Early  Journal  Content  on  JSTOR,  Free  to  Anyone  in  the  World  
This  article  is  one  of  nearly  500,000  scholarly  works  digitized  and  made  freely  available  to  everyone  in  
the  world  by  JSTOR.    

Known  as  the  Early  Journal  Content,  this  set  of  works  include  research  articles,  news,  letters,  and  other  
writings  published  in  more  than  200  of  the  oldest  leading  academic  journals.  The  works  date  from  the  
mid-­‐seventeenth  to  the  early  twentieth  centuries.    

 We  encourage  people  to  read  and  share  the  Early  Journal  Content  openly  and  to  tell  others  that  this  
resource  exists.    People  may  post  this  content  online  or  redistribute  in  any  way  for  non-­‐commercial  
purposes.  

Read  more  about  Early  Journal  Content  at  http://about.jstor.org/participate-­‐jstor/individuals/early-­‐

journal-­‐content.    

JSTOR  is  a  digital  library  of  academic  journals,  books,  and  primary  source  objects.  JSTOR  helps  people  
discover,  use,  and  build  upon  a  wide  range  of  content  through  a  powerful  research  and  teaching  
platform,  and  preserves  this  content  for  future  generations.  JSTOR  is  part  of  ITHAKA,  a  not-­‐for-­‐profit  
organization  that  also  includes  Ithaka  S+R  and  Portico.  For  more  information  about  JSTOR,  please  
contact  [email protected].  
JThe Voc0abulary of Sumerian.-By J. DYNELEY PRINCE,
Professor in Columbia University, New York City.

? I. THE study of the Sumerian vocabulary falls logically

into three divisions. These are 1) the origin of the signs, 2)
the etymology of the phonetic values, and 3) the elucidation
of the many and varied primitive sign-meanings.
Professor Friedrich Delitzsch in his epoch-making work on
the origin of the most ancient Babylonian system of writing'
has paved the way for our thorough understanding of the prin-
ciples which were followed by the fathers of the ancient pre-
cuneiform lineal style of inscription. Previous to Delitzsch's
masterly work in this field, really no one had correctly under-
stood the facts regarding the beginnings of the cuneiform writ-
ing. These are so overwhelmingly in favor of the linguistic
character of Sumerian that they have caused Delitzsch himself
to abandon the Halevyan " ideophonic" cryptographic hypo-
thesis,2 of which the distinguished German Assyriologist had
formerly been an ardent adherent. Delitzsch's work, however,
has thus far included only the study of the signs. He has
made no attempt as yet to combine the examination of the signs
with the investigation of the phonetic values, which it is neces-
sary to do in order to arrive at the true nature of the Sumerian
idiom as it has been handed down to us. The thorough exam-
ination of every sign is imperative, if we are ever to penetrate
the mysteries of the highly difficult problem, but it is equally
imperative that every phonetic value and word-combination be
also studied, both in connection with tbe equivalent signs and
with other allied phonetic values.
? II. A most important point should now be noted. The
etymological study of the Sumerian word-list is attended by

' Die Entstehung des altesten Schriftsystemrs oder der Ursprung der
Keilschriftzeichen, Fried. Delitzsch, Leipzig, 1897.
2 Halevy, JA., vol. vi., ser. 3 (1874), pp. 461 seq.; Comptes rendus, vol.

iv., ser. 3, p. 477; vol. iv., ser. 3, pp. 128, 130; JA., vol. viii., ser. 7, pp.
201 seq. Also his book Recherches critiques sur l'origine de la civiliza-
tion babylonienne, Paris, 1876. See Weissbach, Die sumerische Frage,
p. 183, for further references.
50 T. D. Prince, [1904.

incalculable difficulties, because nearly all the Sumerian texts

which we possess are written in an idiom which is too evidently
under the influence of Semitic. With the exception of certain
very ancient texts, such as those published by Professor Hil-
precht in the OBI. vol. I, i. and ii.1 and probably also the
Gudea inscriptions,2 the Sumerian literature, consisting largely
of religious material such as hymns and incantations, shows a
number of Semitic loanwords and grammatical Semitisms and in
many cases, although not always, is quite patently a translation
of Semitic ideas by Semitic priests into the formal religious
Sumerian language. We must believe that the Sumerian at a
comparatively early date began to be used exclusively in the
temples as the written vehicle of religious thought in much the
same way as was the mediaeval Latin of the Roman Church.
Professor Paul Haupt may be termed the father of Sumerian
etymology, as he was really the first to place this study on a
scientific basis in his Sumerian Fatmily Lawus. Professors
Jensen and Zimmern4 have done excellent work in the same
field and, together with Haupt, have established the correct
method of investigating the Sumerian vocables. These words
should be studied only in relation to the Sumerian literature.
They should on no account be regarded as being etymologically
connected with words in the idioms of more recent peoples, and
practically all such comparisons, for which there is really no
sound basis, must be considered to belong to the realm of pure
conjecture. Sumerian, as we know it tip to the present time,
stands alone, a prehistoric philological remnant, and its etymol-
ogy should be studied only in the light which can be got from
the Sumerian inscriptions themselves. It is, however, permis-
sible to cite grammatical and constructional examples from
other agglutinative idioms, whenever it becomes necessary to
prove the true linguistic character of some Sumerian peculiarity.

