Nyaya Mimansa

The Denotation of Generic Terms in Ancient Indian Philosophy: Grammar, Nyāya, and
Mīmāṃsā
Author(s): Peter M. Scharf
Source: Transactions of the American Philosophical Society , 1996, New Series, Vol. 86,
No. 3 (1996), pp. i-x+1-336
Published by: American Philosophical Society
Stable URL: https://www.jstor.org/stable/1006608
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
American Philosophical Society is collaborating with JSTOR to digitize, preserve and extend
access to Transactions of the American Philosophical Society
This content downloaded from

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
All use subject to https://about.jstor.org/terms
Transactions of the American Philosophical Society
Held at Philadelphia
For Promoting Useful Knowledge
Vol. 86, Pt. 3
The Denotation of Generic Terms

in Ancient Indian Philosophy:
Grammar, Nyaya, and Mimamsa
PETER M. SCHARF
Brown University
American Philosophical Society

Independence Square * Philadelphia
1996

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Copyright @1996 by the American Philosophical Society for
its Transactions series, Vol. 86, Pt. 3. All rights reserved.
Reproduction in any media is restricted.
ISBN:0-87169-863-3 LC: 94-78520

US ISSN: 0065-9746

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Table of Contents
Table of Contents .......................................................................iii

List of Tables ............................................... ...........................vii
Preface .......................................................................................ix
Introduction ............................................................................
Research Review ..................................................................5
Translations ..................................................................7
Studies..........................................................................8
The Meaning of the Term 'adkrti' .................................
Overview of the W ork .......................................................18
G ram m ar ...............................................................................21
1.1. Background ..............................................................21
1.2. Substance (dravya) and Class Property (akrti) ............23

1.2.1. Panini 5.1.119 .................................................23
1.2.2. 'Dravya' Used to Mean
an Individual Object .......................................2
1.2.3. Akrti: Permanent versus Transient ...................27
1.2.4. Shape versus Substance
as the Alternatives...........................................29
1.2.5. Definitions of a jati under 4.1.63 .....................30
1.2.6. Sivasutra 1, A i u n..........................................34
1.2.7. 'Akrti' and 'Jati' are Synonymous
under 1.2.58 and 1.2.64.................................35
1.2.8. Dravya: an Individual Object
under 1.2.58 and 1.2.64 .................................36
1.2.9. Conclusions Regarding the Term 'dkrti'
and Other Term s ....................................... 38
1.3. The Denotation of Generic Terms ..............................39

1.3.1. Denotation is Natural .......................................40
1.3.1.1. Concomitant Presence
and Absence ........................................40
1.3.1.2. One Word Used for Each Object
Denoted .............................................41
1.3.1.3. The Grounds for the Rule
that One Remains ................................42
1.3.1.4. One Word Naturally Denotes
M any M eanings ..................................45
iii

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
iv DENOTATION OF GENERIC TERMS
1.3.1.5. Reasons for Positing

Generic Propertis ................................47
1.3.2. A Generic Term Denotes a Class Property........48
1.3.3. A Generic Term Denotes an Individual
Substance ........................................ ......... 5 2
1.3.4. Denotation of a Class Property Vindicated...
1.4. Equivalence of the Two Views................................

1.4.1. Enjoined Universally, Fulfilled
in a Particular .................................................5 8
1.4.2. Number ...........................................................61
1.4.3. G ender ............................................................62
1.5. Solutions on the View that a Class Property is

D enoted .....................................................................64
1.5.1. Properties are not Innate ................................
1.5.2. Is the Gender Problem Solved
by the Grammarians' Conception
or by a Speaker's Intention Alone? .................65
1.5.3. N um ber ...........................................................72
1.5.4. Agreement in Gender and Number.................
1.6. Patanjali's Solution ....... ... ........... ........................ 7

1.7. Quality Words .............................................79
1.8. Panini 1.2.58 .............................................................84
1.9. Conclusion of the Grammar Section ...........................87
Panini 1.2.58 Translation...........................................................91

Panini 1.2.64 O utline.................................................................99
Panini 1.2.64 Translation.........................................................101
N y y a ....................................................................................1
2.1. B ackground .......................................... ............... 151
2.2. The Nature of a Generic Property, an Individual,
and a Form ............................................................152
2.3. A Generic Term Denotes All Three Together ...........
Gautama 2.2.58-69 Outline .....................................................169

Gautama 2.2.58-69 Translation ............. ................................173
Mimams ............................................. ........................... 95

3.1. Background .............................................................195
3.1.1. Dharma, Injunctions, and Means
of Knowledge ..............................................195
3.1.2. Vedic and Ordinary Words and Meanings.....196
3.1.3. Sentence Meaning .........................................197

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
TABLE OF CONTENTS v
3.2. The Eternal Nature of a Word

and their Relation .....................................................198
3.2.2. W ord .............................................................199
3.2.3. M eaning ......... .............................................203
3.2.3.1. Action ...............................................203
3.2.3.2. A Generic Term Must Denote
an Eternal Object ...............................204
3.3. The Nature of an Individual and of a Class

Property ......... ........ ............................................ 205
3.3.1. General remarks ............................................205
3.3.2. Individual ......................................................206
3.3.3. Class Property................................................209
3.3.4. Class Property: Abstract Generic Property
versus Shape ......................................... 12
3.3.5. The Bearing of the Nature of a Class Property
and an Individual on the Denotation
of a Generic Term ........................................223
3.4. Analysis of Arguments in the Akrtyadhikarana ........22

3.5. Criteria of Denotation ..............................................229
3.6. W ords Denoting Qualities ........................................25
3.7. Conclusions Drawn from Models in Mimamsa
and Nyaya ......... ................................................. 255
Akrtyadhikarana Outline .........................................................263

Jaim ini 1.1.5 Translation .........................................................269
Jaim ini 1.3.30-35 Translation ......... .....................................273
Lokavedadhikarana .........................................................273
Akrtyadhikarana..............................................................2
G eneral Conclusion ...........................................................289

Abbreviations ..........................................................................297
Bibliography ...........................................................................2
Primary Sources...............................................................299
Secondary Sources...........................................................303
Index.......................................................................................325

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
List of Tables
Table 1 ....................................................................................231
Table 2....................................................................................232
Table 3 ....................................................................................234
Table 4....................................................................................236
Table 5....................................................................................239
Table 6....................................................................................241
Table 7....................................................................................243
Table 8 ....................................................................................243
Table 9 ....................................................................................247
Table 10 ..................................................................................249
Table 11 ..................................................................................250
vii

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Preface
This work deals with a topic to which philosophy, most

notably analytic philosophy, has given considerable attention.
Indian thinkers discuss the denotation of generic terms in a very
sophisticated manner at a very early time. The present book seeks
to make these discussions available to philosophers today.
As a philosophy major in college, I once asked my
advisor about Indian philosophy. "You mean Indian religion," he
replied. Most European and American philosophers, especially
analytic philosophers, are under the impression that there is no
philosophy in the history of Indian thought. Nor can one expect
a disciplined thinker to form any other opinion if he judges from
a vague idea of the Orient, the popular interest in India, and the
general publications which respond to them. Moreover, the
barriers of time, distance, and language, which prevent Indian
philosophy from being known in the west, hinder the formation
of any other impression. I hope the present work will come into
the hands of philosophers and contribute to a better
understanding of ancient Indian thought.
I would especially like to give acknowledgment and
thanks to Dr. George Cardona for being so conscientious and
thorough a teacher and advisor, and for conveying his love of
knowledge along with clear and precise understanding. In my
three years of studying Paninian grammar at the University of
Pennsylvania, and in the last year writing my dissertation of
which this book is a revision, he has always inspired the highest
standards of knowledge and expression.
I also owe especially deep gratitude to Dr. Srinarayana
Misra for generously availing me of his unusual combination of
profound traditional learning, analytic ability, and skill as a
teacher. During the year and two months in which I studied a few
hours a day, six days a week with him in Varanasi, I made
incomparable strides in facility with Sanskrit, knowledge of
Indian semantics, and comprehension of the Mahabhasya and
Nyaya texts.
I have benefited much from Dr. Ludo Rocher's
excellence, generosity, and warmth as a teacher, and
Wilhelm Halbfass's insights and comprehensive know
Indian philosophy. I would like to thank them both
comments and suggestions in the preparation of this m
I am also indebted to Dr. Narendranatha Pandeya wit
studied the first three ahnikas of the Mahabhasya in Var
Finally, I am grateful to the American Institute o
studies for giving me a Junior Research Fellowship whic
me to carry out dissertation research, and to the Mellon
ix

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
x DENOTATION OF GENERIC TERMS
Foundation which gave me the opportunity for

trip to India and the freedom to write the origi
uninterrupted.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Introduction
By the late fifth century B.C.E. Panini had composed the

Astadhyayi, consisting of nearly 4,000 rules giving a precise and
fairly complete description of late Vedic Sanskrit.l In the fourth
or third century, Katyayana appended approximately 4,300 brief
statements (vdrttikas) to 1,245 of Panini's rules. Katyayana's
vdrttikas examine the formulations of Panini's rules, their relation
to other rules, suggest modifications, and also address the
fundamental principles presupposed. In the middle of the second
century B.C.E., Patanijali composed his monumental commentary,
the Mahabhasya, on Katyayana's varttikas and independently on
468 sutras of the Astadhyayi.2 In the course of this monumental
commentary, nestled in the midst of detailed grammatical
argumentation, Pataiijali examines various philosophical
presuppositions of the grammar and of ordinary language.
Panini formulates many rules and uses certain procedures
which make sense only if certain philosophical positions are
accepted. The ultimate criterion for the correctness of the
language description are the facts of ordinary usage. As is true in
any science, given the same accuracy in accounting for the facts,
the simpler the description of them, the better. The facts of
language usage are most simply explained by accepting certain
philosophical positions. At numerous points in the midst of a
technical grammatical discussion, the Mahabhasya turns to
consider certain philosophical presuppositions involved, and the
implications of adopting one or other of them for the grammar
and for other philosophical views. One such philosophical
discussion concerns what the object denoted by a word is.
A word has the capacity to generate knowledge of its
object for any competent speaker of the language to which the
word belongs. By virtue of this capacity the word is said to
denote the object, and the object is said to be the meaning of the
word.3 In everyday activity, a person recognizes individual
ISee Cardona (1976: 238).

2Kielhor 1886: 203. Cardona (1976: 260-68) summarizes the evidence for
dating Panini, Katyayana and Patafijali. The date of Patafijali's Mahabhasya is
reasonably accurate; those of Panini and Katyayana are determined relative to
Pataiijali. One should note that the evidence to establish the chronology of
ancient Indian authors is exceedingly scarce in comparison with what is
available for authors in other parts of the world. Hence most of the dates
mentioned are approximate and tentative.
3No distinction between denotation and connotation is made in this work. The
verb 'to denote' is used here in precisely the sense mentioned, namely, to
directly cause cognition of an object. Siderits (1991: 65-110) discusses the
sense-reference distinction in Prabhakara and in Buddhist philosophy of

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
2 DENOTATION OF GENERIC TERMS
objects to be of the same kind. For example, one recognizes the

object with roots, a trunk, branches, and leaves in one's own yard
to be of the same kind as the object with roots, a trunk, branches,
and leaves in the neighbor's yard. The recognition that they are
the same is a certain knowledge or cognition which has, or fo
which one may posit, an object. The object, real or posited, of
the cognition of individual objects as being of the same kind is
called a generic property.
In the ordinary use of language, one uses the same words
on different occasions to refer to different individual objects of
the same kind. One also uses plural forms to refer to many
individual objects of the same kind. For example, one uses the
word 'tree' for the object with roots, a trunk, branches, and leaves
in one's own yard, and one uses the word 'tree' for the object with
roots, a trunk, branches, and leaves in the neighbor's yard. One
also uses the word 'trees' for many such objects. This fact of
speech usage presents a question fundamental to the nature o
language and to the nature of human understanding: Does a
word denote an individual object or a generic property?
The oldest extant systematic consideration of the question
whether a word denotes an individual object or a generic property
in India is in Pataiijali's Mahabhasya on Panini 1.2.64, Sarupdndm
ekasesa ekavibhaktau. The Mahabhasya on 1.2.58,
Jdtydkhydydm ekasmin bahuvacanam anyatarasydm, selectio
from the Paspasahnika and elsewhere supplement this discuss
Patanijali is not, however, the first to address the topic. He him
discusses the question commenting on Katyayana's vdrttikas o
the two sutras mentioned. Katyayana in turn records the view
Vajapyayana (1.2.64, vdrttika 35) and Vyadi (1.2.64, vdrttika 4
and their reasons for them. However, their work has not
survived.4 Both Panini's sutras and Katyayana's vdrttikas are
exceedingly brief and can hardly be understood independent of
Pataiijali's explanations. Hence the Mahabhasya stands as the
most ancient source for the discussion concerning whether a word
denotes an individual or a generic property.5
language.
4Patafjali mentions the Samgraha in the Paspasahnika. Kaiyata says this is the
name of a specific work. Nagega attributes it to Vyadi and describes it as being
the length of 100,000 ilokas (anustubh verses). MB, vol. 1, pp. 54b-55a.
Vyadi, called ddksayana, seems to have been a relative of Panini, who is called
ddksiputra. See Goldstucker (1860: 86-90, 228-32). Nothing more is known of
Vajapyayana.
5Katyayana's vdrttikas are known only as incorporated in the text of the
Mahabhasya. It is not always clear whether a statement belongs to Katyayana or
is Patafijali's own. Concerning criteria for distinguishing the vdrttikas from
their commentary, see Cardona (1976: 247). Yaska mentions the classification
of words into nominals (including both nouns and adjectives), verbs, preverbs
and particles, and briefly discusses the semantic distinction between nominals

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
INTRODUCTION 3
Paninians throughout history base their philosophical

works on Pataijali. The philosophy of Bhartrhari (5th c. C.E.)
whose Vakyapadiya exerted a wide and lasting influence, derives
its substance from the semantic discussions in the Mahabhasya.
Unfortunately Bhartrhari's Mahabhasyadipika commentary on the
Mahabhasya exists only in a single fragmentary and corrupt
manuscript (dhnikas 1-7 with lacunae).6 However, Kaiyata wrote
the commentary Pradipa on the whole of the Mahabhasya in the
eleventh century C.E. incorporating much of Bhartrhari's work.
The more recent works on semantics of Kaundabhatta and Nagesa
are heavily indebted to Patafijali. In the eighteenth century
Nagesa wrote the most comprehensive commentary on the
Pradipa. Numerous other commentaries have been written on the
Pradipa, and on the Mahabhasya. Commentaries directly on the
Astadhyayi, such as Bhattojidiksita's Sabdakaustubha and the
Kasika of Vamana and Jayaditya, are heavily indebted to the
Mahabhasya.7
Patafijali's arguments and conclusions in the area of
semantics in general, and concerning whether a word denotes an
individual object or a generic property in particular, are not only
of central importance to the grammatical tradition. At least two
other major systems of philosophy recognize and react to them:
Nyaya and Purvamimamsa.
Gautama's Nyayasutras, codified perhaps in the second
century C.E., and Vatsyayana's commentary on them, written in
the early fifth century C.E., are the foundation of the Nyaya
system. The most important ancient commentary to follow is
Uddyotakara's Nyayavarttika written at the end of the sixth or
beginning of the seventh century C.E. Vacaspatimisra wrote his
Tatparyatika commentary on the Nyayavarttika early in the eighth
century.8
and verbs. Nirukta 1.1 - 1.2. Nir., part 2, pp. 5-6, part 3, pp.27-29. He also
presents two views concerning the related question of the derivation of nominal
forms. Nirukta 1.12 - 1.14. Nir., part 2, pp. 13-15, part 3, pp. 36-37. Sarup
summarizes the two views in his introduction. Nir., part 1, pp. 68-70.
6Abhyankar and Limaye (1963-69) produced a provisional edition. Palsule
(1985, 1988), Limaye, Palsule and Bhagavat (1985), Bhagavat and Bhate (1986,
1990), Bronkhorst (1987), Devasthali and Palsule (1989), and Palsule and
Bhagavat (1991) have reedited and translated it. Bergdahl (1987) translated
part of the Sadhanasamuddesa of the Vakyapadiya, Houben (1992b) studied the
Sambandhasamuddesa giving a translation of Helaraja's commentary on it, and
Herzberger (1986) studied Bhartrhari in relation to the Buddhist philosophers.
7Filliozat (1973) gives a survey of the commentaries on the Mahabhasya and on
the Pradipa with the known facts concerning their authors and dates. Filliozat
(1993) shows well the relation of the views of the commentators. Regarding
other commentaries on Panini's Astadhyayi see Cardona (1976: 278-86).
8Concerning the dates of early Nyaya see Jacobi (1911: 29), and for later dates
refer to Potter's (1970-) introduction to volume 2 on Nyaya-Vaisesika.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Initially concerned for the most part with means of corr

knowledge and proof, Nyaya grows into a complete system of
philosophy. But right from the beginning, it is concerned with
questions of language and meaning. Gautama treats the subject of
what a word denotes in Nyayasitra 2.2.58-69. At the start of the
section, Gautama 2.2.58, Te vibhaktyantdh padam, makes implicit
reference to Panini 1.4.14, Suptitiantam padam, and Gautama's
conclusion in 2.2.66, Vyaktyakrtijdtayas tu paddrthah, as
explained by Vatsyayana and Uddyotakara, bears some important
similarities to the view Patafijali expresses under 1.2.64 vdrttika
53 and 1.2.58 vdrttika 7. Both solutions to the question of what a
word denotes include all the entities considered in the denoted
meaning, and both involve the element of a speaker's intention
(vivaksd).
Growing out of a long tradition of Vedic exegesis and
performance, the Purvamimamsasutras were codified in about the
second century B.C.E., although they may have reached their
final form somewhat later. They are attributed to Jaimini but the
names of both him and Badarayana, to whom the
Uttaramimamsasitras are attributed, are mentioned in particular
sutras.9 In the fourth or fifth century C.E. Sabara composed his
Bhasya commentary on the Pfrvamimamsasiitras. This is the
most ancient commentary extant on them but Sabara mentions
predecessors and cites a long passage from one Vrttikara in his
commentary on 1.1.5. As would be expected for those concerned
with the analysis and interpretation of statements and injunctions,
the question of what a word denotes is a major concern, and the
ancient literature on the subject in Mimamsa is extensive. Sabara
is followed by Kumarilabhatta, Prabhakara, and Mandanamisra in
the seventh century. Kumarila has been the most influential of
the three but each of them had his distinct ideas and gave rise to
long and active independent traditions.
Jaimini and Sabara themselves deal most explicitly with
the question of what a word denotes in stitras 1.3.30-35 and the
Bhasya on them. They take a view corresponding to the view
which Katyayana attributes to Vajapyayana in Panini 1.2.64
varttika 35 and argues for in vdrttikas 35-44 and 53-59. Sabara
closely paraphrases passages from the Mahabhasya on 1.2.64 and
argues against the conclusion Pataiijali states under varttika 53.
It is impossible to examine all of the ancient Indian
literature on the question of the denotation of a word, even in the
three fields of grammar, Nyaya, and Purvamimamsa. However,
considering the close affinity of Vatsyayana and Uddyotakara's
explanations of Gautama 2.2.66 with Patafijali's statement on
1.2.64 varttika 53, considering the correspondence between
9See Nilakantha Sastri (1921).

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
INTRODUCTION 5
Jaimini's conclusion as explained by Sabara with Katyayana

presentation of Vajapyayana's view, considering the likelihoo
that Sabara cites a passage in the Mahabhasya and directly
responds to it, and finally, considering the possibility of other
direct influence and exchange among these authors, it is judicious
to examine Gautama 2.2.58-69 with Vatsyayana's Bhasya and
Jaimini 1.3.30-35 with Sabara's Bhasya together with the passages
concerning the denotation of a word in Patanijali's Mahabhasya.
Research Review
A considerable amount of scholarship has been do

Indian semantics. Brough (1951, 1953), Ruegg (1959),
Gaurinath Sastri (1959), Kunjunni Raja (1963), Subha Rao
(1969), and Staal (1976) give general overviews of the Indian
philosophy of language. Ruegg (pp. 121-25), Kunjunni Raja
(pp. 321-54), and Staal (pp. 132-36) include extensive
bibliographies of the research in this area. Bhattacharya (1962)
contains some valuable insights on the relationship of
grammatical views with those in Nyaya and Mimamsa.
More recently, books by Matilal (1990) and Siderits
(1991), and a number of essays in the volume edited by Matilal
and Chakrabarti (1994), introduce Indian philosophy of language
to scholars of analytic philosophy. Although some of his general
remarks characterizing Paninian grammar are misleading, for
instance his (pp. 8-9) reference to krt affixes as inflections, his
speculative etymology of the term pratyaya 'affixes' as from a
meaning "condition", and (p. 11) his unfounded assertion that
Patanijali gave no cogent reason for the study of the grammar of
one's first language, Matilal (1990) furnishes an engaging
exposition and critical analysis of classical Indian contributions to
the philosophy of language.
Siderits (1991) thoroughly examines two issues in Indian
semantics comparing them to, and framing them within, questions
current in contemporary philosophy of language. Considering
the problem of sentential unity, he (pp. 32-50) presents the
"related designation theory" (anvitdbhidhdna) of Prabhakara and
(pp. 51-64) compares it favorably with Frege's asymmetry thesis.
Then, taking up the problem of the sense-reference distinction in
Mimamsa and Buddhist philosophy of language, he (p. 76)
rightly recognizes that Indian philosophers of language
distinguish between type of entity denoted rather than between
mode of denotation, and keenly observes that "nominalist scruples
about allowing properties into one's ontology" motivated the
distinction between denotation and connotation. He (pp. 65-86)
argues that the Prabhakaras and (pp. 87-102) the Buddhist

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
logicians approach making the sense-reference distinction.

Although his discussion is clear and engaging, he compromises
precise, detailed description of the objects words denote and how
they are related for the sake of his intended philosophical
audience. As a result, he provides only vague description of the
exact concepts of semantic condition (svdrtha, pravrttinimitta) and
kdraka relations, and avoids discussing the denotation of nominal
terminations so essential for arriving at sentential meaning.
A number of articles in the volume of essays edited by
Matilal and Chakrabarti (1994) deal with speech as a means of
knowledge. Mohanty (pp. 29-49) raises doubts about the status
of speech as a means of knowledge. He examines the arguments
of the Vaisesikas and Buddhists that it is reducible to inference
and of the Buddhists that it does not generate true cognition.
Sibajiban Bhattacharyya (pp. 69-97) surveys a number of topics
in Nyaya concerning speech as a means of knowledge from the
point of view of current discussions in contemporary philosophy
of language. Chakrabarti (pp. 99-124), in an excellent
introduction to the issue, defends the independence of speech as a
means of knowledge by expressing his insights emerging from
primarily Navyanyaya texts. Fricker (pp. 125-61) contributes to
the view that knowledge from speech is inferential knowledge by
arguing against the thesis posited in Nyaya that a speaker may be
assumed to be trustworthy unless special circumstances defeat this
assumption. She argues against the premise of this thesis, that it is
impossible for a hearer to obtain independent confirmation of a
speaker's trustworthiness, by showing how empirical confirmation
of the trustworthiness of a particular speaker is possible. Matilal
(pp. 347-66), on the contrary, defends the Naiyayikas view that
understanding does not constitute an intermediate stage to
knowledge of an utterance. He affirms the view that speech, as a
reliable independent means of knowledge, generates knowledge
immediately unless blocked by contrary evidence over the view
that the content of a proposition, understood by speech, must be
separately justified in order to qualify as knowledge.
The volume includes two articles translated from Sanskrit.
Shukla (pp. 315-24) presents an argument based on Nyaya
against Russell's view that propositions are intentional entities.
Visvabandhu Bhattacarya (pp. 325-46) discusses various views
concerning whether and how an individual is qualified in the
cognition generated by a proper name. Sukharanjan Saha (pp.
367-84) discusses Gafigesa's views concerning the denotation of
cited speech forms.
Rocher (1975), Cardona (1976), and Scharfe (1977)
survey the research done on Sanskrit grammar. More recently,
Coward and Raja (1990), in the fifth volume of Potter's (1970-)
Encyclopedia of Indian Philosophies, summarize major works of

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
INTRODUCTION 7
Indian grammatical philosophy and (pp. 433-548) provide a

extensive bibliography.
In Volume 2 of this encyclopedia (1977), Potter
introduces the philosophy of Nyaya and Vaisesika, giving
summaries of numerous works. Bibliography can be found in
volume 1. Matilal (1977) and Verpoorten (1987) give historical
surveys of Nyaya-Vaisesika and Mimamsa literature respectively
with extensive bibliographies. Gaiganatha Jha (1942) and
Pashupatinath Sastri (1923) give general introductions to the
Mimamsa and its philosophy. For background on generic
properties in Indian philosophy consult Dravid (1972).
Translations
Danielsson (1883), Trapp (1933), Chatterji (1957),

Charudeva Shastri (1968), and Abhyankar and Sukla (1975)
translate initial portions of the Mahabhasya. Danielsson (1883)
translated the Paspasahnika into German. Trapp (1933) produced
a poorly received German translation of the first five dhnikas of
the Mahabhasya. Chatterji (1957) translated the Paspasahnika
with extensive notes. Charudeva Shastri (1968) translated the first
nine dhnikas into Hindi. Finally, Abhyankar and Sukla (1975)
produced a good translation of the first 3 dhnikas.
Paranjpe (1922) translated 273 of Katyayana's vdrttikas
into French. His translation and discussion of 1.2.58, vdrttikas 1-
7, and 1.2.64, vdrttikas 1-59 serve as a good introduction to
Pataiijali's commentary on them. S. D. Joshi (1968) and Joshi and
Roodbergen (1969-81) translate sections concerned with karakas
(1.4.23-55), most of the second adhydya (pddas 1-3), and (1986)
the Paspasahnika. The last contains sections directly concerned
with topics of this book (pp. iv [?2.1.3], viii-xii [?2.4-2.6], 12-25
[?4-15], 78-126 [?56-86]), while the others contain some related
philosophical material. Joshi and Roodbergen reformat
Kielhorn's edition of the text to facilitate comprehension of the
arguments, and supply extensive explanatory notes to accompany
their precise and finished translation.
Strauss (1927a), Subrahmanya Sastri (1944-57), Biardeau
(1964), and Filliozat (1975-86) translate sections of the
Mahabhasya centrally concerned with the question of what a
word denotes. Strauss translated the Mahabhasya on 1.2.64,
vdrttikas 35-59 into German (1927a: 256-71). He also translated
the portion of the Mahabhasya on 4.1.3 preceding the discussion
of the grammarians conception of gender and discusses
epistemological questions in comparison with Carakasamhita,
Saiikhyakarika, and Sabarabhasya on Jaimini 1.1.5 (1927b).
Subrahmanya Sastri (1944-57) translated the Mahabhasya up to

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Panini 2.4.85. However, many terms are left untranslated. He

supplies extensive commentary in the Paspasahnika but this tapers
off quickly and one is left with just a rough translation by the
second pdda. Biardeau (1964) includes translation (French) of
selections from the Paspasahnika and passages from the
Mahabhasya on Panini 1.2.58, 1.2.64, 4.1.63, and 5.1.119 (pp.
43-61). Filliozat (1975-86) translates the Mahabhasya with
Kaiyata's Pradipa and Nagesa's Uddyota on adhydya 1, pddas 1-3
into French. His translations are clear and accurate, and the notes
both supply useful details and explain the drift of the discussion.
A satisfactory English translation of the Mahabhasya discussions
concerning the denotation of a word does not exist.
The Nyaya and Mimamsa sections concerning the
denotation of a word have been translated. Spitzer (1927) studies
Gautama 2.2.58-69 and includes translation (German) of
Gautama's Nyayasutra and Vatsyayana's Bhasya. Ruben (1928)
translated Gautama's Nyayasiitra into German with brief
explanatory notes. Biardeau (1964: 230-34) includes translation
(French) of Vatsyayana's Bhasya on Gautama 2.2.60-69.
Gafganatha Jha (1912-19) translated Vatsyayana's Bhasya and
Uddyotakara's Varttika. This has recently been reprinted. But
Chattopadhyaya and Gangopadhyaya's (1967-73) translation of
the Sutra and Bhasya is far superior. Tiwari (1994: 145-68) re-
translates Gautama 2.2.58-69 with Vatsyayana's Bhasya in his
discussion of generic properties.
Yuddhisthira Mimamsaka (1977-86) translated Sabara's
Bhasya on the Mimamsasutra of Jaimini into Hindi with thorough
notes and an extensive introduction. Ganganatha Jha (1916)
translated and explained the first three adhydyas of Jaimini's
Mimmmsasutras. He later (1933) translated the whole of Sabara's
Bhasya. This is better than his translations of Kumarila's
Slokavarttika (1900-8) and Tantravarttika (1903-24). More
recently, Biardeau (1964) includes a good French translation of
relevant passages from Sabara's commentary on Jaimini 1.1.5 (pp.
162-63), Jaimini 1.1.19 (pp. 164-65) and the Akrtyadhikarana
(pp. 169-73). Devasthali (1959: 27-36) practically paraphrases
Sabara's Akrtyadhikarana giving the corresponding text in notes.
Studies
Jacobi (1929), Hiriyanna (1938), Frauwallner (1960

Narayana Murti (1971), and Matilal (1983) contribute to
study of the denotation of generic terms in the Mahab
Jacobi (1929) includes a brief discussion of the Mahabha
1.2.64 in comparison with Jaimini 1.3.30-35 (p. 149) and
considers Katyayana's vdrttikas 40, 48 and 56 (pp 150-51).

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
INTRODUCTION 9
Hiriyanna considers the implications of Vyadi's and

Vajapyayana's views of what a word denotes for sentence
meaning. Frauwallner (1960b) begins with a useful logical
outline of arguments (pp. 93-94) and a summary of the
Mahabhasya on 1.2.64, vdrttika 53 (p. 95). It may also be
mentioned here that Frauwallner (Nachgelassene Werke 3)
summarizes the Nyaya view of the denotation of generic terms (p.
97) and then proceeds to discuss the Mimamsa's theory of
language and its background. Narayana Murti (1971) evaluates
the grammatical status of Panini 1.2.64-73. In concisely stating
the various factors which impinge on the rule providing the
remainder of one (ekasesa), he brings attention to the related
issues with which Katyayana's and Pataiijali's discussion of the
denotation of generic terms is connected. Matilal (1983) brings
attention to the neglected discussion preceding varttika 35 under
1.2.64. He translates the bhdsya on vdrttikas 28-29 and raises
some probing questions. But after a digression into the theory of
naming he can do little more, in this short article. Mukherji
(1992) argues inadequately, without considering the Mahabhasya
on the sutra, that secondary reference (laksand) motivated the
formulation of 1.2.64. Deshpande (1972) studies Kaundabhatta
on the meaning of nominals. In addition to these studies, Van
Nooten (1983) collects some scattered passages concerning a
speaker's intention (vivaksd), including the bhasya on 1.2.58
vdrttika 7 and 1.2.64 vdrttika 53, to which Scharf (1995)
responds.
A few articles concerning semantics in Nyaya have
appeared recently. Vattanky (1993) gives a technical exposition
of the arguments of later Naiyayikas concerning the denotation of
generic terms in the MuktdvalT, Dinakarl and Rimarudrf
commentaries on the Kdrikdvall. Tarkatirtha (1992) provides a
clear independent exposition of a number of concepts and
examples found in Nyaya semantic discussions. Dash (1991)
briefly peruses some Nyaya and Vaisesika texts concerning the
relation between a word and its meaning. The relation is one of
convention (samaya, he translates 'agreement'). Houben (1992a)
and Matilal (1990: 26-30) clearly distinguish this from the sense
of samaya in Mimamsa and grammar, namely, uncreated
convention based on natural suitability.
Two major contributions to the study of Mimamsa are the
books by D'Sa and Clooney. D'Sa (1980: 83-97) properly
assesses the aim of Sabara's views on the denotation of generic
terms in his exposition of the assertion in Mimamsa that the Veda
serves as an independent means of knowledge. Clooney (1990:
77-79) touches on Mimamsasutras 1.1.6-23 dealing with the
eternal nature of speech in his important recent contribution to the
study of Mimam.sa. While he (p. 78 n. 53) defers to D'Sa's more
thorough treatment on this topic, yet he (p. 78 n. 54) plays down

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
both the claims of Mimamsa regarding the nature of speech an

their importance.
The most penetrating presentation of the content of
arguments in the Mahabhasya concerning the denotation of
word given so far is by Gaurinath Sastri (1959: 143-56).
However, he gives little historical or textual orientation to it.
Spitzer (1927) has more thoroughly examined Vatsyayana on the
denotation of words. A full exposition of the arguments
clarifying the logical presuppositions bound up with each view
Pataiijali discusses has not been done. A similar exposition of
Sabara's Akrtyadhikarana also remains a desideratum.
The most thorough study on the denotation of words in
Patanjali's Mahabhasya (pp. 31-64), Sabara's Mimamsabhasya and
Vatsyayana's Nyayabhasya (pp. 153-247) is by Biardeau (1964).
She brings together an enormous amount of material and gives an
engaging philosophical analysis of these ancient Indian theories
of meaning as part of the background to her study of Bhartrhari.
In contrast to Gaurinath Sastri, she is more concerned in this study
with the history of ideas than with the content of the arguments.
Unfortunately, her concern with evaluating the ontological status
of the entities discussed, the world-view of the authors, and the
implications their conclusions have for their world-view and
sociological circumstances, is sometimes at the expense of
understanding, clarifying, and evaluating the arguments the
authors make to establish their conclusions. It is not fruitful to
superimpose on ancient authors an interpretation of the nature of
entities, the evidence for which is altogether absent in their works,
in order to summarize their conception of the world as the
undeveloped beginnings of more advanced and discriminating
analyses of things. An example of this will be discussed shortly.
In studying the history of ideas, it should not be
overlooked that the ancient authors grapple with issues of current
philosophical interest in a rigorous manner. The arguments of
Patafijali, Vatsyayana, and Sabara seek to account for known facts
of the usage and comprehension of words in a systematic manner
consistent with other aspects of knowledge of the world.
Empirical facts of usage are brought in as evidence in the course
of an argument. As such these authors engage in scientific
activity not categorically different from science as it has
developed in recent centuries.10 Sufficient attention has not been
given to clarifying the content and analyzing the soundness of
their arguments, or to evaluating the relevance of the evidence
adduced. The current book focuses attention on these issues.
10For more on scientific methodology in ancient India see Staal (1988).

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
INTRODUCTION 11
The Meaning of the Term 'dkrti'
Although the current work is not intended primarily as

study of the meaning of the word 'dkrti', the subject will be d
with because it is central to Pataiijali's discussions of the
denotation of generic terms. Moreover, there has been some
disagreement concerning what Pataiijali means by this word.
Most of this disagreement concerns whether it means a concrete
shape or an abstract generic property. Associated with this is
whether it means a class property as opposed to a class.
Matilal (1983), although he does not decide the issue
himself, notes that the tradition held that the term 'dkrti' in the
bhdsya on 1.2.64 means jdti, that it means either a universal or a
form in contrast to a particular. Goldsticker (1860: 168-171,
with notes 180 and 182) maintained that the term 'dkrti' means
'species' in distinction to 'jdti' which means 'genus'. From Ojihara's
(1967) perspicacious analysis of Patafijali's definitions of a
generic property (see my ?1.2.5) it is sufficiently apparent that
Goldstiicker's interpretation was due to a misunderstanding.
Strauss (1927), while recognizing that it may mean either
form or generic property in the Paspasahnika (p. 126), takes the
term 'akrti'in the Mahabhasya on 1.2.64 to be synonymous with
the term 'jdti' (pp. 135-150) as does Jacobi (1929). Renou (1942)
also recognized the use of the terms 'akorti' and 'dravya' in
different senses. These senses include 'forme' and 'genre' for the
former (p. 78) and 'substance' and 'objet' for the latter (p. 162).
Specifically he noted that the term 'dravya' designates an 'objet
individuel' or simply 'individu' in contrast to the term 'akrti' which
designates an 'objet generique' or 'genre', notably in the
exposition in the Mahabhasya under 1.2.64 concerning the
question of which of the two a word denotes (p. 163).11 If the
term 'dkrti' has the sense 'genre', it is synonymous with the term
'jati (espece, genre)' (p. 148).
Sreekrishna Sarma (1954, 1957) surveys the meanings of
the terms 'akrti' and jdti' in different contexts. He concludes that
an akrti is a universal serving as a class property, and that it
should be distinguished from a jdti which is the class and an
abstract entity. D'Sa (1980: 86) repeats Sreekrishna Sarma's view.
Although Joshi and Roodbergen (1986: xi) make concessions to
Sreekrishna Sarma's view, stating, "akrti is interpreted to mean
'(common) shape', 'a generic feature'," and (p. 78), "akrti
originally stands for a perceptible shape or form common to
members of a given class or genus," they (p. 79) recognize the
I ld[ravya] en tant qu' < objet individuel > s'oppose aussi a akrti (q. v.) < objet
generique >, not. dans l'expos6 M[ahabhasya] I 2 64 vt. 35 sqq. concernant la
question si le mot d6signe l'individu (opinion de Vyadi) ou s'il d6signe le
genre (Vajapyayana).

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
double sense of the term in agreement with Strauss and Renou.

They write, "Rather early, however, the terms dkrti and jdti cam
to be regarded as synonyms in the sense of jdti 'class'. This use i
already found in the Mbh.... On the other hand, the same word
akrti is also used in the sense of 'shape'...."
Frauwallner (1960: 95-96) understands an dkrti to be an
eternal form ('ewige Form'). In Sprachtheorie (Nachgelassene
Werke: 116-18) he observes that there is a development of th
concepts of an Jkrti, a jdti, and a dravya. Originally dkrti is seen
as the form of an object. As a distinguishable property of many
individual objects, one identifies it as that by virtue of which an
individual object is a member of a class (jdti). He says that under
the influence of category-theory (Vaisesika) there is a tendency t
identify form and class, or dkrti and jdti.
Scharfe (1961: 129-42) considers the question of the
meaning of the term 'akrti' in Pataiijali's Mahabhasya in detail. He
recognizes that it has a range of meanings. But after translating
and considering numerous passages, he concludes that the
translation 'form' is possible everywhere. He says that the form is
often that which is the common factor in various material things,
and concludes that we should stick with that as a translation of
'dkrti' in distinction to 'dravya' (p. 133). Halbfass's (1992: 91)
translation 'form' (in quotes) for Patanijali's term 'dkrti', eve
where he recognizes that it refers to a "permanent generic"
inhering in each individual object of a class, seems to persist in
this opinion.
Biardeau (1964), in agreement with Sarma, asserts that for
Pataiijali and Sabara the term 'jdti' means 'class' and the term 'dkrti'
means a concrete structure which serves as a class property.
Deshpande (1972: 26-40), after considering the opinions of the
Indian commentators and modern scholars, concurs with Biardeau
and Sarma.
Filliozat (1975-86) appropriately translates 'dravya' as
'individu' in the discussions devoted to the denotation of generic
terms under 1.2.58 and 1.2.64. He translates the term 'akrti' as
'forme' and 'forme generique', and notes that it is the class
property in distinction to the class itself and has this same sense
throughout the Mahabhasya.12 However, he makes one
exception to translating 'akrti' as 'forme generique' translating it as
'universel' instead (vol. 4, p. 354). He also uses the word
'universel' when he supplies the rest of a sentence (p. 365), and
12"Ak.rti, couramment traduit par 'forme', semble etre employ6 plut6t pour
designer le manifestant de I'espece que l'espece meme. C'est dans ce sens qu'il
est pris dans le sutra de Gautama cit6 dessous (voir 2.2.68). On le trouve dans ce
mmem sens dans d'autres passages du bhasya." Filliozat (1975-86: vol 1, p. 21
n. 1). See also p. 78 n. 1, "Nous avons vu ci-dessus que dkrti d6signe la forme a
quoi on reconnait l'espece....

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
INTRODUCTION 13
frequently discusses what 'akrti' and 'dravya' refer to usin

terms 'universel' and 'individu' (vol. 4, p. 343, notes 4-6; p
notes 1-4; 347, notes 1, 2; p. 354, notes 1, 2; p. 358, note 1
Although it is true that the term 'Jkrti' means both
property and a shape, the first meaning is found with
second. It is not proper to tie Patafijali's concept of a
property to a concrete shape. While Scharfe concludes t
should stick with 'form' consistently as a translation of 'd
everywhere, he explicitly recognizes the double meaning o
term 'dravya', used in opposition to it, as both substan
individual object (p. 138). Opposed to Goldsticker, he
recognizes (p. 141) that dkrti is different from jdti in the
the two verses Pataiijali cites under Panini 4.1.3. Howev
does not observe that immediately following the verses Pa
uses the term 'akrti' to refer to the entity to which the ter
refers in the satra and from which the verse distinguishes a
(see ?1.2.5 below). This oversight prevents him from seein
in fact there are two distinct meanings of the term 'akrt
synonymous with the term 'jdti (generic property)' and a
distinguished from it, 'configuration' or 'shape', as in G
2.2.68.
Accepting the interpretation that 'akrti' means a shape,
Biardeau (1964) concludes that Patafijali and Sabara are still
bound to considering concrete objects of everyday use and are
unable to conceive of an abstract generic property as she say
Vatsyayana and even Gautama do. She sums up her discussion of
the passages in the Mahabhasya saying that it is clear that an dkrt
constitutes the permanent perceptual element which permits fixed
speech to apply to transient things. Then she adds that the Vedi
injunction will therefore denote concrete things by the
intermediary of their specific form (p. 61).13
In addition, she draws a distinction between a class and a
class property. She asserts, without adducing any evidence, that
for Patanjali and Sabara the term jdti' means 'class' and the term
'dkrti' means a concrete structure which, as a common trait of
many objects, serves as a class property to identify an object as
one of a class. Because this latter distinction conforms neatly to
the modern concepts of class and class property, it merits close
consideration.
Biardeau refers to Nagesa's comment that for Pataiijali
adkrti is "le genre et le structure" (p. 33).14 She also refers (p. 32,
note 2) to Renou's explanation, based on the commentatorial
13Mais il est clair qu'elle constitue 1'e61ment perceptif permanent qui permet a
un langage fixe de s'appliquer aux choses passageres.... L'injonction v6dique
denotera donc les choses concretes par l'interm6diare de leur forme specifique.
14With regard to the bhasya passage, "Nety dha dkrtir ndma sd," Nagesa: dkrtir
jdtih samsthdnam ca. MB, vol. 1, p. 15b.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
tradition, (1942) giving diverse senses of the term. But she

chooses to ignore the tradition and Renou. She states that
Patanijali uses the term jdti' to mean solely a species or a class of
distinct entities, and the term 'dkrti' to mean solely the
composition of parts, structure, or (specific) form which serves to
make known that an individual belongs to a certain class (p.
48).15 She says that an akrti is the common trait that determines
an individual's class (jdti) (p. 49), and that it is an immediately
visible trait.
However, she retreats from this position in particular
circumstances. Discussing 1.2.64 vdrttika 40, she accepts that an
dkrti is not necessarily visible. She says that an dkrti is the
sdmdnya that is a property which determines an individual entity
as belonging to a certain class (jdti) (p. 52). She notes that not all
samanyas are such properties and not all determine a jdti. A
sadmnya is a dynamic property... expressible by a proposition or
a sensible quality (p. 52, note 3). Discussing 1.2.64 vdrttika 43,
Tatrdsarvadravyagatih, she is forced to admit that the proponent
of the view that an individual substance is denoted considers the
dkrti to be identical to the jdti. He considers it to be the entire
group of individuals of a kind because the injunction to bind a
cow would apply to all cows if it applied to the dkrti. But rather
than depart from her view that a jdti is a class and the dkrti is the
property which identifies an individual as belonging to it, she
concludes from this that just the proponent of the view that an
individual substance is denoted uses the word 'dkrti' to speak of
the class, while the proponent of the view that an dkrti is denoted
uses it to speak only of the class property.
It is true that the two discussants disagree over the nature
of an dkrti, but this disagreement hardly furnishes evidence for
the conclusion that Pataiijali considers an akrti to be a class
property while he considers a jdti to be a class. This would make
a neat and useful distinction if it were the case that Pataiijali
adhered to it. But the fact that the proponent of the view that an
individual substance is denoted uses the term 'dkrti' for the class
here is just one example in which that distinction is not adhered
to. I do not believe Pataiijali makes such a distinction. On the
other hand, I will show that under 4.1.63, where no proponent of
the view that an individual substance is denoted is involved,
Pataijali again uses the term 'akrti' in the sense of a class (jdti) in
distinction to a second meaning of the term 'dk.rti' meaning shape
(see my ?1.2.5).
15Introducing the first verse characterizing a jati under 4.1.63: Cependant, cette
composition de parties repr6sente ce qui fait reconnaitre I'appartenance d'un
individu a une espece ou a une classe d'etres distincts (jdti). Car c'est la
rigoureusement le seul sens possible du terme jdti dans le Mahdbhdsya, tout
comme dkrti n'est jamais que la structure ou la forme (sp6cifique).

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
INTRODUCTION 15
The conclusion that an akrti is always a concrete shape is

due in part to a small but critical error in translating Patanijali on
5.1.119, Tasya bhdvas tvatalau. Biardeau translates 'vartin' as
'subsisting element' taking it to refer to the dravya. She
erroneously connects the passage in which this term occurs with
the previous passage in which Patanijali states that a substance is a
stable element in relation to transient qualities (see my ?1.2.1)
She concludes that words ending in the abstract suffixes 'tva' and
'tal' would denote the substance! She says, "un suffix abstrait
meme s'il est attache a un adjectif, ne signifie pas le genre d'une
qualite.... I1 se rapporte en realite a la substance qui forme le
support de la qualite.... (p. 59)" She then generalizes this to deny
that Pataiijali considers an dkrti to be a condition for the
occurrence of such a suffix (bhdva), "C'est la recherche meme de
l'element stable qui trahit d'ailleurs qu'il s'agit encore d'un aspect
de l'universel, mais Pataiijali ne met le bhdva en rapport ni avec
l'dkrti, 'forme specifique', ni avec la jdti, 'classe', ni avec le
sdmdnya, 'trait commun' (p. 59)."
The general conclusion that an dkrti is a concrete structure
is drawn from the impression that Patafijali says that a substance,
rather than an dkrti inhering in that substance, is the condition for
the occurrence of an abstract suffix. This impression, in turn, is
based on a small but critical mistake in translation. Hence the
general conclusion is undermined by correcting the mistake.
The term 'vartin' does not refer to the substratum of a
property here; it refers to the condition for the usage of th
suffixes tva and tal.16 The present discussion concerns Patanijali'
commentary on Katyayana's vdrttika 5, Yasya gunasya bhdvdd
dravye sabdanivesas tadabhidhdne tvatalau. According to this
vdrttika, 'tva' and 'tal' occur to denote a guna (property). Th
word derived by providing one of these suffixes after a quality
word denotes the quality (tadabhidhdne tvatalau) due to the
presence of which (yasya gunasya bhdvdt) in a substance, the
quality word denotes the substance (dravye sabdanivesah).
Similarly, the word derived by providing one of these suffixes
after a generic term denotes the property (guna) due to the
presence of which in a substance, the generic term denotes the
substance. Hence 'suklatva' denotes the quality white, and
'v.rksatva' denotes the generic property by virtue of having which
a tree is a tree.
Patafijali uses such a term 'vrihitva' under 1.2.58,
Jdtyakhydyam ekasmin bahuvacanam anyatarasydm, as an
example of the single sdmdnya denoted by a generic term on the
view under which 1.2.58 is stated. The sutra itself, by assuming
16See Kaiyata, "Vartyabhdvad iti. Vartanam vartah. Vrttir ity arthah. Varto
'sminnastlti varti pratyayarthah..." and, most explicitly, Nagesa, "...guniparesu
suklddisu guna eva vartl." MB, vol. 4, p. 300b.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
that if a word refers to a jdti it refers to a single entity (ekasmin)

under the most straightforward interpretation, identifies the jdt
with that single entity. Similarly, Katyayana's first varttika,
Jdtydkhydydm sdmdnydbhidhdndd aikdrthyam, refers to that same
single entity using the term 'sdmdnya'. Under vdrttika 7, Patafijal
refers to the single entity denoted by a generic term using the
term 'jdti'. Now in the Paspasahnika Patafijali states that 1.2.58 is
stated considering that a word denotes an dkrti.
Hence, Patafijali himself identifies an dkrti as a jdti and as
a samanya which is in turn what Panini 1.2.58 refers to with the
term 'jdti'. Patanijali uses terms ending in the suffix 'tva' to denot
examples of such entities. On the view adopted under 5.1.119
varttika 5, such entities are properties (guna) of substances
(dravya), and these properties are considered to be the essenc
(bhdva) of an object, namely, that by virtue of having which it is
what it is. Given the synonymous use of the terms dkrti and jdti in
the sense of class under 1.2.64 and 4.1.63 (just mentioned), and
the synonymous use of the two terms I point out in chapter one
(?1.2.7), what evidence is there for the distinction that 'dkrti'
means class and jati' means class property? Given the fact that
these two terms are used for a property which is considered to be
the essence of an individual substance, how can anyone insist that
it is always a concrete shape?
Moreover, the phrase, "That which is known by means of
a shape (dkrti) is a generic property (jati) (Ak.rtigrahand jdtih)," is
so close to Gautama 2.2.68, "A shape is that which makes known
the generic property and its indicatory marks (Akrtir
jdtiliigdkhyd)," that it is more likely that the first verse Patafijali
cites under 4.1.63 and Gautama 2.2.68 refer to the same idea than
that Pataiijali cut a distinction between a jdti and an dkrti in a
manner as different from Gautama's as Biardeau suggests.
Biardeau's statement that an dkrti is the general property
which identifies an individual as belonging to a certain class, is all
right. But her insistence on characterizing it as a visible form, a
structure, rather than as a general property pure and simple,
whether concrete or abstract, is unjustified, and so is her rigid
distinction of an dkrti from a class (which she reserves as the sense
of the term jadti). The discussion on whether an dkrti is eternal
(nitya) or transient (kdrya) in the Paspasahnika may be interpreted
as involving only a concrete form; however, it may also be
interpreted as involving distinct alternatives: an individual
transient concrete form versus an abstract general property.
Therefore, having considered the views of various
scholars concerning the meaning of the term 'dkrti' in the
Mahabhasya, I agree with Strauss, Jacobi and Renou. In the
discussions concerning the denotation of generic terms under
1.2.64, 1.2.58 and in other places, the term 'dkrti' means a class

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
INTRODUCTION 17
property which is clearly distinguished from a shape and as such

is synonymous with the term 'jdti (generic property)'.
As the evidence in chapter one (?1.2) will show, Patafijali
uses the term 'akrti' in at least two distinct senses: 1) the shape
which a material has, and 2) the class property which identifies
what sort of thing an individual object is. In the second sense, it
is synonymous with the term jdti', both of which terms are use
by Patafijali to refer to both the class and the class property. It is
quite clear that a shape may serve as a class property too in the
case of certain classes of material objects, and that a visible shape
makes a good example of a class property. However, this does
not imply that an author who uses the term 'akrti' in both senses i
limited to conceive of a class property as an entity with concret
structure perceptible by the senses.17
Much attention has been paid to the question what
Patafjali, Sabara, and Vatsyayana considered the terms 'akrti' and
'dravya' to denote without explicitly taking up the question of the
nature of the entities. The two questions are intimately connected
In contrast, the recent work of Halbfass (1980, 1992), Matilal
(1973), and Wezler (1985) concerning the nature of substances,
qualities, and universals is exemplary. Bhaduri's (1947) sweeping
yet penetrating presentation of the nature of Nyaya-Vaisesika
entities remains one of the clearest expositions, particularly (pp.
22-51) concerning substance.
There seem to be two traditions of thought concerning
generic properties and qualities which are perceptible by the
senses. According to one, a shape is clearly distinguished from a
generic property. It is the characteristic outward mark of an
object which serves as grounds for the inference that the generic
property is present in a certain object. In this tradition of thought,
Vaisesika clarifies the distinction between a quality (such as a
color) and a generic property (such as treeness) by adopting the
idea that there are individual quality instances. An individual
quality instance is an entity distinct from both the substance in
which it inheres, and from the generic property inhering in
quality instances of the same kind. On the other view, such a
distinction is absent. A single quality recurs in many individual
substances in a manner quite similar to a generic property.
Perceptible qualities such as color and shape are considered to be
general properties much as abstract kinds or types are.18
The former tradition of thought prevails in Nyaya; the
latter in Mimamsa. Both traditions find their place in the
Mahabhasya. It is well known that the discourse in the
17Scharf (1993) collects in a single presentation all the evidence for the
conclusion that Pataijali uses the term 'akrti' to denote a class property.
18Concerning the conception of qualities as individual instances versus their
conception as universals, see Halbfass (1980, 1992: 113-37), Siderits (1982).

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Mahabhasya makes sudden shifts in point of view. In each new

point of view Patafijali takes up in philosophical discussions, he
may represent a different tradition of though prevalent at his time.
For example, in the Paspasahnika when he discusses whether
form is permanent or transient, he states three points of view: 1
It is absolute and immutable (kuitasthanitya). 2) It is permanent in
the sense that it is recognized as the same even with changed
particulars (pravdhanitya). 3) It is transient. It is clear in this
discussion that it is not a single conception of an dkrti that he
means as having these mutually contradictory characteristics. The
same word is used in each case for a distinctly conceived entity.
Therefore, Patafijali uses the term 'akrti' in different meanings. It
is more useful to identify the sense in which a term is used in
particular cases (as Sreekrishna Sarma does in the body of his
work) than to stretch a single term to cover different meanings
and blur the sense of the term in the discussion at hand.
Unfortunately, it is the latter which most of those wh
studied the meaning of the term 'dkrti' in the Mahabh
including Sreekrishna Sarma in his conclusion, attempt to d
Unlike the case in the Mahabhasya, there is a gen
difficulty in Sabara's argument concerning the denota
generic terms which leads one to question whether he
concept of a generic property distinct from a shape. This
difficulty concerns his example of the injunction to build a
falcon-altar (syenacitam cinvita). Biardeau (1964: 177)
recognizes that this weighs in favor of concluding that an dkrti
bound exclusively to sense perception. Chapter three of the
present work discusses the example of the falcon-altar and
Sabara's arguments based on it at length.19
Overview of the Work
The current work consists of three chapters dealing wi

the arguments concerning the denotation of a word in Pataiij
Mahabhasya on Panini 1.2.64 and 1.2.58, Vatsyayana's
Nyayabhasya on Gautama 2.2.58-69, and Sabara's Bhasya on the
Akrtyadhikarana respectively. Each chapter consists of two major
parts, the first being an analysis and the second a translation. In
addition, an outline precedes the major selections translated. The
notes to these selections include translations of passages from the
commentaries on these texts.
The first chapter analyzes the arguments for two extreme
views of what a word denotes: the view that a word denotes just
an individual object, and the view that it denotes just a generic
19Scharf (1993b) briefly analyzes the shortcomings of the example.

111.93.136.226 on Thu, 05 May 2022 05:09:011976 12:34:56 UTC
INTRODUCTION 19
property. This chapter probes the implications of adopting th

views for accepting other presuppositions such as those
concerning the analysis of the meaning of speech forms by
concomitant presence and absence (anvaya and vyatireka), and
the denotation of one and only one object by a word. The first
chapter concludes with an examination of Patafijali's solution to
include both an individual object and a generic property in the
denoted meaning, with one or the other having greater
importance depending on the speaker's intention.
The second chapter examines the arguments of Gautama,
as explained by Vatsyayana, to include three elements in the
meaning denoted by a word, rather than just two as Pataiijali does.
This view includes a generic property, the configuration which
serves as its inferential mark, and an individual object. The
second chapter explores the relationship between the generic
property and its inferential mark with regard to the use of words
for imitations of the objects they ordinarily denote.
The third chapter examines Sabara's arguments in favor of
one of the extreme views considered in the Mahabhasya. It
undertakes to clarify Sabara's argument and the role certain
fundamental principles of epistemology play in it. It explores the
relation that including the means of knowledge known as
presumption (arthapatti) in the procedure to determine a word's
meaning has to the argument against including more than one
element in the denoted meaning. Because these arguments are
sophisticated and subtle, they will be clearly stated in terms of
symbolic logic as well as in English. At the end of chapter three,
the role that examples of the use of generic terms to refer to
imitations play in Sabara's arguments are compared with the role
of similar examples in Nyaya. Finally, the general conclusion
summarizes the relation the different views of what a word
denotes have to the presuppositions concerning the procedur
determine a word's meaning.
The following analysis frequently refers to the translat
and often to other sections of the book. These references are
included in parentheses in the text. If they are preceded b
helix, they refer to a section number as listed in the tabl
contents. Otherwise they refer to the translation accompanyi
the chapter in which they appear.
In the translation of the Mahabhasya on 1.2.64 and
1.2.58, the vdrttikas themselves provide an outline of the subject
matter. Because of the length of its commentary, a brief overview
of the topics in 1.2.64 is given there in outline form. The vdrttika
numbers as in Kielhorn's edition constitute an easy means of
reference, with the following supplements in square brackets. 40a
labels the vdrttika segment separated from its later part 40b,
numbered 40 in Kielhorn's edition. 53a, b, and c label segments
of the long commentary under vdrttika 53.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
A full logical outline precedes the translations of Gautama

2.2.58-69 with Vatsyayana's Nyayabhasya and selections from
Uddyotakara's Nyayavarttika. References to the translation are by
sutra number followed by 'V' for Vatsyayana or 'U' for
Uddyotakara and occasionally a number corresponding to the
lower levels of the outline. References are made in a similar
manner to the minor Mimamsa sections translated. Because of the
length and detail of Sabara's bhdsya on the Akrtyadhikarana, the
outline numbering itself serves as a more precise means to refer to
the text. Hence references are made simultaneously to the outline
and the translation which has been marked with the outline
numbering.
A few conventions used in this book should be
mentioned. When discussing grammatical elements such a
roots and affixes, the markers (it) are set in bold. All
terms and text are set in italics, except the names of sch
which there is no English equivalent, and the titles of wo
chapters. These appear capitalized in roman script. W
exception of Paninian grammatical elements always referr
speech forms, a Sanskrit term referred to as a term is en
single quotes in addition to italics.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Grammar
1.1. Background
Grammar (vydkarana) is the branch of knowledge

associated with the Veda (vedntiga) which instructs the use of
correct speech forms (sabda). Various entities are parts of the
single whole which constitutes an experience. Patafijali singles
out a speech form from among those entities as the sound by
means of which a knower knows an object of experience. It is
that which when uttered is the means to cognition of an object.'
It is distinguished from the objects of which it communicates
knowledge, namely, substances, actions, qualities, and class
properties.
Taking the experience of a cow as an example, Pataiijali
describes four types of objects from which he distinguishes
speech sounds. He calls the object consisting of a dewlap, tail,
hump, hooves, and horns a substance (dravya). He identifies
gesture, movement, and blinking as actions (kriya). He identifies
various colors (white, blue, black, brown, and grey) as qualities
(guna). He calls the general property which is the same in
different things and which is not destroyed when the things in
which it resides are destroyed a class property (dkrti).2
There is an established relationship between words and
their objects, which is known from ordinary usage, such that
certain words are used to denote certain objects. One uses speech
forms in order to convey knowledge of objects.3 When one may
convey meaning equally by the use of correct or incorrect speech
forms, grammar teaches a restriction for the sake of dharma
similar to ordinary and Vedic restrictions for dharma. By
teaching certain speech forms grammar makes known that these
speech forms are correct, they should be used, and by using them
l Yenoccdritena sdsndldntgulakakudakhuravisdnindm sampratyayo bhavati sa

sabdah. MB-K, vol. 1, p. 1, lines 10-11. Filliozat (1991: 33-85) has recently
treated this passage with its commentaries in depth in a lecture entitled,
"Patanijali's Definition of Sabda." Bronkhorst (1987: 46-56) examines
Patanijali's ideas concerning the wholeness of a speech form and (pp. 56-71)
explores its relation to ideas in the early history of Indian philosophy,
specifically in Abhidharma Buddhism. Concerning Patanijali's alternative
explanation of sabda, i.e. pratitapaddrthakah, etc., Wezler (1994) establishes
finally that this means (p. 185), "The word 'sabda' is one the meaning of which
is well-known," namely, (p. 188) sound, including noise as well as language.
2MB-K, vol. 1, p. 1, lines 6-10.
3See 1.2.64 translation, vdrttika 29 note 15.
21

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
one will accrue merit. Other speech forms are incorrect, shoul
not be used and may lead to ill results. Katyayana says this in
very first varttika:
Siddhe sabddrthasambandhe lokato 'rthaprayukte sabdaprayog

sastrena dharmaniyamah, yathd laukikavaidikesu.4
Since speech, its object, and the relationship between the two ar
established (and are known) from ordinary usage, and since one us
speech prompted by meanings in accordance with ordinary usage, th
science (of grammar) restricts (usage to correct speech forms) for the
sake of dharma just as in ordinary and Vedic (instances).
Words are divided into four categories: nominals, verbs,

preverbs, and particles.5 Nominals are again divided into fo
categories according to the type of object which is the semant
condition for their usage (pravrttinimitta): generic terms
(jdtisabda), quality words (gunasabda), action words (kriyasabda),
and arbitrary names (yadrcchasabda).6
4MB-K, vol. 1, pp. 6, 8. Cardona (1988: 633-34) paraphrases the varttika. For
other English translations see Abhyankar and Shukla, pp. 24, 30, 31, Chatterji,
pp. 57, 58, and Joshi and Roodbergen 1986: 90,115,117, 120). In disagreement
with the last, Wezler (1994: 173-74 n. 2) reaffirms that this is the first varttika.
5Yaska says, "Tad ydni catvdri padajdtdni ndmdkhydte copasarganipdtdsi ca
tdnimdni bhavanti. Nirukta, part 3, p. 27. Patanijali says, "catvdri padajdtdni
ndmdkhydtopasarganipdtdi ca." MB-K, vol. 1, p. 3, lines 17, 26.
6Patafijali on Sivasutra 2, vdrttika 1, Lkdropadeso
yadrcchdiaktijdnukaranaplutyddyarthah, says: Catustayi sabddndm
jatisabdd gunasabddh kriydsabdd yadrddhadabdds caturthdh. MB-
19, lines 20-21. There are only three categories of nominals on the v
names are arbitrary. On this view, the semantic condition for th
ultimately a generic property, quality, or action, so they are classed
the first three types. Trayi ca sabddndm pravrttih. d.tisabdd
kriydsabdd iti. Na santi yadrcchsdabddh. MB-K, vol. 1, p. 20, line
are no underived nominals on this view. Under 7.1.2, Ayaneyiniyiyah
phadhakhacchaghdm pratyayddindm, Katyayana and Patanijali conclude that
1.2.45, Arthavad adhdtur apratyayah prdtipadikam, shows that Panini
considered there to be underived nominal bases. The nominal bases whose
derivation is attempted using Unadi affixes comprise these underived b
Pratipadika-vijanadc ca Panineh siddham. Prdtipadikavijiidnc ca
bhagavatah Pdniner dcdryasya siddham. Unddayo 'vyutpanndni prdtipadikani.
MB-K, vol. 3, p. 241, lines 19-21. Both views are represented under 7.3.50,
Thasyekah, "Unddayo 'vyutpanndni prdtipadikdni" and "Athdpy unadayo
vyutpddyanta evam api na dosah." MB-K, vol. 3, p328, lines 20, p. 328 line 24 -
p. 329, line 29, p. 329, lines 4-5. See also 8.3.59, Adesapratyayayoh, and
bhasya. MB-K, vol. 3, p. 439, lines 10-16.
According to this scheme of naming types of words, the semantic condition for
using a generic term for an object is that a certain generic property is present in
the object. The condition for using a quality word for an object is that a certain
quality is present in the object. The condition for using an action word for an
object is that the object have a certain relation to a specific action. Although
the name 'generic term' indicates that a generic property is the condition for
usage of the word, it need not denote a generic property. It may have received
its name because it denotes a substance on the condition that it has a certain

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 23
The first type of word, a generic term, is a common noun

such as 'tree' or 'cow'. Such a term is used to refer either to a
particular individual object of a kind or to all individuals of the
kind. Hence the question arises whether a generic term denotes
the class property or an individual substance of the class.
preparation to investigate Patafijali's view concerning whether
generic term denotes a substance or a class property, we will fir
clarify what he considers the nature of a substance (dravya) and
the nature of a class property (dkrti) to be in this discussion.7
1.2. Substance (dravya) and Class Property

