the sg. nom. common-gender ending is l-sl. On the
cuneiform writing of this lsl on stems ending in nt and Vt, see 1.11 (p. 13) and 1.90
(p. 37).
3.16. Plural common. The OH ending is -e.
31
In the MH Maat letters, the pl.
acc. com. ending -u begins to be used for the nominative (and the pl. nom. com. -e
for the accusative): see Honer forthcoming: 54, 56. The ending -a also appears on
i-stem nouns such as alki- grain, crop: al-ki
I.A
-a HKM 19:6 (Honer forthcom-
ing: 53). Variable use of -e and -u for both pl. nom. com. and pl. acc. com. continues
into NH, but by late NH (i.e., the reigns of attuili III, Tudaliya IV, uppiluliuma
II) a stable distribution is achieved (Melchert 1995). The ending -u becomes the regu-
lar ending for a merged pl. nom.-acc. com. for all stem classes, with the following
exceptions: (1) u-stem adjectives generalize -awe (eliminating the irregular pl. acc.
com. in -amu); (2) stems in -t- (mostly -nt-) use -e with a few exceptions;
32
(3) the
25. See labarnai itoi-i KUB 2.2 + KUB 48.1 iii 9, tabarnai KUB 44.60 iii 15,
MUNUS
tawanannai KBo
17.88 + KBo 24.116 iii 21 (MS),
d
Mammai KUB 27.67 ii 45,
URU
blai KBo 32.19 iii 37 (MHlMS). None
of these examples is found in OS. For an explanation of this ending see 3.24.
26. The vocative and allative did not exist in the plural.
27. To the extent that a distinctive allative form in -a occurs in NH compositions, it is an archaism and
not a form of current speech. In NH the allative was replaced by the dative-locative.
28. The -aza ending is secondary and not attested in Old Hittite (OS). The ablative and instrumental are
indierent to number.
29. The ending -za is conned to consonantal stems, e.g., i-it-tar-za with sundisks and ne-p-i-za
from heaven, r-er-za from the house (3.26, p. 74).
30. To the extent that the instrumental occurs in NH compositions, it is an archaism and not a form of
current speech. In NH the instrumental was replaced by the ablative.
31. There is a single exception [a-an-te-e]z-zi-a rst, older KBo 22.2 obv. 18. The appearance of
-ia for -ie could show an isolated early example (OS) of the change cited for MH in this paragraph and
in 4.19 (p. 87) for NH i-stems (see also 1.68, p. 31), but an archaism is also possible, as per Neu (1979a:
192).
32. nuwama sro
5
-andu tpat pei I do not give them good ones KUB 13.35+ ii 8 (Hatt. III),
and duwarnandu in iv 12; ekkandu KBo 23.114 obv. 23, 25 (oracle questions about Uri-Teub).
3.16 Noun and Adjective Inection
00-HittiteGrammar.indb 70 6/5/08 4:15:17 PM
71
relative-interrogative kui- generalizes kui (8.2, p. 149) (see McIntyre, cited by Mel-
chert 1995). No such regularity is found for i-stems, which show mostly -iu and -au,
but -ie and -ae continue to be found, along with -iya (and rarely even -a and -i; see
Melchert 1995: 271). Contrary to repeated claims, there is no compelling evidence for
the ending -a as pl. nom.-acc. com. other than with i-stems.
33
There is thus no merger
between the plural nominative-accusative and the plural dative-locative in NH. In late
NH,
34
unambiguous writings of the nominative plural common ending as -i occur:
i-da-la-u-i-r
35
t.xr-i-i (= antui) evil persons, al-wa-an-zi-ni-r an-tu-u-i-
i sorcerous persons, i.xr-i (= pini) men, urorn.xr-i (= iuni) deities, r.
xr-i (= negni) brothers, ka-ru--i-li-i itoi.xr-i[] (= aui) former kings,
a-a-ra-ni-i
MUEN
eagles (1.57, p. 28; 4.75, p. 112).
Common-Gender Accusative
3.17. Singular common. The accusative singular of common-gender nouns with
vocalic stems is -n (antuan human < antua + n, alkin grain < alki + n, aun
king < au + n). Consonantal stems use -an (kardimmiyattan anger < kardimmiyatt
+ an).
3.18. Plural common. The OH ending is -u. For developments in MH and NH
see 3.16.
Neuter Nominative-Accusative
3.19. In the neuter nominative-accusative singular, most nouns and adjectives use
the bare stem: ai mouth, uel thread, aal brushwood, alli great, idlu evil,
taru tree, wood, tekan earth. a-stems add an -n: ezzan cha, straw, ekan ice, yugan
yoke, pedan place, kunnan right.
3.20. On the neuter nominative-accusative plural in general in Hittite see Gertz
1982. In older texts we nd a zero ending for neuter plural i-stems and u-stems, such
as
GI
iparuzzi rafters and

idlu uddr evil words (see p. 103, n. 121, and Watkins
1982). There is no tendency to merge neuter nominative-accusative singular and plural
(contra Prins 1997), as shown by the fact that in later texts the zero ending is replaced
by -a (aliya corrals, arliya .-loaves, genuwa knees, idlawa evils), which is
33. If our claim is valid, one has to assume scribal corruption of the following passage: kawa ammel
tuggami nta Now my body parts are warming up VBoT 58 i 24 (OHlNS), where the plene written
-mi-e-e for the pl. nom. com. possessive clitic and the medio-passive pres. pl. 3 a-a-an-ta point to a plural
subject tugga.
34. A possible earlier example occurs in MHlMS: ka utxt.xtts.xr-i [ . . .] KBo 16.45 obv. 8,
which because of k cannot be read utxt.xtts.xr-u[]. In any event, instances of this writing of the pl.
nom. com. ending would be very rare prior to NH.
35. In view of the unambiguous -i ending on the nouns, it is possible that the attributive adjectives also
had this ending, hence, i-da-la-u-i- and al-wa-an-zi-ni- (reading E as , a value that the sign unques-
tionably has in late NH).
Noun and Adjective Inection 3.20
00-HittiteGrammar.indb 71 6/5/08 4:15:18 PM
72
the regular ending for consonantal stems such as uppala (see 4.64, p. 106), kudurra
(see 4.81, p. 114), ikia (see 4.89, p. 119) and aniyatta (see 4.94, p. 121). The
ending -i is attested principally with the verbal substantives of the type arkuwarri
I.A
,
warri, and minumarri
I.A
(see HE 185 b 2), and r-stem adjectives and nouns such as
zankilatarri
I.A
(4.104, p. 127), uppearri
I.A
, kururi
I.A
(4.85, p. 117), alaltumari,
uppari, and aniuri (4.84, p. 117; 4.81, p. 114), but we also nd aniyatti alongside
aniyatta.
36
Some neuter rln-stems that regularly take a zero ending also show plurals
with deletion of the nal -r (e.g., kuduwta calumnies; see 1.133, p. 46). A few
neuter nouns (principally in -rln-) also directly continue an archaic PIE pattern by which
the plural nominative-accusative was distinguished from the singular by a dierent ab-
laut pattern (see 3.37, p. 78): singular wdar (written wa-a-tar) water, plural widr
(written -i-da-a-ar) waters, (bodies of) water, singular uttar word, utterance, matter,
plural uttr (written ud-da-a-ar) words, singular uidar wildlife (written variously:
collective u-u-i-tar, u-u-e-da-ar, u-i-ta-ar, but never with plene writing in nal syl-
lable), distributive plural uidr beasts (written u-i-ta-a-ar, u-i-da-a-ar, u-u-i-ta-
a-ar, always with plene writing in last syllable; for the forms see HED H 35354). In
all three of the above examples, a collective idea is either possible or explicit in the
singular, and an individualizing idea is required for the plural. In the word a-a-u-u
goods, possessions, the old plural of the adjective a-a-u good, the plene writing of
the nal syllable reects the original presence of a laryngeal (*-uh
2
) (see Watkins 1982
and AHP 86, 131, 184).
Ergative
3.21. As described in 3.8 (p. 66), when a neuter noun appears as the subject of a
transitive verb, it takes an ergative case ending: -anza in the singular (rarely written
-az, 1.135 [p. 46]) or -ante in the plural. In cases where there is a dierence between
the nominative-accusative and oblique stem in the base neuter noun, the ergative end-
ings are regularly attached to the oblique stem: atiyanza < atai-latiy- bone,
[uw]alliananza < u(wa)lli(n)- ?, pauenanza < paurlpauen- re, linki-
yanzallinkiaz and linkiyante < lingai-llinkiy- oath(-deity) (3.10, p. 67). There are
rare exceptions:
GI
luttanza < luttai-lluttiy- window. Ergatives built to a collective
stem are known in the case of
[MUNUS].ME
a!-az-ka-ra-a-i-ya-za,
MUNUS.ME
a-az-qa-ra-i-
ya-za, [
MUNUS.ME
]a-az-ka-ra-an-za female temple musicians
37
built to the collective
azkarai (for which see Honer 1998a), and possibly linkiyanza and linkiyante oath-
deityl-deities built to the collective kuie ++-+i-:i
ME
(*lingai) dapianda KBo 16.98 iv
21 (CHD LN 64 pl. neut. nom.) of lingai- oath. One must be careful to distinguish
the ergative endings -anza and -ante on neuter nouns from the homophonous sg. nom.
36. aniyatta is attested already in OS (see Neu 1983: 1516 n. 74), while aniyatti is attested only in
post-OS.
37. Although in this case the ergative singular ending -a(n)za was selected, in all occurrences the verbs
of which the noun is the subject are plurals, showing that the plurality of the referent was kept in mind (see
again Honer 1998a).
3.21 Noun and Adjective Inection
00-HittiteGrammar.indb 72 6/5/08 4:15:19 PM
73
and pl. com. endings of common-gender stems in -ant- (whether these endings are
historically related is a separate question, on which see p. 55, n. 11). One must further
distinguish the ergative ending -anza when it is spelled without the -n- as -az(a) (see
1.135, p. 46) from the ablative ending (see 3.33, p. 77).
Genitive
3.22. Singular. For the genitive singular in PIE the endings: -esl-osl-s are posited
(Sihler 1995: 248; Szemerenyi 1996: 160; Meier-Brgger 2000: 183, 2003: 196). For
the Hittite sg. gen. only the ending -a (from PIE *-os) is widely attested, although
there is a rare use of - with nekuz in the set phrase nekuz meur at twilight time
(originally *time of twilight) (see Rieken 1999b: 84 with n. 400). Other forms such
as tu-az (= iwaz) inlon the day may not be genitives but nominatives. The genitive
personal names
m
Nunnu and
m
Taruu and the DN
d
Taammaz of the old language (so
Neu 1979a: 18586 with n. 22) may be relics of such a gen. in - (opposed by Melchert
1984a) but pose problems that cannot yet be resolved. See also 4.44 (p. 98) and 4.50
(p. 100) (iunzanna- < lsiunsl + lanna-l?).
3.23. Plural. In OH there existed a distinct form for the pl. gen., marked by an
ending in -an < PIE *-m (Laroche 1965; Kammenhuber 1969c: 3045, 311, 313). It
occurs on padnna
GI
oln.otn stool of the feet (OS), iunan antui people of the
gods (OHlNS), r.xr-na r.xr-na itarna among brothers and sisters (OHl
NS),
LU.ME
pian parna to the house of the .-men (OHlMS), utxt.xr-an parna
to the house of the (royal) children (OS), and urorn.xr-nana itarna but (-al-ma)
among the gods (OS), arazenan s+ i.xn xtn.xtn-:t+ the lands of the surround-
ing enemies (OHlNS).
38
As shown by Laroche (1965: 3341), some examples of this
ending must be interpreted as plurals, and all OH occurrences can be so read.
39
Despite
repeated claims, there are no assured OS examples of -an marking genitive singular
(itoi-an ki in Laws 187 may be tolin the gate of the kings, see p. 98, n. 89).
Apparent examples of -an as a singular in post-OS copies of OH archetypes either are
not genitives (see Laroche) or may be errors. One example from a MH text might be
explained by a plural conception of the sky, analogous to the plural form in the Semitic
languages (Akkadian am, WSem mm): nepian [(urorn.xr takna urorn.xr)]
the gods of the heavens, the gods of earth KUB 26.6:67 (the duplicate KBo 8.35 ii
10 reads npia urorn.xr; cited CHD LN 448 as either sg. or pl. gen.). Beginning
already in OH we nd a competing pl. gen. ending -a, which becomes the only genitive
ending in NH. It is impossible to determine whether this ending reects a merger with
the dative-locative plural or the genitive singular.
38. The rst example refutes the claim (Kammenhuber 1969c: 25455 [30.2c] and 3045 [42]) that
the ending was limited to semantically animate nouns.
39. Laroches own conclusion (1965: 40) that -an originally had a collective meaning has no basis in
fact, as his analysis of the attestations shows, but rests entirely on his erroneous prehistoric analysis of the
pronominal pl. gen. ending -enzan. His claim of a prehistoric collective value for -an is refuted by the ex-
clusively plural meaning of the cognate endings Lycian -e and Lydian -a.
Noun and Adjective Inection 3.23
00-HittiteGrammar.indb 73 6/5/08 4:15:19 PM
74
Dative-Locative
3.24. Singular. In Hittite the dative and locative, which were formally distinct in
both singular and plural in PIE (in sg. dat. *-ei, loc. *-i; in pl. dat. *-bh(y)os, *-mos,
loc. -su, see Szemerenyi 1996: 16061; Meier-Brgger 2000: 183, 2003: 19697), have
merged, ending in -i in the singular, -a in the plural. The only sg. d.-l. ending in OS
is -i. This is the regular ending throughout Hittite. In MH and NH we also nd rarely
-e: []manteya HKM 88:5 (MHlMS), KUB 13.2 iv 10 (MHlNS), and the u-stem
forms a-a-a-u-e and i-da-la-u-e cited above in 1.61 (p. 28). In the case of a-stems,
the stem-vowel is usually deleted before the d.-l. ending -i, but in a few cases we nd
the ending added to the stem in -a-, producing an ending -ai (e.g., labarnai). This rare
ending is a sporadic innovation and does not represent an archaism (so, correctly, Neu
1979a: 188).
40
In some cases it can plausibly be explained as inuenced by an im-
mediately preceding i-stem attributive adj. ending correctly in -ai: allai aannai to
the great family. In other cases the a-stem vowel was not deleted in a foreign word or
foreign name:
URU
blai,
d
Mammai, [
d
Apr]itai, labarnai, and
MUNUS
tawanannai (so again
Neu 1979a: 188, followed by Miller 2004: 141).
3.25. From a pre-Hittite IE locative in *-welon-t-i was derived the innitive in
-wanzi, while the supine in -wan may have been derived from the endingless locative
*-welon (Laroche 1970: 4142; Neu 1979a: 189). For another view, deriving the Hittite
innitive from an old ablative, see p. 185, n. 36.
3.26. A zero ending (an endingless locative, also posited for PIE by Meier-Brgger
2000: 183; 2003: 196, but not by Szemerenyi 1996: 160) also exists for a small group
of nouns (Neu 1980, with additions by Oettinger 1982c): iwat on the day, dagan on
the ground, nepi in heaven, r-er (*per) in the house, lamman inlon the name (see
p. 109, n. 142), kear in the hand (see p. 116, n. 183), and possibly also tapuwa on
the sidelrib

(see p. 118, n. 190) and X-er (*ker) in the heart. See below in 3.32 for
ablative forms built on the endingless locative, an archaism (Neu 1980): r-erza (versus
parnaz, 3.32, p. 77; 4.115, p. 130), possibly also nepiza (versus nepiaz, 3.32 and
4.89 [p. 119] with n. 197), and ketkarza and tapuza, whose archaic status is shown
by the fact that they are attested only as adverbs. For examples and uses see below in
chapter 19. The few examples of the dative-locative of u-stems like or-ru are probably
merely scribal errors but in any case do not reect anything archaic in PIE terms.
3.27. The plural always had only one ending (-a) throughout the Hittite period.
Vocative
3.28. The vocative proper exists only in the singular. Plural entities are addressed
using appositives, often but not always in the nominative (see 16.16, p. 245). A voc.
40. Forms of the dative-locative of the i-stem adjectives uppai and allai are not examples of this ar-
chaic ending but can be explained in terms of the behavior of i-stem adjectives (see 4.37 and n. 71, p. 94).
3.24 Noun and Adjective Inection
00-HittiteGrammar.indb 74 6/5/08 4:15:20 PM
75
(sg.) ending in zero is posited for PIE (see Szemerenyi 1996: 160; Meier-Brgger 2000:
18384; 2003: 196). In Hittite, most common nouns and adjectives (including a few
u-stems)
41
the vocative appears as the bare stem (i.e., it has an ending in zero). Com-
pare: (a-stems:) i-mi my lord,
d
rxtn-ta attau[mmi] O Tarunta, ou[r] father,
(u-stems:)
d
Mukianu O Mukianu,
d
Itanui arku itoi-ue O Itanu, pre-eminent
king. In OH, u-stem appellatives and names take the ending -il-e:
d
tft-i (*
d
Itanui) O
Sungod, itoi-ui (`aui) O king (Neu 1979a: 179; Eichner 1974b: 23435).
42
In
at least two cases, a-stem nouns have a vocative in -i with the a-stem vowel deleted be-
fore it: attime O my father and [r]-nimi O my brother (= *negnimi from negna-)
(see CHD sub -mi-e 2 and negna-). For a possible third instance (lli O tongue) see
p. 81, n. 25, for an alternative explanation. With i-stems the evidence is inconclusive:
d
Kumarbi,
d
Impaluri, alli great in
d
Itanu oi-li itoi-ue (i.e., *aue) O Itanu,
great king, [
d
Ala]waimi lea namma zaikei O Alawaimi, dont ght them any
longer. The enclitic possessive pronouns, which show i-mutation (4.17, pp. 8687),
show either -i or -e vocalization in the vocative: i-mi my lord, [r]-nimi O my
brother, attime O my father. The use on vocatives of the clitic possessives in -metl-
mit (
d
tftmet O my Sungod, rmit O my lord, utxtmit O my son, wappumit
O my river bank), which only occur in MS or NS copies of older texts is based on
a misunderstanding of the clitic possessives by copyists who no longer had these as a
living part of their language. Such writings do not represent a real usage of any period
(see Otten 1973: 55; and 6.11, p. 141). Consonantal stems seem to show both the bare
stem and the -i ending:
d
Wiriyanta O Wisuriyant! (read [wisuriyantld]), but pedanti
O place! and xi.nnnn-an-ti O Silver!
43
3.29. No unambiguous example of a vocative form of a neuter substantive is yet
known in Hittite. To serve as the vocative of the neuter nouns pedan place, nepi- sky,
and xi.nnnn silver, Hittite speakers used vocatives in -anti: pedanti O place!, and
xi.nnnn-an-ti O Silver!,
44
or instead of the vocative they used an appositional
direct address, such as in nu atta nepianza rorn-an tarna O Father Sky, release
41. This fact seems to undermine Luraghis attempt (1997a: 2.1.6.2) to distinguish u-stems in nam-
ing constructions (e.g.,
m
Tuttu si+si Tuttu is his name) from u-stem vocatives, since both use the bare
stem.
42. See, however, with -e.
d
tft-e (=
d
Itanue) iami O Sungod, my lord! KUB 31.127 i 1, and
d
tft-e arkui itoi-ue O Sungod, heroic king! (ibid. 15).
43. Although the break before the form in question might be restored as [. . . utxt-i-]i xi.nnnn-
an-ti lmukan kue[i] KUB 36.18 ii 7 and the form in question understood as a dative concluding the
preceding clause.
44. These forms show the vocative ending -i with stems in -ant- either directly or indirectly: directly,
if one interprets pedant- and xi.nnnn-ant- as contemporary examples of the individualizing sux
-ant- (see 2.25, p. 56); indirectly, if one views them as vocatives corresponding to ergatives in -anz(a).
Regardless of its prehistoric source (on which see the refs. in n. 11, p. 66), sg. erg. -anz(a) l-antsl that marks
a subject matches formally the sg. nom. com. of a stem in -ant-. Thus, in cases of genuine personication,
speakers could have analogically created vocatives in -anti by analyzing ergative l-antsl as l-ant+sl and
adding the productive -i to the presumed stem l-ant-l. The absence of assibilation makes it extremely un-
Noun and Adjective Inection 3.29
00-HittiteGrammar.indb 75 6/5/08 4:15:21 PM
76
them KUB 15.34 iv 32 , where the ergative case ending (-anza) on the neuter noun
nepi- is appropriate, since the neuter noun here takes a transitive verb. In the case of
Silver, the form addresses a person bearing that name, not the metal itself (see Honer
1988b: 16364). The place addressed by the vocative pedanti is to be thought of in
some sense as capable of understanding and responding to the address and therefore
animate. Of the situation in PIE, Szemerenyi (1996: 159) writes: In the neuter, nom.
voc. acc. of the same number are not distinguished.
3.30. Neu (1979a: 178) and Luraghi (1997a: 15 2.1.6) claim a vocative plural end-
ing -e but give no example to substantiate it. Nominative pro vocative occurs in rare
instances in the singular but so far not in the plural. They may have in mind an example
of appositional direct address (16.16, p. 245), distinct from the true vocative (16.13,
p. 244).
Allative
3.31. Old Hittite had in the singular a noun case ending in -a indicating motion
to, toward, or into. The allative is exclusively the case of whither, to what place,
not where, in what place (so, correctly, Otten and Souek 1969: 6263; incorrectly,
Kammenhuber 1979a). This case was discovered by Forrer (1928), who called it the
Richtungskasus. It is sometimes called the directive (Laroche 1970; Brixhe 1979;
Kammenhuber 1979a; Luraghi 1997a) or terminative (Starke 1977; Held, Schmal-
stieg, and Gertz 1987). The CHD IIIl1 (1980) p. xvi introduced the term allative as a
deliberate pendant to its opposite, the ablative. So long as the case in -a is kept distinct
from the dative-locative in -i, it is of little consequence which of the several proposed
names one gives it. In this grammar, we use the term allative. On the history of research,
see Starke 1977: 13135. From the allative form of the heteroclitic action nouns in -tar
is derived the innitive in -anna (4.107, p. 128; see also Laroche 1970: 4142; Neu
1970: 55 n. 30). Many scholars think that common local adverbs such as anda, ara,
par, and ar, which express directed movement, originated as allatives (see Laroche
1970; Neu 1974: 67; Starke 1977). Perhaps beginning already in OH (Otten and Souek
1969: 6263; Neu 1979a: 18990) and intensifying in MH, the allative forms in -a (but
not the innitives or local adverbs) were replaced by forms in -i. In NH speech, the al-
lative no longer existed, although allative forms continued to be copied by NH scribes
from older archetypes.
Ablative
3.32. The ablative and instrumental cases do not distinguish singular and plural. The
usual ablative ending is -az, with occasional post-OH spelling variant -aza (see Melchert
1977: 44347). The alternative post-consonantal ending -za (not -aza) is archaic (see
likely that the vocatives in -anti are archaisms of any sort (see the arguments of Carruba [1992: 86] against
Garrett [1990b: 275]).
3.30 Noun and Adjective Inection
00-HittiteGrammar.indb 76 6/5/08 4:15:22 PM
77
3.26 and 4.115 [p. 130]) and thus far found only in r-erza from the house, nepiza
from heaven (OS), ittarza (4.84, p. 117), and the frozen ablative in the local adverbs
tapuza to the side (19.11, p. 291, 20.28, p. 300; compare oin-laz tolon the left,
kunnaz tolon the right), ketkarza at the head (3.26; 7.19, p. 147), andurza inside
(16.92, p. 265, 19.5, p. 289), araza outside (19.5), ppa par ()za backwards
(Neu 1980). The endings -az (following vowels) and -za (following consonants) are
almost in complementary distribution in Old Script manuscripts (Melchert 1977). But
already in these oldest manuscripts -az begins to encroach on -za in the consonantal
stems. There are occasional instances of the use of the Akkadian preposition ts:i from
followed by the inected ablative form of the noun: ts:i xtn Lu--ya-az from the land
of Luwiya KBo 6.2 i 42 = Laws 20 (OS), on which see LH 31 with n. 48.
3.33. There is a rare ablative ending -anza: assured examples include
GI
luttanza
KBo 8.42 obv. 2 (OS), luttianza KBo 21.95 i 11, iananza KUB 39.102 i 1, papran-
nanza, and uddananza KUB 12.58 iv 2627. However, some alleged examples (e.g.,
annenanza, uitnanza, and tuppianza cited by Melchert 1977: 44950) are instead
ergatives (for uddananza in both functions see 4.101, pp. 124125). Since the ergative
ending may also appear as -az(a) (see 1.135, p. 46), caution is needed in interpreting
forms with endings -anza, -az, and -aza.
3.34. For the variant ending -(a)(z)zi that regularly appears before the clitic -ya also,
and and rarely in word-nal position, see 1.116 (p. 42).
Instrumental
3.35. The instrumental, like the ablative, does not distinguish singular and plural.
Alongside the prevailing instrumental ending -it or -et, we also nd a rare ending -dlt(a):
ganut and genut < genu- knee (see p. 101, n. 111), akuit and kuwat
45
< kuwa-
eye, kiarit and kiarta with the hand, itaminit and itamanta with the ear, we-
danda and wetenit with water, aganda with oillgrease, uddanta by the word. See
also kad!-du-ut with (the eagles) talons KUB 43.60 i 17. It is plausible that the simple
dental without preceding i was the earliest Hittite instrumental ending (Neu 1979a:
190), with the i developing by anaptyxis (1.801.81, p. 34) in the environment of a
preceding dental. One also nds this ending with the demonstratives: apedanda with
that, therewith (HED A 87), kedanta with this (see Melchert 1977: 458). For an ex-
ample in a unique innitive form see p. 185, n. 36. The unique form inat in namma
alwanzenan i-e-e-na-at urorn.xn-a iwa[r . . .] [. . .-]andu Then let them [. . .] the
sorceror with iear, like the Mother Goddess KBo 21.12 rev.? 2324 (for sg. nom.-acc.
iar see KUB 9.39 ii 2 and 8.39 2, 4, 5) is probably an imperfect attempt to replace
an older instrumental i-e-e-ni-it with an ablative i-e-e-na-az. It is not evidence for an
instrumental ending -at.
45. KUB 23.72 + rev. 15 (Mita text, MHlMS).
Noun and Adjective Inection 3.35
00-HittiteGrammar.indb 77 6/5/08 4:15:23 PM
78
Stem Variation
3.36. In most cases the stem, consisting of a root or a root plus one or more suxes,
to which inectional endings are attached, is invariant in Hittite. In some instances, there
is variation in the stem, not always reecting the PIE morphological feature ablaut, on
which see 3.37. In nouns and adjectives, such variation usually contrasts the singu-
lar nominative and accusative (sometimes only the singular nominative) with all other
forms of the paradigm, but other patterns also exist. It is dicult to prove, but likely,
that in a few nominal paradigms the variation in the shape of the stem is accompanied
by a shift in the accent (see on this issue 1.46, p. 25). In the demonstrative pronouns,
there is also a contrast between the stem of the nominative and accusative versus the
other cases see (7.3, p. 143).
3.37. Ablaut. Hittite nominal (and verbal) inection shows some traces of the PIE
morphological process of vowel alternation called ablaut (also apophony or vowel gra-
dation).
46
Most PIE roots and suxes were dened by the sequence of consonants they
contained. PIE roots are conventionally cited with an `e vowel, but in inected words
a root or sux
47
could appear with short `e or `o ( full grade), long ` or ` (length-
ened grade), or no vowel (zero grade).
48
It was common in PIE for dierent parts of a
single paradigm to show dierent ablaut grades (e.g., nominative-accusative versus the
other cases, or singular versus plural). Inectional classes that show such alternations
are labeled ablauting (such as i-stem nouns; see sg. nom. wei pl. nom. wea).
As in other IE languages, in Hittite the appearance of ablaut has been changed due to
prehistoric sound changes and the tendency to generalize one ablaut grade throughout a
paradigm. Details will be left to the description of the individual inectional types.
49
46. There was a strong correlation in PIE between the position of the accent and the ablaut pattern, and
it seems virtually certain that the vowel alternations of ablaut were originally phonologically conditioned.
But by the time of PIE, ablaut was a morphological process: a given morphological category had a particular
ablaut. This is still seen in the strong verbs of modern Germanic languages such as English and German:
present singslsingt, past sanglsang, past participle sunglgesungen. For a general overview of PIE ablaut see
Sihler 1995: 10824; Meier-Brgger 2000: 13542; 2003: 14452.
47. There was also ablaut in a limited number of inectional endings (e.g., see 3.22 [p. 73] on the three
PIE forms of the singular genitive ending).
48. In roots or suxes that contained one of the sonorant consonants `m, `n, `l, or `r, the sonorant be-
came syllabic (see the sound of English bottom, button, bottle, butter ) in the zero grade when between two
consonants or between consonant and word boundary. Likewise, the glides `w and `y became respectively
`u and `i in the same position.
49. For the denitions and terminology of PIE inectional types now popular in the eld see, e.g.,
Meier-Brgger 2000: 188201; 2003: 20118. No attempt can be made here to describe the historical de-
velopment of these types in Hittite. For many of the consonant stem nouns see Rieken 1999a.
3.36 Noun and Adjective Inection
00-HittiteGrammar.indb 78 6/5/08 4:15:24 PM
79
Chapter 4
NOUN AND ADJECTIVE DECLENSION
a-Stem Nouns
4.1. Of the Hittite nouns and adjectives with vocalic stems, those with stems in a
equal in number the total of those in i and u combined. They form the largest and most
regular Hittite stem-class. In this class, substantives and adjectives decline identically, in
contrast to the i- and u-stem classes. For a-stem nouns in the Old Anatolian languages
see Kammenhuber 1969b: 19394, 27980 and Meriggi 1980: 540.
123
Common-Gender a-Stem Nouns
4.2. The following are paradigms of a-stem common-gender nouns (HE 6465).
Forms in bold type are in OS. Forms in parentheses are less common variants.
4567
human being
1
father mother lord
Singular
nom.
antua,
antwaa,
antuwai
2
atta, adda anna i, ia
3

acc.
anduan,
antan,
4

antuwaan, attan, addan annan in
voc.
5
atta,
6
atti
7
i, ia
1. The stems antuwaa- and antua- are largely suppletive in OH (perhaps due to ablaut grade, ac-
cording to Rieken 2001: 19091; see here 1.76, p. 33): the former in sg. nom., the latter elsewhere. The
stems begin to compete in MH, and in NH the latter has almost completely displaced the former.
2. The rare form antwaza (KUB 12.44 iii 7), sg. nom. by context, if correct, is presently inexplicable
(HW
2
A 11819 wertlos).
3. This form is found in OS only with attached clitics.
4. The rare writing an-tu-u-u-a-an occurs in the NH fragment Bo 3379 8 (StBoT 19 50).
5. Another a-stem vocative: ne-eg-na O brother (see Honer 1988a).
6.
d
rxtn-ta [a]t-ta-u[m-mi] O Tarunta, our father! KUB 33.66 + KBo 40.333 iii 8, ed. Groddek
1999: 38, 40, 46.
7. at-ti-me O my father! (KBo 12.70 rev.! iii 10, see CHD sub marnan A and parganu-).
00-HittiteGrammar.indb 79 6/5/08 4:15:24 PM
80
human being
1
father mother lord
gen. andua,

antua,
antuwaa
8
atta anna i
d.-l. antui, antue
9
atti anni i
10
(ee, ia
11
)
Numerically Indierent
abl. antuaz
12
attaz annaz(a)
ins.
13
antuet
14
Plural
nom.
15
antue,
antui,
16

antuwae,
antuu
17
atti,
18
atti,
addu
anni i
19
acc .
20
anduu,
antuu,
antue
21
attu annu, anniu
18
r.xr-u
gen. antua adda i
d.-l. antwaa,
antua,
andua
adda ia
22
89 10

11 12 13 14 15 16 17 18 19 20 21 22
8. an-du-u-a-a KBo 17.1 + KBo 25.3 i 23 (OS); an-tu-u-a-a KUB 31.127 + ABoT 44 i 37 (OHlNS).
9. an-tu-u-e KUB 43.58 i 52 (MS), KUB 44.61 rev. 7, 19. For dative-locative in -e see 3.24, p. 74
and 1.61, p. 28. an-tu-u-e is not yet attested in OS.
10. Always written non-plene (e.g., i-i-i-i) when enclitic pronoun is attached.
11. i-a-a-i KUB 24.9 i 51 (TH 2:28 i 57) with var. i-i-e-i KUB 41.1 i 10. See also r-a i in
KUB 36.74 iii 6, 7.
12. an-tu-u-a-az together with (its) inhabitants BrTabl. i 88 (Tud. IV), an-tu-u-a-az-zi-ya (with
clitic -ya and, see 1.16, p. 42; 3.34, p. 77) BrTabl. i 89 (Tud. IV). Also attested in the writing t-az
(NH). Correct LH 270, where sg. acc. i.t
19
.it-na-az (*antuanaz) is incorrectly labeled abl.
13. Other a-stem sg. ins.: teit by means of a dream, patet and oln-it by foot, on foot, unuwait
with decoration, zapzikit with glazed dinnerware, and lalit with the tongue.
14. an-tu-u-e-et 1691lu ii 5 (MS).
15. Other a-stem pl. nom.: ae, anzae, kakkapi, pappanikne, tuekke, and walle in -e or -i,
and kakapu, zo.nr.-u (*iru borders KBo 5.13 i 27) in -u.
16. an-tu-wa-a-e-e KBo 3.60 ii 16 (OHlNS) is possibly sg. nom. by context, although seemingly
pl. nom. in form; an-tu-u-i-i KBo 3.1+ ii 32 (OHlNS), [an-tu-u]-e-e HKM 50:5 (MHlMS).
17. t.xr-u BrTabl. ii 48. For other examples of a-stem nouns with late NH pl. nom. in -u see above
sub 3.16, p. 70.
18. See 4.17, p. 86.
19. i-e-e-e KUB 30.68 obv. 6, KBo 19.88:4, i--e KBo 3.46 obv. 38. There are no examples of
pl. nom. *i-i-i or *i-i-i-i to match at-ti-i, an-ni-i, etc.
20. Other a-stem pl. acc.: nku sisters, r.xr-u (*negnu) brothers, utxt.xr-u sons, chil-
dren, zo.nr.-u borders KBo 4.10 rev. 21, u.xr-u rivers, ntn.so.xr-u mountains.
21. HKM 89:15 (MHlMS) is accusative by context. [u]lle in 89:19 seems also to be pl. acc.
22. i-a-[a?-]a-a-ma-a-a-an (iaamaan) KBo 3.1 i 21 (OHlNS).
4.2 Noun and Adjective Declension
00-HittiteGrammar.indb 80 6/5/08 4:15:25 PM
81
tongue ear
23
roof sea
Singular
nom. lla itamina,
itamana
24
ua aruna
acc. llan itamanan uan arunan
voc. lli
25
gen. lal
24
ua aruna
d.-l. lli i-dam-ma-ne,
26

itamani
27

i aruni
all.
28
ua aruna
29
Numerically Indierent
abl. rxr-az itamanaz uaza,
uazi(ya)
30
arunaz, arunaza
ins. llit itaminit,
itamanta
31
Plural
nom. lle itamane
32
acc. llu itmanu uu arunu
coll. lli
25
ua
gen. rxr.nr.-a aruna
33
d.-l. ua
23 24 25 26 27 28 29 30 31 32 33
23. For the development of the word for ear see Oettinger 2003: 14751. The word was originally a
common gender n-stem of the type of iiman- (see 4.74, p. 111), as still reected in the archaic instru-
mental iimanta. For the most part, however, the word came to be inected as a common gender a-stem
ialimana-. There is no syllabically written neuter form of this noun itaman in KUB 14.13 i 1820, as
HED ElI 458 claims. But such a neuter singular *itaman may stand behind the spelling orft-an KUB
8.83 45 (OHlMS), back-formed from a collective i-ta-ma-a-[mi-it] your (pl.) ears.
24.
tzt
itamanakan KUB 55.20 + KUB 9.4 i 4, cited by HED ElI 458 as genitive, is nominative by
context.
25. For the repeated direct address form la-a-li la-a-li Tongues! Tongues! KBo 20.59:16 and rxr.
nr. rxr.nr. KUB 44.4 + KBo 13.241 rev. 22 (the latter with apparent neuter agreement) as collectives,
see Honer 2003b: 621. Alternatively, la-a-li could be a sg. voc. in -i (compare attime O my father), on
which see 3.28, p. 75, and 16.14 (p. 244).
26. i-dam-ma-n-e-i KBo 10.45 ii 26. For sg. d.-l. in -e see 3.24 (p. 74). The writing -dam-ma- may
not imply geminate m; see 1.24 (p. 19).
27. The unique writing
tzt
i-ta-ma-a-ni KUB 55.20 i 5 is due to a scribal error (read i-ta-ma-a -ni
with Neu apud Rieken 1999a: 406 n. 2042).
28. Among a-stem allatives: laa tolon a campaign (OS), ap to the river, zo-na to the right.
29. a-ru-na KUB 36.110 iii 21 (OS), and fairly frequently in OHlMS or OHlNS.
30. The abl. za (OS) belongs to another stem u()-: see Rieken 1999a: 6566.
31. See 3.35 (p. 77).
32. i-ta-ma-n-e KBo 13.31 ii 11.
33. KUB 36.89 rev. 4 (NH) see HED A 179.
Noun and Adjective Declension 4.2
00-HittiteGrammar.indb 81 6/5/08 4:15:26 PM
82
4.3. For the stem, gender, and paradigm of kear, keara- hand see 4.82
(p. 115).
4.4. For the vocatives see 3.28 (p. 74).
4.5. The common a-stem noun pada- foot is mostly written with the logogram oln.
But some syllabic forms exist: pl. acc. pa-a-tu-u[] KBo 25.46:3, pl. gen. pa-ta-a-an
KBo 17.74 i 9 (OHlMS), pa-ta-a-n(a) (= patna) KBo 20.8 obv.? (4), 19 (OS), [ p]a-
a-ta-an KUB 34.120:6 (OHlNS) , KUB 44.36 ii 14 (OHlNS) , d.-l. pa-ta-a-a(-a-a)
KBo 17.15:10 (OS), ins. pa-te-et KUB 51.20 rev. 5 (NS) and oln.nr.-it (*patelit) in
KBo 10.2 ii 19 (OHlNS).
34
The extremely common a-stem noun son, child is always
written logographically
35
: sg. nom. utxt-(l)a, sg. acc. utxt-(l)an, sg. gen. utxt-
(l)a, sg. d.-l. utxt-li, pl. nom. utxt.xr(-e), pl. acc. utxt.xr-u, pl. gen. utxt.
xr-a, pl. d.-l. utxt.xr-a.
Neuter a-Stem Nouns
4.6. Neuter a-stem nouns are much less common (HE 64), because only common
gender a-stems are productive in Hittite. Examples are: peda- place, ega- ice, and
yuga- pair, yoke. For
GI
eya- (an evergreen tree) see below, 4.70 (p. 110).
place ice
36
yoke
37
Singular
nom.-acc. pdan, pedan ekan yugan
gen. pda, peda eka yuga
d.-l. pdi, pedi eki yuki
Numerically Indierent
abl. pdaz(a),
pedaz(a)
ins. yukit
Plural
gen. pda
d.-l. pda, peda
36 37
34. The Hittite a-stem has replaced an original root noun, whose alternating accent is probably still
reected in the diering position of the vowel length in the pl. acc. versus the pl. gen. and d.-l. See 1.146
(p. 50).
35. The suggestion that pulla- is the syllabic reading for son (Homann 1992) is unconvincing; see
CHD P sub

pulla-.
36. A secondary common gender sigmatic sg. nom. is also attested: e-ga-a KUB 21.18 rev. 19 (HED
ElI 257, citing Laroche 1969a: 37273). See also Honer 1971. No attested form of ega- unambiguously
indicates an n-stem.
37. A rare example of the bare root occurs in
i.xr
nn.nr 3 i--uk -en-zi three pairs of acrobatsldancers
come KBo 25.72:11 (OS). For discussion see Rieken 1999a: 6162.
4.3 Noun and Adjective Declension
00-HittiteGrammar.indb 82 6/5/08 4:15:27 PM
83
4.7. More common than neuter singular nouns are those inected only in the collec-
tive (neuter) plural, indicating items consisting of aggregates of components:
ft
7
aramma
(a kind of stew),
GI
galamma (a part of the door),
(UZU)
uppa consecrated meat,
38
kuata
bride price, tarua (a leather part of the harness). This latter type can be identied by
the fact that there is no clear sg. nom.-acc. form in -an. Only attested in the nominative-
accusative are aramma, galamma, piytta, prana, and tarua. The other two nouns
have the following limited paradigm.
consecrated meat bride price
nom.-acc. uppa kata, kuta, kta
gen. uppa(ya) kuata
4.8. On vocatives of neuter substantives see 3.29 (p. 75).
a-Stem Adjectives
4.9. For primary (i.e., underived) a-stem adjectives see 2.7 (p. 52). See also those
in -ala- (2.20, p 55), -alla- (2.21, p. 55) and -iya- (2.35, p. 58). Other examples can
be found in the lists provided by Reichert (1963: 6574) and Jie (1994: 621).
4.10. The paradigm of a-stem adjectives (HE 65):
39 40 41
empty
39
external right-hand earlier, former
40
Singular
nom. com. arazena
41
kunna annalla
acc. com. tldannattan arazenan zo-an annallan
n.-a. neut. tldannattan arazenan kunnan annallan
gen. arazena zo-na annalla
d.-l. dannatti arazeni kunni, zno-ni annalli
all. arazena zo-na
38.
UZU
uppa is a lexicalized form of the archaic collective plural of uppi- sacred, consecrated; see
4.37 (p. 94).
39. The adjective dannatta- empty is exclusively an a-stem. The forms dan-na-at-ta-te-e and dan-
na-at-te!-in in KUB 36.89 rev. 4142 belong to a noun desolation, wilderness (Haas 1970: 155), not to the
adjective, and a noun is also appropriate for the example dan-na-at-ti-i in KUB 21.29 i 12 (Hatt. III).
40. An i-stem annalli- also exists, for which the sg. neut. annalli, the sg. acc. com. annallin and the
plural form annalliu (annalliu KUB 38.34:7, KUB 40.2 rev. 8, and annaliu HT 4:13) are attested. See
2.21 (p. 55) and 4.17 (p. 86).
41. We do not distinguish the signs zr and zr in the bound transcription of this word, since zr can also
be read ze.
Noun and Adjective Declension 4.10
00-HittiteGrammar.indb 83 6/5/08 4:15:28 PM
84
empty
39
external right-hand earlier, former
40
Numerically Indierent
abl. arazenaza kunnaz, zno-a[z] annallaz
ins.
42
kunnit
Plural
nom. com. dannaddu arazene,
43

arazenu
zo-ni annall(i)e,
annalli
acc. com.
44

43
zo-nu
n.-a. neut. tldannatta arazena kunna annalla
gen. arazena annalla
d.-l. dannatta annalla
42 43 44 45 46 47 48

bewitching rst last exempt, free
Singular
nom. com. alwanzena antezziya appezziya arawa
45
acc. com. alwanzenan antezziyan appezziyan arawan
n.-a. neut. alwanzenan antezzian
46
arawan
gen. alwanzena antezziya arawa
d.-l. alwanzeni arawi
all. alwanzena
Numerically Indierent
abl. alwanzenaz antezzi(y)az apezzi(y)az arawaz
ins.
47
alwanzenit arawit
Plural
nom. com. alwanzene antezzie
48
appezzie
48
arawe
acc. com. alwanzenu antezziu appezziu
n.-a. neut. alwanzena
gen. alwanzena
d.-l. alwanzena appezziya
42. Other a-stem sg. ins.: ne-e-u[-it] and ni-u-i-i [t].
43. Written a-ra-a-z-n-e-e (with the r sign) and transcribed arazenie by HED A 134. arazena
in the NH examples uni arazena xtn.xtn in KBo 3.4 i 26 and k arazena xtn.xtn.xr ibid. i 28 (for
the word order see 18.25, p. 284) cannot be pl. acc. com. as often assumed, as shown by the modiers uni
and k. If we are not dealing with a misspelling for arazena in both places, the construction might be that
of a genitive land(s) of the outside, perhaps modeled on takula utn lands of peace.
44. For the a-stem adjective newa- the pl. acc. ne-mu-u reects the shift *newu > nemu (see 1.126,
p. 44).
45. a-ra-u-a KBo 22.62 + KBo 6.2 iii 22 = Laws 56 (OS), a-ra-u-wa-a KUB 8.41 iii 12 (OS).
46. From the sg. nom.-acc. neut. of appezziya- was formed the adverb appezziyan (from OS) l appezzin
afterwards (see 1.76, p. 32; 19.6, p. 290).
47. Other a-stem sg. ins.: ne-e-u[-it], and ni (or n)-u-i-i [t].
48. See p. 70, n. 31.
4.10 Noun and Adjective Declension
00-HittiteGrammar.indb 84 6/5/08 4:15:29 PM
85
internal, interior central, middle upper, superior
Singular
nom. com. anturiya itarniya
acc. com. itarniyan arzziyan
49
n.-a. neut. itarniyan
gen. anturiya arzzi(y)a
d.-l. anturiya itarniya arazziya
all. arazziya
Numerically Indierent
abl. arzzi(y)az
Plural
nom. com. anturie, anturiu a[razz]i,
50

arazziu
gen. arazziya
d.-l. anturiya itarniya arazziya
49 50sarazziyan paradigm
4.11. The oblique case forms (gen., abl., ins.) and the pl. nom. and acc. of antezziya-,
appezziya-, and arazziya- would be indistinguishable from those of the shorter, i-stem
forms, of these words, antezzi-, appezzi-, and arazzi- (2.53, p. 61; 4.38, p. 94).
i- and u-Stem Nouns
4.12. For i- and u-stem nouns in the Old Anatolian languages see Kammenhuber
1969b: 19495, 27082 and Meriggi 1980: 4192. For the u-stem nouns in Hittite
see Weitenberg 1984 and the reviews by Melchert (1984a), Carruba (1989), and Eichner
(1991).
4.13. There is a strong parallelism between the inection of i- and u-stems in Hit-
tite in terms of where and how they show ablaut. Where the former have -i-, -iy-, and
-ay-, the latter have, respectively, -u-, -uw-, and -aw-. In contrast to the a-stem class,
substantives and adjectives of the i- and u-stem classes for the most part decline dif-
ferently. Substantives regularly have an invariant stem in -i-l-u- (respectively, -iy- and
-uw- before -a-). Adjectives, however, show ablaut (see 3.37, p. 78), usually showing
-i-l-u- in the nominative and accusative singular, but -a(y)-l-aw- in the nominative and
accusative plural and in the oblique cases. There are exceptionally some substantives
with ablaut and some adjectives without it.
49. Only in the broken context of KUB 33.62 ii 4, where Haas (1994: 713) and Glocker (1997: 34)
take it as sg. acc. com. The reconstruction of the context and grammatical interpretation proposed by these
scholars is quite uncertain. The form could also be a sg. nom.-acc. neut.
50. KUB 19.2:47 (NH).
Noun and Adjective Declension 4.13
00-HittiteGrammar.indb 85 6/5/08 4:15:30 PM
86
i-Stem Nouns
4.14. There are ten times as many common gender i-stem nouns as neuter i-stem
nouns. Several productive derivational suxes end in -i-: -alli-, -ati-, -ili-, -ri-, -ulli-,
-uzzi- (see 2.212.49, pp. 5561)
4.15. The nouns pi- ritual pit, erui- basket, and zakki- latch entered Hittite
from Luwian, whence they retain remnants of a Luwian t-stem in the sg. d.-l. eruiti,
zakkit, piti, and abl. pitaz (Starke 1990: 198200, 21112, 221). The Luwian nouns
with nominative-accusative singular in -ai are also t-stems (see DLL 132 5). They are
usually inected as i-stems in Hittite (common gender or neuter), but with the -t- often
in the oblique cases (see the paradigms in 4.23). Luwian inuence is also seen in the
abl. ending -ati occurring on the noun aulati (4.23).
4.16. Many nouns borrowed from Hurrian end in i, the i-vowel often being a part
of a Hurrian sux: -i- (arui-, ubrui-), -i-, -ki-, -ki-, -ugari-, -uri-, -arti-,
etc.
51
Loanwords from Hurrian or Akkadian mediated through Hurrian enter Hittite as
i-stems: tuppi- clay tablet < Akk. uppu, ankunni- priest < Akk. ang < Sum. sanga
(with possible stem inuence from Hurrian as the intermediary of the loan?), aganni- (a
container) < Akk. agannu,
LU
apii- (incantation priest) < Akk. ipu (with metathesis,
1.140, p. 48), atupli- (a garment) < Akk. utuplu, alzi- region, district < Akk. alu,
azzizzi- ear < Akk. assu, kakkari- (a kind of bread loaf) < Akk. kakkaru, nura(n)ti-
pomegranate < Akk. lurindu (`nurindu), etc.
i-Mutation
4.17. Complicating the picture of the i-stem noun is the phenomenon of so-called
i-motion (or i-mutation). Most (but not all) Luwian nominal stems insert an obligatory
-i- between the stem and the case endings of the nominative and accusative of the com-
mon gender, singular and plural (Starke 1990: 5693). In the case of a-stems, the -a- of
the stem is deleted before the -i-. Derivatives show that the -i- is not part of the stem.
The result, however, is in eect a paradigm with -i- in some forms and -a- in others.
When the Hittites began to borrow Luwian nouns and suxes (see 2.55, p. 62), they
not only imitated the Luwian usage in loanwords but also carried it over in some cases
to native Hittite words. In paradigms of such nouns, the -i- forms tend to be found in the
common gender nominative and accusative, and the -a- forms elsewhere (Rieken 1994:
4347). This entailed not only creating new -i- forms to old a-stems (pl. nom. com.
MUNUS.ME
katr to
MUNUS
katra- female musician, atti to atta- father, pl. acc. com.
anniu to anna- mother, gimriu to gimra- eld, sg. acc. com.
GI
natin to nata- reed),
but also by reverse i-mutation new -a- forms to old i-stems (sg. gen.
GI
uluganna
51. For these derivational suxes in Hurrian see the grammatical treatments of Hurrian, such as Speiser
1941; Bush 1964; Neu 1988b; Girbal 1990; Wilhelm 1992a; and Wegner 2000: 4752, 7577.
4.14 Noun and Adjective Declension
00-HittiteGrammar.indb 86 6/5/08 4:15:30 PM
87
and abl.
GI
ulugannaz to
GI
uluganni-, which in OH is a simple i-stem without muta-
tion [see sg. d.-l.
GI
uluganniya, abl.
GI
uluganniaz]).

Predictably, however, there was
some confusion, most notably in stems with suxes containing -l(l)-: e.g., for former,
we nd the stems annala-, annalla-, annali-, and annal li- (Rieken 1994: 4950). In
cases where a noun is poorly attested, especially in OH, we cannot always be certain
whether the stem is originally in -a- or -i- (e.g., pala- and pali- way, road). Our
convention with nouns showing i-mutation is to write the stem as -ali- (e.g., nouns with
stem -aali-).
4.18. In i-stems the inectional endings combine with the stem as follows (see also
Neu 1985): (1) endings which consist of simple consonants or zero are axed directly to
the i; (2) case endings which begin with an a-vowel are axed to the longer stem in -iy-
or -ay-. Case endings that begin with an e or u vowel (pl. nom. com. -e, pl. acc. com.
-u) are axed directly to the i-stem: al-ki-e-e, al-ki-u. A glide may have existed in
speech, but no special spelling (such as *al-ki-i-u) indicates this.
4.19. NH i-stem common-gender nominative-accusative plurals in -i(y)a (e.g.,
alkia) show that examples spelled ambiguously (1.26, p. 19) as -Celi-e-e (e.g., al-
keli-e-e and tu-uz-zeli-e-e) are to be read as -Ci, not -C (see 1.32, p. 20). In NH
there is also occasional contraction of -i to -i (e.g., pl. nom. com. alki and tuzzi;
see nn. 54 and 56, p. 88), producing a pl. nom.-acc. com. ending identical to that of the
sg. nom. com. (see 1.76, p. 32; and Melchert 1995: 27172).
4.20. The allative of the i-stems ends in -iya, and the sg. d.-l. ends in - or -i. Forms
with the ending -iya also occasionally appear in post-OS texts in a dative-locative
function.
52
4.21. Rare forms of the nominative singular such as tu-uz-zi-a (instead of the normal
i-stem noun tu-uz-zi-i army)
53
may be analogical formations with the stems in -iya-
(for the paradigm of iya-stems see 4.10, p. 84).
4.22. On the endings of the plural nominative in late NH see 3.16 (p. 70).
Common-Gender i-Stem Nouns
4.23. Paradigms of common-gender i-stem substantives with non-ablauting stem:
52. For an allative example see either kattawa utniya paitteni you are about to go to the land KBo
22.1:24 (OS), or nankan tuzziya anda uwate Bring him in to the army HKM 35:67 (MHlMS). For a
d.-l. example see
NA4
uwaiya 1 tut appa[(nzi)] they seize one sheep at the stela KUB 53.14 ii 1214
(par. KUB 53.12 iv 1-8); and with an ai-stem, see luttiya 1-i . . . ipanti he libates once at the window
KBo 17.74+ ii 5 (OS).
53. This rare form, tuzzia, is found only in KBo 2.5 ii 13 (NH), ed. AM 182.
Noun and Adjective Declension 4.23
00-HittiteGrammar.indb 87 6/5/08 4:15:31 PM
88
grain army artery(?),
sacricial animal
Singular
nom. alki tuzzi auli
acc. alkin,
d
alkin tuzzin aulin
gen. alki(y)a tuzzi(y)a auliya
d.-l. alkiya, alki,
alk
tuzziya auliya, aul
all. tuzziya
Numerically Indierent
abl. alkiyaz(a) tuzzi(y)az aulati (Luw.)
ins. alkit
Plural
nom. alki,
54
alkia
55
tuzzi,
56
aulie, auliu
acc.
alkiu, alkiya
al-ki-(e-)e,
al-ki
I.A
-a
57
tuzziu, tuzziya auliu
coll.
alki
I.A
gen.
alkia
d.-l.
alkia
54 55 56 57halkis paradigm
lip (a type of priest) (a type of priest)
Singular
nom. pri purapi ptili, pteli,
btili
acc. prin
d.-l. pri, priya purapi
Numerically Indierent
abl. priyaz
ins. prit
Plural
nom. pri purap,
purapiu
patili, ptilie,
batili
acc. priu, pri purapiu
gen. patili, patiliya
d.-l. pri(y)a purapeya patili, ptiliya
54. See 4.19. An additional rare pl. nom., alki , was cited without reference in HW 47. Perhaps nu
al-ki-i iiyante KUB 15.11 ii 6 was meant. This form is not registered in HED H 37.
55. al-ki
nr.
-a in HKM 19:6 (MHlMS); see HED H 37 and Honer forthcoming 53. See 3.16
(p. 70).
56. A rare instance of pl. nom. tu-uz-zi-i occurs in KUB 31.42 iii 19 (MHlNS).
57. al-ki
nr.
-a in HKM 19:10 (MHlMS) is pl. acc. (HED H 38 and Honer forthcoming 57).
4.23 Noun and Adjective Declension
00-HittiteGrammar.indb 88 6/5/08 4:15:32 PM
89
latch observation post spring,
water source
Singular
nom. zakki, zakke auri aldanni, aldani
acc. zakkin
gen. zakki (y)a auri (y)a,
a(u)wari(y)a
d.-l. zakkit,
58
zakkiya auri, auriya aldanni, aldani
Numerically Indierent
abl. auriyaza altaniyaz
Plural
nom. auri, auriu
59
altanni, altannu
acc. zakkiu
(I.A)
,
zakk
auriu, auwari
60
altanniu
coll. za-ak-ki!
I.A
d.-l. auriya altannia,
aldanna
58 59 60 61
belt, sash anger ritual pit eece(?)
Singular
nom. iuzi karpi
SG
ri
acc. iuzzin karpin pin
SG
rin
gen. apia
d.-l. karpi piya, piti
Numerically Indierent
abl. piyaz,
pidaz(a)
Plural
nom. karpiu
acc. iuzzi(y)a
61
apya
coll. pi
SG
ri
d.-l. piya
58. See 4.15 (p. 86).
59. Late NH in BrTabl. III 44.
60. Attested in KUB 13.1+ i 19 (MS).
61. See LH 139f., 279. Only attested in NS. Analyzed by Friedrich 1959: 112, 122 and HED ElI 401 as
free-standing genitive.
Noun and Adjective Declension 4.23
00-HittiteGrammar.indb 89 6/5/08 4:15:33 PM
90
Neuter i-Stem Nouns
4.24. Paradigm of neuter i-stem substantives:
62
stela libation clay tablet
Singular
nom.-acc. uwi ipandu(z)zi tuppi
erg. tuppianza
gen.
uwai(y)a ipantuzzi(y)a tuppi(y)a
d.-l. uwaiya,
uwai
ipanduzzi,
ipantuziya
tuppi, tuppiya
all. uwiya
Numerically Indierent
abl. tuppi(y)az,
tuppiazza
Plural
nom.-acc. uwai
I.A
:ir-r+
I.A
gen. :ir-r+
I.A
-a
63
basket garbage pit,
dump
bridge
(pl. tantum)
Singular
nom.-acc. ri, rui,
erui
uil
62
gen. eriya iliya
d.-l. eriti
63
uili, uiliya,
uulli
Numerically Indierent
abl. eruyaz,
eriyaz
uiliaz
ins. ruit
Plural
nom.-acc. armizzi
d.-l. armizziya
62. This sg. nom.-acc. shows apparent confusion with stems in -il (see 4.65, p. 107). Note also the
sg. d.-l. below with a dierent stem uulli-. The limited evidence makes it hard to determine the original
inection of this noun.
63. The form er-u-i-ti KUB 38.25 i 16, 21, transliterated as er-u-u-i-ti (i.e., scribal error) in HED ElI
283, is rather a real form, showing a Luwian sound change (see AHP 258) in this loanword from Luwian.
4.24 Noun and Adjective Declension
00-HittiteGrammar.indb 90 6/5/08 4:15:33 PM
91
4.25. Other neuter i-stem nouns: appuzzi- tallow, fat, grease, arkami- (a musical in-
strument; on the ghost geminate m see 1.24, p. 19), erimpi- cedar, ri- form, shape,
image, tri- food, li- pen, corral, sector, ariari- bad weather, ariyalli-
pithos vessel, azziwi- rite, cult act, ialli- saliva, spittle, etc.
4.26. In addition to tuppianza (listed above), ergatives of the neuter nouns anna
pili- l stu-li void and
UZU
appuzzi- fat exist in the forms stu-li-an-za (see CHD sub
annapili-) and
UZU
appuzziyanza.
4.27. Perhaps dual or collective in origin: elzi- scales, ulali- dista.
4.28. Paradigm of i-stem substantives with ablauting stem:
64

spear(?) pasture
Singular
nom. com. wei
acc. com. mrin wein
gen. mriya
d.-l. weai
Numerically Indierent
ins. mrat
Plural
nom. com. mri wea
acc. com. mriu weau
4.29. The stem of allative

makziya KBo 22.186 ii 8 cannot yet be determined, since

no case form exclusive to an i-stem is yet attested.
4.30. Paradigm of the mixed r- and i-stem noun alaltumar (i)- corner.
65
Singular
all. or d.-l. alaltumari
65
Numerically Indierent
abl. alaltumaraza
Plural
nom. alaltumariy
n.-a. coll. alaltumr, alaltmari
gen. alaltu(m)mariya,
alaltumara
d.-l. alaltumariya, alaltumara
64. Largely based upon Neu 1985: 25960.
65. KBo 4.1 obv. 19 (so HED H 20); the form +-\+ 4 al-al-du-um-ma-ri-ya KUB 7.41 i 7 (analyzed
as sg. d.-l. alaldummariya in HED H 2021 and HW
2
H 29) is uncertain, because at the beginning of a
new sentence the -ya could be the conjunction and.
Noun and Adjective Declension 4.30
00-HittiteGrammar.indb 91 6/5/08 4:15:34 PM
92
4.31. The noun alaltumar- corner was originally a neuter rln-stem (see 4.112,
p. 130), but underwent a complex series of secondary developments. First, the neuter
noun generalized the r-stem (seen in the collective alaltumr, sg. d.-l. alaltumari,
and alaltumara and alaltumaraz). Second, the alternative collective alaltumari
was created (3.20, p. 72). Finally, alaltumari was taken as the base for creating an
i-stem count plural (see Melchert 2000: 6566).
ai-Stem Nouns
4.32. For the derivation of stems in -ai- see 2.18 (p. 54). Since the oblique stem
of these nouns is identical with that of ordinary non-ablauting i-stem nouns (4.23,
p. 88), examples of the latter inection also occur, such as sg. acc. aklin, ullanzin, and
zain.
66
On the other hand, the older common gender i-stem tuui- smoke (OHlMS
tuui, tuuin) acquires a new stem tuuwai- (NS tuuwai, tuuwain). Nouns in
-ai- show ablaut in the stem (see 3.37, p. 78; HE 68a: Weitenberg 1979, 1984): the
original pattern had -ai- (also spelled -ae-) in the nominative and accusative and -i- (-iy-
before -a-) in the oblique cases. But there is a tendency to generalize -ai- (-ay- before
-a-), so that it appears in the oblique cases as well, hence zaait, lingaya, etc.
4.33. Common-gender action nouns:
67
battle oath curse quarrel
Singular
nom. zaai lingai urti alluwi
acc. zaain,
zain
lingain, lingen,
lingaen
urtin alluwin
erg. linkiyanza,
linkiyaz
67
gen. zaiya linkiya, lingaya urtiya
d.-l. zaiya linkiya, lingai,
lenqai, lingae
urtiya, urtai
Numerically Indierent
abl. zaiyaz(a) lilenkiyaz(a) urtiyaz alluwayaza
ins. zaait
Plural
nom. lingai urda
acc. lingu urtau
66. Compare zaain ~ zain battle and ullanzain ~ ullanzin (VS 28.26 rev. 9) with aklain ~
aklin custom, right.
67. On the status of these ergatives formed to a common gender noun (or possibly to the collective
of such a noun) see 3.10 (p. 67). The rare example of a sg. neut. aptwa \ts urorn-tt+ KUB 26.92:7
reects a sporadic back-formation from the pl. collective lingae (see the similarly back-formed sg. neut.
umanzanan to common gender umanza(n)-, as per Melchert 2000: 64 n. 34).
4.31 Noun and Adjective Declension
00-HittiteGrammar.indb 92 6/5/08 4:15:35 PM
93
battle oath curse quarrel
coll. ling[a]e
erg. linkiyante
67
gen. urdiya
custom, right portent sin song
Singular
nom. akli agai wati iamai
acc. aklin, aklin again watin iamain
gen. aklya akiya
d.-l. or all. akli, akliya
Numerically Indierent
abl. aklayaza
Plural
nom. akl [ e],
aklau
acc. aklau agau watau iama[u]
coll. agae
4.34. Neuter nouns:
68 69 70
bone(s) window(s)
Singular
nom.-acc. atae, atai luttai, luddi
erg. luttanza
gen. ati(y)a
68
lutti(y)a
d.-l. luttiya
Numerically Indierent
abl. luttiyaz, luttanza, luttianza
Plural
nom.
GI
n.xr-u
69
acc. luttu
coll. tai, ate
70
gen. lutti(y)a
d.-l. luttiya
68. Attested as a-a-ti-i-a in KBo 20.8 rev. 7 (OS).
69.
or
n.xr-u and acc. luttu are count plurals (see 3.13, p. 68).
70. Both ate and luddi (in all their spelling variants) reect prehistoric collectives. All clear attested
cases of window are functioning as singulars. It is often hard to tell whether one should translate bone
or bones for ati.
Noun and Adjective Declension 4.34
00-HittiteGrammar.indb 93 6/5/08 4:15:36 PM
94
4.35. An aberrant form of zaai- dream is pl. acc. zaimu, which behaves as
though the stem were *zaiu-; see pl. acc. imu from /siw-us/ to iu- god. See 1.126
(p. 44), 4.45 (p. 98), and 4.50 (p. 100).
i-Stem Adjectives
4.36. The i-stem adjectives (with a few exceptions) show ablauting stems, taking
the zero grade in nominative, accusative, instrumental, allative, and sometimes dative-
locative
71
singular, and the full grade ai or au in all other cases (HE 71, Kammenhuber
1969b: 19495, 28081; Neu 1985). The (stem +) endings for the i-stem adjective are:
Singular Plural
nom. com. -i -ae or -ai
acc. com. -in -au
n.-a. neut. -i (older) -a,
(secondary) -aya
gen. (older) -a,
(secondary) -aya
(older) -a,
(secondary) -aya
d.-l. -ai, -i
all. -iya
Numerically Indierent
ins. -it
abl. (older) -az, (secondary) -ayaz
4.37. The forms in -aya- show a secondary restoration of the intervocalic y, which
was lost prehistorically (see 1.134, p. 46, and Sommer 1932: 357 with literature), either
in Proto-Anatolian (Melchert 1984b: 4445) or in pre-Hittite (AHP 130). The forms that
have been claimed (e.g., HE 15a) to show syncope (1.761.78, pp. 3233) alla
< *allaya, allaz < *allayaz, uppa < *uppaya are actually older than those show-
ing the secondary restoration. The regular system of ablaut given in 4.13 (p. 85) is
sometimes altered by inuence from the i-stem nouns, and adjectives in -ui- are further
aected by the inection of stems in -iya-.
4.38. Paradigms of the i-stem adjectives. Ablauting: alli-, uppi-, mekki-, arki-,
kari-, pali-, daluki-, dankui-, parkui-, warui-. Non-ablauting: karili-, nakk-,
arawanni-, annauli-, arwai-, uelpi-. For atuka-/atuki- and daluka-/daluki- see
4.394.40 (p. 97).

71. Hence forms like uppi from uppi- + -i, beside uppai from full grade uppai- + -i.
4.35 Noun and Adjective Declension
00-HittiteGrammar.indb 94 6/10/08 7:39:06 AM
95
large holy, pure much, many
Singular
nom. com. alli, alle
72
uppi mekki
acc. com. allin uppin
n.-a. neut. alli uppi mekki
voc. `alli (oi-li )
gen. alla, allaya uppaya
d.-l. allai, alli uppai, uppi,
uppa
73
all. alla
Numerically Indierent
abl. allayaz uppaz(a),
uppayaz(a)
meqqayaz
ins. uppit
Plural
nom. com. alla, alle, allau uppa, uppe megga, meqqau
74
acc. com. alliu,
75
allyu
76
uppau meqqu
n.-a. neut. alla, allaya, allai uppa, uppaya meggaya
gen. allaya uppaya meqqaya
d.-l. allaya, alliya [u] ppa, uppaya
72 73 74 75 76 77

karuwili
rough, dense pure dark
Singular
nom. com. warui, war parkui, parkue dankui, dankuwaya
acc. com. warun, warn parkuin dankuin, dankn
n.-a. neut. wari parkui
77
dankui
gen. waruwaya parkuwa,
parkuwaya
dankuwaya
72. See 1.58 (p. 28).
73. In uppa pedi KUB 30.42 iv 22 the allative uppa is functioning as a dative-locative in a post-OH
text.
74. me-ek-ke-e belongs to the stem mekk-, as the CHD points out. It should therefore not be read me-
ek-ki-e (see 4.41, p. 97).
75. Analogical to substantives.
76. In the rare form al-la-mu-u KBo 27.11 obv. 2 (NS) and a semi-logographic oi-la-mu-u KBo
12.89 iii 11 (MS?) the m is secondary, analogical to the stems that take this ending regularly; see 1.127
(p. 45) and 4.45 (p. 98).
77. Written pr-ku-i, pr-ku--i, and pr-ku-u-i. What was read by Archi (1979: 4546) as pr-ku-i in
KBo 22.1:6 (OS) should be read me-ma-i (see CHD mi( ya)u(wa)nt- 2b).
Noun and Adjective Declension 4.38
00-HittiteGrammar.indb 95 6/5/08 4:15:37 PM
96
rough, dense pure dark
Singular
d.-l. parkuwai dankuwi, dankui

all. dankuwa
78
Numerically Indierent
abl. warwayaz parkuwayaz(a) dankuwayaz,
dankuwayazza,
dankuyaz, dankuwaz
ins. parkuwit dankuit
Plural
nom. com. waruwa, wari,
waru
parkuwa, parku,
parkuwe,
parkuwai, parkui
tankuwa, tanku
acc. com. waruwau
n.-a. neut. waruwa, waruwaya parkui, parkue,
parkuwa, parkuwae,
parkuwaya
tankuwaya, tankuwa,
dankuwai
d.-l. parkuwaya,
parkuiya
tankuwa
78 79
white, bright wide old, former
Singular
nom. com. arki pali
79
karili, karile
acc. com. arkin karilin, karlin
n.-a. neut. arki karili
gen. kariliya
d.-l. karili
Numerically Indierent
abl. kariliyaz(a)
ins. kari[li]t
Plural
nom. com. arga pal, pale,
pala, pali
karili, kareli,
karili, kariliu,
kariliya, karila
acc. com. kariliu, karilia
n.-a. neut. arga, arki pali karila, karili,
kariliya
d.-l. kariliya, karilia
78. da!-an-ku-w[a tg]-na KBo 11.14 iii 26.
79. Forms included here are only the adjective pali-, not the noun. Additional forms attested for the
noun are: sg. acc. com. palin, sg. gen. pala, sg. d.-l. pali, pl. nom.-acc. neut. pala, palaea, etc.
4.38 Noun and Adjective Declension
00-HittiteGrammar.indb 96 6/5/08 4:15:38 PM
97
4.39. The adjective atuka-latuki- awesome, frightening has singular nominative
and accusative a-stem forms, but all other case forms appropriate to an i-stem: sg. nom.
com. aduga, sg. acc. com. atugan, sg. nom.-acc. neut. atugan, d.-l. atugai, abl.
atugayaz, pl. nom. com. atugae, adugae, acc. com. atugau. The bolded forms,
which occur in OS, show that these competing stems existed already in OH.
4.40. The adjective daluka-ldaluki- long is so far unattested in OS copies. It has a
late NH a-stem sg. nom. com. form (ou.u-a), several oblique case forms appropriate
to an i-stem (dalugai, dalugayaz, taluga, talugau, talugay[a]), several in MS, and an
atypical pl. d.-l. form dlugauwa (MHlNS),
80
which resembles the corresponding form
of the u-stem adjectives (see 4.51, p. 101). The unambiguous a-stem forms are NH
or NS. And dalugawa, if it is not an error, is probably from the NH scribe. The oldest
occurrences (taluga, talugau, talugay[a], all MS) point to an i-stem.
818283848586
Singular Plural
nom. com. ou.u-a
81
taluga
acc. com. talugau, dalugae
n.-a. neut. daluga
82
talugay[a]
83
gen. taluga,
84
daluga
d.-l. dalugi
85
dlugauwa
Numerically Indierent
abl. dalugayaz
86
4.41. The adjective mekki- much, many also has a k-stem in OH (4.60, p. 105)
that exhibits the following forms (see CHD LN s.v.): sg. acc. com. me-e-ek-kn, me-ek-
kn, sg. nom.-acc. neut. me-e-ek, pl. nom. com. me-ek-ke-e, me-ek-ke-e-e, pl. acc. com.
me-ek-ku-u. These forms are often erroneously cited as forms of the stem mekki- in Hittite
grammars (see Sturtevant and Hahn 1951: 148a; HE 71; and Held, Schmalstieg, and Gertz
1987: 16).
u-Stem Nouns
4.42. The inection of u-stem and au-stem nouns is analogous to that of the i- and
ai-stems (4.18, p. 87; see Weitenberg 1984 and Neu 1985). The ablaut rules (3.37,
p. 78) are the same: (1) endings which consist of simple consonants or zero are axed
directly to the u, (2) case endings which begin with a vowel are axed to the longer
stem in -u(w)- or -aw-.
80. KUB 27.67 ii 40, iii 44 (MHlNS), marked as suspect with ! in HW 206.
81. KUB 26.1 iii 15 (Tud. IV).
82. da-lu-ga fr-tar VAT 7481 iv 6 (OHllate NS). Sg.lpl. neut. as adverb talga KBo 10.7 ii 27, 31.
83. KUB 17.10 iii 29 (OHlMS).
84. The form in KUB 17.10 iii 33 (OHlMS) in broken context could be pl. gen., pl. d.-l., or pl. acc.
85. KBo 10.24 v 9 (OHlNS).
86. dalugayaz x[(t.xx-za)] KUB 15.17 i 5 (NH).
Noun and Adjective Declension 4.42
00-HittiteGrammar.indb 97 6/5/08 4:15:39 PM
98
4.43. Several irregularities must be pointed out: (1) in the sg. d.-l. in NH, an end-
ing -e emerges as an alternative to -i (1.61, p. 28; 3.24, p. 74)
87
and (2) the pl. acc.
com. is -amu, which comes from *-awu by the rule in 1.127 (p. 45). With substan-
tives: eyamu rains, klamu ?, wappamu banks. With adjectives: amu dear,
alluwamu deep, itlamu evil, pargamu high, argamu preeminent, wamu
full (OS), daamu heavy, etc.
4.44. Some authorities arm the existence of a sg. gen. in - (e.g., Neu 1979a:
18586 with n. 22; and Weitenberg 1984: 315 836). Others (e.g., Melchert 1984a) have
declared themselves unconvinced by the sparse evidence.
88
For discussion of possible
cases of the sg. gen. in -, see 3.22 (p. 73) and 4.50 (p. 100).
4.45. In zaimu dreams (< zaai-) Tunn. iii 51 and allamu (also written oi-
lamu; both < alli-) great, the -m- before the ending -u is secondary; see 4.35 (p. 94)
and n. 76 (p. 95).
4.46. The stem + endings for the u-stem (non-ablauting) substantive are:
89 90 91
Singular Plural
nom. com. -u -ue
acc. com. -un -u
n.-a. neut. -u -uwa (older -u)
gen. -uwa -uwa, -uwan
89
all. -uwa
90
d.-l. -ui -uwa
Numerically Indierent
abl. -uwaz
ins. -uit, -ut
91
Common-Gender u-Stem Nouns
4.47. Paradigms for sample u-stem common gender nouns:
87. See Neu 1985: 18788 and Weitenberg 1984: 318.
88. For emending the contextually genitive i
GI
ftxti-u (Laws 40, ms. A) to i
GI
ftxti-a! see
LH 4748. The OS form of A (HZL #158l11, see LH 237, third shape of A in ms. A) is much closer in
form to U than are its post-OS forms.
89. It cannot be determined for certain whether itoi-an (*auwan) is singular of the king (HHw
46 des Knigs) or plural of the kings. Because the Hittite genitive ending -an derives from a PIE pl. gen.
*-m, we have listed it here as plural (on the pl. gen. -an see 3.23, p. 73).
90. See italuwa KBo 18.151 rev. 19 (OS), .si-wa (*welluwa) in 4.46, genuwa in 4.51 (p. 101).
91. For kad!-du-ut with (the eagles) talons KUB 43.60 i 17 (OHlNS) see 3.35 (p. 77). See kad-du-
u-mi-it < *kaddutmit with their talons also in i 17 (Watkins 1995). The latter form incorrectly inter-
preted as nom.-acc. neut. in HED K 138.
4.43 Noun and Adjective Declension
00-HittiteGrammar.indb 98 6/5/08 4:15:40 PM
99
non-ablauting
meadow king
Singular
nom. wllu ttont-u
92
acc. wellun ttont-un
voc. itoi-ue
gen. wlluwa ttont-(w)a, auwa
93
d.-l. wlli ttont-i, itoi-ue
94

all. .si-wa
Numerically Indierent
abl. wlluwaz, wellwaz itoi-waz
ins. .si-it
Plural
nom.
95
itoi.xr-i
acc. itoi.xr-u
coll. wellu,
96
welluwa
gen. ttont-an
97
d.-l. wllua, welluwa itoi.xr-a
92 93 94 95 96 9798 99 100

hassu king paradigm
ablauting
98
rain
99
river bank

Singular
nom. eu
acc. un wappun
voc. wappu
100
gen. ewa, eyawa wappuwa
all. wappuwa
d.-l. wappui
92. The alleged syllabic writing of the sg. nom. a-a-u-u claimed by Gterbock (1957, in KUB
31.100 rev. 910; thereafter cited in HW Erg. 2 12) is shown by the parallel passage KBo 24.57 i 78 to be
pl. acc. of a- ashes (see Puhvel, HED H 210).
93. See namma a dapuza a-a-[u]-w[a-a lamni] 1si ipanti KBo 13.165 iii! 67, restored
following KUB 30.41 iv 5.
94. See 4.42 and 1.61 (p. 28).
95. Other u-stem pl. nom. com. forms:
r
tarnue, uppidanue.
96. Weitenberg (1984 18283) considers this noun neuter in MHlMS (-e-el-lu KBo 5.7 = LS 1), with
a transfer to common gender in NH. See other collective forms -e-el-lu VBoT 58 i 10 (OHlNS), -el-lu
KUB 17.8 iv 27 (pre-NH). -e-el-lu-wa[ in KUB 8.41 (StBoT 25 #109) ii 16.
97. See p. 98, n. 89.
98. Also ablauting is
(TUG)
eknu- (full grade eknau-). On the ablauting sux -nu-l-nau- see Weiten-
berg 1984: 22122.
99. On Hittite eu- see nal 1977 and Neu 1981a, with their literature.
100. For this unusual vocative of a noun denoting an inanimate object see 3.29 (p. 75).
Noun and Adjective Declension 4.47
00-HittiteGrammar.indb 99 6/5/08 4:15:40 PM
100
Numerically Indierent
abl. wappuwaz
ins. eawit
Plural
nom. ewe, eyawe, mu,
-e-u-u
101
acc. -e--u, e(y)amu wappamu,
102
wappu
103

101 102 103
4.48. Weitenberg (1984: 27071) regards warku(i)-,and waku(i)- as secondary u-
stems developed from primary i-stems. For the source of the i-stems see Starke 1990:
18081.
4.49. The noun u- rain already in OH has an ablauting paradigm of the type
shown by the adjective idlu-, but this is far too early to be the result of analogy with
the u-stem adjectives (Kammenhuber 1969b: 281; Neu 1985: 260 2.1). The noun
wappu- river bank has an OHlMS ablauting form wappamu (< *wappawu), but its
other forms show a transfer to the non-ablauting class.
4.50. The noun iu- god has a complex inection.
104
From a stem iu- (< pre historic
`dyu-) are formed: sg. nom. iu,
105
and sg. acc. *iun.
106
A stem iw-(< `diw-) is
the base of pl. acc. i-mu-u (also written urorn.xr-mu-u) (< *iwu; see 1.127,
p. 45). The oblique cases are formed from a stem iun- that was abstracted from the OH
sg. acc. *iun: sg. gen. iuna, d.-l. iuni, pl. gen. iunan, iuna, ins. iunit.
107
This
stem also came to be used for the nominative and accusative singular, with the addition
of i-mutation (see 4.17, p. 86), thus sg. nom. *iuni (urorn-tt+-i), acc. *iunin
(urorn-tt+-in). A rare sg. gen. in lsl (see 3.22, p. 73, and 4.44, p. 98) built to this
longer stem may also be found in the OS compound noun
MUNUS
iunzanna- (lsyunsl +
lannal) gods (or gods) mother, written logographically in OS either in the Sume-
rian and Akkadian order of constituents as
MUNUS(.ME)
x.urorn(-tt+) or in the Hittite
101. The form mu is expected for pl. acc. from lhe:w-usl (see 1.127, p. 45), attested in secondary
use as a pl. nom. A pl. acc. lhe:usl could easily be analogical after other u-stems, where the pl. acc. is the
same as the sg. nom. (e.g., wappu ), but the plene spelling in -u-ul--u remains unexplained.
102. Pl. acc. com. wa-ap-pa-mu-u KUB 33.10 ii 11 (OHlMS).
103. Otten 1961: 12021, zu den Fluufern. Per Weitenberg 1984: 329 an accusative of direction.
104. See Goetze and Pedersen 1934: 7273, 80; Ehelolf 1936; Laroche 1967; Kammenhuber 1969b: 141,
18283, 186, 29091; Steiner 1971; and especially Watkins 1974; Neu 1998; and Rieken 1999a: 3539.
105. In the sg. nom. of the OS noun(!) forms
d
iuummi our god and
d
iumi their god.
106. Preserved in with nal n assimilated to the following sibilant in
d
iu(n)ummin our god (com-
mon noun, not a proper name).
107. For a similar development of the stem - in Greek Zeus and in the interrogative-indenite
`- (, , ) from older accusative singulars see Sihler 1995: 338. The form i--na-an, beginning
a clause in an OS text, could already be an example of this development, if it is not, as Neu (StBoT 26:168)
tentatively suggests, iun + local particle -an.
4.48 Noun and Adjective Declension
00-HittiteGrammar.indb 100 6/5/08 4:15:41 PM
101
order as
MUNUS(.ME)
urorn.x. See Neu (1983: 169 n. 501, 224) for the OS forms. If the
above analysis is correct, which is uncertain, the representation of the lnsl in lsyunsl as
nz could reect the secondary nature of the sg. gen. in lsyun+sl (see 1.136, p. 47, even
though the n here was never syllabic).
108
Neuter u-Stem Nouns
4.51. Paradigms for the (non-ablauting) u-stem nouns of neuter gender, genu- knee
and
GI
tru- (or-ru) tree, wood:
109 110 111
Singular Plural
nom.-acc. genu
109 GI
tru, or-ru genuwa or-ru
I.A
erg. or-ruwanza
gen. genuwa or-ruwa genuwa or-ruwa
all. genuwa
d.-l. or-(ru) i, (or-ru)
110
genuwa
Numerically Indierent
abl. genuwaz
ins. genut, ganut
111
or-ruit
4.52. The phonetically identical singular and plural forms of the nominative-
accusative of the neuter noun or-ru (*tru) are sometimes distinguished graphically in
the texts by means of the plural determinative: or-ru (sg.) and or-ru
I.A
(pl.).
au-Stem Nouns
4.53. Hittite shows a small class of nouns with stems in -au- whose inection is
entirely parallel to that of nouns in -ai- (see 4.32, p. 92; and Weitenberg 1979), except
that the generalization of the full-grade stem is more nearly complete, with only the
noun arnau- birthing stool showing a few traces of ablaut (see 3.37, p. 78). There is
only one common-gender noun attested, arnau- birthing stool. Neuter nouns include
arganau- palm, sole, iau- sweat, and
GI
tanau- (a tree). Most of these nouns are
poorly attested. The only plural attested thus far is the count plural arganawi (KUB
33.66 ii 3).
108. For iunzanna- as showing syncope from `siunas-anna- see Watkins 1974: 105. For further discus-
sion of the dicult iunzanna-liwanzanna- see also Rieken 1999a: 37 n. 160. We take iwanzanna- to be a
case where older -u- undergoes breaking to -wa-; see Rieken 2001 and here 1.79 (p. 34).
109. The alleged common-gender form ge-nu-u abstracted from genuu, ginui, ginuin (see HW
107; and Held, Schmalstieg, and Gertz 1987: 16) has been explained by Eichner (1979) as a false analysis
of genui-.
110. A rare form: KBo 4.2 i 30, 35.
111. This form (KUB 12.63 obv. 26, MS) is an archaism showing zero grade of the root (on ablaut see
3.37, p. 78; for ganut as an instrumental see 3.35, p. 77).
Noun and Adjective Declension 4.53
00-HittiteGrammar.indb 101 6/5/08 4:15:42 PM
102
4.54. Paradigms for au-stem nouns:
birthing stool (com.) poplar (neut.)
Singular
nom. arnu
112

GI
aru
acc. arnan
GI
aru
gen. arnuwa,
arnuwa
113
GI
arauwa
d.-l. arnwi, arnuwi
114 GI
rawilarwi
u-Stem Adjectives
4.55. The normal stem + endings for the u-stem adjective are:
112 113 114115
Singular Plural
nom. com. -u -awe
acc. com. -un -amu
n.-a. neut. -u -u (older), -awa
gen. -awa -awa
all. `-awa
d.-l. -awi, -awe
115
-awa
Numerically Indierent
abl. -awaz
ins. -awit
4.56. Substantivized adjectives sometimes show ablaut (3.37, p. 78) and sometimes
not: e.g., uwaz and awaz with goods in the NH composition Deeds of uppi-
luliuma (Weitenberg 1984: 32223).
4.57. Paradigms for the u-stem adjectives (all ablauting except allu- and dampu-).
For other forms see arku- exalted (ablauting), tepu- insignicant (ablauting), and
alpu- sharp (non-ablauting).
112. The sg. nom.-acc. neut. form arnu is a secondary creation, surely due to the fact that all other
au-stems are neuter. The hapax sg. acc. arnin is an error due to confusion with the substance urnai-. See
Puhvel, HED H 175.
113. KBo 17.65 obv. 15 (MH?lMS).
114. KBo 17.65 rev. 1 (MH?lMS).
115. See 1.61 (p. 28) and 4.43 (p. 98).
4.54 Noun and Adjective Declension
00-HittiteGrammar.indb 102 6/5/08 4:15:43 PM
103
good high bad raw,
uncooked
all, entire
Singular
nom. com. u parku idlu panku,
pangu
acc. com. un parkun idlun pankun
n.-a. neut. u parku idlu uiu, ueu panku
gen. awa
116
nti-uwa
117
pangawa
d.-l. awi,
awe
pargawi
118
itlawi,
idlawi,
idlawe
pangawi,
pangawe
Numerically Indierent
abl. awaza pargawaz idlawaz(a),
nti-uwaz
119
ueawaz(a),
uiawaz,
uiawaza,
uuwaza,
ueuwaza
pangawaz
ins. awet idlawit pankuit
Plural
nom. com. awe pargawe idlawe uiawe,
uwiawe,
ueawe
pangawe
acc. com. amu pargamu,
pr-ga-u-u,
pargaw
120
idlamu panqawe
n.-a. neut. awa,
u
121
pr-ga-u-wa,
parku
122
idlawa,
idlu
121
gen. idlawa,
nti-uwa
d.-l. pargawa idlawa uiawa

116 117 118 119 120 121 122idalu paradigm
116. pr-ku-wa-a KUB 10.11 i 11 (HW 161) for expected *pargawa does not exist, since the form in
question is from parkui- pure.
117. Forms like nti-u-wa-a KBo 4.14 iii 17 need not imply *idluwa, since the complement can
represent the nal signs of i-da-a-la-u-wa-a.
118. pr-ga-u-i. A form pr-ga-u-e-i is also attested once (KBo 3.8 iii 10), a conation of pr-ga-u-i
and pr-ga-u-e. All currently attested sg. d.-l. forms are NH or at least NS.
119. The hapax i-da-a-la-az uddnaz KUB 15.42 ii 9 (HW and HED) may be a mistake for i-da-a-
la-u-az. There is no phonological explanation for the form as it stands.
120. pr-ga-u-e-e KBo 4.4 iv 30 (NH).
121. For pl. nom.-acc. neut. see i-da-a-lu ud-da-a-ar . . . a-a-u ud-da-a-ar bad things . . . good things
KUB 33.68 ii 11-12 (OHlNS). Examples of pl. attributive adjective idlu are cited in Weitenberg 1984:
21112 (531). For these and the substantivized collective pl. goods, property (OS) see 3.20 (p. 71).
122. pr-ku in KUB 24.1 i 25; see Weitenberg 1984: 13233.
Noun and Adjective Declension 4.57
00-HittiteGrammar.indb 103 6/5/08 4:15:44 PM
104
small, few heavy narrow,
tight
deep full
123
Singular
nom. com. tepu dau,
tau
atku u-u-u
acc. com. tepun daun atkun u-u-un,
u-u--un
n.-a. neut. tpu dau, tau u-u, u-u-
gen. tepawa alluwa
d.-l. tepawe
Numerically Indierent
abl. tepawaz(a) atgawaz alluwaz u-u-az
ins.
Plural
nom. com. tepaw,
te-e-pa-u-i-e
124
daawe allu,
allu
125
acc. com. daamu alluwamu,
[al]luwau
126
u-u-wa-mu-u
n.-a. neut. tepawa daawa
gen.
d.-l. alluwa
123 124 125 126
e-Stem Noun
4.58. The noun utne- land, country has an ablauting paradigm in OH, with utn in
the nominative-accusative, both singular and plural, and a weak stem utni(y)- in all other
cases (see Neu 1974: 10914). These weak stem forms (e.g., utniyaz) were preserved
in NH copies of OS originals. But in later Hittite utn- was generalized as the stem for
the entire paradigm, with a -y- often inserted before endings beginning with an -a- (see
1.142, p. 48). In the OS sg. d.-l., *utn-i contracted to utn, thus becoming indistin-
guishable from the sg.-pl. nom.-acc. For preservation of tn see 1.112 (p. 42).
123. Paradigm (for which see Weitenberg 1984: 140) given in transliteration instead of broad transcrip-
tion because of the unusual spellings (1.8, p. 11).
124. KBo 6.5 iv 26; for interpretation as lte:pawisl see LH 55 n. 174.
125. al-lu-u KUB 31.71 iii! 9 (late NH).
126. One would expect *allu (non-ablauting) or *allamu (ablauting). The attested forms are irregu-
lar. See discussion of the similar form uwamu in 1.8 (p. 11).
4.58 Noun and Adjective Declension
00-HittiteGrammar.indb 104 6/5/08 4:15:45 PM
105
Singular Plural
nom.-acc. utn utn
gen. utne(y)a utne(y)a
d.-l. utn utne(y)a, utna
all. utniya
Numerically Indierent
abl. utne(y) az, utniyaz
Consonantal-Stem Nouns and Adjectives
4.59. For athematic nominal suxes in PIE see Watkins 1998: 6263. For nouns with
consonantal stems in the Old Anatolian languages see Kammenhuber 1969b: 19699,
28391, and Meriggi 1980: 93130. For those in Hittite see Rieken 1999a.
4.60. Hittite nouns and adjectives show only a limited number of consonants or
consonant sequences in stem-nal position: , l, n, r, , and (n) t. In OH, however,
alongside the more common adjectival stem mekki- much, many there existed a k-stem
mekk- (see 4.41, p. 97; and CHD sub mekki-). For possible traces of additional stems
in velars see Rieken 1999a: 6166, to which we may also add
NINDA
tnik (sg. nom.-acc.),
NINDA
tninga (sg. gen.), whether or not it is an inherited word, and see n. 30 (p. 81) and
n. 37 (p. 82). There are no b/p-stems, and the only m-stem is giem- winter, attested
in the sg. d.-l. gi-e-mi in winter (Rieken 1999a: 7778). Other PIE root nouns with
bilabial stop stems have migrated to the vocalic stem class (see Goetze 1937 and HED
H 115 [on apa- river]).
4.61. The following derivational suxes (for their meanings, see chapter 2) end in
consonants: -ait-, -al-, -an-, -ant-, -att-,-tar, -ear, -t-, -ul-, -ur-, -want-, -war, and
-zil.
-Stem Noun
4.62. One -stem noun is known, the neuter iqaru (a vessel). Representative de-
clined forms are: sg. nom.-acc. ikaru, ikari (both MS), sg. d.-l. iqarui (NS), ins.
ikaruit (MS).
l-Stem Nouns
4.63. On l-stem nouns see Rieken 1999a: 41998. According to HE p. 54 79, all
l-stem nouns are neuters (see also Kammenhuber 1969b: 297300). This is clearly the
case for stems in -al and -ul (Rieken 1999a: 44572). The situation is less clear for those
in -i/el (Rieken 1999a: 47394). There are no l-stem adjectives, unless takul belongs
to this class.
Noun and Adjective Declension 4.63
00-HittiteGrammar.indb 105 6/10/08 7:39:37 AM
106
al-Stem Nouns
4.64. For nouns formed with a sux -al- see 2.19 (p. 54). Other nouns ending in
-al are: aal brushwood(?), underbrush(?), memal meal, our, uupal (musical
instrument),
127
uppal domestic animal, tawaral (a kind of bread), katral (a metal
harness piece), iyattal blade(?), kurtal (a container), tawal (an intoxicating bever-
age), ml (perhaps intelligence, cleverness, skill). On puriyal (perhaps for puriyalli)
lip cover see 2.21 with n. 8 (p. 55). In memal(l)- (and probably aal(l)-) the gemi-
nate -ll- is original, but because this was simplied to -l in word-nal position in the
nominative-accusative, there was a tendency to generalize -l- (see Rieken 1999a: 71
73). The single example with a geminate in tawal- is probably in imitation of memall-,
but we cannot be sure because the origin of the word is quite unknown. The geminate in
pl. nom.-acc. uuppalli is part of a separate phenomenon whose status remains unclear:
see 1.109 (p. 41). Paradigms for the stems in -al:
128 129
brushwood(?),
underbrush(?)
meal, groats (a beverage) (musical
instrument)
cattle
Singular
nom. com. uppala
128
n.-a. neut. al mmal tawal uupal uppal,
uppalan
erg. uppalanza
gen. aalla mema(l)la tawala
d.-l. aali tawali uupali upli,
uppali
Numerically Indierent
abl. memallaz tawalaz uupalaz
ins. allit mema(l)lit tawalit,
(tawallit)
uupalit
Plural
n.-a. neut. aali uupalli,
uupal,
uupla
uppala
gen. uppalan
129
d.-l. aalla
127. See on this noun Melchert 1988a; Polvani 1988; Gterbock 1995a; Prins 1997: 5758, 117; Dinol
1998.
128. On the problem of the alternative a-stem forms of uppal-, including sg. nom. com. uppala, see
Melchert 1993a: 111.
129. uppalanna annear KUB 31.127 + ABoT 44+ i 43 (OHlNS).
4.64 Noun and Adjective Declension
00-HittiteGrammar.indb 106 6/5/08 4:15:46 PM
107
e/il-Stem Nouns
4.65. Nouns ending in -elil, including those with a derivational sux -zil, are: urkil
unnatural sex act, incest, alil blossom, ower, urutil (kind of stew), dammauel
(variety of pear tree?), uil thread, pa()uil (kind of fodder), tayazzil theft, arnikzil
compensatory payment, gazzimuel (a leather part of the harness), apalzil (kind of
stew?). apalzel is sometimes dissimilated to apalzir. On the problem of apparent
common-gender forms to some stems in -ilel see Melchert 2001, but compare also Rie ken
1999a: 48283 and 49091. The forms with sux -zil are older than those with -zel.
blossom thread unnatural
sex act, incest
compensatory
payment
Singular
nom. com. alila
130
arnikzil,
131

arnikzel
132
n.-a. neut. alil,
133
alel uil, uel urkil, urkel arnikzil,
131
arnikzel
gen. alila urkila
134
arnikzila
d.-l. alili uili
Numerically Indierent
abl. alilaz urkilaza arnikzilaz
ins. uilit
Plural
acc. com. arnikzilu
135
n.-a. neut. alili
I.A
130 131 132 133 134 135

ul-Stem Nouns
4.66. Nouns ending in -ul (for examples see 2.42, p. 59). An NH alternate common-
gender a-stem aula- exists for aul favor, well being, with sg. nom. aula and sg.
acc. aulan. NH forms of the sg. gen. aula and the sg. d.-l. auli could just as well
belong to it as to the older neuter l-stem. A noun takul means peace, alliance.
136
130. Because of the multivalence of the ir sign, all post-OS forms with endings beginning with a vowel
(a-li-la-a, a-li-li, a-li-la-az) could also be read alela, aleli, alelaz.
131. In OS contexts (all in the Laws) this sg. nom. form is ambiguous as to its gender. Explicit common-
gender evidence is post-OH.
132. See [m]n arnikzel kui If (there is) any restitution in KUB 14.8 rev. 30 (Mur. II).
133. Since alel is unattested prior to NS, whereas alil is found in OHlMS, it is probable that alil was the
OH form of the word.
134. ur-ke-la-a and ur-ke-la-za are also possible readings.
135. On this late NS form as evidence of animacy see Melchert 1997a: 713.
136. According to Neu (1979b), both kurur and takul are nouns, not adjectives.
Noun and Adjective Declension 4.66
00-HittiteGrammar.indb 107 6/5/08 4:15:47 PM
108
sin obligation well being peace, alliance
Singular
nom.-acc. watul, wadul iiul aul takul
gen. utula,
watula,
watulla
137
iiula aula takula
d.-l. waduli,
watulli
auli takuli
Numerically Indierent
abl. watullaz,
wadulaza
iiulaza
ins. wadulit,
watulit
takulit
Plural
nom.-acc. watul
I.A 138
iiuli takul

137

138
4.67. There are also a-stem nouns in -ula-, such as aandula- and pittula-. See 2.43
(p. 60).
n-Stem Nouns
Neuter n-Stem Nouns
4.68. Most n-stem nouns (4.684.71) are neuter (Kammenhuber 1969b: 19798
19.5; Oettinger 1980): tkan earth, rman illness, lman name, kuan wage,
aan feudal due, naan fear, reverence, engan death, plague, agan oil,
takan midst, ekan span, ilan step, degree, inan disease.
139

4.69. The a in the nal syllable -an of some n-stems does not appear in their oblique
cases:
140
laman name (gen. lamna, not *lamana ), araman (a kind of bread) (abl.
aramnaz), kuan wage (d.-l. kuni ), tkan earth (gen. takn ), wilan clay (gen.
wilna
141
).
137. The forms transcribed watulla, watulli, and watullaz are written wa-a-tl-lali-, in which the
sign sequence -tl-lali- may not require the positing of a geminate l, as *wa-a-dltu-ul-lali- would (1.25,
p. 19). See also 1.97 (p. 39) and 1.109 (p. 41).
138. wa-a-du-ul-la KUB 18.9 ii 15 in broken context is most likely wadulla with -al-ya and and is
no evidence for a pl. nom.-acc. watulla.
139. The stems karzan- (a weavers tool) and
GI
anzan- ? inect in Hittite as neuter n-stem pluralia
tantum, with forms parallel to araman- (pl. nom.-acc. karza, oblique stem karzan-). For competing ac-
counts of the prehistory of karzan- see Eichner 1973: 98 n. 78 and Rieken 1999a: 39192.
140. This alternation does not reect ablaut but the fact that the prehistoric syllabic `-n of the nominative-
accusative appears as Hittite -an, while `-n- of the oblique cases is preserved. See n. 48 (p. 78).
141. In addition, wilan clay shows an unusual sg. d.-l. -li-ni-i.
4.67 Noun and Adjective Declension
00-HittiteGrammar.indb 108 6/5/08 4:15:48 PM
109
4.70. Paradigms for the neuter n-stems:
142
(a tax or
obligation)
name (a kind of bread)
Singular
nom.-acc. aan laman
gen. aana lamna aramanna,
aramna,
aramma
d.-l. ani lamni, lamman
142
aramni
Numerically Indierent
abl. aanaz(a),
aanza
si+-za
143
aram(ma)naz
ins. anit lamnit aramnit
Plural
nom.-acc. aani aram(ma)na,
aramma,
144

arma
d.-l. aana lamna aramna
143 144 145 146 147 148
Case wage, fee plague oil
Singular
nom.-acc. kuan, kan inkan, enkan agan
145
gen. ingana agn
d.-l. kuani, kuni,
kuan
engani,
ingani, eganni
agni
146
Numerically Indierent
abl. kuanaz, kunaz inganaz(a)
ins. aganda
147
Plural
nom.-acc. ku-u-a-ni
148
142. The so-called endingless locative (Neu 1980). The form la-am-ni-i belongs to the heteroclitic noun
lammar moment, hour.
143. See 31.23 (p. 437).
144. The variant aramma results from aramna by assimilation (see 1.122, p. 44). Here the regular end-
ing -a has been added to the stem aramn-. The variant arma is an archaism (see Gertz 1982: 298, 305).
145. a-ga-an KBo 38.68 obv. 8.
146. a-ag-ni KBo 32.14 iii 11 (MHlMS).
147. a-gn-da KBo 22.2 obv. 2 (OS).
148. This form in -i is a collective, not a count plural, just as in English wage and wages are used
interchangeably.
Noun and Adjective Declension 4.70
00-HittiteGrammar.indb 109 6/5/08 4:15:48 PM
110
(a tree variety)
149
bead
150
Singular
nom.-acc. eyan
151
kuwannan,
kunnanan,
kunnan
gen. eya, eyana kunna,
kunnana
d.-l. eya, eyani
Numerically Indierent
abl. eyaz kuwannanaza
ins. eyanit kunnanit
Plural
nom.-acc. eya,
GI
e-i-e
152
kuwanna
gen.
d.-l. kuwannana
4.71. Paradigm for neuter n-stem tekan with ablaut (3.37, p. 78):
149 150 151 152153
154 155
nom.-acc. tkan
153
gen. takna, takn
d.-l. takn,
154
tagn,
155
tgan
all. takn
abl. taknz(a)
4.72. Alternating vowel length in this paradigm as well as ablaut grades suggest
word stress on the plene position (AHP 185); see 1.8 (p. 11) and 1.146 (p. 50).
149. This word shows a mixture of n-stem (eyana, eyani, eyanit) and a-stem forms (eya, eya, eyaz).
Because it is found in OS only in the nom.-acc. for eyan, which could be either (see 4.6, p. 82, for the a-
stem neuter paradigm), diachronic priority can be established for neither stem.
150. This word as attested shows a mixture of n-stem and a-stem forms. It is found to date only in NH
or NS texts.
151. Sometimes written e-a-an.
152. All nominative-accusative forms written e-ya we take as plurals, even if in some cases they func-
tion as collectives denoting a single object. The form
or
e-i-e is problematic. What appears to be I + E graphi-
cally is probably a scribal lapse for graphic I + A, i.e., r (lyal).
153. This word is conventionally written tkan, takn in broad transcription. Both te-kn and the end-
ingless locative da-ga-an with their single writing between vowels suggest that the velar was voiced. The
nature (in Hittite) of the initial stop (d or t) is uncertain.
154. AHP 185, 187, 361.
155. An identical form tagn (OS) or dagan occurs also in the compound noun daganzipa- (dagan +
ipa-) earth-genius > earth. See 1.136 (p. 47).
4.71 Noun and Adjective Declension
00-HittiteGrammar.indb 110 6/5/08 4:15:49 PM
111
Common-Gender n-Stem Nouns
4.73. All common-gender n-nouns in Hittite except one (4.78, p. 113) show a pat-
tern in which the sg. nom. ends in -a (without -n-!), while all other cases have a stem
containing -n-.
156
In many instances the latter stem is invariant -an-, but some nouns
show ablaut (see 3.37, p. 78), with variants -en- and -n- as well as -an- (see Oettinger
1980; Neu 1981b: 35051; Oettinger 1982b: esp. 17577; Melchert 1983: 312; and,
above all, Oettinger 2003, largely superseding the previous treatments; but compare the
dierent analysis by Rieken 2004b).
4.74. Nouns belonging to the type with sg. nom. -a, oblique in -an-l-en-l-n-, in-
clude: arkamman- tribute, alkitan- branch, ran- eagle, memi(y)an- word, thing,
iimalen- cord, and the ethnicon sux -um(a)n- (for the last item, see 2.45, p. 60).
157

Only the last two show a vocalic alternation ela. There is a tendency in all these nouns
to adopt the inection of the a-stemsin two directions: based either on the sg. nom.
or on the oblique stem. Thus one nds sg. acc. ran beside ranan, and, in fact, only
arkamman, memi(y)an, and
m
uppiluliuman, as if the stems were ra-, arkamma-,
memi(y)a-, and
m
uppiluliuma-. On the other hand, one also nds sg. nom. alkitana,
158

iimana, and
URU
utummana man of uta, as if the stems were alkitana-, iimana-,
and utummana-. The inection of stems in -uman- is further complicated by the loss of
the -a- in the oblique stem (see 4.69, p. 108) and the phonetic changes this produces
(see 1.122, p. 44). For further examples of forms of -uman- see 2.452.47 (p. 60).
4.75. Paradigms for animate nouns with sg. nom. in -a and oblique stem in -an-:
159
eagle tribute cord, rope
Singular
nom. ra arkamma iim,
iimana
acc. ranan,
ran
159
arkamman iimenan,
iimanan,
iaminan
gen. rana arkammana
d.-l. arkammani
156. This type reects a PIE paradigm where the sg. nom. lacked not only the -n- but also the regular
sg. nom. ending -s. See Latin hom, hominis human being, man. Hittite has added the regular ending -s
to the sg. nom.
157. Possibly also
(DUG)
an- lamp, which seems to have developed secondarily in NH a geminate n,
perhaps remodeled on the basis of Luwian loanwords in -annali-.
158. [
GI
alk]itana at KUB 33.117 iv 11.
159. Sg. acc. ranan is OS; ran does not occur before MH and is a back formation, analogical to
sg. nom. ra.
Noun and Adjective Declension 4.75
00-HittiteGrammar.indb 111 6/5/08 4:15:50 PM
112
eagle tribute cord, rope
Numerically Indierent
abl. iimanaz
ins. iimanit,
iimanda,
iimanta
Plural
nom. rani
160
iimne
acc. arkammu,
argamanu
iimanu
gen. arkammana
Note that the adverbial derivative like an eagle shows the n-stem: ranili. For the
instrumental forms iimanda and iimanta see 3.35 (p. 77).
160
4.76. The noun memi(y)an- word also has a variant stem memin- seen in the forms
memini and meminit (see n. 54, p. 33). The unique paradigm of grape cluster may re-
ect mixture of two dierent paradigms (an i-stem mri- and a derived stem mri(ya)n-)
or simply mri(ya)n- with contraction.
161 162
word grape cluster
Singular
nom. memi(y)a mure, mri
acc. memi(y)an
161
murin
gen. memiyana
d.-l. memini,
memiyani,
memieni
Numerically Indierent
abl. memi(ya)naza
ins. memi(ya)nit mrinit
Plural
acc. memiyanu,
memiyane,
memiya
mriu,
162

mriyanu
gen. memiyana
d.-l. memiyana
160. See 1.57 (p. 27), 3.16 (p. 70).
161. memi(y)an is also attested in NH as a secondary sg. nom.-acc. neut.
162. From KUB 43.23 (rev. 21), which we consider to be OS, contra Konk., which lists it as MS.
4.76 Noun and Adjective Declension
00-HittiteGrammar.indb 112 6/5/08 4:15:51 PM
113
4.77. The noun meaning witness originally had a stem in -walen- with sg. nom.
in -wa, like the nouns in 4.74 (p. 111) (Oettinger 1982b), and is likewise subject to
leveling as an a-stem kutruwa-.
163 164 165
Singular Plural
nom. kutruwa
163
kutrue, kutruwa, kutr,
164

kutruwane, kutruwene, kutarwene
165
d.-l. kutrui kutruwa
4.78. The Hittite word for woman (xtts-n-), which is descriptively an n-stem,
shows a unique paradigm with the addition of -s in the sg. nom.
166
If Neu (1990) and
Carruba (1991; 1993a) are right, the underlying Hittite word was kuinn(a)-. But there
are objections worth considering (Gterbock 1992, 1995c).
167 168
Singular
nom. mtts-za, xtts-a
167
xtts-i
168
acc. mtts-nan, xtts-an
gen. xtts-na, xtts-a
d.-l. xtts-ni, xtts-i
Plural
nom. xtts.xr-e, xtts.xr-i, xtts.xr-u
acc. xtts.xr-u
gen. xtts.xr-a
d.-l. xtts.xr-a
4.79. A small class of common-gender nouns has a stem in -anzan- (Melchert
2003b). Examples include:
GI
alanzan- (a kind of tree), itanzan- soul, mind, will,
la()anzan- shelduck, and
(U)
umanzan- (bul)rush. These nouns closely resemble
those in 4.75 in having a sg. nom. in -anza without -n-, while the rest of the cases are
based on a stem -anzan-. But they also have a sg. nom. in -anza without the nal -.
Once again, there is a tendency to remodel this alternating paradigm as an a-stem, either
after the sg. nom. (thus MS sg. gen. alanza as if to alanza- or NS pl. acc. la()anzu
163. KBo 15.25 obv. 35 nuza zik
d
tft-u kutruwa .
164. In kutra (var. kutruwaa) rorn-an iyante aandu KUB 13.4 ii 3637 the form kutr is
pl. nom., not sg.
165. ku-tar--e-ne-e KUB 23.78B ii 9. See on this writing 1.83 (p. 35).
166. On suggestions for the Hittite word underlying xtts (earlier reading si) woman (either
*g(u)wen-, or *g(u)wan-), see Neu 1990; Carruba 1991; Gterbock 1992; Carruba 1993a; Gterbock 1995c;
Weitenberg 1995: 337; HED K 3068; and Rieken 1999a: 3942.
167. As shown by the spelling xtts-na-a KUB 8.66 iii 3 and 10, xtts-a is a secondary a-stem
based on the oblique forms (see 4.74, p. 111).
168. xtts-i KUB 44.4 + KBo 13.241 rev. 27.
Noun and Adjective Declension 4.79
00-HittiteGrammar.indb 113 6/5/08 4:15:52 PM
114
as if to la()anza-; even laanziu [late NS] as if to `laanzi-) or after the other cases
(thus NS sg. nom. itanzana as if to itanzana-). The noun umanzan- also back-forms
a neuter singular based on the collective plural umanza.
r-Stem Nouns and Adjectives
4.80. Of the r-stem nouns (HE

80; EHS; Kammenhuber 1969b: 287), those in -ur
(2.48, p. 61) are all neuters, with the exception of
NA
4
ekur rock sanctuary, which has
common gender agreement (kn
NA
4

ekur) but as yet no unequivocal common gender
inection.
169
There are only two nouns in -er: aduer brushwood, an old collective,
is inected as a neuter, while ater- star is common gender and adds an -s in the sg.
nom., producing atrz(a).
170
The sg. nom. xti-a and the stem form
URU
xti-ra
suggest that a secondary a-stem atelira- arose as it did in the case of kear > keara-
(4.82).
171
Stems in -ar include nouns of both genders and adjectives.
4.81. Paradigms of nouns in -ur and -er.
172 173 174paradigm of hasterza
soup, stew ritual (a body part) brushwood star
Singular
nom. com. aterza
acc. com. xti-an,
URU
atiran,
URU
ateran
n.-a. neut. parr aniur kudur adur
gen. parra aniura
URU
atira,
URU
xti-ra,
xti-a
Numerically Indierent
abl. parrazzi
172
aniuraz aduerraza
173
ins. parrit
Plural
nom. com. mtt.t.n-e
n.-a. neut. aniuri
174
kudur(r)a,
kudurri
I.A

aduer
169. Inected forms of
NA 4
ekur are virtually non-existent. Most examples are part of place names
where the word is written in the stem form. The nal wedge (-a ?) on the pl. nom. form
NA 4
-kur
I.A
-a
cited by HED H 289 is clearly indicated in the copy with a circle of dots as partially erased. Without im-
mediate context, it is a toss-up whether in KBo 17.62 + KBo 17.63 iv 4 to read
NA 4
-kur-u with unusual
syllabication or
NA 4
-kur so.t (with Beckman 1983b: 3839), and nothing in that context suggests
the need for a plural form.
170. See Weitenberg 1995: 33435, citing anterior literature.
171. The earliest examples of sg. nom. xti-a are in MS manuscripts.
172. See 1.116 (p. 42).
173. On the geminate sonorant see 1.109 (p. 41).
174. a-ni--ri is incorrectly registered as sg. d.-l. in HED 1:70.
4.80 Noun and Adjective Declension
00-HittiteGrammar.indb 114 6/5/08 4:15:53 PM
115
4.82. In the case of nouns in -ar we must distinguish between original r-stems and
secondary ones. Of original r-stems, neuter nouns in -ar such as appar- purchase,
sale have an invariant stem and require no further comment.
175
The PIE r-stem poly-
syllabic animate nouns are generally transferred to the a-stem class in Hittite (Rieken
1999a: 26168): e.g., agent nouns in -tara- (< PIE *-tor ; see 2.41, p. 59). There is one
exception, the word for hand (Weitenberg 1995: 334; Rieken 1999a: 26162), which is
once attested with the inherited asigmatic sg. nom. com. kear-i his hand.
176
This
is also the only r-stem noun to show traces of ablaut in its stem. The alternation of e,
i, a, and zero between the and the r of the word for hand probably indicates that
no vowel stood in that position and the word contained a consonantal sequence ssr (so
AHP 29). In the earlier stages of Hittite the word for hand had the following paradigm
(HE 80b): sg. nom. ke-e-ar, sg. all. ki-i-ra-a, sg. d.-l. ki-i-ri or ke-e-ar (3.26,
p. 74), ins. ki-i-ar-ta, ki-i-ar-at.
177
In OH the vowel in the rst syllable of sg. nom.
kear was stressed e, but in the other forms of the paradigm was unstressed i (AHP
101). In the later language, hand became a common gender a-stem noun, as others of
its class (e.g., the agent nouns in -tara-) had previously done. The only forms diagnostic
of this later a-stem are the sg. nom. examples: kiira, keera, and keira. Late NH
pl. acc. kiariu shows the same intrusive i found in anniu (< anna- mother), gimriu
(< gimra- steppe, grassland), ittariu (< ittara-), and laanziu (< laanza [a bird]),
on which see 4.2 (p. 79), 4.17 (p. 86), 4.79, and 4.84.
178 179 180
hand purchase, sale
178
Singular
nom. com. kear, kiira,
179
keera,
keira
acc. com. kieran, kiiran
180
n.-a. neut. appar
gen. kiara
175. See n. 178 for the form appariu. Against HED H 125 (following Laroche), the hapax a-ap-p-ir
in KBo 12.70 rev. 14 (NS) is a back formation based on appiri- beside appari-. See Rieken 1999a: 318
n. 1549, following Kimball.
176. On this correct interpretation of keari see Weitenberg 1995: 334, with references. Contrary to
Kammenhuber (1969b: 281, 287, 310), it is not keari, but rather the form t-a-e-et of the NH parallel
text that is erroneous.
177. For the orthography ki-i-ar-at see 1.11 (p. 12) and 1.12 (p. 13).
178. We follow LH 121, 2078, 276, which takes a-ap-pa-ri-u in Laws 146 (aa ii 11) as a dierent
stem, appari(ya)- sale price and rejects the view that this is to be read as [ant]ezziya appariu wi
he shall buy them at the price of the rst (man), proposed by Neu, StBoT 18 107 and followed by HW
2
H
21516 and HED H 12526.
179. ki-i-i-ra-a VBoT 58 i 39 (OHlNS).
180. ki-e-ra-a-a-an (*kieranan his hand) KBo 3.27 obv. 3 (OHlNS), ki-i-e-ra-an KUB 11.34+
IBoT 4.54 iv 6 (OHlNS), KBo 24.45 obv. 31.
Noun and Adjective Declension 4.82
00-HittiteGrammar.indb 115 6/5/08 4:15:54 PM
116
hand purchase, sale
178
Singular
d.-l. kiar, kiari,
181
kiiri,
kir ,
182
kear
183
appari
all. kir
184
Numerically Indierent
abl. kiaraz apparaz
ins. kearta, kiarta, kiarta, ki-i-
ar-at, kirt, kierit, kiirit
185
Plural
nom. com. kiari
acc. com. kieru, kiariu
gen. kira, kiara
d.-l. kira
181 182 183 184 185kissar hand paradigm end
4.83. Due to a pre-Hittite change involving nal sequences of `-Croslm (Melchert
1993a), a number of common-gender nouns (ittara-, uppara-) came to have sg. nom.
and sg. acc. forms in -ar although their stems were in -ra-. Contrary to Neu (198283:
12526), there is no evidence that there was any gender alternation in any of these stems
in OH. The anomalous sg. nom. and sg. acc. com. forms in -ar, however, were replaced
in later Hittite by a-stem forms in -ara and -aran.
186
4.84. The paradigms of ittar sun disk and uppar (a vessel) follow:
187
sun disk (com.) (a vessel)
Singular
nom. com. ittar uppar, ppara
187

acc. com. ittaralin upparan
gen. uppara
d.-l. ppari
181. ki-i-a-ri-i StBoT 8 i 28 (OS), ki-i-a-ri KUB 13.20 i 18 (MHlNS), ki-i-a-ri-i-mi in(to) their
hand StBoT 8 i 20 (OS), ki-i-[(a-ri-i-i )] in(to) his hand KUB 11.1 ii 5 restored by KUB 11.6 i 4
(= THeth 11:24 i 70) (OHlNS), ki-i-a-ri-mi KBo 3.1 ii 22, KBo 3.23 i 4 (OHlNS).
182. ki-i-ri-i in(to) the hand HKM 44:10 (MHlMS), ki-i-ri-mi in(to) my hand KBo 3.28:7, ki-i-ri-
it-ti in(to) your hand KBo 3.21 ii 5 (Adad hymn, MHlNS), KUB 31.127 i 23.
183. ke-e-ar endingless locative. See Neu 1980: 3335.
184. KBo 8.42 rev. 4 (OS or OHlMS).
185. ki-i-e-ri-it KUB 29.40 iii 32, ki-i-i-ri-it KUB 30.37 i 5, 6 (cult inv., NH?), ki-i-ar-ta KBo 20.8
obv. 11 (OS), KBo 8.35 iii 7 (MHlMS), KUB 12.63 obv. 26, 31 (OHlMS), etc.
186. In the case of the word for horn, the response to the above sound change was the creation of two
separate paradigms: a neuter a-stem awitra- and a neuter r-stem watar (Melchert 1993a: 1056).
187. See e-me-e-na-a u-u-up-pa-ra-a u-u-u KBo 20.8 rev. 4 (OS); ke-e-da-ni-ma
or
u-up-pa-
ra-a kat-ta ki-it-ta KUB 33.59 iii 89 (OS).
4.83 Noun and Adjective Declension
00-HittiteGrammar.indb 116 6/5/08 4:15:55 PM
117
sun disk (com.) (a vessel)
Numerically Indierent
abl. ittaraz, ittarza upparaza
ins. upparit
Plural
nom. ittare
acc. ittariu, ittara upparu
coll. uppari
4.85. Paradigms for r-stem adjectives:
188huppar paradigm end
hostilility full, complete ominous
Singular
nom. com. kurur akuwaara,
akuwaar
acc. com. akuwaaran kallaran
n.-a. neut. kurur akuwaar kallar
188
gen. kurura akuwaara
d.-l. kruri akuwaari
Numerically Indierent
abl. akuwaaraza
ins. akuwaarit kallarit
Plural
nom. com. kurur akuwaaru
acc. com. akuwaaru
n.-a. neut. kurur, kururi
I.A
,
kurura
kallar, kallra
gen. kurura
4.86. We follow Neu (1979b; 1983: 97 with n. 354) and others in analyzing kurur- as
a noun secondarily used as an adjective. The adjectives in -ar (e.g., akuwaar, kallar)
have the same history as the nouns described in 4.83.
s-Stem Nouns
4.87. For the PIE origins of the inection of Hittite s-stems see Rieken: 1999a 171
83; Beekes 1995: 18586; and Sihler 1995: 305. In Hittite most s-stem nouns are neuter,
the exceptions being soap and andai heat. The following nouns belong to the
s-stem class: ai mouth, anki (a kind of plant), andai heat, apu reed stalk(?),
188. [ . . . k]al-la-ar KUB 12.43:4 (MS?; OS according to Konk.).
Noun and Adjective Declension 4.87
00-HittiteGrammar.indb 117 6/5/08 4:15:56 PM
118
penis(?), soap, upalla scalp(?) or skull(?), iki back, kutri reckoning (see
HED K 298), nepi sky, heaven, purpuri (a vessel), kalmu lituus, crook maldani
mushroom(?), danna (a kind of bread), zakarai anus (< zakar excrement + ai
mouth). ai mouth and the compound zakarai containing it are the only s-stem
nouns that show ablaut. The noun u(wa)lli()luwalli (a)n- pine cone shows both
an s-stem, an extended stem -n- (see 4.90 with n. 202, p. 120), and a shorter i-stem
(HW 74, HED H 424).
4.88. Paradigms for s-stem nouns:
mouth side
189
danna-bread
Singular
nom.-acc. ai tapuwa danna
erg. tpuwaanza
gen. ia tapuwaa
d.-l. i, aii tpwa (?)
190
all. ia
Numerically Indierent
abl. iaz
ins. iit dannait
191
Plural
acc. com. xt.nr.-u
192
d.-l. tapuwa
189 190 191192 193
lituus heat soap, ash
Singular
nom. com. andai
193
, aa
acc. com. aan
189. The adverbs tapua to the side and tapuza beside were originally the allative and ablative of
this s-stem.
190. The second ta-a-pu-u-wa-a form in KUB 9.4 i 9, which HW 213 assigned to the genitive, is con-
textually a dative-locative (probably singular). Perhaps it is an endingless locative.
191. KUB 2.8 i 34. The forms
NINDA
tannaza and dannaza do not belong to his word; instead they are
sg. nom. to stems in -ant -.
192. KUB 14.4 ii 10 (NH). This example is not evidence for heterogenicity in this word but rather a
sporadic example of the use of a common-gender form for a count plural in early NH (see 3.13, p. 68).
193. Although no inectional aspect of this word shows it to be common gender, it must be such. Neu-
ter nouns cannot serve as subjects of transitive verbs (see 3.83.9, pp. 6667), whereas andai does so:
mna[(n)] andai walzi zigan ekunimi dai KBo 3.23 i 56 (NS), with restoration from KUB 31.115
(NS).
4.88 Noun and Adjective Declension
00-HittiteGrammar.indb 118 6/5/08 4:15:56 PM
119
lituus heat soap, ash
Singular
nom.-acc. kalmu
gen. kalmua
194
d.-l. kalmui andaii
Numerically Indierent
abl. kalmuaz,
kalmuati (Luw.)
ins. kalmuit
Plural
acc. com. au
4.89. One group of nouns has members with secondary common-gender stems in the
nominative and accusative extended by a thematic vowel a: iki(a)- back, nepi(a)-
sky, ate(a)- axe.
194 195 196 197 198 nepis sky: paradigm beginning
back sky axe
195
Singular
nom. com. [nep]ia, -a
acc. com. ikian nepian atean
nom.-acc. iki npi
196
ate
erg. nepianza
gen. ikia nepia
d.-l. ikii nepii, nepi
197

all. ikia n pia
198
194. If kalmuiya KBo 34.175:6 belongs to the paradigm of kalmu, it resembles kardiya to kard-
heart (see 4.116, p. 131), but it may also belong to a derived adjectival stem *kalmuiya- of the lituus.
195. Against Rieken (1999a: 192), the form a-te-e-a-na-a in KUB 9.31 i 27 is not hopelessly cor-
rupt but merely a sg. nom. used in a list despite its syntactic role as a direct object, for which there are
parallels. This secondary sg. nom. com. to a stem ate (a)na- shows that this noun originally belonged to
the type of s-stem with an -n- in the oblique cases; see 4.90 (p. 120). Only the extended stem in -n- can
explain the consistent geminate -- in this word, which cannot be explained by accent: see AHP 15051.
196. ne-e-p-i (OS), ne-p-i (OS and later), ne-p-e only in NS; see CHD LN 448. Plene writing of
the initial syllable is common in OS.
197. On the endingless loc. nepi see Neu 1980.
198. ne-e-p-a (OS), ne-p-a (NS).
Noun and Adjective Declension 4.89
00-HittiteGrammar.indb 119 6/5/08 4:15:57 PM
120
back sky axe
195
Numerically indierent
abl. ikiaz npiza,
199

nepiaz(a)
ins. ikiitti' ateit, ateet
Plural
acc. com.? ateu
200
n.-a. neut. ikia atia
gen. nepian
201
199 200 201

nepis sky: end
4.90. Another group shows an -n- extension in the oblique cases: ari(n-) lung,
atalki(n-) hawthorn, addare(n-) intersection, fork, u(wa)lli(n-) pine cone,
202

kuppi(n-) stool, camp-chair, take(n-) assemblage, tunnakki(n-) inner cham-
ber, danduki(n-) humanity, the human race, uppi(n-) what is sent (AHP 15051;
Rieken 1999a: 386404).
203
Later, members of this group were regularized by being
tted into the heteroclitic rln-stem class with new nominative-accusative forms such as
tunna(k)kear (KUB 25.15 rev. 4, 15 ; KBo 30.69 iii 4), etc.
inner chamber
nom.-acc. tunnakki
gen. dunnakkena, [t]unnakkena
d.-l. du[nnake]ni
all. tunnakina, tunnakkina
t-Stem Nouns and Adjectives
4.91. While PIE simple t-stems are rare (Beekes 1995: 178), those in -nt- are not.
Due to the wide use and productivity of a derivational sux -att- used to form action
nouns, the number of t-stem nouns in Hittite is fairly large (Pedersen 1938: 4041; Kro-
nasser 1956: 12931; EHS 25456; Kammenhuber 1969b: 196; Rieken 1999a: 83170).
The only neuter t-stems in Hittite are those in -itt- (see 4.95, p. 122).
4.92. The combination of tld with an immediately following (as in the sg. nom.
com.) is represented by the graph z in Hittite cuneiform (see 1.901.91, p. 37). If an-
199. ne-e-p-i-za KUB 43.23 rev. 15 (OS) and ne-p-i-za KBo 3.22:2 (OS). The alternate interpreta-
tions of the latter as sg. gen. (Otten 1951: 40 and EHS 327) and as endingless loc. + particle -za (see n. 9,
p. 363 and 3.26, p. 74) are unfounded.
200. KBo 29.199:7, cited in HED A 22728 in unpublished form as atteu-ma ZABAR 97lb r. 7.
201. This form may be pl. gen. (see 3.23, p. 73, and CHD LN 448).
202. Attested in the ergative form [u-w]a-al-li-i-a-na-an-za KUB 7.58 i 3.
203. For the stems karzan- and
GI
anzan- included in this class by Rieken (1999a: 39192) see n. 139
(p. 108).
4.90 Noun and Adjective Declension
00-HittiteGrammar.indb 120 6/5/08 4:15:58 PM
121
other consonant immediately preceded this ltsl in nal position, it was necessary to write
it in cuneiform as -za, with an unpronounced a. This situation is especially common with
stems in -nt-, where the nal sequence of three consonants lntsl was written -Vn-za.
4.93. Common-gender t-stem nouns include one PIE root noun, the word for year
(wltt- < `wtt-), and several nouns derived with a sux -t-. Of the latter only kar(a)itt-
ood < `g r(o)i-t- shows partially preserved ablaut (3.37, p. 78). Others have gener-
alized the o-grade: kat- hunger < `g(
h
)os-t- (but e-grade is reected in the deriva-
tive kituwant- hungry), a-t- sleep, bed to e-la- to sleep, and kutt- wall
< g
h
(o)u-t- (see HED K 298).
204
The noun almauitt- throne dais is a loanword
from Hattic. There is also a productive class of nouns with the sux -att-, for which
see 2.31 (p. 57).
Simple t-Stem Nouns
4.94. Paradigms of common-gender t-stem nouns:
205 206 207 208 209 210
anger task, ritual famine year
205
day
206
Singular
nom. com. kardimiyaz,
kartimmiyaza
aniyaz
207
gza, kaza `welizza,
xt.xx-za
to.knm-az,
tu-az, waz'
208
acc. com. kardimiyattan aniyattan katan `welittan,
xt-an
tu-an
gen. kardimiyatta aniyatta `welitta,
xt.xx-a
d
watta, tu-a,
waz'
d.-l. ftxt.ftxt-atti,
kartimmi(y)atti
aniyatti kti wtti, we/itti iwatti, wat,
iwt,
209
to-ti
all. `iwatta
210
204. A derived t-stem nekut- evening is attested only in the set phrase nekuz meur evening
time, where nekuz is judged to be an old genitive < `nek
w
-t-s (Rieken 1999a: 84 with n. 400, following
Schindler).
205. The word for year reects a PIE e-grade, but there is evidence for both e- and i-vocalism in the
Hittite word: -e-et-ta-[(an-da-an-ni)] KBo 3.22:64 (OS) vs. -i-it-ti KUB 29.32:2 and KUB 4.72 iv 2.
Spellings like pl. gen. -elit-ta-an are therefore ambiguous. See on this problem AHP 14445.
206. On watt- see Pedersen 1938: 175; Kammenhuber 1969b: 196, 2045; AHP 102, 131; Rieken
1999a: 1025; Yoshida 2000.
207. Against taking a-ni-at-e-et (OS) as evidence for secondary neuter gender see Rieken (1999a:
107), who emends to a-ni-ya-at-ta-e-et. For the possibility that aniat as attested is a correct collective
plural and an archaism see AHP 87.
208. Interpretation as nominative or genitive is possible (see Neu 1983: 1516 n. 74; Starke 1990: 458
n. 1666, Rieken 1999a: 1023).
209. See also the second member of the OH a-ni-i-wa-at today, henceforth (Rieken 1999a: 104).
210. Abstracted from compound [ap-pa- ]i-wa-at-ta KUB 31.64 iv 1 (MHlMS), Rieken 1999a: 103.
Noun and Adjective Declension 4.94
00-HittiteGrammar.indb 121 6/5/08 4:15:59 PM
122
anger task, ritual famine year
205
day
206
Numerically Indierent
abl. xt.xx-za tu(.xx)-az
ins. ktit
Plural
nom. com. xt.nr.-u
211
tu.xx.nr.-u
acc. com. kardimiyattu,
kartimmiyaddu
mt.t.n-u tu.nr.-u
212
coll. aniyatta,
aniyatti,
aniat (?)
207
gen. welittan,
xt.nr.-a
d.-l. aniyatta xt.xx.nr.-
a
tu.nr.-a,
tu.xx.nr.-a
211 212 213 214 215 216shiwatt paradigm end
ood joy fear wall
Singular
nom. com. karaz, kariz,
gerzza
dugaraz(a) naaraz,
213

naarraz
kuzza
acc. com. tugarattan,
dugaratan,
naarattan kuttan
gen. tukaratta kutta
d.-l. karaitti,
karitti
214
dugaratti,
215

duqarati
216
naaratti kutti
Numerically Indierent
abl. dukaratta(z)za naarataza kuttaz
ins.
Plural
nom. com. garitte, garitti naaratte kutte
acc. com. naaraddu kuddu
d.-l. naaratta kutta
4.95. Neuter t-stems are represented by militt- honey and eppitt- (a cereal). Nei-
ther noun occurs in the plural.
211. MSpr. obv. 6 (Mur. II).
212. KBo 4.8 ii 22 (Mur. II).
213. IBoT 1.36 i 59 (MHlMS) with CHD LN 344. Rieken (1999a: 115) argues against the interpreta-
tion as ablative of *naar-.
214. karaitti KUB 36.110:18 (OS), karit[ti ] KBo 22.6 i 27 (OH?lNS).
215. RS 25.421 rev. 58.
216. KUB 22.42 obv. 6 (NH). See Rieken 1999a: 116 n. 533.
4.95 Noun and Adjective Declension
00-HittiteGrammar.indb 122 6/5/08 4:15:59 PM
123
honey eppit-grain
nom.-acc. milit, tXt-it eppit
erg. iXi-anza
gen. milita eppitta, eppida
d.-l. iXi-ti
ins. iXi-it
nt-Stem Nouns and Adjectives
4.96. Common-gender nouns in -nt- include one root noun, kant- (a cereal, oats?),
one probable old participle, uwant- wind, and several types of stems derived with a
sux -(a)nt- (for these see 2.25 and 2.27, p. 56). For the ergative case in -anzal-ante,
see 3.8 (p. 66).
4.97. Adjectives in -nt- for which no shorter bases can be found in Hittite are mant-
all and am(m)iyant- small. For derived adjectives with a sux -ant- see 2.24 and
2.26 (p. 56) and for those with a sux -want- see 2.50 (p. 61). Hittite participles,
which are also formed with a sux -ant- (25.3925.45, pp. 339340), may sometimes
be substantivized.
4.98. Paradigms of Stems in -nt-:
217 218 humant-
night (noun) all (adj.) living (adj.) captured >
captive
Singular
nom. com. ipanza manza ui(u)wanza,
uuwanza
appanza
acc. com. ipandan mandan u(u)wand/tan,
uiwandan
appantan
n.-a. neut. man
gen. ipanda manda uiwanta appanta
d.-l. ipanti, ipandi manti uiwanti appanti
Numerically Indierent
abl. ipandaz mandaz appantaz
ins. mantit
Plural
nom. com.
217
mante,
mandu
ui(u)wante appnte
acc. com. mandu,
mante
uiwandu appandu
n.-a. neut. mand/ta appanti
218
gen. manda appanta
d.-l. manda appanta
217. Other nt-stem plural nominatives are: rx.nr.-u (*uwandu ) winds KUB 24.1 iv 16 (Mur. II).
218. Perhaps this is rather a collective in -i.
Noun and Adjective Declension 4.98
00-HittiteGrammar.indb 123 6/5/08 4:16:00 PM
124
r/n-Stem Nouns
4.99. PIE heteroclite stems have the peculiarity of taking a sux in the nominative-
accusative dierent from the one used in the rest of the cases. The only widespread type
(the so-called rln stems) has a sux ending in -r in the nominative-accusative and a
sux ending in -n- in the rest of the cases. This class is known in other old IE languages
(Beekes 1995: 187; Sihler 1995: 29092; Szemerenyi 1996: 7.3.4). In Hittite it is a
large and productive class (Hrozny 1915: 2425, 1917: 6465; HE 81; Kammenhuber
1955b = Kammenhuber 1993: 110; EHS 278321; Oettinger 1986b: 1115). All rln-
stem nouns are neuter. No adjectives belong to this class.
4.100. There are ve groups of nouns which exhibit this rln alternation in their de-
clension: (1) words built directly to the root, (2) derivatives in -tar, (3) derivatives in
-ear, (4) derivatives in -war, and (5) derivatives in -mar.
Non-Derived r / n-Stem Nouns
4.101. Group 1: Archaic words (some inherited from PIE) built directly to the root
(often no longer recognizable), with and without ablaut (see 3.37, p. 78), some with
mobile accent (see 1.8, p. 11; 1.46, p. 25; and 1.145, p. 49):
219
uttar word, wtar
water, mur time, ar blood, pauwarlpaur re, lammar hour, kuttar
neck(?), strength, ur urine, pankur (a body part) (see Kammenhuber 1969b: 198
99; Sihler 1995: 298302). As Sihler observes (p. 298), the lexicon involved belongs to
the most basic stratum of vocabularybody parts and words like water, re, blood,
and day. In Hittite, however, the word for day (iwatt-) is not an rln-stem.
220 221
word water time blood re
Singular
n.-a. uttar wtar mur ar, iar,
ar
220
paur,
pauwar
erg. uddananza wetinanza,
witenanza
ananza
gen. uddana,
uttana
witena
221
meuna ian,
ina,
na
pau(e)na
d.-l. uddani,
(uddn)
weteni, witeni muni,
meueni
ani,
iani
pau(e)ni
219. The position of the accent in these forms can be determined by vowel gradation and plene writings.
For details on the ablaut and accent pattern of these stems in PIE see Rieken 1999a: 26970; Beekes 1995:
17476, 187; Szemerenyi 1996: 16162, with diering views and terminology. See also Meier-Brgger
2000: 190201; 2003: 20418 on the general framework.
220. Note that the in this word is omissible (see 4.102). It is not due to a scribal error.
221. The vast majority of weten-, witen-, etc. spellings have e in the second syllable (-te-). Spellings
such as -i-ti-ni-it KBo 23.1 iv 24 are exceedingly rare. The initial syllable wi- or we- is inevitably spelled
with -eli-, once -wi
5
-, but not *u-eli-. On the alternation of initial -i- l -e- in this word see Kammen-
huber 1969: 199 with anterior literature.
4.99 Noun and Adjective Declension
00-HittiteGrammar.indb 124 6/5/08 4:16:01 PM
125
word water time blood re
Numerically Indierent
abl. uddanaz(a),
uddananza
222
wetenaz(a),
wetenazza,
witenaz(a)
elianaz,
naza,
iananza
pau(e)naz(a)
ins. uddanit,
uddanta
223
witanta, wetenit,
witenit, witinit,
wedanda,
widandlta
anta,
ianda,
anit
224
pau(e)nit
Plural
n.-a. uttr,
uddr
225
widr, witr,
wedr, uwitr
226
meurri
I.A 227
erg. uttannte,
uddanante
wetenante
gen. uddana .nr.-a meuna
d.-l. uddan,
uttana
uwitena,
228

wetena
229
meuna
222 223 224

225 226 227 228 229

mehur paradigm end watar paradigm
hour neck(?) urine (body part)
Singular
n.-a. lammar kuttar ur, uwar pankur
gen. lamna euna pankuna
d.-l. lamni, lammar kuttani euni
all. euna
Numerically Indierent
abl. kuttanaz
ins. kuttanit pankunit
222. See 3.33 (p. 77).
223. For the instrumentals in -dlta after a resonant (uddanta, wedanda, ianda) see 3.35 (p. 77)
ud-da-an-ta is attested in KUB 30.10 obv. 18 (OHlMS).
224. Listed in HW
2
2:121a without refs.
225. The form uddrramet my words KBo 3.28 ii 20 (OHlNS) appears to show double marking of the
plural (for the ending -a compare kallra in 4.85, p. 117) and the gemination seen with neuter nominative-ac-
cusative plurals in -i (see 1.109, p. 41). The reality of this isolated form in a late copy is open to question.
226. The non-plene spelling of the second syllable in the plural waters is unusual. -e-ta-ar KBo
12.100 rev. 6 may be Luwian (see the immediate context). But -e-da-ar KBo 25.2 ii 8 is not only Hittite
but Old Script! See also -wi
5
-ta-ar in KUB 13.3 iii 23.
227. This nal -i on the neuter plural is more common in the verbal substantives (arkuwarri
I.A
,
minumarri
I.A
) and in the non-heteroclitic consonantal stems (kurur > kururi
I.A
, uupal > uupalli
I.A
,
kurtal > kurtalli
I.A
, alel > aleli
I.A
) than on rln-stems. But see zankilatarri
I.A
in 3.20 (p. 72). On the
subject of this nal -i on neuter nouns see Prins 1997: 5661.
228. KUB 13.3 iii 37.
229. KUB 39.71 i 24.
Noun and Adjective Declension 4.101
00-HittiteGrammar.indb 125 6/5/08 4:16:02 PM
126
4.102. Note that the in ar is omissible in NH. Friedrich (HE 28 b) explained
this as due to a weakly articulated , which may even have been omitted dialectally.
HW
2
E 121 includes only one example, e-e-ar (in its NH paradigm), and attributes it
to a scribal error. HED E/I 313 acknowledges several forms without and considers
them to be due to a secondary development. AHP 71 indicates that the phenomenon is
pre-Hittite, not independent in Hittite (see also Rieken 1999a: 303). AHP 71 considers
gen. na old, with nom. ar analogical to it. The i in ian- (see also iarwant-,
iakant- bloodied, ianittara- blood relative [see Singer 1999], and iarnumi-
to bloody) is explained as a prothetic vowel that originated in the oblique forms of
the collective plural and spread to the nominative-accusative (Melchert 1984b: 109
n. 66, citing Schindler). The alternation paur /pauen- in the word for re shows
ablaut, but the single example meueni for meuni is merely analogical to paueni. It
is quite uncertain whether the variants pauwar and euwar represent ablaut or not.
Ablaut is denitely shown by the alternation in the word for water between wtar,
widr, wedan-, and the set of variants witen-/weten-/wetin-. The alternation of i and e
in witen-/weten- is explicable as i < *e in the unaccented root syllable (Rieken 1999a:
29293). Of these doublets Rieken claims that the former occurs in the older texts.
4.103. Instrumental in -ta. Note too that in three out of the four r/n-stem nouns
whose instrumental is attested (uddanta, ianda, wedanda), the archaic ending -d/ta
is found either instead of or alongside of the normal ending -it. On the form iimanda
(< iima(n)-) see Oettinger 1982a: 235 2.4. See also 3.35 (p. 77) and p. 185, n. 36.
Derived Nouns with Sux -tar
4.104. Group 2: Derived nouns in -tar (Sturtevant and Hahn 1951: 98 159; EHS
29197) show -tar in the nominative-accusative, where -ar reects a PIE syllabic *r (see
n. 48, p. 78) and -tn- or assimilated -nn- in the other cases. They include many abstracts
or words for states or status ( see 2.33, p. 57) , as well as some concrete nouns (4.105).
Hittite has also borrowed a few nouns from Luwian with a sux -(t)tar/-tn-, such as
uitar wildlife, game and kar(a)ttar segment (Starke 1990: 45556 and 56064).
These inect like the nouns in -tar, except that they never show assimilation of -tn- to
-nn- (see 1.112, p. 42).
papratar paradigm beginning
impurity reverence, respect servitude sorcery
Singular
nom.-acc. paprtar nakkiytar -ntar alwanztar
erg. paprannanza alwanzannaz
gen. papranna alwanzanna
d.-l. papranni nakkiyanni -anni alwanzanni
Numerically Indierent
abl. paprannaz(a),
paprannanza
-nanaz
ins. [nakk]i[y]annit
4.102 Noun and Adjective Declension
127
impurity reverence, respect servitude sorcery
Plural
nom.-acc. paprta alwanzta
growth insolence life journey
Singular
nom.-acc. miyatar, miytar ulltar, ullatar uiwatar laiyatar
gen. miyanna uiwanna
d.-l. ullanni uiwanni laiyanni
Numerically Indierent
abl. ullannaz(a) uiwannaza
ins. uiwannit
Plural
nom.-acc. myata ullta
230
laiyatar
230 4.104 end huitar paradigm
ne wildlife segment, parcel
Singular
nom.-acc. zankilatar itar, uetar karatar, karattar
erg. uitnanza
gen. uitna
d.-l. karaddani
Numerically Indierent
ins. uitnit
Plural
nom.-acc. za(n)kilatar
I.A
,
zankilatarri
HI.A
uitr, uidr
4.105. The plural forms of nouns in this class are especially rare because they al-
most never denote concrete, countable objects, but either states (e.g., nakkiytar) or
classes of animate entities (uitar fauna, game, antutar humanity). Some words
that originally denoted a status developed into designation for concrete instances of the
status. Thus iuniytar deity developed into a term for a cult emblem or image, and
i-ntar (*pintar) manhood > manly deed, exploit > military campaign. In the
meaning military campaigns we nd the plural form i-natar
I.A
KBo 12.38 ii 14
(NH). For the ending -i on plural neuters see 3.20 (p. 71). Forms without nal -r are
archaisms and represent originally distinct pl. nom.-acc. forms (Neu 1982; Melchert
1988b). These again represent concrete instances: e.g., attdalattata wise thoughts,
alwanzta acts of sorcery, paprta acts of impurity, kuduwta instances of slander,
230. []u-ul-la-a-ta 78lw 3 (NS).
Noun and Adjective Declension 4.105
00-HittiteGrammar.indb 127 6/5/08 4:16:03 PM
128
iaarwata lordly acts(?), iyata tamta instances of growth and proliferation (=
things that grow abundantly, see HED ElI sub iyatar), etc.
4.106. Almost all nouns in this class show an assimilation of -tn- to -nn- (see 1.112,
p. 42), the earlier forms of the oblique cases of the -tar nouns, *-atna, *-atni, *-atnaz,
assimilating to -anna, -anni, -annaz. But there are some exceptions, such as sg. d.-l.
aratni to artar. Compare also the derivatives iyatniyant- abundant < iytar growth
and takatniya- to make level < taktar plain.
4.107. Although some verbs form their verbal substantive with the sux -tar
(11.20, p. 185), only their sg. nom.-acc. and sg. gen. are attested. These same verbs
form their innitive in -anna, originally the allative of the verbal substantive in -tar.
Derived Nouns with Sux -ear
4.108. Group 3: Derived nouns in -ear (for examples see 2.34, p. 58; Sturtevant
and Hahn 1951: 9899 160) show -ar in the nominative-accusative, where -ar re-
ects a PIE syllabic `r (see n. 48, p. 78) and -n-(see 4.109) in the other cases. As
with the -tar class, the concrete -ear nouns have overt plural forms: uppear
I.A

gifts, mukiar
I.A
invocations, ilear
I.A
omens, portents,
UZU
n-pear
I.A
limbs,
body parts,
TUG
kurear
I.A
womens headdresses, alkuear
I.A
items provided for
the cult, ur.nr. (annear ) cases, and x.nr. (iear ) portions of beer. Verbs
derived from these nouns employ the oblique stem (see 10.10, p. 176).
231 232