1 The Babylonian Expedition of the University of Pennsylvania, I.,

pts. i-ii., H. V. Hilprecht, Philadelphia, 1893-1897.
2 The Great Cylinder Inscription A and B of Gudea, by Ira Maurice
Price, pt. 1, Leipzig, 1899.
3 Die sumerischen Familiengesetze; also see his Sumerian grammar in
ASKT., pp. 133-147.
4 Jensen in ZA. and elsewhere; Zimmern, especially in Beitrdge zur
Kenntniss der altbabylonischen Religion.
Vol. xxv.] The TVoCab?tlaryof Sumerian. 51

For example, it is probable, as I have pointed out elsewhere,'

that tones existed in the primitive spoken Sumerian, not only
for the distinction of similar sounding words, but also for the
purpose of differentiating between the various grammatical
elements. As this latter phase of linguistic tones is practically
unknown and might militate against the real linguistic nature
of the Sumerian in the eyes of the Halevyan school, I have
cited the existence of genuine graminmatical tones in the African
Yoruba language, simply for the purpose of demonstrating a
linguistic parallel.2 In the same manner, the existence of the
incorporated object in the verb may be parallelled by referring
to other languages exhibiting these phenomena. On no account,
however, should one venture to assume a linguistic affinity
between Sumerian and any of these idioms, nor is it possible to
connect Sumerian as yet with any language by dint of probably
accidental verbal similarities.
? III. It is really not at all surprising that the Halevyan
theory as to the cryptographic nature of Sumerian arose. The
first impression given by the distracting tangle of the Sumerian
word-list is the conviction that it would be impossible for such
a vocabulary to exist in any regularly developed language.
Here one finds the same sign denoting pages of meanings, many
of which are seemingly unconnected with any others belonging
to the sign in question. There is also, in a great number of
cases, a multiplicity of meanings attributed, apparently arbitrar-
ily, to the same sound-value or word. For example, some scholars
have asked very pertinently: how could any real language give
the same sound-value to a great variety of meanings 9 Besides
all this, there are, as indicated above, many passages containing
Sumerian words which strikingly resemble the Semitic equiva-
lents and which seem to be mere arbitrary perversions of Semitic
originals.' All these facts taken by themselves would be suffi-

'AJSL., xix., n. 4, p. 205.

2 AJSL., xix., n. 4, pp. 205-6.
3Let us take only two cases cited by Fried. Delitzsch in his Assyrische
Grammatik, p. 62, at the time when he was a follower of the Halevyan
theory: BAL =pal2 ' royal insignium, reign-year.' The original mean-
ing of BAL is ' open, break through,' and from this come the undoubt-
edly Semitic developments ' proceed powerfully, be hostile, oppress,
destroy, cross over; dig, draw water, be sunken; axe, wall-breaker,
spindle,' and finally, ' royal insignium,' = an instrument like an axe (?),
52 1. D. Prince, [1904.

cient to convince most philologists that we have to deal here

with an arbitrarily arranged cryptogram rather than with a
language. I repeat "if taken by themselves," but the main
point is that these phenomena cannot be taken by themselves.
They are mere superficial evidences of deliberate later inter-
ference with the growth of the language, introduced by a priest-
hood who spoke a radically different idiom-Semitic Babylonian
-but who had for religious reasons adopted the Sumerian as
their formal written vehicle. The proof of the original lin-
guistic character of Sumerian is found in the copious evidence
presented by even the latest and most Semitised Sumerian texts.
? IV. Briefly considered, there are four striking proofs that
the Sumerian literature is based on a primitive language:
1. Sumerian has unmistakable internal phonetic variations,
especially between the two main dialects Erne-ku and _Eme-sc8a,
and also within the Erme-kcuitself.1 These show a distinct devel-
opment of sound which may have been peculiar to different
periods in the growth of the language.2 The most noteworthy
sound-changes betweeu EK. and ES. are as follows: EK. g=
ES. mw; thus EK. gir 'foot'= ES. 'meri'; EK. g =ES. b;

which word came to mean later the 'reign-year of the king.' It is

highly probable that palfi is a Sumerian loanword in Semitic from BAL
(bal). Furthermore, there is no reason to believe that the Sum. value
buru, originally 'depression, hollow' is a perversion of Sem. :yN
'a pit,' but, on the contrary, that :yN 'pit' (Assyr. biru, birtu
'well'), owing to its resemblance to Sum. buru started the chain of
ideas which swelled the number of later equivalents of bur (= the
corner-wedge) to more than forty! Mnemonic paronomasia, the delib-
erate association of Semitic words with Sumerian words similar in
sound, played a most important role in the formation of the Sumerian
syllabaries. This point will be treated of in a subsequent paper and in
my forthcoming Lexicon. It is, of course, doubtful in some cases,
whether the Sumerian value is not really a deliberate perversion of the
Semitic word on the part of Semitic priests who had lost the original
Sumerian equivalent. In this connection, cf. egir=arku ' rear,' where
Sum. egir looks like a metathesis of arku. On the other hand, the
Semites borrowed far more from the Sumerian vocabulary than con-
versely, a fact which has been admirably demonstrated by Pontus
Leander (see this article, p. 52, note 2).
1 For my opinions in regard to EK. and ES., see JAOS., xxiv., pp. 105 ff.
(The Hymn to Belit. K. 257). Cf. also Leander, op. cit. (pp. 33 ff).
2 Pontus Leander, Uber die sumerischen Lehnworter im Assyrischen,
Uppsala, 1903.
Vol, xxv.] The Vocabulary of Sumerian. 53

thus, EK. duga ' knee' = ES. zeba. I believe that there were
two g's in EK., i. e. 1) a nasal g (Ag) = ES. m, which was prob-
ably not a clear m, but a nasal labial obscuration as in modern
Gaelic larnh 'hand'; 2) there must also have been a genuine
hard g represented by ES. b, as just indicated. There is also
an EK. g which varies to d in ES., thus, EK. igi ' eye, face ' =
ES. ide (cf. EK. gar = da-ar, V. 11, 28b). In the vowels, we
find also EK. u = ES. e, as EK. tu ' dove' = ES. te. A very
curious consonantal interchange is EK. n ES. s, as EK. ner
'ruler'= ES. ser. EK. n also-=ES. I, as EK. sudun 'yoke'=
ES. suydul. The so-called dialectic variations within the limits of
EK. itself consist chiefly of elision of final consonants, as in the
very numerous instances where we find such double values as
pag, pa _ XU, gig, ge = MI, gud, gu = GUD, etc. Leander
(see above, p. 52, n. 2) has written an elaborate treatise on this
subject, in which he assumes, following Jensen, that the fuller
forms, i. e. jpag, gig, gud are the more ancient ones. In other
words, he believes that the dialectic differentiation within the
EK. is a temporal one. The question is very difficult, as we
find occasionally the full and the apocopated form of the same
word in the same sentence. Thus, ad-a-ni's-gar-ra gd = mltir
gimilit 4bisu 'one who avenges his father' (Br. 7261). Here
,ugarra = gimilmtI 'vengeance,' lit. 'to make (gar) power
(su).' The word occurs in combination with the shorter form
gd (from gar ' he makes'), i. e. ' he makes sgarra:= vengeance.'
If gar belonged to one dialect and gad to another, should we
find them together in this way ? It is possible, of course, that
the older gar-form survived in the compound instead of in the
verb-form. All these phonetic changes are widely different to
those seen in Semitic and evidently depend on quite different
principles. Certainly no cryptogram based on Semitic could
exhibit such phonetic phenomena as we have here.
2. Sumerian has a system of vowel harmony strikingly similar
to that seen in all modern agglutinative languages and it has
also vocalic dissimilation like that found in modern Finnish.
Vowel harmony is the intentional bringing together of vowels
of the same class for the sake of greater euphony, while vocalic
dissimilation is the deliberate insertion of another class of vowel,
in order to prevent the disagreeable monotony arising from too
prolonged a vowel harmony. The following few examples will
VOL. XXV. 6
54 1. 1). Prince, [1904.