(adkrti)
Pataijali uses each of the terms 'dravya' and 'dkrti' in at

least two senses. He uses the term 'dravya' to mean 1) the material
out of which an object is made, and 2) an individual object. He
uses the term 'Jkrti' to mean 1) the shape which a material has,
and 2) the class property which identifies what sort of thing an
individual object is. Although he uses each of these terms in two
senses, he shows that he is aware of a distinction between the two
meanings. In his discussions concerning the denotation of
generic terms he intends each of the terms in their second
meanings: 'dkrti' means a class property and 'dravya' means an
individual object.
1.2.1. Panini 5.1.119
Under 5.1.119 Pataijali juxtaposes these two co

of a dravya in the course of explaining how a dravya is
something other than a particular quality present in it.8 First he
gives the following three examples which characterize a dravya as
the material out of which an object is made: 1) Given the same
volume of cotton and iron, that due to which their weight is
different is dravya. 2) Given the same sharp edge of two things,
that due to which one cuts by just touching and the other does
generic property. In any case, Pataijali's discussion does not concern the name,
nor does it assume that a generic property even exists. I mention the categories
just to identify the words whose denotation is being discussed.
7Concerning the nature of a guna, see MB-K vol. 2, p. 216, line 19 - p. 217, line
5. Halbfass (1980, 1992: 113-37), Matilal (1973), and Subramania Iyer (1942)
discuss the nature of a guna, the latter two in grammar itself. Rocher (1966)
deals with kriya in the Mahabhasya, and Subramania Iyer (1950-51) discusses
the concept more generally among grammarians.
8See the discussion beginning, "Kim punar anyac chabdddibhyo dravyam
dhosvid ananyat?" under 5.1.119, vdrttika 5. MB-K, vol. 2, p. 366, lines 14-26.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
not, even when applied for a long time, is dravya. 3) That due
which one thing causes a break with one blow and another does
not, even with two blows, is dravya.9 The mention of cotton a
iron make it clear that the material is what is considered to be
dravya.
Alternatively Patanijali characterizes a dravya as an
individual whole consisting of a certain set of properties.
Although a whole is made up of its parts, it is not identical to any
one of them. Hence characterizing a dravya as a whole whose
parts are its properties is in accord with the purpose of the section,
namely, to explain how a dravya is something other than one of
its properties. Although a dravya consists of its properties, it is
not identical to any one of them. It is that which remains what it
is even when certain of its qualities disappear and others appear.
For example, the qualities, red, yellow, etc. appear in the fruits,
amalaka, badara, etc. while they ripen, but the dmalaka remains
what it is and so does the badara. A particular amalaka remains
that particular dmalaka while it gets ripe.'0
Patanijali adduces an etymology of the word 'dravya' to
support this second alternative by saying that a dravya is a
gunasamdrava (a confluence of qualities)." The term
'gunasamdrava' should be understood as synonomous with the
term gunasamudaya (aggregate of qualities) which Patafijali uses
to describe the nature of a dravya under 4.1.3, Striydm.12
91ha samdne varsmani parindhe cdnyat tuldgram bhavati lohasydnyat

karpdsanam yatkrto vis'esas tad dravyam. Tatha kascit sprsannevacchinatti
kascil lambamdno 'pi nacchinatti yatkrto visesas tad dravyam. Tatha kascid
ekenaiva prahdrena vyapavargam karoti kascid dvdbhydm api na karoti yatkrto
visesas tad dravyam. MB-K, vol. 2, p. 366, lines 19-23.
'OAthavd yasya gundntaresv api prddurbhavatsu tattvam na vihanyate tad
dravyam. Kim punas tattvam? Tadbhdvas tattvam. Tad yathd. Amalakadindm
phaldndm raktddayah pftddayas ca gundh prddurbhavanty dmalakam badaram ity
eva bhavati. MB-K, vol. 2, p. 366, lines 23-25. Note that I have paraphrased the
passage, "whose...essence is not destroyed (yasya...tattvam na vihanyate)
loosely as "which remains what it is." If 5.1.119 is interpreted according to
Katyayana's vdrttika 5, Siddham tu yasya gunasya bhdvdd dravye sabdanivesas
tadabhidhane tvatalau, then the essence (tattva) of a substance would be the
property by virtue of whose presence in the substance it is called what it is. In
the case of many individual substances of a kind, this essential property by
virtue of which a generic term is used for each of them is a class property.
11"Anvartham khalvapi nirvacanam: gunasamdrdvo dravyam iti." MB-K, vol. 2,
p. 366, lines 25-26. The word 'samdrdva, meaning that to which something
flows, is derived by providing the suffix ghai after the verbal root Vdru, to
flow, when it occurs with the preverb sam. 3.2.23. Sami yudruduvah (ghaf 16,
akartari ca kdrake sajiidydm 19). The word 'dravya' in the same meaning is
derived by providing the krtya-suffix yat after the verbal root vdru which
terminates in a vowel. 3.1.97. Aco yat (dhdtoh 91, k,rtydh 95). The suffix occurs
if the direct object (karman) is to be denoted in accordance with 3.4.70, Tayor
eva krtyaktakhalarthdh. See Kaiyata on 5.1.119. MB, vol. 4, p. 299a.
12"Gunasamuddyo dravyam". MB-K, vol. 2, p. 200, lines 13-14. See his

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 25
In the latter context, the conception of a dravya as an

aggregate of properties is extended to include bhdva (being)
denoted by nouns. It is stated that feminine gender is a property
of a dravya while defending the view that a feminine suffix
denotes feminine gender.13 On this view, if it were the case that
only that which is an aggregate of qualities (gunasamuddya) is a
dravya, then words with krt and taddhita affixes denoting bhdva
would not occur in the feminine because bhdva is not an
aggregate of properties. However, they do occur in the femin
For example, the word 'cikFrsd' with the krt suffix a, and 'g
with the taddhita suffix tal both terminate in the feminine suffix
tap.14 In order to account for words with krt and taddhita affixes
denoting bhdva, an extension is made to treat bhdva as if it were a
dravya.15 Kaiyata comments that the meaning of this extension is
that anything one may refer to with a pronoun is a dravya.16 In
other words, the word 'dravya' means an individual object.
example of the use of the term 'brdhmana' for a gunasamuddya under 2.2.6. MB-
K, vol. 1, p. 411, line 15 - p. 412, line 1 (p. 411, lines 15-21 are cited under
5.1.115, vol. 2, p. 363, lines 12-19). Although Kaiyata explains the etymology
by saying that a dravya is the dsraya of qualities (Gunandam asrayo dravyam ity
arthah.), and likewise explains the examples of fruit by saying that a dravya is
the asraya of color, etc. (Tasmdd asti riupdsrayo dravyam.) in accordance with
Vais'esika notions, Wezler (1985) convincingly argues that Pataijali refers to
Sankhya notions here. Biardeau (1964: 228) mentions that, according to the
Sankhya theory of a substance, a substance endures even with variations in its
formulation. See also Nagesa MB, vol. 4, 299a, and Laghumafijusa, p. 347. He
refers to Vyasa on Yogasutra 3.44, "Sdmdnyavisesasumuddyo 'tra dravyam....
Ayutasiddhdvayavabheddnugata-samiuho dravyam iti Patafijalih."
Patafijalayogasutra, pp. 162-63.
13The alternative is that feminine gender is included in the meaning of the
nominal base as the qualifier of the denoted object. In this case the feminine
suffix co-signifies feminine gender.
143.3.102. A pratyaydt (striydm 94, bhdve 18). The krt suffix a occurs after a
derivative verbal root if bhdva with feminine gender is to be denoted.
5.1.119. Tasya bhdvas tvatalau. The taddhita suffixes tva and tal occur after a
word if the bhdva of the object denoted by the word is to be denoted by the
derivate.
4.1.4 provides the feminine suffix tap. See ?1.2.5, note 36.
15Kasya tdvad bhavdn ekam gunam nydyyam manyate stritvam ndma?
Dravyasya. Dravye ca bhavatah kah sampratyayah? Yadi tdvad gunasamuddyo
dravyam? Kd gatir ya ete bhdvdh krdabhihitds taddhitdbhihitas ca: cikirsd,
goteti? Atha matam etat krdabhihito bhdvo dravyavad bhavatlti.... MB-K, vol.
2, p. 200, lines 12-15.
16Atha matam iti. Idam tad iti sarvandmapratyavamarsayogyam vastu
dravyam ity arthah. MB, vol. 4, p. 30a. Similarly, under 5.3.55, Atis'yane
tamabisthanau, Kaiyata says, "It is intended that anything fit to be referred to by
the pronouns 'this' and 'that' is a dravya. In this way generic properties,
qualities and actions are included too." MB, vol. 4, p. 374b. In the
Bhuiyodravya-samuddesah Bhartrhari says, Vastuipalaksanam yatra sarvandma
prayujyate; dravyam ity ucyate so 'rtho bhedyatvena vivaksitah. That object, to
refer to which a pronoun is used and which is intended as an object to be
qualified, is called a dravya. Vakyapadiya 3.4.3. Abhyankar and Limaye, p. 75.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
1.2.2. 'Dravya' Used to Mean an Individual

Object
Pataiijali uses the term 'dravya' unambiguously to mean a

individual object in other contexts. He gives an example of
use of many names for a single object citing the many names
the single deity Indra. Summarizing the purport of the examp
he says, "There are many names for a single individual ob
(dravya),"17 using the term 'dravya' for the single deity Indr
Similarly, Patanijali uses the word 'dravya' unambiguously
mean an individual object when he states that two propert
being a participant in an action (karakatva) and being the locu
an action (adhikaranatva), mutually specify each other "bec
they both inhere in the same individual object (dravya)."
gives the example, "Gdrgyo Devadattah," in which both wo
refer to the same individual person. The word 'gdrgya' qualifi
the person as a descendant of Garga; the word 'devadatta' qualif
the person as having the name 'devadatta'. The two propert
being a descendant of Garga (gdrgyatva) and being called
Devadatta (devadattatva), mutually specify each other "because
they both inhere in the same individual object (dravya)".18
In the first of these last two examples, a dravya is
understood to be the single object referred to by several words.
In the second, it is the single object which is the locus of two
properties designated by two words. Whenever two or more
words refer to the same individual object, or two or more
properties reside in the same individual object, the words, or
properties, are said to have the same locus. They are
samdnadhikarana. The compound 'samdnddhikarana' is widely
used in grammar and in other disciplines in this meaning. The
word 'adhikarana' in this compound denotes the individual object
which is the locus of properties or the object referred to by the
two words. Patafijali explicitly states that the word 'adhikarana
(locus)' is used as a synonym for the word 'dravya', both in
The Vrtti on 1.64 says: Ydvad idam tad iti prddhdnyenopddlyate tad dravyam.
Whatever (the pronouns) 'this' and 'that' denote as principal is a dravya.
Subramania Iyer, Kanda I, p. 122, line 5.
171.4.1, vdrttika 1. Anyatra samjaisamavesan niyamartham
vacanam. Anyatra samjnd samdveso bhavati. Kvdnyatra? Loke vydkarane ca.
Loke tdvat: indrah sakrah puruhutah puramdarah, kadruh kosthah kusula iti.
Ekasya dravyasya bahvyah samjnd bhavanti. MB-K, vol. 1, p. 296, lines 3-6.
18Under 1.4.23, vdrttika 15: Na hi kdrakam ity anenddhikaranatvam uktam
adhikaranam iti vd kdrakatvam. Ubhau cdnyo'nyavisesakau bhavatah. Katham?
Ekadravyasamavdyitvdt. Tad yathd. Gdrgyo Devadatta iti. Na hi gdrya ity anena
devadattatvam uktam Devadatta ity anena vd gdrgyatvam. Ubhau
cdnyo'nyavisesakau bhavata ekadravyasamavdyitvdt. MB-K, vol. 1, p. 326,
lines 11-15.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 27
ordinary parlance and in grammar.19 Therefore, Patafijali does

use the word 'dravya' to mean an individual object as a locus of a
certain class property and other properties.
1.2.3. Akrti: Permanent versus Transient

o
The phrase, "siddhe sabdarthasambandhe," in Katyayana's

first vdrttika states that the relationship between a word and the
object it denotes is established (?1.1). Patanijali gives the gloss,
"siddhe sabde 'rthe sambandhe ca," for the phrase, and interprets
the word 'siddha' as 'nitya'. In this way he shows that he considers
Katyayana to mean that the word, the denoted object, and the
relation between them are all permanent.20 Whatever object a
generic term denotes, it must be of a permanent nature in order to
maintain a permanent relation with the word which denotes it.
Hence if a generic term denotes an akrti, an akrti must be a
permanent entity. On the other hand, if it denotes a dravya, a
dravya must be a permanent entity. In order to determine which
is of a permanent nature and which is of a transient nature,
Patanijali enters into a discussion on the nature of an akrti and a
dravya at this point in the Paspasahnika.
He gives the following two ordinary examples in which a
substance (dravya) is permanent and its shapes (akrti) are
transient:
Clay connected with a certain shape is a ball. Crushing the ball-shape

small pots are made. Crushing the small-pot-shape cups are made.
Similarly, gold connected with a certain shape is a ball. Crushing the
ball-shape necklaces are made. Crushing the necklace-shape bracelets
are made. Crushing the bracelet-shape auspicious symbols (svastika)
are made. Again it returns to a gold ball. Again connected with
another shape it becomes two golden earrings the color of khadira
19Under 2.1.1, vdrttika 21, dravyam padartha iti cet: Athavd ydvad
bruiydt samdnadravyeneti tdvat samdnddhikaraneneti. Dravyam hi loke
'dhikaranam ity upacaryate. Tad yathd. Ekasmin dravye vyuditam.
Ekasminnadhikarane vyuditam iti. Tathd vydkarane vipratisiddham
cdnadhikaranavdci (2.4.13) ity adravyavdciti gamyate. MB-K, vol. 1, p. 371,
lines 2-5.
20See the discussion beginning, "Kim punar nityah sabda dhosvit kdryah?"
MB-K, vol. 1, pp. 6-7. Katyayana also clearly favors the view that speech units
are eternal, as he states in 1.1.1, vdrttika 9, Siddham tu nityasabdatvat. MB-K,
vol. 1, p. 40, line 26 - p. 41, line 4. See also MB-K, vol. 1, p. 40, line 26 - p. 41,
line 4. See also p. 60, lines 17-26, p. 104, lines 8-25, p. 112, lines 6-17, p. 112,
line 23 - p. 113, line 15, p. 136, line 11 - p. 138, line 26, p. 158, lines 6-8; vol. 2,
p. 113, lines 16-22; p. 234, lines 21-23; vol. 3, p. 251, lines 12-13; p. 420, line
21 - p. 421, line 1. For the posited view that speech units are producible (kdrya)
and transient, see: vol. 1, p. 257, lines 23-25, p. 404, lines 3-5.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
embers. Each shape is different but the substance is the same. By

crushing the shape, only the substance remains.21
Immediately following these examples, Patanijali considers the

alternative that the dkrti is permanent and individual substances
are not. He says, "It is not the case that if the dkrti ceases to be in
one substratum it ceases to be in all of them. On the contrary, it i
found present in another individual substance."22 In this case th
dkrti is considered to be absolutely fixed, eternal and
unchangeable.23 The dravya is considered to be a transient
individual object.
If one considers that a dravya is permanent and an dkrti is
not, then dravya is the material substance out of which an object
is made and an akrti is the shape or the limits of extension which
define an individual body made of that material. On the other
hand, if one considers that an dkrti is permanent and a dravya is
not, then an dkrti is a class property and a dravya is an individual
object of the class. A word must denote that which is permanent.
Hence in the former case it denotes material substance, and in the
latter case it denotes a class property.
Given Patanijali's use of the term 'dkrti' in the two
meanings of a transient shape and a permanent class property, one
may suppose that he uses the term in just one meaning somewhere
in between the two intending that an dkrti has characteristics of
both a shape and a class property.24 One may suppose that a
certain shape, viewed as a single property recurring in many
individual substances, serves to characterize each individual
substance as the object it is. Hence, the shape in a substance is its
class property. One may consider that the shape itself serves as a
class property in the examples of clay and gold formed into
different objects. In these examples, the shape of the object
serves to characterize the object as what it is and to distinguish
one type of object from another. Although the shape is not
precisely the same from one individual substance to another, and
although it is not immutable, it is recognized as essentially the
same shape. Hence it may be said to be permanent.25
21MB-K, vol. 1, p. 7, lines 11-18. Chandogyopanisad 6.1.4-6 has a comparable

passage. Ten Principal Upanishads, p. 505. See Biardeau 1964: 45 note 1 for
references.
22Nitydkrtih. Katham? Na kvacid uparateti krtvd sarvatroparatd bhavati

dravydntarasthd tupalabhyate. MB-K, vol. 1, p. 7, lines 19-20.
23See note 25.
24A survey of the literature on this topic is given in the Introduction.
25Athavd nedam eva nityalaksanam dhruvam ki.tastham avicdly
anapdyopajanavikdry anutpatty avrddhy avyayayogi yat tan nityam iti. Tad api
nityam yasmims tattvam na vihanyate. Kim punas tattvam? Tadbhdvas tattvam.
Akrtdv api tattvam na vihanyate. MB-K, vol. 1, p. 7, lines 21-23. My statement
that the shape is essentially the same paraphrases Patafijali's remark that the

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 29
1.2.4. Shape versus Substance as the

Alternatives
On the supposition that the term 'dkrti' means shape a

class property, it seems possible that Patanijali considere
permanent shape as one alternative in the debate concerning w
the object denoted by a word is, the other alternative b
dravya. We have seen that Pataiijali uses the term 'dravya' in
meanings: 1) a permanent material, and 2) an individual o
The examples of the clay and gold objects illustrate a dravya
the first sense. Dravya, as the material which remains the sam
different objects, is considered to be permanent. Pataiij
remarks in the Paspasahnika that it is as a permanent entity
dravya may be considered to be the object denoted by a
(?1.2.3). Hence it is possible to suppose that the debate over w
the object denoted by a word is, concerns the alternatives of
and material.
In Patanijali's examples depicting a substance remaining
permanent while its shapes vary, the words 'clay' and 'gold' are
mass nouns. Each of them denotes a single substance which is the
same in many different shapes. By the words 'ball, small pot, cup,
necklace, bracelet,' etc., which are generic terms, Patanijali intends
the shape in contrast to the substance. One could consider the
question concerning the meaning of generic terms to be whether
they are mass nouns denoting substance or a separate type of
word denoting shape. For example, the word 'wood', like the
words 'clay' and 'gold', is a mass noun. It denotes the substance
out of which trees and other objects are made. Does the word
'tree' denote only wood? Or does it denote a certain shape?
The supposition is that the debate over the meaning of
generic terms concerns whether they are mass nouns or a separate
type of word denoting shape. It may seem too obvious, on this
supposition, that generic terms denote shape and not substance.
The word 'wood' denotes a single substance whether it is the wood
of one or many trees, or whether one makes it into a table, chair
or house. But the word 'tree' is not used for wood made into a
table, etc. Its use is limited to an object with a certain shape. It
does not refer to a material but to a body. In the plural it refers to
many separate bodies, not to a single material.
essence (tattva) in the shape is not destroyed. Here I concur with Joshi and
Roodbergen 1986: 108-109, n. 444. If 5.1.119 is interpreted according to
vdrttika 5, then Patafijali's statement that there is a tattva in the shape which is
not destroyed means that there is a single generic property in each individual
shape-instance by which each is recognized as being a certain shape. This
accords with the conception that there are generic properties in quality
instances in Vaisesika.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Although it may seem too obvious that generic terms

denote shape and not substance, on the supposition that the debat
concerns whether they are mass nouns or a separate type of word
the debate under this supposition is not trivial. Consider the
English word 'water'. It is a mass noun denoting a single material.
But it is also used for water taking the shape of the glass that holds
it, and in the plural for many separate bodies of water. For
instance, when we ask a waiter for "three waters," he brings three
separate bodies of water in three glasses.
The question whether words such as 'tree' or 'cow' denote
a material or a shape is certainly an interesting question. Patafijal
seems to brush against the question in the examples he gives in
the Paspasahnika, and it may have been a question which his
predecessors considered in works we do not have. However, it is
not the question which Pataiijali and Katyayana deliberate a
length under sitra 1.2.64 and elsewhere. The question which
Patanjali and Katyayana discuss is whether a generic term denotes
a class property or an individual substance of the class. The
debate concerns a class property which is clearly distinguishe
from a shape, and an individual substance which is clearly
distinguished from a material.
The attempt to find a single meaning for each of the terms
'akrti' and 'dravya' is based on the fallacy that authors use a term
in a single meaning consistently. It is quite plain that Patafijali
uses each of the terms in clearly distinguished meanings. First let
us consider how this is true for the term 'akrti'.
1.2.5. Definitions of a jati under 4.1.63
Under 4.1.63, Jdter astrnvisaydd ayopadhdt, Patafijali

quotes two verses which define a generic property (jati). On the
condition that a feminine object is to be denoted, 4.1.63
introduces the suffix his after a nominal base denoting a generic
property which is not restricted to the feminine and which does
not have a penultimate y. Patafijali cites two verses to explain the
nature of a generic property to which the sutra refers:
.Jdter ity ucyate kajatir ndma?

(The sutra) says, "after a word denoting a generic property." What is a
generic property?"
A krtigrahand jdtir lingdndm ca na sarvabhdk;

sakrddkhydtanirgrdhyd gotram ca caranaih saha.
That which is known by means of a shape is a generic property. So
that which does not partake of all genders and which taught once wit
respect to one object is recognizable in others. Lineage is also
generic property, as are branches of Vedic learning.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 31
Apara Jha:
Someone else says:
Pradurbhdvavinddabhyam sattvasya yugapad gunaih;

asarvalingdam bahvarthdm tma jdtim kavayo viduh;
That which manifests, simultaneously with qualities, at the
appearance of an individual substance and disappears,
simultaneously with them, at the disappearance of the substance,
which does not have all genders and which resides in many objects,
that the sages know to be a generic property.
Gotram ca caranani ca.

They consider lineage and branches of Vedic learning generic
properties too.26
The first quarter of the first verse clearly distinguishes a shape

from a generic property. The shape or configuration
(avayavasamnivesavisesa) of an object is a means of knowing its
generic property.27 Here the word 'dkrti' denotes the shape, and
'jdti' denotes the generic property. As a further sign of
distinguishing shape from generic property, the verses classify
lineage and branches of Vedic learning, where shape is not an
identifying factor, as generic properties.
Yet, in commenting on this very sutra, Pataijali uses the
word 'fknri' to refer to a generic property. Just after using the
term 'dkrti' for shape as distinguished from generic property, he
uses it to refer to a generic property-the very meaning from
which he has just distinguished its previous meaning.
According to the second verse, youthfulness (kaumdrya)
is not considered a generic property because it does not appear at
a persons birth and last until the person's death.28 The first two
padas of the second verse are not satisfied. The nominal base
'kumdra' denotes youth. 4.1.20, Vayasi prathame, provides that
the suffix hip occurs after a nominal base denoting the first stage
of life, if a feminine object is to be denoted. This gives the
feminine form 'kumdri (girl)'.29 A bahuvrihi compound is
26MB-K, vol. 2, p. 225, lines 13-19. Referred to in MB-K, vol. 1, p. 303, lines
19, 22. For the use of the term 'jdti', see MB-K, vol. 2, p. 208, line 22 - p. 209,
line 5; p. 219, lines 3-6; p. 222, lines 5-6; p. 226, lines 20-22; p. 227, lines 17-
19; p. 250, line 6 - 251, line 1; p. 259, lines 16-19.
27See Kaiyata: Akrtigrahana. Avayavasamnivesavisesavyangyety arthah.
MB, vol. 4, p. 80a. Ojihara (1967) discusses the meaning of these verses in
detail.
28Under 5.3.55, Atisdyane tamabisthanau, Patafijali says, "That which is

obtained by birth is a generic property and there are no superior and inferior
degrees of this object. (Jananena yd prdpyate sd jdtir na caitasydrthasya
prakarsdpakarsau stah.)" MB-K, vol. 2, p. 414, line 4.
296.4.148 deletes the final a of 'kumdra' before the I of the suffix. See 1.2.64,
vdrttika 44 note 56.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
formed meaning he whose wife is (just) a girl (kumart bhdryd

yasya sa). 6.3.34, Striydh pumvad bhdsitapumskad anuni
samdnddhikarane striyam apuirantpriyadisu, applies so that in the
compound the word 'kumdri' appears like its masculine
counterpart from the same base, 'kumdra'.30 The result is
'kumdrabharyah '.
According to the first verse, youthfulness (kaumdrya) is
considered a generic property. The nominal base 'kumdra'
denotes a generic property, occurs in genders other than the
feminine and does not have a penultimate y. Hence 4.1.63
applies to form 'kumdri' ending in the suffix his in the
feminine.31 When the bahuvrihi compound is formed meaning
he whose wife is a girl, 6.3.41, jates ca, applies providing an
exception to 6.3.34: In the compounds formed on the conditions
under which 6.3.34 applies, a term denoting a generic property in
the feminine does not appear like its masculine counterpart from
the same base. The result is 'kumdrlbhdryah' in which 'kumdr'
retains its feminine suffix.32
In order to determine the scope of the qualification,
"which is not restricted to the feminine (astrivisayad)," in 4.1.63,
Patafijali asks whether the word must occur in the feminine "in the
same dkrti" or "in any akrti":
Athdstrivisaydd iti katham idam vijinyate: samandydm daktau yad

astrivisayam iti, dhosvit kvacid yad astrivisayam iti?
How is (the phrase) "which is not restricted to the feminine"
understood: "which is not restricted to the feminine in the same
dkrti," or "which is not restricted to the feminine in any dkrti"?33
Similarly, in order to determine the scope of the qualification,

"which has a masculine counterpart from the same base
(bhasitapumskad)," in 6.3.34, Pataijali asks whether the first
constituent of the compound must have a masculine counterpart
"in the same dkrti" or "in any dkrti":
Bhdsitapumskad iti katham idam vijnayate: samanayam dlatau yad

bhdsitapumskam dhosvit kvacid yad bhdsitapumskam iti?
How is (the phrase) "which has a masculine counterpart from the
same base" understood: "which has a masculine counterpart from the
30See Cardona 1988: 268-69, 211-12.

31This is identical to the word 'kumcri' ending in nip. Hence in order to show a
different speech form resulting from conceiving youthfulness as a generic
property, Pataijali must derive the compound where the different conception
shows its effect in the retention of the feminine suffix.
32Kah punar etayor jatilaksanayor visesah? Yatha purvam jatilaksanam tatha
kumaribharya iti bhavitavyam. Yathottaram tatha kumdrabharya iti
bhavitavyam. MB-K, vol. 2, p. 225, lines 20-21.
33MB-K, vol. 2, p. 225, lines 22-23.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 33
same base in the same akrti," or "which has a masculine counterpart

from the same base in any dkrti"?34
In both cases, Patafijali uses the word 'dkrti' in the phrase, "in the
same dkrti (samdnayam dkrtau)." The dkrti here is the semantic
condition for usage of the word (pravrttinimitta).35 One of the
conditions for usage for all nominal bases to which 4.1.63 applies
(and all of the bases which 6.3.41 excepts from the application of
the general rule 6.3.34), and with which Pataiijali's question is
therefore concerned, is a generic property. The semantic
condition for usage, here called dkrti, of nominal bases denoting a
generic property is a generic property (jdti).36
34MB-K, vol. 3, p. 150, lines 3-4. See MB-K, vol. 2, p. 408, lines 13-19; vol. 3,
p. 268, lines 4-14.
35See, "Samdnaydm akrtdv ekasmin pravrttinimitte," in the Kasika, part 2,
p. 707. (=KNP, part 5, p. 221.) The term 'dkrti' in the phrase, "samdndyam
dkrtau," is not restricted to just jatis. It extends to the condition for use
(pravrttinimitta) of other nominal speech forms as well. However, Patafijali
always uses it when the condition for use of a generic term is involved. See
under sutras 1.2.66-68 (MB-K, vol. 1, p. 247, lines 17-20; p. 248, lines 4-24)
and 7.1.74, vdrttika 5, Na vd samdnayam dkrtau bhhsitapumskavijfidndt and
bhdsya (MB-K, vol. 3, p. 268, lines 4-14). Under 2.2.28, Tena saheti tulyayoge
(MB-K, vol. 1, p. 429, lines 2-4) and 5.3.28, Daksinottardbhydm atasuc (MB-K,
vol. 2, p. 408, lines 13-14) it is used specifically for directions but the
discussion there refers to the more general discussion under 6.3.34. Hence we
should understand that the meaning pravrttinimitta is an extension of the
meaning jdti rather than that the meaning jdti is a limitation of the meaning
pravrttinimitta.
36Pataijali's discussion under 4.1.63 seems to conclude that the qualification,
"which is not restricted to the feminine (astrivisaydd)," applies to a term in any
meaning. Hence it applies to the words 'droni' meaning trough, 'kutf' meaning
hut, and 'pdtrF meaning a particular kind of pot. Each of these words occurs
only in the feminine in these meanings, but occurs in the masculine and neuter
in other meanings. According to Kaiyata and Nagesa, 'drona' meaning a certain
measure occurs in the masculine, 'kuta' meaning a particular pot is used in the
masculine, and 'pdtra' meaning a pot in general occurs in the masculine and
neuter (MB, vol. 4, p. 81b-82a). If the qualification astrfvisaya limited the
application of 4.1.63 to a term which is not restricted to the feminine in the
same meaning, it would not apply to these. By accepting that the qualification
limits the application of 4.1.63 to a term which is not restricted to the feminine
in any meaning, these words are included within the scope of 4.1.63. However,
with the qualification astrivisaya interpreted in this way the condition for the
application of the sutra is too broad. It includes the word 'mdla' meaning a
garland and the word 'baldkd' meaning a crane. These words also occur in
genders other than the feminine in other meanings. The word 'mala' meaning
rubbing occurs in the masculine, and meaning a certain field occurs in the
neuter; the word 'baldka', as derived from 'bala + dka', meaning braggard, agrees
in gender with the word it qualifies. But it is not desired that 4.1.63 apply to
them. To prevent iils from occurring after the nominal bases to which the suitra
should not apply, he suggests including the derivates 'mlai' and 'baldkd' of
these bases in the set beginning with 'ajd'. 4.1.4 provides the suffix tap for the
members of the set beginning with 'ajd' as an exception to 4.1.63. Panini 4.1.4.
Ajddyatas tap. The affix tap occurs after a base ending in a and specifically in

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Hence Patafijali uses the word 'dk.rti' to refer to a generi

property under 4.1.63. He does so just after citing a verse t
first quarter of which clearly distinguishes between a shape (dkr
and a generic property (jdti). Therefore, he uses the word 'dkr
to refer to two types of objects, a generic property and a shap
which he clearly distinguishes from each other.
1.2.6. Sivasutra 1, A i u n
Finally, Katyayana distinguishes class property from

configuration under Sivasutra 1, A i u n. In vdrttika 13,
Akrtigrahandt siddham, he concludes that mentioning the sound a
refers to all instances of all sounds of the a-class (avarna) "because
it includes the class property (dk.rti)." In vdrttika 16,
Ruipasdmdnydd vd, he gives the alternative solution, "or because
they have the same form (ruipa)." The second vdrttika offers a
solution which is different from the first. It is different only if the
word 'dkrti' in the first denotes something different from what the
word 'rupa' denotes in the second. In the first, 'dkrti' refers to a
generic property common to all the individual a-sounds of
various qualities. In the second, 'ruipa' refers to an aspect of each
individual sound which is similar from one to the other.37 Hence,
'dkrti' in vdrttika 13 denotes a class property as opposed to a
configuration denoted by 'rupa' in vdrttika 16.
the derivates 'ajd', etc. "Astu kvacid yad astrivisayam iti. Katham mdld
baldketi? Ajddisu pdthah karisyate." MB-K, vol. 2, p. 225, lines 25-26.
On the other hand, Patanijali concludes that the qualification, "which has a
masculine counterpart from the same base (bhdsitapumskdd)," in 6.3.34, applies
to a term only in the same meaning. "Astu samanayam dkrtau yad
bhdsitapumskam iti." MB-K, vol. 3, p. 150, lines 6-7.
Haradattamisra concludes in his Padamanijari commentary on 4.1.63 that there
too the qualification applies to a term in the same meaning, not in any meaning.
With this qualification the application of 4.1.63 is too narrow. It does not
apply to form the derivates 'dronl', 'kuti', and 'patri'. he suggests including
them in the set of bases beginning with 'gaura'. 4.1.41, Sidgaurddibhyas ca,
provides for the feminine suffix nis after the bases of the set beginning with
'gaura (white)' as well as after bases ending in an affix marked with s.
"Gaurddipdthdt siddham." KNP, part 3, p. 369. Note that 'drona' does appear
eleventh in the set beginning with 'gaura' in the Kasika, part 1, p.328.
37I have stated the difference as Kaiyata and Nagesa understand it. Kaiyata:
Akrtigrahanad iti. Akdrajdtih sarvapradesesu nirdisyate. MB, vol. 1, p. 88a.
Nagega under Ruipasdmdnydt: Abhedavyavahdra iti. Ruipasdd!sya-
nibandhanapratyabhijn~dkrtdbhedavyavahdra ity arthah. MB, vol. 1, p. 89a.
However, it may be that the configuration (rupa) itself is considered a general
property in vdrttika 16. In that case, the difference is that the class property
(dkrti) in vdrttika 13 is considered to be merely an abstract entity presumed as
the basis for identifying each sound as being of the same kind, while the
configuration is considered to be a perceptible characteristic.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 35
1.2.7. 'Akrti' and 'Jdti' are Synonymous under

1.2.58 and 1.2.64
We have seen that Patafijali uses the term 'dkrti'

distinct meanings: 1) a class property, and 2) a shape. In
sense it is equivalent to a generic property (jdti), and in
to a configuration (avayavasamsthdna). The following
paragraphs show that Pataijali uses the term 'Jkrti' to mean a class
property in his commentary on 1.2.64 concerning the denotation
of generic terms. As such it is synonymous with the term jdti'
meaning a generic property.
Many of the characteristics attributed to a generic
property (jdti) in the verses Pataiijali cites under 4.1.63 are
explicitly attributed to an dkrti under 1.2.64.
The phrase, "Which taught once with respect to one object
is recognizable in others (sakrddkhydtanirgrdhyd)," is used to
describe a generic property in the first verse under 4.1.63
(?1.2.5). Katyayana says, "And taught once, it is known," of an
dkrti in 1.2.64, vdrttika 38, Jndyate caikopadistam. The phrase,
"Which resides in many objects (bahvarthdm)," is used to describe
a generic property in the second verse under 4.1.63 (?1.2.5).
1.2.64, vdrttikas 40 and 56 (Asti caikam anekddhikaranastham
yugapat, etc.) assert being present in many substrata
simultaneously as an essential property of an dkrti, and vdrttika 48
denies the possibility of this in attempting to refute the existence
of an dkrti (?1.3.2-4, ?1.4.1).
The phrase, "Which does not partake of all genders
(linganam ca na sarvabhdk)," in the first verse, and the phrase,
"Which does not have all genders (asarvalingdm)," in the second
verse, speak of a generic property as being associated with fewer
than the three genders. Under 1.2.64, vdrttika 53 [53a], Pataijali
speaks of a generic property as gender specific. He says, "A
generic property is gender specific; from origin to destruction it
does not abandon the gender it starts out with (Avistaliingd
jdtir....)." He uses the word 'generic property (jdti)' to refer to the
very object which he has been calling an dkrti throughout his
commentary on 1.2.64. The reason he uses a different word for
the same object at that point is that the statement is taken from his
commentary on 1.2.52, Visesananam cdjdteh (?1.5.4 note 79). In
that context the word 'generic property (jdti)' is used because it
appears in the sutra itself.
Similarly, both Katyayana and Pataijali discuss the
question of the denotation of generic terms under 1.2.58 in terms
of whether a generic property (jdti) or an individual substance
(dravya) is denoted. The major question at issue is whether the
object denoted is a single object or many objects. The same issue
permeates the debate whether a generic term denotes an dkrti or a

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
dravya under 1.2.64. This is most evident in the discussions

related to the application of the rule that one remains (varttikas 1
3, 28-35, 52, 59) and the occurrence of different number-
terminations (varttikas 27, 46, 53-54). With the same point at
issue on the same topic with the same alternative (that the generic
term denotes a dravya), the discussion under 1.2.58 uses the term
jdti' and the discussion under 1.2.64 uses the term 'dkrti'. One is
forced to conclude that the terms are synonymous. The term
'akrti' under 1.2.64 means a class property. The reason the term
djati' is used instead under 1.2.58 is that Panini himself uses it in
the suttra.
1.2.8. Dravya: an Individual Object under

1.2.58 and 1.2.64
Patanijali uses the term 'dravya' in different meani

as he uses the term 'dkrti' in different meanings. He use
'dravya' in at least three distinct senses: 1) an individual
a material, and 3) wealth. For example, in the phrase, "
wealth (dravya) something is does what he has to with it
the word 'dravya' to mean wealth.38 We have already
the use of the term in the first two meanings in the Pa
(?1.2.3). In the discussions concerning the denotation of
terms under 1.2.58, 1.2.64, and elsewhere, Katyayana and
Patafijali use the term 'dravya' to mean an individual object.
The advantage of the view that a generic term denotes a
dravya over the view that it denotes a class property is that the
former view naturally accounts for differences. The use of dual
and plural terminations is achieved naturally on the view that a
generic term denotes a dravya. 1.2.58, varttika 6, Asisyam vd
bahuvat prthaktvdbhidhdndt, says that there is no need to make a
special provision for the plural terminations on a generic term
because it denotes many objects as separate. Vdrttika 7,
Jdtisabdena hi dravydbhidhdnam, explicitly uses the word 'dravya'
to refer to these separate objects (?1.8).
1.2.64, vdrttika 46, Tathd ca litigavacanasiddhih, also
claims that dual and plural terminations are achieved naturally on
the view that a generic term denotes a dravya. Similarly, vdrttika
46 says that the use of a generic term in different genders is
achieved naturally on the view that a generic term denotes a
dravya (?1.3.3). The use of more than one of the same word in
analytic strings and usages attributing different qualities to the
38Atas cavayavd evdrthavanto na samuddyd yasya hi tad dravyam bhavati sa

tena kdryam karoti yasya ca td gdvah santi sa tdsdm ksaram ghrtam copabhunkte
'nyair etad drastum apy aiakyam. MB-K, vol. 1, p. 217, lines 19-21. For another
use of 'dravya' meaning wealth, see MB-K, vol. 2, p. 340, lines 11-16.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 37
objects they denote are natural if the words denote different

individual objects. Varttikas 50-51 (Asti ca vairupyam. Tatha ca
vigrahah) assert that these usages are accounted for naturally on
the view that a generic term denotes a dravya (?1.3.3).
The use of more than one of the same word in analytic
strings and usages attributing differences to the objects they
denote are difficulties, on the view that a generic term denotes a
class property, because a class property is a single object. As a
single object, it cannot have any number other than the singular
number, nor can it have different properties at once. In defense
of the view that a generic term denotes a class property, under
vdrttika 54 Patafijali explains the use of dual and plural
terminations and different gender terminations based on the dual
and plural number and different genders of the dravyas in which
the class property resides (?1.5.4). After stating in vdrttika 55,
Adhikaranagatih sdhacarydt, that cognition of a substratum arises
because it accompanies the class property, in vdrttika 58,
Vairuipyavigrahau dravyabheddt, Katyayana explains that the
differences one speaks of as belonging to a class property are due
to the differences of the dravyas (?1.3.4).
The passages noted attribute the fact that it naturally
accounts for differences, and for dual and plural number, to the
view that a generic term denotes a dravya. Differences, and dual
and plural number, naturally belong to two or more distinct
individual objects. They do not naturally belong to a single
object. The material out of which objects are made is a single
object; hence dual and plural number, and differences of gender
and other qualities are as foreign to it as they are to a class
property. If Katyayana and Patanijali used the word 'dravya' to
mean the material out of which various individual bodies are
made, the view that a generic term denotes a dravya would have
precisely the same difficulties as the view that a generic term
denotes a class property. The fact that these difficulties are
attributed solely to the view that a generic term denotes a class
property and that different gender and number terminations and
other usages based on the differences of the objects they denote
are naturally achieved on the view that a generic term denotes a
dravya, shows that such a dravya is conceived of as an individual
object, not as the material out of which objects are made.
Therefore, Katyayana and Pataiijali use the term 'dravya' in
1.2.58, 1.2.64, and in other discussions concerning the denotation
of generic terms, to mean an individual object.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
1.2.9. Conclusions Regarding the Term 'akrti'

and Other Terms
We have concluded that in their extended discussions

concerning the denotation of class terms, Katyayana and Patai
use the term 'dkrti' to mean a class property and the term 'drav
to mean an individual object. Therefore, the question concerni
the denotation of generic terms has to do with whether a gene
term denotes a class property or an individual object, not whet
it denotes a shape or a material. Other authors in differen
cultural contexts have raised the same question in terms of
universal versus particular, type versus token, etc. In the
discussions that follow and in the translations of the selections
from the Mahabhasya, the English words 'class property' an
'individual substance' translate the Sanskrit 'dkrti' and 'dravy
respectively. The term 'generic property' translates the Sanskrit
jati'. Even though 'dkrti' and jdti' in these discussions are
synonymous, the distinct use of the Sanskrit terms has been
preserved by translating them consistently with the synonyms
'class property' for 'akrti' and 'generic property' for 'jati'.
Although I regularly translate 'dkrti' as 'class property' and
'jati' as 'generic property', there is one drawback in doing so. We
consider a class to be the collection or set of all objects in which
the class property is present, and clearly distinguish between the
class property and the class. Neither Katyayana nor Patafijali
distinguishes between a class property and the class it serves to
define, or between a generic property and the genus it serves to
define. They do consider an dkrti a property because they speak
of it as residing in substrata just as qualities do. These substrata
are the individual substances of the class. Katyayana in vdrttikas
40, 41, 48, 55, and 56, and Patafijali in his commentary thereon,
speak of it in relation to its substrata. But this does not imply that
it is an entity distinct from the class itself. Indeed, number and
action are also considered to be properties of objects in which
they inhere.
Vdrttika 49, Vinase prddurbhdve ca sarvam tatha sydt,
implies that the class property is identical to the class. This
vdrttika raises the following objection against the view that a
generic term denotes a class property: If a generic term denotes a
class property, to say that the denoted object comes into being or
dies is to say that all of the individuals of the kind do so. This
objection makes sense for a class but not for a class property. It is
true that if a class property is destroyed, then there are no longer
any objects of the class either. But if a class property comes into
being or manifests itself, it does not imply that every object that
ever is, was, or will be of that kind comes into being with it. It
implies only that at least one does. As vdrttika 40a, Asti caikam

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 39
anekadhikaranastham yugapat, states, and Patanjali's discussion

under vdrttika 43 makes clear, the class property has the relation
of being wholly present in each of its substrata. Hence one
individual of the class is sufficient to manifest the class property.
A class does not have the relation of being wholly present
in each of its substrata. A class has the relation with individual
objects of the class of being one object in relation to all of the
individual objects collectively, just as the number 3 is a single
object in relation to three individual objects collectively (it has
vyasajyav rtti). Hence a class requires all of the individual objects
of the class to come into being in order for it to come into being.
So speaking of the class as coming into being implies that all
individuals of the class come into being. Understanding the
object as a class better suits the intent of the objection raised in
vdrttika 47 than understanding it as a class property does.39
Similarly, in the phrase, "if the word refers to a class
(jdtydkhyayam)," in sutra 1.2.58, I translate jdti' as 'class' rather
than 'generic property'. The first objection raised against the sutra
in Patafijali's commentary asserts that the plural terminations are
naturally appropriate because there are many objects. Nothing in
the word 'class property' implies that there are many objects while
the word 'class' does. Translating 'jdti' as 'class' here brings out the
implication which the first objection shows is present in the term
adti'.
Patafijali uses the word 'sdmdnya', which I translate
'general property', to mean anything common to more than one
object. Hence he often uses it to refer to a generic property or
class property. However, he also uses it to refer to an action or a
number (which are not considered generic or class properties).
Commenting on 1.2.64, vdrttika 59, Vyarthesu ca sJmdnydt
siddham, Patanjali mentions action and number as general
properties common to the various meanings of polysemous terms
(? 1.3.4).
1.3. The Denotation of Generic Terms
In the Paspasahnika, Patanjali briefly states that a w

denotes both a class property and an individual substa
says that this conclusion is based on the fact that Panini c
his rules in both ways. He composed 1.2.58, Jdtydkhy
ekasmin bahuvacanam anyatarasydm, considering that a g
term denotes a class property, and he composed 1.2.64,
39See ?1.3.3 note 58 and 1.2.64 translation, vdrttika 49. Note that Kaiyata
states the objection as applying to a class property and reinterprets the vdrttika
to mean as much. See vdrttika 49 note 63.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Saruipanam ekasesa ekavibhaktau, considering that it denotes an

individual substance.40
While this passage does express Patafijali's conclusion
concerning the denotation of generic terms, it is not necessarily
the case that the statement of 1.2.64 implies that a generic term
denotes an individual substance. Under 1.2.64 Patafijali discusses
the denotation of generic terms at length and clarifies his
conclusions on the subject under 1.2.58. We will investigate his
views by reconstructing the nature of the problem as it arises for a
grammarian.41
1.3.1. Denotation is Natural
1.3.1.1. Concomitant Presence and Absence
I have mentioned that Patafijali singles out a speech

from among the various entities which together form
whole of an experience as the means by which a knower
an object of experience. Speech is continuous. A cont
uninterrupted utterance goes alongside a continuous
knowledge of objects. By observing these two parallel
the grammarian sees certain elements of speech which
with certain cognitions of objects. When a certain speech
present, a certain object is known. When that speech
absent that object is not known. By observing the corre
presence (anvaya) and absence (vyatireka) of certain
forms with the knowledge of certain objects, the gra
determines that certain speech forms denote certain
Ordinary usage is the ultimate criterion of the meaning
and observation of the corresponding presence and a
words and knowledge of objects in ordinary usage is the
determine their meaning.42
40Kim punar dkrtih paddrtha dhosvid dravyam? Ubhayam ity dh

jndyate? Ubhayathd hy dacryena sutrdni pathitdni. Akrtim paddrt
Jdtydkhydydm ekasmin bahuvacanam anyatarasydm (1.2.58) it
Dravyam paddrtham matvd Sarupandm (1.2.64) ity ekasesa drabhy
vol. 1, p. 6, lines 8-11.
41Under 5.1.119, Tasya bhdvas tvatalau, and 5.3.55, Atisdyane tam
Pataiijali discusses the question whether a word denotes a prop
individual substance. I cannot deal with these sections in detail her
will mention the essentials of the former discussion. Although in
the latter discussion is interesting for its portrayal of the two ext
which it commences as follows: Kah punar nydpprdtipadikdrthah
Na vai dravyasya prakarsa isyate. Evam tarhi gunah. See MB-K, vol
15.
42See Cardona (1988: 499-503, 1967-68: 313-352, 1981: 79-104, 1983: 37-
116) concering anvaya and vyatireka. Scharfe (1961) deals with varied aspect

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 41
Take for example the generic term 'vrksa (tree)'.

Observing ordinary usage one sees that, when the speech form
'vrksah (vrksa + s)' is used, one understands a certain meaning:
that which has roots, a trunk, fruit, and leaves, and singular
number. When 'vrksau (vrksa + au)' is used, the speech form
'vrksa' remains, s is missing and au is added. Some meaning also
is missing, some is added, and some remains. Singular number is
missing, dual number is added, and that which has roots, a trunk,
fruit, and leaves remains. The meaning which is missing belongs
to the speech form which is missing, the meaning which is added
belongs to the speech form which is added, and the meaning
which remains belongs to the speech form which remains.43
In this manner, by concomitant presence and absence of
certain speech forms with certain meanings, one determines which
meaning belongs to the base and which to the suffix. The use of
the base 'vrksa' corresponds to the meaning, that which has roots,
a trunk, fruit, and leaves. The use of the suffix s corresponds to
the meaning, singular number; the use of the suffix au to dual
number, and as to plural number.44 Hence 'vrksa' denotes that
which has roots, a trunk, fruit, and leaves, and the suffixes s, au,
and as denote singular, dual, and plural number, respectively.
1.3.1.2. One Word Used for Each Object

Denoted
Now consider the fact that speech forms are used in

to convey knowledge of the objects they denote. Eac
knowledge of a certain object is to be conveyed, one spee
denoting that object will be used. There is a one to one
relationship between the knowledge of the object and the use of
the speech form denoting the object.
We have been very careful to say that it is the knowled=g
of the object to which the speech form corresponds rather than to
the object itself. The question is open whether a single cognition
can have 1) many objects or 2) only a single object. Suppose that
of Patafijali's logical procedures, and Staal (1988) deals with several logical
concepts in Indian thought. Halbfass (1983: 54-67) examines Sankara's
references to anvaya and vyatireka.
43See Patafjali under 1.2.45, vdrttika 9, Siddham tv anvayavyatirekdbhyam.
MB-K, vol. 1, p. 219, lines 19-27.
44Yatas tu khalu niyogato v!rksa ity ukte svabhdvatah kasmimscid arthe pratltir
upajdyate 'to manydmahe neme sdmdnya.abdd iti. Na cet samanyasabddh
prakrtih prakrtyarthe vartate pratyayah pratyaydrthe. MB-K, vol. 1, p. 220, lines
6-8. For a more complete discussion on this section of the Mahabhasya see
Cardona 1967-68: 332-36.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
it has only a single object. For each cognition there is a single

object known.
We have just said that there is a one to one relationship
between the knowledge of the object and the use of the speech
form denoting the object. We have also made the supposition that
a single cognition has a single object. The principle may be
formulated on the basis of these two statements that for each
object denoted a single speech form is used. In 1.2.64, vdrttika 1
Katyayana states this principle using the phrase, "words apply pe
object (pratyartham sabdanivesah)."
1.3.1.3. The Grounds for the Rule that One

Remains
Now suppose that a generic term denotes an indivi

substance and not a generic property. We have just determ
by concomitant presence and absence that the nominal
'vrksa' denotes that which has roots, a trunk, fruit, and leaves.
Hence the nominal base 'vrksa' denotes each individual tree. To
denote each individual tree, one nominal base 'vrksa' is used. In
this case a problem arises: To denote many trees, many bases
'vrksa vrksa vrksa' would arise. Although it is accepted usage to
refer to many trees singly by using one word 'vrksah' for each
(each base 'vrksa' with the singular termination s), it is also
accepted that one word 'vrksdh' (one base 'vrksa' with the plural
termination as) refers to many trees. This latter usage would not
arise. In order to obtain this usage, one would have to provide
for it explicitly. 1.2.64, Sariupanam ekasesa ekavibhaktau,
accomplishes this. According to 1.2.64, just one of speech units
of the same form remains if they occur with one inflectional
termination.45
45Cardona (1988: 11-12, 306-8), gives an overview of rules providing for a

single remainder. Patafijali debates whether the rule applies to nominal bases
(prdtipadika) or to words (pada) at length. On each alternative, Pataijali
addresses problems in the derivation on alternative interpretations of the sutra.
I cannot discuss all of these problems and alternatives here. However, whether
the remainder of one applies to nominal bases or to words, it is still true that
one nominal base would arise for each object denoted. The following sketches
of derivations merely illustrate the two alternatives. If 1.2.64 applies to
nominal bases, the derivation may be sketched like this:
vrksa vrksa vrksa One base arises before the termination to
denote each individual tree.
vrksa vrksa vrksa + as The plural termination arises by 1.4.21.
vrksa + as Just one base remains before the termination
by 1.2.64.
vrksas 6.4.102.
If 1.2.64 applies to words, the
vrksa, vrksa, vrksa One base arises to denote each individual tree.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 43
Given the two assumptions 1) that for each object denoted

a single speech form is used, and 2) that a generic term denotes an
individual substance, one cannot account for the use of one
speech form in the plural referring to many individual substances
naturally. One must make an explicit statement in order to bring
this about.46 Although suitra 1.2.64 achieves the use of one
speech form in the plural on these assumptions, this solution is not
satisfactory. It is preferable to account for the usage of certain
speech forms as naturally arising on the basis of certain principles
to convey their meanings rather than to have to state a rule
explicitly.
The grammatical procedure of allowing many speech
forms to occur to denote many objects and then providing that
just one of those speech forms remains to denote those objects is
an artifice to make known that the one word which remains
denotes many meanings. Such procedures are used in Panin
grammar for the purpose of making known correct speech form
The provision and deletion of grammatical elements is such
procedure. It makes known that the meaning attributed to
grammatical element by concomitant presence and absence
sometimes present even in the absence of that grammatica
element. For example, the word 'somasut' is formed by providin
the suffix kvip after the verbal root Vsui when the subordina
word 'soma' denotes the direct object, and the word 'agnicit
formed by providing the same suffix after the verbal root Vc
when the subordinate word 'agni' denotes the direct object. Th
suffix kvip, called krt, occurs to denote the agent of the actio
denoted by the verbal root. The suffix is later deleted. Howeve
vrksa-s, vrksa-s, vrksa-s A singular termination arises after each nomina

base by 1.4.22.
vrksa-s + vrksa-s + vrksa-s A dvandva compound is formed by 2.2.29,
Carthe dvandvah.
(vrksa-s vrksa-s vrksa-s) + as The plural termination arises after the dvandva
base by 1.4.21.
vrksa-s + as Just one word remains before the termination
by 1.2.64.
vrksa + as The termination s of the base is deleted by
2.4.71.
vrksas 6.1.102.
See 1.2.64 translation, note 81
2.4.71. Supo dhdtuprdtipadik
termination (sup) which is part
nominal base (prdtipadika).
6.1.102. Prathamayoh purvas
dirghah 101). A long vowel
single replacement of an a, i, u
second triplet nominal termina
46Note that given the fir
Paspasahnika that Panini com
denotes an individual substance is true.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
the meaning remains. The derivates 'agnicit' and 'somasut' denote

the agents of building a fire and pressing soma even though any
affix denoting an agent is absent.47
Panini has to posit deletion in his grammar because the
relation of a grammatical element and the meaning in which it is
provided are not entirely coextensive. In the example of the
suffix kvip, the absence of the suffix in the words 'somasut' and
'agnicit' is evidence that the suffix is absent when its meaning is
present.48 The relationship of the suffix and its meaning falls
short of strict concomitant presence and absence. The
concomitant presence and absence is posited to enable the
formulation of a general rule that the grammatical element occurs
to denote a certain meaning. Deletion is then posited to account
for the exceptions.
Just as concomitant presence and absence is merely
posited in the case of deletion, similarly, it is merely posited in the
case of the use of one speech form used to denote many objects.
It has been accepted as the second assumption above that a
generic term denotes an individual substance. If one closely
observes ordinary usage one sees that when the nominal base
'vrksa' is present, sometimes the meaning "one tree" is present,
sometimes the meaning "two trees", and sometimes the meaning
"many trees". The usage of the base 'vrksa' is not coextensive
with the meaning "one tree" (nor with the meanings "two trees" or
"many trees"). The concomitant presence and absence of the
nominal base 'vrksa' with the meaning, one tree, is posited in
47The sutra-rejecter argues that there is a difference between the provision and
deletion of lopa and the provision that one remains. He claims that the former is
necessary to derive the correct speech forms whereas the provision that one
remains is not. 1.1.62, Pratyayalope pratyayalaksanam, provides that an
operation conditioned by an affix applies even if the affix has been deleted. In
the example Patafijali gives under 1.2.64, the suffix kvip is a krt affix marked
with p. 6.1.71, Hrasvasya piti krti tuk, provides that the final augment tuk is
added to a short vowel before such an affix. Although the suffix kvip has been
deleted, by virtue of 1.1.62 it provides the condition for the application of
6.1.71 which provides the final t in the words 'somasut' and 'agnicit'. In this
way the provision of the suffix kvip and its replacement by lopa serve a
purpose in deriving the correct speech forms. It is not merely an artifice to make
known that a certain speech form has a different meaning without effecting any
change in the speech form itself. Although this distinction may be valid for
lopa because of 1.1.62, it is not valid for luk. The provision of a grammatical
element and its deletion by luk effects no change in the stem (1.1.63, Na
lumatdngasya); it is merely an artifice to make known that a certain speech form
has the meaning of the deleted element (and to allow the term 'pada' to apply).
See 1.2.64 translation, vdrttika 29 notes 18 and 23. Concerning positing zero
replacements, see Cardona 1988: 503-10.
48Actually, according to the rules of the Astadhyayi, kvip is always deleted by
6.1.67 (see translation, vdrttika 29 note 18). It is termed aprkta by 1.2.41,
Aprkta ekdl pratyayah, because it consists of the isolated sound v. See Cardona
1988: 48 and 1975.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 45
order to abide by the general principle that for each object

denoted a single speech form is used (?1.3.1.2, ?1.3.1.3
assumption 1). The remainder of one is then posited to account
for the cases where the use of the nominal base 'vrksa' does not
correspond to one tree but rather to two trees or many trees.
1.3.1.4. One Word Naturally Denotes Many

Meanings
According to the first assumption, a single speech form is

used for each object denoted. Suppose that we do not accept this
assumption. Although the assumption is based on the fact that
speech forms are used in order to convey their meaning, the fact
does not imply that a speech form necessarily conveys only one
object. The meaning a speech form conveys may include more
than one object. That speech forms are used in order to convey
meaning implies only that the appropriate speech form is used
when the meaning it denotes is to be conveyed. Let us replace the
first assumption with this one, and hold onto the second, that a
generic term denotes an individual substance.
Now let us reexamine the ordinary usage of the generic
term 'vrksa'. We find that if the meaning to be conveyed is one
tree, the speech form used is 'vrksah', if the meaning to be
conveyed is two trees, the speech form used is 'vrksau', if the
meaning to be conveyed is many trees, it is 'vrksdh'. The nominal
base 'vrksa' is never used alone so, properly speaking, it does not
convey meaning by itself. It conveys meaning only when used
together with a suffix.49 Therefore, accepting that the base and
suffix together constitute the meaningful speech unit, and not
attempting to extract a nominal base with a consistent meaning,
even if it is accepted that generic terms denote individual objects
one can account for correct usage naturally.
Grammar is an empirical science. It formulates rules to
account for existing usage. The fact that speech forms are used in
order to convey meaning implies that the speech form used for a
certain meaning arises naturally to convey an understanding
which speakers have. Hence the ordinary use of language has
inherent in it certain commonly held conceptions concerning the
nature of things, and certain usages presuppose certain
conceptions. In the course of formulating and examining rules to
account for ordinary usage, many of these conceptions and
presuppositions come to the fore. Patafijali comes upon the
49See 1.2.45, varttika 7, Arthavattd nopapadyate kevalendvacandt, and vdrttika

8, Na vd pratyayena nityasambandhdt kevalasydprayogah, and bhdsya. MB-K,
vol. 1, p. 219, lines 10-18.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
question whether a generic term denotes a generic property or an

individual substance because certain usages seem to presuppose
that it denotes one or the other. In determining which it denotes,
he weighs the assumption that it denotes one or the other against
other assumptions such as that for each object denoted a single
speech form is used.
In deciding whether to accept an assumption or not,
Pataiijali tests whether grammatical rules can be formulated in a
simpler way to account for all the usages by accepting the
assumption rather than in a more complicated way by rejecting it,
or vice versa.5(? That a certain assumption leads to a simpler
formulation of rules is evidence that that assumption is in fact a
presupposition of ordinary usage of the language. If it leads to an
additional formulation of rules, it is evidence that it is not a
presupposition. Hence, if adopting an assumption entails the
explicit formulation of a rule but rejecting it does not require such
an explicit statement, it is a reason for rejecting the assumption.
Given the assumption that 2) a generic term denotes an individual
substance, to accept 1) for each object denoted a single speech
form is used requires the explicit formulation of the rule that just
one of homophonous speech forms remains.
The explicit formulation of the rule can be avoided even
by accepting that a generic term denotes an individual substance,
if one does not carry out the analysis of word-meaning relation
by concomitant presence and absence fully. If one does not
attempt to attribute a consistent meaning to the nominal base
separately, then one may accept that the word 'vrksdah' and other
plural forms denote many individual substances. In this case one
word denotes many objects. The principle that one word denotes
one object cannot be adhered to as such. It can be accepted only
in the looser sense that each word is used to denote whatever its
meaning is whether that meaning be one object, two objects, or
many objects.
On the other hand, if one does carry out the analysis of
word-meaning relation by concomitant presence and absence
fully and does attribute a consistent meaning to the nominal base,
given the assumption that one word is used for each object
denoted, to accept that a generic term denotes an individual
substance requires one to formulate the rule that just one of
homophonous speech forms remains.
50Similarly Filliozat 1991: 63.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 47
1.3.1.5. Reasons for Positing Generic

Properties
Rejecting the assumption that one word is used fo

object denoted, it is accepted outright that a generic te
plural has a one-many relation with the individual subs
denotes. According to the assumption that one word is
each object denoted, there is a one-one relationship
word used and the object denoted. Given the assumpt
generic term denotes an individual substance, when
many individual substances to be denoted, many w
because of the one-one relationship between the wor
object denoted. By the application of 1.2.64 the sin
which remains has a one-many relation with many
substances.
The fact that words are used to convey meaning implies
that a certain word conveys a certain cognition. A generic term
applies to each of many objects based on the fact that the
cognition which the term conveys is the same for each of many
objects. Whether in actual fact there is a single object that is the
basis of the same cognition is not a matter of concern for a
grammarian. The fact that there is an identical cognition is.
Speakers use words based directly on their knowledge of objects,
not based directly on the objects themselves.
A generic term used in the plural conveys a single
cognition of many objects as identical to each other. Ordinary
speakers have a single cognition of many objects as identical to
each other. There is a one-many relationship between the one
cognition and the many objects. Similarly there is a one-many
relation between the generic term used to convey that cognition
and the many objects.
The innocent assumption regarding every cognition is that
it is correct unless some other evidence contradicts it. When one
has knowledge of an object, one innocently assumes that the
object is as one knows it. Hence it is natural to assume that a
single entity is the basis of a single cognition of many objects as
identical to each other. One assumes that a single forest is the
basis of the cognition of many individual trees as a group. One
assumes that a single number 2 is the basis of the cognition of two
objects as a pair. Similarly, one may assume that a single generic
property or class property is the basis of the cognition of many
objects as being of a kind.
For example, the word 'cow' is used for each of many
individual cows. The word 'cow' is used for each of them based
on the fact that one recognizes each object as a cow. The
knowledge, "x is a cow," is identical regarding the many
individual cows. Not only does one have the identical cognition

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
with respect to each cow, but in addition one recognizes that each
cow is identical in so far as it is a cow. The existence of a single
generic property, cowness, present in each cow is posited as the
basis for the same cognition regarding each cow and the
cognition that each is the same. In any case in which an identical
cognition arises regarding each of many objects and one knows
each of those objects as being the same in a certain respect, one
posits the existence of a generic property present in each of them.
A generic term used in the plural has a one-many relation
with individual substances. The single generic term is used to
convey a single cognition of many individual substances as
identical to each other. There is a one-many relation between the
cognition and the individual substances. If one accepts that a
single class property is the basis of the cognition of many
individual substances as identical to each other, the class property
has a one-many relation to the many individual substances. But
the single class property accepted as the basis of the single
cognition may be accepted as the object denoted by the generic
term. In this case, there is a one-one relation between the word
used and the object denoted.
Therefore, one may uphold assumption one, that for each
object denoted a single speech form is used, and still account for
the use of a single generic term arising in a natural way. The
generic property is a single object so it is natural that one word is
used to denote it. By accepting that a generic term denotes a class
property one accounts for the use of a single generic term to refer
to many individual substances naturally without having to
formulate the explicit rule that just one of homophonous speech
forms remains.
1.3.2. A Generic Term Denotes a Class Property
Patafijali presents two views of what a generic term

denotes under 1.2.64. In varttika 35, Akrtyabhidhdndd vaikam
vibhaktau Vdjapydyanah, Katyayana mentions Vajapyayana as a
proponent of the view that a generic term denotes a class
property, and in vdrttika 45, Dravydbhidhdnam Vyddih, he
mentions Vyadi as a proponent of the view that it denotes an
individual substance. In vdrttikas 36-44 Katyayana presents
reasons for the view that a generic term denotes a class property.
Varttika 35 makes known the implication of the view that
a generic term denotes a class property for the need to state
1.2.64. The suitra is unnecessary on this view because it is natural
that just one nominal base arises before an inflectional
termination. 1.2.58, varttika 1, Jdtyakhydydm samadnydbhidhdndd

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 49
aikarthyam, voices what is understood here, that denoting the

class property means the word has just one object (?1.8).
Vdrttika 36, Prakhydvisesdt, asserts that there is an
identical cognition with respect to each of the individual
substances of a kind. In itself this is evidence for the existence of
a class property. If this identical cognition arises from the use of
a generic term, it also serves as evidence that a generic term
denotes a class property. Pataijali gives an example of identical
cognition arising from the use of a generic term. He says that the
use of the generic term 'cow' does not lead to the cognition of the
specific features of any individual cow.
Vdrttika 37, Avyapavargagates ca, asserts that in addition
to identical cognition with respect to each of the individual
substances of a kind, one knows that they are not set apart from
each other. One knows that each is the same as the other in some
respect. Here Pataiijali gives the same example of the use of the
generic term 'cow'. Hence this reason serves as evidence both for
the existence and denotation of a class property. The knowledge
of individual substances as not set apart from each other is
evidence that they have a common class property. If this
knowledge arises from the use of a generic term, it is evidence for
the fact that a generic term denotes that class property.
Vdrttika 38, Jnidyate caikopadistam, supports the evidence
given in vdrttika 36 for the existence of a class property.
Together with Patafijali's explanation, it states that once one is
taught what one individual substance of a certain class is, one
knows what other individual substances of that class are too.
Once one is taught that one object is a cow, one recognizes oth
animals of different colors and ages at different times and pl
as being cows. If a common property were not present in
different individuals of the class, this would not be possib
Therefore, there must be a class property which they have
common. Having learned the class property in a specific
individual of the class, one recognizes it in any individual
substance in which it resides. The fact that one recognizes a
previously unknown individual substance is evidence that it has a
property which one has known previously in another individual.
The fact that one recognizes a number of objects as being x is
evidence that there is a single generic property, x-ness, belonging
to each of those objects.
Vdrttika 40, Asti caikam anekadhikaranastham yugapat,
etc., supports the argument that there is such a thing as a class
property. A class property is a single entity which resides in
many substrata simultaneously. Vdrttika 40a asserts that it is in
fact possible for one entity to be present in many substrata
simultaneously. Patafijali examines the sun as an example of such
an entity but dismisses the example because it is not analogous to
the present case. In vdrttika 40b, itlndravad visayah, Katyayana

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
proposes Indra (or his name) as such an object.51 Invoked

many Vedic performances at once, Indra is present
simultaneously in them all. A class property is present in many
objects simultaneously as Indra is present in many Vedic
performances simultaneously.
Vdrttika 41, Naikam anekddhikaranastham yugapad iti cet
tathaikasese, strengthens this argument for the existence of a class
property by pointing out that the existence of an object which has
a one-many relation cannot be avoided by denying the existence
of a class property. On the view that a generic term denotes an
individual substance it is accepted that a generic term in the plural
denotes many individual objects. The one word which remains
according to 1.2.64 is a single object which has the relation of
signifier-signified to many objects simultaneously. If a signifier-
signified relation can exist between one word and many objects
then the relation of property-substratum can hold between one
class property and its many substrata.
Varttika 39, Dharmasdstram ca tatha, supports the
evidence given in vdrttika 37 for the denotation of a class
property. Instructions concerning how one ought to behave are
formulated as general injunctions and prohibitions. They are
meant to apply universally within their sphere of application, that
is, to every object that meets their conditions. Hence a generic
term used in such an instruction does not refer to a specific
individual. Rather, it refers to any individual which has a certain
class property. Therefore, the word denotes the class property,
not a particular individual.
Vdrttikas 42, Dravydbhidhdne hy dkrtyasampratyayah,
and 43, Tatrdsarvadravyagatih, explain why it cannot be the case
that a generic term denotes a particular individual. If it denoted
an individual and not the class property, cognition of the class
property would fail to arise. Without cognition of the class
property, that of all the individuals of the kind would not arise;
cognition of just the one particular individual object denoted
would arise.
Under vdrttika 39, Pataijali gives two prohibitions as
examples of general instructions concerning behavior. In the
prohibition, "One should not kill a brahmana," if the generic term
'brdhmana' denoted an individual, the rule would mean that one
should not kill that particular individual. It would not prohibit
killing any other individual because the word 'brdhmana' would
not denote any other individual. However, it is understood that
the rule prohibits killing any individual of the brdhmana class.
Therefore, it must be the case that the generic term 'brdhmana'
51See 1.2.64 translation, vdrttika 40b and notes 45-46, concerning the
interpretation of the example. The interpretation here agrees with Paranjpe
1922: 45-46.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 51
denotes the class property common to all and only brdhmanas.