hannessar
paradigm
judgment (a womans
headdress)
impurity thunder
Singular
nom.-acc. annear kurear paprear tetear
erg. anninanza
231
gen. annena,
anniana
kurena tetina
d.-l. a-an-ne-i-ni,
anneni, annaani
kureni papreni
Numerically Indierent
abl. annenaz tetena(n)za
ins. annenit,
annetnit
232
kurenit, kurinit
Plural
nom.-acc. ur.nr. kurear
I.A
231. Also ununenanza (from unu(n)ear) and annananza (from annaar KUB 43.75 obv. 20).
232. This form is a hapax, with no clear explanation. For one possibility see AHP 273, citing op.
4.106 Noun and Adjective Declension
00-HittiteGrammar.indb 128 6/5/08 4:16:04 PM
129
beer gift hole invocation
Singular
nom.-acc. iar uppear attear mukear
erg.
gen. x-ena attena mukena,
[muk]iana
d.-l. uppeni atteni mukeni
Numerically Indierent
abl. x-za antenaz
ins. ie(a)nit
Plural
nom.-acc. x.nr. uppear
I.A
mukiar
I.A
4.109. The regular form of the sux, -earl-en-, is spelled consistently with -e-
in original texts of all periods (Melchert 1984b: 90, 128, 140). Variant spellings with
-i- are to be viewed as indicated in 1.571.62 (pp. 2729). Since many nouns in
-ear are closely associated with verbs, they can also assume the vocalism of the verbal
stem: e.g., annaar judgment (beside the usual annear) after anna- to judge.
There are also a few nouns in -aar built to unknown roots, such as annaar (part of
a building?). Spellings of the oblique stem in -eli-a-nV- are to be read with an empty
vowel as [-essnV-], where -a- is used to indicate explicitly the geminate [-ss-].
4.110. Note also the examples of secondary nasalization (see 1.121, p. 43) in
tetenanza (for *tetenaza) and antenaz (for *attenaz). Nasalized ablatives, e.g.,
tetenanza with thunder, can be confused formally with ergatives such as anninanza
the judgment (see 4.107). Only the context will reveal which form is intended.
Derived Nouns with Sux -war
4.111. Group 4: Stems in -war, denoting concrete physical objects (Sturtevant
and Hahn 1951: 99 161; HE 85). The oblique stem is -aun-, including gen. -auna
([pa-]ar-ta--na-a KUB 36.49 i 8, OS(?) ). All the following are neuters: awar fold,
pen, karwar horn(s), partawar wing, feather. For the plural forms without -r as
archaisms see 1.133 (p. 46) and 4.105 (p. 127).
Singular
nom.-acc. awar partawar karwar
gen. partauna sr-a
d.-l. aauni, aaunai garauni
Numerically Indierent
abl. aaunaz partaunaz sr-az
ins. partaunit
Plural
nom.-acc. aawa(r) partawa(r) sr.nr.-ar
Noun and Adjective Declension 4.111
00-HittiteGrammar.indb 129 6/5/08 4:16:05 PM
130
Derived Nouns with Sux -mar
4.112. Group 5: Nouns with stems in -mar (HE 86, 185), not to be confused with
the verbal substantives in -mar, for which see 11.20 (p. 185) and 11.22 (p. 186).
All neuters: ilammar gate building, tiyammar cord, arnammar yeast, arpamar
(a pastry), *arummar headwaters (of a river). For alaltumar, which is not het-
eroclitic but retains the stem r in inection and shows an i-stem count plural form
(alaltumar), see 4.30 (p. 91).
Singular
nom.-acc. ilammar tiyammar arnammar,
arnamma
233
erg. ilamnanza
gen. ilamna
all. ilamna
d.-l. ilamni
Numerically Indierent
abl. ilamnaz
ins. tiyammanda
Plural
gen. arumna
d.-l. arumna
4.113. The claim that the form ilanna (sg. gen.) in Laws 82 and 89 belongs to
ilammar can only be valid if one believes in a shift of mn > nn in Hittite. See 1.123
(p. 44). Verbs derived from these nouns employ the oblique stem: arnamniya- to cause
a ferment, stir up, excite (< arnammar ferment, yeast).
233
Verbal Substantives
4.114. Most verbs form their verbal substantive with the sux -war (11.20, p. 185),
which sometimes appears as -mar, as described in 11.22 (p. 186). But with very rare
exceptions cited there, only the nom.-acc. and sg. gen. (in -wlma!) are attested. Verbs
with verbal substantive in -war likewise form their innitive in -wanzi, originally the
abl.-ins. of the verbal substantive in -war (see p. 185, n. 36).
Irregular Consonant-Stem Nouns
4.115. The following is the paradigm of the word *pr house with oblique stem
parn-, but also (only) in the sg. gen. a special form periya (Honer 1995c). Note also
the sg. d.-l. pri and abl. `prza (r-erza) built on an endingless locative `pr.
233. The second form (a-ar-na-am-ma from KUB 7.1 i 26) is spelled like a plural (1.133, p. 46), but a
plural of yeast (kindslpieces of yeast?) seems dicult to maintain. This may be another case of the weak
r (on which see 1.132, p. 46).
4.112 Noun and Adjective Declension
00-HittiteGrammar.indb 130 6/5/08 4:16:05 PM
131
Singular Plural
nom.-acc. r-er (*pr ) r-er, r.nr., r.xr
erg. parnanza
gen. parna, periya
234
r.xr-na (*parna)
d.-l. pri (p-e-ri ), parni, r-er (*pr) parna
Numerically Indierent
all. parna
abl. parnaz(a), r-erza (`prza),
235
r-az
234 235

per paradigm house paradim end
4.116. The PIE neuter root noun for heart appears in Hittite with sg. nom.-
acc. X-er (*kr) heart and oblique stem kard- (except for the sg. gen., where
we nd only kardiya).
236 237
Singular Plural
nom.-acc. X-er (`kr)
(UZU)
X.nr.
erg.
gen. kardltiya
d.-l. karti, ker
237
all. karta, X-ta
Numerically Indierent
abl. kartaz
ins. kardit
234. The form p-e-r [i ?-ya-a i-e-e-e ] or p-e-r [a ?-a i-e-e-e ] (the only trace of the disputed
sign is a single horizontal wedge head which could be either the only horizontal in nr or the upper one at
the beginning of n) in KUB 51.56:4 duplicates r-a ui-ti
ME
-:t+ in KUB 29.1 iii 41. The reading p-e-
r [i-ya-a i-e-e ] proposed by Honer (1995c) was objected to by Rieken (1999a: 306 n. 1471) because
the parallel between gen. kardiya built to the oblique stem is not an exact parallel to *priya built to the
nominative-accusative stem. But since the stem pr- is already securely attested in the oblique cases (d.-l.
p-e-ri, r-ri, abl. r-er-za), and because there is no reason from the duplicate KUB 29.1 to expect pri pr [an
. . . ], which Neu suggested to her as an alternative restoration of KUB 51.56, we would prefer to their res-
toration either pr [iya ie] or pr [a ie ] (pending results of collation of the trace).
235. 3.26 (p. 74).
236. It is the appearance of the i only in the genitive (so also in HLuwian) that is noteworthy. The al-
leged sg. gen. karta cited by some scholars (among others, Sihler 1995 and Rieken 1999a: 52 n. 238; see
below), is debatable. Its existence is claimed on the basis of an analysis of the form kar-ta-a-ma in VBoT
58 i 13 as kartama (allative kartama was proposed by HW 103 and favored by Honer [by letter to
Rieken, cited Rieken 1999a: 52 n. 238]). The matter is further complicated by the apparent duplicate KBo
25.107, which reads [. . . . . . a]ppuzzi karta[-. . .] instead of utxt.xr kartama. It is best not to claim a
genitive karta until the reading and interpretation of VBoT 58 i 13 and its apparent duplicate can be re-
solved to everyones satisfaction.
237. An endingless locative in kerti (tta) minuandu KBo 3.21 iii 12, 16, 22, 26 (Eichner 1979: 45 n. 9).
An alternative view is that kerti is itself an aberrant locative form (see CHD LN 291 sub minu- C section
a (citing Neu) and Rieken 1999a: 5253).
Noun and Adjective Declension 4.116
00-HittiteGrammar.indb 131 6/5/08 4:16:06 PM
132
Chapter 5
PERSONAL PRONOUNS
Distinctive Features of Pronominal Stems and Endings
5.1. It is a characteristic of IE languages that the pronominal paradigms exhibit cer-
tain dierences from the nominal and adjectival ones (Kammenhuber 1969b: 207). Pro-
nouns have some distinctive declensional endings and often show more than one stem
form (see Sihler 1995: 38485; Szemerenyi 1996: 203 8.1; and Rieken 2002). This is
particularly true of the independent personal pronouns,
1
where a dierent stem distin-
guishes the nominative from the other cases: nom. *eg h
2
I versus acc. *me me. In
Hittite we see this in k I (nom.) versus amm- me in the other cases, zik you (sg.)
(nom.) versus tu- in the other cases, or w we (nom.) versus anz- us in the other
cases.
2
The clitic forms of the same pronouns may show the non-nominative stem of
the accented form (-mu, -tu) or yet another stem (-na, -ma).
5.2. In the relatives and interrogatives (see chapter 8) the nominative and accusative
often share one stem, while the other cases show a dierent one: kui who (sg. nom.
com.), kuin whom (sg. acc. com.) versus kue(da)- (see 8.2, p. 149). Compare also the
adjective tamai- other (8.10, p. 152), which has the forms tamai (sg. nom. com.), ta-
main (sg. acc. com.) but oblique forms tamel (gen.), tamedani (sg. d.-l.), tamedaz (abl.).
See 9.7 (p. 154) and 9.11 (p. 156) on numbers which employ the typical pronominal
endings -l and -edani.
3
5.3. Pronouns are often strengthened by means of particles: French ce, celui, celui-ci.
Latin ego-met, ns-met (Szemerenyi 1996: 202). The -ila in zikila you yourself, ukila
I myself, and apila he himself is such a particle, conned to the nominative
forms (see 18.7, p. 279), as is the -i on the pronoun ai, uni, eni (see 7.107.16,
pp. 144146).
1. Also called accented personal pronouns (Luraghi 1997a: 24).
2. In Hittite only the you (pl.) accented pronoun shows a single stem um- for all grammatical cases
(5.8).
3. In the case of the demonstrative pronouns ka- this and apa- that, the normal forms of the sg. nom.
and acc. com. (k kn and ap apn) do not share the same stem. The less common late variants kn
and apn were perhaps formed on analogy with kui, kuin. See 7.3 (p. 143), 7.7 (p. 144).
00-HittiteGrammar.indb 132 6/5/08 4:16:07 PM
133
5.4. There are case endings which are peculiar to pronouns (see also 7.4, p. 143).
In the PIE pronominal systems the neuter nominative-accusative singular often had a
nal dental stop (see Kammenhuber 1969b: 309 45, 1; Sihler 1995: 374, 2; Szeme-
renyi 1996: 2045), which matches apt that, kuit what, and the clitics -met my, -tet
your, and -et his. In the sg. nom.-acc. neut. OH had k (this) without the nal den-
tal, but apt (that) with it. The genitive singular and plural use the ending -el (see 7.8,
p. 144): amml my, tul your, apl hislher, anzl our, uml your (pl.), kl this
ones, kul which ones, whose, while in OH and MH the genitive plural uses -enzan
in the accented pronoun (u(m)menzan your [pl.] 5.10, p. 134) and demonstratives
(kinzan and knzan of these, apenzan of those). In late Hittite, -ell- served once as the
basis for a possessive stem apella- belonging to himlherlit: a-p-el-la-az KUB 14.4 iv
23 , an example of the degenitival adjective class which existed already in OH (on which
see Honer 2006, as well as 6.2, p. 137, and 9.61, p. 170). Sometimes insertions ap-
pear between stem and endings: CLuwian ztali- this (very) (CLL 281) with inserted
-t-, and Hittite sg. d.-l. kdani (i.e., ke+dan+i) this, not *ke-i. The ablative, in addition
to the ending -az, also uses simple -z (as in kz and apz). In the instrumental of the de-
monstratives, the ending ldl, written -da (3.35, p. 77; see ape+dan+da), takes the place
of -it. The ending -e of the common-gender nominative plural (k these, ap those)
reects a pronominal ending *-oi inherited from PIE (see Sihler 1995: 386 374, 5). Hit-
tite also shows an ending -e in the pronominal neuter nominative-accusative plural (k
these, ap those, kue (the ones) which). In the personal and the enclitic possessive
pronouns there is no formal dierence between ablative and instrumental (both cases in
-edaz in the accented personal pronoun, both in -elit in the enclitic possessives), whereas
in the demonstratives the endings dier (abl. -ezl-edaz, ins. -edanda).
5.5. True personal pronouns (i.e., those of the rst and second person) show no
grammatical gender (i.e., animate vs. inanimate), presumably because in speech situa-
tions both speaker and addressee were assumed to be living beings, that is, inherently
animate.
5.6. In Hittite the anaphoric demonstrative pronoun ap- the aforementioned (one)
(see chapter 7) also functions as the accented pronoun for the third person (he, she, it,
they). There is no formal dierentiation of ap- when it is used as a personal pronoun
from when it is used as a demonstrative pronoun. For its forms see 7.3 (p. 143).
Accented (Independent) Personal Pronouns
5.7. The accented rst-person and second-person singular pronouns do not have case
endings in the nominative, dative-locative, or accusative. The endings -l, -enzan, and
-edaz of the genitive and ablative are shared with other pronouns (5.4).
5.8. The following is the paradigm of the independent personal pronouns:
Personal Pronouns 5.8
00-HittiteGrammar.indb 133 6/5/08 4:16:08 PM
134
First Person Second Person
Singular
nom. k, ammuk
4
zik, zk
5
acc. ammuk, uk (late) tuk
gen. ammel, amml, tul
d.-l. ammuk, k (late) tuk
abl. ammdaz(a) tudaz(a)
Plural
nom. w, anz (late) ume, um, umme,
6

um(m) (late)
acc. (anza), anz um(m), um(m) (late)
gen. anzel, anzl u(m)menzan,
7
uml (NH)
d.-l. anz um(m), um(m) (late)
abl. anzdaz um(m)daz
4 5 6 7

paradigm of independent pers. prons.
5.9. Alleged forms of the independent personal pronouns with -a extensions (e.g.,
-ga, -uq-qa, zi-ga, zi-iq-qa; see, e.g., HE 96 and Kammenhuber 1969b: 20910,
250) do not exist. All such sequences represent the pronoun + the clitics and or but
(Houwink ten Cate 1973b; Neu 1997; and chapter 29 here).
5.10. Old Hittite (OH) employs as nominatives only the forms k, w, zik, um,
ap, and ap, whereas in MH and NH various dative or accusative forms have come to
be used as nominatives, and in two cases (k and ume) the reverse also has taken place.
In OH ammuk, anz, and um serve only as datives or accusatives. We have already
seen (3.14, p. 69; 3.16, p. 70; 4.2, p. 79; 4.10, p. 83; etc.) that this mutual contami-
nation of nominative and accusative cases in the plural is common among nouns and
adjectives. Since in the rst- and second-person pronouns (ammuk, tuk) the dative and
accusative forms are identical, ume (originally nominative), which rst was extended
to the accusative, soon was extended further to the dative (KUB 26.12 ii 25). Similarly
k is employed as both accusative and dative in the late NH period. The form umenzan
is OH (but not yet in OS), while u(m)mel is so far attested only in NH, modeled on
ammel, tuel, and anzel. On the other hand anzel our is attested already in OS (see n. 8)
4. Post-OH.
5. Rarely this form has a long i, e.g., zi-i-ik KBo 25.112 i 5; ii 12, 13, 19, 20; iii 3, 4, 8, etc. (OS); and
zi-i-g[a-an] KBo 5.9 iii 8 (Mur. II).
6. The forms with geminate m are all post-OS.
7. KUB 12.43:4 (MS?; OS according to Konk.). A rare variant form [u]minzana occurs in KBo 3.27
obv. 15 (OHlNS), see CHD P sub pankur.
5.9 Personal Pronouns
00-HittiteGrammar.indb 134 6/5/08 4:16:10 PM
135
and is therefore not an NH innovation (contra Sommer and Falkenstein 1938: 77 and HE
97b).
8
Plural apenzan their always contrasts with singular apel his, her, its.
Enclitic Personal Pronouns
5.11. More on the unaccentedlclitic personal pronouns can be found in 18.3 (p. 277)
and 30.19 (p. 411).
5.12. In addition to accented (or independent) personal pronouns, the old Anatolian
family of IE languages possessed a set of clitic forms that ll the role of direct and in-
direct objects of the personal pronouns. The Hittite forms are:
First- and second-person forms:
Singular Plural
1st-person acc.-dat. -mu -na
2nd-person acc.-dat. -ta, (-du) -ma
Third-person forms:
Singular Plural
nom. com. -a -e (OHlMH), -at (MHlNH)
acc. com. -an -u (OHlMH), -a (MHlNH)
n.-a. neut. -at -e (OHlMH), -at (MHlNH)
dat. -e (OH), -i (MHlNH) -ma
5.13. These enclitic pronouns can only occur in certain locations in the clause and
in relation to other clitics on the same word. For this see 30.1530.21 (pp. 410412).
For the rules governing occurrence of the enclitic third-person subject pronoun see
18.1318.19 (pp. 280283).
5.14. The enclitic -mu loses its u-vowel in a morpho phonemic change when im-
mediately followed by -ata or -apa (1.72, p. 32, and CHD -mu a). Geminate writings
-Vm-mu rarely occur in OHlNS texts (CHD -mu a). Enclitic -ta is written regularly with
8. The secure attestation of anzel our in the OS Zalpa text (i+++ i.xr tnt-tt+ ma anze[l
xtts].itoi
URU
Kani 30 xtts.utxt 1si ta KBo 22.2 obv. 1213 OS) makes extremely unlikely
the theory of Forrer (1922, 1926: 17) (transliteration of Bo 2423 = 2 BoTU 10 = KBo 3.27), rejected by
Sommer and Falkenstein (1938: 75), revived by Kammenhuber (1969b: 20910) and Neu (1997: 14041),
and rejected with additional evidence by CHD mn 1 a 2, -mi- d 11, pankur 2 a) that there was in OH a rst
plural independent genitive pronoun *ammenzan, formally parallel to umenzan your (pl.). The true ex-
pected rst plural genitive `anzenzan, if it ever existed, was already replaced by anzel in prehistoric times.
Personal Pronouns 5.14
00-HittiteGrammar.indb 135 6/5/08 4:16:11 PM
136
the f sign. There are only a handful of exceptions, in which the u sign is used.
9
The
particle -ta also regularly occasions a geminate writing of a following -kan (written
-tk-kn), unlike the other dative-accusative clitic pronouns ending in a vowel: -mu-kn,
-el-i-kn. The allomorph -tu (usually written with the ut sign, rarely ft)
9
regularly oc-
curs when this pronoun is immediately followed by either the clitic particle -z(a) or -an
(1.70, p. 31). The form -e is the OH form, replaced later by -i. The initial consonant
is written singly following a consonant or logogram but geminate following a vowel in
-ta, -e, and -na, but not in -mu (see above in this paragraph for rare exceptions in NS).
The clitic -ma occurs following vowels, -ama following consonants or logograms.
5.15. The rst- and second-person pronoun forms (-mu, -ta, -na, -ma) can be
either dative or accusative, while the third-person forms -e, -i, and -ma can only be
dative: -mu me (acc.), tolfrom me (dat.), -ta you (acc.), tolfrom you (dat.) (with the
-du allomorph occurring before -za), -na us (acc.), tolfrom us (dat.), -(a)ma you
(pl. acc.), tolfrom you (dat.); whereas: -eli tolfrom him, her, it, -(a)ma tolfrom
them.
9. On the free variation in principle of flu and ftlut see 1.16 (p. 16), 1.85 (p. 35).
5.15 Personal Pronouns
00-HittiteGrammar.indb 136 6/5/08 4:16:11 PM
137
Chapter 6
POSSESSIVE PRONOUNS
6.1. Of the old Indo-European languages of Anatolia, only Hittite and Luwian clearly
attest enclitic possessive pronouns.
1
OH and MH expressed the possessive relationship
by suxing to the noun a possessive pronoun (my, thy, hislherlits, our, etc.).
In later MH and throughout NH this construction was rst supplemented and then re-
placed by an analytical one, using the genitive independent pronouns ammel my, tuel
your, apel his, her, its, anzel our, and umel your (for the functions of which
see 18.4, p. 278): apl watul his sin (MH), ammel aul my greeting (MH), tuel
xtn-i your land (MH). An even later development during the nal century of NH
was the declinable possessive pronoun tue(da)- your, as in: twarankan tuetaza
memiyanaz kuenner did they not kill it (the bull) at your word? KUB 8.48 i 12 (Gilg.),
tueda iyanta pda
URU
akmi
URU
Nerik in your favorite places, akmi (and)
Nerik KUB 36.90 obv. 1617 , nuza k xtn.xtn i.xtn t\+ xt.10.xx ammdaz t-
az tar (a)un I conquered these hostile lands in ten years with my (own) hand KBo
3.4 iv 4546; see Francia 1996b: 21213. In NS texts, logographically written nouns
and their possessive suxes (e.g., utxt:+ my son) do not indicate the form of the
Hittite possessive. If the text is an NS copy of an OH or MH original, it is possible that
the Akkadian possessive sux covers an enclitic possessive pronoun. But if the text is
an NH composition, it probably covers a genitive independent pronoun.
6.2. The rare form apellaz KUB 14.4 iv 23 suggests the existence of a small class
of degenitival adjectives, including this possessive pronoun apella- (and the analo-
gous kulla-whose [cuius], p. 282, n. 21), built to the genitive form apel (see Honer
2006 and 5.4, p. 133, and 9.61, p. 170). Since apez existed to express from it, it
is likely that in the sentence nu xtts.itoi ammuk ux:+ utxt:+ +\+
d
Iara
[urza]kket nunnakan anda ipanzaket nukan ux:+ apellaz n. And the queen
kept [curs]ing me, my wife, and my son before (the goddess) Iara, and kept making
oerings against us, and my wife died from the (actions) of that one (i.e., the queen)
KUB 14.4 iv 2223 the form apellaz, instead of just being a meaningless alternative to
apez (HW
2
sub ap-2 5.4.g durch jene, and de Martino 1998: 38 with n. 141), actu-
ally meant from (that) of her (so Sommer and Falkenstein 1938: 93 n. 1, 1947: 87;
Houwink ten Cate apud Josephson 1967: 1239 n. 18; Kammenhuber 1969b: 214 durch
1. On the enclitic possessive pronoun in Hittite see HE

1069, 24556; Houwink ten Cate 1967;
Kammenhuber 1969b: 21012; Otten and Souek 1969: 7071; Rosenkranz 1978: 7172; Villar 1988;
CHD LN (1989) 22223; Francia 1995, 1996b; Neu 1997. An attempt to identify this category in Luwian
was made by Carruba (1986).
00-HittiteGrammar.indb 137 6/5/08 4:16:12 PM
138
das von ihr, durch ihre Machenschaften). For the geminate l in apella- see 1.109
(p. 41).
6.3. Another method for indicating the possessive relationship was the use of dative
enclitic pronouns of the type -mu, -ta, -na, -i, -ma (see Luraghi 1997a: 23): i.xn.
xrmukan
LU.ME
aranatallu
d
ts:+u o:+ t-i di Is:+u, my lady, put into my
hand my enemies and my enviers StBoT 24 i 5859; see German etwas mir in die
Hand legen. This method is not yet attested in OS manuscripts (Gterbock 1983: 75).
6.4. The following is the paradigm for the enclitic possessive pronoun:
2

34567 8 910
Case my thy his our
3
your (pl.) their
Singular
nom. com. -mi,
-me
4
-ti, -te
5
-i, -e
6
-ummi -mi -mi
acc. com . -man,
-min
-tan, -tin -an, -in
7
-umman,
8

-ummin
9
-man,
-min
-man,
-min
n.-a. neut. -met -tet -et, -it -ummet,
-ummit
-met,
10

-mit
-met,
10

-mit
2. See HE 65 108; Kammenhuber 1969b: 21112.
3. Otten 1959: 180 and 1973: 35; Neu 1974: 6566, 119, 128; and Francia 1995. The identity of
ummi- as the pl. 1 possessive was not yet known to Friedrich (HE), Kammenhuber (1969b), and Rosen-
kranz (1978), who considered it a variant of mi- your (pl.) or their.
4. Not so far attested as *-me-e, but only with broken writings -mi-e and -me-i (see CHD sub
-mi-), both of which could be interpreted as -mi (using HZLs transcriptional method, -mi- and -m-i).
5. Although the more common, and expected, form is -ti-i, the by-form -te-e is attested in nu
warulate-e ammel katta uwaru Let your (sg.) warula appear to me KUB 36.44 iv 4 (OHlMS), nu tuel
warulate-e paigataru And let your warula go VBoT 58 i 1011 (OHlNS), and
d
r.ii-a attate
xtn-e 4 alaldumari tu[kp]at kiritti t[iy]an arzi Your (sg.) father, Enlil, has put the land (namely) the
four corners in your hand alone KUB 31.127 + ABoT 44 + i 2223 (OHlNS). Also ek-za-te-e your net
KBo 3.21 ii 16. Note that all examples are in post-OH copies.
6. The example of -a in KBo 25.72 right col. 16 (OS) is uncertain.
7. The NH -in is found in at-ta-a-i-in (*attanin) KUB 34.19 iv 8 (OHlNS) (see StBoT 9:5657),
r-i-i-in-ma KBo 6.4 iv 6 (NH) (see LH 54 n. 168 and r-si in identical construction in iv 8, 10),
la-at-ti-en- [i'-in'] KUB 17.18 iii 14 (cited CHD latti- 1 c). We follow Eichners analysis of ge-nu-u-i-in
(1979), which therefore does not contain the clitic pronoun (see 15.13, p. 239, and p. 101, n. 109).
8.
d
tft-um-ma-an in KUB 43.53 i 17 (OHlMS) probably preserves without modernization the read-
ing in an OS archetype.
9. The i vocalization of the sg. acc. -ummin (instead of expected *-umman) is a striking lack of
concurrence with the other OS sg. acc. com. forms -man, -tan, -an, and -man, but it must be correct, if
d
iuummin in KBo 3.22 (OS Anitta) is really the noun + possessive our deity and not just a proper name
that happens to sound like the words our deity, and if
d
tftummin our sun (referring to the king) KBo
20.67 + KBo 17.88 iv 17 (OHlMS?) reects an OS originals spelling.
10. Whenever this ending must immediately follow an unassimilated consonant in its nouns ending
(this normally occurs only in the neuter and with complemented adverbs of the type praamet [for peran
amet] and rorn-anamet), since cuneiform writing cannot represent a sequence of three consecutive con-
sonants without some empty vowel, the enclitic possessive ending -mi- will be written -a-mi- or -e-mi-.
6.3 Possessive Pronouns
00-HittiteGrammar.indb 138 6/5/08 4:16:14 PM
139
Case my thy his our your (pl.) their
voc.
11
-mi, -me,
12

-melit
-um[mi ]
13
gen.
14
-ma -ta -a -umma
d.-l.
15
-mi, -mit -ti -i -ummi -mi,
(-umi )
16
-mi
all. -ma
17
-ta
18
-a
19
Numerically Indierent
abl.-ins. -mit -tit, -tet -et, -it -mit -mit
with
adverbs
-met, -mit -tit -et, -it
20
-met,
10

-mit
Plural
nom. com. -mi -teli -e, -i -umme,
-ummi
-me
21
-mi
-me,
22
-mi
acc. com. -mu -tu -u -ummu -mu -mu,
-umu
n.-a. neut. -met, -mit -et, -it -umet -mit,
-met,
-ummit
22
gen. -man
d.-l. -ta -a
11 12 13 14 15 16 17 18 19 20 21 22 23 enclitic paradigm end
6.5. Both the substantive and the enclitic possessive pronoun are declined (for rare
exceptions in NH see 6.9, p. 141) and agree in number, gender, and case: kiarimi
in my hand (lit. in the hand + in my), tuzziyami my army (sg. nom. com.),
kardiyata of thy heart (lit. of the heart + of thy), attii (lit. to the father +
11. On the forms in -elit see 6.11 (p. 141).
12. Attested in at-ti-me O my father KBo 12.70 rev.! 10b (CHD -mi- e 2 a).
13. KUB 33.66 + KBo 40.333 iii 8.
14. There are no cases of a possessive -man my used with a genitive singular in OS and thus no assur-
ance of the reality of such a usage (pace Kammenhuber 1969b: 212, 1969c: 33; Rosenkranz 1978: 7172).
15. Including with adverbs, such as kattimi, kattitti, katti()i and itarniummi, itarnimi. On the
forms in -elit see 6.11 (p. 141).
16. kat-ti-u-mi HKM 57:21 to you (pl.).
17. In pr-nam-ma (= parnamma), see CHD sub * mi-.
18. For allative -ta, see 16.85 (p. 264).
19. Attested in pr-na-a-a to his house (see CHD P sub per, parn-), i-a-a-a into hislher mouth
KBo 3.38 obv. 4, xtn-e-a-a to his land KUB 23.11 iii 22, an-ta-ga-a-a to her loins(?) KUB 36.44 iv
12, and u-u-um-ma-a-[a] . . . [a-a--n]a-a-a to its sty . . . to its pen KBo 26.136 obv. 1819.
20. [p-]e-ra-a-i-it KBo 17.3 i 26 (OS).
21. See a-re-e-me-e (*areme) your colleagues in KBo 6.3 iii 22 = Laws 55 (OHlNS).
22. See
LU.ME
ka-ru-a-le-e-me-e-a and their k.-men in KBo 6.2 iii 14 = Laws 54 (OS).
23. E.g., KBo 3.1+ i 23 (OHlNS), ed. Homann 1984: 1617.
Possessive Pronouns 6.5
00-HittiteGrammar.indb 139 6/5/08 4:16:15 PM
140
to his), itarnimi in your (pl.) midst, areme your (pl. nom.) companions,
aruwanduu her unborn child (lit. her innards, acc. pl.).
24
For more examples of
this kind of agreement see 15.3 (p. 235). Final n in the substantive regularly assimilates
to the rst consonant of the possessive pronoun (see 1.1181.119, p. 43). Compare
the following singular accusative forms: alugatalla(n) tin your envoy, itamana(n)
an his ear, tuzzi(n) man my army, ara(n) an his border, arra(n)an his anus,
pantua(n)an his bladder.
6.6. In OH (OS), the singular and plural nominative-accusative neuter forms show
consistently -et: me-e-ni-im-me-et, a-a-a-me-et, a-a-ku-wa-a-me-et (more than 40
exx.), with only rare exceptions (u-up-pa-a-mi-it StBoT 25 #13 iv 18, i-da-a-lu-u-
mi-it StBoT 26 #151 ii 4). The ablative-instrumental shows regularly -it: i-a-az-mi-it
StBoT 25 #3 i 18 and dupl. StBoT 25 #4 i 13, ka-lu-lu-p-iz-mi-it StBoT 25 #4 iv 30
(plus 6 more exx.), with two exceptions: ka-lu-lu-p-iz-me-e[t] StBoT 25 #3 iv 34 and
[(ki-i-a-ra-az-s)]e-et StBoT 25 #12 ii 12. The use with adverbs (adpositions) is more
complex. With ketkar at the head, which clearly reects a form containing a noun,
we nd only -et (3 exx.), while with pran in front the regular usage is with -it: p-
e-ra-am-mi-it KBo 3.22:79 (OS), [ p]-e-ra-a-i-it StBoT 25 #4 i 26, p-e-ra-az-mi-it
StBoT 25 #4 iii 44 (plus 3 more exx.), vs. two exceptions p-e-ra-a-e-et StBoT 25 #3
i 33 and p-ra-an-te-et KUB 26.35:9 . The use of the split genitive construction (see
16.38, p. 251) and mostly -et suggests that er above, on behaves like ketkar: e-e-
er-e-me-et StBoT 25 #6 ii 14 (plus 5 more exx.) vs. one exception e-e-er-i-it KBo
6.2 iv 47. See on the split genitive construction Garrett 1998. On the overall problem
of -etl-it spellings see Otten and Souek 1969: 7273; Melchert 1984b: 12226; and
Francia 1996b.
6.7. The same clitic possessive case form in -ilet is shared by ablative and instru-
mental (StBoT 8:71, 73), as is shown by the following examples with ablative nouns:
iaz()mit from their mouths KBo 17.1+ i 18 (OS) (StBoT 8:2021, 58, 66),
aruwantazet from its stomach KBo 6.3 iv 28 = Laws 90 (OHlNS), tuggazt
(sic) from his body KUB 17.10 iii 10 (OHlMS) , tuggazit KUB 43.34:11 , zo-
aztet at your right hand, unumazit KBo 3.8 iii 20, r-az()mit from their houses
KBo 30.20 iii 3, kuttanazmit from their shoulders KBo 20.20 + ABoT 35 + KBo
17.36 + KBo 25.54 i 14 (StBoT 25 #54 OS). See Houwink ten Cate 1967. It has been
claimed by Francia (1996b: 210 n. 2) that no example exists of an instrumental use of
a clitic possessive pronoun, suggesting that the ablative-instrumental forms are all
ablative. But see OS ka-lu-lu-p-iz-mi-it StBoT 8 iv 33 with their ngers
25
and [. . . ]x
tu-u-ri-it-ti-it in broken context of KBo 3.17+ iii 23.
24. Compare Heb. pl. yeladha her unborn child(ren) in the Exodus 21:22 assault law.
25. See Melchert 1977: 16668. It is hard to see how kalulupizmit could be anything but an instrumental
form (so correctly Neu, StBoT 26: 171 nn. 51314). If it were ablative, which is possible in the context, one
would surely have ka-lu-lu-pa-az-mi-it.
6.6 Possessive Pronouns
00-HittiteGrammar.indb 140 6/5/08 4:16:16 PM
141
6.8. In the dative-locative the clitic possessive is found on adverbs and postpositions
which themselves have the form of a dative-locative: itarniummi in our midst (= an-
za itarna), kattitti together with you (= tuk katta), kattiummi with us (= anza
katta). See 20.26 (p. 300) with references.
6.9. In the empire period (NH), when the enclitic possessive was passing out of use,
speakers lost the sense of the bipartite nature of constructions like attami and began
to decline only the nal element. Thus sg. acc. attamin (for *attanmin), attaman (for
*attanman) my father, and attatin (for *attantin) thy father (HE

109b; see also
Francia 1996b: 21011).
6.10. In the singular accusative forms the a vocalism (-man, -tan, -an) is earlier (OH
and sporadically MH),
26
the i vocalism (-min, -tin, -in) is NH. But see above, 6.4,
nn. 8 and 9 (p. 138), on OH -ummin our (sg. acc.).
6.11. There is no pl. nom.-acc. neut. *-ma, `-ta, `-a, etc.; the sg. nom.-acc. neut.
serves instead: kuwamet their eyes KBo 17.1 i 24 (OS). Later scribes wrongly as-
sumed that the forms ending in -it or -et were all-purpose and falsely spread them to
other cases such as vocative singular (3.28, p. 74) or dative-locative singular in copies
of older texts. There is no basis for assuming that the use of -ilet with the vocative or
dative-locative was part of Hittite speech in any period (Otten 1973: 55 and Melchert
1977: 25962, against Houwink ten Cate 1967).
6.12. It is not always easy to decide, especially in post-OS copies, whether -u(m)mi-
indicates the plural 3, plural 2, or plural 1. For analysis of examples see Francia 1995,
1996b. Attached to ordinary substantives, the writing -u(m)mi- refers in OS and MS
originals to our but in NS copies can be either their, your (pl.) or our. In OH
itarniummi always means between us, whereas in NH compositions it may mean
between us, between you (pl.), or between them. In the latter case it is due to a
misunderstanding of older texts where between us was meant.
26. [ukma]
m
Tudaliya ipandaz tuzzi(n)man stu-nun KUB 23.11 ii 22 (NS).
Possessive Pronouns 6.12
00-HittiteGrammar.indb 141 6/5/08 4:16:17 PM
142
Chapter 7
DEIXIS: THE DEMONSTRATIVES
7.1. The function of the demonstratives, called deixis, is to indicate position of a
third party or object relative to the deictic center. In ordinary speech the deictic center
is the person speaking. In what is called situational deixis, a speech is reported and
the deictic center is that of the person whose speech is reported, not that of the reporting
speaker. The two most common systems are a two-way contrast based on distance from
the speaker (this = near, that = far) or a three-way contrast based on primary relationship
to speaker and person spoken to: (1) I-deixis (here, near the speaker, proximal), (2) you-
deixis (there, near the one(s) addressed, medial), (3) yonder-deixis (over there, not near
speaker or hearer, distal) (Szemerenyi 1996: 204 with bibliography in n. 1). In the case
of Hittite, Goedegebuure (20023) has demonstrated a three-way deixis: k-, ap-, and
ai (but see 7.17, p. 146, for a possible minor two-way contrast). In face-to-face speech
situations the spoken demonstratives were probably reinforced by gestures of eye and
hand. Sometimes Hittite texts refer to such demonstrative gestures with phrases such
as ror.nr.-it iezzi he makes (a signal) with (his) eyes or +:++ di he extends (lit.
places) the hand. The proximal demonstrative k- this can refer to the object on which
the writing is found or to a person bringing the tablet to the recipient: kmaza ix
[+ui:+]
m
Tudaliya i[t ui-at] q q
1
at
m
xi.o.f[i-a itoi.oi] . . . iyanun
[My father] Tudaliya [did] not [make] this statue; I, uppiluliyama, [Great King,]
. . . made it (and inscribed it with his exploits) KBo 12.38 ii 410 (cf. comments by
Gterbock 1967: 8081; see also KBo 10.2 iii 2122. Since both KBo 12.38 and KBo
10.2 are copies on clay tablets of the two statue inscriptions, the original function of the
k has been lost). See also maanta k tuppianza anda wemiyazzi when this tablet
reaches you HKM 14:35; mnmawa tma nuwai k
LU
+u:+rri pedi i
earu kuitmana uezzi kuitmana apiya rorn-pa uezzi But if not, then let this chari-
oteer (who brings or accompanies this tablet) stay there in place of him (as a hostage to
guarantee safe return), while he is coming (here, to me) and returning there KUB 14.3
ii 7072 (= Sommer 1932: ii 7173, NH).
7.2. In kian in the following (i.e., yet to be stated) manner and apenian in the
preceding (i.e., just stated) manner (7.18, p. 147) we have the cataphoric (as follows)
and anaphoric (as just stated) use of adverbs based upon deictic pronouns. The third
1. Here we have a case of asyndeton, but apparently with improper gemination. See 29.38 end (p. 400),
and 29.47 (p. 402).
00-HittiteGrammar.indb 142 6/5/08 4:16:17 PM
143
type, inianlenian, is attested as both anaphoric (KBo 5.6 iii 16) and cataphoric (KBo
5.6 iv 4) in the same text, probably because in the latter case it builds upon the rec-
ognitional function (see Goedegebuure 20023) of the distal demonstrative ai l uni l
ini, with reference to something well known to both speaker and addressee (in this case
the Hittite king and his addressee, the Egyptian queen). In other occurrences, ilenian
seems more likely to be anaphoric like apenian (KUB 19.30 iv 1112; KBo 4.4 iv 50;
KBo 5.8 i 1617; KBo 3.4 iii 8384all Murs. II).
k- and ap-
7.3. The proximal and medial demonstratives in Hittite are k- this and ap- that;
the Luwian equivalents are za- and ap-. The medial demonstrative, ap-, also serves as
a third-person independent pronoun (5.6, p. 133). When it does so, its singular nomina-
tive shares with the rst- and second-person singular nominative pronouns k and zik
the ability to take the sux -ila (apila himself; see 18.7, p. 279). The following is
the paradigm for the Hittite demonstratives k- this and ap- that:
2 3 4
Singular
Plural
Singular
Plural
nom. com. k k, k,
2
(ku)
2
ap, (apa) ap, ap
2
acc. com kn, (kn) k, k
3
apn, (apn) ap, ap
3
n.-a. neut. k, k
2
k, k
2
apt, (apat) ap
gen. kl kinzan, kda
2
apl apenzan, apda
2
d.-l. kdani, (kti) kda apedani, apdani,
(apt)
apda
Numerically Indierent Numerically Indierent
abl. kez(za)
3
apz(za)
3
ins. kdanta, kidanda, kt
4
apedanda, apt
7.4. Hittite demonstratives, even in OH, show a mixture of nominal and pronominal
endings. Nominal endings are present in the singular nominative common (-) and ac-
cusative (-n) and the plural accusative common (-u). Pronominal endings are evident
in the neuter forms, the singular and plural genitive (-el, -nzan) and the plural nomina-
tive common (-e). Although the declension of the two primary contrasting demonstra-
tives, k- and ap-, shows broad agreement, there are a few striking dierences. In
the singular nominative-accusative neuter OH had k this but apt that. In the PIE
pronominal systems, singular nominative-accusative neuter often had a nal dental stop
2. Only in LH.
3. Only post-OH.
4. Only in the use to mark relative location on this side, a meaning later expressed by the ablative
kz.
Deixis: The Demonstratives 7.4
00-HittiteGrammar.indb 143 6/5/08 4:16:19 PM
144
(see Kammenhuber 1969b: 309 45, 1; Sihler 1995: 374, 2; Szemerenyi 1996: 2045),
which matches apt. The source of k is unclear. Also unclear is the source of the ele-
ments -d- and -an- in the oblique cases. The rare nom.-acc. neut. kni (e.g., KBo 34.142
i 7) resembles ini (see 7.12). It is unclear whether ke-e-ni attested in broken contexts is
a variant matching eni or a faulty spelling for dative-locative ke-e-da-ni (see Neu 1980:
5051, n. 116, and 1997: 15657).
7.5. The ending -e of the common-gender nominative plural reects a pronominal
ending `-oi inherited from PIE (see 5.4, p. 133). In OH (i.e., OS) we nd only com-
mon-gender nominative plural -e and accusative plural -u. The nominative plural forms
in -u represent contamination from the accusative forms. The isolated example keu
represents a conated formation of the old ending -e + the later ending. On the implica-
tions of these features for the dating of Hittite texts see Houwink ten Cate 1970: 14 and
Honer 1972: 17. In the MH (MS) letters from Maat we nd both -u (ap) and -e
(ap, k) for the common-gender accusative plural (Honer forthcoming 148).
7.6. The regular spelling of ap as a-p-e (i.e., a-bi-e) helps to avoid orthographic
confusion with the Akkadogram +-ut, as well as the Hittite (perhaps originally Hurrian)
noun bi ritual pit (see Honer 1967a) normally written a-a-bi.
7.7. The forms kn and apn are isolated and occur in later texts. The antiquity of
the kn and apn forms in Hittite is supported by the archaic pronoun a-i (sg. nom.),
un-i (sg. acc.) (Laroche 1979), which preserves the same vowel contrast.
7.8. The genitive singular shows the same characteristic pronominal genitive ending
-el that marks the independent personal pronouns (5.4, p. 133) and the plural genitive
ending -enzan that marks the independent personal pronoun umenzan (5.7, p. 133,
and 5.10, p. 134).
7.9. For the instrumentals kidanda and apedanda see 3.35 (p. 77).
ai, uni, ini
7.10. There is a third demonstrativeformally similar to Latin is, ea, idbuilt on
the same stem a- (and variants) as the third-person enclitic (anaphoric) pronoun (see
5.12, p. 135) but showing the same vocalic opposition (a u) in the nominative and
singular accusative common as the demonstratives k- and ap-, a dierent singular
neuter ending (ini vs. -t), and having no plural forms. Laroche (1979), following Peder-
sen (1938), claimed that forms which others had attributed to separate paradigms (with
stems ai-, uni-, ileni-; see HW 36, 41, 234) were actuallyat least during the Old Hit-
tite periodmembers of the same paradigm. He was followed in turn by Neu (1979c:
79, 82), who utilized the diachronic aspect as a dating criterion.
7.5 Deixis: The Demonstratives
00-HittiteGrammar.indb 144 6/5/08 4:16:20 PM
145
7.11. The original unied paradigm resembled the following (Goedegebuure 2002
3: 2528):
sg. nom. com. ai
sg. acc. com. uni
5
n.-a. neut. ini
d.-l.
6
edani
abl.
6
ede(z)za, etez(a), edaza
5 6
7.12. The common-gender nominative and accusative forms consist of an inected
stem a- followed by the deictic element i. The inection can be compared to k and
ap (ending: -a) and kn and apn (ending: -un). The source of nominative-accusa-
tive neuter ini (later eni ) and of the stem ed- of the dative-locative and ablative is not
clear.
7.13. Of the forms in the paradigm above that are not attested in OS, all but the
ablative forms edezza, edaza probably existed in OH speech. The form edi has been
previously regarded as a dative-locative predecessor of NS edani. There is no solid
evidence from its occurrences that it is a dative-locative. On the contrary, its OS use in
complete parallel with the old instrumental kt (see 16.93, p. 266) suggests that it too
had the functions later expressed by the ablative forms edezza, edaza, which represent
post-OH creations.
7.14. It appears that by NH even the speakers of Hittite no longer understood how
these unusual forms t into a single paradigm. New forms arose through analogy to the
customary paradigms, assuming that the stems to be inected were ai-, uni-, and eni-. A
new nominative ai was formed from the stem ai-.
7
The many occurrences of ai
8
in
5. Goedegebuure (20023: 28 n. 55) claims a sg. acc. com. a-i in KBo 17.17 (+) KBo 30.30 obv. 67
(OS!). We have less condence in this reading, especially because it is in OS, and would prefer some other
interpretation. It is notable that, in contrast to the NH paradigm with its example of a singular genitive, the
pre-NH paradigm has no example of a genitive form. It is therefore worth considering that ai is the miss-
ing genitive. Granted that the case-forms preceding the nal -i in the pre-NH paradigm are all pronominal,
whereas -a is the nominal genitive case ending. But there is no *eli form to match the pronominal genitive
ending (in kel and apel ), and the Luwian genitive adjective in -ai- resembles it. The passage would make
as good or better sense if ai is understood as a free-standing genitive: O Sungod of the gods, I hereby
give to the gods (the equivalentlsubstitute) of that one. I give Our Sun Labarna, (inasmuch as) [I] give his
equivalent. Here is his equivalent, his living (substitute)! With this translation we would be opting for
Goedegebuures possibility 2a (on p. 65), namely, that ai anticipates and is coreferential with the posses-
sive clitic -an his (equivalent) in a-a-an-da-a-a-an.
6. These forms function as adverbs marking relative location, as indicated by Goedegebuure.
7. Goedegebuure (20023: 5) considers only two new stems to have been created in NH: uni- and eni-,
both based upon the accusative forms in the pre-NH paradigm. She apparently (following HW
2
A 400a
wertlos) considered ai in its one occurrence as an invention of the scribe (20023: 24).
8. For this form in the singular accusative see Goedegebuure 20023: 4 n. 13.
Deixis: The Demonstratives 7.14
00-HittiteGrammar.indb 145 6/5/08 4:16:21 PM
146
NH as singular nominative common are simply carried over from the pre-NH paradigm,
and the 18 occurrences as singular accusative common exhibit a merger of the nomina-
tive and singular accusative common in late NH (Tudaliya IV and following). For the
new stem uni-, the forms uni (sg. nom. com.), unin (sg. acc. com.), uniya (sg. gen.),
and uniu (pl. nom. and acc. com.) are attested. For the new stem eni-, the forms eni
(sg. nom. com.) and eniu (pl. nom. or acc. com.) are attested. Objections to Laroche by
Kammenhuber (in HW
2
A 400) on the basis of an alleged uniu in OH (reference to the
form was not given!) remain to be conrmed. It appears rather that all uniu forms are
found in NH. That correct forms such as ai (sg. nom. com.), uni (sg. acc. com.), eni
and ini (neut.) continue to occur in NH (Murili II), pace Kammen huber, is not surpris-
ing. They constitute no serious objection to Pedersens and Laroches analyses.
7.15. The ablative and dative-locative forms also continue the pre-NH paradigm, not
being alignable with either of the newly created stems.
7.16. This third demonstrative appears to have been distal and has functions com-
monly found in other languages possessing such a third demonstrative (see Goedege-
buure 20023). Goedegebuure found that ai served in four roles typical of demon-
stratives: (1) situational, (2) recognitional, (3) discourse deictic, and (4) tracking or
anaphoric. To it was built an interjectionladverb identical in formation with kalkma
(Honer 20023). As ka consists of an initial component derived from the near de-
monstrative k- plus the topicalizing -al-ma (ka regularly in OH, rebuilt as kma
in post-OH when non-geminating -a was lost and -ma occurred in all environments),
so OH (i.e., OS) ma must be analyzed as a contracted form from ai plus -al-ma, se-
lecting -ma in post-vocalic environment (29.24, p. 395).
9
On this type of contraction,
compare OH nau ma > post-OH nama (again see Honer 20023 for the details and
above in 1.77, p. 33).
anni-
7.17. From the stem anni- (HED A 5155) there is only the singular nominative
common form anni. To this stem belong the adverbs annaz and annian once, long
ago (compare kez and kian). The meaning of annian suggests a far-deictic mean-
ing that for anni-. There is also a very rare stem ani- (or ana-) seen in the compound
a-ni-i-wa-at today KBo 3.45:12 (see Neu 1980: 1516; HW
2
A 81b; HED A 52).
This meager evidence suggests that an original contrast in spatial deixis between anal
i- *this and annali- *that shifted to temporal deixis: aniiwat on this day > today
vs. annaz *at that time > formerly.
9. Of course, there are examples in NH of ai followed by the particle -al-ma which do not show the
contraction. But these are not examples of the adverb any more than every k followed immediately by
-al-ma is an example of the adverb kalkma, or every kui followed by -al-ya is the distributive kuia
each.
7.15 Deixis: The Demonstratives
00-HittiteGrammar.indb 146 6/5/08 4:16:22 PM
147
Adverbs Built to Demonstratives
7.18. In Hittite a number of adverbs are derived from the demonstrative bases, from
one of which in turn an adjective is derived. As a rule it is the stems keli- and apeli- that
occur in the pronominal adverbs, just as the stem kuwa- (versus kui-) occurs in pronomi-
nal adverbs and conjunctions kuwapi, kuwapikki, kuwattan, etc. (Kammenhuber 1969b:
207). Local and temporal adverbs dier in their frame of reference. In local adverbs,
mirror frames of reference exist between speaker and addressee (each has his own k
by me versus apiya by you): see examples KUB 14.3 ii 7172 (= Sommer 1932: ii
7273) (30.37, p. 417), HKM 18 left edge 25 (30.43, p. 418), and especially KUB
21.38 obv. 1112 (30.69, p. 427), whereas in the temporal a single frame is shared
by both: proximal (indeed, immediate) and mediallremote (kinun now versus apiya
then).
10 11 12 13 14
ka-lke- ap-lape-
k here, hither
10
apiya there, thither
kt fromlin this direction (OH only)
11
kz fromlin this direction (post-OH only)
12
apz from that place
kinun now, at this time apiya then, at that time
kian,
13
knian
14
in the following
manner
apinian,
14
apenian, (apnian) in the
preceding manner
kiuwant- (adj.) this kind of a . . . apeniuwant- (adj.) that kind of a . . .
7.19. The demonstratives k- and ap- also occur in kitpantalaz from this moment
on and apit pantalaz from that moment on. While the rst is always written as one
word and the second in its two occurrences as two words, both are probably univerba-
tions with a single accent. This is reected not only in the spelling ki-it vs. ke-e-et (see
1.46, p. 25, and AHP 139), but also in the placement of -pat in apit pantalazpat
(16.94, p. 266, and 28.118, p. 384). See also ketkarza at the head. The combination
of a demonstrative stem with ending -elit and a noun in the ablative parallels the use
of the clitic possessive ending -elit which serves both instrumental and ablative (see
6.7, p. 140). The demonstratives had no separate ablative form when these adverbial
expressions were created.
10. OH variant of k is kni (KBo 22.2 obv. 9). The CLuwian equivalent of kni is zni (CLL 277) built
to the near demonstrative root z-.
11. In the broken context [. . .] kt nni KBo 17.23 obv. 5 Drive (or: he drives) in this direction [. . .]
seems likely.
12. In OH we also nd kt for on this side.
13. For kian in interrogative clauses see 27.3 (p. 348).
14. Kammenhuber (1969b: 207) considers kinian an analogical formation to apenian, but the exis-
tence of older inian < ini and of knian < kni suggests that apinian was already modeled on the other
two in OH and that the vocalization apenian was later and secondary, just like that of eni and enian.
Deixis: The Demonstratives 7.19
00-HittiteGrammar.indb 147 6/5/08 4:16:22 PM
148
7.20. In some derivatives of the nearlmedial demonstrative that serve the same se-
mantic function, the endings are not the same: kinun now (< k-) versus apiya then
(< ap-).
7.21. One adverb built upon k- (e.g., kalkma, on which see Honer 1968b and
see 24.2724.30, pp. 323324) has no counterpart built upon ap- but intead one built
on the distal demonstrative ai (ma, on which see Honer 20023). Since kalkma
is built upon the near (speaker-proximal) demonstrative, it serves appropriately to add
temporal immediacy to the nite verb forms associated with it. The form ma, on the
other hand, being built on the distal (remote) demonstrative has no such function with
regard to the accompanying verb but instead marks spatial, temporal, or disassociative
distance in the thought expressed (Honer 20023).
7.20 Deixis: The Demonstratives
00-HittiteGrammar.indb 148 6/5/08 4:16:23 PM
149
Chapter 8
RELATIVE AND INDEFINITE PRONOUNS
8.1. In Hittite the common form kui- serves both interrogative and relative functions,
while the indenite function is lled by kuiki someone, anyone. Only in combination
with takkulmn if can kui- have an indenite meaning (see Latin squis).
1

The Interrogative and Relative Pronoun kui-
8.2. The following paradigm of kui- who, which serves for both functions. As with
the demonstratives, there was a stem for the nominative and accusative (kui-) vocalized
dierently from that of the oblique cases (kue-) (see 5.2, p. 132). For the syntax of the
interrogative pronouns see chapter 27 and for relative pronouns see chapter 30.
2 3 4 5 6 7
Singular Plural
nom. com. kui kui, ku (kueu
2
)
acc. com. kuin kuiu, kui,
3
(kueu, kui,
4
kue
5
)
n.-a. neut. kuit kue (kuie)
gen. kul
6
*kuenzan
7
d.-l. kuedani kueda
abl. kuz(za)
The Indefnite Pronoun kuiki and the Distributive kuia
8.3. The indenite pronoun some(one), any(one) is kuiki with inected kui- plus
particle -kkil-kka. The distributive each, every is kuia, with inected kui- plus the
1. Possibly also in parkunuimaza t kuit nuza anda imma atkinui KBo 3.1 ii 44 (OHlNS). See
26.7 (p. 342).
2. Attested in copies of Muw. IIs treaty with Alakandu of Wilua (CTH 76).
3. See 3.16 (p. 70).
4. ku-i KUB 32.123 iii 31 (cited in HW 114).
5. nu-mu xtr.nr. ku-e up-p-e-ta na-at ar-a [ar-]ra-an-te-e e-[ir] na-a e-du-un-na i-it The
birds which (kue, could be collective [3.123.13, p. 68]) you sent to me were spoiled (pl. com.), and I
neither ate them (pl. com.) (nor . . . ) AT 125 obv. 1113 (NH), see ibid. 5 (cited HW 114 as Atch 5, 11).
See also in the paradigm of HE 119.
6. Too rare to be added to the table of forms is ku-u-e-el KUB 12.21 i 15 (OHlNS), ed. Honer 1992a.
7. Expected on analogy with sg. apel pl. apenzan. There may be a vestige of this independently un-
attested form in the indenite ethnic adjective kuenzumna- (see 2.45, p. 60).
00-HittiteGrammar.indb 149 6/5/08 4:16:24 PM
150
conjunction -al-ya also, and. See also adverbial kuwapitta everywhere, in each place.
The plural forms (kuiua, kuedaa) always refers to groups, often numbered (kuedaa
+\+ 3? [ . . . ] HKM 90:6), or collectives with 1-\i:i+ or :+r+t (IBoT 3.114 obv. 8
restored from dupl. KUB 58.83 ii 21 ; KUB 32.123 + KBo 29.206 i 7, 11, 25 [vs. the sg.
kuedaniya 1-edani in line 14]; KUB 58.100 ii 6 ). Forms wholly in parentheses in the
table below are rare.
8 910
kuiki kuia
Singular Plural Singular Plural
nom. com. kuiki kuiqa, kuqa kuia
8
acc. com. kuinki kuiuga kuinna kuiua
n.-a. neut. kuitki kuq(q)a, (kuekki ) kuitta
gen. kulka, kuelka,
kuelqa, kuelga,
(kulki)
9
kuella
d.-l. kuedanikki,
(kuedanikka)
kuedaqa kuedaniya kuedaa
Numerically Indierent Numerically Indierent
abl.lins. kuezqa, (kuezzaqa, kuizzaqa)
10
kuezziya
8.4. kuia is not used as a generalizing relative whoever (contra Sommer and
Falkenstein 1938: 59, HE 120 b and 125). Of Sommers proof texts, several are writ-
ten with ungeminated , thus excluding analysis as kuia, and the others with geminate
(e.g., nu utxt.xr kuia kuwatta utne [(paizzi)] KBo 3.67 i 9) are genuine distribu-
tives. One should not confuse kuia each, every(one) showing geminate -- with
kuia (ku-i-a) whoever, where the non-geminating conjunction -al-ma merely rein-
forces the unspecied sense of clause-initial kui (30.59, p. 424).
8.5. The particle of kuiki regularly appears as -(k)ki when the vowel in the imme-
diately preceding syllable is i (kuiki, kuinki, kuitki, kuedanikki) and as -(k)ka in other
environments. On the other hand, some -(k)ki and -(k)ka forms appear as exceptions to
the above rule: gen. kulki, nom.-acc. neut. kuekki, and d.-l. kuedanikka.
8.6. On the word order in constructions with kuelka + head noun see 16.55
(p. 255). For word order of kuiki with negatives see 26.7 (p. 342), 26.14 (p. 343),
and 26.21 (p. 345). On general principles of word order with kuiki see 18.3318.36
(pp. 286287).
8. See 8.4.
9. ku-e-el-ki KBo 39.219:8, KUB 23.68 obv. 15 (MHlNS), KBo 39.219:8; ku-el-ki KBo 9.137 ii 9,
KBo 19.61 iv 2, KUB 13.23:3. Since all of the non-plene examples are at end of line and cramped for space,
it appears that the plene writing was the norm.
10. Written ku-e-ez-za-qa KUB 22.52 obv. 11 and ku-i-iz-za-qa IBoT 3.122:7.
8.4 Relative and Indenite Pronouns
00-HittiteGrammar.indb 150 6/5/08 4:16:24 PM
151
8.7. The generalizing relative concept is expressed by the repetition of the relative
pronoun or adjective, with or without an intervening imma: nukan kui kui tnt-a
anda srs-ri Whatever town is determined within (that radius of the crime site) KBo
6.4 i 1213 (NH), nu antezziya i-a kuit kuit p[eta] tae arnikzi he shall give as
compensation to him whatever the rst man [gave] Laws 28a; nu tnt.uruir.nr. kui
kui [s]+
m.d
st\-
d
t Whatever cities belong to Armatarunta att. iv 7172. Sometimes
with imma but no repetition of the relative: kul imma or-ruwa aalla alil The
ower of the aal of whatever tree KUB 24.14 i 8 .
8.8. kuiki, kuinki, kuitki, kuedanikki, etc. can be translated someonelsomething
or anyonelanything: nau urorn-tt+-ni kuiki peran wati (if) either someone sins
before a deity KUB 1.16 iii 60 (OHlNS); nu arnikdu itoi-imaapa l kuitki then
let (the oender) compensate, but to the king let there be no responsibility (lit., let there
not be anything) KUB 11.1+ iv 21 (OHlNS).
8.9. The following is a chart of correlatives for Hittite similar to those given in tra-
ditional Latin and Greek grammars (see also Laroche 1979).
Proximal
demonstrative
Medial
demonstrative Interrogative Relative
Generalizing
Relative Indef. pron. Distributive
Pronoun k, kn
(com.)
ap, apn
(com.)
kui, kuin
(com.)
kui, kuin
(com.)
kui (imma)
kui
whoever
kuiki, kuinki
someone,
anyone
kuia,
kuinna each
(com.)
Local
adverb
k here,
hither
apiya there,
thither
kuwapi
where?,
whither?
kuwapi(t),
kuedani
pedi inlto
what place,
where,
whither
kuwapi(t)
kuwapi(t)
wherever
kuwapikki,
somewhere,
anywhere;
it
kuwapikki,
nowhere
kuwapitta
in eachl
every place,
everywhere
Causal kedani er,
kz for this
reason
apadda(n)
(er), apda,
apezza for
that reason
kuwat,
kuedani er,
kuit why?
kuit
because
kuedanikki
er for
whatever
reason
Ablatival kz from
here, on this
side
apz from
there, in that
direction
kuz
whence?,
from where?
kuezka
in any
direction, it
kuezka in no
direction
Temporal kinun now apiya then kuwapi
when?
mn,
maan,
kuwapi
when
kuwapi
kuwapi
whenever
kuwapikki
anytime,
it
kuwapikki
never
Modal kian thus,
as follows
apenian
thus, in the
preceding
manner
maan,
mn how?
maan
as
kuwatqa,
manqa
anyway,
anyhow
Relative and Indenite Pronouns 8.9
00-HittiteGrammar.indb 151 6/5/08 4:16:25 PM
152
Other Stems with Partial Pronominal Infection
8.10. The adjective tamai- other (for the stem see 5.2, p. 132), following a pattern
found in other early IE languages (see Sihler 1995: 385 374.2a on this pronominal-
ization), inects partly like a noun in -ai- (see 4.32, p. 92), partly like the pronouns
k-, ap- and kui-. Also partly pronominal is the inection of dapi(ya)- all, entire. The
following shows their inections:
11 12
Singular Plural
nom. com. tamai, tami,
dami, dammai
ta-ma-e-e,
da-ma-e-e,
damau
acc. com. taman, tamin,
damin
dapian tamu, damau dapiu
n.-a. neut. tamai, dami dapi(y)an,
dapin
11
ta-ma-a-i dapida
gen. dammel, taml
12
dapiya dapida
d.-l. damedani
(damtani,
tamtani)
dap dameda dapiya
all. tamatta, tameda
Numerically Indierent
abl. tamedaz(a) dapiza, dapidaz
8.11. The adjective dapiya- entire, a synonym of mant-, rst appears in texts in
NS.
8.12. In some places in the paradigm, an extended stem dapi(y)ant- takes the place
of dapiya-, perhaps under the inuence of its synomym mant-: sg. nom. com. dapi-
yanza, abl. dapiandaza, nom. com. pl. dapiante, pl. nom.-acc. neut. dapianda, pl. d.-l.
dapianda. Nom.-acc. neut. dapi(y)an could belong either to dapi(y)a- or dapi(y)ant-.
8.13. It was thought by some
13
that the adjective dam(m)eli-, which they translated
another kind of (German andersartig, anderer), was derived from the stem tamai-
(perhaps by way of the genitive pronominal form damel). But since the meaning of
dammeli- is rather virgin (land), unworked or uncultivated (land),
14
it is doubtful if it
can be derived from tamai-.
15
11. See 1.76 (p. 32).
12. The sg. gen. tamda KUB 13.2 ii 29 (in parallelism with
d
) cited by HW 207 is a late rarity.
13. Gterbock 1943; HW 207.
14. For sg. nom.-acc. neut dam-me-el and a discussion of this word see LH 20, 17273. But the correct
analysis of dam-me-el in Laws 6 (late parallel) continues to be ignored and the old incorrect one perpetu-
ated (see Bernab and lvarez-Pedrosa 2000: 181: un campo a otro).
15. Gterbock 1964: 153.
8.10 Relative and Indenite Pronouns
00-HittiteGrammar.indb 152 6/10/08 7:41:56 AM
153
Chapter 9
NUMBERS
9.1. Like most Indo-European languages, Hittite possesssed cardinal numbers (one,
two, three), ordinal numbers ( rst, second, third ), derivative nouns, adjectives, and ad-
verbs to express what we call fractions (half, third, quarter ), multiplicatives (x-times)
and distributives (two each, three each), and factitive verbs based upon numerical-de-
rived adjectives (make one, unify, double, triple). The principal prior study of Hittite
numbers is by Eichner (1992); see now also Honer 2007.
9.2. A major obstacle to modern research into Hittite numbers is the fact that the
scribes made extensive use of ciphers. For this reason the pronunciation of most Hittite
numbers is unknown to us. Furthermore, since the number words are almost always
written with ciphers and the complements consist of only one or two signs, it is ex-
tremely dicult to determine what information the complement provides. Does the
complement consist of: (1) The case ending alone? (2) The end of the sux plus case
ending? (3) The entire sux plus case ending? Or (4) the end of the numerical root plus
entire sux plus case ending? For example, since many dierent numbers can have the
complements -ki-i, -li-i, and -i, are we to assume the presence of three suxes here,
or two? Is the complement -i an abbreviated writing of one of the other two longer
complements? These dicult questions and many more like them beset the study of
Hittite numbers.
Form and Declension of the Cardinals
9.3. Of the few number words which are written syllabically we may mention: the
cardinals one ia- (Goedegebuure 2006), three teri-,
1
and four meu-, the ordinals
antezzi(ya)- rst and dn second, and the compounding form d- in dyuga two-
year-old. The number seven may be attested in the beverage name iptamiya-.
9.4. It appears that in PIE only the words for the cardinal numbers one through
four were declined (Watkins 1998: 67; Szemerenyi 1996: 221; Beekes 1995: 21213).
We have conclusive evidence that one through four also were declined in Hittite.
Complements reecting case inection in some numbers above four may reect declen-
sion of higher cardinals but more likely belong to forms with the individualizing sux
-ant- (see 2.25, p. 56).
1. Inferred from the genitive teriya, the ordinal adverb terin thirdly, and the beverage name
teriyalli-.
00-HittiteGrammar.indb 153 6/5/08 4:16:27 PM
154
9.5. The number two follows exclusively the pronominal declension, as does one,
with a single exception, while three and four follow that of the adjectives, again with
a single exception.
One
9.6. Two words for one are presupposed by Indo-Europeanists (see, e.g., Szeme-
renyi 1996: 222 8.5.2): `oinos, stressing singleness or isolation, and `sem, stressing
togetherness or unity. It is possible that in Hittite also there were two words, but some
candidates that are put forward, such as ani- and -, have proven uncertain. The most
promising candidate is ia-, proposed by Goedegebuure (2006).
9.7. *ia- one declines like the pronouns with the exception of the sg. nom.-acc.
neut. 1-an.
2
There are also i-stem forms, whose formal and functional relationship to the
stem *sia- remains unclear. Compare the endings in the center column with the forms
of the demonstrative pronoun ap- to the right.
3 4 5 6 7
Case Writing Ending apa-
nom. com. 1-a, 1-i
3
- ap
acc. com. 1-an, 1-in
4
-n apn
n.-a. neut. 1-an
5
-n
gen. 1-el,
6
1-e-el, i-i-e-el -el apl
d.-l. i-e-da-ni, i-e-ta-ni, 1-e-da-ni -edani apedani
all. 1-e-da, 1-e-et-ta('), i-e-et-ta(')
7
-eda apeda
2. Compare for this pattern Sanskrit ka- one, which likewise shows pronominal inection except for
the sg. nom.-acc. neut. kam.
3. Though far less common than 1-a, nom. 1-i is attested from OS on. In OS it is found in KBo 25.31
ii 9 (Neu, StBoT 26:283 with n. 1), where two palace men escort the r.urorn, one on each side holding
her (the context is restorable from the NS dupl. KBo 25.42 left col. 18 and two parallel passages in KBo
25.46:67 and KBo 23.74 iii 1516). 1-i precedes its noun in HKM 46:46 (MHlMS); follows its noun in
KBo 18.69 rev. 11 and KBo 18.76 rev. 10; see also KBo 5.2 iii 41; and occurs as a predicate adjective in 1-i
kiat KUB 41.23 ii 20. 1-i replaces 1-a KBo 4.9 v 4 in its dupl. KUB 11.29 v 11. The sg. nom. com. form
1-i is distinguishable from the multiplicative 1-i (i.e., 1-ki) only by context.
4. HKM 47:49 (MHlMS).
5. The form read 1-at(-ta) and claimed as a sg. neut. of the numeral by Eichner (1992) is very likely
a personal name
m
Atta (i ) (see nal 1996: 21 n. 57 and Miller 2004: 450 n. 727). If there was a neuter form
*1-at, no trace of it has yet been found in published texts.
6. For 1-ela see 9.61 (p. 170).
7. The analysis of 1-tta as an instrumental with the geminating conjunction -a also does not account
for its meaning in context, contra Melchert (1977: 37678), Eichner (1992: 3940), and Goedegebuure
(2006). It ought to mean into one, an allative function. We tentatively list it here as a variant of the allative
1-eda. For the alternate ending with geminate -tt- compare apadda beside apda and tamatta beside tameda
(see 8.98.10, pp. 151152). For the e-vocalism versus the other two forms in -tta compare perhaps alter-
nations like itta beside iyatta, but further research is needed to validate this explanation.
9.5 Numbers
00-HittiteGrammar.indb 154 6/5/08 4:16:28 PM
155
Case Writing Ending apa-
abl. 1-e-da-azl-za,
8
1-e-ez, i(-i)-e-ez, e-e-za -edaz, -ez apedaz, apez
ins. i-e-et, 1-e-ta-an-da, e-e-et(') -et apet, apedanda
neut. coll. 1-e -e ap
8
9.8. The sg. nom.-acc. neut. form 1-an is used for ordinary neuter nouns: e.g., 1-
an uttar one thing KUB 24.1 ii 3 and apt 1-an that one (thing) KUB 14.3 iii 62.
The formally plural 1-e in KBo 18.172 obv. 16 is used to modify a noun that is a col-
lective plurale tantum: 2
GI
siui\\i X-u+ 1 2 or.x xi.nnnn 1-ema 1 or.x
xi.nnnn two emblems, of which one (islconsists of) two or.x-trees, while one
(consists of) one or.x-tree.
9
Some instances of 1-an probably represent the sg. nom.-
acc. neut. of the derived stem 1-ant-, which expresses the sense of one as a unit,
10
as
in KBo 5.4 rev. 910 (Targ., NH): [numa] \ts urorn-tt+ kuit 1-an +oot\ numa
\ts urorn-tt+ maan 1-an 1-\i:i[+ y]ama ten because I have given you (pl.)
an oath as one, as you have the oath as one, may you too be one (see also KUB 9.31 iii
21). For the equivalence of Akkadographic 1-\i:i+ and Hittite 1-ant- see further 9.27
(p. 160). Likely instances of this usage thus far attested all appear to be predicative.
9.9. The number one is used in reciprocal constructions: 1-a 1-an walzi the one
hits the other, 1-a 1-edani pi the one gives to the other, 1-a 1-el tut-un di the
one takes the others sheep; and in distributives (18.29, p. 285; 19.10, p. 291; 24.12,
p. 320; 28.119, p. 384): 1-a 1-a (or 1-an 1-an) as a unit means one by one; one at a
time or each in his turn (for the nominative see KUB 13.4 iii 56 in 25.29, p. 336);
the accusative 1-an 1-an is found in: nata r rnr.xr arikuwan 1-an 1-an anda
tarnikezzi and the army commander admits the arikuwa-troops one by one VS 28.30
iii 1617 (fest.). Note that the one by one aspect is further indicated by the imperfec-
tive -ke- verb form (see 24.12, p. 320). For the distributive use with other numbers
compare 2-at 2-at (9.12).
9.10. A special distributive value was assigned to 1-a-a in KUB 43.23 rev. 21 by
Eichner (1992: 41), comparing kuia each. But 1-a-a, both in KUB 43.23 rev. 21
and in KUB 13.4 iii 52 , probably means even one.
8. See 1-e-da-az 1-e-da-az on one side . . . on (the other) one side, reciprocal, on a basis of equality,
one to one KUB 21.1+ iii 76, and 1-e -da-az . . . [1-]e-da-az also in correlation in KUB 30.15 obv. 2728.
Uncorrelated 1-e-az KUB 14.1 rev. 25 (MHlMS) is another example of 1-e -da-az. In the context it should
mean together, jointly, as one.
9. There is no reason to doubt this analysis (against Eichner 1992: 33). As Neu (1992) has shown, col-
lective pluralia tantum are not rare in Hittite. Nor is a plurale tantum for a word of this meaning unexpected:
cf. English insignia which for most current speakers has no singular insigne.
10. In the formulation of the Oxford English Dictionary under one III.7: Designating a complex
whole or entity in which a plurality of components or entities are united or put together; united, joined.
Numbers 9.10
00-HittiteGrammar.indb 155 6/5/08 4:16:29 PM
156
Two to Ten
9.11. The declension of two shows the plural endings of the pronominal declension.
Compare in the following table the plural endings of two with those of apa- that:
ap, ap, apel, apda, and apdanda (see 7.3, p. 143).
11 12 13
nom. com. 2-u (post-OH)
11
n.-a. neut. 2-e
12
gen. 2-el
d.-l. 2-eta, 2-a
ins.? 2-itanta
13
9.12. The form 2-el KBo 6.26 iii 35 (Laws 191, OHlNS) is not, as HE 131 claims,
a rare alternate form to the nom. com. 2-u but in that passage is the regular pronominal
genitive of two meaning in the place of the two (women). The phrase in question
translates as but if (the cohabitation of a man with two sisters occurs) in the place of
the two (2-el pedi), so that he knows (that his partners are sisters): see LH 151 n. 536.
14

The writing 2-an might stand for takan together, or for `2-elan, a sg. neut. used ad-
verbially (see a dierent interpretation in Eichner 1992 63). The distributive form 2-at
2-at KBo 20.8 obv. 11 (OS) and KBo 20.83 i 4 (OHlNS), which appears to mean by
twos or two by two, is perhaps an instrumental form with an ending -at, as in kuwat
(see 3.35, p. 77).
9.13. The following case forms of three are attested:
15
nom. com. 3-e-e (i.e., `tere or `terie )
acc. com. 3-u
n.-a. neut. 3-e
15
gen. te-ri-ya-a
Of these forms, only nom.-acc. neut. 3-e shows the pronominal ending.
11. 2-u in some contexts could be merely a shorter writing of 2-elu (9.61, p. 171).
12. 2-e in the construction itoi-u 2-e ekuzi the king drinks two could either refer to two vessels
(o+t-ri) or be adverbial (two times).
13. 2-i-ta-an-ta in KBo 25.35 ii 9 (OS), in view of apedanda, is probably instrumental. The existence
of this ending in OS suggests that apdanda in KUB 26.71 i 7, the NS duplicate of KBo 3.22: 58 (OS), may
go back to an OS archetype. Cf. also 1-e-ta-[a]n-da KBo 22.203 obv.? 1 (OHlNS) in broken context.
14. The example 2-l in the passage [
URU
akpi(amakan
URU
It)]a[(ar)]aa 2-l iparzer But
akpia and Iara togetherlboth escaped KUB 1.1 ii 14 (NH) awaits explanation, but it clearly has the
meaning together, both.
15. See n. 12 above.
9.11 Numbers
00-HittiteGrammar.indb 156 6/5/08 4:16:29 PM
157
9.14. Four has the following forms (Eichner 1992: 75):
nom. com. mi-e-ya-wa-a, mi-e-wa-a
acc. com. mi-e--u, 4-u, 4-a
gen. mi-i--wa-a
d.-l. mi--wa-a
ins. 4-it
abl 4-az
Since four shows no other trace of pronominal endings, dative-locative 4-ta is probably
not a pronominal ending (*-eda) but rather a number extended by the individualizing
-ant- (9.279.29, p. 160).
9.15. The cardinal number eight has the following complementation in the dative-
locative or accusative 8-ta (OS). In spite of the tempting connection of the -t- to the
PIE nal consonant of eight or the use of a pronominal ending -eda here (unattested
in numbers higher than two in Hittite), this is probably another instance of the indi-
vidualizing -ant- with cardinal numbers.
9.16. 9-an seems to be the cardinal number nine. For example, nu mn
D
tft-u
u kued [ani] paitti tuga 9-an pu kui
LU
x.r.xx nutta 1 tut pu And if you,
O Sungod, give goods to someone, let him give you nine (animals), (but) if it is a poor
man, let him give you one sheep VBoT 58 iv 1112. On the dative-locative form 9-anti
see 9.35 (p. 163).
9.17. Hundreds are written with the Akkadogram +i, thousands with tt-t+, and units
of ten thousand with sro
7
. The rst two of these three logograms are always preceded
by a number, even for the unit one: 1 +i for one hundred, 1 tt-t+ for one thou-
sand.
16
Ten thousand can be written either 1 sro
7
(-an)

or just sro
7
(-an). The phonetic
complement -an is used in OS and MS only.
17
Large numbers are written with combina-
tions of these logograms plus their preceding ciphers: 1 sro
7
5 tt-t+ 5 +i x.n.xr
15,500 captives KUB 14.16 iii 20 , and 1 +i si-st 6 166 KUB 42.33:11 (NH) . The
semi-logographic writing of the adverbial form of thousand, tt-t+-ti-li by the thou-
sands KUB 8.67 iv 13, 18 (ed. Siegelova 1971: 4041), need not indicate that the
Hittite number thousand ended in -t- or -ti-.
18
The complementation of sro
7
-an ten
16. If this rule applied also in compounds, it means that the title
LU
toti tt-t+ cannot mean overseer
of a thousand.
17. Without preceding one in the Song of Release: nu sro
7
-an otu.nr.-un atte . . . sro
7
-an otu.
nr.-un atte 3 sro
7
tut tut. xt.nr.-na atte She slaughtered 10,000 oxen . . . 10,000 oxen . . .
30,000 fat-tailed sheep KBo 32.13 ii 1517 (MS); with preceding one in 1 sro
7
-an xt.xx.Xx 10,000
annually KBo 25.123:9 (OS).
18. See Germanic *ths-hundi- thousand.
Numbers 9.17
00-HittiteGrammar.indb 157 6/5/08 4:16:30 PM
158
thousand reveals relatively little about its stem. Judging from the ascending progression
9-an tt-t+ 9-an sro
7
-an 9-an o+fr
19
in KBo 17.88 + KBo 24.116 iii 9, KBo
20.67 iv 23 (= StBoT 37:318 iii 10; 324 iv 23), o+fr was an even higher num-
ber such as hundred thousand.
20
What appears to be two signs o.fr in this MS
copy is undoubtedly a scribal misreading of a single OS sign, which can be labeled
purely descriptively, using the Mesopotamian system of sign names, as sro
7
-essig (see
Honer 2007).
9.18. All examples of the numbers two through nine + tt-t+ (thousand) take
logograms as their head nouns. The head nouns can be written singular or plural. But
the only attestation of Hittite agreement for such a head noun shows it be singular: nu
uni kuin 9 tt-t+ rnr.xr
m
Pitag gatalli uwatet namu zaiya tiyat nan zaiyanun
That 9,000-man army which Pigattalli led joined battle with me, and I defeated it KBo
5.8 iii 2426 (AM 15859).
The Syntax of the Cardinals
9.19. Until recently, research on Hittite numbers has focused on their morphology.
Their syntax, while unclear in many points, is beginning to come into focus (see Honer
2007).
Agreement in Case
9.20. Cardinal numbers were declined like substantives or pronouns, always agree-
ing with their head noun in case: nui +\+ xtn
URU
Kargami
URU
Kargamiapat 1-a
tnt-a t takulait In the country of Car che mish, only Carchemish, the one town,
did not make peace with him KBo 5.6 ii 910 (NH), 1-an so.ut-an KBo 10.13 +
KBo 10.12 ii 41, 1-an uttar one thing KUB 24.1 ii 3 (see also i 16), 1-edani x-ni
KUB 23.102 i 15 , triya tu-a mi-i--wa-a tu-a xsxi-an pndu Let them
travel a journey of three days (or) of four days KUB 43.60 i 1011 (OHlNS), 21-a
MUNUS.ME
zintuiya s+
URU
Kalpaanaila KBo 10.10 iiiliv 7.
Number Agreement with Non-Collectives
9.21. When non-collective head nouns appeared with the numbers larger than one,
the head noun could occur either in the singular or plural.
19. (Let the gods send years to the king and queen)nine thousands, nine ten-thousands, nine hundred-
thousands(?) (of them). Not only are these ascending numerical units very large, but the number nine
itself in Hittite counting system represents the highest number usable for each of the count-units hundred,
thousand, and ten-thousand. Any number above nine requires switching to a higher count-unit (e.g., ten
hundreds = one thousand, etc.). As a consequence, this pattern may suggest that o+fr is one hundred
thousand and not a still higher unit such as one million.
20. The reading sr[o
7
-a]n in both lines for d[a-a]n in the critical edition by (Klinger 1996: 318, 324)
was conrmed by collation S. Koak (7-27-2005). See Honer 2007.
9.18 Numbers
00-HittiteGrammar.indb 158 6/5/08 4:16:31 PM
159
9.22. Old Hittite syllabic evidence shows declined singulars: 5 gpinan ve threads
KBo 17.1 iv 15 (see Melchert 2000: 60 n. 28), 3 ru n three warm breads, 5 otu
tyuga ve two-year-old oxen Laws 5758 (OS), [(5 otu)] audiza ve wean-
ling oxen Laws 57 (OS) (tyuga and audiza are both sg. nom. com., according
to our understanding), 2
NINDA
wagata (OS), 10
NINDA
arin (OS), 30
NINDA
li (OS), 3
NINDA
kitun (OS), 10
DUG
ari (OS).
9.23. We also nd plurals: 2
D
antaepu two gurines of .-deities (OS), te um-
miu 4-u four cups (OS), 3
NINDA
ara (OS), 2
MUEN
partniu (OS), 20 [pur]puru
(OS), 3
GI
palzau (OS), 3 i-i--u three deities (OS), 3
GI
zaluwaniu (OS), 2
i.xr ululiyante two men ghting (each other), 3 puale three p.-garments,
2
GI
zaluwaniu 2 upparu two z.s (and) two .-vessels, 2 tu[n]inga ru arau
two thick loaves (made) of tuninga , 2
KU
annanuzziu two leather harnesses (OS),
8
TUG
iimane neyante eight i.-garments (are) turned, 2
LU.ME
pi [tt]iyandu two fugi-
tives, ror.nr.x+ lk t!+ laplippu kar (a)p turn (hither) your benevolent eyes; lift
(your) thousand eyelashes KBo 7.28 obv. 11 (OHlMS) (and compare possibly 1 sro
7
l [apliputu kar(a)p] lift your ten thousand eyelids KBo 27.18:4).
9.24. Some syllabic forms are ambiguous as to number, such as neuters: 2 gipear
two ells (of eld), 7
GI
attalu seven bolts, 6 arniar, 10
NINDA
tnik ten t.-breads,
2 ppar two .-vessels (OS), 8 wakur (OS), 3
NINDA
wagear (OS), 2 zipattanni (OS),
100
GI
gipear (OS).
9.25. Logographic writings reinforce the clear evidence of the syllabic writings.
Logograms with numbers above 1 can either have or lack the plural markers xr and
nr.: (a) with: 2 i.xr two men, 3 i.xr three men, 30 tut.nr. thirty sheep,
7 urorn.xr seven gods, 9-it
UZU
n.nr.-it (together) with the nine body parts,
(b) without: 7 so.ut seven persons, 5 tut.f ve male sheep, 2 r rx.iXi two
beehives.
Number Agreement with Collectives
9.26. There is a dierent method of indicating the numbers with collective nouns
(Neu 1992). Collective nouns can be counted as well as count plural ones. But with
the former what is counted are either sets or numbers comprising a single set: logo-
graphically one is written 1-\:i+
21
(Akk. itntum), and all numbers above one
are numeral + :+r+t:
22
2 :+r+t, 3 :+r+t, etc.
23
This way of writing the number with
21. Often abbreviated to 1-\i.
22. The abbreviation of Akkad. :+r+t as :+ in inventory texts (Siegelova 1986: 7045) is due to the
nature of those late texts, as seen by the wholesale abbreviating of words in them. With numbers greater
than one, however, the eect is the same: both x-ta (for -anta) and x-:+r+t indicate that sets are being
numbered.
23. This is not to say, however, that :+r+t never occurs in Hittite with the number one. There are quite
a few post-OS examples: KUB 30.32 obv. 19, 22, KBo 10.36 obv. 14, KBo 37.3 + KUB 28.87 rev. 5, etc.
Numbers 9.26
00-HittiteGrammar.indb 159 6/5/08 4:16:33 PM
160
collective nouns (i.e., numbering sets) is found with both logographically and syllabi-
cally spelled counted nouns: 1-\i:t+ l 4 :+r+t
GI
z.ix.on.nr. one tentlfour tents
and 1-\i:t+ widr one portion of water 14 :+r+t eelliya widr fourteen por-
tions of water of purication. Note that the second examplewidr, the formal plural
of the already semantically collective wtarshows that this practice applies to all
collective plurals, not just to those nouns that occur only as collectives.
9.27. As noted by Melchert (2000: 5960), complementation of the numeral, when
counting collective nouns, shows a stem -ant-.
24
This -ant- stem is the Hittite equiva-
lent of Akkadographic writings with 1-\:i+ or numbers greater than one followed
by :+r+t.
9.28. The endings on the -ant- stem agree in gender and case with the counted collec-
tive noun, showing a neuter form either when marking a formally singular but semanti-
cally collective noun (e.g., paur or appear) or when modifying a marked neuter
plural head noun. But the neuter plural ending -anta marks the number of sets, while
the number of components in a single set is marked by an ending -an, which might be
the neuter singular of -ant-.
9.29. Examples of numbers ending in -anta, marking the number of sets or groups:
GI
arpama 1-anta itoi-a olni kitta xtts.itoi-a 1-anta kitta As for wood-
piles, one (set of unspecied number in the set) lies at the foot
25
of the king, and one
(at that) of the queen KBo 17.3 iv 2526 = StBoT 8 iv 2829 (OS). It can be seen that
in this case the head noun arpa has an overt collective ending -a. Hence, it takes its
number with the stem form -anta.
26
The full complement 1-anta claries shorter writ-
ings of the numeral with simple -ta in older texts. The phrase 4-ta :+r+t rzr
4
rft.
xx kuit karan ta KUB 5.7 rev. 30 means that four monthly festivals had been
neglected. (The four festivals are conceived of as four sets of festival activities, not a
single grouplset of four festivals which would have been written *4-an rzr
4
rft.) In
this passage the collective subject (4-ta :+r+t rzr
4
rft.xx) takes the singular verb
ta by the rule of neuter plural subject taking singular verb. The singular linking verb
(ta) in turn caused the predicate adjective karan to be singular.
27
In NH, however,
But since no example of 1 :+r+t is yet atttested in Old Hittite Script, it is possible that it was introduced
later on (false) analogy with the numbers greater than one. See 2-\i:i+ KUB 38.3 i 14, unattested in Ak-
kadian, which shows the same false analogy.
24. This analysis seems correct for many examples of collectives, but it does not account for all in-
stances of numbers with collectives nor for all uses of 1-\:i+ and :+r+t with nouns. See 9.37 (p. 163)
and 9.40 (p. 164).
25. The possessive clitic -i shows that the complete form was singular, *padii.
26. The form m(i)uwaniyante used in the horse-training texts for a team of four chariot horses (Eich-
ner 1992: 3.4.4.1 and 3.4.4.3), if correctly analyzed as a participle (so CHD LN 3089), in spite of the
-ant- sux, would have nothing to do with the collective (a team of four [horses]).
27. This is the simplest accounting. Alternatively, since taking 4-ta as 4-anta makes it redundant with
immediately following :+r+t, one could translate that four monthly festivals for you had been neglected
9.27 Numbers
00-HittiteGrammar.indb 160 6/5/08 4:16:34 PM
161
and especially in administrative texts, where abbreviated writings are common, we can-
not be sure that the f sign is not an abbreviation for :+-r+t.
9.30. As for examples of numbers ending in -an, marking the number of components
in a set, we nd two with the number 7-an and one with 9-an. The shorter ending -an
occurs in 7-an modifying the head noun paur, which shows no overt collective end-
ing, although it is collective in sense: t\+ tu.3.xxma kz 7-an paur kzziya 7-an
paur pariai On the third day I kindle a (single) group of seven res on one side,
and another group of seven res on the other side KBo 11.11 ii 56.
28
Each group of
seven res is regarded as a set. The number 7-an seems to show the singular collective
ending, not only because groups or sets are so clearly visualized, but also because 7-an
can hardly be the simple cardinal. We adhere to the view of Melchert (AHP 181) and
others that nal -an is not the expected outcome of nal syllabic m in *septm seven.
It is likely that the beverage name iptamiya- shows the a before the m because it was
formed early enough to retain the regular internal treatment. The second example is not
contextually as clear but ts the pattern observed elsewhere: ru.otn
4
.nya 7-an
D
ebat pariyannai memikezziya +:++++ as for thick breads, he breaks a set of
seven for ebat, and speaks the same way KBo 11.14 ii 3031 (MHlNS).
9.31. The examples with 9-an are less certain, since the -an complementing nine
could also be the end of the cardinal number itself. For example, nu mn
D
tft-u u
kued [ani] paitti tuga 9-an pu kui
LU
x.r.xx nutta 1 tut pu And if you, O
Sungod, give goods to someone, let him give you a set lgroup of nine (of animals,
= singular neuter uppal implied?). (But) he who is a poor man, let him give you a
single sheep VBoT 58 iv 1112.
29
A second example is:
GI
zupari 2si 9-an kez! 9-an
lukkanzi [k]ezziya 9-an lukkanzi torches: two sets of nine: on this side they light (one)
set of nine, and on the other side they light another set of nine KBo 20.34 rev. 1011
(MHlNS). A third example is: rorn-andama takna
D
tft-i 1 tut or
6
2 tut nnnn
9-an tut-un tekan paddnzi nukan tut.nr. kattanta ippandanzi But afterward for
the Sungoddess of the Netherworld one black sheep, two white sheep, and a group of
nine sheepthey dig (a hole in) the ground and sacrice the sheep down in it KBo
KUB 5.7 rev. 30. The alternate reading 4ta with d.-l. -ta for you (O god) could be defended on the basis
of i+-++ si++ [+\+] urorn-tt+wa rzr
4
rft.xx s+ rft.6.xx karan KBo 14.21 i 8 (NH). This provi-
sional solution, however, requires assuming that the graphic nexus between the numeral and the logogram
:+r+t (which corresponds only to a stem on the declined number) could be broken by the enclitic pronoun,
which in speech would follow the stem and ending. Both solutions have their diculties.
28. Eichner (1992: 8384) cited 7-an in KBo 11.11 ii 56 and saw that the -an could not be a reex
of PIE septm . He thought it could have contained an -nt- stem meaning seven at a time, but he does
not elaborate this idea. For such an -nt- -times one could invoke the variant ma-i-ya-an-te for maiyanki
however many times, but see CHD LN s.v., where possibilities of paleographic confusion (TE for in-
tended KI) are raised.
29. Here the symbolic meaning of nine in the sense of an unlimited number may be present; see p. 167,
n. 44.
Numbers 9.31
00-HittiteGrammar.indb 161 6/5/08 4:16:35 PM
162
11.10 ii 1719. Note here how the Y number (see the following paragraph) 9-an and
the singular (collective) head noun tut-un point to a closed group of nine sheep sepa-
rate from the enumerated single black sheep and two white ones.
30
9.32. Since only in the nominative-accusative and dative-locative cases can the dis-
tinction between the singular and plural forms of the set-numbering (collective) num-
bers be distinguished, the following table summarizes how these pattern, using either the
number of sets (X) or the number in each set (Y) or both numbers. In the rst two
rows of the table the number represented by X, if it has any Hittite complement, takes
-anta (even if its number is one!), while Y, if it has a Hittite complement, takes -an.
Numbering SetslGroups in Nominative-Accusative
X = number of sets, Y = number in each set
1 X set(s), each set of unspecied number
(Y) of sheep [Y is not expressed]
X-anta (X :+r+t or 1-\i:t+ ) tut
2 One set consisting of Y sheep [X, which
is always understood as one, is not
expressed]
Y-an tut
3 X set(s) of Y sheep [both X and Y
expressed]
Xsi (or X-anki) Y tut
9.33. In row 3 neither X nor Y show collective suxes. X-anki is the regular for-
mation for X-times, while Y would be the normal cardinal number. The third row is
included here not to illustrate use of the collective suxes but to show how the Hittites
indicated both the number of sets and the number of items in each. For this purpose they
did not use the collective sux on either number.
9.34. In the oblique cases no distinctive collective ending shows on the head noun,
but when the sense is collective, the stem of the number is enlarged by -ant-.
9.35. The following two examples with oblique case endings represent numbers of
components in a single set (row 2 in the table above). For an instrumental form see 10-
antit: [ . . . t]s:i 2 +:tx+ 10-antit kalulup[ittitta . . . ] with your two hands, [and]
with your (set of) ten ngers KBo 17.32 obv. 12. In the latter example it is not ten sets
of ngers, but ten ngers comprising a single set, which calls for the collective form of
the number (i.e., with the -ant- sux). Of course, the ablative and instrumental cases
do not show distinctive singular and plural endings, but the semantics here favors the
30. For additional examples see KBo 11.72 iii 28; KBo 13.115:58.
9.32 Numbers
00-HittiteGrammar.indb 162 6/5/08 4:16:37 PM
163
singular meaning, i.e., indicating the number of items in the set.
31
The dative-locative
collective 9-anti appeni KBo 21.14 obv. 13 means on althe set of nine members.
9.36. When more than one set of a specied number is to be indicated, it is done
by NUMBER+-si preceding the number (row 3 in the table and see 9.38 below):
GI
zuppari 2si 9-ankez! 9-an lukkanzi [k]ezziya 9-an lukkanzi torches: two sets of
nineon one side they light a set of nine, and on the other side they light a set of nine.
Since in this case the number is nine, we cannot tell whether the -an ending is the nal
part of the simple cardinal or whether the number is in its singular collective form.
9.37. One also nds numbers without phonetic complement with nouns that appear
to be collectives, both logograms and syllabically written words: 2
GI
x+\\i+ alongside
2 :+r+t
GI
x+\\i+ two rackslstands, 1 galgalturi beside 1-\:t+ galgalturi one (set
of) cymbals. It is uncertain whether such writings are abbreviated for *1-anta, *2-anta
or reect a genuine alternative use of the simple cardinal numbers with collectives. For
evidence of the latter see the example of 1-e cited in 9.8 (p. 155).
32
But the alterna-
tion with :+r+t in the rst instance and 1-\i:t+ in the second show that these are X
numbers, numbering the sets, not indicating how many members were in each set. When
the numbers lack complementation, it is only the context that can show whether they
are X or Y numbers.
Counting Non-Decimal Sets
9.38. The PIE number nomenclature implies a ten-base system of counting. But just
as in some IE languages other groupings can occur ( four-score, two dozen),
33
so also
in Hittite larger numbers were sometimes expressed in multiples other than tens or
hundreds. The multiplier is regularly written with the Akkadian sux -si, while the
multiplied number has either no complement or rarely the regular complement -an (see
9-an and 9.36 above). Note, however, that while the multiplicand varies (the favored
numbers being seven and nine), the multiplier does not exceed four and rarely exceeds
two, and the multiplicand is always larger than the multiplier. Only with nine, the high-
est of the multiplicands, is a multiplier larger than 2si used. A rare exception to this is
the expression 8-anki 8 eight times eight KBo 53.47 obv. 13. Whether the regular
writing with -si for the multiplier in this construction always stands for the adverb in
-anki is quite uncertain. Of the syllabically written numbered items used in these expres-
sions only kappin is unequivocally singular, while kituna, makita, and arma are
unambiguously plural.
31. The form 10-anza in 10-an-za annan[ei-. . .] a set of ten annan[ei-s] KBo 24.11 rev. 10 could
be the sg. nom. com. form of the sux.
32. This use might particularly be expected in the later language; see Melchert 2000: 65 n. 36.
33. Buck 1949: 936 #13.31.
Numbers 9.38
00-HittiteGrammar.indb 163 6/5/08 4:16:38 PM
164
2si 4 = 8
NINDA
wagear
2si 7 = 14 tri
I.A
, kappin, paila,
GI
o+o
2-i 8-ta = 16 kituna, makita
2si 9 = 18 alakar, ru.txnr.nr., tut.nr.
2si 9-an = 18
GI
zupari
3si 9 = 27 rxr.xr, ru.xtn
4
.n
4si 9 = 36 ru.rnr.xr
5.f.Xx 10 = 50
NINDA
arma
Sets of Two (Pairs)
9.39. When numbers of pairs of objects or persons were expressed, scribes wrote
the number + yugan yoke, pair (or logographically as number + ,t+ot) + the noun in
the plural: 9 yugan
LU.ME
unepi
LU.ME
alliyare nine pairs of -men (and) singers; 2
,t+ot otu.nr. tu-ri-ya-an-zi they shall hitch up two pairslteams of oxen Laws 166;
20 ,t+oi r.xtn.n.nr.wakan [k]ar par neun I have already dispatched
20 teams of (chariot) horses HKM 19:2122 (MHlMS). Some words introduced by
1-\:i+ or number + :+r+t have meanings where pair or team (of two) would
be suitable: ntn.nr earrings,
KU
r.srn.nr. shoes,
(GI)
uupal clapper (or similar
concussive musical instrument), otu.nr. oxen. In these cases it is possible that some
Hittite expression such as yugan pair, yoke, team underlies 1-\i:t+ or X :+r+t, rather
than a number in -ant-. But one should not exclude the possibility that, as in Akkadian,
pair was simply conceived as a set (consisting of two) (see 9.26, p. 159).
9.40. In many languages there are nouns referring to objects consisting of multiple
parts that inect only as plurals (the plurale tantum of the classical languages; see
Latin aeds temple or English scissors, pants). In the case of the English examples
cited, one always sayslwrites one pair of scissorslpants. Many such nouns in Hittite
clearly are treated as collective plurals and are counted as described above in 9.26
(p. 159) and following. On the other hand, there are also examples that appear to be
inected only as count plurals. These also are attested with 1-\:i+ and :+r+t:
34
1-
\:i+ mannini one necklace, [x :]+r+t manniniu x necklaces.
35
The contrast
with 1-\:i+ uurtalla, also necklace of some kind, shows that we cannot easily
predict whether the Hittites viewed a particular composite object as a collective or set of
discrete parts. Thus we do not yet know, for many examples attested as logograms with
1-\:i+ and :+r+t, whether we should assume collective or count plurals.
34. Although we have no evidence that the Hittite equivalent in this case was the number suxed with
-ant-, this seems likely.
35. The formal singular mannini merely reects the fact that Hittite can use singular or plural with any
number (9.219.23, pp. 158159) and does not argue against the idea that necklace is a noun viewed as
a unit (or set) consisting of many parts.
9.39 Numbers
00-HittiteGrammar.indb 164 6/5/08 4:16:39 PM
165
Word Order in Counting
9.41. Numbers precede their head nouns. Exceptions are appositional: nata
URU
attuapat tnt-ria 1-a ta attua the city alone (lit., as one) remained (loyal)
KBo 10.2 i 26 (OHlNS); [
m
Ku] panta-
d
ixx-amakan i
URU
Arzauwa 1-a wai
Kupanta-Kuruntiya, the man of Arzawa, ed alone KUB 23.21 obv. 3132 (CTH 143);
numukan
m
Pittaggatallipat 1-a iparzata KBo 5.8 iii 3132; to see that these are
really appositional see the unabbreviated form in [(
m
Tapalazunauli ma k)]an 1-a so.
ut-a iparzata KBo 3.4 ii 77; and
m
Mammalimakan 1-a s[o.ut-a iparzata]
Mammali as the sole per[son escaped] KBo 40.6 + KBo 14.7 i 6; nui +\+ xtn
URU
Kargami
URU
Kargamiapat 1-a tnt-a t takulait KBo 5.6 ii 910; nuza ammuk
1-a t\+ r:+ 15,500 x.n.xr uwatenun And I alone led o 15,500 captives to my
house KUB 14.16 iii 2021 (AM); nu zik 1-a nni You drive here alone! KUB 21.1
+ KUB 19.6+ ii 69 . It is also possible that 1-a here stands for iela alone, single
(9.61, p. 170). Note that all examples above are the number one.
Fractions
9.42. The sign x (HZL #20, which may also be transcribed as ) serves to rep-
resent the noun half, to which if needed the plural marker nr. (x.nr. or .nr.
halves) or the Akkadian singular accusative phonetic complement -++ (x-++ or
-++ ) can be appended.
9.43. The underlying Akkadian noun is milu(m) half. No syllabic writing of the
Hittite word underlying x has yet been identied. All that can be determined at pres-
ent is that it was neuter: kuit x-++ arzi nat KUB 20.99 ii 810 .
36
Although the
cuneiform writing system contained words and signs
37
for other fractions,
38
they are
so far unattested in Hittite texts.
9.44. Attributive-adjective forms of the fractions may be found in the quantitative
notations used following baked goods, which show the endings -i, -li, and -li-e. Eich-
ner (1992: 89) plausibly suggests that these are fractional notations on the order of a
36. One example shows neuter kuit but nan resuming: nu zo-ni-it [+-:t-si] ku-it -++ ar-zi na-
an-a-an
D
ts:+u-a i[-ta-na-n]i zo-az da-a-i oin-l[i-i]t-ta +-:t-si ku-it -+[+ ar-zi] na-an-a-an i[-
t]a-na-ni oin-la-az d [a-a-i ] KuT 53 ii 4346 (ed. Wilhelm 2002a: 34243). The same text just quoted for
common-gender agreement has another example where assimilation of the sequence n > shows a sur-
vival of OH phonology in an MHlMS text: nu [z]o-ni[t +:isi] ku-it -++ ar-zi na-a-a-an
D
rx-n[a-a
i-ta-na-ni] z[o]-az da-a-i oin-li-it-ta +:isi k[u-it] -++ ar-zi na-a-a-an
D
rx-na-a i-ta[-na-ni]
oin-la-az da-a-i KuT 53 ii 1519. Although KuT 53 is an MH composition, not a copy of an older text, the
unexpected resumption with -an (which contradicts the kuit!) may be due to a false resolution of a dictated
na-a-a-an. The assimilated form na-a-a-an dictated to the scribe in two parts of the text would actually
be from na-at-a-an (see 1.111, p. 41). But confused by two possible sources (natan and nanan), he
resolved one instance falsely as na-an-a-an.
37. E.g., nr two-thirds, xrotsri ve-sixths.
38. For the Akkadian see GAG 70 fm.
Numbers 9.44
00-HittiteGrammar.indb 165 6/5/08 4:16:40 PM
166
tenth, twentieth, or fortieth of some unnamed unit of our measure. There also
exists in MH a form 10-ti-li-i modifying 450 ru.rnr.xr (KUB 31.79:6, 17 ). This
is undoubtedly the same sux. If the dental is not part of the fractional sux, it might
reect the t in PIE *dekmt- ten or an ordinal sux *-to-.
Ordinal Numbers
9.45. Only the following primary ordinals are attested in syllabic writings: antez-
zi( ya)- rst (declinable), dn second (indeclinable). Second is also written Ak-
kadographically as 2-\i-i (i.e., Akk. an) or with cipher and Hittite complement as
2-an. The Hittite tldamai- (an extension of the stem d- two) can serve as a declinable
ordinal second (Eichner 1992: 54, 57) but often means only another.
39
9.46. The form ma is not the adverbial ordinal rstly, as used to be thought (HW
and HW
2
sub voce; Eichner 1992: 4344).
40
On the other hand, dn means second
(adj.) or secondly, thereafter (adv.). It is not also a multiplicative (two-times, double,
twice).
9.47. The question arises: were the ordinals greater than second declined? There
are a few examples of ordinals in series, attested up to ten, all ending in -an-na.
41
But
contrary to the claims of Sommer and Friedrich (e.g., HE

133 b), there is no Hittite
ordinal sux -anna. The alleged examples all consist of the ending -an and the gemi-
nating conjunction -a also (attested in a series: . . . also a second (time), and also a
third (time), etc.) (Eichner 1992: 62, 67, crediting Hrozny for this analysis). But in the
attested examples, these ordinal numbers all modify the common-gender noun zaain
dream KUB 17.1 ii 411 or ubruin KBo 5.2 ii 5761, iii 112. There are no ex-
amples with neuter head nouns. This being the case, nothing proves that the nal n is
part of an ordinal stem or that the ordinals greater than second did not decline.
42
9.48. In a passage from the Song of Kumarbi, 3-annatta armaun nakkit
D
Tamit
and thirdly I have impregnated you with the noble god Tamiu KUB 33.120 i 33 , the
form 3-an must be either a singular neuter used adverbally (thirdly) or be indeclinable.
The only other optionthat it refers to the third deitywould require that it agree with
the instrumental noun nakkit
D
Tamit (which it patently does not).
9.49. Other numbers with -an complementation are probably not ordinals but cardi-
nals; some if not all are of the collective type (see 9.30, p. 161).
39. Eichner 1992: 59 also derives takiya . . . takiya to one . . . to another from this stem.
40. On the adverblinterjection ma see 7.16 (p. 146). ma provides no evidence for a stem - one
in Hittite.
41. Since in KBo 5.2 ii 57iii 13 the series antezzin . . . 2-anna . . . 3-anna . . . 4-in . . . 5-na . . .
6-na . . . 7-na . . . 8-na . . . 9-na . . . 10-na continues with 11ma (intending 11-anma) . . . 12ma . . .
13-ma . . . 14-ma, one could say that ordinal forms are attested even higher. But 10-na is the highest ordinal
attested with the explicit complement -an.
42. Another example of agreement with a singular accusative common head noun is 3-an t-an a third
hand KBo 9.79:6.
9.45 Numbers
00-HittiteGrammar.indb 166 6/5/08 4:16:41 PM
167
9.50. The forms third teriyan (also written 3-an), fourth (4-an), and fth (5-an)
also existed in syncopated forms terin 4-in 5-in third, fourth, fth KBo 11.14 ii 15:
43

ziqqa
D
tft-u t nu terin 4-in 5-in wllui triya l dariyantan triyama wariyantan
But you, O Sungod, go! And hitch (them) up in the meadowa third, a fourth, a fth
(time): unhitch the worn out one, and hitch up the fresh one! The appearance of -in
with four and ve indicates that these ordinals also ended in -i(y)an, the i not being a
part of the root of these numbers. This needs explanation. Either analogy with teri( ya)n
is at work, or one has to posit an initial ilya to the ordinal sux (see 9.59, p. 170; and
iptamiya). The form of the ordinal numbers here is singular nominative-accusative used
adverbially. Thirdly, etc., has the force of a third time. They are probably not accu-
sative common-gender forms modifying the animal in question, because an unhitched
animal, once given rest, would be re-hitched up, as the immediately following context
indicates.
9.51. Logographic writings of time expressions such as tu.(numeral).xx, rft. (nu-
meral).xx, and xt.(numeral).xx, and of one non-temporal word utn.(numeral).xx
have an ambiguity that only the context can resolve: they can reect either cardinal or
ordinal numbers. Thus tu.1.xx in one context may mean one day and in another day
number one. Similarly utn.2.xx can be either two tablets or tablet number two
(= the second tablet). The original meaning of .xx in the above constructions was the
Sumerian genitive sux -(a)k + the enclitic copula -am it is. This fused in writing as
the single sign xx. But Hittite scribes, ignorant of the real meaning of xx in such
constructions, abstracted from tu.1.xx rst day or day number one the discontinu-
ous elements tu and xx and created a new logogram tu.xx day. By the same
process they generated xt.xx year and rft.xx month, but not *utn.xx tab-
let. A second indication of their lack of understanding of the (ot).(txrni).xx
construction was their generating by analogy with utn.(number).xx a writing :ir-
(number)-r x tablets or tablet number x. Scribes also abstracted from the prexed
and suxed determinatives of city names (e.g.,
URU
n+-+:-:t
KI
KBo 3.45:3) a logogram
tnt.xr city.
Ordinals in -t-
9.52. The only candidates for ordinals in oblique cases with Hittite complementation
seem to show a stem containing -t-: 9 xt.nr.-a kappwanta
D
Alalu -i itoi-u
ta 9-tian xt-ti
D
Alaluwi
D
Anu menaanda zaain [pa]i For a mere nine years
44

43. Noted as syncopated ordinals by Eichner (1992: 6768), but with no comment on neuter gender.
44. The nine years of these heavenly reigns could be understood as expressing innumerable years,
since the number nine is used in its symbolic sense of an incalculably large number (see Honer 2007:
377). See p. 161, n. 29. But it is not excluded in this one case, especially with the use of kappuwant- count-
able, few (cf. kappuwanta tu.xx.nr.-a in just a few days KBo 10.2 ii 17) and the asyndetically
introduced following clause in the ninth year . . ., that the nine years are not ages but actual years and are
meant to indicate truncated reigns. See Honer 1998b: 42 (for a mere nine years).
Numbers 9.52
00-HittiteGrammar.indb 167 6/5/08 4:16:42 PM
168
Alalu was king in heaven; in the ninth year Anu gave battle against Alalu KUB 33.120
i 1213 (Song of Kumarbi, ed. Garca Trabazo 2002: 16263 with n. 31, compare i 18
(9-tian xt-ti). See also: [(9-an-t)]i xsxi-ima but on the ninth time KBo 10.45
ii 37 with dupl. This occurrence might show that the sux was -ant-, but we cannot
exclude the possibility that the -an- portion is the end of the number nine and the -ti
the regular dative-locative singular termination of a -ta- stem (perhaps an inherited
*-to-
45
).
9.53. The dental suxwhether it is -ant- or -t (a)- serves here to dierentiate
the nal member of a numbered sequence from its predecessors. The ninth year is set
o from the rst eight, as is the ninth time the celebrant speaks, the conjunction -al-ma
marking the contrast. This is the expected function of an ordinal sux in an older
Indo-European language.
46
If, on the other hand, the sux is -ant-, it would serve an
individualizing or demarcating function analogous to its use with nouns for seasons of
the year: as ameanti (in a particular spring) diers from amesi (in springtime
= any spring).
47
In either case the function is that of an ordinal, and we classify these
examples as such. For another view see Eichner 1992: 86.
Multiplicatives
9.54. As multiplicative adverbs in the meaning x-number of times (for both tem-
poral repetition and purely mathematical multiplication) we nd numbers with phonetic
complements in -anki
48
or -kil-i
49
or written with the Akkadographic sux -si
(GAG 71a). In the absence of contrary evidence we suggest that the Hittite sux was
-anki().
50
On possible adjectives in -(l)i modifying bread names see 9.58.
9.55. Repetition focusing on the temporal aspect. txnrn+-i ipanti seems to be
equivalent to txnrn+-si ipanti (txnrn-times). This is particularly clear where
the two writings alternate in immediate context: nu 3si +:++++ memai [t\+] 3
xsxi-\tma 1-i kian memai HelShe speaks three times in the same way, but the
45. For PIE *-th
2
o- as an ordinal sux see Meier-Brgger 2000: 219; 2003: 236.
46. See the description of the ordinal in PIE: Den Ordinalzahlen kam im Uridg. . . . die Aufgabe zu,
den Abschlu einer Reihe anzugeben (vgl. Wir waren neun Nchte unterwegs. In der zehnten aber . . .)
(Meier-Brgger 2000: 219).
47. See Goetze 1951: 46970; followed by Melchert 2000: 5861, 6869.
48. The abbreviated complement -ki is known: 2-ki pi he shall give twiceltwofold Laws 70.
49. The meaning of the form a-an-ki in KUB 4.1 iv 36, 38 remains unclear. Eichner (1992: 4243; with
incorrect citation as KUB IV 2 iv 36, 38) interprets nki as once, the reading for 1-anki. But the adverb
ma no longer provides support for - as a stem for one: see 7.16 (p. 146). On ia- as the stem for one
see Goedegebuure 2006. A sense an equal number of times for nki (see (
LU
)nt- equal [written hyper-
plene as
LU
a-a-an-za], as per Goedegebuure 2002: 6468) does not seem contextually likely, and one would
have expected a hyper-plene writing `a-a-an-ki according to that theory. The most likely scenario is that the
scribe twice wrote a-an-ki for 2-an-ki. The 2 and A signs are quite similar, and the alternation of once
and twice ts well in the context of dance moves.
50. One may compare for this alternation Greek l many times.
9.53 Numbers
00-HittiteGrammar.indb 168 6/5/08 4:16:44 PM
169
third time (s)he speaks once as follows KUB 30.15 + KUB 39.19 obv. 3334 (Otten
1958 6869). Likewise for txnrn+-anki:
d
r.-a
GI
araiya 5-anki inkueni . . . nu
+\+
d
r. 15si inkueni We will bow ve times at the arai- of Ea . . . we will bow
fteen times to Ea KUB 33.106 ii 2223 . Such multiplicative numerals do not occur in
clauses with imperfective (-ke-) forms. See for example: uya 3si arnikzi KBo
6.4 i 5 (late laws), 3si arnikta KBo 12.3 iii 5 (OHlNS), numu i.xr
URU
Ulma
xi-ya menaanda 2si auer KBo 10.2 i 3435 (OHlNS), nuza tu-an 2si 3si
t KUB 1.16 iii 30 (OHlNS), 3i ekuzi IBoT 1.29 obv. 27 (MH). Exceptions occur
when in addition to the multiplicative numeral a distributive expression also is present
(see 24.12, p. 320), for example in nettakkan xt.xx-ti xt.xx-ti peran 3si
[alzikan]du Let them read them (scil., the tablets) aloud before you three times year
by year KUB 21.1 + KUB 19.6+ iii 74 (NH). An indenite relative form based upon
the stem mai- how many exists in the word maiyanki however many times (see
CHD s.v.).
9.56. Repetition of sets in mathematical multiplication. Multiplicative adverbs mean-
ing x-times in the mathematical sense (i.e., two times [i.e., repetitions of ] two is four)
are also written logographically with txrni + Akkadian -si and with txrni+i or
txrni+-anki. Examples: 8-anki 8 eight repetitions of eight (i.e., eight times eight)
KBo 53.47 obv. 13; [kuwapima . . . w]tar alkiyamaat 10-i 20-i naeu[t]
[Where . . . ] there is water, divert it to the crop(s) ten times, twenty times KBo 12.40
iii 12 (= THeth 11:38 iii 6, OHlNS). Note that the ending -i does not agree in gender
or case with the noun wtar. It must therefore be a xed (i.e., indeclinable) form.
9.57. The form 10-pa in r:+ malwaza tepuya [t a]kki tn.so-tarmai
10-pa piyan My brotherhe has not even a little intelligence, but valor is given to him
ten times over KUB 33.113 + KUB 36.12 i 3536 seems to be a multiplicative.
9.58. Because any phonetic complement may exist in longer or shortened form, the
shortened forms of two dierent suxes may appear to be identical. It is therefore not
certain that all numerals complemented by -i are the same. Above we suggested that
one group may be multiplicative and the same as forms showing longer complements
in -ki. But another line of evidence suggests shortened forms in -i equivalent to longer
forms in -li. The sequence: 4si 9 ru.rnr.xr 30-i 2 ru arnantai 10-i
2 ru orurx 10-i 2
NINDA
wakiar 10-li 2 ru.rnr.xr 10-i adanna 9 ru.
rnr.xr 30-i KBo 21.1 i 35 shows the numbers 30-i, 10-i, 10-i, 10-li, 10-i, and
30-i. The only number measuring bread weight or size that modies a clearly neuter
noun (
NINDA
wakiar) is 10-li. This suggests that the -i forms are abbreviated for -li,
and that these are adjectives modifying the bread names. And indeed the following
passage seems to conrm this theory: 10 ru.rnr.xr arpanza 20-li KBo 41.1b
obv. 3. The theory is not without its problems: the -i form is attested with
NINDA
tunik
and
NINDA
kaaret, both of which look like neuters. And the neuter form in -li only oc-
curs with
NINDA
wagear, not with the other mentioned neuter bread names
NINDA
tunik
and
NINDA
kaaret.
Numbers 9.58
00-HittiteGrammar.indb 169 6/5/08 4:16:45 PM
170
9.59. Hittite may have possessed another set of multiplicative adjectives (twofold,
threefold, sevenfold, etc.). This possibility is suggested by the fact that all factitive
verbs in -a- built to such adjectives exhibit a base ending in -(i)ya-, even for numbers
whose stem in the cardinal form does not end in ily. 3-ya- to make threefold, mul-
tiply by three (*teri- + (i)ya- + -a-), 4-ya- to make fourfold, multiply by four
(< me(ya)w- + -(i)ya- + -a-).
51
Two of the three factitive verbs based upon num-
bers known to date have this -ya- ending (10.9 with n. 11, p. 176). The exception,
2-a- could either be a shorter variant of the longer complement -ya- (compare the
free variants sro
5
-yaanlsro
5
-aan and sro
5
-yairlsro
5
-air), or a factitive built to
a special adjective twofold not formed by derivation with -(i)ya-. This behavior would
not be unusual typologically for the number two. An alternate account of 3-ya- and
4-ya- would assume that the former was built directly to the number teri- + -a- to
make three(fold?), and that 4-ya- is remodeled on analogy to 3-ya-. This second
account cannot explain iptamiya- (name of a beverage), which directly attests a sux
-(i)ya-, if as usually assumed it is derived from the PIE cardinal number seven.
Distributives
9.60. Distributives (two each, four apiece, etc.) are expressed with the Sumerian
suxes -Xx (signs: -.) and -f.Xx (signs: -f..) (HZL 274, sign #364).
Derivative Adjectives and Verbs
9.61. A degenitival adjective iela- single; of one and the same is derived from
the genitive il of ia- one (Honer 2006). This word is the Hittite translation of Ak-
kadian dnu single, unattached; unmarried in the lexical text KBo 1.44 + KBo 13.1
i 54 edited by Otten and von Soden (1968: 17). In its nominative plural form iele in
Laws 31 (LH 40, 184), it describes the legal status of a free man and slave woman
prior to their contracting a marriage (nanza +\+ uxsi di). It is a necessary part of
the if clause, because the focus of the law is upon the division of property when the
marriage is dissolved, and this division would be aected if the man (for example) had
another wife with legal claim to part of his property. Because of the equation IoI\i =
iela in the lexical text, we can assume that some form of iela- was the Hittite read-
ing underlying dnu as a logogram elsewhere in Hittite texts. The same derivational
sux, -ela-, occurs in the adjective apella- the . . . of that one (see 5.4, p. 133, and
6.2, p. 137). We also assume an adjective 2-ela- paired, in twos, literally (they) of
two (see Honer 2006),
52
derived from the genitive of the number two (2-el), show-
51. See Szemerenyi 1996: 22728, who shows the IE ordinal sequence *tri-yo- and k
w
tur (i )yo-.
52. na-at-kn 2-e-lu-u-pt +-\+
GI
x a-ra-a [ pa-a-ir ] (19)
d
orioxr-u
m
Ur-a-na-bi-i-a And
theyjust (-pat) the two of them (lit., they of the two)Gilgamesh and Uranabi went up into the boat
KUB 8.50 iii 1819 (NH). 2-e-lu-u is a late Hittite nominative common plural, showing the regular nomi-
native plural ending for a-stems in that period (see 3.16, p. 70).
9.59 Numbers
00-HittiteGrammar.indb 170 6/5/08 4:16:45 PM
171
ing the same sux -ela- attested in iela- of one, single. The forms 1-ela and 2-elila
represent an adverbial use of the nominative-accusative plural neuter of the same adjec-
tives, meaning respectively one at a time and together, as a pair, both.
53
Examples:
LU
so utxt-aa
d
t
d
Kataa
d
x[ ]
d
Telipinu
d
alkin
d
t -i 1-ela akuwanzi The
priest and the prince drink the Storm-god, Kataa, . . . (and) the Storm-god of Heaven
one at a time KUB 53.17 iii 1012 (cited by Eichner 1992: 38 as Bo 2701 iii 12); [
. . . 2]-ilapat kar (ap)pianzi The two lift [ ] together (or They lift the two (nom.-acc.
pl. neut.) [ ] together) KBo 17.30 ii 3 (OS, for restoration see 2-ila in ii 1); [ . . . . . .
(-)t]a
LU.ME
so 2-ila
URU
Zuppara pnzi the priests go to Zuppara in pairs (or both
of them) KUB 60.152 i 6 ; nat 2-ela +\+ +u[i:+] kattan ur And they came to my
father together KUB 19.7 i 89 (D frag. 27, NH), nu 2-ilapat akuwanzi They .
(some form of punishment) them both together KUB 13.9 iii 18. Although this sux
is not yet found on other numbers, there is a priori nothing to prevent Hittite speakers
from having formed *3-ela- in threes, *4-ela- in fours, etc.
9.62. A likely derivative from PIE *sem- is `umuman-, an adjectival derivative (see
2.47, p. 60) in -uman-, seen in the factitive verb umuma- to unite (Rieken 2000b).
For other factitive verbs based upon multiplicative adjectives in -iya- see 9.59.
Numbers in Compounds
9.63. In combination two occurs in t-yuga two-year-old (animal), and the or-
dinal tn second occurs in the name of a cut of meat from an animal,
UZU
dnati
(literally, second bone).
Compound Numbers
9.64. Compound numbers are not the same as numbers in compound words. Rather
they are quantitative expressions involving two or more mathematical units. In Hittite
an example would be 6 rxt 5 gipear .X, where the larger unit is given rst followed
by the smaller one. From variant writings of the same expression we gather that what
is sometimes written in numerical notation as (for example) 19 r+. eppit expresses
what in the spoken language was 19 r+. r+. eppit (see HKM 111:11).
54
In other
words, although the unit (r+utsi) remained the same, the tally proceeded from whole
numbers to the noun for the fraction, itself perhaps preceded by the cardinal number
one (ia-). This seems to be especially common when what is numbered is a unit of
53. The variation in vocalism in 2-elila is unsurprising if the Hittite stem was as expected `dwela-: see
AHP 14445.
54. See also na 2 u uya pennai, 2 o 1 o xi.nnnn, s+ 1 ir\i ir\t, [1?] n
5 ir-\i ir-\i [. . .] KUB 58.49 iii 7; parqati 1 sro.xi sro.xi; but see: 2 o 1 o xi.nnnn
KUB 29.39 i 11, 1 rxt .X SBo I 2 obv. 12; 1 r+. +u:i
GI
uiu+st kinan, where the 1 is of a dierent
commodity.
Numbers 9.64
00-HittiteGrammar.indb 171 6/5/08 4:16:46 PM
172
measure rather than a simple commodity. In expressions containing a whole number and
a fraction, sometimes the Sumerogram for half is preceded by the ur sign (one), in
which case 2 o 1 x should not be read as two plus one and a half, but two plus
one half.
55
55. Why otherwise is 2 o 1 o xi.nnnn iya[tten] KUB 29.39 i 11 not written as 3 o o
xi.nnnn iya[tten] or 3 o xi.nnnn iya[tten]?
9.64 Numbers
00-HittiteGrammar.indb 172 6/5/08 4:16:47 PM
173
Chapter 10
VERB FORMATION
10.1. The nite verb form consists of a stem (a root with optional sux) and an
inectional ending. For example, laknui consists of: lak (verb root) + -nu- (causative
stem-forming sux) + -i (pres. sg. 2 ending) you cause to fall down. Hittite verbs use
a single stem (which may undergo ablaut, see 3.37, p. 78) for all tenses and non-nite
forms,
1
and all verbal categories are expressed by inectional endings (for the endings
see chapter 11).
10.2. Many verbs have root stems: akk- to die, ar- to arrive, ar- to stand, ark- to
divide up, eku- (lek
w
-l) to drink, pp- to seize, e- to be, an- to dip, ar(k)- to
hold, kuen- (lk
w
en-l) to strike, kill, kuer- (lk
w
er-l) to cut, n- to fear, and wal- to
strike. Some of these are ablauting: akk-lek- to die, ar-ler- to arrive, eku-laku-, pp-l
app-, e-la-, kuen-lkun-, kuer-lkur-. For their conjugation see 12.212.17 (pp. 187
196). Many Hittite verbs, however, are derived from other verbs, from nouns, or from
adjectives. Such verbs are characterized by suxes or inxes which enlarge their stems.
The most important of these will be discussed below.
Reduplicated Roots
10.3. Reduplicated verbal roots are encountered in Hittite (van Brock 1964; Oshiro
1995; Oettinger 1998; Rasmussen 2001). This verbal reduplication is of two types.
The rst shows reduplication of the entire root: aa-lae- to seat < a-le- to sit,
alili(ya)- to kneel < aliya- to kneel,
2
ananiya- to be vexed(?), be dissatis-
ed(?),
3
aa- to scrape, katkattiya- to tremble, kikki- < ki- to be placed (see
HED K s.v.), kuwakuwa-lkuku- < kuwa- to crush, lalaiya- and lailaeke-
to be agitated, anxious; mill about, parip(a)rai- < parai- to blow (a horn), partipartike-
?, tataiya- to whisper (onomatopoetic), waltiwalteke- (unknown meaning, perhaps
to rage). For this type of reduplication see EHS 120 and Raggi Braglia 1989: 208. The
mineral name kunkunuzzi- may be derived from a reduplicated form of the verb < kuen-
to strike, perhaps because this particularly hard mineral (basalt?) was used for milling
1. For possible rare exceptions see Melchert 1997c, but the evidence is sparse and debatable. See 14.9
(p. 233).
2. We include under this stem only the intransitive medio-passive form aliliyandri and correspond-
ing iterative aliliskemi. For the transitive verb alila- see below.
3. Compare for the meaning HLuwian ha-ni-ya- malicious. The in front of transcribed Hittite words,
as in the CHD, represents a pair of marker wedges.
00-HittiteGrammar.indb 173 6/5/08 4:16:47 PM
174
and pounding or required more than the usual amount of pounding to quarry. The stem
wariwarant-luriwarant- (< war-) burning denitely shows a connective -i-.
4
Other
instances of connective -i- are uncertain: the transitive verb alila- to atter(?), + ara
to put o (with attery)(?), is reduplicated from alali- to cradle (a child), to roll (a
wheel).
5
If alila- is formed from the base ala-, then the -i- of alila- is like that
in wariwarant-, but alila- could also represent full reduplication of the base ali-. If
one connects arrirra- to scrape o with arra- to wash (both to scrub, rub clean),
it would also show a connective -i-.
6
The status of the rst -i- in the three reduplicated
-ke- forms lailaike-, partipartike-, and waltiwaltike- is also uncertain. It could
be a connective -i-, but it could also reect mere copying of the stem-nal -i- that ap-
pears before the -ke- (for which see 12.32, p. 204).
10.4. The second type, partial reduplication, takes several forms in Hittite: (1) re-
duplication with the vowel e in the rst syllable: wewakk- < wek- to ask, lelaniyale- to
become furious, ed- (< *e-d- with zero grade of the root `sed-) to prosper possibly
from an earlier meaning to sit safely(?);
7
(2) reduplication with i and zero-grade of the
root (see Melchert 1984b: 98100): liluwa- < lau(wa)- to pour out (liquids), tita-
to thunder; (3) reduplication with the vowel of the root: aar- to laugh (prob-
ably also onomatopoetic), pappar- to sprinkle, kukkur- < kuer- to cut up, mutilate,
kun(ni)kunk- < kunk- (?), pupua- to crush(?), kikki- < ki- to become, lillipa- <
lip- to lick, iiya- < ai- to press (as a creditor). The examples with -i- could also
belong to (2). The reduplication pattern in mummiye- (middle) (< mau()- to fall) is
unclear.
10.5. Further research is needed to reach rm conclusions about the meaning and
function of verbal reduplication in Hittite. The task is made dicult by the small number
of attestations for most reduplicated stems. While an iterative or intensive value
may be shown for some cases (Honer 1975: 13940 and Oettinger 1998), this is not
true for all (van Brock 1964: 14445). There is no uniform pattern for the inectional
type of reduplicated stems. For reduplication in noun stems see 2.56 (p. 62).
4. The connective -i- is reminiscent of that in Sanskrit intensives such as gnigam- < gam- to go and
kari-k- to make, for which see Schaefer 1994: 24.
5. The common denominator of the uses of the base verb is to grasp lightly with the two hands. The
precise meaning of the reduplicated stem remains elusive, but to atter ts the context of KBo 3.34 ii 21
and to put o (with atterylexcuses) that of KBo 18.78 obv. 4 with ara (see a similar use of ala- in KBo
5.9 iii 9). For such a meaning from to grasp lightly with the two hands compare English to stroke in the
sense to atter or alternatively the colloquial use of to play in to play (i.e. manipulate) someone. In any case,
the derivation of transitive alila- from ala-lalai- is conrmed by the use of the reexive -za with both.
We reject the widely favored derivation of (transitive) alila- from aliya- to kneel.
6. This derivation of arrirra- < arra- was suggested by Jay Friedman of the University of California
at Los Angeles (in an unpublished paper delivered at the 2001 annual meeting of the American Oriental
Society in Toronto).
7. Probably also lelakk- < lak- to bend, knock over and lelareke- to soothe, conciliate, but these
could also be read as lilakk- and lilareke- with i-reduplication.
10.4 Verb Formation
00-HittiteGrammar.indb 174 6/5/08 4:16:49 PM
175
Verbal Sufxes and Infxes
Overview
10.6. The verbal suxes (and inx) and their known combinations:
8 9 10
added to
Sux nouns verbs adjectives function
-a- no
8
no yes Factitive (to make . . . )
-i- yes no no Denominative
-e- yes no yes StativelFientive (to be(come) . . . )
-e- no
9
no yes Fientive (to become . . . .)
-nu- no
10
yes yes Causative (to cause to . . . ) (< verb)
Factitive (to make . . . ) (< adjective)
-annali- no yes no Imperfective
-kela- no yes no Imperfective
-a no yes no Imperfective
Inx
-ni(n)- no yes no Causative (to cause to . . . )
10.7. As deverbative suxes, the imperfective suxes -kela- and -annali- (here-
after, for simplicity, -ke- and -anna-) can be added to the verbal stems formed with
the other suxes listed and the inx -ni(n)-. Dierent imperfective suxes co-occur:
-anna- + -ke- in oti-anniket, waneket, uittiyannikemi, peiyannekezzi. Other
combinations of more than one sux are for semantic reasons relatively rare. The
motivation for the exceptions is reasonably clear. One can combine the entive sux
-e- and causative -nu-: atk-e-nu- to cause to become narrow(ed), hemmed in >
to besiege, lalukk-e-nu- to cause to become light > to illuminate. The causative
-nu- can also be added to a stem with the imperfective sux -kela- when the latter has
a durative or habitual sense: lalaeke-nu- to cause to keep running > to gallop
(trans.), uke-nu- to cause to inspect.
10.8. Verbs using all but one (-anna-, 13.25, p. 226) of the above suxes inect
according to the mi-conjugation (see chapter 12).
The Individual Suxes
10.9. The sux -a- is added to the stems of adjectives or numerals in order to pro-
duce verbs which mean to make something what the adjective or numeral denotes or
8. Possible exception: iuniya- if this is based on noun iuni- god(dess) instead of a postulated
adjective *iuni(ya)- divine(?).
9. See 10.14 (p. 178) for a possible exception: annitalwae-.
10. The apparent exception earnu- to bloody from ear blood may actually be formed from
eariya- bloody on the model of naarnu- to frighten < naariya- to be(come) afraid.
Verb Formation 10.9
00-HittiteGrammar.indb 175 6/5/08 4:16:50 PM
176
to regard somethinglsomeone as (or declare someone to be, or treat someone as) what
the adjective denotes. See also 2.47 (p. 60), 9.59 (p. 170), 9.62 (p. 171), 10.6.
11
Adjectival Base Factitive Verb
arawa- exempt (from) arawa- to declare someone exempt from
antezzi (ya)- rst antezziya- to make rst
appinant- rich appina- to make rich
*addul (a)- healthy addula- to make healthy
ikuna- cold ikuna- to make cold
idlu- evil idalawa- to treat evilly; harm, injure
kattera- inferior kattera- to declare to be inferior, the loser (in a suit)
mayant- youthful, vigorous mayanta- to rejuvenate
nakki- important nakkiya- to regardltreat as important
newa- new newa- to make new
arazzi (ya)- superior arazziya- to declare to be superior (in a suit or a contest)
3-ya- (*teriya-) threefold 3-ya- to make threefold
11
4-ya- (< meu-, meyau-)
fourfold
4-ya- to make fourfold
11
For a summary of the use of this very productive sux see Oettinger 1979: 23954. For
the formal derivation of appina- from appinant- see 2.2 (p. 51).
10.10. The sux -i- is used to form denominatives from a wide variety of noun
stems:
12
(1) from a-stems such as alpa- excrement (> alpi- to besmear) or
GI
tarma-
peg, stake (> tarmi- to nail down, x); (2) from a single i-stem aranalli- wreath,
diadem (2.21, p. 55), itself a derived form in -alli- (> aranalli- to wreathe);
13

(3) from ai-stems like alluwai- quarrel (> alluwi-to quarrel); (4) from u-stems
like lulu- prosperity (> luluwi- to make prosper), iwaru- dowry, inheritance share
11. [3-i]a-a-a-an-zi 4-ya-a-a-an-zi KUB 9.4 ii 33, [3-i]a-a-a-a-u-wa-ni 4-ya-a-a-a-
u-wa-ni ibid. 35. The 3-ya- is ne for *teri- three, but 4-ya- poses a problem. One would expect
*meuwa- or *meyawa- (i.e., 4-wa or 4-yawa-). The adjectival base is probably derived from the
cardinal by a sux -(i)ya- (2.35, p. 58, and 9.59, p. 170), which explains the y in 4-ya-. Alternatively,
the stem for multiply by four may have been remodeled on that for multiply by three.
12. We cite the sux as -i-. The long diphthong is not always expressed by plene spelling (see AHP
149). Since Ca-(a)-e is attested as a spelling for what assuredly is a long diphthong la:yl in nouns (li-in-
ga-en and ta-ma-a-e), we assume consistently a long diphthong also for all verb forms with the sux -i-.
That some OH instances of -Ca-e- (e.g., tar-ma-e-mi) represent uncontracted [Ca:_e] cannot be entirely
excluded. See Melchert 1984b: 74 75 for discussion.
13. This exceptional example is due to the fact that the verb is modeled on the matching Luwian type
formed from the a-stem sux -alla- that took i-mutation: see arkammanallawi I make tribute-bearing
KUB 23.127 iii 6 in Hittite context < `arkammanallali- tribute-bearing, attested in Hittite form as an i-stem
arkammanalli-. But Hittite speakers surely viewed aranalli- as derived from the i-stem arnalli-.
10.10 Verb Formation
00-HittiteGrammar.indb 176 6/5/08 4:16:51 PM
177
(> iwarwi- to bestow a dowrylinheritance share), genzu- pity, mercy (> genzuwi-
to take pity on), iaru- tears (> iaruwi- to shed tears); (5) from l-stems
such as takul- peace, solidarity, alliance (> takuli- to be allied, at peace) and
ml intelligence, mind (> mali- to approve of; see HED M 2526, 76); (6) from
t-stems such as awitit- weanling(?) (or suckling(?)) (> awititi- to wean(?) [or
to suckle(?)]; (7) from the oblique stem of nouns in -arl-n- like appear body
part (> appini- to dismember), aear fortress (> aeni- to fortify), and
elaar sign, revelation (> elani- to reveal), etc.
14
10.11. The sux -e-, like -e-, is added to adjectival (and a few nominal) stems to
produce verbs which mean to be(come) what the adjective or noun signies. For the
deletion of the sux of the base adjective in some examples, but not others, see 2.2
(p. 51).
15 16
Adjectival or Nominal Base Stative Verb
alpu- sharp, pointed `alpue- (in alpuemar)
arawa- exempt (from) arawe- to declare oneself free (from)
15
`aran(a)- envious arane- to become envious
16
au- king ae- (itoi-(u)e-) to become king
lalukki- bright lalukke- to become bright
marant- false mare- to become false
miyauwant- old miyaunte- to become old, live long
paprant- impure, guilty papre- to be proven guilty by ordeal
parkui- pure, clear parkue- to become clear
*orxr-aliara- female slave orxr-aliare- to become a female slave
10.12. Watkins (1973), who identied this class of verbs in Hittite, called them de-
nominal statives, but recognized (1973: 67) that attested examples may have either
stative or entive meaning (to be . . . or to become . . .). Honer (1998d) has shown
that it is the entive sense that predominates in Hittite and that, in those cases where
both an e- and an e-verb exist for the same adjectival root, the former seems to have
been replaced by the latter in the course of time (arawe- > arawe-, orxr-aliare- >
orxr-aliare-
17
).
10.13. The following are examples of verbs derived from an adjectival base by the
sux -e- which mean to become what the adjective signies. For the inection of
these verbs see 12.20 (p. 197).
14. For these classes and further examples see Oettinger 1979: 36568.
15. For arawe- see LH 13839, 21920; and Honer 1998d.
16. For the original stem of this verb as being in -e- see ar-a-ne-e-i KBo 25.122 iii 2 etc. (OS) and Neu
1983: 2829. The base stem is also seen in aranatalla- envier (2.32, p. 57).
17. For the latter see n. 20 below.
Verb Formation 10.13
00-HittiteGrammar.indb 177 6/5/08 4:16:52 PM
178
Adjective or Noun Base Fientive Verb
arawa- exempt from arawe- to become exempt from
aiwant- poor aiwante- to become poor
appinant- rich appine- to become rich
18
arki- white arke- to become white
atali- brave atale- to become brave
atku- narrow atkue- to become narrow
atukali- awesome, frightening atuke- to become awesome
idalu- evil idalawe- to become evil
mayant- young mayante-to be(come) young
mekki- numerous makke- to become numerous
19
parku- high parke- to become high
18
parkui- pure parkue- to become pure
alli- large alle- to become large, grow
*orxr-aliara- female slave orxr-aliare- to become a female slave
20
18 19 20
10.14. For the deletion of the sux of the base adjective in some examples see 2.2
(p. 51). There is one apparent example derived from a noun: annitalwae- to become
a litigant < annitalwa(n)-.
10.15. Verbs formed by suxing -nu- can be either (1) causatives of other verbs
or (2) factitives of adjectives. See Sihler 1995: 455 (PIE *-new-l*-nu-). Examples
of the causative are: arnu- to cause to arrive (ar-), warnu- to cause to burn (war-),
linganu- to cause to swear (link-), arganu- to cause to perish (ark-), wanu- to
cause to turn (we-). Sometimes one can hardly detect any dierence in meaning
between the root verb and the extension in -nu-: paa- and paanu- to guard,
tekku(i)ye- and tekku(a)nu- to show, reveal, present. For the inection see 12.44
(p. 210, paradigms).
10.16. Examples of factitives are: malikunu- to make weak (maliku-), da(a)nu-
to make strong (dau-), allanu- to make great (alli-), magnify, tepnu- to make
small lfew (tepu-), parkunu- to make pure ( parkui-), parknu-
21
to make high (parku-),
earnu- to make bloody (ear blood). For the deletion of the stem-nal -u- and
18. See 10.14.
19. The verb makke- to become numerous < mekki- much, many shows not only deletion of the
stem vowel but also an alternation in the root vowel reecting an archaic pattern with a zero-grade root
*mk- and inserted vowel.
20. For orxr-aliare-, which was replaced in post-OH by orxr-aliare-, see LH 43, 139, 18586,
263 and 10.12 above.
21. Written pr-ak-nu- and pr-ga-nu-, indicating the absence of the apparent vowel before -nu-; see
1.11 (p. 12) on empty vowels in Hittite cuneiform.
10.14 Verb Formation
00-HittiteGrammar.indb 178 6/5/08 4:16:52 PM
179
-i- of the base adjective in most of the examples see 2.2 (p. 51).
22
On the derivation
of earnu- see also n. 10 (p. 175).
10.17. The verbal inx -ni(n)- is inserted before the nal consonant in certain verbs
ending in -k-: ark- to perish > arnink- to destroy, itark- to get sick > itarnink-
to make sick, *ark- > arnink- to make compensation, `n(e)ik- *rise > ninink- to
mobilize, set in motion. From the examples it is apparent that its force is causative or
transitivizing. A similar value for the PIE nasal inx is seen in examples such as Old
Irish -roind makes red to a root meaning (be) red.
10.18. For the verbal suxes -ke-, -a-, and -anna- see chapter 24, Verbal
Aspect.
22. The verb maknu- to make numerous < mekki- also shows the same alternation in its root vowel as
in makke- (see n. 19 above).
Verb Formation 10.18
00-HittiteGrammar.indb 179 6/5/08 4:16:53 PM
180
Chapter 11
VERB INFLECTION
Types of Verbal Infection
11.1. Hittite nite verbs are inected for the person and number of the subject (sub-
ject marking), and for mood, tense, and voice.
1
There are also non-nite forms: verbal
substantives, innitives, participles, and so-called supines (see chapter 25).
11.2. Subject marking. Finite verbs are inected in the usual three persons and in
two numbers, singular and plural. Hittite has no dual ending. The nite verb is not marked
for gender. Gender distinctions (common [or animate] and neuter [or inanimate], see
3.23.6, pp. 6466) exist in the participle by virtue of its status as a verbal adjective.
11.3. The nite verb expresses only two moods, indicative and imperative (by con-
vention, voluntative and exhortative rst-person forms (23.423.5, p. 313) are in-
cluded in the paradigm of the imperative). Modalities which in other Indo-European
languages would be expressed by subjunctive or optative forms (could, would, must,
ought to, would like to) are conveyed in Hittite by means of modal particles such as
man (23.1023.16, pp. 314316), or adverbs such as imma really (see 19.8, p. 291;
26.10, p. 343; 27.4, p. 350) and kuwatka perhaps (see 16.59, p. 255; 19.8, p. 291;
26.7, p. 342; 27.3, p. 349).
11.4. In the indicative mood there are two tenses, present (for present or future time)
and preterite (for past time). In addition there are analytic constructions (like English
havelhad gone and will go) that explicitly express the present and past perfect and the
future tense. The use of the present, the preterite, and the analytic perfect will be de-
scribed in chapter 22 and the periphrastic future in chapter 24 (see also 22.822.9,
p. 308).
11.5. The nite verb is inected for two voices, active and medio-passive. For their
use see chapter 21. In the active voice nite verbs are inected in one of two conjuga-
tions, the mi- and i-conjugations, named for the endings of the rst-person singular
indicative in each set. The endings of the rst- and third-person plural are identical in the
two sets, while in the second-person plural all mi-verbs and most i-verbs take -ten(i),
while i-verbs with stems ending in -i- take -ten(i). In all nite forms of both voices and
all tenses and moods of the vocalic-stem verbs, endings beginning with a stop or with
regularly appear with geminate spellings. Non-geminate spellings are either faulty of
simplied spellings (see 1.10, p. 12).
1. For the optional explicit marking of imperfective aspect see chapter 24.
00-HittiteGrammar.indb 180 6/5/08 4:16:53 PM
181
11.6. The following is the scheme of endings for the nite forms of the active verb.
For medio-passive endings see 11.17 (p. 184).
2 3 4 5 6 7 8 9 10 11 12 13 15
Present Preterite
mi-conjugation i-conjugation mi-conjugation i-conjugation
sg. 1 -mi -i, (-e in OH)
2
-un,
3
-nun
4
-un
2 -i,-ti -ti, -ti -, -t -ta,
5
-ta,
6
(-t
7
)
3 -zi
8
-i, (-e)
9
-t,
10
(-) -, (-ta
11
, -ta
6
)
pl. 1 -weni,
12
(-wani
13
), -meni,
14
(-mani),
13