suffice to prove that we have here real linguistic phenomena,

although the number of instances might be increased ad
nauseam. Thus, IV. 30, obv. 2, 22: in-di in-di gaba kur-ra =
illak il(l)ak ana irat irpitim ' he goeth, he goeth unto the bosom
of the earth.' Here we find the soft form di = du 'to go,' in
harmony with the vowel prefix in-. In IV. 30, 19-20b-23-4b:
narn-ba-ab-bi-en-Izl taqabbi 'speak not to him," but IV. 11,
31-3: na-m'a-un-nib-bi=ul iqabbi 'he speaks not to him.' Cf.
also HT. 126, obv. 53-4: mu-un-na-ab-bi-e-ne 'they shall speak
to him.' In nambabbien, the neg. prefix nam and the infix bab
are in harmony, and in dissimilation to the stem bi. In numun-
nibbi, the neg. prefix nib and the verbal prefix m-un are in har-
mony, while the infix nib is in harmony with the stem bi, both
being in dissimilation to the prefixes. The same phenomenon
is seen in munnabbiene. Furthermore, in IV. 9, 28a we find
an-sud-dan = kima samn ' like the heavens,' where damnstands
as a dialectic variant for the usual gim (EK. g = ES. d).' In
Turkish and Finnish the vowel harmony is based on similar
principles. Thus, Tk. baqqjak 'he will look,' but sevejek ' he
will love.' In Finnish tyohan, toihin, kyldhan, all occur with
soft vowels. Dissimilation really exists in spoken Turkish in
such forms as al-di-lar for aldular ' they have taken,' but it is
not recognized by the grammarians. In Finnish, however, we
find it in full force as patoja for pataja, annoin for annain (see
Eliot's Finnish Grammar, pp. xii. 9-10). A cryptogram with
vowel harmony could only have been invented by persons who
spoke a language exhibiting this phenomenon, which is certainly
not present in Semitic.
3. The Sumerian postpositions are used almost exactly like
those in Turkish and Finnish. Thus, Sum. e--da 'in (da) the
house (d) - Tk. evde (de I'in'); Sum. ad-da-na-ru ' unto (ru)
his (na) father (adda) = Tk. baba-sy-na; baba = 'father';
sy his'; n is the phonetically inserted consonant and a is the
postposition 'unto.' Here it should be noted that the Sumerian
custom of placing a postposition after a noun and its qualifying
adjective has a perfect parallel in both Turkish and Finnish.
Thus, a-ab-ba-ki-nin-dagal-la-(a) 41e - ' on (s() the sea (a-ab-ba)
and the wide (nindagalla) earth (ki)'; Tk.: deniz (sea) tee (and)
1 Cf. also Prince, AJSL. xix., n. 4, p. 207, s. v. zi-zi-de ' before thee,'
where zi is explained as a probable by-form of za-e ' thou.'
Vol. xxv.] The Vocabelcary of Sionerian. 55

biyuk (great) arz (earth) V2zerinde (upon); Finnish: paljaan

taiwaanV cllat 'under (alla) the open heaven' (see Eliot, Finnish
Gratmmar, p. 205). In all these cases the postposition is
regarded as qualifying both the noun and the adjective. A full
discussion of the Sumerian postpositions will be found HT. p.
141 ? 11. Prepositions, although rare, are found both in Sume-
rian and Finnish (see HT. 141 ? 14 and Eliot, op. cit. pp. 203 ff.).
Postpositive inflexion is as foreign to Semitic as is vowel har-
mony and could not possibly have been invented by Semitic
cryptographers.
4. I have discussed at some length elsewhere the Sumerian
verbal prefixes (see AJSL. xix. No. 4, pp. 206 ff.). It will be
sufficient to note here that the incorporation of the verbal object,
which is the only method in Sumerian of expressing the pronomi-
nal object, has its parallel in other languages of the agglutinative
type. Thus in Basque, the present tense is never without the
object 'it '= d, as dakart-d-ekart ' I bear it' ('I '= t; 'bear'
- ekar; 'it' d).' Practically the same peculiarity is seen
also in the American languages. Thus in the Algonquin Aben-
aki: k'nantiol 'I (1) see (narm) you (-k)' (Prince, Miscellanea
linguistica Ascoliana, p. 357, Turin, 1901). The same peculiarity
appears in the modern Latin idioms as in Frenchje l'airne 'I love
him'; Ital. and Span. io (yo) I'amo, although these are inflex-
ional languages ! To suppose that Semitic priests, who could
have been familiar only with their own tongue, should have
departed in a cryptogram so far as this from their natural suf-
fixed object is the height of absurdity.
? V. Furthermore, one does not have to go far to seek real
linguistic cryptograms and secret idioms. We have several
such in English, perhaps the best known of which is the
so-called costermongers' back-slang which has existed for a long
time as a semi-jocose jargon among the small tradesmen of
certain sections of London. In this dialect they say: cool the
(lelo taoc 'look at the old coat'; 'potatoes' are rattatts (back-
wards for ' tatur ') and a Jew is a Wledge!2 This kind of trans-
position or rebus is peculiar to a number of secret languages.
Such an idiom exists in modern Arabic, especially among school
' See Van Eys, The Basque Langutage, 1883, pp. 29 ff.
2 J have heard vagabonds speak this jargon so rapidly that it was quite
impossible to follow their meaning.
56 JJ -D. Prince, [1904.