Similarly, in the prohibition, "One should not drink liquor," if the
generic term 'liquor' denoted a particular individual drink of
liquor, it would not denote any other drink of liquor. The rule
would mean that one should not drink that particular drink but
would not prohibit drinking any other drink. However, it is
understood that the rule prohibits drinking any liquor at all.
Therefore, it must be the case that the generic term 'liquor'
denotes the class property common to all alcoholic drinks.
Under vdrttika 43, Pataiijali gives the injunction, "One
should bind a cow, a goat for Agni and Soma," as an example of
general instructions concerning behavior. If the word 'cow' or
'goat' denoted an individual cow or goat, it would fail to denote
any other. Only binding that particular animal would fulfill the
injunction and once that animal had been used in a Vedic
performance, no one else could succeed in fulfilling the
command to bind it. It would be useless for any one else to do
the performance or for the original performer to do it again
because they would be unable to bind that particular cow or goat
again, they would be unable to fulfill the command so the fruit of
the performance would be unattainable. However, it is
understood that one may carry out the instruction repeatedly on
any cow or goat. Therefore, each of the generic terms 'cow' and
'goat' denotes a class property, not an individual substance.
Finally, vdrttika 44, Codanayam caikasyopddhivrtteh,
considers that a general injunction associates an attribute with a
class property, not with an individual substance. In the
injunction, "One should set out an astdkapala (eight-cup) rice cake
oblation for Agni," the attribute 'a.s.takapalatva (prepared in eight
cups)' qualifies the oblation to be offered to Agni. The injunction
associates the attribute, prepared in eight cups, with a single thing.
If that single thing were an individual substance, only one person
could carry out the injunction once because the injunction would
attribute only that particular individual substance with being
prepared in eight cups, and no other. However, it is understood
that the injunction is meant to be carried out repeatedly. If the
single thing with which the attribute of being prepared in eight
cups is associated is a class property, the command may be
fulfilled by carrying out the action on any individual substance in
which the class property and its associated attribute reside
together. Hence the command may be carried out repeatedly.
Because the correct meaning of the injunction is arrived at if the
attribute is associated with a class property, and it cannot be
arrived at if the attribute is associated with an individual
substance, it must be the case that a generic term denotes a c
property and not an individual substance.
Patafnjali says (under vdrttika 38) that vdrttika 38 suppor
vdrttika 36 and (under vdrttika 38) that vdrttika 39 supports

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
varttika 37. In addition, from the questions introducing vdrttikas

36 and 37 one can infer that the questioner understands that
vdrttika 36 establishes that a class property exists, and that varttika
37 establishes that it is denoted.52 The question, "But how does
one know that there is one class property and it is denoted?
introduces vdrttika 36, and the question, "Well, even if one know
that there is one class property because there is undifferentiate
cognition, what is the basis for asserting that the word denotes it?
introduces varttika 37. Hence varttika 36 satisfies the questioner
concerning the existence of a class property but does not satisfy
him concerning its being denoted. He asks this of varttika 37.
Although these questions indicate that vdrttika 36 should
serve only as evidence for the existence of a class property and
not as evidence for the fact that it is denoted, nevertheless what I
have stated above is true. Pataiijali's phrase, "When the word
'cow' is uttered," under vdrttika 36, allows one to understand that
the identical cognition referred to in the varttika arises from the
use of a generic term. If identical cognition with respect to many
individuals of a kind arises from a generic term, one can conclude
both that a class property exists and that it is denoted. Since it is
true that both vdrttikas 36 and 37 offer evidence both for the
existence and denotation of a class property, one must understand
that the questioner simply has not understood the full implication
of varttika 36 and its commentary and requires a fuller
exposition.
The argument that a generic term denotes a class property
comprises evidence for the existence of a class property and
evidence for the fact that it is denoted. Vdrttikas 36 and 37 give
evidence of both. The latter is not redundant because the
evidence it offers is different. The first asserts that there is an
identical cognition with respect to many individual objects; the
second that there is the cognition that the individual objects are of
the same kind. Vdrttikas 38, 40 and 41 argue for the existence of
a class property. Vdrttikas 39 and 42-44 argue that it is denoted.
1.3.3. A Generic Term Denotes an Individual

Substance
In vdrttikas 46-52, Katyayana presents reasons fo

view that a generic term denotes an individual substance. V
48 denies the existence of class properties, and varttika 50
evidence to this effect too. Vdrttikas 46-47, 49, 51, and 52
for the view that an individual substance is denoted.
52Note that Kaiyata makes much of this distinction. See translation, note 37.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 53
Varttika 48, Na caikam anekddhikaranastham yugapat,

denies that a single entity can be present in many substrata
simultaneously. Pataiijali gives the example of an individual
person, Devadatta. It is impossible for him to be present in the
city of Srughna at the same time he is present in the city of
Mathura. Just as it is impossible for an individual person to be in
two places at once, it is impossible for a single entity to reside in
many substrata at once. Therefore, there is no such thing as a
class property.
Varttika 50, Asti ca vairupyam, says that individual
substances differ from each other. The fact that one individual
substance is similar to another is not grounds for establishing the
existence of a class property common to both. Although they are
similar, they are nevertheless recognized as distinct objects with
individual differences. Kaiyata explains that it is impossible for
two objects to be both different and the same. He says that the
cognition of objects as the same is based on the fact that they
serve the same function, it is not based on an identical property of
the objects themselves (vdrttika 50 note 64). In fact, the objects
are recognized as different. Since the objects are recognized as
different and the cognition of their sameness is accounted for by
the fact that they serve the same function, there is no reason to
posit the existence of a class property common to the individual
substances. Therefore, there is no such thing as a class property.
As a corollary, the use of a generic term to refer to objects
attributed with different properties implies that it refers to more
than one object. It is impossible for a single entity to have
property a and not property b, and yet have property b and not
property a. For example, "an unwhole bull, a hornless bull
(khando gaur, mundo gaur)," refers to one bull which is crippled
or injured and another which has had its horns clipped. Hence
each use of the generic term 'cow (gauh)' in this example denotes
a different entity. A class property is a single entity; the
individual substances of a kind are multiple. Therefore a generic
term denotes an individual substance.53
Varttika 51, Tathd ca vigrahah, argues that the use of
analytic strings presupposes that a word denotes an individual
substance. for example, the analytic string, "gaus ca gaus ca (a
cow and a cow)," is used to show the meaning of the word 'gdvau
(two cows)'. In the analytic string, each use of the generic term
'gauh (cow)' denotes an individual substance. If both denoted a
class property, both would denote the same entity. Because using
one word is enough to convey cognition of the entity, usage of
the second would not occur. It would be redundant to use both
53See Nagesa, translation, note 64.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
words.54 The fact that there are two usages of the generic ter
'gauh (cow)' shows that each denotes a distinct object. In
addition, the word 'ca (and)' in the analytic string makes known
that there is an aggregate of more than one object. There is just
one class property but there are many individual substances of a
kind. Therefore, a generic term denotes an individual substance
and not a class property.
Vdrttika 46, Tathd ca lingavacanasiddhih, argues that the
occurrence of a generic term in different genders and numbers
presupposes that it denotes an individual substance. If a generic
term denoted a class property, it would occur only in the singular
because a generic property is a single entity and a singular
termination naturally arises when a single object is to be denoted.
A generic term would never occur in the dual or plural.
However, a generic term is used in the dual and plural.
Therefore, a generic term does not denote a class property.
There are many individual substances of a kind. One
taken by itself has singular number, two taken together have dual
number, and many taken together have plural number. A
singular inflectional termination naturally arises to denote singular
number, a dual termination naturally arises to denote dual number
and a plural termination naturally arises to denote plural
number.55 Hence it is natural for a generic term to occur with
different number terminations if it denotes individual substances.
Therefore, a generic term denotes an individual substance.
Similarly, if a generic term denoted a class property, it
would occur in only one gender, the gender which belongs to that
class property. However, the generic term 'brdhmana' occurs both
in the masculine (brahmanah) and in the feminine (brahmant).
The masculine form occurs to denote an object which has the
masculine gender, and the feminine form occurs to denote an
object which has the feminine gender. A single person cannot be
both a male and a female so the male is one person and the female
another. Because their genders are different, there are two distinct
objects denoted. The class property is a single entity; but there
are many individual substances of a kind. Therefore, a generic
term denotes an individual substance and not a class property.
Vdrttika 47, Codandsu ca tasydrambhdt, argues that a
generic term denotes an individual substance because one
performs the action enjoined in a command on an individual
substance. Given the injunction, "One should bind a cow, a goat
for Agni and Soma," one binds an individual cow or goat. One
does not bind a class property cowness or goatness. Therefore,
54See translation, vdrttika 29, note 15 concerning the principle, "One does not
use speech forms whose objects have already been denoted (uktdrthdnam
aprayogah)."
55In accordance with 1.4.21-22. See translation note 81.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 55
each of the generic terms 'cow' and 'goat' must denote an

individual substance. If they denoted class properties, binding an
individual would not fulfill the command. A class property and
an individual are two different entities. The command is fulfilled
only if one does the action on what it says to do it on. However,
it is understood that the command is fulfilled if one binds an
individual cow or goat. Therefore, 'cow' denotes an individu
cow and 'goat' denotes an individual goat.
Varttika 49, Vinase pradurbhdve ca sarvam tatha sydt
points out a difficulty with the view that a generic term denotes
class property. If it were the case that a generic term denotes a
class property, when one spoke of destruction or coming in
being, one would convey that everything of that kind is destroy
or comes into being. In the statement, "A dog has died," if it we
the case that the word 'dog' denotes the class property doghood
then the statement would mean that doghood has ceased to exis
If there were no class property of being a dog, then not a sing
dog would exist. Hence, the statement, "A dog has died," would
mean that every dog has ceased to exist. Conversely, in the
statement, "A cow has been born," if the word 'cow' denoted the
class property cowness, then the statement would mean tha
cowness has come into being. Pataiijali states that there would no
be enough room for everything that ever is, was or will be
cow.56 To manifest the class property cowness, every single cow
would have to be born. However, when one says, "A dog has
died," one means that an individual dog has died, and when one
says, "A cow has been born," one means that an individual co
has been born. therefore, the generic terms 'dog' and 'cow' denot
individuals, not class properties.
Finally, vcrttika 52, Vyarthesu ca muktasams'ayam,
defends the view that a generic term denotes an individua
substance against the accusation that it requires the explic
formulation of a rule which the view that a generic term denotes
class property does not require. This vdrttika argues that adoptin
the view that a generic term denotes a class property does n
spare one from stating 1.2.64. The rule is necessary even on the
view that a generic term denotes a class property. Even on t
view that a generic term denotes a class property there are gener
terms which denote many objects (polysemous terms). For
example, the word 'aksa (die, axle, seed)' would denote the three
56This objection against the view that a generic term denotes a class proper
assumes that it is not the case that one class property resides wholly in each
its substrata simultaneously. It assumes that the class property, like the cl
itself, is fully present only when all the individuals of the class are presen
This mode of relation to substrata is called vyasajyavrtti. The prime example
it is the relation of a whole to its parts. A whole exists only when all of its pa
do. See ?1.2.9.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
class properties, being a die, being an axle and being a see

According to the principle that for each object denoted a singl
speech form is used, one nominal base 'aksa' would be used f
each class property denoted. Hence to refer to all three at once
three bases 'aksa aksa aksa' would arise. Although it is accep
usage to refer to each of these class properties singly by using o
word 'aksah (singular)' for each, it is also accepted that one wo
'aksdh (plural)' can be used to refer to objects of all three types
once. This latter usage would not arise. In order to obtain that
single usage of the generic term 'aksdh (plural)' denotes all thr
class properties at once, one would require a rule to provide th
just one of homophonous speech forms remains if they occ
with one inflectional termination. Hence one would have to state
1.2.64 anyway, even on the view that a generic term denotes a
class property. According to vdrttika 35, the reason for adopting
the view that a generic term denotes a class property rather than
the view that it denotes an individual substance is that one would
not have to state 1.2.64. This reason is invalid because 1.2.64 has
to be stated anyway.
1.3.4. Denotation of a Class Property

Vindicated
In varttikas 53-59 Katyayana responds to each of

arguments against the view that a generic term denotes a
property presented in vdrttikas 46-52. In vdrttika 56, Na
anekddhikaranastham yugapad ity ddityavad visayah, he r
to the objection in vdrttika 48 that a single entity can
present in many substrata simultaneously by reiterati
statement in varttika 40a that there is in fact such an ent
vdrttika 56 Katyayana gives the sun as an example of s
entity. Pataiijali repeats his previous comment that the exam
the sun is not analogous to a class property and reitera
example of Indra. A class property is present in many sub
simultaneously just as Indra is present simultaneously in e
the Vedic performances in which he is invoked.
A generic term denotes a class property on this v
however, as vdrttika 55, Adhikaranagatih sdhacarydt, expl
cognition of an individual substance arises because an indiv
always accompanies a class property. Because it is impos
perform action on a class property, one understands that
takes place on an individual object which accompanies th
property. Thus when one speaks of birth and destruction
connected with a class property, it is an individual substance
associated with the class property which is born or destroyed, not
the class property itself. Vdrttika 57, Avinaso 'nasritatvdt,

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 57
explains that although it accompanies each individual substance

of the class, a class property itself is not dependent on the
individuals. Hence it is not destroyed when an individual
substance is destroyed.
Similarly, according to vdrttika 58, Vairuipyavigrahau
dravyabheddt, one speaks of different properties associated with a
class property because the differences belong to its accompanying
individual substances. Hence it is justified that in analytic strings
there is one occurrence of the generic term for each individual
substance. Just as one speaks of other different properties
associated with a class property because of the differences of its
accompanying individual substances, vdrttikas 53, Litigavacana-
siddhir gunasyanityatvat, and 54, Gunavacanavad va, explain that
different genders and numbers succeed because they are
properties which belong to the individual substances associated
with a class property, not to the class property itself. The
properties of the individual substances are the conditions for the
occurrence of the various suffixes.
Finally, vdrttika 59, Vyarthesu ca sadmnydt siddham,
denies that the rule that one remains would have to be stated even
on the view that a generic term denotes a class property. The
sutra-rejecter adopted the view that a generic term denotes a class
property in order to achieve the use of a single word without
having to state explicitly that one remains. Accepting that one
word denotes a single object, he asserts that the class property is
the single object denoted by the generic term. Because the class
property is a single object, naturally only one word is used to
denote it and there is no need to state that only one remains
(1.2.64). Arguing for the view that a generic term denotes an
individual substance, in vdrttika 52, Vyarthesu ca mukta-
samsayam, the sutra-proponent claimed that 1.2.64 would have to
be stated anyway, even on the view that a generic term denotes a
class property, in order to account for the use of a single nominal
base with a plural termination in the case of polysemous terms
(?1.3.3).
In vdrttika 59, the sutra-rejecter responds that there is a
single general property common to all the objects to which the
word is used to refer. The solution is exactly parallel to the
solution of accepting a single class property common to all the
individual substances of a kind. That is, a "polysemous" generic
term denotes the single general property common to all the
objects for which the word is used. That property may be, for
example, the action denoted by the verbal root from which the
term is derived in its various meanings. Or the property may be a
number which the term denoted in an original meaning no longer
current. Because it is a single object, naturally just one word is

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
used to denote it and there is no need to state that one remains.

1.2.64 is rejected as unnecessary.57
1.4. Equivalence of the Two Views
1.4.1. Enjoined Universally, Fulfilled in a

Particular
Under vdrttika 43, Patanjali points out that there is equ

fault on both views. In vdrttikas 42 and 43 Katyayana attribu
the following fault to the view that a generic term denote
individual substance: Cognition of the class property would no
arise; hence one would fail to understand all the individual
substances to which the term applies (?1.3.2). A generic term
would apply to a particular individual in the same arbitrary way
that a proper name applies to an individual. A proper name does
not apply to all the individuals of a class. Similarly, a generi
term would fail to apply to other individuals of the class. Or if it
did apply to other individuals of the class, one would have t
learn separately, for each individual of the class, that it applies to
that individual; just as one has to learn separately, for each
individual to which a certain proper name applies, that it applies
to that individual. However, it is a fact that cognition of an
individual of the kind may arise from the use of a generic term
And once one has learned the meaning of a generic term it applie
to any individual of the class. There may be innumerable objects
of a kind so one would have to learn innumerable word-meaning
relations.
In later discussions on the topic of the denotation of
generic terms, it is standard to attribute two faults to the view that
57Note that Patanjali does not accept this as the final conclusion. Patafjali
accepts that there are words which denote more than one meaning (polysemous
terms). Under 1.2.45, varttika 9, Siddham tv anvayavyatirekdhhydm, which
along with its commentary is identical to 1.3.1, vdrttika 6, he gives the very
words cited in 1.2.64 as examples of such words: Eka.s ca .abdo hahvarthah.
Tad yathd. Aksdh pdddh masd iti. MB-K, vol. 1, p. 220, line 2; p. 255, lines 21-
22. The fact that a word denotes more than one meaning is inconsistent with the
proposition that it denotes just a single over-arching general property common
to all the objects for which a word is used.
Under 4.1.63 Patafijali concludes that the qualifier 'astrivisayam (which is not
restricted to the feminine)' limits the sutra's application to the nominal base of a
generic term which is not restricted to the feminine in any of iLt meanings, that
is, regardless of the condition for use of the term (kvacid astrivisayam). It does
not limit it to the base of a generic term which is not restricted to the feminine
on a certain condition for use of the term (samdndydm dkrtau). See ?1.2.5, note
38. The discussion accepts that generic terms may have more than one
condition for use (pravrttinimitta), that they denote more than one meaning. If a
generic term denotes more than one meaning, then 1.2.64 has to be stated.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 59
it denotes an individual substance. These two faults are referred

to as deviation (vyabhicdra) and infinity (dnantya). 1) If a
generic term denotes a particular individual it fails to denote any
other. Because one does understand that it refers to others, it
deviates from its denoted meaning. 2) If it does denote each
individual of the class, then there is a separate word-meaning
relation for each of the innumerable individuals of the class.
Hence there would be innumerable word-meaning relations all of
which would have to be learned in order to apply the word
properly to an arbitrary individual of the class.
The view that a generic term denotes a class property does
not have these faults. Cognition of the class property arises
because the term denotes it. The class property is present in all
individuals so one would understand all the individuals. The
word-meaning relation is between one word and one denoted
object so learned once it is known.
However, vdrttikas 47 and 49 indicate complementary
advantages to the view that a generic term denotes an individual.
If a generic term conveyed knowledge of all individuals of a kind
because it denotes the class property, then it would convey
knowledge of every single individual when just a single arbitrary
one is intended. For example, in the command, "One should bind
a cow, a goat for Agni and Soma," if the words 'cow' and 'goat'
conveyed knowledge of all cows and goats because they denote
the class properties cowness and goatness, then one would have to
bind all cows or all goats to fulfill the command. This is
impossible. No one would be able to carry out the action to fulfill
the command and no one would attain the fruit of the action. In
the statements, "A dog has died" and "A cow has been born," if
the words 'dog' and 'cow' conveyed knowledge of all cows and
goats because they denote the class properties, doghood and
cowness, then one would understand that all dogs have died and
all cows have been born. However, that is not the way general
injunctions and statements are understood. Action does not take
place on all of the individuals of a kind at once; it takes place on a
single individual in any given instance. The fact that the action
takes place on particular individuals in particular instances rather
than on every individual of the class in any particular instance i
better accounted for if a generic term denotes an individual.
The solution to these difficulties for both views is
achieved by accepting that a general injunction enjoins
universally but is fulfilled with respect to each individual. O
view that a generic term denotes a class property this is acc
for by accepting the class property as an entity which is en
present in each of many individual substances. Katyayana a
that there is such an entity and gives Indra as an exam
vdrttika 40 while presenting the reasons that a generic
denotes a class property (?1.3.2). He reiterates the same

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
the sun as an example in varttika 56 while vindicating the view

(?1.3.4). Patanjali, discussing both examples in both places,
rejects the example of the sun because it is not analogous b
accepts the example of Indra.
The analogy of a class property to the sun is as follows:
We know that the class property present in different individuals
different times and places is a single entity just as we know th
the sun which appears to be in different places to differen
individuals is a single entity (varttika 40a, note 43, vdrttika 56
Jaimini 1.1.15, Adityavad yaugapadyam, gives the sun as an
example of an object perceived in many locations simultaneousl
Sabara's description portrays the following scenario: To eac
observer facing the sun in the early morning, the sun appears
be directly in front of himself (because of its size and distance
Hence to person A the sun appears to be directly in front
person A and to person B it appears to be directly in front
person B who stands 30 feet to A's right. Comparing their
observations it would seem that the sun is in two places (?3.2.2
note 12).
The example of the sun is not parallel to the case of the
class property. One person can see one class property in ma
substrata at once. However, the appearance of the sun in tw
places depends on two observers standing in different location
(varttika 40a, note 44, varttika 56). But Patafijali does accep
Indra as an example of a single object present in many places a
once. One knower is aware of one class property present in
different individual objects at once just as it is possible for on
observer to be aware of one Indra invoked (or his name
enunciated) in many Vedic performances at once (varttika 40b,
notes 45-46, vdrttika 56).
While presenting the view that a generic term denotes an
individual substance, Katyayana says in varttika 48 that it is not
possible that a single entity be present in many substrata
simultaneously. Patanjali gives the example of an individual
Devadatta. He cannot be present in two different cities, Srughna
and Mathura, simultaneously (varttika 48, note 62).
Although the proponent of the view that a generic term
denotes an individual substance argues that it is impossible for
one entity to be present in many places at once, he himself must
accept that a single entity as a whole is related to each of many
individual objects severally. The single generic term applies to
each of many objects severally. Although the statement of 1.2.64
achieves the relation of one term in the plural to many objects, as
Katyayana points out in varttika 41, it does not achieve the use of
a singular term in a general command or statement for one
individual in one instance and for another individual in another
instance (?1.3.2). In order that the same term apply to a differ
individual in each instance, he must accept that a single entity,

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 61
generic term, is related to each of many individual objects as their

signifier. If he can accept the relation of one to many in the case
of a word denoting each of many individual substances, he can
accept it in the case of a generic property present in each of many
individual substances.
1.4.2. Number
Pataiijali mentions that there is equal merit on bot

in another respect as well. The view that a generic term
generic property accounts for the use of a single nomi
denote many individual substances of a kind in a nat
One speech form is used to denote one object, i.e., t
property.
However, it has difficulty accounting for the d
plural inflectional terminations. The dual termination n
arises when dual number is to be denoted. The plural te
naturally arises when plural number is to be denoted.58
property is a single object; it does not have dual or plura
Hence the dual and plural terminations would not arise
However, one does use generic terms in the dual and
Therefore, if a generic term denotes a class property, a
statement has to be made to account for the correct sp
(end of bhdsya on vdrttika 43).59 To have to make
explicit statement is a shortcoming of the view that a ge
denotes a class property.
The view that a generic term denotes an indiv
substance accounts for the use of singular, dual, and
inflectional terminations in a natural way. The appr
number naturally belongs to however many individual
58According to 1.4.21-22. See translation, note 81. I do not make a

of whether it is the many objects which are the condition for th
plural terminations or the plural number which belongs to them. Pa
up a series of discussions parallel to the debate whether a generic ter
class property or an individual substance: whether number or th
objects (samkhyd or samkhyeya) are the semantic condition of
terminations and number words, whether the participant in an a
relation it has with the action (sakti or saktimat) is the semantic
certain suffixes, whether feminine gender or the object qualified
condition for the use of feminine suffixes, and whether a certain p
as excellence (prakarsa 5.3.55) or reproach (kitsa 5.3.74), or that wh
property is the semantic condition for taddhita affixes. Each
discussions requires a lengthy treatment which I cannot give here.
59Note that 1.2.58 makes such an explicit statement for the use o
terminations. It states that a plural termination occurs when one ob
denoted if that object is a generic property (?1.8). Panini make
statement to provide dual terminations for a word referring to a sin
property.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
are to be denoted. One individual substance has singular numbe

two individual substances have dual number, and many
individual substances have plural number. The singular, dual,
and plural terminations arise naturally to denote singular, dual,
and plural number respectively. Hence there is no need for an
explicit statement to account for the different number
terminations on the view that a generic termination denotes an
individual substance.
However, it has difficulty accounting for the use of a
single nominal base when many individual substances are to b
denoted. To denote many individual substances, many nomina
bases would naturally arise and 1.2.64 would have to be explicitly
stated to provide that only one remains and the others are
dropped (varttika 43).
There is equal fault on both views. Although the view
that a generic term denotes a class property accounts for the us
of just one nominal base naturally, it has to make an explicit
statement to account for the use of dual and plural inflectional
terminations. Conversely, the view that a generic term denotes an
individual substance accounts for the use of dual and plural
terminations naturally, but it has to make an explicit statement to
account for the use of just one nominal base.
1.4.3. Gender
The view that a generic term denotes a class proper

difficulty accounting for the use of a generic term in d
genders just as it has difficulty accounting for dual a
inflectional terminations. Unlike quality words, gener
generally do not agree in gender and number with the w
modify. In the phrase, "Badari vrksah," the word 'badar
feminine denotes a species of tree. The generic term
(tree)' occurring in the masculine does not agree with it in
On the basis of the fact that generic terms occur in a s
gender consistently, the general rule is made that g
properties are permanently associated with a certain
gender. Patanijali expresses this principle under 1.2.52, v
(?1.5.4 note 78) and cites it under 1.2.64, varttika 53 (53a, note
72). He says, "A generic property is gender specific; from origin
to destruction it does not abandon the gender it starts out with."
The verses describing the character of a generic property which
Patafijali cites under 4.1.63 refer to this property of a generic
property to be permanently associated with a specific gender.
The second quarter of the first verse describes a generic property
as "that which does not partake of all genders (litigdndm ca na

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 63
sarvabhak)." The third quarter of the second verse describes it as

that "which does not have all genders (asarvaliingdm)" (?1.2.5).
Gender is considered to be a property of the object a word
denotes. If a generic term denotes a class property, the fact that
the class property is permanently associated with a certain gender
accounts for the consistent use of a generic term in a single
gender. The class property always has its innate gender and the
gender suffixes and inflectional terminations appropriate to that
gender arise naturally after the nominal base of the generic term.
For example, the class property treeness is permanently associated
with masculine gender. Hence the base 'vrksa' never takes
feminine suffixes and is always followed by inflectional
terminations appropriate for the masculine. This accounts for the
use of the term 'vrksa' in the masculine in the example, "Badart
vrksah," cited above.
This also accounts for the consistent use of generic terms
in the masculine when referring to male and females of a kind.
For example, in the phrase, "brahmana ime (these brdhmanas),"
referring to males and females of the brahmana class, the word
'brahmanah' is in the masculine. In this case the consistent use of
the generic term in the masculine could be accounted for by the
fact that the generic property of being a brahmana (brahmanatva)
is inherently masculine. The speech form used to denote a
masculine object naturally occurs in the masculine. Hence the
consistent use of generic terms in the masculine when referring to
both males and females of a kind would be achieved naturally if a
generic term denoted a generic property. Since it is achieved
naturally there would be no need to state the rule that just a
masculine speech form remains.
On the view that a generic term denotes an individual,
'brahmanah (masculine singular)' would arise to denote each male
brahmana and 'brahmanl (feminine singular)' would arise to
denote each female brahmana. 1.2.64 would not provide that one
speech form remains because the condition for its application, that
the speech forms are homophonous, is not met. Hence an
additional rule would have to be stated. 1.2.67 makes this
provision.60
Although on the view that a generic term denotes an
individual 1.2.67 has to be explicitly stated to account for the
consistent use of a generic term in the masculine when male and
female individuals are referred to, different genders are naturally
accounted for on this view. For example, 'brahmant (feminine
singular)' denotes a female brahmana and 'brdhmanah (masculine
singular)' denotes a male brahmana. The feminine suffix hls
occurs after the nominal base 'brahmana' naturally if the
60See 1.2.64 translation, varttika 53a note 72.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
individual to be denoted is a female of the brdhmana class, and

naturally no feminine suffix occurs if a male is to be denoted
(?1.2.5).
Although the view that a generic term denotes a class
property has the benefit of naturally accounting for the use of
generic terms in a single gender consistently, it has difficulty
accounting for the use of generic terms in different genders. If a
generic term denoted a class property which is permanently
associated with an innate gender, the term would always occur in
one gender only. For example, if the nominal base 'brdhmana'
denoted the class property of being a brdhmana (brdhmanatva)
and the class property of being a brdhmana were innately
masculine, then the feminine suffix his would not arise to form
'brdhman'.61 Yet the word 'brdhmant' does occur. In order to
justify its occurrence one would have to accept that there were
two brchmana class properties, one associated with masculine
gender and one associated with feminine gender. But that
contradicts the premise that a single class property is denoted.
And if one accepted that there were more than one class property
denoted, more than one speech form would occur to refer to
males and females of the class together. One would have to state
the rule that only one remains. Hence the very reason for
adopting the view that a generic term denotes a class property
would fail.
1.5. Solutions on the View that a Class

Property is Denoted.
1.5.1. Properties are not Innate
The difficulty in achieving the use of a generic term in

different genders and numbers arises on the view that a generic
term denotes a generic property from accepting that a generic
property is permanently associated with a certain innate gender
and with singular number. In vcrttika 53 Katyayana solves the
difficulty by accepting the reverse: gender and number are not
innate. A class property is not permanently associated with a
certain gender and is not permanently associated with singular
number.
The solution as stated is untenable. First of all, a class
property is a single entity so it has singular number by nature. It
is on the basis of its being a single entity that a single nominal
base is used with the plural termination when many individual
substances of a kind are to be referred to. If it were not a single
61See Nagesa on 1.2.64 vdrttika 53 (translation, vdrttika 53b note 73).

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 65
entity many bases would arise and 1.2.64 would have to be stated
to provide that only one remains. The reason for adopting the
view that a generic term denotes a class property was to spare one
from having to state 1.2.64 (?1.3.2, vdrttika 35). This reason
would now fail (?1.4.2).
Secondly, it was accepted that a class property is
permanently associated with a certain gender. This was accepted
in order to account for the use of generic terms consistently in a
single gender. Generic terms such as 'vrksa' are always used in
the masculine, and generic terms such as 'brahmana' are
consistently used in the masculine when referring to both male
and female individuals of the class (?1.4.3).
Pataiijali revises Katyayana's statement by introducing the
element of a speaker's intention (vivaksa). The statement now
asserts that a speaker's intention of gender and number is not
innate. It is not necessarily the case that a speaker's intention
associates a class property permanently with a certain gender and
number. A speaker may intend the class property as having
various genders and numbers in various instances even if a certain
gender or singular number is innate. The use of the term in
different genders and numbers in different instances arises based
on the varied intentions of the speakers in different circumstances.
A generic term is used in the masculine if the speaker intends the
class property as having masculine gender. The same term is used
in the feminine if the speaker intends the class property as having
feminine gender, and it is used in the neuter if he intends neither.
Similarly, if a speaker intends the class property as having dual or
plural number, dual and plural terminations will arise naturally.
Hence, the use of generic terms in all three numbers is accounted
for, and the use of generic terms such as 'brahmana' in different
genders is accounted for (vdrttika 53a).
1.5.2. Is the Gender Problem Solved by the

Grammarians' Conception or by a Speaker's
Intention Alone?
It is questionable whether or not Patanijali means that th

solution including a speaker's intention as a factor accounts f
the use of a generic term in different genders by itself. After
solution that a speaker's intention of gender and number are
innate is introduced, it is objected to on the same grounds it
objected to before. "A generic property is gender specific; fr
origin to destruction it does not abandon the gender it starts
with (vdrttika 53a)." In answer to this objection, the proponen
the view that a generic term denotes a class property sa
"Therefore, grammarians cannot accept the ordinary gender

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
an object as the criterion for the gender of the words denoting

They have to adopt a conception of their own." Pataiijali th
gives a lengthy citation of the conclusion arrived at under 4.1
Striydm, concerning the grammarians' view of gender. T
objector accepts this response as a solution to the problem sayin
"In that case by saying, 'Gender and number succeed becaus
speaker's intention of a quality is not innate,' it may be possible
avoid the gender problem but it is not possible to avoid t
number problem."
Hence, the line of argument seems to indicate that th
problem of accounting for the use of a generic term in differe
genders is solved by introducing the grammarians' conception
gender and not by the introduction of a speaker's intention alon
However, a speaker's intention is the element in the grammaria
conception of gender which achieves the proper arrangeme
(vyavasthd) of speech forms in the appropriate genders. In ord
to resolve this question let us examine the problem of gender an
the arguments for the grammarians' conception of it under 4.1.
The word 'gender (linga)' refers both to the sex
characteristics of objects and to the characteristics which
determine the use of certain speech forms rather than others (as
well as to the characteristics of speech forms themselves). The
discussion begins by accepting the ordinary conception of gender
which identifies the first and second senses of the word then
adopts the grammarians' view which distinguishes them. Under
the ordinary conception, feminine gender is that by which one
ordinarily identifies something as female, namely, breasts, lon
hair or, more definitely, the female sex organ. Masculine gend
is that by which one ordinarily identifies something as ma
namely, body hair or the male sex organ. Neuter gender is that
by which one ordinarily identifies an object as neuter, namely, t
absence of the male or female characteristic marks.62
Patanjali raises a number of faults that arise accepting the
ordinary conception. Among them is that a bed (khatva) is
feminine and a tree (vrksa) is masculine, while neither a tree nor
bed has either the male or the female characteristic marks and
hence ought to be neuter.63 The only evidence of the presence
female gender in a bed is the feminine affix tap in the wor
'khatva.64 However, it is circular for a grammarian to infer th
feminine gender of the object from the use of the feminine aff
when he justifies the use of the feminine affix by the presence
62Concerning the ordinary conception of gender, see MB-K, vol. 2, p. 196, line
1-5.
63MB-K, vol. 2, p. 196, lines 13-20.

64'Khatvd' has the feminine suffix tap according to Panini 4.1.4. Ajadyatas t
The affix tap occurs after a base ending in a and specifically in the derivat
'aj', etc.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 67
the feminine gender.65 In addition, observing the occurrence of

masculine, feminine, and neuter forms of the word '(river-) bank
(tatah, tatt, tatam)', one would be unable to infer the gender of a
bank.66 If ordinary gender (the sex of the object) were the
criterion for the gender of a word, two words of different gender
could not refer to the same object nor could a word occur with
suffixes of different genders. Therefore grammarians adopt their
own criteria.
Note that this is the case for an individual object too. Any
one of the three forms 'tatah, tati, tatam' may be used to refer to
the same individual river bank. Hence the fault with the ordinary
conception of gender arises regardless of whether a generic term
denotes a class property or an individual substance. A speaker's
intention (vivaksa) must be included in accounting for the correct
use of both gender suffixes and number terminations as arising
from semantic conditions, because words of different genders and
numbers are used for the same individual substance. For
example, a man may use both of the words 'ddrdh (ma
plural)' and 'patni (feminine singular)' with reference to h
wife.
At this point the discussion under sutra 4.1.3 includ
description of the grammarians' theory of gender which P
cites under 1.2.64 (vdrttika 53b). According to this the
manifest entities both become more manifest, in which c
are said to have the quality of increase (prasava) linke
masculine gender (pumliiiga), and become less manifest, in
case they are said to have the quality of decrease (sam
linked with feminine gender (strllihga). The balance of in
and decrease is linked with neuter gender. The link is not
but depends on a speaker's intention.
If the link were direct, no object would be exclu
masculine or feminine because all manifest entities have both the
qualities of increase and decrease. It is a speaker's intention of
increase or decrease with regard to a certain object which
constitutes masculine and feminine gender, and the absence of the
intention of either which constitutes neuter gender. This intention
of increase or decrease, or the lack of this intention, is expressed
by masculine, feminine, and neuter speech forms.
Kaiyata makes clear that the intention of a speaker is not
individual whim commenting on the phrase, "from a speaker's
intention (vivaksdtah)," at the close of the passage cited from the
bhdsya on sutra 4.1.3 (vdrttika 53b). He says, "The intention of a
65MB-K, vol. 2, p. 197, lines 16-18.

66MB-K, vol. 2, p. 197, lines 22-24. See Amarakosa 1.9.7, p. 95b. The feminine
suffix his occurs after the base 'tata' according to 4.1.63 (?1.2.5).

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
speaker is resorted to as an echo of ordinary usage, not

individually."67 In that context Patanjali continues as follows:
Tasyoktau lokato ndma

Tasyoktau ca vacane lokato ndmaitad bhavati stri pumdn
napumzsakam iti.... Na caitan mantavyam svamantsikayocyata iti.
Pathisyati hy dcdryo lingam asisyam lokdsrayatvdl lingasyeti. Punah
pathisyaty ekdrthe sabddnyatvdd drs.tam lihgdnyatvam avayavdnya-
tvac ceti.68
In expressing it one follows ordinary usage69
In expressing conveying it (an object's gender) this (arrangement):
feminine, masculine, neuter, follows ordinary usage.... And
one should not think that (this arrangement) is stated according to
(the grammarians') own theory because the teacher will declare,
"Gender should not be taught because gender is based on ordinary
usage."70 Then he will declare, "One sees that there are different
genders in one object because different words are used for it. And
because different parts (of words) are used for it."71
The ordinary usage of the learned shows the gender an object is

intended to have.
Although Pataiijali describes objective characteristics to
constitute gender as a characteristic of objects, the universality of
the characteristics makes them inadequate to account for a
division of objects into categories of masculine, feminine, and
neuter directly. In fact no objective characteristics of objects can
account for such a division because the usage of some words
attribute different genders to the same objects at once. The most
unambiguous example Pataijali gives is 'tatah, tati, tatam',
masculine, feminine, and neuter for a bank. Commenting on the
last two vdrttikas referred to in the above passage (in quotes),
under sutra 4.1.92 Patanijali gives two examples in which words
67Vivaksdta iti. Lokavyavahdrdnuvddini vivaksdasryate na tu prdyoktri. MB,

vol. 4, p. 25a.
68MB-K, vol. 2, p. 198, lines 11-12, 15-17.
69This is the third pdda of the slokavdrttika: Samstyane stydyater drat stri
suteh sap prasave pumdn; tasyoktau lokato nama guno vd lupi yuktavat. MB-
K, vol. 2, p. 198. Concerning the first line see 1.2.64 translation, vdrttika 53b.
70Nagesa comments that 'teacher' here refers to the author of the vdrttika:
Acarya iti. Vartikakrd ity arthah. Bhargavasastrin (1945: 3) argues that
Katyayana is the author of the varttika, lingam asisyam lokdsrayatvdl lingasya,
while someone other than he and Patafijali authored the slokavdrttika because
Pataijali cites the former in support of the latter. Patafijali cites the vdrttika at
several other points: MB-K, vol. 1, p. 390, lines 18-19; p. 435, lines 2-3; p. 477,
line 4; vol. 2, p. 418, lines 24-25; p. 438, lines 2, 5-6, 12-13, 18; vol. 3, p. 371,
line 7. Astadhyayl 1.2.53, added after Panini but before Patanijali, expresses the
same principle: Tad ai.syam samjnapramanatvat (vyaktivacane 51), "(Gender
and number) should not be taught because the means of knowing them is the
convention (of speakers)."
71pPnini 4.1.92, vdrttikas 6 and 7. See MB-K, vol. 2, p. 246, lines 21-27.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 69
of three different genders are used for the same object, and three
examples in which masculine and feminine forms of the same
word denote the same object. Hence gender is not exclusively an
objective characteristic of objects but has a subjective component:
the intention of a speaker. If a speaker intends an object as
having a certain gender characteristic, he uses a speech form
which conveys that characteristic, that is, one with a feminine
suffix (if appropriate) and with the appropriate inflectional
terminations.
This does not imply that an individual speaker can convey
his intention of a certain object as having a certain gender
characteristic by combining any base denoting the object with any
suffix signifying the intended gender. On the contrary, just as
ordinary usage follows certain patterns regarding the primary
denotation of words (for example, the word 'cow' means a cow
and not a horse) similarly it follows certain patterns regarding
gender. The ordinary usage of the learned is such that certain
nominal bases, such as those in 'khatva' and 'vrksah', are
coextensive with a certain gender characteristic (feminine and
masculine respectively) whenever they denote certain objects (a
bed and a tree). Other nominal bases, such as 'tata (bank)' are not.
A speaker is free to convey a bank as having any of the three
gender characteristics with the nominal base 'tata' by adding the
feminine suffix his (if he intends feminine gender) and the
appropriate nominal inflectional termination.
It is evident that by including the subjective element of a
speaker's intention (to which ordinary usage is the guide) the
advantage of linking the arrangement of gender to the growth
and waning of qualities rather than to ordinary male and female
characteristics vanishes. One can account for the use of the word
'vrksah', which is masculine, for a tree, which has neither male
nor female characteristics, by saying that ordinary usage shows
that speakers intend trees as having male characteristics, just as
well as by saying that it shows that they intend trees as having
growth. The only difference is that under the conception of
gender as the intention of male and female characteristics (or the
lack of either) one superimposes a characteristic which is absent
on an object, while under the grammarians' conception of gender
one highlights one of two characteristics both of which are
already present (or ignores them in the case of neuter gender).72
72This difference is negligible because, as Kaiyata suggests and Nagega

explicitly points out, even under the grammarians' conception one must resort
to superimposition. For example, in the cases of the self and the Lord which do
not have the characteristics of increase and decrease, one uses the words 'dtman'
and 'isvara', which are masculine, by superimposing the increase present in the
manifest elements. (See translation, varttika 53b note 79). Furthermore, in
other contexts Patafijali explicitly states that there can be the lack of intention
of what is present as well as the intention of what is absent (sato 'py avivaksa

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
If this is so, the ordinary conception of gender with the addit

of a speaker's intention can account for usage just as well as t
grammarians' conception of gender.
If Pataiijali's presentation of the grammarians' concepti
of gender under 1.2.64 is intended as solving the gender
problem, it must be because the inclusion of a speaker's intention
as a factor did not solve it. This would indicate a limitation on the
application of a speaker's intention. It highlights a property
already present in the object but it cannot superimpose a property
which is absent. One does not make a horse a cow just by a
speaker's intention that it is so. Similarly, one cannot make an
object have a gender that is not present in it by a speaker's
intention. In other words, gender is still considered to be an
objective characteristic of objects to some extent under the
grammarians' conception. All objects have the properties of
increase, decrease or balance. The masculine, feminine or neuter
speech form a speaker uses conveys which he intends.
On the other hand, if Patafjali's presentation of the
grammarians' conception of gender is not intended as a required
factor in solving the gender problem, it is a grammatically
irrelevant theory concerning the subtle composition of things
about which we have no knowledge. In this case, a speaker's
intention is the factor which solves the gender problem by itself.
The use of a generic term in different genders succeeds on
the view that a generic term denotes a generic property because a
speaker may intend a generic property as having one gender in
one instance and another in another instance. In the case of terms
such as 'brahmana' which occur in different genders, the different
intention is the condition for the use or lack of use of a feminine
suffix. If the speaker intends the generic property brahmanatva
as having feminine gender, the suffix his arises according to
4.1.63 (?1.2.5). Otherwise no feminine suffix arises.
The use of a generic term in different genders contradicts
the assertion, "A generic property is gender specific; from origin
to destruction it does not abandon the gender it starts out with
(vdrttika 53a)." The introduction of a speaker's intention as a
factor allows a class property to be permanently associated with a
single gender in fact and yet have words denoting it occur in
different genders. This accounts for the use of different generic
terms for the same class property in different genders. For
example, even if the generic property of being a hut is
permanently associated with a single gender, the use of the
generic terms 'geham' in the neuter, 'kutt' in the feminine, and
'mathah'in the masculine is accounted for because the gender of
bhavati... asatas ca vivaksa bhavati....). MB-K, vol. 2, p. 118, lines 12-14; vol. 3,
p. 157, lines 22-24. For a discussion on these and other passages concerning
vivaksa see Van Nooten 1983 and Scharf 1995.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 71
the generic property is not a direct condition for the use of a

speech form in a particular gender; a speaker's intention is.73 The
speaker's intention is not permanently associated with a certain
gender even if the class property is.
However, the use of a generic term in different genders
contradicts the assertion, "A generic property is gender specific,
etc," even including a speaker's intention as mediating the
connection of a generic property with a gender. Under 1.2.52
Patanijali states, "A generic property is gender specific, etc," to
explain why generic terms such as 'vrksa' consistently occur in the
masculine gender (?1.5.4 note 79). This explanation applies
equally to explain why the generic term 'geham' consistently
occurs in the neuter, 'kutf' consistently occurs in the feminine, and
'mathah' consistently occurs in the masculine. These three terms
all denote the same generic property. That generic property
cannot in actual fact be associated with any single gender. Yet the
statement that a generic property is gender specific is meant to
explain why each of these terms is consistently used in a single
gender. In order that the statement apply to them it cannot speak
of a generic property as an object independent of the word used
to denote it. It must speak of a generic property as denoted by a
particular word. The real meaning of the saying, "A generic
property is gender specific (avistalitiga jdtih)," is that a generic
property as denoted by a certain word is connected with a specific
gender.74 It does not change gender according to the gender of
the individual substance in which it is present as qualities do.
If this is meant to be true of all generic properties, that a
generic property as denoted by a certain word is gender specific,
then it would also be true in the case of the generic property
brdhmanatva as denoted by the generic term 'brahmana'. The
generic term 'brdhmana' would consistently occur in the
masculine gender, and the form 'brdhmani', in the feminine,
would never occur. However, it does occur in the feminine.
Therefore, it is not true of all generic properties that a generic
property as denoted by a certain word is gender specific. The
statement, "A generic property is gender specific; from origin to
destruction it does not abandon the gender it starts out with
(?1.5.4 note 78)," under 1.2.52 can only mean that a generic
73Patafijali gives the example under 4.1.92 vdrttika 6 which Patanjali cites
under 4.1.3 in the long passage quoted above in this section. MB-K, vol. 2, p.
246, line 23.
74Kaiyata: Avistalinigeti. Avistam lingam yayd sdvistalinig. Niyatalitigety
arthah. Sarvatra sarvesdm lingdndm sadbhdve 'pi kenacic chabdena kincil
lingam pratydyyata iti sabda-visesapeksaydvistalingatvam ucyate. MB, vol. 2,
p. 64a. Nagesa reiterates Kaiyata's remark under 1.2.64. See 1.2.64 translation,
vdrttika 53b note 73.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
property as denoted by a certain word is gender specific as a

fairly regular rule of thumb, not universally.
In summary, different genders succeed on the view that a
generic term denotes a generic property because a speaker may
intend a generic property as having one gender in one instance
and as having another gender in another instance. These different
intentions are the conditions for the use of a generic term in
different genders. Although introducing a speaker's intention as a
factor achieves the use of a generic term in different genders on
the view that a generic term denotes a class property, it does not
achieve the use of a generic term in different numbers.
1.5.3. Number
The solution stating that a speaker's intention of ge

and number are not innate does not achieve the use of a g
term in different numbers because, if a generic prop
intended as having a number other than one, more th
nominal base would arise just as the dual or plural term
arises. It is unable to account for the use of a generic term
dual and plural. A contradiction arises between the int
which determines the number of nominal bases used and the
intention which determines the choice of singular, dual or plu
terminations. A single intention determines both in any instanc
This single intention is the condition for the introduction of b
the nominal base and the inflectional terminations.
There is no problem if a speaker intends a class property
as having singular number. A single nominal base arises and the
singular termination arises. However, if a speaker intends a class
property as having dual or plural number, just as dual and plural
terminations arise, two or many nominal bases would arise as
well. The rule would have to be stated that just one of
homophonous speech forms remains if there is one termination.
And the reason for adopting the view that a generic term denotes
a class property, namely, to spare one from having to state 1.2.64,
would now fail (vdrttika 53a).
Panini 1.4.21-22 state which inflectional terminations
occur for which numbers. It is natural to consider that the
number of the object denoted by the nominal base is the
which is the condition for use of the inflectional terminations. On
the view that a generic term denotes an individual substance, the
number which is the condition for use of the inflectional
terminations is a property of the individual substances de
the base. On this view it has to be explicitly stated that
terminations occur if there is singular number, others if
dual number, and a third set if there is plural number.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 73
On the view that a generic term denotes a class property,

suppose that the number belonging to the individual substances is
accepted as the condition for use of the inflectional terminations,
even though the individual substances are not the object denoted
by the nominal base. One would state rules to achieve the
occurrence of terminations according to the number of individual
substances even though these are not denoted by the nominal
base. Since 1.4.21-22 have to be explicitly stated on the view that
a generic term denotes an individual substance, the same may be
stated on the view that a generic term denotes a generic property.
Hence stating the rules which bring about the use of the proper
terminations in the proper numbers would not constitute an
additional effort required of the view that a generic term denotes
a generic property.
However, on this supposition the problem is that the
number which is the condition for use of the terminations does
not belong to the object denoted by the nominal base after which
they occur.75 Therefore, stating that "a speaker's intention of a
quality is not innate" is not sufficient to account for the use of a
generic term in different genders and numbers on the view that a
generic term denotes a generic property. With this in mind,
Katyayana states vdrttika 54 to explain how gender and number
suffixes occur for generic terms on this view.
1.5.4. Agreement in Gender and Number76
In vdrttika 54, Gunavacanavad vd, Katyayana says that

gender and number terminations occur for generic terms as they
do for quality words. In explanation, Patanijali cites the
established principle of gender and number agreement for quality
75Mimamsa proponents of the view that a generic term denotes just a class
property argue that this is not a problem. It is not necessary that gender and
number suffixes arise to denote gender and number belonging to the object
denoted by the base. They may belong to another object. In the case of verbs
(tinanta), grammarians accept that the verbal root (dhatu) denotes action. The
verbal suffix (tin) denotes the substratum of the action, that is, either the agent
(kartr) or object (karman), and number. Action is a single entity; it naturally has
singular number. The number of the substratum of the action determines which
set of verbal terminations is used. Hence in this case, the property which
determines which set of terminations is used does not belong to the object
denoted by the base (the verbal root) after which they occur. Just as in the case
of verbs the number of the substratum of the action denoted by the verbal root
determines which set of verbal terminations occur, in the case of generic terms
the gender and number of the substratum of the class property denoted by the
nominal base determine whether feminine suffixes occur and which set of
nominal terminations occurs.
76For an introduction to the question of gender and number agreement see

Cardona (1988: 214-17, especially 215).

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
words. "Different genders and numbers occur for words denotin

qualities according to the substrata in which the quality reside
For example, in "suklam vastram (a white cloth), sukla sadt
white sdtl' (a woman's garment)), suklah kambalah (a whit
blanket)," the quality word 'sukla' occurs in the neuter to refer
the quality white residing in a neuter substance, in the feminine
refer to the quality white residing in a feminine substance, and
the masculine singular to refer to the quality white residing in
masculine substance. It occurs in the singular in each of the
three cases to refer to the quality in a single individual substanc
In "suklau kambalau (two white blankets), suklah kambalah
(white blankets)," it occurs in the dual to refer to the quality wh
in two individual substances and in the plural to refer to it
many individual substances (vdrttika 54).
Patafijali's citation of the principle, "Different genders an
numbers occur for words denoting qualities according to t
substrata in which the quality resides," roughly quotes Katyayan
4.1.3, vdrttika 6, "Because gender and number occur for a wor
denoting a quality according to the substrata in which the quali
resides."77 According to Patafijali, Panini's purpose in stati
1.2.52, Visesandndm cdjdteh, is to make known this principle.78
774.1.3, vdrttika 6, Gunavacanasya casrayato lingavacanabhdvdt. MB-K, vol.

p. 200, lines 1-6. The following passages also cite the principle as Katyay
states it: MB-K, vol. 1, p. 410, lines 18-24; p. 422, lines 17-22; p. 430, lines 1
19; vol. 2, p. 414, line 25 - p. 415, line 1.
78Panini 1.2.51, Lupi yuktavad vyaktivacane, provides that when a suffix
deleted (replaced by lup) the object denoted by the derivate is considered t
have the gender and number of the object denoted by the pre-suffixal base. T
result is that gender and number suffixes occur for the derivate as they do f
the base. 1.2.52, Visesandndm cdjdteh, provides that qualifiers of the objec
denoted by the derivate are also considered to have the gender and number
the object denoted by the pre-suffixal base (viiesantndm ca). This extensio
does not apply to qualifiers which are generic properties or to qualifiers tak
to qualify the generic property directly rather than the object denoted by
derivate (ajdteh or djdteh). For example, the word 'panicdldh' originally refers
a certain people. According to 4.2.67, Tadasminnastti des'e tanndmni, the suff
an prescribed by 4.1.83, Prdg divyato 'n, occurs after it if the derivate, mean
that in which the Paficalah people live, is the name of a place. If that place
district (janapada), 4.2.81, Janapade lup, deletes the suffix an leaving no tra
of its occurrence. The derivate 'pancaldh' formed in this manner is the name o
district. By 1.2.51, it takes the masculine plural terminations as does 'panicdld
referring to the people. In the phrase, "Pancdldh subhiksdh (The Pancal
plentiful in food)," the word 'subhiksdh (plentiful in food)' denoting a proper
qualifying the Paficalah place also takes the masculine plural in accordance wit
1.2.52. In contrast, in the phrase, "Pancald janapadah subhiksah," the gene
term for a district, 'janapadah', does not take the plural termination, nor does
word 'subhiksah', because the property of being plentiful in food is understoo
to qualify the place as a generic district.
According to Katyayana, 1.2.52 constitutes a restrictive rule (niyama) rath
than a provision (vidhi). Its purpose is to prevent the gender and number of t
object denoted by the original word from applying to a generic proper

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 75
Although this is not Patafijali's final view on the subject,

there is a view that a quality word denotes just a quality.79 This
view is parallel to the view that a generic term denotes just a
generic property. Gender and number are properties of the
individual substances in which a quality resides.80 It is natural to
assume that the semantic conditions for the occurrence of gender
suffixes and number terminations after a nominal base are the
gender and number which belong to the object denoted by the
nominal base. However, on the view that a quality word denotes
just a quality, this is not so. It is the gender and number of an
object other than that denoted by the nominal base. The gender
and number terminations arise after a nominal base denoting a
quality according to the gender and number of the individual
substances in which the quality resides.
This fact is accounted for by extending the properties of
the substratum to the quality which resides in it. The gender and
number which properly belong to the individual substances in
qualifying the object denoted by the derivate. (1.2.52. vdrttika 1. Visesananam

vacanam jdtinivrttyartham. MB-K, vol. 1, p. 228.) Suffixes naturally occur for
words denoting qualifiers in accordance with the gender and number of the
object denoted by the original word because both words refer to the same object.
(1.2.52. vdrttika 2. Samdnddhikaranatvdt siddham. MB-K, vol. 1, p. 228.) In
the example, the word 'subhiksa' takes the plural, as the word 'paicdldh' does,
naturally because they both refer to the same object. The suftra does not serve to
bring this about. It serves to prevent 'janapada' from taking the plural.
Patanijali notes that if gender and number suffixes for words denoting qualifiers
naturally agree with those for words denoting the objects qualified, and if the
sutra were stated just to prevent such agreement for a word denoting a generic
property qualifying an object denoted by a derivate formed by replacing an
affix by lup, it would not have to be stated. Even when an affix is not replaced
by lup a generic term does not agree in gender and number. For example, in the
phrase, "Badari suiksmakantakd madhurd vrksah," the word 'badari' (feminine
singular) denotes a species of tree. The words 'suksmakantakd' and 'madhurd',
denoting the qualifiers having small thorns and sweet, occur in the feminine.
But the word 'v,rksah' denoting the generic property occurs in the masculine.
The word 'badari' is not derived according to 1.2.51 and involves no deletion.
Since no suitra states an exception for generic terms occurring with 'badarT', there
is no need for 1.2.52 to state an exception for generic terms occurring with
derivates such as 'paicdldh' formed by 1.2.51. Generic properties in general do
not adopt the gender and number of the objects they qualify. They keep their
own gender and number consistently.
Patanjali concludes that the reason 1.2.52 is stated is to inform us of the general
principle of gender and number agreement for words denoting qualities: A word
denoting a quality takes the gender and number of the substratum in which the
quality resides. "Na tarhiddnim ayam yogo vaktavyah. Vaktavyas ca. Kim
prayojanam? Idam tatra tatrocyate, gunavacananam sabddndm dsrayato
litgavacandni bhavantiti. Tad anena kriyate." MB-K, vol. 1, p. 228, lines 19-21.
79Sabara expresses this view commenting on Jaimini 3.1.12. See ?3.6, note 58.
80For example, under 4.1.3, vdrttika 7, Bhdvasya ca bhdvayuktatvdt, Pataijali
states that feminine gender is a property of an individual substance: Kasya
tdvad bhavdn evam gunam nydyyam manyate stritvam ndma? Dravyasya. MB-
K, vol. 2, p. 200, lines 12-13.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
which a quality resides are considered to belong to the quality

too. Considering them to belong to the quality, the gender and
number terminations naturally arise after the word denoting the
quality. Hence, even though the gender and number which ar
the semantic conditions for a quality word's gender suffixes and
number terminations belong to something other than the quality,
by extension they belong to the quality so that the occurrence of
the terminations after the quality word is achieved naturally.
5.2.94, vdrttika 5, Tatha ca litigavacanasiddhih, concludes
how different genders and numbers occur for words denotin
qualities. Commenting on it Patafijali says, "The gender and
number of the substance in which the quality resides occur for th
quality too."81 Different genders and numbers occur for words
denoting qualities according to the substrata in which the quality
resides because the gender and number of the substance in which
the quality resides belong to the quality too.
Similarly, different genders and numbers will occur for
generic terms according to the substrata in which the generi
property resides because the gender and number of the substance
in which the generic property resides belong to the generic
property too.
1.2.58 vdrttika 3, Arthdtidesdt siddham, interprets suitra
1.2.58 to be this sort of extension. According to vdrttika 3, the
number which belongs to many objects is considered to belong t
a single generic property which resides in them (1.2.58, note 5).
On this interpretation, 1.2.58 does not extend the use of the plura
termination by explicit statement to occur where a singular
termination naturally arises. Rather, it extends the meaning
condition on which the plural termination naturally arises. The
sutra allows that a generic property, although a single object, wil
be considered to have the plural number which properly belongs
to the many objects in which it resides. Thus the plural
termination which naturally arises to denote the plural number,
will naturally arise to denote the plural number attributed to a
single generic property.
In vdrttika 55, Adhikaranagatih sdhacarydt, Katyayana
states the reason that the gender and number which properly
belong to the substance in which a property resides belong to the
property too. Cognition of the property's substratum arises
because it accompanies the property. Because cognition of an
individual substance in which the class property resides arises, the
properties of the substance, which by extension belong to the
class property too, are among the semantic conditions for use of
the generic term. It is by virtue of being included among the
81Yad asau dravyam srito bhavati gunas tasya yal lingam vacanam ca tad
gunasydpi bhavati. MB-K, vol. 2, p. 394, lines 14-15.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 77
semantic conditions for use of the generic term that they

determine its gender and number suffixes.
1.6. Patanijali's Solution
Katyayana's argument in vdrttikas 53-59 presents

solutions to the objections raised in vdrttikas 45-52 and establishes
the view that a generic term denotes a class property alone. In
general, Patanjali's comments on these vdrttikas elaborate the
argument as Katyayana presents it. However, under vdrttika 53
Pataiijali presents a lengthy discussion on gender and number
ending with an independent conclusion concerning the denotation
of generic terms.
The manner in which he concluded his commentary on
vdrttika 43 foreshadows his conclusion. There he showed that
there is equal fault on both the view that a generic term denotes a
generic property alone and the view that it denotes an individual
substance alone. I have discussed the equivalence of the two
views above (?1.4). He concludes his commentary on vdrttika 53
by stating that neither view can completely exclude the other. He
says the following:
It is not the case that an individual substance is not denoted for him
for whom a class property is denoted, nor that a class property is not
denoted for him for whom an individual substance is denoted. Both
are denoted for both. But for each something is principal, something
subordinate. For him for whom a class property is the object denoted
the class property is principal and an individual substance is
subordinate. For him for whom an individual substance is the object
denoted an individual substance is principal and its class property is
subordinate (varttika 53c).
The view that a generic term denotes a class property must accept
an individual substance as part of the meaning. Conversely, the
view that a generic term denotes an individual substance must
accept a generic property as part of the meaning. Even on the
view that a generic term denotes a class property, the properties of
the individual substances in which the generic property resides
constitute meaning conditions determining the speech form used.
It is an individual substance's gender, number, difference, and
amenability to birth and destruction which condition the use of
certain suffixes, analytic strings, and allow the use of the term in
syntactic connection with verbs.
Katyayana's defense of the view that a generic term
denotes just a class property solves the problem of different
gender and number terminations by saying that the substratum of
the class property is the determining factor. The substratum of a

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
class property is an individual substance. To admit that an

individual substance is included in the semantic condition for the
use of a speech form is as much as to admit that an individual
substance is included in the meaning.
If one accepts that the gender and number suffixes arise
naturally after a nominal base, conditioned by the gender and
number belonging to the object denoted by the nominal base,
then one must accept that the object to which the gender and
number belong is included in that which is denoted by the
nominal base. Gender and number belong to individual
substances. Therefore, the nominal base of a generic term
denotes individual substances.
Each of the reasons for accepting each of the extreme
views that a generic term denotes a generic property alone and
that it denotes an individual substance alone is a reason for
including the generic property or individual substance in
denoted meaning. If individual substances are included in
denoted meaning, the appropriate number naturally belongs
however many individual substances are to be denoted. Singu
dual, and plural inflectional terminations naturally arise
Similarly, the different genders of different individual substa
of a class are the conditions for the use of a generic term
different genders. The fact that distinct identities belong to
multiplicity of objects, not to a single object, and that multi
instances of a generic term in analytic strings are used to refe
such distinct objects is reason for including individual substa
in the meaning denoted by a generic term. Likewise differen
in quality are due to a multiplicity of objects. If a generic ter
at once syntactically connected with terms for inconsis
qualities, the usage is accounted for by accepting that individ
substances having the different qualities are denoted. The
that particular individuals, not the class as a whole, participat
action in any given instance, and that generic terms are used
refer to such individuals in general statements and injunctio
reason for including an individual in the meaning denoted by
generic term.
Conversely, including a class property in the meaning o
generic term accounts for the use of a single nominal base ev
when referring to a plurality of individual substances. Natur
cognition of the class property arises if the term denotes it. A
this accounts for the universal application of the term. The w
bears a relation with the class property in an arbitrary indivi
substance. Since this relation is consistent, learned once, it is
known.
In any particular usage, either the class property or an
individual substance is principal in the meaning and the other
subordinate. The intention of the speaker determines which. If
the speaker intends the class property as principal, the term will be

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 79
used in the singular but will apply universally to any individual of

the kind. The term will occur in the gender intended as associated
with the class property. Because an individual substance is
included in the meaning as subordinate, action which is
impossible on the class property will take place on the individual
substance.
If the speaker intends an individual substance as principal,
the generic term will be used with gender and number suffixes
according to the gender and number of the individual substances
denoted, and will allow action to apply to particular individuals.
Because the class property is included as subordinate, a single
base will be used even to denote many individuals.
1.2.58, varttika 7, as explained by Patafijali, explicitly
states the conclusion that a generic term denotes both a generic
property and an individual substance. Katyayana says, "Because a
generic term denotes a substance." Pataiijali comments, "Because
a generic term denotes a substance too, as well as a generic
property." Under 5.1.119, Tasya bhdvas tvatalau, Katyayana
refers to the fact that a generic term denotes an individual
substance. In vdrttika 5, Siddham tu yasya gunasya bhdvdd
dravye sabdanivesas tadabhidhdne tvatalau, he says that the
suffixes tva and tal are used in order to denote the property on
account of which a word denotes an individual substance.82 In
stating the condition for the use of tva and tal Katyayana makes
known that the word after which these suffixes are provided
denotes a substance due to the fact that the substance has a certain
property. Patanijali comments that generic terms denote both the
substance and the property and that Katyayana's varttika applies
to generic terms by virtue of their denoting the substance. He
says, "These (generic terms 'tree', etc.), denoting both, signify
both an individual substance and a quality. By virtue of the fact
that they denote a substance, (the suffixes tva and tal) occur
(Ubhayavacand hy ete dravyam cdhur gunam ca yato
dravyavacands tato v.rttir bhavisyati)."83
1.7. Quality Words
1.2.64, varttika 54 made a comparison between generic

terms and quality words. It suggested that a generic term denotes
only a generic property just as a quality word denotes only a
quality. Similarly, Patanijali's conclusion regarding the denotation
of generic terms is parallel to his conclusion regarding the
denotation of quality words. He makes an explicit comparison
82MB-K, vol. 2, p. 366, line 10.