(-uni)
15
-wen, -men
2 -teni, (-tani) -teni, (-tani), -teni -ten
16
-ten, -ten
3 -anzi -er, (-ar )
17
2. In OS rather consistently showing the e-vocalization (sign -); see Otten and Souek 1969: 56.
3. Used with consonantal stems.
4. Used with vocalic stems.
5. See dtta you (sg.) took (MH/MS) and paitta you (sg.) gave (NS).
6. Verb forms showing the -ta ending, either in sg. 2 or 3, are: memita spoke, peta gave, alita
cradled(?), tedaita (meaning unclear), uleta blended(?), merged(?), udata brought, unnita drove
here, pennita drove there, uppeta sent, daita placed, naita turned, tarneta released, and pos-
sibly arueta bowed(?). This ending is attested in MH/MS only in memita. In other cases it is a NH in-
novation. The OS form pal-ku-u-ta (CHD P 68) must contain a verbal stem in -; Oettinger (1979: 333)
has suggested *palkue-.
7. Attested in i-verbs in -i-; see zit and alzait in chapter 13 paradigms (13.21, p. 223).
8. Since the sequence z is elsewhere quite stable, the rare NH form da-ma-a-ti KBo 5.9 ii 26 for
expected da-ma-a-zi may be just a scribal slip.
9. E.g., wa-ar-a-e in KUB 29.30 iii 4 (OS).
10. This ending is spelled -ta in verbs whose stem ends in a stop or fricative: e-ep-ta, e-e-ta, etc. (likewise
when it occurs in the pret. sg. 2). Scholars disagree whether the nal a-vowel is real or merely graphic. For the latter
view see Pedersen 1938: 93, Sturtevant and Hahn 1951: 141, Kronasser 1956: 31, and Yoshida 2002; and for the
former Eichner 1975: 7980, Oettinger 1979: 9, 237, AHP 17576, and Kimball 1999: 193194.
11. This ending replaces - in verbs whose stem ends in -, e.g., ta gave birth (OS), wata bought.
See Oettinger 1979: 41.
12. The ending -weni can stand for either present indicative or imperative rst-person plural. The two
forms are distinguishable only by context.
13. For the endings -wani and -tani see AHP 138 with references to other viewpoints.
14. The endings -meni (-mani) and -men are largely conned to the nu-causative verbs, although one
nds them also in umeni and aumen (from au- to see). The use of the m-variants of -weni, -wani, and -wen
after verb stems in u is phonologically conditioned (1.126, p. 44).
15. There are enough attestations of this ending to suggest that it may be real and not merely the result of
omission of the sign for e. See p. 192, n. 31, p. 207, n. 119, and p. 223, n. 43, and Otten (1969: 26), who adds
kar-a-u-u-ni, -u-ni, and pa-a-i-u-ni.
16. The ending -ten is can stand for either preterite indicative or imperative second-person plural. The
two forms are distinguishable only by context.
17. For rare instances of an ending -ar, such as -e-mi-ya-ar they found (from wemiya-), a-a-ni-ya-ar
they dipped/drew (from niya-), and a-pa-i-ya-ar they scouted (from apaiya-) see Neu 1989a.
Verb Inection 11.6
00-HittiteGrammar.indb 181 6/10/08 7:43:20 AM
182
Imperative
mi-conjugation i-conjugation
sg. 1 -(a)llu, (-lit, -lut) -allu
2 -, -i,
18
-t
19
-, -i
18
3 -d/tu
20
-u
pl. 1 -weni
12
, -wani
21
2 -ten
16
3 -and/tu
18 19 20 21 imperative active paradigm
11.7. The preterite third-plural person ending is typically spelled with the ambigu-
ous rn sign (see 1.28, p. 20). Spellings such as e-e-er and -te-er with unambiguous
Ce signs or da-a-ke-e-er show that the ending is l-erl. We should therefore likewise
read p-i-IR from pi- (the weak stem of pai- to give) or p-e-i-IR from peiya- to
throw as, respectively, p-i-er (beside p-i-e-er ) and pe-e-i-er. That is, the -i- in these
forms serves either to indicate the stem vowel or a glide connecting the stem vowel to
the ending -er. It is not intended to mark the rn sign as containing -i-.
22
Preterite third-
plural active verbs ending in one of the signs rn, rn, frn, xrn, or rrn should always be
read with their alternate e vocalizations: u-un-ner they lled, e-er they were, -ter
they brought, e-ker they died, e-ep-per they seized.
23
For monosyllabic dir and
pir see p. 224, nn. 4850, and 12.41, p. 209.
11.8. In rare instances the pres. sg. 3 mi-conjugation ending -zi and the pres. pl.
ending -anzi are written -za and -anza respectively: e-e-za in KBo 6.2 iv 5355 (Laws
98, OS) for e-e-zi helit is in the duplicate KBo 6.3 iv 5254 (OHlNS),
24
ar-za in
KBo 9.73 obv. 12 (treaty with apiru, OS) for intended ar-zi,
25
a-ku-wa-an-za (for
a-ku-wa-an-zi ) KUB 13.2 iii 16 (MHlNS ) (Melchert 1984b: 97), pan i-i-an-za (for
pan iianzi ) KBo 6.26 i 7 (Laws 158, OHlNS) with dupl. i-a-a-i (LH 126), and
18. E.g., pa-a-i. An example from the i-conjugation is the imp. sg. 2 tar-ni.
19. This ending is conned to the nu-causative verbs (e.g., parkunu-, allanu-, etc.), the imp. sg. 2 t
go! from the unprexed root i- found in uezzi comes and paizzi goes (see 12.42, p. 210), and the vari-
ant uwatet alongside expected uwate from uwate- to bring (see 12.21, p. 199).
20. Old Hittite copies tend to use the cuneiform sign ft for this ending; later, the sign ut was preferred.
This writing convention does not necessarily reect pronunciation.
21. See uwatten
URU
Na paiwani Come, let us go to Nea KBo 22.2 obv. 15 (OS).
22. Rare exceptions, such as p-e-er from peiya- to throw do not invalidate the majority of the
evidence, since forms of that verb such as p-e-e-er and p-e-et indicate that the e vowel in p-e-er was
real. The forms a-a-i-ir they gave birth KUB 21.38 obv. 60 (Hatt. III) from stem a- and wa-a-ti-ir
they sinned BrTabl i 10 (Tud. IV) from stem wata-, both from late NH texts, are genuine exceptions. See
also the atypical pres. sg. 3 wati of this verb in 8.8 (p. 151).
23. The fact that HZL fails to list an e-containing value for some of these Ciler signs is of no
consequence.
24. See CHD nakku loss(?) for discussion and literature.
25. See also another likely occurrence of ar-za for ar-zi in KBo 24.9:5.
11.7 Verb Inection
00-HittiteGrammar.indb 182 6/5/08 4:16:55 PM
183
i-tar-ni-ik-za for i-tar-ni-ik-zi KBo 40.272 left 5. Since two of these examples are
found in Old Hittite manuscripts which do not elsewhere betray a sloppy copyist, it is
likely that the writings with nal -za for -zi reect archaisms showing the regular loss
of nal -i in -zi from *-ti, as in the ablative ending -az (see -azziya; see 3.34, p. 77).
But the single occurrence in a-ar-a-za KUB 19.23 obv. 14 (NH) for intended a-ar-a-zi
it ows, cannot be an archaism. The usual ending -zi has its -i restored after the rest
of the present verbal endings.
11.9. The imperative rst-person singular is formed with the endings -(a)llu, -lut,
or -lit, the latter two forms being much rarer than the rst. The ending -allu clearly
occurs in cases where the verb sterm ends in a consonant: aallu (< - to be), ag-
gallu (< akk- to die), eggallu (< akk- to know), aiggallu (< aik- to satisfy
oneself), karallu (< kar- to cut o), tepaweallu (< tepawe- to become less),
uwallu (< au-lu(w)- to see). When the verb stem ends in a, one cannot detect the
presence of the parenthesized vowel in -(a)llu: memallu (< mema- to say), udallu (<
uda- to bring), wemiyallu (< wemiya- to nd), kuwayatallu (< kuwayata(i)- to be
fearedlformidable). Imperfective stems in -kela- appear as -ka-: ugallu (< au-lu- to
see). The ending -llu occurs with nu-causative verbs: anullu (< a(a)nu- to prepare).
The ending -lutl-lit is attested only for - to be.
Irregularities in the mi-Conjugation
11.10. There is a widespread intrusion of the ending -ti from the i-conjugation dur-
ing the NH period (see 12.3, p. 188; 12.8, p. 191; 12.20, p. 197; 12.33, p. 204). On
rare occasions the pres. sg. 2 (you . . .) ends in -zi instead of the expected -i or -ti:
itamazi you hear
26
(the same form as for he hears) for normal itamati (with the
-ti ending of the i-conjugation).
11.11. In New Hittite (NH) compositions there is a marked tendency to replace the
earlier pret. sg. 2 in - (you . . . ed) with the ending of the sg. 3 -t: earlier ie (OH)
and iya (MH) you didlmade, but later iyat; earlier allanu you raised, later allanut;
earlier paanu you protected, later paanut; earlier aranie you were envious,
tarkummiya you announced, amenu(?),
27
and memanuga you were making (me)
talk (imperfective), later tittanut you made (something) stand; earlier imperfective
stem + ending -ke (MH: atreke you were sending, dake you were taking,
memike you were saying), later -ket (NH). The form pute KUB 30.28 rev. 12
(NS) is by context pret. sg. 3 he led there and is the result of late inuence from the
i-conjugation on an otherwise mi-conjugation verb.
26. KBo 5.9 ii 1617 zikma[an] itamazi, see also iii 7; and KBo 4.3 + KUB 40.34 iv 3132 nu mn
zik
m
Kupanta-
d
ixx-a ap[eda menaanda] idalawezi.
27. a-me-nu-u KUB 31.112:11 (verb pret. sg. 2 according to Oettinger 1976b: MSS 35 99, noun ac-
cording to Pecchioli Daddi 1975: 1089).
Verb Inection 11.11
00-HittiteGrammar.indb 183 6/5/08 4:16:57 PM
184
Irregularities in the i-Conjugation
11.12. In NH, i-verbs with stems in -i- sometimes replace pres. sg. 2 -tti with -ti
after the pres. pl. 2 in -teni: paitilpeti you give, alzeti you call.
11.13. In the pres. sg. 3, an ending -ai occasionally appears where one expects -i:
ipanti and ipandai he libates, oers.
11.14. Similarly, in the pret. pl. 3, -aer occurs alongside the more common -er:
ipanter and ipantaer they libated from the verb stem ipant-. Obviously, in forms
such as dir they took the a is part of the verb stem, and the ending is only -er.
11.15. Conversely, the ending -i occurs in the pres. sg. 3 instead of an expected -ai:
watai he sins (the more usual form, since the normal stem is wata-) and wati (which
shows that the earlier stem was wat-).
28
11.16. The earlier ending - of the pret. sg. 3 was replaced in later Hittite by -ta. See
the paradigms in 13.13 (p. 219), 13.21 (p. 223), 13.23 (p. 224).
Medio-Passive Endings
11.17. The following is the scheme of endings for the medio-passive verb forms of
the present, preterite, and imperative (for a similar table see Neu 1968a: 19). On the
chronological distribution of the shorter and longer endings see 14.7 (p. 233).
Singular Plural
pres. 1 -()a, -()ari, -ri, -aari -wata, -watati
29

2 -ta, -tari, -tati -duma, tuma, -tumari, tu(m)mat
3 -a, -ari, -ri, -ta, -tari, -tri -antal-anda, -antari, -antldri
pret. 1 -ati, -at,
30
-aat, -aati, -aatti -watat, -watati
29
2 -at, -ta, -tat,
31
-tati -dumat, -tumt (?)
32
3 -at, -ati, -ta, -tat,
33
-tati -antat, -antati
imp. 1 -aru, -aaru -watati
2 -ut, -uti -tu(m)mat, -tumati
3 -aru, -taru -antaru
28. N. Oettinger (private communication) has pointed out the same pattern in earlier a-ar-ri he washes
and later (a-)ar-ra-i.
29. Only in NH or NS texts.
30. In paigaat in KBo 17.1+ iv 13 (OS).
31. In kitat KBo 6.2 ii 12 = Laws 37 (OS).
32. In pa-i-kat-tu-ma-a-a[t?] KBo 8.42 obv. 9.
33. In kitanziattat KBo 3.22:46 (OS).
11.12 Verb Inection
00-HittiteGrammar.indb 184 6/5/08 4:16:58 PM
185
Verbal Substantive, Participle, Infnitive, and Supine
11.18. There are four non-nite verb forms: (1) the verbal substantive and (2) the
verbal adjective or participle, both of which are declined; (3) the innitive and (4) the
supine, which are indeclinable. On their use see chapter 25.
11.19. In verbs that show stem variation the verbal substantive, participle, and inni-
tive are regularly formed from the stem shown by the present third plural: pp-/app-
to take > apptar, appant-, appanna, uek-/uk- to slaughter > ugatar, ugant-,
uganna, andi- to prepare, arrange > anduwar, andnt-, anduwanzi, aa-/
ae- to settle > aeuwar, aeant-, aeuwanzi, pai- to give > piyawar/piyatar,
piyant-, piyawanzi. For exceptions to this pattern see the respective paradigms in chap-
ters 12 and 13.
11.20. The scheme of endings for the verbal substantive, participle, innitive, and
supine is as follows:
Set Verbal substantive Participle
34
Innitive Supine
1a -war (gen. -wa)
35
-ant- -wanzi
36
-wan
37
1b -(u)mar (gen. -(u)ma) -ant- -(u)manzi
2 -tar (gen. -anna) -ant- -anna -wan
34 35 3637
11.21. Although the endings of the participle and the supine (-wan) are the same for
all verbs, verbal substantives (forms in -war, -(u)mar, -tar) and innitives (-wanzi,
-(u)manzi, -anna) have dierent sets. Set 2 appears on all monosyllabic mi-verbs show-
ing vowel gradation (ablaut), e.g., - to be, d- to eat, eku- to drink, pp- to
seize, kuen- to strike, kill, kuer- to cut, uek- to slaughter, we- to turn, see
12.4 (p. 189). In addition a few monosyllabic non-ablauting (such as d- to take [see
13.11, p. 218] and wak- to bite [see 13.1, p. 214]) and ablauting i-verbs (such as
au()- to see [see 13.32, p. 229] and pai-/piy- to give) show this pattern. Set 1 ap-
pears on all other verbs. Some verbs switched classes over time and appear sometimes
34. See 4.97 (p. 123) for the paradigm of the participle.
35. Verbal substantives are commonly attested only in nominative-accusative and genitive case forms,
but there may be rare isolated occurrences of other forms, such as the ablative allanumarraza KUB 26.32
i 12 and the dative-locative
D
auwarri KBo 11.1 obv. 8.
36. Plene writings of the innitive ending, -wa-a-an-zi, are extremely rare; only two certain examples
are known to us: pa--wa-a-an-zi KBo 20.8 obv. 9 (OS) and i-pa-an-du-wa-a-an-zi KUB 58.71 obv. right
14 (late NS). A rare form of the innitive ending (-wanta) occurs in i-pa-an-tu-an-ta: maanma -
u ipantuanta iraizzi But when the king nishes making oering KUB 10.21 i 13. This ending is prob-
ably an archaic instrumental ending (see 3.35, p. 77, and p. 125, n. 223), which lends support to the view
that the regular ending -wanzi is an old ablative < *-wen-ti (Jasano 1973 and Melchert 1977: 411, among
others). For another view, deriving it from a locative, see 3.25 (p. 74).
37. On the origin of the ending of the supine see 3.25 (p. 74).
Verb Infection 11.21
186
with -wlmanzi and sometimes with -anna. (See [possibly earlier] arranna and NH or
NS arrumanzi; arrale- was originally a mi-verb but transfers to the i-conjugation in
NH; 13.13, p. 219.)
11.22. Verbs in Set 1 whose stems end in u (especially the nu-causative verbs) take
the 1b endings, e.g., arnumar, arnumanzi, arnuman, arnu(w)ant-. Also i-verbs such as
tarna-, penna-, unna-, arra- to wash, a- to procreate, anna- to conceal, arra-
to divide, unna- to ll, wata- to sin, and mi-verbs of the type peute- to lead
there and wete- to build (see 12.21, p. 198).
11.23. The supine is usually (see 11.24 below) found in the -ke- stem: urzakewan,
karpikewan, uppekewan, memikewan, uikewan, anikewan, mikewan, dakewan,
walannikewan, aruwekewan, tarikewan, punukewan, dukikewan, ukikewan,
wekewan, ippanzakewan, ukkikewan, atrekewan, walukewan, kururiyaikewan,
damekewan, teanikewan, wanukewan, arninkikewan, paigauwan (KBo 5.8
ii 5 ), kappuekewan. In all cases but paigauwan go, wgawan weep, and wikawan
send, the stem is -ke- with e-vocalism.
11.24. Occasionally, instead of the -ke- stem, the verb in the supine takes the
-anna- or -a- stems: uwan, piyanniwan KBo 8.42 rev. 23 (OS), euwan (NS),
iyanniwan (MS), piddanniwan (MS). At least twice it attaches to a reduplicated verbal
root: liluwan dai KBo 32.14 iii 19 (MHlMS); [(na urorn-t)t+-i] kikkiwan di
KBo 3.67 ii 89 (restored by KUB 11.5 obv. 4). See 10.3 (p. 173). Quite rarely the
verb shows no imperfective stem extension (Honer forthcoming: 129): karipwan
dir they began to devour KBo 3.1 i 21 (OHlNS), [(arriyawan)] dir KUB 24.8 iv
2122 (OH?lNS) (restored by KUB 43.70a:2), tarwan di KBo 3.7 iii 25 (OHlNS),
iparruwan da[i (?) . . . ] KBo 14.45:4, annuan di KUB 29.39 iv 8 ; pyawa[n tiyai]
KUB 14.29 + KUB 19.3 i 7 (AM 105). With the exception of the last-cited passage
(Mursili II), these passages were copied from OH or MH archetypes. See too in MHl
MS: apaiyawa[n dir] HKM 7:6. It is possible that in some of these cases the verbs
themselves were felt to be inherently imperfective and needed no external marking (see
Honer and Melchert 2002). Such might be the case with karip- to devour, apaiya-
to scout, and ipar- to spread out, trample. But since others in the above list are
actually attested with the overt imperfective marking (arrike-, tareke-, annike-),
one cannot make this claim for them.
11.22 Verb Inection
00-HittiteGrammar.indb 186 6/5/08 4:16:59 PM
187
Chapter 12
CONJUGATION OF mi-VERBS
12.1. Hittite mi-verbs may conveniently be divided into those with stems ending in a
consonant and those ending in a vowel. Among consonantal stems we may distinguish
those without a sux (so-called root stems), those with inxed -nin-, and those with
suxed -e-.
1
Root stems are subdivided into ablauting and non-ablauting classes (see
3.37, p. 78) for the notion of ablaut). Vocalic stems also may be classied as ablaut-
ing and non-ablauting types.
Consonantal Stems
Root Stems
Ablauting
Stems with ela Ablaut
12.2. Some monosyllabic root stems (ending in a single consonant) show an alterna-
tion between e and a vocalism. This pattern reects PIE ablaut between e-grade and
zero grade respectively (3.37, p. 78). The typical distribution is that e occurs in the
present singular, preterite singular and plural, and imperative second- and third-person
singular and second-person plural, and in the non-nite forms of set 1 verbs (11.21,
p. 185), and a occurs elsewhere. A similar distribution is seen in the ablaut of the im-
perfective sux -ke- (see 12.33, p. 204). But there are exceptions to this pattern, and
competing variants are found, e.g., apteni and pteni you seize, appuen and ppuen
we seized.
12.3. Paradigms of root stems with ela ablaut (- to be, pp- to seize, d- to eat,
eku- to drink):
2
1. Factitive verbs in -a- are mi-verbs only in post-OH. In OH they are i-verbs (13.6, p. 217).
2. The verb e-la- to sleep follows the same pattern. For attested forms see Oettinger 1979:
1819.
00-HittiteGrammar.indb 187 6/5/08 4:17:00 PM
188
Present Indicative
Singular
1 mi pmi tmi ekumi
3
2 i,
4
ti (NH) pi, pti (NH) zi, [zz] ai,
ezatti
5
euki,
6
ekui,
ekutti
7
3 zi, (za)
8
pzi ezazzi, zza(z)zi,
zzi
9
eukzi,
3
ekuzi,
ekuzzi

Plural
1 euwani
10
eppueni, appueni
11
ad/tueni, eduwni,
edue[ni ]
akueni, akuwani,
12

ekueni, ekuwani
2 ptni, apteni,
ptani
13
ezzatteni, azzatni ekutteni
3 aanzi appanzi adanzi, atanzi,
atnzi
aku(w) anzi,
eku(w) anzi

3 4 5 6 7 8 9 10 11 12 13 New n. 3,
3. We follow others in assuming that this stem ends in a labiovelar consonant lg
w
l, not a sequence lgul.
As per Lindeman 1965, the crucial evidence is given by the rst-person plural forms a-ku-e-ni and e-ku-en.
If the stem were *lagu-l ~ *legu-l, the rst-person plural forms would be *a-ku-me-ni and *e-ku-me-en
(1.126, p. 44). The attested rst-person plural forms represent rather lagwenil and legwenl, with loss of
the labial articulation of the labiovelar before lwl. Based on the evidence of the rst-person plural forms,
we may also view the variant OS spellings e-uk- beside e-ku- as attempts to write leg
w
-l, not as evidence for
metathesis (of a nonexistent *legu-l to *leug-l). HED ElI 262 lists e--uk-zi from unpubl. text Bo 2692 V
23, cited by Singer 1975: 90.
4. The OS forms e-e-i and e-e-zi belong to the verb to be sitting, not the verb to be. See 28.30
(p. 362).
5. For the stem as ez-laz- before -t- see 1.125 (p. 44).
6. OH only. See n. 3 above.
7. For intrusive -ti see 11.10 (p. 183).
8. Rare variant e-e-za in Laws 97, copy A (OS). Neu (StBoT 26:39) lists i-e-zi as pres. sg. 3 of to
be, based on [ ] l ku-i-ki i-e-zi [ ] in KUB 34.115 iii 5 (OS; see for text StBoT 26:368). The broken
context leaves the identication uncertain. If it is correct, the unusual spelling might reect a pronunciation
[k
w
iski(y) e:stsi] with an inserted yod, but all solid examples of such insertions involved clitics (see 1.143,
p. 49, end).
9. For the source of the form zzazzi and its pronunciation as le:ts.tsil see 1.125 (p. 44). In later Hittite
the stem in le:ts-l was generalized, and in the process the verb was transferred to the i-conjugation (hence
zzai, zza, zza).
10. Written e-u-wa-ni KUB 44.60 ii 1, HED ElI 285, HW
2
2:93b; no form *eweni (*e-u-e-ni ) is at-
tested yet. On the variant endings -wani and -tani see 1.69 (p. 31). Based on lagwenil (see n. 3 above) and
kuewen < *kuenwen (12.6), we assume likewise leswanil, le:pwenil, etc., with sequences of consonant plus
lwl. However, pronunciations lesuwanil or le:puwenil cannot be denitively excluded (1.143, p. 49).
11. Written e-ep-pu-u-e-ni and ap-pu--e-ni respectively, the latter in KUB 35.18 i 7. See Puhvel, HED
ElI 275.
12. The forms a-ku-e-ni and a-ku-wa-ni we drink with single -k- are not to be confused with ak-ku-
u-e-ni we die with geminate -kk-. In a-ku-e-ni l a-ku-wa-ni the labial component in the labiovelar is lost
before the -w- of the ending.
13. HKM 57:33 (MH), not booked in HED or HW
2
.
12.3 Conjugation of mi-Verbs
00-HittiteGrammar.indb 188 6/5/08 4:17:00 PM
189
Preterite Indicative
Singular
1 eun ppun edun ekun
2 ta
14
pta
15
ezatta ekutta
3 ta pta zta, ezatta, zatta,
ezza, ezzata
eukta, ekutta
Plural
1 euen ppuen, appuen eduen
16
ekuen
2 ten
17
pten
3 eer
18
pper eter ekuer
14 15 16 17 18 19 paradigms: epp-, ed-,
Imperative
Singular
1 lut, lit, aallu
2 p d, zza eku
3 du, tu
19
ptu, pdu zdu, ezzaddu,
ezzadu
ekuddu
Plural
2 ten pten zten, ezatten,
zzaten
ekutten
3 aantu, aandu appantu, appandu adandu, zzandu aku(w)andu
12.4. Verbal substantive: (class 1) euwar, (class 2) apptar, adtar, akuwtar (see
3.31, p. 76; 11.2011.21, p. 185; and 25.2, p. 330). Innitive (class 2): appanna
(with rarer class 1 eppuwanzi), adanna, aku (w)anna.
20
Participle: aant-, appant-l
appnt-, akuwant-. Imperfective: appike-, azzi(k)ke-, akkuke-; this class does not
select the imperfective sux -anna-. See 13.25 (p. 226).
14. KBo 4.14 ii 8, KBo 5.9 i 16.
15. KUB 14.1 rev. 23 (MHlMS).
16. Written e-du-u-en 477lu 13.
17. The second-person plural ending is spelled exclusively -te-en in OS, -te-en or -tn thereafter. We
ignore this dierence in the paradigm entries.
18. Written e-e-er and e-er, both already in OS.
19. OS written e-e-tu. In NS also spelled e-e-tu
4
. See 1.24 (p. 19).
20. KBo 25.72:26 [a-k]u-wa-an-na -e-ek-zi. Despite the question mark on ah? in Konk., the on-line
photo shows without question that this is OS. Although it is the only example with w in OS to date, it must
be admitted.
Conjugation of mi-Verbs 12.4
00-HittiteGrammar.indb 189 6/5/08 4:17:01 PM
190
Stems with uelu Ablaut
12.5. Three Hittite verbs reect PIE e-grade vs. zero-grade ablaut (3.37, p. 78) as
ue vs. u. As attested, these verbs show ue in all nite forms except the present and im-
perative plural third-person. The latter show u, as do the non-nite forms.
21
In the verb
kuen- to strike, hit, kill the nal n of the stem usually
22
is lost when followed by the
w, m, and which begin some verbal endings: kuemi, kuei, kuewen, imperf. kuwake-.
The origin of the variant stem kuennV- remains unclear.
12.6. Paradigms of the monosyllabic mi-verbs containing the sequence -ue- (kuen-
to strike, kuer- to cut, uek- to exorcise
23
):
Present Indicative
Singular Plural
1 kuemi kuermi uekmi,
ukmi
kuennummeni uekuwani
2 kuei,
kuenti,
kueti
kuenatteni
3 kunzi,
kuenzi
kurzi
24
ukzi,
kzi
kunanzi,
kuennanzi
kuranzi kanzi
24 25 kuen: preterite paradigm
Preterite Indicative
Singular Plural
1 kuenun,
kuenunun
kuerun
25
kuewen,
kuinnummen
ugawen
26
2 kuinneta,
kuenta
27
kuenten
3 kunta,
kuenta,
kuinta
kurta,
kuerta
uekta kuener,
kunner
28
kuerer
21. Strictly speaking, in the verbs uek- to slaughter and uek- to exorcise ue- represents an initial
sequence of lhl and lwl, while kuen- to strike and kuer- to cut probably show a single initial labiovelar
lk
w
-l, but for all practical purposes all these verbs pattern alike.
22. A dierent treatment is seen in kuennummeni (NH) for expected *kueweni, and kuinnummen (NH)
for the attested variant kuewen.
23. The verb uek- to slaughter shows forms homophonous with those of to exorcise except in the
-ke- stem, where to exorcise has consistent geminate k versus a usual non-geminate in to slaughter (but
note ukkikanzi they slaughter in KUB 9.3 iv 7). Attested forms of the latter verb are: ukmi/uekmi,
ukziluekzi, kzilukzi, ukku[eni ?], kanzilukanzi, pret. sg. 3 uktaluekta, uekdu, ganna,
ugant-lkant-, uganniwan.
24. The form ku-er-ri in KUB 24.12 iii 19 is faulty. The online photo in Konk. shows extraneous wedges:
the RI is apparently written over an erased sign.
25. KBo 17.23 obv? 6 (OS) ku-e-ru-u[n?] or (less likely) ku-e-ru-e'-n[i'], ku-e-ru-un 315lu 6, cited
by HED K 213. The status of the pret. sg. 1 kurun in KBo 10.2 ii 48 remains unclear. See Oettinger 1979:
119 vs. HED K 21718.
12.5 Conjugation of mi-Verbs
00-HittiteGrammar.indb 190 6/5/08 4:17:02 PM
191
Imperative
Singular Plural
2 kueni,
kuenni
kuenten
3 kuendu,
kuindu
kurdu uekdu kunandu kurandu
26 27 28

New n. 26
12.7. Verbal substantive: kue(n)numar, ktar. Innitive: ganna, kuranna,
kunanna, and kuennummanzi (3.31, p. 76; 11.2011.21, p. 185; and 25.2, p. 330).
Participle: gant-, kunant-, kurant-. Imperfective: kuwarke- (ku-wa-ar-a-ki-iz-zi
169lx i 7 cited in Oettinger 1979: 119), kureke- (from kuer-), ku(w)ake- (from kuen-),
kkike-lukkike-. See also (u)wanikanzi they fuck from wen- to fuck. The im-
perfectives kuwarke- and kuwake- are the older forms, of which kureke- is a NH
regularization. The kuwar-, kuwa(n)- stem is the outcome of the zero-grade form of
the root, as is uk-lk- (see 3.37, p. 78; 12.2, p. 187). This class does not select the
imperfective sux -anna- (see 13.25, p. 226) except for uek- to slaughter, which
shows a supine uganniwan.
Non-Ablauting
12.8. Many mi-conjugation consonantal root stems are non-ablauting. These include
na()- to fear and virtually all monosyllabic stems ending in a sequence of two con-
sonants: wal- to strike, an- to seek, ark- to perish, itark- to get sick, karp- to
lift, link- to swear, warp- to bathe, kurk- to store, alk- to knead, par - to chase,
arp- to associate (with). For syllabary limitations in expressing sequences of three
consonants see 1.11 (p. 12). In order to express precisely the spelling of verbs with
such sequences we have used narrow transliteration instead of the usual broad transcrip-
tion in the following paradigms.
Present Indicative
Singular
1 na-a-mi wa-al-a-mi a-an-a-mi kar-ap-mi
2 na-a-ti wa-la!-a-i a-an-a-i,
a-an-a-ti,
a-an-a-ti,
a-an-a-at-ti,
a-na-a-ti
kar-ap-i
3 na-a-zi, na-a-i,
na-i,
na-a-i-i
29
wa-la-a-zi,
wa-al-a-zi,
u-wa-al-a-zi
a-a-zi,
30
a-an-a-zi
kar-ap-zi

29
26. This form is analogical to the preterite rst-person plural forms of the -(i)- verbs, e.g., andwen.
27. kuinneta KUB 17.3 iii 4 (NH), kuenta KUB 14.1 rev. 23 (MHlMS).
28. Written ku-e-ne-er and ku-e-ner, ku-(e)-en-ne-er and ku-en-ner.
29. Also written logographically as nt-i. Some of these forms were booked as imperatives in CHD
LN s.v. But since the sg. 3 na-a-i, attested already in the early language, probably indicates that this
Conjugation of mi-Verbs 12.8
00-HittiteGrammar.indb 191 6/5/08 4:17:03 PM
192
Present Indicative
Plural
1 na-a-u-u-e-ni,

nt-u-ni
31
wa-al-u-u-e-ni,
wa-al-a-u-e-ni,
wa-al-u-wa-ni
[a-a]n-u-e-ni
32
2 na-a-te-e-ni wa-al-a-ta-ni a-an-a-te-ni,
a-an-a-te-ni,
a-an-a-at-te-ni
3 wa-al-a-an-zi a-an-a-an-zi,
a-an-a-a-an-zi,
a-a-a-an-zi,
a-a-an-a-an-zi,
a-an-a-a-an-zi
kar-pa-an-zi,
kar-ap-pa-an-zi,
kar-(ap-)p-an-zi
30 31 32 33 walah preterite paradigm
Preterite Indicative
Singular
1 na-a-u-un,
na-a-u-un
wa-al-u-un a-an-u-un,
a-a-u-un,
a-an a-u-un
kar-ap-pu-un
3 na-a-ta wa-al-a-ta a-a-ta,
30

a-an-a-ta,
a-an-na-a-ta
ka-ra-ap-ta,
33

kar-ap-ta
Plural
1 wa-al-u-u-en
2 wa-al-a-tn a-an-a-tn kar-ap-tn
3 wa-al-e-er a-an-e-er kar-p-er
Imperative
Singular
2 wa-al-a a-a-a, a-an-a kar-ap
3 wa-al-a-du a-a-du kar-ap-du
end marker
verb was originally a i-verb, ni in lta ni is the regular indicative third-person singular with the
i-conjugation ending in an impersonal construction: Let there not be a fearing to you. This interpreta-
tion avoids both the irregular use of imperative + l and the unparalleled use of a singular enclitic personal
pronoun as a reexive.
30. For the loss of -n- see 1.135 (p. 46).
31. KUB 5.2:6, 10, 15 (div., NH); could be an error for nt-u-e-ni or an example of a real ending -uni ;
see p. 181, n. 15, with references.
32. HKM 48:25 (MHlMS), see Honer 1997b. Accidentally omitted in CHD 163 top left.
33. ka-ra-ap-ta KUB 36.49 i 8 (OS).
12.8 Conjugation of mi-Verbs
00-HittiteGrammar.indb 192 6/5/08 4:17:04 PM
193
Imperative
Plural
1
2 wa-al-a-tn a-a-a-tn,
(a-an-a-at-tn)
kar-ap-tn
3 wa-al-a-an-du a-an-a-an-du
12.9. Verbal substantive: na-a-u-u-wa-a and na-a-u-u-wa-a (gen.). Inni-
tive: wa-al-u-wa-an-zi, a-an-u-wa-an-zi, kar-pu-wa-an-zi. Participle: naant-
34
,
walant-, anant-, karpant-. Imperfective: naeke-, walileke-, walanna/i-,
anike-, aneke-, karpileke-.
12.10. The verb ar (k)- to hold, have is exceptional in that it deletes the nal
-k- before endings beginning with a consonant,
35
while the verbs maz- (< *mat-)
36
and
ipart- show the same dissimilation as et- (1.125, p. 44):
37 38 39 40
41 42 43
Present Indicative
Singular
1 armi
2 ari, arti ma-za-at-ti
3 arzi, arza
37
mazza(z)zi, manzazzi,
mazze, mazzi
iparzazi, iparzizi,
38

iparzai ,
39
iparzaizzi,
ipartzzi
40
Plural
1 arueni,
41
arwani mazzueni
42
2 arten
43
3 arkanzi
34. Sg. neut. nan.
35. For divergent historical accounts of this pecularity see Watkins 1970: 71 and Cowgill cited by Eich-
ner 1975: 90. False HED H 156.
36. The verb stem is written maz- here and in the CHD, because the dissimilation has spread in this word
to all environments (e.g., pl. 1 ma-az-zu-u-e-ni), leaving no attested example of *mat-. The appearance of
mazze in OS shows that in this verb the spread of the stem mazz- and transfer to the i-conjugation began
already in OH. See p. 188, n. 9.
37. KBo 9.73+ obv. 12 (OS).
38. Written i-pa-ar-zi-zi KUB 4.72 rev. 5 (OS).
39. The iparzazi writings are attested in MH. The i-conjugation form iparzai does not occur before
NH.
40. For the present third-person singular forms see Khne and Otten 1971: 42.
41. Written ar-u-e-ni, ar--e-ni, and ar-u-en, not *a-ru-e-ni and *a-ru-en.
42. Written ma-az-zu-u-e-ni.
43. KBo 22.1:31 (OS). This example shows lengthening of the nal syllable vowel as part of interroga-
tive intonationlstress (see 27.2, p. 348).
Conjugation of mi-Verbs 12.10
00-HittiteGrammar.indb 193 6/5/08 4:17:05 PM
194
Preterite Indicative
Singular
1 arkun iparzaun
44
2 arta mazzata
45
iparzata
46
3 arta ma(z)zata iparzata, iparza
Plural
1 aruen
41
2 arten
3 arker iparter, iparzer
Imperative
Singular
2 ark
3 artu,
47
ardu iparzadu, ipartieddu
Plural
2 arten
3 arkandu
44 45 46 47 hark paradigm end
12.11. Verbal substantive: gen. mazzuwa. Participle: iparzant-.
12.12. In the case of tarku- (ltark
w
l) to dance, whirl and watku- (lwatk
w
l) to leap,
the stem ends in a labiovelar k
w
, not a consonantal sequence lkwl. We nd the same kind
of spellings and the same delabialization of the labiovelar as in eku- to drink (12.3,
p. 188, with footnotes).
48
12.13. Root stems ending in -ink- show forms with and without deletion of the -n-
(see 1.135, p. 46) and various ways of spelling l-nk-l and l-ng-l when it is retained.
Forms of link- to swear: pres. sg. 3 li-ik-zi, li-in-ga-zi, li-in-ik'-zi, pl. 1 li-in-ku-u-e-ni,
li-ku-wa-an-ni,
49
pl. 3 li-in-kn-zi, pret. sg. 1 li-in-ku-un, sg. 3 li-ik-ta, li-in-ik-ta, li-in-
kat-ta, pret. pl. 1 li-in-ku-en, le-en-ga-u-en,
50
pl. 3 li-in-ke-er, imp. sg. 2 li-i-ik, li-in-
ki, sg. 3 li-ik-du, pl. 2 li-ik-te-en, le-en-ik-tn, pl. 3 li-in-kn-du, part. linkant-, imperf.
linkike-. The verbal substantive and innitive are unattested.
44. KUB 25.21 iii 14 (HW 90, CTH 524 Kaka treaty). A i-conjugation form (for expected *ipartun), as
is pret. sg. 3 iparza, but these are shown to be secondary by iparzazi, iparzata (OS) and iparzadu.
45. KUB 14.1 obv. 62 (MS).
46. KUB 19.49 i 6 (NH).
47. KUB 31.81 obv. 3 (OS).
48. But the appearance in NS of pret. sg. 3 watkut for earlier watkutta suggests that lk
w
l was eventually
reanalyzed as lkwl in this word (Oettinger 1979: 237).
49. This is a Luwianized form; see Oettinger 1979: 13.
50. See p. 191, n. 26.
12.11 Conjugation of mi-Verbs
00-HittiteGrammar.indb 194 6/5/08 4:17:06 PM
195
12.14. In the case of tak- to put together, construct the combination of the stem-
nal sequence of stop plus followed by the consonant of various consonantal endings
leads to an inserted vowel, either e or i (1.81, p. 34), producing a preconsonantal stem
takkeli- beside a prevocalic stem tak-. Competing eorts to generalize one stem or
the other lead to instances of prevocalic takkeli- (e.g., takkeanzi) and preconsonantal
tak- (e.g., taggateni [taks.steni]). Preconsonantal tak- is necessarily written as tg-ga-
a- (1.11, p. 12). For the geminate -- in takkeli- see AHP 15052.
Present Indicative
Singular Plural
1 tagga[mi']
51
2 takkii, taggai
52
takkiteni, takketeni, taggateni
3 takki[zi ],
53
takkizi, takkezi,
takkzi, takkizzi
54
takanzi, takkianzi, takkeanzi
Preterite Indicative
Singular Plural
3 takkita, takketa, taggata taker, takker
51 52 53 54
Imperative
Singular Plural
3 taggadu
12.15. Innitive taku(w)anzi . Supine takkikewan. Participle takant-. Imperfective
stem takkike-, takkeke-.
12.16. Another group of mi-verbs have stems that end in a sequence with w as its
nal component: tarw- to be able, be in chargelcontrol, anw- to roast. The w
becomes u before endings beginning with a consonant. In the verb tarw- metathesis
to taru- sometimes occurs. We assume a sequence lhwl in these cases (not lh
w
l) on
historical grounds and thus also assume a real metathesis in taru- as opposed to eukzi
and tarukzi (see 1.140, p. 48; and 12.12, p. 194). Proof for this claim is lacking,
since the crucial forms with endings having initial lwl are not yet attested, where one
would expect pres. pl. 1 `anumeni, pret. pl. 1 `anumen, etc. (see p. 188, n. 3 for
the reasoning).
51. Restored arbitrarily as tg-ga-a [-i ] in KBo 5.3 iv 43 (= KBo 5.12 + KBo 5.3 iv 59) (ukk.) by
SV 2.
52. KBo 3.1 ii 15, treated as sg. 3 (by Oettinger 1979: 21718 as an error, by Neu 1974: 91 as assimi-
lated *takzi ) when restoring le-e ku[-i-ki ] in the preceding line. But if one restores le-e ku[-it-ki] there,
tg-ga-a-i can easily be taken as the regular and expected pres. sg. 2. For the second-person formulation
see KBo 3.1 ii 4245.
53. KUB 36.106 rev. 4 (OS).
54. See 1.130 (p. 45).
Conjugation of mi-Verbs 12.16
00-HittiteGrammar.indb 195 6/5/08 4:17:07 PM
196
Present Indicative
sg. 3 a-an-u-uz-zi, ta-ru-u-zi, tar-ru-u-zi, tar-u-uz-zi
pl. 3 a-an-u(-wa)-an-zi, a-an-u-un-zi,
55
tar-ru-u-a-an-zi
Preterite Indicative
sg. 1 ta-ru-u-u-un
pl. 3 tar-u-e-er
Imperative
sg. 3 tar-u-du, tar-u-id-du
pl. 3 a-an-u-u-wa-an-du
12.17. Participle: a-an-u(-u)-wa-an-t, a-an-u-un-t,
55
tar-u-an-t.
56
Axed Stems
Stems with Inxed -ni(n)-
12.18. Paradigms for stems with nasal inx -ni(n)-. For the meaning of the inx see
10.17 (p. 179). These verbs show the form -nink- before endings beginning with a
vowel or lwl and -nik- elsewhere.
57
58
Present Indicative
Singular
1 arnikmi arnikmi ninikmi
2 arnikti niniki itarniki
3 arnikzi arnikzi nnikzi, ninikzi itarnikzi
58
Plural
1 arninkueni nininkuweni
2 arnikteni arnikteni ninikteni
3 arninkanzi arni(n)kanzi,
arnenkanzi
nininkanzi
Preterite Indicative
Singular
1 arninkun arninkun nininkun
2 arnikta
3 arnikta arnikta ninikta
55. See 1.76 (p. 32).
56. KBo 17.105 iii 3 (MHlMS).
57. This ablaut pattern reects a considerably modied PIE ablaut with e-grade inx `-ne- alternating
with zero-grade inx `-n- (see 3.37, p. 78).
58. For the rare form i-tar-ni-ik-za KBo 40.272 left 5, probably for itarnikzi, see 11.8 (p. 182).
12.17 Conjugation of mi-Verbs
00-HittiteGrammar.indb 196 6/5/08 4:17:07 PM
197
Plural
1 arninkuen itarninkuen
3 arninkir arni(n)ker nininkir
harnink: preterite paradigm
Imperative
Singular
2 arnik ninik
3 arnikdu arnikdu
Plural
2 arnikten ninikten
3 arninkandu arninkandu nininkandu
harnink end
12.19. Verbal substantive: arninkuwar, arninkuwa (gen.), nininkuwa (gen.).
Innitive I: arninkuwanzi, arninkuwanzi, nininkuwanzi. Participle: arninkant-,
arninkant-, nininkant-.
Non-Ablauting Stems in -e-
12.20. Verbs in -e- (Oettinger 1979: 23855). For the meaning of the sux see
10.13 (p. 177).
5960
Present Indicative
Singular
2
59
`appineti (o.ftxt-ti), idalawti, `kardimmeti (ftxt.ftxt-e-ti )
3 innarawezi, idlawezi, adukizi, kallarezi, `kunnezi (zo-nezi ), makkezi,
mi(ya)u(wa)ntezi, parkuezi, dannatezi, tepawezi, waruezi, orxr-aarezi, etc.
Plural
1 [ia]arweue[ni]
2 idlaweteni, parkueteni
3 [innar ]aweanzi, idlaweanzi, maninkueanzi, tepaweanzi
Preterite Indicative
Singular
1 parkun, taruileun, oin-liun, adduliun
3 araweta, idlaweta, kartimmta
Plural
2 nakke [ten]
60
3 araweer, mayateer, mareer, alleer
59. The replacement of -i by i-conjugation -ti appears to have been systematic in this class.
60. CHD LN 371 (KUB 14.13 i 46).
Conjugation of mi-Verbs 12.20
00-HittiteGrammar.indb 197 6/5/08 4:17:08 PM
198
Imperative
Singular
2 miye
3 adduledu, mi(y)tu, parkutu
Plural
3 makkeandu
Vocalic Stems
Ablauting

Stems with ela Ablaut
12.21. The following mi-verbs are univerbations with a prehistoric root *dheh
1
- to
put (see 12.48, p. 211, and Oettinger 1979: 12526): uwate- to lead here, peute- to
lead there, wete- to build, werite- to be afraid. The descriptive ablaut ela (reecting
PIE e-grade versus zero-grade; 3.37, p. 78) is fairly predictable:
61 62
63 64
Present Indicative
Singular
1 uwatemi putemi witemi, wedai
61
weritemi
2 uwatei peutei wedai werizzati
3 uwatezzi, uwatezi,
uwadazzi
peute(z)zi,
pute(z)zi,
puttezzi, ptezi
wetezzi, [w] itezzi,
wedai
weritiz[zi],
urizzizz[i ]
62
Plural
1 uwateweni,
uwatewani,
uwatummeni
63
putummni wedumni
2 uwatetteni,
uwatettani,
uwadateni
putetteni,
put ettani
3 uwadanzi pudanzi,
peudanzi,
peutanzi,
peutenzi
64

wedanzi waritanzi,
werita[nzi ]

61. The stem wete- to build comes to be inuenced by the originally separate verb wileda (i)- to bring
and vice versa.
62. The forms werizzati and urizzizzi (if the latter indeed belongs here) suggest a NH reinterpretation of
the stem as *werit(t)-. See the treatment of d- (ezzazzi) to eat (1.125, p. 44), ipart-, *mat- (see 12.10
and n. 36, p. 193) and at- (azzizzi ).
63. See 1.126 (p. 44).
64. Spelled [ p ]-e-u-te-en-zi KBo 25.50 left edge 7, not with the -tn- sign.
12.21 Conjugation of mi-Verbs
00-HittiteGrammar.indb 198 6/5/08 4:17:09 PM
199
Preterite Indicative
Singular
1 uwatenun peutenun wetenun,
wedaun, wetun
2 uwatet peutet weda
3 uwatet peutet, peute wetet, weda weriteta,
wiriteta,
werit[e- . . . ]
Plural
1 uwatewen wetu(m)men
2 peutetten
3 uwater, uwatr peuter weter
65uwatet alongside uwate
Imperative
Singular
2 uwate, uwati,
uwatet
65
peute, peuti wete
3 uwateddu peuteddu weteddu, wedau
Plural
2 uwatetten,
uwatatten,
uwatitten
peutetten wetatten
3 uwadandu peudandu wedandu
12.22. Verbal substantive: wetu(m)mar (-e-tu
4
-mar), `uwatummar. Innitive: we-
tu(m)manzi (-e-tu
4
-ma-an-zi), `uwatu(m)manzi(?). For the (m) in these forms see 1.126
(p. 44). Participle: peudant-, wetant-, widant-. Imperfective: weteke-, weriteke-. No
iterative of uwate- is yet attested, and peuteke- (NH) from peute- is rare.
12.23. Other mi-verbs with vocalic alternating stems ending in ela are of diverse
origin:
66
arkuwe- to make a plea, ulle- to repulse, turn back, defeat (an enemy),
reverse, cancel (an agreement), arne- to besprinkle, ipa(r)re- to spread (something)
out, trample(?), kappuwe- to count, lukke- to ignite, arre- to divide, arkuwe- to
put on (footwear), ulle- to be(come) wanton, disrespectful, ambitious, uwaye- to
look, uwe- to pushldrive away, forfeit, repudiate, divorce, tye- to steal, duwarne-
to break, wae- to clothe, zinne- to nish. In addition to the preceding verbs which
65. See p. 182, n. 19.
66. We prefer to avoid the label simple thematic class (Oettinger 1979: 24, 257314). It is debatable
whether any attested Hittite verbs with stems ending in ela reect PIE verbs that t this denition. It is cer-
tain that many do not: e.g., kappuwe- to count (see Oettinger 1979: 33233), zinne- to nish (Oettinger
2002: p. xx), wae- to clothe (Melchert 1984b: 3135; Oettinger 2002: p. xx) and ulle- to be(come)
wanton, disrespectful (Melchert 2005a). But all these form a single class in Hittite. We use alternating
stems ending in ela merely as a neutral descriptive device.
Conjugation of mi-Verbs 12.23
00-HittiteGrammar.indb 199 6/5/08 4:17:09 PM
200
have no sux from the Hittite point of view, verbs with the sux -e- (for whose mean-
ing see 10.11, p. 177) also are probably inflected as alternating stems ending in
ela.
67
Many of these verbs follow this conjugational pattern only in OH, or in OH and
MH, changing in NH to other patterns (mi-verbs in stem -i-, i-verbs in -a-, etc.). For
some members of the paradigm we have attested only examples of the latter classes,
not the predicted forms of the alternating vocalic stems in -e-l-a-. Footnotes below oer
only a selection of such secondary forms. For a more complete survey see Oettinger
1979: 261312.
12.24. The basic vocalic stems show an alternation between e and a that only partly
reects prehistoric ablaut, and in some cases we nd both e and a (see 3.37, p. 78, and
the idealized schema in Oettinger 1979: 25960):
68
active pres. -ami -elai -ela(z)zi -aweni,
-ewani
-atteni -anzi
active pret. -anun `-e(') -et `-awen
69
-elatten -er
active imp. `-allu `-a
70
-attu `-aweni -elatten -antu
69 70
12.25. Examples for the active voice of the mi-conjugation of alternating vo-
calic stems in ela.
71 72 73 747576
Present Indicative
Singular
1 arnami,
71
zinnami
72
2 arani, ullai,
73
ullei,
74
waai
75
3 arwezzi, aiwantezzi, ullezzi,
76
ullzzi,
77
ullazzi,
78
lukkezzi,
79

paprezzi, arrzzi,
80
ullzzi, waezzi,
81
zinne(z)zi
82
67. We base this claim on the fact that ulle- to be(come) wanton reects the same sux (see Melchert
2005a: 96, after Watkins 1973), and this stem does show ulla- in the participle ullant- (see Oettinger 1979:
29192). Assured forms with stem ending in -a- from the denominative stems in -e- are not yet attested.
The stem paprant- may be the participle to papre- to be proven guilty, but it could also be an adjective in
-ant- derived directly from a base adjective `papra- (see 2.26, p. 56).
68. In view of the diverse origins of this class, it may or may not be signicant that the variants with -e-
in the pres. sg. 2, pres. pl. 1, and pret. and imp. pl. 2 all belong to stems reecting prehistoric `-eh
1
-.
69. The predicted preterite rst plural form is not yet attested. We nd only the replacement form -umen
of the i-verbs in -a-: u-ul-lu-mi-en, arrumen.
70. Contra Oettinger 1979: 259 and AHP 183 the expected outcome of the imv. sg. 2 is *-a < *-e. The
attested example arri (NH) shows the form of a i-verb in -a- (see tarni to tarna-), while waiya (NH)
belongs to a stem waiya-.
71. ar-na-mi KBo 40.272 6, cited when unpublished as 1808lc 5 by Oettinger (1979: 307).
72. NH zinnai.
73. KUB 37.223 A7 (OS).
74. KUB 36.114 iiliv 6 (OHlMS or MHlMS). NH ulliyai.
75. NH waiyai.
76. u-ul-le-ez-zi Laws 173a, copy aa iii 24 (OS).
12.24 Conjugation of mi-Verbs
00-HittiteGrammar.indb 200 6/5/08 4:17:10 PM
201
Present Indicative
Plural
1 uuewani,
83
arraweni, waaweni,
84
zinnaweni
85
2 ullatteni, arratteni
3
ullanzi, lukkanzi, arranzi,
86
zinnanzi
Preterite Indicative
Singular
1 ullanun,
87
[lu] qqanun
88
3 ullet,
89
urnit, lukket, arret,
90
ullt,
91
zinnit
Plural
1
92
2 uiweten,
93
uwatten, ulletten
94

3 uller, marer, arrer
95

Imperative
Singular
3 ulladdu
96
Plural
2 zinnatten
3 arrandu, wandu, waandu
77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 mi-conjucation paradim hulla preterite paradigm
77. NH u-ul-la-(a-)i and u-ul-li-ya-az-zi.
78. u-ul-la-az-zi KUB 37.223 C4 (OS).
79. NH forms lu-uk-zi, lu-ki-i-zi belong to an intransitive verb to become bright; lu-uk-ki-i-zi he sets
re to in Laws 100 is a NH scribes error for correct lu-uk-ke-ez-zi.
80. NH has the additional forms ar-ra-(a-)i, ar-ri, ar-ri-ya-zi, ar-ri-ya-iz-zi.
81. NH and NS also waiezzi, waiyaz(z)i, and waiyaizzi.
82. NH zinni.
83. []u-u-e-wa-ni KBo 3.46 rev. 39 (OHlNS).
84. NH also has waueni, which cannot possibly belong to the inection in -e-l-a- and therefore has to
belong to some other inection class.
85. NH zinnummeni.
86. ullnzi in KBo 43.77: 7 cited when unpublished as 38lg by Oettinger (1979: 291) is irrelevant
here, since it is shown by duplicates to be read [ p ]r-u-ul-la-a-an-zi. Contra Oettinger, ullnun in KBo 5.8
ii 2 belongs to ull (i )- to give (as) a hostage and gives no evidence for an a-stem variant of ulle-.
87. OHlNS u-ul-li-ya-nu-un.
88. NH arraun and zinnaun.
89. OHlNS u-ul-li-i.
90. MHlMS also arra, and NH arriyt.
91. NH also ulliyat and ullit.
92. See p. 200, n. 69.
93. Written u-i--e-te-en.
94. KUB 4.1 ii 11 (MHlNS).
95. NH has additional form ar-ri-i-e-er.
96. NH zinnau.
Conjugation of mi-Verbs 12.25
00-HittiteGrammar.indb 201 6/5/08 4:17:11 PM
202
12.26. Innitive: waauanzi.
97
Verbal substantive.
98
Participle: lukkan, arran,
waan. Imperfective stems: arrake-, arreke-,

arrike-, waeke-, waike-.
12.27. In some cases the dearth of evidence from OS or even MS makes it di-
cult to determine the original stem class of the verb. The verbs ipar(re)- to kick,
tread (on), spread out(?) and mall(e)- to mill may have originally been consonantal
i-verbs (thus Jasano 2003: 7778 contra Oettinger 1979: 26671, 27779). Their
currently known forms, however, are very similar to those of mi-verbs with alternat-
ing stems ending in ela, and they are thus cited here. Forms of ipar(re)- (all NS un-
less marked): ipar(a)i, iparratti, iprilipari, iparrezzi (MHlMS!), iparriyazzi,
iparanzi, iparranzi, iparun, iparraun, iparrer, iparut, iperten, iparrandu,
iparriyawar, iparrant-, iparruwan. Forms of mall(e)- (all examples are in NS):
99

malli, mallai, mallezzi, mallazi, malliyazzi, mallanzi, mallanun, mallt, mlla,
100
mal-
landu,

malluwanzi, malluwar, mallan, mallike-.
Ablauting Stems in -iye/a-
12.28. Ablauting verbal stems in -(i )yela- (henceforth simply -iya-) include basic verbs
such as iya- to do, make, tiya- to step, enter, wemiya- to nd, meet, uett(iya)- to
draw, drag, markiya- to reject, aliya- to prostrate oneself, ariya- to dig, tya- to
steal, and also denominals such as aariya- to rake(?), appariya- to sell, la()iya-
to campaign, lam(a)niya- to name, urkiya- to track down, and zaiya- to battle.
We frequently nd both a stem in -iya- and some other stem for the same verb. In many
cases the stem in -iya- appears to be the innovation, but in some this status is unclear. The
e-grade forms with l-ye-l (-Ci-(i-)e-) are more frequent in the older language and gradu-
ally disappear from use during NH.
101
When Hittite began to borrow Luwian i-verbs
with singular third-person forms in -iyai (which did not exist in Hittite), these verbs were
given singular third-person forms in -iyaizzi (pret. sg. 3 -iyait), and these endings then
came to be used occasionally with native Hittite stems in -iya- (already in OS urkiyaizzi
tracks down). One also nds a few preterite third-person plural forms in -iyaer.
102
97. wa-a-a-u-an-zi KUB 54.76:5, cited as Bo 1709 by Oettinger (1979: 301). In NS we nd
arrumanzi with the form of a i-verb in -a-.
98. The expected verbal substantive in *-awar is not yet attested. We nd only forms in -umar with the
form of a i-verb in -a-: arrumar, ullumar.
99. The restored form m[a-al-l]a-a-i listed in CHD LN 125 from KBo 15.35 + KBo 15.33 i 11 (MS)
does not exist. Read rather with Glocker (1997: 60; line 14) -[d ]a-a-i.
100. So Oettinger 1979: 278 n. 38; but probably to be interpeted as noun ml with and mlla and
ml (CHD LN 124).
101. See on this point, among others, Carruba 1966b: 7980; Houwink ten Cate 1970: 16; Melchert
1977: 32; and Oettinger 1979: 344.
102. On the genesis and spread of the forms in -iyai- see Melchert 2005b revising Oettinger 1979:
38285. An example of a Luwian i-verb with third-person singular in -iyai is taparriyai, altered to Hittite
taparriyaizzi, whence also pret. pl. 3 daparriyar.
12.26 Conjugation of mi-Verbs
00-HittiteGrammar.indb 202 6/5/08 4:17:12 PM
203
12.29. Paradigms for the unmixed mi-conjugation stems in -iya- (see also 12.50
12.51, p. 212): iya- to do, make, tiya- to step, enter, wemiya- to nd, meet, uittiya-
to draw, drag, attract. Another common verb belonging to this class is aniya- to do,
perform, make.
103
Present Indicative
Singular
1 iemi,
103
iyami,
iya(m)mi
tiyami wemiyami uittiyami
2 iei, iyai tiyai wemiyai uittiyai
3 i(z)zi, iya(z)zi,
iyaizzi
tiyezzi, ti(y)zzi,
tiezi, tiyazi
wemiezzi,
wemizzi, wemiezi,
wemiya(z)zi
uittiezzi,
uittiyazi,
uittiyai
Plural
1 iyaweni tiyaweni wemiyaweni
2 iyatteni tiyatteni uittiyatteni
3 ienzi, iyanzi ti(y)enzi, tiyanzi wemiyanzi uittiyanzi
tiya: preterite paradigm
Preterite Indicative
Singular
1 iyanun, iyaun tiyanun wemiyanun uittiyanun
2 iya, iyat tiyat
3 it, iyat tt, tiyat wemi(e)t, wemiyat uitti(e)t, uittiyat
Plural
1 iyawen tiyawen wemiyawen uittiyawen
2 iyatten
3 ir tiyr, tiyer wemier, wemiyar
104
uittier
104
Imperative
Singular Plural
1 iyallu
2 iya tiya, tya uitti iyatten tiyatten
3 iddu,
iyaddu
105

tiyaddu iendu,
iyandu
tiyandu wemiyandu
105
103. i-e-[m(i )] StBoT 8 ii 22 (= StBoT 25 #3 ii 9) (OS).
104. See p. 181, n. 17.
105. Rarely also i-ad-du.
Conjugation of mi-Verbs 12.29
00-HittiteGrammar.indb 203 6/5/08 4:17:14 PM
204
12.30. Verbal substantive: iyawar, tiyawar, uittiyawar. Innitive I: iyawanzi, tiyawa-
nzi wemiyawanzi, uittiyawanzi; Innitive II: tiyanna. Participle: iyant-, ient-, tint-, tiyant-,
uittiant-, uittiyant-.
Stems in -ke-
12.31. Occasionally, a -ke- stem functions as the basic form of the verb (see 24.25,
p. 323): e.g., ike- to smear, anoint. Most stems in -ke- are derived imperfective stems,
for whose use see chapter 24 on verbal aspect. Their formation is as follows. Base verbs
with ela ablaut in their stems add -ke- to the a-form: akkuke- (`ak
w
-ke-) < eku- to
drink; azzikke- (`ad-ke-) < d- to eat; appike- < pp- to seize). Those with welu or
walu use the u-form: ukkike- < uek- to exorcise; urzakke- (`urtke-) < uwart- to
curse). The verbs kuen- and wen-are exceptional: kuwake- < `kwanke-; uwanikke-
< `wan-ke-. i-conjugation verbs with -ai-l-i- use -i-: pike- < pai- to give. au()-l
u- uses u- (uke-). i-conjugation verbs with -a-l-i- use -i-: memike- < mema- to speak.
Base stems in -i- are usually reduced to -e- or -i-: arwike- < ar (u)wi- to do rever-
ence; those in -iyela- to -i-: wemike- < wemiya- to nd. Verbs that have acquired a
non-original stem in -a- revert to the original base: unuke-
106
< unu(wa)- to decorate,
ornament. The stem dai- to place is usually reduced to just d-: zikke- < `d-ke-. For the
-ke-stems of anna- to judge and tarna- to let see 1.81 (p. 34) and 1.120 (p. 43).
12.32. Although the vowel which connects a verbal base ending in a consonant to the
-kela- sux is regularly an -i- in the old language (1.81, p. 34): wekike-, ukkike-,
one also nds -e- later. The confusion between -i- and -e- also is extended to the
-kela- stems formed to verbs in -iya- (Melchert 1984b: 134 35, 14750): tike-lteke-
< tiya- and annike-lanneke- < aniya- (note the special gemination in the -ke- form
of aniya-).
12.33. The -kela- sux shows ela ablaut similar to the stems in -iyela- (12.28
12.29), but the distribution of ela in -kela- has not yet been studied diachronically in
as much detail as in -iyela-. The matter is further complicated by the polyvalence of
the ou sign, which can be read either kat or kilet
6
. The kilet
6
value is attested in Hit-
tite almost exclusively in these -ke- forms. If future study should establish what we
feel the present incomplete assemblage indicatesnamely, that in resolved writings of
the endings where the sign ou is otherwise used those found in pre-NH show an a
vocalization, while those in NH show e, then it would be possible to conclude that the
NH forms written with ou should be read ket
6
while the pre-NH examples should be
read kat. In order not to cloud the issue, forms written with the ou sign are assigned an
indeterminate vocalization in the table below. The following combinations of thematic
vowel and endings are known:
107
106. The form -nu-u-ke-ez-zi is found in KBo 38.265 i 13 . The non-iterative form -nu-uz-zi ibid. 11
shows that in this case unu- was the original base stem and unuwa- secondary.
107. The following table shows the regular inection of -ke- stems. Some forms are transliterated in
order to show unambiguously the nature of the vowel following -k-. In view of plene spellings such as
12.30 Conjugation of mi-Verbs
00-HittiteGrammar.indb 204 6/5/08 4:17:15 PM
205
Present Indicative
Singular
1 -kemi dakmi, dakemi, pekemi, iikmi, memikmi
-kami andiqami, igmi
2 -kei akkuki, akkukei, dakei, pekei, pikei, ukei
-kelatti
108
u-kelat-ti
3 -kezzi dakezzi, pekezzi, kezzi, akkukezzi
Plural
1 -kewani akkukwani, akkukeuwani, dakewani
-keweni dakweni, anikeweni
-gaweni dagaweni, pigaweni
2 -ketteni da-a-ke-et-te-ni, -e-ki-i-ke-et-te-ni
-kettani ak-ku-u-ke-et-ta-ni, up-p-i-ke-et-ta-ni, -e-te-e-ke-ta-ni,
a-a-nu-u-ke-et-ta-ni
-ka(t)teni dqatni,
109
da-a-ga-at[-te-ni],
110
ukateni,
111
u-ka-at-te-ni,
112
a-a-nu-u-ga-at-te-ni, ta-me-e-kelat-te-ni
-kelatteni da-a-kelat-te-ni, p-i-kelat-te-ni, u-kelat-te-ni, u-kelat-te-e-ni
113
3 -kanzi dakanzi, pikanzi, pekanzi, ukanzi, akkukanzi
108 109 110 111 112 113
daske, paradigm
da-a-ke-e-mi we read non-plene forms with ambiguous -keli- consistently with e-vocalism (see 1.64,
p. 30). The example u-u-e-e-ki-i KBo 24.56A i 7 is exceptional. There are a few instances where stems
in -ke- are secondarily inected as stems in -iya-: ikiyazi, ikiyaizzi, ikiyanzi to ike- to smear, anoint,
dukiyazi to duke- to rejoice, ekiyawar to eke- (imperfective stem of e- to sleep).
108. See 11.10 (p. 183).
109. KUB 13.3 iii 6 (MH?lNS). While the ductus of the tablet is NS, many of the spellings indicate a
much older original, to which perhaps this a-vocalization is due. On the strength of this form, other ambigu-
ous writings in this tablet using the ou sign (e.g., u-kat-te-e-ni in i 14) can be vocalized with a. Another
indication of the a-vocalization of the latter form is the plene writing of the following syllable: compare
da-a-a-qa-te-e-ni and the alternation of da-a-ke-wa-ni and da-a-ga-u-e-ni, which suggest that (in older
Hittite) the a vocalization occurred when the stress fell on the following ending instead of on the lskel
syllable.
110. KUB 31.113: 4, 7.
111. KBo 3.33 ii 11 (OHlNS); the form is immediately preceded by a syllabic writing of na-at-ta,
which in post-OH is normally written ideographically.
112. KUB 24.4 obv. 10 (MHlMS).
113. KUB 14.4 ii 4 (NH).
Conjugation of mi-Verbs 12.33
00-HittiteGrammar.indb 205 6/5/08 4:17:16 PM
206
Preterite Indicative
Singular
1 -ganun daganun
-kenun pekenun, ukenun, akkukenun, dakenun
2 -ke, (-ki) dake, zikk (`d+ke < dai- to put),
114
atreke, uki
3 -ket zikkt, daket, peket, uket, akkuket
-kelat a-al-la-nu-u-kelat
115
Plural
1 -gawen ugawen, epuregawen, anigawen, [w]egawen
-kewen wekikewen
116

3 -ker dakr, peker, akkuker
-kar p-i-kar
117
114 115 116 117

formation of ske forms verbs in -ske-: p preterite paradigm
Imperative
Singular
1 -kellu pikellu
2 -ke peke, uke, akkuke
3 -keddu da-a-ke-ed-du, u-ke-ed-du, ak-ku-u-ke-ed-du
Plural
2 -ketten ak-ku-u-ke-et-tn, da-a-ke-et-tn, p-i-ke-tn
-katten dqaten
-kelatten me-mi-i-ke-et-te-en, da-a-kelat-te-en, p-i-kelat-te-en,
p-i-kelat-tn, u-kelat-te-en, ak-ku-u-kelat-te-en
3 -kandu dakandu, pikandu, ukandu, akkukandu
12.34. Supine: -kewan, -kawan (written: --ke-u-an, --ke-u-wa-an, --ke-wa-an,
--ga-wa-an). Verbal substantive: -kewar (written: --ke-u-wa-ar ). Innitive: -kewanzi
(written: --ke-u-wa-an-zi). Participle: -kant-.
Ablauting Stems in -(i)-
12.35. Most mi-verbs in -i- are part of a very productive class of verb stems
formed from nouns and adjectives by means of a sux. The ablaut -i-l-- reects
114. Both of these forms are found in KUB 14.1 (rev. 21 and 35), which is MH in MS. atreke is
found in HKM 30:15 (MHlMS).
115. KBo 22.2 obv. 7 (OS). Here there is a good chance that the a vocalism is to be preferred, since the
-nu-causatives in OH probably carried the stress on the nu syllable. But see the dierent treatments of a-a-
nu-u-ga-at-te-ni and [t]a?-a-nu-u-ke-et-ta-ni in KBo 7.28 + KBo 8.92 2425 (MS).
116. KBo 14.12 iv 17 (NH).
117. KUB 38.3 i 17 (NH). On the reality of the a vowel see Neu 1989a. See also p. 181, n. 17. On the
preterite third-person plural ending in general see Yoshida 1991.
12.34 Conjugation of mi-Verbs
00-HittiteGrammar.indb 206 6/5/08 4:17:16 PM
207
PIE `-ye-l-yo-. It is possible that OS spellings such as tar-ma-e-mi represent a se-
quence ltarma:.e-l not yet contracted to ltarma:y-l. For all the preceding see Oettinger
1979: 3034. The following are some of the verbal stems in -i-:
118
andi- to pre-
pare something, to be tted, matched, joined, atri- to send, write, iri- to make
rounds, mali- to approve, consent to, mugi- to invoke, munni- to hide, harbor,
muti- to remove, discard, dispose of, akti- to tend (medically), dammei- to
oppress, tui- to be short of breath.
12.36. Paradigms of the verbal stems in -i-. The ablaut form -i- is found only in
pres. sg. 3 (atrizzi), pret. sg. 2 and 3 (atre and atrit), and imp. sg. 2 (atri) and
3 (antaiddu).
119
Present Indicative
Singular
1 atrmi andmi irmi
2 atri andi
3 atrizzi andltizzi, antezzi,
andazi, andi
arizzi, iraiz[zi],
irizzi, iri
Plural
1 atrweni, atruni
119
anduni
119
2 atrtt[eni]
3 atrnzi andnzi iranzi, irn[zi]
handai: preterite paradigm
Preterite Indicative
Singular
1 atrnun andanun
2 atre
3 atrit, atrile andit irit
Plural
1 andwen
2 atrtten
3 atrer andltaer
Imperative
Singular
2 atri andi
3 atru andltaiddu
118. Oettinger (1979: 3034) interprets the stem as -ae-.
119. []atruni KUB 14.1 rev. 36 (MHlMS); anduni 1691lu ii 15, cited by Otten (1969: 26) and HED
H 98. See also p. 181, n. 15.
Conjugation of mi-Verbs 12.36
00-HittiteGrammar.indb 207 6/5/08 4:17:17 PM
208
Plural
2 atratten
3 andltandu irandu
12.37. Verbal substantive: andawar, irawar. Innitive I: andawanzi, irawanzi.
Participle: atrant-
120
, andnt-, irant-.
12.38. Aside from the very few singular third-person forms listed above (andai,
irai, atru) there is no confusion in Hittite between mi-verbs in -i- and i-verbs in
-ai- (for which see 13.2013.21, p. 222). Note also pres. sg. 3 ikallizzi alongside
ikallai he tears.
12.39. The following monosyllabic stems in -i- are not formed with the sux -i-
but originally had the same inection as those with the sux: li- to loose, i- to trust,
believe, i- to be angry, rage. But because their stems were monosyllabic, they were
inuenced by the i-conjugation, forms of which are marked below with . Although
Oettinger (1979) assigns l- (pp. 6467) and ae- and ae- (sic) (pp. 36064) to dierent
classes, they conjugate almost identically in most periods of Hittite. Paradigms:
Present Indicative
Singular Plural
1 lmi, mi lweni
2 li, la-[a-i]-i,
121
i
3 lizzi, li , izzi lanzi, lnzi, nzi
121
Preterite Indicative
Singular Plural
1 lnun
,
, ln, nun, nun lwen
2 la, i atten
3 lit, it, it, yit, i ler, er
Imperative
Singular Plural
2 l, li, ltten, tten
3 lu, lddu
12.40. Verbal substantive: war, lwar. Innitive: lwanzi. Participle: lnt-, nt-,
nt-. Imperfective stem: la-a-(i-)i-ke-.
120. mi-verbs in -(i)- in OS texts regularly write the neuter singular participle with a plene nal syl-
lable: ir-a-a-an, etc.
121. If correctly read, see HKM 30:19 (MHlMS).
12.37 Conjugation of mi-Verbs
00-HittiteGrammar.indb 208 6/5/08 4:17:18 PM
209
pai- and ue-
12.41. The extremely common verbs pai- to go and ue-luwa- to come
122
exhibit
many irregularities in inection. They are univerbations of the motion prexes pe- and
u- (see 12.21, p. 198; and 18.32, p. 286) and the inherited motion verb *h
1
ei- to
go, walk, move. The combination of ablaut and contractions of vowels led to compli-
cated paradigms. For paradigms and discussion see Oettinger 1979: 13132, 38889.
In OS to go has a consistent alternation of pai- and p-, except for the pret. pl. 3 pir
([pa:yr]), later replaced by payer ( pa-i-er) and pr based on pai + r and p + r.
123
124pai and uwa: preterite
Present Indicative
Singular Plural
1 paimi, pimi, pmi uwami, uwmi,
uwammi
paiwani, piweni,
pweni
uwaweni
2 paii, pii, pai,
paitti
uwai, uwi paitteni, paittani
123
uwtteni, uwattni
3 paizzi, pizzi uzzi, uezzi,
124

uwazzi
pnzi, panzi uenzi, uwanzi
125
125 126
Preterite Indicative
Singular Plural
1 pun, pn, pnun uwanun,
uwnun
126

piwen, pwen uwawen
2 paitta u, uwa uwatten
3 pat, pit uet, ut pir, payer, pr ur
127 128
Imperative
Singular Plural
2 t eu tten, ( paitten)
127
uwatten,
(uetten)
128

3 paiddu, paittu ueddu pntu, pndu uwandu, uwadu
122. We interpret -eliC- or -e-eC- as monosyllabic lwe-l and -wa- as disyllabic luwa-l. See Melchert
1984b: 40 41.
123. The forms paittni and paittni are found only in MHlNS.
124. In the spelling -ilez-zi it cannot be determined if the form was read lwezzil or lwizzil. The i writ-
ing is the traditional (or default) writing by Hittitologists. The former, however, is more likely in view of
the plene writing -e-ez-zi. See AHP 140.
125. Both uenzi and uwanzi occur in all periods, but uenzi is much less common than uwanzi in NH.
126. -wa-nu-nu-un in BoTU 57 I 10, cited HW1 238, is probably a scribal error for -wa-nu-un.
127. See CHD P 21 for occurrences of imperative paitten.
128. -et-te-en, found only once, to our knowledge, in an NH copy of an OH text, may be a scribal error.
-wa-at-te-en occurs twice in the immediate context.
Conjugation of mi-Verbs 12.41
00-HittiteGrammar.indb 209 6/5/08 4:17:19 PM
210
12.42. The expected imperative second-person singular forms of pai- and uwa- (*pai,
`uwa) are not attested, and the imperative second-person plural paitten occurs just once.
Instead, unprexed forms of `h
1
ei- are used for the imperative of to go, and a special
suxed form for the imperative second-person singular of to come.
129
For pai- the
forms are sg. 2 t (written i-it) and pl. 2 tten (written i-it-tnli-it-te-en). For uwa- the
sg. 2 is eu.
12.43. Verbal substantive: pwar, uwawar. Innitive. pa(u)wanzi, pwanzi, uwa-
wanzi. Participle: pnt-, uwant-.
Non-Ablauting
Stems in -nu-
12.44. Sample causative verbs in -nu- include: arnu- to transport, a(a)nu- to ar-
range, provide, paanu-lpaanu- to protect, and wanu- to turn, change. For the
meaning of the sux see 10.15 (p. 178). Paradigms:
130 131
Present Indicative
Singular
1 arnum(m)i, rnumi wanmi,
wanumi
paanumi a(a)numi
2
130
arnui, arnutti wanui
131
a(a)nui
3 arnu(z)zi wanu(z)zi,
wanuzi
aanu(z)zi,
anuzi, anu(z)zi,
nuzi
Plural
1 arnummeni wanumeni,
wanummeni
anumeni,
aanummeni
2 arnutteni wanutteni paanutteni,
paanutteni
a(a)nutteni
3 arnu(w)anzi,
rnuwanzi,
arnwanzi
wanuwanzi paanuwanzi a(a)nu(w)anzi,
anuwanzi,
aanwanzi
129. There are also two other attested forms of unprexed reexes of `h
1
ei- to go: pres. pl. 3 ya-an-zi
in KBo 22.2 obv. 7 (OS) and pres. sg. 3 i-ez-zi in KUB 33.66 + KBo 40.333 iii 6. The latter is backformed
from the third plural, replacing historically expected but synchronically irregular `ezi.
130. Other verbs: atkinui, arganui, karanui, laknui, nuntarnui, tinnui, tittanui, etc.
131. Also written ni-nu-i KBo 4.14 ii 34; see Honer 1997c.
12.42 Conjugation of mi-Verbs
00-HittiteGrammar.indb 210 6/5/08 4:17:20 PM
211
Preterite Indicative
Singular
2 paanu
3 arnut, rnut wanut paaanut a(a)nut
Plural
1 arnummen wanum(m)en aanummen
2 paanutten
3 arnuer wanuer paanuer,
paaa[nuer]
aanuer
warnu: preterite paradigm
Imperative
Singular
2 rnut wanut paaanut,
paanut
a(a)nut
3 arnuddu wanuddu paanuddu,
paanuddu
aanuddu
Plural
2 arnutten wanutten paanutten,
paanutten
3 arnu(w)andu wanuwandu paanuandu
12.45. Verbal substantive: arnummar, wanumar, wanuwar, paanummar, aa-
nuwawar, gen. a(a)numa. Innitive: arnumanzi, wanummanzi, paanummanzi,
paaanumanzi, aanummanzi, anumanzi. Participle: arnuwant-, arnuant-,
132

wa nuwant-, paanuwant-, paanuwant-, paaanuwant-, a(a)nuwant-.
12.46. Imperfective stem: arnuke-, a(a)nuke-, wanuke-.
12.47. For the forms in -meni, -men, -mar, -manzi, etc., see 1.126 (p. 44).
Suppletive Verb
te-/tar- to say
12.48. The stem te- to say(originally *to put; see 12.21, p. 198) is supplemented
by the stem ter-ltar- in forming a complete paradigm for this verb of speech. For a dis-
cussion see Oettinger 1979: 10910.
132. On this form in Laws 77 manuscript A (OS) see discussion in LH 194.
Conjugation of mi-Verbs 12.48
00-HittiteGrammar.indb 211 6/5/08 4:17:21 PM
212
Present Indicative Preterite Indicative
Singular Plural Singular Plural
1 tmi, temi tarueni tnun
2 tei, ti tarteni, tteni, te
133

3 tezzi, tardi
134
taranzi tt terer
133 134

paradigm of te, end
Imperative
Singular Plural
2 tt tetten, (tten)
3 tddu darandu
12.49. Participle: tarant-. Imperfective stem: tar-i-ke-, tar-a-ke-.
Verbs with Mixed Stems
12.50. A number of Hittite verbs are attested as belonging to more than one stem-
class. In some cases one stem is clearly a chronological replacement of another, but in
other cases the temporal relationship is less clear. Verbs whose original stem class is
reasonably clear have been treated under that class, with later replacement forms duly
noted. What follows here are examples of verbs with multiple stems whose chronology
is indeterminate.
mi-verbs with stems in -i- (12.3512.37, pp. 206208) and -iya- (12.28
12.30, pp. 202204): ari-lariya- to cultivate, appari-lappariya- to do
business, trade, sell, amni-lamniya- to create, arli-larliya- to elevate,
magnify, tarkummi-ltarkummiya- to proclaim.
mi-verbs with stems in consonant, -i-, and -iya-: italk-litalki-litalkiya- to
make smooth, par-lpari-lpariya- to break, tarupp-ltaruppi-ltaruppiya-
to assemble.
mi-verbs with stems in consonant and -iya-: malk-lmalkiya- to spin, ark-l
arkiya- to perish, park-lparkiya- to raise, rise, itark-litarkiya- to sicken,
kar-lkariya- to cut.
12.51. Often, however, one stem of a root belongs to the mi- and other(s) to the
i-conjugation:
Consonantal i-stem and mi-stem in -iya-: n-lniya- to wipe.
i-stems in -a- and mi-stems in -iya-: dala-ldaliya- to leave, duwarna-lduwarniya-
133. See [t]e-e KUB 60.150 obv. 9. In HKM 48:17 instead of i-it te-e-e, read z[i-i]k-ke-e-e.
134. This unusual and unexpected form appears to be a Luwian present third-person singular form
in -ti.
12.49 Conjugation of mi-Verbs
00-HittiteGrammar.indb 212 6/5/08 4:17:21 PM
213
to break,
135
art-lartiya- (and arti-) to wipe, walla-lwalliya- to praise.
See also nai- and neya- 13.21 (p. 223).
i-stems in -ai- and mi-stems in -iya-: iai-liiya- to bind, ipai-lipiya- to be
satiated, ai-liya-.
136
12.52. Not all apparent examples of the above are really the same verb: mark- to
cut up versus markiya- to disapprove of, reject; ar- to ow versus ari-lariya-
to tend, care for, cultivate; and war- to reap, harvest versus wariya- to pacify,
soothe.
135. This verb also shows some forms belonging to a stem in -e-l-a- (see 12.23, p. 199) and to a mi-
stem in -i- (see 12.35, p. 206): duwarnez(z)i, duwarnazi alongside duwarnizzi.
136. In this case two originally separate verbs have merged: ai- to press, seal (i-verb in -ai-) and
iya- to throw, shoot (see peiya- to throw (away), cause to fall).
Conjugation of mi-Verbs 12.52
00-HittiteGrammar.indb 213 6/5/08 4:17:22 PM
214
Chapter 13
CONJUGATION OF i-VERBS
Consonantal Stems
13.1. The stems of many i-verbs end in a single consonant that may be geminate
or non-geminate: akk-lekk- to know, akk-lek- to die, ar-ler- to come to, arrive at,
a-le- to open, aa-lae- to settle, a- to give birth, d- to dry up. Note
that the rst ve stems show an ablaut ale with a distribution nearly completely oppo-
site that of mi -verbs in ela. While there are exceptions such as ekk- to know, there is
also a regular alternation by which stems ending in a geminate consonant show a non-
geminate in the present third-person singular and preterite third-person plural. Note aki
and aleker to akk- to die and the frequent contrast between present third-person sin-
gular and present third-person plural: itpilitappanzi to shutlstop up, ilaanzi
to give birth and zilzaanzi to strike. See AHP 8081. There is a tendency in
this class for -ta to replace - as the preterite third-person singular ending and for the e
vocalism to spread at the expense of a.
to know to die to arrive at to open
Present Indicative
Singular
1 ki, aggai kmi
1
ri
2 akti, ekti kti rti
3 akki aki ri i, azi
Plural
1 ekkueni, ikkueni akkueni
2
erueni
3
aueni
2 aktni, akteni,
ektni
kteni rteni, artni,
erteni
3 akanzi, ekkanzi akkanzi aranzi aanzi, anzi
1 2 3
1. So, instead of expected i-conjugation *ki or *aggai. Other mi-conjugation forms of akk- are
imperative akdu and pret. sg. 3 akta. Since the mi-conjugation forms are all in late texts, we can assume
singular rst-person i-forms in OH and MH.
2. Written ak-ku-(u-)e-ni, in contradistinction to a-ku-e-ni we drink. See p. 188, n. 12, and 1.84
(p. 35).
3. Written e-ru-u-e-ni and er-u-e-ni.
00-HittiteGrammar.indb 214 6/5/08 4:17:23 PM
215
to settle to give birth to dry up
Present Indicative
Singular
1 aai, aae
2 aati
3 aai, ai, aae i ti
Plural
3 aanzi, aianzi aanzi tanzi
shakk- preterite paradigm
to know to die to arrive to open to settle give birth to dry up
Preterite Indicative
Singular
1 aggaun ar (a)un,
r (a)un
aaun un,
un
2 akki,
kta
3 akki,
4

akta, ekta
kki,
a(k)ki,
akta,
agga
ra, ra ta aata,
aeta
ta zta
Plural
1 ekkuen aruen,
eruen
5
uen
6
2 kten
3 ekker aker,
7

eker
arer,

erer er,
eer,
er
aer,
aeer,
aeir
8

er, ir ter
4 5 6 7 8
4. For the inserted -i- in akki, kki, etc. see 1.81 (p. 34). The spellings a-ar-ala-ar-a represent
la:rsl. A form like agga could also be a spelling for la:ksl, with no inserted vowel, but its appearance in
MH (VBoT 1:24, MS) suggests a real ending l-asl borrowed from i-verbs in -a- (see 13.13, p. 220).
5. Written ar--en, e-ru-u-en, e-er-u-en, e-ru-en.
6. Written -e-u-u-en.
7. As a pret. pl. 3, only a-ker is thus far attested in OS. In later texts we nd a-ke-er, e-ke-er, and e-ker.
8. a-e-i-ir in KUB 23.42 obv. 1. For e-e-er KUB 41.1 iv 9 see 1.211.22 (p. 18) and n. 24 there.
Conjugation of i-Verbs 13.1
00-HittiteGrammar.indb 215 6/5/08 4:17:24 PM
216
Imperative
Singular Plural
1 eggallu, iggallu,

aggallu, akkallu
2 k, k,
9
ekten, arten, ten, aeten
3 aku, aru, u, du, akdu, akdu,
du
ekkandu, akkandu, andu
9
13.2. Verbal substantive: aggtar, uwar, aeuwar .
10
Innitive I: aeuwanzi.
Innitive II: aranna (?).
11
Participle: ekkant- (rare akkant-), akkant-, arant-, aant-,
ant-, aeant-, dant-. Imperfective: akkike-, rake-, ake-, eke-, aake-,
aeke-, aike-.
13.3. Some i-verbs end in a sequence of two consonants: pa- l paa- to guard,
i( p)pand- l ipant- to libate, oer.
Present Indicative
Singular Plural
1 paai,
paami
ipantae,
ipandai,
ippantai
paueni ipanduwani
2 paati paateni
3 ipanti,

ipnti,
ipandi, i(p)panti,
ipantai
12
paanzi ipantanzi,

ipantanzi,
ippantanzi,
i(p)pandanzi
12
Preterite Indicative
Singular Plural
1 paa[u]n i(p)pandaun ipantuen,
ipanduen
3 paata ipandlta,
ipanzata
pair i(p)panter,
ippantaer
Imperative
Singular Plural
2 pai paaten
3 paandu
9. All forms in a-le- in this and the following paragraph belong to the verb to open.
10. Written a-e-u-u-wa-ar.
11. KUB 46.40 obv. 3 (NH).
12. Present sg. 3 ipantai and preterite pl. 3 ippantaer are due to late reanalysis as an a-stem modeled
on d- to take.
13.2 Conjugation of i-Verbs
00-HittiteGrammar.indb 216 6/5/08 4:17:24 PM
217
13.4. Verbal substantive: ipantuwar, ippanduwar . Innitive I: ipanduwanzi, ip-
pan duwanzi. Participle: paant-, ipantant-. Imperfective: paake-, ipanzake-,
i ( p)pan zake-.
13.5. The originally consonantal stem - (iter. to iya- to do, make) and the stems
la(w)- and i(w)- are treated in 13.1513.19 (pp. 220222) because their attested
inection is predominantly as vocalic stems.
13.6. Factitive verbs in -a- are inected exclusively as i-verbs in OH (OS), but they
are mostly inected as mi-verbs in the later language. Note the dierent treatment of the
ablauting adjective stem in the u- and i-stem adjectives: idalawa- (not `idaluwa-)
and tepawa- (not *tepuwa-), but uppiya- (not `uppaya-).
131415
Present Indicative
Singular
1 idlawami, uppiyami, manninkuwami, sro
5
-ami, xsxi-iyami
2 idlawati, kururiyati, ln-ati, sro
5
-ati
3 arwai, tauwai, uppiyai, kururiyai, katterai, arazyai,
nti-wai (`idalawai), maninkuwai, kartimmiyai, xsxi-ai ; (mi-
forms:) tauwazi, idlawazi, maniyazi, uppiyazi, xsxi-iyazi
Plural
1 sro
5
-aweni ; dauwawani, [3-y]aauwani, 4-yaauwani
13
2 katterrateni, arazziyateni, sro
5
-ateni, [ak]uwaarateni
3 [3-y]aanzi, 4-yaanzi, xsxi-iaanzi
Preterite Indicative
Singular
1 idlawaun, arwaun, papraun, kururiyaun, uppiyaun,
kutruwaun, newaun, appiaraun, xsxi-ia[un]
2 in -ta
LU
t.or-ata
3 in -a
14
allapaa, antezziyaa, manninkuwaa, newaa, uppiyaa,
taruppiyaa, tatraa
3 in -i ikunai, maniyai, pararai, taruppiai, watarnai,
3 in -ta
15
ln-ata, oin-lata, iiyata, idlawata, katterrata, kuripata, kururiyata,
liliwata, paprata, akiyata, arazziyata, tauwata, tepawata
13. The redundant spelling -a-a-a-u-wa- for l-ahhwa-l is unexplained, perhaps faulty.
14. The third-singular forms in -a and -i co-occur in KUB 28.82 i 1617 and on the same verb
(taruppiyaa [16], taruppiyai [17]).
15. All examples of the third-person singular forms with -ta endings are NH.
Conjugation of i-Verbs 13.6
00-HittiteGrammar.indb 217 6/5/08 4:17:25 PM
218
Preterite Indicative
Plural
1 idlawauen, zo-nauen, [xsxi-]iyauen
2 nti-aten
3 kururiyaer, ulkearaer, appina[er ], pdaaer, ln-naer,
katterraer, [m]iriwae [r], liliwaer, markitaer, arwaer, papraer,
newaer, mayandaer, dauwaer, sro
5
-yaer, dannataer
13.7. Verbal substantive: uppiyauwar . Innitive: maniyawanzi, uppiya-
u wanzi. Participle: idalwaant-, maniyaant-, uppiyaant-. Imperfective: mani-
yaileke-, uppiyaike-.
Vocalic Stems
13.8. The number of possible nal vowels in stems of the i-conjugation is more re-
stricted than in stems of the mi-conjugation. There are basically only two types: stems in
-a- and stems in -i-, but these mutually inuence each other, and remodelings of original
consonantal stems further complicate the situation.
13.9. Stems in -a-: According to the rule given in 1.126 (p. 44) these stems show
rst-person plurals and verbal nouns in -um-. There is, however, a strong tendency to
level out this odd alternation in favor of the regular -a- stem and regular endings (see
dwen, dwanzi, dwa).
a-Stems
13.10. Underived stems in -a-.
13.11. d- to take:
16 17 18 19
Present Indicative Preterite Indicative
Singular Plural Singular Plural
1 de,
16

di,
17
dai
tumni,
18

dummeni,
19

dweni,
20
dwani
21
dun, daun dwen
22
2 dtti, datti, tatti dattni, datteni d datten
3 di danzi d dir
23
16. da-a-a- in CTH 416 (OS).
17. da-a-a-i in StBoT 25 #137 ii 10 (OS) and KBo 17.61 obv. 21, 25 (MHlMS).
18. Written tu-me-e-ni in CTH 416 (StBoT 25 #4 iii 43, iv 25, OS), CTH 752.2 (OS), KUB 43.25
obv. 12 (OS), KBo 17.25 obv.? 8 (OHlMS) and tu-me-ni in KBo 25.8:2 (OS), and tum-me-e-ni KUB
28.79:5 (OHlNS) and du-me-e-ni in CTH 443 (MH).
19. Written du-um-me-e-ni and tum-me-e-ni in Hatt. III and Tud. IV texts, therefore late NH. See 1.24
(p. 19).
13.7 Conjugation of i-Verbs
00-HittiteGrammar.indb 218 6/5/08 4:17:26 PM
219
Imperative
Singular Plural
2 d dtten, datten
3 du, daddu
24
dandu
20 21 22 23 24
13.12. Verbal substantive: dwar, gen. dwa. Innitive I: dwanzi. Innitive II:
danna. Participle: dnt-. Imperfective: dake-.
13.13. Paradigms of verbs with -u(m)- in rst-person plural (1.126, p. 44, but see
also 13.11, tumni above), in the verbal substantive, and innitive I: tarna- to let,
arra- to divide, wata- to sin, uda- to bring here, peda- to carry o. The late Hit-
tite writing -fr-zi and p-fr-zi does not indicate a second vocalization lutenzil or
lpitenzil for the forms udanzi and pedanzi, since the tn (ur) sign in late Hittite often
should be read dan
x
(1.35, p. 21; 1.63, p. 29).
25 26 27 28 29
Present Indicative
Singular
1 tarnae,
tarnai
watai udai ptae,
25

ptai, pdai
2 tarnatti,
tarnai
arratti watatti,
watai
udaddi,
udatti
26
pdatti
3 tarni, tarnai,
tarnaizzi
arrai,
arrizzi
watai, wati
27
udai, udi pdai
28
Plural
1 tarnumeni,
tarnummani
29
arraweni udumni,
utummeni
29
ptumeni,
pedummeni
29
2 tarnatteni arratteni udatteni petatteni,
pdatteni
3 tarnanzi arranzi watanzi udanzi pdanzi, pedanzi
20. Written da-a-[u-]e-ni StBoT 25 #137 ii 16 (OS), da-a-u-e-ni in KUB 16.16 rev. 20 (NH liver
oracle).
21. Written da-a-u-wa-ni KUB 12.63 rev. 8 (CHD LN sub laiyala-).
22. Written da-a-u-en and da-a-u-e-en.
23. We read da-a-ir as monosyllabic dir (i.e., [da:yr]). See p. 224, nn. 4850.
24. Much less common than du (da-a-).
25. OS: p-(e-)t-.
26. OS: -da-ad-di in KUB 33.59 iii 3; NH -da-at-ti KUB 33.121 ii 11.
27. The form wa-a-ti (KUB 1.16 iii 60) suggests that this verb was originally a consonantal stem (see
11.15, p. 184).
28. OS: p (-e)-ta-i, p-e-da-i. Plene writing in the second syllable appears rst in NS.
29. On the graphic mm in these forms see 1.24 (p. 19).
Conjugation of i-Verbs 13.13
00-HittiteGrammar.indb 219 6/5/08 4:17:27 PM
220
Preterite Indicative
Singular
1 tarnaun arraun wataun udaun
30
pedaun
2 tarna wata
3 tarna,
tarneta
31
a-a-ar- a,
32

arra, arret
wata uda pda, peda
Plural
1 tarnummen,
29

tarnuen
33
arrummen
29
utummen
29
ptumen,
34

petummen
29
2 tarnatten
3 tarner arrer water uter peter
30 31 3233 34 tarna preterite
Imperative
Singular
2 tarna, tarni arri uda pta, peda
3 tarnu,
tarnedu
35
udu pedu
Plural
2 tarnatten,
tarniten
36
udatten pedatten,
piteten
36
3 tarnandu udandu ptantu,
pedandu
35 36hi-verbs in a: paradigm
13.14. Verbal substantive: tarnummar, arrumar, wadumar, utummar, petummar.
Innitive I: tarnummanzi, arrumanzi, utummanzi, pedummanzi. Innitive II: arranna.
Participle: tarnant-, arrant-, watant-, udant-, pedant-.
13.15. Imperfectives in -a-. Archaic forms of the stem (a)- to do, make show
that the sux -a-, which marks the imperfective in some verbs instead of -ke/a-,
30. Attested in OS as -da-[u-un] KBo 3.22:58.
31. tarneta shows replacement of the i-conjugation preterite third-person singular - by -ta. See
11.16 (p. 184) and p. 181, n. 6.
32. SBo 1 Text 15:6, Bo 90/732:62 (OS). For the dating of the latter see Carruba 1993b: 81, 83. For the
unusual syllabication see 1.12 (p. 13).
33. tar-nu-en KBo 3.60 iii 7 (OH/NS).
34. OS: p-e-tu-mi-en.
35. tarnedu and eezi are analogical to the preterite third-person singular forms tarneta and eta,
whose -ta ending led to a reanalysis as from stems tarne- and e- with mi-inection. Possibly the same
process can be seen in audu and auzi, reanalyzed on the basis of preterite third-person singular auta (see
13.32, p. 229).
36. This variant is due to inuence of the i-verb stems in -ai- (see 13.21, p. 224).
13.14 Conjugation of i-Verbs
221
originally had consonantal inection.
37
The non-nite forms and derived stem in -ke-
are also based on a stem -(i)-. For the most part, however, the sux is -a-. Other
i-verbs showing this derivational sux are: arpaa- to be unlucky, alzia- to
call, uittea- to pull out, draw out, karmalaa- to suer harm, be incapacitated,
ia- to press, seal, tarpanallaa- to take anothers place, warria- to come to the
aid of.
-ssa preterite
Present Indicative
Singular
1 iai, ai, alziai, war (r)iai
2 iatti, atti, ati (rare), alziatti, i-i-at-ti, warreatti, warriatti,
wa-ar-i-a-at-te
3 ai, ii, i, alzileai, uitteai, karmalaai
Plural
1 ieni, eni
2 tni, ten, iatteni, atteni, warreatteni
3 ianzi, anzi, alzileanzi, warreanzi
Preterite Indicative
Singular
1 iaun, aun, warreaun
2 alzeeta
3 iita, eta, ita, eeta, warreeta, warriita
Plural
1 ien, een, un
2 atten
3 er, er, ier, ir, (eer), alzeer
Imperative
Singular
2 a, a, i, warrea
3 au, addu, alzidu, warreidu
Plural
2 ten, iatten, eten
3 iandu, andu, alzeandu
13.16. Supine: uwan, iuwan, uwan. Verbal substantive: umar, uwar.
Participle: warriant-. Imperfective: iker, iker, iket, eeker.
37. Specically i-i-te-e-ni and i-i-te-en (OS).
Conjugation of i-Verbs 13.16
00-HittiteGrammar.indb 221 6/5/08 4:17:28 PM
222
Stems Originally in -Cw-
13.17. la()w- to pour (liquids), reduplicated lilw-, and iw- to pour (solid
particles). These stems originally inected with a stem in -Cw- before vowel (neces-
sarily written -Cu-wV- or -Cu-ul-V-) and -Cu- before consonant. There was a marked
tendency for these to be replaced by vocalic stems in -C(u)wa- throughout.
38 39 40
Present Indicative Preterite Indicative
Singular Singular
1 laui, lumi
38
1 luwanun
39
2 lautti 2
3 lui, luwai, lawai,
lauzzi, lwiz[zi ]
3 lu, luwa, luwai, lu
40
Plural Plural
1
39
1
2 2
3 lauanzi, luwanzi, lauwanzi 3 lauwer
Imperative
Singular Plural
2
39
2 lauten, luwaten
3 3 lauwandu
13.18. Participle: luwant- (sg. neut. laun); verbal subst. luwar, gen. luwa.
13.19. Just as i-verbs in -i- tend to acquire the inection of mi-verbs in -iya-, based
on forms such as the present third-person plural in -iyanzi, so too there are a few ex-
amples of this verb as a mi-verb stem lauwa(i)- (e.g., lauwanun, lauzzi, lauwaizzi,
etc.).
ai-Stems
13.20. Ablauting verbs in -ai-. We retain the traditional classication of these stems
as i-verbs in -ai-. The paradigms actually show a complicated alternation of stems in
-ai-l-i-, -e-, and -i-l-y-. As the following paradigms show, -i- tends to compete with
the other variants. There is the further important peculiarity that the present and preterite
second-person plural endings in this class are attested from OS as -teni and -ten, with
an initial --. This -- later tends to spread to the present second-person singular ending
38. KBo 32.176 i 16.
39. The pres. pl. 1 la-u-e-ni, pret. sg. 1 la-a-u-un, and imp. sg. 2 la-a-a belong to a stem l- with
identical meaning. See Honer 2003b: 620 (review of HED L).
40. KBo 32.14 ii 42 (see Neu 1994: 19192).
13.17 Conjugation of i-Verbs
00-HittiteGrammar.indb 222 6/5/08 4:17:29 PM
223
(-ti for -ti). The replacement of the preterite third-person singular - with -ta may be
related or reect the inuence of consonant stems ending in -- (see p. 181, n. 11, and
13.1, p. 215).
41
13.21. Paradigms of dai- to put, pai- to give, nai- to turn, zai- to cross, alzai-
to call (originally ii- to bind belonged to this class; see 13.30, p. 227):
42 43 44
45
Present Indicative
Singular
1 te, ti,
tei
pe,
pi, pei
nei neyami
42
[alz]ii, alzii
2 ditti, titti paitti,
paiti, peti
nitti,
neyati
neyai zitti,
zi
alzitti, alziyatti,
alzeti, alziyai
3 di pi, ( pai) ni, (nai) neyazzi zi alzi
Plural
1 tiyaweni,
tiyauni
43

pweni,
piyaweni
naiwani, neyaweni alziwani,
alziyaweni
2 titteni,
titeni
44

piteni,
peteni
naitani, naiteni alziyatteni
3 tianzi,
tianti,
45
tienzi,
tiyanzi
pi(y)anzi nanzi, ne(y)anzi alzianzi, alziyanzi
halzait

dai Preterite
Preterite Indicative
Singular
1 teun peun n[un], neun,
neyaun
zeun alziun
2 paitta, peta naitta zit alzait
3 dai, di,
dita ditta
pai, pi,
paita,
pta
nai, ni, naie,
nita, nitta, nit ,
neyat
zi alzai, alzi
41. For dierent views on the prehistoric source of the -- see Oettinger 1979: 462 with refs. and Kloek-
horst forthcoming.
42. Obviously, this paradigm is not i- but mi-conjugation. It is placed here to call attention to the fact
that nai- conjugates both ways. The i-conjugation is older and more regular for nai-.
43. In ti-ya-u-ni 1691lu ii 18 (CTH 375.1 prayer of Arnuwanda and Amunikal, MHlMS). That this
is not a scribal slip for ti-ya-u-e-ni is indicated by the immediately following na-at sro
5
-a-u-ni for
sro
5
-a-u-e-ni.
44. Note that the -teni, pret. sg. 2 and 3 -ta, and pret. pl. 2 -ten endings are characteristic of the i-,
not mi-, conjugation.
45. This form either represents an unexpected early Luwianism in OS or an error of audition with confu-
sion between ltl and ltsl. Compare for the latter possibility pres. sg. 3 damati for damazi (see p. 181, n. 8)
vs. pres. sg. 2 itamazi for itamati (see 11.10, p. 183).
Conjugation of i-Verbs 13.21
00-HittiteGrammar.indb 223 6/5/08 4:17:30 PM
224
Preterite Indicative
Plural
1 daiwen,
46

tiyawen
pwen,
piyawen
neyawen zaiwen alziwen, alziyawen
2 daiten,
47

tiyatten
3 dir,
48
daier,
dyr,
49

dar,
50
tier
51

pr, pier naer, neyer alzi(y)er
46 47 48 49 50 51

New nn. 4850
Imperative
Singular
2 di pai, pi ni, neya alzi
3 du pu, pedu nu alzu
Plural
2 diten pten,
piten,
paiten
naiten, niten,
neyatten
zitten alziten
3 tiyandu piandu neyandu alzi(y)andu
13.22. Verbal substantive: tiyawar, piyawar, neyawar, (gen.) alziyawa. Innitive
I: tiyawanzi, piyawanzi,
52
alziyawanzi . Innitive II: tiyanna, piyanna. Participle:
tiyant-, piyant-, nant-, neyant-, alziyant-. Supine: uwan, iuwan, uwan. Im-
perfective: zikke-, pike-, naike-, zike-, alzi(a)- (see 13.15, p. 221).
Stems with Mixed Inection in -a- and -i-
13.23. Paradigms for base verbs in -a-/-i-. Disyllabic stems that originally inected
as i-verbs in -ai- (13.20) were inuenced by the i-verbs in -a- (13.13) and show
46. da-i--en (MH), da-i-u-e-en (MH), da-a-i--en (MH).
47. Also iiten and aiten.
48. We read thus da-a-ir ([da:yr]) in KBo 22.1:4 (OS), a form also attested in NH, contracted from
prehistoric *dyr with loss of intervocalic *y.
49. da-i-er KBo 22.2 obv. 16 (OS), KBo 15.10 iii 44 and passim (MH/MS) reects a renewed dai- +
-r (with dai- from forms like dai-wen). See also da-a-i-e-er KBo 16.27 iii 9, iv 9 (MH/MS), KBo 6.34 i 27
(MH/NS), da-a-i-er KBo 15.10 ii 30 (MH/MS), da-i-e-er ibid. iii 47 (MH/MS), KBo 16.27 iii 15 (MH/MS),
KBo 8.35 ii 29 and passim (MH/MS), HKM 57:17; and HKM 63:19 (both MH/MS), sporadically as late as
KBo 16.61 rev.? 17 (late NH).
50. da-e-er HKM 47:11 (MH/MS), perhaps a scribal slip for da-i-e-er, the form found elsewhere in
HKM texts.
51. ti-e-er and ti-i-e-er.
52. See, however, the cautionary remarks in CHD P 42 about this form possibly being from piya- to
send.
13.22 Conjugation of i-Verbs
00-HittiteGrammar.indb 224 6/10/08 7:45:17 AM
225
a mixture of -a- and -i- forms. The stems in -a- had already been generalized in the
rst- and second-singular in Old Hittite. In other persons, the variants with -a- spread
at the expense of those with -i- within the period from Old to New Hittite. Originally
dla- to leave belonged to this class (see 13.30). Stems: mema- to speak, nna- to
drive (here), penna- to drive (there), uppa-to send (here).
Present Indicative
Singular
1 mmae,
mmai, memai
nnai pe(n)nai uppai
2 mmatti nnatti pennatti [up]patti
3 mmi, memai nni, nnai pennai, penni uppi
Plural
1 memiweni,
memaweni,
memiyaweni
unum [ni ]
53
penniweni uppiweni
2 mema(t)teni,
memiteni
nnatteni, nniteni uppatteni
3 memi(y)anzi,
memanzi
nniyanzi, nnanzi penni(y)anzi,
pennanzi, pinnanzi
uppi(y)anzi,
uppanzi
53memali preterite
Preterite Indicative
Singular
1 memaun nnaun pennaun uppaun
2 memita nne, nni uppeta
3 mmita, mema nne, nneta,
nnita
penni, pennita,
penna, pennit
uppa, uppeta
Plural
1 memawin,
memiyawen
nnumen uppiwen
2 memiten
3 mmier,
me(m)mier, memir
nni [er] pennier, penner
54
uppier
54
53. KBo 41.42 i 5, could also be from unu-lunuwa- to decorate.
54. Written p-en-ner, built on the stem penna-. Likewise the unusual p-ner
er
. We read p-en-ni-er as
pennier built on the stem penni-, after mmier, not as a possible p-en-n-er.
Conjugation of i-Verbs 13.23
00-HittiteGrammar.indb 225 6/5/08 4:17:31 PM
226
Imperative
Singular
1 memallu
2 mmi, mema nni penni uppi
3 mmu, memu,
memattu
nnau uppu
Plural
2 memiten nniten penniten uppiten
3 memandu uniandu,
nni(y)andu

13.24. Verbal substantive: memiyawar, memmwar, `nnumar (gen. nnuma ),
`pennumar (gen. pennuma ). Innitive I: memiyawa(n)zi, memiwanzi, memmwanzi,
pennu(m)manzi, pennuanzi, penniyawanzi. Substantive: memant-, memi (y)ant-, nnant-,
penniyant-,
55
pennant-, uppant-. Imperfective: mmike-, penneke-, uppike-.
Verbs with the Imperfective Sux -anna/i-
13.25. As verbs with the imperfective derivational sux -ke- inect according to
the mi-conjugation, so those with the isofunctional sux -anna- inect as i-verbs with
mixed stems in -a- and -i-. There are also some examples of mi-conjugation forms mod-
eled on mi-verbs in -iya-.
13.26. Examples of inection of stems in -anna-.
56
Present Indicative
Singular
1 pariyannai, iwannai, uittiyannai
3 attannai, uttiannai, uittiannai, iyannai, nannai, parannai, pariyannai,
(pariyanni), tuannai, walannai
Plural
3 nannianzi, walannianzi, iyannianzi, allanniyanzi, allannianzi, pariyannanzi
55. Neut. sg. nom.-acc. pennian KUB 12.58 iv 18 (NH) and [ p]ennan KUB 39.49:16 (NS).
56. The occasional failure to spell the geminate -nn- by omitting the -an- sign (e.g., wa-al-a-a-na-i,
wa-al-a-ni-an-da, -e-u-wa-n-e-k[i-it]), pace Neu 1981a, is of no linguistic signicance.
13.24 Conjugation of i-Verbs
00-HittiteGrammar.indb 226 6/5/08 4:17:32 PM
227
Preterite Indicative
Singular
1 [tua]nnaun, iyanniyanun
3 yanni, iyannile, piddanni, (but rarely: iyanniyat, pariyannit )
Plural
2 iyanniyatten
3 attannier, pariyanner
57
57anna preterite
Imperative
Singular
3 walannau
Plural
3 ikarranniandu
13.27. Supine: piyanniwan, walanniuw[an], oti-anniyauwan. Imperfective: wal-
an nilekanzi, walhannileket, walanneker, walanniketten, walannikeuan, weri-
an nikei, laiyannigaueni, uwanek[et ].
13.28. Forms of the verb iyannali- in older texts follow the i-conjugation like other
-anna- duratives. But in NH mi-conjugation forms begin to appear.
13.29. The verb piyani- to reward (with CHD P s.v., contra HW 169) is not a
durative stem of pai-, piya- to give, as its mi-conjugation inection, ungeminated n,
and meaning in context clearly show.
Mixture of mi- and i- Forms
13.30. The following verbs show a more thorough-going mixture of mi- and i- forms
than those in 13.23 (p. 224). The verb dla- to leave probably belonged originally to
the class of memali- (13.23) and ii- to bind to that of dai- (13.21, p. 223), while in
the case of ai-liya- to press, thrust there has been a merger of two originally separate
verbs, one of which was originally of the dai- class belonging to the i-conjugation and
the other a stem in -iya- belonging to the mi-conjugation.
58
The available attested forms
do not permit a condent assignment to a particular inectional class.
57. Written pr-i-ya-an-ner. For the interpretation of a-at-ta-an-ni-er see n. 54 (p. 225).
58. See p. 213, n. 136.
Conjugation of i-Verbs 13.30
00-HittiteGrammar.indb 227 6/5/08 4:17:33 PM
228
Present Indicative
Singular
1 dlai, dliyami iii iyami, iammi
2 dalatti, dliyai, taliyai
3 dlai, tlai, dli, dlaizzi,
dliyazi, taliyazi
ii,
59
iai, iiezzi,
iiyazzi
zi, yizzi, iyaizzi,
iyazzi, i
Plural
1 dliyaweni, tliyaweni iyweni
2 dleteni iiyatteni
3 dliyanzi, dlanzi iianzi, iiyanzi iyanzi
59 60
Preterite Indicative
Singular
1 tlaun, dlaun,
dliyanun, tliyanun
iiun, iiyanun iynun, iyanun
2 ita
60
3 tli, dli, dleta,
dliyat, dla
ii, iiyat t(ta), iyit, yait, a,
i
Plural
1 daliyawen
2 daliyatten iaite[n], iten iten
61
3 dlier iier iyaer
61ishiya paradigm
Imperative
Singular
1 dla, dli ai, iya
2 tledu
Plural
2 dleten iiyatten iten
3 dlandu ii(y)andu iyandu
13.31. Verbal substantive: dalumar, taliyawar, iiyawar, iyawar. Innitive I: iya-
wanzi. Innitive II: iyanna. Participle: daliyant-, iiyant-, iyant-.
59. Laws 158 copy aa (OS).
60. KBo 3.34 i 23 according to Oettinger (1979: 473 352), citing Eichner. See Dardano 1997: 3435,
8788 with n. 64.
61. iten KUB 26.82:9 quoted in HW 175 as imperative is, instead, preterite by context.
13.31 Conjugation of i-Verbs
229
13.32. The conjugation of au()- to see (and the much rarer mau()- to fall) is
classied as irregular. It shows an ablauting stem (full grade au-, zero-grade u-). The two
grades are also reected in nominal derivatives: au+ri- sentinel post, watch and par
uwattalla- overseer, watcher (from par au()- to oversee). The verbal paradigm
shows an inserted -- before endings beginning with -t- (see the verbs in -i- above!).
Unexpectedly, the third-person forms singular show not only the inserted -- but also
mi-conjugation endings. On this phenomenon see p. 220, n. 35.
62 63
Present Indicative Preterite Indicative
Singular Plural Singular Plural
1 i umni, aummeni un aumen
62
2 autti autteni, auteni,
uteni
auta
3 auzi uwanzi
63
auta auer
aus paradigm
Imperative
Singular Plural
1 uwallu
2 au auten
3 audu uwandu
63
13.33. Innitive II: uwanna, Verbal substantive II: uwtar. Imperfective: uke-.
62. See Otten 1973: 30.
63. These forms are identical to the corresponding forms of the verb uwa- to come.
Conjugation of i-Verbs 13.33
230
Chapter 14
MEDIO-PASSIVE CONJUGATION
14.1. We follow HE and Neu (1968b, 1968a) in presenting paradigms of medio-
passives based on whether or not the stem ended in a consonant or a vowel.
Consonantal Stems
14.2. The following are verbs with consonantal stems conjugated in the medio-
passive: ar- to stand, e- to sit, and ki- to happen; become exclusively so, pa- to
protect and tarupp- to assemble (intr.); to be united; to be nished predominantly
so.
Present Indicative
Singular
1 arari,
araari
aari kia,
kiaari
paaa
2 artati, artari tari kita, kitati paata
3 arta, rta,
artari, rtari,
artri
ea, eari,
eri, iari
ka, kia,
kari, kiari,
kiri
paa, paari taruptari,
taruptri,
tarupdari,
tarupta
Plural
1 arwata euwata,
euata,
euwatati
pauwata
2 paaduma
3 aranda,
arnta,
arnta,
[a]randari,
arantri,
arantri,
arntari
eanta,
eanda,
eantari,
eandari,
eandri,
aanta,
aanda
kianta,
kiantari
panta,
paaantari
taruppantari,
taruppandari,
daruppantari
00-HittiteGrammar.indb 230 6/5/08 4:17:35 PM
231
Preterite Indicative
Singular
1 arati,
araat
at, ati,
aat(i),
aatti,
iaat
kiat, kiati,
kiaat
paaat,
rr-aaat,
rr-aat,
paaaat
2 artati, artat tat, eat kitat, kiat
3 artat eat, eadi,
eati, tat
kiat, kiati paatat taruptat,
darupdat
Plural
1 arwatat
2 kidummat
3 arantat,
arandati
eantat kiantat,
kantati,
kiantati
taruppantati,
[tar]uppand [ati],
taruppantat
consonantal paradigm end medio-passive imperative
Imperative
Singular
1 araaru
2 arut, rut ut kiut
3 artaru earu kiaru paaru taruptaru
Plural
2 ardumat,
ar-tum-ma-at
dummat kidumat paadumat
3 arantaru,
arandaru
kiandaru paandaru taruppantaru
14.3. Note: Active and medio-passive verbs take the same forms of verbal substan-
tive, innitive, and participle.
14.4. Verbal substantive: taruppuwar. Innitive: aanna. Participle: arant-, aant-,
kiant-, paant-, taruppant-.
Vocalic Stems
14.5. The following verbs with vocalic stems are conjugated in the medio-passive:
iya- to go and ki- to lie (down) (exclusively so), nai-lneya- to turn (intr.) and
zaiya- to ght (predominantly so; see also 14.11, p. 233):
Medio-Passive Conjugation 14.5
00-HittiteGrammar.indb 231 6/5/08 4:17:36 PM
232
Present Indicative
Singular
1 iyaa, iyaari,
iyaaari
neyaari zaiyaa,
zaiyaari
2 iyattari, iyattati neyattati,
naita(ri)
3 iatta, itta, iyatta,
iyattari iyaddari,
iattari, iyattri,
iyatari
na, neyari zaiyattari kitta, kittari
Plural
1 iyawata zaiyawatati
2 iyadduma zaiyadduma
3 iyanta, iyantari,
ientari, entari
1
nanda,

neyanta,
neyantari
ki(y)anta, kianda,
kiyantari
1

vocalic stem paradigms
Preterite Indicative
Singular
1 iyaat, iyaaat neyaat zaiyaat
2 iyattati
3 iyattat neyattat, neyat zaiyattat,
zaiyatta
kittat, kittati
Plural
3 iyantat neyantat, neyantati kyantati
2

active imperative vocalic: paradigms
Imperative
Singular
1 zaiya[aru]
2
2 iyaut naiut, neut zaiyaut
3 iyattaru neyaru kittaru
Plural
2 iyaddumat naidumat zaiyaddumat kiddumati
3 iyantaru neyandaru
14.6. Verbal substantive: neyawar. Innitive: zaiyawanzi. Participle: iyant-,
nant-, neyant-, zaiyant-.
1. KUB 33.52 ii 12 and KUB 40.28:1. On this spelling see Melchert 1984b: 14.
2. KUB 26.35:7. For restoration as imv. sg. 1 see imv. sg. 2 zaiyaut in line 8 and Neu 1968b: 203.
14.6 Medio-Passive Conjugation
00-HittiteGrammar.indb 232 6/5/08 4:17:36 PM
233
Chronology of the Medio-Passive Endings
14.7. It has been noted that many of the present medio-passive endings without -ri
or -ti (-a, -ta, -a, -wata, -duma, -anta) appear to be especially common in docu-
ments from the earliest period. Preliminary isolated observations were made by Neu
(1968a: 3440). A statistical investigation that paid attention to the chronological order
of the texts and manuscripts was carried out by K. Yoshida. The gist of his conclusions
(Yoshida 1990: 95102) was the following. The locus of -ri in OH was verbs in -ari
but more specically those in -ri. From this starting point the -ri spread to all classes
throughout the history of Hittite, although some resisted longer than others.
3
In the
preterite, on the other hand, the endings without -i spread at the expense of those with
it. According to Neu (1968: 34), the motivation for both developments was probably
the same: to make the medio-passive endings align with those of the active in having
the present with nal -i and the preterite without it.
Medio-Passive Stem Formation
14.8. Few medio-passive forms of mi-verbs that show ela ablaut in the active are at-
tested. For what follows see Neu 1968a: 5051. The verb pp-lapp- to take shows con-
sistent a-vocalism (pret. sg. 3 appattat, pl. 3 appandltat versus active pta and pper),
but for mer- to disappear we nd both martari and mertat. The i-verb a()le- to
open appears with e-vocalism (etat). The i-verbs in -ai- (13.20, p. 222) show the
stem in -i-: miyari grows, iyati was pressed forth. From au()- to see we nd
medio-passive au- with transitive meaning and u(wa)- in the sense to appear, show
oneself. Stems in -kela- appear to show both e- and a-vocalism, as in the active (see
on this problem 12.33, p. 204, and Neu 1968a: 4243).
14.9. In a few cases, verbs that show no sux in the active form their medio-passive
stem by adding -(i)yela-: e.g., ark- to perish, italk- to make smooth. See Neu
1968b: 42 and Melchert 1997c: 8486.
14.10. A fairly large group of verbs occur in the medio-passive only in the marked
imperfective with the sux -ke-: akk-, ariya-, aiwante-, earwa-, iki-, impi-,
inu-, iiya-, kallare-, kunk-, kurkuriya-, makke-, marli-, mugi-, pai- (to go),
pale-, pariparai-, pargawe-, parke-, dala-, dankuwa-, tepawe-, uiya-, uli-,
uaniya-, uwa- (to come), wal-, wallu-.
14.11. Perhaps because of the association of the medio-passive with intransitives,
there is a strong tendency for medio-passives with transitive meaning to be converted
to active i-verbs in New Hittite (the so-called dynamic use; see Neu 1968a: 5456 and
Houwink ten Cate 1970: 1820). Note, e.g., anna- to litigate, judge (with older pres.
sg. 2 annatta(ri), sg. 3 annari, imp. sg. 3 annaru etc., but attested in NH as active
3. Whether there are further conditioning factors on the pattern of its spread as Neu claims requires
further study. On the present endings -tati and -watati see K. Yoshida 1987.
Medio-Passive Conjugation 14.11
00-HittiteGrammar.indb 233 6/5/08 4:17:37 PM
234
i-verb annai, annai, annanzi, annau), par- to chase, expel, and arra- to
cross, transgress.
14.12. The uses of the medio-passive will be discussed in chapter 21.
14.12 Medio-Passive Conjugation
00-HittiteGrammar.indb 234 6/5/08 4:17:38 PM
235
Chapter 15
GRAMMATICAL AGREEMENT
Types of Agreement
15.1. Grammatical agreement is part of a system of marking that enables speakers
to indicate which parts of their utterances correlate with others (possibilities include
number, gender, and case). In Hittite the following correlated speech elements show
agreement in one form or another: (1) nouns with attributive adjectives or participles;
(2) nouns with relative, indenite, demonstrative, and possessive pronouns; (3) nouns
or pronouns and their appositives (see 15.4); (4) antecedent nouns and their adverbial
appositives; (5) antecedent nouns and their compared nouns in like . . . l as . . . clauses
with mn or maan; (6) the subject and the predicate noun or adjective in sentences
with linking verbs such as - to be and ki- to become; and (7) nominal or pronomi-
nal subjects and their verbs.
15.2. (1) Nouns with attributive adjectives or participles: arranza alki washed
barley KUB 1.13 iii 14 , damin antuan another person KBo 4.2 i 25, arran wtar
polluted water KUB 13.3 ii 30 , [(araz)]en (var. arazenante) utnante []mante
all neighboring lands KUB 24.4 rev. 7 (NH) with dupl. KUB 24.3+ ii 49, tarpalliu
waandu clothed substitutes KBo 4.6 obv. 29, k ror-zi i-natarmit this my rst
exploit att. ii 29 (KBo 3.6 ii 13), k xtn.xtn.xr dannatta these depopulated lands
att. ii 56, alli xtn-e large land KUB 23.11 iii 33, ts:i utuut tn.xn 4 arantet
akuwanzi they drink from animal-shaped vessels of lions standing on all fours KUB
10.89 i 2021; dankuwayaz(a) taknaz(a) from the Dark Earth KBo 5.3+ iv 56 and KBo
6.28 rev. 41 (both NH).
15.3. (2) Nouns with relative, indenite, demonstrative, and possessive pronouns
(6.5, p. 139) as attributes: []alla r-er kue ne natta [parkunuttati(?)] The houses
which are great (in Tappaanda) [have] not [been exempted(?)] (from taxes and cor-
vee) (all relevant forms here are nominative-accusative neuter plural) KUB 1.16 ii 66
67 (OHlNS); nunna r-er kuit uwen [(n)]unnakan urorn-ti+ anda artat the
goddess came to us there in the house which we had built for ourselves att. iii 67;
[(nu) kue k]ue aauwar ta [n(at) +\+ rnr.xr aanduli peun whatever sheep-
folds there were, I gave them to the troops for garrison KBo 10.2 i 78; kupiyati . . .
t kuiki not any plot KUB 13.35+ iv 44 (NH), tuel t-an l kuinki not any person of
yours KUB 14.3 i 19 (NH); atkun tu.xx-an kuinki any dicult (lit. tight) day KBo
4.14 iii 19 (NH); uppear kuitki any gift KUB 14.3 i 5455 (NH); appat memia
00-HittiteGrammar.indb 235 6/5/08 4:17:39 PM
236
the aforementioned matter KUB 14.14 obv. 36; k arkuwarri these prayers KUB
6.45 i 26 (NH); kedani pedi in this place KBo 4.14 iv 48 (NH); kardiata of your
heart KBo 3.7 i 26; tuzziami my army KBo 2.5 ii 13; kn xtts-an this woman
KBo 4.6 rev. 12; aataret his procreative power KBo 6.34 ii 31; areme your
companions Laws 55; aruwanduu her unborn children (pl. acc. com.) Laws
17, 18; ammdaz t-az with my own hand(s) KBo 3.4 iv 45 (NH).
15.4. (3) Nouns in apposition show agreement in number and case and normally
supply additional information about a substantive which may or may not be essential
to the meaning of the clause. Nominative: lm[awaran au]zi
d
u-a
URU
Kumm[iya]
tn.so-li-i! itoi-u Let not the Stormgod (Teub), the brave king of Kummiya,
see him! KUB 33.93 + KUB36.7a + KUB 17.7 iii 32 (Ullik., NS);
URU
Tiliura tnt-a
ts:i tu.xx
m
antili dannatti ta Tiliura, the city, was a waste since the time (lit.
day) of antili KUB 21.29 i 1112 (NH); takku i.t
19
.it-a i-a nama xtts-za
takya tnt-ri aki If a person, male or female, is killed in another town KBo 6.2 i 7
(Laws 6, OS). Accusative: nukan
m
Zidantan adda(n)an kuenta he killed Zidanta,
his father KBo 3.1+ i 68 (OHlNS);
m
Telipinuaz
f
Itapariyan antezziyan r-si
arta Telipinu had (as wife) Itapariya, his older sister KBo 3.1+ ii 910 (OHlNS);
+ui:+(a)nnaza . . . . . .
f
urorn.xr-rn-inna utxt.[(xtts-an)] ata My father
begat us (four sons) . . . and urorn.xr-rn, a daughter att. i 911 (NH); dapiy[a
(urorn.xr)-a (addan)]
d
Kumarbin iamii I sing of Kumarbi, father of all the
gods KUB 33.96 i 34 (NS) with dupl. KUB 33.98 i 3 (NS). Genitive: napa urorn.
xr-i attaa
m
Zidanta aret aner then the gods avenged the blood of his
father Zidanta THeth 11 i 6970 (OHlNS) (restored from multiple copies). Dative-
locative: nukan apa pula[z] +\+
d
t
URU
Nerik utxtx+ iyanti []aptat By des-
tiny (lit., by lot) he too was associated with the Stormgod of Nerik, your beloved son
KUB 21.27 i 1112 (NH).
15.5. (4) Other appositional constructions seem to be almost adverbial in force, i.e.,
they serve more to dene or restrict the nature of the verbal action than the noun with
which they are in apposition: numu kappinpat utxt-an
d
ts:+u
URU
amua +\+ +ui:+
wkta ts:+u of amua asked my father for me (when I was) just (-pat) a small child
KBo 6.29 i 78 (NH); ka tuk
m
uqqann appezzin tn.or
7
-an ar dun I took
you up, ukkana, (when you were) a lowly (lit., the last or a young
1
) dog KBo 5.3
i 23 (MHlNS); nu tu[k
m
Ma]ddu[wa]ttan +oi ux.xrx+ [utxt.xr]x+ rnr.
xrx+ya +ut
d
tft-st kiduwn[d]u u[inut] The father of His Majesty saved your
1. The signicance of the adjective appezzi- is debated by those who reckon now with the collated read-
ing tn.or
7
dog. Klinger (1992: 192, 208 n. 42 letzter Hund) reckons with a concept of rank, while T. van
den Hout (personal communication) favors the concept of age (young dog), comparing tn.ftn orx-an as
a puppy in the loyalty oath KUB 26.32 i 8. Since the separate term tn.ftn was available for expressing
young dog, puppy, it seems more likely that Klingers understanding is correct. Regardless of ones deci-
sion on this, the entire context suggests that the phrase as an appezzi- dog is intended to show ukkanas
lowly estate when he was elevated by the Hittite king.
15.4 Grammatical Agreement
00-HittiteGrammar.indb 236 6/5/08 4:17:40 PM
237
life, Madduwatta, together with your wives, your children, and your troops (when
you were) starving KUB 14.1 obv. 89 (MHlMS); if a slave pays the bride-price for a
free young man nan
LU
antiyantan pzi and takes him (as) a live-in son-in-law KBo
6.3 ii 28 (Laws 36, OHlNS). Such examples are dierent from double accusatives with
verbs like iya- to make, alzai- to call, and te-, etc. (on which see 16.21, p. 247).
For double accusatives used in partitive apposition see 16.24 (p. 247)
15.6. Some adverbial examples are participles in the nominative case, appearing im-
mediately before the verb: p[ai]t
2
aan or.-a arkuw[a]nza eket he lay down
on the bed in the bedroom booted (i.e., with his boots one) KUB 24.8 i 2526 (pre-NHl
NS); naza t::t
m
Appu waanza eket and she (his wife) would lie down with Appu
clothed (i.e., with her clothes on) KUB 24.8 i 30; [kuiwattakan] s+ xtn
URU
at[ti
LU
]uyanza anda uezzi whoever of the land of atti takes refuge with you as a fugitive
KUB 14.1 obv. 34 (Madd.);
d
Telipinu llaniyanza uet uwantiwantaz titiketta Teli-
pinu came angrily: with lightning he thunders KUB 17.10 ii 3334; [(nukan ne)]pia
x-a zikpat aanuwanza
d
tft-u arrekei you pass through heavens gate as an
established Sungod KUB 31.127 + ABoT 44+ i 3031 (this case could be an apposi-
tional vocative: O established Sungod (see 15.7 and 16.16, p. 245). Since most
nominative participles are predicates, the less-common non-predicative occurrences
tend to assume the role of apposition to the subject. In some cases they translate best as
adverbs of manner of adverbial phrases.
15.7. An appositional constructionnot to be confused with the true vocative con-
struction (see 16.16, p. 245)is used for direct address.
15.8. (5) Antecedent nouns and their compared nouns in like . . . l as . . . clauses
with mn or maan. Nominative: nat urorn-ti+
LU
r.xr.xtu-a man . . .
warta the deity, like a reaper, has reaped it KUB 29.7 rev. 2021 (MHlMS); nukan
URU
Neriqqa tnt-a
NA
4
aku orx-an [aruni] anda ta the city of Nerik was like a
seashell [in the sea] KUB 21.19 + 1303lu iii 1415 (NH). Nominative-accusative:
nukan
NA
4
t.t-i
d
Upeluriy[a] [zo-n]i
UZU
zo.tut-ni
GI
iyattal maan artat The
Basalt (Monster) was standing on Ubelluris right shoulder like a shaft KUB 33.106
iii 4647 (Ullik.). Accusative: numu urorn-:+ uinut numu wadula kat[t]an ara
iiyandan i-an mn ara l Save me, O my god, and release me, (who am) like a
man bound in sins KUB 36.75 iii 911 (OH?lMS); nuwaru ara dannaru
DUG
fti.
nr. maan duwarnikei you (the king) will break them (your enemies) like empty
vessels KBo 15.52 v 1617 (NS); apnmakan
d
ts:+u . . . xt
6
-un orx-an palaza
rorn-pa itapta Is:+u trapped him (e.g., Uriteup) like a sh in a net KBo 6.29 ii
3334 (NH). Genitive: [u]minzana ln.xr-amman tn.nn.n-a mn pang[ur] 1-i\
tu But let the clan of you, my subjects, be one like (that) of the wolf KBo 3.27:1516
(OHlNS); the rst three words in this sentence are all genitives in agreement.
2. Restoring p[a-i-t]a-a-a-an following pa-a-i-ta-a in KUB 24.8 i 29, against [a-a-t]a-a-a-an in
Siegelova 1971: 4.
Grammatical Agreement 15.8
00-HittiteGrammar.indb 237 6/5/08 4:17:42 PM
238
15.9. (6) In nominal sentences and those with linking verbs such as - to be and
ki- to become the subject agrees with its subject complementwith a predicate
noun in case and number, and with a predicate adjective in case, number, and gender:
annai xt-a his mother (is) a snake KUB 1.16 ii 20 (OHlNS); i r-si aruwan
i.xr n.i[(-si i
LU.ME
\tssi)] arw his house (is) exempt, also his heirs and
relatives (are) exempt KBo 6.9 i 23 (Laws 51, OHlNS) restored from KBo 6.2 iii 2
(OS); xtn-e alli ta The country was large KBo 11.1 obv. 19; pdan mekki nakki
The place (is) very important (or hard to reach) ABoT 60 obv. 2021. For an old pl.
nom. com. see k toti 1 tt-e eer These were overseers of one thousand KBo 8.42
rev. 6 (OS). The example nu t.xr-u marante The people (are) deceitful KUB
19.23 rev. 11 shows good NH agreement. See further examples below in 16.5 (p. 242)
and following.
15.10. (7) A nominal subject agrees with its verb in number: utxt-uiwamukan
akta A son died on me, (and I didnt perform the last rites) KUB 5.9 obv. 26 (NH
oracle inquiry); i+++ itoi-++
m
Zuliyawa ap paiddu Let Zuliya go to the river
(ordeal) KUB 13.3 iii 29 (MHlNS); takku 2 i.xr na[(m)]a 3 i[(.x)]r akkanzi
if two or three men are killed KBo 6.3 ii 30 (Laws 37, OHlNS) with dupl. KBo 6.2
ii 11 (OS), numa urorn.uruir-e taman kartan dayer and the gods gave to them
(e.g., the sons of the Queen of Kane) a dierent appearance KBo 22.2 obv. 16 (OS);
takku i.xr zaanda ta 1
3
-a aki If men ght, and one dies KBo 6.26 ii 16 (Laws
174, OHlNS).
15.11. A pronominal subject agrees with its verb in person and number: uganza
utxt.xtts-:t t alzii I will not call her my daughter KUB 1.16 iii 25 (OHl
NS); kwa itoi-umi kia I will become your king KBo 22.2 rev. 15 (OS);
kwazz()an [(nu)n] I was enraged KUB 33.10 ii 6 (OHlMS); nu andan k
kian memallu And let me truly say the following KUB 6.45 + KUB 30.14 iii 7374
(NH); namaat zik malii or (if) you contemplate it KBo 4.14+ ii 78; zikwa tn.nn.
n-a kitat you have become a wolf KBo 6.2 ii 12 (OS); ziga parkuytar iya as for
you, make purication! (imp.) KUB 1.16 ii 67 (OHlNS); paprezzi kui he who deles
KBo 6.2 i 57 (OS); takku ln-a uwi na +\+ xtn Luwya paizzi if a slave runs away
and (he) goes to the land of Luwiya KBo 6.2 i 51 (OS), apa pait
d
rx-ni tet he pro-
ceeded to say to the Stormgod VBoT 58 i 16 (OHlNS); rorn-pamaa
URU
x.urorn.
n pait afterwards he went to Babylon KBo 3.1 i 28 (OHlNS); nu w
m
Arnuwanta
itoi.oi i
f
Amunikal xtts.itoi.oi otu.nr.-u tut.nr. warkandu sro
5
-
andu . . . rorn-pa pegaweni WeArnuwanda, Great King, and Amunikal, Great
Queenwill give back (to the gods) fat and goodly cattle and sheep KUB 31.124 ii
1417 (MHlMS); kuitmanwaza w (eras.) t\+
URU
attui ewen while we were in
attua HKM 17:56 (MHlMS); ume i.xr
GI
ftxti tamekatteni you (pl.) are
oppressing the ftxti-men KBo 22.1:3 (OS); umea apenian ten and you (pl.)
3. See LH 139 n. 477.
15.9 Grammatical Agreement
00-HittiteGrammar.indb 238 6/5/08 4:17:43 PM
239
must keep doing as before (imp.) KBo 22.62 + KBo 6.2 iii 20 (OS); armawantea
kuie nuza apya t aanzi as for those who are pregnant, even they do not give
birthKUB 17.10 i 15 (OHlNS); takku i-a itti+ orxr-aa iele nat anda aranzi
if a free man and a slave woman are single, and they (at) cohabit KBo 6.3 ii 16 (OHl
NS), apmaan pper nui kutuker but they (the citizens of Tappaanda) seized
him (Prince uzziya) and began to intimidate(?) him KUB 1.16 ii 64 (OHlNS).
Lack of Agreement
In Gender
15.12. It is doubtful that there are more than a few cases of true lack of agreement in
gender. When a resumptive pronoun refers back to two nouns, one of which is common
and the other neuter, the speaker cannot avoid choosing between the two: for example
d
rxwa i.xr i.xn (com.) xtn.xtn.xr i.xn (neut.) tukpat +\+ itoi s+r+t
oln.xr zikkezzi nuwaru (com.) ara dannaru uto.xx.nr. maan duwarnikei,
The Stormgod puts the enemy men (com.) and the enemy lands (neut.) under your feet,
O king, and you smash them (com.) like empty vessels KBo 15.52 v 1417.
15.13. Other cases of disagreement in gender have been claimed for parts of the
body, but there may be better explanations of the forms in question. For example,
genuit his knee KUB 7.1 iii 7 is neuter, but gnuu his knees KBo 3.4 ii 20;
KUB 13.4 ii 58 appears to be common gender. But Eichner (1979) has interpreted the
latter as accusative plural of a noun genui- Kniegelenk (see also p. 101, n. 109, and
p. 138, n. 7). The two forms of the noun hand: kear (previously thought to be neuter)
and keara (common) seem to show inversion in the possessive pronouns: t-aet
(which implies *kearaet) (Laws 4) and keari (Laws 3). But Neu (1983: 97
with n. 354) has plausibly proposed that kear in keari is an asigmatic common
gender r-stem (see 4.82, p. 115, with the paradigm), since there is no attested example
of kear as a (neuter) direct object. The use of the neuter possessive pronoun in t-
aet is simply erroneous. Other possible instances of disagreement in gender may
also require reconsideration: mn antuwaa uppi if a person is pure (neuter!) is
rather: If a person is in a pure state [d.-l.] KBo 5.2 i 3; and alongside correct ipantan
mandan the entire night we can regard or
6
-an man (KUB 1.13 iii 28; KBo 3.5 i
30) as faulty. On the other hand, there is no conict between antuhtar and kuinna in:
nu[kan a(ntuatar)] (neut.) kuinna (com.) apel +\+ tntsi rorn-pa [(peutet)] and
he led back the population, each one (person of the antusatar) to his (own) city KUB
19.11 iv 1415, since the kuinna, apel, and -si refer to units within the antusatar. The
phrase umilin genu KBo 10.37 ii 33 may be explained as regular agreement, with
umilin being a contracted form of *umiliyan (1.76, p. 32), sg. nom.-acc. neut. of
an extended stem *umiliyant- rm beside umili-.
4
4. Catsanicos 1986: 147 n. 154. But his comparison with uppin in KBo 12.89 ii 13 is false, since
appuriyan uppin is common gender.
Grammatical Agreement 15.13
00-HittiteGrammar.indb 239 6/5/08 4:17:45 PM
240
15.14. Sure cases of gender variation are few. We do nd sg. nom.-acc. neut.
memi(y)an word alongside usually common-gender memiya(n)-. orft-an ear is
sg. nom.-acc. neut. in KUB 8.83 4, 5, alongside usual common-gender itamana-.
5
But
the vast majority of alleged cases listed in the dictionaries simply do not exist. Some
examples involve a failure to recognize the Hittite practice of writing the bare stem form
with numbers (see 9.24, p. 159, and 16.3, p. 242). Writings such as 1
GI
intaluzzi or
1
DUG
uppar (4.84, p. 116) are no evidence for neuter gender; both intaluzzi- shovel
and uppar- bowl are consistently common gender. Other putative cases involve com-
mon-gender nouns with collective plurals and neuter nouns with count plurals (3.13,
p. 68). For example, alpa- cloud is only common gender. But in addition to its regular
accusative plural alp clouds it has a collective plural, alpa cloudbank. Similarly,
luttai- window is neuter only. But since the regular (neuter) plural would have had
a collective sense in OH, for reference to discrete individual windows we nd luttae,
luttau. This is a matter of distinction between collective and count plural, not varia-
tion in gender. This number distinction is widespread in Old Hittite (pace Prins 1997:
6263). In cases written with logograms we cannot be certain whether we are dealing
with forms of the same word: e.g., in the case of rman- illness, we nd alongside oro-
an also sg. nom. com. oro-a. The latter may reect real variation in gender or conceal
another word for illness, sickness.
In Number
15.15. Disagreement in number may be accounted for in several ways. A collective
singular can be conceived as a plural: numukan orx-an t.xr-annaza s+
d
ts:+u
o:+ kaniwar s+ r:+ya [a]ulan aur but when the population (sg.) saw
(pl.) the honor of ts:+u, my lady, to me and the favor of my brother toward me
6
att.
i 3032 (NH); [x]n-zakan kui tepawe i [par ]ter aptmakan man a[ra
]apirpat they utterly annihilated it (apt, sg. neut.) all (sg. neut.), the army (com. sg.)
which (com. pl.) escaped (pl.) few in number (com.pl.) KUB 14.1 obv. 48 (MHlMS).
15.16. As in some other Indo-European languages, neuter plural (collective) nominal
subjects take singular predicates (3.123.13, p. 68). When the predicate is an intran-
sitive verb, it will be singular:
m
kalima uddr ari but accusations arose against
Akali KBo 3.34 ii 18. When the predicate is made up of the verb to be and a predicate
nominative adjective, the adjective takes the singular form by attraction to the form of
the verb to be (van den Hout 2001). This pattern is then extended to nominal sentences,
which have no linking verb (e.g., to be). That the singular form of the predicate adjec-
tive is due to its attraction to the verb to be is shown by the fact that neuter plural nouns
take neuter plural attributive adjectives: idlawa uddr, liiyala . . . uddr, akkanta it
uwanda uddr, kue uddr, tami kukki uddr.
5. For a possible explanation of the latter example see Melchert 2000: 66. The source of neuter memiyan
is unclear. Perhaps a neuter singular participle of mema- the spoken thing.
6. The dative -mu receives the verbal action implied in both kaniuwar and aulan.
15.14 Grammatical Agreement
00-HittiteGrammar.indb 240 6/5/08 4:17:46 PM
241
15.17. There are a few apparent examples of common-gender plural subjects
with singular predicates: w ka rains will occur KUB 8.1 iii 89; kwa wall
maan arlipa artari as these w. stand arlipa KUB 29.1 iv 1011;
7
+\+ so.ut utxt
MUNUS
sowa rk idlaweketta nuwa k rk kandati The signs were coming
out unfavorably for the person of the priestess son, and these signs occurred KuT
49:45.
8
All examples attested thus far are with intransitive verbs, a feature that would
be shared with the usage of singular verbs with collective plural subjects. The phenom-
enon calls for further investigation. For the Hittite construction as the possible source
of the Greek Pindaric schema see Watkins 2000: 23.
15.18. Singular forms are often used following numbers greater than 1 (see 9.21
and following, p. 158). And, nally, there are cases where a scribe used a plural verb on
the second of two clauses in which one member of a pair of persons acts: 1 i +isiot
KU
kuran kar (a)pzi
URUDU
n+[,,t\\+ya] 1 i +isiotma s+ urorn-tt+
GI
uppanta
zeriyalliya karpanzi One palace guard lifts a leather shield (and) an axe, while another
palace guard lifts (text: they lift!) the gods uppanta and potstand KUB 11.23 v 812
(Note the extraposition of the
URUDU
n+[,,t\\+ya]).
7. This example supplied by Norbert Oettinger (pers. comm.).
8. The last example cited by Watkins (2000: 14). For the text see Wilhelm 1998: 178.
Grammatical Agreement 15.18
00-HittiteGrammar.indb 241 6/5/08 4:17:47 PM
242
Chapter 16
NOUN CASES
Generalities
16.1. As indicated in chapter 3, Hittite nouns and adjectives are inected in the fol-
lowing cases: nominative, vocative, accusative, ergative, genitive, allative, dative-loca-
tive, ablative, and instrumental.
16.2. Five of these cases (accusative, allative, dative-locative, ablative, and instru-
mental) are always adverbal, i.e., dependent on the verb (including participles), while
the genitive is usually adnominal (for exceptions see 16.57, p. 255, and 16.61, p. 256).
The ergative appears only as the subject of a transitive verb. Nouns and pronouns in the
nominative serve as subjects of nite verbs but as predicates only of the linking verbs
- to be, ki- to become, etc. (16.5). A noun in the vocative (together with any ac-
companying attributive adjectives), and rarely the nominative standing for the vocative
(16.17, p. 245), constitutes a clause in itself (Honer 1998a).
16.3. Stem form. There are several situations in which scribes used the bare stem
form. (a) As a real pronounced form it could represent the vocative case (3.28, p. 74)
or (b) a naming construction (16.15, p. 244). The bare stem represents (c) a purely
graphic entity when the noun in question is a proper name (personal, divine, or geo-
graphical), when it follows an Akkadian preposition (+\+, t\+, s+, etc.), or in listings
immediately following a numeral (nu 2
NINDA
zipinni nu . . . aeanzi KBo 5.1 ii 1617).
Situation (c) is referred to as Akkadographic in the CHD. Some Hittitologists have
such words typeset in uppercase italics like real Akkadograms.
Nominative
16.4. The nominative case is primarily the case of the subject of a clause, regard-
less of the nature of the predicate (noun phrase or verb). Thus in the sentence annai
xt-a his mother (is) a snake (KUB 1.16 ii 20), the noun anna is in the nomina-
tive case, the subject of the sentence whose predicate is xt-a. In the verbal sentence
iian ganezi its owner recognizes it (Laws 61), the noun i is also nomi-
native by virtue of its function as subject of the verb gane-.
16.5. Any demonstrative, attributive adjective, or other such form modifying the sub-
ject of the sentence will also be nominative under the conditions of grammatical agree-
ment (see chapter 15). Nouns or adjectives which serve as the predicate in sentences
00-HittiteGrammar.indb 242 6/5/08 4:17:47 PM
243
with linking verbs (to be, to become, to appear, etc.) also stand in the nominative
case. For example, xt-a snake in annai xt-a. The predicate nominative is
sometimes called a subject complement.
16.6. As in other languages, the verb to be may be omitted in Hittite in the pres-
ent tense. Because of the absence of an expressed nite verb, these sentences are often
called nominal sentences. Examples are: ux uitt-\tmawanna wannummiya
the wife of our master (is) a widow KBo 14.12 iv 1920 (D frag. 28); takku utxt.
xtts i-ni taranza if a daughter (is) promised (in marriage) to a man KBo 6.3 ii 5
(Laws 28); [ pa]rtauwa(r)etwa amiyanta apaauwa amiyanza its wings (are)
small, and it (is) small KUB 17.10 i 38 (Tel. myth, OHlMS). For -(u)wa see 28.2
(p. 354). See 15.9 (p. 238) for additional examples. In at least one example, the verb
be is also omitted when past time has been established in a preceding clause and car-
ries over to that with be: kar kui t\+
URU
Arinna
LU
t.nn k[(at)] i r-si aruwan
i.xr n.i[(-si i
LU.ME
\tssi)] arw Formerly, one who becaame a weaver in
Arinna, his house (was) exempt, also his heirs and relatives (were) exempt KBo 6.9 i
23 (Laws 51, OHlNS) restored from KBo 6.2 iii 2 (OS). See also the examples cited
in 16.15 (p. 244), where be is omitted in the context of mythical past time.
16.7. The predicate noun or adjective will also occur in the nominative case when the
verb - to be is expressed: gga
MUNUS
annanna mi and I am an annanna-woman
VBoT 58 iv 3 (OHlNS); nu k kuit mi and what is this that I am? (or and this which
I am) KUB 57.3:11 (NH); umea [urorn.xr-a u]ddn
1
naante ten You
(pl.) be afraid of the wordlmatter [of the gods] KUB 1.16 iii 4950 (OHlNS); [apa
k]uduwta l l andnpat du And let [those] false accusations never never be
established! KUB 1.16 ii 55 (OHlNS). Similarly when the verb is ki- to become:
zikwa tn.nn.n-a kitat You have become a wolf KBo 6.2 + Hrozny 1922 plate v,
ii 12 (Laws 37, OS).
16.8. Since appositional elements, regardless of their case, can serve for direct ad-
dress (see 16.16, p. 245), we naturally nd many nominatives so used.
16.9. The naming function of the nominative case is reected in its frequent appear-
ance in lists: 3 otu.xt.2 3 otu.xt.1 4 awitiza pi he shall give three two-year-old
oxen, three yearling oxen, and four weanlings (Laws 63). As the example shows, this
usage includes cases where the list as a whole may logically be construed as the direct
object of a verb.
16.10. Partitive apposition, richly attested in the accusative (16.2416.25, pp. 247
248; 16.42, p. 252) and rarely in the ablative (16.94, p. 266; 16.102, p. 267), may
also be found in the nominative: nu maan xt.xx-za meur tiyazi elia unnumanzi
1. The KUB copy is incorrect here; for the correct reading [. . . u]d-da-a-ni-i see already the translitera-
tion from the original by Forrer (1922; 1926: 15), followed by Sommer and Falkenstein (1938: 1415). The
photo published at the back of Sommers book is too dark to be useful for collations.
Noun Cases 16.10
00-HittiteGrammar.indb 243 6/5/08 4:17:49 PM
244
and when the year, (that is,) the seasonltime for the lling of the eli-s arrives KUB
21.17 iii 910 (NH). As in other cases of partitive apposition, the noun denoting the
whole precedes that denoting a part.
2
Vocative and Naming Construction
16.11. The vocative is used in exclamations and in direct address (Gterbock 1945;
Eichner 1974b: 557; and Honer 1998a). Hittite does not employ an interjection (like
English O) to introduce the vocative.
16.12. Form. For the inection of the vocative see 3.283.30, pp. 7476.
16.13. Syntax. The true vocative stands in its own clause (Honer 1998a), and even
in contexts of reported speech often does not take the quotative particle -wa (28.2,
p. 354, and following).
16.14. Occurrences. The true vocative is used principally in prayers, rituals,
and mythological texts:
d
tft-e iami O Sungod, my lord KUB 31.127 i 1 (OHl
NS);
3

d
tft-e arkui itoi-ue O Sungod, heroic king KUB 31.127 i 15;
d
Mukianu
LU
stxxi:+ O Mukianu, my vizier! KUB 12.65 + KBo 26.71 iii 18; iunimi zikmu
iya O my god! You made me KUB 30.10 rev. 1112; inana
d
tft-i katta ssxtn
peun O Sungod of Illness! I have hereby given to you a sacrice KUB 7.1 i 6;
see also ibid. i 15; [
d
Ala]uwaimi la namma zaikei O Alawaimi! Fight them no
longer! KUB 27.67 ii 31; (they bowed down to the Stormgod and asked:)
d
rxtn-ta
attau[mmi] kuwapi ta O Tarunta, ou[r] father! Where were you? KUB 33.66
+ KBo 40.333 iii 89 (MS). In all these examples the vocative stands in a separate
clause of its own and is not included in the immediately following clause with the verb
(see Honer 1998a). Sometimes, perhaps indicating an urgent address, the vocative or
other address form is repeated: rxr.nr. rxr.nr. Tongues! Tongues! KUB 44.4 +
KBo 13.241 rev. 22 and its syllabic equivalent lli lli KBo 20.59:16 (see 3.283.30,
pp. 7476, and p. 81, n. 25).
16.15. Also exhibiting the form of the bare stem (3.19, p. 71; 3.28, p. 75), but
not serving the purpose of direct address, is the form used for the introduction of new
names (Gterbock 1945; Honer 1968a; Hahn 1969; Laroche 1969b; Stefanini 1974;
Honer 1998c): xtts-a
4
si+et
f
intalimeni (there was) a woman, her name was
. KUB 33.121 ii 5; tnt-a si+-anet
UR
[
U
]udul, (there was) a town, its name was
2. The example cited could alternatively be analyzed as showing an accusative of direction (16.27,
p. 248: When the year arrives at the time for lling the eli-s.
3. Immediately followed by appositional direct address (16.16): andanza annena ia O just
lord of judgment KUB 31.127 i 12. For examples from Greek and Vedic of vocative immediately contin-
ued by a nominative form of address, see Meier-Brgger 2000: 249 (S 406). In those languages a nomina-
tive of this sort is not appositional address, as it is in Hittite.
4. On the sg. nom. xtts-a see 4.78 (p. 113) with n. 169.
16.11 Noun Cases
00-HittiteGrammar.indb 244 6/5/08 4:17:52 PM
245
udul KUB 24.8 i 7; and with reverse order of constituents in the second clause: nukan
er i-a
m
Appu si+-anet Up there (there was) a man, Appu (was) his name KUB
14.8+ i 910.
Other Forms of Direct Address
16.16. Appositional Direct Address. In other Indo-European languages, nominative
case forms appear in the role of vocatives (i.e., in direct address; see Meier-Brgger
2000: 24950; 2003: 265). For rare examples of this in Hittite see 16.17. In Hittite,
however, most instances of the nominative of direct address appear in an embedded ap-
positional construction, which may be used not just in the nominative but in whatever
case is appropriate to the syntax, often with the name in apposition to an accented second-
person pronoun used for direct address. Nominative: urorn.xn
d
rx-ni tet iya kuitki
[
d
]rx-a urorn.xn said to the Stormgod, Do something, Stormgod! KUB 17.10 i
3031; andama um uiti
ME
ft
7
mante . . . rft-mi rft-mi linkiketen Further-
more, all you (nom.) who are in charge of (lit. lords of) the stew, . . . you must take
monthly oaths KUB 13.3 ii 2026 (MHlNS). Genitive: nu tuel iuna uddanta natta
sro
5
-aat And did I not prosper by your word, O god? KUB 30.10 obv. 18 (OHlMS).
Here iuna is genitive singular in apposition to tuel your, of you. Dative-locative:
[tuqqawar]a +\+
m
[Maddu]watta +:++++ menaanta krur tu Against [you
(dat.) too], O [Maddu]watta (d.-l.), let him likewise be hostile! KUB 14.1 obv. 31.
16.17. Rarely, however, in the singular (not in the plural
5
) the nominative stands
in its own clause in place of the vocative. Only ve examples are known: (1) ai rxr-
a gag qamakan parkuin miriwantan arkin
GI
orunt t walantan tut-un
ipantaun O mouth, tongue, tooth! Lo, I have sacriced to you a pure, dazzling white
sheep, never struck with a rod KBo 15.10 ii 810 (MHlMS), ed. Szabo 1971; (2) i+++
d
Gula
d
r.ft-a
d
rx-na atta kuitwa uw[(a)] Thus (said)
d
Gula (and?)
d
r.ft: O
Father Stormgod! Why have you come? KUB 33.24 (+) 33.28 i 38 (OHlNS); (3)
d
tft-u
d
rx-a t ullatar O Sungod! O Stormgod! No disrespect (exists) KBo 6.13 i 9 (Laws
169; the last clause is elliptical; see 18.17, p. 282; and 30.13, p. 409); (4) wappuwa
urorn.xn-a kaza 12
UZU
n paprannanza tul t-it apiyan[za] parkunuwanza O
urorn.xn of the River Bank! Be cleansed now (ka) from delement and puried
by your hand with respect to (your) twelve body parts KUB 12.58 iv 13; rxr-a r-
a kuwapi pi Where are you going, O lord of the tongue? KUB 12.62 obv. 10, see
rev. 3. It is inappropriate, however, to claim (as does Kammenhuber
6
) that with com-
mon-gender substantives the vocative corresponds in general to the nominative, since
this implies that such forms are really vocatives that merely share the same endings with
5. Examples such as [um]a i.xr xtn
URU
Imirika mante linkiya ardumat You men of Imerika,
stand under oath! KUB 23.68 + ABoT 58 rev. 11, where the address formula might or might not be a part of
what follows immediately, should be considered appositional. See Honer 1998a: 4244.
6. Der Vokativ bei Subst. commune entspricht im allgemeinem dem Nominativ (Kammenhuber
1969b: 193).
Noun Cases 16.17
00-HittiteGrammar.indb 245 6/5/08 4:17:53 PM
246
the nominative. When Luraghi (1997a: 2.1.5.1) writes, It [the nominative] is also used
as a vocative in the plural, she ignores the fact that any case form in singular or plural
can be used for direct address as long as it does not occupy its own separate clause, as
does the true vocative. The fact that there appears to have been no distinctively plural
vocative ending does not justify her claim (2.1.6.2) that the nominative [plural] is
used in its place. See the treatment of this subject in 16.14 (p. 244).
Accusative
16.18. The accusative case is the primary case governed by the nite transitive verb.
As such its most common use is as (1) the immediate direct object of the transitive verb.
But, like the accusative case in Greek and Latin, the Hittite accusative can also serve as
(2) a second object, both in a predicatival sense (as in English he made me captain of
the team, he called me a coward, they consider him handsome; see 16.21),
7
and in
partitive apposition, from which develops the accusative of respect (see 16.2416.25),
(3) a cognate accusative (the gura etymologica of the classical languages, see 16.26,
p. 248),
8
(4) an accusative of direction (see 16.27, p. 248), (5) an accusative of the way
(see 16.28, p. 249),
9
(6) an accusative of extentlduration (see 16.29, p. 249), and (8)
an adverbial accusative (see 16.30, p. 249).
16.19. The principal use of the accusative case is as the direct object of transitive
verbs: [(takk)]u i.t
19
.it-an kuiki nikzi tan itarnikzi nu apn ktizzi pdiima
i.t
19
.it-an pi If someone injures a man and makes him ill, he shall look after him
and give a man in his place, ( . . . but when he recovers, he shall give him six shekels of
silver) KBo 6.2 i 1617 (Laws 10, OS) restored from KBo 6.3 i 25 (NS).
16.20. The normally intransitive verb -la- is used with -z(a) transitively in the
sense to dwell in, inhabit, occupy and thus takes an accusative direct object (HW
2