children, who say fuj el-ldjir for suf er-rdjil ' look at the man.'
Also in the Spanish thieves language we find such inversions as
tapla for plata 'silver'; demias for medias ' stockings,' etc.
Similar transpositions occur in certain phases of French Argot
as loffe for folle 'mad' (fem.). Most interesting in this con-
nection is the secret idiom of the Irish tinkers which is in use
in this country and in England at the present day. This jargon,
which has long been a puzzle to philologists, is now definitely
established as being fundamentally Irish Gaelic inverted,
although not always with absolute correctness. Thus they say
leichin 'girl' for Irish cailin,; mailya 'hand' for Irish lamh,
etc.' I cannot leave this curious subject of secret languages
without alluding to the incantation language of the Greenland
Eskimo enchanters. In this idiom, which is in conventional
ritualistic use in all incantations, we really find what Halevy
and his followers believe they have discovered in Sumerian, i. e.
a priestly system of disguising the ordinary speech. This
Eskimo shamans' language consists partly of descriptive terms
such as ' boiling place ' for ' kettle,' ' dwelling ' for ' house,'
'tusked-one ' for 'walrus,' etc., and partly of deliberately
chosen archaic expressions which are easily recognizable by com-
parison with other Eskimo dialects.2 In none of these jargons
do wefind any grammatical development other than that of the
language on which they are based.
? VI. This is all to the point in connection with Sumerian,
because these very principles of inversion and substitution have
been cited as being the basis of many of the Sumerian combina-
tions. Deliberate inversion certainly occurs in the Sumerian
inscriptions and it is highly probable that this was a priestly
mode of writing, but never of speaking, at any rate not when
the language was in current use. It is not necessary to suppose,
however, that this device originated with the Semitic priesthood.
'The late Charles G. Leland was the first to bring this jargon to the
attention of scholars (Gypsies, 1886, pp. 354 ff .). Kuno Meyer, in a letter
to me written in 1896, says: " (It) is a kind of back-slang of the Irish
(Gaelic) language. It is evidently not a modern invention, but some-
thing like it has been practised among the Irish ever since the eleventh
century at least. I now believe that the idea was taken from the late
Latin grammarians, but have not worked this out." Mr. Meyer wrote a
paper on the Tinkers' language, which appeared in the Journal of the
Gypsey Lore Society, vol. ii (now extinct).
2 This information was kindly given me by Prof. Franz Boas of
Columbia University.
Vol. xxv.] The Vocabulary of Sumerian. 57

It is quite conceivable that the still earlier Sumerian priesthood

invented the method of orthographic inversion which, as I have
already shown, is the very first device which suggests itself to
the primitive mind when endeavoring to express itself in a man-
ner out of the ordinary. Evident inversions are, for example,
the name of the fire-god Gibil, which was written Bil-gi. Gi
means 'a reed' and bil (NE)' 'fire,' so that the combination
must have meant 'a fire-brand.' We find also Girsu for Sugir,
LLugal-si-kisal for Lugal-kisal-si, Sir-la-bur for Sir9-bur-la and
En-zu-na for Zu-en-na.
Especially worthy of mention is the common inversion found
in the ancient style of writing the word for ' ocean,' i. e. zu-ab,
quite evidently for ab-zau ' sea of wisdom,' the abode of Ea, the
god of wisdom. The Semitic Babylon apssA and all its Semitic
derivatives was in all probability a loanword from the Sumerian
theological vocabulary. The Semitic form apse is in itself defi-
nite proof that the Sumerian word was pronounced abzu and not
zu-ab, as written. Ab (ab-ba) is the usual word for ' sea,' often
written a-ab-ba=tidrntu. It is possible, though not certain,
that this ab (ab-ba) was connected with the Sumerian expres-
sion for 'water'= a (vide infra ? VIII). Zu meant primarily
'knowledge'=nirienqu, V. 30, 48 a; II. 16, 65 a. The original
sign for zu was the element 'great' written inside of an eye,
i. e. 'to be great of eye or perception,' hence ' to know'
(Delitzsch, S~ystemv,p. 138). The well known equivalent bit
nimeqi 'house of wisdom' for ;Dc-ab depends on a later mis-
understanding of this value ab (Leander, op. cit., p. 5). The
scribes evidently confused it with ab2 ' dwelling.'
1 The oldest form. of the sign for ' fire' (probably ca. 6500 B. C.) has re-
cently been discovered in the General Theological Seminary (N. Y.) col-
lection of Assyrian antiquities and brought to my attention by Mr.

Robert Lau. This sign is , which is quite evidently a coii-

firhiationof Prof. Delitzsch's view (System, p. 178) that the fire-sign

represented the generation of flame by turning one piece of wood
against another after the manner of some modern savages. The sign
just given seems to me to be the pictograph of such a primitive fire-
producing instrument, i. e. it represents the straight tinder sticks against
which the bowl-like end provided with a handle was rapidly rubbed.
The sign is used as a suffix -ne in the inscription (see Barton, JAOS.
xxiii. p. 23).
2 ABhas also the value 6 'dwelling,' which was probably the full form
of 6=bitu ' house.'
58 J. -D. Prince, [1904.