83MB-K, vol. 2, p. 367, line 16.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
between the two types of words in the same passage from hi

commentary on 5.1.119 just cited. He says, "Then these (quality
words 'white', etc.) denote both too (Ime 'pi tarhy
ubhayavacandh)."84 A quality word denotes both a quality and
an individual substance just as a generic term denotes both a
generic property and an individual substance. He accepts that
quality words such as 'sukla' have two meanings: 1) the quality
itself, and 2) a substance having that quality.
In his very next statement, Pataiijali accepts the provision
and deletion of matup as a way of making the quality word
denote a substance.85 Similarly, under 2.1.30, Trtlyd
tatkrtdrthena gunavacanena, Pataijali discusses how a word
denoting a quality comes to denote a substance in which the
quality resides. In that context too, he affirms the derivation of a
quality word denoting an individual substance in which the
quality resides by the provision and deletion of matup.86
These two passages cite a derivation Katyayana presents
under 5.2.94, Tad asydstyasminniti matup, to show how a quality
word denoting just a quality comes to denote a substance having
that quality. The original quality word denotes just a quality.
According to 5.2.94, the affix matup occurs after a word to form
a derivate denoting that to which the object denoted by the base is
related or that in which it resides. A quality resides in the
substance it qualifies. Hence matup would arise after a quality
word if the derivate denotes a substance in which the quality
resides. For example, matup would arise after the word 'suklam'
denoting the color white, if a white cloth is to be denoted.87
However, the correct form used to refer to a white cloth, for
example in "suklam vastram," does not contain the suffix matup.
To account for this, Katyayana explicitly states that the suffix
matup is deleted without leaving a trace (replaced by luk).88 The
result is a nominal base denoting a substance identical in form to
the original nominal base denoting a quality.
5.2.94, varttika 4, Avyatirekdt siddham iti ced drsto
vyatirekah, recognizes and responds to an objection against the
derivation of a quality word denoting a substance by the
provision and deletion of matup. The objection is that matup
841me 'pi tarhy uhhayavacandh. MB-K, vol. 2, p. 367, line 17.

85Katham? Arabhyate matublopo gunavacanebhyo matupo lug bhavat?ti. yato
dravyavacands tato v.rttir bhavisyati. MB-K, vol. 2, p. 367, lines 17-18.
86Katham punar ayam gunavacanah san dravyavacanah sampadyate? Arabhyate
tatra matublopo gunavacanebhyo matupo lug iti. MB-K, vol. 1, p. 385, lines 9-
10.
87The nominal termination am after the base 'sukla' would be deleted by 2.4.71
(?1.3.1.3 note 45)
885.2.94, vdrttika 3. Gunavacanebhyo matupo luk. MB-K, vol. 2, p. 394. See
Cardona 1967-68: 327-31.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 81
does not arise in the first place because one identifies the quality
with the substance in which it inheres. Pataijali says, "It is
achieved (that a quality word denotes a substance in which the
quality resides even without the provision and deletion of matup)
because there is no distinction. A quality is never found without a
substratum. (Avyatirekdt siddham. Na guno guninam
vyabhicarati.)"89
Katyayana dismisses this objection because we do find a
quality word used to refer to a quality as something different
from the substance in which it inheres (drsto vyatirekah).90 For
example, in the phrase, "patasya suklah (the white of the cloth),"
the quality white is understood to be an object different from the
cloth and entering into relation with the cloth based on that
difference. Because the quality is understood as different from
the cloth, it is not identified with it so the provision and deletion
of matup is required to make the speech form which naturally
denotes a quality denote a substratum in which the quality resides.
Under 5.1.119, Kaiyata states that the provision and
deletion of matup are a way of making known that a quality
word denotes the substratum in which the quality resides due
merely to the identity of the quality and its substratum. It does
not mean that based on the difference of the quality from its
substratum, and its relation to it, a suffix arises and is deleted.91
He upholds the objection to the provision and deletion of matup.
If we observe usage of a quality word, we see that it is not
the case that there is concomitant presence (anvaya) and absence
(vyatireka) of a quality word and a substance. One does observe
a quality word used without the cognition of a substance. In
Patafijali's example, "patasya suklah," the word 'sukla' gives
cognition only of the quality white, not of that in which white
resides. On the other hand, if we observe usage of a generic term,
we see that there is concomitant presence and absence of a generic
term and a substance. One does not observe any case in which a
89MB-K, vol. 2, p. 394, lines 8-9.

90MB-K, vol. 2, p. 394. See lines 7-15.
91Abhinnarupdah. Suklddayo matubluki vijniyamdne yadyapi taddhitdntds
tathdpi tebhyo bhdvapratyayo guna eva bhavati na tu sambandhe. Gunaguninor
jdtitadvator iva so 'yam ity abhisambandhena lokanirtudhena
bhedasambandhasya nyagbhdvdt 'gunavacanebhyo luk' iti lugvidhdna-
dvdrendbhedenaiva gunino 'bhidhdnam na tu bhedam k,rtvd matvarthotpattir ity
dkhydyate. MB, vol. 4, p. 296a. Nagesa comments at the conclusion of 5.2.94
vdrttika 5 that vdrttika 3 and the derivation by the provision and deletion of
matup is rejected as unnecessary: Bhdsye: Yad asau dravyam ityddind:
matublope 'pi liingdisiddhir drsyate, tena pratydkhydnasyaiva yuktatvam
dariyate, anyathdsya granthasydtrdsangatir ity dhuh. Although the success of
gender, etc. is seen if matup is deleted, the statement, "The gender and number
of the substance, etc.," in the bhdsya shows that the only proper view is to reject
vdrttika 3 as unnecessary, otherwise this statement would not make sense here.
MB, vol. 4, p. 350a. See note 93.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
generic term is used without cognition of a substance in which the

generic property resides. There is no example of the usage of a
generic term parallel to the example, "patasya suklah." One must
employ the suffix tva in the phrase, "gor gotvam," in order to say
"the cowness of the cow."
The usage of a quality word and a generic term reveal the
conception speakers have of the objects they are used to refer to.
The sixth triplet termination signifying a relation in the word
'pata' in the phrase, "patasya suklah (the white of the cloth),"
shows that the object denoted by the quality word 'suklah (white)'
is related to a qualified object, denoted by the word 'pata', from
which it is distinct. In this usage, a quality is conceived as
something different from the substance in which it resides. On
the other hand, in the phrase, "suklah patah (a white cloth)," a
speech form indicating a relation is absent, and there is agreement
of gender and number terminations. The words are coreferential.
Cognition of the quality white and the substance in which it
resides arises from the word 'suklah'. In this usage, a quality is
conceived as identical to the substance in which it resides.
In the case of generic terms, examples such as "patasya
suklah" are lacking. The fact that there is no such usage shows
that a generic term is not used to refer to a generic property as
something different from the substance in which it resides. This
indicates that a generic property is never conceived as something
different from the substance in which it resides. On the contrary,
it is always conceived as identical to the substance in which it
resides.92
When Patafijali proposes that a generic term denotes both
a class property and an individual substance, he has in mind that it
denotes them as identical. Similarly, when he says that a quality
word denotes both a quality and an individual substance in which
it resides, he intends them as identical. He does not intend that the
word denotes two independent objects.93 In the beginning of the
92Under 1.2.64, varttika 54, Nagesa sums up the difference between quality
words and generic terms as follows: lydms tu visesah, suklddisabdavdcya-
gundndm gunind bheddbhedavivakse sabdavdcyajdtes tv abhedavivaksaiveti.
MB, vol. 2, p. 99b. See also varttika 54 note 84 for Kaiyata's comment.
93In the passage quoted in note 91, Nagega points out that Patanjali does not
accept the provision and deletion of matup as the final conclusion regarding
the reason quality words refer to a substance having that quality. The derivate
'sukla' formed by the provision and deletion of matup would not denote a
quality; it would denote a substance in which the quality resides. Gender and
number suffixes would occur naturally according to its denoted object, the
substance. If this solution were accepted as the final one, Patanijali's concluding
statement would not make sense. He says, "The gender and number of the
substance in which the quality resides occur for the quality too." There is no
need to attribute the gender and number of the substance to the quality which
resides in it, if the correct gender and number suffixes occur for the word 'sukla'
because it is a derivate denoting the substance directly. Conversely, the fact

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 83
Paspasahnika in his example of the experience of a cow, Patanijali

illustrates how human conception distinguishes the various
elements of an experience which constitutes a single
undifferentiated whole (?1.1). He shows how human conception
distinguishes a substance, a quality, a class property, etc. When
he says that a generic term denotes both a generic property and a
substance, or that a quality word denotes both a quality and a
substance, he means that in this case human conception does not
distinguish the two elements. This fact makes the objections
which Sabara and Kumarila raise against the view that a generic
term denotes both a class property and an individual substance
vacuous.
Presenting the view that a generic term denotes j

property, Katyayana says in varttika 55, Adhikar
sahacaryat, "Cognition of a substratum arises because it
that Patafijali concludes that a quality adopts the gender and number of the
substance in which it resides implies that gender and number suffixes in
agreement with those occurring after the word denoting the substance occur
naturally for the original word 'sukla' denoting the quality. If that is so, there is
no need for the grammatical mechanism of providing and deleting matup to
make the word denote the substance. Therefore, vdrttika 3 providing zero (luk)
in place of matup is rejected as unnecessary.
Nagesa's point is valid. The derivation by the provision and deletion of matup
derives a word 'sukla' which denotes an individual substance as different from a
quality. If Patafijali's concluding statement is understood to extend the gender
and number of the substance to the quality as a different entity, only then does
it not make sense. If Patafjali's concluding statement is understood to preserve
a distinction between a quality and a substance while extending the properties
of the substance to the quality-and one who accepts the derivation must
understand it this way-then it serves no purpose because the quality word
already achieves the correct gender and number suffixes by virtue of denoting
the substance directly.
If one rejects the derivation by the provision and deletion of matup and instead
understands that the quality adopts the gender and number of the substance by
virtue of being conceived as identical to the substance, then Patanijali's
concluding statement does make sense. However-and this is the point I wish
to make-it makes sense only if it is understood as stating that a quality adopts
the gender and number of a substance in which it resides by virtue of being
identical to it. Otherwise the same fault holds. The reason the statement would
not make sense is that the quality word, by denoting a substance, already
automatically gets the proper gender and number suffixes even if the proper
gender and number don't belong to the quality. This reason is valid irrespective
of whether the original quality word denotes a substance conventionally or an
identical speech form derived by the provision and deletion of matup denotes a
substance.
Nagesa's point is valid, and valid only if Patafijali's statement is taken as
implying that a quality is conceived as identical to a substance in which it
resides. Therefore, his statement, "The gender and number of the substance in
which the quality resides occur for the quality too," must be taken to imply
conception of a quality and substance as identical. In a similar manner, we
should understand that his assertion that a generic term denotes a generic
property and an individual substance in which it resides implies conceiving the
generic property and the substance in which it resides as identical.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
accompanies (the class property)." The very same concomitant

presence and absence of a generic term and an individual
substance which lead Patanijali to conclude that a generic term
denotes both a class property and an individual substance are used
to explain how without denoting an individual substance
cognition of it arises.94
1.8. Pfinini 1.2.58
1.4.21, Bahusu bahuvacanam, and 1.4.22, Dvyeka

dvivacanaikavacane, state the general rules for the occurr
inflectional terminations in the proper numbers. A
termination occurs when singular number is to be den
dual termination occurs when dual number is to be denote
plural termination occurs when plural number is to be den
1.2.58 states an extension (atidesa) regarding the use of th
terminations. A plural termination optionally occurs
object, if the word refers to a class.
As vdrttika 1, Jdtydkhydydm sdmdnydbhidhdn
aikdrthyam, with Patanjali's commentary explain, a gene
denotes a generic property. A generic property is a singl
Hence, if a generic term denotes a generic property, it d
single object. In accordance with the general rules fo
occurrence of inflectional terminations in the proper nu
just a singular termination would occur. However, usag
generic term is desired in the plural too. This usage woul
accounted for by the grammar without stating a specific
the purpose. 1.2.58 does this. 1.2.58 provides for the occu
of a plural termination where only a singular terminatio
naturally arise. It is stated on the assumption that a gener
denotes a single generic property.
94See Cardona 1967-68: 324-27. He concludes, "The relation betw

alternatives is clear. The first assumes that dkrti (orjdti) and dravya are
but that they always accompany each other.... That the two accomp
other then serves as the reason for the second alternative, which consid
identical." He refers to MB-K, vol. 2, p. 367, lines 3-8. On the first a
action takes place on an individual substance which accompanies
class property. The individual substance, which is considered to be
from the class property is not denoted. On the second alternative, ac
place on an individual substance which is identical to a class property
denoted as identical. The very same alternatives are expressed un
Vdrttikas 54-55 express the first alternative that a generic term den
class property but that comprehension of an individual substanc
because it accompanies the class property. Patafijali's statement unde
53 concluding that both a class property and an individual subst
denoted expresses the second alternative that both are denoted because
considered as identical.
95According to 1.4.21-22. See 1.2.64 translation note 81.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 85
In vdrttika 6, Asisyam vd bahuvat prthaktvabhidhandt,

Katyayana states that the rule need not be taught because a
generic term used in the plural denotes many individual objects
separately. Although a single generic property is a single object
present in each of the individual substances of a kind, the
individual substances are many separate objects. If many separate
objects are to be denoted, it is natural that a plural termination
occurs. In vdrttika 7, Jdtisabdena hi dravydbhidhdnam,
Katyayana states that a generic term denotes an individual
substance as well as a generic property. He rejects the assumption
that a generic term denotes just a generic property. Because a
generic term denotes an individual substance, and a plural
termination naturally occurs if there are many objects to be
denoted, it is natural that a generic term be used in the plural to
denote many individual substances of a kind.
A generic term may be used either in the singular or the
plural to refer to a class of objects. If it is understood to denote
the generic property which is a single object, then it is used in the
singular. On the other hand, if it is understood to denote the
individual substances which are many, then it is used in the plural.
The use of a singular or plural termination is an indication of
whether that which is denoted by the term is conceived as a single
entity or multiple entities. The generic property, being a single
entity, has singular number, and singular number is the natural
semantic condition for a singular termination. Individual
substances, being many, have plural number, and plural number
is the natural semantic condition for a plural termination.
One can account for the use of a generic term in both the
singular and plural without stating an additional rule by accepting
that a generic term denotes both a generic property and the
individual substances in which the generic property resides.
Pataiijali explains that a generic term does denote both. When a
speaker intends the generic property as principal then the singular
is used, and when a speaker intends the many individual
substances as principal then the plural is used (1.2.58 vdrttika
7).96
Under vdrttika 7, Patafijali describes a scenario in which a
listener has to understand that the speaker of a statement intends
an individual substance as the principal object denoted by a
generic term. A herdsman sits watching a large herd of cows. A
man approaches him and asks him, "Do you see a cow here?" If
the man meant a cow in general, that is, any cow whatsoever, he
would not have asked the question, because it is obvious that he
sees not only one cow but many. Given that it doesn't make sense
96This is neatly stated by Pataijali under 1.2.69, varttika 2, Akrtivacitvad

ekavacanam: Akrtivdcitvdd ekavacanam bhavisyati. Yadd dravydbhidhdnam
tadd dvivacanabahuvacane bhavisyatah. MB-K, vol. 1, p. 250, lines 8-10.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
to ask the question concerning a cow in general, he must be

asking about a particular cow.
This scenario clarifies what Patanijali means by his
statement under 1.2.64, varttika 53 that a speaker intends either
the generic property or an individual as principal and the other as
subordinate (?1.6, varttika 53c). Various factors in the use of a
generic term and its context show which of the two the speaker
intends as principal. The singular and plural terminations, and the
context are examples of factors that help one to know what the
speaker's intention is, whether he intends the generic property as
principal and an individual substance as subordinate or an
individual substance as principal and a generic property as
subordinate.
We have discussed how the singular and plural
terminations show this. Discussing 1.2.58 with the condition that
a generic term refers to a class, Katyayana and Patanjali have
taken the singular number termination to indicate that the object
denoted has singular number, is a single object, hence is the
generic property. They have also taken the plural number to
indicate that the objects denoted have plural number, are many,
hence are the individual substances.
In his example of the questioner and the herdsman, a
generic term in the singular denotes an individual substance.
Pataijali describes a context which shows that the speaker asks
about an individual. From this example we conclude that, asking
about a particular cow, a speaker intends the individual substance
as principal and its generic property as subordinate. Asking about
a cow in general, a speaker intends the generic property as
principal and an individual substance as subordinate.
In varttika 4, Samkhydprayoge pratisedhah, Katyayana
notes an exception to 1.2.58. There is no option to use a generic
term in the plural to refer to a class if the number one qualifies it.
This exception need not be stated for the same reasons that the
sutra itself need not be stated. The general rules for the
occurrence of inflectional terminations in the proper numbers
account for the correct usage based on whether the speaker
intends the object denoted by the generic property to be a single
generic property or many individual substances.
We can learn a certain fact of usage from the statement of
the varttika with Patafijali's commentary: a generic term is always
used in the singular to refer to a class if the number one qualifies
it. This fact of usage indicates something about the conception of
the speakers: a speaker always intends a generic property as the
principal object denoted by the generic term if he uses the
number word 'eka (one)' with it, and if he refers to a class.
Patafijali gives the example, "One rice grown makes
plenty of food (eko vrlhih samtpannah subhiksam karoti)." In the
example, the contextual information that the rice is plenty of food

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 87
makes it clear that the speaker does not intend a single individual
rice grain as principal in the meaning of the word 'rice' here. He
intends a whole crop as principal in the meaning.97 The use of
the singular inflectional termination and the use of the word 'ekah
(one)' indicates that the speaker intends the object denoted by the
word 'rice' as a single entity. He intends that the crop of rice as a
single class is the principal meaning of the word 'rice' in this
example. This is precisely in accord with the general pattern we
have discussed for a generic term if it is used to refer to a class: if
it is used in the singular, the speaker intends the generic property
as principal in the meaning and the individual substances as
subordinate. Therefore, Patanjali's explanation that a generic term
denotes both a generic property and an individual substance with
the intention of the speaker determining which is principal and
which subordinate in the meaning accounts for the consistent use
of a generic term in the singular to refer to a class if it occurs with
the number one. Because Patanijali's explanation accounts for the
usage based on general principles, there is no need to state
vdrttika 4 to provide for the usage.
1.9. Conclusion of the Grammar Section
Under 1.2.64 Patafijali discusses three views of the

denotation of generic terms: 1) A generic term denotes an
individual substance. 2) It denotes a generic property. 3) It
denotes both. He also discusses the relationship of the principle
that one word denotes one object to the question of the denotation
of generic terms.
The first view discussed accepts that a generic term
denotes individual substances and does not accept the principle
that one word denotes one object at all. In order to account for
the use of a single generic term in the plural for many individual
substances of a kind, the procedure of determining the meaning
of a speech form by concomitant presence and absence cannot be
carried out to its full extent. One must rest content with the
determination that a generic term denotes what it does naturally as
we observe it in ordinary usage. Hence the singular 'vrksah'
denotes one tree, the dual 'vrksau' denotes two trees, and the
plural 'vrksdh', many trees. One cannot divide the nominal base
and suffixes to determine that the base 'vrksa' means an individual
tree. For this reason the view is unsatisfactory.
If one accepts the principle that one word denotes one
object on the view that a generic term denotes an individual
97The fact that Pataijali considers that an exception would have to be stated for
this example if 1.2.58 were stated shows that he considers that the word 'rice' in
this example refers to the class. That is, he considers a crop of rice as a class.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
substance, then one determines that the nominal base 'vrksa'

denotes an individual tree. In order to denote many trees, many
nominal bases would be used. But it is accepted that just one base
with the plural termination denotes many trees. To achieve this
one must state that one remains.
This view is unsatisfactory because the artifice of letting
many homophonous speech forms occur and deleting all but one
is not able to explain the cognition which arises from the single
generic term used. One has the same knowledge regarding many
individual trees that they are trees, and one knows that each is the
same as the other. To convey this knowledge, the nominal base
'vrksa' is used (with an appropriate termination). Whether or not
a generic property actually exists as an object is not the concern
of a grammarian. However, Patanijali seeks to account for the
ordinary usage of a word as arising naturally based on certain
principles from the ordinary conception of an object.
People have identical knowledge regarding each of many
individual substances of a kind and the knowledge that each is the
same as the other. This knowledge is the condition for the use of
a generic term. The object conceived to be the basis of such
knowledge is a generic property. Hence, a generic term is said to
denote a generic property. On this view a generic term denotes a
generic property and the principle that one word denotes one
object is upheld. Upholding the principle that one word denotes
one meaning in order to avoid stating that one remains requires
accepting that a word denotes an over-arching general property
common to all the objects for which a word is used in the case of
words used in different meanings.
The view that a generic term denotes a generic property
has a difficulty complementary to the one belonging to the view
that a generic term denotes an individual substance. The view
that a generic term denotes an individual substance ignores that
there is an identical cognition regarding each of many individual
substances of a kind with the knowledge that each is the same as
the other. As a result it has difficulty accounting for a single
cognition in relation to many individual substances of a kind and
the use of a single speech form to convey that single cognition.
Similarly, the view that a generic term denotes a generic property
has difficulty accounting for the cognitions of dual and plural
number in connection with a single generic property and the use
of dual and plural inflectional terminations which naturally arise
to convey them. To overcome the difficulty it is accepted that the
individual substances which are the substrata of a generic
property are the semantic conditions for usage of inflectional
terminations while still maintaining that the generic property itself
is the only object denoted by the nominal base.
Patafijali accepts the view that a generic term denotes both
a generic property and an individual substance. The generic

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GRAMMAR 89
property included in the meaning accounts for the same cognit

with respect to all individual substances of a kind and the
knowledge that each of them is the same as the others. It is th
semantic condition for the use of a single base and which allow
general statements to apply repeatedly to different individual
The inclusion of an individual substance in the meaning accoun
for the use of a generic term in syntactic connection with acti
including the acts of coming into being and being destroyed. I
accounts for the differences in quality of the object to which t
word applies and for the distinction of one object from anoth
which that implies. Individual substances are the semantic
conditions for the use of different gender and number suffix
and for many instances of a generic term in analytic strings.
One of the two, a generic property or an individual
substance, is principal in the meaning, the other subordin
depending on whether the speaker intends a particular individ
or intends the statement to apply generally. The context a
purpose of the statement help to determine what the speaker'
intention is. In the case of generic terms which are polysemou
different generic properties are the reason they are said to ha
many meanings.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Panini 1.2.58 Translation
1.2.58. Jatyakhydydm ekasmin bahuvacanam

anyatarasydm.
A plural (number termination) optionally occurs for one
(object) if (the word) refers to a class.
Idam ayuktam vartate.

(Objector:) This is not proper.1
Kim atrayuktam?
(Proponent:) What is not proper here?
Bahavas te 'rthds tatra yuktam bahuvacanam. Tad yad ekavacane

sasitavye bahuvacanam sisyata etad ayuktam. Bahusv
ekavacanam iti ndma vaktavyam.
(Objector:) There are many objects so the plural is (naturally)
appropriate. It is not proper that the plural is (explicitly) taught
when it is the singular which has to be taught. "A singular
(number termination) occurs for many (objects)," is what should
be stated.
Ata uttaram pathati:

(Narrator:) Hence he declares the answer:
1. Jatyakhydyam samanydbhidhdndd aikdrthyam.

(Proponent:) If (the word) refers to a class, it has one
object because it denotes a general property.
Jatydkhydydm sdmanydbhidhdndd aikdrthyam bhavisyati. Yat tad
vrlhau vrlhitvam yave yavatvam Gdrgye gargyatvam tad ekam
tac ca vivaksitam. Tasyaikatvad ekavacanam eva prdpnoti. Isyate
ca bahuvacanam sydd iti tac cdntarena yatnam na sidhyatlti
jatydkhydyam ekasmin bahuvacanam. Evam artham idam ucyate.
If the word refers to a class, it will have one object because it
denotes a general property. The riceness in rice, the barleyness in
barley, the gargyaness in Gargya, is a single thing and it is what
the speaker intends. Because it is a single thing, just a singular
number termination obtains. But it is desired that the plural
should occur too and that doesn't succeed without an effort.
Hence Panini states, "A plural (number termination) occurs
IThe objector at first holds the view that a word denotes just an individual
substance; the proponent that it denotes just a generic property. The text is
from MB-K, vol. 1, pp. 229, line 9 - 230, line 21.
91

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
for one (object) if (the word) refers to a class." It is for th

purpose that the present sutra is stated.
Asti prayojanam etat?

(Objector:) This is a reason?
Kim tarhiti?
(Proponent:) What else do you think it is?
2. Tatraikavacanddesa uktam.
(Objector:) In this matter, with regard to the extension
that (a plural) number (termination occurs instead of a singular
termination), it has been stated.
Kim uktam?
(Proponent:) What has been stated?
Vrihibhya dgata ity atra gher hitfti gunah prdpnotlti.

(Objector:) In "vrlhibhya dgatah (come for rice)," the guna vowel
e would obtain (instead of the final i of the stem 'vrlhi) by
7.3.111.2
27.3.111. Gher niti (guna 108). A guna vowel replaces the final sound of a stem
termed ghi before a termination marked with n.
By 1.4.7, Seso ghy asakhi, the class name ghi applies to speech forms ending in
short i or u with the exception of 'sakhi' and 'pati' (outside of a compound). In
addition it applies to feminine speech forms not termed nadl. In the example,
"vrlhibhya dgata," the base 'vrlhi' is termed ghi. Hence, 7.3.111 provides that
the guna vowel e substitutes for i before the fourth triplet singular termination
he (e marked with n).
1.2.58 extends the use of a plural termination so that it occurs where a singular
termination would otherwise occur. The objector understands that it applies
after the singular termination has already arisen. He takes it as a substitution
rule (ddesa) replacing the singular which has already occurred naturally by
explicitly stating that a plural occurs. If 1.2.58 were a substitution rule a fault
would arise in the derivation of 'vrlhibhyah' because the extension rule
(atidesa) 1.1.56 would come into play.
1.1.56. Sthdnivad ddeso 'nalvidhau. Except those which have certain sound
conditions, operations which apply to an item subject to substitution equally
apply to its substitute.
First the fourth triplet singular termination he marked with h would arise after
the base 'vrthi'. Then, because 'vrlhi' refers to rice as a class, 1.2.58 would
optionally replace he with the plural bhyas which is not marked with h and does
not condition guna substitution by 7.3.111. But 1.1.56 would allow 7.3.111 to
replace the final i of 'vrlhi' in 'vrihi + bhyas' with e, as it does in 'vrlhi + he'.
The result is an erroneous form.
Saying, "it has been stated," the objector refers to vdrttika 19 under 1.1.56,
and elaborating he cites a line of Patafijali's comment there: 19.
Jdtydkhydydm vacanatidese sth nivadbhavapratisedha h.
Jdtydkhyydam vacanctidese sthdnivadbhdvasya pratisedho vaktavyah
Vrthibhya dgata ity atra gher niti (7.3.111) iti gunah prdpnoti. Naisa d
uktam etat. Arthdtidesat siddham iti. (Objector:) With regard to th
extension of number, a prohibition against (a substitute) being

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
PANINI 1.2.58 TRANSLATION 93
Naisa dosah.
(Proponent:) This fault does not arise.
3. Arthatidesdt siddham.
It succeeds because the extension applies to the
meaning.
Arthdtideso 'yam. Nedam pdribhdsikasya vacanasya grahanam.
This extension applies to the meaning. This is not a reference to
the technical term.3
Kim tarhi?
(Objector:) What then?
Anvarthagrahanam: Ucyate vacanam. Bahuindm arthdndm

vacanam bahuvacanam iti. Yavad bruiydd eko 'rtho
bahuvadbhavatfti tdvad ekasmin bahuvacanam iti.
(Proponent:) It is a usage in accordance with its etymological
meaning: vacana is that which is stated, bahuvacana is that which
is stated of many objects.4 To say, "A bahuvacana occurs for one
object," is as much as to say, "One object is as if many."5
like its substituend (has to be stated). With regard to the extension that (a
plural) number (termination occurs instead of a singular termination), a
prohibition against (a substitute) being like its substituend has to be stated. In
"vrihibhya dgatah (come for rice)," the guna vowel e would obtain (instead of
the final i of the stem 'vrihi') by 7.3.111. (Proponent:) This fault does not arise.
It has been stated, "It succeeds because the extension applies to the
meaning." MB-K, vol. 1, p. 138, lines 20-23.
Note that the proponent in the cited passage refers to vdrttika 3 under sutra
1.2.58.
3The word 'vacana' in the compound 'bahuvacana' in 1.2.58 is not a usage of the
technical term 'vacana (number termination)'.
4"This is not a usage, etc. (Nedam ...iti)" is cited under 8.2.81, Eta id bahuvacane.
MB-K, vol. 3, p. 415, lines 9-10.
5Vdrttika 3 avoids the fault attributed to 1.2.58 in vdrttika 2. See ?1.5.4.
Kaiyata explains: Bahunam iti. BahuJndm arthdndm bahutvam evocyate ity
ekasminnarthe bahutvam anendtidisyata iti bahutvdd bahuvacanam bhavatiti
ndsti sthdnyddesabhdvah.
Of many. It is the plural number which is stated of many objects; Hence
(1.2.58) extends plural number to one object. The plural number termination
arises because there is plural number. Hence there is no substitute-substituend
relationship. MB, vol. 2, p. 66a-66b.
Panini 1.4.102, Tdny ekavacanadvivacanabahuvacandny ekasah, and 103,
Supah, introduce the terms ekavacana, etc. as technical terms of grammar. They
are class names of certain members of the sets of verbal and nominal inflectional
terminations (tii and sup). Patafijali affirms this under 4.1.21: Evam tarhy
ekavacanam dvivacanam bahuvacanam iti sabdasamjnii etdh. ( MB-K, vol. 1, p.
321, lines 3-4.) Later under the same suitra he cites a slokavdrttika stating that
number (samkhyd) is one of the meanings of the terminations which receive
these class names: Supdm karmddayo 'py arthdh samkhya caiva tathd tiinm.
MB-K, vol. 1, p. 322, line 17.
The fault attributed to 1.2.58 in vdrttika 3 arises from considering 1.2.58 to be a
substitution rule (ddesa). However, it does not supply grounds for denying that
the term 'bahuvacana' names inflectional terminations in 1.2.58. Nor does the

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
4. Samkhydprayoge pratisedhah.
(Objector:) A prohibition has to be stated for when a
number is used.
Samkhyaprayoge pratisedho vaktavyah. Eko vrThih sam
subhiksam karoti.
A prohibition against 1.2.58 has to be stated for when a number is
used. "One rice grown makes plenty of food."
5. Asmado ndmayuvapratyayayos ca.
And for when a name or yuvan-affix (is used with the first
person pronoun) asmad.
Asmado namaprayoge yuvapratyayayoge ca pratisedho
vaktavyah. Ndmaprayoge: Aham Devadatto bravtmi. Aham
Yajinadatto bravlmi. Yuvapratyayaprayoge: Aham Gdrgydyano
bravimi. Aham Vdtsydyano bravimi.
A prohibition has to be stated (against 1.2.59) for when a name is
used or a yuvan-affix is used with the first person pronoun
asmad. When a name is used: I Devadatta say, I Yajnadatta say.
When a yuvan-affix is used: I Gargyayana say, I Vatsyayana
say.6
Yuvagrahanena ncrthah. Asmado ndmapratyayaprayoge nety

eva. Idam api siddham bhavati: Aham Gdrgyo bravimi. Aham
Vdtsyo bravtmi.
(Suggested modification:) There is no purpose for including
'yuvan' in the vdrttika. One should state just, "1.2.59 does not
apply when a name or affix is used with the first person pronoun
suggestion that it is an anvarthagrahana have to be accepted as a final

conclusion. The very fact that vdrttika 6 offers an alternative solution indicates
that it is not the final conclusion. Even if the proponent accepts that the term
'bahuvacana' names a certain set of terminations he can establish his point,
namely, that 1.2.58 is not a substitution rule replacing singular terminations
with plural ones but rather an extension of the meaning in which the plural
terminations initially arise. (See Sivaramendrasarasvati, "Vastutas tv
ekasminnarthe bahuvacanam bhavattti suitrdrthdn na kdpy anupapattih." MPV,
vol. 4, p. 84.) Since there is no sixth case indicating a substituend, 1.1.56
would not come into play. However, a problem would arise in sutra 1.2.63,
Tisyapunarvasvor naksatradvandve bahuvacanasya dvivacanam nityam, where
the presence of the sixth triplet termination makes the explanation of the rule as
an extension (atidesa) rather than a substitute (adesa) less convincing. See
Kaiyata and Nagesa, MB, vol. 2, p. 66b.
61.2.59. Asmado dvayos ca (ekasmin anyatarasydm 58). After the first person
pronoun base asmad, a plural (number termination) optionally occurs for one or
two objects.
4.1.105, Gargddibhyo yafi, provides the suffix yah after the bases 'garga' and
'vatsa' to form the derivatives 'gdrgya' and 'vdtsya' denoting a gotra descendant.
After these derivatives, 4.1.101, Yaniniios ca, provides the affix phak to form the
derivatives 'gdrgydyana' and 'vdtsydyana' denoting yuvan descendants. Panini
4.1.162-167 define the terms 'gotra' and 'yuvan'. See Gautama 2.2.62 V10, note
22, and Panini 1.2.64 translation, note 7. Concerning the status of gdrgyatva as
a generic property, see ?1.2.5.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
asmad. This achieves the following too: "I Gargya say, I Vatsya
say."7
Apara aha: Asmadah savisesanasya prayoge nety eva. Idam api

siddham bhavati. Aham patur bravimi. Aham pandito bravlmi.
Another says: One should state just, "1.2.59 does not apply when
the first person pronoun asmad is used with a qualifier." This
achieves the following too: I, wise, say; I, learned, say.8
6. Asisyam vd bahuvat prthaktvdbhidhdndt.

Asisyo vd bahuvadbhdvah.
(Objector:) Or it should not be taught (that one generic
property) is as if many because (the substances) are
denoted as separate.
It should not be taught that one generic property is as if many.9
70ne need not specify a yuvan affix because the exception holds for affixes
which occur to form derivatives denoting gotra descendants as well.
8If 'patu' and 'pandita' are taken to be underived nominal bases, then the
previous formulation, "(1.2.59) does not (apply) when a name or affix is used
with asmad," cannot account for the two examples, "I, wise, say; I, learned, say."
But if they are accepted as derived, then the phrase, "when an affix is used,"
includes them. 'Patu' is derived by adding the Unddi affix u to the verbal root
Vpat (Kasika, vol. 3, p. 240), and 'pandita' by Panini 5.2.36, Tad asya samjdtam
tarakddibhya itac.
9From here on the discussant labeled 'objector' takes the view that 1.2.58
should be rejected as unnecessary because a generic term denotes both a generic
property and an individual. This discussant need not be the same person as the
one labeled 'objector' above now having shifted to adopt a new view; it may be a
different discussant who presents the final view.
Kaiyata: Ubhayapaddrthapaksdsrayena pratydkhydtum aha: asisyam iti. Tatra
kaddcij jdteh prddhdnyam kaddcid dravyasyeti yathestam prayoga upapadyate.
He says, "it should not be taught ," to reject (1.2.58) as unnecessary by
resorting to the view that a word denotes both (a generic property and an
individual substance). On that (view) sometimes the generic property is
principal, sometimes an individual substance. Hence usage arises as desired.
MB, vol. 2, p. 67a.
Nagesa summarizes the view on which 1.2.58 is unnecessary: A word gives
cognition of both a generic property and an individual. In any particular usage,
a speaker's intention determines which of the two is the qualified object
(principal) and which the qualifier. Inflectional terminations, accompanying
words, etc., make the speaker's intention understood. In the domain to which
the present sutra would apply, the singular number termination conveys that the
speaker intends a generic property and the plural that he intends individuals.
When a word such as 'gauh (cow)' is used, the generic property is intended as
being identical to an individual, even if it is intended as the qualified object.
Hence it makes sense that the action carried out on an individual fulfills the
intent of the command, "Gam vdhaya (drive a cow)," even though an individual
is not intended as principal. In contrast, a word ending in the suffix tva gives
cognition of just a generic property as distinct from any individual. Hence it
does not make sense to say, "Gotvam vdhaya (drive cowness)." See MB, vol. 2,
p. 67a.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Kim kdranam?
(Proponent:) What is the reason?
Prthaktvdbhidhdndt. Prthaktvena hi dravydny abhidhtyante.

Bahavas te 'rthds tatra yuktam bahuvacanam.
(Objector:) Because the individual substances are denoted as
separate. Because individual substances are denoted as separate.
There are many objects so the plural is naturally appropriate.
Kim ucyate prthaktvdbhidhdndd iti Ydvateddntm evoktam

jdtydkhydydm sdmdnydbhidhdndd aikdrthyam iti.
(Proponent:) How can you say, "Because (the substances) are
denoted as being separate," since just now it has been said, "If (the
word) refers to a class, it has one object because it denotes a
general property (vdrttika 1)."
7. Jatisabdena hi dravyabhidhanam.
(Objector:) Because a generic term denotes a substance.
Jdtisabdena hi dravyam apy abhidhlyate jdtir api.
Because a generic term denotes a substance too, as well as a
generic property.
Katham punar jnadyate jatisabdena dravyam apy abhidhtyata iti?

(Proponent:) But how does one know that a generic property
denotes a substance too.
Evam hi kascin mahati gomandale gopalakam dslnam p rcchati

"Asty atra kamcid gdm pasyastti?" Sa pasyati: Pasyati cdyam gah,
prcchati ca "Kdmcid atra gdm pasyasiti." Nunam asya dravyam
vivaksitam iti. Tad yadd dravydbhidhdnam tadd bahuvacanam
bhavisyati; yadd sdmanydbhidhdnam tadaikavacanam bhavisyati.
(Objector:) In just this way: Someone asks a herdsman sitting by
a large herd of cattle, "Do you see a cow?" The herdsman sees,
"He sees cows and yet asks, 'Do you see a cow here?' Certainly he
intends a specific individual substance." Therefore, when a
generic term denotes an individual substance, the plural will
occur; when it denotes a general property, the singular will
occur.10
10The content of the last sentence is repeated in 1.2.69, vdrttika 2 and

commentary: Akrtivacitvdd ekavacanam. A krtivdcitvdd ekavacanam
bhavisyati. Yada dravydbhidhdnam tadd dvivacanabahuvacane bhavisyatah.
MB-K, vol. 1, p. 250, lines 8-10.
In order to convey the intent of the question, "Asty atra kdmcid gdm pasyasi?" it
should be translated, "Do you see a certain cow?" rendering 'kdmcit' by 'a
certain'. I have deliberately translated the question, "Do you see a cow?"
rendering 'kdmcit' by the vague 'a' in order to allow Patafijali's own discussion
to show the intent of the question. To translate "kdmcid gdm" by "a certain
cow," gives it away that 'cow' refers to a particular individual.
Nagesa: Bhdsye mahati gomandala iti. Atinibidagomandale 'viviktatayd

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
drste ittham uktih. Tatra jater eva sabddrthatve jater drstatvdt prasndnupapattir
ity arthah. In the hhdsya, by a large herd of cattle: The statement is such
that (the questioner) sees a very thick herd of cattle without distinguishing any
of them. In such a situation, the question wouldn't make sense if just a generi
property were the object denoted by a word because he sees the generic property.
MB, vol. 2, p. 67b.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Panini 1.2.64 Outline
I. Introduction to the purpose of the rule

A. The function of each of the terms in defining the sco
of the rule
B. The grammatical purpose of the rule (vdrttikas 1-3)
II. Examination of the rule's mode and scope of application (not

included)
III. The rule is rejected as unnecessary on the view that a word

denotes an individual substance by not accepting the
principle that one word denotes one object. Words denote
what they are seen denoting in ordinary usage naturally.
(varttikas 27-34)
IV. The rule is rejected as unnecessary on Vajapyayana's view

that a word denotes a single class property even accepting
that one word denotes one object. (varttikas 35-44)
V. Vyadi's view that a word denotes an individual. (varttikas 45-

52)
VI. Defense of Vajapyayana's view that a word denotes a class

property (varttikas 53-59).
99

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Panini 1.2.64 Translation
1.2.64. Sarupdanam ekasesa ekavibhaktau.

Of those speech units whose form (rupa) is the same,
just one remains, if there is one termination.1
Rupagrahanam kimartham?
(Sutra-rejecter:) For what purpose does the suitra include the
word 'form (rupa)'?
Samanadndm ekasesa ekavibhaktdv itiyaty ucyamdne yatraiva

sarvam samanam sabdo 'rthas ca tatraiva sydt. Vrksd plaksd iti.
Iha na sydt: aksdh pdddh mdsd iti. Ruipagrahane punah kriyamane
na doso bhavati. Rupam nimittatvenasrtyate srutau ca
ruipagrahanam.
(Sutra-proponent:) If the sutra said just, "Of the same just one
remains if there is one inflectional termination (samdndndm
ekasesa ekavibhaktau)," then the rule would apply only where
everything is the same, the speech form and the meaning, for
example, 'vrksah (trees), plaksdh (fig trees)'. It would not apply
here: 'aksdh (die, axle, seed), padad (foot, verse segment,
quarter), mdsdh (bean, monetary unit, fool)'. But if the suttra
includes the word 'form (ruipa)' the fault does not arise. The form
is resorted to as the condition upon which just one remains, and
form (ruipa) is taken in the domain of sound (sruti).2
l"Samdnam rupam esdm iti sarupdh. Saruipdnam sabdhdndm ekavibhaktau

parata ekaseso bhavati. Ekah sisyate itare nivartante." Kgaika, vol. 1, p. 48. The
word 'sarupa' is a bahuvrihi compound derived according to 2.2.24. Anekam
anyapaddrthe. 'Sa' replaces 'samdna' according to 6.3.85. Jyotirjanapadardtri-
ndbhindmagotraruipasthdnavarnavayovacanabandhusu (sah 78, samdnasya 84).
The Kasika (vol. 1, p. 48) says, "before one termination (ekavibhaktau paratah),"
rather than, "if there is one termination." Patafijali discusses the meaning and
significance of the term 'ekavibhaktau' in 1.2.64 at length. It has different
meanings depending upon whether the rule applies to nominal bases
(pratipadika) or words (pada). He determines justifiable meanings for the term
which would yield the correct results on each interpretation and then proceeds
to consider the more fundamental semantic issues bearing on whether it is
necessary to state the rule at all. The present dissertation is concerned just with
these latter issues and cannot dwell on the fine points of the formulation of the
rule.
2Pataijali explains that by mentioning form (rupa) in 1.1.68, Svam rupam

sabdasyasabdasamjhi, Panini informs us that both sound and meaning belong
to a speech form (sabda) and that reference to a speech form in grammar includes
the meaning (artha) connected with the speech form as well as the speech form
itself. MB-K, vol. 1, p. 175, lines 19-23. However, just the sound is a factor, not
the meaning, in the condition for 1.2.64 to apply.
I have translated according to the first of two interpretations of the phrase,
"srutau ca rupagrahanam," which Kaiyata offers. He comments: Sabddnuidsana-
101

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Athaikagrahanam kimartham?
(Sutra-rejecter:) For what purpose does the sutra include the
word 'one (eka)'?
Saruipadnam sesa ekavibhaktav itiyaty ucyamdne dvibahvor api

sesah prasajyeta. Ekagrahane punah kriyamdne na doso bhavati.
(Sutra-proponent:) If the sutra said just, "Of those speech units
whose form is the same, there is a remainder, if there is one
termination (sarupadndm sesa ekavibhaktau)," it would be possible
for two or many to remain. But if the sutra includes the word
'one (eka)' the fault does not arise.
Atha s'esagrahanam kimartham?

(Sutra-rejecter:) Then for what purpose does the sutra include the
word 'remains (sesa)'?
Sarupdndm eka ekavibhaktdv ittyaty ucyamdna ddeso 'yam

vijndyeta.
(Sutra-proponent:) If the sutra said just, "Of those speech units
whose form is the same, one occurs, if there is one inflectional
termination (sarupadndm eka ekavibhaktau)," the rule would be
understood as a substitution (ddesa).3
Tatra ko dosah?
(Sutra-rejecter:) If that were so, what fault would arise?
As'vas casvas ca dsvau. Antaryato dvyuddttavatah sthdnino

dvyuddttavdn ddesah prasajyeta. Lopyalopitd ca na prakalpeta.
(Sutra-proponent:) As'va (horse) and asva (horse) asvau (two
horses). Because of its similarity, in place of a substituend with
two high pitched sounds a substitute with two high pitched sounds
would obtain.4 And that which ought to be deleted would not
succeed in being deleted.
Tatra ko dosah?
prastdvac ca sabdasambandhirutpam grhyate. And because the presentation is

instruction of (correct) speech (sabda), it is the form connected with speech
sound (sabda) which is meant here (grhyate). Nagesa clarifies: Adye grahanam
visayatd. On the first (of the two views) taking (grahana means) being in the
domain of. MB, vol. 2, p. 71a.
Sabara mentions the example of the word 'aksa' under Jaimini 6.8.10. MD, part
5, p. 370.
31.1.49. Sastht sthdneyoga. A sixth triplet ending designates the relation "in
place of".
41.1.50. Sthdne 'ntaratamah. In place of (a substituend) the (element) most
similar (to it occurs).

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Gargah Vatsdh. Biddh Urvah. An yo bahusu yan yo bahusv ity

ucyamdno lug na prapnoti.
(Sutra-proponent:) Gargah, Vatsah. Bidah, Urvah. The zero
(luk) stated for the affix al which refers to many descendants and
for the affix yai which refers to many descendants would not
obtain.5
Ma bhud evam. Aniantam yad bahusu yanantam yad bahusv ity

evam bhavisyati.
(Sutra-rejecter:) It would not be so. It will be like this: The base
ending in yai which refers to many, the base ending in al which
refers to many.6
Naivam sakyam. Iha hi dosah sydt. Kdsyapapratikrtayah kasyapd

iti.
(Sutra-proponent:) It can't be that way because there would be a
fault here: 'kasyapdh' meaning many images of Kasyapa.7
5With 4.1.92, Tasydpatyam, 4.1.104, Anrsydnantarye vidddibhyo 'n (gotre 98),

provides the affix al after the bases 'vida, urva,' etc., and 4.1.105, Gargddihhyo
yan (gotre 98), provides the affix yai after the bases 'garga, vatsa,' and most of
the others in the list beginning with 'garga', to form derivates denoting gotra
descendants. Concerning the term gotra see note 7.
2.4.64. Yahanio ca (bahusu tenaivdstriydm 62, gotre 63, luk 58). Zero (luk)
replaces the affixes yah and al referring to many gotra descendants in a gender
other than the feminine (and nullifies any operations which those affixes
condition).
Kaiyata explains as follows: Since there are many descendants of one
predecessor Garga, according to the principle that one speech form occurs for
each object denoted, one base will be followed by many suffixes to yield 'garga
+ yayaya'. Based on the assumption that 1.2.64 applies only to speech units
whose form is the same, by applying the metarule 1.1.49 by which 1.2.64 would
be a substitution rule, one suffix ya would be the substitute for all of the
suffixes ya ya ya. On this view the fault which the proponent states would not
obtain. However, in vdrttika 23, Ekdrthdndm api virupdndm (MB-K, vol. 1, p.
239, line 6), Katyayana extends the rule that one remains to speech units whose
forms are different. Based on this extension, applying 1.1.49 would result in
'garga + ya' substituting for 'garga + ya ya ya'. In this case, the stated objection
would apply. It would not be the case that the suffix ya alone refers to many;
rather, the base + affix together would refer to many. Hence the zero (luk) which
2.4.64 states as applying on the condition that the suffix refers to many would
not apply. MB, vol. 2, p. 72a.
61.1.72. Yena vidhis tadantasya. (A qualifier) x by means of which a provision
is made refers to that which ends in x. Hence, yai in 2.4.64 refers to that which
ends in yan. As long as the aggregate 'garga + ya' which replaces 'garga + ya ya
ya' refers to many, luk will apply to the affix yan.
75.3.96. Ive pratikrtau (kan 95). The affix kan occurs after a base if an image
similar what the base denotes is to be denoted.
5.3.99. Jivikdrthe cdpanye (lup 98). Zero (Lup) occurs in place of kan if the
object denoted is an image (used for worship), is not for sale, and serves the
purpose of sustenance (for beggars). See Padamanijari, KNP, vol. 4, p. 311.
7.2.117, Taddhitesv acdm ddeh (angasya 6.4.1, vrddhih 114, nniti 115). Vrddhi
replaces the first of the vowels of the stem after which a taddhita affix marked

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Ekavibhaktdv iti kimartham?

(Sftra-rejecter:) For what purpose does the sutra include the
word 'if there is one termination (ekavibhaktau)'?
Payah payo jarayati. Vdso vasas chddayati. Brdhmandbhydm ca
k rtam brdhmandbhydm ca dehlti.
(Sutra-proponent:) Milk curdles milk. Cloth covers cloth. Done
by brdhmanas and give to brdhmanas.8
with A or n is provided.
'Kasyapa' is 3rd in the list beginning with 'vida'. The derivation of 'kadyapdh' is
as follows:
kasyapa + an, 4.1.104.
kdsyapa 7.2.117.
kasyapa + kan 5.3.96.
kdsyapa + 0 5.3.99.
One 'kasyapa' substituting for 'kdsyapa kdsyapa kdsyapa' ends in the affix
and refers to many objects which are not feminine. Hence 2.4.64. woul
erroneously provide luk in this case. Later in his commentary on this sut
Patanijali cites this passage (MB-K, vol. 1, p. 238, lines 22-26) and answers t
last objection as follows: Naisa dosah. Laukikasya tatra gotrasya grahanam
caital laukikam gotram (p. 238, lines 26-27). Tatra' refers to 2.4.64 and 'eta
refers to the use of 'kdayapa' for images. The term 'gotra' has three meanings
The technical term 'gotra' (sistrTyam gotram) defined in 4.1.162-67 is confi
to the provision of gotra affixes. See Gautama 2.2.62 V10, note 22, and Cardon
(1988: 52-53). 2) Elsewhere such as in 2.4.63, Yaskddibhyo gotre, and the
sitras in which it recurs, the term 'gotra' has its ordinary sense of descendant
(apatya). 3) The term's sense is extended in practice to mean family line. In
weddings and other ceremonies the participants' family line (gotra) is read out.
2.4.64 provides luk of the affixes yai and af wherever they occur in the second
sense of the term 'gotra'. Patafijali mentions 'laukika' just to reiterate that sense;
it is not relevant to the example of 'kdsyapa' used for an image of a descendant.
Such usage does not refer to gotra in any sense technical or ordinary, except
metaphorically.
2.4.64 provides luk of an only if the affix aft occurs in the plural. There must be
many descendants. In the example, however, there are many images of only one
descendant. Hence 2.4.64 will not apply (as it ought not) to 'kdiyapa' as a
single substitute for 'kdsyapa kadyapa kdsyapa'. Commenting on 2.4.62,
Patanijali repeats much of the passage and clarifies that in order that luk occur
the affix ah (or yai) itself must refer to many: Tenaiva cet krtam bahutvam iti
va vaktavyam....Tasmdt pratyaydrthahahutve lug ity esa eva pakso jyaydn. See
MB-K, vol. 1, p. 492, lines 12-19. For Kaiyata and Nagesa's comments see MB,
vol. 2, p. 72b and p. 84a-b, and vol. 2, p. 534b. For the Kasika, Nyasa, and
Padamafijari on 2.4.63 see KNP, vol. 2, pp. 307-8.
8The examples show different terminations of the same form. In the first
example, the first 'payah' is treated as if it ends in a first triplet termination; the
second as if it ends in a second triplet termination. The second example is
similar. In the third example the first 'brdhmana' ends in a third triplet
termination; the second in a fourth triplet termination. Even though the speech
forms are of the same form, the rule that one remains does not apply because the
terminations are different. Later under the present sutra Pataijali gives the last
of these three examples again and then commenting on vdrttika 14, Na
vdrthavipratisedhdd yugapadvacandbhhvah, accounts for the fact that the rule
that one remains does not apply to it even without the word 'ekavihhaktau'. The
meanings of different terminations conflict and so cannot be expressed
simultaneously. MB-K, vol. 1, p. 236 lines 12-19.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Kimartham punar idam ucyate?

(Sutra-rejecter:) But for what purpose is this sutra stated?
1. Pratyartham sabdanivesdn
naikendnekasydbhidhdnam.
(Sftra-proponent:) Because words apply per object, one
word does not denote many objects.
Pratyartham sabdd abhinivisante.
Words apply per object.9
Kim idam pratyartham iti?

(Sutra-rejecter:) What is this, 'per object'?
Artham artham prati pratyartham. Pratyartham sabdanivesdd

etasmdt kdranan naikena sabdendnekasydrthasydbhidhdnam
prdpnoti.
(Sutra-proponent:) 'Per object' means "to each object".10 Because
words apply per object, due to this cause, it would not obtain that
one word denotes many objects.
Tatra ko dosah?
2. Tatrdnekdrthabhidhdne 'nekasabdatvam.
(Sitra-proponent:) In that case, to denote many objects
many words would be used.
Tatrdnekdrthdbhidhdne 'nekasabdatvam prdpnoti. Isyate
caikendpy anekasyabhidhdnam sydd iti tac cdntarena yatnam na
sidhyati.
In that case, to denote many objects the use of many words would
obtain. But it is accepted that even one word may denote many
objects, and that doesn't succeed without an effort.
3. Tasmdd ekasesah.
Therefore, the rule that one remains.
9To be precise I should continue to translate 'sabda' by 'speech form'. I use

'word' only because it is less cumbersome and because my paper is not
concerned with distinguishing different types of speech forms.
10I have translated the doubling of the word 'artha' as 'each' in accordance with
8.1.4, Nityavipsayoh (san,asya dye 1), a whole (syntactic item) is doubled to
convey repetition or pervasion. The compound 'pratyartham' is formed by 2.1.6,
Avyayam vibhakti-samipa-samrddhi-vyrddhy-arthbhhdvdtyaydsamprati-
sabdapradurbhdva-pa. scd-yathdnupurvya-yaugapadya-sddrsya-sampatti-
sdkalydnta-vacanesu (sulp 2, supd 4, avyayibhhva 5), an indeclinable conveying
the meaning of a termination (vihhakti), ..., the appearance of a sound or word
(sabdaprddurbhdva), ..., certain meanings of yathd (fitness, repetition or
pervasion (vipsa), and accordance), etc., combines with a syntactically
related nominal word to form an avyayibhdva compound. The Kasika (vol. 1, p.
101) gives the present passage, "artham artham prati pratyartham," as the
example of an avyayfbhdva formed from 'yathd' conveying pervasion.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Evamartham idam ucyate.

For this purpose the present sutra is stated.
Asti prayojanam etat?

(Sutra-rejecter:) This is a reason?
Kim tarhtti?
(Sutra-proponent:) What else do you think it is?
Kim idam pratyartham sabdd abhinivesanta ity etam drstantam

dsthdya sarupdndm ekasesa drabhyate na punar apratyartha
sabdd abhinivisanta ity etam drstdntam dsthdya virupd.na
anekasesa drabhyate?
(Sutra-rejecter:) Why assume the principle that words apply pe
object and make the rule that of those speech units whose form
the same just one remains rather than assume the principle th
words apply other than per object and make the rule that of th
speech units whose form is different many remain?
Tatraitat sydl laghiyasi saruipanivrttir garlyasf viruipapratipattir it

(Sutra-proponent:) This would be pertinent: Omitting those
whose form is the same is easier; understanding those whose for
is different is more difficult.11
Tac ca na. Laghlyast viruipapratipattih.

(STitra-rejecter:) That's not right either. It is easier to understan
those whose form is different.
Kim kdranam?
(Sutra-proponent:) What is the reason?
Yatra hi bahuindm sarupdndm ekah sisyate tatrdvarato dvayoh

sarupayor niv.rttir vaktavyd sydt.
(Sutra-rejecter:) Because where one of many speech units whose
form is the same remains one must state the omission of at least
two whose form is the same.12
11Once one knows the sequence of sounds of one of words of the same form one
knows them all. On the other hand, one must learn the sequence of sounds of
each of words whose form is different separately.
12If there are two similar speech forms just the last one is omitted, if there are
three then the last two, ... if there are n then the last n-l. Understanding speech
units with a different sound sequence is easier because the mind is consistently
engaged in producing speech forms whereas if just one of similar speech forms
remains the mind is sometimes engaged in producing and sometimes in
omitting. The dual type of behavior is more complicated. See Kaiyata MB, vol.
2, p. 74a.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Evam apy etasmin sati kimcid acaryah sukaratarakam manyate

sukaratarakam caikasesarambham manyate.
(Sutra-proponent:) Even though that is true the teacher gives
regard to something that is easier and he considers making the
rule that one remains easier.13
27. Dvivacanabahuvacantprasiddhis caikdrthatvdt.

Dvivacanabahuvacanayos cdprasiddhih.
(Sutra-rejecter:) And the dual and plural terminations
would not obtain because one object is denoted.14
And the dual and plural terminations would not obtain.
Kim kdranam?
(Sftra-proponent:) What is the reason?
Ekdrthatvdt. Eko 'yam avasisyate tendnena tadarthena

bhavitavyam.
(Sftra-rejecter:) Because one object is denoted. Just one (speech
form) remains; that same (remaining speech form) should occur
to denote its object.
Kimarthena?
(Sutra-proponent:) What object?
Yadartha ekah.
(Sutra-rejecter:) The object which one speech form denotes.
Kimarthas caikah?
(Sutra-proponent:) What object does one speech form denote?
Eka ekdrthah.
(Sutra-rejecter:) One speech form denotes one object.
Naikdrthyam. Nayam ekarthah.

(Suitra-proponent:) It is not so that one object is denoted. This
one speech form which remains does not denote one object.
Kim tarhi?
(Sftra-rejecter:) What then?
13In this way the section 1.2.64-73 remains consistent throughout. One must
make the rule that one remains anyway for 1.2.65, Vrddho yuna tallaksanas ced
eva visesah, etc. which concern speech units of differing forms. In these sutras
just one of speech units of different forms remains.
14Varttika 27 deals with the topic of number which is introduced in and recurs
throughout the technical discussion concerning the term 'ekavibhaktau' (if
there is one termination). See MB-K, vol. 1, p. 235, line 21-24, p. 236, lines 19-
20 and p. 237, line 27 - p. 238, line 1.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Dvyartho bahvarthas ca.

(Suitra-proponent:) It denotes two objects and many objects.
28. Naikdrthyam iti ced drambhanarthakyam.

Naikdrthyam iti ced ekasesdrambho 'narthakah sydt. Iha hi
sabdasya svdbhdviki vdnekdrthatd sydd vdcaniki vd. Tad yadi
tdvat svabhdviki
29. Asisya ekasesa ekenoktatvdt.
Asisya ekasesah.
(Sitra-rejecter:) If it is not so that one object is denoted,
there is no purpose in stating sutra 1.2.64.
If it were not so that it denotes one object, stating that one remains
would have no purpose. The fact that the speech form denotes
many objects is either natural or has to be explicitly stated. If it is
natural,
It should not be taught that one remains because one
speech form has already denoted the many objects.
It should not be taught that one remains.
Kim kdranam?
(Sutra-proponent:) What is the reason?
Ekenoktatvdt tasydrthasya dvitiyasya prayogena na bhavitavyam

uktdrthdndm aprayoga iti. Atha vdcanikf tad vaktavyam eko 'yam
avisisyate sa ca dvyartho bhavati bahvarthas ceti.
(Sutra-rejecter:) Because one speech form has already denoted all
of the many objects, usage of a second speech form denoting any
of them would not occur, "One does not use speech forms whose
objects have already been denoted."15 Alternatively, if it has to
be explicitly stated, it must be stated that one speech form remains
and it denotes two objects and many objects.16
15Kityayana lays the basis for this basic principle in his first vdrttika with the
phrase, "arthaprayukte sabdaprayoge (since it is the case that one uses words
prompted by the objects they denote)." See my discussion ?1.1. Patanijali
explains the principle more fully as follows: Arthagatyarthah sabdaprayogah.
Artham sampratydyayisydmiti iabdah prayujyate. Tatraikenoktatvdt
tasyarthasya dvitiyasya prayogena na bhavitavyam uktdrthdndm aprayoga iti.
The use of words is for the purpose of the comprehension of the objects they
denote. With the intention, "I will give the understanding of an object" a word
is used. This being the case, because one (word) has already stated it, usage of a
second (word) denoting that object would not occur, "One does not use words
whose objects have already been denoted." MB-K, vol. 1, p. 105, lines 2-3.
Patanijali cites the principle earlier under the present suttra MB-K, vol. 1, p. 238,
lines 10-11, and in many places throughout the Mahabhasya, for example, MB-
K, vol. 1, p. 252, lines 6-11; p. 370, lines 17-19; vol. 2, p. 199, lines 1-3. Many
collections of these kinds of principles include it. See Abhyankar (1967: 470,
index no. 12).
16The present suttra must add that the one remaining speech form denotes more
than one object. The suitra would say, "Of those (speech units) whose form
(ruipa) is the same just one remains and it denotes more than one object, if there

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Na vaktavyam. Siddham ekasesa ity eva.

(Sftra-proponent:) It doesn't have to be stated. It succeeds just
stating that one remains.17
Katham punar eko 'yam avisisyata ity anena dvyarthatd

bahvarthatd va sakya labdhum?
(Sutra-rejecter:) But how is one able to get by stating, "This one
speech form remains," that it denotes two objects or that it denotes
many objects?
Tac caikasesakrtam. Na hy antarena tadvdcinah sabdasya

prayogam tasydrthasya gatir bhavati. Pasydmas ca punar
antarendpi tadvdcinah sabdasya prayogam tasydrthasya gatir
bhavatfti agnicit somasud iti yathd. Te manydmahe lopakrtam
etad yendtrdntarendpi tadvacinah sabdasya prayogam tasydrthasya
gatir bhavatiti. Evam ihdpy ekasesakrtam etad yendtraiko 'yam
avasisyata ity anena dvyarthata bahvarthatd va bhavati.
(Sutra-proponent:) The provision that one remains achieves that
too. For without the use of the speech form denoting a certain
object, comprehension of that object does not occur. And yet we
see that even without the use of the speech form denoting a
certain object comprehension of that object occurs, as in 'agnicit
(one who has built an altar), somasut (one who has pressed
soma)'. We consider that deletion (lopa) achieves that in this case
even without the use of the speech form denoting a certain object
comprehension of that object occurs.18 Similarly here too the
provision that one remains achieves that in this case seeing, "This
is one termination. So says Kaiyata: Sa ca dvyartha iti. "Sarupadnam

ekaseso 'nekdrthas caikavihhaktdu" iti vaktavyam ity arthah. MB, vol. 2, p. 88a.
17Kaiyata explains: Siddham iti. Ekasesdrambhasdmarthydd ayam artho
gamyate. sisyamanah sabdo 'nekdrthdbhidhdyiti. Ekena hy
anekdrthdbhidhdanyaikasesah kriyate. It succeeds. By virtue of stating that
one remains this meaning is understood: the remaining speech form denotes
many objects. (The rule) that one remains is made for the very purpose that one
(speech form) may denote many objects. MB, vol. 2, p. 88a.
18Panini 3.2.90. Some sunah (bhute 84, kvip 87). When the subordinate word
'soma' occurs as direct object, the affix kvip occurs after the verbal root vlsuh
(to press).
Panini 3.2.91. Agnau cehi (bhute 84, kvip 87). When the subordinate word 'agni
(altar)' occurs as direct object, the affix kvip occurs after the verbal root v/ci (to
pile up).
Panini 3.1.93. Affixes provided for in the section beginning with 3.1.91,
Dhdtoh, which are not called tin (3.4.78, Tiptasjhi....) are called krt.
Panini 3.4.67. Kartari krt. (In general) affixes termed krt occur to convey the
agent (kartr).
Panini 6.1.67. Ver aprktasya (lopah 66). Zero (lopa) occurs in place of an
isolated vi.
The affix kvip occurs after the given root to convey the agent (3.4.67) and then
is replaced by zero.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
one speech form remains," we comprehend two objects or many

objects by it.19
Ucyeta tarhi na tu gamyeta. Yo hi gdm asva iti bruiydd asvam va

gaur iti na jdtucit sampratyayah sydt. Tendnekdrthdbhidhdne
yatnam kurvatdvasyam lokah prs.thato 'nugantavyah: Kesv
arthesu laukikdh kdan abdan prayuijata iti. Loke caikasmin vrksa
iti prayunjate dvayor vrkscv iti bahusu vrksd iti. Yadi tarhi loko
'vasyam sabdesu pramdanam kimartham ekasesa drabhyate?
(Sftra-rejecter:) It could be stated (that the speech form which
remains denotes two or many objects) but it would not be
understood.20 If one calls a cow a horse or a horse a cow there's
19The proponent does not use the verb 'achieves', literally, 'to make, do ("/kr)',
in its literal sense. A grammatical rule does not make a word or create its
relation with the object it denotes; it makes us understand what is the correct
speech form to denote a certain object. The word, the object it denotes, and the
relation between them are eternal according to Katyayana's very first varttika
(?1.1). Hence the rule for deletion does not make the word 'somasut' denote a
certain object; it makes us understand what the object is that it already denotes.
Similarly, the rule that one remains does not make one word denote more than
one object; it makes us understand that certain words do denote more than one
object.
For the foregoing reason Kaiyata explains that 'krta (made, done)' means made
understood: Tac caikasesakrtam iti. Ekasesavidhdnenaitat pratipaditam
ity arthah. That one remains achieves that too. The meaning is: The
provision that one remains makes it understood. Lopakrtam iti.
Lopadvdrenaitat samgrhftam ity arthah. Deletion achieves. It is
understood by means of deletion.
Nagesa makes it clear that the provision of the rule makes the usage understood:
Etat pratipaditam iti. Bhdsye krtam iti karotir ekasesavidhdnasdmarthya-
labdhapratipddane vartate, tena sabddrthasambandhanityatve 'pi na ksatir iti
bhdvah. Makes it understood. 'Done (krta)' in the text: 'to do' occurs in
the meaning "to make understood what is gotten by virtue of the provision that
one remains." In this way there is no harm to a word, its meaning and their
relation being eternal. That is what he has in mind.
Nagesa again expresses that Pataiijali's discussion concerns how to account for
the way we understand speech; grammar does not create that manner of
understanding: Bhdsye bahvarthata va bhavat!ti. ....Vidyamdnaiva
bhodhyate ity arthah. In the text, "or the fact of denoting many objects
occurs." ....The meaning is: being present is made known. MB, vol. 2, p. 88b.
I have translated the conclusion concerning ekasesa exactly parallel to the
example of lopa to which it is compared. My words, "by it," translate 'anena'
and refer to the one remaining speech form (anena si.syamdnena sabdena) which
the phrase, "This one remains," characterizes. Hence, I depart from Kaiyata who
takes the phrase, "This one remains," as stating the provision of the rule a
second time: Aneneti. Ekasesavidhdnena. By this. By the provision that
one remains. MB, vol. 2, p. 88b.
20The objection arises from taking the proponents use of the word 'achieves' (in
the text, 'krta) in its literal sense of to make. He objects that stating a rule does
not make the word have a meaning it did not have already even without the rule.
Since that is so, why bother to undertake the effort of making the rule?
Kaiyata comments: Ucyeteti. Vacandd anekdrthakdrydni pravartantam yathd
jdtydkhydydm ekasminniti, na tv ekdrthah sabda ekasesavidhdnend

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
no way that it would be understood.21 Hence with respect to one

word denoting many objects, one has to make an effort and
follow ordinary usage closely-to denote which objects do
ordinary speakers use which words? And in ordinary usage for
one tree they use 'vrksa (tree sg.)', for two 'vrksau (trees du.)', and
for many 'vrksah (trees pl.)'. So if ordinary usage is necessarily
the authority, for what purpose does Panini commence the
remainder of one?
Atha kimartham lopa drabhyate?