E 97, 10910): [k]awatta xtn
UR.SAG
Zippal +o[ot\] nuwaza apnpat ei
nammamawa[z] par tamin aptin tamai xtn-e zr-it l tari I have herewith
(ka) given you the mountain country Zippal. Occupy only it. Beyond (this), do not
intentionally occupy any other river country or land KUB 14.1 obv. 1920 (MHlMS);
[
d
tf]t-uza
URU
zrxnrn-an zi,
d
si\-amaza
URU
Kuzinan [zi],
d
t-aza
URU
Kummiyan
[zi ],
d
ts:+u-imaza
URU
Nenuwan [zi ],
d
Nanayamaza
URU
Kiinan [zi]
URU
x.
urorn.n-anmaza
d
xn.tft-a [zi ] the Sungod dwells in Sippar, the Moongod
dwells in Kuzina, the Stormgod dwells in Kummiya, (the goddess) ts:+u dwells in
Nenuwa, Nanaya dwells in Kiina, and in Babylon Marduk dwells KUB 24.8+ iv
1318 (NS); [mnwara tm]a uezzi nuwa-kn xtn-e peran ten [If ] he does
7. See on this type van den Hout 1992.
8. See HE 119 sub 199a. Note, however, that we view Friedrichs second type of inneres Objekt as
an adverbial accusative (19.14, p. 292).
9. See HE 120 sub 201b. Example: mankan
UR.SAG
Teinan ar pun I would have gone up Mt.
Teina.
16.18 Noun Cases
00-HittiteGrammar.indb 246 6/5/08 4:17:55 PM
247
[not] come, occupy the land in advance! KUB 14.16 i 1718 (annals of Mur. II, NH).
Note that ei in KUB 14.1 is formally active transitive, and tari in the same passage
is a transitive medio-passive. This usage ( pace HW
2
E 10910) is at least as early as
MHlMS, not coined in the reign of Murili II.
16.21. Double accusatives (van den Hout 1992) are particularly common with verbs
which are causatives of transitive verbs: u-an zi he crosses the river and otusi
u-an znukezzi (zinu- is the causative of zai-) he makes his ox cross the river, both
in KBo 6.2 ii 3032 (Laws 43, OS). But verbs other than causatives of transitives also
take double objects, among them: iya- to make something (rst acc.) into something
(second acc.); nu aruna iru it he made them (u) borders (iru) of the sea
(KBo 3.1+ i 8, OHlNS); nu itoi-u . . .
LU.ME
rr.i iyanun I, the king, made them
farmers (
LU.ME
rr.i) KBo 3.1+ ii 2930 (OHlNS); ln:+mawa nwn par dai
nuwaranzakan
LU
+i:t:+ iyami I refuse to (nwan; see 26.1, p. 341; and 26.19,
p. 344) take my servant and make him my husband KBo 5.6 iii 1415 (NH); nuza
URU.
d
t-an al[(li) +su]i iyat he (Muwatalli II) made the city Taruntaa the great place
(i.e., the new capital city) KBo 6.29+ i 32 (NH).
16.22. From the fundamental meaning to make into constructions of this kind de-
veloped into a means of expressing the idea to treat like: i +\+ utxt.xr
URU
N[(a
id )]lu natta kuedanikki takkita [ln'u] annu attu it He harmed none of the
citizens (lit., sons) of Nea, but treated [them] like mothers (and) fathers KBo 3.22
obv. 79 (OS) with dupl. KBo 22.5 obv. 7 (NS); kuiwaranzan [di nuwar ]an
uppiyaar ui-zi who will [take] him (the speakers son) and treat him like a gift?
KUB 33.93 iii 2829 (Ullik. I A).
16.23. Other verbs which like iya- take a double object are: alzai- to call te-
ltar- to designate as, declare to be, allanu- to raise up to be, iai- to provide
someone with something. Examples: [itoi-uanza] utxt-la(n)man alziun I,
[the king,] have proclaimed [him] (to be) my son KUB 1.16 ii 34; [numa ftn-lan]
labarnan tenun [to you] have I declared [the young man] to be the Labarna KUB 1.16
ii 23; takku . . . .X.nr. s+ i
GIS
ftxti arkantan taranzi if they declare the elds
of a craftsman to be vacated Laws 40; nu nti-lun (i.e., idlun) tu.xx-an i.nti-an
(i.e., uwappan) allan[(ukezzi)] and he raises up a bad day to be (or which is) an
evil KUB 33.96 i 6 (Ullik. I A); xtts.xr wannummiu xr-an ii he provides
the widow women with a task KBo 16.54+ 15.
16.24. Another double accusative construction in Hittite is partitive apposition,
which corresponds to the Greek . It is found almost exclu-
sively in relation to the human body: takku ln-an nama orxr-an xrn
14
et kuiki wki
if anyone bites o the nose of a slave boy or a slave girl (lit., bites a slave boy or
slave girl, namely hislher nose) KBo 6.3 i 35 (Laws 14, OHlNS); (if any child is sick,)
nan tukku igai and I anoint him, (namely his) members KUB 7.1+ i 40. That this
Noun Cases 16.24
00-HittiteGrammar.indb 247 6/5/08 4:17:58 PM
248
construction only became popular after Old Hittite is clear from passages in the laws
where the Old Script copy KBo 6.2 uses the genitive construction and the NH copyist of
KBo 6.3 replaced it with the (Laws 1113). Partitive apposition overlapped the
function of the so-called split genitive (see 16.38, p. 251) and eventually replaced it.
16.25. When the partitive accusative construction undergoes passivization, the
whole (scil., the person) becomes the subject, while the part (scil., the body part)
remains in the accusative, producing the so-called accusative of respect. Examples: na
genzu amikta t na
UZU
pantaan amiktat he was bound with respect to (his)
genzu (scrotum?), he was bound with respect to his bladder KUB 7.1 iii 56 (CHD P
95); urorn.xrzakan xtxi\i idlauwaz uddnaz . . . rrante ten Be washed
(clean), O gods, with respect to your mouths, from evil words, . . . KUB 43.58 ii 39
41; wappuwa urorn.xn-a kaza 12
UZU
n paprannanza tul t-it apiyan[za]
parkunuwanza O urorn.xn of the River Bank! Be hereby (ka) cleansed from de-
lement and puried by your hand with respect to (your) twelve body parts KUB 12.58
iv 13; 2 foma! so.uti z.ol prin tittalitaimenzi Two garments, head cover(s),
blue, tittaliti-ed (i.e., decorated in a certain way) with respect to the borderlhem KUB
12.1 iv 43 (inventory, NH), see CHD P 386. In rare cases one nds the accusative of re-
spect extended beyond the passive construction:
d
t-a ror.nr.-wa [i]aruwanza The
Stormgod, tearful of eyes, . . . KUB 33.113 + KUB 36.12 i 3031 (Ullik. II B).
16.26. The cognate accusative, the gura etymologica of the classical languages and
of several Semitic languages,
10
exists in Hittite as well. The characteristic of this con-
struction is a verb which takes as direct object a noun from the same root or stem (hence,
cognate accusative): annear anna- to judge a judgment; uittear uittiya- to
pull out tufts; ukmae uek- to incant incantations; ullanzan ulle- to ght a ght;
iiul iiya- to tie a tie; iuear iuwa- to heap a heap; kupiyatin kup- to plot a
plot; memiyan mema- to speak a word; arnikzel arnink- to make compensation;
atan e- to sleep a sleep; uppear uppa- to send sendings.
16.27. Accusative of Direction. As the dative-locative case answers the question
in what place? so the allative and occasionally the accusative case answer the ques-
tion to what place? (HE

201): nuzakan arunan tapua
11
[iyanniyanun] [I went]
to the side (allative) of the sea (acc.) KUB 17.7 ii 1112 (Ullik. I); numakan ntn.
so-an paranzi and they will chase you into the mountain (acc.) KUB 13.3 ii 11;
tu alki[t]n tarnae and I release them (scil., birds) to the branch KBo 17.1+ iv
38. This is what Neu (1980: 30 n. 67) calls Akkusativ der Richtung (accusative of di-
rection); see also Francia 1996a, especially 14950 and Zeilfelder 2001: 2539. It is an
archaic feature seen in other old PIE languages. It is unclear whether there is a functional
10. Compare the paronomastic innitive constructions of Akkadian (Ungnad 1992: 109d) and the
schma etymologicum of Biblical Hebrew (Gesenius, Kautsch, and Cowley 1910: 117 p and 113 lx).
11. ta-pu-a was accidentally left out of the hand copy in KUB 17.7; see Otten 1950: 13 n. 2.
16.25 Noun Cases
00-HittiteGrammar.indb 248 6/5/08 4:18:00 PM
249
distinction between the accusative of direction (i.e., of place-to-which) and the allative,
or merely a replacement of the former by the latter (Neu 1980).
16.28. Distinct from the accusative of direction is the accusative of the way: laan
kuwatta rnr.xr-u paizzi ne appa t sro
5
-in ukanta on whatever expedition the
troops went (actually, both verbs are present tense) they returned unsuccessful KBo
3.1+ ii 34; taknaat
d
tft-a xsxi-an paiddu let it (i.e., the evil) go the way
of the Sungoddess of the Netherworld! KUB 17.10 iv 13; nukan +it parkuin ku-
wapi rorn-an ar uwai when you come up behind the bare(?) high ground KUB
19.37 ii 3031 (AM); mankan
UR.SAG
Teinan [ar] pun I would have ascended Mt.
Teina KUB 19.37 iii 49. Francia (1997: 13940) lists these as accusatives of exten-
sion in space. Her additional examples we would also consider accusative of the way,
although the noun in the accusative is not pala- (= xsxi) way.
16.29. The accusative of extent, dimension, or duration is seen in the following
examples. Of time: nuza tu-an 2si 3si t eat twice (or) thrice in the course of a
day! KUB 1.16 iii 30; nu ipandan mandan iyaat And I marched throughout the
night KBo 5.8 iii 21 (AM 15859); [(peran par)]yazzi apn o[(r
6
-a)]n ts:i xtts-
:t [(tea)] And beforehand ( peran par) throughout that night (accusative of extent)
(the king) abstained from (intercourse with) a woman KBo 4.2 iii 5859, ed. Lebrun
1985 (obv. 1920); nu kar xt.10.xx itoi-ueznanun I had already reigned for ten
years KBo 3.4 iii 62 (AM). The accusative tu.xx.nr. in KBo 4.4 iii 62, cited by
Francia as accusative of duration, could be understood rather as during the daytime
(see 16.95, p. 266), especially since the expression lacks a number. Extent of space:
t\+ tu.1.xxwara ++++oi pargawkad [(d)] aru t\+ rft.1.xxmawara rxt-
an pargaw [(kadd )]ari In one day may (Ullikummi) grow a cubit, in one month may
he grow an rxt KUB 33.98 iii 1516 (Ullik. I); see CHD nai- 15. In the last-cited
example it is not the time that is in the accusative but the height (++++:i, rxt-an).
This last example refutes the claim by Francia (1997: 145) thatunlike Greek and
LatinHittite does not use the accusative to express extension in space as a measure.
For the dative-locative occasionally designating extent of time see 16.75 (p. 261).
16.30. The plural nominative-accusative neuter of adjectives is regularly used adver-
bially (see 19.13, p. 292). Also adduced as adverbial accusatives: antezzi in the rst
place, rst of all, lammar immediately, and karuwariwar in the morning.
16.31. The construction with the experiencer as a grammatical accusative object of
an impersonal verb is found with verbs of sickness (itark-, irmaliya-) and emotional
states such as fear (na-; see 16.32). Since none of the secure examples are from
OS or MS copies, we cannot claim its existence earlier than the New Hittite period.
Some of the better examples use dative-accusative clitic pronouns, which are ambigu-
ous: mnmu itarkzi kuwapi whenever I became ill (historical present; 22.622.7,
pp. 307308) att. i 44; tuk itarkta you became ill (lit., it ill-ed you). The follow-
ing example with a transitive medio-passive shows an unambiguous accusative object:
Noun Cases 16.31
00-HittiteGrammar.indb 249 6/5/08 4:18:03 PM
250
[(nu)]
m
itoi-
d
st\-u-un [r:]+ [ . . . itar (kiyattat na)] n. And arri-kuu,
my brother, became ill, and he died KBo 4.4 i 56 with dupl. KUB 14.29 + KUB
19.3 i 31 (both NH). We see the origin of the impersonal verb itark- in the following
NH example, where the grammatical subject (illness in the nominative) is expressed:
nu
GI
kalmina paitpat [nu
URU
A]paan s+
m
Ua-i tnt-an oti-ata
m
a-i-
inna oti-ata nan idlu oro-a itarkta And a comet(?) went and struck Apaa,
Uazitis city; it struck Uaziti, and a serious (lit., bad) illness befell him KUB
14.15 ii 46 (NH), see ibid. ii 13.
16.32. We also interpret -ta in lta ni fear not KUB 30.36 ii 8 (OHlNS) and in
lwatta ni KUB 33.24 i 43 (OHlNS) as accusative and ni as an impersonal pres-
ent third-person singular instead of imperative second-person singular. We reject the lat-
ter analysis because singular dative clitics are nowhere else attested in reexive function
in Hittite (see 28.18, p. 358), and because the l + imperative construction is quite rare
(see 26.17, p. 344, and CHD l). Compare also: nuwaan mn pimi n[(aimu)]
par(a)ni tn.nn.n-ni (Above, on the ankuri- (are) a . (and) a lamb;) when I go up
there (-an), I fear (lit., it fears me) with respect to the leopard (and) the wolf KBo
21.90 rev. 5152 (fest. of Teteabi, OHlMS), with dupl. KBo 21.103 rev. 5152.
12
For
the impersonal construction compare German mich frchtet, Latin m vertur.
16.33. Note, however, that in NH both the verbs of sickness and fear can also be
construed with a personal subject: (Arnuwanda, son of uppiluliuma, sat down on the
throne of his father,) rorn-anmaa irmaliyattatpat but later he became ill as well
KBo 3.4 i 56 (AM 1415). For examples of na()- see CHD LN 33841.
Ergative
16.34. For the use of the ergative case with neuter nouns see 3.8 (p. 66).
Genitive
16.35. The Hittite genitive is for the most part an adnominal case: a noun in the geni-
tive modies another noun in various ways. Evidence for an adverbal genitive is very
sparse and uncertain (16.57, p. 255). The genitive is also used in OH with postpositions
(16.5816.59, p. 255), and the genitive may appear in a free-standing construction
(16.61, p. 256). On the Hittite genitive see Otten and Souek 1969: 6062; Carruba
197879; D. Yoshida 1987.
16.36. The Hittite adnominal genitive has the following uses: (1) to indicate the pos-
sessor (possessive genitive); (2) to indicate the whole of which the head noun denotes
a part (partitive genitive); (3) to indicate the material, contents, or container of the head
noun; (4) to indicate the actor (subjective genitive) or acted upon (objective genitive)
12. On this passage see CHD sub a- A.
16.32 Noun Cases
00-HittiteGrammar.indb 250 6/5/08 4:18:04 PM
251
when the head noun denotes an action; (5) to indicate the length, width, capacity, or
weight of the head noun (genitive of measure); and (6) to indicate the purpose for which
the head noun is intended.
16.37. The most common genitive use is the possessive genitive (in OH, see D. Yo-
shida 1987: 1246): kuela ari aki on what (person)s property he dies KBo 6.2 i 7
(Laws 6); orxr-a aruwanduu the slave womans unborn child KBo 6.3 i 43
(Laws 18); attaaa r-ri aki and (if) she dies in her fathers house KBo 6.3 ii 3
(Laws 27); ammelpatwaza otu-un dai (I have decided that) I will take my own
ox KBo 6.3 iii 71 (Laws 74);
URU
Nea itoi-u Neas king KBo 3.22:4 (OS);
attaa r-ri in the house (dative-locative) of his father (lit., of father, of his) KUB
33.24 (+) 33.28 ii 12; itoi-an ka to the gate (allative) of the kings (archaic genitive
in -an) KUB 13.4 ii 49.
16.38. Often in Old (and rarely as an archaism in Middle) Hittite the dependent geni-
tive is doubled by a clitic possessive pronoun on the head noun (called by Otten and
Souek 1969: 70 pleonastischer Gebrauch). Garrett (1998), who calls it a split geni-
tive, has shown that this construction is normally restricted to inalienable possession
(the head noun, which is the thing possessed, is a body part, etc.): i.t
19
.it-a itt++-
a xrn
14
et KBo 6.2 i 24 (Laws 13, OS); orxr-a aruwan[du]u KBo 6.3 i 43
(Laws 18 (OHlNS); kl me-n[e!]-i-i-it this ones face KBo 6.26 i 36 (Laws 166
OHlNS
13
); [a]mmella lm n mit KUB 1.16 iii 13 (OHlNS); labarnaa xt.nr.
e the Labarnas years KBo 17.22 iii 9 (OS); ammel tuggami my members
VBoT 58 i 24 (OHlNS). The NS examples are mere recopying from the OH archetype,
since the clitic possessive was moribund in NH. There is one assured MHlMS example:
umenzanpat keremet your (pl.) hearts KBo 8.35 ii 21; and another very likely
one in a MH composition containing OH archaisms:
14
[(weawa) -a xr-a(a)
zio.o]-a( ) mi (var. la-lu-uk-ki-mi-i) we are the illumination (var. luminaries
[pl.]) of heaven and earth KUB 33.91:8 restored from dupl. KUB 36.19 iv? 1314 (see
CHD lalukkima- 1). The scribe of the duplicate has made the singular zio.o-a
into a plural because of the two genitive nouns and has modernized the construction by
deleting the doubling clitic -mi. It is unclear whether the Akkadogram clitic posses-
sive attests this usage in other MHlMS examples such as: tuel . . . rx+ HKM 63:12
(MHlMS).
16.39. In some cases, however, even as early as OS, this construction seems not
to express inalienable possession:

mn utxt-a
d
rx-na aantii uekzi when the
prince recites an incantation to the concubine of the Stormgod KUB 8.41 ii 7 (OS);
13. Preserved only in NH copy KBo 6.26. In OS KUB 29.30 iii 23 it is usually restored, but there is
unfortunately no way to prove its presence there.
14. Such as the occurrences of weawa (without -za) alongside the NS scribes modernizations with
weawaz in nominal sentences.
Noun Cases 16.39
00-HittiteGrammar.indb 251 6/5/08 4:18:05 PM
252
itoi-a xtts.itoi-a . . . paprtaramet their (i.e.) the kings and queens
impurity StBoT 8 ii 1011.
16.40. In the same way in Old Hittite (OS) the doubling possessive can be suxed
to a primitive noun-become-adverb (for ketkar, see 3.26, p. 74; 3.32, p. 77; 7.19,
p. 147; for er, see 16.5816.59, p. 255; 20.26, p. 300): itoi-a xtts.itoi-
a ketkaramet at the heads of the king and queen KBo 17.1 + KBo 25.3 ii 2728
(OS); itoi-a xtts.itoi-aa [(remet)] overlabove the king and queen
ibid. ii 20.
16.41. The partitive genitive (genitive of the whole; D. Yoshida 1987: 8890)
is weakly attested in OH by the following: s+
LU
tntut.on t kuiki arauwa no
one of the copperworker(s) is exempt KBo 6.6 i 31 (Laws 56, OHlNS) and a few
other examples adduced by Yoshida, which are somewhat doubtful. One of these, also
proposed by Neu (1974: 46, 56): ardia(n)anna kuin uwatet of the Anitta text (Neu:
wen von seinen Helfern [. is plural genitive]), is also doubted by Houwink ten Cate
(1984: 79 n. 51). This clause could be interpreted as his auxiliary troop (acc.) which
he brought, taking ardian as singular accusative. In most clear cases of the partitive
genitive in Old Hittite, the genitive seems to depend on the indenite pronoun kuiki as
the head pronoun. As a partitive genitive in the NH period Friedrich (HE 1 211) cites:
nuza mandapat rorn-ezzi utxt-a eun I was the last child of all att. i 11.
This example, however, like other forms in -a in the superlative constructions (17.17,
p. 275), may be taken as either genitive or dative-locative plural. The example tnt-a
pupulli the ruins (neut. nom.-acc.) of a city
15
could be an instance of a possessive
genitive, but the city does not really possess the ruins; rather, the ruins are all that is
left (i.e., a part) of the city. Further examples that express superlatives can be found in
17.17 (p. 275).
16.42. When the relationship of possessor to possessed is also one of whole to part
(as with the human body), possession can instead be construed as a partitive relation-
ship. In such cases an OH possessive genitive may be replaced in post-OH by partitive
apposition (the so-called , 16.24, p. 247): takku i.t
19
.it-a itt++-a +ssi
nama olnsi kuiki tuwa [rnizzi] If anyone breaks a free persons arm or leg KBo 6.2
i 20 (Laws 11, OS). Here the OS copy KBo 6.2+ (A) uses the genitive, while the NH
copies B and C replace it with partitive apposition (16.10, pp. 243244) in the accusa-
tive:
16
takku i.t
19
.it-an itti+ +ssi nama olnsi kuiki tuwarnizzi. This partitive
apposition replaced the older split genitive (16.38, p. 251) and is consequently only
used with inalienable possession (head nouns are body parts, etc.; Garrett 1998).
15. (Sum.) gu- bal = (Akk.) te-lu tell, mound of ruins = (Hitt.) tnt-a p[u-pu-ul-li ] r [uins] of a city
KBo 1.42 iii 6, ed. Honer 1967b 303 with n. 23; see CHD P s.v. pupulli-.
16. See HE 12324 213; Kammenhuber 1969b: 2034; D. Yoshida 1987: 3435; Neu 1997: 14243;
LH 2425, 176 on 9).
16.40 Noun Cases
00-HittiteGrammar.indb 252 6/5/08 4:18:09 PM
253
16.43. The genitive of material (made) of (D. Yoshida 1987: 7076) is attested in:
uto.orn
4
-a oi-in goblet (made) of red clay KUB 13.3 ii 26; 1
GI
unila
GI
laura
ftn one small laura- table (made) of unila- wood VBoT 58 iv 18 (OHlNS); rx-[a]
upuwi .-vessel (made) of clay KBo 39.8 ii 55 (MH?lMS); [(eppitta)]
NINDA
danna
danna-loaves (made) of eppit-cereal KUB 9.16 i 1920 (NS) with dupl. KUB 10.48
i 3; and a nal example: .fr-amai
4
-a
NA
4
ku[nk]unuzziya ui-anza its body
(was) made of k.-stone KUB 33.95 + KUB 36.7b + KUB 33.93 iv 12 (Ullik., NS). For
more examples in OH see Otten, StBoT 8 61.
16.44. The genitive of contents (head noun is the container, genitive is the contents;
D. Yoshida 1987: 8387) is exemplied in: nu man mekki andait orfr-a
DUG
pali
marnuwanda
DUG
pali [w]aliya
DUG
pali (The goddess Inara) prepared everything
on a grand scale: large vessels of wine, large vessels of marnuwan-beer, large ves-
sels of wali-beer KBo 3.7 i 1517 (OHlNS); zz-a
DUG
ariyalliya peran in front
of the storage vessel of wheat KBo 15.36 + KBo 21.61 ii 13 (OHlMS); 1
DUG
uirx
oi
TU
7
ampukkiya one large bowl of ampukki-stew KBo 5.1 iii 1011; z.ntx
xi.nnnn s+ +I one silver z.ntx-vessel of water KUB 39.70 i 17; alkiyai
DUG
ariyalli for him a storage vessel of grainlbarley KUB 21.17 ii 12 (NH). Another
pattern in Hittite for indicating a (vessel) of (liquid content) is not with a genitival
construction but with case concord: 12 uto marnuan KUB 53.14 i 6, 11 (OHlMS?);
1 uppar marnuan i 27. This may be purely graphic, another example of the Akkado-
graphic writing of a Hittite noun. Note that, unlike the genitival construction, here the
noun of content always follows the head noun, as in Akkadian syntax. If the construction
is real, one may compare German uses such as eine Flasche Milch = a bottle of milk
(T. van den Hout, pers. comm.).
16.45. The opposite of the genitive of contents is the genitive of the container: s+

garupaia . . . [(ez)]zan or-ru the cha (and) wood of (i.e., contained in) the gra-
nary att. iv 8283.
16.46. When the head noun denotes an action, the genitive can indicate the subject
or object of the action (so-called subjective and objective genitive). Subjective: kmu
kuit s+
m
Marrwa i
URU
immuwa aliyatar atre This homage of Marruwa, the
ruler of immuwa, about which you wrote to me HKM 13:45 (MHlMS). Objective:
xtts-a duar the taking of a woman KBo 1.53:10, s+ xtn
URU
Amka oti-auwar
the attacking of the land of Amqa KBo 5.6 iii 56 (D frag. 28).
16.47. In the genitive of measure the units of weight or measure always follow
their head nouns: 4
NINDA
arau tarna four thick breads of (i.e., weighing each) a
tarna i.e., not 4 tarna
NINDA
arau or the like. Perhaps it was the presence of the
numeral 4 preceding the
NINDA
arau which caused the genitive tarna to move to
a position following the governing noun, to remove the ambiguity (is it 4 loaves each
one tarna or an indeterminate number of loaves each of 4 tarna ?). Also dependent on
kuia each: nu 9 ru.otn
4
.n.nr. anda neyandu tarna 9 ru.otn
4
.n.nr. l
Noun Cases 16.47
00-HittiteGrammar.indb 253 6/5/08 4:18:10 PM
254
tarna 3 ru n kuia tarna udanzi And they bring (in) nine thick loaves braided
together (each weighing) a tarna, nine oiled thick loaves (each weighing) a tarna, and
three warm loaves (each weighing) a tarna KUB 17.23 i 24, ru.iXiyaan
mandu kuia par tarna kittari KUB 32.128 i 56; 3 ru-an azzila KUB 7.53
i 22;
NINDA
iduri zu.u azzila KBo 15.37 i 8; 1
DUG
uirx.oi ft
7
emeuna azzila
KBo 16.49 iv 6. No example of a syllabically written unit of measure occurs in the plural
with a numeral (e.g., *2 tarna, *2 azziliya), but logographic examples occur: s+ 2
ir\t KUB 13.4 i 6; KBo 21.1 i 8; s+ 3 ir\t ibid. i 9; s+ 3 s':t KUB 30.15+ obv. 26.
16.48. The genitive sometimes indicates the person or object for which the head
noun is intended: t.xr-a wtar water (neut. nom.-acc.) of (i.e., for) the hands,
genuwa ou-an cloth (accusative) of (i.e., for) the knees; tayezzila arnikzel com-
pensation of (i.e., for) theft KUB 13.9 ii 89 (MHlNS); txutut
I.A
s+
d
ts[:+u] votive
oerings of (i.e., for)
d
ts:+u KBo 16.83 ii 12 (NH); s+
d
r.oi ssxtn the ritual of
(i.e., for) r.oi KUB 45.47 iv 36.
16.49. The use of the genitive ablativally in the phrase or
6
-ya xr-a xr.nu-a
(i.e., tankuwaya daganzipa tuwala) remote from the Dark Earth KUB 33.106 iii
3839 is isolated (see 30.42, p. 418). See also [m]n r-er ana . . . kurkurima
linkiya parkunuwa[nzi ] When(ever) [they] purify a house oflfrom bloodshed, . . . cal-
umny, (and) perjury KUB 7.41 i 12, where one might expect ablatives (cf. KBo 10.45
iv 3941 [the same ritual] and VBoT 111 iii 1718).
Word Order in a Genitival Phrase
16.50. In the normal word order the genitive precedes its head noun (see Sommer
and Falkenstein 1938:

4748), except for the genitive of measure (on which see 16.47,
p. 253).
16.51. If a demonstrative adjective modies a head noun governing a genitive, the
word order is urxosfnfrvr + orrfrvr + nru: nu ini s+
GI
t\ut
I.A
r [i] That im-
age (made) of fruits KUB 39.11:40 (NH); k tandukena utxt-a this mortal (lit.,
this child of mortality) KUB 7.5 i 8.
16.52. If an attributive adjective modies a head noun governing a genitive, the se-
quence is ffnrntfrvr + orrfrvr + nru: andnza maniya[]aya i zi[k] You
are a just lord of rule KUB 31.127 i 20 (prayer).
16.53. Certain conditions describe the less common cases in which the sequence
nru + orrfrvr is found:
16.54. Principally, it is the presence of a Sumerogram (or Akkadogram) in the con-
struction which causes the reversal of the sequence: urorn.xr r.xr nepia xr-
aa the gods, lords of heaven and of earth KUB 6.45+ i 36; .nn or
6
nepia
black iron of (i.e., from) the sky (= meteorite iron) KBo 4.1 i 39.
16.48 Noun Cases
00-HittiteGrammar.indb 254 6/5/08 4:18:13 PM
255
16.55. When the genitive is the indenite pronoun kuelka, it normally follows the
head noun: uppalaet kulqa someones animals KBo 6.26 i 22 (Laws 163, OHl
NS); parnama kulka to someones house KBo 6.2 ii 35 (Laws 44b, OS); takku
otu-a .X.nr.-ni kuelqa aki If an ox dies on someones eld KBo 6.3 iii 68 (Laws
72, OHlNS); wadul kulqa[ au]tti you see someones sin KUB 1.16 iii 5960. This
pattern persists even when the kuelka is negated: [(ln-na)]nata orxr-an natta kuelka
daun I took no ones [male slave] (or) female slave KUB 31.4 + KBo 3.41 6 restored
from KBo 13.78 obv. 5. For further examples see the following paragraphs.
16.56. The use of the genitive kuelqa as an indenite pronoun described in 16.65
is not to be confused with kuelka as an indenite adjective modifying a noun in the
genitive, for which see 18.3518.36 (p. 287).
16.57. One uncertain example of an adverbal partitive genitive is nata
DUG
palan
mandan ek[uer] They (Illuyanka and sons) drank (some) from (lit., of) all the ba-
sins KUB 17.5 i 11 (OHlMS) (thus Melchert 1989: 182). The lack of any other attested
examples leaves the reality of this usage in doubt. This usage shares features of the rare
partitive ablative (see 16.102, p. 267).
16.58. Certain postpositions governed the genitive in Old Hittite but the dative-
locative in New Hittite (see 20.15, p. 298, and following for examples): attama
ppan after (the death of) my father KBo 3.22:30 (OS), itoi-wa peran in front
of the king KBo 20.12 i 5 (OS), annama katta with my mother KBo 3.40:14,
npia kattan under the sky KUB 29.3 i 11 (OS); ammel katta with me or (with
verbs of sending) to me, to where I am KUB 36.44 iv 4 (OHlMS), a katta next
to the brazier (KBo 17.15 rev.! 15 (OS) (for postpositions with enclitic possessives see
20.26, p. 300).
16.59. In New Hittite one postposition continues to govern the genitive case, iwar
in the manner of, like (see the dictionaries; Honer 1993; and 19.9, p. 291). Ex-
amples: tn.or
7
-a iwar . . . aki he will die . . . . . . like a dog KBo 32.14 left edge
6 (MS); i-a iwar like a man KUB 31.69 obv. 5; r.xr-a iwar like brothers
KBo 10.12 iii 10; i.xn-a iwar like an enemy KBo 5.4 rev. 15; s+
m
Mauiluwa
iwar like Mauiluwa KUB 6.41 iii 55 (Mur. II); itoi-uzna iwar KBo 16.17 +
KBo 2.5 iii 3031 = s+ itoi-i::t iwar in the manner of royalty, like a king KBo
3.4 iii 75 (AM 8889); s+
LU
t.urn iwar like a captive KBo 3.6 iii 62 (= att. iv 31
with variant
LU
tt);
URU
attua iwar in the Hittite manner KUB 21.17 iii 7; anzel iwar
like us KUB 17.21 i 17; nu ammel iwar kuwatqa iya[i] Perhaps you will act like me
KUB 60.136:4 compared with nu ziqqa kuatqa s+
m
Maturi iwar iyai Perhaps you
too will act like
m
Maturi KUB 23.1 ii 2930 (Tud. IV); +\+ xtn i.xn-x+ zakkiya
iwar tiya[zi] he will enter (literally, step to) the land hostile to you like a key (enters a
lock) KUB 29.11 ii 4 (translation of Akkadian lunar omen); nuwa uga n-a iwar
uiyami and I will squeal like a pig KUB 14.1 rev. 93 (MHlMS); nuwakan apl zr-an
Noun Cases 16.59
00-HittiteGrammar.indb 255 6/5/08 4:18:15 PM
256
ororn.xr: uwitena iwar ara luwaten pour out his soul, O gods, like water KUB
13.3 iii 12 (ins.).
16.60. In later NH iwar is occasionally written (not spoken!) as a preposition before
logograms or proper names written as logograms (without case endings; Honer 1993):
k iwar itoi
KUR
Karg[ami] I, like the King of Carchemish KBo 18.48 obv. 19 (let-
ter); iwar
m
Manini like Manini KUB 5.1 i 43 (see ibid. 87, 90, 93); iwar xtts :t.or
like an Old Woman IBoT 1.33:84; iwar
URU
Arzawa in the manner of Arzawa KUB
18.67 obv.? 12; iwar
LU
:+iitsi like his colleague KUB 8.48 i 1516 (Gilg.). In this
pseudo-Akkadographic construction it simply imitates the word order of the Akkadian
preposition kma like.
16.61. The independent (or free-standing) genitive he of . . . (HE 212, D. Yo-
shida 1987: 111), or genitivus absolutus (Carruba 1966a: 4142), is attested from Old
Hittite on. Examples: watara he of the horn; horn-player KBo 17.1+ i 37 (OS);
LU
arka(m)miya (=
LU
arka(m)miyala) arka(m)mi-player (see Otten and Souek 1969:
6162); watula he of the sin; the oender Laws 14748; tuk . . . linkiyaa it
he made you . . . (a person) of his oath KUB 14.1 obv. 13 (direct obj.); tayazila he
of the theft; the thief Laws 73, the (compensation) for theft Laws 9495 (direct
obj.); kardiata that of your heart, your wish KBo 3.7 i 26 (direct obj.); auwa
memiyana (he) of good words = he who is on good terms att. iv 50; fr-anna
(= uiwanna) he of (long) life, long-lived KBo 3.6 i 13 (= att. i 15) (NH); with the
genitive of the verbal substantive (see 25.5, p. 330): nawa he of being reverent
KBo 3.6 iv 15 (NH); parnauwa the (cupbearer) of squatting KUB 10.89 i 19 and
passim; ar puwa (a person) of going up (into the temple) KUB 13.4 ii 12; and
of innitives ara tarnumma he who is to be released (from military service) KUB
13.20 i 11; kuitma or-ar umel t tar(a)wa what legal case is of your not-being-
able = whatever case you cannot settle KUB 13.20 i 36. Note that the genitive of the
verbal noun may have the force of he who (or that which) can (or must) be . . . -ed, as
illustrated in the last two examples. See also Yakubovich 2006.
16.62. Free-standing genitive independent pronouns also exist: apl his (water)
Laws 162; anzel our (people) KUB 9.31 iii 47; ammel dai I will take (what is)
mine KUB 13.35 ii 3. See Sommer and Falkenstein 1938: 9293 with notes.
16.63. The same construction is attested in Akkadian texts (Ungnad 1992: 30 and
114), and is attested also by Akkadograms in Hittite texts: s+ +++i:t he of the oath,
the liegeman (see linkiyaa above); s+ xsxi-\t+ that of the road = provisions
for the trip KUB 9.31 iii 13.
16.64. Related to the free-standing genitive is the usage in the expression itoi-
u otn-a = aranda (or ft:-a = aanda) ekuzi the kingin a standing (or sitting)
positiondrinks . . . which occurs frequently in the festival texts. This often has been
discussed, for instance by Holland (1986: 165 note). If otn-a stood for the genitive of
the verbal substantive (arawa or aawa), it would be easier to understand: he of the
16.60 Noun Cases
257
action of standinglsitting. But since it is the genitive of the participle, it is problematic:
he of the one standinglsitting.
Dative-Locative
16.65. The dative and locative cases have merged in Hittite, in both form and func-
tion. There is no basis for the claim of Starke (1977: 6368) of dierent functions ac-
cording to an alleged distinction between nouns referring to persons and those referring
to things (3.7, p. 66). The functional merger of the two cases is shown: (1) by use of the
merged dative-locative to express place from which with inanimate referents (16.69,
p. 259); (2) use of the dative-locative to express place to which with inanimate ref-
erents already in OH (in competition with the allative; 16.70, p. 259); (3) use of the
dative-locative to express location with persons in combination with postpositions, also
already in OH (16.72, p. 260); (4) use of the dative-locative singular of the enclitic
personal pronouns to refer to inanimate objects (16.71, p. 260). The third usage is still
rare in OH in comparison with the genitive (see 16.35, p. 250), but it does exist, and
the last usage is exceptional, but the rst two are not, despite claims to the contrary.
For obvious pragmatic reasons, nouns referring to persons are more frequent in some
uses, and nouns referring to things in others, but there is no grammatical contrast. We
group the uses in what follows into those related to the dative and those related to the
locative purely for expository purposes, to make clearer the relationship to uses in other
languages.
Indirect Object
16.66. The use of the dative-locative for the indirect object, answering the question
to whom, derives from the dative case. This usage occurs with nouns, independent
and enclitic pronouns, and with nouns represented by logograms. In the last case, when
phonetic complements are lacking, one can only detect the dative-locative case by the
employment of the Akkadogram preposition - to, for. Examples: (nouns) takku
otxt.xtts i-ni taranza if a young woman is promised to a man KBo 6.3 ii 5
(Laws 28, OHlNS); takku otxt.xtts-a i-ni amenkanza nui kata piddaizzi
if a girl is betrothed to a man, and he gives the bride price for her KBo 6.3 ii 11 (Laws
29, OHlNS); [ka] otxtmi Labarni r-er pe[un] I have [hereby] given a house
to my son Labarna KUB 1.16 ii 3031 (OHlNS); ta itoi-i kardimiyattu pikatteni
and you are giving to the king reasons to be angry KBo 22.1: 20 (OS); (independent
pronouns) apedaa tamain or:fr-an piyr But to them they gave a dierent wine
KBo 3.34 ii 34; (enclitic pronouns) nu antezziya i-a kuit kuit p[eta] taa(t) e
arnikzi He shall make compensation to the rst man of whatever [he paid] KBo 6.3
ii 67 (Laws 28, OHlNS); (logograms) nu ororn- :t.xr:-a wtar par pzi
He holds out to the deity water for the hands KUB 39.70 i 19. Rarely, the ideographi-
cally written noun, which appears to be a dative, has no to introduce it: nue 6 o
xi.nnnn pi
LU
.ztya kuan appat pi he will give to him six shekels of silver
and to the physician that same (man) will pay the fee KBo 6.2 i 19 (Laws 10, OS).
Noun Cases 16.66
258
Indicating Possession
16.67. The use of the dative-locative to indicate possession is attested not only in
predicates of clauses with the verb to be[k(zzama ma)an (+\+ r:+) . . . ( . . .
uiuwali)] utxt-a [(t) (ta)] Now at this time (kzza) since there was not a son
entitled to rule belonging to my brother (i.e., since . . . my brother had no son) att. iii
4041 (on which see Koak 1996), but even non-predicatively without that verb.
17
The
co-occurrence of a dative clitic pronoun and a clitic possessive pronoun in MS or NS
copies of OH texts may reect a real usage or merely a faulty updating in which the new
use of the dative clitic was added while the old clitic possessive was retained: umu
urorn.uruir utxt
URU
Pur[uanda] kirimi daier The gods put the Puruandaean in
my hand (lit., to me in the hand; see German mir in die Hand legen) KBo 3.28:67
(OHlNS); [uka +\+] urorn:+ kuit iyanun numu r:+ inani peran pittuliya r-er
kiat But what have I done to my god, that because of illness my house has become
a house of fear? KUB 30.10 rev. 1314 (OHlMS). A dative of possession using the
relative pronoun kui- is found in the following: paltanu kuedani awan katta kiyantari
one whose shoulders are hanging down idly (approximate and clumsy Hittite attempt
to translate Akkadian au nad he is idle, lit., (his) hand(s) hang down) KBo 1.42 ii
32 (NH). For the possessive genitive see above, 16.37 (p. 251).
Dative of Disadvantage
16.68. The so-called dative of disadvantage is reected in the use of the dative-
locative translatable as from with verbs such as (-kan ara) da- to take (from), ( peran
ara) munni- to hide (something from), anna- to hide (something from), and tu-
to separate (something from), wek- to ask (something from someone). In all the fol-
lowing examples the dative-locative represents a person, or at least a living being (deity,
animal, etc.), and in the vast majority of occurrences a local particle, either -an or
-kan, occurs (see 28.76, p. 371). With local particle: [(k)]atamakan utniyandan
llu dun [(erm)]a(n)makan dun I have just taken the slanders of the inhab-
itants from you. I have taken the illness from you KBo 17.1 i 1112 (OS) (see ibid. iv
38 with -ata: [(du(n)mata an wn [pittulium(u)]a I have taken from
you the woe, pain, and your anxieties); tmakan dn kuedaniki kuitki armi Nor
have I taken anything from anyone HKM 68:56 (MHlMS); nuian
GI
uean
GI
ulaliya ara dai nui
GI
n . . . pei I take the spindle and dista from him
(-i) and give him (-i) a bow KUB 9.27 + KUB 7.8 i 2325 (rit.);
18
apelmaikan
17. As per Gterbock (1983), the non-predicatival use of the dative clitic pronouns to mark possession
is not attested in OS. This usage does appear in MS and NS copies of OH compositions, as shown by the
examples cited below. It is unclear whether these examples reect OH grammar or are merely the result of
modernization by copyists.
18. Here the disadvantageladvantage polarity is particularly sharp in the opposition of the two verbs,
with -an accompanying (ara) d- to take away and no particle accompanying pai- to give. See also
BrTabl. iv 1819 (Tud. IV) in 28.76 (p. 371).
16.67 Noun Cases
00-HittiteGrammar.indb 258 6/5/08 4:18:20 PM
259
ara di I take his away from him KBo 17.61 obv. 21; namaankan +\+
d
tft-
stma munni or if you hide (munna-) him from (+\+) His Majesty KBo 5.4 obv.
9 (NH; see CHD munnai- b); (if someone reveals a plot to you,) zikmaankan +\+
d
tft-st annatti and you conceal (anna-) it (an) from (+\+) His Majesty KUB 21.1
iii 2628 (NH); (if a girl is betrothed to a man, and he brings the bride price for her, but
after wards the parents contest it,) nankan i-ni tuanzi katama 2si arninkanzi
they may separate (i.e., withhold) her from the man (i-ni), but they must pay back his
bride price twofold KBo 6.3 ii 1213 (Laws 29, OHlNS); see nanikan tuanta
(var. tuanzi) they may withhold her from him (-i) KBo 6.3 ii 10 (Laws 28, OHlNS);
(if you hear evil concerning My Majesty,) natmukan mn annatteni and conceal it
from me KBo 5.3 i 28 (MHlNS); without particle: (the gods were angry at Kei) nui
itar []man peran ara munner and hid all the game from him KUB 33.121 ii
1314 (NH) (see CHD munnai- a 2); numu . . .
d
ts:+u
URU
amua +\+ +ui:+ wkta
Is:+u of amua requested me from my father KBo 6.29 i 78; see Gterbock 1943:
154 and HE 207e. Possibly also the following (with verb peiya- mng. 8 to cause to
fall), if we adopt the translation favored by CHD P: takku tut-un tn.nn.n-ni kuiki
peiezzi If someone causes a sheep to fall from (the mouth of) a wolf (i.e., snatches it
away) KBo 6.2 iv 14 (Laws 80, OS), see CHD P 322, with literature cited there. But
much is still uncertain about this passage from the laws. The dative of disadvantage is
also seen in clauses with the verb akk- to die and the dative of persons most aected
thereby (see 28.76, p. 371).
16.69. This usage to express that from which something is removed is extended
in Hittite to include things as well as persons, already in OSwith local particle:
[(irm)]a(n)makan dun kardimiyaatkan du[n (ara)]nimiyaatkan
dun I have taken the sickness from you. I have both taken it from your heart(s) and
taken it from your head(s) KBo 17.1 i 1213 (OS); in the later language: (they cook
a piglet and bring it back;) nukan kuwapitta appeni tepu dai and I take a little
bit from each body part (and sacrice it to the Sungoddess of the Earth) KUB 17.28 i
1718; without particle: nan itoi-eznanni ara tittanut (When my father attuili
(had) made war against Uriteub, son of Muwattalli,) he removed him from kingship
BrTabl. i 8. For further examples of this usage see Melchert 1977: 18889, 289, 310,
and 352. Not all of the examples listed there, however, are valid, in particular ##46 and
47 on p. 189.
19
Goal
16.70. The dative-locative also is regularly used to express place to which (goal)
with inanimate objects already in OS: itoi-u ppari ipanti The king libates into
a bowl KBo 25.61 rev. 9 (thus with Klinger 1996: 440 against Starke 1977: 5758);
aliea ar (a)pta strays intolassociates with (another) pen KBo 6.2 iii 48 (Laws 66,
19. In example #46 genuwa is genitive, not dative-locative, and in #47 the context shows rather that
carries it (i.e., the divine image) out to the tent is meant. See lines 1521 in the following context.
Noun Cases 16.70
00-HittiteGrammar.indb 259 6/5/08 4:18:22 PM
260
OS) (with Friedrich 1959: 39 and LH 77 against Starke 1977: 51); takku otu.nr. .X-
ni pnzi if oxen go into (another mans) eld KBo 6.2 iv 12 (Laws 79, OS; with
Friedrich 1959: 43, Souek 1970: 274, and LH 84 against Starke 1977: 53; takku i-an
paueni kuiki peiezzi na aki if anyone makes a man fall into a re, so that he dies
KBo 6.2 ii 33 (Laws 44a, OS; with Friedrich 1959: 31 and LH 52 against Starke 1977:
56). These examples show, contra Starke (1977: 5166), that in OS the dative-locative
as well as the allative expressed place to which with inanimate objects. Compare .X-
ni pnzi they go to the eld with parna paimi I go home. We do not have enough
examples of each case with the same verb to determine whether there was any contrast
in meaning. Old Hittite speakers may have used the allative selectively with verbs of
perambulation or gross movement, while they used the dative-locative with verbs of
non-perambulatory and minor movements (see 16.82, p. 263).
Purpose or Result
16.71. In OH, where the dative-locative and allative are always written dierently,
we nd a clear example of the dative-locative of purpose or result: uwa
URU
attua
engani pun I went to attua to die (lit., for death) KBo 22.2 rev. 5 (see comments
in Francia 1996a: 139, 148); in the famous phrase from the laws (OS) we also nd a
rare use of the enclitic dative-locative pronoun used to express purpose with a thing:
parnaea uwayezzi and (the plaintilvictim) shall look to the (accuseds) house
for it (i.e, the compensation) (Gterbock 1983; LH 16869 with refs.; also see here
30.17, p. 411). Post-OS examples: nu panku par ingani arta and the assembly
held them for the death penalty KBo 3.1 ii 28 (OHlNS); numu +\+ urorn-tt+ ln-anni
peta and he (my father) gave me to the deity for servitude att. i 18; [orx-anmaat
URU
tu.xrn.t].xr erer natz[(a
d
tft-i! ur-eni tiyr)] [But when] they arrived at
[Sipp]ar, they stood before Shamash for judgment ABoT 48:4 (OHlNS) with dupl.
KBo 19.108:3 (NS). The Akkadographic preposition is normally +ssi+ but occasion-
ally +\+: nui apel utxt.xttssi r:+
f
Muwattin +ssi+ ux-i::t+ pai And
he gave to him his own daughter, my sister Muwatti, in marriage KUB 6.44 + KUB
19.53 i 67 (NH), versus +\+ ux-i::tsi describing the same event in KUB 14.15 iv
41 (AM). That anneni KBo 6.2 ii 13 and kuani KUB 29.29+ ii 31 are examples of
the dative-locative of purpose seems unlikely ( pace Francia 1996a: 148).
Location
16.72. The dative-locative reects the locative case in answering the question where?
in what place? takku i.t
19
.it-a i-a nama xtts-za takiya tnt-ri aki kuela
ari aki 1 +i gipear .X kar (a)iyzzi nanza di If a personmale or femaleis
found dead in another village, the one in whose boundary (s)he dies, he shall cut o
100 gipear of eld, and he (the surviving heir) shall take it KBo 6.3 i 1415 (Laws
6, OHlNS); nu

alentuwa alli aear Then in the alentuwa-buildings a large as-

sembly (convenes) KUB 9.16 i 1516; genuwaa on his knees KBo 17.61 obv. 22;
16.71 Noun Cases
00-HittiteGrammar.indb 260 6/5/08 4:18:23 PM
261
tepauwanni in disgracelobscurity KBo 3.34 ii 12; apedani pedi in that place HKM
47:53 (MHlMS); nepii in the sky KBo 3.28:16; aruni in the sea KBo 3.4 ii 51, 52
(AM). Nouns in the dative-locative referring to persons are also used in a locative sense
with postpositions, already in OH. In OH this usage with certain postpositions ( peran,
er, etc.) competes with the genitive, which is more common (16.5816.59, p. 255);
in later Hittite one nds only the dative-locative. Example: utxt.r.oi itoi-i peran
uwi the palace ocial runs ahead of the king KBo 17.15 rev. 18 (OS).
16.73. The dative-locative is never used to express the agent with a passive (pace HE
207b). In the two examples zikzakan ammuqqa 1-edani x-ni aante Were
you and I begotten in one mother? KUB 23.102 i 1415, and iyl ln.xrsi [t\+ 1
x] aante The servants of Number One (i.e., the king?) have been begotten in
one mother KUB 1.16 ii 47 (OHlNS) the mother is viewed as the place of birth, not the
agent (contra Mora and Giorgieri 2004: 190: generati dalla stessa madre, citing Starke
1993: 23; 1997: 46465, 482). For the locative with the verb to give birth, beget see
Sanskrit jamaymsa Nalo Damayantym . . . Nala begat inlon Damayant . . . Hittite
shows variable syntax for aant- born, begotten plus the noun for mother: r.xr
d
tft-stya kui akuwaara( ) s+ xtts.itoi awan ox aante Also the
brothers of His Majesty who are begotten also of the legitimate queen KUB 21.42 iv
1617 (restored from KUB 40.24 rev. 5) with a genitive, but also r.xr
d
tft-stya
kui ts:i xtts.itoi aante KUB 26.12 i 1213. The last example with ts:i
could be either an ablative of origin (16.89, p. 264) or an instrumentallablative of agent
(16.107, p. 269, and 16.99, p. 267, respectively). See also 21.14 (p. 305).
Temporal Uses
16.74. The dative-locative is used for temporal expressions, usually simply express-
ing a period within which something occurred (Francia 1997: 14344): ipandi at
night; antaii meuni in the hot time (of the day); otu puugarinma kuedani tu-ti
unuer nuza
d
tft-st apedani tu-ti war (a) pta His Majesty bathed on the day on which
they adorned the ox of ritual substitution KBo 4.2 iii 56-57, with dupl. KUB 43.50
obv. 8 + KUB 15.36 obv. 10; gimmanti in winter HKM 17:10 (MHlMS). As in other
languages, Hittite location in time is expressed in terms of place (Honer 2002). Simple
dative-locative of nouns expressing units of time: witti in the year; iwatti on the day;
lamni in the hour. In combination with postpositions: -mu peran before me; -eli
ppan after him (i.e., his death); etc.
16.75. In other examples it is the extent or duration of time which is indicated by the
dative-locative (or its equivalent, Akk. t\+): [(nu)] t\+ xt.7.xx tauiyaa For seven
years I endured KUB 1.4 + KBo 3.6 iii 30 (= att. iii 62);
m
Kei rft.3.xx-a ntn.so.
xr-a anda weatta Kei wandered for three months in the mountains KUB 33.121
ii 15; t\+ tu.1.xxwara ++++oi pargawkad[(d )]aru t\+ rft.1.xxmawara
rxt-an pargaw[(kadd )]ari In one day may (Ullikummi) grow a cubit, in one month
Noun Cases 16.75
00-HittiteGrammar.indb 261 6/5/08 4:18:24 PM
262
may he grow an rxt KUB 33.98 iii 1516 (Ullik. I); nuza k arazena xtn.xtn.xr
i.xn t\+ xt.10.xx tar (a)un and I conquered these lands in ten years KBo 3.4 i
2829; 1-i\ xt-anti ta (The Queen of Kanesh) bore (children) in the course of one
particular year KBo 22.2 obv. 1 (OS). The notion of a particular year comes from the
fact that the noun for year shows the -ant- derivational sux (2.25, p. 56). For the
accusative designating extent of time see 16.29 (p. 249).
16.76. For the use of dative-locatives of time words repeated in distributive expres-
sions, see 19.10 (p. 291).
Units of Measure and Dimensions
16.77. The dative-locative is also used with units of measure or degree.
d
ts:+u
[(ma mukan o:+ ilani ilani )] namma tiket But ts:+u, my lady, kept on ad-
vancing me by degrees (ilani ilani, distributive expression, see 19.10, p. 291) att. iv
3940; nutta
GI
ororn.nr. [ . . . ] . . . r.nr. txt.nr. iueni piket x.orfr.
nr.yatta niito nrrrn.nr. i+,+ [o.xr.o.nr]. iueni pi[ket] And he kept
giving you chariots, [ . . . ] . . . grain and seeds in large amounts (iueni); and [he kept]
giving you beer-wine, malt, beer-bread, rennet (and) [cheese]s in large amounts KUB
14.1 obv. 78 (MHlMS).
16.78. Occasionally the dative-locative indicates not the unit of measure but the
dimension to be measured: na dalugati 3 u palati[maa . . . x u]
uya she was three miles in length and x and a half miles in width KUB 33. 98
+ KUB 36.8 i 1516 (Ullik. I B).
Additive-Incremental
16.79. The OH expression andaea (as well as its NH modernized form anda
ya i ) in addition to it (see fuller discussion in 30.17, p. 411) supports an additive-
incremental use of the dative-locative (e it) already in OH.
Allative
16.80. The allative case (identied by its characteristic ending -a) in Old Hittite
exclusively indicates place to(ward) which, in contradistinction to the dative-locative,
which usually (not always; see 16.70, p. 259) answers the question in what place?
The allative occurs with directive verbs (go, come, etc.), as opposed to stative
ones (lie) (Starke 1977; Rieken 2005a: 106). On the form and name of the allative see
3.31 (p. 76).
16.81. Comparison of the pairs nepia to the sky KUB 17.8 iv 19 and nepii in the
sky, as well as aruna to the sea KUB 12.60 i 11 and aruni inlat the sea makes this
distinction quite clear: [(eu)]
d
Telipinu zikza [. . . a(runa t) ne]pia
d
tft-un arunaz
[rorn-p(a uw)ate]
d
Telipinu aruna pait Come, Telipinu! You go . . . to the sea. Bring
16.76 Noun Cases
00-HittiteGrammar.indb 262 6/5/08 4:18:25 PM
263
back the Sungod of Heaven from the sea. So Telipinu went to the sea KUB 12.60 i 911
(myth, OHlNS) with dupl. KUB 33.81 iv 67 (NS).
16.82. The allative expresses that object
20
to(ward) or into which a movement is
directed ( parna to the house, laa to a campaign, aruna to the sea, ap to the
river, gimra to the open eld, i into the mouth, tunnakina into the bedroom,
nepia to the sky, takn into the groundlearth). With many of the same nouns of loca-
tion, the dative-locative, on the other hand, designates the object in or on which a thing
is situated ( parni in the house, i in the mouth, nepii in the sky, takni on earth).
Whereas the allative implies movement, the dative-locative implies rest. When describ-
ing the role of the allative, it is important to distinguish verbs of perambulation or gross
movement pai- to go, uwa- to come, iyanni- to set out, iya- (mid.) to travel,
arnu- to transport, carry, uda- to bring, peda-, uwate-, peute-, tarna- to release,
etc., all of which govern the allative and those which describe non-perambulatory
and minor movements such as ipant- to libate, such as iuwa- to pour (solids into),
lauwa- to pour (liquids into), dai- to place (upon), which in OS mostly construe with
the dative-locative (but note the example with ariya- to bury cited below). Conclu-
sions drawn from NH copies of older texts, where the scribes no longer fully understood
the function of this case, are particularly dubious.
16.83. Speakers could use, even if they only rarely did so, the allative for movement
through time: He shall feed them (i.e., cattle) [nua]n par ameanda ar [nuzi]
and shall bri[ng] them (i.e., the cattle) through to the following spring KBo 6.2 iv 60
(Laws 100, OS; see Starke 1977: 136 and LH 9798, 198), where ameanda is an
allative based on the -ant- individualizing sux (see 2.25, p. 56). Compare OH xt-
anti (*wittanti also an -ant-suxed time-word in the dative-locative) which indicates
during the course of a particular year (see 16.75, p. 261).
16.84. The use of the allative of place to which is attested in texts showing the Old
Script: [appe]zziyana
m
Anitta itoi.oi
d
iuu[m(min
U
)
RU
Z ]lpuwaz ppa
URU
Na
p[daun] But subsequently I, Anitta, the Great King, brought our deity back from
Zalpuwa to Nea KBo 3.22:4142 (OS); mn tunnakinama paizzi When he goes
into the inner room KBo 3.22:78 (OS); takn ariemi I bury (it) in(to) the ground
StBoT 8 iii 89, utxt.xr(-an) parna paimi I go into the house of the children ibid.
ii 3, 16f., iii 17; u u-a tarna and she let them (the baskets containing her babies)
into the river KBo 22.2 obv. 3 (OS). But even in Old Hittite texts which were recopied
in the New Hittite period (OHlNS) the true use of the allative survives: na laa
paizzi and he goes to a campaign KBo 6.3 ii 48 (Laws 42, OHlNS);
m
Zuliyawa ap
paiddu let Z. go to the river ordeal KUB 13.3 iii 29; naza parnaa iyanni and
he (Appu) went to his house KUB 24.8+ i 2425; nat r.X-na pdi and he carries it
to the inner room KUB 2.6+ ii 8, etc.
20. Rare instances of what may be persons are all late and quite likely due to confusion on the part of
speakers for whom this case was no longer a part of their living speech.
Noun Cases 16.84
00-HittiteGrammar.indb 263 6/5/08 4:18:26 PM
264
16.85. The allative ending is found also on clitic possessive pronouns accompany-
ing nouns in the allative: parnaa to his house, parnamma to my house, tuel
parnat[ta] to yo[ur] house KBo 23.9 i 10; see 6.4 (p. 139).
16.86. The local meanings of the allative and dative-locative are given greater preci-
sion by the addition of the local adverbs anda(n), appa(n), anda, itarna, katta(n),
peran, ar, and er.
21
And although most of these constructions are attested only in
the New Hittite period, when no formal distinction is made between allative and dative-
locative singular, it is often possible on the basis of meaning (motion toward or rest
inlon) to assign them to either allative or dative-locative. For instance, er above, upon
goes with the dative-locative, while ar up onto goes with the allative.
16.87. The allative goes out of use in NH, surviving only in a few lexicalized expres-
sions such as tapua aside, to the side and dameda elsewhere.
Ablative
16.88. The fundamental study of the ablative and instrumental cases in Hittite is
Melchert 1977. For the ablative endings attested for nouns see 3.32 (p. 76). For the
ending on clitic possessive pronouns accompanying nouns in the ablative see 6.4
(p. 138).
16.89. The ablative most often designates the origin of a movement and answers the
question from what place?: iaz out of the mouth, nepiaz from the sky, wetenaz
out of the water. Often the notions of source and separation are equally stressed in an
utterance. Examples where source or origin (Melchert 1977: 19293) is more prominent
are: kar
m
na . . .
d
iu(n)umm[in
U
]
RU
Naz
URU
Zlpuwa pd [a] Previously (King)
Una had carried o (the statue of) our deity from Nea to Zalpuwa KBo 3.22 obv.
3940; namma
URU
Paluiaz ppa
URU
xi.nnnn-ti uwanun Then I came back home
from Paluia to attua KBo 3.4 ii 7 (annals of Mur.); nata l-an aruwantazet
xn-ezzi he may recover the lard from its (i.e., the dogs) stomach KBo 6.3 iv 28
(Laws 90, OHlNS; see 16.111, p. 270); kamu PN
URU
Kaepraz atr[it] PN has
just written to me from the city Kaepura HKM 25:45 (MHlMS); 5 r.xtn.n.nr.
wa eker nuwara parnaza arninkun ve horses died, and I replaced them from
(my own) house (i.e., personal property) KUB 13.35+ ii 40; nukan
d
tft-u n[(epiaz
k)atta] akwaiyat The Sungod looked down from the sky KUB 24.8+ iv 2223,
3031; nu urorn.xr-az ariyaer so they sought information by oracle from the gods
KBo 4.6 obv. 26; written with ts:i: (he does not render corvee labor himself,) ts:i r
+utsima k[arpianzi] but [they] re[nder] (it) from the house of his father KBo 6.2 ii
40 (Laws 46, OS). Only rarely is this sense of the ablative employed with a noun or
pronoun denoting a person: r:+ma ammdaza o.ftxt-ti kuitki My brother, you
21. On the derivation of local adverbs in -a see 3.31 (p. 76).
16.85 Noun Cases
00-HittiteGrammar.indb 264 6/5/08 4:18:27 PM
265
are enriching yourself somewhat at my expense (lit., from me) KUB 21.38 obv. 16
(see 18.5, p. 278).
16.90. Examples where separation (from what place?) is more prominent are:
urorn.uruir-a utxt.xr-u .n.n-az ar dir But the gods picked up the
boys from the sea KBo 22.2 obv. 45 (OS), ed. StBoT 17:6; see Melchert 1977: 149);
GI
ftxti.nr.-u muta zo.urn-za daun I removed (lit., took) their(!, text
his) weapons from (their) shoulders KBo 3.1+ ii 30 (OHlNS); itoi.oi tabarna
s+ orxr.xrsi t.xr-u ts:i
4
n
5
daun s+ ln.xrya t.xrsi\i ts:i xr
daun I, the Great King, the Tabarna, took the hands of his (i.e., the enemy kings)
servant girls from the millstone; I took the hands of (his) man servants from the sickle
KBo 10.2 iii 1517 (annals of Hatt. I, OHlNS); 1-amakan . . . arunaz ara uet (one
of the two brothers stayed in their island refuge, lit. in the sea) but the (other) one . . .
left the island (lit., came outlaway from the sea) KBo 3.4 ii 5354 (annals of Murs. II).
Verbs used in clauses with the ablative of separation are (1) intransitive motion verbs
( pai- to go, uwa- to come), (2) transitive motion verbs (uittiya- to pull, draw,
peda- to carry, d- to take, uda- to bring, uwate- to lead, conduct), and at least
three verbs which themselves express separation: amen- to withdraw from, tu- to
separate (someonel-thing) from, and tilea- to abstain from. A rare example where
a person is denoted by the ablative is [(peran par)]yazzi apn o[(r
6
-a)]n ts:i xt-
ts-:t [(tea)] And beforehand throughout that night (accusative of extent, 16.29,
p. 249) (the king) abstained from a woman KBo 4.2 iii 5859 (so Melchert 1977: 348,
correctly).
16.91. From the notion of origin or source it is easy to derive the idea of cause
(Melchert 1977: 192; LH 166): [takku i-an n]ama xtts-an ullannaz kuiki kuenzi
If someone kills a (free) man or woman out of (sheer) wantonness KBo 6.3 i 1 (Laws
1, OHlNS), see also Laws 2, 27, orx-anma r:+
m
rn.oi ts:i +++: (for +w+: )
urorn-tt+si t\+ xtn s+rtt:t katta pait but when my brother Muwatalli because of
a word (Akkad. ts:i covers the Hitt. ablative) from his god went down (from attua)
to the Lower Land att. i 7576 (NH); apzkan uddanaz ara akkikanzi people die
because of that behavior KBo 5.3 iii 3839 (NH); twarankan tuetaza memiyanaz
kuenne[r] otu aln
d
uwawainna Was it not because of your wordlcommand that
they (scil., Gilgamesh and Enkidu) killed the Bull of Heaven and uwawa? KUB 8.48
i 1213 (Gilg.).
16.92. The ablative is also used for indicating directions (Melchert 1977: 15152,
19596). See 18.5 (p. 278; for independent pronoun in ablative), 20.27 (p. 300) and
20.28 (p. 300). A similar usage of a from construction can be found in the Semitic
languages: Hebrew miss
e
ml onlto the left, miafn to the north. Hittite examples
are kunnaz on the right, oin-laz on the left, antezziyaz in front, ikiaz in back,
backwards, appezziyaz in the rear, ammedaza on my side, tuedaza on your (sg.)
Noun Cases 16.92
00-HittiteGrammar.indb 265 6/5/08 4:18:28 PM
266
side, anzedaza on our side,
m
Uriteupaz on Uriteups side. The postposition
tapuza is an ablative of direction in origin (see 20.28, p. 300). This kind of ablative is
widely used to indicate the boundaries of a territory:
URU
azamai
URU
Zarta zo-
a
URU
Zartamakan +\+ xtn
D
laya anza In the direction of a, Zarta is his
boundary, but Zarta belongs to the land of the ulaya River KBo 4.10 obv. 21 (NH).
16.93. A few OH examples depart from the normal function of the ablative form,
which serves to express direction with reference to a point usually expressed by the
dative-locative (on the left [abl.] (of) point of reference [d.-l.]); these unusual exam-
ples use the ablative to indicate the point of reference: (land located)
NA
4
uwaiaz anda
. . .
NA
4
uwaiaz araza inside of the stela . . . outside of the stela SBo I 4 = LSU 4 obv.
1516 (where the ablative uwaiaz is the point of reference). See discussion of SBo I
4 = LSU 4 obv. 1516 in Melchert 1977: 15354. An OS example (unfortunately partly
broken) has the same construction of ablative and anda meaning inside of, i.e., toward
the speaker from the ablative point of reference: utn manda
URU
Zlpuaz anda arunaz
[anda(?) . . . ] All the lands on this side of Zalpuwa, [on this side(?)] of the sea [. . .].
KBo 3.22:38 (Anitta, OS). Two other Old Hittite examples where the ablative marks
the reference point, but where kt and edi ll the role of anda and araza, are: takku kt
(later var. kz) u-az . . . takku edi u-az if he is on this side of the river . . . if he is on
that side of the river KBo 6.2 i 4950 (Laws 22, OS) (where the river [abl.] is the point
of reference: the pair does not mean in the direction of this river . . . in the direction of
that river) and
[LU.ME]
xtniurx ipantuziaaru edi taranzipaz [x-x] pdanzi The
cooks carry o the libation vessels to the other side of the taranzipa-platform KUB
43.30 (StBoT 25 #30) ii 67 (OS). In these instances there is no reason to take edi as a
dative-locative any more than OS kt (see 7.13, p. 145).
16.94. The ablative is used in temporal constructions. The forms kitpandalaz from
this moment on KUB 11.1 iv 5, apit(-)pantalazpat from that very moment KUB
33.118 iliv? 24, utxt-annaz from childhood KUB 30.10 obv. 10 (prayer, OHlMS),
annazma kartaz from (my) mothers womb (lit., heart) KUB 30.10 rev. 20, and
(apezlkez) tu(.xx)-az from (thatlthis) day are clear examples of time from which
(Melchert 1977: 289, 310, 35253). It appears that speakers distinguished this usage
from the following one by the obligatory demonstrative (apezlkez). The construction
annaz kartaz from the mothers womb also shows partitive apposition (16.10, p. 243;
16.24, p. 247; 16.38, p. 251; 16.42, p. 252).
16.95. But beginning in MHlMS the ablative is also used for the kind of time, like
the German genitive nachts = by night: tu(.xx)-az (iwattaz) by day KBo 5.8
i 24, KUB 19.37 iii 11 (Mur. II) and ipandaz(a) (or
6
.xx-za) by night HKM 25:7
(MHlMS), KUB 23.11 ii 22 (MHlNS), KBo 5.6 i 22, KUB 19.37 iii 12, KBo 4.4 iii 63
(Mur. II), just as rarely the accusative may appear (KBo 4.4 iii 62 in 16.29, p. 249).
For the accusative of extent of time see 16.29 (p. 249); for dative-locative of location
in or extent of time see 16.74, p. 261, and 16.75, p. 261).
16.93 Noun Cases
00-HittiteGrammar.indb 266 6/5/08 4:18:28 PM
267
16.96. A perlative ablative (translated through; Melchert 1977: 157, 208) is found
in
d
tft-u[(kan
GI
lutt)]anza anda tarnai The Sungod slips in(?) through the window
KUB 33.70 ii 2 (NS); nammaakan
GI
n-az andan 2s[i appanzi] Next [they hold]
them (the laanza-ducks) in through the window twice KUB 39.7 ii 22 (late NS).
16.97. For the one-time occurrence of the ablative for comparison see 17.15
(p. 274).
16.98. In the post-OH period we see uses of the ablative taken over from the in-
strumental for the expression of means (Melchert 1977: 20912): na rzr-az kattan
[lukkun] And [I burned] them down with re KBo 10.2 iii 39 (OHlNS); n+,,t\\iwa
t-za p take the axe with (your) hand! KUB 8.50 iii 11, zaiyaz katta daun I
captured (the cities) through battle KBo 3.4 iv 40, apnmakan
d
ts:+u
URU
amua
o:+ xt
6
-un orx-an palaza rorn-pa itapta the amuian ts:+u, my lady,
ensnared him like a sh in (i.e., by means of) a net KBo 6.29+ ii 3334 (NH), nata
itoi-u
URU
Ta[ur]pi a[r]
GI
ororn-az paizzi the king goes up to Taurpa by chariot
KBo 10.20 i 1617 (OHlNS).
16.99. Similar to the ablative of means is the ablative of agency (with passive
verbs), likewise rst used in post-OH texts (Melchert 1977: 21415, 367): t\+
URU
Ta
nipiya
A.A
kuera itoi-waz piyanza a eld in Tanipiya was given by the king KBo
3.7 iv 2223 (OHlNS);
URU
attua utne inganaz tamatat the land of attua has
been oppressed by a plague (abl.) PP 1 i 89; ts:i urorn-tt+mu par par sro
5
-
ikattari through the agency of the goddess (abl.) things went better and better for me
KBo 6.29 i 1011 (NH), ed. Goetze 1925: 4447 (ts:i urorn-tt+ might stand for iunit
or iunaz); kinuna ammuk
m
rn.oi itoi-u tuedaz [ts]:i
d
t piaai allanuwanza
arkuwkemi Now, I, King Muwatalli, who have been raised up by you, O Stormgod
of P., am oering (this) prayer KUB 6.45 iii 3233 (Muw. II).
16.100. Less common is the comitative ablative (Melchert 1977: 39798): nakan
tetenaza katta aruni ara (the Stormgod sprang up from his wagon,) and came down
with thunder (abl.) to the sea KBo 26.65 iv 2122 (Ullik. III),
d
ts:+u-imakan -az
tn.so-annaz ara uet Is:+u came out of the sky with bravery (abl.) KUB 33.113 +
KUB 36.12 + KBo 26.64 i 1718 (Ullik. II B). Some examples of this type function
virtually as adverbs: parenaz with haste = hastily.
16.101. Uncertain cases with ts:i could be covering either instrumental or ablative:
[(nu ts:i xn)] [(r.xtn.n).xr pun] [I went(?)] accompanied by infantry and
chariotry (abl.) att. iii 4546. For the instrumental of accompaniment see 16.108
(p. 269).
16.102. A rare partitive ablative can be observed in: nata ts:i
UZU
o.oro uia
waz wkuen we have bitten o (some) of the raw liver (abl.) HT 1 i 44. Usually, the
partitive function is conveyed by apposition rather than by a particular case form. See
an example in the ablative: annaz kartaz from (my) mothers womb KUB 30.10 rev. 20
Noun Cases 16.102
00-HittiteGrammar.indb 267 6/5/08 4:18:29 PM
268
(see 16.94), the much better attested partitive apposition in the accusative (16.24
16.25, pp. 247248) and a possible example in the nominative (16.10, p. 243).
16.103. For the ablative with various postpositions see 20.27 (p. 300) and 20.29
(p. 300).
Instrumental
16.104. For the instrumental endings see 3.35 (p. 77). The instrumental designates
the tool, means, instrument, or material by or through which an action is accomplished
(otu.nr.-it by the oxen, aganda with grease, paunit with re, llit with the
tongue, natit with an arrow, wetenit with water,
GI
atalkenit with a (branch of)
hawthorn, etc.). It is used to express cause (ktit (die) from hunger), and, with pas-
sive verbs, the agent (see iunit by the god). There is also an instrumental of accompa-
niment (appantet kunantit together with captured and killed (men). By New Hittite the
instrumental is no longer productive, its functions having been assumed by the ablative
(Melchert 1977: 42425). When a noun in either the ablative or instrumental case is
represented by an ideogram, it is specied as to case by the Akkadian preposition ts:i
from, with, by means of. See chapter 31.
16.105. The instrumental case is employed chiey to denote the means or instru-
ment by which an action is performed (Melchert 1977: 16264, 22745, 300301,
32732, 37173): tuppu akanda (instrumental of agan oil, grease) unna she
lled (i.e., caulked) the baskets with grease KBo 22.2 obv. 2 (OS); andakan alna
teummiu tarlipit wamu 2-:++ ptumini Twice(?) we bring in cups (made) of
alina, lled with tarlipa KBo 17.1 + KBo 25.3 i 2627 (OS); nu tuiyattit kti
and you will die by means of suocation KBo 7.14 obv. 5 (OS); nukan rzr er wetenit
kitanuanzi they extinguish the re on top with water (wetenit) KBo 4.2 i 12, nu
nepia
d
tft-un ror.nr.-it ukezzi she sees the sun of heaven with (her) eyes KBo 4.8
ii 19 (NH), ed. Honer 1983a: 188; idlawa ror.nr.-wa
GI
atalkenit katta tarmn du
let the evil eyes be pinned down by the hawthorn KUB 12.44 iii 89; [(nu
d
ts:+u)]
o:+ +\+
m
Murili +ut:+ i-it
m
rn.oi-in r:+ uiyat Is:+u, my lady, sent my
brother Muwattalli to my father by means of a dream (i.e., in a dream of the father)
att. i 1314; nu
MU
illuyankan iiman[(ta)] kallit (upaiya) tied up the serpent
with cord(s) KUB 17.5 i 1516 (OHlNS); nukan oln-it ar pun I went up on foot
KUB 14.3 i 24 (Hatt. III); itoi-u uezzi nau
GI
ororn-it nama
GI
ulugannit t\+ r-
:t+ oi paizzi The king proceeds to travel to the Great House either by chariot or by
cart KBo 19.128 i 25, ed. StBoT 13:2f.
16.106. There are also instances of the instrumental used to express cause: mankan
mn +\+
m
Attariya uiwetenna ktitaman kten and even if you had escaped
from Attariya (contrary-to-fact condition), you would have died from hunger KUB
14.1 obv. 12 (MHlMS). The phrase man mn . . . man expresses the contrary-to-fact
condition (HE, p. 166, 330; see also CHD sub man). See also [k]itantit arkueni we
will die of hunger KUB 17.10 i 29 (OHlMS). For an attempt to motivate the use of the
16.103 Noun Cases
00-HittiteGrammar.indb 268 6/5/08 4:18:29 PM
269
instrumental instead of the more common ablative (16.91, p. 265) see Francia 2002a.
It is possible, of course, that the Hittites viewed hunger (katit) not as a cause but as
a means of dying.
16.107. The source of the ablative of agency used with passive verbs was the instru-
mental of agency:
GI
ftxti.nr.-iwatta iunit piyante weapons have been given to
you by the god KBo 22.6 + KUB 48.98 i 2425 (OH?lNS); [utxt.utxt].utxtsi
s+
m
attuili itoi.oi s+ itoi
URU
Kuar
d
iunit k[ania]nda (gen.) txt-a
(nom.) [great-grand]son of Hattuili (II), Great King, seed of the king of Kuar (i.e.,
attuili I) who was recognized by the god KBo 6.28 obv. 45 (Hatt. III) [on this last
passage see Goetze 1968; Carruba 1971: 78; and especially Gterbock 1970: 75]. Since
the instrumental is replaced in all uses by the ablative in NH (see 16.9816.100,
p. 267, and Melchert 1977: 42330), the use of the instrumental to mark the agent with a
passive participle cannot be an innovation and must reect OH practice, despite the lack
thus far of OS examples. This absence may easily be due to chance, given the relative
rarity of expressed agents with the passive. Despite the claim of Starke (1977: 1024),
the fact that agency is expressed once in OS by the circumlocution t\+ +:t + genitive
by the hand of in no way proves that agency marked by the instrumental of a noun
referring to a person did not occur in OH. For the instrumental as the case of the agent
in PIE (inherited in Hittite) see Jamison 1979.
16.108. The instrumental is also used to express accompaniment (comitative)
or inclusivity:
LU
n
GI
unzinarit sln-ui The singer sings to the accompaniment of
the unzinar instrument KUB 30.24 ii 910; maanmaat eelliya witenit rorn-pa
uwanzi But when they come back with the water of purication KUB 7.20 obv. 1011;
(I eliminated 16 men of the enemy) appantet kunantit with (or including) = count-
ing both) captured and killed (men) HKM 10:40 (MHlNS).
16.109. In a very few cases the instrumental expressing separation is attested. For
example, nakan aanit luzzit arawaun I freed them from s. and l. KBo 10.2 iii
1819 (annals of Hatt. I). Such examples are attested only in copies of older composi-
tions, and it is unlikely that this usage of the instrumental is genuine for any period of
the language (see Melchert 1977: 426 with refs.).
16.110. Since the Akkadograms ts:i and +oi could conceivably reect either the
ablative or the instrumental, we cannot identify the case of the following comitative ex-
amples with certainty: i
URU
al[(atiwara +oi utxt.xrsi ari) numu ror]-anda
uet The king (lit., man) of alatiwara arose together with his sons, and he came [ag]ainst
[me] (in battle) KBo 3.22:6566 (OS, restorations from NS dupl.); s+ orxr:+ata
kuedani uddan watarnaikenun tuelmaankan maan maniyaante ts:i zu.
u[] ara dayaer nanan [. . .] t\+
URU
Taazzimuna rorn-pa put [e]r Concerning
what I informed you about my maidservant, when your subjects stole her away together
with the our (she had milled), they carried her back to Taazzimuna HKM 36:4248
(MHlMS), ed. Alp 1991a 184187.
Noun Cases 16.110
00-HittiteGrammar.indb 269 6/5/08 4:18:30 PM
270
16.111. In OH the instrumental of the enclitic possessive pronouns (see chapter 11)
is used with nouns in the ablative, since no ablative forms of the enclitic possessives
existed (HE 218c): aruwantazet from its stomach KBo 6.3 iv 28 (Laws 90,
OHlNS; see 6.7, p. 140); iaz()mit out of their mouths KBo 17.1 i 18 (OS), zo-
aztit on your right KUB 31.127 i 66; kartazmit from my heart IBoT 3.135:6. To
be similarly understood is kitpantalaz from this moment on KUB 11.1 (BoTU 23B)
iv 5. There is no basis for the claim of Houwink ten Cate (1967) of an old locative
form in -(e)d.
Cases with Particular Verbs
16.112. A number of verbs can be construed with dierent cases, in many instances
showing a slightly dierent meaning.
16.113. wtarna- (1) with the accusative of the person means to command or
commission (someone), while (2) with the thing spoken in the accusative and the per-
son in the dative it means to send a message or announcement to (someone) (AU 223).
See KBo 18.48 obv. 1011 (interpret contra Hagenbuchner 1989: 9).
16.114. n- to fear, be afraid of governs two dierent cases: (1) the accusative:
namiu I fear them KUB 8.65:22; (2) the dative-locative: pauenaa uddan me-
kki naante ten be very careful also in the matter of re KUB 13.4 iii 44; (3) the
ablative: nuza alluwayaza mekki naante ten be very careful of strife KUB 13.4
iii 43. In addition there is the impersonal use discussed in 16.32 (p. 250).
16.115. punu- to ask, inquire can be construed in two ways. (1) One can say: to
question someone (accusative) about something (dative-locative), e.g., na
d
tft-st
+\+ ur.nr. punumi and I, My Majesty, will question them regarding the legal dis-
putes KBo 3.3 iv 910; or (2) one can say: to investigate something (accusative) for
someone (dative-locative): numakan ur.nr. punukeddu Let him investigate legal
disputes for you (-ma) KBo 3.3 iii 2829.
16.116. Similarly the verb wek- to ask, request construes with the accusative of the
thing requested and the dative of the person asked: kinunawama kuit wkmi What
now shall I ask from them (i.e., the gods)? KUB 34.53 rev. 8 (CTH 370).
16.117. arra-: (1) (active) with the accusative and no -kan (often with -za) means
to divide something; while (2) (active or medio-passive) without -za, but with -kan
or -ata and the accusative or more often the dative-locative, it means to pass through,
cross over, transgress.
16.111 Noun Cases
00-HittiteGrammar.indb 270 6/5/08 4:18:31 PM
271
Chapter 17
ADJECTIVES
17.1. The inection of adjectives is essentially the same as that of nouns.
17.2. An adjective can be modied by an adverb, as in ntn.so.xryawa mekki
nti-u The mountains too are very bad KUB 23.103 rev. 21 (NH).
17.3. Adjectives can be attributes or predicates. Most attributive adjectives precede
their head nouns (alli xtn-e the great land), while predicate adjectives regularly fol-
low a noun subject (xtn-e alli ta the land was great). For more on the word order
of predicate adjectives see 30.28 (p. 413) and following.
17.4. Examples of the normal word order with attributive adjectives: nakki urorn-
tt+-i honored city (sg. nom. com.); nakk pedan honored place (sg. nom.-acc.
neut.); nakkit
d
Tamiu with the honored Tamiu (ins.); pangawi rnr.xr-ti to all
the troops (sg. d.-l.); panqauwaz rxr-azza from every (evil) tongue (abl.); pargamu
ntn.so.xr-mu the high mountains (pl. acc. com.); pargauwa auriya into the
high watchtowers (pl. d.-l.); parkuin inan cleanlpure dough (sg. acc. comm.). For
further examples see the respective articles in the CHD for the adjectives cited here and
others.
17.5. Attributive participles regularly follow their head nouns. For examples of the
regular order and some of the few exceptions see 25.4125.42 (p. 339).
17.6. The universal quantiers mant- and dapiant- (both translatable as all, en-
tire) also regularly follow their head nouns. This may be due to their meaning or to their
formal resemblance to participles in -ant-. Examples: nuzaapa utniyanza manza
(10) ikimet anda
URU
attua lagan ard[u] and let the entire land keep their backs
bowed down (lit., inclined) toward attua KUB 36.110 rev. 910 (OS), a[ima]
utn man tinut The freezing wind paralyzed the entire land VBoT 58 i 78 (OHl
NS), nu ux.xrsi [utxt.xrsi x.n.nr.]si-\i ya mantapt d
and he took his wives, [his children,] his [civilian captives] and all the goods KUB
14.1+ obv. 50 (MHlMS); nu ipandan mandan iyaat and I traveled the entire
night KBo 5.8 iii 21 (NH);
GI
noryawa
DUG
inuri orn
4
.nr. dapianda s+
LU
x.
r.xxpt dagaweni We take nor-vessels, kneading troughs, and all the pottery
belonging to that very poor man KUB 16.16 rev. 1415;
LU.ME
Gaga
I.A
maankan
dapiante ox tot n-anzi But all the Kaska men will strike it (the city) up from
below KUB 5.1 iii 6162.
00-HittiteGrammar.indb 271 6/5/08 4:18:31 PM
272
17.7. The quantier mant- rarely precedes its noun: mand[az x]tn.xtn.xr-za
eu come from every land KUB 36.90 obv. 39; arazena mantaz xtn-yaz KBo
20.107 + KBo 23.50 iii 26 (vs. normal utn manda KBo 3.22:38 [OS], utn man
VBoT 58 i 8 [OHlNS]). These exceptions are only a very small fraction of the thousands
of examples of the regular word order, where mant- follows its head noun. Some ap-
parent exceptions actually form part of the rule. For example, preposed mant- is nor-
mal when used with body parts: mantet kardit with the whole heart KUB 30.10 obv.
24, 27 (see analogous dapian zr-an the entire zr KUB 5.1 i 2, etc.); mandaz tueggaz
KUB 24.9 i 45 (vs. normal [].frsi man KUB 26.8 ii 3); mandaza appenaz
from every limb KUB 30.40 i 17; mantaza tarnaza from every tarna- KBo 15.37
iii 5152. Preposed mant- is also regular when it is substantivized and modied by
a predicatival adjective: man u everything that is good KUB 17.10 i 22 (but see
u man every good thing KBo 10.37 iii 43); man ornri-:t+ everything that is
new KBo 17.65 obv. 24; but idlu man alllevery evil KBo 15.21 i 22.
1
17.8. Attributive adjectives in -want- may precede or follow their head nouns. The
postposed examples may or may not be due to their formal resemblance to participles
in -ant-.
2
Preposed examples: iarwantu fo.nr.-u blood-red garments KBo 17.1
i 25 (OS); iarwandan
d
t.otn bloody Nergal KUB 9.34 i 26; pakkuawan eppit
cracked eppit-grain KBo 21.1 i 15; iwanduakan laanzu live l.-birds, on the
other hand (adjective fronted and with contrastive -al-ma) KUB 39.7 ii 33. Postposed
examples:
MUEN
aranan uuwandan a live eagle KBo 17.4 ii 14 (OS); ranan
MUEN

lliwandan the swift eagle KUB 17.10 i 24 (OHlMS); luliya naduwanza a reedy
marsh KUB 52.91 iii 3.
17.9. The original syntax of nouns with the quantiers mekk(i )- much and tpu- a
little, few was that of an appositional phrase, with the quantied noun in its appropri-
ate case followed by the adjective as a neuter substantive. This syntax is still rarely
attested for tpu-:
4
z.o tepu . . . tia tepu lapis lazuli a little bit . . . and
uti-material (sg. nom. com.!) a little bit KBo 5.2 i 3738. It is better preserved
in the partitive expression kuitta par tpu (of) each in turn a little bit following lists
(see CHD par 9a). The terms tpu- in most cases and mekk(i)- in all cases
3

appear as
attributive adjectives that agree with their head noun in gender, number, and case and, as
expected, they regularly precede their head noun. But unsurprisingly the original syntax
1. The example [(nutta )]man upp[i (parkui p)]ikanzi KUB 24.1 ii 15 (restorations from parallels
such as KUB 24.2 obv. 1819) may involve an adverbial use of the adjectives, as per Gurney (1940: 21):
and everything they present to thee holy (and) pure. See also remarks in CHD P 165 sub parkui- a2o.
2. Some stems in -want- function as virtual participles (see 2.50, p. 61) but, as the examples cited here
show, postposing aects both these and possessive -want- adjectives.
3. Since the word for eld written .X is common gender in KBo 6.2 (demonstrated in LH 311), it
is likely that the phrase takku .X-na mekk wi in KBo 6.2 ii 46 (OS) is likewise literally But if he
buys a elda large portion of it . . ., with the older syntax. Elsewhere mekk(i)- is always an attributive
adjective.
17.7 Adjectives
00-HittiteGrammar.indb 272 6/5/08 4:18:33 PM
273
also is partially preserved in occasional instances where mekk(i)- and tpu- follow the
head noun, but show agreement in number and gender: mriu mekku many (grape)
clusters KUB 43.23 rev. 22 (OS);
LU.ME
ipantuzziyaliu tpaw few libation servers
KUB 14.8 rev. 19 (NH). The semantic essence of the old syntax is caught in the OH
quoted speech in colloquial language:
GI
ftxti ou.u wkmi x [ . . . mu
GI
ftxti ou.
u] pu nuwa mekki I will demand long weapons (as follows): Let him give [me long
weapons] plenty (of them)! KBo 3.38 obv. 3233 (OHlNS), ed. StBoT 17:89.
17.10. There are also some examples where ordinary attributive adjectives appear
to follow their head nouns: riya nakkya inaccessible valleys KUB 30.36 ii 4; so
mitali- red wool (frequent; examples in CHD mit(t)a- 1). For two attempts to dene
the conditions for this postposing see Laroche 1982 and Francia 2001.
4
17.11. More complex rules of word order are treated in 16.47 (p. 253), 18.2418.28
(pp. 284285), 18.37 (p. 287), 26.3 (p. 341), 26.21 (p. 345), 30.1. (pp. 406.).
Comparison of Adjectives
17.12. Hittite does not express the comparative (better) and superlative (best)
degree by suxes, as is the case in Greek - and - or - and -, Latin
(-ior and -issimus), and English (-er and -(e)st), but by syntactic means.
Comparative Degree
17.13. The comparative degree was expressed by the predicate adjective governing
the object of comparison in the dative-locative case (HE 22123). Thus: nuwakan
+\+ rnr.xrx+ rnr.xr:+ mek[ki], +\+ r.xtn.n.xrmawatta r.xtn.
n.xr:+ mekki my troops (rnr.xr:+)
5
are more numerous (mekki)
6
than your
troops (+\+ rnr.xrx+); my horses are more numerous than your horses KUB 19.29
iv 1820; nammakan anzel fr-anni [(t s)]+ uiti
ME
-\t fr-tar nakk if the life of our
lord (i.e., the king) (s+ uiti-\t fr-tar) is not dearer (to us) (it + nakk) than our own life
(anzel fr-anni) KUB 31.42 ii 1819; +\+ itoi xtn
URU
t-taakan 1-a
LU
tukanti
alli du nammamaikan l kuiki alli Only the crown prince (tukanti)
4. A very high percentage of such postposed examples involve cases like so mitali- red wool where
the head noun is written logographically (with or without phonetic complement). We must therefore con-
sider seriously the possibility that many or most such examples do not reect the genuine Hittite word order
but merely imitate graphically the Sumerian word order in phrases written entirely logographically like so
s
5
red wool. Further investigation of genuine postposed adjectives in Hittite must take into account this
possible confounding factor. In assessing postposed attributive adjectives one must also be careful to set
aside examples of adjectives used appositionally (see 15.515.6, pp. 236237).
5. The Hittite noun underlying rnr.xr is common gender, rnr.xr-t-.
6. The nouns for troops and horses are common gender. The form mekki may be construed as a col-
lective noun a large amount serving as the predicate. As a noun and not an adjective, it need not show
agreement.
Adjectives 17.13
00-HittiteGrammar.indb 273 6/5/08 4:18:34 PM
274
shall be greater than the King of Tarhuntaa; let no one else be greater BrTabl. II
8081 (Tud. IV). Note how the local particle -kan is employed in all of the previous
(NH) examples. Yet in OHlNS this particle is not always present: arietta ari
7

oi-li so.utsi +\+ so.uts[i oi-li] [x]xxsi +\+ xxxsi oi-li ror.
nr.si +\+ ror.nr.si oi-li (21) [a]eta i oi-li rxrsi +\+ rxrsi oi-li
(22) [kap]ruetata kapruaz oi-li mielieta (23) [mi]elia oi-li ikietata
ikii oi-li (24) [ paltan]ai!ata palta[n] alli onsi +[\+] (25) [onsi a]lli
Xsi +\+ Xsi oi-li this ones (lit., his) stature is greater than that ones (lit., his)
stature; his head is greater than his head; his nose is greater than his nose; his eyes . . . ,
his mouth . . . , his tongue . . . , his mieli . . . his back is greater than his back KUB 43.53
i 1925 (quoted by Sommer and Falkenstein 1938: 219 in unpublished state).
17.14. Several observations are pertinent to the KUB 43.53 passage which was just
quoted: (1) the local particles -kan or -ata are missing in some cases in OH (NS), as
shown by the rst six, eighth, eleventh, and twelfth clauses, but -ata appears in the sev-
enth (kapru), ninth (iki), and tenth (paltana) clauses; (2) the gender of the predicate
adjective is in most cases neuter,
8
although in line 21 rxrsi +\+ rxrsi oi-li his
tongue is greater than his tongue the usual gender of rxr (= lala-) is common;
9
and
(3) although the possessive pronoun is employed with both members in all logographic
writings, it never appears with the second member (i.e., the dative-locative) in the syl-
labic writings. An instance of the common gender in this kind of predicate adjective
employed comparatively is in: [paltan]ai!ata
10
paltan alli his shoulder is
larger than his shoulder KUB 43.53 i 24.
17.15. In line 22 of KUB 43.53 Sommer found a rare example of the ablative of
comparison: kapruetata kapruaz oi-li his kapru is larger than his kapru. See
also Melchert 1977: 21516, who regards it as secondary at best. Note the presence
of the particle -ata. It would be convenient if we could show that -kan was used in
comparisons with the dative-locative, while -ata was used in those with the ablative.
But -ata seems also to occur with dative-locative comparisons (see KUB 43.53 i 2324
in the preceding paragraph). So far, only -kan (no -ata) occurs in the NH examples of
this comparative construction. Since this is the only example of an ablative in this func-
tion, we suspect that the text is not in order. An Old Script (hyp)archetype could have
had the shape of i (HZL #217) confusable with the Old Script z sign (HZL #92l6), in
which case a scribe at some point in the chain of transmission wrote original kap-ru-i
7. Judging from the parallel so.utsi +\+ so.utsi which follows, the rst eari is nominative and
the second dative-locative. Note that the order of cases diers here from the earlier cited NH examples.
8. Since in this case we are not dealing with any form of plurality, a collective interpretation of the
forms in -i is not possible (see p. 273, n. 6).
9. Note, however, that in two passages cited in CHD sub lala-, namely, KBo 2.3 i 50 (MHlNS) and
KBo 23.72 rev. 24 (OH or MHlMS), there is a nominative form rxr-an, which has to be neuter. Perhaps in
KUB 43.53 too rxr is neuter.
10. Emend -i-ta-a-ta here to -i-a!-a-ta for the correct common-gender form.
17.14 Adjectives
00-HittiteGrammar.indb 274 6/5/08 4:18:36 PM
275
as kap-ru-az. At least this is a possible explanation, if we hesitate in accepting this rare
construction on the basis of a single example.
17.16. The syntactic construction employed for comparison is the same when the
predicate is the stative verb takk- to be like (active and medio-passive):
GI
ftxti.nr.-
w[atta] alli wanti atugai kari[tti . . . . . .] weteni takkantari [Your] weapons are
like the great wind, the terrifying ood, and the [. . .] water KBo 22.6 i 2628, and a
long series of clauses involving the body parts in KUB 43.53 i 215. In the latter pas-
sage, as a few lines later in the long passage cited above from i 1925, the occurrence
or non-occurrence of local particles (in this case -apa) seems arbitrary. It occurs in 4
out of 23 clauses (lines 4, 5, 7, 15). In two of the four instances the noun has the clitic
possessive, and in two it does not.
Superlative Degree
17.17. To express the superlative degree in post-OH one construes the adjective
with a plural form in -a, either the dative-locative or genitive (see 16.41, p. 252):
allayakan urorn.xr-a kui alli he who among the great gods is the great(est)
KUB 31.141:3. Often the dative-locative plural is further strengthened by mant- all:
nu urorn.xr-na manda s+
URU
Katama
d
Zaapun alli among all the gods
of Katama Zaapuna is the great(est) KBo 3.7 iv 1517 (Illuyanka myth, OHlNS);
nuza mandapat rorn-ezzi utxt-a eun among (or perhaps genitive, of?) all
(the children) I was the young(est) child att. i 11; [(nuza
d
t maa)]n arzziya
arzzi [(zik ante)]zziyaaz antezzi zi [(k)] As you, O Stormgod, are the highest
of the high, and the foremost of the foremost KBo 34.34:1112 (OHlMS) with dupl.
KUB 33.68 iii 68 (OHlNS). The rst example, showing otherwise unmotivated -kan,
favors the dative-locative interpretation. The remaining examples without -kan favor
the genitive one.
17.18. What are clearly dative-locatives are occasionally also strengthened by itarna
in the midst of: xtn-ekan itarna ap appinanza in the midst of the land (i.e., in all
the land) he was the rich(est) KUB 24.8+ i 1011 (Appu story); even with an adjectival
verb: urorn-ti+ urorn.xr-a itarna taruilezi the god is the bravest among the
gods KUB 36.55 ii 3132.
17.19. When the adjective is not a predicate, the objects with which comparison is
made stand not in the dative-locative but in the genitive: dauai
d
Anu urorn.
xr-a antezziyami peranet arta mighty Anu, foremost of the gods, stands be-
fore him KUB 33.120 i 9 (Song of Kumarbi myth).
11
11. antezziyami their foremost is in apposition to the grammatical subject dau
d
Anu in this
example. It is also possible that antezziya is here substantivized (the leader of the gods), in which case
we are not dealing with a superlative.
Adjectives 17.19
00-HittiteGrammar.indb 275 6/5/08 4:18:38 PM
276
17.20. Since all the examples cited in the preceding three paragraphs stem from
post-OH, where -a can be either dative-locative or genitive plural, caution is needed in
the case identication. Note that some examples show the particle -kan, while others do
not. It may be that (originally) the dative-locative construction required -kan, while that
with the genitive did not. But we cannot prove this.
17.20 Adjectives
00-HittiteGrammar.indb 276 6/5/08 4:18:39 PM
277
Chapter 18
PRONOUNS
Independent Personal Pronouns
18.1. The use of the cases (see chapter 16) is the same for nouns, adjectives, and all
classes of pronouns. As in the case of nouns (16.68, p. 258), the dative of the personal
pronouns is used to indicate not only to and for but also from with verbs like to
take (the so-called dative of disadvantage). The merger of dative and locative into a
single case form (16.65, p. 257) is reected in both the independent and enclitic dative
pronouns, which do not formally distinguish dative from locatival use.
18.2. Since the nite forms of Hittite verbs already indicate that the subject is I,
we, you, (s)he, or they, the presence of an independent personal pronoun in the
nominative case indicates a special emphasis on the subject or a conscious contrast
with the subject of a neighboring clause. Thus, kwa itoi-umi kia I will
become your king KBo 22.2 rev. 15 (OS) could have been written itoi-umiwa
kia without ambiguity as to the subject. Use of the independent pronoun as sub-
ject in order to mark contrast with a dierent actor in a neighboring clause can be
seen in:
MUEN
ranan it[oi-a xtts.itoi-a]a ramet 3[(si)] utxt.r.oi
wanu[zzi ]gamaan rnr.xr-an [(r)] 3si wanmi The palace servant
waves an eagle three times over the king and queen, but I wave (a model of) troops
over them three times StBoT 8 ii 3032 (= KBo 17.1 + KBo 25.3 ii 1618) (OS). See:
zikwa tn.nn.n-a kitat You have become a wolf KBo 6.2 ii 12 (Laws 37, OS),
1

nanza zik d (No one else has accepted my advice,) but you must accept it KUB
1.16 iii 27 (OHlNS).
18.3. The same is the case with forms of the dative and accusative, since a corre-
sponding clitic form of the pronoun (-mu, -ta, -i, -na, etc.; 5.12., pp. 135.) could
have been used: wea kuwapit aumen nu xtts-za [30] utxt [1si ]i nuzza anza
1si ta Where have we seen (it),
2
that a woman bears [thirty] sons [at one time]?
Yet she bore us all at once! KBo 22.2 obv. 1112 (OS); andanzakan a[(nt)]ua
tukpat u nan zikpat ar [l]ikei It is to you (in particular) that the just person
is dear, and you (in particular) always elevate him KUB 31.127 i 810 (OHlNS) with
dupl. KUB 31.128 obv 78 (NS). In these last examples the natural emphasis involved
in using the independent form is strengthened by the particle -pat.
1. The zik is necessary to disambiguate in the NH copy (KBo 6.3 ii 29), where the nite verb ki-a-at is
identical with the third singular form.
2. Reading a--me-en instead of a-ru-me-en (Eichner 1974a: 185). Alternatively, one may retain the
reading arumen and understand the sentence as Where have we (ever) been (lit., arrived), that . . ..
00-HittiteGrammar.indb 277 6/5/08 4:18:39 PM
278
18.4. Likewise emphatic or contrastive in at least in pre-NH is the independent geni-
tive pronoun for which a clitic possessive could have been used. In NH, when the clitic
possessive pronoun was no longer current in speech, the independent genitive pronoun
was often used without implication of contrast or emphasis. Examples of emphasis or
contrast: kui ammel ppan itoi-u kar [i] He who becomes king after me KBo
3.22:22, 49 (OS). For a clear opposition of ammel and tuel see: zilatiyata katta [(a)]a
anzaa tul utxt-an ammel utxt.xr:+ utxt.utxt.xr:+ [ paa]antaripat
In the future, to the third and fourth generations, my sons and grandsons will likewise
loyally protect your son KUB 21.1 + KUB 19.6 i 7375 ; see ammel ln.xr:+ . . . tul
ln.xrx+ my servants . . . your servants HKM 74:10, 1516 (MHlMS); zik ammel
r-na l uwai uga tul parna t uwmi You must not come to my house, and I will not
come to your house KUB 29.1 i 1920 (OHlNS). In the last example the independent
pronouns in both nominative (zik, uk) and genitive cases (ammel, tuel) stand opposed
to each other, and the two clauses are connected using the contrastive particle -al-ma
(uga).
18.5. The ablative forms of the independent pronoun usually convey the directional
sense of the ablative (16.92, p. 265), with the resultant meaning on mylyour side
or in mylyour favor: nu urorn.xr ammdaza [(tiandu) nuwa ur-e-ar ammed]az
annandu Let the gods stand on my side and adjudicate the case in my favor KUB
26.79 i 1112 with dupl. KUB 14.17 iii 1819;
d
r.-a oio-a itoi-u tue[d ]aza
artari Ea, the king of wisdom, is standing on your side KBo 26.118 + KBo 26.65 iv
24 (Ullik. III); anzedaza tiyr urorn.xr anzedazza memier The gods took our side.
They spoke on our behalf KUB 7.58 i 1314. Only rarely do they have the normal
ablatival force of from in the sense of source: r:+ma ammedaza o.ftxt-ti kuitki
My brother, you are enriching yourself somewhat at my expense (lit., from me)
KUB 21.38 obv. 16 . Another possible example would be: tukmawa ammetaz sro
5
-izi
from me good will come to you BrTabl. ii 52 (Tud. IV). This use is rare, because the
ablative case itself is so rarely applied to human beings (16.99, p. 267). But ablatives
of independent pronouns can also convey the more usual idea of agency with pass-
sives and participles (see 16.99, p. 267): i-imakan uxsiya kui iya[ntari]
numakan iyatar zo[-a]an ar [nuwanzi] nat tued azza
d
ts:+u-liazza [t]arn
(as for) the man and wife who love each other and bring their love to completion, it is
decreed (lit., spoken) by you, O
d
ts:+u KUB 24.7 i 3840; kinuna ammuk
m
rn.oi
itoi-u tuedaz ts:i
d
t pihaai allanuwanza arkuwkemi But now I, Muwattalli,
the king raised up by you, O Storm god Piaai, am making petition KUB 6.45 +
KUB 30.14 iii 3233.
18.6. Ablative forms can be found in double case constructions, where they func-
tion like adjectives (Luraghi 1997a: 24 2.1.11, 2.3.1.1): tuedaza memiyanaza on the
authority of your word, ammedaz t-az with my hand, ammedaza
GI
ni.ftn-az with
my spindle, anzedaz memiyanaz on the authority of our words, anzedaz
d
Gulaz on
the authority of our fate-deity.
18.4 Pronouns
00-HittiteGrammar.indb 278 6/5/08 4:18:40 PM
279
18.7. The sux -il(a) added to the independent pronouns expresses the idea of -self,
not in a reexive
3
but an emphatic sense (like Latin ipse), always nominative (see also
p. 357, n. 2): ukila I myself, zikila you yourself, umila you yourselves, apila
he himself, apila
4
and apilu
5
they themselves (see above in 5.3, p. 132, and
Sommer 1922: 48 n. 1, 1938 141 n. 3; HE 99, HED A 88; and HW
2
I 16566). Forms
in -el (or -il)
6
without the nal a are also attested: -kile-el KUB 40.1 rev.! 14; zi-kile-el
KBo 16.41 i 14; a-pa-i-el KUB 36.89 obv. 19; a-pa-a-i-el KBo 22.260 obv. 7 (for
-i-el see 1.62, p. 29). Although apila is properly used only with third-person sub-
jects, on rare occasions in NH one even nds it employed with a second-person singular
verb: namaan apila alziyatti Or you yourself call him (thus with HED A 88 and
Beckman 1999: 67) KBo 5.4 rev. 26 (Muw. II). This shows a developing tendency to
use the third-person form independently in the subject-reinforcing sense self. To claim
(with HW
2
A166b sub 2) that apila is singular accusative here (oder du rufst ihn
selbst) seems to betray a basic misunderstanding of the role of the pronouns in -il(a)
which reinforce only subjects, not direct or indirect objects.
Enclitic Personal Pronouns
18.8. The dative plurals of the enclitic personal pronouns may be used reexively
(28.18, p. 358).
18.9. The dative of possession can be expressed with enclitic personal pronouns
(16.67, p. 258).
18.10. The use of postpositionsladverbs with a spatial locatival sense to govern (en-
clitic) personal pronouns, not yet attested in OS, can be seen in the following examples:
nui peran ara tarna and you gave way before him KUB 14.1 obv. 62 (MHlMS);
nammamani rorn-an xsxi.xr-:t+ t [,u+:] Then he wanted to seize the roads be-
hind him KUB 14.1 rev. 27 (MHlMS); numu k r.xtn.n.xr kuiki kattan Some
chariotry is here with me HKM 30:9 (MHlMS); tmamu
GI
ftxti i.xn kuwapikki
er wanut Nor did she ever let the weapon of the enemy whirl over my head att. i
4142; nui
d
ts:+u
URU
amua o:+ palaan er pta And ts:+u of amua, my
lady, held her p. over him (KUB 31.20 iii 67 + KBo 16.36 iii 910).
3. For the Hittite expression of the reexive (the other use of English -self) see 28.16. (pp. 357.).
4. Although in the following passage apila is used with a plural subject, the discordant numbers be-
tween the serial verb uezzi and the main verb mallanzi show that the speaker was confused as to the number
of his compound subject: nu uezzima
d
ts:+u-i
d
ebatua [
NA
4
nn].nn apila mallanzi Then it will
happen that
d
ts:+u and ebat will themselves mill grain KUB 33.103 ii 78 (NS); see HW
2
A 166b.
5. A NH scribe has axed to apil(a) the plural nominative common ending used (in NH) with a-stem
nouns, namely -u (3.14, p. 69, 3.16, p. 70).
6. Since all forms occur to date only in NS texts, it is impossible to determine if the nal ri sign is to
be read el or il
5
.
Pronouns 18.10
00-HittiteGrammar.indb 279 6/5/08 4:18:41 PM
280
18.11. The temporal meaning before, which may have been derived from an earlier
spatial concept transferred to time (on this see Honer 2002 and see 16.74, p. 261), is
seen with a personal object in passages such as: peranmaatmu before me, however,
(so-and-so governed) it att. i 27.
18.12. Rarely also the dative-locative of a third person personal pronoun is used
to express the ideas to it and for it ( parnaea and to (his) house for it Laws
120; see 16.71, p. 260). The pronoun -e in reference to a thing (it) also occurs in
antayae in addition to it KUB 29.28 i 10 (Laws 128, OS).
(Third-Person) Subject Clitics
18.13. Third-person subject clitics never occur in sentences with transitive verbs
with or without explicit direct objects (Garrett 1990a, following Watkins). Their oc-
currence with intransitive verbs is basically lexically determined (Garrett 1996: 90).
Intransitive verbs that take subject clitics select e- as the auxiliary verb in the analytic
perfect, while those that do not occur with subject clitics use ark- (see 22.19, p. 310,
and Garrett 1996: 1026). Many, but by no means all, of the intransitive verbs that oc-
cur with subject clitics fall into well-dened semantic classes. The behavior of a given
intransitive verb must ultimately be determined on a case-by-case basis.
7
Intransitive Verbs with Subject Clitics
18.14. (1) Intransitive verbs of travellperambulation such as ar- to arrive, uya-
to run, iya- (m.-p.) to go, march, pai- to go,
8
uwa- to come (but neither pai- nor
uwa- take the clitic subject when used as serial verbs, 24.31, p. 324), par- to ee, take
ight, piddai- to run, ee, and arra- (m.-p.) to traverse, cross a distance, (2) stative
and change-of-state verbs such as akk- to die, arai- to get up, stand up, aanduli- to
encamp, go into garrison, - to be,
9
- (m.-p.) to sit (down),
10
ark- to perish,
die, get lost, arp(iya)- (m.-p.) to reassociate, change ones association, aik- to
7. Hittite intransitive verbs that take subject clitics and select the auxiliary e- correlate with the class
commonly labeled in linguistic literature unaccusative and those without subject clitics that select ark-
as auxiliary with the class termed unergative. For a full discussion of split intransitivity in Hittite see
Garrett 1996.
8. The inner accusative construction (16.18, p. 246) xsxi-an pai- to travel the way containing
the so-called accusative of the way (16.28, p. 249), behaves like an intransitive verb of perambulation in
taking the clitic subject pronoun: taknaat
d
tft-a xsxi-an paiddu Let it go the way of the Sungoddess
of the Earth KUB 17.10 iv 13 (OHlNS).
9. An apparent exception is: arku i.xr (var. i-e) ta he was an outstanding man KBo 3.34 ii
1112 (OHlNS), which lacks the subject clitic -a. Not an exception is: ka otu.x[n] [dau]' ta Lo,
there (once) was a [powerf ]ul bull KUB 31.4 + KBo 3.41:1718 (see 18.27, p. 282).
10. E.g., ne aa katta eanta They sit down beside the hearth StBoT 25 #54 iii 10 (OS). Exceptions:
ta eanda KBo 17.74 + KBo 21.25 + ABoT 8 i 18 (OHlMS, StBoT 12, 19), possibly [
d
Kuma]rbiakan
iyawaniyawanza
URU
Nipp[uri . . .] [. . .] pait iaarwantiyazaan i[ . . .] [. . .] eat KUB 33.120 i 4244
(MHlNS), nu gimr[i . . .] paikanta eanda KUB 53.15 ii 910.
18.11 Pronouns
00-HittiteGrammar.indb 280 6/5/08 4:18:42 PM
281
become satised with drink, igi- (m.-p.) to freeze, irmaliya- (m.-p.) to get sick,
ipai-lipiya- to become satised with food, iyanni- to set out, ki- to lie (down),
karuiya- to fall silent, ki- (m.-p.) to become, happen, kitanziya- to get hungry,
lzziya- do well, ourish, mai- to grow,
11
mayante- to grow up, mer- to disap-
pear, neya- (m.-p.) to turn, putkiya- to swell up, alliya- to melt, e- to lie down,
sleep, upp- to fall asleep, tiya- to take up a position, war- (m.-p.) to burn, we-
(m.-p.) to turn, change, zeya- (m.-p.) to become cooked, and the -e- and -e- en-
tives such as miyaunte- to grow old, mie- to become sweet, parkue()- to
become pure, alle- to grow up, (3) medio-passives which function as true passives
of transitive actives such as ai- to seal and wemiya- to nd,
12
(4) medio-passives
with (usually active) transitive counterparts, but occurring with the reexive clitic -z(a)
such as munni- to hide,
13
(5) verbs expressing emotions such as: araniya- to be
upset, but also to be upset with someone, envy someone, aar- to laugh or to
laugh at, kartimmiya- (m.-p.) to get angry, na- to become afraid and to fear (some-
thing), i- to belget angry but also belget angry at (both usually with -za), ulle- to
become wantonldisrespectful, takupike- to wail, lament, duke- to become happy,
rejoice but also to amuse, entertain, uwaya- to arouse pity, wike- to weep, wail.
In their transitive uses these verbs take no clitic subject.
18.15. Some intransitive verbs with clitic subjects fall into no clear category: aliya-
to kneel, e(n)k- to bow,
14
untariya- to fart.
Intransitive Verbs without Subject Clitics
18.16. Intransitive verbs used impersonally without referential subjects do not take
subject clitics: ppai it is nished, idlawezi it turns bad,
15
mai it prospers, ezi
it thrives,
16
and titai it thunders.
17
11. KUB 39.41 rev. 10, cited CHD LN 114b.
12. E.g., nammaat iyttari Then it is sealed KBo 25.163 v 6 and nat wemiyattaru Let it (cause of
the plague) be discovered! KUB 14.10 iv 19.
13. E.g., naza munnaittat He hid himself KUB 33.120 i 38.
14. This verb actually appears both with and without a clitic subject: na inga He bows KBo 20.11
iii 4 (OS) vs. +\+ itoi ekta a-ap-pa-e-a (i.e., ppaya) ekta He bows to the king, and he bows back
KBo 20.10+ i 56 (OS). Against Garrett (1996: 9798) these examples cannot be assigned to two dierent
verbs. For the variation in e(n)k- see 1.135 (p. 46).
15. nu ue[(zzi)] apdanipat t-i r-riipat id [(alawzi )] it will turn out badly for that person in
hislher own house KBo 3.67 iv 1314 with dupl. KUB 11.1 iv 2526 (OHlNS). See also: apedani t-i
kallarezi it will be unlucky for that person KUB 4.1 iv 29. As a change of state verb with referential sub-
ject, idlawe- does take a subject clitic: nti.xr-ermaat kdani memiyani But they became estranged
in this matter KUB 21.17 i 4 (NH).
16. E.g., nu mn mi ezi And when prosperity and abundance come KBo 3.7 i 67 (OHlNS) (for
the nature of this clause see 30.34, p. 415). Contrast the use of a subject clitic with mai- with a referential
subject: na mikewan di It (a vineyard) will begin to prosper KUB 12.44 ii 28.
17. orx-an ameanza ui-ri tetai When it becomes spring (and) it thunders KBo 2.7 rev. 1618.
Pronouns 18.16
00-HittiteGrammar.indb 281 6/5/08 4:18:43 PM
282
Other Clauses without Referential Subjects
18.17. No clitic subject pronoun regularly occurs in elliptical to be (-) clauses
with non-referential subjects such as urkil (it is) urkil Laws 189, 191, 195;
18

t artar (it is) no oense Laws 191, 193194 and 200; t ullatar (it is) not
(a case of) disrespect KBo 6.13 i 9 (Laws169, OHlNS); xtts-na watai (it is)
the womans sin KBo 6.26 iv 7 (Laws 197, OHlNS); alwanzatar o!\ itoi (it is)
sorcery; (it is) a case for the king Laws 111; or natta i.xr ,tot:txi\i (is it)
not your provisions-bearers? KBo 22.1 1718 (OS);
19
takku tma
A.A
.oXn dam-
mel pdan But if (it is) not a cultivated eld, (but it is) uncultivated land KBo 6.4 i
11 (Laws IV, NH); mn
LU
x.r.xxma But if (it is) a poor man KUB 7.1 ii 9,
KUB 7.54 iii 12; nan punu mn ki()an mn t ki()an Just ask him whether (it
is) so or if (it is) not so KUB 21.38 obv. 12 (NH); also existential sentences such as
arnikzil (There islwill be) compensation KBo 6.26 i 27 (Laws 163, OHlNS; with LH
131). Other similar one-word clauses may be cases of ellipsis of a nite verb (30.13,
p. 409). In the past tense the verb ta occurs: t ta was it not (so)? att. i 51. In-
stances where a subject clitic pronoun is used are rare: nuwaratmu allakartata[r]
kupiyatima[wara] t kuiki There was negligence on my part, but there was no
intent to defraud KUB 13.35+ iv 4344 (NH);
20
nu urorn.xr t [ekteni k]ulla
(6.2, p. 137)
21
dammea Or do you gods not know whose is the injury? KBo 4.8
iii 34 (NH). It appears that Hittite occasionally (so far only in NH) permitted a subject
clitic for a nonreferential subject, and that when this happened, it was marked by the
feature of gender attraction.
Special Cases
18.18. Finally, there are some verbs whose lack of a subject pronoun seems idiosyn-
cratic: aruwi- to bow (vs. e (n) k- to bow!), gimmantariya- to spend the winter,
18. Compare also nakku It is nakku in the Hittite laws, which shows no subject clitic marker. Note
also that takku ln-i KBo 6.4 i 1819 (Laws VII, NH; see ibid. i 26 [Laws IX]) must be translated (with
LH 22, 24) But if it (not he) is a slave, since without -a the subject is non-referential.
19. A borderline case, which may show the use of a subject clitic for a non-referential it subject is: [l
kui]ki tezzi itoi-a duddumili kardiyaa [iezzi n]at parkunumi takkuwat zi takkuwat t.oi
KUB 1.16 ii 5354. The clause [ pa-a]-a-a-nu-wa-an ta HKM 89:25 (MHlMS) is problematic on
several grounds.
20. The striking dierence in gender (-at to -a) in the clitic subject pronouns of this pair of tightly
joined clauses shows clearly that we have gender attraction to the two predicates.
21. We choose this analysis based on the preceding cited example where an enclitic subject pronoun
denitely is used for a nonreferential subject and shows gender attraction to the noun predicate. This solu-
tion does present a problem in that indirect speech requires that the content clause should be introduced
asyndetically, whereas geminate ll here would seem to point to the presence of the conjunction -a-l-ya-.
However, other such examples do occur (e.g., KBo 12.38 ii 6, cited in 7.1, p. 142), however they are to be
accounted for. An alternative solution for the present case would be to assume kulla dammea whose is
the injury with a degenitival pronominal adjective of the type established by Honer 2006.
18.17 Pronouns
00-HittiteGrammar.indb 282 6/5/08 4:18:45 PM
283
a- to open, uwapp- to behave badly toward, kururiya- to be hostile, make war,
nuntarnu- to hurry, alik- to approach, uwaya- to look at, wanu- to turn, change
(intrans.), wata- to sin.
22
18.19. A special problem is also posed by the verb ar- (m.-p.) to stand. Like e-
(m.-p.) to sit it belongs with the stative verbs that take the subject clitic, but it fails to
take the subject clitic in the Kikkuli hippological treatise (e.g., KUB 1.13+ iii 11, 21),
a text which in many other aspects of its grammar shows that it was not composed by
a native speaker of Hittite.
Demonstrative Pronouns
k-, ap-, and ai: Word Order
18.20. The demonstratives k- this (proximal, near the speaker), ap- that (me-
dial, near the addressee), and ai that (distal, far) can be used either as pronouns or
adjectives. If the clause is verbal, the pronoun may occupy whatever position in the
clause is appropriate to its syntactic function (subject, object, etc.).
18.21. In post-OH nominal sentences the pronoun ap or k normally precedes
its predicate noun or adjective: numu k ror-zi i-natarmit This was my rst heroic
deed att. ii 29 (KBo 3.6 ii 13); nuwanna k[( r)]-apat this one alone is our
lord KUB 26.12 + KUB 21.42 i 1718 (von Schuler 1957: 23).
18.22. In OH (OS), however, a dierent pattern applied, where the accented pronoun
subject followed the predicate (see 30.29, p. 414).
18.23. There is also a rare use of the demonstrative in an appositional construc-
tion, in which case the demonstrative immediately follows its noun: (Queen Puduepa
writes to Ramesses II of Egypt:) nu +\+ r. lump[ati ]n apn ui-mi Am I do-
ing that to My Brother (Ramesses) as a grief ? KUB 21.38 rev. 12 (NH), eds. Helck
1963; Stefanini 1964);
KUR
Alaiyanmazakan pedei [ ln-au]n arkammanaun,
arkammanmaikan [k]n iiyanun But I subjected Alaiya and made it tributary
on the spot; and as tribute I imposed the following (= kn) KBo 12.38 i 79 (NH);
ed. Gterbock 1967: 75, 77). (Because there are many herbal medicines, I will inquire
which herb is determined by oracle. I will also inquire about physicians) kuimu
LU
.
zt srs-ri nukan +\+
d
tft-st X ror.nr. apt parritti And as medicine the physi-
cian who is designated to me by oracle will apply that to His Majestys eyes KUB 22.61
i 1819 (NH), ed. Burde 1974: 4 (soll . . . selbiges Kraut streichen); it is not clear why
22. In view of the attested transitive active use of watanzi (if) they omit (so much as a thread) in KBo
16.47:8 the absence of a subject clitic with wata- to sin may reect that in Hittite terms it is a generically
used transitive without expressed object to make an omission, like d- to eat.
Pronouns 18.23
00-HittiteGrammar.indb 283 6/5/08 4:18:46 PM
284
the apt is separated from the in this clause; see ibid. i 6. It is unclear whether it is
signicant that our only attested examples come from late NH.
18.24. The unmarked position of the adjectival demonstrative is preceding its sub-
stantive. When that substantive is further modied by either an attributive adjective or a
noun in the genitive case, the latter intervenes between demonstrative and substantive.
18.25. With intervening genitive (1) with k-: k
URU
ina[u]wa
URU
Ubariyaa
[uttar l ] iyatteni Do not do this thing of the cities X and Y! KUB 1.16 ii 5051
(OHlNS); ki[a]nmamu kuit atre kuitmanwamu
d
tft-st k s+ i.xr
URU
Gaga
takula uttar atri ammugawa memian t\+ xtn
URU
Iupita ukemi But con-
cerning what you wrote me as follows: Until you, Your Majesty, write me about this
matter of the peace of the Kakaean men, I will await word in the land of Iupitta
HKM 10:2332 (MHlMS); kmu kuit s+
m
Marrwa i
URU
immuwa aliyatar atre
This homage of Marruwa, the ruler of immuwa, about which you wrote to me HKM
13:45 (MHlMS); kma namma s+ xtn
URU
attipat xtn.xtn.nr.-:t+ xtn
URU
Gaga
[n]at
LU.ME
srr n i i.xr irts ou er (var. eer) Moreover, the following
(lit., these) lands which belong to atti, the Kaka land they were swineherds
and weavers (Arawanna, Kalama, Lukka, etc., have declared themselves free from
the Sungoddess of Arinna) KUB 24.3++ ii 3839 (pr. of Murs. II); antutarma
kuit nu[i]an tya kuit uttar wemiyaweni nat [an ked ]ani linkiya tuppiya t
kitta[ri] Whatever matter we do not nd out about the population will not be placed on
this tablet of the oath KBo 5.3+ ii 6062 (MHlNS); kwa na uttar tuppiyaz au Just
consider (lit., see) this record of bloodshed KBo 3.1++ (THeth 11) ii 4748 (OHlNS);
(2) with ap-: apn s+ orurx ru-an parkunuer They puried that breadlfood of the
deceased KBo 11.1 rev. 11 (pr. of Muw. II); (3) with ai: nu uni s+
HUR.SAG
Aarpaya
URU
Gagan zaiyanun So I fought that Kakaean (group) of Mt. Aarpaya KBo
3.4 iii 41, ed. AM 7677; nuwamukan uni arazena xtn.xtn-a? i.xn peran
kuenni Strike down before me the aforementioned enemy of the surrounding lands
KBo 3.4 i 26, ed. AM 22.
18.26. With intervening attributive adjective or participle used attributively: k
600 rnr.xr these (grammatically sg.) 600 troops KUB 14.15 iv 48 (AM 74); see
kyata warn paur orx-an kitati just as this blazing re was extinguished (par-
ticiple warn) KBo 6.34 iv 56 (MHlNS), ed. Oettinger 1976a; kma idlauwa uddr
kuie anzi those who are performing these evil things KBo 3.1 ii 61; ammel kpat
1-a dammea This (is) my only punishment (for her); (in this way only I punished
her: just the fact that I expelled her from the palace) KBo 4.8 ii 1213, 16 (NH) ed.
Honer 1983a (where the number is used as attributive); k[( 3 i.xr t )]\ [+]
URU
Tanip-
iya a[(anzi)] these three men remain in Tanipiya KBo 3.7 iv 2021 (OHlNS); edani
pangawi i.[(xn)] that entire enemy (force) KBo 14.3 iv 29; nukan uni pankun si:t
kuenta he struck that entire tribe KUB 19.18 i 56 (NH).
18.24 Pronouns
00-HittiteGrammar.indb 284 6/5/08 4:18:47 PM
285
18.27. When, however, one or both of the words following the demonstrative is a
logogram (but see above in 18.25), divergence from this order is permitted: (1) geni-
tive follows head noun:
23
nuza k ix:+ s+ xi.or iyanun so I constructed this gold
statue of myself KBo 10.2 iii 21 (annals of Hatt. I, OHlNS); kma )irr+
I.A
iilla
kar aniyan ta these tablets of the treaty were already executed KBo 4.10 i 38, see
50 (treaty with Ulmiteub of Taruntaa); nu aptaya uttar s+
f
Danuepa iya[t k]ui
nuza apaa urorn-tt+-i kar kiat He too who committed that deed against (lit.,
of) Danuepa has already died (lit., become a god) KUB 21.19 + 1303lu (+) KUB
14.7 ii 1213 (NH); nu k )irri s+ ur.nr. apiya iyanzi They will seal these legal
documents (lit., these tablets of cases) there KBo 3.3+ iv 1213 (NH); nuza maan
eni )irr+ s+ xtn
URU
Mizri peran wemiyanun When I found before myself that tablet of
(i.e., about) the land of Egypt KUB 14.8 i 3132 (NH); ai xtts.itoi
URU
uruma
That queen of uruma KBo 3.28:21. (2) Attributive adjective follows head noun:
numu k xtn.xtn.xr dannatta +ssi+ +ituoi::t+ peta she (the goddess) gave me
these depopulated lands to govern att. ii 56, see lines 6364.
18.28. The demonstrative naturally can also be followed by a substantivized ad-
jective or a free-standing genitive (16.61, p. 256): [kuw]atmu k tepu paitta why
did you give me this little bit? KUB 1.16 iii 10 (OHlNS); knmawaza s+ 2 oln.
xrsi kuwat un why have I borne this two-legged (child) (lit., this one of its
two legs)? KUB 24.7 iii 2324 (story of cow and sherman); kpat s+
f
Tunnawiya
xtts t.or +:t this (ritual) of Tunnawiya, the Old Woman, is nished KUB 7.53
iv 7 (colophon).
Declinable k . . . k as a Distributive Expression
18.29. As an equivalent to the English distributive this . . . and that Hittite coor-
dinates either k . . . kma (. . . kma) or ka . . . ka (. . . ka). We have
arbitrarily cited the nominative common singular, but correlated forms can show any
case, gender, or number. Examples: ka . . . ka this one . . . that one StBoT
8 ii 3536 (OS); k
GIS
ftxti-limet kma aa(n)met this is my craftsmans ef,
and that is my aan eld Laws 40 (OH); kl . . . klla this ones . . . that ones
(OH); kedani . . . kedaniya to this one . . . to that one. The distributive k- also occurs
in the expression kez . . . kezziya on this side . . . and on that side, or in kian . . .
kiana in this way and in that way. Notice that Hittite sometimes views such pairs as
contrastive, marking the second member with -al-ma (k . . . kma) and sometimes as
coordinated, using -al-ya (ka . . . ka, kez . . . kezziya, etc.). Note that although
(especially with the use of -al-ma) the pairing has the function of contrast, the same
(near) demonstrative is used for both members. Similar in force is kuie . . . kuie (pl.
nom.) some . . . others.
23. In the following examples the presence of logograms either as head noun or genitive or both renders
the Hittite word order less certain.
Pronouns 18.29
00-HittiteGrammar.indb 285 6/5/08 4:18:49 PM
286
18.30. Sometimes k k (or even simple k) is used to represent a name to be
supplied: [itoi-ik]an k k itarningain rorn-pa d So-and-so has taken back
[from the king(?)] the illness KUB 29.1 ii 32 (rit., OHlNS); kwarati itoi-u
pai (If from the palace they give him [i.e., a temple servant] silver, gold, . . . , let it be
enumeratedlspecied:) King So-and-so (lit., this king) gave it to him KUB 13.4 ii
34 (pre-NHlNS), cf. CHD lamniya- 2b. Note that k k takes a singular verb. On the
other hand, k ka this one and that one takes a plural verb: When they gave it
to him, nuwa k ka arantat so-and-so and so-and-so were present KUB 13.4+
ii 3738 (instructions for temple ocials, ed. Sturtevant and Bechtel 1935). The latter
construction also occurs in the accusative: kya kya iyami I am about to do such-and-
such KBo 3.7 i 22 (Illuyanka, OHlNS).
Correlated k-Forms Having Dierent Cases
18.31. When correlated k-forms have dierent cases, this is sometimes used (like
ara ari and ara aran or 1-a 1-an
24
) to express reciprocal action (one . . . the other):
kman kn epzi one might seize the other Laws 49; [lmazaka]n k kn rorn-
pan azketta let one [not] . . . the other! KUB 1.16 ii 58 (OHlNS).
Other Deictic Elements
18.32. Although the primary demonstrative pronouns of Hittite are k-, ap-, and
ai, they are not the only deictic elements in Hittite speech. The contrasting inseparable
prexes u- and p(e)- on verbs of movement (uda-, peda-, uwate-, peute-, ue-, pai-) also
indicate movement relative to the speaker (the I): u- toward the speaker and pe- away.
See 12.21 (p. 198, uwate-, peute-), 12.41 (p. 209, ue-, pai-).
The Indefnite Pronoun kuiki
18.33. The constructions with kuiki are formed as follows. When kuiki is used
substantivally, it tends to take a position quite close to the nite verb (i.e., toward the
end of the clause): nata uttar [l k]uiki arratta let no one transgress the command
KUB 1.16 ii 4950; nau urorn-tt+-ni kuiki peran wati or if someone (kuiki) sins
before a deity (urorn-tt+-ni peran) KUB 1.16 iii 60 (OHlNS) (note how kuiki inter-
rupts urorn-tt+-ni peran); aanna(n)anzakan le kuinki kuenti of his clan do
not kill anyone KBo 3.1 ii 45 (OHlNS); nui i.xn zaiya menaanda namma
t kuiki mazzata no enemy (i.xn . . . t kuiki ) dared any longer (to go) against
him in battle D frag. 28, A i 78 (NH); nu mn kian kuiki memai And if someone
speaks thus KUB 1.4 iii 40(28) (= att. iii 7374; dupls. A and F read kui instead
of kuiki ); aaniya luzzi l kuiki pz[(i)] Let no one (l kuiki ) seize them for land
tax (or) for corvee att. iv 85.
24. For reciprocals expressed with correlations of the numeral one: 1-a 1-an (or 1-edani, 1-edaz,
etc.), see 9.9 (p. 155).
18.30 Pronouns
00-HittiteGrammar.indb 286 6/5/08 4:18:50 PM
287
18.34. In some cases the form of substantival kuiki even follows the nite verb at
the end of the clause: [takku i-a]n nama xtts-an itt++ walzi kui[k]i na aki If
anyone strikes a free man or woman, so that slhe dies KBo 6.3 i 6 (Laws 3, OHlNS);
[n]amama rorn-ziaz itamazi kuiki kuitki t\+ r.oi-tt+maat t memai or if
anyone of you subsequently hears anything, but doesnt tell it to the palace KUB 21.42
+ i 2627; mnna +\+ txt
md
ixx nakkizi kuitki And if anything weighs upon
the progeny of Kurunta BrTabl. ii 7576 (Tud. IV); mn tukma
m
Kupanta-
d
ixx-
an waggariyazzi kuiki if anyone revolts against you, Kupanta-ixx KUB 19.54
iv! 8 (Murs. II treaty with Kupanta-ixx); nammamu mn uppi kuiki mnmu t
kuiki uppi numu )irr+
I.A
atrtten nu igallu Furthermore, write me letters and let
me know whether someone will send (oil) to me or no one will send (it) to me KBo
18.2 rev. 812 (NH).
18.35. When kuiki is used as an attributive adjective and the modied noun has
no other modifying adjective, kuiki immediately follows the noun: numa ardiya
kuiki paizzi And some supporter goes to them KBo 6.3 ii 31 (Laws 38, OHlNS);
numu k r.xtn.n.xr kuiki kattan (There is) some chariotry (lit., horses) here
with me HKM 30:89 (MHlMS); mn ot\i++ kuitki allezi But if some legal case
becomes (too) large KBo 3.3 iii 2930 (NH); mn utxt.itoi kuiki watai If any
prince sins KBo 3.1+ ii 59 (OHlNS); see also HKM 46:1516 for a kuinki example.
18.36. When the attributive adjective is tamai-, kuiki follows it directly and pre-
cedes the head noun: +\+ xtn-:tman tamin kuinki r-an iyanun KBo 5.13 i 2122
(see also KBo 4.3 i 11); see also KBo 16.97 rev. 32 (MHlMS) tamama kuiki
d
ts:+u
some other Is:+u vs. rev. 30 Is:+u of my father; mna apel s+ xtts-:t utxt-a
mna taml kulqa xtts-a utxt-a Whether he is the child of that woman or he
is the child of some other woman BrTabl. ii 9092 (Tud. IV), but this also occurs oc-
casionally without tamai-: nti-lunma kuinki rrx KUB 21.5+ iii 1; tuel kuinki ln-:i+
KUB 14.3 iv 47; kuiki idlu memia[ s ]+ ni KBo 5.4 obv. 10. See also KBo 5.13 i
36 apailama kuiki tnt-a.
tamai-
18.37. When used adjectivally, tamai- another (for pronominal forms in its declen-
sion see 5.2, p. 132; for the paradigm 8.10, p. 152) like most other adjectives precedes
its noun: tamai utne another land, tamai i another lord, tamai pedan another
place, tamai kupiyati another plot, tamai tnt-a another city. Occasionally, when
its head noun is a logogram without Hittite phonetic complement, tamai- as attribu-
tive adjective follows its noun: maanmaan itoi-u +\+ xsxi.oi par ari
GI
orornma tamai triyan antan but when the king comes forth upon the Great Road,
another chariot is harnessed (and) ready KUB 10.18 i 2426;
LU.ME
i.zt
9
tamae
fo oi. waan arkanzi the other performers wear brightly colored garments KBo
4.9 i 4546; ru.xtn
4
.n damau in KUB 7.5 ii 27.
Pronouns 18.37
00-HittiteGrammar.indb 287 6/5/08 4:18:51 PM
288
18.38. When used substantivally, tamai- means another (one) and is construed
within its clause just like any ordinary noun: itarna tu.nr.-tima ru.xtn
4
.n
damau pariyannai nekuz meurra damau pariyannai at midday I break other
loaves, and in the evening I break (still) others KUB 7.5 ii 2630; takku utxt.xtts
i-ni taranza tamaiaan (or tamaian) pittenuzzi If a young woman is promised
(in marriage) to a man, and another (man) runs o with her KBo 6.3 ii 57 (Laws
28a, OHlNS); kur-zu-na-a (var. lam-na-a) [ito]i-u aki
GI
ot.zsimazakan
tamai(!) eari a . . . king will die, and another will sit down on his throne KUB 8.1
ii 1819 (omens, OHlNS).
18.38 Pronouns
00-HittiteGrammar.indb 288 6/5/08 4:18:51 PM
289
Chapter 19
ADVERBS
19.1. Whereas an adjective is a word that qualies a noun, an adverb is one that
qualies a verb, an adjective, or another adverb. The italicized words in the following
examples are English adverbs: he walks slowly, the day has ended too soon, she is
less talkative than her neighbor.
19.2. Interrogative adverbs can serve as conjunctions introducing subordinate clauses
(as interrogatives they can also introduce indirect questions; see 30.68., pp. 427.):
kuwapi whenlwhen? or wherelwhere?, maan when, as soon aslwhen? or asl
how?, kuan when (rare).
19.3. Adverbs can be derived from other parts of speech by the addition of a sux
or by the use of a particular noun case (see 16.30, p. 249). On adverbs formed on a
demonstrative base see 7.18 (p. 147).
19.4. In the examples cited below note that the same adverb may be assigned to
several categories. For example, apiya is listed below in local (there) and temporal
(then) meanings, as is kuwapi where, when.
Local Adverbs
19.5. Local Adverbs: k here, hither (OH kni), kt on this side,
1
kz from
here, on this side (often in correlation kz . . . kzziya on this side . . . on that side),
apiya there, thither, a-pd-da, a-pd-da-an (post-OH) there, thither, to that place,
2