There are five distinct words ab in the Sumerian vocabulary;

viz., ab6 (AB) 'enclosure, dwelling,' from which we have the
Sem. aptu 'birdsnest' (llwb. 111). The sign AB was archaic-
ally a pictograph of a space. It is probable that aptu is a Semitic
loanword from this a6b. AM2(AB) ==tamtu 'sea,' Sc. 95 et passim,
which meaning may be an extension of the original sense of
space, i. e. the sea is the space par excellence. When written
a-ab-ba we must translate it 'water-space.' Closely connected
here is abl= arax Tebet -=Sum. ituc ab-ba-ud-dut ' the month of
the coming forth of water, the month of floods.' TeJbetis from
Sem. tebPd ' dip, sink in' cf. II. 49, 4 e, for variants of this
month-name, and see Muss-Arnolt, JBL. xi. 170. Also ab2=
amel irris.u ' cultivator' or 'irrigator' (in this sense ApV has
the meaning 'to plant,') but am 6e irris'u has also the Sum. equiv-
alent PIN= engar=Sem. ikkaru ' husbandman.' It is possible,
as I have just indicated, that ab= tdlntut ' sea ' was really a cog-
nate in fuller form of a-' water' and that it was associated in
the AB group owing to the earlier scribes arbitrarily connecting
ab=td'mtu with the idea of 'space, dwelling.' The third ab3
(AB)= abut 'father' in several passages. Here the question at
once arises as to whether this al)3 may not be a Semitic loanword
in Sumerian, but too hasty a judgment regarding it should not
be formed. It is probable that ab3=aabu 'father' is the ab
seen in ab-gal abkallurn? 'leader' (cf. 32-8-16, 1. col. i. 31:
niumn-ne=ab-ga-al). This ab means also neasikd ' prince,' Br.
3820, and also, probably owing to Semitic influence, sbIbu ' an
old man,' Br. 3821 (cf. nain-ab-ba= sibfitu 'old age,' II. 33,10 c).
I am strongly inclined to believe that the original sense of this
a43 was nasika ' prince,' and that this word was transferred to
abi ' father, leader' by Semitic analogy (ab3= sibyu 'an elder,' as
noted above). The fourth ab4 is the value attached to the sign
LIT, Sb. 254; ab = arxn 'road, way.' This sign probably indi-
cated the lines of a road + the corner-wedge denoting compact-
ness, i. e. ' a highway' - ar}xn. The same sign with the value
litrt Sem. littt ' offspring.' Here the sign must have been applied
with the idea of 'pairing,' obtained from the same double lines.
The value lit for this sign is undoubtedly of Semitic origin from
littt, as is also the value rimn no doubt suggested by rtnu ' the
womb,' owing to association with littu, although LIT is not the
sign for ' womb.' Finally, in this connection, the fifth ab5 (AB)
Vol. xxv.] The T`ccbulatry of Stenerian. 59

is the ab of the pronominal suffix of the third person, probably

cognitive with the verbal prefix ab and the possessive suffix -bi.
These five values of ab certainly seem to indicate that when
the language was a spoken idiom there must have been tone
differentiation in this case. In Chinese, for example, we find
the following four tones for the syllable mo, i. e. mo' 'feel,
touch'; Mno2 4grind. rub'; mo3 'obliterate,' and mo4 'after-
wards, at last' (Stent's Chinese and English Vocabulary, pp.
417-18).' Here the first three mo's are certainly variations of
the-same fundamental idea. In the same way, it is quite possi-
ble to believe that ab' 'dwelling,' abh 'sea,' ab3 'leader,' and
ab4 ' road' formed four distinct tones. Ab4 ' dwelling' and ab2
'sea,' however, may have been uttered with the same tone.
This would leave the grammatical ab, prefix and suffix, to have
the fourth tone (eight tones are physically possible) or else we
may assume that it was pronounced like any one of the ab-tones
just conjectured, since its position in the sentence-construction
could leave no doubt as to its meaning. Of course nothing cer-
tain can be postulated with regard to Sumerian tones. If they
existed at all, as they must have done if Sumerian was ever
spoken, they must have died out at a very early date, so soon
as the language began to be the purely written ceremonial lan-
guage of the Semitic priests. In cases where an astounding
multiplicity of meanings are assigned to the same phonetic
value, the tone theory does not, of course, solve the problem
entirely. In such instances, as the meanings in the following
word-list attributed to the value a (e), the most we can do is to
collect the fundamental ideas belonging to the sign and word and
assume that each of these ideas was originally represented by a
distinct tone. It is curious to note that in every such example
the number of conjectural fundamental tones never exceeds the
possible number eight, the physical tone limit.
I This information
was kindly given me by Prof. Fried. Hirth of
Columbia University. Sentence tones occur in English and other Euro-
pean languages expressing doubt, interrogation, surprise, etc. Perhaps
the best illustration of tones in a European idiom is seen in modern
Swedish, where actual differentiation in meaning is expressed by differ-
ence of tone. Thus, han ar i brunnen (rising and falling tone) ' he is
in the well," but han ar brunnen (falling and rising tone) 'Ihe is burned,'
although this distinction is not invariably observed. The speakers often
leave the sense to be understood by the context.
60 J. -D. Prince, [1904.

? VIII. In order to illustrate more thoroughly these perplex-

ing difficulties, let us take the sign A, which in both Babylonian
and Assyrian is a later development of the earlier convention-
alized pictograph of water moved by the wind, i. e. the wave
motif (Delitzsch, System, p. 130). The sign A, which is named
au (V. 22, 43 a) after its most important sound value, has six
different phonetic equivalents; viz., a, me, bur (pur), e, dur
(duru) and id, arranged in the order of their respective impor-
tance. There can be no doubt that the sign originally meant
simply 'water' and yet in Brfinnow's List there are two full
pages of meanings devoted to the simple A, a careful analysis
of which will serve to elucidate the principles followed by the
Semitic compilers of the later Sumerian syllabary.
A. I divide the meanings of A (a, me) into four groups as
follows
1. The water-group (from a=m,, ' water'; see below).
A (a=butuqtu 'overflow,' only V. 22, 49a (Br. 11331). -Bu-
tuqtu also=a-tar V. 31, 30e (Br. 11382); lit. 'water cutting
through' tar= bataqu 'cut through' (Del. System, p. 102); a-xul
(IV. 26, 19a); lit. ' evil (xul) water.'
A (a)=dimtuI 'tear,' only DT. 67, obv. 12; usually a-igi
(er) 'water of the eye' (Br. 11609).
A (a)=masqqitu 'irrigation,' only V. 50, 52 (u-a=ritu u
masvqitu 'food and drink').
A (a-a) = milu, from e-Ig ' go up,' 'high water,' usually a-lig
'strong water' (Br. 11538), with value ela, perhaps e-ba (?).
If the Sum. word is ela, it may be a loanword from Sem. e1A
'go up.' MJfu also- a-kur w water of the land'; mnku ma'du
'great flood, II. 39, 9g. There is unquestionably a Semitic
paronomasia between kur (read mcat) in Sem., and mna'du!
A (me) =rni 'water,' Br. 11347 passinm.
A (a) =naqu 'lament' connected with dimntu 'tear' (only
II. 45, 35e; Sa. vi. 25). The phonetic value a is not given,
but is probably understood. Note the value cas(til) = tani4atuin,
V. 40, l0gh 'a lament.'
A (mne)=rax4au 'inundate, wash out,' only V. 22, 76a, Br.
11351 (also RI-RI and RA, the latter probably mnemonically
associated with rax au).
A (me)=rutbu 'moisture' (Br. 11358).
All these meanings derived from A -mg are evidently Semitic
attempts to synonymize the idea 'water,' and this appears
Vol. xxv.] The Vocabulary of S5umerian. 61