(Sutra-proponent:) Well, for what purpose does he commence
deletion?22
Pratyayalaksanam acdryah prarthayamdno lopam arabhata

ekasesdrambhe punar asya na kimcit prayojanam asti.
(Sutra-rejecter:) The teacher commences deletion seeking to
achieve operations caused by an affix where no affix is present,
but there is no reason for him to commence the remainder of
one.23
Nanu coktam pratyartham sabdanivesan naikendnekasydbhi-

dhdnam iti. Yadi caikena sabdendnekasyarthasydbhidhdnam sydn
na pratyartham sabdanivesah krtah sydt.
(Sutra-proponent:) But I have already said, "Because words apply
per object, it is not the case that one word denotes many
'nekdrthah sampadyata ity arthah. It could be stated. The meaning is:

From the explicit statement (it is understood) let the operations which pertain
to (a speech form) denoting many objects apply (to the one remaining), just as
(from the statement of) 1.2.58, (the operations which pertain to a speech form
denoting many objects apply to a speech form denoting one object if it denotes
a generic property). But a speech form denoting one object does not come to
denote many objects by the provision that one remains. MB, vol. 2, p. 88b.
21Nagesa explains: Sampratyayah syad iti hhdsye. Asvatvddipramd sydd
ity arthah. (There's no way that (na jdtucit)) cognition of (the cow)
being a horse would arise. MB, vol. 2, p. 88b.
22Nagesa: bhdsye atha kimartham lopa iti. Agnicid ity ato lokata eva
kartrarthapratttisiddher iti bhdvah. Well, for what purpose is deletion.
Cognition of the meaning 'agent' succeeds from 'agnicit' just by virtue of
ordinary usage. That is what he has in mind. MB, vol. 2, p. 88b.
23kvip, which is provided (3.2.90-91) and then deleted (6.1.67) in the
derivation of 'somasut' and 'agnicit' (See note 18), is a krt affix marked with p.
By virtue of 1.1.62, 6.1.71 provides the augment tuk, even though kvip has
been deleted.
Panini 1.1.62. Pratyayalope pratyayalaksanam. An operation conditioned by an
affix applies even if the affix has been deleted.
Panini 6.1.71. Hrasvasya piti krti tuk. The final augment tuk is added to a
short vowel before a krt suffix marked with p.
Nagesa: Pratyayalaksanam iti. Anena lopendnutpatter evdnvdkhydnam ity
uktam. Operations caused by an affix. By this he states that lopa is just
an explanation that (the missing element) doesn't arise. MB, vol. 2, p. 89a.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
objects."24 And if it were the case that one word denotes many
objects, it would not be true that words apply per object.
30. Pratyartham sabdanivesad ekenanekasydbhi-

dhdndd apratyartham iti cet tad api pratyartham eva.
Pratyartham sabdanivesad ekendnekasydbhidhdndd apratyartham
iti ced evam ucyate, yad apy ekendnekasyabhidhdnam bhavati tad
api pratyartham eva. Yad api hy arthdv arthau prati tad api
pratyartham eva. Yad api hy arthdn arthan prati tad api
pratyartham eva.
(SQtra-rejecter:) If you say this: "From the fact that
words apply per object (it follows that it is not the case that
one word denotes many objects), hence from the fact that
one word denotes many objects it would follow that it
is not true that words apply per object," (I answer:) That
too is precisely per object.
If you say this: "From the fact that words apply per object (it
follows that it is not the case that one word denotes many objects),
hence from the fact that one word denotes many objects it would
follow that it is not true that words apply per object," even one
word denoting many objects is still precisely per object. That
which is per two objects is still per object, that which is per many
objects is still per object.25
Ydvatdm abhidhdnam tdvatdm prayogo nydyyah. Ydvatdm

arthdndm abhidhdnam bhavati tdvatam sabddndm prayoga ity esa
pakso nyayyah.
(Sutra-proponent:) It is reasonable that one uses as many as are
denoted. This view is reasonable, that one uses as many words as
there are objects denoted.
31. Ydvatdm abhidhdnam tdvatdm prayogo nydyya iti

ced ekendpy anekasyabhidhanam.
Ydvatdm abhidhdnam tdvatdm prayogo nyayya iti ced evam
ucyate, eso 'pi nyayya eva yad apy ekendpy anekasyabhidhdnam
bhavati.
24See vdrttika 1.
25Kaiyata: Tad api pratyartham eveti. Artho hy abhidheyam ucyate. Yas
ca dvyarthah iabdo ndsau dvyarthatam jahdti bahvarthas ca bahvarthatdm....
That too is precisely per object. An object (artha) is said to be that
which is denoted (abhidheya). And a word which denotes two objects does not
depart from denoting two objects, nor does one which denotes many objects
(depart from) denoting many objects.
Nagesa: Tad apiti. Sviyasviydrthdtydga eva pratyarthatvam iti bhdvah. That
too. Being per object is just not abandoning whatever its own denoted object
is (regardless of how many are its own). That is what he has in mind. MB, vol. 2,
p. 89b.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
(Sutra-rejecter:) If you say: "It is reasonable that one

uses as many as are denoted," (I answer:) one denotes
many too.
If you say this: "It is reasonable that one uses as many words as
there are objects denoted," (I answer:) this is reasonable too, that it
is also the case that one word denotes many objects.26
Yadi tarhy ekendnekasydbhidhdnam bhavati plaksanyagrodhau

ekenoktatvdd aparasya prayogo 'nupapannah. Ekenoktatvdt
tasydrthasydparasya prayogena na bhavitavyam.
(Sltra-proponent:) If it were the case that one word denotes
many objects, in the dvandva compound 'plaksa-nyagrodhau (the
Plaksa and the Nyagrodha species of Fig trees)', because one word
has already denoted the object denoted by the other, usage of the
other would not make sense. Because one word has already
denoted the object, usage of another word for that object would
not occur.27
26Kaiyata: Ekendpiti. Sahavivaksayam eka eva sabda

tadbhutdvayavabhedam samuddyam abhidhatta ity arthah. One also. The
meaning is: When (the speaker) desires to express (the objects) together
(sahavivaksdyam), just one word denotes an aggregate (samuddya) the parts of
which are distinct. MB, vol. 2, p. 89b.
Patanijali discusses sahavivaksd under vdrttika 19 of the present suitra. He says:
Sahavivaksayam ekasesah. One remains when (the speaker) desires to express
(the objects) together. He then explains that the usage, "vrksas ca vrksas ca (tree
and tree)," is not equivalent to the usage, "vrksau (trees (dual))." In the former
there is no desire to express the objects together; in the latter there is. The facts
of usage are that these different usages appear to express their respective
meanings and each occurs only in its own meaning. We know which forms are
used in which meanings from ordinary usage. Hence there is no need to state the
rule 1.2.64, that one remains. See MB-K, vol. 1, p. 238, lines 2-5.
27A dvandva compound also gives simultaneous expression (yugapadvacanatd)
to more than one object either as the mutual relation of individual objects
(itaretarayoga) or as an aggregate (samdhdra). In the course of discussing the
role of the word 'ekavibhaktau' in the present sutra, Patanjali considers 1.2.64 as
an exception to the former type of dvandva compound. Under Panini 2.2.29,
Cdrthe dvandvah (MB-K, vol. 1, pp. 430-35), and 2.4.12, Vibhasa
vrksamrgatrnadhdnyavyaijanapasusakunyasvavadavapurvdparddharottardndm
(MB-K, vol. 1, p. 475, line 15 - p. 476, line 5 and p. 476, line 23 - p. 477, line 5),
Patafijali discusses the meaning condition for the formation of a dvandva
compound and the example 'plaksanyagrodhau'. It conveys the two distinct and
different objects in relation. (The Plaksa is different from the Nyagrodha, and
the dual compound is not a samdhdra. 2.4.12, vdrttika 1, Bahuprakrtih
phalasendvanaspatimrgaiakuntaksudrajantudhanyatrndndm, limits ekavad-
bhdva to dvandvas denoting many of certain objects (bahuprakrti) including
trees (vanaspati) and prevents it applying to duals) 'Plaksau', however,
expresses together two distinct objects which are yet the same.
See varttika 12 under this sutra: Dvandvapratisedhas ca (And a negation (to
prevent the formation) of a dvandva compound would have to be stated). MB-K,
vol. 1, p. 234, lines 24-26, p. 235, lines 18-21, p. 236, lines 1-5 and p. 238, lines
5-12. Also , p. 236, line 20, and p. 237, line 27 - p. 238, line 1.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Kim kdranam?
(Sutra-rejecter:) What is the reason?
Uktdrthdndm aprayoga iti.

(Suitra-proponent:) "One does not use words whose objects have
already been denoted."28
32. Ekenoktatvdd aparasya prayogo 'nupapanna iti

ced anuktatvat plaksena nyagrodhasya nyagrodha-
prayogah.
Ekenoktatvdd aparasya prayogo 'nupapanna iti ced anuktah
plaksena nyagrodhdrtha iti krtvd nyagrodhasabdah prayujyate.
(Sutra-rejecter:) If you say: "Because one word has
already denoted the object denoted by the other,
usage of the other would not make sense," (I answer:)
there is usage of the word 'nyagrodha' because the
word 'plaksa' has not denoted the object denoted by
'nyagrodha'.
If you say: "Because one word has already denoted the object
denoted by the other, usage of the other would not make sense, (I
answer:) the word 'nyagrodha' is used based on the fact that the
word 'plaksa' has not denoted the object denoted by
'nyagrodha'.29
Katham anukto ydvateddntm evoktam ekendpy

anekasydbhidhdnam bhavattti?
(Sitra-proponent:) How has it not denoted it since just now y
said that it is the case that one denotes many too.30
Sarutpdndm ekendpy anekasydbhidhdnam bhavati na viriupadn

(Sutra-rejecter:) For words whose form is the same it is the
that one word denotes many too, not for words whose form
different.
Kim punah kdranam saruipdndm ekendpy anekasydbhidhd

bhavati na punar virupanam?
(Sutra-proponent:) But what is the reason that for words wh
form is the same it is the case that one denotes many too but
for words whose form is different?
33. Abhidhdnam punah svdbhdvikam.

Svdbhdvikam abhidhdnam.
34. Ubhayadarsanac ca.
28Concerning this principle, see varttika 29 note 15.

29Equals vdrttika 14 and bhasya under Panini 2.2.29. MB-K, vol. 1, p. 433,
lines 17-21.
30See vdrttika 31 and bhasya.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Ubhayam khalvapi dirsyate. Virupdcnam apy ekenanekasyabhi-

dhanam bhavati. Tad yatha dyava ha ksdmd. Dyava cid asmai
prthivf namete iti. Viriupana.m kila ndmaikendnekasydbhidhanam
sydt kim punah saruipdnam.
(Sitra-rejecter:) And denotation is natural.
Denotation is natural.31
And because we see both.
Both are seen. Even for words whose form is different it is the
case that one denotes many, for example, "Dyacvd ha ksama...
(May heaven and earth...),"32 "Dyavd cid asmai prthivi namete...
(Heaven and earth bow to him...).33 When for words whose form
is different it is possible that one denotes many, how much more
so is it the case for words whose form is the same?34
31Equals vdrttika 15 and bhdsya under Panini 2.2.29. MB-K, vol. 1, p. 433,
lines 24-25. The fact that denotation is natural is often cited to explain why
certain speech forms allowed by the grammar do not occur. See 3.2.1, vdrttika 5,
Anabhidhdndt. MB-K, vol. 2, p. 94, line 14. See also Cardona (1988: 645).
32Rgveda 10.12.1. RVS, vol. 4, p. 303. Atharvaveda 18.1.29 (p. 342).
33Rgveda 2.12.13. RVS, vol. 2, p. 50. Atharvaveda 20.34.14 (p. 407). Pataiijali
cites both mantras again under 2.2.29, vdrttika 6, and the first a third time under
2.2.29, vdrttika 15. MB-K, vol. 1, p. 432, lines 4-5; p. 434, line 4. Reference:
Rau (1985: 40, #346, #347). 'Dyavd' and 'ksamd' in the first example, and
'dyavd' and 'prthiv{' in the second example, are vedic duals. (The Padapatha
gives, "prthiv{ iti." The 'iti' tells that the final long f is pragrhya, and according
to 1.1.11 a dual form in final i is termed pragrhya. A nominative singular is not
so termed.) In vedic, according to 7.1.39, a occurs in place of the expected dual
termination au, and by 6.1.106 a long vowel homogenous with the final vowel
of the base optionally occurs to prevent the ordinary dual termination au. In the
singular the former (dyauh) means heaven and the latter (prthivi) means earth.
The dual makes either word refer to both heaven and earth. Hence one of two
words whose forms are different refers to both objects.
Panini 1.1.11. Iduided dvivacanam pragrhyam. A dual speech form ending in
long ', u, or e is termed pragrhya.
Panini 6.1.106. Vd chandasi (dirghah 101, puirvasavarnah 102, na ici 104,
dirghdt jasi ca 105). In vedic after a long vowel, before a nominative or
accusative termination beginning in a vowel other than a and the nominative
plural termination jas, a long vowel homogenous with the preceding vowel
optionally does not occur.
Panini 7.1.39. Supdm sulukpiirvasavarndccheyddddydydjdlah. In vedic the
following substitutes occur in place of the nominal terminations, sup: su, luk, a
sound homogenous with the preceding, a, at, se, yd, da, .dy, ydc, dl.
34Kaiyata sums up the point of the section: Tad evam dravydbhidhdnapakse 'pi
pratydkhydta ekasesah. Anekdrthaprakrame anekdrtha eka eva vrksasabdah
prayujyate, na tu dvandvah, anabhidhdndt. In this way the rule that one remains
is rejected as unnecessary even if words denote individual substances (dravya).
To convey many objects (trees) just one word 'v.rksa (tree)' denoting many
objects is used, not a dvandva compound because it is not used to denote that
meaning (anabhidhdna).
Nagesa comments: Idam ca pratydkhydnam arthasyanyapramdnatvdd ity
anena sutrak,rtdpi darsitaprdyam eva. Nanv evam api dvandvo durvdro 'ta dha:
na tv iti. Sahavivaksdydm ekajdtiydrthdnekasabdaprayogddarsanena tad-
asddhutvapratipddandrtham suttram iti bhdvah. And the sutra author himself as

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
35. Akrtyabhidhandd vaikam vibhaktau

Vdjapyayanah.
Ak.rtyabhidhdndd vaikam sabdam vibhaktau Vdjapydyana acaryo
nyayyam manyate. Ekdk.rtih sd cdbhidhlyata iti.
(Sftra-rejecter:) Alternatively, because it denotes a class
property, Vajapyayana holds that there is one speech
form before an inflectional termination.
Alternatively, because it denotes a class property, the tea
Vajapyayana considers one speech form before an inflect
termination proper: There is one class property and i
denoted.35
Katham punar jndyata ekdkrtih sd cdbhidhlyata iti?

(Sutra-proponent:) But how does one know that there is one
property and it is denoted?
36. Prakhydvisesat.
Na hi gaur ity ukte visesah prakhydyate sukla nlla kapila
kapotiketi.
(Sutra-rejecter:) Because there is no difference in the
cognition.
When the word 'cow' is uttered, one does not apprehend a
difference: a white cow, a blue one, a brown one, a grey one.36
much as shows this rejection by (saying in sutra 1.2.56) "something other than
grammatical rules is the authority concerning meaning." But even so it would
be difficult to avoid a dvandva compound occurring; hence, he says, "But
not." Because we do not see many words used for objects of one kind if the
speaker desires to express them together (sahavivaksdydm), the present sutra
(1.2.64) is for the purpose of expressing that such usage is incorrect. That is
what he has in mind. MB, vol. 2, p. 90b.
35Kaiyata comments: Iddnim dkrtipaksdsrayena pratydkhydnam karoti: dkrty-
abhidhdndd iti. Jdtih sabddrthah, tasyd ekatvdd anekasabdaprayogaprasanga
eva ndsti kim ekasesenety arthah. Now he shows 1.2.64 to be useless by
resorting to the view that a word denotes a class property: because i t
denotes a class property. The meaning is: A generic property is the
object denoted by a word. Because it is one, the possibility of using more than
one word doesn't even arise so of what use is the rule that one remains? MB,
vol. 2, p. 90b.
36Kaiyata comments: Prakhydvisesdd iti. Prakhyd buddhih, tasyd
avisesdd ekaruipatvdt tadvisayasyapy aikyam pratiyate. Gunapramdnddi-
bhinnesv api gopindesu gaur gaur ity ekdkdrapratyayodaydd avasyam
ekendlambanena sdmdnyena bhdvyam iti jdtisadbhdva ekatvam cavaslyate.
Gaur ity ukta iti. Gaur ity etena sabdenokte pratydyite samanyalaksane 'rthe
visesdnavadhdrandd aikyam sdmdnyasydvastyata ity arthah. Because there
is no difference in the cognition. Cognition, awareness, because
there is no difference in it (because it is of one form) we apprehend that its
object is one too. Because a single type of cognition 'cow' arises with respect to
each of many cow-objects having different qualities, sizes, etc. there absolutely
has to be a single general property as its basis. Hence it is determined that a
generic property exists and is one. When (the word) 'cow' is uttered. The
meaning is: When the word 'cow' states, conveys, an object whose character is

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Yadyapi tdvat prakhydvisesdj jndyata ekdkrtir iti kutas tv etat

sdbhidhTyata iti?
(Sftra-proponent:) Well, even if one knows that there is one class
property because there is undifferentiated cognition, what is the
basis for asserting that the word denotes it?
37. Avyapavargagates ca.

Avyapavargagates ca manydmaha dk.rtir abhidhtyata iti. N
gaur ity ukte vyapavargo gamyate sukld nlld kapild kapotiket
38. Jidyate caikopadistam.
Jiadyate khalvapy ekopadistam. Gaur asya kaddcid upadisto
bhavati. Sa tam anyasmin dese 'nyasmin kale 'nyasyam ca
vayo'vasthaydam drstvd jdndty ayam gaur iti.
(Sutra-rejecter:) And because one understands that there
is no setting apart.
And because one understands that there is no setting apart, we
consider that the class property is denoted. When the word 'cow'
is uttered, one does not apprehend a difference: a white cow, a
blue one, a brown one, a grey one.37
And taught once, it is known.
Moreover, taught once, it is known. At some time a bull is
pointed out to him. Seeing one in another place, at another time,
at another stage of life, he knows that it is a bull.38
a general property, because one does not determine any difference one
concludes that the general property is a single thing. MB, vol. 2, pp. 90b-91a.
Filliozat (1975-86: vol. 4, p. 333) and Subrahmanya Sastri (1944-62: vol 4, p.
163) translate 'sukla', etc. as substantives denoting the qualities white, etc,
rather than as adjectives. As such they refer to the specific properties to which
the word 'visesa' previously is taken to refer. But see Vatsyayana on Gautama
2.2.60, example 8, varna, "iukld gauh kapild gaur iti."
37Kaiyata: Avyapavargagates ceti. Avyapavargo 'bhedah avicchedah
avisesas tasya gatih pratltir ity arthah. Kaiyata then clarifies the difference
between vdrttikas 36 and 37: Atha prakhydvisesad avyapavargagates
ceti kimartham ubhayor updddnam, ekendpi hetunaikatvdbhidhdnayoh
siddhatvdt? Naisa dosah. Anabhidhiyamdndpi jdtih samnidhimdtrena prakhyd-
'visese nimittam bhavatfti prakhyd'visese[na] pratyabhijidpratyayaruipenaika-
tvam eva pratipdditam ndbhidhdnam. Now for what purpose are both "Because
the cognition is not different" and "And because (one)
understands (that there is) no setting apart" stated-one is cause en
to achieve both unity and denotation? This fault does not arise. Even
being denoted, just by being present, a generic property is a cause of
being not different; hence, cognition being not different, in the fo
awareness which is a recognition (pratyabhijid-pratyaya), explains onl
not denotation.
Nagesa clearly sums up the point of vdrttikas 36 and 37: Gunddibhinnesv
anusyutaikdkdrapratftyd jdtisiddhau tatra saktigrahah. Sabddt tasyd bodhas ca.
Since the recurrent uniform cognition of objects having different qualities, etc.
proves that there is a generic property, one locates the signifying capacity in it.
And cognition of it arises from the word. MB, vol. 2, p. 91a.
38Sivaramendrasarasvati: Ekopadistam iti. Ekavdram upadistam sakrd-
upadistam ity arthah. Against Kaiyata's two alternatives Sivaramendrasarasvati

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Kah punar asya vis'esah prakhydvisescd ity atah?

(Sutra-proponent:) But how is this (varttika 38) different from
"Because there is no difference in the cognition (varttika 36)"?
Tasyaivopodbalakam etat. Prakhydvisesdj jniyate caikopadistam

iti.
39. Dharmasastram ca tatha.
Evam ca krtvd dharmasastram pravrttam. Brdhmano na
hantavyah surd na peyeti brdhmanamdtram na hanyate
surdmdtram ca na plyate. Yadi dravyam paddrthah sydd ekam
brdhmanam ahatvaikdm ca surdm apltvdnyatra kdmacdrah sydt.
(Sutra-rejecter:) This (vdrttika 38) supports that (varttika 36):
Because there is no difference in the cognition, and taught once, it
is known.39
most closely follows Patafijali's own comment here. MPV, vol. 4, p. 111.
Although Patafijali uses masculine forms here (ayam, tam) and feminine
adjectives under vdrttikas 36 and 37 (sukla, etc.), in both places he intends the
word 'gauh' to refer to any generic bovine animal without regard to gender or
any other subclassification. Whenever I use the word 'cow', I intend it in the
same generic sense. Patafjali uses the masculine form generically in accordance
with Sanskrit usage as codified in 1.2.67, Pumdn striyd. In English usage, 'cow'
may be used either generically or for the female; 'bull' occurs only for the male.
39Kaiyata: Prakhydvisesdd ity anena pratyaksam jdtydlambanam pramdnam
upanyastam. Tato vipratipannapratipddandyanumanam iha jatisadbhdve
jidayate caikopadistam ity anenoktam, desakalavasthdpindantaresv
abddhitapratyabhijinpratayodaydnyathdnupapattyd sdmdnyasadbhdvo
'numlyate ity arthah. (Vdrttika 36) "Because there is no difference i n
the cognition" presents the direct perception whose basis is a generic
property as evidence. Then, in order to explain to one who suspects error,
(varttika 38) "And taught once, it is known" states an inference here for
the existence of a generic property: One infers the existence of a general
property because the occurrence of an uncontroverted recognition-awareness
(pratyabhijni-pratyaya) with respect to different bodies at different places,
times and stages, does not make sense otherwise (i.e., if there is not a single
generic property). That is the meaning.
Kaiyata states this after giving the following interpretation of the term
'upodbalaka': udgatam balam pratyaksam tatpirvakatvdd anumanasya
udbalasya samtpam anumanam ity arthah. The meaning is: inference is that
which is near to that whose strength is greater (udbala); that whose strength is
greater is direct perception, because inference is preceded by it. Then he offers a
second interpretation of the term: Atha vd balam uddlpayaty upodbalakam
upabrmhakam ucyate. Or being that which lights up the strength (of x) an
upodbalaka is called a support (upabrmhaka).
Nagesa: Prayogas tu: Vimatapratyayo jativisayo desidibhede 'py
abddhitaikdkdrapratyayatvdt sammatavad iti. The statement of the inference is:
The doubted cognition (which vdrttika 36 presents as evidence) does have a
generic property as its object because it is an uncontroverted cognition whose
form is the same even when the place, etc. are different, just like the cognition
which it is agreed (has a generic property as its object) (the topic of varttika 38).
... Upabrmhakam iti. Bhrdntatvsdankdnivartanena pramdtvanirvdhakas tarka
ity arthah. An argument to establish that it is correct knowledge by removing
the doubt that (the direct perception) is erroneous. MB, vol. 2, p. 91b.
Sivaramendrasarasvati: Upodbalakam iti. Sahakdrlty arthah.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
And the instruction of dharma (dharmasastra) is in this

manner.
And the instruction of dharma proceeds on this ba

denotes a class property): The instruction, "One s
brdhmana. One should not drink liquor," means: o
any brahmana and one doesn't drink any liquor. I
substance were the object denoted by a word, r
killing one brdhmana and abstaining from drinki
one could do as one pleased with the rest.40
Kah punar asya viseso 'vyapavargagates cety atah.
Sivaramendrasarasvati rejects Kaiyata's statement that vdrttika 36 offers the

evidence of direct perception. He holds that undifferentiated cognition with
respect to many individual objects is grounds (hetu) for the inference
(anumdna) that there is one generic property; it is not direct perception of the
generic property. Hence both vdrttika 36 and 38 offer inferential evidence, and
the latter is a helper (sahakdrin) of the former. MPV, vol. 4, p. 111 bottom - 1 12
top.
Kaiyata expresses the view that perception is based in reality. If we have one
undifferentiated perception, there is one object at its basis which is the object of
that one perception. See the translation of Sabarabhasya on Jaimini 1.1.5.
40Dharmasastram ca tatheti. anenaitad darsayati: na sdd.rsyaikdrthakriyd-
kdritvddinimittd bhrdntd pratyabhijnd sarvatreti grahyam, kim tarhi, abhinna-
visayanimittd smrtikdrdndm apy avigftapramdnabhdvasmrtopanibaddha-
vacandnam jdtydirayena sarvavyavahdrapravartandd ity arthah. And t h e
instruction of dharma (dharmas'astra) is in this manner. By this he
shows the following: One should not accept that recognition (pratyabhijna) is
everywhere erroneous, caused by the fact that objects are similar in that they
perform the same purposeful activity. Rather, it is caused by a single object
because, in addition, all the usage of the authors of smrti texts, whose
statements based on received tradition have undisputed authority, proceeds by
resorting to generic properties. That's the meaning.
Nagesa: Dravyasya sabddrthatve hy ekabrdhmanavadhanisedhady eva pratlyeta.
Ekah ka iti cet, yo yasya tadvdkyasravanakdle buddhisthah sa iti g,rhna. Jdteh
paddrthatve tu na dosa iti tdtparyam. Anyatra kdmacara iti. Sarvesdm iti
sesah. Ekenaikasydhanane jdtah sdstrdrtha ity anyabrdhmanavisaye sarvesdm
kdmacdrah sydd ity arthah. The point is: If an individual substance were the
object denoted by a word, one would understand a prohibition against slaying
just one brdhmana, and so on. If you ask "Which one?" take that for each hearer
it is he who comes to mind at the time of hearing the statement. But if a generic
property is the object denoted by a word the fault doesn't arise. As one
pleased with the rest. Add "Everyone." When one person doesn't kill one
(brdhmana) the purpose of the instruction is achieved. Hence, concerning other
brdhmanas everyone could do as he pleased. That's the meaning. MB, vol. 2, p.
92a.
Patafjali discusses these examples again under 6.1.84, vdrttikas 4-5. MB-K,
vol. 3, p. 57, line 7 - p. 58, line 17; especially p. 57, lines 16-19, and p. 58, lines
3-5. Manusmrti 11.93 says, "na surdm pibet," and 11.94 has "surd... na pdtavyd."
See 9.235 and 11.55. Buhler, pp. 383 and 441, and his "Synopsis of parallel
passages." Manusmrti, p. 318. The Kathakasamhita 12.12 (p. 117) says,
"Brdhmanah surdm na pibati."

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
(Sutra-proponent:) But how is this (varttika 39) different from

"And because one understands that there is no setting apart"
(varttika 37)?
Tasyaivopodbalakam etat. Avyapavargagates ca dharmasastram

ca tatheti.
(Sutra-rejecter:) This (varttika 39) supports that (varttika 37):
And because one understands that there is no setting apart, and
the instruction of dharma is in this manner.41
[40a.] Asti caikam anekadhikaranastham yugapat.
Asti khalvapy ekam anekddhikaranastham yugapad upalabhyate.
And it is the case that one thing is present in many
substrata (adhikarana) simultaneously.
Moreover it is the case that one thing is found present in many
substrata simultaneously.42
Kim?
(Sutra-proponent:) What is?
Adityah. Tad yathd. Eka Adityo 'nekddhikaranastho yugapad

upalabhyate.
(Sutra-rejecter:) The sun. For example, the one sun is found
located in many places (adhikarana) simultaneously.43
Visama upanydsah. Naiko drastddityam anekddhikaranastham

yugapad upalabhate.
(Sutra-proponent:) The example given is not the same. It is not
the case that one observer sees the sun present in many places
simultaneously.44
41Kaiyata: Tasyeti. Avyapavargagatir abhrdntd, smrtikdrdnam api tathaiva

vyavaharadarsandd ity arthah. "This..." The meaning is: understanding (that
there is) no setting apart is not erroneous because we see that the usage of
authors of smrti texts is in this manner (their usage assumes that words denote
generic properties). MB, vol. 2, p. 92a.
42Kaiyata: Nanv ekam anekastham katham bhavati, na hi Devadatto yugapan
Mathurdydm Srughne ca bhavatity ata dha: asti ceti. But how is one thing
present in many? After all Devadatta is not simultaneously in Mathura and
Srughna. MB, vol. 2, p. 92a. Kaiyata anticipates the objections to the notion of
a class property raised in the argument to establish that a word denotes an
individual substance. See Patafijali on varttika 48.
43Patafijali gives this passage (including the example of the sun, and its
rebuttal in the following line) in commenting on Sivasutra 1, varttika 12. MB-
K, vol. 1, p. 18, lines 16-19. Sivaramendrasarasvati says: Ekasminneva pradese
vartamdna ddityo ndnddesasthair yugapad drsyata iti.... People present in
various places simultaneously see the sun which is present in just one place.
MPV, vol. 4, p. 115.
44Under Sivasutra 1, vdrttika 12, the objection appears as follows: Naiko
drastddityam anekddhikaranastham yugapad desaprthaktves upalabhate. It
includes the additional word, 'desap,rthaktvesu (in different places)'. It is not a
single observer in different places who sees the sun present in many places

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Evam tarhi
40b. Itindravad visayah.
Tad yathd. Eka indro 'nekasmin kratusata dhuto yugapat sarvat
bhavati. Evam adkrtir api yugapat sarvatra bhavisyati. Avasyam
caitad evam vijneyam ekam anekddhikaranastham yugapad
upalabhyata iti.
41. Naikam anekadhikaranastham yugapad iti cet
tathaikasese.
Yo hi manyate naikam anekddhikaranastham yugapad
upalabhyata ity ekasese tasya dosah sydt. Ekasese 'pi naiko
vrksasabdo 'nekam artham yugapad abhidadhlta. Avasyam caitad
evam vijneyam dkrtir abhidhlyata iti.
42. Dravydbhidhane hy dkrtyasampratyayah.
Dravydbhidhdne saty dkrter asampratyayah sydt.
(Sutra-rejecter:) If that is so, then
The domain is like that of 'Indra'.45
Just as one Indra, invoked in many hundreds of Vedic
performances, is simultaneously everywhere, thus a class property
will be simultaneously everywhere too.46 And it absolutely has to
simultaneously. MB-K, vol. 1, p. 18, line 18. See ?1.4.1.

45With Kayata's suppletion of the sentence, this could be translated: (One is
related to more than one) object as 'Indra' is. With Nagesa's alternative
reading of Kayata's suppletion: One object (is present in many places) as
'Indra' is.
Kaiyata: Itindravadvisaya iti. Ekasydpy aneka iti vdkyas'esa
related to more than one) object as 'Indra' is. The remainder of
is: even one is related to more than one.
Ndgesa: Ekasydpy aneka iti pdthah. Ekasydpindrasabdasya yathdneko ydgo
'ngitvena visaya ekabodhajidnavisayas ca, ekadesivacchinndnekaydgesu tathd
darsanasambhavdt tathaikasyd api jdter aneko visaya asrayatayety arthah.
Anekastha iti pdthe eko 'pi yathd indrasabdo 'nekaydgastha ekasydpi boddhur
visayas tathd jdtir aplty arthah. The reading is: even one is related t o
more than one. The meaning is: Just as even one word 'Indra' is related to
more than one Vedic performance, which is a related object (visaya) as the
principal (action to which 'ndra' is subordinate) and is also related as the
(common) object of one cognition (of both the word 'Indra' and the Vedic
performance), similarly-because it is possible to see many Vedic performances
in one place-similarly, even one generic property is related to the many
objects which are its substrata. If the reading is, "present in many places," just
as even one word 'ndra' present in many Vedic performances, is an object for
one perceiver; similarly, a generic property is too. MB, vol. 2, p. 92b.
Panini 2.1.6 provides for the indeclinable compound 'itindra'. See vdrttika 1
note 10. According to 5.1.115, the suffix vat follows to form the indeclinable
'itindravat' meaning "like the word 'indra'. The Kasika (vol. 1, p. 101) on 2.1.6
says: Sabdaprddurbhavah prakdiatd sabdasya: itipdnini tatpdnini.
Pdninisabdo loke prakadate ity arthah. A speech unit's manifestation the
being known of a speech unit: itipdnini tatpdanini. The meaning is: The speech
unit (name) 'pdnini' is well known in the world.
Panini 5.1.115. Tena tulyam kriyd ced vatih. After a word syntactically
connected with a word x ending in a third triplet nominal termination, the affix
vatih occurs to mean "like x", if that which is alike is an action.
46Kaiyata: Sarvatreti. Sarvesu yagesu dravyavad angatdm pratipadyata ity

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
be understood this way, that one thing is found present in many

substrata simultaneously.
If (you say:) it is not the case that one thing is present
in many substrata simultaneously, (I answer:) likewise if
one remains.
A fault arises for him who thinks that it is not the case that one
thing is present in many substrata simultaneously. Even if it were
the case that one remains, one word 'tree (vrksa)' would not
denote many objects simultaneously.47 And it absolutely has to
be understood this way, that a class property is denoted.
arthah. Everywhere. The meaning is: In all the Vedic performances (the word
'indra' or Indra) attains to being a subordinate part, like a substance.
Nagesa: Prddurbhavavisayendrasabdaruipd tadbodhyd vd devatd vrihyddi-
dravyavat sarvaydgesv angatdm yugapat pratipadyata ity arthah. Bhdsye 'pi
Indrah indrasabdah. Ahutah prddurbhuitah. Yugapat sarvatrdagam bhavatity
arthah. The meaning is: The deity, which is either the manifesting word 'indra'
or that which it conveys, simultaneously attains to being a subordinate part of
all the Vedic performances, like the substances rice, etc. In the bhdsya too,
Indra is the speech unit 'indra'. Invoked manifested. The meaning is:
participates in every performance simultaneously. MB, vol. 2, p. 92b.
Sivaramendrasarasvati takes the view that a class property is compared to the
diety Indra, not the speech unit 'indra' which names him: Itindravad iti.
Itlndreti sabdaprddurbhdve 'vyaylbhdvah, tasmdd arthaprakdsalaksakdt 'tena
tulyam' iti vatih.... Purdnddiprasiddham idam: Devatd sarvatra kratusu
samnihiteti. As 'Indra' is. The indeclinable (compound) 'itindra' occurs
meaning the manifestation of a speech unit. It secondarily refers to the
appearance of the object the speech unit denotes. After the (compound 'itfndra',
secondarily referring to the appearance of the deity Indra), the suffix vat occurs
according to 5.1.115.... This is well attested in the Puranas, etc., that the deity is
present in all the Vedic performances. MPV, vol. 4, p. 112.
On the view that a deity is a speech form, the word 'invoked (dhuta)' must be
interpreted to mean manifested. On the view that a deity is denoted or named by
a speech form, the word 'invoked' makes sense in its usual meaning.
Patafijali presents the view that it is one and the same speech form which is
enunciated at various places and times in commenting on Sivasutra 1, vdrttika
12. MB-K, vol. 1, p. 18, lines 19-20. Sabara presents the same view concluding
his comment on Jaimini 1.1.15. Adityavad yaugapadyam. See ?3.2.2, note 12,
and ?1.4.1. In the Mahabhasya, however, the explanation of the awareness of
one speech form in many individual enunciations is not content with this
solution but offers two further solutions: Vdrttika 13, Akrtigrahandt siddham,
and vdrttika 16, Ripasdmdnydd vd. These solutions are the two debated in the
present discussion. The former accepts a generic property in many individuals;
the latter does not. See MB-K, vol. 1, p. 18, line 20 - p. 19, line 8.
Note that vdrttika 40 is parallel to vdrttika 56 in wording. This has raised the
question of interpolated text. Jacobi (1929: 151) considers that vdrttika 40
with its commentary is added later, as is possibly vdrttika 41, so that vdrttika 42
follows vdrttika 39 with a discussion on the same topic.
47Kaiyata: Ekasese 'piti. Yady ekasydnekasambandho ndbhyupagamyate
tadaikah sabdo 'nekam artham sambandhdbhdvdn na pratydyayed ity arthah.
Even it it were the case that one remains. The meaning is: If it is not
accepted that one thing has many relations, then one word would not give the
understanding of many objects because of the absence of any relation (to them).
MB, vol. 2, p. 92b.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Because if it were the case that a word denotes an

individual substance, there would be no cognition of
the class property.
Tatra ko dosah?
(Suitra-proponent:) What is the fault if that were so?
43. Tatrdsarvadravyagatih.
Tatrdsarvadravyagatih prdpnoti.
(Sutra-rejecter:) If that were so, one would fail to
understand all the individual substances.
If that were so, failure to understand all the individual su
would obtain.
Asarvadravyagatau ko dosah?
(Sutra-proponent:) What is the fault if one fails to understand all
the individual substances?
Gaur anubandhyo 'jo 'gnlsomya iti. Ekah sastroktam kurvitdparo

'sdstroktam. Asastrokte ca kriyamane vigunam karma bhavati
vigune ca karmani phaldnavdptih.
(Sutra-rejecter:) "One should bind a cow, a goat for Agni and
Soma." One person would perform that which the instruction has
stated, the rest would perform something other than what the
instruction has stated. If something other than what the
instruction has stated is performed, the action is deficient and if
the action is deficient, one does not obtain the result.48
48Patafijali cites this Vedic injunction immediately below under vdrttika 47,
and in MB-K, vol. 1, p. 71, lines 18-19; p. 354, lines 12-13; p. 393, lines 15-18;
vol. 2, p. 246, lines 5-11; p. 357, lines 4-7; vol. 3, p. 161, lines 2-3; p. 445, line
21. Aitareyabrahmana 2.24 (part 1, p. 235): pasur anubandhyah,
Tandyabrahmana 21.14.11: ajo 'gnisomtyah, Sabara on Jaimini 10.4.32 (MD,
part 6, p. 386): Asti jyotistomah. Tatra sruyate, gaur anubandhyah, ajo
'gnlsomiya iti. Kathakasamhita 13.11 (p. 124): Agneyam ajam dlabheta
saumyam rsabham...., Maitrayanisamhita 2.5.5 (p.137): Agneyam ajam dlabheta,
saumyam babhrum rsabham pinigalam bhutikdmam ydjayed. See also
Taittiriyasamhita 5.5.24.6 (p.233): dgneyah krsno 'jah. Kathakasamhita 24.7 (p.
261) speaks of an aja which is agntsomtya. Sabara on Jaimini 6.8.30 says that
some recension states: ajo 'gnisomfyah. MD, part 5, p. 368. See also Jaimini
6.3.31, Chdgo vd mantravarndt, and Sabara's commentary. MD, part 5, pp. 368-
69.
Nagesa: Aparo 'sdstroktam iti. Ekah sabdah pratyarthanivesitvdd ekam
dravyam gorupam bodhayet, tasya kenacid dlambhe k.rte parais tena vd
gavdntardlambhe 'sdstrdrthah sydt. The rest would perform s o me t h i n g
other than what the instruction has stated. Because words apply per
object, one word would make known one individual cow substance. Once
someone has sacrificed it, if he or anyone else should sacrifice another cow, it
would be other than the intent of the instruction. MB, vol. 2, p. 93a.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Nanu ca yasyapy akrtih padarthas tasyapi yady anavayave

codyate na cdnubadhyate vigunam karma bhavati vigune ca
karmani phaldnavdptih. Ekakrtir iti ca pratijnd hiyeta. Yac cds
paksasyopdddne prayojanam ekaseso na vaktavya iti sa ceddnTm
vaktavyo bhavati.
(Sutra-proponent:) But isn't it so that even for him who holds th
a class property is the object denoted by a word, if the instruct
enjoins the action universally (anavayavena)49 and one does
bind all cows,50 then the action is deficient, and if the action
deficient, one does not obtain the result? And, to avoid this, on
would abandon the thesis that there is one class property, and, a
result, the very reason for adopting this view, namely, that o
would not have to state that one remains, would fail-now one
would have to state it.51
49Sdkalyena anubandhanddikam karma vidhfyate ity arthah. Raghunatha

Sarman, MB, vol. 2, p. 93b, note 2.
50Sakalo gaus tarhi iti sesah. Raghunatha Sarman, MB, vol. 2, p. 93b, note 3.
51Kaiyata: Nanu ceti. Sarvavisayd ced dkrtis codyate tadd sarvavyakti-
visayam anusthdnam vina vaigunyam eva sydd ity arthah. Athdnusthdna-
siddhyartham ekadravyavisayam eva karma kriyate tatrdpy ucyate: ekdkrtir iti
ceti.... Sa ceddnim iti. Ekatve jdter ekasabdapravrtteh siddhatvdd
ekaiesdrambhah pratydkhydyate, anekatve tv drabdhavya evaikasesa
But isn't it so. The meaning is: If the instruction enjoins (actio
class property which belongs to all (individuals of the class), th
performing (it) on all individuals, certainly deficiency would oc
other hand, if, in order to accomplish the performance, one does the
regard to one individual substance-concerning this he says: (one would
abandon the thesis that) the class property is one. Now one would.
The meaning is: Stating (the rule) that one remains is rendered useless
(pratydkhydyate) because the remainder of one is achieved due to the fact that
one word is used if a generic property is one (entity). But if there are many
(generic properties within the class) (the rule) that one remains certainly does
have to be stated.
Ndgesa: Athdnusthaneti. Ayam bhdvah: Jdteh sdksdd anubandhanddy-
asambhavdd vyaktidvdrd tad vdcyam, tatrdnusthdnasiddhyartham ekaprayoge
ekavyaktidvdrakam eva tad adsryate. Tatrakrter ekatve vinigamanavirahdt
sarvavyaktidvdrakam eva tad isrlyeta. Akrter anekatve tu pratijhidhdnir iti. On
the other hand, if, (in order to accomplish) the performance. This is
what he has in mind: Because (an action such as) binding, etc., is impossible on
the generic property directly, one must say that it involves individual objects.
If that is so, in order that the performance succeed in one instance, one accepts
that it involves just one individual object. Hence, if there were one class
property, because one would not decide (on one individual to the exclusion of
the others), one would resort to (the action) involving all individuals. But if
there were many class properties, the thesis (that there is one class property)
would be abandoned.
Bh.sye: anavayaveneti. Jdtydirayd vyaktivisesd avayavd ity ucya[n]te.
Anavayavena. Vyaktivisesdndlambanenety arthah. In the bhdsya:
universally. Specific individuals which are the substrata of the generic
property are called parts (avayava). Universally. The meaning is: without
regard to specific individuals. MB, vol. 2, p. 93a-b.
Annambhatta: Akrter abhidheyatvam anglkrtyaikadravyavisaye codandnglkare

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Evam tarhy anavayavena codyate pratyekam ca parisamdpyate

yathadityah.
(Sutra-rejecter:) If that is so, then an instruction enjoins action
universally and is fulfilled with respect to each individual, as is
the sun.52
Nanu ca yasydpi dravyam paddrthas tasydpy anavayavena

codyate pratyekam ca parisamdpyate.
(Sutra-proponent:) But an instruction enjoins action universally
and is fulfilled with respect to each individual of the class for him
dk,rteh pratyekaparisamaptir vdcya, tatra yugapad ekasydnekaparisamdpti-

v.rttikatvdyogdd dkrter aikyam bhajyeteti ekakrtir iti ca iti bhdsyam.
Accepting that a class property is denoted, if one accepts that an injunction
concerns one (individual) substance, one has to admit that the class property is
complete in each. If that were so, the unity of the class property would fail
because it is impossible for one thing to occur complete in many things
simultaneously. Hence, the Bhasya, "But (if one abandons the thesis) that
there is one class property." MPV, vol. 4, p. 112.
52Kaiyata: Evam tarhiti. Sarvdbhir vyaktibhih sambandhasya tulyatvdt
sarvatraivdbhinnabuddhyutpddandt pratyekam parisamdptatvdd ekasminnapi
dravye tat karma kriyamdnam jdtau k.rtam eva bhavatlty arthah. Yathaditya
iti. Pratidesam yathd sarvatmandditya upalabhyate tathd jdtir apity arthah. If
that is so. The meaning is: Because (the generic property's) relation to each
individual is the same, because the same cognition arises with respect to each
(individual), and because the generic property is complete in each, the action
done even on one individual substance is in fact done on the generic property.
As is the sun. The (whole) generic property (is present in each individual) as
the whole sun is seen in each place. (In the scenario described by Sabara (see
?3.2.2 note 12) and discussed in ?1.4.1, each observer sees the whole sun
directly in front of himself.)
Nagesa: Nanv dsrayabheddt tadbhedo 'ta aha: sarvdbhir iti. Bhedakdbhdvdn
na bheda iti bhdvah. Pratyaksavirodhdd dharmigrdhakamdnavirodhdc ca
bheddnumdnam asangatam iti bhdvah. In answer to the objection that the
generic property is different (in each individual of the class) because its
substrata are different, he says: to each. What he has in mind is: There is no
difference (between the class property in one individual of the class and in
another) because difference (of the substratum of a property) is not a cause of
difference (of the property). What he has in mind is: The inference ("The
generic property is different because its substrata are different") is unsound
because it contradicts direct perception (varttikas 36-37) and because it
contradicts the evidence by which one knows the subject of the inference
(paksa, i.e., the generic property). MB, vol. 2, p. 93b.
Nagesa raises a technical point: Inference establishes a conclusion (sddhya) on
the basis of an established (siddha) common ground (paksa or dharmin 'subject
of the inference'). The inference is faulty if the conclusion contradicts the
established common ground. In the present case, the conclusion that the
generic property in one individual of the class differs from that in another
contradicts the direct perception which establishes that there is a generic
property in the first place. The same cognition which occurs with respect to
each of many individuals is direct evidence (pratyaksa) (or according to
Sivaramendrasarasvati, grounds for inference) of the generic property. It is in
conflict with the conclusion that it is a different generic property in each of the
many individuals of the class.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
who holds that an individual substance is the object denoted

too.53
Ekasesas tvayd vaktavyah.

(Sutra-rejecter:) You have to state that one remains.54
Tvaydpi tarhi dvivacanabahuvacandni sddhydni.

(Sutra-proponent:) Then you have to justify the occurrence of
dual and plural inflectional terminations.55
44. Codandydm caikasyopddhivrtteh.

Codandyam caikasyopddhivrtter manydmahe adkrtir abhidhiyata
iti. Agneyam astdkapdlam nirvapet. Ekam nirupya dvitlyas
trtiyas ca nirupyate. Yadi ca dravyam paddrthah sydd ekam
nirupya dvitlyasya trtiyasya ca nirvapanam na prakalpeta.
(Sutra-rejecter:) And because in an injunction an
attribute (upddhi) applies to one thing.
And because in an injunction an attribute (upddhi) applies to one
thing, we consider that a class property is denoted. "One should
set out an astdkapala (eight-cup rice cake oblation) for (the deity)
Agni."56 Having set out one, he sets out a second and a third.
53Kaiyata: Nanu ceti. Dravye 'pi paddrthe jdter anabhidheydyd

upalaksanatvendsrayandt parityaktabhedam dravyamdtram visistajdtiyuktam
sabdenocyate. Pratyekam ca parisamapyata iti. Sdstram ity adhydhdrah.
Atha vd jdtir upalaksanabhutd pratyekam parisamdpyata iti vydkhyeyam. B u t.
If an individual substance is the object denoted by a word, because the generic
property, although not denoted, is accepted as an identifying mark
(upalaksana), a word denotes any individual substance connected with a
specific generic property regardless of its differences (from others connected
with that generic property). And is fulfilled with respect to each
(individual of the class). Supply 'an instruction'. Or it should be explained like
this: The generic property as an identifying mark is complete with repect to
each. MB, vol. 2, p. 93b.
54Kaiyata: Ekasesas tvayeti. Jdtyupalaksitasya dravyasya
sabdendbhidhdne saty abhidheydnekatvendnekasabdatvaprasangdd ity arthah.
You (have to state) that one remains. Because, if it were the case that a
word denotes an individual substance identified by a generic property, due to
there being many objects to denote it would be possible that many words occur.
MB, vol. 2, p. 93b.
55Kaiyata: Tvaydpi. Jdter ekatvdd dvitvabahutvanibandhandni dvivacana-
bahuvacanani na sidhyantiti bhdvah. Then you. Because a generic property
is one (object), the dual and plural (terminations) which depend on there being
two or many (objects) would not succeed. That is what he has in mind. MB, vol.
2, p. 93b.
56"Agneyam astdkapdlam nirvapet": Kathakasamhita 10.9 [13] (p. 92), 11.1 [1]
(p. 18), 11.2 [2] (pp.18-19), 13.3 [7] (p. 123), 22.3 [7] (p. 236). Maitrayani-
samhita 2.1.3, (p. 105), 2.2.7 [29] (p.115), 2.2.13 [35] (p. 118), 2.3.1 [1] (p. 119),
2.3.5 [19] (p. 123), 2.3.6 [22] (p. 124), 2.4.6 [22] (p. 131), 2.5.9 [11] (p. 140).
Taittiriyasamhita 2.2.5.5 [7] (p.70), 2.3.2.3 [10] (p. 77), 5.6.5.1 (p. 236).
Taittiriyabrahmana 3.9.17.1 (vol. 4, p. 141). See also Taittiriyasamhita 1.8.1.10,
1.8.2.1 (p. 51), 1.8.7.2, 1.8.17.1 (p. 56), 1.8.19.1, 1.8.20.1 (p. 57), "dgneyam

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
But if an individual substance were the object denoted, having set

out one, setting out a second and a third would not succeed.57
astakapalam nirvapati." Reference: Rau (1985: 18, #108).

The word 'dgneya', meaning that which is offered to the deity Agni (4.2.24), is
formed from the word 'agni' by adding the suffix dhak (4.2.24, 33),
strengthening the first vowel (7.2.118) and dropping the final i of the base
(6.4.148).
4.2.24. Sdsya devatd. Affixes occur as provided in the meaning of a sixth triplet
termination after a word ending in a first triplet termination, if it names a deity.
4.2.33. Agner dhak. The suffix dhak occurs after the word 'agni' to form a
derivative meaning that whose deity is Agni.
6.4.148. Yasyeti ca (lopah 147, taddhite 144, bhasya 129, angasya 1). Zero
replaces the final sound of a stem termed bha (1.4.18) which ends in a sound of
the class i or a and which is followed by a taddhita affix or the sound f. For the
term bha see Cardona (1988: 49).
7.2.118. Kiti ca (taddhitesv acdm ddeh 117, aco nniti 115, vrddhih 114 angasya
6.4.1). A vrddhi vowel replaces the first vowel of the stem after which a taddhita
affix marked with k is provided.
The word 'astdkapala' means an oblation (havis) prepared (samskrta) in eight
(asta) cups (kapala). It is derived by forming the dvigu compound (2.1.52,
Samkhydpuirvo dviguh) in the meaning "prepared food" (4.2.16), one of the
meanings in which the taddhita (2.1.51, Taddhitdrthottarapadasamdhdre ca) affix
an occurs (4.1.83, Prdgdlvyato 'n). No trace of the affix remains in the resulting
dvigu compound (4.1.88, Dvigor lug anapatye).
4.2.16. Samsk.rtam bhaksdh (tatra 14). After a word ending in a seventh triplet
termination, affixes occur as provided to mean that which is prepared, if it is
food.
Kasika (vol. 1, p. 113) on 4.2.16 explains the details of derivation of
'astdkapala'.
57Kaiyata: Evam bhdsyakdrena paksayoh sdmyam darsitam. Vdkyakdras tu
jdtipaksdsrayanasya prayojandni samuccayenacaste: codanayam caikasyeti.
Sdmdnyasyagneyddeh. Upddhivrtter iti. Astdkapalatvddir upadhih tasya
vrttih pravrttih. Etad uktam bhavati: Yadi dgneyddisabdena samdnyam
abhidhlyate tadd tendiesadravydndm dksepdt pratidravyam astdkapalatvam
upddhir yujyate dravye tu paddrthe ekenaiva yajamanena sakrd evdstdkapdlah
puroddso nirupyeta ndnyena ndpy anyadety arthah. In this way Patafijali shows
the equivalence of the two views. But Katyayana states all the reasons for
accepting the view that a generic property is denoted together: And (because)
in an injunction (an attribute) applies to one (thing). To one general
property: being for Agni. An attribute applies. The attribute of being
astdkapdla, etc., applies occurs. What is stated is this: If a word such as 'being
for Agni', etc., denotes a general property, then because (the general property)
requires any individual substance, the attribute of being astakapala would be
joined to each individual substance. But if an individual substance were the
object denoted by a word, only one performer would set out an astdkapala rice
cake oblation only once; no one else (would) nor (would he) another time. That
is the meaning.
Nagesa: Ekasyagneydder ya upddhir astdkapalatvddis tasya punah punah
pravrtter iti vdrtikdrthah.... Astdkapdlatvddityddipadena nirvdpah. Ata eva
bhdsye sa eva darsita ekam nirupyetyddind. The meaning of the vdrttika is:
Because an attribute such as being astdkapala, etc., occurs again and again for
one (constant qualifier) such as being for Agni. By the word 'etc.' in 'astdkapala,
etc.', (he intends) setting out (nirvdpa). Hence, that is what Pataiijali shows by
(saying) "having set out one." MB, vol. 2, p. 94a.
If the word 'dgneya' referred to a single individual oblation, it would refer to the

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Kah punar etayor jdticodanayor visesah?

(Sutra-proponent:) But what is the difference between these two
generic property injunctions?58
Eka nirvrttendpard nirvartyena.

(Sitra-rejecter:) The first enjoins by means of something which
already exists, the other by means of something which has to be
made.59
particular oblation set out by one performer on one particular occasion. It

would not refer to the oblation which he might set out on any other occasion or
which another performer might set out. Hence the injunction which refers to an
oblation having the attribute astdkapalatva could only be carried out once.
However, if the word 'dgneya' denotes the general property being for Agni, it
will refer to any individual substance joined to that general property. Hence the
injunction to set out such a substance having the attribute astdkapalatva could
be satisfied by setting out any oblation for Agni having the attribute
astdkapalatva, that is, any oblation prepared in eight cups for Agni.
Annambhatta: Sdmdnyasyeti. "Ekam anekdnugatam sdmdnyam" iti
laksandnusdrdd ekam sdmdnyam ity arthah.... Jdtisabdena sdmdnya-
mctrdbhidhdndd upddhirupasdmdnyasydpy abhidheyatvdvirodhah. To one
general property. The meaning is: One is a general property following the
definition, "One thing which occurs in many things is a general property."...
Because the word 'generic property (jdti)' denotes any general property
(sdmdnyamdtra), there is no conflict in (the word 'one') denoting a general
property (sdmdnya) in the form of an attribute (upddhi). MPV, vol. 4, p. 113.
Although 'dgneya' is also an attribute applied to the oblation, it is taken as
constant in the present example. Astdkapalatva varies with other attributes such
as ekddasakapdlatva (being prepared in eleven cups).
Sivaramendrasarasvati: Ekasyopadivrtter iti. Ekasyaiva vastuna
upddhindstdkapalatvddind vrtteh sambandhdt tadvastuno dravyatve ekenaiva
sakrd eva niruptam sastriyam sydt, ndnyena, na punas ca. Akrtitve tu sarvair
niruptam punar niruptam ca sastriyam sydd iti bhdvah. Because a property such
as being astdkapala, etc., applies is related to just one thing, if that thing
were an individual substance, that which is set out by just one (person) just
once would fit the instruction, not (that which is set out) by anyone else, nor
(by the same person) another time. But if that thing were a class property, that
which is set out by anyone and set out (even by the same person) another time
would fit the instruction. That is what he has in mind. MPV, vol. 4, p. 115.
58Kaiyata: Kah punar iti. "Gaur anubandhyah" iti "Agneyam
astdkapalam" iti cdnayor ity arthah. But what. He means of these two:
"One should bind a cow, (a goat for Agni and Soma)," and "(One should set out)
an astdkapala for Agni." MB, vol. 2, p. 94a.
59Kaiyata: Eka nirvrtteneti. Nispanno hi pasur ydgasddhandyopddlyate.
Apara nirvartyeneti. Agneyo 'stdkapalas codanayaivotpddyate na tu
codandydh purvam nispanna ity arthah. The first (enjoins) by means o f
something which already exists. One takes an already existing cow for a
means to accomplish the Vedic performance. The other by means of
something which has to be made. The injunction itself causes the
oblation for Agni to come into being.; the oblation does not exist prior to the
injunction. That is the meaning. MB, vol. 2, p. 94a-b.
Kaiyata and Nagesa go on to give a reason for the distinction. This is
unconvincing for the reasons which Sivaramendrasarasvati gives:
Sivaramendrasarasvati: Astdkapalddikam hi sad yo nispddya tena ydgah
kriyate. Etena "Tatra nirv.rtte bheddgrahandt dravyantare 'pi pravrttih sydn na tu

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
45. Dravydbhidhanam Vyddih.

Dravydbhidhdnam Vyddir acdryo nyayyam manyate. Dravyam
abhidhiyata iti.
46. Tatha ca lingavacanasiddhih.
Evam ca krtvd lingavacandni siddhdni bhavanti. brdhmani
brdhmanah brdhmanau brdhmand iti.
47. Codandsu ca tasyarambhdt.
Codandsu ca tasydrambhdn manydmahe dravyam abhidhiyata iti.
Gaur anubandhyo 'o 'gnfsomlya iti. Ak.rtau coditdydm dravya
drambhandlambhanaproksanavisasanddtni kriyante.
48. Na caikam anekadhikaranastham yugapat.
Na khalvapy ekam anekddhikaranastham yugapad upalabhyate.
Na hy eko Devadatto yugapat Srughne bhavati Mathurayydm ca.
49. Vindse prddurbhdve ca sarvam tathd sydt.
(Sutra-proponent:) Vyadi holds that a word denotes an
individual substance.
The teacher Vyadi considers it appropriate that a word d
individual substance: A substance is denoted.60
nirvartye sddrsydagrahandd ity avasydsrayanlyd jdtih," iti Kaiyato nirastah,

ndndvidhapakvdnnddnadm nirvartydndm sddrsyagrahanapurvakam loke
kriyamdnatvendnubhavaviruddhatvdt. One does the Vedic performance with the
astakapala, etc., entity which is to be brought into existence. This refutes
Kaiyata (saying) "Since this is so, although it may be possible for one to act on
another already existing individual substance because one is not aware of its
differences (from the individual the word denotes), it would not be possible on
an object which has to be made because one would not be aware of the
similarity. Therefore, one must necessarily resort to a generic property,"
because it is opposed to the ordinary experience of various types of cooked
food and so on which have to be brought into existence being made with a prior
grasp of the (mutual) similarity (of the individuals of each type). MPV, vol. 4,
p. 115.
60Kaiyata: Dravydbhidhanam iti. Jdter vrttivikalpdksamatvendbhdvam
manyamano Vyddir dravyam eva sabdendbhidhlyata iti manyate. (Vyadi holds)
that (a word) denotes an individual substance. Considering that there
is no generic property because it is incapable of any of the possible modes of
occurrence, Vyadi considers that a word denotes just an individual substance.
Nagesa: Nanu vyaktisaktdv dnantyavyabhicdrayoh sattvdt jdtau saktir ity
dsankya jdtir eva ndstfty aha: vrttivikalpeti. Jdtir na tdvad dravyesu
vydsajyavrttih, ekdsrayandse 'pratitydpatteh. Kim ca gdm alabhetetyddi-
nodandsu sarvadravydlambhdpattih. Pratyekasamdptau tu dravydntare tad-
andpattis tatrdpi sattve ekk!Itir iti pratijnidhdnir ity arthah. Tattatsamsthadnddy-
upalaksitam adhisthdnacaitanyam eva dravyam iti ndnantyavyabhicdrdv iti
tdtparyam. Answering the objection, "Because there would be infinity (of
signifying relations) or deviation (of a word's usage from the object it signifies)
if the signifying relation were to an individual, the signifying relation is to a
generic property," he says that there is no such thing as a generic property:
(Because it is incapable of) any of the possible modes of occurrence.
(1) A generic property does not occur in individual substances collectively
because then upon the destruction of one substratum, it would not be perceived.
Moreover, in commands such as, "Sacrifice a cow," etc., the sacrifice of all
individual substances would result. (2a) But if a generic property were
complete in one (individual substance), it could not be in another individual

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
And in this manner gender and number succeed.