apz from that place,
3
edez on that side of, kuwapi where, whither, kuwapiya
1. takku kt (var. B kz) u-az 2 o xi.nnnn pi takku edi u-az nue 3 o xi.nnnn pi If (it
happens) on this side of the river, he shall pay two shekels of silver; if on that side of the river, he shall give
him three shekels of silver KBo 6.2 i 4950 (Laws 22, OS); [(toti
LU.ME
xtniurx)] a katta kt
arta The foreman of the cooks stands next to the replace on this side KBo 17.15 obv. 1415.
2. apda (OHlNS) occurs securely as therefore (see 19.8 with p. 291, n. 7) and possibly once in NH
broken context (KUB 6.48 ii 3, see SV 2:148, cited HW
2
A 180) as thither. We read the forms with the
ptlpt sign as -pd- based on the contrast da-ma-at-ta vs. da-me-e-da to another (place), which appears
to show that the forms with geminate ddltt had a vocalism.
3. kinuna apz d nu ap txt.nr. anniya Now take from there and sow those seeds HKM 54:25
26; [m]n
SG
maitana maiwantan watanzi [n]u
d
tft-st kezza zaiyami ziku apezza zaiyai If
they fail (to give) so much as a tiny bit of wool, I, My Majesty, will ght them from this side, and you shall
ght them from that side KBo 16.47:89.
00-HittiteGrammar.indb 289 6/5/08 4:18:52 PM
290
everywhere;
4
kuwapit where, whither; kuwapitta everywhere; kuwatta wherever;
dameda elsewhere; 1-eda alone; anda(n) inside, within; andurza on the inside;
from inside; araza on the outside; from outside, arazanda around, about, ara
away; kaz outside; menaanda opposite, against (usually a preverb or postposi-
tion, not an adverb); par further, in addition; peran in front; er above; ar up-
wards; katta down; kattan below; kattanta downward; tapua, tapuza alongside,
laterally, to the side; arayan separately; twa and twaz at a distance, remotely;
maninkuwan nearby, in the vicinity; duwan . . . duwan here . . . there; in this direction
. . . in that (the opposite) direction. On the possible derivation of local adverbs in -a
from old allatives see 3.31 (p. 76). For more on the use of local adverbs such as katta
down and kattan below see chapter 20.
Temporal Adverbs
19.6. Temporal Adverbs: mn when (OH), maan when (NH); kuwapi when
(8.9, p. 151; 19.2, p. 289; 30.34, p. 415); kuitman while (30.37, p. 416); kuan
when, as soon as; kinun now; apiya then; kuwapi(k)ki anywhere, anytime; natta
kuwapikki nowhere, never; kuwapiya always, whenever; kar formerly, long ago,
already; annian formerly; annaz(a) at one time, formerly (7.17, p. 146); tapeni
previously(?); lukkatta on the morrow; zilatiya, ziladuwa (Luwian loanword) subse-
quently, in the future; kattama later on, afterwards; namma next, nally; nwa still,
yet; nawi not yet; peran par provisionally, previously; duwan par until now;
dk immediately, suddenly; lammar or lamn instantly, immediately;
5
lammar
lammar at any moment; 1-anki immediately, at once; pareni quickly; kuanka
anytime, whenever; it kuanka never; appan (rorn-an) and appanda afterwards
(both can be written rornsi ); appezziyan or appezzin afterwards; appaiwatti (rorn-
pa-tu(.xx)-ti) in the future; karuwariwar early in the morning; andaii meuni
in the heat of the day, at noon; nekuz meur in the evening; mitiliya meur (a time
of day); ipandaz at night; tu.xx-az in the daytime.
19.7. The Hittite manner of referencing temporal relationships using terms of spatial
reference is a complicated subject in its own right (see Honer 2002). Does the speaker
face the future, with the past behind him, or does he face the past, which he can and
does know, with the yet unknown future behind him? The second theory is supported
by the normal use of peran for before and ppan after, and appaiwatt- (lit., after-
day) for future. Also antezzi (ya)- older, earlier-born (< ant- front, forehead)
versus appezzi(ya)- younger, later-born. Yet there are a few expressions suggesting
4. menaanda[(yaa)n k(urakki)] tapuwa zo-ni oin-li nu kuwapiya +:++++ 4 wallu[ di] Op-
posite the pillar, alongside, on the right, on the left everywhere in the same way [he deposits] four
wallu KBo 4.1 rev. 34. The term kuwapiya also means whenever; see 19.6.
5. nan ++n+u
d
[tft-st ] lammar uwada[ndu] Let them conduct him immediately to My Majesty
HKM 12 rev. 35.
19.6 Adverbs
00-HittiteGrammar.indb 290 6/5/08 4:18:53 PM
291
the opposite, such as -mu peran before me KUB 13.4 ii 5859 (MHlNS) in the sense
of tasks awaiting my attention in the future.
Adverbs of Manner
19.8. Adverbs of Kind and Manner and Causality: kian thus, as follows;
6

apenian thus, as just seen or said; enian thus, in the previously mentioned man-
ner (see 7.18, p. 147); apda (OHlNS), a-pd-da, a-pd-da-an for that reason, there-
fore;
7
kuwat or kuit why?, kuwatka somehow, perhaps, it kuwatka by no means,
in no way, arumma highly, exceedingly, imma actually, really, katta accordingly,
subsequently, kari really, truly, genuinely, maan how? as, manka (usually it
manka) in some way, in any way, kuwatta kuwatta in every respect.
8
One occurs in
both main and subordinate (i.e., dependent) clauses: maan as, how. For additional
manner adverbs see 16.30 (p. 249) and 19.1319.14 (p. 292)
19.9. Adverbial phrases of manner are formed with the postpositional iwar in the
manner of, like, which governs the noun preceding it in the genitive case (
LU
sor-a
iwar in the manner of a cupbearer) (see Honer 1993 and 16.59, p. 255), and the post-
positive conjunctions mn (CHD mn mng. 1) and maan like (see 20.32, p. 301).
19.10. Distributive adverbs (24.12, p. 320; 28.119, p. 384) are formed by repetition:
usually dative-locative, ilani ilani gradually, step by step, by degrees; uddan uddan
word by word; witti witti or xt(.xx)-ti xt(.xx)-ti year by year; rft(.xx)-mi
rft(.xx)-mi month by month; tu(.xx)-at tu(.xx)-at day by day; or
6
-ti or
6
-ti
night by night; lammar lammar
9
moment by moment, at any moment; xsxi-i
xsxi-i trip by trip; but also other cases: xrsin-az xrsin-az from each threshing
oor KBo 16.72:5; zo.on.n zo.on.n altar after altar; 1-an 1-an one by one
(acc.); etc. See also r.urorn-tt+ r.urorn-tt+ Bo 6251 obv. 3, and gipeni gipe [ni]
KUB 45.78 iv? 3. For the use of distributives of time words with -ke- verbs see 24.12
(p. 320). This construction must be carefully distinguished from the use of the repeated
vocative (or direct address form) such as lli lli (= rxr.nr. rxr.nr.) O tongues!
Tongues! (see p. 81, n. 25).
Denominative Adverbs
19.11. As denominative adverbs one nds either (1) frozen case-forms of nouns
or (2) frozen case forms of adjectives, including derivatives with the productive sux
-ili.
6. For kian in interrogative clauses see 27.3 (p. 348) and 30.69 and 30.71 (p. 427).
7. See kuit anda . . . apda why? . . . for that reason KBo 3.41 + KUB 31.4 obv. 1619, cited HW
2
A
180. The meaning thither for this word is not assured (see 19.5, p. 289).
8.
m
Akaliya
URU
urmi eta apa kuwatta kuwatta i.xr (var. i-e) ta Askaliya lived (lit.,
was) in Hurma, and he was a man in every respect CTH 8 (A = KBo 3.34) ii 89.
9. We follow the CHD (s.v. lammar) in taking these forms as endingless dative-locatives.
Adverbs 19.11
00-HittiteGrammar.indb 291 6/5/08 4:18:54 PM
292
19.12. Examples of frozen case-forms of nouns are: the accusatives karuwariwar
in the morning, nekuz meur in the evening; the dative-locative forms auli in a
friendly manner, pareni in haste, tapeni at the rst; the ablatives of direction
kunnaz (zo-az) on the right, ikiaz backwards, antezziyaz forwards (on these see
16.92, p. 265), twaz at a distance, tapuza to the side, iwattaz (tu.xx-az) in the
daytime, ipandaz (or
6
.xx-az) at night.
19.13. The neuter nominative-accusative of adjectives can serve as an adverb: (1) in
the singular: tepu somewhat, a little, a bit, mekki very, greatly, antezzi (at) rst (op-
portunity) (16.30, p. 249), kari in i.xn kari zaiyadduma[t] reallyltruly ght
the enemy! KUB 13.20 i 19; (2) in the plural: atuga (< *atugaya) in a terrifying
manner, munnanda secretly, hidden. The ablative annaz of the demonstrative stem
*anna- that, yonder is used as an adverb formerly.
19.14. Neuter accusatives of pronouns and numerical adjectives can also serve an ad-
verbial function (HE 199b regarded these as inner objects): tuk t kuitki idlawaun
I harmed you in no way KBo 5.13 i 23 (Mur. II); aptta[ya] [\t ]s urorn-tt+ arratti
in that way too you transgress the oath KBo 5.4 rev. 2324 (Mur. II); kyan 1-an
dammeanun (only) in this one way did I harm her KBo 4.8 ii 13.
The Adverbial Sufx -ili
19.15. A productive denominative adverbial sux is -ili (originally plural nomi-
native-accusative neuter of adjectives such as karuili- old). It is suxed directly
to consonantal stems: pittiyantili like a fugitive (pittiyant-); karuiyantili quietly
(karuiyant-); mayantili like a young man; xt.xx-li or xt-tili (witt-) annually,
yearly; tu.xx-tili (iwatt-) daily; aranili like an eagle (aran-), swiftly; xtts-
nili of a womanly nature (xtts-n-). The sux elides the a of a-stem nouns: nilaili
in the (speech) manner of (the city of) Nea;
URU
urlili in the Hurrian language
(urla-); luwili (without determinative) in the Luwian language (<
URU
Luwiya). Since
we lack a good example of an adverb based on a u-stem, it is unclear how -ili was
suxed to such nouns, although probably the u was not elided. One assumes that it
would have been analogous to the formation of the adjective karili- primordial from
the adverb kar long ago and the adjective taruili- valiant. When based on geo-
graphic names, -ili often denotes the language (or some other cultural characteristic)
of that region:
URU
niili,

naili, and neumnili (the latter two without determinative)
in the language of the city of Nea, in Hittite;
URU
attili in the (pre-IE) language
of the Hattians, in Hattic; luwili (without determinative) in the Luwian language
(<
URU
Luwiya);
URU
palaumnili in the language of the land of Pala, in Palaic;
URU
urlili
in Hurrian (< urla-),
URU
pabilili in Akkadian; laanili in the modelmanner of the
city Laana (a dance style); uppianili in the modelmanner of the city ubina
(a dance style). For a full list of the adverbs in -ili (without translations) see Reichert
1963: 120 and Jie 1994: 2728.
19.12 Adverbs
00-HittiteGrammar.indb 292 6/5/08 4:18:55 PM
293
Adverbial Circumlocutions
19.16. For cases where a nite verb in a construction with an innitive describes the
manner of action see 25.34 (p. 337).
19.17. For cases where an innitive describes the manner of action of a nite verb
see 25.35 (p. 337).
19.18. An adjective (or participle) in the nominative (appositional to the subject) can
function like an adverb in English translation. See 15.6 (p. 237).
Adverbs 19.18
00-HittiteGrammar.indb 293 6/5/08 4:18:55 PM
294
Chapter 20
LOCAL ADVERBS, PREVERBS, AND POSTPOSITIONS
20.1. Hittite has a number of words expressing various spatial relationships, such as
up, down, in(to), out, beside, away. With a few exceptions, these occur in a
variety of syntactic roles. One may compare for this exibility the uses of English up:
I found the enemy up in the citadel, I climbed up the stairs, I threw up my hands.
In the rst example up is a free-standing adverb that describes where the action of the
predicate takes place. In the second, it is a preposition forming a phrase with the stairs.
In the third it is a verbal particle that delimits the meaning of the verb (the latter being
typically a verb of motion). It is also commonplace for the verbal particles to occur in
combinations with verbs in an evolved, non-spatial sense: I turned up the heat, I tore
up the letter. Hittite has equivalents of all of these uses. The challenge is to determine
which are which. In what follows we draw upon the results of Zuntz 1936; Goetze 1963;
Starke 1977; Boley 1985; Tjerkstra 1999; and Francia 2002b.
20.2. We follow the CHD P 10930 (sub par) and Francia (2002b: 29) against
Starke (1977: 12731
1
) in assuming that most Hittite locallspatial adverbs occur in
three distinct syntactic roles: (1) as free-standing adverbs; (2) as preverbs; (3) as post-
positions. As per Tjerkstra (1999: 15875) and Francia (2002b 6, 214), one cannot
distinguish these dierent uses solely on the basis of word order; one must also use
functional criteria. Inevitably, some attested examples are ambiguous and open to com-
peting interpretations. The existence of such examples does not refute the basic claim
of a three-way contrast. In what follows we purposely cite only enough unambiguous
examples to establish each of the three usages. For more complicated and problematic
cases see the works cited in 20.1 and CHD sub par, pran, and ar.
20.3. Starke (1977: 13187) established (based on OS evidence) that in Old Hittite
there was a contrast between two paired sets of adverbs, one of which expressed place
toltoward which and one place where:
2
Those of Set 1 are more common as preverbs,
and those of Set 2 as postpositions. But this is a tendency, not a rule.
1. For some arguments against Starkes claim of a single syntactic category of adverb see Salisbury
1999: 6165.
2. We prefer to avoid the labels directive or allative for Set 1 and locatival for Set 2. We also take
no stand on the question of the origin(s) of this system. On the latter see among others Laroche 1970: 3541;
Neu 1974: 5961; and Starke 1977: 132.
00-HittiteGrammar.indb 294 6/5/08 4:18:56 PM
295
Set 1: Place to Which Set 2: Place Where
anda into andan within, inside
ppa back ppan behind, after
katta down
3
kattan below
3

par out, forth pran in front, before
ar up(ward) er above, over
3 Chap. 20, new n. e
20.4. It is important to recognize the use of the place where set to express a loca-
tion as the result of movement: [utx]t.xr itoi pnzi i.xr +isiot-an ppan
tienzi The princes proceed to station themselves behind the bodyguard KBo 20.12+
i 34 (OS) (Salisbury 1999: 64 and Francia 2002b: 79, 57 and 68; overlooked by
Starke 1977). This system mostly remains intact in New Hittite (Francia 2002b, but see
Salisbury 1999: 64 on a limited change). There is a strong tendency for Set 1 to be used
mostly as preverbs and Set 2 most often as postpositions, but at least in New Hittite there
are exceptions to this pattern. Both sets function as free-standing adverbs. While many
questions of detail remain to be answered, certain main facts seem clear.
Free-standing Adverbs
20.5. The following examples of syntactically free-standing local adverbs are merely
illustrative and do not exhaust the possible dierent meanings of each: andamakan
tut-un kuwpi kuewn nu linkiya kattan kian daiwen in addition when we had killed
a sheep (as part of the treaty-making ceremony), we placed the following (obligations)
under oath KBo 16.47:1516 (MHlNS); takku r.t.u andan t.[oi] if there
[is] no straw inside KBo 6.2 iv 61 (Laws 100, OS); ppama itoi
URU
[atti ] . . .
ullanun But afterwards . . . I struck the king of attua KBo 3.22 1415 (OS); rorn-
anmaa irmaliyattatpat But afterwards he in fact became ill KBo 3.4 i 56 (NH);
nama katta utxt.xr
d
tft-st kuiki waggariyazi Or if subsequently someone revolts
against the sons of His Majesty KBo 5.13 iii 13 (NH); ta or-ru kattan And a tree (is)
below KBo 17.1 iv 16 (OS); takku .X.nr. kulla par wi If he buys in addition
someone (else)s land Laws XXXIX (OHlNS); nui pargatar peran 1 tt+ 9 +iya
u[] At rst his height was 1900 us KBo 26.65 i 17 (Ullik. III); arma or-
ru azata But above, the tree dried up KUB 12.62 obv. 8 (pre-NH); nukan gurtan
er wetenun And up there I built a citadel KBo 4.4 ii 6162 (NH).
Preverbs
20.6. In many Indo-European languages the basic meaning of the simple verbal
stem is modied by the addition of various prexes. In some languages these prexes
3. With Starke (1977: 18187) and against Tjerkstra (1999), Francia (2002b), and others, one must
distinguish a separate kattalkattanlkatti- which functions solely as a postposition meaning beside, next to,
with (see 20.21, p. 299).
Local Adverbs, Preverbs, and Postpositions 20.6
00-HittiteGrammar.indb 295 6/5/08 4:18:57 PM
296
occasionally appear separated from the verb. In Hittite most of these elements are freely
separable and are thus labeled preverbs.
20.7. In Hittite only two truly inseparable preverbs exist: u- here and pe- there,
attested in the pairs:
4
uda- to carry here and peda- to carry away; uiya- to send here
and peya- to send away; unna- to drive here and penna- to drive away; uwate- to
lead here and peute- to lead away; which are always written as single words. An
exception is p in p ar(k)- to bring, present, in which the preverb is clearly separable:
ppat arkanzi they shall also present (it). On the particle -pat see chapter 28. All
other Hittite preverbs are separable. More research is needed on the word order of Hit-
tite preverbs, but with few exceptions they occur: (1) immediately preceding the verb;
(2) separated from the verb by one element (usually a negation, indenite pronoun, or
locatival phrase); (3) fronted to clause-initial position.
5
20.8. The following examples of preverbs are merely illustrative and do not exhaust
the possible dierent meanings of each: andakan alna teummiu tarlipit wamu
2-:++ ptumini We bring in clay vessels twice(?)-lled with tarlipa-liquid KBo 17.1
i 2627 (OS); lukkattama t\+
URU
Iyaria rotu-pa uwanun But on the next day I
came back to Iyaria KBo 4.4 iii 40; r:+wakan edaza ara aru May My Lord
get away from there KBo 4.14 iii 34; nukan x.n.xr katta ur natmu oln.nr.-
a k[(att)]an liyanda[(t)] The colonists came down, and they prostrated themselves
at my feet KUB 14.16 iii 16 restored from KUB 14.15 iii 4647 (AM); nankan t\+
xtn
URU
Nuai par neun And I sent him out to the land of Nuhai KBo 4.4 i 40
(AM); nan par uittiyanun nan r-t++ iyanun I picked him out and made him a
lord KBo 16.17 + KBo 2.5 iii 25; itoi-i par 1si paizzi He goes forward to the
king one time KBo 17.43 i 1112 (OS); tnt-rimaakan ar l uezzi Let him not
come up into the city! KUB 21.29 ii 12; less common: nu utxt.xrsi andan zikt
u u-a tarna She put her sons inside, and released them to the river KBo 22.2 obv.
3 (OS); alkimam a apiya aniyanza kuit nu rotu-an tiyatten because grain has
been sown for you there, get behind (the matter) HKM 18 left edge 23 (MHlMS); nu
s+ utxt-ut kattan t,u+: He undertook (concerned himself with) (the matter) of a son
KBo 5.6 iv 15 (D, frag. 28), nunna uezzi [ap]pat peran uwkeuwan di he
himself proceeds to begin running in front of us KBo 3.40:89; erpat aruwnzi (the
king and queen) do reverence upward KBo 17.74+ iii 19 (OHlMS);
URU
amuanna
tnt-ti+ urorn-tt+ alwanzenaza er unnita And he lled up amua, the city of
the goddess, with sorcery att. ii 7879 (NH).
20.9. The force of each preverb can be best detected through a study of the various
verbs with which it combines and the resulting meanings of each combination. Espe-
4. This opposition corresponds neatly to German her and hin.
5. We nd the use of the terms initial position and rst position in the literature on Hittite word
order confusing and prefer to use the unambiguous terms clause-initial (i.e., absolute initial position in the
clause) and following clause-initial conjunction and attached clitics.
20.7 Local Adverbs, Preverbs, and Postpositions
00-HittiteGrammar.indb 296 6/5/08 4:18:58 PM
297
cially helpful in this respect are verbs of perambulation or motion: iya- to go; pai- to
go; uwa- to come; unna- and penna- to drive herelthere; ar- to arrive at, reach;
tiya- to step, enter; piddai- to run; uwai- to run, ee; nai- to turn, go; etc. It can
be seen from the above list that, even without recourse to preverbs, the Hittite language
possessed a stock of verbs of motion which distinguished direction of movement. The
preverbs rened the directional aspect inherent in the verb. Understanding the use of the
preverbs is complicated by the fact that there are changes in their use from OH to NH
(for the pair anda and andan see Salisbury 1999). These changes mean that the usage
in later copies of earlier texts may be inconsistent. Further study is needed.
20.10. Two preverbs can occur with a single verb: appan anda pai-, er ara da-,
appa ar da-, peran katta, itarna ara. In some cases the rst of two adverblpreverbs
indicates the starting point of the motion (e.g., er katta down from above), while in
others it merely species the direction indicated by the second ( peran katta down in
front of). Limited to the rst position in such a chain of two preverbs is awan, which
occurs in awan ara, awan katta, and awan ar. Since awan does not occur without
another preverb following it, some believe that awan has no special meaning of its
own but merely intensies the following preverb (HW
2
, HED). Melchert (1996: 135)
proposed that awan ara pai- meant to go past (the side of). For a wide range of these
constructions with tentative translations for each, see CHD P 1819 (table of contents
of pai- article).
Postpositions
20.11. Instead of prepositions Hittite has postpositions, called place words by
some. They serve the same function as the prepositions of the other Indo-European
languages: they lend greater precision to the local or temporal distinctions provided by
the case-forms of the noun. Most of the words that function as postpositions in Hittite
are the same as those that function independently as local adverbs: er, peran, katta(n),
etc. There is broad agreement that when a local adverb co-occurs with a preceding noun
or pronoun in the genitive (in OH), the adverb is a postposition (see examples below in
20.16, p. 298, and following; and see 16.35, p. 250). We follow the majority view that
in many, if not all, cases where the dative-locative is followed by a local adverb express-
ing location, the latter is also a postposition (examples below); but there are dissenting
opinions (see Starke 1977: 17277). Whether the local adverbs expressing directionl
place to which (anda into, ppa, ara, katta down, par, ar) function as postposi-
tions in combination with a preceding noun is a much-debated question on which even
the present authors do not fully agree. Examples of this sort cited below (e.g., all those
in 20.15, p. 298) are thus interpreted dierently by individual scholars. Fortunately, for
understanding the meaning of the texts it usually does not matter whether we assume
that the local adverb in such cases is a postposition or an independent adverb. One
should, however, note the dierence between ntn.so-i er on top of the mountain
and t\+
URU
Puranda er uplabove in Puranda (not on top of P.!).
Local Adverbs, Preverbs, and Postpositions 20.11
00-HittiteGrammar.indb 297 6/5/08 4:18:59 PM
298
20.12. Often the bare case-form without postposition expresses the spatial relation-
ship which we would indicate in English with a prepositional phrase: r-ri (*peri ) or r-ni
( parni ) in the house, parnaz out of (or from) the house. See chapter 16.
20.13. In answer to the questions in what place? (kuwapi or kuedani pedi ) or at
what time? (kuwapi ) the Hittites used the dative-locative case of a place or time word,
often in combination with a postposition.
20.14. The following postpositions govern the dative-locative, at least in Middle and
New Hittite (see 16.86, p. 264). Some governed the genitive in Old Hittite (16.58
16.59, p. 255).
20.15. anda or andan in, into: r-ri anda(n) in the house; xtn-e anda uet he came
into the land; auwa anda among (his) possessionslgoods; aawa antua anda
among good people;
UR.SAG
apadduini anda in Mt. .; xtn kurur anda into an
enemy land; ulin anda into the clay; kiari anda in the hand; xsxi-i anda on
the road; kuedanikki anda itama- to hear fromlvia someone; tuekkii anda on
his body; aanduli anda in garrison.
20.16. ppan (spatially) behind, (temporally) after:

tarnui ppan behind the

tarnu-building, tnt-ri rorn-an behind the city, +\+ xtts-:t unuwanti rorn-an
behind the dressed-up woman, taranzipi rorn-an behind the platform(?), antezzi
ubrui rorn-an behindlafter the rst .-vessel; in Old Hittite governing the genitive:
[nu
m
Pi]tna attama ppan after (the death of) Pitana, my father KBo 3.22:10,
30 (Anitta text, OS); kui ammel ppan itoi-u kiar [i] whoever becomes king after
me KBo 3.22:22, 49 (OS).
20.17. Juxtaposing the readings of an OH text with its NS copy, one can see how
the NH scribe modernized the genitive construction ammel rorn-an after me of his
archetype to the dative one ammuk rorn-an (THeth 11 iii 49).
20.18. anda for the sake of: s+ r-:+ nakkiyanni anda for the sake of the
dignity of my brother; ammel utotu-ni anda for the sake of my dignity; r-:+
anda for the sake of my brother; si+-ni anda for the sake of namelreputation;
[k]edan[ipa]t memini anda for the sake of this very matter.
20.19. itarna in the midst of, among: iuna (urorn.xr-a) itarna among the
gods, in the midst of the gods; +\+ xtn
URU
atti itarna in the midst of attua; in
OH governing the genitive: r.xr-na r.xr-na itarna among the brothers and
sisters KBo 3.1+ ii 50; aannan itarna among the (royal) family (Homann 1984:
54 iv 31l23, genitive in -an).
20.20. ka (OH only) instead of (takes the genitive): kinuna 1 tut i-na
kaaa uittianta But now they shall substitute one sheep for the man KBo 6.26
i 41 (Laws 167, NS).
20.12 Local Adverbs, Preverbs, and Postpositions
00-HittiteGrammar.indb 298 6/5/08 4:18:59 PM
299
20.21. katta or kattan with, alongside of, and (with verbs of motion) to. Although in
MH and NH katta(n) governs the dative-locative, in OH it governs the genitive katta:
ka[ttimi] sro
5
-in tugga katta sro
5
-i [n] [t]u It is well with me; may it be well with
you too ABoT 65 obv. 34 (MHlMS); annama katta with my mother, where my
mother is; takku i-i otu-a katta [wat]ai If a man has sexual relations (lit., sins)
with a cow KBo 6.26 iii 20 (Laws 187, OHlNS); tut-a katta with a sheep KUB
29.34 iv 5 (Laws 188, OHlNS); aplpat annaa katta with her mother, . . . utxt.
xtts-a katta with the daughter . . . utxt.rf-a katta with a son KBo 6.26 iii
2628 (Laws 189, OHlNS); orxr-a nama
MUNUS
xn.xru-a katta with either a slave
girl or a prostitute KBo 6.26 iii 47 (Laws 194, OHlNS);
6
[(toti
LU.ME
xtniurx)]
a katta kt arta the overseer of the cooks stands on this side, next to the brazier
KBo 17.15 iv 1415 (OS); ammel katta with (or: to) me KUB 36.44 iv 4; ammel katta
arnut to me, apl katta to him, annama katta arnut bury me with my mother.
kattan. numu k r.xtn.n.xr kuiki kattan nanda uppai (There is) some
chariotry here with me. I will dispatch it to you HKM 30:810 (MHlMS); ammuk kat-
tan with me KUB 23.77 rev. 68. nu
d
Telipinu[
d
tft-un] arunaa utxt.xtts-si
arunaz rorn-[ pa uwatet] nu
d
rx-a kattan pe[utet] Telipinu brought the Sungod and
the daughter of the Sea(god) back from the sea and conducted them to the Stormgod
(
d
rx-a kattan) KUB 12.60 i 1315 (myth, OHlNS).
20.22. kattan under, below, mostly with the dative-locative:
GI
ntn-i kattan
under the table; andamakan tut-un kuwpi kuewn nu linkiya kattan kian daiwen
in addition when we had killed a sheep (as part of the treaty-making ceremony), we
placed the following obligations (lit., placed as follows) under oath KBo 16.47:1516
(treaty);
GI
atalkini kattan tuli[ya pedi] under the hawthorn tree, in the place of as-
sembly; in OH with the genitive: nepia ka[(ttan ulilikeddumat )] under heaven you
(pl., addressing trees) were growing KUB 29.3 i 11 (OS), restored by KUB 29.1 i 28
(OHlNS); see also Neu 1968b sub uliliya-).
20.23. peran
7
before, in NH with the dative-locative: kartimi peran before their
hearts (i.e., to themselves),
d
tft-i peran before the Sungod, iunimi peran before
my god,
GI
ntn-i peran in front of the table; in Old Hittite with the genitive or da-
tive-locative: itoi-wa pran or itoi-i peran before the king; also in New Hittite
in the sense of because of: kati peran akki He died from (i.e., because of) hunger,
inani peran on account of illness, pittuliyai peran because of fear.
20.24. ar up onto (answering the question to what place?), mostly with the
dative-locative: ui ar up onto the roof;
GI
uo-ti ar up onto the dais,
6. NH copies of the laws substituted dative-locative (r.xtn.n-i) or accusative (arnuwalan) for
these genitives (see Laws 200 in KBo 6.26 iv 23).
7. This word is written either p-e-ra-an or plp-ra-an. Many Hittitologists, considering the translitera-
tion p to be unmarked, represent the word even in broad transcription as piran. But only once (in an Old
Script tablet) is there a writing p-i-ra-an, and this passage is suspect.
Local Adverbs, Preverbs, and Postpositions 20.24
00-HittiteGrammar.indb 299 6/5/08 4:19:00 PM
300
NA
4
uwaiya ar up to the stela,
URU
Taurpi ar up to the city of Tahurpa,

ilamni
ar up to the portico (or: gate structure); sometimes with allative: urorn-ti+kan
x.oi-a ar ari the (statue of the) god comes up to the main gate, ua ar paizzi
he goes up onto the roof.
20.25. er upon, over, above, on behalf of, concerning, usually with the dative-
locative: ui er on top of the roof; ntn.so-i er on top of the mountain; txt-ni
er txt-an seed upon seed; ln-i er on the slaves behalf; +\+ i.xr
KUR
Amurra
er because of the people of Amurru; kuedani er for the sake of what?; itoi-wa
zr-ni er for the sake of the kings lifelsoul; memiyani er concerning the matter.
With the genitive, as in OH: kuel er on behalf of whom.
20.26. Just as some postpositions take the genitive of nouns and independent pro-
nouns in OH, so also they can occur with the enclitic possessive pronouns (see Houwink
ten Cate and Josephson 1967; Otten and Souek 1969: 7073; Otten 1973: 4445; Neu
1983: 96): rorn-anamet afterlbehind them; ret on his behalf; perantet be-
fore you. When this kind of peran takes enclitic pronouns beginning with m or , the
nal n assimilates: peraet before himlher; perammet before me. In the preceding
examples the place words look like nominative-accusative neuter substantives, but see
6.6 (p. 140). In other examples a dierent group of postpositions look like dative-
locative forms: itarniummi in our midst (= anza itarna), kattitti together with
you (= tuk katta), kattiummi with us (= anza katta). See 6.8 (p. 141). The use of
the genitive or an enclitic possessive pronoun with peran, er, etc., does not alter the
fact that they are postpositions (see Salisbury 1999: 6162 against Starke 1977: 131).
Whether any of them are derived historically from nouns is a separate question. See in
addition to the above authors Melchert 1984b: 12226.
20.27. In answer to the question from what place? the Hittites used the ablative,
often in combination with one of the above-mentioned adverbs: tnt-az katta down
from the city (ancient Anatolian cities were built on hills or mounds); .n.n-az ar
up from the sea;
URU
Gaipuraz ar up from Gaipura; antenaz ar up from the
hole (in the ground). But see tuzziyaz ppa behind the army (not from behind the
army!). Once again the issue of whether the adverbs here are postpositions is debated,
mostly without consequence for the meaning.
20.28. Also governing the dative-locative is tapuza (tapua) alongside of: a (or
otr-i ) tapuza alongside the brazier; patteni tapuza beside the hole; tabarwaui
tapuza beside the t.-bread;
GI
uo-ti tapuza beside the dais; luttiya tapuza by
the window; uwaiya tapuza beside the stela. The form tapuza is an ablative of
direction (see 16.92, p. 265) from the -stem noun tapu- side: .n.n tapua to
the (other?) side of the sea; overseas.
20.29. Perhaps governing the ablative is par forth from, out of:
GI
z.ix.on-
az par out of the tent; yet one also nds x-a par out of the gate, in which the
20.25 Local Adverbs, Preverbs, and Postpositions
00-HittiteGrammar.indb 300 6/5/08 4:19:00 PM
301
noun in -a is not an ablative. See for this use of dative-locative to indicate place from
which 16.69 (p. 259).
20.30. parranda and pariya(n) both mean over, across, beyond; out in, yet parranda
governs the dative-locative, while pariya(n) governs the accusative: aruni parranda or
arunan pariyan across the sea or out in(to) the sea.
20.31. The postposition iwar in the manner of; as governs the genitive (see
16.5916.60, pp. 255256, and 19.9, p. 291).
20.32. An older view claimed that the postpositives mn and maan like, as gov-
erned the genitive case (Goetze and Pedersen 1934), but it has since been shown that
they can govern any case, namely, the case of the noun or pronoun to which they make
a comparison (Gterbock 1943: 154). Postpositional comparative mn (see CHD LN
mn 1 with literature) is attested from Old Hittite onwards (Neu 1974: 1034 with
n. 227; contra Kammenhuber 1969a: 28081).
Local Adverbs, Preverbs, and Postpositions 20.32
00-HittiteGrammar.indb 301 6/5/08 4:19:01 PM
302
Chapter 21
VERB VOICE
21.1. Functionally, we may distinguish three verbal voices: active, passive, and
middle.
1
In the active voice the subject is the initiator of the action (the agent) ex-
pressed by the verb with no further implication. In the passive the subject is the recipient
of the verbal action (the patient). In the middle voice the subject both initiates and is
somehow aected by the verbal action.
2
21.2. In Hittite, active verb forms express the active voice, while medio-passive
forms can express either active, middle, or passive voice. Some individual verbs show
only one of these functions in their medio-passive forms, but others more than one ac-
cording to context (see Neu 1968a on the functions of the medio-passive system). The
passive is most often expressed not by a nite (medio-passive) verb form but by a con-
struction with the verb to be plus the participle (see 21.1021.12, pp. 304305).
Active Voice
21.3. Active verbs may be divided into transitive and intransitive. Transitive verbs
regularly take an accusative object, while intransitive do not. Sample transitive active
verbs are pp- to seize, d- to take, dai- to put, pai- to give, wal- to strike, hit,
and kuer- to cut. These and other transitive verbs can occur without a direct object in
suitable contexts: e.g., (My father sent infantry and chariotry, and they attacked Egyp-
tian Amqa) nammaya uiyat nu namma wal (a)er He sent again, and they struck
again KUB 14.8 i 20 (NH). Sample intransitive active verbs are: - to be, - to be
left over, remain, e- to sleep, up- to rise (of the sun), ipart- to escape, survive,
and ark- to perish. These and a few others are always intransitive, that is, they never
take a direct object (except for cognate objects, au atan e- to sleep well). A
number of active verbs can either take a direct object or not: akuwai(a)- to look or
to see (someone), tar- to prevail or to defeat (someone), na()- to be afraid or
to fear (someone), penna- to drive. The dierence in usage is sometimes marked by
the addition of -za (see 28.25, p. 361). With na()- to be afraid there is also a third
option: the impersonal use (see 16.32, p. 250)
1. We follow here, for descriptive purposes, the traditional conception of voice or diathesis preva-
lent in Indo-European studies, for which see Neu 1968a: 13 and Meier-Brgger 2000: 243, 2003: 259. We
cannot enter into the controversy over the proper cross-linguistic denition of the categories middle and
passive. See among others on this issue Kemmer 1993, Fox and Hopper 1994, and Rubio 2007: 3.13.3.
2. See the characterization of the Greek middle by Goodwin (1965: 267): in the middle voice the sub-
ject is represented as acting upon himself, or in some manner which concerns himself.
00-HittiteGrammar.indb 302 6/5/08 4:19:01 PM
303
21.4. Hittite medio-passive forms that show active (instead of middle) function may
be transitive or intransitive. Examples with transitive meaning (so-called deponents):
pa- to guard, arra- (with -kan) to transgress,
3
par-, pariya- to break (bread).
Most of these are also attested with active forms in the same meaning. The latter tend
to spread at the expense of the former from OH to NH. Often the medio-passive forms
of such verbs are preferred in OH and MH, and the active forms in NH (see 14.11,
p. 233).
21.5. Examples with intransitive meaning: ar- to be standing, aiya- to be dear,
loved, aliya- to prostrate oneself, - to take ones seat, sit down, iya- to go, igi-
to become cold, iduwa- to become known, ki- to be situated, lie, ki- to occur, hap-
pen, become, pangariya- to be widespread, general, dukk- to be seen, be observed,
regarded, important, etc. Some of these are semantically transformative or change
of state (-, ki-, aliya-, igi-, iduwa-) and others stative (ar-, ki-, aiya-, dukk-,
pangariya-). As a rule, verbs of this class are not attested in both active and middle
forms with no apparent dierence in meaning, but there are some exceptions such as
aliya- to prostrate oneself, uya- to ee, run, and naariya- to be afraid.
Middle Voice
21.6. Some medio-passive forms of transitive active verbs, often accompanied by
-z(a), are reexives, expressing an action performed by the subject on the subject: nai-
(act.) to turn something, (m.-p.) to turn oneself; unuwi- (act.) to decorate, (m.-p.)
to decorate oneself; see also aeke- (m.-p.) to seat oneself; arra- (m.-p.) to wash
oneself; munni- (m.-p.) to hidelconceal oneself; ln-(n)a- to subject oneself.
21.7. Some medio-passives, when plural, can be understood as reciprocals, with the
action performed by the subjects on each other: zaiya- (m.-p.) to ght each other;
pp-lapp- (m.-p.) to take each other by the hand; arra- (m.-p.) to part from each
other; anna- (m.-p., and -za) to contend with each other (in court).
Passive Voice
21.8. Some medio-passive verbs function as passives to corresponding actives:
armizziya- (act.) to bridge something, (m.-p.) to be bridged (of a river); alzai- (act.)
to call for, name, (m.-p.) to be called; amenk- (act.) to bind, (m.-p.) to be bound;
and andi- (act.) to match, correlate (as transitives), (m.-p.) to be matched, corre-
lated.
4
Other verbs in the medio-passive form which belong to this category and may
3. Note, however, that even in NH the verb arra- in its meaning to separate, divide has a passive
meaning to be separated, divided in the medio-passive.
4. The medio-passive use is largely conned to oracle texts. The usual translation is be determined,
but the idea behind this use is that an oracular eventlobservation is matched to a verdict (favorablel
unfavorable). The other active meaning to prepare, of course, has no correlation to the meaning of the
medio-passive forms.
Verb Voice 21.8
00-HittiteGrammar.indb 303 6/5/08 4:19:03 PM
304
have a passive sense are: a(a)nu- to be brought into order; arai- to be stopped,
ariya- to be investigated by oracle, anna- to be judged, itama- to be heard, au()-
to be seen; immiya- to be mixed, kariya- to be covered, l- to be loosed, lauwa-
to be poured (liquids), and lag- to be knocked down, knocked crooked, tarupp- to
be assembled, uwa- (< au-) to be seen, ln-na- to be subjugated. It is not claimed
that every medio-passive form of the above-listed verbs must be interpreted as passive,
only that at least one example of each verb is.
21.9. Some agentless passives are eectively equivalent to intransitives: iri- (act.)
to bound or limit, (m.-p.) to come to an end; zinne- (act.) to put an end to, (m.-p.)
to come to an end, etc.
21.10. The medio-passive form of the verb is not the preferred choice for express-
ing the passive of verbs that exist in the active. Usually a construction employing the
(passive) participle in -ant- of a transitive verb plus a nite form of the verb to be is
employed. On predicative participles see Houwink ten Cate 1973a; Laroche 1975; Boley
1984; Wagner 1985; and also 25.43 (p. 339).
21.11. The following are examples of passive participles of transitive verbs used as
predicates with the verb to be expressed: nu utn maniyaeker nu tnt.uruir.nr.
oi.oi-:t+ tittiyante eer They governed the land, and the large cities were assigned
(to them) KBo 3.1+ i 1011 (Telipinu proclamation, OHlNS); napa r.xrsi utxt.
xrsi
LU.ME
gaenai i.xr aannaa i rnr.xrsi taruppante aandu Let
(the future kings) brothers, sons, in-laws, the men of his clan and his troops be united
KBo 3.1+ ii 4042 (OHlNS); nuza r+\t i.xn mekki paaanuanza Be very
much on your guard (lit., be protected) before the enemy! HKM 1:1113 (MHlMS);
tnt.uruir.nr.ma [(kui s )]+ xtn
URU
atti itappante eer What cities of the Land
of atti were blockaded att. ii 4142; (The birds which you sent to me,) nat ara
arrante eer they were spoiled, (so I didnt eat them) AT 125:12 (NH); (Let those who
prepare the daily bread be clean;) war [ p]antiat kartante aandu iniumakan
txn[r.xry]a dn du let them be washed and clean, and let the body hair and
ngernails be removed from them KUB 13.4 i 1516 (pre-NHlNS). In most cases
the construction with the participle and be represents a state passive: nat ara
arrante eer they (scil., some killed birds) were spoiled (= rotten) AT 125:12 (NH),
but in some of the examples cited it is impossible to tell whether they refer to a state or
to an eventlaction (see 22.25, p. 312, for more on this problem).
5
21.12. With the verb to be unexpressed: iyl ln.xrsi [ts:i 1 x] aante
the servants of Number One (i.e., the king) (are) born [of one mo]ther KUB 1.16 ii
47 (OHlNS); takku kuan piyn if the wage has been paid KBo 6.2 + KBo 19.1 ii 28
(OS); takku utxt.xtts i-ni taranza if a girl has been (or: is) promised to a man
5. The distinction here is the same as German die Tr ist genet the door is opened (= open) versus
die Tr wird genet the door is (being) opened.
21.9 Verb Voice
00-HittiteGrammar.indb 304 6/5/08 4:19:04 PM
305
KBo 6.3 ii 5 (OHlNS); eknumet anda nan their .-garments are turned inside-out
KBo 3.34 i 21 (OHlNS); [
GI
n]tnka[n o]u-it kariyan the table (was) covered with
the linen cloth KUB 24.8 i 22 (NS); [(t)]ukwa ror-anda
GI
rnr-p kar [(duw)]arnan
ft
7
.nr.yatta r[or-anda] kar zanuwan Before you the cedar (is) already broken,
before you meals (are) already cooked KUB 33.102 ii 1012 (Ullik. I); xtn
URU
a [tti
ts ]:i txt itoi-i::t uwan the land of atti (is) lled with the seed of kingship
KUB 26.1 i 1011 (edict of Tud. IV); du iia tukpat piyan a weighty lord-
ship(?) (is) given to you KUB 31.127+ i 19 (hymn to ama); nuza kar maan an-
naza X-za aanza eun when long ago I was born from my mothers womb KUB
31.127 + ABoT 44+ iv 24 (OHlNS). For the use of -za in the last example, see 28.34
(p. 363). The verb to be is often unexpressed when the time referred to is the present
or the statement is a general, timeless one (22.3, p. 306).
21.13. Surrogates for passives. Some transitive verbs form no passive at all but em-
ploy forms of other verbs of related meaning (so-called suppletives). Thus forms of
akk- to die are used to express the passive of kuen- to kill. Similarly forms of ki- to
be situated, lie, be placed are used instead of passive forms of dai- to put, place; er
tiya- appears as a passive for er dai- to place upon; and -za ki- to become as a
passive for iya- to make.
21.14. In passive constructions it is customary to leave the agent unspecied. When,
however, an agent is specied, the noun or pronoun denoting the agent can appear in
either ablative (see 16.99, p. 267) or instrumental (see 16.107, p. 269) cases, never
the dative-locative (see 16.73, p. 261).
Verb Voice 21.14
00-HittiteGrammar.indb 305 6/5/08 4:19:04 PM
306
Chapter 22
VERB TENSE
22.1. Not only does the Hittite language lack the subjunctive and optative moods
of other old Indo-European languages but like the Germanic languages it has only two
synthetic (non-phrasal) tenses: (1) the present and (2) the preterite. On the analytic
construction with participle and the verbs to have and to be compare 22.19 (p. 310)
and following.
Present
22.2. As its name indicates, the present tense sometimes indicates an action si-
multaneous with the time of speech. These examples translate best as English present
progressives (to be . . . -ing): antutarmawanna arantallin[(zi )] the people
are murmuring against us KUB 6.41 i 25 (Mur. II);
1
nuza -ua utnemet
-ermitta paami (the gods have entrusted to me the land and my house,) and I,
the king, am (now) protecting my land and my house KUB 29.1 i 1819 (rit., OH/NS);
-pa!ma tezzi nu kuit ianzi api -pa tezzi
GI
lali arzi
GI
u uwadu
arkanzi But he replies (to the eagle): And what are they doing? (For the use of nu
here see 29.48, p. 403, and examples in 27.8, p. 351, and 27.10, p. 352.) (The eagle)
replies: She is holding a dista, (and) they are holding full spindles KUB 29.1 ii
57; kuedaniwaza menaanda iamikei To whom (lit., facing whom) are you
singing? KUB 36.12 ii 9 (Ullik. II). This immediate present can also be made more
explicit by the addition of the adverbs ka (tta) and kma (see 24.29, p. 324).
22.3. Certain other syntactic choices correlate with the bipartite presentpreterite
tense system of Hittite: The verb - to be is often unexpressed when it would be in the
present tense but never in the past tense. In OH the choice of ta versus u as a sentence
connective corresponds to present (including the analytic perfect [22.1922.23,
pp. 310311] which uses a present tense auxiliary verb) or preterite tense verbs (29.3,
p. 390).
22.4. Hittite uses the present tense to express a present state with the adverbs kar
already and nawi not yet. English employs the present perfect: na kar paizzi he
has already gone; and na nawi paizzi he has not yet gone. For these constraining
adverbs with the preterite see below, 22.16 (p. 309).
1. The present pl. 3 arantallinzi is an unreal hybrid form resulting from imperfect Hittitization of a
Luwian *arantallinti. There is no genuine Hittite third plural ending -inzi.
307
22.5. Statements of general validity (sometimes called gnomic) are expressed in
Hittite by the present tense (see also 24.15, p. 321): ror-zian orx-an
GI
urkin rorn-
zi anda t wemiyazi As the rear (wheel) does not overtake the front wheel KBo
11.14 ii 2223 (OHlNS); xtr-izakan
GI
taptappan rorn-pa pzi nan
GI
taptappa
u[inuzi ] A bird takes refuge in the nest(?), and the nest(?) saves it KUB 14.8 rev.
22 (PP 2, NH); +\+ utxt.x.i.t
19
.itpatkan anda memian kian memikanzi
arnuwawa xtts-n urorn-ti+ kri tiya[zi] Among mortals they have a say-
ing (lit., they speak as follows): A god yields to the wish of a woman of the birthing
stool KUB 21.27 ii 1516; lalawa armizzi the tongue is a bridge KBo 11.72 iii 5
(MH?lNS; CHD LN 22); and other proverbs (for a collection of which see Beckman
1986 and CoS 1:215 exx. 5 and 9).
22.6. Hittite also uses the formal present tense to indicate past time, especially in
narrative. This is what is usually called a historical present. Since all identied ex-
amples are found in New Hittite texts, it is possible that this construction was not
used in Old Hittite.
2
In the following examples present tense forms translatable as past
are underlined: maanmaa
URU
Twanuwa ari nu s+r+t
URU
Tuwanuwa di nuza

URU
Twanuwan zaiyauwanzi pzi When he reached Tuwanuwa, he positioned (him-
self) below Tuwanuwa
3
and began to ght against Tuwanuwa KUB 19.18 iv 1618
(D frag. 15); [luk]kattimakan +ui:+
URU
Tiwanzanaza katta xtn-ekan anda pennai
rorn-annan
LU.ME
+u:+rrtsi 6 ,t+oi+ r.xtn.n.xr arzi nu +ui:+ maan
nannai nukan edani pangawi i.xn 1-ankipat anda andaizzi nanzaan +ui:+
zaiyauwanzipat pzi The following day my father drove down from Tiwanzana into
the land, (while) in the rear his charioteers and six teams of horses were supporting him.
And as my father was driving, he encountered that whole enemy at once, and my father
engaged him in battle KBo14.3 iv 2631 (D frag. 15); +\+
m
[Piu]niyama orx-an
nakketa [na tamedani] namma [t] kuedanikki paizzi na[mu kattan uet] n[amu
ue]t oln.xr-a ox-an aliyat[tat] When it became dicult for [Piu]niya, he did[nt]
go (lit., doesnt go) to anyone [else], but [he came to me,] and he [proceed]ed to pros-
trate himself at [my] feet KUB 19.30 i 1618; nu
m
Aitaqqama kui itoi
URU
Kinza ta
nui
m
\-++-
d
t-a [kui] [ant]ezzi utxt-la ta nu maan auta [anda]kan kuit
atkenuwante numa alkiu namma [tepaw]ezi nuza
m
\-++-
d
t-a
m
Aitaqqaman
+uisi kuenta When Niqmadda, the eldest son of Aitaqqama, King of Kadesh, saw
that they (were) blockaded, and that furthermore the crops (i.e., food supplies in siege)
were in short supply (sg. verb!), Niqmadda killed Aitaqqama, his father KBo 4.4 ii 36
(AM 12223). Note that in the last example the historical present tepawezi (and the
implied verb aanzi; see 16.6, p. 243) are in a kuit that clause subordinate to auta,
2. The pattern of use of the present tense in past narratives in Old Hittite seems to be regulated dier-
ently. For a preliminary description see Melchert 1998: 41617 (but the examples he cites for the historical
present in Old Hittite are all dubious).
3. That is, he besieged the city. See kattan dai- with the same meaning. Perhaps s+r+t GN dai- is merely
another way of writing GN kattan dai-.
Verb Tense 22.6
00-HittiteGrammar.indb 307 6/5/08 4:19:06 PM
308
that is, they concern what the young man saw (auta) and therefore were matters pres-
ent to him at the time he saw.
22.7. Historical presents are by far most common when the subject is grammati-
cally third person. In historical narratives which are expressed in rst person (e.g.,
Murilis annals or attuili IIIs Apology), the rst-person forms are all preterites, and
the only historical presents are in third person, referring to people other than the narrator.
Of course, rst-person present forms which actually express future actions occur fre-
quently in direct discourse recorded in these narratives, but that is an entirely dierent
matter: nu
m
Kantuzilin [kian watarnaun i.xr]
URU
Nuaiwa kuit kururiyaer
[nuwara ara arnik nuwama m]n rnr.xr
URU
Mizr [warre]a[nzi nuwamu
]a[trai nuwa] ammuk [ pimi ] nuwaran zaiyami And I i[nstructed Kantuzzili
as follows:] Because the men of Nuhai have made war, [destroy them.] If Egyptian
troops come to their aid, [write to me,] and I [will go] and ght them KUB 14.17 ii
2226 (AM 86).
22.8. The present tense can also express an English future (shall, will): iuu++
siu++ kui ammuk rorn-anda itoi-u kiari whoever in the future shall become king
after me KBo 3.1 ii 40 (Tel. pr., OHlNS); +\+
LU
+i:t-x+wa ammuk peran iyami
nuwazakan
URU
xi.nnnn-a manza ts:i s+
LU
+i:tx+ neyari I will march be-
fore your husband, and all atti land will turn to the side of your husband att. iv
911.
22.9. Commonly the temporal ambiguity of a present-tense form is resolved by the
use of a temporal adverb (19.6, p. 290) such as iuu++ siu++, zilatiya, or appaiwatti or
by the employment of the phrasallserial construction with uwa- or pai-, which when in
the present-future tense always refers to the future (see 24.3124.42, pp. 324329).
22.10. Speakers could constrain a present progressive or immediate future meaning
of the present tense with the introductory words ka or kma (see 24.29, p. 324).
22.11. In if clauses (introduced by either takku or mn) the present-tense form is
always hypothetical or future in meaning (see 30.49., pp. 420.): takkuwaan k
azzizi tawa urorn-ti+ takkuwaan nattama a[zzizi] tawa antuwae If he
achieves this, he is a deity; but if he doesnt achieve it, he is a mortal KBo 3.60 ii 1417;
takku utxt.itoi antezzi t.oi nu kui tn pda utxt-ui nu itoi-u ap
kiaru If there shall not be a rst-rank prince, then let a prince of second rank become
king KBo 3.1 ii 3638; takku utxt.itoima watai nu so.ut-azpat arnikdu If
a prince sins (i.e., attempts homicide), let him pay only with his own person (not with
those of his extended family) KBo 3.1 ii 5556. For if clauses with the preterite see
30.53. (pp. 422.).
22.12. Like the English future tense, the Hittite present tense used for future time
often expresses armation. In the rst person it indicates willingness to comply with a
request or simple expression of intention to act: i+++ +ut +ut:+++ [kuiwa p]aizzi
22.7 Verb Tense
00-HittiteGrammar.indb 308 6/5/08 4:19:08 PM
309
i+++ +ui:+++ ammukwa pimi [nukan +]ut +ut:+ +ui:+ par naita My
grandfather said, [Who] will go? and my father replied: I will go. [Whereupon]
my grandfather sent my father out KBo 14.3 iii 911 (D frag. 14); s+
d
ts:+u par
andandatar memai nat otxt.x.i.t
19
.it-a itamadu I shall declare (lit.,
speak) the divine providence of (the goddess) Is:+u, and let (all) humanity hear it!
att. i 56; kuitmanmaza otxt-a eun nuza xtn.xtn i.xn kue tarikenun nat
)iiii ant oi-mi nat i+\t ororn-tt+ tei I shall make a separate tablet (in which
to enumerate) those enemy lands which I conquered while I was (only) a prince, and
I shall deposit it before the goddess att. i 7374;
d
ts:+umamu o::+ i-a[(t)]
numu i-it k memita ororn-tt+-niwatta ammuk tarnai nuwa l nati Is:+u, my
lady, appeared to me in a dream and in the dream said to me: I will turn you over to a
(friendly protective) deity, so dont be afraid! (or perhaps rhetorical question: Would
I turn you over to a (hostile) deity? Dont be afraid!) att. i 3638.
22.13. In the second person the eect can approximate a command (HE 136 258,
2):
4
ro-an azzateni wtarra ekutteni You shall (i.e., must) eat (only) bread and
drink (only) water KUB 1.16 iii 48 (OHlNS).
Preterite
22.14. The preterite form, when not marked by constraining adverbs, is used freely
to refer to any past event. According to context, the preterite may refer to a simple past
action (perfective), an ongoing action (imperfective; see 24.3, p. 317), an action
prior to another action in the past (pluperfect), or an action completed in the past
(usually the recent past) whose eects carry over into the present (so-called present
relevance). English may render these meanings respectively with its simple past (he
went), past progressive (he was going), past perfect (he had gone), and present
perfect (he has gone).
22.15. Speakers marked a present perfect meaning of the preterite form with the
introductory words ka or kma. This adverb has a constraining function (see 24.28,
p. 324) just as does the use of the phrasallserial construction with uwa-lpai-, which
marks the future (24.38., pp. 327.).
22.16. Hittite uses the preterite with the adverbs kar and nwi (26.11, p. 343, and
following) to express a state anterior to another reference point in the past. English usage
requires a past perfect translation for such cases: na kar pait he had already gone.
na nwi pait he had not yet gone.
22.17. For the use of the preterite in conditional clauses with mn if see 30.53
30.55 (pp. 422423).
4. German Heischefutur.
Verb Tense 22.17
00-HittiteGrammar.indb 309 6/10/08 7:50:08 AM
310
22.18. Temporal mn (when) with the preterite very rarely corresponds to the Eng-
lish future perfect (see Sommer 1932: 163, cited by HE 258a1 as similar to futurum
exactum of Latin). The example cited by Sommer is mnmaiyatkan uia na
rorn-pa t\+ xtnx+ ue[z]ziwhen I (shall) have denied it (i.e., the land he attacks) to
him, he will come back into your land KUB 14.3 iii 6061 (NH). Standard current
English omits the shall in this construction: when I have defended it from him. The
present-future verb uezzi in the apodosis shows that a future perfect is intended by the
(Luwian) preterite form uia.
The Analytic Perfect Construction
Function
22.19. Hittite expresses the completion of an action prior to the time of speech (pres-
ent perfect has gone) or prior to another action in the past (pluperfect had gone)
(22.14, p. 309) either with the simple preterite form, with the simple preterite plus the
adverb kalkma (see 22.15, p. 309), or by means of a construction with the (past)
participle followed by an indicative
5
form of the verb ar(k)- to have or e- to be.
6

In Old Hittite the conjunction u correlates with analytic perfect constructions using
a preterite auxiliary verb, while ta patterns with the present tense auxiliary verbs (see
29.3, p. 390).
Form
22.20. There are two types of this construction (initially observed by Benveniste
1962a: 63, 65): one with the auxiliary verb ar(k)- and another with the auxiliary verb
e-. Transitive verbs as well as some intransitive verbs employ the rst type, while other
intransitive verbs use the second.
7
As the examples cited below show, in the construction
with ar (k)- the participle is always in the neuter nominative-accusative singular, while
in that with e- the participle agrees with the subject in gender, number, and case.
Uniqueness within the Old Anatolian Indo-European Group
22.21. No analogous construction exists in Luwian (Melchert 2003d: 206; Dardano
2005: 93 n. 4). Since it antedates the development of similar constructions in the Ro-
mance and Germanic languages, caution is needed in making hasty comparisons (un
rapprochement sommaire) with those constructions (see Benveniste 1962a: 41).
5. When the imperative of the auxiliary verb is used, the construction has a dierent force. See 22.24
(p. 311).
6. See Boley 1984; Dardano 2005.
7. Intransitive verbs that take e- as the auxiliary verb also require subject clitic pronouns (see 18.13
(p. 280) and belong to the class often labeled unaccusative, while those that take ar(k)- do not occur with
subject clitics and belong to the unergative class. See Garrett 1996: 1026 revising Sommer and Ehelolf
1924: 30 and Boley 1984. Dardano (2005: 99 and 103) follows Garrett.
22.18 Verb Tense
00-HittiteGrammar.indb 310 6/5/08 4:19:11 PM
311
Examples
22.22. Present perfect examples with present of ar (k)- and with unexpressed e-
(22.3, p. 306): ta mn utxt.xr r.oi kuzzi parnan arkanzi nat apezza peiyazi
If he throws (it) to the side on which the palace ocials have crouched KBo 4.9
vi 68; natzakan ka +\+
d
t
URU
atti r:+ i +\+ urorn.xr uiti
ME
:+ peran
tarnan armi (The sin of my father also reached me), and I have confessed it before the
Stormgod of atti, my lord, and the gods, my lords KUB 14.8 rev. 1416 (PP 2, NH);
nuan mn alki arante nakan ara warten When the crops have ripened
(lit., arrived), harvest them! (Bring them to the granary!) HKM 25:1519 (MHlMS);
ammellamukan kui s+
d
tft-st peran ara watkuwante Also those (subjects) of
mine, His Majesty, who have ed from me, (let them be an enemy to you, as they are
an enemy to Mauiluwa) KBo 19.70:2224 (Manapa-Tarunta Treaty).
22.23. Past perfect examples with preterite of ar(k)- and e-: numu itamaan
kuit arker (I went into the land of Taggata.) Because they had heard about me, (I
no longer went after them with force by day, I marched at night) KBo 5.8 i 2325
(D); s+ i.xr
URU
Taggatama kui rnr.xr \+u'uI anda warriante eer nat
ara pareer But the auxiliary troops of Taggata who had come to help (my op-
ponents) dispersed KBo 5.8 i 1820 (AM 148149); +\+
URU
Katariyaya kui rnr.
xr
URU
Gaga manza wa[rr ]i panza ta The whole Kaka army that had gone to K.
also to help KUB 19.11 iv 3637 (D frag. 13); nukan antutar kuit t\+ tnt.uruir.
nr.si\i rorn-pa pn ta Because the population had gone back into their cities
KBo 5.6 i 1920 (D fr. 28).
Stative Constructions with ar (k)- and -
22.24. The (neuter singular) participle + ar(k)- construction with transitive verbs
can also have the meaning to holdlkeep (someonel-thing) . . . -ed). This is simply the
literal meaning of the two components: ar (k)- to hold + the passive participle of the
transitive verb: nu i.xn-an utn kuttanit tar (a)an arta (the king) held the land(s)
of the enemy subjected by (his) strong arm KBo 3.67 i 6 (Tel. pr., OHlNS); nu xtn-
e paanuwan arker They held the land protected (not the pluperfect: They had
protected the land) KUB 14.16 i 24, ed. AM 28; nuwarakan i.xr
URU
niuut anda
wanuwan arkanzi the Hurrians are keeping them surrounded KBo 5.6 ii 25 (D frag.
28). This construction is more common with the auxiliary verb ar(k)- in the impera-
tive. When the particle -za occurs in this construction, it construes not with the auxiliary
verb ar(k)- but with the verb in the frozen neuter participle. The construction participle
+ imperative of ar(k)- does not express a tense but should be translated keep (some-
thing) . . . -ed: nutta urorn.xr fr-an arkandu nutta auli paandaru May the
gods keep you alive and protect you in wellbeing HKM 10:4546 (MHlMS); numu
urorn-ti+ itamanan lagn ar (a)k O god, keep your ear inclined to me KUB 24.1
i 1617 (NH); nu s+ i.xn kui xsxi.nr. naza uit ++oo+t:t kappwan ardu
Verb Tense 22.24
00-HittiteGrammar.indb 311 6/5/08 4:19:11 PM
312
naza gul (a)an ardu let the district commander keep track of (lit., keep counted)
the routes of the enemy (incursions) and keep a written record of them KUB 13.2 i 910
(MHlNS); tnt.uruir.nr. nXukan kui maniyaiya anda nuza urupan rorn-an
kappwan arkandu Let them keep urup-ed and keep track of the fortied cities in the
administrative district KUB 13.2 i 2223; nuwa karuiyan ar (a)k so keep quiet!
KUB 14.4 iv 11 (NH); nuwazakan ikia ppa t\+ xtn
UR.SAG
Zippal tiyn ar (a)k
So keep yourself placed with your back towards the land of Mt. Zippal KUB 14.1
obv. 1617 (MHlMS); parkuwaya fo.nr. waan arkandu Let them have on clean
clothes (lit., keep clean clothes worn) KUB 13.4 i 16 (pre-NH).
22.25. As in the case of ar (k)- with a frozen neuter singular participle, the con-
struction with - (unexpressed in the present tense, 22.3, p. 306) and a nominative
participle of an unergative intransitive verb can also indicate a state rather than an
event or action: (The birds which you sent to me,) nat ara arrante eer they were
spoiled (= rotten), (so I didnt eat them) AT 125:12 (NH); nukan 1-a 1-edani er
mauanza (if there are two kidneys) and one is fallen on top of the other KUB 4.1 iv
26 (extispicy). In many contexts it is hard to tell with intransitive verbs if a completed
event or a state is intended.
22.25 Verb Tense
00-HittiteGrammar.indb 312 6/5/08 4:19:12 PM
313
Chapter 23
VERB MOOD
23.1. The Hittite language has only two moods: indicative and imperative. There are
no separate sets of verbal forms with the force of a subjunctive or optative. The poten-
tial, unreal, or wished for are expressed by the indicative together with the optative and
potential particle man (usually written ma-an or -ma-an in contrast to the temporal or
conditional conjunction ma-a-an).
Indicative
23.2. The indicative mood is used to report real events in the past or present as well
as to refer to anticipated future events and to make statements of general validity. For
examples see illustrations of the present and preterite tense in chapter 22.
23.3. Negative commands (both prohibitive and inhibitive) are expressed by the
negative l and the present indicative (CHD L N l a; see 26.16, p. 344): nukan
x.n.xr katta uer natmu oln.xr-a k [(atta)]n aliyanda[(t)] uitt\twanna
l arnikti The civilian captives came down and fell down at my feet, (saying,) Our
lord, do not destroy us! KUB 14.16 iii 1617 restored from KUB 14.15 iii 4647
(NH); uitt \t wana kar [ku]it arnikta nuwa uitt\t l namma uwai Our lord,
because you destroyed us once before, do not come again! KBo 4.4 iv 4546 (NH).
Rarer and so far only in NH copies of OH texts is the use of the imperative with l (CHD
l c; see 26.17, p. 344).
Imperative
23.4. What are conventionally termed rst-person singular imperative forms (those
ending in -(al)lu in the active) are actually voluntative (HE

263b), expressing the
desire or strong intention of the speaker, equivalent to English let me . . . , may I . . .
, or I will surely . . .: [(natzak)]an nau teit uwallu Either let me see it in a
dream . . . KUB 14.8 rev. 42; numu atri nu eggallu Write to me and let me know
KUB 23.88 obv. 6; nu andan k kian memallu Truly I wish to say the following
KUB 6.45+ iii 7374 (NH), tr. ANET 398 (Truly I wish to speak thus); nu srsxtn s+
D
[Mala] iyallu natkan aa[(n)]ullu And I intend to perform the ritual of the [Mala]
River and complete it KUB 14.11 iii 1920 (NH); nuwa u[g]a n-a iwar wiyami
[nammaw]a akkallu I will squeal like a pig, [and then] I will surely die KUB 14.1
rev. 9394 (MHlMS).
00-HittiteGrammar.indb 313 6/5/08 4:19:12 PM
314
23.5. The rst-person plural imperative (properly a cohortative or exhortative,
HE

263c) is formally identical with the present-future indicative (endings -weni, -wani,
-meni, -mani, -watati). The presence of the imperative can only be determined by
context. The presence of eu or uwatten in the exhortative sense come on! indicates
the imperatival use. The rst-person plural forms express a wish of the speaker that the
addressee join him in the performance of an action (let us . . .): eu +\+
d
t r:+ . . .
ur-eni tiyaweni Come, let us go before the Stormgod, my lord, . . . for judgment KBo
6.29 ii 24 (NH); kinunawa eu nuwa zaiyauwatati Come now, let us do battle
with each other KBo 3.4 ii 1314 (NH); uwatten
URU
Na paiwani come, let us go to
Nea KBo 22.2 obv. 15 (OS). For further examples see 23.7 and 24.32 (p. 325).
23.6. The second-person forms are direct commands or requests: nuwamu
d
tft
URU
Arinna o:+ kattan tiya nuwamukan uni arazena xtn.xtn i.xn peran
kuenni O Sungoddess of Arinna, take my side and slay before me that enemy of the
surrounding lands KBo 3.4 i 2526 (NH, Murs. II); nuwara pten nuwaramu
par p[eten] Arrest them and extradite them to me KUB 14.15 i 14 (NH); nammamu
urorn-ti+ o. i-at [+oi ] r-:twamu ln-aut then the goddess, my lady,
appeared to me in a dream, (saying,) Serve me with (your) house! att. iii 45.
23.7. The third-person forms are employed when the speaker expresses to a second
party the wish that a third party may perform some action. Occasionally, there is either
the implied seeking of the consent of the second party for the third party to do this or the
implication that the second party joins the speaker in this wish: nuwa zaiyauwatati
nuwanna
d
t uitt:+ ot\++ annu Let us do battle, and may the Stormgod, my
lord, decide our case KBo 3.4 ii 1314, ed. AM 46; +\+
d
tft-stma r:+ s+
m
adu-
ur[orn-t]t+ kuit atrmi
d
tft
URU
fi-na o:+ ukedu Let the Sungoddess of
Arinna, my lady, examine what I shall write to His Majesty regarding aduili KUB
40.1 obv. 3738; iuneda auli paantaru May the gods kindly protect you AT
125:3 (from salutation in a letter); utnewa mu edu May the land prosper and have
rest KBo 3.7 i 5 (OHlNS); nu +\+ urorn.xr r.xr:+ zr-anza namma wardu To
the gods, my lords, may the mind again be gentle KUB 14.14 rev. 14 (PP 1 8). When
a second party is not addressed, a strong wish for a third partys action can be expressed
by using the present tense and the (speaker-) optative particle man (see 23.11).
23.8. Imperative forms are rarely used in prohibitions with l (see 26.17, p. 344).
23.9. For the isolated use of present indicative forms in positive commands see
22.13 (p. 309).
Optative, Potential, and Irrealis
23.10. For the expression of the potential, the unreal, or the wished for, notions as-
sociated with the subjunctive and optative moods of Greek and Latin, Hittite scribes
23.5 Verb Mood
00-HittiteGrammar.indb 314 6/5/08 4:19:13 PM
315
employed the indicative together with a particle man, which they usually (but not al-
ways) distinguished from the conjunction mn (written ma-a-an) by the shorter spell-
ings ma-an or ma-nalu-. All examples but one (KBo 5.8 iii 15)
1
of the plene writing of
potential man are clause-initial and followed by a vowel: ma-a-nu-u-kn KBo 3.1 ii
11 (OHlNS); ma-a-na-an KBo 3.4 iii 11 in AM 66. Further examples may be found in
CHD L N sub man. Unlike the conjunction mn when, if, man is not found in clauses
beginning with nu (HE 310 f), nor with ta or u.
Optative
23.11. The particle man can express a desire of the speaker (so-called speaker opta-
tive [Honer 1982]): [nu kuit (r.xr utxt.i)]toi danna ilaliyanzi nu kian [da]ranzi
[(aimanwa tnt-a a)]mmel kiari [Because] they covet the princes houses, they
speak as follows: That city ought to be mine KBo 3.1 ii 6364 (Tel. pr., OHlNS);
kamu k tet
m
Kalbaya xt-az memita manwana ianittartar iyaweni Kal-
baya has just said this to me he conveyed it orally We ought to make for our-
selves a blood relationship VBoT 2:13 (letter); manwa
d
tft-st fr-ezi I hope His
Majesty recovers KUB 15.30 iii 5.
23.12. The subject optative (Honer 1982), which expresses a wish not of the
speaker but of the grammatical subject of the clause, can be seen in: mnukan
m
uzziya kuenta nu uttar iduwti uzziya wanted to kill them, but the matter be-
came known KBo 3.1 ii 11; and in the following example the grammatical subject is
also the speaker: man t\+
URU
ayaa punpat nuza xt.xx-za er tepaweanza ta
I wanted to go also to ayaa, but the year was too short for that KBo 4.4 iii 2223
(Mursili II annals). On the dual use of man for speaker- and subject-optative see also
CHD man, a.
Potential and Irrealis
23.13. Future potential is expressed with man and the present tense: mnwamu 1-an
utxt-x+ paiti manwaramu
LU
+i:t:+ kiari If you would give me one of your
sons, he would become my husband KBo 5.6 iii 1213 (NH); nama m anwakan
uniu r.xr alla [ll] pnzi qqamanwa pudanzi Or those lords might defect
and also take me away (with them) KBo 4.14 ii 7980 (NH). For possible further ex-
amples see 26.23 (p. 346), 30.52 (p. 421).
23.14. Past potential is expressed with man and the preterite tense: [m]antakkan
r +ut-x+ xtn-x+ya t ara dir manat [(da)]mdani kuedanikki piyr Could they
not have taken from you your fathers house and your land and given them to someone
else? KBo 5.13 i 2021 (NH).
1. But note that the scribe of this text corrected ma-a-an-mu in iii 18 to ma-an-mu because it would have
violated this rule.
Verb Mood 23.14
00-HittiteGrammar.indb 315 6/5/08 4:19:14 PM
316
23.15. For hypothetical future actions deemed undesirable or unlikely to occur be-
cause of undesirable results expressed in the apodosis,
2
Hittite uses man with a preterite
in the protasis and a present or mixture of preterite and present (marer . . . kantati
. . . pzi . . . pzi ) in the apodosis: ma-a-am[-ma-an-a-ma-a?] (i.e., *mnmanma)
3

tayazzil piker man mantepat marr [man]e'
LU.ME
.zt kantati kman kn
pzi [k]aman kn pzi If they [were] to charge [them] with theft, all of them might
dissemble or become thieves; this one might seize that one, and that one might seize
this one KBo 6.2 ii 5355 (Laws 49, OS; see LH 5960). The reason for the alternat-
ing tenses is unclear. For another example with present tense in the apodosis see KUB
30.10 obv. 2223 in 30.54 (p. 422).
23.16. Past contrary-to-fact conditions are expressed with man and the preterite in-
dicative, both in main clauses alone and in combinations of conditional and main clause:
mankan mn +\+
m
Attariya uiwetenna ktitaman kten Even if you (pl.) had
survived Attariya, you would have died from hunger KUB 14.1 obv. 12 (MHlMS);
manza r]nr.xr] rorn-an kar p[a]anut antuaakan uttar kiar anda kar
dai If you had already protected the troops, (the god) would have already put into
your hand the matter of the persons HKM 38:37 (MHlMS); numa
LU
auriyalu kuit
arantat mnkan mn +\+
m
Pittaggatallipat warpa teun manmu
LU
auriyalu kuit s+
m
Pitaggatalli aur manmu t duuiyait And because sentries were in place for them,
if I had tried to surround Pittaggatalli himself, because the sentries of Pittaggatalli would
have seen me, he would not have waited KBo 5.8 iii 1418 (NH). Note the complex
syntax in the last example: three dependent clauses (causalconditionalcausal) fol-
lowed by the main clause. For the omission of nu or other conjunction to introduce the
nal clause see 29.4629.58 (pp. 401405).
2. Sometimes called non-past contrary to fact. We choose a dierent formulation, since future actions
cannot be judged as factual or non-factual, only as likely or unlikely.
3. See LH 5960, 19192.
23.15 Verb Mood
00-HittiteGrammar.indb 316 6/5/08 4:19:15 PM
317
Chapter 24
VERB ASPECT
24.1. Whereas tense marks the time of a reported action or state relative to the time
of speech, aspect in its broadest sense describes the internal temporal structure of a
verbal predicate. This topic is complex and controversial, and the use of terminology
varies widely (some scholars, especially Slavicists, reserve the term aspect for a much
narrower application than that used here). Here our goal is to make clear the meaning of
various Hittite verbal forms as they occur in the texts and to correct certain widespread
misinformation on their use. For a more detailed but by no means exhaustive descrip-
tion see Honer and Melchert 2002. The preliminary treatment of Melchert 1998 is also
useful but in need of revision.
24.2. For Hittite it is sucient to distinguish imperfective aspect, in which an event
is viewed as having an internal temporal structure, from perfective aspect, in which it is
not.
1
Perfective aspect is the unmarked partner in Hittite. Imperfective aspect, whether
it is explicitly marked or not, takes on various specic meanings outlined below, based
on the combination of the lexical meaning of the verb and certain contextual markers.
Imperfectives
Adverbial Markers
24.3. Any basic verbal stem in Hittite may be read as perfective or imperfective,
provided that its inherent meaning and the context are appropriate. The latter value is
assured by the presence of markers such as the conjunction kuitman while, the adverb
maiyanki as many times as, or multiplicative expressions such as txnrn-anki (or
txnrn-si) X-times (see 9.55, p. 168; 24.12 end, p. 320). Compare nammaa t\+
HURSAG
Zukkuki rorn-pa uet Then he came back to Mt. Zukkuki KBo 5.6 i 1 (perfective)
versus nu kuitman
m.GI
orunt-i-i ts:i xtn
URU
Mizri rorn-pa uet While attuaziti
was coming back from the land of Egypt KBo 5.6 iii 26 (imperfective). The aspect
indicator in the second sentence is kuitman: nothing in the form uetneither sux nor
inherent lexical component of uwa- dictates or indicates aspect. Likewise, nu otu-
li turier They yoked them like cattle KBo 3.34 i 16 (OHlNS) versus triyanzimaa
maiyanki nu xsxi-i xsxi-ipat t\+ 7.rxt.nr. anda pennekezzi But as many
times as they hitch them up, each time he drives them seven rxts KBo 3.5 ii 1315.
1. Perfective aspect, in which an action is viewed as an indivisible whole, should not be confused with
the perfect tense, which refers to a completed action.
00-HittiteGrammar.indb 317 6/5/08 4:19:15 PM
318
Here triyanzi refers to iterated action just as much as the sux-marked pennekezzi
and the contextual marker xsxi-i xsxi-i.
Sux Markers
24.4. Most but not all Hittite verbs may also optionally mark imperfective aspect
explicitly by the addition of one of three suxes -kela-, -(a)-, and -annali- (hence-
forth for simplicity -ke-, -a-, and -annali-). Certain classes of verbs, however, do not
take the sux, because their inherent meaning either already implies imperfectivity or
is incompatible with it. The stative verbs akk- to know, ar(k)- to hold, keep, and
ar- to stand and the entire class of denominal statives in -e- (many of which mean
become X in Hittite; see 10.1110.12, p. 177) and the denominal entives in -e-
(which mean become X; see 10.1310.14, pp. 177178) are not attested with an
imperfective sux. The verb iya- (m.-p.) to be in motion, be walking takes -annali-
only in the inceptive sense to set out, start to walk (see 24.18, p. 322), and the verb
akk- to die takes a -ke- medio-passive form only in the distributive sense, with and
without a subject: akkikettari there is much dying, that is, many people are dying.
See 24.16 (p. 321).
24.5. The suxes -ke- and -a- are often labeled iterative or iterative-durative,
while -annali- is usually said to be durative (e.g., HE 7375 137, 141). This descrip-
tion is misleading in two respects. First, as established by Bechtel (1936) and Dress ler
(1968), the use of the suxes -ke- and -a- is much broader than implied by the name
iterative-(durative). Friedrich (HE 141 and 269) actually recognizes many, though
not all, of the variety of meanings cited below. Second, contrary to the claim of Friedrich
(HE 137 and 269e) and others, the use of -annali- is completely equivalent to that of
the other two suxes; it is not limited to durativity.
Nuances of Imperfective Aspect
24.6. At least the following dierent meanings are attested for the three suxes
-ke-, -a-, and -annali-.
24.7. They may have a progressive meaning, describing an ongoing action, in some
instances setting the scene for another actionso-called backgrounding (see Bechtel
1936: 5257, against Dressler 1968: 18588, and others). For example: takku i-a
otusi u-an znukezzi If a man is driving his ox across a river (and another person
pushes him o, seizes the oxs tail and fords the river [all non-ke-]) KBo 6.2 ii 30 =
Laws 43 (OS); nu annikemi kuin nankan si+si alzii I call by his name the
one whom I am treating (KUB 12.63 obv. 28). Here the -ke- forms provide the time-
frame in which the other events described by the non-ke- verbs take place. In a further
example the contemporaneity of the two actions (backgrounding) is further marked
by kuitman while: kuitmanmaz( )an uit ssxtn ts:i so.utsi ttan lapl[ pan]
enranna ttiyannai
MUNUS
t.orma luwili kian kkikezzi While the ritual client
24.4 Verb Aspect
00-HittiteGrammar.indb 318 6/5/08 4:19:16 PM
319
is pulling the hair, lash(es), and brow(s) from his own (-za) head, the Old Woman is
reciting in Luwian as follows KUB 32.8 iii 610. Note the functional equivalence of the
two suxes in ittiyannai and kkikezzi, both of which are marking ongoing actions.
The adverb ka, which always implies an element of temporal immediacy, may be used
to emphasize that the action is going on at the moment of speaking (contemporaneity):
kattawa
LU.ME
\+st ,tot:txi\i damekatteni You are oppressing your provisions
bearers! KBo 22.1 obv. 1819 (OS). For more on the grammaticalization of ka see
24.2724.30 (pp. 323324).
24.8. Imperfective forms are also used to indicate duration of an activity over an
extended period of time. In some cases the period of time is explicitly indicated: nu
or
6
-an mandan uzurin nu.ut. azzikkanzi (the horses) eat [hay] throughout the
entire night (16.29, p. 249) KUB 1.13 ii 1011; nu
URU
[anauit]tan t\+ rft.5.xx
zaekenun and I fought against anauitta for ve months (16.75, p. 261) KBo
10.2 i 47 (OHlNS). In other cases a non-ke- verb is used to mark the endpoint of
the extended action expressed by the marked imperfective stem: nu r-rii annikezzi
kuitmna lzziatta He will (continue to) work in his (the victims) house, until he (the
victim) recovers KBo 6.2 i 1718. Sometimes the period of time is left unexpressed,
with no explicit endpoint: takku
GI
ftxti-lima mimmai i
GI
ftxti-a(!, text -u)
.X.nr. arkanta taranzi nanza i.xr tnt-tt+ annikanzi But if he refuses (to
perform) the craft, they shall declare the elds to be those of a vacating craftsman, and
the men of the city shall work them KBo 6.2 + KBo 19.1 ii 2021 = Laws 40 (OS),
with dupl. KBo 6.3 ii 4041 (OHlNS), ed. LH 4748, 18788. Alternatively, this last
example could show the inceptive use of the imperfective, translatable as and the men
of the city shall begin to work them (see 24.4, p. 318; 24.18, p. 322).
24.9. The meaning of the marked imperfectives in imperatives is often keep on . . .
-ing (which can be viewed either as progressive or durative): uitimu aul atreke
nammaan +\+ r:+ ror.nr.-wa ar (a) k Keep sending greetings to me, O lord.
And further keep (your) eyes on my house HKM 27:2325 (MHlMS); nuan par
kalnkanza nutta kuit memikemi numu urorn-ti+ itamanan lagn ar (a) k nat
it[am]ake Be appeased, and with respect to what I am saying to you, O god, keep
your ear inclined to me and keep listening KUB 24.1 i 1517. Note that the inherently
stative verbs - be and ar (k)- hold, keep do not take the -ke- form, including in
the periphrastic construction with the participle plus ar (k)- (see 24.4, p. 318).
24.10. Likewise, the use of a -ke- form with the negative l sometimes carries the
meaning of an inhibitive; that is, it means stop . . . -ing versus the usual prohibi-
tive meaning dont . . . ! (see 26.16, p. 344): umemaaza zr-az kuwat dakatteni
kinunakan ap x.n.xr +\+
m
Duppi-
d
t ara l dakatteni Why do you keep
willfully taking them for yourselves? Now stop taking those civilian captives away from
Duppi-Teub! KBo 3.3 iii 2426 (NH); fo.nr.wa kue zik w[aiyai nu]warat
+\+ utxtx+ l pekei apmaati piket Stop giving the clothes you w[ear] to
Verb Aspect 24.10
00-HittiteGrammar.indb 319 6/5/08 4:19:17 PM
320
your son! But she kept giving them to him KUB 22.70 rev. 3536 (NH). But note
that not every imperfective form used with l is inhibitive. The context shows that the
following example refers to a habitual action (see 24.13): nani dugarauwanza
pikellu pidduliyauwanzamada l pekemi (But the leavened bread which I give to
you), may I (always) give it to you joyfully. May I not (ever) give it to you under con-
straint KUB 6.45 + KUB 30.14 iii 6667 (NH).
24.11. The sux -ke- and equivalents may mark iterativity, repetition of an ac-
tion either continually or intermittently: rornsi
LU
meneya itta ktta ktta or-an
uittianni tarnimaan natta alzii The m.-man walks behind. He keeps draw-
ing (back) the arrow towards this side and that, but does not let it go. He keeps crying
out ee ee (imitating a bird or animal cry? 1.8, p. 11; and 1.144, p. 49) KBo 17.43
i 1011. Note that despite the translation with keep . . . -ing the meaning here is itera-
tive, not durative (the action is repetitive, not continuous). For an example involving
separate occasions see again the example triyanzimaa maiyanki nu xsxi-i xs-
xi-ipat t\+ 7. rxt.nr. anda pennekanzi But as many times as they hitch them up,
each time they drive them seven rxts (KBo 3.5 ii 1315). Note again the equivalence of
the three suxes for the same usage. On this same sentence see above 24.3 (p. 317).
24.12. The last example cited also illustrates the fact that the use of the marked im-
perfective stem is virtually obligatory with distributive expressions such as tu-at tu-at
day after day, rft-mi rft-mi month after month, or
6
-ti or
6
-ti night after night, xt-ti
xt-tiyear by year, lammar lammar moment by moment, uddan uddan word by
word, etc.: +\+ urorn-tt+ anda tu-at tu-at memi[kezzi nu ur]orn-t++ wallikezzi
(the scribe who reads the royal prayer to the cult statue of the deity) says it to the deity
daily and repeatedly praises the deity KUB 24.2 i 12, see KBo 3.5 i 4; uddrmetta
peun nu k [tupp]i rft-mi rft-mi perantit alzeandu I have given you my words.
Let them read this tablet aloud in your presence monthly! KUB 1.16 iii 5657 (OHl
NS). The verb iya- (m.-p.) be in motion, however, does not take the imperfective suf-
x (except in the inceptive sense, 24.4, p. 318): nata i.xn +:++++ kuit xtn-e
anda lammar lammar iattar [i ] And because the enemy likewise marches moment by
moment into the land HKM 8:1214 (MHlMS). Marked imperfectives are also used
with 1-an 1-an one by one (see 9.8, p. 155). Imperfectives are normally not used
with multiplicatives in -anki unless there is also a distributive expression in the clause
(9.55, p. 168).
24.13. The marked -ke- stem or equivalents may also express habitual, regular, cus-
tomary, or characteristic behavior: ts:i
GI
ntnmazakan kuezza azzikkenun ts:i
oiyakan kuezza akkukenun atiyazakan kuedani ekikenun ts:i
URUDU
ut
10

yazakan kuiezza arrekenun the table from which I regularly ate, also the cup from
which I regularly drank, also the bed in which I regularly slept, also the wash bowl from
which I regularly washed myself (see Goetze and Pedersen 1934: 1011, lines 1619).
Note the contrast with KBo 4.2 iv 3739 (a few lines later), where the king speaks
of what happened on a single occasion (non-ke- forms). With the preterite tense the
24.11 Verb Aspect
00-HittiteGrammar.indb 320 6/5/08 4:19:18 PM
321
habitual meaning may be accompanied by the adverb kar: takku i.t
19
.it-an itt++
kuiki dauwai nama zt
9
si lki kar 1 x. xi.nnnn piker kinuna 20 o
xi.nnnn pi If someone blinds a free person or knocks out his tooth, formerly they
used to give one mina of silver, but now one gives twenty shekels of silver KBo 6.2
i 910 = Laws 7 (OS). But the use of kar is not obligatory for this meaning: nuwa
+\+ urorn-tt+ ts:i r.oi-tt+ 1 o xi.nnnn so s
5
so z.ol 1-\i:i+
KU
o.
nXn.nr.ya peker kinunawa rzr
4
araitain ir xi.nnnnmawa so s
5
so
z.ol
KU
o.nXn.nr.ya t piyr From the palace they used to give to the deity one
shekel of silver, red and blue wool, and a set of curtains, but now they celebrated the
Festival of Womanhood, and they didnt give the silver, red, and blue wool and curtains
(KUB 5.10 i 813, NH). Note again the contrast with the non-ke- forms referring to a
specic single action.
24.14. The habitual sense of -ke- is eectively lexicalized in maniyaeke- to
administer, govern, from maniya- to handle (always with -ke- in this sense except
as an innitive; see CHD s.v.). See also the remarks on nanna- to drive below, 24.22
(p. 322).
24.15. Likewise -ke- or equivalents may used for so-called gnomic statements
of general validity (see 22.5, p. 307): alkiwa maan x.i.t
19
.it otu tut
uitarra man uinukezzi just as grain keeps all humans, cattle, sheep and wild
game alive KBo 4.2 i 5859. In particular, memikanzi is used for they say, one says,
it is said. It thus introduces proverbs or popular sayings expressing customary behav-
ior (Beckman 1986), but the verb expressing the customary behavior in this case does
not appear in the marked form: +\+ utxt.x.i.t
19
.itpatkan anda memian kian
mem[i ]kanzi arnuwawa xtts-n urorn-ti+ kri tiya[zi ] Among mortals they
have a saying: A god yields (non-imperfective present tense tiyazi, 22.5, p. 307) to
the wish of a woman of the birthing stool KUB 21. 27 ii 1516.
24.16. The marked imperfective stem may be used distributively, to refer to actions
performed on a series of objects (the action is thus from a certain point of view iterated)
(Dressler 1968: 17282):
NINDA
n
I.A
ru.xt
7
.nr. kue pariyannianzi nukan anai
dakanzi natkan rorn-pa zikkanzi They take a taste of the warm breads (one by one)
(and) the sweet breads which they break (one by one) and put them back (one by one)
KUB 25.32 iii 2325. Here the focus is on doing something to the multiple objects one-
by-one. This usage is very common in ritual contexts. Note again the equivalence of
the -annali- and -ke- forms. Likewise, an action may be viewed as unfolding in several
stages or consisting of repeated movements: nukan xtts t.or
DUG
upuwaya a
anda laukezzi
DUG
upuwayama tuwarnikezzi The Old Woman pours out little by
little (from) the .-vessel onto the hearth, and she breaks the .-vessel piece by piece
KBo 39.8 ii 3233 (MH?lMS).
24.17. It is less certain whether Hittite ever similarly marks a single action performed
by multiple subjects, but a likely example is: i+++
m
Uamuwa i
URU
Arzawa mnkan
Verb Aspect 24.17
00-HittiteGrammar.indb 321 6/5/08 4:19:18 PM
322
X xtn-:t akkikettari Thus (says) Ua-muwa the man of Arzawa: If there is (much)
dying in the land (i.e., a plague in which many people die) HT 1 ii 1718; see Dressler
1968: 16263. Since the context surely suggests that there are multiple deaths, it is likely
that this is the basis for the -ke- form.
24.18. In the case of verbs that refer to prolonged activities or to processes, the
-ke- form or equivalent may be inceptive, focusing on the beginning of the activity:
manmakan
GI
uluganni par iyannai When the carriage begins to move IBoT
1.36 ii 2324 (MHlMS) (stem iya- to walk, march, move). The case of iyanna- is
complicated by the likelihood that it was eventually lexicalized as a separate verb to
go. A further example of inceptive meaning is: p[ai ]taan [
GI
]-a arkuwanza
eket He proceeded to go to sleep in bed with his shoes on KUB 24.8 i 2526 (stem
e- to sleep). Here the meaning of eke- is not habitual (see 24.13, p. 321) but refers
rather to entry into the activity of sleeping. For the serial use of pai- see 24.31.
(pp. 324.) An example of inceptive value with a process verb is: urorn.uruir-a
utxt.xr-u .n.n-az ar dir u allanuker the gods took the boys up out of
the sea and began to raise them KBo 22.2 obv. 45 (OS). Note, however, that the same
verb in its -ke- form can indicate the entire protracted course of the action: (O my god,
ever since my mother gave birth to me) numu amml [urorn:+] [al]lanukei you,
[O my god], have been raising me FHG 1 ii 1617 (OHlNS); similarly KUB 30.10 obv.
67 (OHlMS).
24.19. The inceptive value is also seen in the construction of the so-called supine
with the verbs dai- to put and tiya- to step meaning to begin tolundertake to (see
25.3725.38, p. 338). With extremely rare exceptions the supine is always formed
from the marked imperfective stem (see 11.24, p. 186; 25.37, p. 338): uwan daiten
. . . piyanniwan dait[en] you began to do . . . you began to give KBo 8.42 rev. 23
(OS); numu rnr.xr pekewan dir They began to give me troops KBo 3.4 iii 24.
Note again the equivalence of the three suxes in this usage.
Choice of Sux
24.20. The productive sux for all the functions just described is -ke-. Only a hand-
ful of verbs take one of the other suxes.
24.21. The verbs alzai-, iya- to make, do, ai-, and warrai- form their marked
imperfective stems in -a-. In NH the stem often appears as -ea- instead (see on this
problem Melchert 1984b: 15355, and Rieken 1996: 29497).
24.22. The verbs atta-, uittiya-, ikar-, iya- to walk, go, pariya-, alla(i )-, and
tu- regularly employ the sux -annali-. In OH and MH wal- also takes -annali-. A
few other verbs are attested with only one or two examples of -annali- beside more com-
mon -ke-: uek- to slaughter, ulla-, iuwa-, ipar-, laiya-, pai-, par-, peiya-,
piddai-, dai-, talliya-, weriya-. The infrequency of occurrences with -annali- in some
24.18 Verb Aspect
00-HittiteGrammar.indb 322 6/5/08 4:19:20 PM
323
cases may be due to the small number of occurrences of any imperfective sux with
that particular verb. In the case of ullale- all occurrences of -ke- are NH, while the
only occurrence in OH or MH is -annali-, which may indicate that it (like some of the
other verbs) had an original preference for -annali- but changed to the only remaining
productive sux -ke- in NH. A special case is nanna-, which has been lexicalized as
a separate verb to drive, while only naike- serves as the marked imperfective form
of nai-.
24.23. As the productive form of the category, the sux -ke- spread at the expense
of the other two. In NH one nds walike- for walanna- (also rarely alzike- for
alzi (a)-).
Sux Redundancy
24.24. As described, the added sux explicitly marks a verbal stem as expressing one
of the values given above, versus the simple verb that is functionally neutral. By a com-
mon linguistic process, there is a tendency to reinforce the expressive value of the sux
by repeating it: ukike- beside uke-, [ap] pikikemi (HKM 89: 18) alongside usual
appike-, and ekeke- and ekike- in NH texts alongside eke-. By the same process,
the productive -ke- is often redundantly added to -annali-: ewaneke-, uittiyannike-,
laiyannike-, pariyannike-, peiyannike-, allannike-, takannike-, walannike-,
weriannike-. Examples like the one in KBo 3.6 i 7172, where walikewan and
walannikewan (dai) began to beset are used equivalently in successive lines, show
that the extra sux adds no functional value.
24.25. Somewhat dierent are cases where the sux -ke- lost its expressive value
already in prehistoric times and became what is in Hittite the basic verbal stem: ike- to
smear, anoint, pake- to x, fasten, duke- to rejoice. Naturally, these verbs can, like
any other, add the sux in its productive function: dukike-.
24.26. On the formal aspects of reduplication in Hittite verbs see 10.3 (p. 173).
The precise function of the dierent types of reduplication and their relationship to the
marked imperfective stems requires further study. In some cases there seems to be a
functional overlap between the two.
ka(tta) and kma
24.27. It is customary among translators of Hittite texts to render these forms with
a mechanical behold, lo (or their equivalents in other languages). But as Honer
(1968b) pointed out, this approach misses both the clear pattern of their use with the
Hittite verbal tenses and the fundamental function of these adverbs. Previous interpret-
ers of these forms have rightly signaled their ultimate derivation from the base of the
proximal demonstrative k- this and the adverb k here denoting spatial proximity.
The implication of the proximal demonstrative origin of the forms should not be lost
Verb Aspect 24.27
00-HittiteGrammar.indb 323 6/5/08 4:19:21 PM
324
in the approach to interpreting their force in combination with the verbal tenses. Thus,
ka (and its post-OH form kma) renders both past and non-past verbal tenses more
immediate.
24.28. With preterite verbal forms the eect of the adverb is to express so-called
present relevance, a meaning carried in English by the present perfect have ( just) . . .
-ed (Honer 1968b): [(k)]atamakan utniyandan llu dun I have ( just) taken
the slanders (tongues) of the population from you KBo 17.1 i 11 (OS); kinuna ka
f
Gauliyawia tul orxr-:i+ itarkiat Your servant Gauliyawiya has become ill
KBo 4.6 obv. 27 (NH). See also HKM 25:45 (16.89, p. 300). The presence of ka
is not obligatory in order for a Hittite preterite verb to have the meaning of a present
perfect (22.14, p. 309), but its use does eectively constrain such a reading (22.15,
p. 309). In other cases the just completed past act can be rendered with have herebyl
herewith . . . -ed (see KUB 7.1+ i 6 in 16.14, p. 244). The adverbs hereby and here-
with are appropriate to some uses of ka with either the present or the preterite (see
next paragraph).
24.29. With present verbal forms the meaning conveyed by ka and equivalents
varies by context (Honer 1968b). In some cases it reinforces an immediate present
(contemporaneous with the speech act) reading is . . . -ing (right now) (see 24.7,
p. 319, for an example). attuili I wrote: [k]ma
m
Murili utxt:[+] Murili is
hereby my (adopted) son (and designated heir) KUB 1.16 ii 37 (OHlNS). The adverb
here underscores the performative force of the statement (it is the verbal declaration
itself that carries out the action described, as in I (hereby) swearlpromiselforbid . . .).
In still other cases the adverb points to the immediate future (be on the point of or be
about to): i+++
d
Inar
m
paiya kawa kya kya iyami Inara spoke as follows to
Hupasiya: I am about to do such-and-such a thing KBo 3.7 i 2122 (Illuyanka Myth,
see the translation by Honer 1998b: 12).
24.30. The precise signicance of its use with the imperative is harder to determine.
It may add immediacy (Do this right now); see the example KUB 12.58+ ii 13 cited
in 16.25 (p. 248).
The Serial Use of pai- to go and uwa- to come
24.31. The serial or phrasal
2
construction consists of a nite form of pai- to go or
uwa- to come that normally agrees in person, number, tense, and mood

with another
nite verb that follows it in the same clause. The pai-luwa- verb appears either in clause-
initial position or immediately following a clause-initial conjunction and any attached
clitics.
3
Any sentential particles or enclitic pronouns associated functionally with the
2. For a discussion of the terms serial and consecutive to describe syntactic structures in Hittite see van
den Hout forthcoming.
3. The fronting of other elements occasionally results in the appearance of the pai-luwa- verb after two
other constituents: ziga[( mn pi)]i apn memiyan rorn -pa mematti But if you proceed to repeat that
24.28 Verb Aspect
00-HittiteGrammar.indb 324 6/5/08 4:19:22 PM
325
main verb are attached to the pai-luwa- verb or the preceding conjunctiona conrma-
tion that the two verbs are in the same clause. The second verb normally carries its full
lexical meaning, while the rst only modies the meaning of the second verb or that of
the entire clause. The second verb (the full verb) also governs the use of third-person
enclitic subject pronouns (see 18.13., pp. 280.). When pai- and uwa- function as
full verbs, they require subject pronouns (see 18.14, p. 280). But when they serve as
seriallphrasal verbs, if the main verb in the serial construction is transitive or unerga-
tive, the clitic subject is omitted. This provides a further conrmation that the two verbs
are in the same clause and the rst is a seriallphrasal verb. The seriallphrasal pai-luwa-
construction is never negated, nor does the pai-luwa- verb ever take a preverb or the
marked imperfective -ke- sux. The full verb, however, may take this sux.
24.32. The dening features just cited are important in separating the serial construc-
tion as attested from two other usages.
4
First, the agreement between the pai-luwa- verb
and the full verb distinguishes the serial construction from the exhortative use of the im-
perative second singular and plural of to come, as in: euwa zaiyauwa[tati] Come
(on), let us ght one another! (KUB 34.23 ii 17); uwatten
URU
Na paiwani Come (on),
let us go to Nesa (KBo 22.2 obv. 15; OS). Second, the consistent fronting of the pai-l
uwa- verb distinguishes the serial construction from cases of two or more nite verbs
in asyndeton, including those where one verb is semantically subordinated to the other:
nu t taruzzi i He cannot open (it) (KUB 17.10 i 33); nuza maan urorn.xr
iyazzi zinni As soon as he nishes worshipping the gods (KBo 18.15, 1113). For
further examples of nite verbs in asyndeton see van den Hout forthcoming.
24.33. There is no agreement on the precise meaning of the Hittite serial construc-
tion or its origin.
5
This usage is typologically comparable to colloquial English to
go and . . . (Why did you go and do that?lLets go and get married) and similar
constructions in other languages. The English usage very often has the connotation that
the action described is a turn of events; that is, it is surprising (often but not always
unpleasantly so), or culminating, following upon, and logically completing a preceding
action. As noted below, one can often discern similar connotations in the Hittite serial
construction. However, the very colloquial status of the English construction and its re-
striction to certain dialects makes go and . . . unsuitable as a standard translation. The
best English equivalent applicable to most examples is probably the one adopted by the
CHD: to proceed to . . . (see the examples cited below). What is clearly false and to be
avoided is the frequently encountered interpretation of the phrasal verb as an impersonal
matter to him KBo 5.3+ iii 2728 (MHlNS).
4. Whether there is a historical connection between the serial construction and either of the other usages
(see respectively Dunkel 1985 and Hock 2002) is a quite separate question. For a very dierent view of the
historical development see van den Hout forthcoming.
5. For a range of views see Disterheft 1984: 22223; Dunkel 1985; Garrett 1990c: 7274; Neu 1995;
Luraghi 1997a: 4244, 1997b: 31316; Dunkel 1998; Hock 2002; van den Hout 2003; Rieken forthcoming
and van den Hout forthcoming.
Verb Aspect 24.33
00-HittiteGrammar.indb 325 6/5/08 4:19:22 PM
326
it will happenlhappened that . . .. As the agreement of the phrasal verb with the main
verb shows, there is nothing impersonal about the construction.
24.34. The suggested English translation to proceed to . . . is intended to convey
that the serial construction underscores a transition from one action to another. That
is, it seems to emphasize some direct connection between the clause containing the
serial construction and the immediately preceding context (see van den Hout 2003). It
is important to stress that this connectiontemporal, causal, or otheris inherent in
the context and would remain even if the serial construction were not employed.
6
The
characterization just given is frustratingly vague but, as the following examples show,
no other common denominator can be identied in all the various uses of the serial
construction.
24.35. The serial construction is common in past narratives, where events naturally
often occur in (temporal) succession:
d
rx-a
d
tft-i piyt ttenwa
d
tft-un uwatet[ten
p]ir
d
tft-un anekanzi nan t wemiya[nzi] The Stormgod sent for the Sungod:
Proceed to fetch the Sungod! They proceeded to search for the Sungod, but they did
not nd him VBoT 58 i 2122 (the search is a direct consequence of the preceding com-
mand; on the serial construction in the imperative sentence see 24.41, p. 329, below);
7

nu xtn.xtn.xr ma[nda a]ra ar (a) kta nat uet +ut +ut:+
m
uppiluliuma rorn-pa
wetet All the lands had gone to ruin. My grandfather uppiluliuma proceeded to re-
build them KUB 19.9 i 68 (the rebuilding grows out of the stated need to rebuild; nu
MUNUS
Itapariya xtts.itoi n. rorn-pama uet
m
Ammuna utxt.itoi n.
Itapariya the queen died, and afterwards Ammuna the prince proceeded to die KBo
3.1+ ii 3132. The second death compounds the rst. Here is the element of the un-
expected (and the culminating!) turn of events that we note is inherent in the go and,
went and in colloquial American English. As in the case of American English proceed
to, the Hittite serial usage may also underscore here the nality of the event.
8
24.36. It is clear in the last example that the serial verb uet cannot have its usual
meaning of spatial movement. In other cases the context does imply movement, but it
is hard to decide whether this meaning can or should be attributed to the serial form of
uwa-lpai-: numukan xsi mena[anda par] naita namu uet oln.xr-a kat-
tan aliyattat . . . numu xtts-:i+ kuit menaanda uet namu oln.xr-a kattan
6. It is therefore not surprising to nd nearly identical passages with and without the serial construction,
such as example 50ab cited by van den Hout 2003.
7. The use of the marked imperfective form in -ke- in this example emphasizes either the beginning
of the search or its duration. This and other counterexamples cited below contradict the claim of Dunkel
(1985: 63) that the serial construction marks perfective aspect (on which see 24.224.3, p. 317). For the
use of the present tense here see Melchert 1998: 41617.
8. This frequent nuance of the serial construction likely reects the inherent lexical aspect of the motion
verbs pai-, which typically means to go in the sense to leave, depart or to go to (a goal), and uwa- to
come (forth) or to come to (a goal). For the terminative sense of these verbs see Starke 1977: 202 with
references.
24.34 Verb Aspect
00-HittiteGrammar.indb 326 6/5/08 4:19:24 PM
327
aliyattat He sent his mother to meet me, and she proceeded to prostrate herself at my
feet . . . Because the woman came to meet me, and prostrated herself at my feet KUB
14.15 iv 2829. It is clear that genuine movement is involved in both descriptions of this
single event, but one may argue that it is expressed in -mu . . . menaanda uet in the
second instance but by -mu . . . menaanda par naita in the rst, not by the serial uet.
The latter may instead underscore that the womans act was an immediate consequence
of being sent by her son. Note also once again the likely connotation of surprise: the
womans act is presented as being almost unheard of and quite unanticipated. That
is, she actually did this, and the king is so impressed that he complies. The fact that the
second time the action is described it is not in the serial construction ts this idea. It is
only a real surprise when it is rst mentioned.
24.37. The serial construction also is well attested in ritual contexts, where a suc-
cession of events is common: 12
UZU
n.nr.yaan +:++++pat iraizzi naza uezzi
rorn-pa par(a)za ezi She (the practitioner) also makes the rounds of his twelve
members in the same fashion. He (the client) proceeds to lie down backwards KUB 9.4
ii 2830 (the clients action permits the practitioner to repeat the treatment in another po-
sition). Here there is no element of surprise, but there is a culminating aspect: the serial
construction was not used to describe his rst lying down; only his complementary lying
down backwards, which completes all possibilities. Another example: nu
LU.ME
alliyar
LU.ME
i.zt
9

LU
palwatalla
LU
ktaa t::t
GI.d
r.nr.pat iyantari nuza pnzi
+s+usi\i appanzi The halliyara-singers, the performers, the clapper, and the kita-
man walk with the aforementioned Is:+u-instruments, and they proceed to take their
places KBo 4.9 v 4245 (the positioning of the various participants not only follows
directly upon their entrance together but also is a culminating act marking the end of the
preceding walking). One may note again that the actual movement involved is expressed
overtly by the iyantari of the preceding sentence.
24.38. One future event can also be viewed as directly dependent on another. For
reasons that are not yet clear, all examples of declarative statements attested thus far
show uwa- in this usage of the serial construction (for one account of this see Rieken
forthcoming, but see also 24.42, p. 329, below): nui eni oro awan ara namma
tittanut na adduledu namma nu uezzi utxt.xtts oi zilatiya tuk urorn-t++
wallikezzi si+-anna tulpat s+ urorn-t++ memikezzi Remove that illness from
her and let her become healthy again, and the princess will henceforth proceed to praise
you, O deity, and speak your name, O deity KBo 4.6 obv. 1720 (the action of the
princess is promised as a direct quid pro quo if the deity restores her to health. Note
again that there is a dramatic turn of events after she has become healthy. Now she is
able to do the following two actions: for the rst time she will begin to praise the deity,
etc.);
9
mnza 1-i\ tnt-ti+ma kuinki aati nukan kma \tis urorn-tt+ arratti
9. The use of the marked imperfective in -ke- may once again be taken either in an inceptive sense
(will begin to praise, see 24.1824.19, p. 322) or as expressing duration (will praise continually and
indenitely, 24.8, p. 319). Likewise imperfective and contradicting the claim of Dunkel (1985: 63) is the
Verb Aspect 24.38
00-HittiteGrammar.indb 327 6/5/08 4:19:25 PM
328
nan uwammi i.xn-a iwar oti-ami But if you occupy some one city, you will
transgress the oath, and I will proceed to strike it in the fashion of an enemy KBo 4.3+
i 2224 (breaking the oath will bring certain retribution; notice how often in our ex-
amples this construction occurs where there is a prominent change of subject); nankan
tuzziya anda uwate mn tma uwai ar (a) ki Bring them (heavily armed troops) to
the camp. But if not (= if you dont), you will proceed to die! HKM 35:69 (MHlMS).
The addressee is threatened with sure punishment if he fails to carry out the order, both
a dramatic turn of events and terminative in a very concrete sense.
24.39. In another set of uses of the serial construction the connection is not to a
previous action or event but to an expressed or unexpressed wish of a person or de-
ity. This usage takes several forms. One is seen in deliberative questions,
10
mostly
commonlybut not exclusivelyin oracle questions of the following sort: paimikan
d
tft-st antuan t\+
URU
ama par nei nuan paizzi t\+
URU
ama +\+
d
ts:+u
,iut mukear pedipat pi nammai rzr
4
-an izzi memiyanua r+\t urorn-tt+
auli memai Should I, My Majesty, therefore dispatch a person to amua? Should he
therefore give a mukear rite for
d
ts:+u of the Battleeld on the spot in amua, then
make a festival for her and speak friendly words before the deity? KUB 32.130 49
(MHlMS). Although the phrasal verb pai- must be repeated in the second clause, since
it changes from paimi to paizzi (another turn of events), it is unnecessary in the two
clauses which follow, since there is no further change of subject. This deliberative use
is also seen in the following series of questions: nu panzi +\+ urorn-tt+ ts:i xi.or

4

TUG
termazziya pianzi (answer: no) nu pnzi +\+ urorn-tt+
TUG
termaz t-azziya
pianzi (answer: no) nu pnzi +\+ urorn-tt+ 1 fo pianzi Should they therefore
make a gift to the deity with gold, jewels, and ne garments? (answer: no). Then
11

should they make a gift to the deity with ne garments and persons? (answer: no).
Then should they make a gift to the deity of one garment (answer: no) KUB 22.70
obv. 5658 (series of oracle questions). The restriction here to deliberative questions is
signicant, because in non-deliberative oracle questions of the type Will the enemy do
such-and-such? the phrasal construction is not used. In the deliberative oracle questions
the proposed actions are presented as following from the suspected will of the deity
being questioned: in eect, the meaning is Do you wish that one . . . ? Questions us-
ing the serial construction are restricted to the deliberative type, but not all deliberative
questions are required to do so.
24.40. Still another example has nothing to do with oracular inquiry but shares the
notion should Ilwelhelthey in the sense of do you wish that Ilwelhelthey . . . ?. It is
serial construction in the nominal sentence: nuwa u[(ezzi)] tulp[a(t

karimmi)] nakkiyaan And your

temple will proceed to be revered KUB 30.19+ iv 45. It is clear that the state of reverence is to continue
indenitely. See also the example with expressed be in KBo 17.88+ iii 2324 (MS).
10. On questions in general see chapter 30.
11. This English then is not intended to be temporal but in that case, continuing the previous there-
fore in alternative clauses. See p. 329, n. 12.
24.39 Verb Aspect
00-HittiteGrammar.indb 328 6/5/08 4:19:26 PM
329
a question put to attuili III, who is being pursued by Uri-Teub with lethal intent:
paiweniwarankan kuennumni Should (= do you wish that) we kill him (scil., Uri-
Teup)? KBo 6.29 ii 25 (ed. Goetze 1925: 109). This construction is another in the
type of the man and nman clauses (23.1023.16, pp. 314316) which supply the
lack of a subjunctive or optative mood in Hittite with which to express the modalities
of potentiality and wish (see 22.1, p. 306; 23.1, p. 313; 23.10, p. 314).
12
24.41. The use of the serial construction to mark an action that ows directly from
someones will naturally also occurs with imperatives: 5 r.xrt numa r.xr
taggata pnduwaz aandu nuwaza azzikkandu akkukandu idlumamakan
l ku[itki] taggai As for his ve brothers, he built houses for them: Now let them
proceed to settle down. Let them eat and drink. Let one do no evil to them. KBo 3.1+
ii 1315; ln
m
apartayakan kuin t\+ xtn
URU
Gaga par neun uedduwa utxt
m
aparta anda w[e]miyaddu And let the servant of aparta whom I sent out proceed
to nd the son of aparta (MHlMS) HKM 66:2023; paidduwa dugarazpat tu
NowlSo let there be joy KBo 17.88 + KBo 24.116 iii 2526. Note again the clearly
imperfective sense of the last example. See also the example of a second-person impera-
tive in 24.35 (p. 326) above.
13
24.42. Yet to be discovered is the rationale for the choice between pai- and uwa- as
the phrasal verb. Rieken (forthcoming) has argued that pai- marks an action in proxim-
ity to a reference point, while uwa- indicates that there is a temporal distance between
the action and the reference point. This generalization seems to hold for cases of pai-
but by no means for all examples of uwa-, as shown by several passages cited above.
It is not credible that there is any signicant temporal gap between the sending of the
mother to fall at the knees of the Hittite king and her doing so (KUB 14.15 iv 2829;
24.36, p. 327). The same applies to the ritual treatment described in KUB 9.4 ii 2830
(24.37, p. 327). And surely the threat of capital punishment for failure to carry out an
order is meant to apply immediately (HKM 35:69; 24.38, p. 327). There is likewise
no reason to think that the king does not want the missing person found forthwith (HKM
66:2023; 24.41). These and other counterexamples still leave the basis for the choice
of pai- or uwa- unclear.
12. Here too one can capture the desired nuance with English then: Shall we then kill him?. But it is
important to note that this is an implicational use of then, implying a preexisting intention or wish (in this
case of the king, in the oracle questions of the deity). In this example, which begins direct speech and where
there is no preceding action, a temporal then, thereupon makes no sense (contra Disterheft 1984 et al.).
13. We have included now in our translations of these sentences because we believe it captures the
force of the serial construction that the content of the expressed command is to follow directly upon the wish
of the speaker. Compare the use of deliberative then in the oracle questions above.
Verb Aspect 24.42
00-HittiteGrammar.indb 329 6/5/08 4:19:28 PM
330
Chapter 25
NON-FINITE VERB FORMS
25.1. The Hittite language has four non-nite verb forms (11.18, p. 185): the verbal
substantive, the innitive, the supine, and the participle.
1
25.2. The forms of the verb in -war, -mar (singular genitive -wa, -ma ) and -tar
(singular genitive -anna) preliminarily discussed in 4.114 (p. 130) and 11.20 (p. 185)
mark what is called the verbal substantive. The verbs which form their verbal substan-
tive with -warl-mar form the innitive with -wanzil -manzi; those with verbal substan-
tives in -tar form the innitive in -anna (see 3.31, p. 76; 11.2011.21, p. 185). The
choice of ending is wholly a matter of the class of the verb and in no way reects a
dierence in meaning or usage.
25.3. The terminology of the various verbal nouns in Hittite is summarized by the
following table.
Verbal Substantive Innitive Supine Participle
Ending -war, -mar, -tar -wanzi, -manzi, -anna -wan -ant-
Verbal Substantive
25.4. The verbal substantive names an action, as in English Walking is a healthy
activity. With extremely rare exceptions (see n. 35, p. 185) the Hittite verbal substan-
tive is restricted to the nominative-accusative and genitive cases.
25.5. The verbal substantive in the nominative-accusative may function as the sub-
ject: k kuit daliyawar srs-at (now with regard to) this omission which has been
determined KBo 14.21 i 28; wekuwar (this composition is) a wekuwar [an askingl
requesting] in the colophon that follows KUB 15.5+ iv 40. The verbal substantive is
the citation form used to translate Akkadian innitives in Sumerian-Akkadian-Hittite
lexical texts: (Akk.) alu = (Hitt.) punuuwar, (Akk.) itaalu = (Hitt.) punukewar,
(Akk.) uzzuzu = (Hitt.) kattaan arnumar KBo 1.44 i 1113, ed. MSL 17:101.
25.6. In the nominative-accusative it may express the direct object of a transitive
verb: [nuza m]aan +ui:+ wal (a) uwar auzi when my father saw the attacking,
(he drove up to Tuwanuwa) KBo 14.3 iv 39, ed. Gterbock 1956; (with regard to the
1. On the rst three of these see Kammenhuber 1954, 1955a.
00-HittiteGrammar.indb 330 6/5/08 4:19:28 PM
331
shipments which I have been sending to you, if I send you an aul-gift,) zikmaatza
par dammenkuwar alzeatti you call it forming an attachment (perhaps even ingra-
tiating oneself) KBo 18.24 i 47 (see Otten 196869: 11213 and Mora and Giorgieri
2004: 92, 95); nuza +\+ xn uwtar . . . iyanun I made for the army an inspection
(lit., a seeing) KBo 4.4 iii 28; ror.nr.-amu uwtar pi give me the seeing of (my)
eyes! KUB 27.67 ii 65.
25.7. It may appear as an adnominal genitive: taknaza duwa ssxtn the ritual
of taking (something) from the earth KUB 17.18 iii 20, see Taracha 1990; maanma
d.KU
kura . . . irauwa pedi ari but when the (deied) hunting bag reaches the place
of the irawar rite KUB 20.25+ i 45 (JNES 20:9293); r.xtn.n triyawa a
draft horse (lit., horse of hitching up) KBo 6.2 + KBo 19.1 iii 43 = Laws 64 (OS).
25.8. The verbal substantive occurs frequently as a free-standing genitive without
expressed head noun (see 16.61, p. 256). In this usage it often has a potential or deon-
tic sense (. . . -able or to be . . . -ed), especially when negated: mn
LU
iu+ui+ma
kuedani (var. [kued ]ani [k]i ) uezzi [(na mn) t\+] r.urorn-tt+ ar puwa but if a
privileged outsider comes to someone and is one of going up (i.e., one allowed access)
to the temple KUB 13.5 ii 1112 (with restorations from KUB 40.63 i 17); mna
arkanna if he is one of perishing (i.e., one deserving the death penalty) KBo
4.10 i 10; nwa one of fearing (i.e., respectful) att. iv 55; kui ara tarnumma
he who is to be released (from military duty) KUB 13.20 i 11; nu [(annanu)mm]a 6
o xi.nnnn pi he shall give six shekels of silver as the (price) of training KBo
6.26 iv 2930 (Laws 200b) with dupl. KUB 13.14 + KUB 13.16 rev. 67, ed. LH
15859; note the following negated examples: kuitma ur-ar umel t tar (a) wa
whatever litigation (proves) impossible for you (plural) (to resolvelsettle) KUB 13.20
i 36;
D
Aranzait t mazzuwa (I have impregnated you) with the Tigris River, not to
be resisted KUB 33.120+ i 32 (NS); nuwa memiya[()] n[akki i-itm]awara
ara p[e]iyauwa[()] the matter is important; it is something not to be cast aside
KUB 33.93 + KUB 33.96 iv 46 (Ullik. I A). This use of the verbal substantive has a
parallel in the use of the action nouns in -tar as free-standing genitives (see fr-anna
(= uiwanna) he of (long) life, long-lived KBo 3.6 i 13 (= att. i 15), 16.61,
p. 256).
25.9. The verbal substantive, having both a nominal and a verbal aspect, can take its
logical object in either the genitive or accusative case. When its nominal aspect is prom-
inent, it takes its object in the genitive: ammukmanwa ammel [u]+++\t:+ ammella
xtn-ea tepnumar tametani xtn-e atranun Would I have written to another country of
my own and my countrys humiliation? KBo 5.6 iii 54-iv 2 (D frag. 28); r:+mamu
kuit ki()an :+sriu s+ utxt.xttswa zaluganum[ar it ra] Concerning what you,
my brother (referring to Ramesses II), wrote to me, saying: Detaining the (prom-
ised) daughter [is not permitted] KUB 21.38 obv. 34 (letter of Puduepa); i.xr
xtn
URU
Mizrama maan s+ xtn
URU
Amqa oti-auwar itamaanzi but when the
people of Egypt heard (lit., hear) the attacking of the land of Amqa KBo 5.6 iii 56;
Non-Finite Verb Forms 25.9
00-HittiteGrammar.indb 331 6/5/08 4:19:29 PM
332
nuttakkan
d
tft- kuitki kuggurniyauwar rorn-pa anda udai and he brings up
again some slander of His Majesty KUB 23.1 iii 1213 (aukamuwa); :r:-ma

UR.SAG
Ammana uwauwar kuit namma memekei Why do you continue to speak yet
again of brotherhood and of coming (to) Mt. Amanus? KUB 23.102 i 56 (NH). When
its verbal aspect is prominent, it takes its object in the accusative: nu
m
Ipudainara
maniyaekezzi or-an
GI
txnr aauwar
GI
ftxti apptar Ipuda-inara taught
(historical present shows or teaches) them (how) to smoothlsharpen an arrow (or) a
wheel (and) to hold a weapon KBo 3.34 ii 2829 (Palace Chronicle, OHlNS). Perhaps
even the dative-locative: nuza xn: uwtar . . . iyanun I made for the army an
inspection (i.e., I inspected the army) KBo 4.4 iii 28. The logical subject of the verb
contained in the verbal substantive can also occur in the genitive case: ror.r.-amu
uwtar pi give me the seeing of (my) eyes! KUB 27.67 ii 65.
Infnitive
25.10. The Hittite innitive is unmarked for voice and may equate to the active or
passive innitive of other languages (English to give or to be given respectively).
25.11. The innitive, both when it depends on a nite verb and on a noun, expresses
purpose. Depending on a nite verb: nuwakan . . .
m
Pipitain a[ p]iyawanzi [ par]
neun I sent Pipitahi out to do reconnaissance HKM 17:1617 (MHlMS); takku
ppatriwanzi kuiki [ p]aizzi If anyone goes to requisition KBo 6.26 i 28 (Laws 164,
OHlNS); sometimes the verb governing the innitive is itself non-nite (e.g., a parti-
ciple): allall pwanziwaza tot linkanuwanza I have been made to swear with the
regard to defection KBo 4.14 ii 4647 (treaty). Depending on a noun: 1
DUG
ania
or:fr ipanduwanzi one jug of wine for libating KUB 7.53 i 23 (rit. of Tunn.).
25.12. The innitive can depend on a (predicate) adjective: eki nXo-ni itoi-a
xsxi-a takuwanzi
GI
sn.or:fr-a tuuuanzi (var. KBo 6.3 iii 24 tuuwanzi)
[(
LU
tntot.o)]n natta kuiki arawa No one of the metalworkers shall be exempt
with respect to making ice, a fortication wall, and a kings road, or harvesting vine-
yards KBo 22.62 + KBo 6.2 iii 2122 (Laws 56, OS);
URU
Akitumamaa srsxtn-
ear anda ukiyawanzi kuit anezzi she (i.e., my mother) is an oering of the Akiti
festival which is pleasant to look at RS 25.421:5456 (ed. Laroche 1968).
25.13. The innitive can be construed so that its logical direct object stands in the
accusative case (note also the variety of auxiliary verbs and their force): anzelzaka[n]
rnr.xr:-an rnr.xr: i.xn waluwanzi zikkezzi The enemys troops will begin
to attack our troops KBo 10.7+ iii 1516;
GI
armiz[ziwa]
4
wedumanzi kar
zinnanda They have already nished building the bridge with stones HKM 72:46
(MHlMS); nuza
URU
Tuwanuwan zaiyawanzi pzi he began to ght Tuwanuwa KBo
14.3 iv 22 (D frag. 15); [ ppa]makan
m
Attariiya i
URU
iy ara uet nu rorn-
an tukpat
m
Madduwattan kunanna an[ike]t Afterwards, Attariiya, the man of
25.10 Non-Finite Verb Forms
333
Aiy, came, and was seeking to kill you, Madduwatta KUB 14.1 obv. 60 (MHlMS);
nuza pait
URU
Alminan wetummanzi t,u+: he proceeded to begin fortifying Almina
KBo 5.6 i 9 (D frag. 28A); nuwaza utxt-a kui[ zik] [nuw]arankan taparuna
kuwapi paii And you who are but a child, where will you go to rule him? KUB 19.29
iv 2021, ed. AM 2021;
md
ixx-aa kue xn.nr. t\+ xtn
URU
Nuai alkiu
arninkuwanzi peudan arta The troops that Kurunta had led to Nuae in order to
destroy (their) crops KBo 4.4 ii 6364, ed. AM 120121; nu uwai s+
UR.SAG
Taa
NA
4
perunu ts:i so.utx+ oti-auwanzi zinnikei You will end up striking the
rocks of Mt. Taa with your head KUB 33.120 i 3436 (Song of Kumarbi); mn
NINDA
taparpau arrumanzi taruptari when the distribution of taparpau-breads is
completed KUB 10.89 i 1617.
25.14. The preceding examples with common-gender nouns show that the following
with neuter nouns are also direct objects in the accusative: u tzt.l uwappanna
tzt.l uwanna to see the favorable inner organ and the unfavorable one KBo 3.21
ii 910 (MHlNS); [nu kuit] r.xr utxt.itoi danna ilaliyanzi [And because] they
want to take the estates of the prince(s) KBo 3.1 ii 63 (OHlNS); liiyalattama
nepia daganzipaa uddar kattan ara petummanzi It is for you to take along the
oracular words of heaven and earth KBo 3.21 ii 67 (hymn to Adad, MHlNS); mnza
LU.ME
xrsi.itn r.xr ornri anewanzi appanzi When the courtyard washers begin
plastering the new building(s) KUB 29.1 iii 29 (without +\+ the logogram r.xr ornri
is probably accusative, not dative-locative); maanmaza alkuear anduwanzi
zinni nuza rzr
4
namma iyauwanzi pzi when he nishes preparing the materials
(for the festival), he will then begin performing the festival itself KUB 27.59 i 2325;
nat +\+
d
tft-si uwanna ander and (the lords) arranged for His Majesty to see them
(aforementioned chariots) HKM 63:2021 (MHlMS) note that the logical subject
here is in the dative-locative. The uses of MH liliwauwanzi in the Maat letters (HKM
13:1012, HKM 15:1013, HKM 20:612, etc.) are adverbial (hastily), with the ac-
cusative objects depending upon the main verb in the clause.
25.15. More often, the logical direct object of an innitive stands in the dative-
locative:
d
Anumata
d
r.ii-aa . . . +\+ i.xr xnsi\i wemiyawanzi tuk
wtarnaer Anu and Enlil . . . commissioned you to nd their enemies KBo
3.21 ii 1213 (MHlMS); kuedamakan +\+ rzr
4
.nr. oi.nr. s+ rzr
4
.xr rft
anumanzi andandat The great festivals of the monthly festivals that were determined
to be performed ABoT 14+ v 911 (CTH 568, oracles); rxr.nr. rxr.nr. kuwapiwa
paitte[ni]
NA
4
peruni paluna paiweni . . . +\+
NA
4
x duwarnuma[nzi p]aiweni tn.xn
or-ruanzi xr.xr tn.nn.n patalauna xr.xr .zammanti utxt.rf lalauna xr.xr
Tongues, tongues, where are you going? We go to pala- the rock, . . . we go to break
the x-stone, we go to tarwi- the lion, we go to patala- the wolf, we go to release
the zammant-boy (from a spell) KUB 44.4 + KBo 13.241, rev. 2225 (ritual incan-
tation); xtts.itoi
f
Pudu[(epakan kuwapi)]
m
tn.xn-i-in [(oi utn.sn.
xr
U
)]
RU
attu[(i +\+ )irr+
I.A
)]
URU
Kizzuwa[(tna )]anwanzi [(weriyat)] nata k
Non-Finite Verb Forms 25.15
00-HittiteGrammar.indb 333 6/5/08 4:19:32 PM
334
)ir[(r+
HI.A
s+)] rzr
4
iu[(wa ap)]iy[(a tu-at ara aniyat)] When Queen Puduepa
commanded Walwaziti, (who is) chief of the scribes, to seek in atti for tablets of
Kizzuwatna, on that day he copied out these tablets of the Festival of () iuwa KBo
15.60 vi 311 restored from duplicates KBo 7.45 rev. left 69, KBo 40.65:17 and
KBo 15.52 vi 3945 (colophon); mn t\+ tu.2.xx lukkatta nu xtts.xr uktu[riy]a
atia lewanzi pnzi At dawn on the second day women go to the pyre to collect
the bones (dative) KUB 30.15 i 12 (OHlNS).
25.16. In some cases we cannot be sure if the object is in the dative or accusative:
apmamu arkanna . . . anta but he sought to kill me (lit., he sought me for
perishing) att. iii 63; mnmu idalauwanniya kui waggariya[w]anzi anazi if
someone in wickedness seeks to rebel against me KUB 21.47 + KUB 23.82 rev. 23; nu
kui
LU
on or-,t ntn.so
GI
innaa karuwanzi paizzi He who is a carpenter goes
to the mountain to cut beams KUB 29.1 iii 1415 (rit., OHlNS) (innaa in NH copy
could be accusative plural, but is probably dative here); see also KBo 22.62 + KBo 6.2
iii 2122 above sub 25.12 (p. 332). The following could be an example of the logical
object in either the dative or the accusative, depending on whether +\+ or the ending
-an is to be taken more seriously: nama +\+
m
Duppi-
d
rx-upan t\+ xtn
URU
Amurri +\+
itoi-i::t+ ara tittanummanzi takkezi of (if) he plots to remove Duppi-Teup
(logical object, but dative) from kingship in Amurru KBo 5.9 ii 3436 (Mur. II).
25.17. In some instances the logical object of the innitive appears in Hittite as the
grammatical subject of the clause, while the innitive is dependent on the verb to be
or other predicatival expressions: ru.xtn
4
.n pariyawanzi t oi There are no
loaves for breaking KUB 12.12 v 32; [k]zmakan
URU
Tiyailta +oi .X .oXn si
[t\]+
URU
Timmuala ipanduwanzi anda appanza in this direction (the city) Tiyailta
with its elds and meadows (is) combined with (the city) Timmuala for the purpose
of being made sacrosanct (to a deity) KUB 19.37 ii 2122 (NH); ts: [i xtn]
URU
xi.
nnnn-:t
LU
+i\\+u:i+ rorn-pa piyanna t ra a fugitive (nom.) is not right for
giving back from the land of atti (i.e., it is not permitted to give back a fugitive from
the land of atti) KBo 5.4 obv. 38 (NH);
LU
+i\\+u:i+ rorn stx-wanzi t ra It is
not permitted to give back fugitives KUB 19.55 rev. 4 (NH); nui otu pugari
piyawanzi srs-at A substitute ox was determined for him for giving (if from pai-;
or, for sending there, if from piya-) KBo 4.2 iii 50 (NH); mn tnt-ti+ kui . . . +\+
m
Ulmiteup piyanna it zr-anza (lit.:) If some villagelcity is not the wish to give to
Ulmi-Teup (i.e., If it is not (His Majestys) wish to give some villagelcity to Ulmi-
Teup) KBo 4.10 ii 18.
25.18. When there is an impersonal verb such as it doesnt succeed that or it
doesnt happen that, the logical subject of the innitive can stand in the dative: mn
tukma [warr ]iuwanzi t kiari But if you are unable to [as]sist KBo 5.9 ii 19 (Mur.
II); nunnakan epurawanzi t apdat We couldnt epura- KBo 18.54 rev. 1415
(letter); according to one possible interpretation, the passages BrTabl. i 9697 and ii 12
(Tud. IV) would also serve as examples.
25.16 Non-Finite Verb Forms
00-HittiteGrammar.indb 334 6/5/08 4:19:32 PM
335
25.19. Often the innitive combines with a nite verb idiomatically, as if the latter
were an auxiliary verb.
25.20. The innitive combines with the auxiliary verb pp- and -za with the mean-
ing to begin to do something (Goetze 1925: 89; SV 2:154; and HW
2
2:6465). This
construction is attested once in a New Hittite copy of an Old Hittite ritual: mnza
LU.ME
xrsi.itn r.xr ornri anewanzi appanzi When the courtyard-washers begin
plastering new buildings KUB 29.1 iii 29 (OHlNS). But since there is no other example
in either OH or MH, we should probably attribute this construction to the NH scribe
rather than his OH archetype. A few securely datable New Hittite examples: nuza pait
URU
Alminan wetummanzi t,u+: He went and began to fortify Almina KBo 5.6 i 9 (D
by Mur. II), not found in Murili IIs own military annals; numuza alwanzawanzi
namma +oi uxsi utxtsi pper They (he) together with his wife and son
began again to bewitch me att. ii 7778. In addition, the construction is found
routinely in NH ritual and cult texts: nuza urorn.xr ittiyawanzi appanzi And they
begin to draw the gods KUB 15.31 i 33, iii 48; nuza rzr
4
namma iyauwanzi pzi And
he begins to celebrate the festival again KUB 27.59 i 2425.
25.21. The innitive with dai- and -za also means to begin to (do something) (lit., to
place oneself to (do something)): andamazkan mn i.xn-a kuwpi uwaluwanzi
di Then when the enemy begins (lit., sets himself) to attack KBo 16.50 (oath of
Aapala, MHlMS); anzelzakan rnr.xr-an rnr.xr i.xn waluwanzi zikkezzi
Enemy troops will begin to attack our troops KBo 10.7 + HSM 3645 iii 1516, see ibid.
1920. See the same verbal construction without the innitive: i.xnzakan maan
URU
Kaan
URU
Taazzimunanna waluwanzi zikkezzi How the enemy begins to at-
tack the (towns of) Kasasa and Tahazzimuna HKM 27:36 (letter MHlMS).
25.22. The innitive with tiya- also means to begin to (do something): nu urorn.
xr mante +\+
d
Ullukummi . . . otu.nr. maan uwayawanzi ter And all the gods
like cows began to low towards Ullukummi KBo 26.65 iv 1920 (Ullik. III). See also:
[ . . . i]-ti+ kuwapi wawanzi tianzi KUB 31.69 obv. 8 (NH); 1 ru.xt
7
n.n.
z s+ 2 ir\[t] 1
DUG
alwatalla iXi 1
DUG
nn.nn o.ou.u +\+ +s[ut
I.A
] irauwanzi
tiya[nzi ] nu +sut
I.A
iranz[i ] KUB 25.19 vi 1317 + IBoT 4.80:48 (NS); itoi-u
irawanzi tiyazi 3si! iraizzi
d
t
URU
Zippalanda
UR.SAG
Da
d
amyu KUB 11.30 +
IBoT 4.197 iii 2225. This construction is strikingly similar to the more common use
of tiya- with the supine (see 25.37, p. 338).
25.23. The innitive with the verb zinne- is translated to nish (doing something):
maanma
LU
n +w+[:i
ME
] memiyauwanzi zinna[i ] but when the singer nishes
speaking the words KUB 39.84 obv. 12; nuza orx-an [ . . . ] srsxtn.nr. iyawanzi
zinnai And when I nish performing the rituals AT 125:2224; maanmaza
alkuear anduwanzi zinni nuza rzr
4
namma iyauwanzi pzi when he nishes
preparing the materials (for the festival), he will then begin performing the festival
itself KUB 27.59 i 2325.
Non-Finite Verb Forms 25.23
00-HittiteGrammar.indb 335 6/5/08 4:19:33 PM
336
25.24. Similar in meaning is the innitive with ar tittanu-
2
and a(a)nu-: nu tuel
s+
d
tft
URU
Arinna immu ssxtn.nr. rzr
4
.nr. iyauwanzi ar tittanukanzi They
perform fully your immu, rituals (and) festivals, O Sungoddess of Arinna KUB 24.3
i 2325 (Mur. II). With a(a)nu-: maankan xtts t.or mugauwanzi anuzi nu
aruwaizzi When the Old Woman nishes invoking (the deity), she bows down KUB
17.23 i 12; nata orx-an tuppa
I.A
-aa memiyanu anda memiyauwanzi aanuwanzi
When they nish speaking the words of the tablets KUB 17.18 ii 1516.
25.25. The innitive with andalliya- means to dare to (do something) KBo 4.4 iii
6263 (AM 13233).
25.26. The innitive with mazz- also means to dare to (do something): [m]n
apma memiyawanzi it mazzazzi But if he doesnt dare to tell (his superior) KUB
13.4 iii 76. See CHD mazz- a 1 d.
25.27. The innitive with tar- means to be able to (do something): nuwaratza
namma iyatnuwan uwi
SAR
[ pu]uwanzi l kuiki tarzi And let no one be able to
[cr]ush it, the iyatnuwan auwai plant, again KUB 29.7 + KBo 21.41 rev. 2728, but
also to defeat someone (accusative) in doing something: sisst i.xr 70
LU
otnt[z]a
iiyauwanzi tarta he defeated sixty men (and) seventy young men in shooting KUB
36.67 ii 23 (Gurparanzahu myth).
25.28. The innitive with wakkar- means to almost do something, not quite do some-
thing (lit., to lack to do something): (When ebat saw Tamiu,) nukan
d
ebadu
uaz katta mauuwanzi waqqare she almost fell down from the roof (where she
was standing) KUB 33.106 ii 78 (Ullik. III A). See also 25.34 (p. 337).
25.29. The innitive with kar- means to fail to (do something), neglect to (do
something): (Whoever crosses the threshold of the gods,) nu 1-a 1-a t\+ r.urorn-
tt+ ar wanzi lpat kartari Let him not neglect each in his turn (lit., one by
one) to spend the night up in the temple KUB 13.4 iii 56 (pre-NHlNS).
25.30. The innitive with mimma- means to refuse to (do something):
m
Maddu
wattaaz xtn
UR.SAG
riyati anna mimma Madduwatta refused to occupy the
Mt. riyati district KUB 14.1 obv. 18; ntn.so-aza arauwanzi memmai ariyaza
appnna memmai xsxi-aza karipuwanzi memmai The mountain will refuse re-
straining (or: to be restrained). The valley will refuse seizing (or: to be seized). The road
will refuse devouring (or: to be devoured) KUB 12.62 rev. 36 (ritual). It is unclear if
2. See also i.xr r.ur.ma urorn-t++ xi.or ienzi +\+ urorn-tt+ma aniuret maan nan aniy-
auwanzi +:++++ ar tittanuwanzi the smiths make a gold (statue of) the god, and they performltreat
himlit fully just as hislits protocol is KUB 29.4 i 68. The phrase ar tittanu- means to complete, nish
(HW 225 vollenden). This interpretation is to be preferred to Puhvels undertake (HED H 315), if by that
is implied an ingressive aspect.
25.24 Non-Finite Verb Forms
00-HittiteGrammar.indb 336 6/5/08 4:19:33 PM
337
the -z(a) . . . it meli (m)ma- that means to refuse is a usage of mema- to say (hence:
to say no; see CHD LN mema- 12b) or an unusual extension of the meaning of the
non-negated mimma- to refuse to its negated form: s+ r.oi-tt+maz txt.nr.
aniyauwanzi t mimatti (How can you make haste in regard to the seed of the lords
and plant the seed of the lords,) but refuse to sow the seeds supplied by the palace (lit.,
of the palace)? HKM 55:2728 (MHlMS); and see CHD LN mema- 12b for further
examples.
25.31. The innitive with tarna- means to allow (something to be done): nu
namma
d
tft-st
URU
Dukkaman tnt-an []ruwawanzi it tarnaun And then I, His
Majesty, did not permit the city Dukkama to be plundered KBo 4.4 iv 2324 (AM);
ammukmazakan pariyan puwa[nzi] t tarna But he did not allow me to go be-
yond KUB 23.87:1011 (letter).
25.32. The innitive with nuntarnu- means to (do something) hastily or precipi-
tately: urorn.xrma zr-anza dau [(nu ppuwan)]zi t nuntarnuzi But the deter-
mination of the gods is strong. It is in no hurry to seize (an oender), (and when it does
seize, it does not let go again) KUB 13.5 ii 3031 (pre-NHlNS) with restorations from
dupl. KUB 13.6 ii 14.
25.33. The innitive with iri- means to perform (an action) in a circular way: 1
DUG
xixiu xya ipanduwanzi iraizzi he libates also one pitcher of beer in a circle
or he makes the rounds libating a pitcher of beer KBo 24.45 obv. 12.
25.34. In some cases the nite verb describes the manner of the action expressed by
the innitive and is equivalent to what would be an adverb in languages such as English:
nu [k]uitman akkukanzi kuitman a[kuan]na mante irnzi [i ]amikanma ap [t]
sln nat iamiyawanzi rorn-pa a[rku]kanzi and while they are drinking, while all are
completing the drinking, that song is being sung, and they are singing it responsively
(lit., they are responding [arkuke-] to sing) KUB 25.37 + KUB 35.131 + KUB 51.9
i 3840 (CTH 771.1); maniyaia ian kuin uitsi iezi nuan apedani tnt-ri
rorn-an ror.ut
8
.nr. dnna maknut (This is the man) whom his lord makes a district
governor, and behind (the back of) the city he took tribute excessively (lit., behind the
city he increased to take tribute) KBo 32.14 iii 1315 (Song of Release). See also
25.28 (p. 336) on the use of wakkar- plus innitive.
25.35. Occasionally the innitive describes the manner in which the action of the
main verb is executed: liliwauwanzi nai Send quickly! ABoT 60 rev. 34, VBoT
2:1012; liliwauwanzi arnutten Move quickly! HKM 15:1213; liliwauwanzi
nni Drive here quickly! HKM 14:7; nu t\+ 8 +ist penniyawanzi 6 u arnuanzi
And for eight nights they move them six us at a trot KBo 3.5 i 5758 (Kikkuli
horse training manual); nu akurwawanzi aiknuanzi And they satisfy them (i.e.,
the horses) by watering KUB 29.40 ii 5.
Non-Finite Verb Forms 25.35
00-HittiteGrammar.indb 337 6/5/08 4:19:34 PM
338
25.36. The innitive with natta ra means It is not permitted to (do something):
LU
+i\\+u:i+ rorn stx-wanzi it ra It is not permitted to give back fugitives KUB
19.55 rev. 4 (NH); nu itma ra tot uanzi Or is it not permitted to spend the night
up there? KUB 5.1 i 38 (NH). The innitive with the negated impersonal (third-person
singular) verb form t apdatlapdari and the dative of the actor means couldntl
cant (do something) (lit., it waslis impossible for (the person) to (do something)):
nunnakan epurawanzi t apdat We couldnt epura- KBo 18.54 rev. 1415 (NH);
epurawanzimakan it apdari it will be impossible to epura- KBo 18.54 rev. 19.
Supine
25.37. The form of the verb in -(u)wan is called the supine. It is employed only in
constructions with an auxiliary verb, either dai- to put or tiya- to step, arrive, enter
(see HE 184c, 273, and 259c). Principal studies of the supine are Ose 1944 and Kam-
menhuber 1955a. The supine construction is never negated, and no word breaks the
nexus between the supine and its auxiliary verb (dai- or tiya-). This is in contrast to the
innitive (see nu namma kel s+ xtn.xtn-:t+ i.[xr )i+t] ++n+u
d
tft-st uwawanzi
t tarnai he doesnt let messengers from these lands come before My Majesty KUB
14.1 rev. 3031 (MHlMS); and s+ r.oi-tt+maz txt.nr. aniyauwanzi t mimatti
HKM 55:2728 above in 25.30, p. 337). In the supine the verb almost always takes an
imperfective stem (see 11.2311.24, p. 186; 24.19, p. 322). The older texts show a
preference for dai- to put as the auxiliary verb; later texts can employ either dai- or
tiya- to step, enter, although tiya- is most commonly conned to the plural: tiyaweni,
tier. The form ti(-ya)-an-zi is ambiguous, as it could be parsed as third-person plural
present of either verb. Rarer (mostly OH or MH) plural forms of dai- as auxiliary verb
are: daiten, dir and daier ; otherwise it is employed in the singular: tei, daitti, di,
dai.
25.38. One usually translates the supine construction to begin to do something, to
be ready to do something, to be willing to do something. It is found occasionally in
almost all types of text, but it is used quite extensively in only two types: the NH mytho-
logical narratives (see DeVries 1967) and the NH military annals (see Goetze 1933a
and Gterbock 1956), but especially in the former. There is a noticable complementary
distribution of the two semantically similar constructions (both meaning to begin to
. . .), supine with dai- or tiya- and innitive with pp-: where the supine construction
is found, the innitival one is not, and vice versa. Thus the innitval construction is not
found in the myths or in the annals of Murili II, but the supine construction occurs,
and in the Deeds of uppiluliuma, authored by Murili II, the innitival construction oc-
curs, but the supine construction is not found. In the Apology of Hattuili III the supine
construction is quite common, while the innitive + pp- occurs once. The innitive +
pp- construction occurs rarely prior to New Hittite, while the supine construction was
in normal use from Old Hittite times. For dai- with the innitive see 25.21 (p. 335).
For tiya- with the innitive see 25.22 (p. 335).
25.36 Non-Finite Verb Forms
00-HittiteGrammar.indb 338 6/5/08 4:19:34 PM
339
Participle
25.39. The participle expresses a state. With transitive verbs it usually corresponds
to the passive participle of other languages, expressing the state of a person or thing
acted upon: appant- seized, taken, pariyant- broken, piyant- given, aant- born,
taruppant- assembled, united). With intransitive verbs it may indicate an attained
state: akkant- having died, pnt- having gone, uwant- having come [a river in ood
stage], but sometimes also an ongoing state equivalent to an English present participle:
arant- standing, uwant- running, naant- afraid. Participles of transitive verbs
used generically can be either active (ekkant- knowing, itamaant- hearing [ear],
uwant- seeing [eye], adant- having eaten, akuwant- having drunk,
d
Wiuriyant-
the strangleress) or passive. One even nds the very same verbs used in both ways:
ekkant- both knowing (spirit) and known (person).
25.40. Like any adjective, the participle can function as an attribute of a noun, as a
predicate, or substantivized as a noun itself.
25.41. Like other adjectives, the participle is used attributively to modify a noun.
But when it does so, it rarely (see the next paragraph) precedes the noun, as do most
adjectives (17.4, p. 271), but follows it, as do the universal quantiers mant- all
and dapiyant- all (17.6, p. 271): [g]aggapan zanuandan tianzi they set out a cooked
gaggapa-animal KUB 20.11 ii 22; i
GI
ntn 2 ru.xtn
4
.n xt
7
pariyandu
uppaza
GI
ntn-za di A table-man takes two broken up sweet loaves from the
the sacred table KUB 10.52 vi 1314;
LU
sor-a waanza a fully clothed cupbearer
KUB 10.21 ii 11;
NA
4 pailu ndu warm pebbles KUB 7.53 ii 22, KUB 17.23 i 13;