especially true in cases where the synonym has a more usual

Sumerian equivalent of its own, as dimtu ' tear,' ordinarily=A.
SI (er).1
2. Closely connected with this idea of 'water, moisture' is
the second group of meanings applying to copulation. Here it
should be noted that the word 'water' is used in Arabic for
semenely,i. e. Egypt. Arab. rnoiyet el ab 'seed of the father' (cf.
also Heb. nKID, probably=the same). I consider that this
whole second group was of Semitic origin.
A (a-a) =abu Br. 11324, passim. The word for 'father' is
also ad-da=AD, the primitive sense of which is ' dwelling-pro-
tector,' also axu 'brother' means 'protector' (Br. 1142); viz.,
'the protector of the house' (see Delitzsch, System, p. 58). I
am inclined to see in ad-da=a-a ' father "a paronomasia. A-a
' father' may be the softened form for ad-da, a phenomenon
seen in modern Cuban Spanish hablao for hablado ' spoken.'
This a-a pronunciation for ad-da may have occurred dialecti-
cally. Then the pronunciation a-a=ad-da may have suggested
to the scribes the word a ' water, semen' and they accordingly
wrote it with the water-sign A, i. e. 'father'-' the seed pro-
ducer.' Cf. here a-a-a=abi abi 'grandfather,' II. 32, 61c.
As a-a also= I grandson ' (see below in this section), this deriva-
tion was probably made at a time when the language had become
purely orthographic.
A (a) -amettu ' human kind ' (Br. 11326), plainly a deriva-
tive from the idea ' seed' (see Akk. Spr. xxxviii). The usual
ideogram has the value gisgal=ES. rntul= amneltu.
A (a)= aPlu 'son' (Br. 11328). In 11344 A( )a = mndru
'son' and AL. 314 a-a=binbini 'grandson.' These, like the
preceding word, are variants of the idea ' seed.'
A (a-a) = 6r'A 'be pregnant' Br. 11333 (Hquab. 130). The
origin of this in connection with ' seed' is perfectly patent.
The regular ideogram is the sign having the value fves- (Br. 8101),
i. e. s-4 ' interior ' with the water-sign A written inside

'A. 9I.= a-igi 'water of the eye,' hence 'tear'=er. This er is a by-
form of the word es, which is also a value of A. AI. The etymological
connection between er and es is clear, i. e. rzs.
2 Even though
nNI: might have been a derivative and not a com-
pound word (Gray, Proper Names, p. 25), the narrative of Gen. xix. 34 ff.
shows that its popular etymology at least was ' seed (water) of the
father,' a proof that ' water' was used in this sense in ancient Hebrew, as
it is to-day in Arabic.
62- J. -. Prince, [1904.

A (a-a, me) =baml 'beget' (V. 22, 60; 72 abd). Another

word is niud=band, Sc. 51. The Heb. TMZ also has the sense
'beget."
A (a) =ipratu, V. 21, 4 (Br. 11338), probably 'figure, image,'
connected with Aidt (D. Prol. 33).
A (a, me)=lubsu, V. 22, 62a (Br. 11341). This word does
not mean 'garment,' but must denote ' offspring, brood' (cf.
Uibistu, Hwb. 372). Also V. 37, lOdef, we find the corner-
wedge = su-us-= s-aqd sa lubs ' to moisten, said of lubsu.' This
must have a sexual signification.
A (a) rix4tu ' sexual love,' Br. 11353 (cf. marxitu 'wife,'
the usual ideogram of which is XI-NIR, perhaps=' excellent
(XI.) in size (NIR).'
A (a) =rikibtu, Br. 11354, a synonym of rix4tuI 'sexual love'
(Hfwb. 620). The word rikibtu, lit. ' mounting,' is from rakdbu
'ride astride of.' Hence we have the next value:
A (a)=rakaba 'ride astride,' only IV. 11, 41a and HT. 220;
ba-a=irkab and ba-an-da-a=irtakab. See Br. 11352.
A=pallu s-a raxe 'one who copulates,' Br. 11359. callut is
from faldlut ' to lie down to sleep,' here especially with a
woman, a syn. of uttulu ' sleep' in general.
A (a)= mailuI 'seat, bed, Br. 11343, a variant of the idea
'lying down.' Usual ideogram is nab and ki-na ' place for
lying down' (ffiwb. 406).
A (a) =nilxu 'rest' (Br. 11349), a word from the same idea
as rnailu.
A (a) =paseadx 'to be pacified, quiet, at rest.' This a is
probably a synonym here in connection with reails, nu'xu. The
value d (ud-du) ' go out' also=pacsa'xu in the sense ' recover from
a sickness' (cf. French; sortir de mnaladie) and it is highly likely
that ct=pasdxu is a paronomasia from e=pcpsaxu, especially as
the values a and e interchange (see below, this section D).
3. A third group of meanings formed more directly from the
wave idea is represented by the following word denoting
' effulgence.'
1 ;5-J:1 ' build a house,' i. e. ' form a family,' a paronomasia on In son;
cf. Gen. xvi. 2; xxx. 3; 1n::: Adnb 'I shall be built up by her' (a
childless wife by means of a concubine). This use of In: probably
occurs in V cxxvii. 1: 'except Jhvh build the house, they labour in vain
that build it.' This, according to Prof. Haupt in his lectures on the
Psalms of the Return, was an allusion to Nehemiah's inability to 1&1 1
ftln, following the theory that Nehemiah was a eunuch.
Vol. xxv.] The Vocablary of' Stmerian. 63