And on this basis, different genders and numbers are achieved:
brdhmani (feminine singular); brdhmanah (masculine singular),
brdhmanau (masculine dual), brdhmanah (masculine plural).
And because in injunctions one commences the action
on a substance.
And because in injunctions one commences the action on a
substance, we consider that a substance is denoted. "One should
bind a cow, a goat for Agni and Soma." Even if the instruction
enjoins the action on the class property, one does the tying,
killing, washing, cutting up, etc., on a substance.61
And it is not the case that one thing is present in many
substrata simultaneously.
Moreover it is not the case that one thing is found present in
many substrata simultaneously. For the single individual
Devadatta is not present simultaneously in Srughna and in
Mathura.62
substance. (2b) If it were in another too, that would destroy the thesis that there
is one generic property (for the class). The purport is: There would not be
infinity or deviation because a substance is just underlying consciousness
characterized by a certain configuration, etc. MB, vol. 2, p. 94b.
Nagesa suggests that a word denotes brahman, the single basis of the world full
of diffferences. Since there is just one brahman, the faults of infinity and
deviation would not apply. They are possible only if there are many individual
substances. However, this is not the view of Vyadi as Patanijali presents it. A
single substance would be devoid of gender and number so would not be
capable of the solution to the gender and number problem stated in varttika 46.
In addition, it would be subject to the faults Vyadi attributes to the view that a
word denotes a generic property in varttika 49. From Patanijali's presentation it
is clear that Vyadi takes the view that a word denotes each of many individual
substances.
61Kaiyata: Akrtau coditaydm iti. Ak,rtivddipakse sabendkrrtis codyate

dravye tu kdryam pravartate. Ayuktam caitat. Na hy anyacodane 'nyasya kdrye
kriyamdne yathoktam k,rtam bhavati. Tasmdd dravyam eva sabdenocyata iti
nydyyam.
If (the instruction) enjoins (the action) on the class property. On the
view that a word denotes a class property, a word enjoins (an action to be done
on) a class property but the action to be brought about takes place on a
substance. And that is not right. Because when an instruction enjoins (an
action) on one thing and the action to be brought about is done on another, it is
not done as stated. Therefore, it is appropriate that a word denotes just a
substance.
Nagega: ...arambhanam krayo bandhanam vd. ...tying is either purchase o
binding. MB, vol. 2, p. 95a.
62The passage appears under Sivasutra 1, A i u n, vdrttika 12, Yugapac ca
desaprthaktvadarsandt: Na hy eko Devadatto yugapat Srughne ca bhavati
Mathurdyam ca. MB-K, vol. 1, p. 18, lines 11-12.
Kaiyata: Na caikam iti. Na ca heturahitenddityadrstantena sddhyasiddhir
bhavati, devadattadrstdntena viparyayasydpi sadhandt.
And it is not the case that one thing. One does not establish the
(inferential) conclusion (that one generic property is present in many
substances) by the example of the sun without an inferential mark, because by

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
When something perished and when something came

into being, all of it would likewise.
Kim?
(Sutra-rejecter:) What?
Vinasyec ca praduhsyac ca. Svd mrta iti svd ndma loke na

pracaret. Gaurjdta iti sarvam gobhutam anavakasam sydt.
50. Asti ca vairupyam.
Asti khalvapi vairuipyam. Gaus ca gaus ca khando munda iti.
51. Tatha ca vigrahah.
Evam ca krtvd vigraha upapanno bhavati. Gaus ca gaus ceti.
52. Vyarthesu ca muktasamsayam.
Vyarthesu ca muktasamsayam bhavati. Akrtdv api paddrtha
ekaseso vaktavyah. Aksdh pdddh mdsd iti.
(Sutra-proponent:) Perish and come into being. "A dog has
died." No dog whatsoever would remain in the world. "A bull
has been born." There would not be enough room for everything
that ever is (was or will be) a cow.63
the example of Devadatta one establishes the opposite. MB, vol. 2, p. 95a.
63Kaiyata: Svd mrta iti. Abhivyaijakavinise jdtes tirobhdvdt pinddntare
sveti pratyayo na sydt. Satapratyaya ivaikdpdya ity arthah. Yad vd asraydpdye
dsritasydpy apdyah avayavdpdye avayavina iveti jdter vindiaprasanga ity arthah.
Gaur jdta iti. Jdtena gopindena gotvam abhivyaktam tac ca sarvas'rayavartlti
ekagopindapratyaksatve sarvagopindapratyaksatvaprasangah.... Atha vd sarva-
vastuvydpi gotvam estavyam.... Tatas ca...sarvesu paddrthesu gopratyaya-
prasangah. A dog has died. The meaning is: Because the generic property
would disappear when an object manifesting it perished, the cognition, "dog,"
would not occur with respect to any other object just as the cognition of a
hundred (does not occur) when one is subtracted. Or when the substratum
ceases to be, that which rests on it ceases to be too just as the whole (ceases to
be) when a part ceases to be; hence, the destruction of the generic property
would result. A bull has been born. An individual cow having been born
manifests cowness and that is present in all its substrata; hence, when direct
perception of one individual cow occurred, direct perception of all individual
cows would result.... Or it should be accepted that cowness pervades all things...
and then... the cognition of cow in all objects would result. MB, vol. 2, p. 95b.
Kaiyata cites Dharmakirti's Pramanavarttika (p. 304-05), Pariccheda 3, verses
152, 155b-56a, to elaborate on the faults in the concept of a generic property.
Nagesa: Abh ivy ajaketi. Vydsajyavrttir jdtih sarvavyaktibhih
sambhiuydbhivyajyata iti mate idam dusanam, jdter ndSas ca tirobhdva evety
arthah.... Evam vindsapadam vyaktisu dhvamsaparam tatheti pardmrstam dkrtau
tirodhdnaparam ity ayuktam iti matvdha: yad veti. .... Jdteneti. Vydsajya-
v,rttigotvam sarvasrayajhdnair abhivyanktavyam anyathd gotvajiinam eva na
sydd ity arthah. Astu tarhi tatpratyaksatvdya taddarayasarvasya pratyaksatvam
ata aha bhdsye: anavakasam iti. (An object) manifesting. This fault
occurs on the view that all individuals combined manifest a generic property
which is present in them collectively, and the perishing of the generic property
is just its disappearance.... Considering it incorrect that the word 'perishing',
meaning destruction with regard to individuals, which is referred to by
'likewise' (vt. 49), should mean disappearance with regard to a class property, he
says: Or. .... (An individual cow) having been born. The meaning is:

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
And there is difference.

Moreover, there is difference: a bull and a bull; an unwhole
a hornless bull.64
And in this manner, analytic strings (make sense).
And on this basis, the analytic string, "a bull and a bull," (for the
dual word 'two bulls (gavau)') makes sense.65
cowness present in all its substrata collectively is manifested by knowledge of

all the substrata (together); otherwise, knowledge of cowness simply would not
arise. Since that is so, in order to bring about direct knowledge of it, let there be
direct perception of all its substrata. Hence he says in the bhasya: not
enough room. MB, vol. 2, p. 95b-96a.
Sivaramendrasarasvati explains the word 'anavakdsam' saying, "niravakdsam
ddhdrastinyam (without space, devoid of ground)," and elaborates that even the
entire earth would not be able to be ground for innumerable cows. Na hy
asamkhyeycdndm gavdm prthivl samastdpy ddhdro bhavitum alam. He rejects
Kaiyata and Nage?a's interpretation of the word as a bahuvrthi bearing the
meaning of a fifth triplet termination saying: Anavakdsasabdo hi prdyena
sasthyarthagarbhitabahuvrfhir evdtra sdstre asritah. The word anavakdsa is
usually used in this discipline only as a bahuvrfhi enveloping the meaning of a
sixth triplet termination. See MPV, vol. 4, p. 115, lines 22-23 and 26-27.
64Kaiyata: Asti ca vairupyam iti. Tasmdd dravyam evdstu na tu samanyam.
Na hy ekasya bheddbhedau viruddhau upapadyete ity arthah. Ekdrthakriydkdri-
tvddinimittas tu pratyabhijncipratyayah. Gaus ceti. bhedddhisthdnah
samuccayah sdmdnyasyaikatvdbhideyatvayor na sydd ity arthah. And there
is difference. Therefore, let just an individual substance be (the object
denoted by a word) and not a general property. Because the opposites,
difference and non-difference cannot belong to one thing. That is the meaning.
And the identity-cognition (pratyabhijndpratyaya) is caused by the fact that
different objects perform the same purposeful activity. A bull and (a bull).
The meaning is: There could not be an aggregate if there were one general
property and it were denoted, because an aggregate depends on the aggregated
objects being distinct.
Nagesa: Na hiti. Evam ca tattacchabdapratipddye ekasmin viruddha-
dharmopapatter abhdvdd dravyam eva paddrtha iti bhdvah. Bhdsye gaus ca
gaus cety anendbhipretam vairuipyam upapddayati: bhedeti. Evam khando
gaur mundo gaur iti ca na sydd iti bhdvah. Because (one thing) cannot. And
so just a substance is the object denoted by a word because it is impossible for
opposite properties to be in one single object which each of many words
expresses. Kaiyata explains the difference intended by the bhasya passage, "a
bull and a bull": difference. Similarly (because one thing cannot have
property a and not property b, and yet have property b and not a) "an unwhole
bull, a hornless bull" wouldn't occur either. That is the sense. That is what he
has in mind. MB, vol. 2, p. 96a.
I have edited Pataijali's examples here in accordance with Kaiyata and Nagesa's
commentaries. However, Sivaramendrasarasvati takes the example as showing
both similarity and difference. The repetition of the word 'cow' shows
similarity; the different qualifiers 'unwhole' and 'hornless' show dissimilarity.
He says: Sdrupyam tdvad darsayati: gaus ca gaus ceti. Vairtupyam darsayati:
khando munda iti. MPV, vol. 4, p. 114.
65Nagesa: Tathd ca vigraha iti. Ekarthatve paryayanam iva sahaprayogo na
sydd iti bhdvah. And thus analysis. They would not be used together
because they would denote the same object, like synonyms. That is what he has
in mind. MB, vol. 2, p. 96b.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
And for words having different meanings, it is

indubitable.
And for words having different meanings, stating
remains is indubitable. Even if a class property is th
denoted by a word, one has to state that one remains: 'ak
axle, seed), paddh (foot, verse segment, quarter), mds
monetary unit, fool).'66
53. Lingavacanasiddhir gunasyanityatvat.

Lingavacandni siddhdni bhavanti.
(Satra-rejecter:) Gender and number succeed because a
quality is not innate.
Different genders and numbers are achieved.
Kutah?
(Sutra-proponent:) Why?
Gunasydnityatvdt. Anityd gund apdyina updyinas ca.

(Sutra-rejecter:) Because a quality is not innate. Qualities are not
innate; they come and go.67
66Kaiyata: Akrtav apiti. Yadartha dkrtipaksaparigrahah tad eva na sidhyati

na hi sakataksadevandksddisv ekdkrtisadbhdvah. Even if a class
property. The very reason for adopting the view that a class
denoted (i.e., that the rule that one remains would not have to be
not succeed because (the same) one class property does not exis
cart-axle (sakata-aksa) and a gaming die (devana-aksa), etc.
Nagesa: bhdsye: vyarthesv iti. Vijdtiydnekdrthasarupasabdesv
Muktasamsayam iti. Ekasesadngkaranam iti sesah.
In the bhhsya: For (words) having different meanings. The m
for words of the same form having many meanings of diffe
Indubitable. The remainder is: accepting that one remains. MB
96b.
67Kaiyata: Iddnim dkrtivddi dosdn parihartum dha: lifigeti. Akrtisadbhdvah

purvam eva pratyaksdnumdnagamaprav.rttibhih pratipdditah. Sarvagatatve 'pi
paddrthdndm vicitrasaktiyogdt kascit paddrthah kdmcid evdkrtim abhivyanakti
na tu sarvam sarvah. Gunds cdtra dvitvastritvddayo vivaksitdh. Tais cdkrter
ekdrthasamavdyalaksanah sambandha ity adkrtdv api paddrthe lingasamkhyd-
siddhir ity arthah. Anitya iti. Kaddcit strltvayuktena pindendkrtir vyajyate
kaddcit pumstvayuktena, kaddcid ekddhdrasthd pratipadyate kaddcid
anekddhdrasthety arthah. Now the proponent of the view that a class property is
denoted, to avoid the faults (which Vyadi had attributed to it), says: gender. It
was shown previously by direct perception, inference, authoritative texts and
behavior that class properties exist. Even if class properties are all pervading,
because entities have varying capacities, a certain entity manifests a certain
class property; not all (entities manifest) every (class property). He intends the
qualities here to be duality, femininity, etc. A class property has the relation of
inhering in the same object with them; hence, even if a class property is the
object denoted by a word, gender and number succeed. That is the meaning.
Not innate. Sometimes a class property is manifested by a body connected
with feminine gender, sometimes by one connected with masculine gender;
sometimes it occurs in one substratum, sometimes in many substrata.
Nagesa: Anityd iti. Aniyatd ity arthah Apdyitvopdyitvdbhydm aniyatatvam

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Kim ya ete suklddayah?

(Sutra-proponent:) What, white and so on?
Netyaha. Strtpumnapumsakdni sattvagund ekatvadvitva-

bahutvdni ca. Kaddcid dk.rtir ekatvena yujyate kaddcid dvitvena
kaddcid bahutvena kaddcit strftvena kaddcit pumstvena kaddcin
napumsakatvena.
(Sutra-rejecter:) No. The qualities of a substance: feminine,
masculine, and neuter gender, and singular, dual, and plural
number. Sometimes a class property is connected with singular
number, sometimes with dual number, sometimes with plural
number; sometimes with feminine gender, sometimes with
masculine gender, sometimes with neuter gender.68
Bhavel lingaparihdra upapanno vacanaparihdras tu nopapadyate.

Yadi hi kaddcid dkrtir ekatvena yujyate kaddcid dvitvena kaddcid
bahutvenaikdkrtir iti pratijnid hyeta yac cdsya paksasyopdddne
prayojanam uktam ekaseso na vaktavya iti sa ceddnrm vaktavyo
bhavati.
(Sutra-proponent:) It may be possible to avoid the gender
problem but it is not possible to avoid the number problem.
Because if a class property is connected sometimes with singular
number, one would abandon the thesis that there is one class
property, and, as a result, the very reason mentioned for adopting
this view, namely that one would not have to state that one
remains, would fail-now one would have to state it.69
[53a.] Evam tarhi lingavacanasiddhir gunavivaksdnityatvdt.

Lingavacandni siddhdni bhavanti.
eva bhdsye darsitam. Not innate. The meaning is impermanent. The fact of
going and coming (mentioned) in the bhdsya shows just impermanence. MB,
vol. 2, p. 96b.
68The word 'gender' like the term 'linga' it translates, refers 1) to the sex
characteristics of objects, 2) to the characteristics which determine the use of
certain speech forms rather than others and 3) to the characteristics of speech
forms themselves. It has the third sense in the vdrttika, meaning the correct
affixes would occur. In the present paragraph the sutra-rejector intends the
second sense but the sutra-proponent understands the first sense.
69Kaiyata: Vacanaparihdras tv iti. Liigaparihdre na kimcid dusyati,
vacanaparihdre tv abhidhlyamdne ekdkrtir iti pratijidhdnih, dvitvabahutva-
sambandhasydpy abhyupagamdt. Dvivacanabahuvacanavad anekaiabdatvam
api prdpnotity ekaseso 'pi vaktavya ity arthah. But it is not pertinent t o
number. Nothing is wrong with the rebuttal sparing gender but if one puts
forward the rebuttal regarding number, it destroys the thesis that there is one
class property because it accepts that a class property has a relation with dual
and plural number too. The occurrence of many words would obtain just as the
dual and plural inflectional terminations do and so (the rule) that one remains
would have to be stated. MB, vol. 2, p. 97a.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
(Suitra-rejecter:) In that case, gender and number succeed because

a speaker's intention of a quality is not innate. Different genders
and numbers are achieved.
Kutah?
Gunavivaksdyd anityatvdt. Anityd gunavivaksd. Kaddcid dkrtir

ekatvena vivaksitd bhavati kaddcid dvitvena kaddcid bahutvena
kaddcit stritvena kaddcit pumstvena kaddcin napumsakatvena.
(Sutra-rejecter:) Because a speaker's intention of a quality is not
innate. A speaker's intention of a quality is not innate.
Sometimes a speaker intends a class property with singular
number; sometimes with feminine gender, sometimes with
masculine gender, sometimes with neuter gender.70
Bavel lingaparihdra upapanno vacanaparihdras tu nopapadyate.

Yadi kaddcid dkrtir ekatvena vivaksitd bhavati kaddcid dvitvena
kaddcid bahutvenaikdkrtir iti pratijnii hiyeta yac casya
pakasyopdddne prayojanam uktam ekaseso na vaktavya iti sa
ceddnlm vaktavyo bhavati. Liigaparihdras cdpi nopapadyate.
(Sutra-proponent:) It may be possible to avoid the gender
problem but it is not possible to avoid the number problem. If it
were the case that sometimes a speaker intends a class property
with singular number, sometimes with dual number, sometimes
with plural number, one would abandon the thesis that there is
one class property, and, as a result, the very reason mentioned for
adopting this view, namely that one would not have to state that
one remains, would fail-now one would have to state it.
Moreover, it is not possible to avoid the gender problem either.71
70Kaiyata: Gunavivakad'nityatvad iti. Svata dkrter ekatvam dravyagata-

dvitvabahutvavivaksdydm tu dvivacanabahuvacane bhavatah. Asrayadharmair
asritdndm vyapadesah.
Because a speaker's intention of a quality is not innate. In and of
itself a class property has singular number but the dual and plural (inflectional
terminations) occur if a speaker intends the dual and plural number which
belong to the substance. One speaks of the properties which reside in a
substrate (the class property) as having the properties of the substrate (gender
and number). MB, voj. 2, p. 97a.
71Sivaramendrasarasvati: Pratijid hiy e teti. Vivaksamdtrendrtha-
visesdsiddheh. Anyathd vivaksayd gavy asvasabdah, asve 'pi goiabdah
prayoktavyah sydt. Ata dkrtibhedam antarena vacanopapddandsambhavdt
pratijidhdnyddikam tadavastham iti bhdvah. (One) would abandon the
thesis. Because a certain meaning does not occur just by a speaker's intention.
Otherwise, if it did, the word 'horse' would be used for a cow and the word 'cow'
for a horse just by a speaker's intention (see Patafijali on vdrttika 29). Hence the
problem of abandoning the thesis and so on would still remain, because it is
impossible to explain (dual and plural) number without splitting up a class
property. MPV, vol. 4, p. 118.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Kim kdranam?
(Sutra-rejecter:) What is the reason?
Avistalitigd jdtir yal lingam updddya pravartata utpattiprabhrtyd

vindadt tal litigam na jahdti.
(Sutra-proponent:) A generic property is gender specific; from
origin to destruction it does not abandon the gender it starts out
with.72
[53 b .] Tasmdn na vaiydkaranaih sakyam laukikam lingam

dsthdtum. Avasyam kascit svakrtdnta dstheyah.
(Sutra-rejecter:) Therefore, grammarians cannot accept the
ordinary gender of an object (as the criterion for the gender of
the words denoting it). They have to adopt a conception of their
own.73
72The passage, "A generic property... (dvistalingd jdtir...)," cites the commentary
on sutra 1.2.52, Visesandndm cdjdteh. MB-K, vol. 1, p. 228, lines 18-19. See my
discussion ?1.4.3.
Kaiyata: Tatas caikam evasrayagatam lingam upddatte sarvadd jdtih yathd
gramyapasujatayah samghe stritvam evopddadate gdva imd iti, anyds tu
pumstvam eva brdhmana ime iti yathety arthah. And so a generic property
always takes just one of the genders present in its substrata. In this way the
generic properties of domestic animals in herds take only feminine gender, (for
example) "these cows (feminine by 1.2.73)," while in contrast other (generic
properties take) only masculine gender, (for example) "these brdhmanas
(masculine by 1.2.67)." MB, vol. 2, p. 97b.
1.2.73 is an exception to 1.2.67.
Panini 1.2.67. Pumdn striyd (tallaksanas ced eva visesah 65). In stating (a male)
together with a female just the masculine speech form remains if that is the only
difference between them.
Panini 1.2.73. Grdmyapasusamghesv atarunesu stri. For herds of domestic
animals other than just young, if a speaker intends the animals together, just a
feminine speech from remains.
The word 'dvistalingd' is a bahuvrihi compound literally meaning that in which
gender (linga) has entered (dvista). Formed in accordance with Panini 2.2.24,
Anekam anyapaddrthe, the past passive participle (word ending in the suffix
kta) falls first according to Panini 2.2.36, Nisthd (purvam 30, bahuvrihau 35).
Sivaramendrasarasvati: Yal liigam iti. Avistalingd jdtir iti dravyapaddrtha-
vddindpy darayantyam eva, vrksasabdaprabhrtteh sarvathd pumllingddy-
anupapatteh. Tathd cobhdbhydm eva kimcil lokasiddhalingavilaksanam lingam
dstheyam ity aha: Tasmad iti. The gender. The proponent of the view that
an individual substance is the object denoted by a word has to accept that a
generic property is gender-specific too because (otherwise) it would not make
sense that the word 'tree', and other words, always occur in the masculine gender.
And so both (the proponents of both views) have to accept a conception of
gender which is different from the one ordinarily accepted. Hence he says:
Therefore. MPV, vol. 4, p. 118-19.
73Kaiyata: Laukikam iti. Stanakesddisambandha ity arthah.
Tasminndsriyamdne lingdntarena sambandho na sydt. Tasmdt svasiddhdnta-
vyavasthasriyate.
Ordinary. The meaning is: Having a relation to breasts, (long) hair, etc. If
grammarians did accept ordinary gender, a generic property's relation to another
gender would not occur. Therefore, they resort to a classification (based on) a

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Ko 'sau svakrtdntah?
(Sutra-proponent:) What is this conception of their own?
Samstydnaprasavau lingam.
Samstydnaprasavau lingam dstheyau.
(Sutra-rejecter:) Samstyana and prasava have to be accepted
as gender.
Samstyana and prasava have to be accepted as gender.
Kim idam samstydnaprasavdv iti?

(Sitra-proponent:) What is this 'Samstydna and prasava'?
Samstydne stydyater drat stri suteh sap prasave pumdn.

(Sutra-rejecter:) The word 'strf (feminine)' is formed from the
verbal root Vstyai (to combine) with the suffix drat in the
meaning samstydna; the word 'pums (masculine)' is formed from
Visu (to procreate) with p for s of the root in the meaning
prasava.74
conception of their own.

Nagesa clarifies that if the gender of words for the generic property brdhmanatva
were determined by the sex of its substrata, the word 'brahmana' would take just
the masculine (1.2.67) and 'brdhmant' with the feminine affix his (4.1.63,
?1.2.5) meaning a brdhmana woman would never occur: Ligadntareneti.
Brahmani ddrd ity adau dravyadvdrd jdter litigntarena yogo na sydt tatas
tatityddau ca kasydpi lihgasya yogo na sydd iti bhavah. To another gender.
In "a brdhmana (feminine) wife," etc., the generic property would not have a
connection with another gender by way of the substance and in "a bank
(masculine), a bank (feminine)," etc., it would not have a connection with any
gender at all. That is what he has in mind.
Nagesa concludes: Sabdavisesdpeksayd tu jdter dvistalingatvoktir iti
tatparyam. The point is: A generic property is said to be gender specific only in
so far as its gender depends on the specific word used for it. MB, vol. 2, p. 97b.
74The word,'strt' is derived as follows:
V/Styai + drat Pancapadyunidisutra 4.166.
V/Styai + dra + nip 4.1.15.
Styd + dra + iip 6.1.45.
Sty + ra + hip 6.4.143.
Stra + hip 6.1.66.
Stri The final a of 'stra' is deleted before t by 6.4.148.
See note 56.
Dhatupatha 1.959. v/Styai sabdasamghdtayoh. Bohtlingk, vol. 2, p. 69.
Paicapadyunadisutra 4.166. Stydyater drat. SK #615 p. 559. See Amarakosa
2.6.2, p. 203.
4.1.15. Ti.ddhdnandvayasajdaghnaimdtractayapthakthafikaikvarapah (nip 5).
The feminine suffix tip occurs after a base ending in an affix marked with t, etc.
6.1.45. Ad eca upadese 'siti. a substitutes for the final sound of a verbal root
ending in e, o, ai, or au, unless an affix marked with s follows.
6.4.143. Teh (diti 142, lopah 134). Zero (lopa) replaces the ti (the last vowel
and following sounds 1.1.64) of a base followed by an affix marked with d.
6.1.66. Lopo vyor vali. Zero (lopa) replaces v or y followed by a consonant
other than y.
Pataijali derives the nominal base 'pums' from the verbal root V/sui. V'suni

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Nanu ca loke 'pi styayater eva str sutes ca pumdn.

(Sutra-proponent:) But 'strt (a female)' is from Vstyai and 'pums
(a male)' from Vsu in ordinary usage too, isn't it so?
Adhikaranasddhana loke str. Stydyaty asydm garbha iti.

Kartrsddhanas ca pumdn. Sute pumdn iti. Iha punar ubhayam
bhdvasddhanam. Stydnam prav.rttis ca.
(Sutra-rejecter:) In ordinary usage 'strl (a female)' means that
which brings about the action of combining as its locus, "that in
which an embryo combines."75 And 'pums (a male)' means that
which brings about the action as its agent, "a male procreates."76
But here (in the grammatical conception of gender) both words
mean the occurrence of the action itself, decrease (samstydna) is
feminine and increase (pravrtti) is masculine.77
Kasya punah stydnam strt pravrttir vd pumdn?

(Sutra-proponent:) But of what is feminine gender the decrease
or masculine gender the increase?
Gundndm.
(Sutra-rejecter:) Of the qualities.
Kesdm?
(Sftra-proponent:) Of which qualities?
Sabdasparsariuparasagandhdndm. Sarvas ca punar murtaya evam

dtmikdh sa mstydnaprasavagundh sabdasparsaruparasagandha-
vatyah. Yatrdlptydmso gunds tatrdvaratas trayah sabdah spargo
rupam iti. Rasagandhau na sarvatra. Pravrttih khalvapi nity. Na
hlha kagcid api svasminndtmani muhurtam apy avatisthate.
prdnigarbhavimocane. Dhatupatha 2.21. Bohtlingk, vol. 2, p. 71. Kaiyata gives

the suffix masun following Vsuni and justifies the shortening of the u. Nagesa
remarks that "Pater dumsun" (Pancapadyunadisutra 4.178. SK #627 p. 559. See
Amarakosa 2.6.1, p. 203) belongs to another grammatical school and has no
authority. In his Ramasrami commentary on the Amarakoga, Bhanurji Diksita
gives the alternate reading "Punho dumsun." But see MPV, vol. 4, p. 116, note 9:
Pumso 'sun iti sutrlyasekhare tu msunpratyaya iti drsyate.
75Kaiyata: Samghatarupam prdpnoti. Obtains a combined state. MB, vol. 4, p.
22b. Carakasamhita, Sarirasthana 3.3 (p. 308) speaks of the soul (jiva) entering
the combination of the male's sperm and female's blood
("...sukrasonitasamsargam... jivo 'vakramati....")
76Kaiyata: Sute 'patyam janayati. Procreates, generates offspring. MB,
p. 22b.
77Kaiyata: Samstydnam iti tirobhdvah, pravrttir dvirbhdvah, samyavasthd
sthitih, etds cdvasthdh sabdagocard evety avaseyam. "Samstydnam"
disappearance, pravrttir manifestation, state of balance: maintenance. And
one should conclude that these states are the meanings of the words (feminine,
masculine and neuter respectively). MB, vol. 4, p. 22b.
Nagesa: Stydnam apacayah. Pravrttir vrddhih. Stydnam waning.
Pravrttir growth. MB, vol. 2, p. 98a.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Vardhate yavad anena vardhitavyam apacayena va yujyate. Tac

cobhayam sarvatra.
(Sitra-rejecter:) Of sound, touch, visual appearance, taste and
smell. And what's more, all manifest entities (miirti) are of this
nature: the qualities, viz., decrease and increase, are in them and
so are sound, touch, appearance, taste and smell. Where there are
fewer qualities there are at least three: sound, touch and
appearance.78 Taste and smell aren't everywhere. Moreover,
evolution is continuous. Nothing in the world remains in its own
state even for an instant. It grows as much as it can grow or it
begins to wane. This pair (growing and waning) is everywhere.79
Yady ubhayam sarvatra kuto vyavasthd?

(Sutra-proponent:) If it is everywhere, whence the arrangement?
78Kaiyata: Avarata iti. Antata ity arthah. MB, vol. 4, p. 23b.

79Growth and waning are two of the six changes which entities undergo
according to Varsydyani. See Nirukta 1.2 (part 2, pp. 6-7; part 3, p. 29).
The list of the five basic elements (tanmdtra) is representative of all manifest
entities because all such entities consist of these basic elements according to
Sankhya. Original matter (prakrti), comprised of three fundamental tendencies
sattva, rajas and tamas, after becoming intelligence (citta, or buddhi and then
ahamkdra, ego) becomes the five basic elements (tanmdtra) sequentially. These
in turn manifest all that we ordinarily know. See Nagesa MB, vol. 4, p. 22a, and
EIP, vol. 4, pp. 23-29.
Kaiyata: Gundndm sattvarajastamahparinamarupdna m sabdddindm vrddhihrdsa-
madhyasthdvasthdh sabdaikagocardh lingatvena parigrhyante. A tmanas t v
aparindmirupatvdd aprav.rttidharmatve 'py anyaprav.rttyanuydyitvdl lingayogah.
The growth, diminution and intermediate states of the qualities sound, etc.
which are transformations of sattva, rajas and tamas and which can only be
known by means of words, are taken to be gender. But, even though it does not
have the property of increase because it is nonchanging, the self (dtman) is
associated with gender because it takes on the (property of) increase belonging
to other things. MB, p. vol. 2, p. 98a-b.
Patafijali explains the theory of gender as applying to all manifest entities
(miirti) whether such entities be considered to include only those objects
consisting of all five basic entities, or those objects which have at least three of
them (fire, water and earth). See Kumarila ?3.3.4 note 36.
Nagesa: Traya iti... Vastutas tejobannariipanam sarvamuirtlnam trivrtkarana-
mdtrdbhiprdyenedam uktam, vdyvdkdsasabdavdcyayor lingavyvahdras tu isvara
iva svakdryatejobannagatopacaydropena. ... Evam ca prakrtiparindmavisesa-
rupajdteh svata eva lingayoga iti tdtparyam. Actually, Patafjali says 'traya'
intending merely the three-fold occurrence of all manifest entities consisting of
fire, water and earth. But the usage of masculine and neuter terminations as if
the objects denoted by the words 'air (vdyu, masculine)' and 'space (dkdsa,
neuter)' had gender proceeds just as it does in the case of the Lord (isvara,
masculine) by superimposing on him the increase present in fire, water and earth
which are his effects. ... The point is that in this way a generic property, which
is a specific transformation of original matter (prakrti), in and of itself is
associated with a certain gender. MB, p. vol. 2, p. 98b.
Chandogyopanisad 6.3 mentions three causes of entities. Sankara refers to
these (trivrtkarana) as fire, water and earth (tejahprabh.rti). Ten Principal
Upanishads, pp. 511-12.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Vivaksdtah. Samstydnavivaksdydm stri prasavavivaksadym

pumdn ubhayor apy avivaksayam napumsakam.
(Sfitra-rejecter:) From a speaker's intention. When a speake
intends decrease, the object is treated as feminine; when a speak
intends increase, masculine; when a speaker does not intend eith
of them, neuter.80
Tatra lingavacanasiddhir gunavivaksdnityatvdd iti lingaparihdr

upapannah. Vacanaparihdras tu nopapadyate.
(Sutra-proponent:) In that case by saying, "Gender and number
succeed because a speaker's intention of a quality is not innate,"
may be possible to avoid the gender problem but it is not possib
to avoid the number problem.
[53c.] Vacanaparihdras capy upapannah. Idam tdvad ayam

prastavyah. Atha yasya dravyam padarthah katham
tasyaikavacanadvivacanabahuvacandni bhavantiti. Evam sa
vaksyati. Ekasminnekavacanam dvayor dvivacanam bahusu
bahuvacanam iti. Yadi tasyapi vacanikani na svabhdvikany aham
apy evam vaksyamy ekasminnekavacanam dvayor dvivacanam
bahusu bahuvacanam iti. Na hy dk.rtipadarthikasya dravyam na
paddrtho dvavyapaddrthikasya vdkrtir na padarthah. Ubhayor
ubhayam paddrthah kasyacit tu kimcit pradhdnabhutam kimcid
gunabhutam. A k.rtipadarthikasydkrtih pradhdnabhutd dravyam
gunabhuitam. Dravyapaddrthikasya dravyam pradhdnabhuitam
dkrtir gunabhutd.
(Sfitra-rejecter:) It is possible to avoid the number problem too.
The proponent of the view that an individual substance is denoted
has to be asked this: "How do singular, dual, and plural
inflectional terminations occur for him for whom an individual
substance is the object denoted by a word?" He will state, "The
singular occurs for one object, the dual for two objects, and the
plural for many objects."81 If they have to be explicitly stated
80The passage, "Therefore, grammarians cannot accept the ordinary

gender...when a speaker does not intend either of them, neuter, (Tasmdn na
vaiyakaranaih sakyam laukikam lingam asthatum. ...ubhayor apy avivaksdyam
napumsakam)," cites Patafijali on 4.1.3, Striyam. MB-K, vol. 2, p. 197, line 25 -
p. 198, linelO.
81Panini 1.4.21. Bahusu bahuvacanam. The nominal terminations (sup) called
plural (bahuvacana 1.4.102) occur if there are many objects denoted.
1.4.22. Dvyekayor dvivacanaikavacane. The nominal terminations (sup) called
dual (dvivacana 1.4.102) occur if there are two objects denoted, and those called
singular (ekavacana 1.4.102), if there is one.
The first vdrttika with Pataiijali's commentary gives the reason for stating the
suttras as follows:
Panini 4.1.2 and 3.4.78 provide that nominal and verbal affixes occur after
nominal bases and verbal roots respectively without specifying which of the
terminations of each set occur in which meanings. Yet Panini does not sanction
using the terminations in the wrong meanings. However, occasionally we do see

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
and don't occur naturally for him, then I too will state, "Th
singular occurs for one object, the dual for two objects, and the
plural for many objects." Because it is not the case that an
individual substance is not denoted for him for whom a class
property is denoted, nor that a class property is not denoted f
him for whom an individual substance is denoted. Both are
denoted for both. But for each something is principal, someth
subordinate. For him for whom a class property is the o
denoted the class property is principal and an individual subst
is subordinate. For him for whom an individual substance is the
object denoted an individual substance is principal and its class
property is subordinate.82
the wrong terminations used. For example, in "My eyes (aksini) are beautiful
(darsanlydni); my feet (pddah) are tender (sukumardh)," the plural termination is
used instead of the dual when there are only two eyes and two feet rather than
many. Because we do see such usage and because the terminations are provided
without specification, incorrect usage would result. To prevent that, Panini
states 1.4.21-22 as a restrictive rule. MB-K, vol. 1, p. 322, lines 1-8. See
Cardona (1988: 41-45, 174, 183, especially 177-78).
82Kaiyata: Vdcanikaniti. Samkirnaprayogopalambhdt asamkardrtho yathd
dravyapaddrthikasya vdcaniko vacananiyamah tathdkrtivddino 'pi. Sa hy evam
vydcaste ekas'rayasambandhavivaksdyam dk,rter ekavacanam yathd pasuna
yajeteti. Anekisrayasambandhavivaksaydm dvivacanabahuvacane. Nanu
dravyasydnabhidheyatvdt katham tatkrta dkrter vacanopakdra ity dha: nahiti
Tatra yathd pacati pacatah pacantiti kriydprddhdnye 'pi sddhanasamkhydkrto
vacanabhedah evam dkrtiprddhdnye 'pi dravyasamkhydkrta ity adosah. They
have to be explicitly stated. Because we do find mixed usage (such as
the dual termination on 'godau' meaning a single village near two ponds and the
plural termination on 'pancdalah' meaning a certain single district (4.2.81,
1.2.51)), the proponent of the view that an individual substance is the object
denoted by a word has to state explicitly a restrictive rule concerning number in
order to prevent mixing them up (i.e., the use of one set of endings where
another should occur). Similarly, the proponent of the view that a class
property is the object denoted by a word has to state it explicitly too. He
explains (sutras 1.4.21-22) as follows: A singular termination occurs if a
speaker intends the class property as associated with a single substratum, for
example, "One should sacrifice an animal (singular)." Dual and plural
terminations occur if a speaker intends the class property as associated with
more than one substratum. To answer the objection, "But if a substance is not
denoted, how can it help (determine) number for the class property?" he says:
Because it is not the case. In "(he) cooks, (they two) cook, (they) cook,"
although action is the principal meaning, the number of participants in the act
causes different number-terminations. Similarly, although a class property is
the principal meaning, the number of substances causes (different number-
terminations). Hence the view that a class property is denoted is without fault.
Nagesa summarizes the section as follows: Ity adosa iti. Evam ca lingayogah
svata eva. Samkhydyogas tv aropena. Ekd dkrtir iti pratijna naropita-
samkhydbhiprdyd, ndpy ekasesarambhah. Anekasabdaprayoge mdndbhdvdd iti
tatparyam. Atra bhdsye dravyapaddrthavdde dkrter visesanatvoktyd tajjdty-
anugatlkrtdnekavyaktibodhasyaikasmdd eva sabddt sambhavena ndndsabda-
prayogdprdptyd tatpakse 'py ekaseso na kdrya iti stcitam. Jdtivdde vyaktir
dravyavdde jdtih sakyatdvacchediketi ca sucitam. Hence it is without
fault. And in this way a class property has a connection with gender in and of

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
54. Gunavacanavad vd.

Gunavacanavad vJ lingavacanani bhavisyanti. Tad yathd
gunavacandndm sabddndm dsrayato litigavacanani bhavanti.
Suklam vastram sukld sadtt skuklah kambalah suklau kambalau
s'uklah kambald iti. Yad asau dravyam srito bhavati gunas tasya
yal linigam vacanam ca tad gunasydpi bhavati. Evam ihdpi yad
asau dravyam sritdk.rtis tasya yal lingam vacanam ca tad dakrter api
bhavisyati.
55. Adhikaranagatih sdhacarydt.
Akrtdv drambhanddlndm sambhavo ndstiti krtvdkrtisahacarite
dravya drambhanaddni bhavisyanti.
Or as they do for words denoting qualities.
Or different genders and numbers will occur as they do for words
denoting qualities, that is, as follows: Different genders and
numbers occur for words denoting qualities according to the
substrata in which the quality resides: suklam vastram (a white
cloth), sukla s.ti (a white sari), suklah kambalah (a white blanket),
suklau kambalau (two white blankets), suklah kambalah (white
blankets).83 The gender and number of the substance in which
the quality resides occur for the quality too. Similarly here too,
the gender and number of the substance in which the class
property resides will occur fof the class property too.84
Cognition of a substratum arises because it
accompanies the class property.
Considering that tying, etc., are impossible on a class property,
tying, etc., will take place on a substance which accompanies the
class property.85
itself but has a connection with number by superimposition. The thesis that a
class property is one does not intend the superimposed number (which
determines the use of number-terminations. Hence the thesis does not have to
be abandoned) and (hence) one need not commence (the rule) that one remains
because there is no evidence that many words would be used. That is the point.
It is implied in (Pataiijali's) present comment that (the rule) that one remains is
not necessary on the view that an individual substance is the object denoted
either. The use of many words does not obtain because, due to the fact that the
class property is denoted as a qualifier, one can know the many individuals
which include that generic property as a recurrent feature from just a single
word. It is also implied that an individual delimits the generic property's being
signified on the view that a generic property is denoted and a generic property
(delimits an individual's being signified) on the view that an individual is
denoted. MB, vol. 2, p. 99a.
83The examples show agreement of the word 'sukla (white)' with words in the
neuter and feminine singular, and with words in the masculine singular, dual,
and plural, respectively.
84Kaiyata: Yad asdv iti. Jdtitadvator abhedddhyavasdydd dravydtma jatih
pratlyamana tadgatalitgasamkhydgrahintty arthah. The meaning is: From
determining the generic property as identical to that in which it resides, the
generic property appears to have the nature of the substance and assumes the
gender and number which belong to it. MB, vol. 2, p. 99b.
85Varttika 55 answers vdrttika 47 as well as completing vdrttika 54's answer to

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
56. Na caikam anekadhikaranastham yugapad ity

adityavad visayah.
Na khalvapy ekam anekadhikaranastham yugapad upalabhyata
ity adityavad visayo bhavisyati. Tad yathd. Eka ddityo
'nekddhikaranastho yugapad upalabhyate.
(If you object) "And it is not the case that one thing is
present in many substrata simultaneously," (I answer:)
The domain is like that of the sun.
(If you object) "Moreover it is not the case that one thing
present in many substrata simultaneously," (I answer
domain will be like that of the sun. For example, the one
found located in many places simultaneously.
Visama upanydsah. Naiko drastdnekddhikaranastham d

yugapad upalabhate.
(Sitra-proponent:) The example given is not the same.
the case that one observer sees the sun present in man
simultaneously.
vdrttika 46.
Kaiyata: Codandsu ca tasydrambhdd iti yad uktam tatparihdrdrtham dha:
adhikaranagatir iti. Yathagnir dniyatdm ity ukte kevalasydgner
dnayandsambhavdn ndntarlyakatvdd acoditam api pdtram dntyate etad evdgner
dnayanam yat pdtrasthasya, tathdk,rtdv drambhanddini codyamdndni sdmarthydt
sdhacarydd dravyam abhinivisante. Sarva evdkrteh kriyayogo
'ntarbhdvitadravydyd eveti dravyadvdrakah sampadyate. He says (vdrttika 55),
"Cognition of a substratum arises because of it s
accompaniment," to rebutt (vdrttika 47) "And because in in
(one) commences (the action) on (a substance)." When someon
fire," one brings its pot even though it is not enjoined because it i
included due to the fact that it is impossible to bring fire b
bringing of fire consists precisely in bringing it in a pot. Similarl
enjoined as if on a class property, apply to a substance becaus
(on a substance) and because the substance accompanies the class
the action connected with a class property is connected with it
substance and so takes place by way of the substance. MB, vol. 2,
Nagesa: Sdbdabodhe tu na bddhajfidnam pratibandhakam.... But the
controverting knowledge (that it is impossible to perform action on a class
property) is not an obstacle to verbal cognition (of the class property).... MB,
vol. 2, p. lOOa.
Sivaramendrasarasvati: A krtau coditcydm dravya drambha.ndikaranam
asdstriya[m iti] duisanam dattam, tat pariharati: adhikaranagatir iti.
Sahacarydd iti. Akrtivyaktyor atyantabheddbhdvdd vyaktau krtam dkrtdv eva
krtam iti na tasys'astriyateti bhdvah. (Under vdrttika 47) the following fault
was attributed (to the view that a class property is denoted): Performing the
tying, etc., on a substance when a class property has been enjoined (as that on
which to perform it) is not in accordance with the instruction. To avoid it he
says: Cognition of a substratum arises. Because it accompanies.
Because a class property and an individual are not entirely distinct, what is
done on an individual is done on the class property. Hence it certainly is in
accord with the instruction. That is what he has in mind. MPV, vol. 4, p. 119.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Evam tarhitindravad visayah. Tad yathd. Eka Indro 'nekasmin

kratusata ahuto yugapat sarvatra bhavati. Evam dkrtir yugapat
sarvatra bhavisyati.
(Sutra-rejecter:) If that is so, then the domain is like that of
'Indra'. Just as one Indra, invoked in many hundreds of Vedic
performances, is simultaneously everywhere, thus a class property
will be simultaneously everywhere.86
57. A vinaso 'ndsritatvdt.
Dravyavinasa dkorter avindasah.
A class property is not destroyed because it is not
dependent.
A class property is not destroyed when a substance in which it
resides is destroyed.
Kutah?
Anas.ritatvat. Andaritadkrtir dravyam.

(Sutra-rejecter:) Because it is not dependent. A class property
does not depend on a substance.87
Kim ucyate 'nasritatvdd iti yad iddnlm evoktam adhikaranagatih

sdhacarydd iti.
(Sutra-proponent:) How can you say, "Because it is independent,"
when just now you said, "Cognition of a substratum arises because
it accompanies the class property (varttika 55)."
Evam tarhy avinaso 'naikdtmydt. Dravyavindaa dkrter avindsah.

(Sutra-rejecter:) If that is so, then it is not destroyed because it is
not identical. A class property is not destroyed when a substance
in which it resides is destroyed.
Kutah?
(SFtra-proponent:) Why?
Anaikdtmydt. Aneka dtmdk.rter dravyasya ca. Tad yathd.

Vrksastho 'vatdno vrkse chinne 'pi na vinasyati.
86vdrttika 56 answers vdrttika 48 and reiterates vdrttika 40.

Nagesa: Itindravad iti. Svabhdvdd dheturahitadrstantenapy anekavrttitvam
jdter abhyupagamyata ity arthah. Like that of 'Indra'. The meaning is: It is
accepted that a generic property is present in many objects because of its nature,
even though the inference given to establish that fact consists of an example
devoid of an inferential mark (hetu). MB, vol. 2, p. lOOa.
87Vdrttika 57 answers vdrttika 49.
Kaiyata: Anyatrdpi pratyaydbhidhdnayor dkrtinimittayor bhdvdn nityatvam
tasyd ity arthah. Because designation and cognition conditioned by a class
property occur with respect to other individuals, a class property is eternal. MB,
vol. 2, p. 100b.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
(Sutra-rejecter:) Because it is not identical. A class property and

a substance have separate identities. For example, a vine hanging
on a tree is not destroyed even when the tree is cut down.88
58. Vairipyavigrahau dravyabhedat.
Vairupyavigrahdv api dravyabheddd bhavisyatah.
Difference and analytic strings are due to the
difference of the individual substances in which a class
property resides from each other.
The difference among individuals of a kind, and analytic strings
illustrating the meaning of words in the dual and plural, will be
due to the fact that the individual substances in which a class
property inheres are different from each other.89
59. Vyarthesu ca sdmdnyat siddham.
Vibhinnarthesu ca scmanyat siddham sarvam. Asnoter aksah.
Padyateh pddah. Mimfter mdsah. Tatra kriyd-sdmdnydt siddham.
Aparas tv dha. Purdkalpa etad adst sodasa mdsah kdrsdpanam
sodasaphalas ca masasamba.tyah. Tatra samkhyasamanydt
siddham.
And for words having different meanings, the
occurrence of just one word is achieved because there
is a general property.
And for words having different meanings, the occurrence of just
one word for all of the meanings is achieved because there is a
general property which they all have in common. In all
meanings, 'aksah (a die, an axle, a seed)' is derived from Vas (to
pervade), 'pddah (a foot, a verse segment, a quarter)' from V/pad
(to step) and 'mdsah (a bean, a monetary unit, a fool)' from V/ma
(to measure).90 Since that is so, the occurrence of just one word
88Kaiyata: Anaikdtmyad iti. ... nityd dkrtir ekd sarvatragd, tadviparltam

dravyam iti bheddd dravyavindse 'pi na jdter vina'ah. Because it is not
identical. ... A generic property is not destroyed even when a substance (in
which it resides) is destroyed because of this difference: a class property is
eternal, one and all pervading; a substance is the opposite. MB, vol. 2, p. 101a.
89Because the class property is not identical to the substances in which it
inheres, the fact that each substance has its individual differences does not
prevent it from having the class property. Vdrttika 58 answers vdrttikas 50-51.
Kaiyata: Vairupyavigrahav iti. Dravydndm svagatabhedapratibaddhau
vairupyavigrahdv ity arthah. Tatra dravyagatabhedopacdrdd ekasydm apy dkrtau
samuccayopapattyd vigraho na virudhyate. Difference and analytic strings.
Difference and analytic strings are connected with differences present in the
substances themselves. Since that is so, even though there is a single class
property, conjunction makes sense by attending to the differences present in the
substances. Hence an analytic string is not prevented.
Nagesa: Dravyagatalihgddigrdhitvavat tadgatabhedagrdhitvam api tasyd iti
bodhyam. One should understand that a class property assumes the differences
present in the substances (in which it resides) as it assumes the gender and so
on present in them. MB, vol. 2, p. 101a.
90As.u vydptau samghdte ca. Dhatupatha 1265. SK, p. 735. Bohtlingk, 5.18;
vol. 2, p. 75.
Pada gatau. Dhatupatha 1169. SK, p. 734. B6htlingk, 4.60; vol. 2, p. 73.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
is achieved because the action constitutes a general property

(kriyd-sdmanya).91
Man mcne sabde ca. Dhatupatha 1088. SK, p. 734. Bohtlingk, 3.6; vol. 2, p. 72.
According to the Amarakosa, a masa is a measure (5.1.34), a bean (5.2.4) and an
equal part (5.4.59).
91Varttika 59 answers vdrttika 52. It does not concern the view that a word
denotes a substance at all. Varttika 52 posed the objection that the rule that one
remains would have to be stated even if a word denoted a generic property
because some words refer to objects of more than one kind. Those words would
denote more than one generic property. The rule that one remains would have to
be stated in order to achieve the desired result that one occurrence of the word
conveys the cognition of all of the generic properties concerned. Varttika 59
answers that such words denote a more general property broad enough to
include all the kinds of objects to which the word refers.
Kaiyata explains this passage as meaning that there is one general property in
all of the action instances of the action denoted by a certain verbal root. The
general property inheres in each action instance, and each action instance
inheres in an individual substance. This explanation justifies understanding
'kriyd-samdnya' as a saptamitatpurusa compound. Nagesa seems to favor certain
others who take 'kriyd-sdmdnya' as a karmadhdraya compound meaning an
action which is a general property. They consider that coextensive with that
general action there is a generic property which inheres in the individual
substances in which the action occurs. Finally, Sivaramendrasarasvati
understands that the single general action inheres directly in each individual
substance. The action itself constitutes a general property because it is a single
entity residing in many substrata.
Kaiyata: Kriyasamanyad iti. Bhinndsu kriydsu abhinnapratyaydbhidhdna-
nimittam yat sdmdnyam tad eva dravyesv api tannimittam bhavati. Yathd
pdcaka ity israydntaragatam api hi sdmanyam samavetasamavdydd dravye
upakaroti. Yathd gairikddigatam lauhityam samyuktasamavdydt pate lauhitya-
pratyayam ddadhdti. Because there is a general property in the
actions (kriyd-sdmdnya). The general property which causes identical
cognition and denotation to arise with respect to distinct actions causes
identical cognition and denotation to arise with respect to (the different)
substances (in which those actions take place) too. It is like 'cook (pdcaka)'.
The general property (of cooking, pdkatva), even though present in a different
substratum (a cooking action instance which inheres in a cook rather than in the
cook himself), serves in the substance (the cook) because it inheres in that
which inheres in it. It is like this: The red color present in red chalk (used as
dye) is the basis for the cognition of red in cloth because it inheres in that
which is in contact with it (the color inheres in the chalk which is in contact
with the cloth). MB, vol. 2, p. lOla.
Nagesa: Evam ca vydptyddigatasdmdnyasya paramparayd dravye sattvena tattaj-
jdtivisistasya dravyasya vdcyatvam iti bhdvah. Na caivam ndndrthocchedah,
ndndjdtnadm sambandhaghatakatvena ndndrthatvopapatteh. Tattatkriydriupa-
sdmdnyasamaniyatasydksatvdder jdtirutpasydksapaddrthatrayasddhdranasydpi
sakyatdvacchedakatvasakyatvdnyatardngfkdrdt siddham iti bh.sydrtham anye.
Sutrdrambhe 'pldam dvasyakam ekaruipenaiva sarvdrthabodhdd iti bhdva iti
bodhyam. And so for example, the word 'aksah' denotes a substance qualified
by a certain specific generic property (either being a die, being an axle or being
a seed) by virtue of the fact that the general property of pervasion (vydptitva)
which inheres in the pervasion action instances is present indirectly in the
substance (a die, an axle or a seed). That is what he has in mind. Nor is it the
case that in this way the different meanings are undermined, because having
different meanings attains by including the relation to the various specific

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
generic properties. Others say that the bhdsya passage means the following: It
succeeds by accepting, for example, the generic property aksatva, which is
coextensive with the general property in the form of a certain action, as the
universal meaning of the word 'aksah' either as the signified (sakya) or as the
delimiter of what is signified (sakyatdvacchedaka). (If aksatva is signified, then
either being a die, being an axle or being a seed is the delimiter of what is
signified; alternatively, if aksatva is the delimiter of what is signified, then
either being a die, being an axle or being a seed is signified.) This (vdrttika 59)
is necessary even if sutra (1.2.64) is undertaken, because, from the use of the
word 'aksa', one apprehends all the objects as being of the same type. It should
be understood that this is what he has in mind. MB, vol. 2, p. 10lb.
Even when the word 'aksah' is used in the singular it gives rise to all three
meanings. The rule that one remains (ekasesa) assumes that one word 'aksah'
occurs in just one of the meanings, a second word 'aksah' occurs in the second
meaning, and a third word 'aksah' in the third meaning. Applying the rule that
one remains, one attains the word 'aksah' in the plural denoting all three
meanings. By a similar procedure, 'aksau' in the dual could denote no more than
two meanings and 'aksah' in the singular could denote no more than one of the
meanings. Since the rule that one remains does not achieve the desired result,
one must accept that there is a single general property common to all the
individual substances to which the word refers in all of its meanings. That one
arch general property is the object denoted by the word. Note that this is indeed
Patafijali's intent as is shown by the fact that he uses the words 'aksah', 'pddah',
and 'masah' in the singular rather than the plural as he had when previously
discussing these so-called polysemous examples before vdrttika 1 and under
vdrttika 52.
Both Kaiyata and Nagesa depart from Patafijali's statement in the following
respect: They do not accept that the action itself constitutes the general
property inherent in all the individual substances to which a polysemous term
refers. Instead Kaiyata accepts that the action general property has an indirect
relation to the individual substances by way of action instances. Nagesa
accepts a different general property coextensive with the action general
property. According to him this other general property inheres in the
individual substances. They do not accept that the action itself is a generic
property inhering directly in the individual substances for the following
reason: They wish to conform to certain notions concerning general properties
which are accepted in order to avoid having generic properties of different
categories in a single individual. With regard to the case at hand, this implies
that an action general property can only occur in action instances and the action
instances can inhere in substances, but an action general property cannot inhere
in substances.
However, Patafijali accepts that a single action occurs again and again in many
instances. The action itself has the character of a general property present in the
substances which are its substrata. See "ekd hi kriyd" under vdrttika 19 of the
present suitra. MB-K, vol. 2, p. 238, line 10. For another passage in which this
conclusion is cited see MB-K, vol. 2, p. 418, lines 15-16. Patafijali specifically
shows that one action inheres in many substances commenting on sutra 3.1.67.
He says, "all those who perform a certain action are its substrata (ydvantas te tdm
kriydm kurvanti sarve te tasyd dsrayd bhavanti)." MB-K, vol. 2, p. 56, lines 20-
21. The agent (kartr) or direct object (karman) are considered to be the
substratum (dsraya) of action. See the discussion beginning, "Bhdvah punar eka
eva," under 3.1.67. MB-K, vol. 2, p. 56, line 18 - p. 57, line 12.
Sivaramendrasarasvatl gives the correct interpretation of the bhdsya passage:
Kriydsdmdnyad iti. Pdcakddisabdavad aksddisabddndm kriydiabdatvam
asritya vydptyddikriydrupam sdmdnyam pravrttinimittam.... Because the
action constitutes a general property (kriya-samanya). Accepting

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
But someone else says: In a previous age (kalpa) there were

mdsas to a kdrsapana and beanpods having 16 masas (beans)
Since that is so, the occurrence of just one word is achieve
because the number (16) constitutes a general property (samkhy
sdmanya).93
the word 'aksa', for example, as a kriydiabda (word denoting action) like t
words 'pdcaka (a cook)', etc., the general property in the form of the action
pervading is the word aksa's condition for usage (pravrttinimitta).... MPV,
4, p. 120.
92A kdirsapana is a coin or weight. As a weight of gold it equals a karsa which
Monier-Williams defines as follows: A weight of gold or silver (= 16 mdsas =
80 rettis = 1/4 pala = 1/400 of a tula). In common use 8 rettis are given to the
mdsa. Monier-Williams, p. 259, 276. Nagesa, calling a retti a guijd, confirms
that while previously there were 5 guijds to a mdsa and 16 mdsas to a kdrsapana,
now there are 8 guijds to a mdsa and 10 mdsas to a kdrsapana. See the next note.
93Kaiyata: Pura kalpa iti. Purd kalpe prasiddho 'rthah samprati nimittdbhdve
'pi tathaiva vyapadisyate. Masasambatya iti. Tatra sodasabhdgatvena mdsah
prasiddhah tathaiva kdrsdpane mdsah [sodasabhdga iti] sodasa[bhdga]tvam
ndma sadmnyam upddatte masasabda iti siddham ekas'abdatvam.
Samkhydsamanyad iti. Samkhyaiva sdmdnyam abhinnapratyayahetutvdt.
Nanu sodasasamkhyd sodasasu paddrthesu vyasajya sthitd na tv ekasmin
sodase. Evam tarhi sodaasasmkhydpuranatvam ndma yat sdmanyam
tatsamkhydmualatvdt samkhydsdmanyam ucyata ity adosah. ... Tad evam
sarvatrdkrtisadbhdvdt siddham ekasabdatvam. In the previous age. In the
previous age the meaning was well known; now it is signified in the same way
even though the cause is absent. Beanpods. In (a beanpod) a nmdsa (bean) is
known as a sixteenth part. Similarly a mdsa is a sixteenth part of a kdrsdpana.
Hence the word 'mdsa' receives the general property being a sixteenth part, so
the occurrence of just one word is achieved. Because the number
constitutes a general property. The number itself is the general property
(samkhyd-sdmdnya is a karmadharaya compound). (Objection:) But the number
16 stands encompassing 16 entities not one sixteenth. (Answer:) In that case
the general property, namely, being what completes the number 16 (i.e. being
the ordinal number, a sixteenth), is called a number-general property (samkhyd-
sdmdnya) because it is based on that number. Hence it is without fault. ... In
this way the occurrence of just one word is justified because of the existence of
a class property (common) to all (the word's) meanings. MB, vol. 2, p. 101b.
Nagesa: Masasambatyam mdsah sodasabhdgapalaruipah. Kdrsdpane mdsah
panicaguijdmitah sodaso bhdgas tayoh sodaiabhdgatvena sdmdnyena sarvesu
masatvavyavahdrah. Iddnim tv astaguijamitamdse sodasatvam aropya tatha
prayoga iti bhdvah. In a beanpod, a mdsa, in the form of a bean, was a sixteenth
part. In a kdrsapana a mdsa measured by 5 gunjds (a berry or small weight) was a
sixteenth part. Being a mdsa applied to both because both had the general
property of being a sixteenth part. But now the usage is the same
superimposing being a sixteenth on a mdsa measured by 8 guijds. That is what
he has in mind. MB, vol. 2, p. 102a.
It is clear from the fact that ancient weights were named for the berrys, beans, etc.
which had those weights, and from the fact that coins were named for certain
weights of precious metals, that a word such as 'mdsa' had just one primary
meaning on which the others are based.
In his critical apparatus Kielhom notes that all MSS. except one have '-palds',
which the other editions prefer, instead of '-phaldi'. One might infer from this
that 'palds' indeed should be accepted, and that Patafijali is not referring to
beans and beanpods here but to other measures of weight.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Sivaramendrasarasvati notes that even in the present age a karsdpana is equal to

16 mdsas but there are not 16 mdsas in a mdsasambati. In Trailinga language
(Telegu) a sambati is called gifjd, and a mdsasambati, minayagiiijd. There are
16 gold mdsas in a kdarsdpana and 16 gold palas in a mdsa. MPV, vol. 4, p. 120.
But Guruprasadasastri notes that a sambatl is known as 'chimi (a peapod)' or
'phali (a beanpod)'. MBPU, vol. 1, part 2, p. 140, note 1. Filliozat (1975-86:
vol. 4, p. 370 note 4, p. 374 note 1) discusses the possibilities thoroughly and
(1991: 24) more briefly. Kielhor, in the preface to his first edition (vol. 3, p.
15, note 2 of p. 14), comments that the passage should read as given in the
revised edition which is repeated and translated accordingly here.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Nyaya
2.1. Background
The Nyaya system of Indian philosophy analyzes the

means of gaining knowledge (pramana) and the objects of
knowledge (prameya) in order to arrive at correct knowledge
(tattvajndna). From correct knowledge one attains the highest
good (nihs'reyasa).1 The elimination of false knowledge leads to
absolute liberation (apavarga) from suffering (duhkha) and its
causes.2 Nyaya recognizes four means of gaining knowledge, the
fourth of which is verbal evidence (sabda).3 The instruction
(upadesa) of someone who directly knows something (apta) and
just wants to tell what he knows constitutes verbal evidence.4
Depending upon whether he tells about ordinary things or about
transcendental things, verbal evidence is ordinary or Vedic.5
Sound (sabda) is the one of the five basic qualities (guna)
which belongs to space (akdas) and which one perceives by
means of the ear (srotra).6 It is not eternal but is created and
destroyed and if it appears to change into a similar sound, in fact,
it is a new sound of similar quality which appears in its place.7
Sound is of two types, plain noise (dhvani) and speech sound
(varna).8 Speech sounds singly and together constitute
11.1.1. Pramana-prameya-samsnaya-prayojana-drstdnta-siddhdntavayava-tarka-
nirnaya-vdda-jalpa-vitandd-hetvdbhdsacchala-jdti-nigrahasthdndndm tattva-
jnanan nihsreyasadhigamah.
21.1.2. Duhkha-janma-pravrtti-dosa-mithydjnadnndm uttarottarapdye
tadanantarpadydd apavargah.
31.1.3. Pratyaksdnumdnopamana-s'abddh pramdnani. Perception, inference,
comparison and verbal testimony are the means of valid knowledge.
41.1.7. Aptopadesah sabdah. Vatsyayana comments: Sdksdtkaranam
arthasydptih, taydpravartata ity dptah. ND, p. 174. Biardeau (1964: 209 note
1) mentions that the expressions 'dptavacana (parole des dpta)' and 'dptdgama (la
tradition des dpta)' are used in Sfankhya.
51.1.8. Sa dvividho drstddrstarthatvdt. Vatsyayana: Yasyeha drsyate 'rthah sa
(sabdo) drstdrthah, yasyamutra pratlyate so 'drstdrthah. Evam
rsilaukikavdkyndam vibhdga iti. ND, p. 179.
61.1.14, 3.1.63. See trans. under 2.2.67 note 35.
7Vatsyayana introduces the final view saying, "Anityah sabda ity uttaram." ND,
p. 594. 2.2.13, Adimattvdd aindriyakatvdt k,rtakavadupacdrdc ca, and 2.2.18,
Prdg uccdrandd anupalabdher, dvarandd anupalabdhes ca, give most of the
reasons for this conclusion.
8Vatsyayana introduces 2.2.40 saying, "Dvividhas cayam sabdo varndtmako

dhvanimdtras ca." ND, p. 638.
151

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
grammatical constituents such as verbal roots, nominal bases, an

verbal and nominal terminations. Ending in a verbal or nominal
termination, speech sounds constitute a word. Words have
meaning.9
The relation of a word to its meaning is conventional
(samayika), not innate (svdbhdvika).'l Speech is a valid means of
knowledge only in so far as it is spoken by a trustworthy
authority.1" A person is a trustworthy authority because he has
direct knowledge and out of compassion for other creatures wants
to tell them.12 One infers that speech is authoritative by
confirming that what it says is correct. One infers that the speaker
is knowledgeable and benevolent from the beneficial results of his
instruction. One infers that speech regarding transcendental
matters (such as Vedic texts) is authoritative if it is spoken by such
a person.13
2.2. The Nature of a Generic Property, an

Individual, and a Form
In Nyaya, the discussion of what the object denoted by a

word is centers around three entities, a generic property (jati), an
9See 2.2.58 trans.