karimna sro
5
-and [a] sanctuaries in good condition, tnt.uruir.nr. . . . aandu
settled . . . cities KBo 11.1 obv. 33.
25.42. Exceptional preposed participles are: witantu tnt.nr.-u fortied cit-
ies KUB 36.108 obv. 6, 8 (OS); warn paur a burning re (KBo 6.34 iv 6), ts:i
+ut:+ kaneanza t-a a person recognizedlhonored by my father KBo 4.12 obv. 9;
anda ppandu
4
.nr. stones gathered together VBoT 24 ii 20 (Anniwiyani rit.),
ed. Sturtevant and Bechtel 1935: 100126; pariyandu ru.xtn
4
.n.nr. broken
breads KUB 7.5 ii 9 (Pakuwatti rit.), ed. Honer 1987;
GI
orunt 0i walantan tut-
un a sheep never struck with a rod KBo 15.10+ ii 910 (MHlMS); and a-a-an-du-u
NA
4
pailu KBo 16.54 + KUB 32.131 + KBo 20.73 iv 18, if (as is usual) nt- is viewed
as the participle of a verb - (Neu 1968b: 1; Friedrich and Kammenhuber 197584
4445).
25.43. Like other adjectives, the participle can ll the role of a predicate (the land
is large, the situation is grave), where it can be regarded as the second component of
an equation, X = Y: k )irri ara arran ta (the prototype of) this tablet had dete-
riorated (and was recopied) KUB 33.120 + KUB 48.97 iv 3233 (colophon); arkia
NINDA
ari kar pariyanza and the white .-bread (was) already broken up KUB
Non-Finite Verb Forms 25.43
00-HittiteGrammar.indb 339 6/5/08 4:19:35 PM
340
10.52 vi 89. Negated: [apa k]uduwta l l andnpat du And let [those] false
accusations never never be established! KUB 1.16 ii 55 (OHlNS). In the preceding
examples the verb either expressed or implied is e- to be. But the neuter singular
participle of many verbs is used in the predicate with the auxiliary verb ar (k)- to have,
hold. On this see 22.1922.25 (pp. 310312).
25.44. The participle can be substantivized: uninkanza 3 o xi.nnnn di the
injured party takes three shekels of silver Hittite Laws 9; akkant- dead person,
LU
pitteyant- fugitive,
LU
maniyaant- (CHD s.v.); see also the divine name
d
Wiuri-
yanza the strangleress (see Carruba 1966a: 4952 and the review by Honer 1968b).
25.45. Since the verbal origin of the participle was not forgotten, it continued to be
used with preverbs (anda ppandu
4
.nr. stones gathered together VBoT 24 ii 20)
and with complements qualifying its verbal element, such as nouns in the instrumental
case: ts:i +ut:+ kaneanza t-a a person honored by my father KBo 4.12 obv. 9;

lima zeriyalli ou-it kariyanda kar artari potstands covered with linen cloths are
already standing in the courtyard KUB 10.21 ii 78;
3
nuan
d
Telipinu l.uio.o-it
papparanta xsxi-a iyanni O Telipinu, set out on the road sprinkled with ne oil
KUB 17.10 ii 2930 (Tel. myth, OHlMS); ai rxr-a gag qamakan parkuin
miriwantan arkin
GI
orunt 0i walantan tut-un ipantaun O mouth, tongue,
tooth! Behold I have sacriced to you (plural) a pure, gleaming white sheep never (lit.,
not) struck with a rod KBo 15.10 + KBo 20.42 ii 810 (rit., MHlMS). This recalls
similar constructions involving pure adjectives whose meanings resemble passive parti-
ciples: marnuwantet n full (i.e., lled) with marnuwant-beer KBo 21.72 i 13). This
use is more common as a predicate than as an attribute.
3. kariyanda (pl. neut.) shows that zeriyalli is a plural. The verbal force of kariyanda is complemented
by ou-it with linen.
25.44 Non-Finite Verb Forms
00-HittiteGrammar.indb 340 6/5/08 4:19:35 PM
341
Chapter 26
NEGATION
26.1. During the Empire Period Hittite possessed ve negative words: (1) the nega-
tive of assertions natta (usually written as an Akkadogram it or i-it, on which see
31.9, p. 432), rarely as Sumerogram t), (2) the negative of prohibition le-e, (3) the
negative of not yet, nwi (older spelling na(-a)--i continuing into latest periods, later
na(-a)-wi
5
unattested before New Hittite), (4) the negative of wish or potential nu-u-ma-
(a-)an (or nu-u-wa-(a-)an), and (5) nekku not . . . somehow? (HE 27983). All ve
of these words have been comprehensively treated in the CHD, volume LN (1990).
natta
26.2. In Hittite the primary negative of assertions na-at-ta (Akkadian it) may
modify any sort of predicate except a verb in the imperative: (1) present or preterite
indicative nite verbs, (2) adjectives or participles, (3) nouns (including verbal sub-
stantivessee 25.8, p. 331), (4) adverbs. For examples see Honer 1986: 9192 and
CHD LN sub natta b. In addition, natta may modify elements other than the predicate:
(subject) nattaan k tarnaun itoi-an xtts.itoi-a tarna It was not I
who released it, the king and queen released it KBo 17.3 iii 45 (OS); (direct object)
natta apn orfr-an piyr It was not that wine that they gave KBo 3.34 ii 5 (OHl
NS); (other) utxt.i.t
19
.it t innar uwanun tma ullanni uwanun O mortal, it
was not rashly that I came, not in wantonness that I came. See CHD LN sub natta c.
Finally, natta may also modify non-predicatival verbal substantives and participles: t
uwawa one of not coming (i.e., one not permitted to come); t naariyawanza the
unafraid; t pittuliantan the unintimidated; t zanuwanti uncooked.
Word Order with natta
26.3. The usual word order is natta (it) (or other negation: l, nwi, nman) imme-
diately before the verbal form (Honer 1986: 86): takku ume natta aktni If you do
not know KBo 22.1:5 (OS); luzzi natta karpiyezzi he shall not perform corvee work
KBo 6.2 ii 40 (OS). With preverbal constructions natta (and other negatives; for l see
26.22, p. 345) breaks the nexus between preverb and verb: nu namma ara it tarni
he doesnt let go again KUB 13.4 ii 2324 (pre-NH); nuwaramu par t petni
and (if) you dont hand them over to me KUB 14.15 i 15 (NH). There are occasional
exceptions to this rule with anda: t anda eun I was not present KUB 21.19 i 35, ii
5; nukan itma anda alikmi KUB 5.1 i 29, but no exceptional examples with ara,
00-HittiteGrammar.indb 341 6/5/08 4:19:36 PM
342
katta(n), par, or ar. For further examples, exceptions, and discussion see Honer
1986: 8689.
26.4. The negative natta can be fronted and separated from its nite verb, often
in rhetorical questions (Honer 1986: 8990): natta k u.xr-u ntn.so.xr-u
arunua [rorn-p]a tarmakimi Is it not I who restrain the rivers, mountains and
seas? KUB 31.4 + KBo 3.41:1213 (OHlNS); twa itoi-wa arami zik Are
you not the friend of me, the king? KUB 29.1 i 35 (OHlMS); emphatic assertion:
tatikan ara danzi They shall not take it away from him BrTabl. ii 99 (Tud.
IV). See further examples in 26.9 (p. 343). For other elements fronted to clause-initial
position see 30.5 (p. 407).
26.5. In contrastive sentences (often with the nite verb fronted), the negation comes
at the end, after the nite verb (Goetze 1927: 114): nammamakan xtn
URU
aplla
kuentaya t ptayaat t But then you neither attacked the land of apalla nor seized
it KUB 14.1 rev. 23 (MHlMS); [nuwaran] annattiya l mu[nn]iy[aw]aran l
you must neither hide him nor conceal him KUB 14.1 obv. 35 (MHlMS); ot\++
arazzi katterai l katterra arazziyai l kuit andan apt a Dont declare a su-
perior case to be inferior, dont declare an inferior case to be superior, do what is right
KUB 13.2 iii 2728 (MHlNS); paimi nwi i nwi I havent yet gone, nor have I
yet seen KBo 3.34 i 23 (OHlNS), warpzimaza nwi but he hasnt yet bathed KUB
13.4 iii 79 (MHlNS). See also the threefold example below in 26.20 (p. 345; categori-
cal negative).
26.6. Negatives immediately precede indenite pronouns or adverbs (Honer 1986:
8889): kuanna natta kuiki i[zzi] No one pa[ys] a wage KBo 6.2 iii 17 (Laws 55,
OS); [i ]dlu natta kuedanikki takkita He did harm to no one KBo 3.22:8 (OS); nu
LU
ippari ppar l [(ku)]iki iezzi No one shall do business with a ippara-man KBo
6.2 ii 4950 (Laws 48, OS); kinunawa r.xtn.n.nr. nawi kuiki uezzi But now
no cavalry has yet come HKM 19:1517 (MHlMS).
26.7. When an indenite pronoun or adverb (kuiki, kuinki, kuitki, kuwatqa) is de-
layed until after the verb, it takes with it any negative associated with it: arrua puwar
anzi l kuiki Let no one seek to defect KUB 26.12+ ii 1617; u kuiu ia[un
numu'] uinuzzi t kuiki Of those whom I treated well no one saves [me] KUB
30.10 rev. 2425 (OHlMS); and with kuit for kuitki: (You must not say:) arawa
parkunummi parkunuimaza t kuit nuza anda imma atkinui I will thoroughly
clean up, while however you yourself clean nothing up, but you yourself rather op-
press KBo 3.1+ ii 4344 (OHlNS).
Negative Rhetorical Questions (with Negatives of Assertion)
26.8. A rhetorical question is an utterance which has the outward form of a question
but which instead of seeking information seeks to make an emphatic contrary assertion.
As such, a negative rhetorical question only employs negatives of assertion, and never
the imperatival negative l. And although all examples identied to date use the nega-
26.4 Negation
00-HittiteGrammar.indb 342 6/5/08 4:19:37 PM
343
tion natta (it), there is no reason to doubt that Hittite speakers could have employed
any of the negatives of assertion in rhetorical questions. A sentence such as *nwi i
utneya enkan (or: *utneya enkan nwi i) Have I not yet see the lands plague?
is completely possible, as is *nman pei utxt.xtts :+ (or: *utxt.xtts :+
nman pei ) Do I not wish to give my daughter? For discussion of negative rhe-
torical questions see Honer 1986: 8991 with the earlier literature cited there; also
Melchert 1985; and chaper 27 below.
26.9. Negative rhetorical questions (Has he not done so-and-so?) are the functional
equivalents of emphatic positive assertions (He has certainly done so-and-so!). In Hit-
tite they frequently show the negative in clause-initial position: nattaama
LU.ME
utotu
tuppi azzian arzi Has (my father) not inscribed a tablet for you dignitaries? KBo
22.1: 23 (OS); twarankan tuetaza memiyanaz kuenner Was it not at your word
that they killed it (the Bull of Heaven)? KUB 8.48 i 12 (Gilgamesh, NH). See further
in 26.4 (p. 342) and 30.5 (p. 407).
26.10. But in even more cases (see tabulation in Honer 1986: 89) the negative is
not clause-initial: (If a mortal were to live forever and the unpleasant illness of such
a man were to continue,) manati natta kattawatar Wouldnt it be a grievance for
him? KUB 30.10 obv. 23 (prayer, OHlMS); man zik t araniee Wouldnt you be
upset? ABoT 65 rev. 6 (MHlMS); often with it imma: ugawaz t imma i-a Am
I not indeed a man? KUB 23.72 obv. 40 (MHlNS); nuwann[a ap] t imma utxt
r-i\t Is he not indeed the son of our lord? KUB 26.1 i 2223 (NH); nuwata it
imma pei peitta Will I indeed not give it to you? I will (certainly) give (it) to
you! VBoT 2:89 (letter to an Egyptian pharaoh).
nawi not yet
26.11. A second negative of assertion is nwi, which is usually translated not yet.
When the accompanying verb is present tense, it is translated with an English present
perfect (haslhave not yet . . . -ed), while when accompanied by the preterite, it is
translated with a past perfect (had not yet . . . -ed). See 22.4 (p. 306).
26.12. A common use of nwi is in the construction with kuitman while (19.6,
p. 290; 30.34, p. 415): kuitman . . . nwi uezzi while he has not yet come = before he
comes (CHD nwi a 2); kuitman . . . nwi uet while he had not yet come = before
he came (CHD nwi b 2 and 30.37, p. 416).
26.13. When nwi occurs in a sentence with unexpressed verb to be (- see 22.3,
p. 306) (CHD nwi c), the implied tense of - is present, and the construction is trans-
lated havelhas not yet . . . -ed.
26.14. As in the case of the other negatives, certain classes of words may intervene
between nwi and the nite verb (CHD nwi d): indenite (kuiki) and relative pronouns
(kui-), conjunctions such as kuitman, and adverbs such as anku and kuwapikki.
Negation 26.14
00-HittiteGrammar.indb 343 6/5/08 4:19:38 PM
344
26.15. Like the other negatives, nwi breaks the nexus between preverb and verb
(Honer 1986: 8689): andaaan parna nwi paizzi (but) he hasnt yet gone into
the house KBo 6.3 iv 36 (Laws 93, OHlNS); kuitmanwa annear [a]ra nwi
ariyaweni before we seek a decision (by oracle) KBo 16.47:1617 (MS); nata kuit
man [aliy]az . . . [ . . . ] katta nwi uwan[zi ] before they come down from the district
KUB 13.1 + 885lz i 67, ed. Pecchioli Daddi 2003: 70.
Imperatival and Categorical Negative l
26.16. The imperatival l, used as both prohibitive (Dont do that!) and inhibitive
(Stop doing that!),
1
regularly occurs with the present indicative (CHD l a), occurring
with all three persons: l aggai I dont want to know CHD l a 1; l umni Let us
not see ibid. a 4; nuwa uitt\t t\+
URU
ayaa l pii Our lord, dont go to the city
of ayaa! KBo 4.4 iii 2526 (AM 12425); nu
LU
ippari ppar l [ku]iki iezzi Let
no one do business with a ippara-man KBo 6.2 ii 4950 (Laws 48, OS).
26.17. The negative l rarely associates with the imperative (see Sommer and Falken-
stein 1938: 9192 and CHD l c; all examples are from NH copies of Old Hittite docu-
ments): l andnpat du Let it not be determined KUB 1.16 ii 51 (OHlNS); nutta
i.xr t.or
URU
xi.nnnn-:t l memikandu Let the elders of attua not speak to
you ibid. ii 60. Another alleged (CHD l c) example, lwatta ni Dont be afraid
KUB 33.24 i 43, KUB 30.33 i 15, and lta ni KUB 30.36 ii 8, KUB 30.33 i 15, is
better understood as an impersonal third-person singular present with accusative -ta let
it not make you afraid (see 16.32, p. 250).
lman and nman
26.18. The combined form lman (CHD l b) consists of l and the speaker opta-
tive, indicating a wish of the speaker (-man, see 23.1, p. 313; 23.10, p. 314; 23.11
23.12, p. 315; Honer 1982). It is slightly more polite than l without the optative
particle and is appropriate when the person addressed is a superior: lmane itoi-u
kian tezzi I wish the king would not speak thus to her KUB 1.16 iii 6566 (OHlNS),
and +ui:+manwakan xtts.itoiya l annetalwan ammuqq amanwa l
kuitki nti-wzi . . . lmanwamu kuitki nti-wzi I wish that my father and the
queen would not be opponents-at-law. I wish it would also not do me any harm. . . . I
wish it would not do me any harm KUB 31.66 iii 58, 19 (NH).
26.19. The compound nman (2.59, p. 63) is the negative subject-optative coun-
terpart (see 23.12, p. 315; and Honer 1982), indicating a negative wish of the sub-
ject of the sentence. The NH KUB 31.66 passage with its separation of l and man
(in ammuqqamanwa l kuitki nti-wzi) shows that that combination had not been
1. See 24.10 (p. 319). The prohibitive is the negation of the perfective aspect (do), while the inhibi-
tive expresses the interruption of the imperfective aspect (is doing).
26.15 Negation
00-HittiteGrammar.indb 344 6/5/08 4:19:38 PM
345
frozen into a single word as *lman, which seems to have been the case with the com-
pound nman. Whereas the special sense of l + man as the negation of the speaker
optative was so close to simple l as to be almost indistinguishable (perhaps accounting
for its sparse use), the peculiar force of nman as the negation of the subject optative
could not be expressed by any of the other negatives. See the table in 26.26 (p. 347).
Sommer and Falkenstein (1938: 198) compared l man to Latin utinam ne would that
. . . not. . .. On the other hand, nman (nwan) expresses the wish of the subject of the
sentence (Honer 1982). While itoi-u l uezzi means may the king not come (i.e.,
I the speaker do not wish the king to come), itoi-u nman uezzi means the king
does not wish to come (regardless of whether I wish it).
Categorical Negative l
26.20. A rare, but clearly attested usage of l is the categorical negative (see Honer
1977: 15152 and CHD l ): In a meadow there stands a iiyamma- tree. Beneath it sit a
blind man and a deaf man. tawanza auzi l duddumiyanzama itamazi l ikniyanza
piddai l tn
7
.nr.-aa uddanante r.srsxtn +:++++ l uwanzi The blind man
certainly does not see. The deaf man certainly does not hear. The lame man certainly
does not run. In the same way the words of sorcery will certainly not see the man for
whom this ritual is performed KUB 12.62 + KBo 53.3 rev. 810. This use of l, the
usual prohibitive negative, to express an emphatic negative assertion is paralleled by the
use of the rst-person singular imperative (voluntative) to express a positive assertion
of strong intention (see 23.4, p. 313). On the post-verbal (contrastive) position of l
here see 26.5 (p. 342).
Word Order with l
26.21. When l precedes the nite verb, it usually allows only certain types of words
to separate it from the verb, notably indenites such as kuiki, kuitki. Its occurrence in
nal position in nominal sentences (CHD l d) lnmi l Let him not be my subject! is
possible only through the suppression of the implied verb to be: lnmi l zi.
26.22. Like the other negatives (26.3, p. 341), l usually breaks the nexus between
preverb and verb (Honer 1986: 8689): anda l tarnikanzi KUB 31.86 ii 25; tat
ppa ar l uzzi StBoT 8 iii 1213 (OS); ara l autti KUB 17.6 i 20 (OHlNS);
par l uwanzi they shall not come out KUB 13.8:8 (MHlNS). The rare exceptions
are late (Muw. II) and involve anda: +\+ xtn-:tmaatkan l anda anti KBo 11.1
obv. 39 (contra rev. 8!); tapa[riyawamu]zakan l anda kitati KUB 14.3 i 2021
(attuili III).
ne / ikku
26.23. There is a kind of negative question that unlike the rhetorical question
(26.8, p. 342) does not necessarily assume either a positive or negative answer
Negation 26.23
00-HittiteGrammar.indb 345 6/5/08 4:19:39 PM
346
but strongly suggests a positive one. This type of question in Hittite is expressed with
nekku. xtn-ewa nikku kuwapikki arkan manwa tnt.uruir.nr. nikku kwapikki
dannateante mnwa
LU
rnr.xr nikku kuwapiki ullante The land is not destroyed
somewhere, is it? The cities wouldnt be devastated somewhere, would they? The troops
wouldnt be defeated somewhere, would they? (If they arent, then why has this deity
come to me?) KUB 24.8 ii 1618 (OH?lNS), ed. Siegelova 1971: 89. We take both
the plene ma-a-an and the non-plene ma-an in this sequence to be the irrealis man (see
23.10., pp. 314.). The term nelikku (see CHD LN s.v. ne(k)ku for ni-ik-ku, ni-ku,
ne-ek-ku, and ne-ku writings) contains the -(a)ku seen in the disjunctive marker -(a)ku,
used in whether . . . or constructions (29.60, p. 405; and Eichner 1971).
Double Negatives
26.24. In Hittite a double negative does not equal a positive but rather an intensied
negative. Although there seem to be no examples of nwi, nman, or nekku in double
negative constructions, both natta (it) and l are so used: nuwaran . . . uwappi
urorn-tt+-ni t par t kuwapikki tarnaun I never turned him over . . . to a hostile
deity, never! att. iv 1213; [apa k]uduwta l l andnpat du And let [those]
false accusations never never be established! KUB 1.16 ii 55 (OHlNS), see ibid.
ii 30. But a negated non-nite verb and a negated nite verb in the same clause are
independent of each other: ektamadduan iraz t naariyawanza ara t uezzi
But even he who is not afraid will not escape (lit., come out of) the circle of your
net KBo 3.21 ii 1718.
Carry-Over of Negative Force
26.25. The force of a negative in one clause can continue into the following one
(Sommer 1922: 8 n. 3; Goetze 1925: 93), as in the English translations: (As malt has
no ability to germinate,) tan .X-ni pdanzi nan txt-an ienzi they dont carry
it into the eld and use it as seed KBo 6.34 ii 3233 (Soldiers Oath); see [niito]
man tepu tan gimra pd [anzi] nan txt-an iyanzi tmaan ru-an
iya[nzi] r
NA
4
xrrn tianzi As [malt] is sterile, they dont carry it into the eld and use
it as seed, nor do they make it into bread and deposit it in a storehouse KUB 17.10 iii
1619 (OHlMS); aanna utxt-an idalu l kuiki iyazi nuian on-an takkezi
Let no one mistreat a son of the (royal) family and drive a dagger into him KBo 3.1
ii 35 (OHlNS).
Summary
26.26. To summarize, the following table lists the negatives of Hittite together with
translations:
26.24 Negation
00-HittiteGrammar.indb 346 6/5/08 4:19:39 PM
347
natta (it) makes a negative assertion he
2
has not (does not, will not)
come
nwi denies action prior to the time of
the action in the main clause
(until now) he has not yet come
or (until then) he had not yet
come
l expresses negative wish or
command of the speaker; or
(rarely) states a categorical
negative
may he not come or (rare) he
never comes
nman states a negative wish of the
subject of the sentence
he does not (did not) want to
come
nekku in negative rhetorical questions
this suggests a positive answer,
but without the unavoidable
positive aspect of natta imma
he isnt coming, is he?
2
Table in 26.26
2. For the sake of convenience we have chosen the third-person masculine singular present tense to il-
lustrate these usages. But obviously any person can serve as the subject. And in the indicative either present-
future or past tense verbs can be used.
Negation 26.26
00-HittiteGrammar.indb 347 6/5/08 4:19:40 PM
348
Chapter 27
QUESTIONS
27.1. As in other languages, questions in Hittite may be direct or indirect. Direct
questions are full sentences that either pose queries answerable by armation or denial
or that contain interrogative pronouns, adjectives, or adverbs, whose answers require
supplying addition information. Indirect questions are subordinate clauses depending
on verbs such as to know or to see. Yesno questions are marked only by prosodic
features such as voice intonation (see 27.2; HE 284; and Mascheroni 1980: 5354).
As such they are usually formally unmarked in writing. Marking these questions by in-
verted word order, as in English Are you going? versus You are going, is not a feature
of Hittite, although it remains true that, as in declarative sentences, marked elements
within a question could be fronted.
Intonation Marking
27.2. Scribes in Assyria and Babylonia who wrote Akkadian in cuneiform script
sometimes indicated the interrogative intonation by a plene spelling of the vowel in the
nal syllable of the central word in the interrogative clause (von Soden 1995: 153d;
Ungnad 1992: 24d). Since the Hittite cuneiform writing system was initially derived
from scribes who used it to write Akkadian, it is no surprise that in one of the oldest
Hittite tablets, KBo 22.1, an instructions text written in the Old Hittite ductus, one nds
the same phenomenon in writing Hittite yes-no questions: nu kian +w': +ut:+ arn
ar-te-ni-i Is this the way you hold my fathers word (as) a limitation? (lines 3031;
see chapter 12, n. 43, p. 193 above). One further example may be present in a NH letter:
[ts:i'lr+\t?] +uix+ya orx-an eer nu me-ma-a-i-i And shall I tell (you), how they
were [with(?)] your father? KBo 18.22 obv. 6. This is the only known example of the
form memai with a plene writing of the nal vowel. Since this phenomenon is found
to date only rarely, it apparently did not become a regular scribal method of indicating
interrogative stress. See Honer 1995a: 88. On the possible functions of plene writing
see 1.461.47 (p. 25).
YesNo Questions
27.3. Sample yes-no questions are: r:+za mali Are you agreed, my brother?
KUB 14.3 iii 62 (Tawaglawa letter, NH); urorn-ti+za kda wakuwa r ftxt.
ftxt-uwanza O deity, are you angry on account of these oences? KUB 5.10 i 12
(oracle question, NH); kian +w': +ut:+ paanutten Is this the way you (pl.) have
00-HittiteGrammar.indb 348 6/5/08 4:19:40 PM
349
kept my fathers word (i.e., command)? KBo 22.1:45 (note the fronting of kian in
this way); peiwar ati mnwai t pe[i ] Shall I give it to him? (What) if
I do not give it to him? KUB 12.60 i 21 (OHlNS); and see KBo 22.1:3031 in the pre-
ceding paragraph. In Hittite one probably answered such a question not with a yes or
no word but with a conrming or denying sentence: r:+za mali Are you agreed,
my brother? was answered with malmiza I agree or itza malmi I dont agree
(so assumed by Honer 1995a: 89). But since our only examples of yesno questions
are either rhetorical ones which do not elicit an overt reply, or questions in letters for
which we possess no specic reply (KUB 14.3 iii 62), or questions in oracles, where the
answer only comes in the standard manner of favorable or unfavorable, we cannot
verify this supposition. As just stated, some yesno questions are also rhetorical ques-
tions which expect a denial of the situation claimed in the query: nu ammel dammea
s+ ux:+ inkan sro
5
-yattat My punishment is the death of my wife. Has (this) gotten
any better? (implied answer: No, it has not gotten any better) KBo 4.8 ii 2021 (NH),
ed. Honer 1983a: 188; zikzakan ammuqqa 1-edani x-ni aante Were you
and I born of one mother? (implied answer: No, we were not) KUB 23.102 i 1415
(NH); nankan +\+
GI
ororn waggariyanun namaankan X r-:t waqqariyanun Did
I rebel against him in the chariot, or did I rebel against him in the house? att. iii
6768; (man doesnt live forever; his days are numbered; if a mortal were to live forever,
(and) the painful ills of man were to remain,) manati natta kattawatar wouldnt that
(lit., it) be a punishment (lit., grievance) for him? KUB 30.10 obv. 23; nu ziqqa
kuwatqa s+
m
Maturi iwar iyai Will you perhaps also act like Maturi? KUB 23.1+
ii 2930, ed. Khne and Otten 1971: 1011. Further examples in Honer 1995a: 90.
Some rhetorical yesno questions serve as the apodosis of a contrary to fact condi-
tional clause: [utxt.itoi]ma[nwa]nna kuwapi ta anzmanwa [da]mdani
xtn-e uwawen mnwana [an]zel uitt wekikewen If we had [a prince] anywhere,
would we have come into a foreign land and kept requesting a lord for ourselves?
KBo 14.12 iv 1517 (D frag. 28, NH), ed. Gterbock 1956: 9798 (dierently re-
stored
1
); mnwaramukan ulliyat kuwapi t mn andn itoi.oi +\+ itoi
ftn katterraer If he (Uriteub) had not at some time become wantonldisrespectful
towards me, would (the gods) have truly subjected a Great King to a minor king? att.
iii 7677. Others introduce a clause giving the grounds: kinuna apel fr-tar idalawta
fr-anza<a> kuit nu nepia
d
tft-un ror.nr.-it ukezzi Has her (i.e., Tawannannas)
life become bad, just because she is (still) living and sees with her eyes the Sungod of
Heaven? KBo 4.8 ii 1819 (NH), ed. Honer 1983a: 188; see also
LU
X.fx [ . . . nti]-
lu [
GI
.-an] [kui] di nuza it eat i
GI
ntn-a nti-lu [
GI
ntn-un
kui] [d ]i nuza t ezatta
LU
or.-a nti-l [u oi-in] kui pi nu t ekutta Is the
1. Gterbock restored [utxt.itoi-wa] ma[-a-a]n-na-a, but making utxt.itoi-wa a separate
clause and leaving ma[-a-a]n-na-a without the particle -wa makes for very awkward syntax, and we can-
not read [utxt.itoi-wa-]ma[-a-a]n-na-a as one word, because both a-i-ma-an-wa KUB 11.6 ii 11
(= Homann 1984: 36 ii 64) (OHlNS) and anzamanwa in this very same passage show that the proper
sequence requires -man to precede -wa.
Questions 27.3
00-HittiteGrammar.indb 349 6/5/08 4:19:42 PM
350
chamberlain [wr]ong [w]ho set [the chair] (for you), that you did not sit down? Is the
table-man wrong [who] set [the table] (for you), that you did not eat? Is the cupbearer
wrong who gave (you) [the cup,] that you did not drink? KUB 33.96 + KUB 36.7a iv
5558 (Ullik. I), ed. Gterbock 1952: 26.
27.4. Negated questions are usually rhetorical questions, which do not seek infor-
mation but presume a positive answer. See above 26.8. (pp. 342.). Although these
are often marked by fronted negation, this is not always the case (contra HE 282c;
see Honer 1986). For fronted examples see 26.9 (p. 343); for non-fronted examples:
26.10 (p. 343). Because of their emotional content, rhetorical questions often (but not
always) employ imma actually, really, indeed (see Otten apud Rost 1956: 33233;
Melchert 1985; and above sub 26.10, p. 343). For negative questions that do not neces-
sarily assume a positive answer see 26.23 (p. 346).
27.5. A special class of yesno question is the deliberative question, on which see
24.39 (p. 328).
Direct Questions with Interrogative Pronouns,
Adjectives, and Adverbs
27.6. The following are some examples of direct questions using interrogative pro-
nouns, adjectives, and adverbs. Mascheroni (1980: 6061) calls them domande com-
plende (German Ergnzungsfragen), because the reply cannot be merely yes or
no but must supply the information requested (where, how, why, when) (see Ho-
ner 1995a: 92104). Hittite, like English and many other languages, has so-called wh-
movement, by which interrogative forms are placed at the beginning of their clauses, re-
gardless of the role they play in the sentence (subject, direct object, indirect object, etc.)
(Garrett 1994, following Hale 1987). Most interrogatives thus appear in clause-initial
position (as in English) or immediately following sequences of clause-initial conjunc-
tions and sentential clitics (which do not count as full words for this purpose). Hittite,
however, does with some frequency allow one word (or constituentsee the last two
examples in 27.7) to occur before the interrogative. Compare the similar phenomenon
with relative pronouns (30.5., pp. 407.), but the motivation for the variation with
interrogatives is much less clear. We may guess that some pragmatic considerations of
emphasis or focus are involved (see the example in 27.11, p. 352), but no obvious
dierence in meaning is discernible for many examples. As in the case of relative pro-
nouns, there are also a few instances where the interrogative word occurs still later in
the clause, preceded by more than one constituent (see the examples in 27.12, p. 352,
and others below). These require further study.
27.7. Direct questions are formed with the interrogative pronoun (kui- who, what).
Clause-initial: [mn
d
tft]-un pmi nan munnmi nu kuit izzi
d
rx-a If I seize the
Sungod and hide him, then what will the Stormgod do? KUB 36.44 i(!) 56; nu kuit
so what? KUB 1.16 ii 9 (OHlNS); kuit iyanun kuit What did I do? What? KUB 31.4
27.4 Questions
00-HittiteGrammar.indb 350 6/5/08 4:19:46 PM
351
+ KBo 3.41:3; kuedaniwaza menaanda iamikei Before whom are you sing-
ing? KUB 36.12 ii 9 (Ullik. II); nuza kul walliyatar tza s+
d
t piaai r:+
walliyatar Then whose glory am I? Am I not the glory of the Stormgod piaai
my lord? KUB 6.45+ iii 4849 (pr. of Muw. II), note that -za excludes the translation
Then whose is the glory? Is it not the glory of the Stormgod p. found in ANET 39798;
kuelwakan zr-anza uri dandukewakan zr-anza uri Whose soul is great? The
mortal soul is great KUB 43.60 i 2728. Not clause-initial: uku punukemi [kwa
k]uit walkuwan I ask them, What is this mob? KBo 3.40:15; apa pait
d
rx-ni tet
k kuit kiat He proceeded to say to the Stormgod, What is this that has occurred?
VBoT 58 i 16 (OHlNS); [(xtn.xtn.nr.)m]a man kui arzi Who holds all the
lands? KUB 31.4 + KBo 3.41:12. In the just-cited example the noun and its attribute
[(xtn.xtn.nr.)m]a man all lands are counted as one position. Note that in k
ara kui peutet Who led these (people) out (here)? KBo 3.34 i 19 (OHlNS), ed.
Dardano 1997: 3436, kui appears to be in third position.
27.8. With interrogative adjective kui- what?. Clause-initial: kuitwaikan
si+-an [tei
d
]Guluwamu urorn.xn.nr.-u kuin utxt-an stx-er What name
[shall I give] to him, the child whom (lit., what child) the gula- and kunutalla-deities
have given me? KUB 33.93 iii 1617 (Ullik. I). If the second clause in the preceding
example is not another exception to the word-order rule, it would indicate that the coor-
dinated subject [
d
]Guluwamu urorn.xn.nr.-u was counted as one position. In
the following dialogue we see how the interrogative adjective kui- is used to narrow pos-
sibilities: nu kuz uwai uppazwa uwami nuwa kuz uppayaz zaanittennazwa
nuwa kuz zaanittennaz
d
tft-wawa r-az nuwa kuz
d
tft-az Open! So, where
are you coming from? I am coming from the holy thing. So, from what holy thing?
From the z. So, from what z.? From the Sungods temple. So, from what Sun-
god? KBo 21.22:2225 (OHlMS). Both in this dialogue, in the rst example in 27.10
(p. 352), in the second example cited in 27.7, and in the second example cited in 27.9
the nu implies that the course of thought (perhaps the logic) is consciously carried
on by the new speaker (see 29.48, p. 403). In most cases this requires the word so in
English. Not clause-initial: i+++ st++ [k]wa kuit walkuan un She said: What
mob is this that I have borne? KBo 22.2 obv. 12 (OS);
2
[tue]lla utxt.xrx+
kuin again iyanzi And what miracle do (i.e., can) your sons perform? VBoT 58 i 7;
[zi(qqaz)]a kui urorn-tt+-i nu t [ . . . ] What (kind of a) deity are you, that [ . . . ]
not [ . . . ]? KUB 33.86 ii 12 with dupl. KUB 36.56 iii 9 (StBoT 14:54).
27.9. With maan how? Clause-initial: man iyaweni kitantit arkueni What
shall we do (lit., how shall we act)? We will die of starvation! KUB 17.10 i 2930
(OHlMS). Not clause-initial: nuwa wattaru man iyan So, how is the fountain
made? KBo 21.22:4142 (OHlMS).
2. One could also translate What is this mob that I have born? with kuit as interrrogative pronoun.
Questions 27.9
00-HittiteGrammar.indb 351 6/5/08 4:19:46 PM
352
27.10. With kuit anda why? Clause-initial: [
d
]rx-a tezzi nuwaran kuit anda
it wemiya[tten] The Stormgod says: So, why didnt you (pl.) nd him? VBoT 58
i 23. Not clause-initial: [uku] punukimi karwaret kuit anda lipan I ask [them]:
Why is its (i.e., the bulls) horn bentlcracked? KUB 31.4 + KBo 3.41:16 (OHlNS).
27.11. With kuedani er on what account? why? not clause-initial with a fronted
adversative element: utxt.xr itoima kuedani [(e)]r arkikantari But as for the
princes, why are they dying (i.e., being put to death)? KBo 3.1+ ii 5658 (OHlNS).
27.12. With kuwat, kuit why? Clause-initial: kuitwa uw[(a)] Why have you
come? KUB 33.24(+)28 i 38; [i+++ st++ ku]watmu k tepu paitta She said: Why
have you given me this little bit? KUB 1.16 iii 10 (OHlNS); itoi-ua memaun
[kuwa]tware akkanzi And I, the king, said: [Wh]y should they die? KBo 3.1
ii 2829 (OHlNS); (The Egyptian queen wrote back to Suppiluliuma I:) kuwatwa
apenian :+ut appalekanziwamu Why do you speak thus: They are deceiving
me? KBo 5.6 iii 5253 = D frag. 28 (NH); (Pharaoh Ramses II writes to Puduepa:)
kuwatwaranmu kinun t peta Why have you not given her (scil., the bride) to
me now? KUB 21.38 obv. 8 (NH). Not clause-initial: i.xr )i+t:+mu kuwat t
uiekettani Why are you not sending my envoys (back) to me? HKM 53:2930 (MHl
MS); zikwakan apn anda kuwat auta Why did you look at that (woman)? KBo
5.3+ iii 71 (MHlNS); nu kn memiyan kuwat iyatten Why did you do this thing? KBo
3.3+ iii 34 (NH); annianwaran [ito]i-eznanni kuwat tittanut kinunmawai
kururiyauanzi [k]uwat atrikei Why did you previously put him on the throne, and
why now are you writing to him to declare war? KUB 1.4 + KBo 3.6 + iii 4042 (att.
iii 7576). Many other examples in Honer 1995a: 97100.
27.13. With kuwapi Where? Clause-initial: nu kuwapi
d
tft-u mumizzi [ . . . ]-
iku appnikku or-ikku aalikku mumizzi So where will the Sungod fall? Will
he fall into the [ . . . ], or the ame(?), or the tree(s), or the brush? KUB 36.44 iv 89
(OHlMS?); kuwapi[(warat)] andan piddaikanzi 2
LU.ME
+:ni:t+ Where are they
running to, (these) two brothers? KUB 33.113 + KUB 36.12 + KBo 26.64 i 1920
(Ullik. II B, NS); rxr.nr. rxr.nr. kuwapiwa paitte[ni] Tongues! tongues! Where
are you going? KUB 44.4 + KBo 13.241 rev. 22 (NS). In rhetorical question interroga-
tive kuwapi preceded by the fronted topic: s+ i.oi.oi-:t+ rsi\i kuwapi tat
arkr Where are the houseslestates of the grandees? Have they not perished? KUB
1.16 iii 45 (OHlNS).
27.14. With kuwapit Where? To what place? kuwapit aumen nu r-i ar-
katta Where have we seen (it), that a donkey will climb? (expressing incredulity; see
18.3, p. 277, with n. 2) KBo 22.2 obv. 10 (OS); [(mnwa +)]\+
d
rx
URU
Nerik piwani
nuwaan kuwapit (variant: kuwapi) ewatati When we go to the Stormgod of Nerik,
where shall we sit down? KBo 3.7 iv 67 (OHlNS).
27.10 Questions
00-HittiteGrammar.indb 352 6/5/08 4:19:47 PM
353
27.15. An example with a mixture of ways of posing questions: nu mai xt.nr.
pir [maiea]kan uwir s+ i oi.oi-:t+ r-si\i kuwapi tat arkr How
many years have passed? [How many] have ed? And where (are) the houseslestates of
the grandees? Have they not perished? KUB 1.16 iii 4445 (OHlNS).
27.16. The interrogative clause can be followed by a clause expressing the reason
for the query: nuza kui . . . [nu anz]el .si ara weiekei Who are you . . . that
you devour our meadow? KUB 24.7 ii 5657 (NS); kuitwa watultit nuwaratt[a
sro
5
-ami ] KUB 24.8 i 45 (pre-NHlNS) could be translated What is your problem (lit.,
sin), [that I may remedy] it for you?
3
See 27.18 for another example.
Questions Posing Alternatives
27.17. Questions which pose alternatives take a special form (Sommer 1932: 7778;
Honer 1995a: 101). The second question of the pair frequently uses -ma to mark the
word which best expresses the alternative (see CHD sub -ma a 1 b 4, where it is shown
that the initial nu is only optional): ni andurza kuiki ui-yazi . . . nu ni arazama
kuiki ui-zi Will someone revolt from inside (the kingdom), or will someone revolt
from outside? KUB 5.4 i 3335; nuwarat r[:+ nuw]arat ma Does
my brother know it or not? KUB 14.3 i 52 (NH); kuitat r-i::+ nat kuitma s+
UR.SAG
Ammana uwawar What is it, (this) brotherhood? Or what is it, (this) coming
(to) Mt. Amanus? KUB 23.102 i 78 (letter to Adad-nirri of Assyria); see Sommer
1932: 78.
27.18. A special kind of alternative question is that which is also rhetorical in na-
ture, implying a denial of both stated alternatives: ux:+ xtts.itoi idalawata
ku[i ]t [k]i nan tepnuttama kuitki nukan
f
Tawannanna ux:+ ku[e]n[d ]a Did my
wife harm the queen in some way, or did she demote her, so that Tawannanna killed my
wife? KUB 14.4 iii 2122 (prayer of Mur. II), with collation by Rster conrming
ku-[e]n-[d ]a (cited by de Martino 1998: 28 n. 65).
Indirect Questions
27.19. For indirect questions see 30.68. (pp. 427.).
3. Because the ku-i restored in nu-wa-ra-an-k [n or
6
-i ] xr-p-ya an-d [a ku-i (') p-e-d ]a'-i' And
who will carry him o to the dark netherworld? KUB 33.93 + KUB 36.7a + KUB 17.7 iii 30 (Ullik. I A,
ed. Gterbock 1952: 18) would occur later than the second position, the line should probably be restored
without ku-i and the clause understood as carrying over the force of the ku-i in line 28: Who will . . . and
(will) . . . and (will). . . ?
Questions 27.19
00-HittiteGrammar.indb 353 6/5/08 4:19:48 PM
354
Chapter 28
PARTICLES
28.1. Hittite has a wide variety of particles, most of them sentential, but several not.
Sentential particles are attached (following clitic conjunctions such as -al-ma [29.23.,
pp. 395.], -al-ya [29.38, p. 399] or the non-sentential particle -pat [28.115.,
pp. 384.]) to the rst word in the clause. For the order in which these and other sen-
tential clitics appear see 30.15 (p. 410). One must base all conclusions regarding the
true positioning of both sentential and word clitics on instances where the phrases in
which they occur are written entirely phonetically. Hittite scribes for the most part do
not write clitics in the middle of phrases written partially or entirely logographically.
We thus nd examples such as:
MUNUS.ME
zintuiy
URU
Taurpamakan (KBo 11.73 iii
78) for what was likely spoken as *Taurpamakan zintuiy or t\+ r
KU
kurama
(KUB 59.2 iii 3) for *kurama parni, etc. See further p. 396, n. 10.
The Particle -wa (r-)
28.2. Direct quotes are marked by a particle -wa(r-) suxed to the rst word of the
quoted utterance (Hrozny 1917: 98 n. 1, 144; Pecora 1984; and Fortson 1998). Rarely,
when the particle follows an a-vowel, it is spelled -u-wa-(r . . . ): [ pa]rtauwa (r)etwa
amiyanta a-pa-a-a-u-wa amiyanza its wings (are) small, and it (is) small KUB 17.10
i 38 (Tel. myth, OHlMS); for evidence see Honer 1985.
28.3. The longer form -war- is used when another enclitic beginning with a vowel
is added: akuwara KUB 11.1 iv 20 (OHlNS); ts:i urorn-tt+wara n. Laws
75; nuwaranzaan peutet KUB 12.60 i 18 (OHlNS); kmu kuit s+
m
Marrwa
i
URU
immuwa aliyatar atre parwarankan neun HKM 13:35 (MHlMS);
nuwaratapa KBo 3.1 ii 49 (OHlNS); ugawaru ibid. ii 15; [kuw]atware akkanzi
ibid. ii 29.
28.4. When followed by a consonant or in word-nal position, the form is -wa:
nuwa passim; attatawa KUB 1.16 ii 65 (OHlNS); [l]n-iwaan KUB 1.16 ii
71; [k]awaz KUB 1.16 iii 67; kinunwaz ibid. iii 68; apwakan HKM 6:7
(MHlMS).
28.5. For the likely source of the form -wa see 1.1321.133 (p. 46).
28.6. Rarely (about 26 times) wa appears for war before a vowel: pei
war ati mnwai t pe[i ] KUB 12.60 i 21 (OHlNS); apwakan itarna
ara uet nammamawar a kuwapi pait nuwara t toi HKM 6:710 (MHl
MS); stx-erwar ati kuwapi nuwa k ka arantat KUB 13.4 ii 3738
00-HittiteGrammar.indb 354 6/5/08 4:19:49 PM
355
(pre-NHlNS); nuwamu 1-i\ utxtx+ pi nuwara ammuk
LU
+i:t:+ t\+ xtn
URU
Miz rimawar a itoi-u KBo 5.6 iv 1012 = D frag. 28 (NH) (note, how-
ever, the evidence of other mistakes in this context); [t]tanumawar at KUB 33.41
ii 5; namawar atzakan KUB 14.8 obv. 2 (PP 2); see also mnwar at[a] KUB
34.53 obv. 2, which is otherwise always warata. Note that this omission always oc-
curs before an enclitic beginning with a, never before those beginning with e or u. This
restriction tends to suggest that these are examples of the weak articulation of intervo-
calic -r- (see 1.132, p. 46).
28.7. Quoted speech is frequently introduced by a verbum dicendi, such as mema-,
te- or tar-, alzai-, tataiya-, punu-, or mimma-: nu kian mematti eniwa idlu
kiaru And you say as follows: Let that evil thing happen! KUB 6.41 iii 4647
(Kup.); takku uit otu tezzi ammelpatwaza otu-un dai otusi dai If the owner
of the ox says: I will take my own ox, he shall take his ox KBo 6.3 iii 7172 (Laws
74); i+++
m
utxt. tu.20.xx +\+
m
Uzz r.uio.o. tut++ r+\t
m
Pullkan
ammel aul alzai otuyawamu kuin tet nuwaranmu uppi Thus says Mar-er:
Speak to Uzzu, my dear brother: Give my greeting to Pulli: Send me also the ox
you promised me! HKM 22:916; nutta uezzi pe[(ran kui)]ki t [(ataiyaizzi)]
d
tft-
stwaduzakan nti-lauwanni [kia]n kiann[(a)] zikkezzi Someone will proceed to
whisper in your presence: His Majesty is planning this and that (lit., thus and thus)
for evil against you KUB 21.1+ iii 1719 (Alak.), ed. SV 2: 6871; Otten 1957;
nutta mn urorn.xr kian punuanzi kwa kuit ier If the gods ask you as
follows: Why were they doing this? KBo 17.105 ii 1718 (incant., MHlMS); nan
punukewan d[i k]uitwa watultit (The Sungod) began to ask him (i.e., Appu):
What is your trouble? KUB 24.8 i 4445 (OH?lNS), ed. StBoT 14:6f.; [(m)]n utxt.
xr
URU
atti i.xr ttxt ur +\+ +ut itoi aruwa[nzi] nu tarikanzi kannaza
natta kuiki iz[(zi)] nuwannaza mimmanzi i.xr ttxtwa ume When the sons
of atti, men owing ttxi, came, they bow to the father of the king and say: No one
pays us a wage; they refuse us, (saying:) You are men owing ttxi KBo 22.62 +
KBo 6.2 iii 1618 (Laws 55, OS).
28.8. Occasionally, quoted speech is introduced abruptly, without any verb of speech:
nukan x.n.xr katta ur natmu oln.xr-a k[(att)]an liyanda[(t)] uitt-\t
wanna l arnikti The civilian captives came down and prostrated themselves at my
feet, (saying:) Our lord, do not destroy us! KUB 14.16 iii 1617 (AM 5657) restored
from KUB 14.15 iii 46 47; other examples of verbs which do not denote speech acts
followed by direct discourse: after arnu- KUB 36.90 obv. 7; au- KBo 24.124 rev. 1,
KBo 24.128 rev. 2, KUB 5.24+ ii 12, KUB 44.4+ rev. 9; iya- KBo 11.10 iii 1617;
pp- KBo 5.3+ iii 7071, KUB 33.106+ ii 5, IBoT 1.36 i 56; ea- KBo 14.21 ii 58;
kappuwe- KUB 36.51 rev. 3, KUB 17.10 i 21; piya- KUB 12.60 i 16, KBo 13.228 i
56, RHA 77:83f. B i 21; peda- KUB 13.4 iv 70; akuwai (a)- KUB 24.8 iv 31; ua-
KUB 12.26 ii 22; tarna- KUB 22.70 rev. 44; dai- KBo 3.60 ii 1317, KUB 13.4 iv
27; teaniya- KUB 1.1+ iii 45, iv 9; uiya- KBo 3.40:10, KBo 4.14 iii 73, KUB 1.1
Particles 28.8
00-HittiteGrammar.indb 355 6/5/08 4:19:51 PM
356
i 1415, KUB 17.10 i 24, KUB 24.2 i 5, KUB 26.89:5, KUB 33.52 ii 4; wanu- KBo
9.106 ii 1, KUB 44.4+ rev. 25; zuwi- KUB 44.4+ rev. 1011. See also the comments
in CHD sub palwai-.
28.9. Once a saying is quoted which justies a legal decision: takku xtts-nan
kuiki pittinuzzi n[ukan ard ]iye ppan anda p[n]z[i ] takku 3 i.xr nama 2
i.xr ak[(kanzi )] arnikzil t oi zikwa tn.nn.n-a kitat If anyone runs o
with a woman, and helpers (of the brides family) go after them, if three men or two
men are killed, there shall be no compensation, (because:) You have become a wolf
KBo 6.2 ii 1012 (Laws 37, OS) dupl. KBo 6.3 ii 2930 (NS).
28.10. In at least one case, this indicates that the words constitute a popular saying
(on which see CHD lala- 1b, followed by Beckman 1986): lalawa (var. rxr-awa)
armizzi The tongue is a bridge KBo 11.72 iii 5 with dupl. KBo 11.10 iii 17.
1
28.11. Occasionally, especially when there is an introducing verb of speech, the par-
ticle -wa can be omitted. Friedrich (HE 292), quoting himself and Sommer, claims
that this is particularly true in the mythological texts and cites a passage from OHlNS
Illuyanka: nu ar nepii attii alzi ammugga rorn-pa anda p lmu genzuwaii
He called up to heaven to his father: Take me together with (them)! Have no mercy on
me! KBo 3.7 iii 2730; see edition by Beckman 1982. But the particle is also omitted
in Old Hittite (OS) rituals such as: ta kian mmae tuwattu urorn.xr-e And I say
as follows: Mercy, O gods! KBo 17.5 ii 9 (StBoT 8); and in letters, where quota-
tions are sometimes embedded within quotations: ki[an]mamu kui [t ]atre rnr.
xr
GI
za[l ]tay[awa k]ui
URU
Kae[ p]ra pait [. . .] . . . [ki]an memikanzi maan
[i.xr
U
]
RU
Gag[a uwan]zi nuwak[an par] rorn[-anda pai]wani nuw[aruka]n
ar [arn]u[m]eni Concerning what you wrote to me as follows: The troops of the
(wooden) zaltaya who went to Kaepura, they are speaking thus . . . : When the men of
the Kaka [come], we will [go] forth after (them), and we will bring them up HKM
24:410; [tnt-r]imakan anda kian memiyawen [nu uezz]i
d
tft-st
URU
Taggatan
walzi alkiu[]akan [ar-]a (collated) warzi But in [the city] we said: [Then] His
Majesty will [proceed to] strike Taggata and harvest the crops HKM 47:1719; in a
formal prayer, where the rst lines consist of a long series of divine names and epithets:
nu kia[n m]emai
d
tft s++I i
d
tft
URU
fi-na o:+ xtts.itoi o:+
xtts.itoi s[+ xt]n
URU
atti KUB 6.45 i 1011; in the instructions text ABoT 53 +
KBo 16.54:1718 the clause nu i.xr
GI
ftxti i.xr t.or anda alza[i] [nu]ma
kian memi Summon the ftxti-men and the elders and speak as follows to them is
followed by a lengthy quote without any use of -wa; similarly [nu . . . kia]n memai
KUB 57.105+ ii 24 and nan utxt.xtts uppiari pai nu wiyekezzi andamakan
kian memikezzi VBoT 24 iii 3436 and nata
NA
4
pailu ts:i x kitanumi nu
kian memai ibid. iv 12 and erraan zu.u zz xtya uai nu kian memai
KBo 11.14 i 20 and nu kian memai KBo 15.21 i 17 in rituals. Examples are also given
by Mascheroni (1980: 5758), who claims that in texts from later periods -wa is much
1. Incorrectly read as la-la-na-a in HW 3. Erg. 22; on HW
2
sub armizzi bridge see correction in
Honer 1983b: 412.
28.9 Particles
00-HittiteGrammar.indb 356 6/5/08 4:19:53 PM
357
less likely to be omitted, and cites KBo 16.61 obv.? 4, where the scribe has inserted the
particle as an afterthought above the line.
28.12. The particle can be omitted when the quoted speech consists of a single word:
nu
URU
Niili kian [alza]i tapua (The guard) calls out in Hittite: To the side! IBoT
1.36 iii 6465 (+isiot instruction, MHlMS), ed. Gterbock and van den Hout 1991;
duddu alziai numu urorn-tt+:+ itama I call Help! Hear me, O my god!
KUB 30.11 rev. 21 (OHlMS). See 16.1316.14 (p. 244) for its frequent omission
with the vocative.
28.13. It is not uncommon in quoted speeches which stretch over several clauses to
nd -wa omitted in one or more of the consecutive clauses.
28.14. Quoted speech very rarely begins with a conjunction: [
d
]rx-a tezzi nuwaran
kuit anda it wemiya[tten] And the Stormgod said: AndlSo (nu) why didnt you nd
him? VBoT 58 i 23, tr. Pecchioli Daddi and Polvani 1990: 5770; Honer 1990:
2628. See 29.48 (p. 402).
28.15. The above discussion, which is based on the current generally-held view,
assumes that consistent use of the particle at the beginning of each clause is the norm
and omission is to be explained case by case. This view has been challenged by Fortson
(1998), who argues that the use of the particle was limited in spoken Hittite and condi-
tioned by various discourse factors, some traces of which are still visible in more infor-
mal text varieties. The consistent use of the particle in every sentence of quoted speech
represents in his view a stylistic standardization of the more formal literary standard of
Hittite administrative texts.
The Particle -z(a)
Form
28.16. This particle is conventionally referred to as -za, but there is evidence to
show that its real form was merely l-tsl. First, in OH and MH it is regularly spelled as
-z after a morpheme ending in a vowel: zi-ik-wa-az, lu-uk-kat-ti-ma-az, ta-az, nu-uz, ne-
ez, ma-a-n-ez, and (not nal) ta-az-kn, an-da-ma-az-kn, a-ap-pa-ma-wa-az-kn (see
Khne 1988). Second, when it occurs before the local particle -an, the sequence of the
two is spelled simply -za-an, i.e., l-ts-sanl (see 28.83, p. 374). Elsewhere it is spelled
-za. Under no circumstances (contra Sturtevant 1933: 105 and Kammenhuber in HW
2
)
does it take the form -az (see Sommer and Falkenstein 1938: 114 n. 4). The a which
precedes it always belongs to the preceding morpheme (-ma-az = maz).
Reexive Function
28.17. A reexive pronoun is one which refers back to the subject of the sentence or
clause in which it stands. English reexive pronouns contain the element -self: my-
self, himself, herself, yourself, ourselves.
2
These may function as direct objects, as in
2. The reexive pronouns are not used to express the -self forms in English used for emphasis, as in
He said it himself. For this purpose Hittite uses the sux -il(a) (see 18.7, p. 279) or the particle -pat (see
Particles 28.17
00-HittiteGrammar.indb 357 6/5/08 4:19:56 PM
358
he hit himself, as indirect objects, as in I bought myself a sandwich, or as objects of
prepositions, as in he compares others to himself. For the rst two of these functions
Hittite uses either the particle -z(a) or the enclitic dative pronoun (only in the plural)
of the same person as the grammatical subject of the clause.
3
Reexivity may also be
indicated without a particle by the medio-passive forms of certain verbs (see 21.6,
p. 303).
28.18. In Hittite (as in Latin) the dative forms of the personal pronouns can serve
as reexives. In Hittite this procedure is applicable to the plural: (rst person dat.-
acc.:) nunna utxt.rf.xr utxt.xtts.xr iyawen and we begat for ourselves
sons (and) daughters (att. iii 4); (third person dat.:) warpanzimawama t but
they do not wash themselves KUB 16.16 i 28. The unexpected choice of the dative
in the last example for a direct object is in order to assure the reexive meaning, since
warpanzimawaru t with accusative -u in place of the dative -ma would mean
they do not wash them (i.e., a second group).
28.19. The usual means of expressing the reexive idea in Hittite is the enclitic par-
ticle -za (alternate form -z), whose position in the chain of sentence enclitics is noted
in 30.15 (p. 410), and which serves without inection for all persons, numbers, and
genders.
28.20. When the verb is transitive and the clause has no other accusative form,
-za may indicate that the direct object of the verb is the same person as its subject:
warpanzimawaz t but they do not wash themselves (-z) KUB 16.34 i 89; kinuna
maza itoi-u labarna ulanun Now (kinuna) I, the labarna, have united (ulanun)
myself (-za) to you (-ma) KUB 29.1 i 3334 (OHlNS).
28.21. When the verb is transitive and the clause contains a direct object (an accusa-
tive), -za may indicate that the indirect object of the verb is the same person as its sub-
ject (tolfor himself) (see van den Hout 1992): nuza utxt.rf.xr utxt.xtts.
xr ui-nun and I made (iyanun) for myself (-za) sons and daughters KUB 21.38
obv. 58; nuza PN . . . ux-anni arta and he had PN for himself in wifeship KBo
3.1+ i 3233 (Tel. pr., OHlNS); ru-anza wemiyanun . . . wtarmaz wemiya nun
I found for myself bread . . . I found for myself water KUB 30.10 obv. 1617 (OHl
MS); natzakan pedipat ln-ata he made them (-at refers back to a neuter collec-
tive noun for people) subject to himself on the spot KUB 19.9 i 15; nuza . . . x.
n t\+ r.itoi uwatenun I brought civilian captives into the kings house for myself
KBo 3.4 ii 41 (AM); nuzakan t\+ xtn
URU
atti
d
tft
URU
fi-na si+-an dait[a] but
in the land of atti you have placed upon yourself the name Sungoddess of Arinna
KUB 21.27 i 4; nuza k ix:+ s+ xi.or iyanun I made for myself this gold statue
of myself KBo 10.2 iii 2122 (OHlNS); nanza +\+ uxsi di nuza r-er i utxt.
28.115., pp. 384.).
3. It is likely, but not yet fully established, that Hittite (only NH?) may also use -z(a) as the object of a
postposition. A likely example is: nuza rnr.xr \+u'ui s+ xtn-:t tepauwaza ox-an ppn I took with
me auxiliary troops of the country in small numbers KBo 3.6 ii 7. See Francia 2002b: 53.
28.18 Particles
00-HittiteGrammar.indb 358 6/5/08 4:19:57 PM
359
xr ienzi and he takes her to himself as his wife, and they make for themselves a
house and children KBo 6.3 ii 17 (Laws 31, OHlNS); [kui ]za
LU
ippari ppar iezzi
he who makes a business transaction for himself with a ippara-man KBo 6.2 ii 51
(Laws 48, OS); oi-iza
d
tft-u rzr
4
-an it The great Sungod made a party (lit.,
festival) for himself KUB 17.10 i 19 (OHlMS); nammaz uet +ut
d
tft-st tuk
m
Maddu-
wattan linkiyaa it then my father (lit., the father of My Majesty) proceeded to
make you, Madduwatta, a sworn ally (lit., one of his oath) for himself KUB 14.1
obv. 1314 (MHlMS); nata urorn.xr-apat zr-ni iyatten nu ru-an zzatteni
wtarma ekutteni r-erraza iyatteni Act in the will of the gods, and you will eat
bread, drink water, and make a house for yourselves KUB 13.4 ii 6971 (instructions
for priests, pre-NH). When the verb iya- to make is construed with a double object (to
make something into something else),
4
it sometimes takes -za and sometimes does not.
In these cases the-za refers back to the subject as a dative of interest.
28.22. Beginning in Old Hittite, the particle -za can serve in a possessive construc-
tion to identify the possessor with the grammatical subject of the clause (his own,
her own, our own, etc.; Honer 1973a: 52324): nuzza utxt.xtts.xrs+ +\+
utxt.rf.xrs+ pai (The queen) gave her own daughters to her own sons (in mar-
riage) StBoT 17 obv. 17 (OS); nuzakan utxt.xrsi par uwiezzi (if a mother
removes her garment from her son,) she thus disinherits her own son (pl. sons is a
scribal error) Hittite Laws 171 (OHlNS); New Hittite examples: nammaanzaan
+\+ so.utsi er anza [+\+ utxt].xrsi er linganut he made him swear loy-
alty to himself (lit., to his own head), and us he made swear loyalty to his sons KBo
4.4 iv 5960; ammukwaza r.urorn-tt+:+ paai I am guarding my own temple
KUB 13.4 iii 2526 (pre-NH), (when someone has paid compensation for homicide,)
nuza() ta so.ut-si wata he has purchased his own life (lit., head) KUB 13.9
ii 34, nuzakan 2 r.ssxtn wtar t\+ so.ut.xrsi\i ar lauwanzi and the
two oerers pour the water on their own heads KBo 2.3 iv 45. The construction can
even omit the possessive: nuza r-a r-an kattan peket [
LU
a]ramaza
LU
aran
kattan peket brother betrayed his own brother, friend betrayed his own friend KBo
2.5 iv 1617 (AM); nuzakan ror.nr.-wa kuwattan +\+ xtn i.xn andan nikenun
numukan ror.nr.-wa i.xn rorn-pa t kuiki ni toward whatever enemy land I
directed my own eyes, no enemy was able to turn my(!) eyes back.
5
att. i 6769;
6

nuza
d
ts:+u o:+ par andandatar . . . tikkuanut ts:+u, my lady, revealed
her own sovereign power att. iii 1516; taz kieru r[ri ] and he washes his own
hands KUB 20.96 iii 6. In all these cases the -za has primary reference to the subjects
being the possessor of something mentioned in the clause; it does not construe primarily
with the nite verb. See German Sie waschen sich die Hnde.
4. On the double-accusative constructions see van den Hout 1992.
5. For the thought, compare
d
Kumarbiya ror.nr.-wa t namma manzazzi he was no longer able to
resist Kumarbis eyes KUB 33.120+ i 2021.
6. If the ror.nr.-wa in the second clause had belonged to the enemy, who is the grammatical subject of
nai, the particle -za would have appeared again in the second clause.
Particles 28.22
00-HittiteGrammar.indb 359 6/5/08 4:19:59 PM
360
Lexical Use of -z(a)
28.23. Some verbs exhibit signicant dierences in meaning, depending upon
whether or not they are construed with -za (Honer 1973a: 521):
7 8za + halzissa
Verb without -za with -za Bibliography
au()- to see (physically) to see with insight,
experience, understand,
see in a dream
Friedrich 1952: 38 sub
au-; Honer 1973a:
523; Boley 1993:
10616
pp- to seize, grasp to betake oneself to,
take refuge in (with -za
+ -kan)
7
Honer 1970, 1973a
- to sit, remain seated to take a seat, sit down Goetze 1933b: 45; Neu
1968b: 2728; Boley
1993: 7376
alzi(a)- to summon to call in the sense of
give a name to
iya- to do, make to celebrate (festivals),
worship (gods)
Boley 1993: 7786
ki- to occur, happen to become (something) But see Neu 1968b: 97
98; and Honer 1973a:
522 n. 14; Boley 1993:
4954
mald- to recite, speak to take a vow Laroche 1964: 812;
CHD mald-
ninink- to move, stir up
(something)
to muster (troops) CHD sub ninink-
8
peda- to dispose of to carry o withlfor
oneself
ak- to be aware of, know to know, master (a
skill), to recognize
(authority)
akuwaye- to look (toward
something, with d.-l.)
to see (something,
with acc.)
Honer 1973a
tar- to have the upper hand,
prevail, be able
to conquer (someone)
tepnu- to demote, curtail to belittle (verbally) Honer 1977
7. HW
2
E 5152, 6364.
8. In meaning 1 (mobilizing or mustering troops) the verb can occur with or without -za. But in the
other meanings (28) -za does not occur.
28.23 Particles
00-HittiteGrammar.indb 360 6/5/08 4:20:01 PM
361
28.24. The connection between the reexive meaning of -z(a) and its idiomatic use
with particular verbs is far from clear. Some, though by no means all, of the idiomatic uses
can be classied under the labels of (1) transitivity toggle and (2) transformative.
28.25. The particle -za often serves to distinguish a transitive from an intransitive
use of a verb (transitivity toggle; Honer 1973a: 52426). Note that in this transitivel
intransitive opposition the transitivity toggle -za always marks the transitive use. The
best-established case is with tar-: (1) trans. with -za to overcome, subject, (2) intrans.
without -za to have the upper hand, be superior. A second case is akuwai(a)- to look,
see (Honer 1973a: 524). The verb akuwai(a)- to look, see like its synonym au()-
occurs sometimes with -za and sometimes without. But the factor determining the need
for -za is not the same as with au()-, where it has to do with seeing something closely
associated with the seer. Rather it is the same factor which operates with the verb tar-:
when the verb takes a direct object, the particle -za will regularly accompany it: [(nukan
d
tft-u nepiaz) katta ]akuwait nuza
d
Ullukummin akuikezzi [(
d
Ulluk)ummiaza
(nepi)]an
d
tft-un akuikezzi The Sungod looked (without -za) down from the sky
and saw (with -za) Ullukummi, and Ullukummi saw (with -za) heavenly Sungod KUB
33.93 iv 2829 (Ullik. I A iv 3334). Beginning in MH, - also developed a transitive
use to occupy, settle in (a land, area). In this usage it shows -z(a): [k]awatta xtn
UR.AG
Zippal +o[ot\] nuwaza apnpat ei I have herewith given you the land
of Mt. Zippala; so occupy (or: settle in) it KUB 14.1 obv. 19 (MHlMS); see Goetze
1933b: 4 sub e; Honer 1973a: 525 with n. 25; and Boley 1993: 73.
28.26. In a few instances the particle -z(a) may be said to have a transformative
value. That is, it underscores a change of state. This usage, which is attested only in
post-OH, is clearest with the verbs ki- to happen; become and - to sit (down.
28.27. The occurrence of -za with ki- in post-OH correlates rather consistently with
its use as a linking verb (i.e., with a second, predicate nominative). When no second
nominative occurs (and in these cases the subject is always third person), the meaning
is to happen, take place, occur and there is usually no -za: nat uezzi umma +\+
i.xr xn
URU
Imirika watul kiari And a sin will occur for you men of Imirika
(i.e., you men will become guilty of an infraction of the treaty) KUB 23.68 rev. 6 (MHl
NS); gaza kiat A famine occurred HKM 113:10 (MHlMS), rorn-annakan t\+ X
xn inkan kiat And afterwards a plague occurred (i.e., broke out) in the army
KBo 5.6 i 910 (D frag. 28); iniwamu uttar karuwiliyaz pe[(ran)] t kuwapikki kiat
Such a thing never occurred before to me KBo 14.9 iii 78 (D frag. 28); nu maan
ameanza kiat But when spring came (lit., when spring occurred) KBo 2.5 ii 1
(annals of Mur. II).
28.28. When a second nominative occurs (and in these cases the subject can be any
of the three persons), the meaning is to become, turn out to be and there is usually a -za
or equivalent dative clitic pronoun agreeing in person with the subject: nu xt-ti mnia
armala maan n[uza ug]a ap kiat Like one sick throughout the cycle of the
Particles 28.28
00-HittiteGrammar.indb 361 6/5/08 4:20:01 PM
362
year, I became that one KUB 30.11 rev. 1213 (OHlMS); nuza k ap[ k]iaa[t]
I became that one KUB 36.79a + KUB 31.127 + KUB 31.132+ iii 19 (OHlNS); nuza
+ui:+ kuwapi urorn-tt+-i ui-at And when my father died (lit., became a god)
KBo 3.4 i 4 (NH); ammukmaza +\+ r+\t r:+ r xn kiaat But during the
reign of my brother I became an army commander att. i 24 (NH).
28.29. Exceptionally, in predicate nominative clauses with ki- no reexive particle
occurs: manwaramu
LU
+i:t:+ kiari he would become my husband KBo 5.6 iii
13 (NH); nukan +ui:+ kuin i.xn
URU
Gagan t\+ X xtn-:t txsio na 12 si:t kiat
The hostile Kaka which my father encountered in the land turned out to be twelve
tribal groups KBo 14.3 iii 1516 (NH).
28.30. The particle -z(a) does not occur with - to sit in Old Hittite (i.e., OS).
One distinguishes a transformative medio-passive sit down, seat oneself and a non-
transformative active sit, be sitting meaning of the verb - (the latter formally indis-
tinguishable from - to be, which probably survives in NH in the sense to abide,
remain (see n. 12, p. 366). Beginning in MH, the medio-passive alone is used for non-
transformative sit, be sitting and the medio-passive plus -z(a) for transformative sit
down. See Boley 1993: 7376.
28.31. Very probably the use of -z(a) with na- also marks the transformative use
to become afraid as opposed to the non-transformative to be afraid: (1) without -za
to be afraid of somethinglsomeone: itoi-uwa l nati O king, stop being afraid
KBo 15.52 v 14, see 21; namiwa I am afraid KUB 14.3 ii 26 (NH); nu nun and
I was afraid KUB 12.27 i 3 (NH); (2) with -za: to become afraid nuwazakan er
nun I became afraid on (that) account KUB 13.35 ii 33 (NH). Admittedly, it is some-
times impossible contextually to decide which meaning is appropriate. For this reason
we should concede that the use of -z(a) to mark the transformative use of na- may not
have ben carried through with the thoroughness that it was with ki- to become and
- to sit down.
Use in Nominal Sentences and with the Verb To Be
28.32. The particle -za often occurs in clauses with the verb to be expressed or
merely understood. Friedrich (HE [1960] 243) called such clauses Nominalstzen
(nominal sentences), although strictly speaking only the second type deserves that name.
The conditions for the use of -za or equivalent dative enclitic pronoun in such sentences
are the following (Honer 1969; see also 1973a: 52021): (1) In Old Hittite no -za or
enclitic reexive dative was employed, regardless of the person of the subject; (2) in
New Hittite either -za or a dative enclitic reexive pronoun is required when the subject
of the sentence is I, we, or you (sg. or pl.) i.e., a rst- or second-person subject;
(3) Middle Hittite is a transitional period for this rule: earlier texts follow the OH pat-
tern, later ones the NH pattern. Note that this rule applies only to sentences whose verb
is to be. Other rules govern clauses using other verbs and -za. Boley (1993) underesti-
28.29 Particles
00-HittiteGrammar.indb 362 6/5/08 4:20:02 PM
363
mated the validity of this rule, and many of her alleged counterexamples were explained
by Honer (1996b).
9
28.33. Since the period when -za is used as the distinguishing mark for rst- and
second-person subjects of the verb to be is New Hittite, we shall consider the New
Hittite evidence rst. And since most of our texts date from that period, it has the most
examples.
28.34. First-person singular subject: ammukmaza par andnza kuit t-a
eun but because I was a divinely-guided person att. i 4647 (NH); [amm]ukmaza
nwa ftn-a eun but I was still (only) a child KUB 19.29 i 10 (NH); nuza
d
tft-st
apez linkiyaz . . . parkui lit then let me, the emperor, be pure from that oath KBo
5.3 iv 32 (MHlNS); ammukmaza
f
Puduepa annalli orxrx+ but I, Puduepa,
am your long-standing maidservant KUB 21.27 i 7 (NH); ammuqqaza
f
Puduepa
arnwa xtts-za I too, Puduepa, am a woman of the bearing stool KUB 21.27
i 17 (NH).
28.35. First-person plural: anzmawanna ln.xr
d
tft-stpat but we are sub-
jects of His Majesty too KBo 4.3 iv 15 (Mur. II, ed. SV 1:144 145); anzamana
utxt.i.t
19
.it kuit because we are human beings KBo 11.1 rev. 15 (prayer of Muw. II),
ed. Houwink ten Cate and Josephson 1967: 109, 119 with comments on pp. 13435.
28.36. Second-person singular: fr-anzawaza Be alive! (a greeting) (KUB
33.106 iii 7); [xtts]-anzawaza xtts-niliyawa z zik you are a woman and of
a womanly nature (KUB 24.8 i 36);
10
zikza kui who are you? KBo 5.11 iv 24.
28.37. Second-person plural: nuza aliy[a] uddan mekki paaanuwante
ten so be very vigilant concerning the matter of the night watch! KUB 13.4 ii 7374
(pre-NHlNS); andamaza uma kui i.xr r urorn-tt+ furthermore you who are
temple servants KUB 13.4 ii 52; ummema kui i.xr so you who are eu-
nuchs KUB 26.1 i 6 (NH); see also ibid. iii 45; nuza mn parkuwa if you are pure
KUB 13.4 iv 53 (NS); takkuza paprantema but if you are deled KUB 13.4 iv 54
(NS); manwaza nte if you are angry KUB 15.32 i 46 (NS).
28.38. The following NH clause sequence contains one rst-person subject and one
second-person: appanwamuzakan [namma'w]aduza tuk rorn-pa mi be be-
hind me, and I will be behind you! KUB 36.35+ i 1213 (NS), see i 1 (nam-m]a-wa-
du-za tu-uk), ed. Otten 1953a; 1953b; Honer 1965.
9. We reject the arguments of Nowicki (2000), followed by Josephson (2003b: 21517), that ne-pi-i-
za in KBo 3.22:2 (OS) is to be analyzed as an endingless locative nepi + reexive -za in a sentence with to
be in the third person. None of the other alleged examples presented by Nowicki and Josephson for such
use of -za from any period are valid. We therefore retain the analysis of nepiza as an ablative. See p. 120,
n. 199.
10. Note that there are two clauses in this cited material: xtts-niliyaz contains a new set of sentence
particles.
Particles 28.38
00-HittiteGrammar.indb 363 6/5/08 4:20:03 PM
364
28.39. The particles absence with rst- or second-person in NH is extremely rare
and perhaps intentionally archaizing: utxt i.t
19
.it-a eun I was a mortal KUB
6.45 iii 26 (NH).
28.40. Contrast the following examples of NH to be sentences with third-person
subjects, in which neither -za nor a dative enclitic pronoun is required: nakan X
urorn.xr +\+
d
ts:+u
URU
amua naanza du so let him be reverent to ts:+u of
amua among the gods (att. iv 8889); kuitwa watultit what is your problem
(lit., sin)? KUB 24.8 i 45; kuiwara ai utxt-a who is he, this child? KUB 33.95
+ KUB 36.7b iv 14 (Ullik. I A); kwa kuit what is this? KBo 6.34 i 30 (MHlNS).
28.41. In Old Hittite and early Middle Hittite this use of the reexive to distinguish
rst- and second-person subjects from third-person ones in to be clauses was not yet
operational. Observe the non-occurrence of the reexive in OH and early MH to be
sentences with rst- and second-person subjects: First-person singular: [(
m
appi )]
+\+ i.xr
URU
Zalpa tarikizzi kwa a[(tti)]m[(i ) natt]a u appi says to the men
of Zalpa: I am not dear to my father KBo 22.2 rev. 45 (OS); gga
MUNUS
annanna
mi I too am an a.-woman VBoT 58 iv 3 (OH or MHlNS); []k uit
GI
ftxti eun I
was a Lord-of-the-Weapon KBo 3.35 i 9 (OHlNS). First-person plural: [xi.nnnn-
wa uitt-\t l]wannakan kuei weawa [nepia takna]a lalukkimi O Silver,
our lord!] Do not kill us! We are the luminaries [of heaven] and [earth]! KUB 36.19
iv? 1314 (OH?lNS) (see Honer 1988b: 160). Second-person singular: zigawa
GI
ftxti you are a ftxti(-man) KBo 22.1:21 (OS); maranzawa zik you are treach-
erous KBo 3.34 ii 20 (OHlNS). Also in a NH ritual but imitating archaic syntax: utxt-
annai uxsi zik you are his child-producing wife KUB 7.8 ii 34 (see Honer
1987: 280, 285). Second-person plural: umea [urorn.xr-a u]ddani naante
ten Be respectful toward the word [of the gods] KUB 1.16 iii 49 (OHlNS); i.xr
ttxtwa ume you (pl.) are men bound to render ttxi-tax. Notice how in OH the in-
dependent pronoun (zik, ume) often occurs in nal position, when the verb - is left
unexpressed (22.3, p. 306).
28.42. A similar use of dative pronouns referring back to the subject is used in nomi-
nal sentences in Hieroglyphic Luwian, where one nds in sentences with rst-person
singular subjects (Latinogram EGO I) the regular use of the enclitic pronoun -mi:
EGO-mi U+rali-hi-li-na I am Urhilina (Restan and Qalat el Mudiq = Hawkins 2000:
(2:)409. On this subject see F. Josephson in Houwink ten Cate and Josephson 1967:
13536 with anterior literature cited there.
The Local Particles -an, -apa, -ata, -kan, and -an
28.43. The particles -an, -apa, -ata, -kan, and -an belong to a single class of sen-
tence particles which modify the action expressed by the main verb and its adverbal
adjuncts. The nature of that modication is disputed: some thinking that it marks pri-
marily local relationships (Goetze 1933b; Carruba 1964; HE 28788), and another
28.39 Particles
00-HittiteGrammar.indb 364 6/5/08 4:20:03 PM
365
(Joseph son 1972) that it marks verbal aspect. These two viewponts are not mutually
exclusive (Carruba 1964: 43236; Boley 1989; Honer 1992b; and Josephson 1995).
We retain here the term local particles as a convenient means of distinguishing this
group from other sentence particles such as -za (likewise Luraghi 1997a: 5253).
28.44. The normal position of these particles is last in the chain of sentence enclitics,
which chain is suxed to the rst accented word of the clause. Less commonly, these
particles can be suxed to a dative-locative in the interior of the clause (Neu 1993).
11

When this happens, the same particle or another local particle in the normal position
may also occur at the head of the clause: nata par uian ar urorn.i.xr-a
pdai He carries (it) out (ata with par) up onto (an with ar) the roof for the male
deities KBo 21.33+ iv 2021 (Neu 1993: 138); nukan ma[an] ru.otn
4
.n +\+
LU.ME
xtniurx andizzi nata aranan akan anda luwi When he prepares
thick bread(s) for the cooks, he pours out (ata . . . lauwai) the arana-liquid into the
brazier (kan anda) KUB 27.69 ii 48. As Neu correctly notes, the two local particles
in the same clause refer to distinct local notions.
28.45. The rst two of these particles (-an, -apa) are found chiey in documents
composed during the Old Hittite period. With the disappearance of these two, and even-
tually of -ata, their functions were assumed by -an and -kan. This fact makes it im-
practical to use the majority of Hittite texts, which either originate or were recopied
with modernizations during this late period, to determine reliably the original functions
of even -an and -kan, since many of their uses in the late period originally belonged
to the now defunct -an, -apa, and -ata. In our discussion of each particle we present
representative examples of their usage, organized into coherent groupings. Speculations
about a possible unitary or basic meaning and etymology are reserved for the end. The
presentation of the evidence is illustrative, not exhaustive (except for the very rare -an),
and even the classication of the material is tentative.
28.46. These ve particles occur in complementary distribution, and thus the pres-
ence of any one in a given chain of sentence particles excludes the others (Honer
1973a; Kammenhuber 1979b).
28.47. Of the ve, -kan by far occurs most frequently. It also has the least restrictive
sphere of meaning, including in its own semantic range the nuances of several of the
other four. For this reason, especially in NH copies of OH or MH compositions, one can
nd almost identical sentences, one of which contains -kan and the other -an, -ata,
-apa, or -an. The same cannot be said for any of the other four particles, which dier
markedly from each other. The rst clues to the meaning of these particles were found in
sentences containing verbs of movement (Goetze 1933b; HE 295). This does not, how-
ever, indicate that the conditioning factor for the use of the particles is always the main
verb in its clause. Good evidence exists that often the more pertinent consideration is
11. For the same occasional irregularity in Hieroglyphic Luwian see -mi-ya-za-ta with local particle
-tta in KARATEPE XXX and comments by Meriggi 196675: IIl1, p. 93.
Particles 28.47
00-HittiteGrammar.indb 365 6/5/08 4:20:04 PM
366
the presence or absence of a local expression (noun phrase in the locative, local adverb,
postpositional phrase, or preverb) (Goetze 1933b: 1630; Tjerkstra 1999: 131 40).
The Particle -kan
The Local Value
In To Be Sentences
28.48. The local value of -kan is clearest where it co-occurs with a local adverb
in nominal sentences and in those with the verb to be. This usage is not yet attested
before MH, although it is found in OS with the particle -an. For one interpretation of
this fact see Boley 2001.
28.49. With anda with, in: r.xtn.n.nr.yamakan kui anda And also
the horses that are with you (move them into His Majestys presence with haste in three
days!) HKM 15:89 (MHlMS); numukan kui idalu memia zr-ni anda The evil
thing that is in my soul . . . KUB 6.45 iii 4546 (prayer of Muwatalli II); nu
m
a-i-
i kuit oro-at na[ka]n aruni anda ta and because Uaziti became ill, he remained
on the island (lit., in the sea) KBo 3.4 ii 5051 (NH).
12
28.50. With ppan behind:
GI
ftxti s+
LU
xt
7
.xi.[o]rmakan kui rorn-an But
the weapon of the gold-spear man that is behind . . . BrTabl. i 4445 (Tud. IV). See
with -an: ln.xr urorn-tt+yaan kui t\+
URU
Kappri rorn-an eer And also the
servants of the deity who were behind Kapperi KUB 19.37 iii 3839 (NH) (and ibid.
42 and 44).
28.51. With er up: rnr.xryakan panku er ta And the entire army was
up (there) KUB 14.16 iii 11 (NH); wtarmakan kuit +\+
UR.SAG
Arlanta er But the
water that is up on Mt. Arlanta . . . KBo 4.10 obv. 25 (NH).
28.52. But no particle seems to co-occur with kattan below, down in such sen-
tences: nu kuitman +ui:+ t\+ xtn
URU
Kargami
i
kattan ta And while my father
was down in Kargami KBo 5.6 iii 12 (NH). Lack of a particle also seems regular
with peran: [. . . . . .]mamu peran arpuwante e [er] [The roadslthe mountains] were
arduous ahead of me KUB 19.37 iii 50 (NH). We do nd -kan with the combination
peran - when it means be responsible for (lit., remain before
12
): nuwakan xtn!-e
peran ten Be responsible for the land! KUB 14.16 i 17 (NH).
With Local Adverbs and Non-motion Verbs
28.53. The particle -kan likewise occurs with the same local adverbs in other sen-
tences with non-motion verbs and some local expression. In none of these cases is the
12. This active - is the stative (non-eventive or non-transformative; see 28.30, p. 362) counterpart to
medio-passive - take ones seat, sit down (Hrozny 1917: 165; Goetze 1927: 100105), and not the verb
to be. Supporting this analysis is the example in KUB 14.16 i 17 cited below in 28.52. In the case of the
imperative ten the verb to be would have taken the particle -za.
28.48 Particles
00-HittiteGrammar.indb 366 6/5/08 4:20:04 PM
367
local adverb a preverb construed with the verb, since it in no way modies or limits the
meaning of the verb.
28.54. With anda in (beginning in MHlMS): (The carpenter heard himself being
cursed,) nuikan anda i [dl ]aweta and (his mood) became sadlbad within him
KBo 32.14 rev. 44 (MHlMS); also in KUB 33.120 iii6768 (NS) with X-ni added; and
in KBo 26.65 iv 10 (NS), KUB 36.25 iv 2 (NS) and KUB 36.35 i 2425 (NS) with ex-
plicit subject zr-anza (NS); the same expression but with [anda]n KBo 32.14 left edge
12 (MHlMS);
13
nukan
m
a-i-i aruni anda n. Uaziti died on an island
(lit., in the sea) KBo 3.4 ii 52 (NH); nuzakan ntn.so.xr-a anda du eta
He made resting places for himself in the mountains KBo 5.8 ii 28 (NH).
28.55. With ppan behind and ppa back: nukan (var. -an) alkiu rorn-an
ma[(knunun)] and I multiplied crops behind(?) KBo 3.1 iii 44 (OHlNS), ed. TH 11:38
39; [namma]atakkan rorn-anpat kittat [then] he kept after you (and kept pursuing
you) KUB 14.1 obv. 2 (MHlMS); nuttakkan
m
Attariiyan rorn-an ara ka[rta] and
I cut Attariiya o from (going) after you KUB 14.1 obv. 4 (MHlMS); nammakan
m
Madduwatta i.xr
URU
Dalawa +\+ xtn
URU
atti rorn-an arapat nai Then Mad-
duwatta turned the men of Talawa away from following (lit., from behind) the land
of atti KUB 14.1 obv. 73 (MHlMS); nukan antutar
URU
atti +\+ tnt.uruir.nr.
URU
Gaga rorn-an kui

arzanan arker kuimakan ppiri rorn-pa pnte Some

of the Hittite population had hostel(s) behind the Kaskean towns, while some had gone
back into town KBo 5.6 i 1517 (NH).
28.56. With er up: When I arrived in Atata, I went up in the city Atata, nukan
gurtan er wetenun And I built a citadel up (there) KBo 4.4 ii 6162 (NH).
With Local Adverbs and Motion Verbs
28.57. The particle -kan is generally absent when a motion verb has no associated
preverb, even when an explicit local expression is present: nu t\+ fin pun and I went
to the sheepfold KBo 17.61 rev. 8 (MHlMS); t\+

inapiwara pait he went to the

inapi-building KUB 30.28 rev. 4 (NS);
d
tft-st kuit t\+ xsxi
URU
Nerik t pait Be-
cause His Majesty did not go on a trip to Nerik KUB 2.15 v 2223 (NS); na t\+ xtn
URU
ayaa pait And he went to the land of ayaa KUB 19.11 iv 41 (NH). But there
13. Aside from nuikan zr-anza anda nti-ue[ta] and she became deeply distressed KUB 36.35
i 2425 (Aertu and Elkunira myth), which to our knowledge no one has dated earlier than NH, the NH
continuation of this construction lacks anda (n) and the local particle: (If anyone disposes of contaminated
materials from a purication ritual on anothers property, he must purify that house,) mn r-riya kuitki
idlawzi and if anything goes wrong in the house, (he shall purify it again and make compensation for
the loss) KBo 6.4 iv 12 (NH); see also Laws 31 NS copy B, and KUB 11.1 iv 26 (OHlNS). The mean-
ing of idlawe- in this construction is dierent from that expressing hostility, which is without anda and
local particle: nu 1-a 1-edani idalwezi KBo 5.4 rev. 3 (NH), or is a plural reciprocal verb without dative-
locative or anda but with -kan: nukan l idalwtni dont be angry with each other KBo 5.4 rev. 21
(NH).
Particles 28.57
00-HittiteGrammar.indb 367 6/5/08 4:20:05 PM
368
are exceptions, especially with the verb ar-ler- to arrive: maanwakan ammuk
t\+ xtn
URU
Iupitta run When I arrived in the land of Iupitta HKM 10:3435
(MHlMS); versus mn
m
Tudaliya itoi.[oi]
URU
attui run When I, Tudaliya,
[great] king, arrived in attua KUB 23.11 iii 1213 (MHlNS). The verb uwai- to
run also takes -kan in combination with aan, accusative of hearth: aankan
1si yanzi they run around(?) the hearth once KBo 25.46:5, etc. (OH).
28.58. Use of -kan (or alternatively -ata or -an) is frequent with intransitive or
transitive verbs of motion when these are accompanied by a preverb, but there are also
examples where there is no particle, and the frequency of the use of the particle can vary
markedly from one preverb+verb combination to another.
28.59. Examples of anda in(to), to + motion verb with -kan: +\+ otxt.xtts
oimakan anda auli namma neut Turn again towards the princess in benevolence
KBo 4.6 obv. 16 (NH); +\+ i.xr:
URU
Pauwamaakan anda pait But he went
to the men of Pauwa KUB 23.72 rev. 17 (MHlMS); andakan alna teummiu
tarlipit wamu 2-:++ ptumini We twice(?) bring in two clay vessels lled with t.;
KBo 17.1 i 2627 (OS) (see ibid. i 3132, 34 with particle -an!); mn rorn-pa kuwatga
wanuzi nakan xtn-ya anda uezzi If he turns back somewhere and comes into the
country HKM 46:1012 (MHlMS).
28.60. Examples of anda + motion verb without -kan: tu anda

alentiu [(pdanzi)]
And they bring them into the palace KBo 17.11+ i 47 (OS); nu
m
imu-ororn-tt+
kuit +\+
m
Kapiya i +\+
m
Zilapiya pai nat anda arnunun What imuili had given to
Kapiya and Zilapiya I collected (and gave it to the men of Kaipura) HKM 68:1921
(MHlMS); numu xn: t\+
URU
arrana anda ra My army arrived in arrana KBo
4.4 iii 27 (NH); nunna +ui:+ anda arnut My father brought us together BrTabl. ii
46 (Tud. IV); nui +ui:+ zo anda uittiat My father moved in (i.e., reduced) his
boundaries BrTabl. i 23.
28.61. A verb of directional movement with andan towards as a preverb appears to
require -kan in NH: nuzakan ror.nr.-wa kuwattan +\+ xtn i.xn andan nikenun
Towards whichever enemy land I turned my eyes . . . att. i 6768.
28.62. No particle appears in NH when andan occurs as a postposition with a place-
name (see Salisbury 1999: 69) and the motion verb thus is without a preverb: nu t\+ xtn
URU
Takkuwaina andan pun I went to(wards) the land of Takkuwaina (and while I
was crossing the land of Italubba . . . ) KBo 5.8 i 31 (NH). Note that the example cited
makes clear that andan in this construction does not mean specically in(side).
28.63. The preverb ppa back plus a motion verb does not regularly take a particle:
(And when the King of Egypt had conquered the Land of the Hurrians) [(nu ro)]rn-pa
xtn Aba pait He went back to the land of Aba KUB 21.17 i 17 (NH); lukkattama
t\+
URU
Iyaria rorn-pa uwanun But on the next day I came back to Iyaria KBo
4.4 iii 40 (NH).
28.58 Particles
00-HittiteGrammar.indb 368 6/10/08 7:52:49 AM
369
28.64. But a few examples of ppa + motion verb with -kan do occur: numakan
d
z.n.n-a
GI
ftxti.nr.xi\ [i] ppa nu May Zababa turn your weapons back
against you! KBo 8.35 ii 1920 (MHlMS); +\+ tnt.uruir.nr.ma kueda sro
5
-r [i
n]ukan inkan rorn-pa l paiz[zi ] (In the cities where people are dying from plague,
let there be recovery,) and let the plague not come back to the cities in good health
KUB 14.13 iv 2021 (NH); nankan rorn-pa t\+ xtnsi putezzi And he leads him
back to his own land KUB 24.5 obv. 26 (NH).
28.65. Examples of ara away, o + motion verb with -kan: r:+wakan edaza
ara aru May my lord get away from there KBo 4.14 iii 34 (NH); Because the booty,
civilian captives, cattle, and sheep had become too much for me, nukan namma ammuk
ts:i xn ara t pun I could no longer get away from the camp KBo 5.8 iv 1415
(NH);

1-amakan
m
T [apala]zunauli arunaz ara uet T. came away from the sea
alone KBo 3.4 ii 54 (NH); nankan ara uwater And they brought him away KBo
3.4 + KUB 23.125 iii 8 (NH) (vs. nan ara uwater ibid. iii 12!); in addition with the
verb kuer- the particle -kan serves to distinguish to cut o from (with -kan) from to
cut up or to cut apart (with a double object)
14
(both without -kan). See nukan ts:i 9
UZU
n ara ku[ermi] [I] cut o from nine members KBo 13.101 i 13 (NS). See 28.110
(p. 382) for a similar use of -ata.
28.66. Examples of ara + motion verb without -kan. The one clear generalization
is that the construction -za ara pai- go home never takes -kan: naza ara t\+
xtnsi pait He went home to his own country KUB 14.1 obv. 65 (MHlMS). Other
examples: na arapat pait And he went away, (he didnt come near me at all) att.
ii 49 (NH); nan ara putet He carried them (troops and chariotry) away att. ii 51
(NH); namma ara uwanun nu maan t\+
URU
ammaa run Then I came away, and
when I arrived in ammaa KBo 3.4 iii 45 (NH); nu ara [t\+]
URU
Itaara uwanun I
came away to Itaara KBo 2.5 i 78 (NH).
28.67. Examples of katta down + motion verb with -kan: nukan x.n.xr katta
uer The civilian captives came down (and they prostrated themselves at my feet) KUB
14.16 iii 16 (NH); nankan s+
LU
t.un iwar katta uwatenun I brought him down like
a prisoner KBo 3.6 + iii 73 (NH).
28.68. Examples of katta down + verbs of directional movement without -kan:
[ito]i
URU
Kuara tnt-az katta [ pa]ngarit u[et] The King of Kuara came down
from the city en masse KBo 3.22:5 (OS); orx-anma r:+
m
rn.oi ts:i +++: ur-
orn-tt+si t\+ xtn s+rtt:t katta pait But when my brother Muwattalli went down to
the Lower Land at the word of his god att. i 7576; na t\+
URU.d
t-taa katta pda
And he carried them down to Taruntaa att. ii 53; na ts:i xtn
URU
tot-:t katta
uet He came down from the high country KUB 19.10 i 8 (D, NH).
14. uerma -i kuwapi tekanna
URUDU
kuruzzit ara kuerer When they proceeded to cut apart heaven
and earth with a cutting tool KBo 26.65 iii 42, see iii 5253 (Ullik., NS).
Particles 28.68
00-HittiteGrammar.indb 369 6/5/08 4:20:06 PM
370
28.69. Examples of par out, forth + verbs of directional movement with -kan:
par arnu- to send nakan +\+ xrsin par arnutten And send it (grain, pl.) to the
threshing oor HKM 25:1819 (MHlMS) (but see also with -ata in HKM 52:4446);
par nai- to send: nankan ka par nei I am sending him right now HKM
2:1213 (MHlMS) (and often in the MH Maat Letters, but see with -ata in KUB
14.1 obv. 61, MHlMS); nankan t\+ xtn
URU
Nuai par neun And I sent him to
the land of Nuhassi KBo 4.4 i 40 (and often in NH). See nata +\+ xt
URU
rmana
tamai antua par l paizzi Let no other man go out for the salt of Sarmana! BrTabl.
ii 1415 (Tud. IV).
28.70. Examples of par out, forth or forward, further on + verbs of directional
movement without -kan: nat +\+
d
rxtn +utx+ [i +\+
d
tft
URU
r-n]a xx+ par
arnut Pass it (my plea) on to the Stormgod your father and the Sungoddess of Arinna
your mother! KUB 21.27 iv 2122 (prayer of Muwatalli); nan par uittiyanun nan
r-t++ iyanun I picked (lit., pulled) him out and made him a lord KBo 16.17 + KBo
2.5 iii 2526; lukkattima t\+
URU
Taptina par iyaat But on the next day I marched
on towards Taptina KBo 4.4 iii 43; itoi-i par 1si paizzi He goes forward to the
king one time KBo 17.43 i 1112 (OS); numuza r:+
m
rn.oi par uket My
brother Muwattalli used to send me forth att. i 66. The adverb par with the meaning
forward, further on appears never to occur with a local particle (Goetze 1933b: 128;
Tjerkstra 1999: 64; CHD par 3).
28.71. Examples of ar up + verbs of directional movement with -kan (MH and
NH): [ar ]atkan rkandu let them keep coming up KUB 40.56 + KUB 31.88 iii
8 (MHlMS); nakan
URU
Gaipraz ar sro
5
-in arnut Send them (the prisoners) up
well from Kaipura HKM 65:1112 (MHlMS); nanz[ak]an tnt-ri ar peuteddu
And let them (troops) carry it up into the city for themselves HKM 24:50 (MHlMS);
[
URU
Tma]nnazmaakan
UR.SAG
Ka ar pait But from Tumanna he went up Mt.
Kau KUB 19.13 i 42 (NH); na rorn-pa paiddu tnt-rimaakan ar l uezzi Let
him go back! Let him not come up into the city! KUB 21.29 ii 12 (NH); tntankan
ar l uwadazzi Let one not bring him up to the city! KUB 21.29 iii 3738 (NH) (see
with -ata in HKM 43:3, MHlMS). For other exx. (. uittiya-, iya-
ttari
, ipart-, penna-,
peda-, tarna-, uiya-, up-, uda-, watku-) see CHD ar B. The adverb ar + some mo-
tion verbs, such as arai- to arise and pp- holdllift (up), take -kan only when there
is an ablative in the clause (see CHD ar B 1 a 3).
28.72. Examples of ar +motion verb without -kan: a ar tnt-ya pait And he
went up to the city KBo 22.2 rev. 14 (OS); nat t\+ r.urorn-tt+ tot pnzi And they
go up to the temple KBo 2.8 i 21 (cult inv., NH); tat ppa ar l uzzi And let them
not come back up again! KBo 17.1 iii 1213 (OS).
28.73. The particle is also found in certain other verbal constructions intended to
convey local restrictions (Goetze 1933b: 30 n. 2): pedi ninink- to muster on the spot,
28.69 Particles
00-HittiteGrammar.indb 370 6/5/08 4:20:07 PM
371
pedi daliya- to leave on the spot, A-ta tarna- to take to heart, kiari dai- to put in
the hand, xsxi-i dai- to put on the road.
28.74. An interesting contrast is between an- with and without local particle (Ho-
ner 1992b). Without -kan or -ata, an- denotes to search for (someone or some-
thing), but when the verb has one of the local particles, it means to search (an area or
place). This usage certainly reinforces the other evidence for a local dimension to the
presence or absence of this class of particle.
28.75. Goetze (1933b: 30) noted another small class of verbs which regularly em-
ploys -kan or -(a)taverbs of crossing over: arra- and zai-. To these should also be
added itarna ara pai- to pass through, traverse (with acc.). In fact, there is an oppo-
sition between arra- without local particle, which usually means to divide, separate,
and with local particle (-ata or -kan), which means to cross over, transgress (see
CHD ar (r )a- D). This distribution shows that the earlier assumption (compare the
translation verletzen in HW 183) that with oaths and commands it meant to break,
violate (< to divide) was incorrect: the semantic development was from to go across,
traverse to to transgress. The command or oath is also a kind of boundary or limit
in an extended sense. In HW it was not even noted that the usage with oaths, like that
physical traversing, was associated with -kan. Somewhat in the same semantic domain
is the verb amen- which with -kan means to pass by and, by extension as a transitive,
to forfeit, relinquish: (Let no one do business with a ippara-man . . . whoever does
business with him) nakan pparaz [eme]nzi shall forfeit his purchase price KBo
6.2 ii 5859 (Laws 48, OS) restored from KBo 6.4 iv 39 (NH), and other exx. in CHD
120. Other exx. are post-OS.
Accompanying the Dative of Disadvantage
28.76. The verb akk-lekk- to die is a change-of-state verb. Normally it does not
take a local particle (-kan, etc.). But when there is a dative expression indicating the
person(s) most aected by the subjects death (almost certainly the dative of disadvan-
tage, 16.68, p. 258), -kan regularly occurs (compare informal American English then
my wife died on me), all exx. NH: numakan rsi\i kuit
m
Pipuriya immakku
n. And because their lord, Pipuriya (throne name of the pharaoh), had just died on
them, (the men of Egypt became afraid) KBo 5.6 iii 78 (NH) (for immakku here see
Honer 1995b); i-awamukan n. utxt:+mawamu t.oi My husband
(lit., man) died on me, and I have no son KBo 5.6 iii 1011 (NH); and ammelwa
LU
+iot:+ kui ta nwaramukan n. He who was my husband died on me KBo
5.6 iv 45 (NH); nukan BE-an +\+ itoi ln.xr zr akkanzi If (his) favorite servants
die on the king KBo 4.14 iii 3738 (NH), and i+++
d
tft-st++ +\+
m
u+u+-+nt-toot\
tut++ rxi\i makan kuit n. . . . +uisiikan kuit n. Thus says His
Majesty: Say to
m
Baba-ai-iddin: Because your (pl.) lord (the Assyrian king) has died
on you (pl.) . . . because his father has died on him KUB 23.103 + KUB 23.92 rev. 89,
Particles 28.76
00-HittiteGrammar.indb 371 6/5/08 4:20:07 PM
372
16 (NH), nutta
d
Gula nti-ada natakkan n. your luck turned against you,
and he died on you KUB 23.85 rev. 6 (NH). The adverb -kan (or -ata) also regularly
occurs with the verb d- when it means take away from and with the combination ara
d- to take away from. This too is the dative of disadvantage: namai piyan kuit
armi nuikan ara kuitki di Or takes away from him any of what I have given
him BrTabl. iv 1819 (Tud. IV), even if in some instances what is removed is an evil
(e.g., ABoT 4a i 1114, OS). Similar is its use with the verb tu- to separate (from):
(If the parents disapprove of the marriage of a young woman to the man who abducted
her,) nanikan tuanta they may separate her from him KBo 6.3 ii 10 (Laws 28,
OHlNS). See further exx. in 16.6816.69 (pp. 258259).
The Terminative Value
28.77. Some uses of -kan have no discernible local dimension. The clearest case is
that of the verb kuen-, which is not a verb of movement, never occurs with a preverb,
and only rarely and coincidentally appears in clauses with an expression for a location.
When kuen- occurs without -kan, it means to strike, hit; with -kan it means to strike
dead, kill. Here, then, -kan seems to have a terminative force; it adds nality to the
action described by the verb. One nds a similar use of -kan with (ara) ap- to de-
stroy, annihilate. Possibly the OHlOS use of of -kan in rnr.xr-:tkan mauzi the
army will fall (be defeated) KUB 4.72B:3, and the OHlNS use of -kan with wak- to
bite o (takkan wki KBo 17.74+ iv 4) should be explained on this basis. But the
nuance added by -kan can be quite subtle. The verb (ara) arnink- to destroy takes
-kan only when the verbs subject is a deity, whose destroying can be seen as having
a special degree of nality (Honer 1992b).
With Verbs of Hostility
28.78. Another class of verbs regularly employing -kan has a common semantic ele-
ment of hostility: pran wal- to strike rst, preempt (a competitor) (already in Laws
146, OS), ulle- to show disrespect (only NH; wlo particle in MHlMS), idalawe-
to become evil, hostile, uwapp- to be ill-disposed toward, zammuri- to alict,
oppress. See 28.9328.94 (p. 377) on -an with similar verbs.
Aspect Independent of Verbs Lexical Meaning
28.79. Since many studies of the particles -kan, -ata, -apa, etc. focus on the lexi-
cal, in that they list all verbs which take a particular particle, it is important to note
that the semantic aspect conveyed by -kan (for instance) is shared by all verbs in a
lengthy sequence: ts:i
GI
ntnmazakan kuzza azzikkenun ts:i oiyakan
kuzza akkukenun atiyazakan kuedani ekekenun ts:i
URUDU
ut
10
.yazakan
kuzza arrekenun the table from which I regularly ate, the cup from which I regularly
drank, the bed in which I regularly slept, the wash-bowl from which I regularly washed
myself KBo 4.2 iv 2832 and duplicates (see Goetze and Pedersen 1934: 1011, lines
28.77 Particles
00-HittiteGrammar.indb 372 6/5/08 4:20:08 PM
373
1619). This suggests that the aspect shared by all these clauses is independent of each
particular verbs meaning. Note that the same is not true of the particle -za in the above-
cited sequence: it is omitted in the akkukenun clause. For tentative observations on the
presence or absence of -z(a) with eku- see Honer 1973a: 52425.
Absence of Expected -kan
28.80. Not fully claried is the circumstance that -kan fails to occur where expected,
when the clause begins with takku if (HE 298,2): takku
LU
ux.oXn
URU
atti kuiki
kunzi if someone kills a Hittite merchant KBo 6.2 i 3 (Laws 5, OS); takkuu i-i
wemiyazi tu kuenz

i if the man (i.e., husband) nds them (his wife and her lover in the
act of adultery) and he kills them, (he will not be accused of murder) KBo 6.26 iv 89
(Laws 197, OHlNS). In both cases the verb kuen- in the meaning kill occurs, which
normally requires -kan. Both examples which Friedrich cites from the laws in clauses
without takku which use -kan kuen-, namely, Laws 90 and 199, are found in New
Hittite copies. In the latter there is even alternation: kuenziu itoi-u . . . nankan
kunanzi KBo 6.26 iv 17, 21 (OHlNS). All examples of takkukan occur in NH transla-
tions of Akkadian omen texts.
Summary
28.81. It is hard to make generalizations about the overall use of -kan that are valid
for all the available data. We have not handled all problematic attestations. For example,
the signicance of the use of -kan with OHlOS gank- to hang up (an object) and lip- to
lick up is still unclear, as is its use with d- to take and an instrumental but no da-
tive (from) in KUB 30.30 (OS). Contrary to what was once claimed, the use of -kan,
-ata, or -an with ipant- to oer, libate, consecrate cannot easily be distinguished in
meaning or in the nature of the direct object from its use without a particle. There does
seem to be some correlation between the lack of a specic starting or end point for an
action and the absence of -kan: note examples such as anda pai- to converge, anda
uittiya- to pull in, reduce, par pai- to go (further) on. This pattern suggests that,
conversely, the presence of a starting or end point (explicit or implicit) that limits the
action described is associated with the use of -kan. However, the positive evidence for
this association is not as regular as one would expect. This problem remains even if one
excludes OH on the basis that the usage of -kan was not yet fully established then (see
Tjerkstra 1999: 144, following others). It does seem reasonably certain that one must
allow for both local and aspectual functions for -kan, but the boundary between these
is not entirely clear. Starting and end point may be construed in spatial or temporal
terms, and development of a terminative value from a particle with an original local
meaning is commonplace. One may therefore derive -kan from a PIE aspectual particle
`-ken seen in Sanskrit (n ) kam and Greek (n ) ken (Josephson 1972: 418 and others),
from a local adverb `kom by, with, next to seen in Latin cum with but also in Latin
con-cd to cut uplto pieces, etc. (Josephson 1976: 173), or from both (see Dunkel
Particles 28.81
00-HittiteGrammar.indb 373 6/5/08 4:20:09 PM
374
1990: 11522). On the other hand, the view of Boley (2001), that -kan originally had
an entirely non-local sense which led to the other uses, is hard to reconcile with the
evidence we have presented here.
The Particle -an
28.82. The particle -an, like -kan, expressed a relationship largely spatial and so
closely related to that conveyed by -kan that it occasionally substituted for -kan in
sentences of the type outlined above for -kan. Particularly dicult is the distinction in
coordinated sentences such as: nuan paraya [l] naitti rorn-payakan l mauta
Neither postpone(?) (the oerings to the gods); nor fall behind(?) KUB 1.16 iii 5152
(OHlNS) (so tentatively CHD LN 363). Yet these two particles are by no means syn-
onymous. Since -apa and -ata (particularly -apa) were used principally in Old and
Middle Hittite and -apa became rare in New Hittite, their functions tended to be as-
sumed by -kan andlor -an. In particular, -apa, whose function was to convey the notion
of movement up against something, was replaced by -an. Some functions of -ata, on
the other hand, seem to have been assumed by -kan.
28.83. In Old Hittite texts the writing -za-an can be interpreted either as (1) -za
and -an (local particle) or (2) -za and -an. In New Hittite texts, since the local par-
ticle -an no longer existed, the same sequence either had to be (2) or in the case of
nanzaan, which one can compare with natiat the second an is a meaningless
repetition of the rst (him) (see 30.19, p. 411).
28.84. In Old Hittite texts, where the clitic possessive pronoun his on singular
common-gender nouns has the vocalization -an, one must be careful not to confuse
this with the local particle, especially in cases where the possessive pronoun occurs on
the rst word of a clause.
28.85. We noted above in 28.48 (p. 366) that the local value of -kan is clearest
where it co-occurs with a local adverb in to be sentences. From this point of view one
should note in OS an example of -an in a to be sentence without er above, upon
but with the meaning upon: nean
[N]A
4
pruni wetan and it (Labarnas palace) is
built upon a rocklcli KUB 36.110 rev. 1516 (OS). Its early date, combined with the
absence in the clause of complicating factors, such as a nite form of a particular verb
of movement or a local adverb, gives this example particular importance.
Co-occurring with the Adverb er
28.86. The particle -an is particularly common from OS onward in clauses con-
taining the adverb er above, over or verbs of putting, placing, or standing, which
can occasionally also be construed with -kan. The occurrence of -an probably gives
to these verbs the particular nuance stand over, place upon, etc. Examples with er.
[u]gamaan rnr.xr-an [(r )] 3si wanmi but I wave the soldier(-bread) over
28.82 Particles
00-HittiteGrammar.indb 374 6/5/08 4:20:09 PM
375
them KBo 17.1 ii 1718 (StBoT 8 ii 3132) (OS). In addition to physical superposition
(upon), this construction is found from OS on for relative conceptual position (upon,
in addition to, instead of): (if someone is in the process of selling a house, a village, a
garden, or a pasture, but another (seller) goes and strikes rst,) taan [(appari ) ]r
appar izzi and makes a deal of his own instead (lit., makes a sale upon a sale) KUB
29.29 ii 910 (Laws 146, OS) restored from NS copy; see Honer, LH 121.
28.87. Without er but with main verbs implying superposition. Physical superpo-
sition: taan almauitti tizzi and he (i.e., the king) steps upon the throne dais
KUB 43.30 ii 16 (OS);
NINDA
arruim[(a )]an rnr.xr-az zi the soldier-bread
is sitting on the arruwa-bread KBo 17.1+ i 30 (OS) restored from KBo 17.3 i 24
(OS); nuan [ . . . ] ai and seats [ . . . ] upon [ . . . ] KBo 3.22:2728 (OS);
[k]ma
m
Murili utxt:[+ nu]za ap[n ekten] [nu]an apn aete![n] Murili
is hereby my son. Recognize him (as such), and seat him (on the throne)! KUB 1.16
ii 3738 (OHlNS); attatawa [
GI
.i utxt.rf t oi l]n-iwaan eari
[There is no son for the throne] of your father; a (mere) subject (lit., slave) will sit
down upon it! KUB 1.16 ii 7071 (OHlNS); walanzian they beat on (them, i.e.,
musical instruments) StBoT 13 passim (OS); naan
GI
priya nnnn ki [tta] and
it (a bowl) lies on white wooden puriyas KBo 20.8 rev. 5 (OS) (compare raa[n
. . . ]
GI
pria nnnn kit[ta] ibid. 1011); numuan k yugan i[(aite)n] [you
have] bound this yoke upon me KUB 31.4 + KBo 3.41 obv. 7 (OHlNS). But without
explicit er some instances of -an refer to spatial relationships that do not unambigu-
ously refer to superposition: (water is poured into a cup,) andan tuuiar iiyan [k]itta
na-a-[(a-a)]n (natan > naan) a r+\t
d
t.otn kitta inside lies tuuiar bound,
and it (all) lies onlat the hearth in the presence of the god Nergal KBo 17.40 iv 1112
(OHlMS) with OS dupl. KBo 17.15 rev.! 17 conrming naan; naan katta
DUG
oln.
xr-ya laweni and we pour it down into a oln.xr-vessel KBo 17.25 obv. 9 (OS);
utxt-la maan [tuel] kuit kardi nuza apt i My son, do whatever is uponlin
(-an) [your] heart! KUB 1.16 iii 6263 (OHlNS); KUB 1.16 ii 7071.
Physical or Conceptual Movement toward an Object
28.88. As common as this co-occurrence with er and notion of superposition may
have been for -an and in OS texts it is virtually the only usage of the particle it
is by no means the entire range of possible meanings. In post-OH, including in later
copies of Old Hittite texts, -an was used to express other relationships. First among
these is movement directed toward an object. With verbs of movement: (The Great
Seagod quarreled with the other gods,) na(n)[an nepia
d
tft-un] katta putet and
he brought him, [(namely) the Sungod of Heaven,] down (to his home in the sea, and
hid him) KUB 12.60 i 34 (OHlNS), restored after ibid. i 10;
d
tft-umaan ar [(
nep)]ii iyanni but the Sungod went up to heaven KUB 24.8 ii 1112 (OH?lNS), see
also KUB 33.120+ i 37 (Song of Kumarbi); nuan
d
Wiuriyandan katta [u-i pedai]
and [I take] (the goddess) Wiuriyanza down [to] the river KBo 15.25 obv. 5 (ritual);
Particles 28.88
00-HittiteGrammar.indb 375 6/5/08 4:20:10 PM
376
rorn-pamaan tnt-si annapili nman paizzi kti kaninti he did (lit., does) not
wish to go back to his city empty-handed, in hunger and in thirst KUB 33.121 ii 1516
(Kei story); numaan ka ot\++ kuit arnummeni nat itamaten Hear the case
which we are bringing tolbefore (-an) you KUB 4.1 ii 56.
28.89. With other verbs, expressing a kind of conceptual movement (temporal, audi-
tory, visual, attentive, etc.): (If someone sets re to anothers storage shed, he must feed
that ones cattle,) nuan par ameanda arnuzi and bring (i.e., sustain) them to
the following spring KBo 6.3 iv 6061 (Laws 100, OHlNS); mn +w+: itoima t
paanutteni [x x x x] x-aan t uiteni nu arkteni If you do not keep the word of
the king, you will not survive (lit., live) unto (-an) [ . . . . . . ]; you will perish KUB
1.16 iii 3637 (OHlNS); nunnaan anda miy nunnaan anda talliy be gentle
toward us, be pleasant toward us VBoT 24 iii 3839. With -an, +\+ followed by a place
designation, and the medio-passive form of nai-, the meaning is to turn toward. Whether
one uses the translation toward or against depends upon the qualifying adverb of
manner: auli in a favorable manner or i.xn-li in a hostile manner: [(nu)]an
+\+ xtn
URU
atti auli anda namma naiut turn again toward the land of atti in
favor KUB 9.31 ii 5960; see also KUB 15.32 i 5455; nuz( )an +\+ xtn
URU
atti
i.xn-li ror.nr. di (whoever) sets his eyes against the land of atti in enmity KBo
6.34 i 1718, 3637, ii 1213. With +\+ or anda and verbs of speech, -an focuses upon
the person addressed:
d
tu.sro
5
yaan anda tezzi he also speaks to the god Favorable
Day KUB 39.33 iii 1112, cited as 120lp in Otten 1958: 114 15; nammaan uddr
par +:++++ memai Next he speaks words in the same way (to her) KBo 2.9 iv
1516. Interestingly, both speaking to and listening to someone show the same syntax,
but the latter with par instead of anda: nuan
m
Kei par +\+ uxsipat ts[+i]
now Kei listened only (-pat) to his wife KUB 33.121 ii 6, see also line 8.
28.90. In some cases (mostly with anda) the implication of -an is more than direc-
tion (toward) and practically amounts to movement which reaches the goal and enters
into it (into). Old Hittite references (including one OS ex.): anda [(a-an parna nwi
paizzi )] (If they capture a free man at the beginning of his act of breaking into a house,)
that is, he has not yet gone inside KBo 6.2 iv 37 (Laws 93, OS) restored from NS cop-
ies but clearly showing the beginning of -an in the OS original; nuzan (if this is zan
instead of zan)
d
rx-a
UZU
X akuwae t ta rorn-pa d and so the Stormgod took
back into himself his heart and eyes KBo 3.7 iii 1819 (OHlNS). Post-OH references:
(You come up from the sea, O Sungod) nuan nepii ti[( ya)]i and you enter heaven
KUB 6.45 iii 1415 (NH); nammaan nXu.xr-ni anda l kuiki [( padd)]i let no
one dig into the city wall KUB 31.86 + 1203lu ii 22 (MHlNS).
28.91. Related to the idea of into or inside is that of things which rightfully pertain
to or belong to an object or person. The passage concerning the Stormgods eyes and
heart cited above (KBo 3.7 iii 1819) certainly shares this notion. Another example is a
rare instance of -an in a to be clause (for an OS instance with a dierent nuance of -an
28.89 Particles
00-HittiteGrammar.indb 376 6/5/08 4:20:10 PM
377
see 28.85, p. 374): [s+ xnm]uan 4 oln.xrsi knmawaza s+ 2 oln.xrsi
kuwat un my [calf] should have four feet (lit., my [calfs] four feet should be inlon
it); why then have I borne this one with only two feet? KUB 24.7 iii 2224 (NS).
28.92. A sexual idiom, similar to biblical Hebrew b' 'el to go in unto (a woman),
is: +\+ tut.t
19
maan tut.rn-a kuedani nwi pizzi a ewe to which a ram has not
yet gone (sexually) KUB 9.32 rev. 2122.
28.93. When physical or conceptual movement toward is hostile, we translate
against. Old Hittite references: ia []amaan tataekewan dir and they be-
gan to whisperlconspire against their lords KBo 3.1 i 22 (OHlNS); +\+ rsimaian
i +\+ utxt.xrsi idalu l takkianzi but let them not plan evil against his house
or his sons ibid. ii 5556; ukwaz( )an [(nun)] I became angry at (them) KUB
33.10 ii 6 (OHlMS), restored from KUB 33.9 ii 13, where -an is lacking; mnan
d
Telipinua kuedanikki nakkezi but (-ma) when Telipinu becomes dicult against
anyone VBoT 58 iv 89 (OHlNS). Post-OH references: mna llet nuan rorn-pa
tnt-ri idlu takkikewan dai When he became ambitious, he began to plot evil against
(his new) city KBo 32.14 ii 1920 (MHlMS); numuan t\+
URU
Paluia rorn-an
i?.xn?
URU
Peuru xi-ya tiyat behind Paluia, the hostile (city) Peuru advanced
against me for battle KBo 3.4 ii 23 (NH).
28.94. The particle -an also occurs in clauses containing verbs of contending, con-
testing, quarreling. and striving. Middle Hittite example: nuwaa[n] annenanni
[ ]annuwanzi t tarratta you are not able to judge lawsuits KUB 40.62 + KUB 13.9
i 78 (MHlNS).
28.95. The only common denominator of the contexts in which -an appears is that of
physical or conceptual proximity, usually but not always produced by movement toward
one person or thing by another, a very broad characterization that fails to distinguish -an
suciently from -(a)pa (28.100., pp. 378.), on the one hand, and -kan (28.48.,
pp. 366.) on the other. Nevertheless, such a meaning is quite compatible with the most
likely derivation of the particle, from PIE `som (together) with (Josephson 1972: 419
and others). The problem of the precise dierence in the nature of the proximity ex-
pressed by PIE `elopi, `kom, and `som extends far beyond the use of Hittite -(a)pa,
-kan, and -an. It is hard to see how the many uses of -an in cases with no reference to
verticality could have developed from a specic original value (up)on, over, as implied
by the alternative account of -an < `saran (Carruba 1964: 429). One may still wonder
whether the favoring of -an in contexts referring to vertical contact is due to a secondary
association of -an with ar and er based on their phonetic resemblance.
The Particle -an
28.96. -an, the rarest of the local particles, was rst identied by Otten and Souek
(1969: 81). Following the plural clitic pronoun -e it takes the form -n (KBo 17.1 i 20,
OS). That this change is morphophonemic is shown by the sequence nuean KBo 6.2
Particles 28.96
00-HittiteGrammar.indb 377 6/5/08 4:20:11 PM
378
iv 1011 (OS). It is attested scarcely a dozen times in a handful of OH compositions. Its
usage falls into two broad categories.
28.97. With anda and various verbs: tan anda 3-i itoi-u xtts.itoi-a
zriya allapaanzi And the king and queen spit into the cup three times StBoT 8 iv
3435 (OS); andaadan arakzi And it perishes therein KUB 17.10 iv 17 (OHlMS)
(see andaatkanl[a]ndaatan arakdu in the NS copies KUB 33.8 iii 14 and IBoT
3.141 iv 13);
d
Andaliyaan anda iyanni . . .
d
tftan utxt-a anda pait Andaliya
started to go in . . . the son of the Sungod(dess) went in KUB 12.63 rev. 19,22 (OHl
MS); rn.xr-nan kui anda ptai The one who brings in the troops . . . KBo 17.1
i 32 (OS) (also with anda peda- ibid. i 31, 33; ii 12; and iii 29), but see andakan
. . . ptumini ibid. i 26; takku otu-un kuiki kuaniezzi nuean
KU
uan nama
KU
tarua anda di If anyone rents an ox and then puts on it a leather . . . or a leather
. . . KBo 6.2 iv 1011 (Laws 78, OS; -an replaced by -kan in NS copy).
28.98. Elsewhere: [n]en kiarimi di And he puts them in their hand(s) KBo
17.1 i 20 (OS); arkanzimaan
d
antaepe andua ar [(rr)]a
GI
txtn.nr.
ya The antaepe-deities hold human heads and spears KBo 17.1 i 2223 (OS);
takkuwatan parnama kulka peiezzi But if he disposes of them in someones
house KBo 6.2 ii 35 (Laws 44b, OS); appnimaan [(ii-it l-na) pei]emi But
in(to) the open ame I [throw] the honey and fat KBo 17.5 ii 78 (OS); 9-tian xt-ti
in the ninth year KUB 33.120 i 12,18 (Song of Kumarbi); tan kardama al[(ikti)]
and you shall penetrate their hearts KUB 31.4 obv. 89 (OHlNS).
28.99. Since the notion of in(side) is clearly present even in the examples without
anda (the gurines of the deities may be assumed to be holding the objects in their
hands), there seems little reason to doubt that -an underscores this meaning and is ety-
mologically related to PIE `en in (and also to anda). See Oshiro 1990 [92].
The Particle -apa
28.100. The particle -apa shows the following writings: The rarest writing is -aba
in
LU
iti-i-ya-a-ba (for
LU
iti-i-ya(a')aba, i.e.,
LU
taraiya(a?)apa) KBo 17.43
i 9 (OS) with dupl. KBo 17.18 ii 10 (OS), i.xr
GI
ntn-a-ba KBo 20.32 ii 6, i
GI
ntn-a-a-ba KBo 20.32 ii 11 (compare i
GI
orunt-a-a-pa KBo 20.32 ii 16),
itoi-a-ba KBo 41.101 6. The writing -ap is found in u-a-ap (uap) KBo 3.60 iii
3 (OHlNS), compare a-na-ap (anap) ibid. ii 3, 5, 18, iii 9. Following certain vowels
the a elides, giving -pa (1.74, p. 32): e-pa KUB 36.27:2 (OH) and e-e-pa KUB 43.36
(OH) (i.e., epa), nu-u-e-pa < nueapa KUB 36.110 obv. 11 (OS), na-a-i-pa
KUB 33.21 iii 19, see KUB 33.31:2; and as late as the Appu myth KUB 24.8 i 43 (OH?l
NS); nu-u-i-pa KUB 24.8 i 15, 16 (OH?lNS); na-at-i-pa KUB 24.9 + JCS 24 37 iii 8,
na-an-i-pa KUB 35.148 iii 14. The most usual writing is -apa: na-pa (for n(u)apa)
15

15. The unique nu-pa KUB 35.148 iii 29 is probably to be emended to na!-pa.
28.97 Particles
00-HittiteGrammar.indb 378 6/5/08 4:20:11 PM
379
KBo 3.67 ii 5 and KBo 3.1 ii 20 (both Tel. pr., and passim in that text), KUB 43.60 i
11, na-ta-pa (natapa) KUB 43.36:5; an-da-ma-pa (for andam(u)apa, 1.72, p. 32)
KBo 3.7 i 13 (OHlNS), see Carruba 1964: 420 and in CHD sub -mu b 4; a-an-za-pa
KBo 12.18 i 6; nu-za-pa KBo 19.92:7; nu-a-pa (nuapa and them + -apa) KUB
1.16 ii 24 (OHlNS), nu-wa-ra-ta-pa (nuwaratapa) KBo 3.1 ii 49; the alleged nu-un-
na-p[a'] of KBo 12.63 ii 5 claimed by HW
2
sub -a2 is worthless because, among other
reasons, the context is broken and the -p[a'] is highly uncertain; one expects something
like nu-un-na-[a- . . . ] (< nunna).
28.101. So far as dating is concerned, -apa occurs often in OS and continues to
be written in NH copies of OH texts down to the end of the empire. It is not yet at-
tested, however, in MH historical texts (treaties and letters). The alleged example in
the ukkana Treaty (e.g., HW
2
sub -apa) does not exist (see Honer 1973b and CHD P
s.v.). In addition to assured OH texts, -apa does appear in a few compositions that may
be OH or MH (notably the Appu myth and the Ritual of Alli).
28.102. Most uses of -apa may be assigned to one of two broad categories, but there
are a few examples that do not seem to t either of them.
28.103. As noted by Carruba (1964: 42224 and especially 430), many instances of
-apa appear to share the notion of convergence: bringing one thing into physical contact
with another (transitive verbs) or coming into physical contact with something (intransi-
tive verbs). Examples: with an- to wipe (something) onto, besmear, daub:
LU
erala
rt[a] []eran arzi napa pri nkezzi The .-man stands, holds . (a substance),
and wipes his lips (with it) KBo 19.163 i 2223; with arnu- to bring: napa arnu-
wanzi KBo 21.95 i 15; with anda pp- to include: ammuggazaapa anda p lmu
genzuwaii Include also me (with your enemies); dont spare me KBo 3.7 iii 2930 (OHl
NS); with amank- to bind, intertwine: uiwatarm(u)apa anda ingani aminkan
inganam(u)apa anda uiwanniya aminkan Life for me is bound up with death;
death, however, for me is bound up with life too KUB 30.10 obv. 20 (OHlMS); with -za
andi- to t (something) to (something else), mate, match, combine: nu kuin xsxi-
an arzi uran xsxi-an arzi marnuwalan xsxi-an arzi anzaapa xsxi-i
i.xsxi-la andit And what road does (the human soul) travel (lit., hold)? It
travels the great road; it travels the invisible(?) road. The traveler has tted it to his
road KUB 43.60 i 2830; with (anda) ar(k)- to hold close (verb conjecturally re-
stored): [nue] uddrmet [ak]ikemi sro
5
-an[zamu rri] [taggan]iyatamuzaapa
and [a ark] numu tagga[niyata] taknaz pai I am making my words known [to her]:
Wash [me] well, [hold] me tightly to your breast, and protect me at your breast from the
earth KUB 1.16 iii 7172 (OHlNS); with anda iyannali- to go to: nai(a)pa anda
iyann[i] nan punukewan da[i] he went to him and began to question him KUB 24.8
+ KUB 36.60 i 4345 (Appu story); with anda lag- to lean againstlupon: [l ]abarna
itoi
URU
atti aearumme[t] tu nuzapa utniyanza manza ikimet anda

URU
attua lagan ard [u] labarna itoi-u inarawanza nue(a)pa utniyanza
Particles 28.103
00-HittiteGrammar.indb 379 6/5/08 4:20:12 PM
380
manza anda inarai Let the Labarna, the King of atti be our fortress. The entire
land should lean its(!, text their) back against attua. The Labarna, the king, is vigor-
ous; and the entire land shall be vigorous with him (i.e., by the contact with him; the
force of anda is carried over from the sentence with anda + lag-) KUB 36.110:912
(OS); with anda + mugi- to invoke: paruen[a kitta] naipa anda mugnza
[du] p. [lies here; may] he [be] invoked by it (i.e., be attracted to it) KUB 33.21 iii
1819; with andan +pai- to give in addition (to): andanapa ap [(niwan pi )] he
shall give the same amount in addition to it KUB 29.23 :14 = Laws 110 (OHlNS);
with dai- to place: kwa na
16
uttar tuppiaz au karwa ar
URU
attui makketa
nuwaratapa urorn.xr-i allai aannai dir Behold this tale of bloodshed from
the tablet! Formerly bloodshed in atti was rife, and the gods laid it on the royal family
(lit., Great Family) KBo 3.1 ii 4749 (OHlNS); with takk- to match: aitapa
xt-i dk[ki ] His mouth matches (his) mouth KUB 43.53 i 4 (OHlNS) (multiple
examples ibid. i 515); with tarupp- to gather, bring together, unite: n[(apa) ut(xt.
xrsi) r.x(si)] [(
LU.ME
gaenaea
LU.ME
a)]anna (var. + a) i [(rnr.
xrsi)] [(taruppante ee)]r And his sons, his brothers, his in-laws, his family, and
his troops were united KBo 3.1 i 13 (Telipinu pr.); with anda tiya- to come together,
convene: (The Storm god convoked all the gods:) andam(u)apa tiyatten Come (lit.,
step) together to me! KBo 3.7 i 1213 (OHlNS); with anda turiya- to unite, hitch to-
gether: kinunam(u)apa [urorn:]+ innarau[w]r i
d
ixx anda triya But now,
[O] my [god], unite with me vigor and good luck (lit., the Patron Deity) KUB 30.10
rev. 1920 (Kantuzili prayer), the plene writing of innarawar here is unique; with uda-
to bring: rx.iXi triya tu-a muwa tu-a xsxi-an pndu napa iyatarmit
udandu Let the bee(s) go a journey of three days (or) four days, and let them bring
(to me) my prosperity KUB 43.60 i 1012; in a nominal sentence: itoi-imaapa l
kuitki Let there be nothing for the king = Let the king have nothing to do with itlno
involvement KUB 11.1 iv 21, ed. THeth 11:5253.
28.104. In other cases -apa appears to have a terminative sense.
17
In most of these
it underscores that an action is brought to a conclusion (endterminativ). Compare
28.77 (p. 372) for this use with -kan. Examples: with appai- to nish: napa aear
pankua ppai the seated group and the entire congregation nish (worshiping) KUB
53.14 ii 1 (OHlMS); napa utxt.itoi appai and the prince nishes KUB 53.14 ii
2 (OHlMS); tapa ppai KBo 21.93 ii 8; with a- to open (up): [m]n lukkattama
tap[(a aanzi
KU
o.nXn-a)]n [(uia)nz]i When the morrow comes, they open up
and pull back the curtains (in the temple) ABoT 9 + KBo 17.74 + KBo 21.25 i 3132
(OS?); napa [ . . ] rorn-pa akei KBo 11.14 iii 2527 (OHlNS); mnapa

alentuwa
aanzi
KU
o.nXnata uiyanzi when they open the residential palace and pull back
the curtains KUB 25.16 i 13 (OHlNS); with iya- to make into (with double obj.)
16. See 4.101 (p. 126).
17. For the notion terminative, including both endterminativ and anfangsterminativ see among
others Garca Ramon 2002: 111.
28.104 Particles
00-HittiteGrammar.indb 380 6/5/08 4:20:12 PM
381
nuwarat ta karandu nat!apa
GI
ntn.xr iyandu Let them cut (a tree) down
and make it into tables KBo 22.6 iv 1718 (OHlNS); with pai- to go: nu tuliyan
alziten mnapa uttaret paizzi nu so.ut-naz arnikdu Convene the assembly. If
his plan (lit., word) goes to its conclusion, then let him pay with his head KBo 3.1 ii
5152 (OHlNS); with an- to seek for, avenge (bloodldeathlmurder):
18

m
Zidantaa
itoi-ut napa urorn.xr
m
Pieniya iar aner Zidanta became king, and the
gods sought (i.e., avenged) the bloodlmurder of Pieni KBo 3.1 i 6668 (OHlNS),

m
Ammunaa itoi-ut napa urorn.xr-i attaa
m
Zidanta aret aner
KBo 3.1 i 6971 (OHlNS); with medio-passive tarupp- to be completed: [(m)]napa
NINDA
[(taparpau arrumanzi taruptari)] when the distribution of taparpau-breads is
completed KUB 41.42 iii 5, restored from dupl. KUB 10.89 i 1617, which lacks -apa;
with uwa- to come or rorn-pa uwa- to return: mnaapa laazma rorn-pa uezzi
Whenever he returns (to home base) from an expedition KBo 3.1+ i 18 (OHlNS); [nu
s ]+
f
Zi. alwanzataret idlu uddret +:++++ [ar]kdu natapa rorn-pa l uezzi
Let the sorcery (and) evil word of the woman Ziplantawiya likewise get lost, and let it
not come back here KBo 15.10 + KBo 20.42 ii 1516 (MHlMS); mnapa itoi-u