A (a)-ebbut 'shining, glistening, effulgent' (Br. 11335) per-

haps also with the value dur (see below this, section E). This
notion must have been developed from the shining ripples of
water. In Turkish si=zboth 'water' and also 'the lustre of
a jewel, (Redhouse, Turkish and English Lexicon, p. 1188).
In English also we speak of gems of thefirst water. Here it
should be noted that za=abnt 'stone' (Br. 11721) is also a
derivative like A (a) from the wave-motif and probably had the
meaning 'shining stone' or 'jewel' (see Delitzsch, System, p.
130).
4. Finally in this connection in the last A (a, me) group we
find A (a) as a mere ending of the statu8spronomiflalis.
A (a-a) =-an dku 'I,' V. 22, 69a only (Br. 11327).
A (a-a)=atta 'thou,' V. 22, 70a only (Br. 11329). A (a)
occurs also passinm in the third person (see Br. p. 548). Here
should be placed the indefinite form a, seen most commonly as
a verbal prefix (Prince, AJSL. xix. p. 211). It is unnecessary
with Jensen (ZA. i. 61) to connect this a with me=ren'ro the
verb ' to be' used with all three persons, because this a is also
a common verbal prefix, evidently with the value a and not me-,
and may indicate all three persons; cf. HT. 123, rev. 1: a-rab-
tag-tag= aptabsilki 'I implore thee' (fem.) (Prince, loc. cit.).
The ending -a==ana-ina 'unto, into,' Br. 11364-5 passim,
is an element which is probably connected with a= axldap ' how
long,' Br. 11325, mostly in ES. texts (Hwrb. 44). The usual
ideogram for axulap- is sux-a, HT. 122, obv. 12 (Zb. 28; HT.
115, rev. 5). This -a seems to denote ' duration' and to be
cognate with the status prolongationis (HT. 135 ? 4). It should
be noted that when the sign A was used in making combina-
tions, its sound value was always a.
A (a) =xanat2u only V. 30, 62a (Br. 11384). This is an
error, as the a here is only the a of prolongation for surub-a=
xamdtu sa kab bi 'to burn, said of consuming by fire.'
The original tones of A (a) must have been identical with
the four idea-groups just indicated (see below, this section F, on
IDrca).
B. The value A (me) must now be considered. This has the
meanings banil 'beget' (Br. 11330), lubsec 'offspring' (Br.
11341), mil 'water' (Br. 11347), raxcpu 'inundate' (Br.
1 On men = all three persons, see Prince, AJSL. xix., n. 4, pp. 206; 226
64 J. -D. Prince, [1904.

11351), rimxu meaning unknown, but must be connected with

'water' (Br. 11355), and rutbu 'moisture' (Br. 11358), all of
which except rimxu have been considered under A (a). I
believe that this value nm is a later Semitic loan-value from mil,
pl. me 'water.' Note that banii, iub&Su, rutbu also have the
value a.
C. The value bur=A Br. 11318 is seen only in the name of
the Euphrates A-rat (Br. 11444; AL. 313), which was evidently
pronounced Burat. The regular Sumerian ideogram for this
river was ID. UD. KIB. NUN. KI (Br. 11662), i. e. ' the
river of Sippar'=UD. KIB. NUN. KI= Sum. Zimbir, V. 23,
29, probably the original form of the Semitic name Sippar.
Another Sumerian word for the Euphrates was -Buraqunu ' the
great (nuntu) river (bura)'. This word bura' is undoubtedly
the same bur=BUR and also-=-the corner-wedge. It means
'vessel, receptacle, hollow,' hence 'river-bed.' From -Bura-
gmmu no doubt comes the Semitic form Purattu, i. e. simply
'river''=bura+the feminine ending -tu. Therefore, when we
find A-rat=Purattu, it must be assumed that the water-sign A
was pressed into service here to denote the water KaT' $ox'q)v,in
lieu of the longer ideogram. The Heb. TWO, not attested with
certainty before DJe ;2 Greek 'Ev4pa'rvq, Old Persian (4fratu
(Spiegel APK. 211), must all be derivatives from Assyrian
Pturattu. The modern Turkish ]Iburad-st ' water of Murad '=

Euphrates is undoubtedly a popular variation of the original

Arabic Frat! In II. 48, 47; 50, 8; 51, 26, we find the form
Uruttu=Purattu. Uruttu is probably a derivative from araddu
'descend" and the form may be regarded as a later Semitic
paronomasia on Puratta.
D. A has the value e in a number of passages, owing to
vowel harmony and also to dissimilation. In K. 4225, dupl. 6,
a=e, which simply indicates that a may occasionally have the
value e as circumstances require.
A (e) occurs in II. 29, 20a, ac-nigin (a=e), probably=palgu
'canal,' lit. 'collection (nigin) of waters' (Br. 11676).
I The Babylonian Expedition of the University of Pennsylvania, I.,
pts. i-ii., H. V. Hilprecht, Philadelphia, 1893-1897.
2 Cf. Francis Brown (Robinson's) Hebrew and English Lexicon, p. 832

s V. 1t)-
3' See Muss-Arnolt, Assyrian Lexicon, s. v. uruttu.
Vol. Xxv.] The Vocabulary of Sumerian. 65