102.1.55. Na sdmayikatvdc chabddrthasampratyayasya. Vatsyayana: Yat tad
avocdma asyedam iti sasthlvisistasya vakyasydrthaviseso 'nujndtah
sabddrthayoh sambandha iti samayam tam avocdma iti. Kah punar ayam
samayah? Asya sabdasyedam arthajdtam abhidheyam iti abhidhdndbhidheya-
niyamaniyogah. ND, p. 543. Biardeau (1964: 208). Vaisesika sutra 7.2.24,
Sdmayikah sabddd arthapratyayah, expresses the same. VS, p. 60. Gautama
2.1.56 and commentary argue against the view that a word and its object have a
natural relation. Vaisesika sutra 7.2.19, Sabddrthdv asambaddhau, explicitly
rejects such a natural relation. VS, p. 59. See Dash (1991), Houben (1992a), and
Matilal (1990: 26-30).
112.1.52. Aptopadesasdmarthydd arthasampratyayah. Vatsyayana: Svargah,
apsarasah, uttardh kuravah, sapta dvipdh, samudrah, lokasannivesah ityevam-
dder apratyaksasydrthasya na sabdamdtrdt pratyayah. Kim tarhi? Aptair ayam
uktah sabda ity atah sampratyayah. Viparyayena sampratyaydbhdvdt. ND, p.
536. 2.1.68. Mantrayurvedapramdnyavac ca tatpramanyam dptaprdmdnydt. The
suttra states an inference for the authority of the Veda, but for all speech it is the
case that the authority of the speech depends on the authority of the speaker.
Vatsyayana concludes his commentary saying, "Aptaprdamnydc ca prdmdnyam,
laukikesu sabdesu caitat samdnam iti." ND, p. 570. Matilal (1990: 49-74) and
the several articles in Matilal 1994 discussed in the literature review above deal
with this topic.
12Under 2.1.68 Vatsyayana describes the authority of an authoritative speaker
in detail beginning with the following brief statement: Kim punar dptdnam
prdmdnyam? Sdkstk.rtadharmatd bhutadayd yathdbhutdrthacikhydpayiseti. ND,
p. 565-66. See Matilal (1990: 6).
13See Vatsyayana on 2.1.68. ND, p. 567.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
NYAYA 153
individual (vyakti), and a form (akrti). Gautama defines them in

the following three sutras:
2.2.67. An individual is the physical body which is the

substratum of specific qualities.
2.2.68. A form is that which makes known the generic
property and its indicatory marks.
2.2.69. A generic property is that whose nature is to
produce the same (cognition).
Vatsyayana explains that a physical body (muirti) is a

finite substance consisting of parts which is the substratum of
certain qualities. Only a substance consisting of earth, water, fire,
and air which is not an atomic constituent (paramdnu) fits this
description. Such a substance is termed a manifestation (vyakti)
because one can perceive it with the sense organs (2.2.67).
Form (dkrti) is that by which one knows the generic
property of an individual (vyakti) or knows the parts of the
individual which in turn serve as inferential marks of the generic
property. It is the configuration of the parts (avayavasamsthdna)
of an individual or the configuration of the parts of the parts of
an individual. If one sees a cow at a distance, the figure itself,
including such features as a dewlap, head, feet, tail, and hump, is
the inferential mark that the generic property cowness inheres in
that individual cow. If one sees the head of a cow peering around
a corner, or its foot, one can infer from the configuration of the
parts of that part that it is the head or foot of a cow. That head or
foot, being a part of the whole cow is an inferential mark that the
generic property cowness is in the individual whose head or foot
it is (2.2.68 VI, U1).14
Although every form indicates a certain generic property,
not all generic properties are indicated by a form (2.2.68 V2,
U2). A form serves as an inferential mark of a generic property
only if it is fixed (niyata) for the life of the individual substance.
If the configuration changes but the individual substance remains
what it is, then the configuration does not indicate the generic
property. For example, clay, gold, and silver have no special
shape. One can mold them into many different shapes. Whatever
shape they assume they remain what they are: clay, gold, or
14Vacaspatimisra comments as follows: Sirahpdnyddivyuha k.rtir jdtim

manusyatvddikam acaste. SirondsikdlaldtacibukddTnam siro'vayavandm vyuho
manusyatvajdtilingam sira dcaste. The form, which is the configuration of the
head, hands, etc., makes known the generic property humanness, etc. The
configuration of the parts of the head, the head's nose, forehead, chin, etc.,
makes known the head which is an inferential mark of the generic property
humanness. ND, p. 692.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
silver. The configuration does not indicate what the substance is;
other qualities, such as color, taste, smell, etc., do.15
According to Uddyotakara, a configuration is a special
conjunction (samyoga) and is classed as a quality (guna) (2.2.67
U). He considers qualities to be individual instances which inhere
in an individual substance. A generic property inhering in all the
quality instances of a kind is the basis of the same cognition with
respect to each of them. Therefore, a form, being a special
conjunction, inheres in a single individual. Although the form of
one individual may be similar to the form of another, it is not
common to both individuals (2.2.64 U4a). Furthermore, as a
quality, it is immaterial and has no connection with action (2.2.60
U).
A generic property (jiti) is that which is the same in many
individuals and differentiates those individuals from others. It
generates the same cognition with respect to different individual
objects. It is what Vaisesika terms a general specific property
(sdmdnyavisesa), considered and called general or specific
depending on whether it serves to identify an individual as the
same as others or to distinguish some from others.16 Vaisesika
defines a general property as a single eternal entity which inheres
in more than one substance, quality, or action.17 In Vaisesika, the
15Vacaspatimisra makes it clear: Na punah sarvd jatir dkrtyd lihgyata

iti. Mrtsuvarnarajatddikd hi rupavisesavyangyd jdtir ndkrtyd, brdhmanatvddi-
jdtis tu yonivyangyd, djyatailddndam jdtis tu gandhena vd rasena vd vyajyate.
Ata eva na sdrsapddfndm tailatvam asti tadvyaijakayor gandharasayor abhdvdt,
bhdktas tu tailasabdaprayogah. But not every generic property i s
indicated by a form: A specific color manifests the generic property of
clay, gold, silver, etc., the family descent manifests (the generic property)
brahmana-ness (brdhmanatva), etc., the smell or taste manifests the generic
property of clarified butter, sesame oil, etc. Hence being sesame oil does not
belong to mustard oil, etc. because the smell and taste which manifest it are
absent, but the use of the term 'sesame oil' (for it) is secondary. ND, p. 692. See
Kumarila's Slokavarttika, Vanavada 26-29. SV, pp. 439-40.
16Vaisesika sutra 1.2.3. Samdnyam visesa iti buddhyapeksam. VS, p. 8. See
Halbfass (1992: 239, 256-58). Though a general property may be called
specific, it should not be confused with the unique properties of eternal entities
which are also called specific properties. See Halbfass (1992: 71-72). Halbfass
(1992: 113-37) discusses universals and (pp. 237-75) translates passages
dealing with them.
17This is most succinctly stated in the Tarkasamgraha (p. 114): Nityam ekam
anekdnugatam sdmdnyam, dravyagunakarmavrtti. The Dipika notes: anugata-
tvam samavetatvam. Vaisesika sutra 8.6, Sdmdnyavisesdpeksam dravyaguna-
karmasu, attributes knowledge of the ontological status of entities to general
properties. VS, p. 63. 1.1.7, Sad anityam dravyavat kdryam kdranam sdmdnya-
visesavad iti dravyagunakarmandm avisesah, includes a general property as one
of the entities which belong to all three classes of entities: substances, qualities
and actions. VS, p. 3.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
NYAYA 155
relation by which a general property is present in each object of a

kind is known as inherence (samavdya).'8
Because Vatsyayana has defined the individual as a
complex entity consisting of earth, water, fire, and air, his
discussion concerning the object denoted by a word concerns
only those generic properties which inhere in these substances.
He considers only those generic properties which have the same
substratum as some form or other. He does not mention generic
properties in qualities or actions.
Although every generic property generates the same
cognition in different individual objects, the occurrence of the
same cognition in different individuals is not always due to a
generic property (2.2.69 U, 2.2.64 U5-U6). Even without a
generic property one may have the same cognition with respect to
different objects. For example, one has the same cognition with
respect to every cook that he is a cook, a cognition which the
word 'cook' conveys. Yet, as Uddyotakara says, it is not a generic
property which is responsible for the same cognition in this case.
A cook is the agent or principal participant in the act of
cooking. The action of cooking and the relation of agency inhere
together in every cook. The relation of agency involves being
the primary participant in an adtion as opposed to the direct object
(the food), the substratum (the pan), etc. This being primary in
the action is part of what one knows in the cognition of a cook.
Hence, in the case of cognition of a cook, two entities, the action
of cooking and the property of being principal, present together,
are responsible for the recurrent cognition with respect to many
individuals. Because there are two entities, not one, which
generate the recurrent cognition, the cause of the general
cognition is not a generic property (2.2.64 U6).
The Buddhist objector in Uddyotakara's commentary on
2.2.64 finds fault with the conception of a generic property as
defined by Nyaya. He raises the following objection: Nyaya
does not require that every recurrent cognition have a generic
property as its basis. Uddyotakara gives two examples: 1) The
cognition that x is a generic property is common to all generic
properties but is not caused by another generic property inherent
in them all. Inherence in many substrata, which is common to
them all, is the basis of the same cognition with respect to them
(2.2.64 U5). 2) As has just been discussed, the cognition that x is
a cook is common to all cooks but is not caused by a generic
property (2.2.64 U6). Just as Nyaya accepts a cause for the same
cognition with respect to many objects which is not a generic
18Vaisesika sutra 7.2.29. Iheti yatah kdryakdranayoh sa samavdyah.

Candrananda comments, " A generic property inheres in an individual, specific
properties in eternal substances (jdter vyaktau visesdndm ca nityadravyesu
samavdyah)."

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
property in these cases, something other than a generic propert

may be the cause of the same cognition with respect to man
objects in other cases as well.
Gautama 2.2.69 defines a generic property as the cause of
a recurrent cognition. Vatsyayana comments with respect to th
ultimate general property that it is "that object which causes th
same cognition to recur with respect to many objects (yo 'rtho
'nekatra pratyaydnuvrttinimittam)." Gautama 2.2.68 define
form as an inferential mark of a generic property. In whateve
individual substance one sees a certain form, one infers the
presence of a certain generic property, unless a contradictory
cognition arises to block such an inference. In a substance such
as clay, color or something else serves as an inferential mark by
which to infer the generic property. The objector to the
conception of the generic property argues that whatever serves as
the inferential mark of the generic property may itself account for
the recurrent cognition with respect to many objects. Hence one
can always account for the recurrent cognition without accepting
that there is a generic property besides the inferential mark.
In Uddyotakara's discussion, the objector to the notion of
the generic property asserts that one can account for the recurrent
cognition that x is a cow by the fact that the form is the same
(2.2.64 U4). By this he means that the arrangement of the head,
feet, etc. is the same in all cows (the head up front, four feet
below, hump over the shoulders, dewlap below the neck). In
response, the proponent of the notion of a generic property offers
two possible conceptions of a form: either 1) it inheres in just one
body (2.2.64 U4a) or 2) it inheres in many (2.2.64 U4b). 1) If it
inheres in just one then something else which inheres in all of
those bodies is required to be the basis of the same cognition with
respect to them (2.2.64 U4c). That something else is the generic
property. 2) If the form inheres in many bodies, then it differs
from the generic property only in name. He says, "Argument
over a name (ndmni vivddo) is pointless (2.2.64 U4b)." If it
differs from a generic property only in name, then it does not
differ in fact. For all practical purposes it is a generic property.
Hence there are two possible meanings of the word 'form (dkrti)'
in this discussion: 1) a configuration of parts as belonging to a
single individual substance, and 2) a configuration of parts as a
common property belonging to all of the many individual
substances of a kind, in which case the form itself is a generic
property.
This dichotomy is complicated by another dichotomy.
The form may be considered either as a rough figure or as an
extremely detailed and precise arrangement of every particle of
an individual substance. Although neither Vatsyayana nor
Uddyotakara dwell on this distinction, Kumarila brings it up in
defending Sabara's example of the use of a generic term for a

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
NYAYA 157
brick falcon altar and the role this example plays in his argument
for the view that a generic term denotes just a class property
(Sabara IIIA note 12). A falcon altar has the shape of a falcon
but is not a real individual falcon. Similarly, the model cow used
as an example by Gautama has the shape of a cow, but it is not a
real individual cow nor does it have the generic property
cowness. In both Sabara's and Gautama's arguments, a model
plays a decisive role; but it plays a different role in each. Hence it
is important to clarify the conception of a shape, its relation to a
generic property and the role it plays in each of their arguments.
This is especially important given the different use of the term
'dkrti' by these authors and the difference of opinion concerning
its translation in modern scholarship.
Whether the form is considered to be a rough figure or a
detailed arrangement, it may be considered to belong just to a
single individual or it may be considered common to many
individuals. Hence there are four distinct possible conceptions of
a form (dk.rti): 1) It has all the precision of the tendons, sinews,
blood vessels, etc., and is common to many individual substances.
2) It is a rough figure with a hump, head, etc., and common to
many individual substances. 3) It has all the precision of the
sinews, etc., but is not common to many individual substances. 4)
It is a rough figure with a hump, head, etc. as a specific
conjunction, a quality inherent in a single individual substance.
I have already mentioned that in Nyaya the form is
considered to be a specific conjunction which is a quality and
inheres in just one substance. Hence, it does not mean form in the
sense of 1 or 2. In addition, the form is not considered to be an
extremely detailed and precise arrangement including all the
precision of the sinews, etc. I have just mentioned Uddyotakara's
discussion in which the objector to the notion of the generic
property asserts that one can account for the recurrent cognition
that x is a cow by the fact that the form is the same. The
proponent of sutra 2.2.64 elaborates what the objector means by
this as follows: the arrangement of the head, feet, etc. is the same
in all cows (the head up front, four feet below, hump over the
shoulders, dewlap below the neck) (2.2.64 U4). Commenting on
the definition of form, Uddyotakara explains that the arrangement
of the head, hooves, etc. indicates the generic property of a cow
(2.2.68 U1), and Vacaspatimisra explains that the arrangement of
the head, hands, etc. indicates the generic property of a human
being, namely, humanness (see note 14 above).
Furthermore, Nyaya definitely considers the form to be
present in the clay model. Because the model does not have
sinews, etc. but has the dewlap, hump, head, tail, and hooves,
Nyaya must mean the form to be the arrangement of these parts.
In Uddyotakara's commentary, when asked why the word 'cow' is
(sometimes) used for a clay model cow, the proponent of suitra

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
2.2.64 answers, "Because the form is the same, just as (the form
shown) in a picture is (the same as the form of the object i
depicts) (2.2.64 U2b)." Here too the form is conceived as an
arrangement of parts such as the hump, etc. (2.2.64 U2c). Th
form of the clay model cow is the same as the form of real cow
Finally, sutra 2.2.64, as well Vatsyayana and Uddyotakara's
explanations of it, asserts that the form of a cow is present in a
clay model cow while the generic property cowness is not. Th
absence of the generic property accounts for not carrying ou
actions generally prescribed for cows on a clay model cow
(2.2.64). If Gautama had considered that the form of a cow was
not in the clay model, he could not have said in the sutra, "even
though it is an individual and has a form." By including this
concessive clause Gautama himself shows that the form is present
in the clay model. Therefore, in Nyaya a form (ak.rti) i
considered to be a rough figure showing the arrangement of part
such as the head, hands, etc. It means this rough figure as a
quality inherent in a single individual, the last of the four possibl
meanings offered above.19
In commenting on Gautama 2.2.64 Uddyotakara argues at
length to defend the conception of a generic property as th
single abstract entity which accounts for the same cognition
recurring with respect to different individuals of a kind. Th
Buddhist objector contends that it is unnecessary to conceive that
there is such an entity because one can account for the same
cognition recurring with respect to different individuals without
it. By this he means that the form, which Gautama accepts as the
inferential mark of the generic property (2.2.68), would directly
serve as the basis for a recurrent cognition (2.2.64 U4).
Uddyotakara considers two alternatives as to how the objector
conceives a form when he asserts this. The form is either present
in a single individual or is common to many individuals:
alternatives 4 and 2 of the above list, respectively. Rejecting
alternative 4, Uddyotakara states that if the form occurs in just
one individual it is not held in common (2.2.64 U4a). He implies
that because it is not held in common it cannot be the basis for the
same cognition recurring with respect to different individuals.
The second alternative Uddyotakara considers is that the
objector conceives a form as common to many individuals
(alternative 2). He charges the objector with arguing over a name
if he accepts that a form, as common, is the basis of a recurrent
cognition. His statement is briefly, "If you accept that a form
does occur in other bodies, argument over a name is pointless
(2.2.64 U4b)," but the implication is clear: If the form inheres in
19Halbfass (1992: 257-58) translates a passage in which the Vaigesika

commentator Sridhara discusses the issue of a number of recurrent
characteristics implying the presence of a universal.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
NYAYA 159
many bodies, then it differs from the generic property only in

name.
The example of the clay model cow contradicts

Uddyotakara's statement here. If the statement, "If the form
inheres in many bodies, then it differs from the generic property
only in name," were true, then whatever has the form of a cow,
including a clay model cow, would have the generic property
cowness. Now whatever has the generic property cowness, is a
cow. Hence, whatever has the form of a cow would be a cow. If
this were so, the following statement would be true: x is a cow if
and only if x has the form of a cow. Symbolically, we may state
this as follows: (Ax)(Cx = Fx), where x is any individual
substance, Cx is the statement, "x is a cow," and Fx is the
statement, "x has the form of a cow." This implies the following:
It is not the case that there is an x such that x has the form of a
cow and x is not a cow {(-(Ex)(Fx & -,Cx)}. But the case of the
clay model cow offers a falsifying example. The clay model cow
(xm) has the form of a cow and is not a cow {(Fxm & -Cxm)}.
In the clay model cow, although the form is present, it does not
succeed in implying the presence of the generic property.
That a clay model cow does not have the generic property
cowness, and is not a cow, is made clear by Gautama in suttra
2.2.64 itself, as well as by both Vatsyayana and Uddyotakara. In
answer to the objector's question, "Well, why isn't cowness the
generic property in a clay model cow?" Uddyotakara says,
"Because a clay model cow is not a cow, and cowness is not
present in that which is not a cow, just as it is not present in a
horse (2.2.64 U2a)."
The example of a clay model cow falsifies the statement,
"If the form inheres in many bodies, then it differs from the
generic property only in name." Therefore, it falsifies
Uddyotakara's answer to the objector under the second alternative
interpretation of the nature of a form. The charge that the
objector is arguing over a name is false. The reason that a form
cannot serve, in place of the generic property cowness, as the
basis of the recurrent cognition "x is a cow" with regard to any
individual substance x, is not that a form is not held in common.
Something besides a form is required to make known cowness.
The example of the clay model cow, in serving as a
counter example to the statement: x is a cow if and only if x has
the form of a cow (Ax)(Cx = Fx), leads us to examine the status
of the form of a cow as an inferential mark of the generic
property cowness. The example also serves as a counterexample
to the statement: x is a cow if x has the form of a cow (Ax)(Fx ->
Cx). Yet this is equivalent to the statement of pervasion (vydpti)
in the inference with cowness as the object to be inferred (sadhya)
and the form of a cow as the inferential mark (hetu): In whatever
individual substance the form of a cow is present, the generic

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
property cowness is present (Yatra yatra gor ak.rtih, tatra tatra

gotvam).
It is acceptable to call something an inferential mark even
if there are exceptions to the statement of pervasion. It is still true
that the form of a cow is the inferential mark of cowness.
However, in an exception, another cognition controverts
cognition of cowness. In the case of the clay model cow, t
knowledge that it is a clay model blocks the knowledge that it
cow. When the objector asks, "Well, why is a clay model cow
a cow?," Uddyotakara answers, "It is not a cow just for the rea
that it is a clay model cow (2.2.64 U2a)." One knows that it
clay model because it is clay, that is, the generic property clay
is present in it. One knows clayness is present in it because o
directly perceives the inferential marks of clayness.
A cow has other properties besides the form which tak
together with the form are coextensive with the generic prop
cowness. These also serve to indicate that the generic prop
cowness is in an individual. The statement that a form is an
inferential mark of a generic property does not deny that ot
things besides a form indicate generic properties. Vatsyayana a
Uddyotakara have both stated that there are substances for wh
the form does not indicate the generic property (2.2.68 V2, U
Vacaspatimisra gives examples in which other qualities indi
the generic property (see note 15 above). A specific color
indicates that a certain substance is clay, gold, or silver; the smell
or taste indicates the generic property of clarified butter, sesame
oil, etc. His family descent is an inferential mark of the class of a
brahmana, etc. Hence qualities such as the sound ('moo'), touch
(soft and warm), color, smell, weight, size, etc. in addition to the
form, and knowing that a bull and cow gave it birth, indicate that
a substance has the generic property cowness. Similarly, the
lightness, touch (cold and hard), etc. of a clay model cow indicate
that it is a clay model. Knowledge of the generic property
clayness there serves to block the inference of cowness based on
the form. Hence one concludes that it is not a real cow.
Let us return to the argument of the Buddhist objector that
one can account for the same cognition recurring with respect to
different individuals without conceiving a single abstract entity, a
generic property, as its basis. We have mentioned above that
Nyaya accepts a cause for the same cognition with respect to
many objects which is not a generic property in certain cases.
Uddyotakara mentions two such cases in the argument under
discussion. The first case concerns the cognition which recurs
with respect to each generic property cowness, horseness, etc. that
it is a generic property. Vaisesika suitra 8.1.5 denies that there is a
generic property generic-property-ness at the basis of this
cognition. Uddyotakara defends this suitra using the first person
pronoun as if it belonged to his own school Nyaya in qualifying

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
NYAYA 161
the meaning of the sutra. He says "I do not accept (na maya...)
that the same cognition with respect to (the general properties)
cowness, potness, and horseness is without cause (2.2.64 U5)."
The basis of the cognition is the inherence in more than one
object. Every generic property is recognized as a generic
property because it inheres in more than one object (2.2.64 U5).
The second case concerns the cognition which recurs with
respect to each cook that he is a cook. Uddyotakara himself
clearly states the example, "the words 'cook, etc.' and the
(accompanying) recurrent cognitions occur, yet there is no
general property cookness (2.2.64 U6)." Uddyotakara goes on to
explain how accepting that a generic property is not the cause of a
recurrent cognition in these two cases does not conflict with
accepting that there are generic properties at the basis of other
recurrent cognitions. The reason he accepts generic properties in
other cases is, "Because distinct cognitions are not without basis."
He explains that this "means that the cognition which is distinct
from the cognition of a body arises from a different cause. It
does not mean that all recurrent cognition occurs only because of
a general property."
Uddyotakara's argument stops there. However, he has not
got at the heart of the objection. The objector asserts that just as
something other than a single abstract generic property is at the
basis of recurrent cognition with respect to all generic properties
and with respect to all cooks, something other than a single
abstract cowness is at the basis of recurrent cognition with respect
to all cows. Although it is not true that the form is coextensive
with the generic property cowness, the form together with the
qualities of a certain sound, touch, sight, smell, weight, etc. all
taken together may be. If so, then that set of qualities may be the
cause of the recurrent cognition "x is a cow" with respect to each
and every cow. Then it would be unnecessary to presume the
existence of an additional entity, a generic property. What we call
cowness would in fact be the aggregate of all these qualities.
Naiyayikas deny that a generic property can be reduced to
an aggregate of concrete properties such as a form, color, or
principal relation to an action like cooking. After Uddyotakara,
they, as well as Kumarila, assert that perception of an object
includes cognition of its generic property. They conceive two
stages of perception: The first, unqualified (nirvikalpaka)
perception, is a simple awareness of the object. The second,
qualified (savikalpaka) perception, is cognition of the object
qualified by its generic property and other properties. They
consider that cognition of the generic property must occur even at
the first stage, i.e. in the unqualified perception of the object,
although at this stage it is not grasped as the qualifier of the object

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
nor can it serve to relate the object to other objects or distinguis

it from them.20
According to the Vaisesika philosopher Prasastapada (550
C.E.), the first to describe these stages explicitly, unqualifie
cognition of the object produces unqualified cognition of it
general properties which then serve to produce qualified
cognition of the object.21 He states that the intuition
(alocanamdtra) of a substance is the means of knowledge
(pramdna), perception, in the arising of cognition of the general
properties of the object.22 Although Prasastapada does not use
the term 'knowledge' (pramd) for the cognition of the general
properties (sdmdnyavisesanajnidna) nor the term 'objects of
knowledge' (prameya) for the general properties in such
cognition, his commentator Sridhara does use the former term.23
Furthermore, Prasastapada explicitly asserts that yogis, with their
settled awareness focused on the object, directly see general
properties (and other objects beyond ordinary perceptual
knowledge).24 The claim that we have direct perception of them
supports the Naiyayikas claim that there are such entities as
generic properties.
Buddhists deny that there are such entities as general
properties as part of their denial of the duration of aggregate
objects and of the substantiality of their constituents.25 They
deny that the identification of an object is due to a particular
20Vacaspatimigra ND, p. 307-308. Kumarila, Slokavarttika, Pratyaksasutra 118:

nirvikalpakabodhe 'pi dvydtmakasydpi vastunah grahanam. 119 states the
limitations. 111-13 introduce nirvikalpaka describing the lack of comparison
to other objects. 120 describes savikalpakapratyaksa.
21Vatsyayana mentions intuition followed by representation by the mind but
gives no description of the character of these stages under NS 1.1.4, ND p. 123
sarvatra pratyaksavisaye jnidtur indriyena vyavasdyah, pascan manasd
'nuvyavasdyah. Clear (1990) discusses two-level perception in the Yuktidipika
after an overview of the subject in early Indian epistemology.
22Prasastapada, Padarthadharmasanigraha, p. 472-73: sdmanyavisesa-
jndnotpattau avibhaktam dlocanamdtram pratyaksam pramdnam. Sridharabhatta,
Nyayakandali: ato visesanajinna[e] indriydrthasannikarsamdtram eva pramdnam
ity arthah. Hence the means of knowledge for the cognition of the qualifier is
just the mere relation of the sense with the object. Halbfass (1992: 100) briefly
discusses this passage.
23p. 472: yadd nirvikalpakam sdmdnyavisesajnanam api pramdruipam
arthapratitiriipatvdt tadd tadutpattau avibhaktam dlocanamdtram pratyaksam.
When unqualified cognition of general and specific properties arises in the form
of knowledge (prama) because they appear as objects of cognition, then, with
respect to that arising, simple awareness unattached (to any other cognition) is
(the means of knowledge) perception.
24Prasastapada, Padarthadharmasafigraha, p. 465: asmadviiistandm tu yogindm
yuktdndm yogajadharmdnugrhitena manasd ... sdmdnyavisesesu ... avitatham
svaruipadarsanam utpadyate.
25Dhammapada v. 277: sabbe sankhara anicca, and v. 279: sabbe dhamma
anatta.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
NYAYA 163
property of a special sort (a generic property) rather than to a

temporary collocation of constituents. They consider it false
belief to attribute identity through time to an aggregate of
concrete properties which in fact is constantly changing. Belief in
the permanence of objects results in suffering while realization o
their impermanence brings about freedom from suffering.
As early as Nagasena's dialogue with king Milinda
regarding the identity of persons, Buddhists try to demonstrate
that the only reality to the identity attributed to an aggregate
through time is the term that designates it. Usage of the term in
this case is not based on the existence of a single substance. On
similar grounds Dinnaga argues that the use of a generic term is
not based upon a single general property belonging to the
denoted objects. Instead, he argues that the basis of the use of the
term is the negation of the complement class. Buddhists consider
that knowledge arises from words through inference rather than
through an independent means of knowledge. The negation of
the complement class determines the set of objects subject to the
inference from the word to its denoted object. Further discussion
of Buddhist theories of meaning will lead us away from the texts
under consideration here.26
Now that we have discussed the Naiyayika's views
concerning the nature of a generic property, a form and an
individual, let us consider their view regarding which of them
generic terms denote.
2.3. A Generic Term Denotes All Three

Together.
Gautama, as elaborated by Vatsyayana, concludes that a

generic term like 'cow' denotes an individual, its form and its
generic property together, one being principal and the others
subordinate. A detailed outline of the argument for this
conclusion precedes the translation.
The discussion comprises an argument for each of the
three elements, the individual, form, and generic property, singly
being the object denoted by a word, and an argument against
each element being denoted by itself. The arguments concerning
the individual and the generic property are comparable to those
which Sabara and Patanjali explain. The conclusion is very close
to one which Pataijali prefers and which Sabara rejects. That the
object denoted includes a third element, the form, requires
260n the Buddhist view of meaning known as apohavdda see Siderits, 1985 and
1991: 87-110, Matilal 1986: articles by Matilal, Hayes, Herzberger, Kunjunni
Raja, Tillemans and Bhattacharya.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
detailed investigation because it will shed light on Sabara's

example of the falcon altar. The numerous examples which
Gautama, Vatsyayana, and Uddyotakara give throughout th
section supply us with examples of usages in which it is clea
which element of the meaning is principal and which subordinate
Our discussion of how 2.2.66 applies to these examples not only
will elucidate Vatsyayana's explanation of this sutra but will also
shed light on Patanijali's view of how one element of the denoted
object is principal and the other subordinate.
Gautama argues that the form should be accepted as the
object denoted by the word because one depends on the form to
determine what the individual substance is (2.2.63). We have
shown that the form is not a reliable inferential mark of the
generic property. A clay model cow has the form but is not
cow. However, the failure of an inference is not necessarily due
to the fact that the inferential mark is inadequate. Even if the
inferential mark is adequate the inference will fail if there is a
obstacle (badhaka). In the case of a clay model cow, the
configuration is an inferential mark of the generic property
cowness. However, the color (and numerous other qualities)
indicate that it is clay. The generic property clayness is present
there. The knowledge that it is clay blocks the inference that it is
a cow. Therefore, the generic property cowness is absent even
though the inferential mark, the form of a cow, is present.
In the case of clay, gold, silver, clarified butter, etc., one
depends on qualities other than a form to determine what the
object is. If form is included in the object denoted by a word for
an individual substance whose generic property form identifies,
then other qualities should be included in the object denoted by a
word for a substance whose generic property they identify. Yet
they are not. This consideration led the new school of Nyaya
(navyanydya) to exclude the form from the object denoted by the
word, just as other properties which may indicate the presence of
a certain generic property in an individual are excluded from the
object denoted by the word. The form as well as other qualities
may indicate the presence of a certain generic property in an
individual, but just the individual and the generic property are
denoted. One of the two is principal and the other subordinate in
the manner which Vatsyayana states under sutra 2.2.66.27
Vatsyayana states that the form by itself is not the object
denoted by a word because the word denotes that which is
qualified by the generic property (2.2.63 VB). The substance
which has a certain form is that which is qualified by the generic
property. In this statement, he repeats the statement by which he
rejected that the individual alone is the object denoted by a word.
27See note 30 below.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
NYAYA 165
Uddyotakara states explicitly that Gautama 2.2.61, "Na

tadanavasthdnat (No because one would not determine it)," should
be understood to be present here after 2.2.63 too. In its original
occurrence, 2.2.61 gives the reason that the individual alone is
not the object denoted by a word, namely, one does not
determine the intended object of speech just by knowing that it is
an unspecified individual. Recurring here it states that in fact one
does not determine the intended object of speech just by knowing
the form of an individual substance. The configuration alone
does not succeed in identifying what the individual substance is.
One relies on other properties too. The clay model cow has the
form of a cow, yet one determines that it is not a cow. Its color,
etc. indicate that it is clay. When the word 'cow' is used, one
(usually) understands a real cow, not a model. Gautama's
argument for accepting the generic property as the object denoted
by a word shows that the form of the cow is not coextensive with
the generic property cowness. The form does not by itself
succeed in identifying what an individual substance is; it does not
succeed in identifying the intended object of speech.
Gautama argues that the generic property is the object
denoted by a word because due to its absence in a model the word
does not refer to the model (2.2.64). Concomitant presence and
absence determine what a word denotes. The form and individual
are present in the model; the generic property is not. The word
does not refer to the model. Therefore, there is a concomitant
absence of the word and the generic property. Hence the generic
property is the object denoted by a word. It is not the case that
there is concomitant absence of the form and the word or of the
individual and the word; hence, neither the form nor the
individual is the object denoted by a word. The argument is
entirely negative. It succeeds in establishing that the generic
property is denoted but not that it alone is denoted. Gautama
rejects the argument in the next suitra because the generic property
depends on the individual and form in order to be manifest. One
never finds the generic property cowness without an individual
substance and the form of a cow (2.2.65).
The argument that the generic property alone is denoted is
just the converse of the argument which Sabara presents with
respect to the altar in the shape of a falcon (IIIA). Sabara argues
that there is concomitant presence between the class property and
the generic term. As we will see in the next chapter, the deciding
case in Sabara's argument is the example of the falcon altar. It is
worth noting that Gautama's argument to show that the generic
property is the object denoted implies that the form is not denoted
alone, while Sabara's argument to show that just the class property
is denoted is based on a word referring to a shape. As has been
shown, Nyaya clearly distinguishes the form and generic
property. As we shall see, Sabara does not.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
After explaining the sutra, Uddyotakara shows in his

commentary that Nyaya also recognizes that the word 'cow'
sometimes used for the clay model. The proponent of the suitra
asked why one does use the word 'cow' for a clay model cow. H
answers, "Because the form is the same, just as (the form shown
in a picture is (the same as the form of the object it depict
(2.2.64 U2b)." Complementing Gautama's use of the example
the clay model cow, the fact that the word is used for the mode
implies that the form is denoted even if the generic property
not. The presence of the generic property along with the form
and individual in a real cow ordain the use of the word 'cow' for
it. The clay model is sometimes called a cow, sometimes not.
That is, when the word 'cow' is used, one sometimes understands
the clay model, sometimes not. The usage of the word for the
model is based on the form. The absence of usage is based on the
absence of the generic property.
After stating arguments for and against accepting each of
the individual, generic property, and form singly as the object
denoted by a word, Gautama concludes in 2.2.66 that a word
denotes the three of them together. Vatsyayana explains that
when possible, all three are the object denoted by a word
together. One of the elements is principal; the others subordinate.
He points out while defining form that the form is not included in
the object denoted by words for substances such as clay, gold,
and silver (2.2.68 V2). In these substances, the form does not
indicate the generic property. The form is included in the object
denoted by a word only when it indicates the generic property the
word is used for. For example, the word 'cow' denotes cowness
and an individual substance in which cowness inheres. Wherever
the individual and cowness are present, so is the form of a cow.
Hence, the form of a cow is also the object denoted by the word
'cow'. In contrast, a word referring to a substance such as clay,
gold, or silver, which remains what it is even while taking various
forms, does not denote the form. It denotes the generic property
and the individual substance. In a sentence like, "This ornament
is silver," the generic property is principal and the individual
ornament is subordinate. The word 'silver' there intends to
communicate the being silver of the ornaments. In a sentenc
like, "Put the silver on the table," the individual silver utensils ar
principal and the generic property is subordinate.
Uddyotakara explains and gives examples of what
Vatsyayana means when he says that one of the elements in the
meaning is principal and the others subordinate (2.2.66 U1). The
individual is principal when it is clear that the word refers to
different objects. He gives the first example given in suitra 2.2.60
as an example, "The cow which is standing, ... ; the cow which is
sitting, ...." The injunctions, "Let the cow loose. Tie the cow,"
which are used indicating a specific individual, also serve as

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
NYAYA 167
examples.28 The generic property is principal when it is clear that

the word refers to any of the kind without distinction. For
example, "One should not kick a cow. One should honor cows."
The same word in the same sentence may sometimes mean the
individual as principal, sometimes the generic property. It
depends on the context. The configuration is principal when it is
clear that that is what is meant. For example, in the sentence,
"Make cows consisting of flour," the form is principal because
one cannot make the generic property cowness. In fact the
generic property cowness will be absent from the object
produced; flourness will inhere in the flour model cows. The
word 'flour' makes known the substance, flour. Making means
forming. That which one can form is the form. Hence the
principal element in the meaning is the form. The individual is
subordinate, and the generic property is absent.29 By analogy to
Uddyotakara's example, the examples of the use of the word 'cow'
for a clay model cow and Sabara's example of a brick falcon
would also qualify as use of a word for a meaning in which the
form is principal and the generic property and individual are
subordinate.
Because the form is already one of the three elements in
the object denoted by a word, it does not make sense to speak of
secondary reference to a meaning in which the form is principal
and the generic property and individual subordinate. In the
example which Uddyotakara has given, however, there is no
generic property at all in the meaning. The flour cow is not a
cow. The generic property cowness does not inhere in it, just as
the generic property cowness does not inhere in the clay model
cow. As I have said above, the concessive clause of sutra 2.2.64,
"even though it is an individual and has a form," implies, and
Vatsyayana states, that the generic property is absent from the
clay model cow (2.2.64 VI). On the occasions that the word
'cow' is used for the model, the case is strictly analogous to the
example Uddyotakara gives of a flour cow. Hence the meaning
of the word in this usage is different from the meaning in
ordinary usage. Ordinarily the meaning has three elements; in
this usage there are two. So it is possible that the word refers to
this different meaning by secondary reference.
On the other hand, secondary reference is the use of a
word for that which it does not denote (2.2.62). The form and
individual are two of the three elements in the primary meaning
of the word. Hence, use of the word for them does not constitute
28G&m munfca, gdm badhdneti niyatdm kdiicid vyaktim uddisya prayujyate.

NyiyamaijarT, vol 2, p. 42, lines 1-2.
29Kvacid dkrteh prddhdnyam vyakter angabhdvo jdtir ndsty eva, yathd
pistakamayyo gdvah kriyantdm iti tatsannivesacikirsayd prayoga iti.
Nyiyamafijari, vol. 2, p. 42, lines 2-3.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
"use for that which it does not denote." It is use for that which it
already denotes minus one element. In addition, the use of the
word determines what it denotes (2.2.60 V). If the word is used
for these two elements they are denoted. Lastly, the whole gist of
Gautama's and Vatsyayana's argument has been to show that each
of a generic property, an individual, and a form must be denoted,
yet none of the three elements alone can be. The three are
together, each taking a turn as principal. If the flour cows, clay
model cow, and similar objects do not exemplify what
Vatsyayana intended by the form being principal and the generic
property and individual subordinate, what does? Therefore, it
seems to me that the ancient Naiyayikas intended such examples
to show a less common but primary use of the word for the form
and individual.
The new school of Nyaya considers the use of the word
'cow' for a real cow to denote the individual and generic property.
The use of the word for a model denotes the form and individual.
Hence the word has two word-meaning relations.30
We have discussed how qualities besides the form indicate
generic properties. The color of the model indicates that it is clay,
that the generic property clayness is present in it. The knowledge
that it is a clay model cow blocks the inference based on the form
that it is a real cow. The color, size, weight, sound, etc., of the
real cow taken together with the form indicate that the generic
property cowness is present in it. Hence, the quality which the
model shares with real cows, the form, allows the use of the word
for it. The absence of other qualities characteristic of cows
prevents it.
30Navyds tu samsthandnupasthitdv api gotvddind gavddyanvayabodhdj

jativisistavyaktdv eva saktih. Samsthdne ca prthag eva saktih. Dinakari comm.
on Karikavalimuktavali, p. 378.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Gautama 2.2.58-69 Outline
I. Preliminaries
A. Definitions (2.2.58)
1. Speech sounds ending in an inflectional termination
constitute a word.
2. A word is the basic unit of speech ordinarily used to
express meaning.
B. Whenever one uses a generic term like 'cow', three
objects are present together: the individual, the
configuration of parts of the individual, and the generic
property (2.2.59).
II. An individual
A. Usage shows that the word denotes an individual
(2.2.60).
1. The word 'cow' refers to different objects.
Individuals are many and distinct from one another.
The generic property is one and the same.
a. It refers to the same object as a pronoun
referring to a specific individual (1).
b. It refers to objects of which a plurality is
mentioned:
i. An aggregate (2)
ii. A number greater than one (5).l
c. It refers to objects which belong to different
owners (4).
d. It refers to a succession of objects (10).
2. The word refers to an object which participates in
action. Only a finite mass can participate in action.
The generic property is insubstantial (so is the
form). The individual is a finite mass.
a. It refers to an object given (3).
b. It refers to an object which grows (6).
c. It refers to an object which decays (7).2
3. The word refers to an object which is the
substratum of qualities. Qualities inhere only in
substances. The individual is a substance. The
generic property is not (nor is the form).
IThis example also fits under 3 below because number (samkhyd) is a

(guna).
2In these last two cases the action is possible only on an object which consists
of parts. Growth is the addition of parts. Decay is the subtraction of parts. Only
finite masses consist of parts.
169

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
a. It refers to an object given (IIA2a). The

qualities disjunction and conjunction are
essential in the act of giving (3).
b. It refers to colored objects (8).
c. In compounds it refers to an object which
enjoys happiness, etc., which are qualities (9).
B. One does not use the word 'cow' for the pure individual
devoid of all properties, but rather, for that which is
qualified by the generic property (2.2.61).
C. Even if the individual were not denoted by the word,
one could account for use referring to the individual by
secondary reference (the use of a word for that which it
does not denote when there is some cause). The
following exemplify secondary reference (2.2.62):
1. The word 'staff refers to a brdhmana because he
always carries a staff.
2. The word 'platform' refers to a man because he
stands on it.
3. The word 'mat' refers to the reeds being woven
because their purpose is the mat.
4. The names of deities refer to a king because he
behaves like them.
5. The word for a measuring container refers to the
item measured in that amount.
6. The word 'balance' refers to sandalwood powder
because one usually buys an amount which doesn't
weigh enough to tip the scales.
7. A river name refers to its bank due to its proximity.
8. The word for a color refers to that which has that
color because it is connected with it.
9. The word 'life-breaths' refers to food because food
is their cause.
10. The words 'family' and 'gotra' refer to the head of
the family or extended family because he rules it.
III. The configuration of parts

A. The word denotes the configuration of parts because one
depends on it to determine what the substance is
(2.2.63).
B. The word does not denote the configuration alone;
rather, it denotes the object connected with the generic
property, viz. the substance which has a certain
configuration.
IV. The generic property

A. The word 'cow' denotes the generic property because
one usually does not use it to refer to a model cow
(2.2.64).

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GAUTAMA 2.2.58-69 OUTLINE 171
1. The generic property cowness is absent in a clay

model cow.
2. A clay model cow is an individual and has the
configuration of parts of a cow.
3. The meaning of the word is that object due to the
absence of which the word 'cow' is not used for the
clay model cow.
B. The word does not denote the generic property alone.
The generic property depends on the configuration and
individual in order to be manifest (2.2.65).
V. The individual, form and generic property together are the

object denoted by the word (2.2.66).
A. When the speaker intends to distinguish an object from
others of its kind and one understands a specific
individual, then the individual is principal and the
generic property and configuration are subordinate.
B. When the speaker does not intend to distinguish an
object from others of its kind and one understands
generality, then the generic property is principal and the
individual and configuration are subordinate.
C. One should understand the use of a word intending the
form as principal in a similar manner.
VI. Definitions
A. An individual is a finite substance consisting of parts
which has certain qualities (2.2.67).
B. Form (2.2.68)
1. A form is the configuration of the parts of a
substance which indicates the generic property and
other inferential marks of the generic property.
2. A form is not the object denoted by the word for
objects in which a form is not an inferential mark of
its generic property, for example, clay, gold and
silver.
C. A general property is that which is responsible for the
fact that we know many things as the same (2.2.69).
1. The ultimate general property (existence) produces
just the cognition of sameness.
2. Specific general properties (generic properties)
produce the cognition that some things are the same
as each other and yet different from other things.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Gautama 2.2.58-69 Translation
2.2.58. Te vibhaktyantah padam.1

Those speech sounds which end in an inflectional
termination constitute a word.
Yathddarsanam vikrtd varnd vibhaktyantah padasamjii bha
Vibhaktir dvayl: namikl dkhyatikt ca, brdhmanah pac
uddharanam.
1) (Proponent:) Speech sounds, changed according to
grammatical rules, ending in an inflectional termination, are
termed a word (pada).2 Inflectional terminations are of two
types, nominal and verbal. 'Brdhmanah' and 'pacati' are examples
(of words ending in nominal and verbal terminations
respectively).
Upasarganipdtds tarhi na padasamjnidh. Laksandntaram vacyam

iti.
(Objector:) Then prefixes and particles wouldn't be termed
words. Another definition of a word should be stated.
Sisyate ca khalu: ndmikyd vibhakter avyaydl lopas, tayoh

padasamjnidrtham iti. Padendrthasampratyaya iti prayojanam.
Ndmapadam cddhikrtya parTksd. Gaur iti padam khalv idam
udaharanam. 58.
(Proponent:) As a matter of fact, Panini teaches that zero
substitutes for the nominal ending occurring after an indeclinable
just so that the term 'word (pada)' may apply to prefixes and
particles.3 2) The reason that Gautama states the present sutra is
1The text is from ND, pp. 654-94. Occassionally I have preferred the reading of
Vatsyayana's Bhasya in ND2, pp. 180-88.
21n the previous suttra Vatsyayana has explained that 'change (vikdra)' means
substitution or the use of one sound where another would have been used.
Modification of the sound is impossible because sounds cease to exist shortl
after they arise. In Mimamsa, Sabara similarly asserts that a word consists ju
of speech sounds. "Tasmdd aksarany eva padam." MD, vol. 1, p. 58. However
he considers the sounds to be non-changing, eternal.
3Panini 2.4.82. Avyaydd dpsupah (luk 58). Luk substitutes for (the feminin
affix) dp and (nominal termination) sup after an indeclinable.
Patanijali concludes his commentary on 2.4.82 by quoting the following
conclusion arrived at under 1.4.21 (see 1.2.64 trans. under vdrttika 53c note 81)
Atha vdcdryapravrttir jnidpayati, utpadyante 'vyayebhyah svddaya iti yad aya
avyaydd dpsupa ity avyaydl lukam sdsti. MB-K, vol. 1, p. 322, lines 24-25;
398, lines 11-12. Or the teacher's (Panini's) procedure, namely, that he teach
zero (luk) in place of them after indeclinables, informs us that nominal ending
arise after indeclinables.
Patanijali notes that the reason Panini teaches the terminations and then zero in
173

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
that one knows an object by means of a word.4 The following

investigation concerning the object denoted by a word proceeds
making a nominal word the topic. The word 'cow' is the
example.5
Tadarthe
The following sutras concern the object it denotes:
2.2.59. Vyaktydkrtijatisamnidhav upacdrdt samspayah.
The doubt arises because one uses (a word like 'cow') for
an individual, its form, and its generic property
present together.
Avindbhdvavrttih samnidhih. Avindbhdvena vartamdndsu
vyaktydkrtijdtisu gaur iti prayujyate. Tatra na jniyate:
anyatamah paddrtha uta sarva iti. 59.
Presence together means coextensive occurrence.6 One u
the word 'cow' for an individual, its form, and its generic prop
occurring coextensively. Hence one does not know wheth
certain one of the them is the object denoted by the word or
are.
Sabdasya prayogasdmarthydt paddrthdvadhdranam. Tasm

(Objector:) One determines the object denoted by a
force of the word's use. Therefore,
their place is to utilize the principle that rules apply as if an affix

zero were still present: Pratyayalaksanam dcdryah prarthayamanah
mrsyati. The teacher permits zero in place of the nominal te
(following indeclinables) intending Panini 1.1.62 (see 1.2.64 trans.
note 23) to apply. In the present consideration rules that shoul
indeclinables apply only to a speech form termed word (pada). The
(pada)' applies to speech forms ending in a nominal or verbal termina
Panini 1.4.14. Suptinantam padam. (A speech form) ending in a no
or verbal termination (tin) is termed a word (pada).
Gautama's sutra is equivalent to Panini's definition of a word, and
accepts it.
4Katyayana in his first varttika (?1.1) and Patanijali (1.2.64 trans. under vdrttika
29 note 15) express that the pupose of using words is to convey meaning.
5Although the present discussion concerns the object denoted by a nominal
word (pada), not a nominal base (pratipadika), the entities discussed are in fact
considered to be denoted by the base in later Nyaya discussions which clearly
distinguish the meaning of the base from that of the termination. In the present
discussion, Vatsyayana means only to distinguish nominals from verbs and to
state that the discussion concerns the former rather than the latter. The example
given is 'gauh' which is a generic term.
60ne occurs if and only if the others do. When one uses a generic term, one has
cognition of all three at once, not of one of them without the others. Vigvanitha
gives this interpretation saying that sannidhi is the milana of an individual,
generic property and form. Others interpret sannidhi as the relation of the word
to these three objects. Ruben (1928: 55) translates "Weil [das Wort] im
Zusammenhang mit Individuum, form und genus gebraucht wird," and
Chattopadhyaya and Gangopadhyaya (p. 147) translate, "because it [term] is
found to be in use as 'invariably related' (sannidhi) to [i.e. as conveying] the
individual (vyakti), the shape (akrti) and the class-essence (jati)."

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GAUTAMA 2.2.58-69 TRANSLATION 175
2.2.60. Ydsabda-samiuha-tydga-parigraha-samkhyd-
vrddhy-apacaya- varna-samdsa-anubandhdndm vy a kt v
upacdrdd, vyaktih.
Because the relative pronoun, a group, giving,
acquisition, number, growth, decay, color,
compounds, and succession pertain to the individual,
an individual (is the object denoted by a word).
Vyaktih paddrthah.
An individual is the object denoted by a word.
Kasmdt?
(Proponent:) Why?
Ydsabdaprabhrtindm vyaktdv upacarat. Upacdrah prayogah. 1)

Yd gaus tisthati yd gaur nisanneti nedam vdkyam jater
abhidhayakam, abhedat. Bheddt tu dravyabhidhayakam. 2)
Gavdm samuha iti bheddd dravyabhidhanam; na jdter, abheddt.
3) Vaidydya gdm dadatiti dravyasya tyago; na jdter, amuirtatvdt,
pratikramdnukramanupapattes ca. 4) Parigrahah
svatvendbhisambandhah. Kaundinyasya gaur brdhmanasya g
iti dravydbhidhdne dravyabhedat sambandhabheda ity
upapannam. Abhinna tu jdtir iti. 5) Samkhyd: dasa gavo
vimsatir gava iti bhinnam dravyam samkhydyate; na jdtir,
abhedad iti. 6) Vrddhih: kdranavato dravyasyavayavopacayah,
Avardhata gaur iti. Niravayavd tu jatir iti. 7) Etendpacayo
vydkhydtah. 8) Varna: sukla gauh kapild gaur iti dravyasya
gunayogo, na sdmanyasya. 9) Samdsah: gohitam gosukham iti
dravyasya sukhddiyogo, na jdter iti. 10) Anubandhah
sarupaprajananasamtdnah. Gaur gam janayati iti,
tadutpattidharmatvdd dravye yuktam na jdtau viparyaydd i
Dravyam vyaktir iti hi narthdntaram. 60.
(Objector:) Because the relative pronoun, etc. pertain to th
individual. To pertain to is to apply to. 1) The sentence, "Th
cow which is standing, ... ; the cow which is sitting, ... ," does n
refer to the generic property because a generic property is not
differentiated. But it does refer to an individual substance
because one individual substance differs from another of the same
kind.7 2) The phrase, "a group of cows," refers to individual
7The sentence given is just a couple of relative clauses indicating agents of

different activities. The full sentence would read something like the following:
The cow which is standing gives milk, the cow which is sitting doesn't. The
same generic property, cowness, is present in both cows. The relative pronoun
'which (ya)' points out a particular thing. The attribution of different activities,
standing versus sitting, informs us that the two occurrences of the pronoun refer
to two different objects. The word cow is in apposition with the pronoun in
each of the two clauses so it refers to two different objects too. Hence cowness
cannot be the object denoted by the word 'cow' because it is just one thing in
both. The particular substance or individual standing is different from the one

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
substances because they differ; not to a generic property because

it is the same.8 3) "He gives a cow to the dyurvedic physician."
One gives a substance, not a generic property because it is
immaterial and because (the cow's) disjunction (from the giver)
and conjunction (with the physician) wouldn't make sense.9 4)
Acquisition is the connection of something to oneself as
property. The phrases, "Kaundinya's cow, a brahmana's cow,"
make sense if the word 'cow' refers to a substance because the
relation differs according to the difference of the substance. But
the generic property is undifferentiated.10 5) Number: "Ten
cows, twenty cows." Individual substances, which are distinct
from each other, are enumerated, not the generic property,
because it is the same.11 6) Growth: A substance which has
parts as its material causes accumulates parts. "The cow grew."
But a generic property has no parts.12 7) This explains decay
too. 8) Color: "White cow, brown cow." A quality has a
connection with a substance, not a generic property.13 9)
Compounds: "cow-benefit, cow-pleasure." Pleasure, etc. have
a connection with a substance, not with a generic property.14 10)
sitting, hence the individual substance is the object denoted by the word.
8If the word 'cow' denoted the generic property, it would not make sense to
speak of a group of cows because there is only one generic property, not many.
A group requires more than one. Therefore, the word 'cow' must denote an
individual because it makes sense to speak of a group of individuals.
9A generic property is present in an individual of its type by the relation of
inherence (samavaya). It has no direct relation with any other substance or
action. According to Vaisesika, only substances which are not all-pervasive are
substrata of action, and only substances are substrata of qualities. Action
brings about conjunction and disjunction which are qualities. See Vaisesika
suttra 1.1.5. Ruparasagandhasparidh sankhydh parimdndni prthaktvam
samyogavibhdgau paratvdparatve buddhayah sukhaduhkhe icchddvesau
prayatnas ca gundh. VS, p. 2. 1.1.14. Kriydvad gunavat samavdyikdranam iti
dravyalaksanam. VS, p. 5. Hence for two reasons the word 'cow' means the
substance here and not the generic property: 1) The action of giving inheres in
a substance, not a generic property. 2) The qualities of conjunction and
disjunction inhere in substances, not generic properties.
10Different owners do not own the generic property cowness; they own
individual cows. They have an ownership relation with the individual, not with
the generic property.
11Vaisesika sutra 1.1.5 lists number (satikhyd) among the qualities. See note 9.
12In Vaisesika, parts are the material cause of all conglomerate substances.
13See Vaigesika sutra 1.1.14 in note 9.
14Happiness (sukha), which is listed as a quality in Vaisesika sutra 1.1.5,
resides in a substance. The compounds mentioned are provided for by Panini
2.1.36. Cathurthl tadarthdrthabalihita-sukharaksitaih. A word ending in a
fourth triplet nominal termination is compounded with a word for an original
object which serves the purpose of that which the word ending in the fourth
triplet termination denotes. It is also compounded with the words 'artha
(purpose), bali (tribute, offering), hita (benefit), sukha (happiness) and raksita
(protection)'. The compound so formed is a tatpurusa.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Succession is a series of generations of like things. The

sentence, "A cow gives birth to a cow," makes sense with respect
to substances because a succession of them can occur, not with
respect to a generic property because a succession of a generic
property can't. 'Substance' means nothing other than 'individual'.
....Akrter apy abhidhdyakam etan na bhavati, niskriyatvdd
akrteh. Na hi niskriya gacchati tisthati bhavati ceti....15
60.
....This (the word 'cow' in the phrase, "the cow which is
standing, etc.,") does not denote the form either because
the form is without activity. That which is without
activity does not go, stand or become.
Asya pratisedhah:
(Proponent:) The following sutra refutes this.
22.261. Na tadanavasthandt.
No because one would not determine what the object
is.
Na vyaktih paddrthah.
It is not the case that just the individual is denoted by a word.
Kasmdt?
(Objector:) Why?
Anavasthdndt. Yasabdaprabhrtibhir yo visesyate, sa gosabddrtho,

yd gaus tisthati yd gaur nisanneti, na dravyamdtram avisistam
jatya vindbhidhlyate.
(Proponent:) Because if it were, one would not determine what
the object referred to is.16 That which the relative pronoun, etc.
specify in the examples, "The cow which is standing, ... ; the cow
which is sitting, ... ," etc. is the object denoted by the word 'cow'.
The word 'cow' does not denote a mere unqualified substance
without any generic property.
Kim tarhi?
(Objector:) Then what does it denote?
Jdtivisistam. Tasmdn na vyaktih paddrthah. Evam samuihddisu

drastavyam. 61.
(Proponent:) That qualified by the generic property. Therefore,
the individual alone is not the object denoted by a word. The
case is the same with respect to a group and the rest of the
examples in the last sutra.
15ND, pp. 660-61.

160ne would fail to put the object in the right ontological category.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
....Ndnena gosabdena vyaktimatram suddham ucyate.

Yady ayam vyaktimdtrabhidhayako 'bhavisyat, tena
yasyam kasyamcid vyaktau pratyayo 'bhavisyad iti
sutrarthah.17 61.
....The suitra means: The word 'cow' does not denote just
the pure individual. If it did denote just the individual,
then one would understand any individual whatsoever.
Yadi na vyaktih paddrthah, katham tarhi vyakdv upacdra iti?

(Objector:) If the individual is not denoted by a word, then how
is it that one uses a word for an individual?
Nimittdd atadbdve 'pi tadupacarah. Drsyate khalu

(Proponent:) A word is used for that which it does not denote
when there is some cause. One sees the following:
2.2.62. Sahacarana-sthana-tadarthya- vrtta-mdna-
dhdrana-sdmipya-yoga-sadhana-adhipatyebho
brdhmana-mahnca-kata-raja-saktu-candana-gahgd-
s.taka-anna-purusesv atadbhdve 'pi tad-upacdrah.
A word is used for that which it does not denote
because of accompaniment, place, being for that
purpose, behavior, measure, weight, proximity,
connection, causation and seniority, in (the examples
of) a brdhmana, a platform, a mat, a king, barley
meal, sandalwood powder, the Ganges, cloth, food,
and a man respectively.
Atadbhdve 'pi tadupacdra ity atacchabdasya tena
sabdendbhidhdnam iti. 1) Sahacarandt: Yastikam bhojaya i
yastikdsahacarito brahmano 'bhidhfyata iti. 2) Sthdndt: Mancad
kros'anti iti mancasthdh purusd abhidhiyante. 3) Tddarthyd
katarthesu vlranesu vyuhyamanesu, Katam karoti iti. 4) Vrttd
Yamo rajd Kubero rTjeti tadvad vartata iti. 5) Mdndt: ddhakena
mitdh saktava ddhakasaktava iti. 6) Dhdrandt: tulayd dhrta
candanam tuldcandanam iti. 7) Sdmipydt: Ganigdydm gdv
carantiti deso 'bhidhlyate samnikrstah. 8) Yogdt: krsnena rdge
yuktah sdtakah krsna ity abhidhtyate. 9) Sddhandt: Annam
prdnd iti. 10) Adhipatydt: Ayam purusah kulam ayam gotr
iti tatrdyam sahacarandd yogdd vd jctisabdo vyaktau prayujya
iti. 62.
"A word is used for that which it does not denote"
means: A certain word refers to that which the word does not
denote. Because of 1) accompaniment: The word 'staff
(yastikd)' in the sentence, "Feed the staff," refers to a brdhmana
accompanied by a staff. 2) Presence: The word 'platforms
(manicdh)' in the sentence, "The platforms are shouting," refers to
17ND, p. 662.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
the men located on the platforms.18 3) Being for that

purpose: The word 'mat (kata)' in the sentence, "He makes a
mat," refers to the reeds which are being woven together for the
purpose of a mat. 4) Behavior: The names 'Yama' and 'Kuber
in the descriptions, "The king is Yama, the king is Kubera," refer
to the king because he behaves like Yama or Kubera.19 5
Measure: One refers to the barley meal measured by an a.dhaka
as an a.dhaka of barley meal (.dhakasaktu).20 6) Weight: One
refers to the sandalwood powder (candana) weighed by a balance
as balance-sandalwood (tulacandana).21 7) Proximity: The
word 'Ganges' in the sentence, "The cows roam on the Ganges,"
refers to a place beside the river Ganges. 8) Connection: On
refers to cloth connected with black dye as black. 9) Causation:
The word 'life-breaths (prandh)' in the sentence, "Food is the life-
breaths," refers to the food which is a cause of them. 10)
Rulership: The words 'family (kula)' and 'gotra' in the sentence,
"This man is the family, he is the gotra," refer to the man who is
the head of the family or gotra.22 Here one uses the generic term
for the individual because of accompaniment (1) or connection
(8).23
18The English equivalent is "The bleachers are roaring." Patafijali mentions

this and three other reasons for the use of a word for what it does not denote
with examples under Panini 4.1.48: Caturbhih prakarair atasmin sa ity etad
bhavati tdtsthydt tdddharmyat tatsdmipydt tatsdhacarydd iti. tdtsthydt tdvat
manca hasanti girir dahyate. Tdddharmyat jatinam yantam brahmadatta ity aha.
Brahmadatte ydni kdrydni jatiny api tdni kriyanta ity ato ja.t brahmadatta ity
ucyate. Tatsdmipyat gangdyam ghosah kupe gargakulam. Tatsdhacarydt kuntdn
pravesaya yastlh pravesayeti. MB-K, vol. 2, p. 218, lines 14-19. And a fifth
under Panini 6.1.37: ...Tddarthydt tdcchabdyam bhavisyati. Tad yathd indrarthd
sthunendra iti. MB-K, vol. 3, p. 32, line 12.
19The lord of death and the lord of wealth. Calling the king Kubera is
equivalent to the saying "The king is as rich or magnanimous as Kubera."
20An ddhaka is a measure of volume equivalent to approximately 7 Ibs., 11 ozs.
of grain. In the example, the word properly denotes the measure but is used to
refer to the item measured, just as the word 'cup' in "a cup of flour" refers to the
flour the cup measures.
21The word 'tuld' properly denotes a balance but in this usage refers to the
standard weight of sandal-wood powder.
22Vatsyayana takes the word 'gotra' properly to denote the extended family but
to refer secondarily to the head of the extended family. It is this second
meaning which Panini states in sutras 4.1.162-65. 4.1.162. Apatyam
pautraprabhrti gotram. A descendant beginning with a grandson is (termed)
gotra. 4.1.163. Jivati tu vamsye yuvd. But if a direct male ancestor of his lives,
he is (termed) yuvan. 4.1.164. Bhrdtari ca jydyasi. Or if an elder brother of his
lives (he is termed yuvan). 4.1.165. Vdnyasmin sapinde sthaviratare jivati.
Optionally if another elder kinsman (descended from the same great
grandfather) lives (he is termed yuvan).
23The suitra states that secondary reference occurs when there is a cause for it
and lists ten causes with corresponding examples. Note that the sttra has a
mixed mode of referring to an example of the secondary use of a word. In some
cases the sutra uses a word which has primary and secondary meanings, in other

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Yastikdydm tdvad ayam yastikasabdo jatinimittah.

Yastikdtvam jdtih. Sd yastikdydam varttate. tayd
yastikdtvayuktaya yastikayd brahmanasya yogah.
Sdhacaryat samyuktasamavetdm jatim brdhmane
'dhydropya brdhmanam yastikety dha. Evam sesany
upacdrabfjdni svayam utpreksanlydnTti.24
The word 'staff used for a staff has a generic property as
its semantic condition. That generic property is staffhood.
It is present in the staff. A brdhmana is connected to a
staff which is connected to the generic property staffhood.
Due to accompaniment, one superimposes the generic
property inherent in that to which he is connected, and
hence one calls the brdhmana a staff. The reader should
understand the rest of the reasons for secondary usage in a
similar manner himself.
Yadi gaur ity asya padasya na vyaktir arthah, astu tarhi

(Objector:) If the object denoted by this word 'cow' is not an
individual, then let it be
22.263. Akrtis tadaptksatvdt sattvavyavasthana-
siddheh.
A form is the object denoted because one depends on
that to succeed in determining what the substance is.
A krtih padarthah.
A) A form is the object denoted by the word.
Kasmdt?
(Proponent:) Why?
Tadapeksatvdt sattvavyavasthdnasiddheh. Sattvdvayavdnadm tad-

avayavanam ca niyato vyuiha dk.rtih. Tasydm grhyamadnayadm
sattvavyavasthdnam sidhyati: ayam gaur ayam asva iti,
ndgrhyamnadydm. Yasya grahanat sattvavyavasthdnam sidhyati,
tam sabdo 'bhidhdtum arhati, so 'sydrtha iti. 63.
(Objector:) Because one depends on that to succeed in
determining what the substance is.25 Form is the fixed
cases it refers directly to the secondary meaning by using a word which has that
meaning as its primary sense. The first group includes examples 2, 3, and 7; the
latter group includes examples 1, 4, 5, 6, 8, 9, 10. Ganganatha Jha (1912-19:
1020 note 1) not recognizing this, erroneously asserts that all but the first are
terms "that are figuratively applied to things other than those directly denoted
by them." Although he interprets examples 5 and 6 so that the word for the
measured item (prameya) applies to the the unit of measure (pramana) (the
reverse of what I have said); even doing so, only half of the examples have the
characteristic he asserts all but the first have.
24ND, p. 664.
25Only seeing the form can one determine what type of substance it is
(dravyasvaruipanirnaya).

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
configuration of the parts of a substance and of their parts. If on

grasps that, one succeeds in determining specifically what th
substance is, "This is a cow, this is a horse," but one doesn't
succeed if one doesn't grasp that. A word should denote that due
to the comprehension of which one determines specifically what
the substance is. That is the object denoted by the word.
Naitad upapadyate. Yasya jdtyd yogas, tad atra jdtivisistam

abhidhiyate gaur iti. Na cdvayavavyiuhasya jdtyd yogah.
B) (Proponent:) This doesn't make sense. Whatever is connected
with the generic property, that, qualified by the generic property,
is denoted by the word 'cow' here. And the configuration of parts
is not connected with the generic property.
Kasya tarhi?
(Objector:) What is (connected with the generic property)?
Niyatdvayavavyuhasya dravyasya. Tasmdn ndkirtih paddrthah.

(Proponent:) The substance whose parts are in a fixed
configuration. Therefore, it is not the case that the form is the
object denoted by the word.
....Atrdpi tad evopasthitam, Na tadanavasthdndd iti.
Yasya ca jdtyd yogah sa gosabdasya visayah, ndkrtir
jdtyayogdt....26 63.
Here too the (above sutra 2.2.61), "No because one
would not determine what the object is," is
present. That which has a connection with the generic
property is the object of the word 'cow', not the form
because it does not have a connection with the generic
property.
Astu tarhi jdtih paddrthah.

(Objector:) Then let the generic property be the object denote
by the word.
22.264. Vyaktydkrtiyukte 'py aprasaigdt
proksanadinam mrdgavake, jdtih.
Because one does not erroneously wash, etc. a clay
model cow, even though it is an individual and has a
form, the generic property is the object denoted by a
word.
Jdtih paddrthah.
The generic property is the object denoted by a word.
Kasmdt?
(Proponent:) Why?
26ND, pp. 665.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Vyaktydkrtiyukte 'pi mrdgavake proksanadlndm aprasangdd iti.

Gdm proksaya gam anaya gam dehlti naitdni mrdgavake
prayujyante.
(Objector:) Because washing, etc. do not erroneously apply to a
clay model cow, even though it is an individual and has a form.27
The sentences, "Wash the cow, bring a cow, give a cow," are not
used for a clay model cow.
Kasmdt?
(Proponent:) Why?
Jater abhdvdt. Asti hi tatra vyaktih. Asty dkrtih. Yadabhavdt

tatrdsampratyayah, sa paddrtha iti. 64.
1) (Objector:) Because the generic property is absent. 2) There is
an individual present; there is a form. 3) That due to the absence
of which one does not comprehend the object is the object
denoted by the word.
Yasman mrdgavake vyaktyakrtl stah, na ca tatra
proksaniddni prayujyante. Na hi gdm anaya, gam dehfti
mrdgavake praisasampratipatti bhavatah. Yadabhavdt
tatrdsampratyayah, sa gosabdasydrthah.
U1) Because there is an individual substance and a form
in a clay model cow, yet one does not perform washing,
etc. on it. In the sentences, "Bring a cow, Give a cow," the
speaker does not intend, nor does the hearer comprehend,
that the action be performed on the clay model cow. That
due to the absence of which (in the model) one does not
understand the model as the object of action is the object
denoted by the word 'cow'.
Kasyabhdvat tatrasampratyayah?
(Buddhist:) Due to the absence of what does one not
understand the model as the object of action?
Jateh.
(Jativadin:) The generic property.
Atha mrdgavake kasmdd gotvam na jdtih?

U2a) (Buddhist:) Well, why isn't cowness the generic
property in a clay model cow?
Yasmdn mrdgavako gaur na bhavati, yas ca gaur na, na

tatra gotvam vartate, yaths'va iti.
27The word 'gavaka' meaning an imitation of a cow, is formed by Panini 5.3.96.

See 1.2.64 trans. note 7.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
(Jativadin:) Because a clay model cow is not a cow, and

cowness is not present in that which is not a cow, just as it
is not present in a horse.
Atha mrdgavakah kasmad gaur na bhavatfti?

(Buddhist:) Well, why is a clay model cow not a cow?
Ata eva gaur na bhavati, yasmdn mrdgavaka iti. Na

cayam yuktah prasnah: Mrdgavakah kasmcd gaur na
bhavattti.
(Jativadin:) It is not a cow just for the reason that it is a
clay model cow. Nor does your question make sense,
"Why is a clay model cow not a cow?"
Kim kdranam?
(Buddhist:) Why?
Pratipattibddhitatvdt. Bhavdn apy enam artham

pratipadyate: Ndyam gaur iti. Patipattyd ca prasno
bddhyata iti.
(Jativadin:) Because your understanding blocks it. You
also understand this fact that it is not a cow. And
understanding the answer blocks the question.
Atha mrdgavake gosabdah kutah?

b) (Buddhist:) Then why does one use the word 'cow' f
the clay model cow?
Akrtisdmdnydt citrddivad iti.