URU
Lawazzantiya uwanun When I, the king, came to Lawazzantiya KBo 3.1+ ii 2021
(OHlNS); with zinne- to destroy, wipe out: nuapa uezzi zin[nai] she will proceed
to wipe them out KUB 1.16 ii 24 (OHlNS). The use with d- to eat up and karap-
could either be the terminative use (eat up) or an extension of the convergence aspect:
complete assimilation or envelopment of the eaten object. Examples: anap azzikanzi
and they eat him KBo 3.60 ii 3 (OHlNS); anap atnzi KBo 3.60 ii 5 (OHlNS);
anap zta KBo 3.60 ii 18 (OHlNS); with karap- to devour: w [k]a ntnt
14
.
nr. sro
5
-anta daganzipa [u]idr pari napa alkin karapanzi Rains will occur,
19

crops will thrive, wildlife of the earth will appear, and they
20
will devour the grain
KUB 8.1 iii 810 (OH?lNS).
28.105. The motivation for the use of -apa with the verb arai- to arise remains
unclear: [ar]aia(?)pa
21

m
App[u] naza parnaa [iy]anni Appu arose and set
out for his house KUB 24.8 + KUB 36.60 i 2426 (OHlNS);
22
araia(?)pa
m
Appu
GI
-az nuza arkin sri
4
-an d na
d
tft-i kattan iyanni [] Appu arose from his
bed, took a white lamb, and set out for the Sungod KUB 24.8 + KUB 36.60 i 3840
(OHlNS); zipat nattaaa[ pa a]ri (s)he keeps on sitting; (s)he does not get up
KBo 19.163 ii 3334 (OHlNS) (reading na-at-ta-a![-pa a-r]a-a-i); napa r.urorn
ara[i . . . ] KBo 21.90 obv. 1112 (OHlMS); [m]n lukkattama napa r.urorn-a
18. Note, however, that this construction also exists with -ata and without sentence particle (CHD sub
an- mng. 5).
19. Although w is common-gender plural, its verb ka(ri) is singular here. See 15.17 (p. 241).
20. Although uitr is formally neuter singular, semantically it is a collective, indicating many animals;
hence, the shift to the plural verb here.
21. For the ambiguity of clitic boundaries here see 29.47 (p. 402).
22. The very fact that -(a) pa apears to have gone out of use after OH points to an OH archetype for this
text.
Particles 28.105
00-HittiteGrammar.indb 381 6/5/08 4:20:13 PM
382
ari when morning comes, the r.urorn priestess arises (and . . . ) KBo 21.95 i 17
(OHlNS).
28.106. The use of -apa with waqqar- to be lacking does not seem to t into any of
our categories: nuipa t kuitki waqqari nuipa 1-an uttar waqqari utxt.rfi
utxt.xtts-i t.oi Nothing was lacking to him; only one thing was lacking to
him: he had neither son nor daughter KUB 24.8 + KUB 36.60 i 1517 (OH?lNS).
28.107. The grouping of the examples just given obviously is quite tentative. The
notion of physical contact of one object with another seems to be shared with -an (see
28.95, p. 377), and most of the verbs cited in 28.104 (p. 380) refer to actions that are
inherently telic. Verbs such as pai- to go and uwa- to come occur dozens of times
without -apa. Nevertheless, the frequent co-occurrence with anda specically in its de-
rived sense (up) to is noteworthy (contrast the usage with -an in 28.97, p. 378), and
in the example with pai- it appears to be -apa that adds the nuance to its conclusion.
Development of the terminative sense from that of physical contact would be trivial.
See Latin op-prim to press against but also to overwhelm, destroy. As suggested by
Carruba (1964: 429), an etymological connection of Hittite -apa with Latin op-lob- and
related forms is likely (and with Hittite ppa).
The Particle -ata
28.108. Like -apa (-pa) and -an (-n), the particle -ata elides its a-vowel after an
e- or i-vowel, thus: ta-a-i-i-ta KUB 31.1 + KBo 3.16 ii 14 (OHlNS); nu-u-i-i-ta
KUB 17.10 iv 2 (OHlMS); peran aramaita KUB 23.11 iii 17, nu-u-e-e-ta
KUB 33.5 iii 12 (OHlMS).
28.109. But (as with -apa KBo 3.7 i 13 in 28.100, p. 378) an immediately preced-
ing u-vowel elides, leaving an a vocalization (see 1.72, p. 32): nu-um-ma-a-ta KUB
31.130 rev. 7 (OHlMS) and nu-ma-a-ta KUB 36.75 iii 12 (OHlMS) and KBo 14.74:3
(from nu + -mu + -ata).
28.110. Most discussions of -ata have emphasized its use to mark separation or
divergence, often with the preverbs ara and par (see, e.g., Carruba 1964: 40812).
Examples of this kind are in fact frequent: nata par paiwani And we go outlleave
KBo 17.3 i 15 (OS); teta pnzi They leave KBo 17.9 iv 9 (OS); rananata par
ptanzi They remove the eagle KBo 17.1 ii 3738 (OS); tu() ta ar tumni We
pick them up KBo 17.1 iv 25; anata ara puter They led him o KBo 3.34
ii 67 (OHlNS); nata zo-an
U
[
ZU
or]ft-an ara kurzi He cuts o the right ear
KUB 27.67 ii 4849. See 28.65 (p. 369) for a similar use of -kan. As already noted by
Goetze (1933b: 128),
23
it is the presence of -ata (or another particle; see CHD par)
that gives the sense to go out, leave to par + pai-, which without the particle means
to go forward or to go furtherlon.
23. See also CHD P par mngs. 1 and 3.
28.106 Particles
00-HittiteGrammar.indb 382 6/5/08 4:20:15 PM
383
28.111. From the basic idea of separation one may easily derive a partitive sense,
selection from a group, explicit or implicit (see Carruba 1964: 41213, who aptly com-
pares Italian ne): nata
URU
attuapat tnt-ria 1-a ta Out of (all other cities)
the city of attua alone remained KBo 10.2 i 26 (OHlNS); perhaps also nepizata
d
rx-unni u ta To the Stormgod from heaven he was dearest (of all) KBo 3.22:2
(OS). Similar is the use of -ata to mark comparison: iki etata ikii oi-li
[ paltan]ai!ata paltan alli His back is bigger than (his) back; his shoulder is
bigger than (his) shoulder KUB 43.53 i 2324 (OHlNS).
28.112. In many cases, however, -ata occurs in contexts referring to arrival at a
goal, including entry (with anda): [m]nata oisi\[i itoi-i a]nditta When
their leader draws even with the king KBo 20.14 + KBo 25.33 i 9 (OS) (ibid. i 13
without particle!); nata man arkantin +\+
LU.ME
+isioi:t+ andnzi When they
bring the petitioner in line with the guards IBoT 1.36 iii 20 (translation with Gterbock
and van den Hout 1991: 25 and CHD 58); nata X xtn.xtn.xr anda
d
tft-u tiyat
In the midst of the lands the Sungod stood KBo 10.2 ii 5253 (either OHlNS or NH);
nu azzikkeddu akkukedd [u mna u] nata ar uikettaru Let him continue to
eat and drink. [So long as he is on good behavior,] let him continue to come up (to the
palace) KUB 1.16 ii 3334 (royal edict, OHlNS).
28.113. With verbs of movement it seems to be associated with the idea of transition
or crossing: [lma] arkaliyatumari l ku[i ]ki [ku]rur nata uttar [l k]uiki arratta
Do[nt] elevate yourselves (over your brothers)! Let there be no hostility! And let [no]
one transgress the word (of the king)! KUB 1.16 ii 4950; nata nepia x-u zikpat
[aanuw]anza
d
tft-u arraketta you, [establish]ed Sungod, pass through the gate
of heaven KUB 36.75 + Bo 4696 i 78 (OHlMS); (The Kaskaeans) come and take
the gifts, then they swear (oaths), but when they arrive back (home), nata lingu
arranzi They transgress the oaths KUB 17.21 iv 1617 (prayer , MHlMS); nata
kui k \ts urorn-tt+ arriezzi Whoever transgresses these oaths KBo 6.34 ii 46
49 (Soldiers oath, MHlNS), ed. Oettinger 1976a: 1011; lingainnata t kuanka
arraat I never transgressed an oath KUB 30.10 obv. 1214 (OHlMS); mnata
ka lingu arradduma If you transgress these oaths KBo 8.35 ii 16 (treaty with
the Kaka, MHlMS); nata uttar [l k]uiki arratta Let no one transgress the words
KUB 1.16 ii 4950 (OHlNS); laiyauwaza [uttar] t imma ekteni erata kuitki
arran Do you actually not know [the matter] of campaigning: that something concern-
ing (it) has been transgressed? KBo 16.25 i 3334 (MHlMS); nata u-an zaitten You
crossed the river KUB 31.101:67, but see with -kan in ibid. 1011.
28.114. One possibility for uniting the apparently diverse meanings just cited is to
suppose that -ata originally marked passage from one spatial domain across a bound-
ary into another domain (see Josephson 1995: 171, comparing the Russian verbal prex
za-). The specic uses with reference to separation, transition, and arrival would merely
reect varying focus on one aspect of the movement according to context (and choice
Particles 28.114
00-HittiteGrammar.indb 383 6/5/08 4:20:15 PM
384
of the speaker). It is also easy to see why overlap with -kan and -an might occur when
-ata was used with reference to arrival.
24
Luwian -ta shows a similar range of uses: see
especially [(w)]ratta u-ti [nan]amman . . . [w]ratta zl [a u-i] anda [(n)]wa
iti The water is led from the river . . . the water does not henceforth go into the river
KUB 35.54 iii 1720. This correspondence makes attractive the suggestion of Josephson
(1972: 419) that Hittite -(a)ta is etymologically derived by a resegmentation from `-te
seen in Luwian and Palaic -ta, Greek when etc. (see also Josephson 1997: 5052).
For another very dierent view see Carruba 1964: 42728.
The Particle -pat
28.115. Hittite -pat is an enclitic particle of specication, limitation, and identity,
attested since Old Hittite (Hart 1971; Honer 1973b; and CHD P with full discussion
and anterior literature).
28.116. Normally it was written with the nr sign (- HZL sign no. 13, pp. 9697),
which in Boazky texts has several syllabic values. The problem of the correct tran-
scription of HZL sign no. 13 is particularly important for this particle, since the particle
is almost always written with the same sign. For evidence establishing the reading lpatl
see Honer 1973b and CHD P s.v.
28.117. Its position. Since it is not a sentence particle, -pat does not occur within
the chain of such sentence particles as -wa(r), -a, -mu, -za, and -kan (see Friedrich,
HE 288; Hart 1971: 102; Honer 1973b: 104). If a chain of sentence particles attaches
itself to any word which also carries -pat, -pat will precede the entire chain: e.g., in OH
ammelpatwaza KBo 6.3 iii 71 (Laws 74, OHlNS).
28.118. When -pat is construed with a word and its modiers (noun + attributive
adjective, noun + genitive modier, verb + preverb, verb + adverb, verb + negative),
it will usually attach itself to that modier which serves to particularize or dene the
construction to the highest degree (Honer 1973b: 105). Hart (1971: 102) describes the
distribution dierently: -pat is regularly attached to the rst word in nominal groups.
Hart points out (1971: 103) that a consequence of this xed position . . . is that posi-
tion cannot be used to decide whether -pat belongs functionally with the adjective etc.
or to the noun.
28.119. With pairs of nouns in distributive expressions it occurs on the second (i.e.,
last) noun, suggesting the phrase is felt to be a unit: xsxi-i xsxi-ipat KBo 3.5
ii 1315 (NH); rft-mi rft-mipat KUB 22.7 obv.? 3; tu-at tu-atpat KUB 1.13+ iii 7.
See 19.10 (p. 291).
28.120. As established by Hart (1971), Honer (1973b), and others, the basic func-
tion of -pat is to call attention to the marked items identity with a previously stated or
24. On its alternation with -kan see already Gterbock 1964.
28.115 Particles
00-HittiteGrammar.indb 384 6/5/08 4:20:16 PM
385
implied entity. But this meaning is realized in a number of ways. In the following dis-
cussion we have tried to classify the dierent uses according to inner Hittite criteria and
then present the various alternative translations suitable for each. We have made every
eort to avoid the trap of imposing an outside standard conditioned by our own na-
tive language. Nevertheless, some usages that we have kept apart would probably have
appeared indistinguishable to native speakers of Hittite, and they may have perceived
further distinctions that we overlook. In many instances one simply cannot decide be-
tween competing interpretations. Hart (1971: 11621) discusses many examples that are
ambiguous. For a listing of possible uses see CHD P s.v.
Anaphoric
28.121. The rst major use of -pat may be termed anaphoric. It is used to indicate
that a person, object, circumstance, or action is in some sense identical to one mentioned
previously in the discourse (Sommer 1921). The fact that various translations are used to
express this function (the (very) same, the aforementioned, likewise, continue to . . .)
should not obscure the fact that the basic meaning is the same. In some cases, however,
the notion of identity seems to have been relaxed to include an action that is not identi-
cal to a preceding one but a natural continuation of extension of it. These examples are
best expressed in English by also, furthermore or the like. Finally, since one reason to
stress that someone continues to do something is that the action is unexpected, -pat can
acquire a meaning nevertheless, anyhow.
28.122. When the particle is attached to a noun or attributive adjective, it is best
translated in English as the (very) same, the aforementioned. When it is attached ei-
ther to a verb or to the predicate noun or adjective, one may wish rather to translate it
likewise. The fundamental signicance of the particle, however, remains the same in
these cases.
28.123. With nouns we may cite the following: 1 xtts-:i+
f
Kattittai-
m
Tatilni
s[i+si] 1 utxt.xtts rsi
f
Kattittaipat s[i+si] one woman, her name Kat
tittai of Tatileni. One daughter of her brother, her name Kattittai likewise Otten, St-
BoT 1:20; When a slave has stolen, and his lordlowner says: I will make the compen-
sation in his behalf, he shall make the compensation [takku m]immaima nu ln-anpat
[(uwezzi)] But if he (the owner) refuses, he forfeits the aforementioned slave KBo 6.3
iv 47 = Laws 95 (OS), restored from KBo 6.2 iv 48 (OS).
28.124. When a verbal phrase is repeated, and the speakerlwriter wishes to call at-
tention to the repetition, he may ax -pat to the second occurrence to the verb. And,
although this is simply the verbal equivalent of the usage noted above (28.122) with
substantives, one cannot translate into English smoothly as the aforementioned but
must use likewise, or continues to . . ., or . . . s as before. This is the usage discussed
by Hart (1971: 13536) and Honer (1973b: 11112). For example, aruwizzi . . . ta
namma aruwizzi nakan kuitman ilana ar ari aruwikezzipat He bows . . . he
Particles 28.124
00-HittiteGrammar.indb 385 6/5/08 4:20:17 PM
386
bows again; and all the while he is mounting the stairs he keeps on bowing KUB 20.46
iii 1116.
28.125. But one can also use the English word also to translate Hittite predicates
marked with -pat, where the second action is no repetition of the rst but a new and
additional action. While in the examples adduced above the particle could be translated
by German ebenfalls, in those which follow one would have to employ German auch
noch; And, when my father had become a god, my brother, Arnuwanda, seated him-
self on the throne of his father rorn-anmaa irmaliyattatpat but afterwards he also
[auch noch] became (deathly) ill KBo 3.4 i 56, ed. HE I 293c; If from the palace
they give to (some)one silver, gold, garments, bronze utensils as a gift for him, let it be
named: The king gave it to him, and however much is its weight nat iyanpat edu
let that also [auch noch] be recorded [lit., made] KUB 13.4 ii 35.
28.126. The particle -pat can mean nevertheless in the sequence although . . . ,
nevertheless . . . (contrary to expectation; Tenner 1926: 101; Pedersen 1935; Honer
1973b: 11213); on the semantic development from ebenfalls to trotzdem in Tenner
and Pedersens thinking see Hart 1971: 107; [nu]za mn irmalanzaa ta
d
tft-
stmata [(+\+)] +s+u +utx+ tittanununpat although you were also ill, nevertheless
I, My Majesty, installed you [in] the place of your father KBo 5.9 i 1617 with dupl.
KBo 16.19: 23 (Mur. II), ed. SV 1:1011.
Particularizing
28.127. The second major use of -pat is to restrict a particular exemplar or represen-
tative of a class of things. One may call this emphasizing, and in some cases -pat may
be be translated with the English emphatic reexive (we ourselves), but the implied
restrictive contrast with other members of a conceived class is always present (we and
no one else). -pat may also be used to underscore an overt contrast between actions:
natta X . . . Y-pat not X . . . (but rather) Y. Finally, the stricture X, and only X can
lead to a meaning even X (compare English the very thought of = even the thought
of, the mere thought of).
28.128. When the particle is attached to proper names, it can occasionally be trans-
lated himself, herself, itself, Latin ipse, German selbst:
URU
attuipat in attua
itself (= even in attua) KBo 6.2 i 39 (Laws 19b);
m
Pittaggatallipat Pittaggatalli
himself KBo 5.8 iii 15, ed. AM 15657, see also line 31).
28.129. Sometimes also with personal pronouns: numu kma umepat kuit
atrtten concerning what you yourselves have just written to me HKM 17:1315
(MHlMS), ed. Alp 1991a: 14243; apilapat KUB 14.1 rev. 44, 48 (MH), KUB 21.37
i 38 (Hatt. III), KBo 11.14 iv 21 (MHlNS), see Hart 1971: 11314 for discussion.
28.130. With the possessive pronouns -pat corresponds to English (his, her, its)
own (i.e., and no one elses), German eigen (HE 293b; Hart 1971: 12223; Honer
28.125 Particles
00-HittiteGrammar.indb 386 6/5/08 4:20:18 PM
387
1973b: 115); ammelpatwaza otu-un dai I will take my own ox KBo 6.3 iii 71
(Laws 74); ap [(lpa)]t annaa katta with his own mother KBo 6.26 iii 26 (Laws
189); so.ut-x+pat your own head KBo 5.3 ii 19 (ukk.).
Restrictive
28.131. Also exceedingly common are passages in which the particle imparts a re-
strictive or exclusive meaning. In most cases this thought is best translated into English
by means of the word only. This usage is particularly frequent in legal texts (laws,
instructions, treaties), in which the precise limits of obligation and liability must be
drawn.
28.132. In the treaty between Zidanza and Pilliya of Kizzuwatna the allotment of
cities between the two powers is made quite precise by means of -pat: ne s+
d
tft-st
pt aantu let these (cities) belong exclusively to His Majesty and ne
m
Pilliyapat
a[antu] let these b[elong] exclusively to Pilliya! KUB 36.108 obv. 45.
28.133. The precise limits of Madduwattas feudal holding are expressed with the
words: I have given to you the mountain land of Zippala nuwaza apnpat ei oc-
cupy only it! KUB 14.1 obv. 19.
28.134. The limits of liability are often dened with it: nu so.ut-azpat arnikdu
Let him pay with his (own) person only, (but let no one harm his house or his children)!
KBo 3.1 ii 5556; If someone steals the slave of a Luwian from the land of Luwiya
and transports him to the land of atti, and his owner recognizes him, nuzza lnsipat
di arnikzil t.oi he shall take only his slave, there will be no compensation KBo
6.2 i 47 (Laws 21).
28.135. In the Telipinu Proclamation the restrictions as to which persons qualify for
succession to the throne make good use of -pat: itoi-uan antezziyapat utxt.
itoi utxt-ui kikk[(i)]taru let only a rst-rank prince, a son, become the king!
KBo 3.1 ii 36 (royal edict, OHlNS).
28.136. Exceedingly common in the oracle inquiries is the use of -pat to eliminate
all possible causes for divine wrath other than the one under consideration (HE 293d):
nr-an kpat xr.xr nu xtr niuut sro
6
-ru, if this and only this ditto (i.e., is the
cause of the gods wrath against us), then let the xtsr niuut omen-taking be favor-
able! AT 454 i 16, ii 12, 24; for conrmation of the sense only see also nammama
otn-i it kui [tki] And there is noth[ing] else in addition AT 454 i 27.
28.137. Occasionally, restrictive -pat modies an entire clause: ammel kpat 1-a
dammea kyaan 1-an dammeanunun ts:i r.oi-tt+patkan kuit katta uiyanun
This was my only punishment (for her); in this way only I punished her: just the fact
that I expelled her from the palace KBo 4.8 ii 1214 (Mur. II); na kuwapipat
s
5
-ezi Only when it becomes red KBo 21.20 i 8 (NH).
Particles 28.137
00-HittiteGrammar.indb 387 6/5/08 4:20:19 PM
388
Contrastive
28.138. The particle -pat may also highlight a positive statement contrasted with
a negative statement of identical import. First the negative, then the positive: t\+ xtn
URU
atti t iezzi akipa-a[t ]an (he who does such a deed) will not go on living
in the land of atti: he will die there (an) KBo 5.3+ iii 31 (ukk.), ed. SV 2:12425
with note Zu S. 151 on p. 170; tmamu
GI
ftxti i.xn kuwapikki er wanut
d
ts:+umuzakan o::+ mandazapat daket At no time did she let the weapon
of an enemy be waved (threateningly) over me. In all this Is:+u my lady took me to
herself att. i 4143.
28.139. First the positive, then the negative: nanza akuaara[(npat d )]i
LU
.
ztan natta pzi (If the owner of an animal which has strayed to another mans pen
nds it there,) he shall take it as his rightful property, but he shall not seize him (i.e.,
the owner of the pen) as a thief KBo 6.2 + KBo 19.1 iii 4950 (Laws 66, OS); [nuza

d
tftst] tukpat
m
Alakandun aggai apnmaza [it agga] i I, My Majesty rec-
ognize you alone, Alakandu; I do not recognize him KUB 21.5 ii 910 (Muw. II).
28.140. Even modifying a single word within a clause: nuzakan irmalapat s+
ororn-tt+ andandatar er ukenun Even as an ill person (although I was ill,) I kept
seeing the divine power of the deity over me att. i 4445;
LU.ME
appinantepat
t aiwantekantari do not even the rich become poor? KBo 4.14 ii 5253; numu
kappinpat otxt-an
d
ts:+u
URU
am[u]a +\+ +ui-:+ wkta even when I was but a
small child,
d
ts:+u of Samua requested me from my father KBo 6.29 i 78; nukan
URU
Zippanan or
6
-azpat ar pun Even at night (i.e., one expects military operations
during the day, but at night they are unusual and unexpected) I went up to Zippana
KBo 10.2 ii 4950 (OHlNS).
28.138 Particles
00-HittiteGrammar.indb 388 6/10/08 7:53:07 AM
389
Chapter 29
CONJUNCTIONS
29.1. Clause linkage is marked either by conjunctions or by simple juxtaposition
(what is traditionally called asyndeton). There are ve clause-linking conjunctions in
Hittite: n(u), u (-), t(a), -al-ya, and -al-ma. Two of these (-al-ya and -al-ma), however,
can serve other purposes in the sentence: contrasting a single word in one clause with
one in a corresponding position in an adjacent clause (-al-ma) and linking members
of a series of words (X, Y, and Z) in a single clause (-al-ya). These two therefore do
not connect clauses in all their occurrences. The conjunction u (-) occurs only with
attached enclitics and never stands alone,
1
while free-standing nu and ta are common.
The possible signicance of this dierence is not yet clear.
29.2. The Akkadogram i (written with the sign i) can also serve graphically to
represent nu or -al-ya, but only when it immediately precedes a logogram, including
logographically written proper names
2
and numerals. When i connects nouns in a se-
ries, it stands for -al-ya on the following word; rarely, a syllabically written noun (or
even another logogram) immediately following it can redundantly indicate and by an
explicit -al-ya:
LU.ME
xt
7
i
LU.ME
karualimea KBo 6.2 iii 14 (OS),
3
ts:i xtr.
nr. i ts:i st.xrya KBo 4.4 ii 50 (NH). Akkadian prepositions are normally repeated
after i in coordinated logograms, whether nor not the coordination indicates multiple
personslobjects or apposition (on the latter see Sommer 1932: 130): +\+ . . . i +\+,
4
s+
. . . i s+, ts:i . . . i ts:i, t::t . . . i t::t, ++n+u . . . i ++n+u, r+\t . . . i r+\t, +oi . . .
i +oi.
5
But i+++ is not repeated: i+++ PN
1
i PN
2
.
6
1. A possible exception, u-u in the broken context of KBo 12.14 rev. 4 (OHlNS), was suggested to
Weitenberg (1992: 338 n. 9) by Houwink ten Cate.
2. Rarely even a name with its case ending:
m
Zinwaeli i
m
Lelli [] meggae KBo 3.1+ ii 2425
(OHlNS).
3. Also nu
URU
Tapigga
URU
Anziliya
URU
ariya i +\+
URU
aninqawaya r-++ i zzya [k]ue anniyan ta
HKM 54:8-13 (MHlMS), orx-anmamu-kn
m.d
Sin-
d
t-a utxt
m
Zid s+
d
ts:+u o:+ i s+ r:+ya
ulan auta att. ii 7475 (NH) and i?
URU
Tiyailtaa +oi
A.A
.oXnsi [+\+]
d
rxtn ipandaza
KUB 19.37 ii 23 (AM 168).
4. E.g., HKM 23 obv. 12 and 25:13 (MHlMS). But see Hagenbuchner-Dresel 1999: 54 for diering
conventions at attua and Tapikka (Maat).
5. E.g., HKM 33:3536 (MHlMS); but not in HKM 82:3.
6. E.g., HKM 48:23 (MHlMS).
00-HittiteGrammar.indb 389 6/5/08 4:20:20 PM
390
Clause-Initial Clause-Linking Conjunctions
Distribution
29.3. Old Hittite uses all three of these clause-linking conjunctions. The conjunc-
tions *u (-) and ta (t-) are in complementary distribution, the former with preterites
and the latter with present-futures (Weitenberg 1992: 327), as well as with imperatives
(see 15.11, p. 239, last example)
7
and the analytic present perfect ar (k)-construction
(cf. KBo 4.9 vi 67 [OHlNS]). As per Weitenberg, the preponderance of ta in ritual texts
and its rarity in historical texts is merely a function of the typical use of tenses in these
genres, and likewise its extensive use in the Laws together with the total absence of u
there. In post-OH copies of OH texts both u and ta are quite frequently preserved by
scribes, and there is no noticeable tendency for ta to be preserved more frequently than
u. Claims that u was eliminated from active use earlier than ta (Weitenberg 1992: 309;
Rieken 2000a: 41112) must therefore be regarded as unproven.
29.4. Weitenberg (1992: 325) stresses that nu, u, and ta always serve to link clauses
and thus never occur in discourse-initial position. This is strictly true for u and ta (see
29.48, p. 402). But nu is used discourse-initial when a new question follows logi-
cally upon an earlier answer: So . . . ? or If that is the case, then . . . ? See 27.8
(p. 351).
29.5. Establishing dierences in the usage of nu, u, and ta has proven dicult.
The complementary distribution of u and ta according to tense suggests that they are
functional variants, and in many respects their use is parallel. Many of their uses indeed
are also shared with nu. For attempts to discern patterns in the use of u and ta see
Weitenberg 1992 and for a very dierent account of ta see Rieken 1999b. We content
ourselves here with illustrating use of the conjunctions with dierent clause types, with
special attention to cases where they seem to overlap (e.g., 29.15, p. 393).
nu
29.6. nu connects independent clauses from OS onward. In the following exam-
ples the actions are denitely sequential (at times the second or nal clause express-
ing a result, so that . . ., therefore): kalulupi(t)mitata ig[(ara)]nta di [n]en
kiarimi di nat [(a pa)]r paiwani He takes the things fastened to their ngers
and puts them in their hand(s), and we leave KBo 17.1 i 1920 (OS); 100 gipear
.X kar (a)iyezzi nanza di he shall cut o 100 gipear of eld and take it for
himself KBo 6.2 i 8 (Laws 6, OS); kinuna itoi-u s+ r.oi-tt+ peiet nuza
nikanzapat 3 o xtnnnn d [i] But now the king has waived the palaces share,
7. There is only one exception to this distribution in OS: raan ou-an peiyami u i-a natta
auzi I throw a cloth over them, and no man will see them KBo 17.3 iv 1819. Even when one includes
copies, the complementary distribution remains clear: of slightly more than 100 examples of (u) where the
context can be determined, only ve occur with present-futures. Examples of t (a) with preterites are also
rare even in post-OH copies, less than a dozen out of several hundred.
29.3 Conjunctions
00-HittiteGrammar.indb 390 6/5/08 4:20:22 PM
391
so that only the injured party shall take 3 shekels KBo 6.2 i 15 (Laws 9, OS); [takku
i-a]n nama xtts-an itt++ walzi kui [k]i na aki [If ] someone strikes a free
[man] or woman, so that slhe dies KBo 6.3 i 6 (Laws 3, OHlNS); kawa kya kya
uttar iyami nuwamuan ziqqa arput I am about to do such-and-such, so you
too join me! KBo 3.7 i 2123 (OHlNS); nukan kma r.xtn.n.nr. par neun
nuza r+\t i.xn mekki paaanuanza I have just dispatched chariotry, so be
much on the alert toward the enemy HKM 1:813 (MHlMS); zikmawaza utxt-a
nuwa it kuitki akti You are a mere child, and so know nothing at all KUB 19.29
iv 16 (NH).
29.7. nu connects two subordinate clauses: takku ln-a uwi na +\+ xtn Luwiya
paizzi . . . If a male slave runs away and goes to the land of Luwiya . . . KBo 6.2 i
51 (Laws 23, OS); kui ammel ppan itoi-u kar [i] nu
URU
attuan ppa a [i]
. . . Whoever becomes king after me and resettles attua . . . KBo 3.22:4950 (OS);
s+
m
Piinakkimu kui [t] uttar atre
m
Piinakkiza maan
URU
Liipran ki [tt]ari
nuwaza karu 30 r-:i+ aean []arzi Concerning the message you sent me about
Piinakki: how Piinakki is settling the city Liipra, and (how) he has already settled
(there) 30 households HKM 10:36 (MHlMS); [mn] uma t p arzi nankan
ullannaza [ku]iki kuenzi But if he (scil., the merchant) does not have goods in his
possession, and (the oender) kills him out of wantonness KBo 6.4 i 67 (Laws III,
NH);
d
tft-stma maan iyaat nu maan +\+
URU
allapa [aru]n When I, My
Majesty, marched, and when I [reac]hed allapa KUB 14.15 ii 7 (NH, Mur. II). See
29.18 (p. 394), where ta, -a-l-ya-, and i also ll this role.
29.8. nu connects a subordinate clause to a following independent clause. This usage
occurs already in OS with temporal clauses (see 30.36, p. 416): mn
URU
Tamar[mara]
arer nu tarikanzi When they arrived in T., they said (historical present; see 22.6,
p. 307) KBo 22.2 obv. 8 (OS); m[]n lukkattama nu
LU
.zt gga paiwani But when
it dawns, the physician and I go KBo 17.1 iv 7 (OS); mn t\ [+ t]u.2.xx [ ppa
paiwa]ni napa aueni When we go back on the second day, we open up KBo 8.74+
ii 9 (OS); mn
MUEN
ranan u [(uwandan appanzi)] nan udanzi When they catch a
live eagle, they bring it StBoT 8 ii 1920 (OS); also in OHlNS: nu mn (7) mi ezi nu
rzr
4
purulliya (8) iyanzi And when prosperity and abundance come, they celebrate
the festival of purulli KBo 3.7 i 68 (OHlNS). But in OH an independent clause fol-
lowing a subordinate temporal clause can also be introduced by u (29.12, p. 392), by
ta ([m]n lukkattama tapa aanzi KBo 17.74 + KBo 34.10 i 31 [OHlMS], see also
KBo 17.4 ii 1011 [OS]), or by asyndeton (marked in this and the immediately fol-
lowing paragraphs): mn
MUEN
ranan rnr.xr-anna 3si wanumi rananata
par ptanzi When (i.e., after) I wave the eagle and the troops three times, they carry
the eagle out KBo 17.1 + KBo 25.3 ii 3738 (OS), see also KBo 3.22 7879 (OS).
29.9. nu also connects a preposed relative clause to a main clause, already from OS
onward:
d
tft-az utn [kuit k]uitpat arai nu mandup[at ]u[llanu]n Whatever
lands rebelled. . . , I fought all of them! KBo 3.22:1112 (OS);

nu kuit [(it)]oi-u
Conjunctions 29.9
00-HittiteGrammar.indb 391 6/5/08 4:20:23 PM
392
tezzi nu apt iyami And what the kings says, that I do KBo 17.4 ii 1213 (OS); kuian
ppama uwatezzi nanza appat di whoever brings him back shall keep him for
himself Laws 23 (OS). This usage is also attested for u, ta, and asyndeton: for u
see example in 29.14 (p. 393), for ta see 29.20 (p. 394), for asyndeton see: kuiat
ull [ezzi]
U
[
RU
N]a i.xnsi [tu] Let whoever defaces it (scil., the inscription)
be the enemy of the city Nea KBo 3.22:35 (OS).
29.10. In OH and MH texts conditional clauses (if clauses) are often connected
to following main clauses without conjunction (i.e., asyndeton; marked with ): takku
LU
ux.oXn (dupl. KBo 6.3 i 10 adds
URU
atti) kuiki kunzi 1 +i x. xi.nnnn
pi If someone kills a (Hittite) merchant, he shall give 100 minas of silver KBo 6.2
i 3 = Laws 5 (OS); takku i.t
19
.it-an itt++ kuiki dauwai nama zt
9
si lki
kar 1 x. xi.nnnn piker If someone blinds a free person or knocks out his
teeth, they used to pay 40 shekels of silver KBo 6.2 i 910 (Laws 7, OS); takku ume
natta aktni kni i t.or-ea t oi numa memai +w': +ut:+ If you do not
know (my fathers instructions), is there not here (with you) even an old man that he
may tell you my fathers word? KBo 22.1:56 (OS); mn itma uwai ar (a) ki
But if not (= if you dont bring the prisoner), you will proceed to die HKM 35:89
(MHlNS); mnata
LU
pitteanza
URU
attuaz
URU
[Kizzuwatni] paizzi
m
unauraan
ptu If a fugitive goes from attua to [Kizzuwatna] unaura must seize him KUB
8.81 ii 1112 (MHlMS). But there are cases in OS and copies of OH manuscripts where
a clause-initial conjunction occurs: takku i.t
19
.it-a itt++-a 'ssi nama olnsi
kuiki tuwa[rnizzi] nue 20 o xi.nnnn pi If someone breaks a free persons arm
or leg, he shall give to him 20 shekels of silver KBo 6.2 i 2021 (Laws 11, OS); mn
(var. nama) t\+ xtn
URU
atti nuza (var. nu-uz-za) unattallanpat arnuzzi If (var. or
if) it is in the land of atti, he shall bring (the dead body of) the merchant himself
KBo 6.3 i 1213 (Laws 5, OHlNS), dupl. KBo 6.2 i 56 (OS). The conjunction nu is
used mainly when a clitic is attached, but see also the rst example in 29.18 (p. 394).
For an example with ta and no clitic see 29.21 (p. 394).
29.11. In post-OH (including most OHlMS and OHlNS manuscripts) nu is normally
used to introduce main clauses following subordinate ones: (But if it is not cultivated
land, but uncultivated (steppe), (they shall measure) 3 us in this direction and 3
us in that,) nukan kui kui tnt-a anda srs-ri nu appat di takku tnt-
a t oi nakan amenzi and whatever town is determined within (that measured
area), he shall take them. If, however, there is no town (in that area), he will forfeit (his
claim) KBo 6.4 i 1213 (Laws IVb, NH); (upaiya said to
d
Inar:) m(n)wa kattiti
em[i nu]wa uwami kardiata iyami If I may sleep with you, I will proceed to do
what you wish KBo 3.7 i 2526 (OHlNS).
u
29.12. u connects independent clauses, which typically refer to sequential actions:
tuppu akanda unna nu utxt.xrsi andan zikt u u-a tarna She lled (the
29.10 Conjunctions
00-HittiteGrammar.indb 392 6/5/08 4:20:24 PM
393
interstices of) the baskets with grease (in order to waterproof them), put her sons in-
side, and released them to the river KBo 22.2 obv. 23 (OS);
m
armun
m
Nunnunna
UR.SAG
Taayai peuter nu otu-li turier
m
Nunnua
LU
kainan an pper an
m
armuwi
m
Nunnuwiya akuwama ukta They led S. and N. to Mount T.
and (nu) yoked them like oxen. They also seized the in-law of N. and (u ) slaughtered
him before the eyes of S. and N. KBo 3.34 i 1518 (OHlNS). Note the apparently
parallel use of nu and u in the last example, usage assured as OH by the OS duplicate
KUB 36.104 obv. 1415; itoi-a ts+i a iyanni The king heard (about it), and
he set out KBo 22.2 rev. 7 (OS); i i.xr tnt-tt+ natta pianzi u tameer e aker
But the men of the city would not give (them), so they oppressed them, and they died
KBo 22.2 rev. 1213 (OS); an
d
alma [uiz]
d
u( )mi par pai an ipandi na-
kkit dun Then alma uit, their god, handed it over, and I took it at night by force
KBo 3.22:4648 (OS). Note that the last clause in each of the last two examples may
also be read as nal clauses: so that they died and so that I took it. . . . We also
nd one example in a post-OH copy with the imperative: ntn.so-ikuwatan [r]x.
iXiat udau an pedii du If it is on the mountain, let the bee bring it and put it in
its place KUB 43.60 i 56 (OHlNS). There are no examples of u or ta linking clauses
with imperatives in OS, while there are several with nu.
29.13. u connects a subordinate (temporal) clause to a following independent clause:
mna appezziyana kitanziattat an
d
alma [uiz]
d
u()mi par pai But when
subsequently (the city) suered famine, their god almauit handed it over KBo
3.22:4547 (OS). The absence of u connecting conditional clauses to main clauses is
expected, given the restriction of u to occurrence with the preterite.
29.14. u connects a preposed relative clause to a main clause: ardian anna
kuin uwatet u
URU
al[(amp)i . . . ] And also the auxiliary troops that he brought, them
in alampi [ . . . ] KBo 3.22:37 (OS).
ta
29.15. ta connects independent main clauses in a chronologically sequential series of
actions: [(uwa)]mi kdanda pattanit ekan utikemi ta zaikemi [(taza)] utn arnikmi
kdanda natt
!
(text na-ti-i-da) tan kardama al [(ikti)] I will proceed to carry ice
with this basket. And I will ght, and I will destroy the land with this arrow, and you
(the arrow?) shall penetrate their hearts KUB 31.4+ obv. 89 (OHlNS); itoi-u 3si
aet rri [(tat)] urtiya[(li)]ya lui xtts.itoi-a 3si aet [()]rri nat
urtiyaliya lui The king washes his mouth three times and pours it (the water) into
the basin. The queen also washes her mouth three times and pours it into the basin KBo
17.1 i 1516 (OS). Note the apparently completely parallel use of nu and ta here; [ . . . ]
d
antaepu teummiua di t [ata] par paiw[a]n[i] [ . . . ] takes the .-deities
and the vessels, and we leave KBo 17.1 I 3839 (OS). See the use of nu in [ . . . . . . ]
kiarimi di nata par paiwani He puts [ . . . . . . ] in my hand, and we leave KBo
17.3+ i 15 (OS).
Conjunctions 29.15
00-HittiteGrammar.indb 393 6/5/08 4:20:24 PM
394
29.16. ta introduces the last clause of a multi-clause apodosis: takku .X.nr. o.
n itoi kuiki arzi aan luzzi nat[ta karpiezzi] itoi-u
GI
ntn-az ru-
an di tae pi If someone holds elds as a gift from the king, he shall not [render]
aan (or) luzzi: the king shall take food from his table and give it to him KBo 6.2 ii
4344 (Laws 47a, OS).
29.17. ta introduces a result clause: kinuna itoi-u idalu mekki un ta itoi-
wa uddrramet l arrattuma But now I, the king, have seen much evil, so you
(plural) must not transgress even (-al-ya) the words of me, the king KBo 3.28:2021
(OHlNS). In the following example, the NH scribe has used nu to introduce the second
result clause: ntn.so-an tarmaemi tata edi natta nari [(arun)]an tarmmi nu ppa
natta lui I will x the mountain in place, so that it may not move; I will x the sea in
place, so that it may not ow back KUB 31.4 + KBo 3.41 1314 (OHlNS) with dupl.
KBo 13.78 obv. 15.
29.18. ta connects two subordinate clauses: [(takk)]u i.t
10
.it-an kuiki nikzi
tan itarnikzi nu apn ktizzi If someone injures a person, and (ta) he falls ill, he
must care for him KBo 6.2 i 1617 = Laws 10 (OS) with dupl. KBo 6.3 i 2528 (OHl
NS); a wide variety of conjunctions join the following multi-clause protasis: takku i-a
otusi u-an znukezzi tamaiaan [(uwaizzi)] nu otu-a xt-an pzi ta u-an zi i
uit o[(tu u-a pdai)] If a man is making his ox cross a river, but (-a-l-ma-) another
(person) shoves him o and (nu) seizes the oxs tail, and (ta) crosses the river, and (i)
the river carries o the owner of the ox, . . . KBo 6.2 ii 3032 = Laws 43 (OS) with
dupl. KBo 6.3 ii 5253 (OHlNS).
29.19. ta connects a subordinate temporal clause to an independent clause: tu.2.xx
mn lukta t [a . . . ] On the second day when it dawns, . . . KBo 25.68 + KBo 17.13
rev. 9 (OS); [m(n lukkat)]tama tapa aanzi But when it dawns, they open up
KBo 17.11+ i 31 (OS); nu mn itoi-un xtts.itoi-a[nna (aniem)i] ta itoi-i
kian tmi And when I treat the king and queen, I say to the king as follows: StBoT
8 ii 1516 (OS).
29.20. ta connects a preposed relative clause to a main clause: [k]ui aga kari
ta itoi-i xtts.itoiya tarweni We tell the king and queen what portent occurs
KBo 17.1 iv 9 (OS); nu antezziya i-a kuit kuit p [eta] tae arnikzi He compen-
sates the rst man for whatever he gave KBo 6.3 ii 67 (Laws 28); [kui]za
LU
ippari
ppar iezzi nakan pparaz [eme]nzi
LU
ippara kuit apparit ta(t)z ppa di
Whoever makes a purchase from a ippara-man shall forfeit his purchase price, and the
ippara-man shall take back what he sold KBo 6.2 ii 5152 (Laws 48). Note the use of
both nu and ta connecting a relative clause to a main clause in the same passage.
29.21. ta connects a complex conditional clause to a following main clause: takku
iwarua i
A.A
kuli rki nam[ae] i.xr tnt-tt+ .X.nr.-an pianzi ta luzzi
karpiezzi If an heir cuts out for himselflherself idle land, or the men of the village give
himlher land, (s)he shall render the luzzi-services KBo 6.2 ii 4142 (Laws 46, OS).
29.16 Conjunctions
00-HittiteGrammar.indb 394 6/5/08 4:20:26 PM
395
29.22. ta connects a single conditional clause to a following main clause (so far
only in post-OH copies): takkuwaan k azzizi tawa urorn-ti+ takkuwaan
nattama a[zzizi] tawa antuwae tawa [ . . . ] zauweni If he guesses this cor-
rectly, it is a god; if he does not guess correctly, it is a human being, and we will ght [
. . . ] KBo 3.60 ii 1417. The absence of a subject clitic in both apodoses may be due to
the subject being non-referential (it is, not he is: meaning essentially we are dealing
with a god or a human being).
Enclitic Clause-Linking Conjunctions
29.23. In addition to the clause-initial clause-linking conjunctions nu, u, and ta,
Hittite has two enclitics, topicalizinglcontrastive -al-ma (cf. CHD -ma) and conjunc-
tiveladditive -al-ya, which among other uses can serve to link clauses.
-a/-ma
Form
29.24. In OH the topicalizinglcontrastive enclitic conjunction appears as -a after
words ending in a consonant and as -ma after vowels (see Rieken 2000a: 412, following
Melchert).
8
As rst pointed out by Houwink ten Cate (1973b), topicalizinglcontrastive
-a does not geminate a preceding consonant, whereas conjunctiveladditive -a does. See
below, 29.38 (p. 399), for more on this feature.
29.25. In MHlMS -ma is generalized as the topicalizinglcontrastive marker also after
words ending in a consonant. The non-geminating -a is regularly still used only after
independent personal pronouns (see end of 29.30, p. 397, below, on this usage): e.g.,
uga KUB 14.1 obv. 26, KUB 23.72 obv. 40; ammuga HKM 10:26; ziga HKM 10:30;
tuga HKM 52:10; wea KUB 23.77 lower edge 50; umea KBo 16.27 i 7; uma
HKM 75:22. It is also still found occasionally with personal names, but alternating with
-ma:
m
Madduwattaa KUB 14.1 obv. 18 etc. but also
m
Madduwattama KUB 14.1
obv. 72 etc. Other use of non-geminating -a in MH compositions is extremely rare:
antuaakan HKM 38:5; memiyanuakan KBo 16.27 iv 27; katitaman KUB
14.1 obv. 12; inaamaan KBo 39.8 ii 7 (MS, cf. Miller 2004: 219). In NH only
-ma is used productively. kinuna serves as a mere variant of kinun now, and an isolated
example such as ugakan +\+
m
rn.oi . . . In me, Muwatalli . . . in KUB 6.45 iii 60
is surely an imitation of older usage in the context of a prayer.
Position
29.26. -al-ma is usually attached to the rst accented word in the clause. There are
two systematic exceptions (see the CHD under -ma f for further details). First, in OH and
8. We do nd in OS a few examples of -a after -e: e.g., ka KUB 31.143a iii 23 and kata in KUB
8.41 ii 9 vs. kmata ibid. ii 13 in parallel contexts. See also p. 399, n. 14.
Conjunctions 29.26
00-HittiteGrammar.indb 395 6/5/08 4:20:28 PM
396
MH -al-ma is not attached to the subordinating conjunctions takku, mn, and disjunctive
nama
9
but is postponed to the second accented word:
10
takkuatan parnama kulka
peiezzi But if he disposes of them in someones house KBo 6.2 ii 35 (Laws 44b,
OS). In NH one nds instead regularly mnma. Second, when it is used to mark alterna-
tives in double questions (see CHD -ma a 1 b 4 and 29.31, p. 397, below), -al-ma is
attached to that word in the second question which constitutes the alternative, wherever
that word appears in the clause (see the rst example cited in 29.31).
11
29.27. -al-ma occurs only on the second and succeeding members of a sequence
of topic changes. In what at rst glance seems to be an exception: uiwatarmapa
anda ingani aminkan inganamapa anda uiwanniya aminkan (so HED H 66,
CHD -ma a 2; same mistaken analysis on ki-nu-na-ma-pa by HW
2
A 128b) we must
rather segment uiwatarm(u)apa . . . inganam(u)apa and render Life for me is
bound up with death; death on the other hand (ingana) is for me bound up with life
too (-ya) (so CHD -mu a, correcting the above earlier analysis). The loss of u before
the a in -apa is in accord with the rule stated in CHD -mu a (and see above in 1.72,
p. 32), and explains the a between ingan and -ma-pa. The -a is the correct form for
topicalizing -al-ma following a consonant.
Function
29.28. The main function of -al-ma is to signal a change of topic with respect to
some constituent in the preceding clause. Although there is a semblance of clause con-
trast, in most cases it can be seen that what is being contrasted is a particular constituent
in each of the two clauses.
29.29. Thus (as in the CHD article on -ma a) the rst and primary function is mark-
ing the correlation of single words in adjacent clauses: kar . . . kinuna formerly . . .
but now Laws 81 (OS) and passim in the laws; kar . . . [app]ezziyana formerly . . .
but subsequently KBo 3.22:39, 41 (OS), see ibid. 46; arzzi . . . kattirrama above
. . . below KUB 31.127 ii 1, 3 (OHlNS), zo-aztet . . . oin-lazmatta on your right
9. nama, being itself a contracted form of nauma (1.77, p. 33), would not be expected to host an
additional -ma, although one does nd nammama.
10. We are unaware of examples of postponement beyond the second position. Apparent examples
involving logograms are misleading, since the underlying Hittite word order diers, e.g.: mn r
A.A
.
oXnma fr-anza but if the elds owner is (still) living KBo 6.4 iv 18 (Laws XXXVII, NH), ed. LH 57,
where the underlying Hittite would have the genitive elds (
A.A
.oXn) preceding the sg. nom. owner
(r = i). From a Hittite perspective the entire group r
A.A
.oXn is a unit. See also [(takku .X.nr)].
s+ s+\i+ma arzi But if he holds elds of another KBo 6.2 + KBo 19.1 ii 16 = Laws 39 (OS). The rule
that -al-ma is not postponed beyond the second position helps to demonstrate that in takku
A.A.I.A
kulima
rki But if he divides up kule-elds KBo 6.2 ii 47 = Laws 47B (OS) .X.nr. is a determinative and kuli
a nominative-accusative rather than a dative-locative; see LH 56 n. 187.
11. There are some instances of postponed -ma whose motivation remains to be explained: e.g., nukan
mai imma tu.nr. andama srs-ri nakan tot otn-ri He (the king) will stay up there however
many days are determined for that(?) by oracle KUB 5.1 i 6061 (NH).
29.27 Conjunctions
00-HittiteGrammar.indb 396 6/5/08 4:20:29 PM
397
. . . on your left ABoT 44 i 5960 (OHlNS); otu-uza xn-un it kappuwaizzi
[tu]t-umaza sri
4
-an it kappuwaizzi The cow doesnt take care of her calf; the
sheep doesnt take care of her lamb KUB 33.37 + KUB 33.39 iv 4 5 (OHlNS); nuza
kuia (var. [ku]i) utxt.rf-li [ru-]an tzt.lya pi [kui]maza utxt.rf-li
aku[(wa)]nna pi One (of the men in the community) gives bread and meat to his son,
and another gives something to drink to his son KUB 24.8 i 1921 (OHlNS) with dupl.
KUB 43.70b:68. Several of the above marked pairs are lexically contrastive (right-left,
upper-lower). But we also nd the particle marking identical pronouns or pronominally
based adverbs in a sequence where each has a dierent referent: 3 wattaru it kdani
. . . kdanima . . . kdanima . . . He made three wells: at this one . . . , at this (other)
one . . . , (and) at this (other) one . . . KUB 33.59 iii 79 (OHlMS); kman kn pz[i]
[k]aman kn pzi This one might seize that one, and this (other) one might seize
that (other) one KBo 6.2 ii 5657 (Laws 49, OS); kz . . . kzma on this side . . . but
on this (other) side KUB 19.37 ii 20, 21, 24 (NH).
29.30. From the above examples and the others cited in the CHD -ma a 1 it can be
seen that the translation but rarely ts and that, while there is contrast, the primary con-
cern is signaling a change of topic: from the cow to the sheep, from the right hand to the
left, from the spring to the harvest, from the one well to the other two in the sequence,
or the one father giving food to the other father giving drink. Sometimes a change
of subjects is marked by -al-ma. iai tzt.l di apa xt tut di his (the
shepherds) master shall take the meat (of the sheep), but he himself (i.e., the shepherd)
(-al-ma) shall take the sheepskin KBo 6.2 iv 1415 (Laws 80, OS); maanmaza
+ui:+
m
Murili urorn-tt+-i kiat r:+mazakan
m
rn.oi +\+
GI
ot.z +utsi
eat ammukmaza +\+ r+\t r:+ r.xn kiaat But when my father Murili
died (lit., became a god), and my brother Muwatalli sat down on the throne of his father,
I during my brothers reign became commander of the army KUB 1.1 i 2224 (NH). It
is this common use of -al-ma to mark a change of subject that accounts for its popularity
in being axed to (often fronted) independent pronouns: -ga but I StBoT 25 #3 ii 6
(OS); StBoT 25 #4 iii 37 (OS); zi-ga but you HKM 58:30 (MHlMS), KUB 1.16 ii 67
and KUB 31.110 10 (OHlNS); -e-a but we KBo 22.2 obv. 11 (OS), KBo 16.50:11
(MHlMS); u-me-a but you (pl.) KUB 1.16 iii 49 (OHlNS); u-me-e-a KUB 23.72
rev. 64 (MHlMS); [u-me-]e-a KUB 23.68 rev. 11 (MHlNS); a-pa-a-a but she KBo
6.2 iv 15 (OS); a-pa-a KBo 3.22:72 (OS), VBoT 58 i 16 (OHlNS); a-p-e-ma but they
KBo 20.26 + KBo 25.34 obv. 14 (OS).
29.31. Obviously a construction so well tted to mark a change in topic and signal
a contrast was perfectly adapted to mark alternatives in double questions which pose
contrary or at least mutually exclusive alternatives: nat pnzi +\+ urorn-tt+ ts:i
4

pianzi . . . nat +\+ urorn-tt+ ts:i xi.orma pianzi Shall they proceed to give it to the
deity with gems . . . or shall they give it to the deity with gold? KUB 22.70 obv. 5153
(NH); especially common when the second alternative is a simple negation: nuwarat
r[:+ toi nuw]arat tma toi Does my brother know it, or does he not know it?
Conjunctions 29.31
00-HittiteGrammar.indb 397 6/5/08 4:20:30 PM
398
KUB 14.3 i 52 (NH). That the -ma in these cases is not a clause connective is obvious
in the cases where its clause is already introduced by nu.
29.32. It may well be that from the established use of -al-ma to mark change of topic
within two adjacent clauses there developed a more general use to contrast the adjacent
clauses themselves. One can probably see this in the rst -ma occurring in the example
from KUB 1.1 i 2224 (but when . . .) cited above in 29.30 (p. 397). But most of the
examples cited in CHD -ma b (pp. 9394) to support this usage can instead be used as
illustrations supporting the former use. What makes some of them appear to mark clause
contrast is the fact that the element within the clauses that is contrasted is the verb itself,
even identical verbs, or the negation: aliyawa t wi tmawa wki tmawa
iparrezzi The deer makes no (loud) cry, nor does it bite, nor does it trample KUB 14.1
rev. 91 (MHlMS) (further exx. in CHD -ma b 4); t otu-u tmawa tut-u there
were no cattle, nor were there sheep KUB 36.51 obv. 7 (OHlNS). Nevertheless, there do
appear to be rare examples where the contrast is probably
12
the entire clause: If a girl
is betrothed to a man, and he pays the bride-price for her, and subsequently the parents
abrogate it and withhold her from the man, they shall repay double the aforementioned
bride-price,) takku i-a (-al-ma) utxt.xtts nawi di nanza mimmai but if the
man has not yet taken the girl, and he refuses her, (he must forfeit the bride-price
which he paid) KBo 6.3 ii 14 (Laws 2930); takkuza xtts-za i-an m[immai ]
. . . takkuza i-a xtts-an uw[ezzi] If a woman rejects a man (i.e., wishes a di-
vorce) . . . If on the other hand a man expe[ls] (i.e., divorces) a woman KUB 26.56 ii
1, 4 (Laws 26, OHlNS); If in a village someone holds elds as an inheritance share
(iwaru), . . . takkue .X.nr.-a te[-e-pu-u piyanza] But if the elds [are given]
to him f[ew in number] KBo 6.2 ii 39 (Laws 46, OS); takku addaea uiwanza
but if his father is (still) living KUB 29.34 iv 13 (Laws 190, OHlNS).
29.33. The use in the expression nammama . . . tamai- (CHD -ma b 8) may have
to do with the implied topic change signaled by the adjective tamai- another. See also:
[(tami )]akan pran wal[az]i but another man strikes rst KUB 29.30 + KUB
29.29 ii 18 (Laws 148, OS) with restoration from KBo 6.10 iii 26 (OHlNS).
29.34. The use with anaphora (CHD -ma d) also grows out of the topicalizing func-
tion, although the change of topic may not always be very clear. That it does grow out
of this topicalizing function is seen in the fronting of the repeated word, to which then
-al-ma is suxed:
d
tft-at aruni pda[] arunima tntut-a palae kianda[ri']
The Sun-god took it o to the sea, and in that sea copper vessels are lying KUB 33.66
ii 89 (OHlMS?); +\+ urorn-tt+ pran
GI
eyan a[rta]
GI
eyazmakan tut-a kura
kank[anza] Before the deity a r tree st[ands], and from that r tree a hunting bag is
han[ging] KUB 33.38 iv 67 (OHlMS).
29.35. The CHD proposes to see another extension of the root use of -al-ma in its
use to show concurrent actions, with or without kuitman in the immediately preceding
12. The motivation here too could be change of subject from girl and parents to the man.
29.32 Conjunctions
00-HittiteGrammar.indb 398 6/5/08 4:20:32 PM
399
clause (-ma b 7): nu kuitman +ui:+ t\+ xtn
URU
Kargami
i
kattan ta
m
Lupakkin
ma kan . . . par naita And while my father was in the land of Carchemish, he sent
out Lupakki . . . KBo 5.6 iii 13 (NH); nat par tiyanzi
LU
l.ut
8
maakan a[tti ]li
lamnit alziai (those who spend the night up in the palace) step forward, (while) the
gateman calls them by name in attic KBo 5.11 i 67. And although here too the actual
motivating factor may sometimes be either a change of subject, verb, or topic, there does
seem to be good reason to believe that this extended usage exists.
29.36. The root use of marking a change of topic is further extended in the use
of clause-initial andama now in addition,
13
which (often immediately following a
paragraph break) signals a change to another matter for discussion (see CHD -ma e
2 a). In the Akkadian letters from Amarna the same function is performed by the word
antam now in addition (e.g., on another subject; lit., secondly). Another good ex-
ample is found in a MH letter from Kuakl-aria: +\+ utxt.xtts.xrya lara
(42) xtr.nr. aumen nu lara (43) xtr.nr. ara pe[]ier (44) andamamu
+\+ o[i] xi.or (45) kuit atre nan kankaun (46) nu 30 o.o.t xr.iisi
(47) nan +\+
m
Walwalli +oot\ And about the young women we observed the (oracle)
birds of lara, and the birds gave a negative answer. On another subject: regarding
what you wrote to me about the gold cup: I weighed it, and its weight was 30 shekels.
And I gave it to Walwalli KuT 50 4147, ed. Wilhelm 1998: 18184.
29.37. For -al-ma in kuia see 8.4 (p. 150).
-a/-ya
Form
29.38. Conjunctiveladditive -al-ya takes the form -a after syllabically written Hittite
words ending in a consonant, which consonant is then geminated, but -ya after syllabi-
cally written words ending in a vowel: a-pa-a-a-a (apa) (s)he too, a-p-e-ya
(ap ya) those too.
14
In Old Hittite the topicalizinglconstrastive particle would be
written a-pa-a-a (apa) he on the other hand and a-p-e-ma those on the other
hand. In NH there occasionally arises a double writing of the conjunction (-aya) fol-
lowing a word ending in a consonant, usually with the gemination of that consonant
13. This is similar to the use of anda in andayai and in addition to it KBo 6.3 iii 61 (OHlNS, older
andae KBo 6.2 iii 57, OS) = Laws 70. The example an-da-e-ya Laws 110 in copy d (LH 107, 269)
shows the same order of clitics as parnaea uwayezzi.
14. There is only one small set of exceptions to this rule. We do nd in the older language a very few
examples of -a after -e (HW
2
A 43b): (As he chants before the alputi), ka +:++++ these things too
(he chants) in the same way KBo 25.112 ii 14-15 (OS); nu ka +:++++[pat l]inker These (men) too
have sworn an oath in the [very] same way KBo 8.35 ii 28 (Treaty; MHlMS); t-\+ ru.nr. a-ku-wa-an-
na-e-a
LU.ME
ipunn[ale] [ . . . ]-zi the ipunala-men [] in food and drink KBo 11.34 i 9-11, where the
extra -a shows that the -e is not simply a misdrawn -ya. See also LH 168 on parnaa in the Laws (and
in 30.17, p. 411, below). These examples may reect the use of -e-a to spell l-yal in word-nal position.
See AHP 35 and 141. Apparent exceptions such as me-ma-al-ya KBo 15.34 iii 8 and i-ya-at-ya KUB 21.38
obv. 54 are the result of weary, careless, or unskilled scribes.
Conjunctions 29.38
00-HittiteGrammar.indb 399 6/5/08 4:20:33 PM
400
(HW
2
43b): a-pd-da-ya that too, a-ra-wa-an-na-a-a-ya and the a.-ritual, a-a-
da-an ar-ra-a-an-na-ya dried and crushed, ou-it-ta-ya and with a linen cloth, but
rarely without the gemination: ki-nu-na-ya now too, a-pa-a-a-ya and that one. In
Old Hittite the gemination or lack of such before -a, which distinguished the conjunc-
tive-additive from the topicalizinglcontrastive marker, was often obscured with logo-
grams by the occcasional practice of writing only one syllabic Hittite sign after the
logogram. If itoi and xtts.itoi were nominative, the pair could be written either
itoi(-u) xtts.itoi-a-a or itoi(-u) xtts.itoi-a (standing for *au
auaraa) the king and queen. But in post-Old Hittite the same combination
would be written itoi(-u) xtts.itoiya. Because the topicalizinglcontrastive -a
had disappeared in New Hittite (see 29.25, p. 395), copyists did not always understand
its usage in older texts and replaced it incorrectly with geminating -a rather than with
-ma. One must therefore sometimes interpret geminating -a in such copies as the topi-
calizinglcontrastive marker. Rarely the opposite is true: intended conjunctiveladditive
-al-ya is wrongly written without gemination of the preceding consonant: [ku-i]t da-a-
an da-ya-na what was taken and stolen HKM 30:24 (MHlMS).
Position
29.39. When it connects clauses, conjunctiveladditive -al-ya regularly is attached
to the rst accented word in the clause. Like -al-ma (see 29.26, p. 395) it can be post-
poned to the next accented word following takku, mn, etc. Since these subordinating
conjunctions are regularly clause-initial, this means that postponed -al-ya and -al-ma
must attach to the second accented word in the clause. If for some reason the subordinat-
ing conjunction itself is postponed, -al-ya and -al-ma will appear on the next accented
word following the conjunction: [nu]za mn irmalanzaa ta
d
tft-stma[t]ta []
+s+u +utx+ tittanununpat Even though you were ill, I His Majesty still installed you
in the place of your father KBo 5.9 i 1617 (NH). For further examples of -al-ya post-
poned to the second word of the clause see CHD mn 7d. When -al-ya connects a series
of words in the same clause, it attaches to the last word in the series (see 29.40).
Function
Connecting Individual Words
29.40. -al-ya is the only connective which joins individual words. Some examples
are: nutta urorn.xr
d
r.-aa attanna itoi-u auli paantaru May the gods
and Ea, the (divine) king of wisdom, keep you in good health HKM 3:1820; kawa
LU.ME
apaalli +sriu nuwa
URU
Malazzian
URU
Taggatanna apa
!
iyar I have just
sent scouts, and they have scouted the cities Malazzia and Taggata HKM 6:1822
(MHlMS).
Connecting Clauses
29.41. -alya- can also be used to connect clauses, when actions are viewed as paral-
lel to one another: (he shall pay so many shekels of silver) uya 3si arnikzi and
he shall replace the goods threefold KBo 6.4 i 5 (laws parallel series III, NH), (he
29.39 Conjunctions
00-HittiteGrammar.indb 400 6/5/08 4:20:34 PM
401
shall give six shekels of silver to the injured man,)
LU
.ztya kuan appat pi and
he shall give the fee to the physician KBo 6.2 i 19 (Laws 10, OS); (If a man deles
a vessel, previously they gave 6 shekels of silver, he who deles (used to) give three
shekels of silver,) [itoi]-anna parna 3 o xtnnnn daker and they used to take
three shekels of silver for the house of [the kings] KBo 6.2 i 58a (Laws 25, OS). It is
important to note that Hittite -al-ya, unlike English and or German und, is not used
to connect consecutive actions (see correctly Friedrich HE I 304). That is, it does not
stand for and (then), for which Hittite uses one of the sentence connectives nu, u, or
ta (see 29.6, p. 390; 29.12, p. 392; 29.15, p. 393).
Correlative both . . . and
29.42. The Hittite equivalent of both . . . and is a pair of correlative -al-ya, whether
connecting individual words or clauses. Examples: ppanna peranna both behind
and in front KBo 7.14 i 9 (OS); itanana ketta ktta on both this side and that
side of the altar KUB 32.117 rev.
!
12 + KBo 19.156 iii 5 (OS); erma(n)makan
dun kardimiyaatkan du[n] [(ara)]nimiyaatkan dun I have taken
the sickness from you. I have both taken it from your heart(s) and taken it from your
head(s) KBo 17.1 i 1213 (OS). Double -al-ya plus the negative is equivalent to nei-
ther . . . nor: nammamakan xtn
URU
aplla kuentaya t ptayaat t But then
you neither attacked the land of apalla nor seized it KUB 14.1 ii 23 (MHlMS).
Additive also, too and Concessive even
29.43. In some cases clause-connecting -al-ya has an additive force best expressed
by also or too: (If I, My Majesty, go to war . . . ) numu ziqqa +oi rnr.xr r.
xtn.n.xr kattan laekei you too shall go to war with me with your troops and
chariots KUB 21.5 iii 21 (NH); nu-za
d
ts:+u o-:+ par andandatar apdaniya
meuni tikkuanut Is:+u My Lady displayed her providence also at that time att.
iii 1516 Also att. iv 1819 in a variant of the same sentence. See the position of
apiyaya then too in att. ii 37, 38, 45.
29.44. Sometimes -al-ya has the concessive sense even (see German auch): nama
kan kl tuppia 1-anna memian wanuzi or changes even one word of this tablet
BrTabl. iv 1920 (Tud. IV). This use is especially common with the conjunction mn
in the meaning even if (see 30.46, p. 419).
29.45. When -al-ya functions in one of these two senses, it occurs on the focus word,
wherever that word occurs in the clause.
Asyndeton
Form
29.46. Asyndeton is the simple juxtaposition of coordinate discourse elements with-
out conjunction. Our concern here will not be with juxtaposed nouns and noun phrases
Conjunctions 29.46
00-HittiteGrammar.indb 401 6/5/08 4:20:34 PM
402
(e.g., ezzan or-ru straw (and) wood, and the Hittite equivalents of otu tut cattle
(and) sheep
15
), but with clauses.
29.47. The absence of a conjunction is normally obvious and therefore easily rec-
ognized. But there is at least one situation in which it is masked. When a clause-initial
word ending in a consonant hosts a clitic pronoun beginning with the vowel a, a situ-
ation is created in which the presence of the clitic conjunction -a-l-ma-
16
in OH or
the clitic conjunction -a-l-ya- in NH is uncertain. In Old Hittite a combination such as
zi-ga-an could be analyzed either as zigaan (with -a-l-ma-) or as zigan (asyndeton).
One can disambiguate such cases only by comparison with clear examples in parallel
constructions: kuelaa ari aki on whoevers property he is killed KBo 6.2 i 78
(Laws 6, OS) can be shown not to be asyndeton on the basis of clear examples such
as kuiama i.xn i )i+t uizzi whatever enemy sends an envoy to you KUB
23.72 rev 22 (MH); kuitmna lazziatta until he recovers KBo 6.2 i 1718 (OS) is
shown to be asyndetic by other examples of postposed kuitman clauses (30.37, p. 416).
The situation is even more obscure when the host word is a logogram, in which case
gemination of the nal consonant might not be shown by the writing of the phonetic
complement. In OS itoi-a-an can stand for *itoi-uaan, *itoi-uaan, or
*itoi-uan! In post-OH, when -ma assumed the role of its allomorph, there remained
no possibility of confusing such clauses in which topicalizing -ma occurred with cases
of asyndeton: the former was written zi-ik-ma-an (e.g., KBo 10.13 + KBo 10.12 ii
32); rorn-anmaas KBo 6.4 ii 21); itoi.oimaankan (KBo 10.2 ii 47), LUGAL-
(u-)ma-an;
17
and the latter zi-ga-an KBo 5.9 ii 41; itoi-u-at-za BrTabl. iv 23, or
urorn-ti+-an KUB 5.1 i 65, 68, iii 62. While further study is needed, it appears that
in NH ambiguous spellings with logograms such as itoi-a-an were avoided in favor
of indicating asyndeton either with a fuller phonetic complement (e.g., itoi-u-an) or
no complement at all (e.g., itoi-an).
Function
29.48. The most obvious situation in which asyndeton is appropriate is where there
is no connection to a previous clause. One nds this use primarily at the beginning of
compositions, or at the beginning of recorded speech. Beginning entire compositions:
15. The combination itoi xtts.itoi king and queen, on the other hand, is shown by OH ex-
amples to have contained a clitic conjunction.
16. Weitenberg (1992: 319) failed to notice this, when he claimed i-a-a-a- [i- ]a-an KUB 29.25 i 8
(Laws 121, OS) as asyndetic *iian, whereas in its context as a fronted contrasted subject it clearly
contained topicalizing -al-ma (*iiaan).
17. But there are examples of -a-l-ma- erroneously written with gemination by NH scribes, such as a-ki-
i-a-an KBo 3.36 18, where its duplicate A ii 12 has a-ki-i-ma-a tepauwanni but he died in poverty,
showing that the OH original had *a-ki-a-an (akiaan), with topicalizing -a-l-ma-and OH local particle
-an (but see Dardano 1997: 9899, who takes it as asyndeton akian). Both Dardanos interpretation and
Kammenhubers interpretations (the latter reading akian starb ihn) require asyndeton. But since asyn-
deton is not attested to introduce strong contrast (but, rather), one may doubt these analyses.
29.47 Conjunctions
00-HittiteGrammar.indb 402 6/5/08 4:20:35 PM
403
d
tft-e iami andanza annena ia . . . xtn-e zik duddukei O Sungod, my
lord, just lord of judgment . . . you have mercy upon the land(s) KUB 31.127 i 13.
Beginning a long narrative following a brief introduction: karan karil[iy]a xt.nr.
-a
d
Alalu -i itoi-u ta Long ago Alalu was king in heaven KUB 33.120
i 78. Beginning recorded speech: karwa ar
URU
attui makketa Formerly in
attua bloodshed was common THeth 11 ii 48 (OHlNS); even in a dialogue where the
speakers change several times: see KUB 30.28 rev. 113 with dupl. KBo 34.80:18, ed.
Otten 1958: 9697, or when a single speaker simulates a dialog for rhetorical purposes,
asking a question and then answering it (see KUB 14.3 ii 727 in 29.51). But note that
discourse-initial nu does occur, when there is a conscious allusion to something said by
the previous speaker (then, so): see the examples cited in 27.7, p. 350 (KUB 1.16
ii 9); 27.8, p. 351 (KBo 21.22:2225); 27.9, p. 351 (KBo 21.22:4142); and 27.10,
p. 352 (VBoT 58 i 23).
29.49. When there is a preceding clause, the relationship of the two clauses de-
termines whether or not a conjunction is called for. If the subsequent of two or more
juxtaposed clauses expresses the same thought in a dierent form, usually no con-
junction is employed: ammukmaza par andnza kuit t-a eun r+\t urorn.
xr kuit par andandanni iyaaat s+ utxt x.i.t
19
.it-i::t nti-lu uttar it
kuwapikki iyanun But because I was an upright person, because I conducted myself
in uprightness before the gods, because I never committed the evil deeds of (other)
mortals, (you, O goddess, my lady, rescued me from all harm) att. i 4650 (NH). The
most common form of this repetition of the same thought in other words is when one
of the two clauses is negative and the other positive (or vice versa): exx. takku tma
A.A
.oXn dammel pdan But if it is not cultivated ground, (that is to say) it is virgin
terrain KBo 6.4 i 11 (Laws IVb, NH), see LH 1113, 20, 17273; takkui .X.
nr.-u manza it pianza tepui piyan If the cultivated land is not given him in
(its) entirety, (that is) what is give to him is partial KBo 6.4 iv 2324 (Laws XXXVIII,
NH). See also KBo 12.38 ii 410 (NH) cited in 7.1 (p. 142).
29.50. A circumstantial clause essential to the subordinate conditional clause can be
asyndetic (further exx. in HE 310e): takku i itti+ arawanniu annaneku anna(n)
manna uenzi k takiya utn ka takiya utnya t ara[tar ] If a free man has
sex with free maternal sisters and their mother the one (being) in one country and
the other in another, there is no crime KBo 6.26 iii 3234 (Laws 191, OHlNS). Of-
ten such clauses serve as inserted explanatory glosses (HE 310d): kuia imma k[ui]
s+ r.oi-tt+ xtts mna itti+ mna xtts.stntn.i Whatever palace
woman whether she is free or she is a hierodule KBo 5.3 iii 4445 (MHlNS).
29.51. Another variation of this use is when a question is asked and then answered
by a rhetorical question: s+ i.oi.oi-:t+ rsi\i kuwapi tat harkr Where are
the estates of the great men? Have they not perished? KUB 1.16 iii 45 (OHlNS);
kma
LU
+u:+rri kui s+ xtts.itoiza kuit s+ x-:t arzi t\+ xtn
URU
atti
Conjunctions 29.51
00-HittiteGrammar.indb 403 6/5/08 4:20:35 PM
404
s+ xtts.itoi x-:i+ mekki alli namu t imma
LU
n+o+\i But who is this
chariot driver? Because he belongs to the family of the queen in the land of atti
the queens family is extremely important is he not my son-in-law? KUB 14.3 ii
7274 (= Sommer 1932: ii 7375, NH).
29.52. Conditional clauses, both with takku and mn if, even if they do not start a
new section of the composition, often take no introductory conjunction: mni!kan
uitsi (!) er t arnikzi If his master does not make compensation for him (i.e., the
slave who has incurred a penalty) KUB 8.81 + KBo 19.39 iii 6 (MHlMS); mn 2 zt
8
nama 3 zt
8
lki If he knocks out two teeth or three teeth KBo 6.4 i 1819 (NH).
29.53. In Old Hittite it is common for the apodosis following a conditional clause
using takku if to lack an introductory conjunction (see 29.10, p. 392, and 30.48,
p. 420, and the rst example in 29.50, p. 403). Old Hittite compositions, even in NS
copies, and where mn instead of takku is the if word, sometimes preserve this pattern:
mnamata [ar ] iyri ap tanduki If [blood] spurts from them, they are
mortals KUB 31.1 + KBo 3.16 ii 89 (OHlNS). In NH compositions, however, this is no
longer the case, even when the NH author has used takku as conscious archaizing: takku
tnt-a t.oi nakan amenzi If there is no town, (the heir of the murdered man)
shall forfeit them (i.e., the indemnifying payments) KBo 6.4 i 13 (Laws IV, NH).
29.54. Sometimes the main clause following a subordinate temporal clause with mn
is asyndetic (see 30.36, p. 416; KBo 6.2+ iii 1617).
29.55. Postposed kuitman clauses show no advance in time and dispense with any
conjunction, just as is the case with foregrounded clauses following a background-
ing imperfective (see Hittite Laws 43). Virtually all relative clauses (30.5830.64,
pp. 424426) in Old Hittite are asyndetic, whereas those introduced by nu are less rare
in Middle Hittite and become rather common in New Hittite. Postposed temporal mn
clauses are asyndetic (for preposed ones see CHD mn 5): nata aptta 3si d
natmu u[ ppi] mn damai kuik[i] uezzi then take three times that (property stolen
from me by the slave woman) and send it to me whenever any other (messenger) comes
(here) HKM 30 left edge 14 (end of the piggyback letter).
29.56. Clauses with potential-irrealis manlmn are introduced asyndetically (HE
310311 and exx. in CHD man): mnwaramukan ulliyat kuwapi t mn
andn itoi.oi +\+ itoi.ftn katteraer If (Uri-Teup) had not at some time
treated me disrespectfully, would (the gods) have subjected a great king to a minor
king? att. iii 7677 (NH).
29.57. When describing a situation by a sequence of disjointed clauses that are not
intended to convey a temporal sequence, asyndeton can be used:
GI
luttu kammar
t,u+: r-er tuui [t,u+: ] mist seized the windows; smoke [seized] the house KUB
17.10 i 5 (OHlNS). Perhaps it is this timeless quality of a prescriptive ritual that lent
itself to the extensive use of asyndeton (HE 309b).
29.52 Conjunctions
00-HittiteGrammar.indb 404 6/5/08 4:20:36 PM
405
29.58. Some clauses with double nite verb forms and idiomatic meaning (24.32,
p. 325, nal two exx.) can be classied as asyndeton.
Disjunction
nau . . . na(u)ma
29.59. Disjunction is regularly expressed in Hittite by nama or or by nau . . .
nama either . . . or. nama developed by syncope from nauma (1.77, p. 33), which
contains the topicalizinglcontrastive conjunction -al-ma. These may conjoin clauses or
individual words: (clauses) nankan +\+
GI
ororn waggariyanun namaankan X r-:t
waqqariyanun Did I rebel against him in the chariot, or did I rebel against him in the
house? att. iii 6768; nui nau adanna pekezzi namai akuwanna pekezzi
He either gives him to eat or he gives him to drink KUB 13.4 i 24; (individual words)
takku i.t
19
.it-a i-a nama xtts-za takya tnt-ri aki If a person, male or fe-
male, is killed in another town KBo 6.2 i 7 (Laws 6, OS); itoi-u uezzi nau
GI
ororn-it nama
GI
ulugannit t\+ r-:t+ oi paizzi The king proceeds to travel to the
Great House either by chariot or by cart KBo 19.128 i 25. For further examples see
the CHD articles for nama and nau.
-(a)ku . . . -(a)ku
29.60. A pair of enclitic disjunctive markers is -(a)ku . . . -(a)ku whether . . . or,
attested from the oldest written stage of the language: i 4 so.ut pi i-naku xtts-
naku and he shall give four persons, whether man or woman KBo 6.3 i 2 (Laws 1,
OHlNS); [(andan)]a r-ri kuit arkzi i.t
18
it
LU
ku otuku [tu]tku zi and what-
ever perishes in the house the estate, whether it be human being, ox or [she]ep KBo
6.3 iv 53 (Laws 98, OHlNS); nu kuwapi
d
tft-u mumizzi [ . . . -]ikku appnikku
or-ikku aalikku mumizzi So where will the Sun-god fall? Will he fall into the
[ . . . ], or in the ame(?), or in the tree(s), or in the brush? KUB 36.44 iv 89. Very
likely it is this construction, not nau . . . nama, that underlies the Akkadographic t
. . . t . . . t whether . . . or . . . or in KBo 3.22:6063 (OS), which is only attested
in OS (see 31.35, p. 440). On the Akkadian construction see GAG 117 cld, AHw
55960 l C 23, CAD L 22427, or CDA 184 l C.
29.61. In some cases the pair -(a)ku . . . -(a)ku is equivalent to takku . . . takku if . . .
(or) if: i+++
D
r.ft lwatta ni tu[(lku wata)i ] ugat sro
5
-ziyami t-akku
(var. [na]ttakku) tu[(el w)atai] ugat sro
5
-ziyami Thus says (the goddess) r.ft:
Dont be afraid! If it is your fault, I will make it right; (or) if it is not your fault, I will
(still) make it right KUB 33.24 (+) KUB 33.28 i 4345. For -(a)ku in the negative
nekku see 26.23 (p. 346).
Conjunctions 29.61
00-HittiteGrammar.indb 405 6/5/08 4:20:36 PM
406
Chapter 30
CLAUSES
Word Order
30.1. The following description of Hittite word order applies to both independent
and dependent clauses. The latter dier only in the presence of a subordinating conjunc-
tion, for whose positioning see 30.31. (pp. 414.). In nominal sentences (30.22,
p. 412), which contain no nite verb, the subject regularly precedes the predicate (but
see 30.29, p. 414).
Major Constituents
30.2. The functionally neutral or unmarked word order in Hittite is S(ubject)
O(bject) V(erb).
1

The clause negation immediately precedes the verb (for details and
exceptions see 26.3, p. 341). In nominal sentences (30.22, p. 412) the clause negation
precedes the predicate noun, adjective, or adverb (for an example see 30.24, p. 412).
Within the predicate the order of the direct object and indirect object relative to each
other and to other constituents such as postpositional phrases, dative-locatives, and ad-
verbs is yet to be established.
30.3. Within noun phrases genitive modiers precede their head noun except for the
genitive of measure (see 16.50, p. 254, and 16.47, p. 253). Demonstratives and adjec-
tives regularly precede the noun and any genitive modier (see 18.2418.25, p. 284,
and 16.5116.52, p. 254). Attributive participles, however, and a few other adjectives
in -ant- regularly follow their head noun (see 25.41, p. 339). On the word order of the
indenite pronounladjective kuiki see 18.3318.36 (pp. 286287).
30.4. Various discourse factors not infrequently lead to deviations from the neutral
S-O-V word order.
2
In formal terms, we may distinguish at least four processes: front-
1. As usual in this formula, O(bject) actually represents any constituent of the predicate except the
nite verb, not only the direct object of transitive verbs but also such things as adverbs, negations, postposi-
tional phrases, and innitives. S(ubject) includes nouns (and noun phrases), accented personal pronouns,
demonstrative pronouns, but not the indenite pronoun-adjective kuiki (18.33, p. 286).
2. We use the vague term discourse factors intentionally. It is clear that such things as emphasis,
focus, and topicalization play a signicant role in Hittite word order, but we cannot yet systematically
correlate these with particular congurations. Furthermore, some patterns are almost certainly governed
by other considerations: See 30.58 (p. 423) on the obligatory fronting involved in determinate relative
clauses.
00-HittiteGrammar.indb 406 6/5/08 4:20:37 PM
407
ing, extraposition, and left- and right-dislocation.
3
By far the most common of these is
fronting, by which any constituent can be moved either to absolute initial position in
the clause or to the position immediately following clause-initial conjunctions and any
clitics dependent upon them.
30.5. Examples of fronting to clause-initial position: (verb; see also 26.5, p. 342)
eu piweni t\+
URU
Apzuwa ++n+u
d
r. Come, let us go to Apzuwa into Eas presence
KUB 33.106 ii 19 (Ullik. III); wanumiankan kuwapi When should I alter it? KUB
21.38 obv. 40; (negative; see 26.9, p. 343) tatikan ara danzi They shall not
take it away from him BrTabl. ii 99; (preverb) andakan alna teummiu tarlipit
wamu 2-:++ ptumini Twice(?) we bring in cups (made of) alna- lled with tar-
lipa-drink KBo 17.1 i 2627 (OS); (temporal adverb) annaz
URU
Iyaruwata tnt-a s+
xtn
URU
Parga ta Formerly the city Iyaruwata had belonged to the land of Parga KBo
3.3 i 34; (direct object) 13 r+ xi\+si
m
atipa-i
URU
Narapiduwa arzi atipaziti
holds 13 measures of emmer wheat in Narapiduwa HKM 111:56 (MHlMS); (indirect
object) urorn-tt+-niwatta ammuk tarnai I will hand you over to a deity att.
i 3738; (noun in dative-locative)
URU
Kuzuruwi kakkapu marakta
URU
Ankuwa kakkapi
maklante In Kuzuru he butchered kakkapa-animals; in Ankuwa the kakkapa-animals
are emaciated KBo 3.34 ii 1214 (OHlNS).
30.6. Examples of fronting to rst position following clause-initial conjunction:
(verb) numu memiket orx-an As she had been telling me, (so it also happened) att.
iv 17; (negative) nu tya kuit iyan ta And even that which had not been made, (she
gave that away) KUB 14.4 ii 1011; (preverb) nat rorn-pa s+ xtn
URU
atti xtn-e
iyanun I made it once again Hittite territory KBo 3.4 + KUB 23.125 iii 95; (temporal
adverb) nu ziladuwa kn memian l kuiki ulliyazzi Henceforth let no one contest this
word KBo 4.10 obv. 46; (direct object) nu r
d
ts:+u
m
Dudaliya utxt:+ tapardu
Let Tudaliya my son administer the house of ts:+u att. iv 7778; (indirect object)
nuza +\+ urorn-tt+
LU
ankunniyanza ni-aun I sacriced to the goddess as priest
att. i 1819.
30.7. When fronting occurs, usually the fronted constituent is marked by the topi-
calizinglcontrastive conjunction -al-ma (see 29.28, p. 396), but fronting also occurs
without -al-ma, as examples in the preceding paragraph show. Subjects may also un-
dergo fronting, as shown by examples in determinate relative clauses, where the sub-
ject appears before the relative pronoun that has itself undergone wh-movement (see
30.59, p. 424, and see Hale 1987: 4748 and Garrett 1994: 4546): utxtsimawa
i zakan kui +\+
GI
ot.z +utsi eat But his son, who sat down on the throne of
his father, (was already a mature man) KBo 3.4 i 1112.
4
3. The following account couched in terms of movement of constituents is intended to be purely de-
scriptive and makes no claim regarding how these facts are to be analyzed in terms of syntactic theory.
4. It is possible, but by no means proven, that any constituent marked by -al-ma has been fronted. If this
is true, then all subjects thus marked have been fronted, even those in non-relative clauses where otherwise
there is no visible sign of fronting.
Clauses 30.7
00-HittiteGrammar.indb 407 6/5/08 4:20:38 PM
408
30.8. Fronting of more than one constituent in a single clause is well attested
(against the claim of Garrett 1994: 39): nui ziladuwa s+ xtn
URU
atti laiyanni
2 +i iyattaru Henceforth let 200 of his (men) go on a military campaign of the land
of atti KBo 4.10 obv. 4344; 28 r+. zz.nr. t\+
URU
ariya
m
imu-urorn-tt+

m
Maruwa
m
Tiwa-i
LU
xt
7
xi.[o]r unikanzi imuili, Maruwa, and Tiwaziti sow
twenty-eight measures of barley in ariya HKM 111:2427 (MHlMS); )irr+ +\\t:++
t\+
URU
Twa +\+ r+\t
m
Neriqqaili utxt.itoi . . .
m
alwaziti
LU
utn.sn utxt
m
Lu-
pakki i
URU
Ukkiya it)iu alwaziti, the scribe, son of Lupakki of Ukkiya, inscribed
this tablet in Tawa in the presence of Prince Neriqqaili . . . BrTabl. iv 3043.
30.9. Far less frequent than fronting is extraposition, by which a constituent other
than a subject or direct object is moved to the right of the nite verb:
5
2
d
antaepu
arwani or-a We are holding two antaepa-deitiesof wood KBo 17.1 i 22; alkin
. . . ipiyatarra pda gimri ullui marmara andan (Telipinu went away), and he car-
ried o grain . . . and abundanceto the steppe, to the meadow, into the m. KUB 17.10
i 1012; utxt.r.oi uppi wtar par pzi itoi-i xtts.itoi-i The palace o-
cial holds out pure waterto the king and queen KBo 17.1 i 1415; numukan zr-:+
dankui daganzipi kattanta pnza apdani uddni peran My soul has gone down into
the Dark Netherworld because of that matter KBo 13.62:1011. Noun phrases that are
in apposition to the subject or direct object may also be extraposed, but the latter remain
in their regular position: i 4 so.ut pi i-naku xtts-naku and shall give four
personsmale or female KBo 6.3 i 2 (Laws 1, OHlNS). See also ta kalulupumu
gpinit ulliemi s+ +:tsi\i And I wrap their ngers with yarn(the ngers) of
their hands StBoT 8 iv 2627.
30.10. By left-dislocation a subject or object is moved outside its clause to the left.
Its extra-clausal position is shown by its appearance before a clause-initial conjunction
or another clause-initial constituent to which sentential clitics are attached (see Garrett
1990c: 26669 and 1994: 38): (Subject) ntn.so.xr-a widr
GI
xrnr
6
.nr. wllu nu
tuel warulate paigataru (var. [ pa-i-k]at-ta-ru) nu l tinnuzi The waters of the
mountains, the gardens, the meadow(s) let your refreshment keep going (singular),
and let (aima, the Freeze) not paralyze them VBoT 58 i 1011 (OHlNS);
d
ts:+u
urorn-tt+amu (As for) Is:+u, she is my goddess att. iv 74;
m
Pya
m
unupai 5
i.xr kattimi
URU
Talmaliya nuza linkiya takula uttar kattan +:++++pat dair
(As for) Piya, unupai (and) ve men with them of Talmaliya, they placed the word
of treaty under oath for themselves in the very same way KBo 8.35 ii 3031 (MHlMS);
(Direct Object)
m
uidudduwalli nan
URU
allana aaer (As for) uidudduwalli,
they settled him in allana HKM 113:1415 (MHlMS); (Indirect Object) 5 r.
xrsi numa r.xr taggata (As for) his ve brothers, for them he fashioned
houses KBo 3.1 ii 13 (OHlNS). As the preceding examples illustrate, whenever pos-
sible the left-dislocated constituent is resumed by an enclitic pronoun (so-called clitic
5. See McCone (1979), who does not, however, make the distinction between extraposition and
right-dislocation, referring in general to amplied sentences.
30.8 Clauses
00-HittiteGrammar.indb 408 6/5/08 4:20:40 PM
409
doubling).
6
The last two examples cited appear to show that left-dislocated objects
appear in the nominative case, but further study is needed to determine whether this is
a regular feature of left-dislocation.
30.11. By right-dislocation either a subject or object is moved outside the clause to
the right, following the nite verb (see Garrett 1990c: 25256). As in the case of left-
dislocation, there is clitic doubling of subjects of unaccusative intransitive verbs and
of direct objects. Right-dislocated direct objects appear in the accusative case: [(na)]n
l annatti memian Do not conceal it, the matter KUB 26.1 iv 14; nat kutruw aandu
d
tft s++I
d
tft
URU
Arinna Let them be witnesses, the Sungod of Heaven, the Sungod-
dess of Arinna BrTabl. iii 81 (Tud. IV); na ar t [(yat
d
Ku)]marbi He stood up,
(did) Kumarbi KUB 33.102 ii 17; kuitwara uezzi -a
D
tft-u [(xtn-e)a
LU
srr.
tut-a] Why is he comingthe Sungod, shepherd of the lands? KBo 26.58 iv 4445
(JCS 5:15859) (with dupl. KUB 33.104 iv 34). Right-dislocation is especially fre-
quent in translations of Hurrian literature (see McCone 1979: 470), where it is probably
used to mirror Hurrian word order,
7
but the rst two examples cited here show that this
usage appears also in native Hittite compositions.
8
Deletion Processes
30.12. Hittite shows various types of discourse-conditioned deletion or omission.
None has received systematic investigation, and the description here is limited to a few
illustrative examples.
Ellipsis
30.13. Ellipsis refers to the omission of any element of a sentence that can be recov-
ered from the context. It may, but need not be, identical to undeleted material. Examples
from Hittite: ar t\+ xtn
URU
xi.nnnn-ti ra nat t Is bloodshed in the land of
atti permissible? It is not! KUB 14.3 ii 8 (NH); itoi-i::+wamu k pedii pi
mnwa tma nuwa t uwami Give me the kingship here on the spot! If (you do) not
(give me the kingship here on the spot), I will not come! KUB 14.3 i 1415 (Taw.); sro
5

dn nukan anda stu-li Well-being is taken and (is put) into stu KUB 5.24 i 2122
(xr oracle); see sro
5
dn nukan urorn-tt+-ni karpi on-ri Well-being is taken and is
put with the deitys anger KUB 5.24 ii 1011; takku otu.xn-a li kuiki amenuzzi
o!\ itoi apparranzi If someone dispenses with a bulls pen (i.e., lets it run free), it
is a case for the king; they shall put (the bull) up for sale KBo 6.26 ii 2122 = Laws
6. The clitic doubling is impossible when the left-dislocated element is the subject of a transitive verb:
See 18.13 (p. 280).
7. In at least one case where the Hurrian archetype is preserved, the order of constituents is identical
( peiyanduyaan aliyanan
LU.ME
,'to:t+ Let them, (namely) the hunters, fell it, (namely) the deer KBo
32.14 ii 1314 (StBoT 32:113)). In most cases in the Hittite translations the Hurrian original is unavailable
to us.
8. Another example of right-dislocation in a native Hittite composition (in this case with fronting of the
verb) is: paiyaan
d
tft-st Also protect him, His Majesty KBo 5.3 i 16.
Clauses 30.13
00-HittiteGrammar.indb 409 6/5/08 4:20:42 PM
410
176 (OHlNS) (see LH 220); arnikzil (arnikzi) he shall make compensation
9
KBo
6.26 i 27 = Laws 163 (OHlNS).
Gapping
30.14. Gapping is the deletion of a repeated verb or verb plus other elements from
the second and any subsequent clauses in a sequence related by coordination. The phe-
nomenon is known from many languages, both ancient and modern. It is illustrated in
English by the following example: I will be giving some help to mother on Monday,
and Bill on Tuesday. In this case the elements gapped are both the verb and its
objects (will be giving some help to mother). Here are two examples from Hittite:
kuptarrakan [
GI
z.ix.o]n-a[z] [katt]a udanzi nanna
GI
z.ix.on-az katta
udanzi nan u-i [iun]nanzi The remnants they bring down out of the tent, and the
gure they bring down out of the tent , and they [thro]w it into the river KUB 12.58
iii 1416, ed. Goetze 1938: 1617; nu mn i-i akkanza nuan eliya anda s+
GI
t\ut
ix gulanzi mn xtts-zama akkan za nuan zz-a eliya anda ts:i
GI
t\ut
ix gulanzi And if it is a man who has died, they depict (his) gure out of fruits
in a grain pile, but if it is a woman who has died, they depict (her) gure with fruits
in a pile of wheat KUB 30.24 iii 3841, ed. Kassian, Korolv, and Sideltsev 2002:
39899, 410 (commentary).
Sentential Clitics
30.15. Hittite makes widespread use of sentential clitics, which are attached to the
rst word in a clause (the so-called Wackernagel position). These clitics include: (1) the
quotative particle -wa(r-), (2) the reexive particle -za, (3) the enclitic personal pro-
nouns, and (4) the local particles -an, -ap(a), -(a)ta, -kan, and -an. When they co-
occur, there is a specic sequence in which they appear (see Laroche 1958: 161, HE
288). This ordering is illustrated below, rst by means of a chart showing the slots
in the sequence and the clitics capable of lling each. Comments on the occupants of
each slot and various co-occurrence restrictions follow.
Host 1 2 3 4 5 6
Accented Word (+ -al-ya) -wa(r ) -na -a- (e.g., -mu -za -an
Accented Word (+ -al-ma) -ma -a, -an, -ta l-du -apa
nu -at, -e, -u) -eli -ata
u -kan
ta -an
30.16. The rst column contains the host on which the clitics of slots 16 depend. If
an accented word through fronting appears in initial position, it may take either of the
9. If one assumes ellipsis of a verb arnikzi. But one may also assume (with LH 131) an existential
sentence compensation (will exist), i.e., there will be compensation.
30.14 Clauses
00-HittiteGrammar.indb 410 6/5/08 4:20:43 PM
411
conjunctions -al-ma or -al-ya followed by the sentential clitics. If the sentence begins
with nu, u, or ta, however, neither -al-ma nor -al-ya can follow. Although -al-ma and
-al-ya are enclitics, they are not properly sentential ones. They owe their occasional
position before the chain of sentential clitics (slots 16) entirely to their dependence on
a fronted accented host word. This is not, however, the only position where they may
appear (see 29.26, p. 395, and 29.39, p. 400).
30.17. There is one situation, attested in two OH constructions, in which the enclitic
conjunction -al-ya and
10
follows rather than precedes the third-person singular dative-
locative clitic pronoun -e of slot 4. One sees it in the archaic expression parnaea
uwayezzi and he shall look to the house for it that concludes a number of laws (see
LH 29293) and in the equally archaic expression anda()eya and in addition to it
in Laws 110 (NH copy d
11
). Already in OS (Laws 128, copy aa) the expression was
altered to andayae, and in post-OH to andayai (70, 110, and 128; see LH
7879, 107 [with n. 346], 11617, and 201), but no modernized version of the former
expression (*parniyai) is attested.
30.18. Slot 1 contains the quotative particle -wa(r) that indicates direct speech. For
more on its form and use see 28.2. (pp. 354.).
30.19. Slots 2 to 4 are occupied by the enclitic personal pronouns, for whose form
see chapter 5 and for whose use chapter 18. The dative-accusative pronouns of the rst-
and second-person plural and the dative of the third-person plural appear in Slot 2, the
nominative and accusative third-person pronouns in Slot 3, and the dative-accusative
pronouns of the rst and second singular, as well as the dative pronouns of the third-
person singular, occur in Slot 4. The pronouns of Slots 2 and 4 are mutually exclusive
and never co-occur. Slot 2 contains the plurals; slot 4 the singulars (Honer 1986: 93
94). In late New Hittite the nominative-accusative pronoun of Slot 3 may be repeated
after the dative-locative pronouns (so far attested with -i, -na, and -(a)ma) or -za:
e.g., na-at-i-at, na-a-i (-ya)-a, na-an-za-an, na-at-za-at (Goetze 1930b: 1920; for
further examples see HW
2
I 41b sub -a2

to which add the rare examples with the plural
dative-locative: with -na in nuwarannaan anzel zr-ni piyawn KUB 13.4 iv 7273
(Honer 1986: 9394); with -(a)ma in nam-ma-at-a-ma-a-at-kn KUB 27.67 ii 36
with dupl. KBo 53.21 i 8). The sequence nanzaan occurs for the rst time in the
Deeds of uppiluliuma I, written by Murili II,
12
although it may have been introduced
10. Since in OH the expected form of this expression with -al-ma would have been *parnaema, it
is certain that we are dealing with -al-ya here.
11. Although found only in a NH copy, the very strangeness of this form and its exact correspondence
to the admittedly OH parnaea guarantee that it is an isolated survival of an OH formation antecedent to
andayae in Laws 128 in OS copy aa.
12. nanzaan in D frag. 15 F iv 30 and D frag. 26 KUB 34.23 ii 8, and D frag. 28 A KBo 5.6 iv
67 utxt:+wamu t.oi ln:+mawa nman dai nuwaranzaan
LU
+i:t:+ iyami. See in KUB
14.16 ii 15-16 (AM) maanmaanzaankan rorn-pa uun, which has the same author as D. It also
occurs in NH copies of OH and MH compositions: nanzaan in the NH manuscript KBo 6.3 (B) of Laws
23, where the OS copy A reads nanza, in nanzaan in KUB 1.16 ii 44 (OHlNS), and namaanzaan
Clauses 30.19
00-HittiteGrammar.indb 411 6/5/08 4:20:45 PM
412
by the later scribe who recopied the text. This double writing is a late conation of the
two congurations DAT-LOC-PL + Slot 3 and Slot 3 + DAT-LOC-SG. There are ex-
ceptional cases (but not many; see Honer 1986) where a Slot 2 plural pronoun occurs
in Slot 4. For example,
LU.ME
soryaatna paryaana pikanzi (text erroneously
pikiwani)
LU.ME
xtniurx-atna arrakanziyaana On the one hand, they are
our cupbearers, and they give them (-a, i.e., cups) to us. They are our cooks, and they
wash them (-a, i.e., dishes) for us KBo 32.15 ii 2729 (MHlMS); and tetwaratna
iyaweniwarat Tell it to us! We will do it! KUB 12.63 obv. 24 (OHlMS). Again:
kuwapit . . . uwatettani nanama ppa apiyakku pei Whither will you escort him,
that I should give him back to you there? KBo 8.35 ii 67 (MHlMS), and natama
[rorn-an ara]nta (The guards who hold spears) stand behind them IBoT 1.36 iii 46
(MHlMS). In all the above cited examples the plural pronoun, whether rst-, second-,
or third-person, is a dative (never accusative), and all examples are from texts originally
composed in the OH or MH periods. On the overall issue see Honer 1986: 9394.
30.20. Slot 5 contains the reexive particle -za, for which see 28.16. (pp. 357.).
For the late NH repetition of Slot 3 pronominal clitics between Slots 5 and 6 see above
in 30.19 (p. 411).
30.21. Slot 6 is occupied by the local particles -an, -ap(a), -(a)ta, -kan, and -an,
on whose use see chapter 28.
Nominal and To Be Sentences
30.22. If a clause has as its verb am, is, are, was, were, will be (see Cotticelli-
Kurras 1992a; 1992b), its predicate will be either (1) a noun or pronoun, (2) adjective,
(3) adverb, or (4) adverbial phrase. For example, the king is (1) a warrior, (2) brave,
(3) nearby, (4) in the palace. When the verb to be (-) would appear in the present
tense, it is often omitted, producing so-called nominal sentences (see 22.3, p. 306).
30.23. Adjective predicates of to be sentences (2) agree with the subject in num-
ber (singular, plural), gender (common or neuter), and case (nominative) (see above in
chapter 15).
30.24. Noun, pronoun, and numeral (the cited 1-edaz 1-edaz) predicates can be any
number or gender and any case except accusative, ergative, allative, or instrumental.
Examples of to be or nominal sentences with predicates in various cases: nuwaza
s+ 1-i\ n+t,t itoi-u I am king (nom.) of one fortress KUB 1.7 iii 38 (att.);
umeaikan takula but you are at peace (lit., of peace) with him KUB 26.40:98
(MHlMS) (gen.); kima tuel tu let this be yours (gen.) KBo 7.28 + KBo 8.92 17,
KUB 31.115 11 (OHlNS), and nammaanzaan in KUB 23.11 iii 14 (MHlNS). But none of these is in
OS or even MS; they are NH copies, subject to the same modernizing that we see in KBo 6.3 of Laws 23,
where the OH copy is extant.
30.20 Clauses
00-HittiteGrammar.indb 412 6/5/08 4:20:46 PM
413
etc. (OHlMS);
13
na apl it (scil., an irrigation ditch and its ow) is his (gen.) Laws
162 (OHlNS); tuqqa t kuitki zi also to you is nothing = you also have nothing
KUB 21.38 obv. 16 (dat.); rnr.xrwakan r.xtn.n.xr kui t\+
URU
Murmuriga
er What army (was) up in Murmuriga (d.-l.) KBo 5.6 ii 24 (NH); kma +w+:i
ME
t
kuit [ki ] 1-edaz 1-edaz ts:i xtn
URU
attiat These matters (i.e., treaty regulations) (are)
by no means reciprocal (lit., from [this] one, from [that] one (abl.)); they are (solely)
from the land of atti (abl.) KUB 21.1+ iii 7677 (Mur. II).
30.25. Sentences where the verb to be predicates the existence of the subject may
consist merely of the subject and the verb, but the verb may take a complement: mn
txt utxt.rfma t zi but if there is no male progeny KBo 4.10 obv. 12; tnt
nXu rorn-pa appanna +sui t.oi kuiki ta there was no fortied city, no place
of retreat KBo 5.8 ii 2425; ru.xtn
4
.n pariyawanzi t.oi there are no loaves
for breaking KUB 12.12 v 32. The verb to be may be omitted in the present tense: t
otu-u tmawa tut-u there were no cattle, there were no sheep (historical pres-
ent) KUB 36.51 obv. 7 (OHlNS); tnt-a tx-anet
URU
udul . . . nukan er i-a
m
Appu si+-anet There was a city, its name was udul . . . and there was a man up
there, his name was Appu KUB 24.8 + KUB 36.60 i 710.
14
30.26. The adverb kalkma may be employed (always clause-initial), if it is in-
tended to convey the notion that the equation of subject and predicate has just taken
eect (performative use; see 7.21, p. 148; 22.15, p. 309; 24.7, p. 319; 24.27
24.30, pp. 323324) or that the situation described is presented vividly: nuza ka
ume i.xr xtn
URU
Imirika mante t: [:t
d
tf]t-[st l]i [n] kiyaa All you men of
Imirika are hereby (men) of his oath (gen.) with respect to His Majesty KUB 23.68
+ ABoT 58 rev. 7 (MHlNS); kaza
URU
attui ++n+u
LU.ME
:+rrt\t I am now in
attua in the company of our colleagues (d.-l.) HKM 36 left edge 3 (MHlMS).
30.27. If the subject is either rst- or second-person (I, we, you), the reexive
particle -za (28.16., pp. 357.) must be employed in Middle and New Hittite (see
examples in the previous paragraph and 28.32., pp. 362.). The particle is not used
in this function in Old Hittite.
30.28. It is normal for the subject to precede the predicate, whether that predicate
be a noun, accented pronoun, adjective, or adverb(-phrase): see the examples in 30.24
(p. 412). For an example of a predicate adjective see kma uttar nakk This matter,
however, is important KUB 7.1 iii 27. However, nominal sentences and those with to
be are subject to the same movement rules as other clauses (see 30.4., pp. 406.):
(fronting) ammel kpat 1-a dammea Mine is only this one punishment KBo 4.8
ii 1213; kya t\+ rx+ man sro
5
-in also here in your house all is well HKM 2 left
13. There is disagreement on the reading of the rst sign: as xr or srirx. See Singer 2002b: 27 n. 2.
14. It seems more likely to take tnt-a as nominative as understood here and translated by Honer
(1998b: 83 2; likewise Siegelova 1971: 5), although it is also possible to take it as genitive in a split geni-
tive construction: the citys name was udul (for which see 16.38, p. 251).
Clauses 30.28
00-HittiteGrammar.indb 413 6/5/08 4:20:48 PM
414
edge 12 (MHlMS); (left-dislocation)
d
ts:+u urorn-tt+amu (As for) Is:+u, she is
my goddess. Examples of right-dislocation appear to be limited to accented pronomi-
nal subjects. In the following example the rst and last clauses show right-dislocation
and fronting, the second merely fronting: parkuia ap miriwanza ap arkia
ap She is pure, that one. Lovely is that one. She is white, that one KBo 4.6 obv.
1314. The only example with a right-dislocated noun subject (also with fronting of
the predicate adjective) is from a translation of a Hurrian sentence and may only mirror
Hurrian word order (30.11, p. 409): dauwara alluwai strong is it, the strife
KUB 33.96 iv 10 (Ullik. I).
30.29. Old Hittite appears to show a dierent pattern in nominal sentences with
accented pronominal subjects, where the regular word order is predicate + subject: it
utxt:+ ap KUB 1.16 ii 14 (OHlNS), (but nattaa utxt:+ is also permitted);
maranzawa zik you are treacherous; kinunawakan ara namma l nyanza zik
Now do not any longer be turned away! KBo 20.82 ii 1819 (pre-NH); andnza
maniya[]aya i zi[k] you are a just lord of government KUB 31.127+ i 20 (OHl
NS); dankuwaya xtn-e[a] atta anna zi[k] you are father and mother of the dark
land(s) KUB 31.127+ i 21 (OHlNS); annena i zik you are a lord of judgment
KUB 31.127+ i 24 (OHlNS).
15
See further examples in 26.4 (p. 342), 28.36 (p. 363),
and 28.41 (p. 364). On this as a characteristic of OH see Honer 1987: 280, 285. The
reverse sequence (accented personal pronoun subject followed by the predicate) is found
when there are two virtually identical adjacent clauses with contrasted pronoun subjects:
zikkawa
GI
ftxti apaa
GI
ftxti both you are a ftxti-man, and he is a ftxti-
man KBo 22.1:20 (OS); attami uga utxt-ati lit (You) be my father, and I
shall be your son KUB 26.35:6 (OS); takku otu

audiza natta o[tu
.
xn]-a . . . takku
otu tyuga apa otu.xn-a If it is an unweaned calf, it is not a bull . . . (but) if it is
a two-year-old bovine, that (contrastive!) is a bull KBo 6.2 iii 2324 (OS).
30.30. In nominal sentences with non-referential subjects the third-person clitic sub-
ject pronoun is usually omitted, and if no conjunction is used, the result is sometimes a
one-word nominative case predicate (18.17, p. 282; and 30.13, p. 409).
Dependent Clauses
30.31. Most Hittite dependent clauses are formally marked by various subordinat-
ing conjunctions, never by word order or verbal mood (as in English Were I rich . . .
or German Wre ich . . .). For occasional unmarked conditional clauses see 30.56
(p. 423). Sentences may be composed either of one or more coordinate independent
clauses, or by the combination of one or more dependent clauses and an independent
one. The apodosis can be implied but unexpressed, such as may the gods destroy mel
15. The large number of examples in CTH 733 (KUB 8.41 and similar texts) such as dandukini taat
tanuiti zi [k] KUB 8.41 ii 2 (OS) do not add support, since the dandukini is fronted for contrastive purposes
versus urorn.xr-na itarna.
30.29 Clauses
00-HittiteGrammar.indb 414 6/5/08 4:20:49 PM
415
uslyoulhimlthem (see CHD mn 7e). Example: namakan mn
d
tft-st kuedani anda
idlu itamati natmukan mn annatti natmu t mematti apnnamu antuan
t tekkuanui nan anda imma munni Or if you hear evil about My Majesty from
someone, (if) you keep it from me and dont tell me and do not reveal that person to
me and even hide him, (implied: the gods will punish you) KBo 5.3 i 2730 (MHlNS).
Since this type of suppressed apodosis is conned to divine punishments, it is likely
that the suppression was prompted by a fear of mentioning the punishments. See HKM
34:1317 (HBM, pp. 18083) and CHD LN mn 7e2 for further examples.
30.32. In the terminology of traditional grammar one can identify the following
types of dependent clauses in Hittite: (1) temporal (when, while, or until), (2) causal
(because), (3) concessive (although), (4) conditional (if), (5) relative (who or
which), (6) indirect statements or questions (that, how, or whether).
30.33. The manner of constructing such clause types varies over time in the Hittite
textual record. The details have not yet been worked out for many of these syntactic
structures.
30.34. Each of the six types of dependent clause mentioned above in 30.32 is
marked by a characteristic subordinating conjunction. The following table summarizes
the Hittite subordinating conjunctions.
Clause Type Old Hittite New Hittite
1. temporal when; until mn; kuitman maan, kuwapi ; kuitman
2. causal because kuit kuit
3. concessive although mn mn
4. conditional if takku, (rarely mn) mn
5. relative who, which kui- kui-
6. indirect statement that kuit, maan, mn
A seventh type frequently found in other languages, namely nal or result clause, is not
a dependent clause in Hittite, but rather a coordinate independent one (HE 322), or the
same meaning is conveyed by means of an innitive (see 25.1025.36, pp. 332338).
Compare the following example: na t tarnaun nankan t kuenner I did not
allow them; and they did not kill him > I did not allow that they should kill him KBo
6.29 ii 2728. Or this example from the laws: [takku i-a]n nama xtts-an itt++
walzi kui[k]i na aki [If ] someone strikes a free [man] or woman, so that he dies
KBo 6.3 i 6 (Laws 3, OHlNS). The suggestion of Stefanini (1988: 255), followed by
Pecchioli Daddi and Polvani (1990: 50 n. 8) and Garca Trabazo (2002: 85 n. 26), that
in KBo 3.7 i 67 one should assume a use of mn introducing a dependent clause to
express purpose seems unlikely, because no other example of this alleged use has ever
been found.
Clauses 30.34
00-HittiteGrammar.indb 415 6/5/08 4:20:50 PM
416
30.35. A complex sentence containing a preposed subordinate clause can itself
naturally be linked to its preceding context by the connective nu: nu +ui:+ maan
nannai nukan edani pangawi i.xn 1-ankipat anda andaizzi And as my father was
driving, he encountered that whole enemy all at once KBo 40.293 + KBo 14.3 (D
frag. 15 F) iv 2831; or by the more contrastive -ma: nammaa rorn-pa gimmandari
ya [(uwanz)]i
URU
attui uet maanmazaka[n (rzr
4
xt.xx-:t )] karpta na t\+
xtn
URU
Itaara pait Then he came back home to Hattua to spend the winter; but
when he had nished the annual festival, he went to Itaara KBo 5.6 (D frag. 28 A)
i 4042.
Temporal Clauses
30.36. Temporal clauses in Old Hittite (OH) are frequently marked by mn when,
which in OH was only beginning to gain the secondary meaning if that it acquired in
NH (see CHD mn, and see 30.47., pp. 419.). Post clause-initial kuwapi when
rst appears post-OH. But since in OH mn had uses other than when for example,
postpositional like, just as, interrogative how?, and indirect question whether
(CHD mn mngs. 14), the temporal use when was marked by a consistent fronted
position in the clause. Many examples can be found in the CHD mn article sub mng. 5.
Here we cite only a few examples: mn utxt.xr
URU
atti i.xr ttxt ur +\+ +ut
itoi aruwa[nzi] nu tarikanzi when Hittite men subject to ttxi-tax came, they did
reverence to the father of the king and said KBo 6.2 + KBo 22.62 iii 1617 (Laws 55,
OS); mn
URU
Tama[rmara] arer nu tarikanzi when they arrived at Tamarmara, they
said
16
KBo 22.2 obv. 8 (OS); mna alletama nuza utxt.xtts
d
Illuyanka
ux-anni d when he grew up, he took the daughter of the serpent as his wife KBo
3.7 iii 68 (OHlNS).
30.37. Other temporal clauses meaning while or until are marked by kuitman (see
19.6, p. 290; 26.12, p. 343; 30.34, p. 415; and 24.3, p. 317). Until clauses follow
the main clause asyndetically: nu r-rii annikezzi kuitmna lazziatta and he shall
work on his estate until (the injured man) recovers KBo 6.2 i 1718 (Laws 10, OS);
see also KBo 6.3 iv 68 (Laws 79, OHlNS). While clauses usually precede the main
clause. They are connected to what precedes by nu or another conjunction, unless they
are the rst sentence in a discourse (see 29.4, p. 390; see HKM 10:2428 cited below
for an example beginning a discourse). Although the following main clause is usually
asyndetic (e.g., nuza kuitman nwa ftn-a eun s+
KU
x.fn.rza eun While
I was still a child, I was a bridle-holder KUB 1.2 i 11 = att. i 12), it is sometimes
introduced by nu or other conjunction (-a in HKM 10:24:2428 below): nu kuitman
+ui:+ t\+ xtn
URU
Mittanni ta nu i.xn
URU
Arauwanna kui xtn
URU
Kiiy oti-
annekit nat mekki tamaan arta and while my father was in the land of Mittanni,
the Arawannaean enemy, who was attacking the land of Kiiya, had greatly oppressed
it KBo 3.4 iii 4749 (AM); kuitmanwamu
d
tft-st k s+ i.xr
URU
Kaka takula
16. Present tense verbs here describe past acts; see 22.6 (p. 307).
30.35 Clauses
00-HittiteGrammar.indb 416 6/5/08 4:20:50 PM
417
uttar atri ammug awa memian t\+ xtn
URU
Iupitta ukemi While you, Your
Majesty, send me this word about making peace with the Kaka people, I am awaiting
word in the country of Iupitta HKM 10:2428 (MHlMS); parma xt.xx.nr.-
a kuitman r:+
m
rn.oi-i t\+ xtn
URU
atti ta nu xtn
URU
Gaga
I.A
mante
kururiyaer But further, in the years while my brother Muwattalli was in the land of
atti, all the Kaskean lands become hostile att. ii 1618; kuitman in copies B and E,
kueda in copy A). Several kuitman until clauses in a row are asyndetic, when they
are synonymous, merely expressing the same idea in dierent terms (see LH 16768):
This ocial will remain where he is kuitmana uezzi kuitmana apiya rorn-pa uezzi
until (the person whose safety is guaranteed) comes (to me) and returns there KUB
14.3 ii 7172 (= Sommer 1932 ii 7273, NH); on apiya here see 7.18, p. 147). Before
is expressed by kuitman . . . nawi (see 26.12, p. 343). In a few cases kuitman occurs in
main clauses with a meaning something like in the meantime: (If someone whispers
to you His Majesty is plotting thus and so against you) zik[m]a apn memiyan +\+
d
tftst kuitman atri In the meantime write of that matter to His Majesty KBo 5.4
obv. 2930 (Mur. II).
30.38. The regular when word in NH is maan (also written orx-an). It usually
appears clause-initially but may also be preceded by nu (with any clitics) or by a fronted
constituent (see examples below). The most common pattern is temporal clause followed
by a nu-introduced main clause: maanma +ut +ut:+ apz rorn-[ pa u]et na t\+ xtn
URU
ayaa pait But when my grandfather came back from there, he went to the land of
ayaa KUB 19.11 iv 4042 (D); [nukan] maan +ui:+ X xtn-:t ra [nuka]n
i.xn
URU
Gaga kui t\+ X xtn
URU
atti uwanza ta nu xtn-e mekki idalawata
[And] when my father arrived in the land, the enemy Kakaean who had come into
the land of atti had greatly ravaged the land KBo 14.3 iii 1214 (D); maanma
tnt.uruir.nr. wetummanzi zinnit na
URU
Almina andan pait But when he nished
fortifying (lit., building) the cities, he went to Almina KBo 5.6 i 57 (D); maanma
ameanza kiat But when it became spring KUB 14.15 i 23; with several when
clauses in a row:
d
tft-stma maan iyaat nu maan +\+
URU
allapa [aru]n But
when I was marching, and when I [reached] allapa KUB 14.15 ii 7 (AM).
30.39. In New Hittite one also nds temporal use of kuwapi in the second (Wac ker-
nagel) position as when (19.6, p. 290): apeneuwantanma memian kuwapi [(ita
mati )] nuzakan memiyani (var. memiyanni ) er l k [(aruiyai)] But when you hear
of such a matter, do not be silent about the matter KUB 21.1 + KUB 19.6 + KBo 19.73
+ KBo 19.73a ii 8283 (Alak.); +ui:+ kuwapi
m
attuili +\+
m
Uriteupa utxt
m
Mwattalli menaanda kururiata nan itoi-eznani ara tittanut When my father
attuili became hostile towards Uriteub, son of Muwattalli, and he removed him
from the kingship BrTabl. i 68; nu )irri utxtt:t kuwapi ir apnmaza xtts-an
md
ixx-a +\+ r+\t +ut:+ dattapat nawi And when they made the treaty tablet, Ku-
runta had not yet taken (in marriage) that woman in the presence of my father BrTabl.
ii 8687; andamakan tut-un kuwpi kuewn nu linkiya kattan kian daiwen But in
Clauses 30.39
00-HittiteGrammar.indb 417 6/5/08 4:20:51 PM
418
addition when we had killed a sheep (as part of the treaty-making ceremony), we placed
the following obligations (lit., placed as follows) under oath KBo 16.47:1516 (treaty);
see 25.15 (p. 333) for KBo 15.60 vi 311.
30.40. Instead of a single temporal adverb, a form of the relative kui- modifying a
noun of time (hour, day, year, night, time) may occur: kakan k tuppi kuedani tu-ti
par neun nata rnr[].xr xtn tot apdani tu-ti ara uittiyanun numuan
ziqqa xnpat dk arnut On what day I have dispatched this tablet, on that day I
have drawn forth troops of the Upper Land, and you too must promptly bring troops to
me HKM 71:2431 (MHlMS); nata +\+ i.xr
URU
Pauwa kuedani tu-ti kurura
memian anda itamateni nu apedani [tu-ti] rten On what day you hear a word of hos-
tility among the men of Pauwa, on that [day] get there, (and strike Pauwa!) KUB
23.72 + rev. 2728; nui kattan rorn-pa kuedani muni run nXu-earma s+ or-
, 40 gipear katta uet at what time I caught up with him, a fortication wall (made)
of wood (measuring) 40 gipear (ca. 20 m) came down KBo 6.29 ii 3133; pzi[ma
kued ]ani muni nu namma ara it tarni at what(ever) time (an angry deity) seizes,
he doesnt ever let go KUB 13.4 ii 2324; nu
d
s!\-a kueda[ni] or
6
-ti akiyazi nu
apdani or
6
-ti t kuitki ienzi And on what night the Moon gives a sign, on that night
they do nothing (i.e., all normal activities are suspended) KUB 17.28 iii 2123.
Causal Clauses
30.41. Causal clauses regularly precede the main clause and contain the word kuit
because in the second position (HE 323). In a sequence of two consecutive causal
clauses, kuit is optional in the second clause (30.73, p. 428).
30.42. In second (Wackernagel ) position: xtn
URU
Kizzuwatnaya kuit antezzi
auri And because the land of Kizzuwatna is the rst border watch point HKM 74:12
14 (MHlMS); numu
d
ts:+u o:+ kuit kanian arta r:+yamu
m
rn.oi u
arta Because Is:+u, my lady, had honored me, and my brother Muwatalli held me
in favor KUB 1.1+ i 2830 (Hatt. III); nuza
d
Upelluri kuit or
6
-ya xr-a xr.nu-a
nuza ai nu[t]tariyan urorn-tt+-in t akti Is it because you are remote from (see
16.49, p. 254) the Dark Earth, O Upelluri, that you do not know that upstart deity?
KUB 33.106 iii 3839 (Ullik. III); xtn-ema man kuit ts:i i.xn [dann(attaan)]
ta Because the whole land had been depopulated by the enemy KUB 19.11 iv 1213
(D); and
URU
Gagama manza kuit takl ta Because the entire Kaka (group)
was at peace (with us) KBo 5.6 i 14 (D).
30.43. Later than the second position: alkima m a apiya aniyanza kuit nu
rorn-an tiyatten nan anda pten nankan rso-i anda iitten because grain has
been sown for you there (see 7.18, p. 147), get behind (the matter): gather it in and
store it in a storage pit! HKM 18 left edge 25 (MHlMS); [z]igazza
m
ulli (37)
[+]+[n]+u
d
tft-st kuit ta But because you, uilli, were with His Majesty HKM
55:3637 (MHlMS); xtn
URU
Neriggamaz i.xr
URU
Gaga kuit dn arkanzi But
30.40 Clauses
00-HittiteGrammar.indb 418 6/5/08 4:20:51 PM
419
because the Kaka-men have taken for themselves the land of Nerik KUB 17.21 iv
67 (MHlMS).
30.44. Note that in the rst example cited in 30.43 kuit is preceded by three constit-
uents (subject alki, adverb apiya, and predicative participle aniyanza). The following
example appears to show four constituents before kuit: ammukma itoi-i::+
d
ts:+u
o:+ annianpat kuit memiket But because Is:+u, My Lady, had previously
promised me the kingship . . . att. iv 78.
30.45. Either the causal clause or its result clause or both may be asyndetic, but
in NH usually both are introduced by nu: numu
d
ts:+u o:+ kuit kanian arta
r:+yamu
m
rn.oi au arta numukan orx-an t.xr-anna n za s+
d
ts:+u
o:+ kaniuwar s+ r:+ya [a]ulan auer numu araniyer Because Is:+u,
my lady, had recognized me, and my brother Muwatalli held me in favor, when (mean-
spirited) men saw the recognition of Is:+u and the favor of my brother, (nu) they envied
me KUB 1.1 i 2832 (Hatt. III); na katta aanna kuit srs-at nan katta aaun
And because she (i.e., the Tawananna) was indicated by oracle to be removed (from
her oce) (lit., made to sit lower), I removed her KBo 4.8 ii 67 (prayer, Mur. II).
When the causal clause follows the main clause, it is introduced asyndetically: +\+ r+\t
X[x+]waz dukiketta i-natarmitwa kuit pata Are you rejoicing because you
have swallowed my manhood? KUB 33.120 i 2829 (myth, MHlNS).
Concessive Clauses
30.46. Concessive clauses (although, even if) can be introduced with mn . . . -(y)a
or man mn . . . -(y)a even if:
m
Uri-
d
U-upamamu mn nti-lua ta ammukma
. . . tpat karuiyanun But although Uri-Teup was hostile toward me, yet I was
not silent KBo 4.12 i 2426; [nu]za mn irmalanzaa ta
d
tft-stmatta [+\+]
+s+u +utx+ tittanununpat Although you were sickly, nevertheless (-ma . . . -pat) I, My
Majesty, installed you in the place of your father KBo 5.9 i 1618. When the apodosis
expresses a hypothetical situation, both protasis and apodosis are introduced by potential
man (irrealis): manza t manga wadula ta [m]antakkan r +utx+ xtnx+ya
t ara dir manat damdani kuedanikki piyr Even if you had been in no way a party
to the crime, couldnt they have taken away from you the house of your father and your
land and given it to someone else? KBo 5.13 i 1920 (see CHD man b 2 [p. 142]); see
manmu [mn xsxi.xr pera]n arpuwantea eer a[mmukmankan] [t\+]
UR.
SAG
Teina arpat pu[n] Even [if the roads before] me had been impassable, I would
still (-pat) have ascended Mt. Teina KUB 19.37 iii 5153 (AM 176179.).
Conditional Clauses
30.47. A conditional clause expresses the premise that must hold if the action or
state of the main clause is to be fullled. Conditions may be fulllable or non-fulllable.
Clauses 30.47
00-HittiteGrammar.indb 419 6/5/08 4:20:52 PM
420
Simple fulllable conditions may refer to past, present, or future time.
17
Contrary-to-
fact conditions are based on a premise that is known to be false and are thus unfulllable.
Simple conditions are introduced either by takku (in Old Hittite) or mn (in post-Old
Hittite), both meaning if (see 30.48 on the diachrony). Contrary-to-fact conditional
clauses (CHD man b 2) are introduced by the potential particle man (usually written
ma-an or ma-n
o
) and mn if. Often these two words are blended as ma-am-ma-a-an
(from *man+mn).
30.48. Preliminary study indicates that in the case of conditional sentences the man-
ner of connecting dependent if clause (protasis) with result clause (apodosis) was
signicantly altered between Old and New Hittite. But even in OH the method seems
to have varied. In the OH Hittite laws and elsewhere
18
the protasis was introduced by
takku if, and the apodosis followed without introductory conjunctive particle (Fried-
rich 1959: 88; LH 12). That is, the main clause was joined asyndetically to the preced-
ing dependent clause. This type is also found in other passages in OHlMS and OHlNS
texts (e.g., takku ln.xr-amman (see 1.118, p. 43) itarna si+s[i\i] kuiki tezzi ln
mi l If among my servants someone speaks their name, let him not be my servant
KBo 3.27 obv. 1011, see also KBo 3.28:12, VBoT 58 i 4041). These include OHlNS
omen texts showing the sequence of takku . . . (asyndeton) . . . (e.g., takkuza riio
tut-a tetan kia apedani t-i kallarezi If the kidney (in extispicy) resembles (lit.,
becomes) a sheeps tetan, it will be unlucky for that person KUB 4.1 iv 29).
19
In
post-OH the if word changed to mn, but the asyndetic introduction of the apodosis re-
mained in isolated examples in MHlMS and even in NH: mn tma uwai ar (a)ki
But if not (i.e., if you do not bring the prisoners), you will surely die (see 24.31.,
pp. 324.) HKM 35:89 (MHlMS), mnat [i]yazi lat kuiki akki If he does it, let
no one know it KUB 21.17 iii 3132 (Hatt. III). In some Old Hittite rituals, however,
a dierent pattern is found: the conditional clause is introduced by mn or takku, and
the apodosis by nu (so Otten and Souek 1969: 9192). In MH and NH compositions
(not copies of OH texts made by post-OH scribes), on the other hand, nu almost always
introduced the apodosis following a protasis containing mn if.
Simple Conditions
30.49. By far the most common type of conditional clause in Hittite is the simple
future (open or fulllable) condition, attested especially in the Laws, in omen texts, in
instructions, and in treaties. In Old Hittite: takku i.t
19
.it-a i-a nama xtts-a
takya tnt-ri aki kuela ari aki 1 +i
GI
gipear .X kar (a)iyezzi nanza di
17. Conditions may also be either particular (If the king sendslwill send 1,000 troops, we will defeat
the enemy) or general (If anyone violateslshall violate the oath, he will die), but this distinction does not
appear to be relevant for Hittite grammar.
18. takku ume natta akteni kni
LU
t.or-ea t.oi numa memai +w': +ut:+ If you do not
know, is there here not even an old man, that he may tell you the word of my father? KBo 22.1 56 (OS).
19. Many of the above observations on asyndesis in Hittite texts were made by Honer (1996c).
30.48 Clauses
00-HittiteGrammar.indb 420 6/5/08 4:20:53 PM
421
If a person, man or woman, is everlshall be killed in another(?) city, (the victims heir)
shall deduct three acres from the land of the person on whose property the person was
killed and shall take it for himself KBo 6.2 i 78 (Laws 6, OS); note that the apodoses
clauses contain present-future tense verbs (kar (a)iyezzi and di) and are introduced
asyndetically and by nu.
30.50. Other examples from OH texts copied by NH scribes (OHlNS): takku utxt.
itoi antezzi t.oi nu kui tn pda utxt-ui nu itoi-u ap kiaru If there
shall not be a rst-rank prince, then let a prince of second rank become king KBo 3.1 ii
3638 (OHlNS); takku utxt.itoima watai nu so.ut-azpat arnikdu If a prince
sins (i.e., attempts homicide), let him pay only with his own person (not with those of
his extended family) KBo 3.1 ii 5556 (OHlNS); takku xtts-[za ]i nu annazpat
X-az [a] ara i nu memai
d
t-a xtn-e zi If a woman gives birth, and right
from the mothers womb it (the infant) opens its mouth and speaks, the Stormgod will
strike the land KBo 6.25+13.35 iii 57 (birth omen).
30.51. Simple future conditions with mn if: mn t\+ tu.12.xx
d
st\-a aki
xtn-e anda ntnt
5
ari If the Moongod ever dies (i.e., is eclipsed) on the twelfth day,
a swarm of locusts will arise in the land KUB 8.1 iii 3 (lunar omen); mnwa [i]
URU
Aur uezzi nuwaran za[i]yatten If the Assyrian comes, ght him KUB 14.16
i 16 (AM 2627); mnwarn tma ptni nuwaranmu par t ptni nuwa
uwami nuwama +oi xtnxi\i ara arnikmi If you do not seize him (the fugi-
tive) and do not give him over to me, I will come and destroy you together with your
country KBo 5.13 i 79 (Treaty with Kupanta-Kurunta). Here, as often, the force of
mn if carries over into an additional protasis introduced by nu.
30.52. Hittite also has the equivalent of what are traditionally termed less vivid
future conditions, marked with man to express greater uncertainty. Compare the Greek
construction with optatives in both protasis and apodosis and the irrealis particle in
the apodosis (Goodwin 1965: 301 1408). With man in both the protasis and apodosis
(for the omission of if see 30.56, p. 423): mankan
d
tft-st uitt:+ uiti kuinki par
naitti man xtn-i i.xn t dammiaizzi (If) you, Your Majesty, were to send a lord
(to lead an army), the enemy would not oppress the land HKM 46:1517 (MHlMS).
By his use of the irrealis man particle the speaker is not excluding the possibility of
the kings sending an army; instead he appears to be making a greater concession to
the kings freedom to choose: If you were to send . . . , this would be helpful, but you
may have good reasons for not doing so. With the rst man-irrealis written plene (see
CHD man b 2 c): mnwamu 1-an utxt-x+ paiti manwaramu
LU
+i:t:+ kiari
If you were to give me one of your sons, he would become my husband KBo 5.6 iii
1213 (D).
20
20. We cannot accept Puhvels interpretation of this example as an optative If you give me a son of
yours, may he become my husband! (HED M 37). Indeed we explicitly deny his claim that there is insuf-
cient evidence to sustain a potential category with man and present tense . . . or a subject-optative . . .
Clauses 30.52
00-HittiteGrammar.indb 421 6/5/08 4:20:53 PM
422
30.53. Examples of a simple present condition: nuan mn alki arante na
kan ara waraten If the crops are ripe, harvest them! HKM 25 1517 (MHlMS); na
mn kar pnza If he islhas already gone . . . HKM 66 2324 (MHlMS); (Does my
brother then have nothing?) mn +\+ utxt
d
tft nama utxt
d
t t kuitki zi nama
aruni t zi tuqqa t kuitki zi If the son of the Sungod or the son of the Stormgod
has nothing, or the sea has nothing, (then) you also have nothing! KUB 21.38 obv.
1516 (Letter of Puduepa to Ramses II). For an additional OH example see p. 420,
n. 18. An example of a simple past condition is: ru-anza wemiyanun nanza
+nt:t:+ natta kuwapikki edun wtarmaz wemiyanun nat +nt:t:+ t kuwapikki
ekun (if ever) I found bread, I did not eat it secretly; (if ever) I found water, I did not
drink it secretly KUB 30.10 obv. 1617. The following example of a simple past con-
dition employs the historical present (see 22.6, p. 307): mnmu itarakzi kuwapi
nuzakan irmalapat s+ urorn-tt+ andandatar er ukenun If I ever became ill,
even (while) ill I experienced the divine power of the deity att. i 4445.
Contrary-to-Fact Conditions
30.54. Unfulllable (contrary-to-fact) conditions or undesirable future situations
(23.15, p. 316) are marked with man or mn + man in the protasis and man in the
apodosis. One attested example of a present contrary-to-fact condition shows a pret-
erite tense verb in the rst protasis, but present tense in the second protasis and in
the apodosis: mmman (= mn man) dandukinaa utxt-a uktri uiwanza ta
manata mn antuwaa idluwa inan arta manati natta kattawatar If a mortal
were to live forever, and the evil sickness of man were to remain (lit., stand), would that
not be a grievance for him? KUB 30.10 obv. 2223.
21
Another has a preterite verb
in the protasis and the rst two clauses of the apodosis, but present verbs in the last
two clauses of the apodosis: ma-a-am[-ma-an-a-ma-a?] (i.e., *mnman (a)ma)
22

tayazzil piker man mantepat marr [man]e
LU.ME
.zt kantati kman kn
pzi [k]aman kn pzi If they [were] to charge [them] with theft, all of them might
dissemble, they [mig]ht become thieves; this one might seize that one, and that one
might seize this one KBo 6.2 ii 5355 (Laws 49, OS). A third has a preterite verb in
the protasis and a present verb in the apodosis: mammanzakan kuiki r-er tamai
arnut man zik t araniee If someone else were to conscate your house, would you
not be upset? ABoT 65 rev. 56 (MHlMS). The generalization appears to be that the
construction must begin with a preterite and end with a present tense, but the rationale
for the distribution of the tenses of any intermediate verbs remains unclear.
alleged examples have found better integration in the optative and contrary-to-fact constructions discussed
above (HED M 39).
21. The parallel text KUB 36.79 ii 50 has artariya in the second apodosis, conrming the present tense.
The omission of the verb be in the apodosis shows that it is in the present tense (see 22.3, p. 306).
22. See LH 5960, 19192.
30.53 Clauses
00-HittiteGrammar.indb 422 6/5/08 4:20:54 PM
423
30.55. Past contrary-to-fact conditions show consistently preterite verbs in both
the protasis and apodosis. Examples: mankan mn +\+
m
Attariya uiwetenna
ktitaman kten Even if you had survived Attariya, you would have died of hun-
ger KUB 14.1 obv. 12 (MHlMS); numa
LU
auriyalu kuit arantat mnkan mn
+\+
m
Pittaggatallipat warpa teun manmu
LU
auriyalu (var.
LU
auriyatallu) kuit s+
m
Pitaggatalli aur manmu t duuiyait And because sentries were in place, if I had
tried to surround Pittaggatalli himself, because the sentries of Pittaggatalli would have
seen me, he would not have waited for me KBo 5.8 iii 1418 (NH). As the second
example shows, when the potential-irrealis man is employed together with conditional
clauses, no clause connective (nu or -ma) is used. For further examples of past contrary-
to-fact conditions see CHD man b 2 c.
30.56. Sometimes mn if is unexpressed in conditional clauses: t\+ rft.12.xx
utxt-a miyari ap utxt-a
LU
t.or-ezi (if) a child is born in the twelfth month,
that child will live to an advanced age KUB 8.35 i 9 (menology); wadul kulqa autti . . .
nuza pankun rorn-pa punuke (if) you see a crime, always consult the panku KUB
1.16 iii 5961 (OHlNS); ru-anza wemiyanun nanza +nt:t:+ natta kuwapikki
edun wtarmaz wemiya nun nat +nt:t:+ t kuwapikki ekun (if ever) I found
bread, I did not eat it secretly; (if ever) I found water, I did not drink it secretly KUB
30.10 obv. 1617. A similar use with the present-future tense (both in real future condi-
tions and past contrary-to-fact conditions) exists in informal USA sports-TV jargon: He
catches that ball, and the game is over.
30.57. Often, in sequences of if clauses, instead of takku or mn, the word nama
or (if) introduces the conditional clause without mn: nama i.xn oti-azi nu p
arzi zikma peran ar t wa[(rri)]atti nu i.xn t zaiyai or (if) the enemy
attacks, and holds (the acquired position), but you dont come to assist in advance, and
you dont ght the enemy KUB 21.1 iii 5052 (NH). On occasions, however, the more
complete nama mn is employed: nama mn xtn-:i+ kuitki zaiyaza itoi xtn
URU
atti anda atkinuzzi of if the King of atti besieges some land in battle KBo 5.9
iii 2324 (NH); nama mn +\+ ln-ot kuedanikki watul watulmazzakan +\+ r+\t
rsi tarni or if some slave sins, he shall confess his sin before his master KUB
14.8 rev. 2526 (NH); nama mn uiti kuinki [(kz)] [(xtn-az)] laiyauwanzi uiyami
or if from this land I send out some lord to campaign KUB 21.1 iii 78; nama mn
urorn-tt+-ni kuedani rzr
4
o zi or if for some deity there is a Festival of Milk
KUB 13.4 iv 41 (pre-NH).
Relative Clauses
30.58. The most common type of dependent clause in Hittite is the relative clause
(see chapter 8 for the relative pronouns). There are two fundamental dierences between
relative clauses in Hittite and those in languages such as English. First, since relative
clauses in English must immediately follow the noun they describe or modify, they are
Clauses 30.58
00-HittiteGrammar.indb 423 6/5/08 4:20:55 PM
424
most often inserted into the main clause: The horse that he rode was strong or The
man who (m) I saw walked away, especially when the modied noun is the subject of
the main clause. When the modied noun is the direct object of its clause, the unmarked
word order is subject-verb-object (SVO), with the object last, and therefore the relative
clause does not interrupt the main clause: I know a woman who likes cats. Most Hit-
tite relative clauses precede the main clause, while a few follow; relative clauses never
interrupt a main clause. Second, while in English the noun phrase referred to by the
relative clause (the antecedent) always stands in the main clause (underlined in the
examples given), in Hittite preposed relative clauses (by far the most common type) the
antecedent stands in the relative clause (in the appropriate case) and is resumed in the
main clause by (1) an enclitic pronoun, (2) a demonstrative pronoun, or (3) a demon-
strative plus the repeated noun, again in the appropriate case. In rare instances (4) there
is no resumptive word, but the antecedent still stands in the relative clause. Examples:
(1) nu utxt-an kuin ukkikemi nankan si+si tmi Which child I am saying spells
over, him I say his name = I call by name the child whom I am saying spells over
KUB 7.1 i 7; (2) xtn.xtn.nr. kue dannatta ammuk rorn-pa aeanunun numukan
apya manda ara d Which empty lands I resettled, all those too he took away
from me = He also took away from me all those empty lands that I had resettled att.
iii 5758; (3) [(ot)]u pugarinma kue[(d )]ani tu-ti unur nuza
d
tft[(st ape)]tani
tu-ti war[( p)]ta On which day they adorned the substitute ox, on that day His Maj-
esty bathed = His Majesty bathed (on) the day on which they adorned the substitute
ox KUB 43.50 + KUB 15.36 obv 1819 (NH); (4) [
LU
x]
4
.r tarzi kui 1 x.
xi.nnnn . . . pianzi Which runner wins, they give one mina of silver = They give
one mina of silver to the runner who wins KBo 20.33:12 (OS). For the asyndeton in
(4) see 29.55 (p. 404).
23
30.59. Hittite and English relative clauses do share one important feature: the rela-
tive pronoun must undergo wh-movement, that is, it is fronted in its clause, regard-
less of its function (subject, object, locatival expression, etc.): see Berman 1972a: 45.
But this pattern is complicated in Hittite by another factor not present in English: Hit-
tite distinguishes between determinate and indeterminate preposed relatives (see Held
1957; Hale 1987: 4649; and Garrett 1994: 4349). Determinate relative clauses refer
to specic things whose existence has already been established in the discourse, as in
all four examples given above in 30.58. Indeterminate relatives, on the other hand,
refer to indenite things whose existence is not necessarily established: (5) kui ammel
ppan itoi-u kari nu
URU
attuan ppa ai nan nepia
d
rxtn-a azzit [(tu)]
Who(ever) becomes king after me and resettles attua, let the Stormgod of Heaven
strike him! KBo 3.22: 4951 (OS); (6) nuian kuit aan itoi-u di nu apt
ai What(ever) service the king imposes upon him, he shall do it KBo 6.4 iv 1617
(Laws XXXVII, NH). Notice that since indeterminate relative clauses do not estab-
23. The present description could not take into account the newly published analysis of OH relative
clauses by Probert (2006).
30.59 Clauses
00-HittiteGrammar.indb 424 6/5/08 4:20:55 PM
425
lish the reality of what is said, they are always equivalent to conditional clauses (= If
someone becomes king after me and . . ., and If the king imposes some service upon
him . . .) and take present-future verbs. As rst shown by Held (1957), in Hittite in-
determinate relative clauses, the relative pronoun must occur in clause-initial position
or follow directly a clause-initial conjunction and any attached clitics (for the latter
possiblity see example 6 above); it cannot be preceded by any accented constituent.
In determinate relative clauses the relative pronoun must be preceded by at least one
accented constituent, as the result of fronting (see 30.530.6, p. 407). Sometimes
only one constituent of a noun phrase is fronted, resulting in a noun being separated
from its modiers or from a coordinated noun: (7) nuza k kue +w+:i
ME
+\+
d
rxtn