A (e) in ASKT. 75, n. 4, a-girn has as its gloss e-qi-me=

pas4xau (see this section A. 2, a=pas-dixu), a plain instance of
vowel harmony (Br. 11321).
The following cases of A==e are owing to vocalic dissimila-
tion:
A (e) occurs in V. 40, 1 and 4 e; u-a-(e)=s'apuuan (?) and
saua. This is doubtful, as the meanings of the equivalents are
unknown (Br. 6092).
A (e) is seen in II. 32, 13 g; a-gi(MI)-a=e-ga-a=agA 'flood';
lit. 'black (MI) water' (Br. 11593).
A (e) is found II. 39, 7 g; a-lig(KAL)=e-la or e-ba=emil
'high water' (see above, this section A. 1), Br. 11538.
A (e) appears II. 32, 52: a-ma-e-du (du=TU) 'the womb
that bears.' We expect ama 'womb,'+a-du. This sign for
'womb' aata (dagal) also-=rimu 'womb,' IV. 9, 24a. Now
the real sign arma(AM) =rrnu 'a bull,' so that there is proba-
bly a deliberate paronomasia in this instance.
E. The fifth value of A is dur=labdku, found only II. 43,
30 e, a doubtful word which is associated with azal and narabu
(cf. the form tulabbak, C. 45, v. 2). Here it should be noted
that a and ku both=dur. The god Sin=dingir a-ku, II. 48,
48 a and also dingir tur-kut (dumugu). This seems to prove the
dur value for A. In V. 29 44 g; za-gin-a=zagindur; V. 22,
11: TAK zagindara-zagin-du-ur (cf. ZA. i. 62, n. i.). Now
zagindur=uknit ebbu 'shining crystal.' Hence dur=a seems
to contain the idea ' shine,' seen also above in A= a. It is pos-
sible, therefore, that a=ebbu (Br. 11335) may have had the
reading dur which would seem to give the meaning ' shine' to
labadku, grouped together with azal and Scarabu, the meanings
of which are unknown.
F. A=id, only Sa. vi. 25 (i[?]-id) but A-idu 'hand,' K.
4870, 43 (Br. 11336); a-ni-su=ana idisuun ' unto their hands,'
so A=id is clearly a Semitic combination. This is plainly a
case where the later Semitic scribes confounded ID= a the reg-
ular sign for ' hand, arm, strength' (Br. 6542) with the water-
sign A. Such an interchange could only have taken place after
the Sumerian had ceased to be a spoken idiom, because we must
assume a difference of tone between ID-a and A-a. This
would give us the fifth a- tone (see above ? VII, and this see-
tion-A. 4).
66 J. D. Prefce, [1904.

? IX. No better example than A can be had of the manner

in which the original Sumerian syllabary was treated in the
course of centuries. Here we find a sign which primitively
meant only 'water' and most probably corresponded to the
simple vocable a-' water,' from which meaning, as just shown,
were developed; 1) almost every possible conception directly
connected with ' water'; 2) a number of ideas suggested by
the secondary sense of ' semen'(=' water'); 3) a word denot-
ing effulgence (=' shining water,' also with the value dur);
and 4) the a which was probably an arbitrary vowel used in
grammatical relations, having no connection with a= ' water.'
The makers of the syllabary were not content, however, with a
single value for this overworked sign. Still having 'water,
moisture' in mind, they added the me-value, most probably a
derivative from their own Semitic word me ' waters.' Then A
had to serve with the value bur, used with this sign originally
only of the Euphrates (Bunrt-nunu). True to the inherent
principles of vowel-harmony and dissimilation, A is pronounced
e in a number of cases, and finally, we find A=dur in the sense
'shine.'
? X. It is clear from the above study that many of these
evident accretions to the original meaning could only have been
due to a later interference with and an arbitrary development
of the primitive syllabary. This idea is confirmed more and
more as the entire Sumerian vocabulary is studied, every word
of which must be examined separately. With this object in
view, I am at present engaged in preparing what I trust shall
be as exhaustive a study as possible of the extant Sumerian
word-list. The work will be a Sumerian Lexicon and at the
same time a Prolegomenon for the further investigation of this
important subject.
? XI. The Sumerian literature, as we have it to-day, presents
a most curious phenomenon to the philologist; viz., a practi-
cally monosyllabic agglutinative idiom exhibiting all the marks
of this class of language, adopted by an alien priesthood as a
sacred tongue. In the course of two thousand years, this for-
eign priesthood, having lost at quite an early date the use of
Sumerian as a living language, continued to employ it as a
purely written vehicle, filling in its vocabularies with countless
synonyms and variations, which could only have grown out of
Vol. xxv.] The Vocabutcry of 8umeriamn. 67

an orthographic system. Of course, I do not mean to imply

that this priestly idiom was not even at the very latest date
recited orally in incantations, but in all probability not in the
ancient manner with tones. In spite of all this superimposition
of extraneous matter, however, the genuine linguistic character
of the Sumerian sentence can never for an instant be disre-
garded. If every word in the vocabulary were pure Semitic,
the grammar of the language even in its latest most garbled
form would point to its non-Semitic origin.

INDEX TO MOST IMPORTANT SUMERIAN WORDS.

The numbers and letters refer to the sections and divisions.
a, VIII. A, 1-4. ge, IV. 1.
a-ab-ba, VI. gibil,.VI, and p. 57, note 1.
a-ab-ba-ki-nin-dagal-la-(a)-9ft, IV. gig, IV. 1.
3. gir, IV. 1.
ab, VI-VII. Girsu, VI.
ab-ba, VI. gu, IV. 1.
ab-zu, VI. gud, IV. 1.
ad-da-na-ru, IV. 3. id, VIII. F.
A. Al.. p. 61, note 1. ide, IV. 1.
bab, IV. 2. igi, IV. 1.
bal, p. 51, note 3. lit, VI.
bi, IV. 2. Lugal-kisalsi, VI.
bur, VIII. C., p. 51, note 3. me, VIII. B.
bura, VIII. C. men, P. 63, note.
dam, IV. 2. meri, IV. 1.
dar, IV. 1. mun, IV. 2.
di for du, IV. 2. nab, IV. 2.
duga, IV. 1. nam, IV. 2.
dur, VIII. E. ner, IV. 1.
duru, VIII. E. pa, IV. 1.
e, VIII. D. pag, IV. 1.
e-da, IV. 3. rim, VI.
egir, p. 51, note 3. ser, IV. 1.
emeku, IV. 1. girburla, VI.
emesal, IV. 1. gudul, IV. 1.
Enzuna, VI. gudun, IV. 1.
er, p. 61, note 1. sugarra, IV. 1.
ek, p. 57, note 2 and p. 61, note 1. te, IV. 1.
gah IV. 1. t A, IV. 1.
gar, IV. 1. zeba, IV. 1.