+utx+ i +\+
d
tft
URU
fi-na xx+ arkuwar ai These words which I make as
a plea to the Stormgod your father and to the Sungoddess of Arinna your mother . . .
KUB 21.27 iv 4445; (8) nu txt.nr. kue man anuta all the seeds which were
roasted KBo 4.2 i 62 (rit.); (9) x.n.nr. xtn
URU
Nuai kui i x.n.nr. xtn
URU
Kinza +ui:+ arnut The civilian captives of the land of Nuai and the civilian
captives of the land of Kinza that my father removed . . . KBo 5.9 ii 3839 (NH). See
also example 2 cited above: xtn.xtn.nr. kue dannatta ammuk rorn-pa aeanunun
the empty lands that I resettled.
30.60. In the vast majority of determinate relative clauses, the relative pronoun is
preceded by only one syntactic constituent (Hale 1987: 49; Garrett 1994: 46). But there
remain genuine exceptions in which more than one constituent precedes the relative pro-
noun. In addition to example 4,
LU
x
4
.r tar (a)zi kui cited above, where the relative
pronoun follows the subject noun and the verb, note also the following: (10) [(numu
itoi)].xr ++nui:+ auwa memiyana kui eer The kings senior to me who
had been on good terms with me . . . att. iv 50; and (11) nu
m
Uri-
d
t-upa uiti
I.A
kui kuwapi ara uiyat The lords whom Uriteub had sent away somewhere . . .
(att. iv 1920). For similar exceptions with interrogatives see 27.7 (p. 351) and
27.12 (p. 352), and see also 30.43 (p. 418) on the position of the causal conjunction
kuit. Further research and some revision of the formal description in Hale and Garrett
may be needed. For understanding the crucial distinction between indeterminate and
determinate relatives, it is enough to determine whether or not the relative pronoun is
preceded by at least one accented constituent.
30.61. In extremely rare instances Hittite does employ a pair of preposed relative
clauses where the rst must be understood as subordinate to the second: urorn-tt+kan
kuedani +\+
DUG
oln.o anda arranzi nata wtar kuit +\+
DUG
oln.o anda nat r+\t
urorn-tt+ apzpat ts:i
DUG
oln.o di In which vessel they wash the god, the water
which is in the vessel, it he places before the god with that very same vessel = He puts
the water which is in the vessel in which they wash the god before the god with that
very same vessel KUB 27.16 i 3033 (NS).
30.62. Occasionally, Hittite relative clauses follow the main clause. Their structure
supercially matches that of English: the antecedent is in the main clause and is not
Clauses 30.62
00-HittiteGrammar.indb 425 6/5/08 4:20:56 PM
426
resumed in the relative clause. Garrett (1994: 47) distinguishes two types. The rst is the
non-restrictive relative, which, like those in English, does not identity the antecedent
noun phrase but merely adds further incidental information (such a clause is usually
marked by a pause in spoken English and is set o by commas in writing). For example,
urorn.i.xr urorn.xtts.xr s+ itoi i xtts.itoi-i::t kui darante kui
t darante Gods and goddesses of the King and Queen, those who (are) named, (and)
those who (are) not named, (those whose temples the King and Queen frequent, and
those whose temples they do not frequent) KUB 6.46 iii 4243 (NH).
30.63. Also postposed after the main clause are indenite relatives. Unlike pre posed
indeterminate relatives, these presuppose the existence of the noun phrase referred to.
But unlike preposed determinate relatives, the noun phrase is indenite, not something
specic that has already been mentioned: e.g., nu 8 utxt.xr-u uwadanzi xtts-
nian kui nwi pnzi They bring eight boys who have not yet gone to a woman
KUB 9.31 ii 910 (NS); mn ta[m]aima kuiki uiti+ anditta kui antezzi [a]nni
arta But if some other ocial is available who is standing in the front line IBoT 1.36
iv 2021 (MHlMS).
30.64. Very rarely, forms of relative kui- stand for unexpressed clauses of the type
who (islare there), in which case kui- stands in the main clause itself: i+++
d
tft-st-
++ +\+
m
Kuikuianduwa kueda tut-++ thus says His Majesty: say to Kuikuiandu
(and) who(ever are also there)! (Ortaky 90l1400 13 [letter], cited in Sel 1992: 491;
the lines which follow contain second-person plural imperative verb forms). Note: the
absence of clitic -al-ya on kueda may indicate that Kuikuianduwa is included.
Indirect Statements and Questions
30.65. Indirect statements and indirect questions generally follow the main clause
and are introduced asyndetically (the point of asyndeton is indicated below by ). Al-
most all of our examples come from NH or at least NS texts. Indirect statements usu-
ally contain kuit that: ammel kpat 1-a dammea kyan 1-an dammeanunun
ts:i r.oi-tt+patkan kuit katta uiyanun This was my only punishment (of the
Tawan nanna). I punished her with this one thing: only (-pat) that (kuit ) I sent her down
from the palace KBo 4.8 ii 1214; often after verbs of perception (seeing, hearing,
knowing): maanma i.xr
URU
Aur aur tnt.uruir.nr. nXukan kuit zaiyaz
katta dakeuwan teun When the Assyrians saw that I had begun to capture fortied
cities in battle KBo 4.4 iv 2829 (NH); [(nu
m
Aitaqq)]ama kui itoi
URU
Kinza ta
nui
m
Niqma-
d
t-a [kui] [ant]ezzi utxt-la ta nu maan auta [anda]kan
kuit atkenuwante numa alki
I.A
-u namma [tepaw]ezi nuza
m
Niqma-
d
t-a
m
Aitaqqaman +uisi kuenta When Niqmadda, the oldest son of Aitaqqama, king of
Kadesh, saw that they were besieged, and that their grain supplies were low, Niq-
maddu killed Aitaqqama, his father KBo 4.4 ii 36 (AM 112113).
30.63 Clauses
00-HittiteGrammar.indb 426 6/5/08 4:20:56 PM
427
30.66. Instead of a kuit clause, verbs of perception or speaking can take an adjective,
noun, or participle as a second direct object: ammukwaran akkantan tut he told me
it was dead KUB 13.35 iii 17; maanma xtn.xtn.xr i.xn
m
Arnuandan r:+
irman itamaer But when the enemy lands heard that my brother Arnuanda was ill
KBo 3.4 i 67 (AM).
30.67. With the verb to know that . . . one can also use a construction in which the
indirect statement precedes the main clause asyndetically: kinunwaz nwa
MUNUS.ME
t.
or[-u] [punukez]zi t aggai I dont know (if) she [is] now (kinun) still [consult-
ing] Old Women KUB 1.16 iii 6869 (OHlNS).
30.68. Indirect questions (Mascheroni 1980: 5859) are often formed in conjunction
with clauses containing the verb ak- to know, au- to see, or atri- to write, as
well as punu- to ask. They usually follow their main clause but occasionally precede,
as the examples cited below show.
30.69. Using mn if, whether (HE 333; Mascheroni 1980: 58; and CHD sub mn
4) with single mn: nu uiti i.xr xtn-:tpat punu mn ammuk aan luzzi
iaun O lord, just ask the men of the land if I ever had to perform aan and luzzi
obligations! HKM 52:3839 (MHlMS); with mn . . . mn whether . . . or: nu
m
[Urite]upa kuit apiya nan punu mn kian mn t kian Since Uri-Teub is
there (see 7.18, p. 147), just ask him if it is so or not so KUB 21.38 obv. 1112 (NH);
nu t agga[un] mnza itoi xtn
URU
Mizri +\+ [+ut]. eda +\+ xtn.xtn.xr
er ar [kamma]n iyat mnza it kuitki [iyat] I dont know if the King of Egypt has
rendered tribute to my father for those lands or if he has rendered nothing KUB 31.121a
ii 1215; cf also KUB 31.121 iii 1114.
30.70. Using a form of kui- who? what? (all exx. NH or NS): nu urorn.xr t
[ekteni k]ulla dammea O gods, dont [you know] whose is the injury? KBo
4.8 iii 34, ed. Honer 1983a: 188; t aqqa[]i kuia ai urorn-tt+-i I dont
know who that god is KUB 33.106 iii 44 (Ullik.). Not clause-initial: umewa [u]rorn.
xr t ukatteni kianwamu kui iyan arzi dont you gods see who has done this
(lit., thus) to me? KUB 54.1 i 2021 (NH). With subject matter of inquiry preceding:
xtn
URU
Wiluama +\+ xtn
URU
atti kuedani itoi-i auwan ara tyat nu memya
kui [(t)] [(i)]tantanza nan t aqqa[i] Since the incident occurred long ago, I dont
know from what Hittite king Wilua defected KUB 21.2 + KUB 48.95 i 69 (Alak.,
NH); kuedaniwa uddan uwanun numu fi-anza punudu witenanza Let the spring,
the water, ask me why I have come KBo 10.45 ii 2324 (OHlNS).
30.71. Using kuwapi where?: kianmu kuit atre kawa i.xn uet
nuwazakan
URU
aparan inian tamata
URU
Kaipuranmawakan kz tamata
apwakan itarna ara uet nammamawar a kuwapi pait nuwara t toi nu
ap i.xn alwanzaanza imma ta nan t a-a-ak-ta-a (Concerning) what
you wrote me as follows: The enemy has come and has besieged the city apara on that
Clauses 30.71
00-HittiteGrammar.indb 427 6/5/08 4:20:57 PM
428
side, and has besieged the city Kaipura on this side; and he has passed through (and got-
ten away). But I do not know (text: he doesldid not know) where he was going. Was
that enemy perhaps bewitched, that you did not know him? HKM 6:314 (MHlMS),
ed. Alp 1991a: 12629. With question preposed: kuwapiwa paii ammukmawatta
l aggai Let me not know where you are going KBo 5.9 ii 4445 (NH).
30.72. Using maan how? (see CHD LN, maan mng. 3) with maan
clause preceding: []antezziumaat itoi.xr maan [a]ra pittaler nat
d
tft
URU
fi-na o:+ [ ]akti How the earlier kings neglected it (i.e., Nerik), you, Sun-
goddess of Arinna, my Lady, know it KUB 21.27 i 1618 (prayer, NH), ed. Srenhagen
1981: 108111, see ibid. i 43; numakan mandaz maan [ . . . . . . ] nu aptta
urorn.xr-epat ekteni You gods also know how [ . . . . . . -ed] you fromlwith
everything KBo 53.10 ii 1112 (MHlMS); with maan clause following: zik
d
ts:+u
URU
Nenuwa o-\t t akti xtn
URU
atti orx-an dammean Dont you know
Is:+u of Nineveh, our Lady, how the country of atti has been oppressed? KBo 2.9 i
3839 (prayer in a rit., NH); nu tuel i )i+i kuwat t punuta memauni orx-an
Why did you not ask your messenger how I spoke to him? KUB 23.101 ii 56 (NH).
Multiple Dependent Clauses
30.73. The force of a subordinating conjunction can continue in multiple subse-
quent clauses: (conditional) [(takk)]u i.t
19
.it-an kuiki nikzi tan itarnikzi nu
apn ktizzi If someone injures a (free) person and incapacitates him, he shall treat
him KBo 6.2 i 1617 (Laws 10, OS); mn kpat namma[ma] dami t.oi kuitki
nu ror-zi [st.xr sro
5
-ru] If it is only this, and there is nothing additional, let the rst
[exta be favorable] KUB 22.70 rev. 1 (NH); (temporal) +ui:+ kuwapi
m
attuili +\+
m
Uriteupa utxt
m
Mwattalli menaanda kururiata nan itoi-eznani ara tit-
tanut +\+
m.d
ixxmakan watul t kuitki ta When my father Hattusili became
hostile towards Uriteshub, son of Muwattalli and removed him from the kingship, no
sin inhered in Kurunta BrTabl. i 69 (NH); (causal) ammuk t kuitki kuit dammin
armi tmakan dn kuedaniki kuitki armi +:++++mamu kuwat dammiikanzi
Because I have damaged nothing, and (because) I have taken nothing from anyone,
why are they injuring me in that way? HKM 68:47 (MHlMS) (and 30.75).
30.74. A temporal clause can be followed by a causal clause (here connected by -ma)
and then its main clause: nuza maan tnt-an tar[ta] [+ui:+m]akan urorn.xr-
a kuit naa[nza ta] [(nu)] arzzi gurti s+
d
[Kubaba? (i s+)]
d
ixx ma[- . . . (?)]
t kuinki tarna When (i.e., after) my father had conquered the city (of Carchemish),
because he feared the gods, he allowed no one [to spoil(?)] the upper citadel [or the
temples(?)] of [Kubaba] and the Patron Deity KBo 5.6 iii 3135 (D frag. 28.A). The
reason for the adversative ma here may perhaps be made clear by a slightly dierent
translation: (Although) my father had conquered the city, yet (-ma) because he feared
the gods, he allowed no one. . . . The same pattern can be seen with kuwapi: nu +ui:+
30.72 Clauses
00-HittiteGrammar.indb 428 6/5/08 4:20:57 PM
429
kuwapi
m.GI
r-i-in t\+ xtn
URU
Mezri tsriu nan kian kuit watarnata When my
father sent attuaziti to Egypt, because he commanded him as follows . . . KBo 5.6
iii 4547 (D frag. 28.A).
30.75. The opposite order of causal followed by temporal clause also occurs: numu
d
ts:+u o:+ kuit kanian arta r:+yamu
m
rn.oi u arta numukan
otm-an t.xr-anna n za s+
d
ts:+u o:+ kaniwar s+ r:+ya [a]ulan aur
numu araniyr And because Is:+u, my lady, had recognized me, and my brother
Muwatalli held me in favor, when people saw the recognition by Is:+u and the favor of
my brother, they envied me att. i 2832.
30.76. A conditional clause can be followed by a causal clause, a temporal clause,
and then its main clause: mn otu pugarima rorn xsxi ak[i xsxi? k (uit)]
twa nat otm-an apiya [ara (nzi nu tamain otu pugarin . . . apzza . . . [n (ann)]anzi
But if the substitute ox dies on the way, because it (i.e., the destination city) is far,
when they [arrive], they will drive another . . . KBo 4.2 iv 4649 = MSpr rev. 3235.
30.77. Relative clauses also co-occur in the same sentence with other dependent
clauses: nu ammel kuwapi +w+:i
ME
urorn.xr itamaanzi numukan kui idalu
memia zr-ni anda nanmu urorn.xr rorn-pa sro
5
-aanzi arlanzi When the gods
hear my words, the bad thing that is in my soul, they will make it right and lift it from
me KUB 6.45 iii 4547.
Clauses 30.77
00-HittiteGrammar.indb 429 6/5/08 4:20:58 PM
430
Chapter 31
SUMERIAN AND AKKADIAN
31.1. Much in Hittite language and literature is elucidated by Assyriology. Ancient
Mesopotamian civilization (law, religion, economy, etc.) can be studied in any number
of popular handbooks. Those aspects of Akkadian grammar and writing system which
aect the reading of Hittite texts can be found in HE, pp. 170182 (grammar) and HZL
(writing system).
Sumerograms
31.2. One needs to know relatively little about Sumerian grammar in order to prop-
erly understand the Sumerograms in Hittite texts. They consist of the root forms of a
limited number of nouns, adjectives, and verbs.
Nouns, Adjectives, and Participles
31.3. Noun case is normally not indicated by means of the common Sumerian suf-
xes, nor are any of the Sumerian verbal suxes employed. For details see 31.17
(p. 434). When the plural markers are axed to Akkadograms or syllabically written
Hittite words, they are postpositional determinatives and are written in superscript (e.g.,
+-:i
I.A
-:+ my hands). When they occur on Sumerograms, we follow the convention
of the CAD and CHD in not superscripting them (e.g., tut.nr. sheep); see 31.11
(p. 432). Plurals of nouns or adjectives are sometimes indicated by double writing: xtn.
xtn(.xr) the lands, urorn.xr oi.oi urorn.xr ftn.ftn the great gods (and)
the lesser gods. None of the overt plural markings is obligatory: the simple noun or
adjective root can stand for the plural (see 31.18, p. 435, for details).
Pronouns
31.4. Sumerian pronominal suxes are limited to the third-singular -nr its, seen in
xr.ii.nr its weight.
Verbs
31.5. Rarely a verbal prex, proper in Sumerian texts only to the preterite but em-
ployed more broadly in Hittite, n- occurs in the complex n. he died ( to die).
Akkadian Grammar
31.6. The best preparation for understanding Akkadograms in Hittite is a formal
course in Akkadian. If only self-study is possible, either H. A. Honers English trans-
00-HittiteGrammar.indb 430 6/5/08 4:20:58 PM
431
lation of A. Ungnads concise outline of Akkadian grammar (1992) or the two-volume
introductory grammar by J. Huehnergard (2005a; 2005b) is recommended. The com-
prehensive grammar is by von Soden (1995; abbr. GAG
3
). There are two grammars of
the Boazky corpus of Akkadian, that is, the Akkadian unilingual texts written by
scribes in the Hittite capital. The older is by Labat (1932); the more recent is by Durham
(1976). The latter was a Ph.D. dissertation at Harvard University and is not available
commercially. It is not the purpose of this chapter to treat the material examined by these
authors. Instead, what follows is a brief synopsis of Akkadian grammar as it appears in
logograms found in Hittite contexts. It may seem strange to treat the grammar of what
are essentially logograms scattered in sentences controlled by Hittite grammar. But it is
important to have an orientation in Akkadian grammar in order to properly understand
the Hittite scribes use of these logograms.
31.7. Akkadian words and forms in Hittite context are called Akkadograms and are
transcribed in printed publications in uppercase italic (1.13, p. 14). For the nature of
the cuneiform writing system see above, 1.5. (pp. 10.).
Orthography
31.8. Hittite scribes did not use separate cuneiform signs to indicate voiced and
voiceless stops (see above in chapter 1). For example, the orthographic system they in-
herited used both the u and f signs for either ldal or ltal. Akkadian blti-ni our lady
was usually written ui-it-ot-\t, using the ur sign for what in Akkadian is a voiceless
stop. Akkadian labka moist (masc. sg. acc.) was written t+-+u-o+, using the o sign
for voiceless x+, which required more strokes to form. In transliteration it is possible to
use the fl value of the ur sign to write ui-it-: and the xX value of o to write t+-+u-x+.
The transliterator in this way aids readers in identifying the intended Akkadian word,
but such writings may tend to mislead readers into thinking that the ancient scribe has
sought to represent the voiceless phoneme by a particular sign, when in fact he has not.
In this grammar,
1
when representing Akkadograms in bound transcription, we consider
it proper to render Akkadograms in a way that they can be correctly identied in an
Akkadian word list or dictionary. We willingly sacrice the accuracy of reecting the
signs the Hittite scribe wrote for the sake of clarity of meaning. Thus the Akkadogram
ui-it-ot (my) lady intends Akkadian blti (or blt) and in this grammar is written in
bound transcription as uIt:t. Similarly, :i-st spring comes from the Akkadian plural
d,
2
and
GI
o+-+\-\i(+) pot-stand represents
GI
x+\\i(+). In restricted contexts rare
values of signs were used: in writing Akkadian bli-ni our lord, Hittite scribes did not
write ui-tt-\t but ui-t-\t, using the r sign with its unusual value t.
1. In Hittitological literature in general it is more usual to nd the Akkadograms rendered according to
the main values of the cuneiform signs: uitot, :ist, o+\\i+, etc.
2. There is a very real possibility that the similarity in shape of TE and DI caused Hittite scribes to
confuse the two signs in ot-st as :i-st, for at no time and in no geographic region in which Akkadian was
written was this word vocalized as *d but rather as d (Babylonian) or da (Assyrian).
Sumerian and Akkadian 31.8
00-HittiteGrammar.indb 431 6/5/08 4:20:59 PM
432
31.9. The negative word it (26.1., pp. 341.), is regularly written i-it with
two signs in Akkadian texts, not because there is vowel length but because this was
the accepted spelling of the word in the writing system of the Akkadian scribes who
introduced cuneiform writing at the Hittite capital. It is the regular way of spelling the
word in Old Babylonian Akkadian. It is likewise written with two signs in all but the
last century of scribal practice at Boazky. In texts written there during the thirteenth
century the word was written with only one sign (it, not i-it). In the convention of the
CHD, when representing the two-sign writing in broad transcription the u-vowel is writ-
ten with a long mark, t. We follow this convention in order to indicate the spelling. But
by this writing we do not mean that we believe there was any dierence in vowel length
indicated by the dierent writings. In Hittite texts no phonetic complement indicates the
underlying Hittite equivalent natta. That is, we nd no writing *i-it-ta.
Superscripting
31.10. In transcribing Hittite texts superscripting is employed mostly with deter-
minatives (1.39, p. 23, 1.43, p. 24).
3
Sumerian words such as or wood, tree, tnt
city, xtn country, tntut copper,
4
stone, fo cloth, garment, ru bread,
i man, xtts woman, ntn.so mountain, u river, uto vessel, container, etc.,
are often prexed to words in order to identify their semantic class, category, or material.
A few determinatives are not prexed but rather suxed to their nouns (1.42, p. 24):
xtr bird, sn vegetable, and xr place (in
URU
atti
KI
,
URU
Arinna
KI
,
URU
alpa
KI
and
x.urorn.n
KI
). The occasional writing tnt.xr (in xxrx tnt.xr and r tnt.xr) is a
(false) Hittite abstraction from the fuller form with specic place names, just as tu.xx
day and utn.xx tablet were abstracted by Hittite scribes from longer expressions
with numerals (see 9.51, p. 167; and p. 434, n. 5).
31.11. The most common post-positioned determinatives are the plural markers
xr and nr.. Limited to the earliest written phases of Hittite is uruir, often combined
with nr. as uruir.nr.. Although none of these was originally a plural marker in Su-
merian itself but were rst used to indicate the plural of Sumerograms in Akkadian
texts, we follow the CAD and CHD convention and do not superscript them when they
pluralize Sumerograms. Thus we write tnt.uruir.nr., not tnt
DIDLI.I.A
, but str:i+
I.A

incantations, niu!:i+
I.A
ditches, \++++\:i+
I.A
measuring vessels. While the
Sumerian plural marker nr. was occasionally employed on Hittite words (kururi
I.A
,
lapanalli
I.A
-u, alki
I.A
-u), the other plural markers xr and uruir were not.
3. Rarely, superscripting is used in transliterations to indicate the function of a sign as the scribes hint
as to the correct reading of the immediately preceding polyvalent sign. An unusual example of this, which
seems to have no purpose in Hittite, is in the sequence of signs t + aliu + a, which we transcribe t
u
-
a, superscripting the lrltn sign, in order to show that it intends to show that the fn sign is to be read with
a CVC value ending in vowri + . This seems unnecessary, since in Hittite the fn sign only has values
ending in vowri + (i.e., ta, tu, da, du).
31.9 Sumerian and Akkadian
00-HittiteGrammar.indb 432 6/5/08 4:21:01 PM
433
Phonetic Complements
31.12. We follow the procedure of the CHD in not superscripting phonetic com-
plements, either Hittite or Akkadian. We therefore write urorn-tt+ god(dess), and
urorn-tt+-i. Even such apparently awkward forms as r.nr.-:t+ houses (for Ak-
kadian bttim), x.oi.nr.-:t+ city gates (for Akk. abulltim), xsxi.nr.-:t+
roads (for Akk. arrntim), utn.nr.-:t+ clay tablets (for Akk. upptim), ftn.
nr.-:t+ small (ones) (for Akk. ertim), o.n.nr.-:t+ gifts (for Akk. qtim),
and xtn.xtn.nr.-:t+ countries (for Akk. mttim) are not essentially dierent from
lapanalli
I.A
-u and alki
I.A
-u cited above. And there is no doubt that the nal -:t+ in
ui-ti
I.A
-:t+ lords was intended to be read as part of the Akkadian word. Admittedly,
there were a few inept scribes who falsely regarded -:t+ (or -:t, without mimation; see
31.20, p. 436) as a kind of universal plural for Akkadian words. When this happened, it
was axed to words where a nal syllable -tim or -dim cannot reasonably be assumed:
GI
ntn.nr.-:t+ tables is inappropriate for the Sumerian loanword in Akkadian
paru, which inects in the plural as a masculine noun: parl. But since this hap-
pened rarely, it does not justify our treating -:t+ like a postpositional determinative.
31.13. While Hittite phonetic complements on Sumerograms are quite common
(e.g., itoi-u king), those on Akkadograms are understandably rare, because with
an Akkadographically written noun there is little chance of ambiguity of either num-
ber or case, and with nite verb forms the person of the subject and the tense are also
normally quite clear. There are few instances of unnecessary Hittite complements on
Akkadograms, and never on those in the construct state: natmukan talza +sriu-zi it
he(!)
4
will not send it to me talza KUB 40.1 rev! 17; it-t++-a 'ssi his hand (that)
of a free man KBo 6.3 i 29 (Laws 11). For ui-ti-u (-a-an) lord (sg. nom.) HKM
52:25 and HKM 80 obv. 5 there may be a better explanation (see p. 22, n. 30).
Phonology
31.14. Akkadian possesses three basic vowels (a, i, and u) and a secondary vowel e
derived from either a or i. These all occur both short and long, although the long vowels
are rarely written plene. Real diphthongs are rare. There are two semi-vowels in Ak-
kadian: y and w. A glottal stop (transcribed ' but often omitted in cuneiform writing)
reects a prehistorical merger of ve distinct Proto-Semitic sounds. The presence in
the noun or verb root of a Proto-Semitic laryngeal ( or ) in contact with an a-vowel
colored that a (and sometimes also any a in the neighboring syllable) to the secondary
vowel e. This shift was followed by the loss of the laryngeal and the compensatory
lengthening of the preceding vowel: `bal- lord > *bel- > bl; feminine *balat- lady
(= queen) > *belat- > blet-; *malqat- that which is taken (as a tax or impost) >
*malqet- > melqt-. In Hittite texts the rst two nouns are seen in the singular as uIti
4. The Hittite complement -zi shows that the verb is third-person singular, but the Akkadian form +sriu
(versus tsriu) should be rst-person singular!
Sumerian and Akkadian 31.14
00-HittiteGrammar.indb 433 6/5/08 4:21:02 PM
434
(uIt:i), construct uIt (uIti:), and plural as uIt:t (uItI:t) (written r-(i:-):t+); the
third noun appears in the construct with clitic possessive pronoun as +itIssi (often
inadequately transcribed as +i-it-xt-tz-zi) its taxlperquisite.
31.15. The consonants of Akkadian are represented in the following table:
voiced voiceless emphatic nasal
dentals d t n
labials b p m
velars g k q (velar)
sibilants z s,
liquids l, r
glottals '
velar fricatives
31.16. The sibilant immediately preceding a dental stop became l, as seen in MB
iltuu whip from OB ituu and in ilur he wrote from iur (innitive aru).
In the opposite sequencedental followed by the rst shifted to s, whereupon the
dental fully assimilated to the s, producing ss. This development is most frequently seen
in the nominative-accusative construct forms of nouns whose stems end in a consonant
when followed by a clitic possessive pronoun beginning with (e.g., -u his, -i her,
-unu their, etc.). Among Akkadograms in Hittite one nds it routinely in the word
+-+s-si (less commonly written t-si) for Akkadian qssu his hand (nom. or acc.),
which derives from the construct state qt of the noun qtu hand and the third-person
masculine pronominal sux -u his. Other examples where the gemination of the sibi-
lant is masked by a Sumerogram are: xtn-si for Akkadian mssu (< mt-u) his land,
so.ut-si for Akkadian qaqqassu (< qaqqad-u) his head, r-si for Akkadian bssu (<
bt-u) his house, ln-si for Akkadian warassu (< warad-u) his man-slave, r-si-\i
their house(s), fo-si-\i (ubssunu < ubt-unu) their garment(s), x.n.nr.-
si-\i (allassunu < allat-unu) their captives, prisoners of war, r-si-\i (assunu
< at-unu) their sister, utxt.xtts-si (mrassu < mrat-u) his daughter.
Morphology
Nouns and Adjectives
31.17. Noun case is normally not indicated by means of the common Sumerian suf-
xes, nor are any of the Sumerian verbal suxes employed. Only six Sumerian suxes
are employed on nouns or adjectives: (1) the genitive in -a(k),
5
seen in r.xtn.n
5. The velar (k) is preserved only when immediately followed by another vowel-initial sux, which
in Hittite texts occurs only in the xx immediately following a number (e.g., tu.10.xx day of ten, i.e.,
tenth day, utn.2.xx second tablet).
31.15 Sumerian and Akkadian
00-HittiteGrammar.indb 434 6/5/08 4:21:03 PM
435
(donkey [ane] of the foreign counry [kur + -a(k)]), (2) the adjectivallparticipial end-
ing -a, seen only occasionally, as in l uio.o (oil [] that is nelgood [dug + -a]), and
(3, 4, 5, and 6) the plural indicators xr, nr., r.r, and uruir, as well as the combina-
tions xr.nr. and uruir.nr.. The sux uruir alone (i.e., without nr. added) is very
rare and occurs only in the oldest texts. In a small number of cases a Sumerian word is
doubled in the plural, such as xtn.xtn.xr or xtn.xtn.nr. countries, and in the case
of urorn.xr oi.oi the great gods, tnt.uruir.nr. oi.oi-:t+ the large cities,
4
GI
n ftn.ftn four small bows, the attributive adjective alone is doubled. This us-
age is occasionally extended even to a syllabically written Hittite noun: 5 gapinan ftn.
ftn ve small threads.
31.18. Sumerograms did not need to have plural markers, especially when the con-
text or preceding numbers made the plurality clear. Since the Hittite nouns behind x.
n deportees, otu cattle, and tut sheep, when conceived in the plural, were gram-
matically singular but with collective meaning, it was natural for them to often dispense
with overt plural markers. And since, even in Hittite, the syntax of numbering permitted
the counted item to be in the singular (9.22, p. 159), it is not surprising to nd 5 otu
instead of 5 otu.nr.. The Akkadian noun inects through three cases in the singu-
lar (nominative, genitive, accusative) and two in the plural (nominative and genitive-
accusative). Only in the earlier phases of Akkadian (Old Akkadian, Old Assyrian, and
Old Babylonian) does the ending of the nominative and accusative singular consistently
dier. Beginning in Middle Babylonian, both cases were written with the ending of the
nominative (-u or -um with mimation).
31.19. arr-
6
is the masculine noun stem meaning king, while arrat- is the femi-
nine noun stem meaning queen. Similarly, bl- is the masculine noun stem lord, while
blt- is the feminine equivalent meaning lady (= queen). In the singular these nouns
decline as follows:
nominative arrum arratum blum bltum
genitive arrim arratim blim bltim
accusative arram arratam blam bltam
31.20. From this it is apparent that arr- and bl- are the basic root of the words,
-at- (or -t-) is a feminine stem-forming sux, and the endings of the singular are -um,
-im, and -am. Another Akkadian noun whose feminine sux is -t- is mrtum daugh-
ter (compare mrum son). The nal m on these endings is omissible in the Hittite
time-frame. The nal m in the feminine noun forms is also omissible (sg. -(a)tu, pl.
6. When we discuss Akkadian forms in their own right (for instance, in the following paradigms), we
will often write them in lowercase italic, as we would if we were discussing an Akkadian text. When Ak-
kadian words or forms occurring as logograms in a Hittite context are discussed, the Akkadian parts of the
logogram will be written in uppercase italic, while the Sumerian parts will be written in uppercase roman
letters.
Sumerian and Akkadian 31.20
00-HittiteGrammar.indb 435 6/5/08 4:21:03 PM
436
-tu). This omissible nal m is called mimation. In the plural the forms in classic Old
Babylonian are:
nominative arr arrtum bl bltum
genitive-accusative arr or arr arrtim bllbl bltim
31.21. In the Hittite usage the genitive-accusative forms were often used in place of
the nominative. What makes the masculine noun-stem plural is the long vowel in the
endings, which is not indicated in the Hittite writings of Akkadian nouns. What makes
the feminine noun-stem plural is the long vowel in the feminine root-forming sux,
-t- (in some cases -t-), likewise normally unrepresented by plene writing in Hittite.
Some Akkadian nouns which are masculine in the singular form feminine plurals: nru
river (pl. nrtum), arrnu road (pl. arrntum), ikkaru farmer (pl. ikkartum).
These can also be found in Hittite texts: u.xr-:t+, xsxi.xr-:t+,
LU.ME
ron.
xr-:t+.
31.22. Akkadian adjectives decline like the noun, except that in the masculine plural
the ending is not -, -, but -tu(m) and -ti(m). Hittite scribes, like their contemporary
Middle Babylonian and Middle Assyrian scribes, often used this adjectival masculine
plural ending on nouns. Thus uIt:t (written r.xr-:t+ or ui-ti
ME
-:t ) was the plural
of uIti lord. And since in Hittite masculine and feminine are not distinguished in any
grammatical category, it is rare especially in post-OH that any scribe uses a femi-
nine form of the Akkadian adjectives. Common Akkadian adjectives used in Hittite texts
as logograms are: ellu free, emu sour, aartu green, labku moist, rqu empty,
and almu complete. As used by the Hittite scribes, whose knowledge of Akkadian
was rudimentary, the properly nominative ending -u(m) is often used for the accusa-
tive, while the proper endings for accusative and genitive are much less frequently used
for the nominative. For example, the Akkadian nominative urorn-ti+ (= Akk. ilum)
god can be a direct object in Hittite context, while the Akkadian genitive urorn-tt+
is often used as a nominative, when there is also a clarifying Hittite nominative case
ending added (urorn-tt+-i for iuni). The Akkadian accusative form urorn-t++ is
extremely rare as anything but a direct object.
31.23. The combination of two nouns in a possessor + possessed relationship (the
kings son) is constructed dierently in Akkadian than in Hittite. In Hittite one wrote
auwa utxt.rf (i.e., of the king + the son). In Akkadian the noun denoting the
possessed comes rst, assuming a form called the construct state. This construct form
is followed by the possessor in the genitive case: mr son (in construct) + arri(m) of
the king (gen.). The construct state of a noun is formed by removing the case ending:
mru(m) son (sg. nom.), mr son (of) (construct); arratu(m) queen, arrat queen
(of) (construct); bltu(m) lady, blet lady (of) (construct); and thus mtt lands
(of), amlt men (of). When the resulting stem would end in a doubled consonant,
sometimes the doubling is preserved by adding a short i ( uppi tablet (of), libbi heart
31.21 Sumerian and Akkadian
00-HittiteGrammar.indb 436 6/5/08 4:21:04 PM
437
(of)) and at other times, when the nal doubled consonant is a liquid or resonant, it is
simplied (arrum king, ar king (of). Hittite phonetic complements are never added
to Akkadograms in the construct state. The most frequent construct forms encountered
in Hittite contexts are uIt lord (of), )irrt tablet (of), r'\ face (of), before, ':
hand (of), s+r+t underside (of), below, +ssi+ (contracted in Akk. itself from earlier
ana m) for the sake of, and +s+u place (of). The construct of mru son is found in
the personal name of a Hittite scribe:
m
Mar-er son of (i.e., born on) the twentieth (day
of the month). The logogram for name has become a frozen form in Hittite, written
with the single sign si+, as though it were always in the construct state, never as si-+i,
si-++ or si-+t (see 31.37, p. 441). The familiar Akkadian prepositions t\+ in(to), +\+
to, for, ts:i from, with, +oi with, t::t together with, and ++n+u before are
followed by nouns in the Akkadian genitive case or by a Sumerogram with an Akkadian
complement to indicate case, e.g., ts:i xtn-:t (= Akk. itu mti) from the land.
Pronominal Suxes
31.24. Akkadian nouns take possessive suxes. With singular nouns:
7 8
nominative-accusative genitive
mri-ya
7
my son mri-ya of my son
mr-ka your (masc. sg.) son mri-ka of your son
mr-ki
8
your (fem. sg.) son mri-ki of your son
mr-u his son mri-u of his son
mr-a her son mri-a of her son
mr-ni our son mri-ni of our son
mr-kunu your (masc. pl.) son mri-kunu of your son
mr-kina your (fem. pl.) son mri-kina of your son
mr-unu their (masc. pl.) son mri-unu of their son
mr-ina their (fem. pl.) son mri-ina of their son
31.25. Most examples of possessive suxes in Hittite texts are axed to Sumero-
grams, much less often to Akkadograms (ui-t-\t our lord,
9
i+-++-\t our mother),
and not at all to Hittite words. And since in Hittite there was no masculine-feminine
distinction in inectional forms, the suxes -xt your and -s+ her are quite rare, and
-xt\+ your (pl.) and -st\+ their virtually nonexistent. See 31.16 (p. 434).
7. In classical Old Babylonian this would be mr-, but Hittite scribes always wrote the sux my as
-:+. Similarly, Hittite scribes wrote nominative-accusative bl- my lord as ui-ttlt-:+.
8. Often in Hittite texts the distinctively feminine Akkadian suxes (-ki, -a, -kina, -ina) are replaced
by the more common masculine forms: -ka, -u, -kunu, -unu.
9. The correct form of our lord in the nominative and accusative is blni, genitive blni. The Hittites,
however, used the genitive form for all cases.
Sumerian and Akkadian 31.25
00-HittiteGrammar.indb 437 6/5/08 4:21:04 PM
438
Numbers
31.26. Numbers. The following Akkadian cardinal numbers are found in syllabic
writing in Hittite texts: 1-i\ (itn) one, er twenty (gen. of er; see 31.23, p. 437).
The following ordinal numbers are found: an second, alu third (in
LU
s+tsi third
man). Akkadian multiplicative expressions (x times) are formed with the genitive of
the ordinal numbers plus the third-person singular possessive sux -u its: al-u
three times, am-u ve times. One nds this in semi-logographic form (e.g., 1si,
2si) everywhere in Hittite as the equivalent of the Hittite multiplicatives in -anki (see
9.54., pp. 168.).
Verbs
31.27. Verbs. Since, especially in NH, the scribes utilized Akkadian verbs as logo-
grams for their Hittite counterparts, a brief summary of the Akkadian verb is necessary
here. But since the Hittite scribes used only a few of the most common Akkadian verbs
as logograms and usually in the preterite tense, even when in good Akkadian the
present-future is called for the student of Hittite does not require the detailed control
of Akkadian verbal paradigms which would be required for reading Akkadian texts.
The Akkadian language belongs to the Semitic family. As such, its verbal conjugation
is based upon a combination of prexes, suxes, and patterns of internal vowels. Each
Akkadian verb has a consonantal skeleton which we call a root. Most of these are tricon-
sonantal (sometimes called strong roots), e.g., pr to send. The dictionaries list the
verbs by their ininitve form. For those attested in the G- or Basic Stem this innitive
takes the form C
1
a C
2
C
3
u (m), where C
1
,

C
2
, and C
3
represent the rst, second, and third
consonants of the strong root. The verb to send would be listed as apru(m).
31.28. The combination of prexes and suxes used for the preterite (past tense)
paradigm follows:
10

11
Translation Prex Stem Sux
I sent a- pur
10
You (masc. sg.) sent
11
ta- pur
He
12
sent i- pur
We sent ni- pur
You (pl.) sent ta- pur -
They (masc.) sent
13
i- pur -
31.29. The present-future tense forms in Akkadian paradigm forms aappar,
taappar, iappar are extremely rare as logograms in Hittite. In the land grant texts
10. pur is the preterite stem of apru. The corresponding present-future forms of the paradigm contain
the present-future stem -appar-.
11. The corresponding feminine form tapur is unused in Hittite logograms, since Hittite did not dis-
tinguish masculine from feminine grammatically.
31.26 Sumerian and Akkadian
00-HittiteGrammar.indb 438 6/5/08 4:21:05 PM
439
the form tu+ooi+ from ragmu to (legally) contest (a claim) is found, going back to
OH or MH usage. But otherwise, such forms are not to be found. You (pl.) forms of
Akkadian verbs (tapur) as logograms are rare, if they exist at all.
1213
31.30. So-called strong roots, triconsonantal with no weak consonants, inect
largely as indicated in the above simplied paradigm. See +xuiu I vowed, txuiu he
vowed, ++i: I fell, +sxi: I became silent, +sriu I sent, tsriu he sent, \tsxi\
we placed, laid. The so-called weak verbs substitute a semi-vowel (y or w) or a long
vowel for one or more of the three consonants of the root.
14
Some weak verbs are
actually biconsonantal. In traditional Akkadian grammars it is customary to identify
the position of the so-called weak radical in a hypothetical triconsonantal pattern.
Hence, the verb mtu(m) to die is traditionally identied as middle weak (as if the
triradical skeleton were *mwt). The verb qab to say is nal weak, and epu to do,
make is initial weak. The presence of a weak consonant in one or more positions of
the root results in an inectional paradigm that is quite dierent from that given above.
For students of Hittite who cannot take a formal introductory course in Akkadian it is
more practical simply to learn the individual inected forms of weak verbs which
are limited in number in Hittite texts as vocabulary items, than to seek to explain the
paradigms of all types of weak verb.
31.31. Middle weak verbs occurring in Hittite logograms are: is+!: (causative -
stem preterite of mtu to die). Final weak verbs include qab to speak (nite forms
tut, +ut, impv. tui, etc.), leq to take (forms ttI, written tt-i-i), em to hear
(forms +s+i, ts+i), ban to build, fortify (forms +u\t, tu\t). Initial weak verbs include
amru to see (forms !+iu he saw), epu to make (participial form Irts maker
of). Initial and nal weak verbs include id to know (form toi).
31.32. The paradigm given above in 31.28 is of the Basic (or Primary) Stem (called
G-Stem, for German Grundstamm). There are also derived stems in Akkadian: the so-
called intensive-factitive D-Stem (with doubled middle radical),
15
the causative -
Stem,
16
the passive N-Stem,
17
and variants of each of these with inxed -ta-
18
and
-tan-
19
syllables.
31.33. Some verbs color the a vowels in the prexes of the G- and N-Stems to e:
tpu (from root p ). In the we form of the G-stem the vowel i replaces a: niddin we
12. In Akkadian there is also a she sent form tapur, but in Hittite texts the he sent form is used for
both genders.
13. The corresponding feminine form ipur is unused in Hittite logograms, since Hittite did not distin-
guish masculine from feminine grammatically.
14. This statement represents an oversimplication of the complex nature of the Akkadian verbal sys-
tem, but it will serve the purpose of this brief survey.
15. One of whose basic roles is to form factitive verbs.
16. Principal use to form causatives.
17. Principal use to form passives of the G-Stem.
18. Usually adding a reexive or reciprocal idea.
19. Usually adding an iterative or frequentative idea.
Sumerian and Akkadian 31.33
00-HittiteGrammar.indb 439 6/5/08 4:21:06 PM
440
gave, nilqe we took, nikun we put. In the D- and -Stems the prex always contains
a u-vowel, e.g., umt he killed, tumt you killed, uknn they bowed.
31.34. The most common Akkadian verbs used as logograms in Hittite are abtu
(t,u+:
20
he seized from Akkadian abtu to seize, used for Hittite pta), nadnu to
give (toot\ he gave, +oot\ I gave, toot\ they gave, \toot\ we gave from Ak-
kadian nadnu, used for forms of pai-lpiya- to give), qeb (nal weak) to say (tut
he said, :+ut you said), and em to hear (ts+i he heard, :+s+i you heard,
\ts+i we heard). These forms occur so frequently and in such predictable contexts
that they pose no serious problem to recognition. Imperative forms of Akkadograms
are limited to upur send!, qib speak! (sg.), normally at the beginning of letters, and
mis wash! (written +t-t-st) in Babylonian-inuenced rituals. Participles are limited to
LU
Irts maker (of),
LU
sIu brewer,
LU
\'st ,to!:t provisions bearer, +i\\+u:i fugi-
tive, and
LU
,'toi hunter. A fairly complete list of all inected Akkadian verb forms
found in Hittite texts can be found in HZL 36269. A commonly occurring Sumerogram
verb which contains the verbal prex n is: n. he died.
Conjunctions
31.35. The principal conjunction of Akkadian is u and (Semitic wa), which von
Soden (AHw, GAG) distinguished in transcription from the disjunctive or (Semitic
*aw, Hebr. ). In many periods of Akkadian dierent u-signs are conventionally and
more or less consistently used to distinguish u and from or. And was written i
in most periods and dialects, but often in Old Assyrian and t in Middle Babylonian
and Middle Assyrian. In Hittite texts the Akkadogram i and is always written with
the sign i. Akkadian or occurs in Hittite texts only as a component of the Akkadian
disjunctive combination l . . . l . . . l whether . . . or . . . or) (see 29.60, p. 405),
written ti-i . . . ti-i . . . i ti[-i] in KBo 3.22:6162 (OS), but ti-i . . . ti-i . . . i
ti-i in the NS dupl. KUB 26.71 i 9 . An Akadian enclitic conjunction -++ (GAG 123a;
AHw 56970) occurs much less frequently, principally in the combinations i+-++ PN-
++ thus sayslsaid PN and i+-++ si-i-++ thus (s)he sayslsaid.
Calques
31.36. Calques are literal translations of foreign idioms. Hittite possessed a small
number of calques derived from Akkadian idioms based upon uIt lord of: annena
ia- opponent in court (lit., lord of the judgment) (< uIt o!\t), auriya ia- dis-
trict commander, provincial governor (lit., lord of the watch-post) (< uIt ++oo+t:t ),
ana ia- heir of a murdered man, kinsman avenger of blood (lit., lord of the
blood-money) (< uIt o'+!
21
), iwarua ia- holder of a dowry or inheritance share
20. The uppercase italic forms represent forms actually found in Hittite Akkadograms. They are also
found in Akkadian texts.
21. Koroec (1931: 38, followed by CAD D 80 with literature) thought that only in Hittite did the ex-
pression derived from Akkadian bl dami have the meaning avenger of blood, since Akkadian bl dami
31.34 Sumerian and Akkadian
00-HittiteGrammar.indb 440 6/5/08 4:21:07 PM
441
(uIt siutx:t ), parna ia- home owner, householder (uItluIti: r-:t+ ), tuzziya
ia- military commander (< r xn), maniyaaya ia- administrator, judge,
uddana ia- opponent in court(?), enemy(?) (< uIt +w+:t+ ), etc. The syllabic Hittite
writings of other similar combinations, such as uIt urorn-tt+ worshiper, r.ssxtn
(uIt \tI) sacricer, r ':t craftsman, have not yet been found in extant texts.
And some of the Hittite idioms with ia- seem modeled on the Akkadian calques yet
have no known Akkadian counterparts: mukina ia- person who commissions the
m.-ritual. Several of the syllabic Hittite examples take the i determinative on the geni-
tive, as though the genitive + head noun were regarded as an indivisible compound (on
such forms see Neu 1986).
Prepositions
31.37. Prepositions. The following Akkadian prepositions are attested in Hittite
contexts (see 31.23, p. 436, for the cases they govern): +\+ to, for, +ssi+ for, as,
t\+ in, ts:i from, with, t::t with, ++n+u before, in the presence of, r'\(t) before,
+oi together with, along with, s+r+t under, below. Some of these originated as
constructs of nouns: ++nui, r'\, s+rti. They and other nouns often combine with a
preceding preposition to form compound prepositions: +\+ r'\t and t\+ r'\t before,
t\+ ++n+u in(to) the presence of. The preposition +ssi+ for the sake of, on account
of is such a formation, being derived from ana m inlfor the name of. The Sumero-
gram rorn without Hittite complement often stands for the Akkadian pseudo-preposi-
tion (w)arki behind, after. In general +\+ covers the Hittite allative and dative cases,
t\+ generally the locative, ts:i the ablative and instrumental, r'\ and t\+ r'\(t) the
postpositional peran with preceding noun in dative-locative, and s+r+t the postposition
kattan with preceding noun in dative-locative.
only means slayer, murderer. This claim is now disputed. Mishali (2000) describes the functionary in
Neo-Assyrian legal proceedings as an intermediary responsible for carrying out the redemption of blood or
redemption payment . . . (it) does not refer to the murderer or the avenger. Barmash (2004: 5256), on the
other hand, claims that Akkadian bl dami encompassed the two meanings: killer and avenger.
Sumerian and Akkadian 31.37
00-HittiteGrammar.indb 441 6/5/08 4:21:08 PM
442
REFERENCES
Alp, Sedat. 1991a. Hethitische Briefe aus Maat-Hyk. Atatrk Kltr, Dil ve Tarih Yksek
Kurumu, Trk Tarih Kurumu Yaynlar VIl35. Ankara: Trk Tarih Kurumu.
______ . 1991b. Hethitische Keilschrifttafeln aus Maat-Hyk. Atatrk Kltr, Dil ve Tarih
Yksek Kurumu, Trk Tarih Kurumu Yaynlar VIl34. Ankara: Trk Tarih Kurumu.
Andre-Salvini, Beatrice, and Mirjo Salvini. 1999. A New Trilingual Vocabulary from Ras Shamra
and the Relationship between Hurrian and Urartian. Pp. 26776 in Nuzi at Seventy-Five,
edited by D. I. Owen and G. Wilhelm. Bethesda, MD: CDL.
Archi, Alfonso. 1979. Lhumanite des hittites. Pp. 3748 in Florilegium Anatolicum. mlanges
oerts a Emmanuel Laroche, edited by E. Masson. Paris: de Boccard.
Balkan, Kemal. 1957. Letter of King Anum-hirbi of Mama to King Warshama of Kanish. Ankara:
Trk Tarih Kurumu.
Barmash, Pamela. 2004. Homicide in the Biblical World. Cambridge: Cambridge University
Press.
Bechtel, George. 1936. Hittite Verbs in -sk-. a Study of Verbal Aspect. Ann Arbor: Edwards
Brothers.
Beckman, Gary M. 1982. The Anatolian Myth of Illuyanka. JANES 14: 1125.
______ . 1983a. A Contribution to Hittite Onomastic Studies. JAOS 103: 62327.
______ . 1983b. Hittite Birth Rituals. Studien zu den Boazky-Texten 29. Wiesbaden: Harras-
sowitz.
______ . 1986. Proverbs and Proverbial Allusions in Hittite. JNES 45: 1930.
______ . 1999. Hittite Diplomatic Texts. Second edition. Writings from the Ancient World 7.
Atlanta: Scholars Press.
Beckman, Gary M., and Harry A. Honer Jr. 1985. Hittite Fragments in American Collections.
Philadelphia: American Schools of Oriental Research.
Beekes, Robert S. P. 1995. Comparative Indo-European Linguistics. An Introduction. Amster-
dam: John Benjamins.
Benveniste, mile. 1962a. Hittite et indo-europen. tudes comparatives. Paris: Maisonneuve.
______ . 1962b. Les substantifs en -ant du hittite. BSL 57: 4451.
Berman, Howard. 1972a. Relative Clauses in Hittite. Pp. 18 in The Chicago Which Hunt. Pa-
pers from the Relative Clause Festival, April 13, 1972, edited by P. M. Peranteau, J. N.
Levi and G. C. Phares. Chicago: Chicago Linguistic Society.
______ . 1972b. The Stem Formation of Hittite Nouns and Adjectives. Ph.D. diss., The University
of Chicago.
Bernabe, Alberto. 1983. Hethitica. AuOr 1: 3943.
Bernabe, Alberto, and Juan Antonio lvarez-Pedrosa. 2000. Historia y leyes de los hititas. Tex-
tos del imperio antiguo, El cdigo. Akal lOriente 3. Madrid: Akal.
Bittel, Kurt. 1970. Hattusha. Capital of the Hittites. New York: Oxford University Press.
Black, Jeremy A., Andrew George, and J. Nicholas Postgate. 1999. A Concise Dictionary of Ak-
kadian. Santag 10. Wiesbaden: Harrassowitz.
00-HittiteGrammar.indb 442 6/5/08 4:21:08 PM
443 References
Boley, Jacqueline. 1984. The Hittite hark-construction. Innsbrucker Beitrge zur Sprachwissen-
schaft 44. Innsbruck: Institut fr Sprachwissenschaft der Universitt Innsbruck.
______ . 1985. Notes on Hittite Place Word Syntax. Hethitica 6: 543.
______ . 1989. The Sentence Particles and the Place Words in Old and Middle Hittite. Inns-
brucker Beitrge zur Sprachwissenschaft 60. Innsbruck: Institut fr Sprachwissenschaft
der Universitt Innsbruck.
______ . 1993. The Hittite Particle -z l -za. Innsbrucker Beitrge zur Sprachwissenschaft 79.
Innsbruck: Institut fr Sprachwissenschaft der Universitt Innsbruck.
______ . 2001. Suggestions for the Original Function of -kan. Pp. 1928 in Anatolisch und In-
dogermanischlAnatolico e Indoeuropeo. Akten des Kolloquiums der Indogermanischen
Gesellschaft Pavia 22.25. September 1998, edited by O. Carruba and W. Meid. Innsbruck:
Institut fr Sprachwissenschaft der Universitt Innsbruck.
Brixhe, Claude. 1979. Le directif du vieux-hittite et ses ascendances indo-europeennes. Pp. 65
77 in Florilegium Anatolicum. mlanges oerts a Emmanuel Laroche, edited by E. Mas-
son. Paris: de Boccard.
Brock, Nadia van. 1964. Les themes verbaux redoublement du hittite et le verbe indo-europeen.
RHA 22 (75): 11965.
Bryce, Trevor R. 1998. The Kingdom of the Hittites. Oxford: Clarendon.
______ . 2002. Life and Society in the Hittite World. Oxford: Oxford University Press.
Buck, Carl Darling. 1949. A Dictionary of Selected Synonyms in the Principal Indo-European
Languages. Chicago: University of Chicago Press.
Burde, Cornelia. 1974. Hethitische medizinische Texte. Studien zu den Boazky-Texten 19.
Wies baden: Harrassowitz.
Bush, Frederick W. 1964. A Grammar of the Hurrian Language. Ph.D. diss., Brandeis University.
Carruba, Onofrio. 1964. Hethitisch -(a)ta, -(a)pa und die anderen Ortsbezugspartikeln. Or NS
33: 40536.
______ . 1966a. Das Beschwrungsritual fr die Gttin Wiuriyanza. Studien zu den Boazky-
Texten 2. Wiesbaden: Harrassowitz.
______ . 1966b. Die Verbalendungen auf -wani und -tani und das relative Alter der heth. Texte.
Die Sprache 12: 7989.
______ . 1968. Die I. und II. Pers. Plur. im Luwischen und im Lykischen. Die Sprache 14: 13
23.
______ . 1971. Hattusili II. SMEA 14: 7594.
______ . 197879. Sul genitivo in anatolico. Atti del Sodalizio Glottologico Milanese 20: 31
32.
______ . 1981. Pleneschreibung und Betonung im Hethitischen. KZ 95: 232248.
______ . 1986. Die 3. Pers. Sing. des Possessivpronomens im Luwischen. Pp. 4952 in Kani-
uwar. A Tribute to Hans G. Gterbock on his Seventy-fth Birthday, May 27, 1983, edited
by H. A. Honer Jr. and G. M. Beckman. Chicago: The Oriental Institute of the University
of Chicago.
______ . 1989. Review of J. J. S. Weitenberg, Die hethitischen U-Stmme (Amsterdam, 1984).
OLZ 84: 42227.
______ . 1991. Searching for Woman in Anatolian and Indo-European. Pp. 15581 in Perspec-
tives on Indo-European Language, Culture and Religion. Studies in Honor of Edgar C.
Polom, edited by R. Pearson. McLean, Virginia: Institute for the Study of Man.
00-HittiteGrammar.indb 443 6/5/08 4:21:09 PM
444 References
______ . 1992. Le notazioni dellagente animato nelle lingue anatoliche (e lergativo). Pp. 6198
in Per una grammatica ittita. Towards a Hittite Grammar, edited by O. Carruba. Pavia:
Iuculano.
______ . 1993a. Hethitisch kuinna- Frau; kuwansa- weiblich und anderes. HS 106: 26271.
______ . 1993b. Zur Datierung der ltesten Schenkungsurkunden und anonymen Tabarna-Siegel.
IM 43: 7185.
Carruba, Onofrio, Vladimir Souek, and Reinhard Sternemann. 1965. Kleine Bemerkungen zur
jngsten Fassung der hethitischen Gesetze. ArOr 33: 118.
Catsanicos, Jean. 1986. A propos des adjectifs hitt. u-mili- et ved. s-mya-: quelques re-
marques sur le traitement du groupe V-H
x
C la jointure des composes. BSL 81: 12180.
______ . 1996. Lapport de la bilingue de Hattusa la lexicologie hourrite. Pp. 197296 in Mari,
bla et les Hourrites. Dix ans de travaux, edited by J.-M. Durand. Paris: ditions Re-
cherche sur les Civilisations.
Comrie, Bernard. 1998. The Indo-European Linguistic Family: Genetic and Typological Perspec-
tives. Pp. 7497 in The Indo-European Languages, edited by A. G. Ramat and P. Ramat.
London: Routledge.
Cotticelli-Kurras, Paola. 1992a. Das hethitische Verbum sein. Syntaktische Untersuchungen.
Texte der Hethiter 17. Heidelberg: Winter.
______ . 1992b. Die hethitischen Nominalstze. Pp. 99136 in Per una grammatica ittita. To-
wards a Hittite Grammar, edited by O. Carruba. Pavia: Iuculano.
Daddi Pecchioli, Franca. 1975. Il azan(n)u nei testi di Hattusa. OA 14: 93136.
Dardano, Paola. 1997. Laneddoto e il racconto in eta antico-hittita. La cosidetta cronaca di
palazzo. Biblioteca di ricerche linguistiche e lologiche 43. Rome: Il Calamo.
______ . 2005. I costutti perifrastici con il verb ar(k)- dellittito: stato della questione e prospet-
tive di metodo. Or NS 74: 93113.
DeVries, Bert. 1967. The Style of Hittite Epic and Mythology. Ph.D. diss., Brandeis University.
Dinol, Belks. 1998. Beobachtungen ber die Bedeutung des hethitischen Musikinstruments
GI
uupal. Anatolica 24: 15.
Disterheft, Dorothy. 1984. Non-nal Verbs in Hittite. KZ 97: 22127.
Dressler, Wolfgang. 1968. Studien zur verbalen Pluralitt. Iterativum, Distributivum, Durati-
vum, Intensivum in der allgemeinen Grammatik, im Lateinischen und Hethitischen. s-
terreichische Akademie der Wissenschaften. Philosophisch-historische Klasse: Sitzungs be-
richte, 259l1. Vienna: Bhlau.
Dunkel, George. 1985. IE hortatory `y, `yte: Ved. ta . . . stvma, Hitt. eu=wa it, Hom.
. MSS 46: 4779.
Dunkel, George E. 1998. Back to the Future in Hittite. MSS 58: 5155.
Durand, Jean-Marie, and Emmanuel Laroche. 1982. Fragments Hittites du Louvre. Pp. 73107
in Mmorial Atatrk. tudes darchologie et de philologie anatoliennes. Paris: ditions
recherches sur les civilisations.
Durham, John W. 1976. Studies in Boazky Akkadian. Ph.D. diss., Harvard University.
Edel, Elmar. 1948. Neue keilschriftliche Umschreibungen gyptischer Namen aus den Boazky-
Texten. JNES 7: 1124.
______ . 1973. Hethitische Personnamen in hieroglyphischer Umschrift. Pp. 5970 in Festschrift
Heinrich Otten. 27. Dez. 1973, edited by E. Neu and C. Rster. Wiesbaden: Harrassowitz.
Ehelolf, Hans. 1936. Hethitische-akkadische Wortgleichungen. ZA 44: 17095.
00-HittiteGrammar.indb 444 6/5/08 4:21:10 PM
445 References
Eichner, Heiner. 1971. Urindogermanisch `K
w
e wenn im Hethitischen. MSS 29: 2746.
______ . 1973. Die Etymologie von heth. mur. MSS 31: 53107.
______ . 1974a. Indogermanische Chronik 20b II. Anatolisch. Die Sprache 20: 18386.
______ . 1974b. Untersuchungen zur hethitischen Deklination (Diss. Teildruck). Erlangen.
______ . 1975. Die Vorgeschichte des hethitischen Verbalsystem. Pp. 71103 in Flexion und
Wortbildung. Akten der V. Fachtagung der Indogermanischen Gesellschaft, Regensburg,
9.14. September 1973, edited by H. Rix. Wiesbaden: Reichert.
______ . 1979. Hethitisch gnuu, ginui, ginuin. Pp. 4161 in Hethitisch und Indoger-
manisch. Vergleichende Studien zur historischen Grammatik und zur Dialektgeographie,
edited by E. Neu and W. Meid. Innsbruck: Institut fr Sprachwissenschaft der Universitt
Innsbruck.
______ . 1980. Phonetik und Lautgesetze des Hethitischen ein Weg zu ihrer Entschlsse-
lung. Pp. 12065 in Akten der VI. Fachtagung der Indogermanischen Gesellschaft, Wien,
24.29. September 1978, edited by M. Mayrhofer, M. Peters, and O. Pfeier. Wiesbaden:
Reichert.
______ . 1985. Das Problem des Ansatzes eines urindogermanischen Numerus Kollektiv
(Komprehensiv). Pp. 13469 in Akten der VII. Fachtagung der Indogermanischen Ge-
sellschaft, Berlin, 20.25. Februar 1983, edited by B. Schlerath and V. Rittner. Wiesbaden:
Reichert.
______ . 1988. Anatolisch und Trilaryngalismus. Pp. 12351 in Die Laryngaltheorie, edited by
A. Bammesberger. Heidelberg: Winter.
______ . 1991. Review of J. J. S. Weitenberg, Die hethitischen U-Stmme (Amsterdam, 1984).
ZA 81: 15557.
______ . 1992. Anatolian. Pp. 2996 in Indo-European Numerals, edited by J. Gvozdanovi.
Berlin: Mouton de Gruyter.
Ertem, Hayri. 1973. Boazky metinlerinde geen corafya adlar dizini. Ankara niversitesi
Dil ve Tarih-Corafya Fakltesi Yaynlar 230. Ankara: Ankara niversitesi Basmevi.
Fincke, Jeanette. 2000. Augenleiden nach keilschriftlichen Quellen. Untersuchungen zur altori-
entalischen Medizin. Wrzburger medizin-historische Forschungen 70. Wrzburg: Knigs-
hausen & Neumann.
Forrer, Emil. 1922. Die Inschriften und Sprachen des Hatti-Reiches. ZDMG 76 (Neue Folge 1):
174269.
______ . 1922, 1926. Die Boghazki-Texte in Umschrift, Zweiter Band. Geschichtliche Texte.
Wissenschaftliche Verentlichungen der Deutschen Orient-Gesellschaft 42, Hefte 12.
Leipzig: Hinrichs.
______ . 1928. Ein siebenter Kasus im Alt-Kanisischen. Pp. 3035 in Altorientalische Studien.
Bruno Meissner zum sechzigsten Geburtstag am 25. April 1928 gewidmet von Freunden,
Kollegen und Schlern. Leipzig: Harrassowitz.
Forssman, Bernhard. 1994. Zu hethitisch ipand- und ipand-. Pp. 93106 in In honorem Holger
Pedersen. Kolloquium der Indogermanischen Gesellschaft vom 26. bis 28. Mrz 1993 in
Kopenhagen, 93106, edited by J. E. Rasmussen and B. Nielsen. Wiesbaden: Reichert.
Fortson, Benjamin. 1998. A New Study of Hittite -wa(r). JCS 50: 2134.
Fox, Barbara A., and Paul J. Hopper. 1994. Voice. Form and Function. Typological Studies in
Language 27. Amsterdam and Philadelphia: Benjamins.
Francia, Rita. 1995. Costruzione e valore del pronome possessivo enclitico di prima plurale in
hittito. SMEA 35: 93100.
00-HittiteGrammar.indb 445 6/5/08 4:21:10 PM
446 References
______ . 1996a. Funzioni sintattiche nei testi dell antico ittita: il locativo di meta e scopo e
laccusativo di direzione. IncLing 19: 13753.
______ . 1996b. Il pronome possessivo enclitico in antico ittita: alcune riessioni. VO 10: 209
59.
______ . 1997. Laccusativo de estensione nello spazio e nel tempo in ittita. Annali del Diparti-
mento di Studi del Mondo Classico e del Mediterraneo Antico, Napoli 19: 13945.
______ . 2001. La posizione degli aggettivi qualicativi nella frase ittita. Pp. 8191 in Anatolisch
und IndogermanischlAnatolico e Indoeuropeo. Akten des Kolloquiums der Indogermani-
schen Gesellschaft Pavia 22.25. September 1998, edited by O. Carruba and W. Meid.
Innsbruck: Institut fr Sprachwissenschaft der Universitt Innsbruck.
______ . 2002a. Lespressione della causa in ittita. Pp. 27784 in Anatolia antica. Studi in me-
moria di Fiorella Imparati, edited by S. de Martino and F. Pecchioli Daddi. Florence:
LoGisma.
______ . 2002b. Le funzioni sintattiche degli elementi avverbiali di luogo ittiti anda(n), ppa(n),
katta(n), katti-, peran, par, er, ar. Rome: Studia Asiana 1. Herder.
______ . 2005. Lineamenti di grammatica ittita. Studia Asiana 4. Rome: Herder.
Friedrich, Johannes. 1926. Staatsvertrge des atti-Reiches in hethitischer Sprache, 1. Teil. Mit-
teilungen der Vorderasiatisch-Aegyptischen Gesellschaft 31l1. Leipzig: Hinrichs.
______ . 1930. Staatsvertrge des atti-Reiches in hethitischer Sprache, 2. Teil. Mitteilungen
der Vorderasiatisch-Aegyptischen Gesellschaft 34l1. Leipzig: Hinrichs.
______ . 1952. Hethitisches Wrterbuch. Kurzgefate kritische Sammlung der Deutungen hethi-
tischer Wrter. 1st ed. Heidelberg: Winter.
______ . 1959. Die hethitischen Gesetze. Documenta et Monumenta ad Orienti Antiqui 7. Leiden:
Brill.
______ . 1960. Hethitisches Elementarbuch, 1. Teil. Kurzgefate Grammatik. 2nd ed. Indoger-
manische Bibliothek, 1. Reihe: Lehr- und Handbcher. Heidelberg: Winter.
______ . 1991. Kurzgefasstes Hethitisches Wrterbuch. Kurzgefasste kritische Sammlung der
Deutungen hethitischer Wrter. Indogermanische Bibliothek, 2. Reihe, Wrterbucher. Hei-
del berg: Winter.
Friedrich, Johannes, Annelies Kammenhuber, and Inge Homann. 1975. Hethitisches Wrter-
buch. 2nd ed. Indogermanische Bibliothek. 2. Reihe: Wrterbcher. Heidelberg: Winter.
Garca Ramon, Jose-Lus. 2002. Zu Verbalcharakter, morphologischer Aktionsart, und Aspekt
in der indogermanischen Rekonstruktion. Pp. 10536 in Indogermanische Syntax. Fragen
und Perspektiven, edited by H. Hettrich and J.-S. Kim. Wiesbaden: Reichert.
Garca Trabazo, Jose Virgilio. 2002. Textos religiosos hititas. Mitos, plegaries y rituales. Biblio-
teca de Ciencias Bblicas y Orientales 6. Madrid: Editorial Trotta.
Gardiner, Alan. 1957. Egyptian Grammar. 3rd ed. London: Oxford University Press.
Garrett, Andrew James. 1990a. Hittite Enclitic Subjects and Transitive Verbs. JCS 42: 22742.
______ . 1990b. The Origins of NP Split Ergativity. Lg 66: 26196 (Anatolian discussed on 265
80, 28791).
______ . 1990c. The Syntax of Anatolian Pronominal Clitics. Ph.D. diss., Harvard University.
______ . 1994. Relative Clause Syntax in Lycian and Hittite. Die Sprache 36: 2969.
______ . 1996. Wackernagels Law and Unaccusativity in Hittite. Pp. 85133 in Approaching
Second. Second Position Clitics and Related Phenomena, edited by A. L. Halpern and
A. M. Zwicky. Stanford, CA: CSLI.
00-HittiteGrammar.indb 446 6/5/08 4:21:11 PM
447 References
______ . 1998. Remarks on the Old Hittite Split Genitive. Pp. 155163 in Mr Curad. Studies in
Honor of Calvert Watkins, edited by J. Jasano, H. C. Melchert, and L. Oliver. Innsbruck:
Institut fr Sprachwissenschaft der Universitt Innsbruck.
Gelb, Ignace J., Thorkild Jacobsen, Benno Landsberger, A. Leo Oppenheim, Erica Reiner, Miguel
Civil, and Martha T. Roth, eds. 1956. The Assyrian Dictionary of The Oriental Institute of
The University of Chicago. Chicago: The Oriental Institute of the University of Chicago.
Gertz, Janet E. 1982. The Nominative-Accusative Neuter Plural in Anatolian. Ph. D. diss., Yale
University.
Gesenius, Wilhelm, Emil Kautsch, and Arthur Ernest Cowley. 1910. Gesenius Hebrew Gram-
mar. Translated by A. E. Cowley. 2nd English ed. Oxford: Clarendon.
Gessel, Ben H. L. van. 19982001. Onomasticon of the Hittite Pantheon, Parts 13. Handbuch
der Orientalistik 1l33: 13. Leiden: Brill.
Giorgieri, Mauro. 2000. Schizzo grammaticale della lingua hurrita. Pp. 171277 in La civilta dei
hurriti. Naples: Macchiaroli.
Girbal, Christian. 1990. Zur Grammatik des Mittani-Hurritischen. ZA 80: 93100.
Glocker, Jrgen. 1997. Das Ritual fr den Wettergott von Kuliwina. Eothen 6. Florence: Lo-
Gisma.
Goedegebuure, Petra M. 2002. KBo 17.17+: Remarks on an Old Hittite Royal Substitution Rit-
ual. JANER 2: 6173.
______ . 20023. The Hittite Distal Demonstrative ASI (UNI, ENI, etc.). Die Sprache 43: 132.
______ . 2006. A New Proposal for the Reading of the Hittite Numeral 1: ia-. Pp. 6588 in
The Life and Times of attuili III and Tutaliya IV. Proceedings of a Symposium held in
Honour of J. de Roos, 1213 December 2003, Leiden, edited by T. P. J. van den Hout and
C. H. van Zoest. Leiden: Nederlands Historisch-Archaeologisch Instituut te Istanbul.
Goetze, Albrecht. 1925. attuili. Der Bericht ber seine Thronbesteigung nebst den Parallel-
texten. Mitteilungen der Vorderasiatisch-Aegyptischen Gesellschaft 29. Leipzig: Hinrichs.
______ . 1927. Madduwatta. Mitteilungen der Vorderasiatisch-Aegyptischen Gesellschaft 32.
Leipzig: Hinrichs.
______ . 1930a. Die Pestgebete des Murili. KlF 1: 161251.
______ . 1930b. Neue Bruchstcke zum grossen Text des attuili und den Paralleltexten. Mit-
teilungen der Vorderasiatisch-Aegyptischen Gesellschaft 34. Leipzig: Hinrichs.
______ . 1930c. Verstreute Boghazki-Texte. Marburg an der Lahn: Published by the editor.
______ . 1933a. Die Annalen des Murili. Mitteilungen der Vorderasia tisch-Aegyptischen Ge-
sellschaft 38. Leipzig: Hinrichs.
______ . 1933b. ber die Partikeln -za, -kan und -an der hethitischen Satzverbindung. ArOr 5:
138.
______ . 1937. Transfer of Consonantal Stems to the Thematic Declension in Hittite. Pp. 48895
in Mlanges linguistiques oerts a M. Holger Pedersen, edited by L. Hjelmslev. Copenha-
gen: Levin & Munksgaard.
______ . 1938. The Hittite Ritual of Tunnawi. American Oriental Series 14. New Haven, CT:
American Oriental Society.
______ . 1951. On the Hittite Words for Year and the Seasons and for Night and Day. Lg
27: 467476.
______ . 1954. Review of J. Friedrich, Hethitisches Wrterbuch, Lieferungen 14 (Heidelberg,
195254). Lg 30: 4015.
00-HittiteGrammar.indb 447 6/5/08 4:21:11 PM
448 References
______ . 1963. Postposition and Preverb in Hittite. JCS 17: 98111.
______ . 1968. The Predecessors of uppiluliuma of Hatti and the Chronology of the Ancient
Near East. JCS 22: 4650.
Goetze, Albrecht, and Holger Pedersen. 1934. Murilis Sprachlhmung. Det Danske Videnskab-
ernes Selskab. Hist.-lol. Meddelelser 21l1. Copenhagen: Levin & Munksgaard.
Goodwin, William W. 1965. A Greek Grammar. London: MacMillan.
Groddek, Detlev. 1999. CTH 331: Mythos vom verschwundenen Wettergott oder Aitiologie der
Zerstrung Lizinas? ZA 89: 3649.
Gurney, Oliver R. 1940. Hittite Prayers of Mursili II. AAA 27: 3163.
______ . 1990. The Hittites. 2nd ed. Baltimore: Penguin.
Gterbock, Hans Gustav. 1940. Siegel aus Boasky I. Archiv fr Orientforschung Beiheft 5.
Berlin: Weidner.
______ . 1943. Review of J. Friedrich, Hethitisches Elementarbuch, 1. Teil. Kurzgefate Gram-
matik (Heidelberg, 1940). Or NS 12: 15155.
______ . 1945. The Vocative in Hittite. JAOS 65: 24857.
______ . 1952. The Song of Ullikummi. Revised Text of the Hittite Version of a Hurrian Myth.
New Haven: American Schools of Oriental Research.
______ . 1955. Zu einigen hethitischen Komposita. Pp. 6368 in Corolla linguistica. Festschrift
Ferdinand Sommer zum 80. Geburtstag am 4. Mai 1955 dargebracht von Freunden, Sch-
lern und Kollegen. Heidelberg: Winter.
______ . 1956. The Deeds of uppiluliuma I as Told by His Son, Murili II. JCS 10: 4168,
7598, 10730.
______ . 1957. Review of J. Friedrich, Hethitisches Wrterbuch (Heidelberg, 1952). Oriens 10:
35062.
______ . 1958. Kane and Nea, Two Forms of One Anatolian Place Name. Eretz-Israel 5 (Mazar
Volume): 4650.
______ . 1959. Towards a Denition of the Term Hittite. Oriens 10: 23339.
______ . 1964. Lexicographical Notes II. RHA 22 (74): 95113.
______ . 1967. The Hittite Conquest of Cyprus Reconsidered. JNES 26: 7381.
______ . 1970. The Predecessors of uppiluliuma again. JNES 29: 7377.
______ . 1983. Noch einmal die Formel parnaea uwaizzi. Or NS 52: 7380.
______ . 1992. Ist das hethitische Wort fr Frau gefunden? HS 105: 13.
______ . 1995a. Reections on the Musical Instruments arkammi, galgalturi and uupal in Hit-
tite. Pp. 5772 in Studio historiae ardens. Ancient Near Eastern Studies Presented to Philo
H. J. Houwink ten Cate on the Occasion of his 65th Birthday, edited by T. P. J. van den Hout
and J. de Roos. Istanbul: Institut historique et archeologique neerlandais de Stamboul.
______ . 1995b. Resurrecting the Hittites. Pp. (4: ) 276578 in Civilizations of the Ancient Near
East, edited by J. M. Sasson. New York: Scribners Sons.
______ . 1995c. The Hittite Word for Woman Again. HS 108: 1215.
Gterbock, Hans Gustav, Harry A. Honer Jr., and Theo P. J. van den Hout. 1989. The Hittite
Dictionary of the Oriental Institute of the University of Chicago. Chicago: The Oriental
Institute of the University of Chicago.
Gterbock, Hans Gustav, and Theo P. J. van den Hout. 1991. The Hittite Instruction for the Royal
Bodyguard. Assyriological Studies 24. Chicago: The Oriental Institute of the University
of Chicago.
00-HittiteGrammar.indb 448 6/5/08 4:21:12 PM
449 References
Haas, Volkert. 1970. Der Kult von Nerik. Studia Pohl 4. Rome: Ponticium Institutum Bibli-
cum.
______ . 1994. Geschichte der hethitischen Religion. Handbuch der Orientalistik 1l15. Leiden:
Brill.
Haas, Volkert, and H. Thiel. 1978. Die Beschwrungsrituale der Allaitura()i und verwandte
Texte. Alter Orient und Altes Testament 31. Neukirchen-Vluyn: Neukirchener Verlag.
Hagenbuchner, Albertine. 1989. Die Korrespondenz der Hethiter, 2. Teil. Texte der Hethiter 16.
Heidelberg: Winter.
Hagenbuchner-Dresel, Albertine. 1999. Bemerkungen zu krzlich edierten Briefen. ZA 89: 50
64.
Hahn, E. Adelaide. 1969. Naming Constructions in Some Indo-European Languages. Philologi-
cal Monographs of the American Philological Association 27. Cleveland: American Philo-
logical Association.
Hale, Mark. 1987. Notes on Wackernagels Law in the language of the Rigveda. Pp. 3859 in
Studies in Memory of Warren Cowgill. Papers of the Fourth East Coast Indo-European
Conference, Cornell Univ. 1985, edited by C. Watkins. Berlin: de Gruyter.
Hallo, William W., and K. Lawson Younger. 2003. The Context of Scripture, Volume 1. Canoni-
cal Compositions from the Biblical World. Leiden: Brill.
Hararson, Jon Axel. 1994. Der Verlust zweier wichtiger Flexionskategorien im Uranatolischen.
HS 107: 3041.
Hart, Gillian R. 1971. The Hittite Particle -PAT. TPS: 94162.
______ . 1980. Some Observations on Plene-Writing in Hittite. BSOAS 43: 117.
______ . 1983. Problems of Writing and Phonology in Cuneiform Hittite. TPS: 100154.
Hawkins, J. David. 1979. The Origin and Dissemination of Writing in Western Asia. Pp. 12866
in The Origins of Civilization, edited by P. R. S. Moorey. Oxford: Clarendon.
______ . 1986. Writing in Anatolia: Imported and Indigenous Systems. World Archaeology 17:
36376.
______ . 2000. Corpus of Hieroglyphic Luwian Inscriptions. Untersuchungen zur indoger mani-
schen Sprach- und Kulturwissenschaft 8.1. Berlin: de Gruyter.
______ . 2003. Chapter Four: Scripts and Texts. Pp. 12869 in The Luwians, edited by H. C.
Melchert. Leiden: Brill.
Hawkins, J. David, Anna Morpurgo Davies, and Gnter Neumann. 1974. Hittite Hieroglyphs
and Luwian. New Evidence for the Connection. Nachrichten der Akademie der Wissen-
schaften in Gttingen, Philologisch-historische Klasse 73 No. 6. Gttingen: Vandenhoeck
& Ruprecht.
Hecker, Karl. 1992. Zur Herkunft der hethitischen Keilschrift. Pp. 5363 in Uluslararas 1. Hiti-
toloji Kongresi Bildirileri (1921 Temmuz 1990). orum, Turkey: Uluslararas orum Hitit
Festivali Komitesi Bakanl.
______ . 1996. Zur Herkunft der hethitischen Keilschrift. Pp. 291303 in Richard F. S. Starr
Memorial Volume, edited by D. I. Owen and G. Wilhelm. Bethesda, MD: CDL.
Helck, Wolfgang. 1963. Uri-Teup in gypten. JCS 17: 8797.
Held, Warren H., Jr. 1957. The Hittite Relative Sentence. Language Dissertations 55. Baltimore:
Linguistic Society of America.
Held, Warren H., Jr., William R. Schmalstieg, and Janet E. Gertz. 1987. Beginning Hittite. Co-
lumbus, OH: Slavica.
00-HittiteGrammar.indb 449 6/5/08 4:21:12 PM
450 References
Hock, Hans Henrich. 2002. Vedic ta . . . stvma: Subordinate, Coordinate, or What? Pp. 89
102 in Indo-European Perspectives, edited by M. R. V. Southern. Washington, DC: Insti-
tute for the Study of Man.
Homann, Inge. 1984. Der Erla Telipinus. Texte der Hethiter 11. Heidelberg: Winter.
______ . 1992. Das hethitische Wort fr Sohn. Pp. 28993 in Hittite and Other Anatolian and
Near Eastern Studies in Honour of Sedat Alp, edited by H. Otten, E. Akurgal, H. Ertem,
and A. Sel. Ankara: Trk Tarih Kurumu.
Honer, Harry A., Jr. 1965. The Elkunirsa Myth Reconsidered. RHA 23 (76): 516.
______ . 1966. Composite Nouns, Verbs, and Adjectives in Hittite. Or NS 35: 377402.
______ . 1967a. Second Millennium Antecedents to the Hebrew OB. JBL 86: 385401.
______ . 1967b. Ugaritic pwt : A Term from the Early Canaanite Dyeing Industry. JAOS 87: 300
303.
______ . 1968a. Birth and Namegiving in Hittite Texts. JNES 27: 198203.
______ . 1968b. Review of Carruba, Das Beschwrungsritual fr die Gttin Wiuriyanza [StBoT
2] (Wiesbaden, 1966). JAOS 88: 53134.
______ . 1969. On the Use of Hittite -za in Nominal Sentences. JNES 28: 22530.
______ . 1970. Remarks on the Hittite Version of the Naram-Sin Legend. JCS 23: 1722.
______ . 1971. Hittite ega- and egan-. JCS 24: 3136.
______ . 1972. Review of P. Houwink ten Cate, Records of the Early Hittite Empire (c. 1450
1380 u.t.) (Istanbul, 1970). JNES 31: 2935.
______ . 1973a. Studies of the Hittite Particles, I. Review of O. Carruba, Die satzeinleitenden
Partikeln in den indogermanischen Sprachen Anatoliens (Rome, 1969). JAOS 93: 520
26.
______ . 1973b. The Hittite Particle -PAT. Pp. 99118 in Festschrift Heinrich Otten. 27. Dez.
1973, edited by E. Neu and C. Rster. Wiesbaden: Harrassowitz.
______ . 1974. The Arzana House. Pp. 11322 in Anatolian Studies Presented to Hans Gustav
Gterbock on the Occasion of His 65th birthday, edited by K. Bittel, P. H. J. Houwink ten
Cate, and E. Reiner. Istanbul: Nederlands Historisch-Archaeologisch Institut in Het Nabije
Oosten.
______ . 1975. Hittite Mythological Texts: A Survey. Pp. 13645 in Unity and Diversity. Essays
in the History, Literature, and Religion of the Ancient Near East, edited by H. Goedicke
and J. J. M. Roberts. Baltimore: Johns Hopkins University Press.
______ . 1977. Studies in Hittite Vocabulary, Syntax and Style: Hommage M. Emmanuel La-
roche. JCS 29: 15156.
______ . 1982. Hittite mn and nman. Pp. 3845 in Investigationes philologicae et comparati-
vae. Gedenkschrift fr Heinz Kronasser, edited by E. Neu. Wiesbaden: Harrassowitz.
______ . 1983a. A Prayer of Murili II about his Stepmother. JAOS 103: 18792.
______ . 1983b. Review of J. Friedrich and A. Kammenhuber, Hethitisches Wrterbuch
2
Il45
(arazena-arzanatar) (Heidelberg 197980). BiOr 40: 40717.
______ . 1985. Review of H. Otten, Die Apologie Hattusilis III. [StBoT 24] (Wiesbaden, 1981).
JAOS 105: 33739.
______ . 1986. Studies in Hittite Grammar. Pp. 8394 in Kaniuwar. A Tribute to Hans G. Gter-
bock on his Seventy-fth Birthday, May 27, 1983, edited by H. A. Honer Jr. and G. M.
Beckman. Chicago: The Oriental Institute of the University of Chicago.
______ . 1987. Paskuwattis Ritual Against Sexual Impotence (CTH 406). AuOr 5: 27187.
00-HittiteGrammar.indb 450 6/5/08 4:21:13 PM
451 References
______ . 1988a. A Scene in the Realm of the Dead. Pp. 19199 in A Scientic Humanist. Studies
in Memory of Abraham Sachs, edited by E. Leichty, M. deJ. Ellis, and P. Gerardi. Philadel-
phia: University Museum.
______ . 1988b. The Song of Silver. Pp. 14366 in Documentum Asiae Minoris antiquae. Fest-
schrift fr Heinrich Otten zum 75. Geburtstag, edited by E. Neu and C. Rster. Wiesbaden:
Harrassowitz.
______ . 1990. Hittite Myths. Writings from the Ancient World 2. Atlanta: Scholars Press.
______ . 1992a. Advice to a King. Pp. 295304 in Hittite and Other Anatolian and Near Eastern
Studies in Honour of Sedat Alp, edited by H. Otten, E. Akurgal, H. Ertem, and A. Sel.
Ankara: Trk Tarih Kurumu.
______ . 1992b. Studies in the Hittite Particles, II: On Some Uses of -kan. Pp. 13752 in Per una
grammatica ittita. Towards a Hittite Grammar, edited by O. Carruba. Pavia: Iuculano.
______ . 1993. Hittite iwar and Related Modes of Expressing Comparison. IM 43: 3951.
______ . 1994. The Hittites. Pp. 12756 in Peoples of the Old Testament World, edited by A. J.
Hoerth, G. L. Mattingly, and E. M. Yamauchi. Grand Rapids, MI: Baker.
______ . 1995a. About Questions. Pp. 87104 in Studio historiae ardens. Ancient Near Eastern
Studies Presented to Philo H. J. Houwink ten Cate on the Occasion of His 65th Birthday,
edited by T. P. J. van den Hout and J. de Roos. Istanbul: Institut historique et archeologique
neerlandais de Stamboul.
______ . 1995b. Hittite immakku: moreover or just now? Pp. 19194 in Atti del II Congresso
Internazionale di Hittitologia, edited by O. Carruba, M. Giorgieri, and C. Mora. Pavia:
Iuculano.
______ . 1995c. The Stem of the Hittite Word for House. HS 108: 19294.
______ . 1996a. Hittite Equivalents of Old Assyrian kumrum and epattum. WZKM 86: 15156.
______ . 1996b. Review of J. Boley, The Hittite Particle -z l -za (Innsbruck, 1993). BiOr 53:
75061.
______ . 1996c. Syntactical Perspectives on the Hittite Laws. Unpublished paper read at the
206th Annual Meeting of the American Oriental Society, Philadelphia, PA.
______ . 1997a. Hittites. Pp. 8488 in volume 3 of The Oxford Encyclopedia of Archaeology in
the Near East, edited by E. M. Meyers. New York: Oxford University Press.
______ . 1997b. On Safari in Hittite Anatolia. Pp. 521 in Studies in Honor of Jaan Puhvel, Part
One. Ancient Languages and Philology, edited by D. Disterheft, M. Huld, and J. A. C.
Greppin. Washington, DC: Institute for the Study of Man.
______ . 1997c. On the Hittite Use of Sumerian BAL in the Expression BAL-nu-. ArAn 3: 191
98.
______ . 1997d. The Laws of the Hittites. A Critical Edition. Documenta et Monumenta Orientis
Antiqui 23. Leiden: Brill.
______ . 1998a. From the Disciplines of a Dictionary Editor. JCS 50: 3544.
______ . 1998b. Hittite Myths. 2nd ed. Writings from the Ancient World 2. Atlanta: Scholars
Press.
______ . 1998c. Name, Namengebung. C. Bei den Hethitern. Pp. 11621 in Reallexikon der As-
syriologie, vol. 9, edited by E. Ebeling, B. Meissner, E. Weidner, W. von Soden, and D. O.
Edzard. Berlin: de Gruyter.
______ . 1998d. On the Denominative Verb arwe-. Pp. 27584 in Mr Curad. Studies in Honor
of Calvert Watkins, edited by J. Jasano, H. C. Melchert, and L. Oliver. Innsbruck: Institut
fr Sprachwissenschaft der Universitt Innsbruck.
00-HittiteGrammar.indb 451 6/5/08 4:21:13 PM
452 References
______ . 2000. The Treaty of Tudaliya IV with Kurunta of Taruntaa on the Bronze Tablet
Found in attua (2.18). Pp. 100106 in volume 2 of The Context of Scripture, edited by
W. W. Hallo and K. L. Younger. Leiden: Brill.
______ . 2001. Some Thoughts on Merchants and Trade in the Hittite Kingdom. Pp. 17989
in Kulturgeschichten. Altorientalistische Studien fr Volkert Haas zum 65. Geburtstag,
edited by T. Richter, D. Prechel, and J. Klinger. Saarbrcken: Saarbrcker Druckerei und
Verlag.
______ . 2002. Before and After: Space, Time, Rank and Causality. Pp. 16371 in Silva Ana-
tolica. Anatolian Studies Presented to Maciej Popko on the Occasion of His 65th Birthday,
edited by P. Taracha. Warsaw: Agade.
______ . 20023. Hittite a-a-ma. Die Sprache 43: 8087.
______ . 2003a. Daily Life among the Hittites. Pp. 95120 in Life and Culture in the Ancient
Near East, edited by R. Averbeck, M. W. Chavalas, and D. B. Weisberg. Bethesda, MD:
CDL.
______ . 2003b. On a Hittite Lexicographic Project. Review of J. Puhvel, Hittite Etymological
Dictionary, Vol. 5: Words beginning with L. Indices to Volumes 15 (Berlin, 2001). JAOS
123: 61725.
______ . 2005. Placenta, Colostrum and Meconium in Hittite. Pp. 33758 in Antiquus Oriens.
Mlanges oerts au Professeur Ren Lebrun, volume 1, edited by M. Mazoyer and O. Casa-
bonne. Paris: Harmattan.
______ . 2006. The Hittite Degenitival Adjectives iela-, 2-ela- and apella-. Pp. 18997 in The
Life and Times of attuili III and Tutaliya IV. Proceedings of a Symposium Held in Hon-
our of J. de Roos, 1213 December 2003, Leiden, edited by T. P. J. van den Hout. Leiden:
Nederlands Historisch-Archaeologisch Instituut te Istanbul.
______ . 2007. On Higher Numbers in Hittite. Pp. 37785 in VI Congresso Internazionale di
Ittitologia Roma, 59 settembre 2005, edited by A. Archi and R. Francia. Studi Micenei ed
Egeo-Anatolici 49. Rome: CNR Istituto de Studi sulle Civilt delEgeo ed del Vicino
Oriente.
______ . forthcoming. A Grammatical Prole of the Middle Hittite Maat Texts in la-a-ma-an-
te-et na-ak-ki-i. Studia Anatolica in memoriam Erich Neu scripta, edited by C. Rster,
J. Catsanicos, and R. Lebrun. Louvain-la-Neuve: Peeters.
Honer, Harry A., Jr., and H. Craig Melchert. 2002. A Practical Approach to Verbal Aspect in
Hittite. Pp. 37790 in Anatolia antica. Studi in memoria di Fiorella Imparati, edited by
S. de Martino and F. Pecchioli Daddi. Florence: LoGisma.
Holland, Gary B. 1986. Nominal Sentences and the Origin of Absolute Constructions in Indo-
European. KZ 99: 16384.
Hout, Theo P. J. van den. 1991. A Tale of Tissaruli(ya): A Dramatic Interlude in the Hittite
KI.LAM Festival? JNES 50: 193202.
______ . 1992. Remarks on Some Hittite Double Accusative Constructions. Pp. 275304 in
Per una grammatica ittita. Towards a Hittite Grammar, edited by O. Carruba. Pavia:
Iuculano.
______ . 2001. Neuter Plural Subjects and Nominal Predicates in Anatolian. Pp. 16792 in Ana-
tolisch und IndogermanischlAnatolico e Indoeuropeo. Akten des Kolloquiums der Indoger-
manischen Gesellschaft Pavia 22.25. September 1998, edited by O. Carruba and W. Meid.
Innsbruck: Institut fr Sprachwissenschaft der Universitt Innsbruck.
00-HittiteGrammar.indb 452 6/5/08 4:21:14 PM
453 References
______ . 2003. Studies in the Hittite Phraseological Construction. I. Its Syntactic and Semantic
Properties. Pp. 177204 in Hittite Studies in Honor of Harry A. Honer Jr. on the Oc-
casion of His 65th Birthday, edited by G. M. Beckman, R. H. Beal, and J. G. McMahon.
Winona Lake, IN: Eisenbrauns.
______ . 2006. Institutions, Vernaculars, Publics. Pp. 22162 in Margins of Writing. Origins
of Cultures, ed. by S. L. Sanders. Chicago: The Oriental Institute of the University of
Chicago.
______ . forthcoming. Studies in the Hittite Phraseological Construction, II: Its Origin in la-a-ma-
an-te-et na-ak-ki-i. Studia Anatolica in memoriam Erich Neu scripta, edited by C. Rster,
J. Catsanicos, and R. Lebrun. Louvain-la-Neuve: Peeters.
Houwink ten Cate, Philo H. J. 1967. The Ending -d of the Hittite Possessive Pronoun. RHA 24
(79): 12332.
______ . 1970. The Records of the Early Hittite Empire (c. 14501380 u.t.). Uitgaven van
het Nederlands Historisch-Archaeologisch Instituut te Istanbul 26. Istanbul: Nederlands
Historisch-Archaeologisch Instituut in het Nabije Oosten.
______ . 1973a. Impersonal and Reexive Constructions of the Predicative Participle in Hit-
tite. Pp. 199395 in Symbolae Biblicae et Mesopotamicae Francisco Mario Theodoro de
Liagre Bhl Dedicatae, edited by M. A. Beek, A. A. Kampman, C. Nijland, and J. Ryck-
mans. Leiden: Brill.
______ . 1973b. The Particle -a and Its Usage with Respect to the Personal Pronoun. Pp. 11940
in Festschrift Heinrich Otten. 27. Dez. 1973, edited by E. Neu and C. Rster. Wiesbaden:
Harrassowitz.
______ . 1984. The History of Warfare according to Hittite Sources: The Annals of Hattusilis I
(Part II). Anatolica 11: 4783.
Houwink ten Cate, Philo H. J., and Folke Josephson. 1967. Muwatallis Prayer to the Storm-God
of Kummanni (KBo XI 1). RHA 25 (81): 10140.
Hrozny, Friedrich. 1915. Die Lsung des hethitischen Problems. MDOG 56: 1750.
______ . 1917. Die Sprache der Hethiter. Boghazki-Studien 12. Leipzig: Hinrichs.
Huehnergard, John. 2005a. A Grammar of Akkadian. 2nd ed. Harvard Semitic Studies 45. Wi-
nona Lake, IN: Eisenbrauns.
______ . 2005b. Key to A Grammar of Akkadian. 2nd ed. Harvard Semitic Studies 46. Winona
Lake, IN: Eisenbrauns.
Hutter, Manfred. 1988. Behexung, Entshnung und Heilung. Das Ritual der Tunnawiya fr ein
Knigspaar aus mittelhethitischer Zeit (KBo XXI 1 KUB IX 34 KBo XXI 6). Orbis
biblicus et orientalis 82. Gttingen: Vandenhoeck & Ruprecht.
Ivanov, V. V. 1963. Xettskij jazyk. Moscow: Izdatelstvo vostonoj literatury.
Jakob-Rost, Liane. 1972. Das Ritual der Malli aus Arzawa gegen Behexung (KUB XXIV 9+).
Texte der Hethiter 2. Heidelberg: Winter.
Jamison, Stephanie. 1979. The Case of the Agent in Indo-European. Die Sprache 25: 12943.
Jasano, Jay H. 1973. The Hittite ablative in -anz(a). MSS 31: 12328.
______ . 2003. Hittite and the Indo-European Verb. Oxford: Oxford University Press.
Jie, Jin. 1994. A Complete Retrograde Glossary of the Hittite Language. Uitgaven van het Ned-
erlands Historisch-Archaeologisch Instituut te Istanbul 71. Leiden: Nederlands Instituut
voor het Nabije Oosten.
Josephson, Folke. 1967. Pronominal Adverbs of Anatolian: Formation and Function. RHA 24
(79): 13354.
00-HittiteGrammar.indb 453 6/5/08 4:21:15 PM
454 References
______ . 1972. The Function of the Sentence Particles in Old and Middle Hittite. Acta Universi-
tatis Upsaliensis. Studia Indoeuropaea Upsaliensia 2. Uppsala: Skriv Service AB.
______ . 1976. On the Function of the Gothic Preverb ga-. IF 81: 15275.
______ . 1995. Directionality in Hittite. Analecta Indoeuropaea Cracoviensia 2: 16576.
______ . 1997. Joint Celtic, Italic and Anatolian Evidence for the Expression of Deixis, Direc-
tionality. Pp. 4952 in Miscellanea Celtica in memoriam Heinrich Wagner, edited by S. M.
Mathuna and A. . Corrain. Uppsala: Uppsala University Press.
______ . 2003a. Semantics and Typology of Hittite -ant-. Pp. 91118 in Indo-European Word
Formation, edited by J. Clackson and B. A. Olsen. Copenhagen: Museum Tusculanum.
______ . 2003b. The Hittite Reexive Construction in a Typological Perspective. Pp. 21132 in
Language in Time and Space. A Festschrift for Werner Winter on the Occasion of His 80th
Birthday, edited by B. L. M. Bauer and G.-J. Pinault. Berlin: Mouton de Gruyter.
______ . 2004. Singulative and Agentive in Hittite and Germanic. Pp. 25762 in Per aspera
ad asteriscos. Studia Indogermanica in honorem Jens Elmegrd Rasmussen sexagenarii
Idibus Martiis anno MMIV, edited by A. Hyllested. Innsbruck: Institut fr Sprachwissen-
schaft der Universitt Innsbruck.
Kammenhuber, Annelies. 1954. Studien zum hethitischen Innitivsystem IIII. MIO 2: 4477,
24565, 40344.
______ . 1955a. Studien zum hethitischen Innitivsystem IVV. MIO 3: 3157, 34577.
______ . 1955b. Zur Genese der hethitischen -rl-n-Heteroklitica. Pp. 97106 in Corolla linguis-
tica. Festschrift Ferdinand Sommer zum 80. Geburtstag am 4. Mai 1955 dargebracht von
Freunden, Schlern und Kollegen. Heidelberg: Winter.
______ . 1969a. Die Sprachstufen des Hethitischen. KZ 83: 25689.
______ . 1969b. Hethitisch, Palaisch, Luwisch und Hieroglyphenluwisch. Pp. 119357 in Alt-
kleinasiatische Sprachen, edited by B. Spuler. Leiden: Brill.
______ . 1969c. Hethitisch, Palaisch, Luwisch, Hieroglyphenluwisch und Hattisch. Altkleinasi-
atische Indices zum Handbuch der Orientalistik. MSS Beiheft 4. Munich.
______ . 1979a. Direktiv, Terminativ undloder Lokativ im Hethitischen. Pp. 11542 in Hethi-
tisch und indogermanisch. Vergleichende Studien zur historischen Grammatik und zur Dia-
lektgeographie, edited by E. Neu and W. Meid. Innsbruck: Institut fr Sprachwissenschaft
der Universitt Innsbruck.
______ . 1979b. Zwischenbilanz zu den hethitischen Enklitikaketten. Pp. 18596 in Florilegium
Anatolicum. Mlanges oerts a Emmanuel Laroche. Paris: de Boccard.
______ . 1985. Zum Modus Injunktiv und zum Drei-Genus-System im Ur-Indogermanischen
(c. 30002500 v.Chr.). Pp. 43566 in Studia linguistica diachronica et synchronica Wer-
ner Winter sexagenario, edited by U. Pieper and G. Stickel. Berlin: de Gruyter.
______ . 1993. Kleine Schriften zum Altanatolischen und Indogermanischen. 2 vols. Texte der
Hethiter 19. Heidelberg: Winter.
Kassian, Alexei S., Andrej Korolv, and Andrej V. Sideltsev. 2002. Hittite Funerary Ritual.
alli watai. Alter Orient und Altes Testament 288. Mnster: Ugarit Verlag.
Kemmer, Suzanne. 1993. The Middle Voice. Amsterdam: Benjamins.
Kimball, Sara E. 1983. Hittite Plene Writing. Ph.D. diss., University of Pennsylvania.
______ . 1999. Hittite Historical Phonology. Innsbrucker Beitrge zur Sprachwissenschaft 95.
Innsbruck: Institut fr Sprachwissenschaft der Universitt Innsbruck.
00-HittiteGrammar.indb 454 6/5/08 4:21:15 PM
455 References
Klengel, Horst, and Evelyn Klengel. 1975. Die Hethiter und ihre Nachbarn. Eine Kultur-
geschichte Kleinasiens von atal Hyk bis zu Alexander dem Groen. Leipzig: Koehler
and Amelang.
Klinger, Jrg. 1992. Fremde und Aussenseiter in Hatti. Pp. 187212 in Aussenseiter, edited by
V. Haas. Konstanz: Universittsverlag Konstanz.
______ . 1996. Untersuchungen zur Rekonstruktion der hattischen Kultschicht. Studien zu den
Boazky-Texten 37. Wiesbaden: Harrassowitz.
______ . 1998. Wer lehrte die Hethiter das Schreiben? Zur Palographie frher Texte in akka-
discher Sprache aus Boazky: Skizze einiger berlegungen und vorluger Ergebnisse.
Pp. 36576 in Uluslararas Hititoloji Kongresi bildirileri. orum, 1622 Eyll, 1996. Acts
of the IIIrd International Congress of Hittitology, orum, September 1622, 1996, edited
by S. Alp and A. Sel. Ankara: Uyum Ajans.
Kloekhorst, Alwin. forthcoming. The Hittite 2 pl.-ending -sten(i). In Atti del 6 Congresso Inter-
nazionale di Ittitologia, edited by A. Archi and R. Francia. Rome: Herder.
Knudtzon, Jrgen Alexander, Sophus Bugge, and Alf Torp. 1902. Die zwei Arzawa-Briefe. Die
ltesten Urkunden in indogermanischer Sprache. Leipzig: Hinrichs.
Koroec, Viktor. 1931. Hethitische Staatsvertrge. Leipziger rechtswissenschaftliche Studien
60. Leipzig: Weicher.
Koak, Silvin. 1993. Die Stadtwerke von Hattua. Linguistica 33: 10712.
______ . 1996. Ein hethitisches ghost word entgeistert. AoF 23: 9597.
Kronasser, Heinz. 1956. Vergleichende Laut- und Formenlehre des Hethitischen. Heidelberg:
Winter.
______ . 1966. Etymologie der hethitischen Sprache I. Wiesbaden: Harrassowitz.
Khne, Cord. 1988. ber die Darstellung der hethitischen Reexivpartikel -z, besonders in post-
vokalischer Position. Pp. 20333 in Documentum Asiae Minoris antiquae. Festschrift fr
Heinrich Otten zum 75. Geburtstag, edited by E. Neu and C. Rster. Wiesbaden: Harras-
sowitz.
Khne, Cord, and Heinrich Otten. 1971. Der augamuwa-Vertrag. Studien zu den Boazky-
Texten 16. Wiesbaden: Harrassowitz.
Kmmel, Hans Martin. 1967. Ersatzrituale fr den hethitischen Knig. Studien zu den Boazky-
Texten 3. Wiesbaden: Harrassowitz.
Kuryowicz, Jerzy. 1958. Le hittite. Pp. 21651 in Proceedings of the Eighth International Con-
gress of Linguists (1957), edited by E. Sivertsen. Oslo: Oslo University Press.
Labat, Rene. 1932. LAkkadien de Boghaz-Ki. tude sur la langue des lettres, traits et vocabu-
laires akkadiens trouves a Boghaz-Ki. Bordeaux: Delmas.
______ . 1988. Manuel dpigraphie akkadienne. 6th ed., edited and expanded by Florence
Malbran-Labat. Paris: Geuthner.
Laroche, Emmanuel. 1947. Recherches sur les noms des dieux hittites (= RHA 7 1947). Paris:
Maisonneuve.
______ . 1951a. Fragments hittites de Geneve. RA 45: 13136, 18494.
______ . 1951b. Recueil donomastique hittite. Paris: Klincksieck.
______ . 1952. Fragments hittites de Geneve (2). RA 46: 4250.
______ . 1955. Onomastique hittite: additions et corrections. RHA 13 (57): 4158.
______ . 1956. Hittite -ima-, indo-europeen -m-. BSL 52: 7282.
______ . 1957a. tudes de vocabulaire VI. RHA 15 (60): 929.
00-HittiteGrammar.indb 455 6/5/08 4:21:16 PM
456 References
______ . 1957b. Notes de toponymie anatolienne. Pp. 17 in . Gedenkschrift
Paul Kretschmer, 2. Mai 18669. Mrz 1956, edited by H. Kronasser. Wiesbaden: Harras-
sowitz.
______ . 1958. Comparaison du louvite et du lycien. BSL 53: 15997.
______ . 1959. Dictionnaire de la langue louvite. Bibliotheque archeoligique et historique de
linstitut franais darcheologie dIstanbul 6. Paris: Maisonneuve.
______ . 1960. Les hiroglyphes hittites, I. Paris: CNRS.
______ . 1962. Un ergatif en indo-europeen dAsie Mineure. BSL 57: 2343.
______ . 1964. La priere hittite: vocabulaire et typologie. Annuaire de lcole Pratique des
Hautes tudes (V
e
section): 829.
______ . 1965. tudes de linguistique anatolienne. RHA 23 (76): 3354.
______ . 1966. Les noms des Hittites. Paris: Klincksieck.
______ . 1967. Les noms anatoliens du dieu et leurs derives. JCS 21: 17477.
______ . 1968. Textes de Ras Shamra en langue hittite. Pp. 76984 in Ugaritica V. Nouveaux
textes accadiens, hourrites et ugaritiques des Archives et Bibliotheques prives dUgarit,
commentaires des textes historiques, edited by J. Nougayrol, E. Laroche, C. Virolleaud,
and C. F. A. Schaeer. Paris: Imprimerie Nationale.
______ . 1969a. Fragments hittites du traite mitannien de Suppiluliuma I
er
. Pp. 36973 in Uga-
ritica VI, edited by J. Nougayrol, E. Laroche, C. Virolleaud, and C. F. A. Schaeer. Paris:
Imprimerie Nationale.
______ . 1969b. Vocatif et cas absolu en anatolien. Athenaeum NS 47: 17378.
______ . 1970. tudes de linguistique anatolienne, III. RHA 28: 2271.
______ . 1971. Catalogue des textes hittites. tudes et Commentaires 75. Paris: Klincksieck.
______ . 1973. Fleuve et ordalie en Asie Mineure hittite. Pp. 17990 in Festschrift Heinrich Ot-
ten. 27. Dez. 1973, edited by E. Neu and C. Rster. Wiesbaden: Harrassowitz.
______ . 1975. Noms daction en indo-europeen dAnatolie. Pp. 33946 in Mlanges linguis-
tiques oerts a Emile Benveniste. Paris: Societe de linguistique de Paris.
______ . 197879. Glossaire de la langue hourrite. 2 vols. Paris: Klincksieck.
______ . 1979. Anaphore et deixis en anatolien. Pp. 14752 in Hethitisch und indogermanisch.
Vergleichende Studien zur historischen Grammatik und zur Dialektgeographie, edited
by E. Neu and W. Meid. Innsbruck: Institut fr Sprachwissenschaft der Universitt Inns-
bruck.
______ . 1982. Epithetes e predicats en hittite. Pp. 13336 in Serta Indogermanica. Festschrift
fr Gnter Neumann zum 60. Geburtstag, edited by J. Tischler. Innsbruck: Institut fr
Sprachwissenschaft der Universitt Innsbruck.
Lebrun, Rene. 1985. Laphasie de Mursili II = CTH 486. Hethitica 6: 10337.
Lindeman, Fredrik Otto. 1965. Note phonologique sur hittite eku- boire. RHA 23 (76): 2932.
Luraghi, Silvia. 1987. Reconstructing Proto-Indo-European as an Ergative Language: A Test.
JIES 15: 35980.
______ . 1997a. Hittite. Languages of the WorldlMaterials 114. Munich: Lincom Europa.
______ . 1997b. I verbi ausiliari in ittita. Pp. 299320 in Ars linguistica. Studi oerti a Paolo
Ramat, edited by G. Gernini, P. Cuzzolin, and P. Molinelli. Rome: Bulzoni.
Macqueen, James G. 1986. The Hittites and Their Contemporaries in Asia Minor. 2nd ed. Lon-
don: Thames and Hudson.
00-HittiteGrammar.indb 456 6/5/08 4:21:16 PM
457 References
Marazzi, Massimiliano. 1996. Hethitische Lexikographie: CHD L-N, Facz. 4. Or NS 65: 157
65.
Martino, Stefano de. 1998. Le accuse di Murili II alla regina Tawananna secondo il testo KUB
XIV 4. Pp. 1948 in Studi e Testi I, edited by S. de Martino and F. Imparati. Florence:
LoGisma.
______ . 2003. Gli Ittiti. Le Bussole 96. Rome: Carocci.
Mascheroni, Lorenza M. 1980. Il modulo interrogativo in eteo I: note sintattiche. SMEA 22:
5362.
McCone, Kim R. 1979. The Diachronic Possibilities of the Indo-European Amplied Sentence:
A Case History from Anatolian. Pp. 46787 in Studies in Diachronic, Synchronic, and
Typological Linguistics. Festschrift for Oswald Szemernyi on the Occasion of his 65th
Birthday, edited by B. Brogyanyi. Amsterdam: Benjamins.
Meier-Brgger, Michael. 2000. Indogermanische Sprachwissenschaft. Berlin: de Gruyter.
______ . 2003. Indo-European Linguistics. Berlin: de Gruyter.
Melchert, H. Craig. 1977. Ablative and Instrumental in Hittite. Ph. D. Diss., Harvard University.
______ . 1983. A New PIE *men Sux. Die Sprache 29: 126.
______ . 1984a. Review of J. J. S. Weitenberg, Die hethitischen U-Stmme (Amsterdam, 1984).
Kratylos 29: 7982.
______ . 1984b. Studies in Hittite Historical Phonology. Ergnzungsheft zur Zeitschrift fr Ver-
gleichende Sprachforschung 32. Gttingen: Vandenhoeck & Ruprecht.
______ . 1985. Hittite imma and Latin immo. KZ 98: 184205.
______ . 1988a. Luvian Lexical Notes. HS 101: 21143.
______ . 1988b. Word-nal -r in Hittite. Pp. 21534 in A Linguistic Happening in Memory of
Ben Schwarz. Studies in Anatolian, Italic, and Other Indo-European Languages, edited by
Y. L. Arbeitman. Louvain-La-Neuve: Peeters.
______ . 1989. Review of D. Yoshida, Die Syntax des althethitischen substantivischen Genitivs
(Heidelberg, 1987). Kratylos 34: 18182.
______ . 1992. Hittite Vocalism. Pp. 18196 in Per una grammatica ittita. Towards a Hittite
Grammar, edited by O. Carruba. Pavia: Iuculano.
______ . 1993a. A New Anatolian Law of Finals. JAC 8: 10513.
______ . 1993b. Cuneiform Luvian Lexicon. Lexica Anatolica 2. Chapel Hill, NC: Published by
the Author.
______ . 1994. Anatolian Historical Phonology. Leiden Studies in Indo-European 3. Amsterdam:
Rodopi.
______ . 1995. Neo-Hittite Nominal Inection. Pp. 26974 in Atti del II Congresso Internazio-
nale di Hittitologia, edited by O. Carruba, M. Giorgieri, and C. Mora. Pavia: Iuculano.
______ . 1996. Review of H. G. Gterbock and T. van den Hout, The Hittite Instruction for the
Royal Bodyguard [AS 24] (Chicago, 1991). JNES 55: 13435.
______ . 1997a. Review of H. G. Gterbock and H. A. Honer, eds., The Hittite Dictionary of the
Oriental Instiute P, fasc. 2 (Chicago 1995). JAOS 117: 71314.
______ . 1997b. Syncope and Anaptyxis in Hittite. Pp. 17780 in Sound Law and Analogy. Pa-
pers in Honor of Robert S. P. Beekes on the Occasion of his 60th Birthday, edited by
A. Lubotsky. Amsterdam: Rodopi.
______ . 1997c. Traces of a PIE Aspectual Contrast in Anatolian? IncLing 20: 8392.
00-HittiteGrammar.indb 457 6/5/08 4:21:17 PM
458 References
______ . 1998. Aspects of Verbal Aspect in Hittite. Pp. 413418 in Uluslararas Hititoloji Kon-
gresi bildirileri. orum, 1622 Eyll, 1996. Acts of the IIIrd International Congress of
Hittitology, orum, September 1622, 1996, edited by S. Alp and A. Sel. Ankara: Uyum
Ajans.
______ . 2000. Tocharian Plurals in -nt- and Related Phenomena. TIES 9: 5375.
______ . 2001. Hittite Nominal Stems in -il. Pp. 26372 in Anatolisch und IndogermanischlAna-
tolico e Indoeuropeo. Akten des Kolloquiums der Indogermanischen Gesellschaft Pavia
22.25. September 1998, edited by O. Carruba and W. Meid. Innsbruck: Institut fr Sprach-
wissenschaft der Universitt Innsbruck.
______ . 2003a. Hittite antaka- Loins and an Overlooked Myth about Fire. Pp. 28188 in Hit-
tite Studies in Honor of Harry A. Honer Jr. on the Occasion of His 65th Birthday, edited
by G. M. Beckman, R. H. Beal and J. G. McMahon. Winona Lake, IN: Eisenbrauns.
______ . 2003b. Hittite Nominal Stems in -anzan-. Pp. 12939 in Indogermanisches Nomen.
Derivation, Flexion, und Ablaut, edited by E. Tichy, D. S. Wodtko, and B. Irslinger. Bre-
men: Hempen.
______ . 2003c. PIE thorn in Cuneiform Luvian? Pp. 14561 in Proceedings of the Fourteenth
Annual UCLA Indo-European Conference, edited by K. Jones-Bley. Washington: Institute
for the Study of Man.
______ . 2005a. Latin insolesco, Hittite ulle()- and PIE Statives in --. Pp. 9098 in Hd
mnas. Studies Presented to Professor Leonard G. Herzenberg on the occasion of his 70-
birthday, edited by N. N. Kazansky, E. R. Kryuchkova, A. S. Nikolaev and A. V. Shatskov.
Saint Petersburg: Nauka.
______ . 2005b. The Problem of Luvian Inuence on Hittite: When and How Much? Pp. 44560
in Sprachkontakt und Sprachwandel. Akten der XI. Fachtagung der Indogermanischen Ge-
sellschaft, 17.23. September 2000, Halle an der Saale, edited by G. Meiser and O. Hack-
stein. Wiesbaden: Reichert.
, ed. 2003d. The Luwians. Handbook of Oriental Studies, Section One: The Near and
Middle East 68. Leiden: Brill.
Meriggi, Piero. 1962. Hieroglyphisch-hethitisches Glossar. Wiesbaden: Harrassowitz.
______ . 19661975. Manuale di Eteo Geroglico. Rome: Ateneo.
______ . 1980. Schizzo grammaticale dellAnatolico. Accademia nazionale dei Lincei. Memorie.
Classe di scienze morali, storiche e lologiche, serie 8, 24l3. Rome: Accademia Nazionale
dei Lincei.
Miller, Jared L. 2004. Studies in the Origins, Development and Interpretation of the Kizzuwatna
Rituals. Studien zu den Boazky-Texten 46. Wiesbaden: Harrassowitz.
Mishali, Ayala. 2000. The bl dms role in the Neo-Assyrian legal process. Zeitschrift fr alt-
orientalische und biblische Rechtsgeschichte 6: 3553.
Monte, Giuseppe F. del. 1992. Die Orts- und Gewssernamen der hethitischen Texte, Supple-
ment. Repertoire Geographique des Textes Cuneiformes 6l2. Beihefte zum Tbinger Atlas
des Vorderen Orients, Reihe B 7l6. Wiesbaden: Reichert.
Monte, Giuseppe F. del, and Johann Tischler. 1978. Die Orts- und Gewssernamen der hethi-
tischen Texte. Repertoire Geographique des Textes Cuneiformes 6. Beihefte zum Tbinger
Atlas des Vorderen Orients, Reihe B, 7. Wiesbaden: Reichert.
Mora, Clelia. 1991. Sullorigine della scrittura geroglica anatolica. Kadmos 30: 128.
Mora, Clelia, and Mauro Giorgieri. 2004. Le lettere tra i re ittiti e i re assiri ritrovate a Hattua.
History of the Ancient Near East Monographs 7. Padua: Sargon.
00-HittiteGrammar.indb 458 6/5/08 4:21:17 PM
459 References
Morpurgo Davies, Anna. 1986. Forms of Writing in the Ancient Mediterranean World. Pp. 5177
in The Written Word. Literacy in Transition, edited by G. Baumann. Oxford.
Neu, Erich. 1968a. Das hethitische Mediopassiv und seine indogermanischen Grundlagen. Stu-
dien zu den Boazky-Texten 6. Wiesbaden: Harrassowitz.
______ . 1968b. Interpretation der hethitischen mediopassiven Verbalformen. Studien zu den
Boazky-Texten 5. Wiesbaden: Harrassowitz.
______ . 1969. Review of W. Kastner, Die griechischen Adjektive zweier Endungen auf -o (Hei-
delberg, 1967). IF 74: 23541.
______ . 1970. Ein althethitisches Gewitterritual. Studien zu den Boazky-Texten 12. Wies-
baden: Harrassowitz.
______ . 1974. Der Anitta-Text. Studien zu den Boazky-Texten 18. Wiesbaden: Harrassowitz.
______ . 1979a. Einige berlegungen zu den hethitischen Kasusendungen. Pp. 17796 in He thi-
tisch und indogermanisch. Vergleichende Studien zur historischen Grammatik und zur Dia-
lektgeographie, edited by E. Neu and W. Meid. Innsbruck: Institut fr Sprachwissenschaft
der Universitt Innsbruck.
______ . 1979b. Hethitisch kurur und takul in syntaktischer Sicht. Pp. 407428 in Studia Medi-
terranea Piero Meriggi dicata, edited by O. Carruba. Pavia: Aurora.
______ . 1979c. Zum sprachlichen Alter des Hukkana-Vertrages. KZ 93: 6484.
______ . 1980a. Studien zum endungslosen Lokativ des Hethitischen. Innsbrucker Beitrge zur
Sprachwissenschaft 23. Innsbruck: Institut fr Sprachwissenschaft der Universitt Inns-
bruck.
______ . 1980b. Althethitische Ritualtexte in Umschrift. Studien zu den Boazky-Texten 25.
Wiesbaden: Harrassowitz.
______ . 1981a. Noch einmal hethitisch heu- Regen. Pp. 20312 in Bono Homini Donum. Es-
says in Historical Linguistics in Memory of J. Alexander Kerns, edited by Y. L. Arbeitman
and A. R. Bomhard. Amsterdam: Benjamins.
______ . 1981b. Review of J. Tischler, Hethitisches Etymologisches Glossar, Lieferung 2 (Inns-
bruck, 1978). IF 86: 34751.
______ . 1982. Hethitisch lrl im Wortauslaut. Pp. 20525 in Serta Indogermanica. Festschrift fr
Gnter Neumann zum 60. Geburtstag, edited by J. Tischler. Innsbruck: Institut fr Sprach-
wissenschaft der Universitt Innsbruck.
______ . 1982l1983. Zum Genus hethitischer r-Stmme. Linguistica 28l29: 12530.
______ . 1983. Glossar zu den althethitischen Ritualtexten. Studien zu den Boazky-Texten 26.
Wiesbaden: Harrassowitz.
______ . 1985. Zur Stammabstufung bei i- und u-stmmigen Substantiven des Hethitischen.
Pp. 25964 in Sprachwissenschaftliche Forschungen. Festschrift fr Johann Knobloch
zum 65. Geburtstag am 5. Januar 1984 dargebracht von Freunden und Kollegen, edited by
H. M. lberg, G. Schmidt, and H. Bothien. Innsbruck: Institut fr Sprachwissenschaft der
Universitt Innsbruck.
______ . 1986. Zur unechten Nominalkomposition im Hethitischen. Pp. 10715 in o-o-pe-ro-si.
Festschrift fr Ernst Risch zum 75. Geburtstag, edited by A. Etter. Berlin: de Gruyter.
______ . 1988a. Das Hurritische. Eine altorientalische Sprache in neuem Licht. Akademie der
Wissenschaften und Literatur, Mainz. Abhandlungen der geistes- und sozialwissenschaft-
lichen Klasse, 1988l3. Stuttgart: Steiner.
______ . 1988b. Zur Grammatik des Hurritischen auf der Grundlage der hurritisch-hethitischen
Bilingue aus der Bogazky-Grabungskampagne 1983. Pp. 95116 in Hurriter und Hur-
00-HittiteGrammar.indb 459 6/5/08 4:21:18 PM
460 References
ritisch. Konstanzer Altorientalische Symposien, vol. 2, edited by V. Haas. Konstanz: Uni-
versittsverlag Konstanz.
______ . 1989a. Zu einer hethitischen Prteritalendung -ar. HS 102: 1620.
______ . 1989b. Zum Alter der personzierenden -ant-Bildung des Hethitischen: Ein Beitrag zur
Geschichte der indogermanischen Genuskategorie. HS 102: 115.
______ . 1990. Das hethitische Wort fr Frau. HS 103: 20817.
______ . 1992. Zum Kollektivum im Hethitischen. Pp. 197212 in Per una grammatica ittita.
Towards a Hittite Grammar, edited by O. Carruba. Pavia: Iuculano.
______ . 1993. Zu den hethitischen Ortspartikeln. Linguistica 33: 13752.
______ . 1994. Zur mittelhethitischen Sprachform der hurritisch-hethitischen Bilingue aus
attua. Pp. 189202 in Iranian and Indo-European Studies. Memorial Volume of Otakar
Klma, edited by P. Vavrouek. Prague: Enigma.
______ . 1995. Futur im Hethitischen? Pp. 195202 in Verba et Structurae. Festschrift fr Klaus
Strunk zum 65. Geburtstag, edited by H. Hettrich, W. Hock, P.-A. Mumm, and N. Oet-
tinger. Innsbruck: Institut fr Sprachwissenschaft der Universitt Innsbruck.
______ . 1996. Das hurritische Epos der Freilassung I. Untersuchungen zu einem hurritisch-
hethitischen Textensemble aus attua. Studien zu den Boazky-Texten 32. Wiesbaden:
Harrassowitz.
______ . 1997. Zu einigen Pronominalformen des Hethitischen. Pp. 139169 in Studies in Honor
of Jaan Puhvel, Part One. Ancient Languages and Philology, edited by D. Disterheft,
M. Huld, and J. A. C. Greppin. Washington, DC: Institute for the Study of Man.
______ . 1998. Hethitisch i-mu-u. HS 111: 5560.
Neumann, Gnter. 1964. Drei luwische Wrter. MSS 16: 4753.
Neve, Peter J. 1987. Die Ausgrabungen in Boazky-attua 1986. AA: 381412, 75455.
Nowicki, Helmuth. 2000. Zum Einleitungsparagraphen des Anitta-Textes (CTH 1,14). Pp. 347
55 in 125 Jahre Indogermanistik in Graz. Festband anlsslich des 125 jhrigen Bestehens
der Forschungsrichtung Indogermanistik an der Karl-Franzens-Universitt Graz, ed-
ited by C. Zinko and M. Otsch. Graz: Leykam.
Oettinger, Norbert. 1976a. Die Militrischen Eide der Hethiter. Studien zu den Boazky-Texten
22. Wiesbaden: Harrassowitz.
______ . 1976b. Indogermanisch `s(h
2
)neu_ln Sehne und `(s)men- gering sein im Hethi-
tischen. MSS 35: 93103.
______ . 1979. Die Stammbildung des hethitischen Verbums. Erlanger Beitrge zur Sprach- und
Kunstwissenschaft 64. Nrnberg: Carl.
______ . 1980. Die n-Stmme des Hethitischen und ihre indogermanischen Ausgangspunkte. HS
94: 4463.
______ . 1981. Hethitisch ganenant- gebeugt, drstig. MSS 40: 14353.
______ . 1982a. Die Dentalerweiterung von n-Stmmen und Heteroklitika im Griechischen, Ana-
tolischen und Altindischen. Pp. 23346 in Serta Indogermanica. Festschrift fr Gnter
Neumann zum 60. Geburtstag, edited by J. Tischler. Innsbruck: Institut fr Sprachwissen-
schaft der Universitt Innsbruck.
______ . 1982b. Reste von e-Hochstufe im Formans hethitischer n-Stmme einschliesslich des
umna-Suxes. Pp. 16277 in Investigationes philologicae et comparativae. Gedenk-
schrift fr Heinz Kronasser, edited by E. Neu. Wiesbaden: Harrassowitz.
00-HittiteGrammar.indb 460 6/5/08 4:21:18 PM
461 References
______ . 1982c. Review of E. Neu, Studien zum endungslosen Lokativ des hethitischen (Inns-
bruck, 1980). BiOr 39: 36468.
______ . 1986a. Anatolische Kurzgeschichten. HS 99: 4353.
______ . 1986b. Indo-Hittite-Hypothese und Wortbildung. Innsbrucker Beitrge zur Sprach-
wissenschaft 37. Innsbruck: Institut fr Sprachwissenschaft der Universitt Innsbruck.
______ . 1987. Bemerkungen zur anatolischen i-Motion und Genusfrage. HS 100: 3543.
______ . 1988. Hethitisch -want-. Pp. 27386 in Documentum Asiae Minoris antiquae. Fest-
schrift fr Heinrich Otten zum 75. Geburtstag, edited by E. Neu and C. Rster. Wiesbaden:
Harrassowitz.
______ . 1993. Mittelhethitisches hatte sie schlachtete. Linguistica 33: 15355.
______ . 1994. Etymologisch unerwarteter Nasal im Hethitischen. Pp. 30730 in In honorem
Holger Pedersen. Kolloquium der Indogermanischen Gesellschaft vom 26. bis 28. Mrz
1993 in Kopenhagen, edited by J. E. Rasmussen and B. Nielsen. Wiesbaden: Reichert.
______ . 1995. Griech. , heth. kuli und ein neues Kollektivsux. Pp. 21128 in Verba
et structurae. Festschrift fr Klaus Strunk zum 65. Geburtstag, edited by H. Hettrich,
W. Hock, P.-A. Mumm, and N. Oettinger. Innsbruck: Institut fr Sprachwissenschaft der
Universitt Innsbruck.
______ . 1998. Skizze zur Funktion der Reduplikation im Hethitischen. Pp. 45156 in Uluslararas
Hititoloji Kongresi bildirileri. orum, 1622 Eyll, 1996. Acts of the IIIrd International
Congress of Hittitology, orum, September 1622, 1996, edited by S. Alp and A. Sel.
Ankara: Uyum Ajans.
______ . 2000a. Der hethitische Lautwandel kr > hr. Pp. 35762 in 125 Jahre Indogermanistik
in Graz. Festband anlsslich des 125jhrigen Bestehens der Forschungsrichtung Indo-
germanistik an der Karl-Franzens-Universitt Graz, edited by M. Otsch and C. Zinko.
Graz: Leykam.
______ . 2000b. Heth. udn, arm. getin Land und lyk. wedreli. Pp. 18187 in Anusantatyai.
Festschrift fr Johanna Narten zum 70. Geburtstag, edited by A. Hintze and E. Tichy.
Dettelbach: Rll.
______ . 2001a. Hethitisch -ima- oder: Wie ein Sux aektiv werden kann. Pp. 45677 in Akten
des IV. Internationalen Kongresses fr Hethitiologie, Wrzburg, 4.8. Oktober, 1999, ed-
ited by G. Wilhelm. Wiesbaden: Harrassowitz.
______ . 2001b. Neue Gedanken ber das nt-Sux. Pp. 30116 in Anatolisch und Indoger-
manischlAnatolico e Indoeuropeo. Akten des Kolloquiums der Indogermanischen Gesell-
schaft Pavia 22.25. September 1998, edited by O. Carruba and W. Meid. Innsbruck: Insti-
tut fr Sprachwissenschaft der Universitt Innsbruck.
______ . 2002. Die Stammbildung des hethitischen Verbums. Nachdruck mit einer kurzen Revi-
sion der hethitischen Verbalklassen. Dresdner Beitrge zur Hethitologie 7. Dresden: Tech-
nische Universitt Dresden.
______ . 2003. Zum Ablaut von n-Stmmen im Anatolischen und der Brechung > ya. Pp. 141
52 in Indogermanisches Nomen. Derivation, Flexion, und Ablaut, edited by E. Tichy, D. S.
Wodtko, and B. Irslinger. Bremen: Hempen.
Ose, Fritz. 1944. Supinum und Innitiv im Hethitischen. Mitteilungen der Vorderasiatisch-
Aegyptischen Gesellschaft 47l1. Leipzig: Hinrichs.
Oshiro, Terumasa. 1990 [92]. The Function of the Sentence Particle -an in Old Hittite. Pp. 3
10 in Folia Linguistica Historica. Acta Societatis Linguisticae Europaeae, XIl12. The
Hague: Mouton.
00-HittiteGrammar.indb 461 6/5/08 4:21:19 PM
462 References
______ . 1995. The Verbal Reduplication in Hieroglyphic Luwian. Orient 3031: 294301.
Otten, Heinrich. 1950. Mythen vom Gotte Kumarbi. Neue Fragmente. Verentlichungen des In-
stituts fr Orientforschung der Deutschen Akademie der Wissenschaften 3. Berlin: Akad-
emie Verlag.
______ . 1951. Zu den Anfngen der hethitischen Geschichte. MDOG 83: 3345.
______ . 1953a. Ein kanaanischer Mythus aus Bogazky. MIO 1: 12550.
______ . 1953b. Kanaanische Mythen aus Hattusa-Bogazky. MDOG 85: 2738.
______ . 1957. Zustzliche Lesungen zum Alakandu-Vertrag. MIO 5: 2630.
______ . 1958. Hethitische Totenrituale. Verentlichungen des Instituts fr Orientforschung der
Deutschen Akademie der Wissenschaften 37. Berlin: Akademie Verlag.
______ . 1959. Zur Kontinuitt eines altanatolischen Kultes. ZA 53: 17484.
______ . 1961. Eine Beschwrung der Unterirdischen aus Boazky. ZA 54: 11457.
______ . 19681969. Die Tontafelfunde von 1967. AfO 22: 11113.
______ . 1969. Sprachliche Stellung und Datierung des Madduwatta-Textes. Studien zu den
Boazky-Texten 11. Wiesbaden: Harrassowitz.
______ . 1973. Eine althethitische Erzhlung um die Stadt Zalpa. Studien zu den Boazky-
Texten 17. Wiesbaden: Harrassowitz.
______ . 1981. Die Apologie Hattusilis III.. Das Bild der berlieferung. Studien zu den
Boazky-Texten 24. Wiesbaden: Harrassowitz.
______ . 1983. Der Anfang der HAZANNU-Instruktion. Or NS 52: 13342.
______ . 1988. Die Bronzetafel aus Boazky. Ein Staatsvertrag Tuthalijas IV. Studien zu den
Boazky-Texten Beiheft 1. Wiesbaden: Harrassowitz.
______ . 1990. Bemerkungen zur berlieferung einiger hethitischer Texte. ZA 80: 22327.
______ . 1992. Erwgungen zur Kontinuitt altanatolischer Kulte. Pp. 3442 in Uluslararas 1.
Hititoloji Kongresi Bilderileri (1921 Temmuz 1990). orum, Turkey: Uluslararas orum
Hitit Festivali Komitesi Bakanl.
Otten, Heinrich, and Wolfram von Soden. 1968. Das akkadisch-hethitische Vokabular KBo I 44
+ KBo XIII 1. Studien zu den Boazky-Texten 7. Wiesbaden: Harrassowitz.
Otten, Heinrich, and Vladimir Souek. 1969. Ein althethitisches Ritual fr das Knigspaar. Stu-
dien zu den Boazky-Texten 8. Wiesbaden: Harrassowitz.
Payne, Annick. 2004. Hieroglyphic Luwian. Elementa Linguarum Orientis 3. Wiesbaden: Harras-
sowitz.
Pecchioli Daddi, Franca. 2003. Il vincolo per i governatori de provincia. Studia Mediterranea 14.
Pavia: Italian University Press.
Pecchioli Daddi, Franca, and Anna Maria Polvani. 1990. La mitologia ittita. Testi del Vicino
Oriente antico 4.1. Brescia: Paideia.
Pecora, Laura. 1984. La particella -wa(r)- e il discorso diretto in antico-eteo. IF 89: 10424.
Pedersen, Holger. 1935. Zum Lautwert des Zeichens - im Hettitischen. ArOr 7: 8088.
______ . 1938. Hittitisch und die anderen indoeuropischen Sprachen. Det Danske Videnskab-
ernes Selskab. Hist.-lol. Meddelelser 25l2. Copenhagen.
Plchl, Reinhold. 2003. Einfhrung ins Hieroglyphen-Luwische. Dresdner Beitrge zur Hethi-
tologie 8. Dresden: Technische Universitt.
Polvani, Anna Maria. 1988. Appunti per una storia della musica cultuale ittita: lo strumento
huhupal. Hethitica 9: 17179.
00-HittiteGrammar.indb 462 6/5/08 4:21:19 PM
463 References
Popko, Maciej. 1995. Religions of Asia Minor. Translated by I. Zych. Warsaw: Academic Pub-
lications Dialog.
Prins, Anna Adriana Maria. 1997. Hittite Neuter SingularNeuter Plural. Some Evidence for a
Connection. Leiden: Research School CNWS.
Pritchard, James B. 1969. Ancient Near Eastern Texts Relating to the Old Testament. 3rd ed.,
with suppl. Princeton: Princeton University Press.
Probert, Philomen. 2006. Clause Boundaries in Old Hittite Relative Sentences. TPS 104: 1783.
Puhvel, Jaan. 1984. Hittite Etymological Dictionary. Berlin: Mouton.
Raggi Braglia, Maria S. 1989. IE *per- volare nelle lingue anatoliche. OA 28: 20111.
Rasmussen, Jens E. 2001. From the Realm of Anatolian Verbal Stem Formation: Problems of
Reduplication. Pp. 35568 in Anatolisch und IndogermanischlAnatolico e Indoeuropeo.
Akten des Kolloquiums der Indogermanischen Gesellschaft Pavia 22.25. September 1998,
edited by O. Carruba and W. Meid. Innsbruck: Institut fr Sprachwissenschaft der Univer-
sitt Innsbruck.
Reichert, Pierre. 1963. Glossaire inverse de la langue hittite. RHA 21 (73): 59145.
Rieken, Elisabeth. 1994. Der Wechsel -a-l-i- in der Stammbildung des hethitischen Nomens. HS
107: 4253.
______ . 1996. Beitrge zur anatolischen Sprachgeschichte. AoF 23: 28997.
______ . 1999a. Untersuchungen zur nominalen Stammbildung des Hethitischen. Studien zu den
Boazky-Texten 44. Wiesbaden: Harrassowitz.
______ . 1999b. Zur Verwendung der Konjunktion ta in den hethitischen Texten. MSS 59: 63
88.
______ . 2000a. Die Partikeln -a, -i _a, -ma im Althethitischen und das Akkadogramm U. Pp. 411
19 in 125 Jahre Indogermanistik in Graz. Festband anlsslich des 125 jhrigen Be stehens
der Forschungsrichtung Indogermanistik an der Karl-Franzens-Universitt Graz, ed-
ited by C. Zinko and M. Otsch. Graz: Leykam.
______ . 2000b. Hethitisch umuma-. HS 113: 17175.
______ . 2001. Einige Beobachtungen zum Wechsel u-l(u)u_a- in den hethitischen Texten.
Pp. 36980 in Anatolisch und IndogermanischlAnatolico e Indoeuropeo. Akten des Kol-
loquiums der Indogermanischen Gesellschaft Pavia 22.25. September 1998, edited by
O. Carruba and W. Meid. Innsbruck: Institut fr Sprachwissenschaft der Universitt Inns-
bruck.
______ . 2002. Ein Lautgesetz und der Obliquusstamm des urindg. Personalpronomens. Pp. 407
16 in Novalis Indogermanica. Festschrift fr Gnter Neumann zum 80. Geburtstag, edited
by M. Fritz and S. Zeilfelder. Graz: Leykam.
______ . 2004a. Merkwrdige Kasusformen im Hethitischen. Pp. 53343 in arnikzel. Hethi-
tologische Studien zum Gedenken an Emil Orgetorix Forrer, edited by D. Groddek and
S. Rle. Dresden: Technische Universitt Dresden.
______ . 2004b. Reste von e-Hochstufe im Formans hethitischer n-Stmme? Pp. 28494 in Indo-
European Word Formation. Proceedings of the Conference held at the University of Co-
penhagen October 20th22nd 2000, edited by J. Clackson and B. A. Olsen. Copenhagen:
Museum Tusculanum.
______ . 2005a. Hethitisch. Pp. 80127 in Sprachen des Alten Orients, edited by M. P. Streck.
Darmstadt: Wissenschaftliche Buchgesellschaft.
00-HittiteGrammar.indb 463 6/5/08 4:21:20 PM
464 References
______ . 2005b. Zur Wiedergabe von hethitisch lol. Pp. 53749 in Sprachkontakt und Sprach-
wandel. Akten der XI. Fachtagung der Indogermanischen Gesellschaft, 17.23. September
2000, Halle an der Saale, edited by G. Meiser and O. Hackstein. Wiesbaden: Reichert.
______ . forthcoming. Die periphrastischen Konstruktionen mit pai- gehen und uwa- kommen
im Hethitischen. In la-a-ma-an-te-et na-ak-ki-i. Studia anatolica in memoriam Erich Neu
scripta, edited by C. Rster, J. Catsanicos, and R. Lebrun. Louvain-la-Neuve: Peeters.
Riemschneider, Kaspar K. 1958. Die hethtitischen Landschenkungsurkunden. MIO 6:32181.
______ . 1970. Babylonische Geburtsomina in hethitischer bersetzung. Studien zu den Boazky-
Texten 9. Wiesbaden: Harrassowitz.
Rosenkranz, Bernhard. 1952. Beitrge zur Erforschung des Luvischen. Wiesbaden: Harrasso-
witz.
______ . 1978. Vergleichende Untersuchungen der altanatolischen Sprachen. Trends in Linguis-
tics, State-of-the-Art Reports 8. The Hague: Mouton de Gruyter.
Rle, Silvester. 1998. Heth. karza Gert des Webers und kazzarnul eine bestimmte Tuchart.
MSS 58: 11128.
Rost, Liane. 1956. Die ausserhalb von Boazky gefundenen hethitischen Briefe. MIO 4: 328
50.
Rubio, Gonzalo. 2006. Writing in Another Tongue: Alloglottography in the Ancient near East.
Pp. 3164 in Margins of Writing, Origins of Cultures, edited by S. L. Sanders. Chicago:
The Oriental Institute of the University of Chicago.
______ . 2007. Sumerian Morphology. Pp. 132779 in Morphologies of Asia and Africa, edited
by A. S. Kaye. Winona Lake, IN: Eisenbrauns.
Rumsey, A. 1987. Was Proto-Indo-European an Ergative Language? JIES 15: 1938.
Rster, Christel. 1988. Materialien zu einer Fehlertypologie der hethitischen Texte. Pp. 295306
in Documentum Asiae Minoris Antiquae. Festschrift fr Heinrich Otten zum 75. Geburts-
tag, edited by E. Neu and C. Rster. Wiesbaden: Harrassowitz.
Rster, Christel, and Erich Neu. 1989. Hethitisches Zeichenlexikon. Inventar und Interpretation
der Keilschriftzeichen aus den Boazky-Texten. Studien zu den Boazky-Texten, Beiheft
2. Wiesbaden: Harrassowitz.
Salisbury, Donna. 1999. anda and andan in Neo-Hittite. JCS 51: 112.
Schaefer, Christiane. 1994. Das Intensivum im Vedischen. Historische Sprachforschung Ergn-
zungsheft 37. Gttingen: Vandenhoeck & Ru precht.
Schindler, Jochem. 1967. Das idg. Wort fr Erde und die dentalen Spiranten. Die Sprache 13:
191205.
Schuler, Einar von. 1957. Hethitische Dienstanweisungen fr hhere Hof- und Staatsbeamte.
AfO Beiheft 10. Graz: Weidner.
Sideltsev, Andrej. 2002. A New Morphophonological Rule for Middle Hittite? Pp. 2180 in
Memoriae A. A. Korolv dicata, edited by A. S. Kassian and A. V. Sideltsev. Moscow:
Languages of Slavonic Culture.
Siegelova, Jana. 1971. Appu-Mrchen und Hedammu-Mythus. Studien zu den Boazky-Texten
14. Wiesbaden: Harrassowitz.
______ . 1986. Hethitische Verwaltungspraxis im Lichte der Wirtschafts- und Inventardoku-
mente. Prague: Narodn Muzeum v Praze.
Sihler, Andrew L. 1995. New Comparative Grammar of Greek and Latin. New York: Oxford
University Press.
00-HittiteGrammar.indb 464 6/5/08 4:21:21 PM
465 References
Singer, Itamar. 1975. Hittite ilammar and Hieroglyphic Luwian ilana. ZA 65: 69103.
______ . 1996. Muwatallis Prayer to the Assembly of Gods through the Storm-God of Lightning
(CTH 381). Atlanta: Scholars Press.
______ . 1999. A New Hittite Letter from Emar. Pp. 6572 in Landscapes, Territories, Fron-
tiers and Horizons in the Ancient Near East. Papers presented to the XLIV Rencontre As-
syriologique Internationale, Venezia, 711 July 1997, volume 2: Geography and Cultural
Landscapes, edited by L. Milano, S. de Martino, M. F. Fales and G. B. Lanfranchi. Padova:
Sargon.
______ . 2002a. A Hattian whistle? NABU 2002 No. 3: 6667 #67.
______ . 2002b. Hittite Prayers. Writings from the Ancient World 11. Atlanta: Society of Biblical
Literature.
Soden, Wolfram von. 1952. Grundriss der akkadischen Grammatik. Analecta Orientalia 33.
Rome: Ponticium Institutum Biblicum.
______ . 196585. Akkadisches Handwrterbuch. Wiesbaden: Harrassowitz.
______ . 1995. Grundriss der akkadischen Grammatik. 3rd revised ed., edited by Werner R.
Mayer. Analecta Orientalia 33. Rome: Ponticium Institutum Biblicum.
Soden, Wolfram von, and Wolfgang Rllig. 1991. Das Akkadische Syllabar. 4th ed. Analecta
Orientalia 42. Rome: Ponticio Istituto Biblico.
Sommer, Ferdinand. 1920. Hethitisches I. Boghazki-Studien 4. Leipzig: Hinrichs.
______ . 1921. Hethitisch aruna- und die Partikel -p. OLZ 24: 197201.
______ . 1922. Hethitisches II. Boghazki-Studien 7. Leipzig: Hinrichs.
______ . 1932. Die Aijava-Urkunden. Abhandlungen der Bayerischen Akademie der Wissen-
schaften, philosophisch-historische Abteilung NF 6. Munich: Bayerische Akademie der
Wissenschaften.
______ . 1947. Hethiter und Hethitisch. Stuttgart: Kohlhammer.
Sommer, Ferdinand, and Hans Ehelolf. 1924. Das hethitische Ritual des Papanikri von Komana.
Boghazki-Studien 10. Leipzig: Hinrichs.
Sommer, Ferdinand, and A. Falkenstein. 1938. Die hethitisch-akkadische Bilingue des Hattusili
I. (Labarna II.). Abhandlungen der Bayerischen Akademie der Wissenschaften, philoso-
phisch-historische Abteilung NF 16. Munich: Bayerische Akademie der Wissenschaften.
Souek, Vladimir. 1970. Zur Sprache der hethitischen Gesetze. ArOr 38: 26976.
Soysal, Ouz. 2004. Hattischer Wortschatz in hethitischer Textberlieferung. Handbook of Ori-
ental Studies, Section One: The Near and Middle East 74. Leiden: Brill.
Speiser, Ephraim Avigdor. 1941. Introduction to Hurrian. Annual of the American Schools of
Oriental Research 20. New Haven: American Schools of Oriental Research.
Starke, Frank. 1977. Die Funktionen der dimensionalen Kasus und Adverbien im Althethitischen.
Studien zu den Boazky-Texten 23. Wiesbaden: Harrassowitz.
______ . 1979. Zu den hethitischen und luwischen Verbalabstrakta auf -a-. KZ 93: 24761.
______ . 1980. Das luwische Wort fr Frau. HS 94: 7486.
______ . 1985. Die keilschrift-luwischen Texte in Umschrift. Studien zu den Boazky-Texten
30. Wiesbaden: Harrassowitz.
______ . 1990. Untersuchung zur Stammbildung des keilschrift-luwischen Nomens. Studien zu
den Boazky-Texten 31. Wiesbaden: Harrassowitz.
______ . 1993. Zur Herkunft von akkad. talurgumannu(m) Dolmetscher. WO 24: 2038.
00-HittiteGrammar.indb 465 6/5/08 4:21:21 PM
466 References
______ . 1997. Troia im Kontext des historisch-politischen und sprachlichen Umfeldes Klein-
asiens im 2. Jahrtausend. Studia Troica 7: 44687.
Stefanini, Ruggero. 1964. Una lettera della regina Puduhepa al re di Alasija (KUB XXI 38).
AIPHOS 29: 369.
______ . 1974. Ancora sul vocativo ittita. AGI 59: 3742.
______ . 1988. Alcuni problemi ittiti, lessicali e sintattici. Pp. 25156 in Studi di Storia e di Filo-
logia Anatolica dedicati a Giovanni Pugliese Carratelli, edited by F. Imparati. Florence:
Librarie Italiane Estere.
Steiner, Gerd. 1971. Gott. D. Nach hethitischen Texten. Pp. 54775 in Reallexikon der Assyriolo-
gie, vol. 3, edited by E. Ebeling, B. Meissner, E. Weidner, W. von Soden, and D. O. Edzard.
Berlin: de Gruyter.
Sturtevant, Edgar H. 1933. A Comparative Grammar of the Hittite Language. New Haven: Yale
University Press.
Sturtevant, Edgar H., and George Bechtel. 1935. A Hittite Chrestomathy. Philadelphia: Linguis-
tic Society of America.
Sturtevant, Edgar H., and E. Adelaide Hahn. 1951. A Comparative Grammar of the Hittite Lan-
guage. 2nd ed. New Haven: Yale University Press.
Sel, Aygl. 1992. Ortaky: Eine hethitische Stadt mit hethitischen und hurritischen Ton tafel-
entdeckungen. Pp. 487492 in Hittite and Other Anatolian and Near Eastern Studies in
Honour of Sedat Alp, edited by H. Otten, E. Akurgal, H. Ertem and A. Sel. Ankara: Trk
Tarih Kurumu.
______ . 2002. Ortaky-apinuwa. Pp. 157166 in Recent Developments in Hittite Archaeology
and History, edited by K. A. Yener and H. A. Honer, Jr. Winona Lake, IN: Eisenbrauns.
Srenhagen, Dietrich. 1981. Zwei Gebete Hattusilis und der Puduhepa: Textliche und literatur-
historische Untersuchungen. AoF 8: 84168.
Szabo, Gabriella. 1971. Ein hethitisches Entshnungsritual fr das Knigspaar Tutaliya und
Nikalmati. Texte der Hethiter 1. Heidelberg: Winter.
Szemerenyi, Oswald J. L. 1996. Introduction to Indo-European Linguistics. Oxford: Clarendon.
Taracha, Piotr. 1990. More about the Hittite taknaz da- Rituals. Hethitica 10: 171184.
______ . 2001. Hethitisch

kippa- und das Sumerogramm (.)GI.PAD mesopotamischer Texte.

AoF 28: 132146.
Tenner, E. 1926. Ein hethitischer Annalentext des Knigs Murili II (KBo V 8 = Bo 2022). Text,
bersetzungsversuch, sprachliche Anmerkungen. Pp. 83106 in Jubilums-Festschrift der
II. Stdtischen Realschule zu Leipzig. Leipzig: Zweite Stdtische Realschule.
Thomsen, Marie-Louise. 1984. The Sumerian Language. An Introduction to Its History and
Grammatical Structure. Mesopotamia: Copenhagen Studies in Assyriology 10. Copen-
hagen: Akademisk Forlag.
Tischler, Johann. 2001. Hethitisches Handwrterbuch. Innsbrucker Beitrge zur Sprachwissen-
schaft 102. Innsbruck.
Tjerkstra, Franoise A. 1999. Principles of the Relation between Local Adverb, Verb and Sen-
tence Particle in Hittite. Cuneiform Monographs 15. Groningen: Styx.
Tropper, Josef. 2000. Ugaritische Grammatik. Alter Orient und Altes Testament 273. Mnster:
Ugarit-Verlag.
nal, Ahmet. 1977. Naturkatastrophen in Anatolien im 2. Jahrtausende v. Chr. Belleten 41: 447
72.
00-HittiteGrammar.indb 466 6/5/08 4:21:22 PM
467 References
______ . 1996. Studies in Ancient Anatolian Magical Practices. The Magic Ritual of antitau
from the City of urma against Troublesome Years. Ankara: Trk Tarih Kurumu.
Ungnad, Arthur. 1920. Das Personalpronomen der dritten Person in der hethitischen Satzverbin-
dung. ZDMG 74: 417422.
______ . 1992. Akkadian Grammar. Translated by H. A. Honer Jr., 5th edition. Society of Bibli-
cal Literature Resources for Biblical Study 30. Atlanta: Scholars Press.
Villar, F. 1988. On the 2nd Person Singular Pronoun in Hittite and in Indo-European. JIES 16:
18.
Villar, Francisco. 1983. Ergatividad, acusatividad y gnero en la familia lingistica indo-
europea. Salamanca: Universidad de Salamanca.
Wagner, Heinrich. 1985. Das Hethitische vom Standpunkte der typologischen Sprachgeographie.
Testi linguistici 7. Pisa: Giardini.
Watkins, Calvert. 1970. On the family of arce, and Hittite ark-. HSCP 74: 6774.
______ . 1973. Hittite and Indo-European Studies: The Denominative Statives in --. TPS 104:
5193.
______ . 1974. god. Pp. 10110 in Antiquitates Indogermanicae. Studien zur indogermanischen
Altertumskunde und zur Sprach- und Kulturgeschichte der indogermanischen Vlker, ed-
ited by M. Mayrhofer. Innsbruck: Institut fr Sprachwissenschaft der Universitt Inns-
bruck.
______ . 1975. Die Vertretung der Laryngale in gewissen morphologischen Kategorien in den in-
dogermanischen Sprachen Anatoliens. Pp. 35878 in Flexion und Wortbildung. Akten der
V. Fachtagung der indogermanischen Gesellschaft, Regensburg, 9.14. Sept. 1973, edited
by H. Rix. Wiesbaden: Reichert.
______ . 1982. Notes on the Plural Formations of the Hittite Neuters. Pp. 25262 in Investigatio-
nes philologicae et comparativae. Gedenkschrift fr Heinz Kronasser, edited by E. Neu.
Wiesbaden: Harrassowitz.
______ . 1995. How to Kill a Dragon. Aspects of Indo-European Poetics. Oxford: Oxford Uni-
versity Press.
______ . 1998. Proto-Indo-European: Comparison and Reconstruction. Pp. 2573 in The Indo-
European Languages, edited by A. G. Ramat and P. Ramat. London: Routledge.
______ . 2000. A Distant Anatolian Echo in Pindar: The Origin of the Aegis Again. HSCP 100:
114.
Weber, Otto, and Erich Ebeling. 1915. Die El-Amarna Tafeln. Anmerkungen und Register.
Leipzig: Hinrichs.
Wegner, Ilse. 2000. Einfhrung in die hurritische Sprache. Wiesbaden: Harrassowitz.
Weidner, Ernst. 1923. Politische Dokumente aus Kleinasien. Die Staatsvertrge in akkadischer
Sprache aus dem Archiv von Boghazki. Boghazki-Studien 8. Leipzig: Hinrichs.
Weitenberg, Joseph Johannes Sicco. 1979. Einige Bemerkungen zu den hethitischen Diphthong-
Stammen. Pp. 289303 in Hethitisch und indogermanisch. Vergleichende Studien zur his-
torischen Grammatik und zur Dialektgeographie, edited by E. Neu and W. Meid. Inns-
bruck: Institut fr Sprachwissenschaft der Universitt Innsbruck.
______ . 1984. Die hethitischen U-Stmme. Amsterdam: Rodopi.
______ . 1992. The Use of Asyndesis and Particles in Old Hittite Simple Sentences. Pp. 30553
in Per una grammatica ittita. Towards a Hittite Grammar, edited by O. Carruba. Pavia:
Iuculano.
00-HittiteGrammar.indb 467 6/5/08 4:21:22 PM
468 References
______ . 1995. Sigmatization and Thematization in Hittite. Pp. 33344 in Studio historiae ar-
dens. Ancient Near Eastern Studies Presented to Philo H. J. Houwink ten Cate on the Oc-
casion of His 65th Birthday, edited by T. P. J. van den Hout and J. de Roos. Istanbul: Institut
historique et archeologique neerlandais de Stamboul.
Werner, Rudolf, and Barbara Lscher. 1991. Kleine Einfhrung ins Hieroglyphen-Luwische. Or-
bis biblicus et orientalis 106. Freiburg, Schweiz: Universittsverlag.
Wilhelm, Gernot. 1992a. Hurritische Lexikographie und Grammatik: Die hurritisch-hethitische
Bilingue aus Boazky. Or NS 61: 12241.
______ . 1992b. Notizen zum hurritischen Wrterbuch. SMEA 29: 24548.
______ . 1992c. Zum viersprachigen Vokabular aus Ugarit. SMEA 29: 24954.
______ . 1995. Die Tontafelfunde der 2. Ausgrabungskampagne 1994 in Kuakl. MDOG 127:
3742.
______ . 1998. Zwei mittelhethitische Briefe aus dem Gebude C in Kuakl. MDOG 130: 175
87.
______ . 2002a. Die Keilschriftfunde der Kampagne 2001 in Kuakl. MDOG 134: 34251.
______ . 2002b. Die Sprachen des Hethiterreiches. Pp. 4649 in Die Hethiter und ihr Reich.
Das Volk der 1000 Gtter, edited by Kunst- und Ausstellungshalle der Bundesrepublik
Deutschland. Stuttgart: Theiss.
, ed. 1997. Keilschrifttexte aus Gebude A. KuaklSarissa Il1. RahdenlWestf.: Lei-
dorf.
Yakubovich, Ilya. 2006. Freestanding Genitive and Hypostasis in Hittite. JNES 65: 3949.
Yoshida, Daisuke. 1987. Die Syntax des althethitischen substantivischen Genitivs. Texte der
Hethiter 13. Heidelberg: Winter.
Yoshida, Kazuhiko. 1987. The Present Mediopassive Endings -tati and -watati in Hittite. Die
Sprache 33: 2933.
______ . 1990. The Hittite mediopassive endings in -ri. Untersuchungen zur indogermanischen
Sprach- und Kulturwissenschaft 5. Berlin.
______ . 1991. Reconstruction of Anatolian Verbal Endings: The Third Person Plural Preterites.
JIES 19: 35774.
______ . 1998. Hittite Verbs in -Vzi. Pp. 60514 in Uluslararas Hititoloji Kongresi bildirileri.
orum, 1622 Eyll, 1996. Acts of the IIIrd International Congress of Hittitology, orum,
September 1622, 1996, edited by S. Alp and A. Sel. Ankara: Uyum Ajans.
______ . 2000. The Original Ablaut of Hittite iu_att-. MSS 60: 17584.
______ . 2002. Observations on Some Cuneiform Spellings: Epithetic or Graphic? Pp. 16576 in
Acts of the Thirteenth Annual UCLA Indo-European Conference, Los Angeles, November
910, 2001, edited by K. Jones-Bley, M. E. Huld, A. Della Volpe, and M. R. Dexter. Wash-
ington, DC: Institute for the Study of Man.
Zeilfelder, Susanne. 2001. Archaismus und Ausgliederung. Studien zur sprachlichen Stellung
des Hethitischen. Heidelberg: Winter.
Zuntz, Leonie. 1936. Die hethitischen Ortsadverbien ara, par, piran als selbstndige Adver-
bien und in ihrer Verbindung mit Nomina und Verba. Speyer: Pilger.
00-HittiteGrammar.indb 468 6/5/08 4:21:23 PM