(Jativadin:) Because the form is the same, just as the f
shown in a picture is the same as the form of the obje
depicts.
Sarvagateti cet,
(Buddhist:) It is all-pervading.
Atha manyase: Sarvagatd jdtih. Sotpadyamanena yathd

gavd sambadhyate tatha mrdgavakendplti asti mrdgavake
gotvam. Na, anabhyupagamdt. Kena sarvagatatvam jdter
abhyupagamyate, api tu svavisaye sarvatra varttata iti
sarvagatety ucyate.
(Jativadin:) You think the following: The generic
property is all-pervading. Just as it is connected to a cow
upon coming into being, it is connected to a clay model
cow; hence, there is cowness in a clay model cow. This is
wrong because no one accepts it. No one accepts that the
generic property is all pervading; however, they say it is

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
all-pervading because it is present everywhere in its own

domain.
Kah punar gotvasya svo visayah?

(Buddhist:) But what is cowness's own domain?
Yatra gotvam varttate.

(Jativadin:) Where cowness is present.
Kva punar gotvam varttate?

(Buddhist:) But where is cowness present?
Yatra gotvanimitto 'nuvrttapratyayo bhavati.

(Jativadin:) Where there is the recurrent cognition caused
by cowness.
Kva punar anuvrttapratyayam gotvam karoti?

(Buddhist:) But with respect to what does cowness cause
the recurrent cognition?
Yat tasya sddhanam.

(Jativadin:) With respect to that which is the means of
establishing it.
Kah punar nitye gotve goh sddhandrthah?

(Buddhist:) But if cowness is eternal, how does a cow
serve as a means of establishing it?
Yat tena vyajyate. Na hi kakudddimadarthavyatirekena

gotvasya abhivyaktir asti.
c) (Jgtivadin:) In that it is manifested by a cow. Cowness
is not manifest in the absence of an object which has a
hump, etc.
Na hi pindebhyo 'rthdntaram gotvam, pin.dntardlesv

adarsanad iti Kecit.
U3) (Buddhist:) Some say that cowness is not a separate
object from the bodies themselves because it is not found
between the bodies.
Na vikalpdnupapatteh. Pinidantardlesv adarsandd iti

bruvdnah pinddntardlam paryanuyojyah: Kim idam
pin,ddntardlam, kim akasam aho abhdva uta
dravydntaram? Yady akasam, na tatra gotvam; na hy
akdsam gaur iti. Etendbhdvo dravydntaram ca
vydkhydtam. Visesapratyayanam andkasmikatvcc ca.
Ayam ca pin.dapratyayavyatirekabhdk pratyaya
upajayamano nimittdntardd bhavati. Dr.sta khalu pinda-

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
pratyayavyatirekabhdjam pratyaydndam nimittdntardd

utpattih. Yathd vastracarmakambalesu n[lapratyaya iti.
(Jativadin:) No a) because the alternatives don't make
sense. He who says, "because it is not found between the
bodies," must be challenged with regard to what is
between the bodies: What is this stuff between the bodies?
Is it space (akas'a), absence or another substance? If it is
space, cowness is not there because space is not a cow.
This explains absence and another substance too. b) And
because distinct cognitions are not without any basis. The
cognition which arises as distinct from the cognition of a
body occurs from a different cause. One does see
cognitions separate from the cognitions of bodies arising
from other causes, for example, the cognition of blue in a
cloth, a skin and a blanket.28
A krtisdmdnydd iti cet,

U4) (Buddhist:) (The same cognition arises with respect
to different individual bodies) because their form is the
same.
Atha manyase: Satyam ayam astu: pindavyatire

anuvrttipratayo bhavati na punar asya jdtir nim
tu yac chirahpdnipdddkrtisdma nyam tasmdd a
pratyaya iti. Na, pindavad asJdhdranatvdd
Yathaiva pinddh pin.dntaravrttayo na bhavanti,
api ekapindavarttitvdt pinddntaravarttint n
A krteh pinddntaravarttitvam abhyupagamy
vviddo 'ndrtha iti. Atha pindavyatiriktam sd
ndbhyupagamyate, tathdpy ekasdmanyanird
samdnapratyayabtjam anyad vaktavyam sydt. N
samanye samdnapratyayam pasyama iti.
(Jativadin:) Suppose you think the following: I
true that something different from the body
cognition of recurrence, but a generic propert
rather, the fact that the form (arrangement) of
hands, and feet is the same causes the cognition of
recurrence. a) This is wrong, because each form is
unique, just as a body is. Just as bodies do not occur in
other bodies, a form does not occur in another body
either, because it occurs in only one body. b) If you
280ne sees the quality blue in different substances. The fact that one has the
same cognition with respect to different substances shows that the cognition of
the quality is different from the cognition of the substances. Different
cognitions have different bases so the quality blue is not identical to the
substance in which it inheres. Just as a quality is not simply identical to the
substance in which it inheres so the generic property is not either.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
accept that a form does occur in other bodies, argument

over a name is pointless. (What you call a form; we call a
generic property.) c) If you do not accept that there is
any general property distinct from the body; nevertheless
because you have denied that there is one general
property, you have to state another cause of the same
cognition. We just do not see the same cognition if there
is no general property.
Asaty api sdmdnye drstah sdmanyaprataya iti cet,

U5) (Buddhist:) One does see the same cognition even if
there is no general property.
Atha manyase: Yathd samdnyesv asati sadmdnydntare

'nuvrttapratyayo drstah, gotvddisu idam sdmdnyam, idam
samanyam iti; tasmdd anaikdntikatvdd, anuvrttipratyayo
'sddhanam arthdntarapratipddana iti. Na, an-
abhyupagamdt. Na mayd gotvaghatatvas'vatvesu
nirnimittah sdmdnyapratyayo 'bhyupagamyate. Tasmdd
anuttaram etat.
(Jativadin:) If you think the following: Because one d

see the recurrent cognition with respect to each o
general properties cowness, etc., that it is a gene
property, even when there is no other general propert
each which causes the same cognition with respec
each); therefore, because it is not consistent, the cogn
of recurrence does not succeed in implying that ther
another object. This is wrong because we do not ac
that. I do not accept that the same cognition with re
to the general properties cowness, potness, and horse
is without cause.
Sutravyaghdta iti cet,

(Buddhist:) You contradict your sutra.
Yadi manyase: Samadnyesv api sdmdnyam astiti nanu

suitravydghdtah, "Sdmdnyaviesessu samanyavis'esdbhdvdt,
tata eva jnanam" iti? Na suitrdrthdparijidndat. Ayam
suitrdrthah: Yathd dravyam iti pratyayo dravyatvavis'esano
dravye bhavati na punah sdmanyavisesesv evam iti. Na
punar ayam sutdrthah: Nirnimitto 'nuvrttipratyaya iti.
(Jativadin:) If you think the following: Accepting that
there is a general property even in general properties,
don't you contradict the suttra, "Because there is no
specific general property in specific general properties,
one knows (that it is a general property) from those very

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
things"?29 This is wrong, because you have not

understood what the sutra means. The suttra means the
following: The cognition "substance," qualified by th
general property substanceness, occurs with respect to a
substance, but it is not so that the cognition "genera
specific property" which occurs with respect to specific
general properties is qualified by another general
property, namely, general-property-ness. But the sutra
does not mean that the cognition of recurrence with
respect to general properties is without any cause.
Kimnimittam iti cet,

(Buddhist:) What is its cause.
Atha manyase: Gotvddisv anuv.rttipratyayasya kirm

nimittam? Yatah samdnyam iti bhavati.
(Jativadin:) If you wonder, "what is the cause of the
cognition of recurrence with respect to the general
properties cowness, etc.?" the cause is that because of
which one knows that it is a general property.
Kutah sdmdnyam iti bhavati?

(Buddhist:) Because of what does one know that it is a
general property?
Anekdrthasamavdydt. Yathd gotvam anekdrthasamavdyi

tathad'vatvddity anekdrthais tatsamavdyasya samdnatvdt,
gotvddisv anuvrttipratyaya iti.
(Jativadin:) Because it inheres in more than one object.
Just as cowness inheres in many objects so do horseness,
etc. Therefore, because each general property is the same
in that it has the inherence relation with many objects,
there is cognition of recurrence with respect to general
properties cowness, etc.
Pdcakddisabdavad iti cet,

U6) (Buddhist:) It is just like the words 'cook', etc.
Atha manyase: Yathd pdcakadisabdd anuvrttapratyaya?S ca

bhavanti, na ca pdcakatvam ndma scdmdnyaviseso 'sti tathd
gavadisv anuvrttapratyaya iti. Na hetvarthdparijiidndt.
Visesapratyayandm andkasmikatvdd ity asya hetoh
pin.dapratyayavyatiriktasya pratyayasya nimittdntardd
utpdda ity ayam arthah, na punah sarvo 'nuvrttipratyayah
sdmdnydd eva bhavatiti. Evam sati pacanakriyadyaim yat
29Vaisesika sutra 8.1.5. VS, p. 62.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
pradhdnam sadhanam tat pacakasabdenocyate. Tac ca

prddhdnyam pdcakdntare 'py asttti na dosah.
(Jativadin:) If you think the following: Just as the words
'cook', etc., and the accompanying recurrent cognitions,
occur, yet there is no general property cookness, similarly
there is recurrent cognition with respect to cows, etc.
(even without a general property cowness). You are
wrong, because you do not understand what the reason
(for recurrent cognition which we stated above) means.
The reason, "Because distinct cognitions are not without
basis," means that the cognition which is distinct from the
cognition of a body arises from a different cause. It does
not mean that all recurrent cognition occurs only because
of a general property. This being the case, the word
'cook' denotes the most important participant in the action
of cooking. That importance occurs in other cooks too.
Hence there is no problem.
....tasmdd upapannd jdtih. Na cdsau mrdgavaka iti.
Yadabhcvdt tatrdsampratyayah sa paddrtha iti. 64.30
(Jativadin:) Therefore, the concept of a generic property
is sound. And the generic property cowness is not present
in a clay model cow. That due to the absence of which,
one does not understand the model as the object of action
is the object denoted by the word 'cow'.
2.2.65. Na, dkrtivyaktyapeksatvdj jdtyabhivyakteh.

(Proponent:) No, because the manifestation of the
generic property depends on the form and individual.
Jdter abhivyaktir dk.rtivyaktl apeksate. Ndgrhyamdndydm dkrtau
vyaktau, jdtimdtram suddham grhyate. Tasmdn na jdtih paddrtha
iti. 65.
The manifestation of the generic property depends on the form
and individual. One does not grasp just the pure generic
property, unless one grasps the form and individual. Therefore it
is not the case that just a generic property is the object denoted by
a word.
...A krtivisesanatvam vyakteh. Yasmdd vyaktyd
visesyamand akrtir jdter lingam bhavati, tasmdt
pradhanam akrtih....31 65.
....The individual qualifies the form. Because the form
qualified by the individual is an inferential mark of the
generic property, therefore the form is principal....
30ND, pp. 666-68, 669.

31ND, pp. 670.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Na vai paddrthena na bhavitum sakyam. Kah khalv iddnlm

paddrtha iti?
(Objector:) Well, it can't be the case that there is no object
denoted by a word. So then what object is denoted by a word?
2.2.66. Vyaktyakrtijatayas tu paddrthah.

(Proponent:) But the individual, form and generic
property (together) are the object denoted by a word.
Tusabdo visesandrthah.
The word 'but' is for the purpose of a qualification.
Kim visisyate?
(Objector:) What is the qualification?
Pradhdndngabhdvasydniyamena paddrthatvam iti. A) Yadd hi

bhedavivaksa visesagatis ca, tadd vyaktih pradhdnam, angam tu
jatydkrtl. B) Yadd tu bhedo 'vivaksita samanyagatis, tadd jatih
pradhdnam, antgam tu vyaktyJkrtl. Tad etad bahulam prayogesu.
C) Akrtes tu pradhdnabhdva utpreksitavyah. 66.
(Proponent:) The object denoted by a word is the aggregate in
which any one of the three is principal and the other two
subordinate. A) When a speaker intends an object as
distinguished from others of its kind (bhedavivaksd) and the
hearer understands a specific object (visesagati), then the
individual is principal, and the generic property and form are
subordinate. B) When a speaker does not intend an object as
distinguished from others of its kind (bhedo 'vivaksitah) and the
hearer understands generality (sdmdnyagati), then the generic
property is principal and the individual and form are subordinate.
Which is principal and which subordinate varies according to
each usage. C) One should understand the form as principal in a
similar manner.
Pradhdnopasarjanabhdvasydniyamena paddrthatvam:
Yadd bhedavivaksd visesagatis ca, gaus tisthati gaur
nisanneti, tadd vyaktih paddrthah, angam jatydkrtf; jatav
adkrtau ca sthdnagamanddlTnm abhdvdd yatra sambhavah
sa paddrtha iti. Yadd punar bhedo na vivaksitah
sdmdnydvagatis ca, tadd jdtih paddrthah, yathd gaur na
padd sprastavyeti. Akrtes ca pradhdnabhdva
utpreksitavya iti svayam utpreksya ity ayam arthah.
U1) The object denoted by the word is the aggregate in
which any one of the three is principal and the other two
subordinate. a) When a speaker intends an object as
distinguished from others of its kind (bhedavivaksd) and
the hearer understands a specific object (visesagati), for
example, "The cow is standing, the cow is sitting," then
the individual is principal, and the generic property and
form are subordinate. Because standing and going, etc.,

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
do not subsist in a generic property or a form, that in

which they are possible is the principal object denoted by
the word.32 b) But when a speaker does not intend an
object as distinguished from others of its kind (bhedo
'vivaksitah) and the hearer understands generality
(sdmdnyagati), then the generic property is the principal
object denoted by the word, for example, "One should not
kick a cow." c) (Vatsyayana's statement) "One should
understand the form as principal in a similar manner"
means one should understand it as similar oneself.
Kva punar dkrteh prddhdnyam drstam ?

(Objector:) But where does one see the form being
principal?
Yathd pistakamayyo gdvah kriyantdm iti. Tad etad

bahulam prayogesu pradhanopasarjanabhdvah svayam
utpreksya iti.33
(Proponent:) An example is "Make cows consisting of
flour." Because it varies according to each usage, one
should understand which one is principal and which
subordinate in a similar manner oneself.
....Na hi kascit tdntriko 'sti yo jatisabdavdcydn bheddn
pratipadyate. Jdter apy avdcakatve vyaghato jatisabda iti.
Na hi yo yasydvdcakah sa tacchabda iti sakyam vaktum.
Na jdtisabdo bheddnam vdcaka iti....34 66.
U2) There is no logician who thinks that a generic term
denotes individuals alone. If it did not denote the generic
property too, it would fail to be a generic term. You
simply cannot call that which does not denote x an x-
word. A generic term does not denote individuals alone.
Katham punar jadyate, nana vyaktydkrtijdtaya iti?

(Objector:) But how does one know that the individual, form,
and generic property are different from each other.
Laksanabheddt. Tatra tdvad:

(Proponent:) Because they have different definitions.
Concerning these,
2.2.67. Vyaktir gunavisesdsrayo murtih.
An individual is the physical body which is the
substratum of specific qualities.
32See example 1 under Gautama 2.2.60.

33ND, p. 671.
34ND, pp. 679.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Vyajyata iti vyaktir indriyagrdhyeti na sarvam dravyam vyakti

Yo gunavisesdndam sparsantdndm gurutvaghanatvadravat
samskdrdndm avydpinah parimanasydsrayo yathdsambhavam,
dravyam miurtir muirchitdvayavatvdd iti. 67.
An individual, literally that which is manifested, is perceptible
the senses; hence not every substance (dravya) is an individ
(vyakti). That substance which is the substratum, as possible, o
the specific qualities ending with touch, as well as weight, densit
fluidity, resilience and finite measure, is a physical body becau
it is composed of parts.35
Gunavis'esagrahanam kimartham? Gunapaddrthena
saingrhftdya dkrter nirdkarandrtham: A krtih
samyogavisesah, sd ca gunagrahanena sagrh-hteti
tannirdkarandrtham gunavisesaagrahanam...36 67.
Why mention the word 'gunavisesa' in the definition of an
individual (vyakti)? In order to exclude a form (dkrti)
which is included in the class of entities qualities. A form
is a specific conjunction, and mentioning qualities
includes it. Hence, mentioning qualities and specific
properties in the definition of an individual serves to
exclude a form from being an individual.
2.2.68. Akrtir jatiliigadkhyd.
A form is that which makes known the generic
property and its indicatory marks.37
Yayd jdtir jdtilinigdni ca prakhydyante, tdm dkrtim vidydt. Sd
ndnyd sattvdncdm tadavayavdndm ca niyatdd vyuihdd iti.
Niyatdvayavavyuhdh khalu sattvdvayavd jdtilingam, sirasd pcden
gdm anuminvanti. Niyate ca sattvdvayavdndm vyuthe sati gotva
35Gautama 1.1.14 states that the qualities smell, taste, color, touch and soun
which are the qualities of the substances earth, water, fire, air and space, are t
objects of the five senses. Gautama 3.1.62-63 state that all five qualities belo
to earth, the last four belong to water, the last three belong to fire, etc. 1.1.1
Gandharasariupasparsasabddh prthivyddigunds tadarthadh. ND, p. 202. 3.1.
Gandharasarupasparsasabddndm sparsaparyantdh prthivydh. 3.1.63.
Aptejovcyunam purvam purvam apohydkasasyottarah. ND, p. 795. Because
they are perishable and are characteristic of certain substances, they are called
specific qualities (vis'esaguna or vaisesikaguna). See Frauwallner (1956: vol. 2,
86-89). Nevertheless, it is possible that the word 'gunavis'esa' here means just
certain qualities. Chattopadhyaya and Gangopadhyaya (p. 152) translate "some
qualities" in the sutra, and just "qualities" in the Bhasya.
Because Vatsyayana mentions "finite measure" he excludes the all-pervading
substances: dkdsa, kala, dik, and dtman. The last clause, "composed of parts,"
excludes manas and the paramdnus of the other four elements. That leaves
substances which are aggregates consisting of earth, air, fire, and water, to be
called vyakti.
36ND, p. 691.
37Ruben, disregarding Vatsyayana's interpretation, takes the sttra to say just
that a form is known as the mark of the generic property. He translates (1928:
57), "Die Form nennt man: Merkmal des Genus."

111.93.:ffff:ffff:ffff:ffff on Thu, 01 Jan 1976 12:34:56 UTC
prakhydyata iti. Andk.rtivyangyaydm jatau, mrt, suvarnam,

rajatam, ity evamddisv dkrtir nivartate, jahdti paddrthatvam. 68.
1) Know that, by means of which one knows the generic property
and its indicatory marks, to be the form. And that is none othe
than the fixed configuration of the parts of a substance and of
their parts. The parts of the substance, which (in turn) have
fixed configuration of their parts, are the indicatory mark of th
generic property. People infer, by its head or by its foot, that an
object is a cow. Cowness is also known if the fixed configuration
of parts is present.38 2) If one does not know the generi
property by means of a form, as in clay, gold, and silver, th
form recedes, gives up being the object denoted by the word.
....Niyatavyiuhdh khalv avayava jdtim anumdpayanti:
sirahpdnyadimaddkdarasdmnyena gotvam vyajyate iti.
Sira.hpdnyddtnam vdvayavavisesandm darsanena.
U1) ....The parts in a fixed arrangement cause one to infer
the generic property: the general form consisting of a
head, hooves, etc. manifests cowness. Or seeing certain
parts, the head, hooves, etc. manifests cowness.
Kaddcid punar iyam dkfrtih paddrthatvam jahdti.
U2) But sometimes the form gives up being the object
denoted by the word.'
Kva punar jahdti?

Where does it give it up?
Yatrdk.rtivyatgyd jctir na bhavati yathd mrtsuvarnarajatam

iti. Akrtau niyamo na jdtau: sarvdk.rtir jatilingam iti na
punah sarvd jdtir akrtya lingyate.39 68.
Where the generic property is not manifested by a form,
for example, clay, gold, and silver. The suitra restricts the
form, not the generic property: every form is an
indicatory mark of a generic property, but not every
generic property is indicated by a form.
2.2.69. Samdnaprasavatmikd jdtih.
A generic property is that whose nature is to produce
the same cognition.
Ya samanam buddhim prasutte bhinnesv adhikaranesu, yaya
bahumnitaretarato na vydvartante, yo 'rtho 'nekatra pratyaydnuvortti-
nimittam, tat sdmdnyam. Yac ca kesadmcid abhedam kutascid
bhedam karoti, tat sdmdnyaviseso jdtir iti. 69.
1) That which produces the same cognition in different substrata,
that by which many things are not differentiated from each other,
38Vaisesika sitra 2.1.8, Visdni kakudmdn prdntevdladhih sdsndvdn iti gotve

drstam lingam, states that in the case of a cow, having certain features is the
inferential mark of the generic property cowness. VS, p. 12.
39ND, p. 692.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
that object which causes the same cognition to recur with respect
to many objects, is the ultimate general property. 2) And that,
which makes some things the same as each other, and yet
differentiates them from other things, is a specific general
property (samanyavisesa), a generic property (jdti).
...Jatau niyamo na samdnapratyayotpattau, jdtim
antarendpi drstatvat: jdtim antarendpi samdnapratyayo
drstah, yatha pacakddisu....40 69.
....The sutra restricts the generic property, not the
occurrence of the same cognition, because one sees the
same cognition even without a generic property: One sees
the same cognition even without a generic property, for
example, in a cook, etc.
40ND, p. 693.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Mimimsa
3.1. Background
3.1.1. Dharma, Injunctions, and Means of

Knowledge
The Mimamsa system of Indian philosophy investigates

dharma. Dharma is that object (artha) conducive to the highest
good (nihsreyasa) which is indicated by a Vedic injunction.2 As
an object indicated by an injunction, dharma is duty, an action to
be performed.3 Because it is yet to be brought about, it is not
present and so cannot be perceived by means of direct perception
(pratyaksa).4 The other means of knowledge, inference
(anumdna), analogy (upamana) and presumption (arthdpatti)
depend on direct perception.5 Speech (sabda) does not
11.1.1 Athato dharmajijiidsd.

21.1.2. Codandlaksano 'rtho dharmah. Sabara introduces the suitra saying,
"(Dharma) unites a man with the highest good (Sa hi nihsreyasena purusam
samyunakti)." MD, part 1, p. 13. Similarly he summarizes the meaning of the
sutra as follows: Tayd (codanayd) yo laksyate so 'rthah purusam nihsrayasena
samyunakti. The object indicated by an injunction unites a man with the
highest good. MD, part 1, p. 15. Further on he states that the Vedic
performances Jyotistoma etc. are such objects (Ko 'rthah? Yo nihs/rayasaya
jyotistomddih). MD, part 1, p. 21. The exact nature of dharma, whether it is the
action itself or an object beyond the range of the senses manifested by the
action, how it is indicated by an injunction, how it brings about the highest
good, what the nature of the highest good is, and other questions are all debated
at length throughout the history of Mimamsa. We just touch on the tip of such
questions here as background to the perspective from which Mimamsa
approaches the question of what a generic term denotes. For a general
introduction to various topics in Mimamsa see Gainganatha Jha (1942) or
Pashupatinath Sastri (1923).
3Sabara: Codaneti kriydydh pravartakam vacanam dhuh. They say an
injunction (codand) is an utterance which urges one to action. MD, part 1, p. 14.
41.1.4. Satsamprayoge purusasyendriydndm buddhijanma tat pratyaksam;
animittam, vidyamdnopalambhanatvdt. Direct perception is the cognition
which arises for a person (purusa) when the senses come into contact with an
existing object. It cannot cause knowledge of dharma because it is a perception
of an existing thing.
5Concluding his commentary on 1.1.4, Sabara says, "Because they presuppose
perception, inference, analogy and presumption do not cause knowledge of
dharma (Pratyaksapurvakatvdc cdnumdnopamdndrthdpatttndm apy
akdranatvam)." But it is not the case that dharma does not exist because verbal
instruction gives knowledge of it. Sabara introduces the next sutra saying,
"Abhdvo 'pi ndsti, yatah..." MD, part 1, p. 27. Kumarila takes non-perception
195

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
necessarily; it can be independent. Hence it is the only means of

knowledge (pramana) of objects beyond the range of direct
perception (pratyaksa), and in particular, of dharma (1.1.5).
Cognition is inherently positive and subject to doubt only
as regards its cause. Although the validity of human speech may
be doubted because of contradictory evidence from other source
or suspicion of the author of the speech; the Veda is not subject t
doubt because, being authorless (apauruseya) and eternal, it is no
caused, and, being the sole authority regarding dharma, it cannot
be contradicted. Hence, the Veda has self-evident authority
(svatahprdmdnya) regarding dharma.
3.1.2. Vedic and Ordinary Words and Meanings
The ability of the Veda to give knowledge of duty rests

on the ability of an ordinary word to convey knowledge of an
object. The Veda makes known duty by enjoining an action to be
performed. If the words in the Veda were unintelligible, they
would be unable to enjoin one to act. One can only act upon the
statements made in the Veda, if one understands its sentences, and
can only carry out its injunctions if the acts it enjoins and the
objects used in those acts are familiar. One can only understand
its injunctive sentences if one understands the words of which
they are composed. Understanding the words depends on the fact
that they are the same as ordinary words and on the fact that their
meanings are the same which one knows ordinary words to have.
Therefore, the words that occur in the Veda and their meanings
are the same as the words that occur in ordinary usage and their
meanings (1.3.30 Lokavedadhikarana).
Once it has been determined that Vedic words and
meanings are the same as ordinary words and meanings,
Mimamsa can raise the question whether Vedic words denote the
generic property or the individual substance. If Vedic words and
meanings were different from ordinary words and meanings the
question would be pointless because there would be no way to
determine the meaning of Vedic words. Because they are the
same as ordinary words and meanings, the same methods used to
determine what ordinary words denote can be applied to Vedic
words. Furthermore, because their meanings are the same, if
ordinary words denote the generic property so will Vedic words,
or if ordinary words denote the individual substance so will Vedic
words (I).
(anupalabdhi) to be a separate means of knowledge regarding something's

absence.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MiMAMSA 197
3.1.3. Sentence Meaning
One comprehends the meaning of a sentence based upon

one's comprehension of the meanings of the individual words.
Sentences do not have an independent relation to a sentence-
meaning as an independent entity.6 One cannot presume such an
independent relation because one can account for comprehension
of the sentence-meaning otherwise.7 Vedic sentences are eternal;
61.1.25. Tadbhitdndam kriyarthena samamndyo 'rthasya

tannimittatvat. A sentence consists of words denoting their
word-meanings together with a word denoting action because
the sentence-meaning is caused by (the word-meanings). Sabara: Tesv
eva paddrthesu bhutandam kriyarthena samuccdranam. Ndnapeksya paddrthdn
pdrthagarthyena vdkyam arthdntaraprasiddham. Words occurring in the same
meanings they denote independently are spoken together with a word denoting
action (to form a sentence). A sentence is not known in another meaning
independent of and not based on the meanings of its words. MD, part 1, p. 115.
7Nanv arthdpattir asti yat paddrthavyatiriktam artham avagacchdmah, na ca
saktim antarena tad avakalpyata iti. (Objector:) Doesn't one presume (that the
sentence has a capacity, independent of its constituent words, to denote an
independent meaning) as follows: We do understand a meaning over and above
the meanings of the words, and we can't imagine (how that meaning can occur)
without (the sentence having) the capacity to denote meaning. Tan na. Arthasya
tannimittatvdt. Bhaved arthdpattir yady asatydm api saktau nanyan nimittam
avakalpyeta. Avagamyate tu nimittam. (Proponent:) That is not so, because the
sentence-meaning is caused by the word-meanings. One could presume that the
sentence had a signifying capacity independent of its constituent words if
without such a signifying capacity no other cause of the sentence-meaning
would be possible. But a cause is understood. Kim? (Objector:) What?
Paddrthdh. Paddni hi svam svam paddrtham abhidhdya nivrttavydpdrani.
Athedjnlm paddrthd avagatdh santo vdkydrtham gamayanti. (Proponent:) The
word-meanings. Each word denotes its own meaning and finishes its work with
that. Then at that stage the word-meanings being understood convey the
sentence-meaning. Katham? (Objector:) How? Yathd hi sukla iti vi krsna iti va
gunah pratito bhavati. Bhavati khalv asdv alam gunavati pratyayam ddhdtum.
Tena gunavati pratyayam icchantah kevalam gunavacanam uccdrayanti.
Sampatsyata esdm yathdsamkalpito 'bhiprdyah. Bhavisyati visistdrtha-
sampratyayah. Viisistrthasampratyayas ca vdkydrthah. Evam ced avagamyate
'nyata eva vdkydrthah, ko jdtucid: adrstd padasamuddyasya saktir arthdd
avagamyata iti vadisyati? (Proponent:) In just the following manner: From the
word 'white' or the word 'black' the quality is cognized. That (quality) is
sufficient by itself to supply the cognition of the qualified (substance) having
that quality. Because that is so, those desiring (to convey) the cognition of the
qualified (substance) speak just the word denoting the quality. Their intended
import will occur. There will be cognition of the object qualified (by the
denoted object). This cognition of the qualified object (visistartha-
sampratyaya) is the sentence-meaning (vikyartha). If in this way the sentence-
meaning is accounted for even without (the sentence having an independent
capacity to denote), who will ever say that one presumes an invisible capacity of
the word-group from the presence of the sentence-meaning? MD, part 1, p. 116.
Sabara's explanation given of how the sentence-meaning arises is called
abhihitdnvaya. Kumarila expands this view. Concerning Prabhakara's
anvitdbhidhdna, the view that words denote meanings already linked, see
Siderits (1991: 32-50, 79-86). Concerning Bhartrhari's view that the sentence is

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
they are not composed. Therefore, if words, their meanings and

the relation of the two are eternal, and the sentences of the Veda
are eternal, so are the Vedic-sentence-meanings. Hence, the Veda
is the sole authority concerning duty owing to the fact that words
their meanings and the word-meaning relations are eternal.
3.2. The Eternal Nature of a Word, its Meaning,

and their Relation
Mimamsa considers words, their meanings, and t

relation between a word and its meaning to be eternal (1.1.5
word's capacity to convey knowledge of an object (sakti) con
in a permanent relation between the word and the known o
Unless both the speaker and the listener already know the
meaning relation, speaking the word will fail to conve
speaker's intention. In order that the relation be eternal, the
must be eternal too.
The view that speech sounds, and their meaningful
combinations in the form of words and Vedic sentences, are
eternal is central to Mimamsa. It is on the basis of the eternal
nature of sound and the authorless (apauruseya) character of
Vedic sentences that the Vedic injunctions are eternal. It is this
eternal character which allows them to express a meaning entirel
independent of the input of an author. This in turn gives them a
status independent of perception and the other means of
knowledge by which they are the sole authorities concernin
dharma. Because of the importance of the eternal nature o
speech sound in Mimamsa, the major arguments for this view in
the Tarkapada as elaborated by Sabara are included in the
following section. Selections from Sabara's commentary appea
in the notes.8
the primary denoter, see Biardeau (1964).

8Consult Frauwallner (1961b) and D'Sa (1980) for further exploration of this
topic. Jacobi (1929:148-49), although supposing that the Mimamsakas tak
the doctrine of the eternal nature of speech from the grammarians, considers
unresolved whether the discussion in Mimamsasutra 1.1.6-23 concerning
whether speech is eternal or not rests on Vyadi's Sangraha or is original. Th
view that speech units, their objects and the relation between them is
established is upheld in the Mahabhasya. In the first varttika, Katyayana states
that the relation between a word and the object it denotes is established (?1.1
Patafijali in commenting on it states that the word and object are also
established, and Katyayana clearly holds this view. The view that they ar
transient is also considered. See ?1.2.3 and the references in note 20.
Vatsyayana considers that the relation between a word and the object it denot
is conventional, and that speech sounds are producible and transient. See ?
and note 7.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MiMAMSA 199
3.2.2. Word
Speech sound (sabda), which in the form of

injunctions is the means to know duty, is distinct fr
(ndda) which is a producible and perishable substance.
must be single and eternal because one uses it to convey
to others. If it were new each time it was spoken, it
convey meaning because a single utterance cannot form
with a meaning. However, if the same word occurs repe
can form a relation which, if known by both the user an
allows the word to convey its meaning.10
One argument for the view that a word is new
different each time it is spoken deserves mention:
appears simultaneously in many places. It is imposs
single thing to be in many places at once. Hence, a word
Unless there is a difference an eternal entity is not cons
be many, but a producible entity is naturally as man
produced. Therefore, because it is many, a word is not e
91.1.17. Nddavrddhipard. When one speaks of increase of

volume one refers to the increase of sound (not speech sound).
Compare this with Sabara's remark on Jaimini 1.1.5 that a word is merely the
speech sounds, nothing more, and his rejection of the sphota doctrine. See the
discussion, "Atha gaur ity atra kah sabdah....Tasmdd aksarany eva padam." MD,
part 1, pp. 54-58. Ganigantha Jha (1933: vol. 1, 19-20).
101.1.18. Nityas tu sydd darsanasya pardrthatvdt. Speech
sound must be eternal because it appears in order to convey
meaning to others. Sabara: Nityah sabdo bhavitum arhati. Kutah
Darsanasya pardrthatvdt. Darsanam uccaranam. Tat pardrtham param art
pratydyayitum. Uccaritamdtre hi vinaste sabde na cdnyo 'nyan art
pratydyayitum saknuydt; ato na pardrtham uccdryeta. Atha na vinastas
bahusa upalabdhatvdd arthdvagama iti yuktam....na caikenoccaranayat
samvyavahdras cdrthasambandhas ca kartum. Speech sound must be eter
Why? Because it appears for the purpose of others. Its appearance mean
utterance. That is for the purpose of others, i.e. in order to convey meanin
others. If it were the case that a speech sound is destroyed as soon as it has
spoken, one would not be able to convey meaning to others. Hence, one w
not utter speech for the sake of others. Now if it is not destroyed, then be
it is perceived repeatedly, the comprehension of a denoted object (is poss
Hence this is correct. ...and by the effort of a single utterance, verbal interc
cannot be carried out nor can a signifying relation with an object be for
MD, part 1, pp. 101-102.
111.1.9. Sattvantare ca yaugapadydt. Because a speech unit i s
found in different substances simultaneously. Sabara: Ndnddesesu
ca yugapac chabdam upalabhdmahe; tad ekasya nityasyanupapannam iti. Asati
visese nityasya ndnekatvam. Kdrydndm tu bahundm ndnddesesu kriyamdndndm
upapadyate 'nekadesasambandhas; tasmdd apy anityah. We perceive a speech
unit in various places simultaneously; that doesn't make sense for one eternal
entity. An eternal entity is not more than one unless each one is different from
the others. However, it does make sense that many products produced in
various places be connected to more than one place (at once). For this reason
too, a speech unit is not eternal. MD, part 1, p. 89. See the Mahabhasya on

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Mimamsa answers this objection as follows: People use

and perceive the same single word in many places at once just as
people see the sun in many places at once. In the morning th
sun appears directly ahead of each of two people standing side by
side facing east. Yet neither sees two suns, one directly in front o
himself and a second directly in front of the man beside him
Hence the sun is one. It appears to be in more than one place
because one does not realize how far away it is. Similarly, one
seems to hear a word produced in various places because one
overlooks the imperceptible air vibrations occurring in the ear
and the fact that the ear does not go to the place of sound
production but rather remains right where it is. Actually, the
place of a word is space which is just one. Hence, a word is
one.12
vdrttika 48, ?1.3.3, and SivasFtra 1, vdrttika 12, Yugapacca deiaprthaktva-

darsandt. MB-K, vol. 1, p. 18, 9-21. For an English translation see Abhyankar
and Shukla (1975: 62-69).
121.1.15. Adityavad yaugapadyam. (One word) is simultaneously
(in many places) just as the sun is. Sabara: Yat tv ekades'asya sato
ndnddesesu yugapad darsanam anupapannam ity ddityam pasya devdndmpriya.
Ekah sannanekadesdvasthita iva laksyate. (Proponent:) Concerning the
objection that it doesn't make sense for one eternal entity to be perceived
simultaneously in various places, consider the sun, blessed fool! Being one it
appears as if present in more than one place. Katham punar avagamyata eka
dditya iti. (Objector:) How do we know that there's one sun? Ucyate:
prdnmukho Devadattah puirvdhne samprati purastdd ddityam pasyati. tasya
daksinato 'vasthito na dvau pasyati, dtmdnas ca samprati sthitam, tirasclnam
devadattasydrjave. Tasmdd eka dditya iti. Duiratvdd asya deso ndvadhdryate.
Ato vydmohah. Evam sabde 'pi vydmohdd anavadhdranam desasya. Yadi
srotram samyogavibhdgadesam dgatya sabdam grh.niydt, tathdpi tdvad
anekadesatd kaddcid avagamyeta. Na ca tat samyogavibhagadesam dgacchati.
Pratyaksd hi karnasaskull taddesd grhyate. Vdyavydah punah samyogavibhdgd
apratyaksasya vdyoh karnasaskulipradese prddurbhavanto nopalabhyanta iti
ndnupapannam. Ata eva vydmoho yan ndnddeiesu sabda iti. Aksdadesas ca
sabda iti. Ekam ca dkdaam. Ato 'pi na ndnddesesu. Api caikaruipye sati
desabhedena kdmam desd eva bhinnd, na tu sabdah. Tasmdd ayam apy adosah.
(The Proponent) answers: Devadatta facing east in the morning sees the sun
directly in front of him. The person standing at his right doesn't see two, one
directly in front of himself, another askance from himself straight in front of
Devadatta. Therefore, there is one sun. Because it is far away we can't determine
the exact position (of the sun). Therefore, we are deluded (that it is in more than
one place). In the same way, we are deluded that a speech unit (is in more than
one place) so can't determine in which place it is. If the ear, going to the place of
contact and disjunction (of the tongue with a place of pronunciation), perceived
a speech sound, then one might be able to understand that a speech sound was
in many places. But it doesn't go to the place of contact. The ear-drum is
obvious right where it is. But the contacts and disjunctions of the
imperceptible air manifesting at the ear-drum are not perceived. Hence, it makes
perfectly good sense (that a speech unit is one yet seems to be in many places).
Also a speech unit is located in space and space is just one. For this reason too
a speech unit is not in various places. Furthermore, (a speech unit) being of a
single form, let the places be different, but the speech unit is not different.
Therefore, this (view that a speech unit seems to be perceived in many places

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MIMAMSA 201
Other reasons for concluding that a word is single and

eternal are the following: 1) A word must be eternal because, in
the case of a generic term, it conveys the cognition of all
individual objects with respect to which it is used at once. Hence
it refers to an element common to all those objects, the class
property.13 However, every object has many class properties.
For example a cow is also a substance, and an entity; hence,
besides the generic property cowness, it has substancehood and
existence. If it were true that a word is new each time it is uttered,
it could not form a relation with the generic property cowness,
nor could one learn that relation. It would be present only once
and then absent. Concomitance could not single out its relatum.
One would be uncertain as to whether the speaker meant cowness,
substancehood or existence. Only if the same word recurs can
concomitance serve to establish the word's relation with a
meaning. Therefore, a word is eternal and not new each ti
uttered.
2) We do not associate a word with a number greater than
one. We count the number of occurrences of a word, while
recognizing the word in each occurrence to be the very same
identical word.14 The fact that we do not perceive the word
simultaneously) is without fault too. MD, part 1, pp. 97-99. See ?1.4.1; the
Mahabhasya on 1.2.64 varttika 40, ?1.3.2; 56, ?1.3.4, and especially on
Sivasutra 1 vdrttika 12. MB-K, vol. 1, p. 18, lines 9-21.
131.1.19. Sarvatra yaugapadydt. Because of simultaneity with
respect to all. Sabara: Gosabda uccarite sarvagavisu yugapat pratyayo
bhavati. Ata dkrtivacano 'yam. Na cdk,rtya sabdasya sambandhah sakyate
kartum. Nirdisya hy dk,rtim kartd sambadhnfydt. Gopinde ca bahuindm dkrtlndm
sadbhdvdc chabdam antarena gosabdavdcydm vibhaktdm dkrtim kena
prakdrenopadeksyati? Nitye tu sati gosabde bahukrtva uccaritah irutapurvas
cdnydsu govyaktisv anvayavyatirekdbhydm dk.rtivacanam avagamayisyati.
Tasmdd api nityah. When somenone says the word 'cow', one cognizes all cows
simultaneously. When the word 'cow' is uttered, cognition arises
simultaneously with respect to all cows. Hence, it denotes a class property
(dkrti). And (if the word 'cow' were new each time it was uttered) one could not
create a (word-meaning) relation between the word and the general property. For
the creator (of the word-meaning relation) would relate (the word to the general
property) by pointing at the general property. But because there are many
general properties in any particular cow, by what means other than the word
itself would he indicate the general property (cowness) singled out (from all the
other properties in the cow) as the one denoted by the word 'cow'? However, if
the word 'cow' were eternal, uttered many times and heard before in connection
with different individual cows, it could be cause to the understanding that it
denotes the general property by means of its concommitant presence and
absence (with cowness). Therefore too, speech sound is eternal. MD, part 1, p.
104-05.
141.1.20. Samkhydbhdvat. Because there is no number. Sabara

Atha matam, anyatve sati sdd.rsyena vydmiudhdh: sa iti vaksyanti, tan na. na hi
te: sadrsa iti pratiyanti kim tarhi: sa evayam iti. The view that each utteranc
constitutes a new and different word but that people say it is the same word
because they are confused due to the similarity (of one utterance to another)

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
between occurrences does not establish that the word itself is

many. We leave the house in the morning and do not doubt
while we do not perceive our family members that they do not
exist. When we return we do not doubt that the mother, father
wife, etc. with whom we meet are different from and mere
similar to those whom we left that morning. Rather, we greet
them certain that they are the very same identical persons we lef
Similarly, we identify the same word in each of its occurrences.
Therefore, a word is eternal and not new each time uttered.
3) Under 1.1.23, Lingadarsandc ca (And because we see
indicative text), Sabara refers to the following verse as scriptural
evidence that speech is eternal:16
wrong. People do not understand that the word of one utterance is similar to the
word of another utterance but that it is that very same word. MD, part 1, p. 106.
15Hyastanasya sabdasya vindisd anyo 'dyatana iti cet, naisa vinastah, yata enam
punar upalabhdmahe. Na hi pratyaksadrstam muhutrtam adrstva punar
upalabhyamdnam pratyabhijdnanto vinastam parikalpayanti. Parikalpayanto
dvitlyasamdarsane mdtari jayaydm pitari vd nasvasyuh. Na hy anupalambha-
mdtrena, ndstity avagamya, nasta ity eva kalpayanti. Apramdnataydm viditdydm
ndstity avagacchdmah. Na hi pramdne pratyakse saty apramdnatd sydt. Astiti
punar avydmohenavagamyamdne na kvacid apy abhdvah. Na cdsiddhe 'bhdve
vydmohah. Na ca siddho 'bhdvah. Tasmdd asati vydmohe ndbhdvah. Tad etad
dnupurvyd siddham. Tasmdt purastdd anuccdritam anupalabhamdna api-na
vinasta ity avagantum arhanti. Tathd g.rhdn nirgatdh sarvag.rhajanam apasyantah
punah pravisyopalabhamdnd api na: prdk pravesid vinasta ity avagacchanti.
Tadvad enam api: ndnya iti vaktum arhanti.... Sa iti pratyaksah pratyayah:
sad.rsa ity dnumdnikah. na ca pratyaksaviruddham anumdnam udeti. Svakaryam
vd sddhayati. Tasmdn nityah. If one objects that, because yesterday's word has
been destroyed, today's is a different one, (I reply that) it has not been destroyed
because we perceive it again. When people recognize a directly perceived object
which is perceived after not being seen for a moment, they do not assume that it
was destroyed. Those assuming so, when they saw their mother, wife or father a
second time, wouldn't be sure that it was their mother, wife or father. It is not
the case that just from not perceiving something they understand that it does
not exist and assume that it has been destroyed. We understand that something
does not exist when there is no means to know it. And it cannot be the case that
there is no means to know it when direct perception is a means of knowledge of
it. On the contrary, when we clearly understand that it exists, there's no chance
that it doesn't exist. And when its lack of existence has not been established, we
are not confused (as to whether it exists or not). And (a speech sound's) lack of
existence has not been established. Therefore, because there is no confusion (as
to whether it exists or not), it doesn't not exist (it definitely exists). That is
proven by the consequence. Therefore, even while not perceiving the word not
being uttered before them, they cannot properly understand that it is destroyed.
Similarly, even leaving the house and not seeing anyone of the household, then
reentering and seeing them, they do not understand that the members of the
household had been destroyed in the meantime. In the same way, people are
able to say that a word is the same and not other (than the one they perceived
before).... The recognition that it is the same word as before is a direct
perception; that it is similar (yet different) is an inference. Inference doesn't
stand up against direct perception or accomplish anything contradictory to it.
Therefore, speech is eternal. MD, part 1, p. 106-7.
16MD part 1, p. 110.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MIMAMSA 203
Tdsmai nundm abhidyave, vaca virupa nityaya; vfsne codasva

sustutim.
O Virupa, send forth eulogy to the radiant showerer (bull, Agni)

eternal speech! (Rgveda 8.75.6)17
3.2.3. Meaning
3.2.3.1. Action
In addition to the word being eternal, the meani

be eternal too in order that the relation between the two be
eternal. One object which words in the Veda denote is dharma
the form of the Vedic performances to be brought about (sadhy
(?3.1.1, note 2). In fact, it is the speciality of Vedic speech that
alone conveys knowledge of duty. Direct perception cann
provide knowledge of duty because perception is limited to wh
exists, whereas duty does not exist at the present time but
produced, brought into existence in the future (sadhya) (?3.1.1
note 4). In contrast, speech can provide knowledge of what is
in the range of direct perception, that is, an action to be broug
about in the future (sadhya), duty (?3.1.1, 1.1.5). However,
order for speech to convey meaning, the object denoted canno
be something to be brought about (sadhya), but rather, must
something established (siddha). Just as speech must be eternal
order to be able to establish a relation with its object and so b
able to denote it (?3.2.2, note 10), the object denoted must
eternal too in order to establish a relation with the word and so be
able to be conveyed. Duty is action, namely the Vedic
performances (yajfia). If it were the case that action is new each
time it is brought about, it could not form a lasting relation with a
denoting word, speech would not be able to denote action, so the
Veda would not be able to enjoin duty.
Jaimini considers that there is one action which does not
differ according to different performances of it. The act is an
established entity (siddha) and so can form an abiding connection
with the word denoting it. Vedic injunctions enjoin its
performance which is producible (sadhya). Duty is both an
established entity (siddha) and producible (sadhya). It exists
eternally in unmanifest form which is not subject to direct
perception and has a relation with a word. Its performance
manifests it. Although the Veda denotes the eternal action which
cannot be brought about because it is already established, since it
enjoins its performance, one determines that it means one to act
17RVS, vol. 3, p. 867. Concerning the word 'nitya', see Hara (1959).

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
upon what one can act upon, i.e. the performance. Jaimini shows
this to be his view stating his conclusion concerning the
designation of names for Vedic performances in the following
suttra:
Jaimini 6.3.12. Api vavyatirekad rupasabddvibhagdc ca gotvavad

aikakarmyam sydn namadheyam ca sattvavat.
On the other hand, because one performance of an action is no
different from another, and because there is no difference in the word
used for it, action is one, just as cowness is, and the name of the
action refers to an individual occurrence of the action just as a generic
term refers to a substance. 18
A name of a Vedic rite denotes the single eternal action but refers
to an individual performance of it just as a generic term denotes a
generic property yet refers to an individual.
3.2.3.2. A Generic Term Must Denote an

Eternal Object
In the case of generic terms too, the word

relation will be eternal only if the object denot
Hence one must consider the nature of a class prope
individual in order to determine which of the two is eternal and
hence able to be denoted by a generic term.
18MD, vol. 5, p. 250. The sutra rejects an initial view expessed in the previous
sitra. Jaimini 6.3.11. Kriydndm asritatvdd dravydntare vibhdgah sydt. Because
actions inhere in substances, they are different when different substances are
used (in different instances). Kumarila discusses these suitras in the
Akrtyadhikarana. Karmandm api ydgatvahomatvddivibhdgatah; aparydyasmrtair
uktam dhdtubhih pravibhajyate. Punarvidhdnasamkhydkhydgunaprakarand-
'ntaraih; anye tu bhedam dcdrydh karmandm eva manvate. Api vdvyatirekdt sydd
desdder ekarupatd; rupaiabddvibhdgdc ca vaksyaty etad dhi Jaiminih. The
action denoted by non-synonymous verbal roots is divisible according to the
division of the properties inherent in it, being a Vedic performance, being an
offering, etc. Other teachers considered the action itself to differ according to
the chapter discussing the Vedic procedure, subordinate acts, the words for it,
number and repeated provision (6.3.11). On the other hand Jaimini himself will
say that it is the same act because no more or fewer (participants) such as place
are involved in the act and because there is no difference in the word used for it
(6.3.12). MD, part 2, p. 236-37.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MIMAMSA 205
3.3. The Nature of an Individual and of a Class

Property
3.3.1. General remarks
At the beginning of the Akrtyadhikarana Sabara

what he means by a class property and an individual. He
A class property is any general property of substances, quali

actions. An individual is that in which there are exclusive
distinguishing properties (IA).
Kumarila says that the term 'class property (dkrti)' mean

the same thing as the terms 'generic property (jati)' and 'gener
property (sdmdnya)'. It is that which identifies or characterizes
individual as what it is and is the basis of the single gene
cognition which recurs with respect to distinct individua
objects.19 Two distinct natures, sameness (sdmdnya) and
difference (visesa) are required to account for two types of
cognition in all things, one which identifies an object with othe
(anugama) and one which differentiates an object from oth
(vydvrtti).20 The same property which is the basis for discernin
that some things are the same is the basis for discerning that ot
things are different. With the exception of universal existence
(satta) (and ultimate special properties according to Vaisesika),
properties have both natures.
General properties (sdmdnya) range in breadth from
existence (satta) or objecthood (vastutva) which belong to al
entities, to more specific general properties of substances, qualit
and actions.21 Examples of specific general properties o
substances include being a pot (ghatatva) and being a cow
(gotva). Properties of qualities include being a quality (gunata),
being a color (rupata), and whiteness (Suklata), and of actions,
being an action (karmatva) and occurrence (janmata). The
general property of a quality such as whiteness inheres in each
19Akrtivdda 3. Jdtim evdkrtim prdhur vyaktir dkriyate yayd; sdmdnyam tac ca

pinddndm ekabuddhinibandhanam. It is the generic property (jdti) alone that
one calls the class property (using the term dkrti derived etymologically from a
+ V/kr, to mean that) by which an individual (vyakti) is drawn to (knowledge);
and that (generic property) is the general property of things which is the basis
of the same cognition with respect to (distinct) objects. SV, p. 385.
20Akrtivdda 5. Sarvavastusu buddhis ca vydvrttyanugamatmika; jayate
dvydtmakatvena vind sa ca na siddhyati. With respect to all objects, cognition
has the nature of differentiating (an object from others) and identifying (it with
others). Such cognition would not succeed unless objects had two natures. SV,
p. 386.
21See Kumarila on 1.3.30. MD, part 2, p. 236. The broadest general property is
vastutva according to Prabhakara; sattd according to Kumarila.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
individual instance of the quality white.22 The individual quality

instance, in turn, inheres in the white substance. As Jaimini
6.3.12 implies, each verbal root, except synonymous ones,
denotes a single action, instances of which occur at various times
and places (?3.2.3.1 and note 18).
3.3.2. Individual
In his Tantravarttika on the Akrtyadhikarana, K

states that Sabara means an individual (vyakti) to be the
substratum (asraya) of exclusive specific properties (visesa), not
the exclusive specific properties themselves.23 He rejects the view
that an individual is identical to a specific property, because a
specific property belongs to many individuals, even to individuals
of different classes.24 It is not true, he argues, that a unique
specific property specifies each individual. A set of properties,
each of which taken singly is common to a number of
individuals, may isolate a unique individual. One different
property in the set makes the set unique. Hence, it makes sense to
call those properties exclusive. The word 'exclusive-specific-
properties' is not a karmadhdraya compound meaning exclusive
specific properties, but rather, a bahuvrihi compound meaning
that in which specific general properties (taken as a group) are
unique. He cites Sabara stating that the individual is the
substratum of general and specific properties.25
22Tantravarttika on 1.3.30. Tatah suklddi tadvyaktigunesu pratitisthati. Next

(the general qualities) white, etc. are present in their individual quality
instances. MD, part 2, p. 236.
23Tasmdd: asddhdrana visesd yasyam sd vyaktir ity evam vydkhyeyam.
(Proponent:) Therefore, (Sabara's definition of a vyakti) should be explained
like this: The individual is that in which there are exclusive specific properties.
MD, part 2, p. 238.
24Asadhdranavisesd vyaktir iti. kecid dhur: asddhdrand viiesa eva
vyaktir visesavyatiriktavyaktyabhdvad iti. Some say that the individual is the
exclusive specific properties themselves because there is no individual separate
from its specific properties. MD, part 2, pp. 237. Te cdnyatrdpi drsyante tatha
jatyantaresv api. (Proponent:) The so called specific properties are seen in
other individuals and even in (individuals of) different classes. MD, part 2, pp.
238. The question whether an individual object is identical to its properties or
separate from them is raised by Patanijali in commenting on 5.1.119,
"anyacchabdddibhyo dravyam dhosvid ananyat?" See the discussion of
Patafijali's term 'gunasamdrdvo dravyam' in ?1.2.1, and the references in note
12. Vyasa's term, "Sdmdnyavisesasumuddyo 'tra dravyam," comes closest to the
formulation of the objector here.
25Ucyate: Naiva visesanam pratyekam asddhdranatvam diritya vyaktivisesana-
tvenopdddnam; katham tarhi, sddhdranarupanam apy ekadvitrddibhedena kvacid
upalabdhdndm yad ekatra pindik.rtndam grahanam, tadapeksam asddhdrana-
viiesatvdbhidhdnam. (Proponent:) We answer: Specific properties are not
taken to qualify individuals relying on the fact that each specific property is

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MIMAMSA 207
Vaisesika holds the view that a substance is different in

nature from a quality or action which inheres in it by the relation
of inherence (samavaya) and of which it is the substratum.
Similarly, substances, qualities, and actions are different in nature
from general and specific properties which inhere in them and of
which they are the substrata. A unique specific property inheres
in each individual eternal entity (in each atom and in akasa,
dtman, etc.), making it unique.26 Kumarila rejects this view
because it involves an overabundance of entities. Say there are 3
properties, a, b, and c, and 7 individuals. Kumarila intends that
one need not posit the additional properties d, e, f, and g, in order
to uniquely specify 7 individuals; seven sets of properties can
uniquely specify the seven individuals: 1) a, 2) b, 3) c, 4) ab, 5)
bc, 6) ac, 7) abc.27
Sanikhya holds the view that a substance is nothing other
than the aggregate of its qualities. It does not assert that an
individual substance is identical to each specific quality; rather it
is identical to the set of them.28 Hence, in the above example, the
seven substances are identical to the seven sets of properties a, b,
unique. Rather, although taken singly or a few at a time specific properties,

being found in several individuals, are common; when taken together as a group
they are found in just one individual then specific properties are called unique.
Pravibhaktd hi ye drstat drsiyante samhatdh punah; pinddsddhdranatvena tair
vyaktir upalaksyate. For the same (properties) seen separately are seen again
grouped together. Because the group is unique, they identify an individual.
Na ca-ekasydm vyaktau ye samuditd dr.stds ta eva vyaktyantare 'py
anyundnatiriktd drsyante. Ya evddrstapurvas tasmin samuddye samprati drsyate
sa evdsddhdranatdm dpddayati. Anantabheddsv api vyaktisu ndtyantdpurva-
vis'esanopalaksanopdddnam, kimcid utkarsdpakarsamdtrena ca
sarvatrdsddhdranopalaksanaldbhdt. Parasdmdnydpeksayd ca sarvany evavantara-
sdmdnydni visesavyapadesam prdpnuvanti. Samghdtdvasthdydm ca vyakti-
visesanatvdt tesu visesagabdah. Etac ca vyakter visesandnam anyatvam
uparistdd bhdsyakdro 'pi vaksyaty eva: yo 'rthah sdmanyasya visesanam
cdsrayah sd vyaktir iti. Nor does one see the very same properties which are seen
grouped in one individual without any additions or subtractions in another
individual. The property which one has not seen previously in the group, which
one sees now, causes the uniqueness. We cannot accept that a completely new
qualifier identifies (each individual) even if each individual is unique, because
just by some increase or decrease we obtain a unique identification for each.
Properties are called specific because in comparison to the ultimate universal
(existence, satta), all narrower general properties are specific, and because
bunched together they specify an individual. Later the commentator himself
will confirm outright that the individual is different from the specific
properties, saying, "That object which is the substratum of the general property
and specific properties is the individual (IIIC2b)." Kumarila, MD, part 2, pp.
238.
26See Prasastapada (p. 78): Sd ca (p.rthivl) dvividhd nityd cdnityd ca.

Paramdnulaksand nityd, kdryyalaksand tv anityd.
27If n is the number of individuals, n factorial (n!) is the minimum number of
properties required to specify each of them uniquely.
28See Wezler (1985).

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
and c, and ultimately only the three properties a, b, and c exist.

Kumarila is dissatisfied with this view because it does not
recognize the difference between an individual and a property.
In the Akrtivada section of his Slokavarttika, Kumarila
rejects both extremes. Properties are neither identical to t
individuals in which they are present nor distinct from th
There is a natural relation between them whereby the property
part (ams'a) of the individual which prevents them from being
entirely distinguished. Yet different natures of the individual
property, namely, the fact that the individual is unique but t
the property is common to other individuals, prevents them fr
being entirely identified.29
Given that Kumarila himself accepts a natural bond
between an individual and its properties, his rejection of t
identity of an individual and its properties in his Tantravarttika
the Akrtyadhikarana is overstated. In particular, his rejection
the interpretation of the compound 'exclusive-specific-properti
in Sabara's definition of an individual is unjustified. He say
"Later the commentator himself will confirm outright that t
individual is different from the specific properties, saying, 'Th
object which is the substratum of the general property a
specific properties is the individual. (IIIC2b)' (?3.3.2 note 25
Sabara states the passage Kumarila quotes through the mouth o
an objector.
Immediately preceding this statement of the objector, the
proponent says, "If it (a word) denotes another individual devoid
of all general and specific properties, then that which it denotes is
precisely a general property (IIIC2)." The proponent intends to
say that an individual devoid of all its properties is exactly the
same as any other. In fact, without even the property of being a
distinct entity, it is purely a thing in itself, which is nothing other
than the ultimate general property, existence (sattd). Hence
Sabara recognizes that the line of distinction between properties
and substrata is not absolute.
29Kumarila states one aspect of the relationship as follows: Tasmdt svdbhdviko

bhedo jdtivyaktyoh pratlyate. Therefore, there is understood to be a natural
difference between the general property and the individual. Akrtivada 50b. He
qualifies this in verse 65a. ....amatvam tu na vdryate. ...but (the properties) are
not thereby prevented from still being parts (of the thing). Parthasarathimisra
clearly states the qualification in his commentary on verse 50b-51: Na tdvad
Vaisesikddivaj jdtivyaktyor bhinnayoh samavdyasambhavah, pratlty-
ananugunatvdt. Na hi: iha gotvam iti kasyacit kaddcid avagatih: ayam gaur ity
aparydyabuddhidvayasdmdnddhikaranyadarsanat. But it is not possible to
relate the general property and the individual which are different by the relation
of inherence as Vaisesika and others do. Because no one ever has the
understanding, "There is cowness in this." Instead one sees the simu
presence of the pair of opposing cognitions (distinguishing the in
from the generic property, yet identifying the individual with it),
cow." SV, p. 395.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MIMAMSA 209
While proving the exist

shows that he does not cons
distinct from the individual
shortly, the relation betwee
which it inheres bears some resemblance to the relation between a
whole and its parts. Sabara entertains the following objection:
The class property does not exist because it is nothing other than
the individuals just as the forest does not exist because it is
nothing other than the trees. In his answer he says the following:
Moreover, just because we don't perceive the forest to be distinct

from the trees doesn't mean the forest doesn't exist. If for some other
reason a cognition arises which contradicts the existence of the forest,
then certainly the cognition of the forest is false. Then we understand
that the forest does not exist.... On the other hand, if a cognition
does not arise to contradict the existence of the forest and other
(aggregates) then it is not true that they do not exist. (1.1.5 S4b).
Hence, a whole exists because we directly perceive it. The fact

that it is not distinct from its parts does not imply that it does not
exist. The conclusion that it does not exist follows only upon a
cognition which contradicts its existence. Similarly, a property
exists because we directly perceive it. The fact that it is not
distinct from the individual to which it belongs does not imply
that it does not exist. Therefore, Sabara recognizes that a
property is not entirely distinct from the individuals to which it
belongs.
However, he is not of the opinion that the individual is
merely identical to its properties. In the dialogue in the
Akrtyadhikarana just discussed, he shows that he considers the
individual to be other than its properties. Just before his statement
quoted above, the proponent says, "If it is used for another
individual, then the individual is not the object denoted by the
word because only that which is devoid of all general and specific
properties is called an individual (IIIClb)." This statement, very
similar to the one just quoted, intends to say something quite
different: If the same word refers to more than one object, it does
so by denoting a property common to the objects, not by
denoting the individual, because an individual is unique, not
common to more than one thing. Hence the fact that they are
common gives properties an existential status which is not
reducible to the individuals to which they belong.
3.3.3. Class Property
In his commentary on Jaimini 1.1.5, Sabara cites a long

passage from the Vrttikara including a discussion in which he

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
establishes that class properties exist.30 The examination of th

discussion in the next few pages makes frequent reference to th
passage. The reader may find the text and translation und
Jaimini 1.1.5 in the translation section of this chapter.
The fact that cognition is inherently positive (?3.1.1)
means that our cognitions are based on real objects. If one s
cloth, cloth is there; it is not an illusion (1.1.5 Si). Similarly
general cognition (the single cognition which recurs with respe
to more than one individual object) is based on a general
property. One can only establish that a cognition is illusory if
another cognition shows it to be false (1.1.5 S3). Direct
perception (pratyaksa) is the strongest means of knowledge with
regard to things which are directly perceivable. Hence if direct
perception testifies to the existence of something, nothing can
contradict it. The class property is directly perceptible
(pratyaksa).31 It is not inferable (sadhya) from the presence of
certain features or marks on an object (1.1.5 S2). Therefore, the
class property exists.
At the beginning of this argument, the Vrttikara states that
the word 'cow' denotes the class property qualified by the dewlap,
etc. Then, asked whether the class property is something to be
inferred (sddhya) or not, the proponent says, "Being directly
perceptible (pratyaksa) it cannot be something to be inferred
(1.1.5 S2)." Kumarila confirms that the word 'pratyaksa' here has
more than its ordinary sense of 'obvious'. He says that a general
property is not known by inference. Being directly perceptible
one knows it without inference (laksana). An inferential mark
(hetu) is not required to bond a generic property to an individual.
The bond is natural.32
30Akrtivada 2. Tatsadbhdvaprasiddhyartham atra tdvat prayatyate; vdcyatve

vaksyate yuktir vyaktyd saha baldbale. Here (under sutra 1.1.5) we try to show
that the general property exists. Later (in the Akrtyadhikarana) we will state the
reason that it is the object denoted, a reason regarding its relative strength in
comparison with an individual. SV, p. 385.
31Pratyaksa does not require that the direct knowledge come through the
senses. It may be a sensual or mental intuition. I translate with the word
'perception' in this broader sense.
32Akrtivada 30b. Dharmas cdvyabhicaryasya na mrgya upalaksane. 31.
Ndnumanavagamyam tat, pratyakse laksanena kim; svdbhdvikas ca sambandho
jdtivyaktyor na hetumdn. 32. Tenaitasya prasiddhyartham ndnyat samdnyam
isyate; iaktisiddhivad etasya svabhdvo 'tra na vdryate. 30b. One should not
seek a property to indicate the undeviating (general property). 31. A general
property is not known by inference. Being directly perceptible, what is the need
for an indicative mark? And the relation between a generic property and an
individual is natural not caused. 32. Therefore, there is no need for a second
general property (to determine the set of individuals) in order to achieve the
relationship of the individuals to the first general property. Its own nature is
not excluded (from achieving the relation between the individual and the
general property) just as (the individual's) nature achieved (the relation between

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MIMAMSA 211
Saying that the class property is directly perceptible, the

Vrttikara gives the example of three golden ornaments. One
directly perceives that a golden necklace is a necklace, a golden
svastika a svastika and a golden cup a cup (1.1.5 S2). The fact
that all three items are gold yet are different indicates that they are
what they are because of something other than the gold which is
their substance. That by virtue of which they are recognized as
being the type of object they are is their class property. Being a
necklace is the class property of necklaces, etc. The example
shows the class property to be something other than the substance
and yet directly perceptible. One has the same cognition with
respect to one gold piece which one identifies as a necklace as one
has to any other piece so identified. Kumarila explains in
Akrtivada 3 that the single cognition which one has with respect
to different individuals has a basis, a cause. That cause is the class
property (?3.3.1 note 19).
The perception of the class properties, being a necklace,
etc. is not delusion. First of all, it is not delusion in the sense that
it is completely without any basis in reality. The cognition has
some basis. We do not perceive a forest when no trees are
present, or a cloth when no threads are present. Similarly, we do
not perceive the individual necklace's class property, being a
necklace, if there is no golden ornament present (1.1.5 S4a).
Secondly, it is not delusion in the sense that we perceive
more than is there. The objector implies that just as we falsely
perceive the forest given only trees, we falsely perceive the class
property given only individual substances. Just as the forest has
no existence distinct from the trees, the class property has no
existence distinct from individual substances. The Vrttikara
answers that just because it is not distinct doesn't mean it doesn't
exist. Just because the forest is not distinct from the trees, does
not mean there is no such thing as a forest. The whole is not
distinct from the part but still exists. Similar to an aggregate such
as a forest, the class property which many substances have in
common does not exist distinct from the substances but still exists
(1.1.5 S4b).33
In his Akrtivada, Kumarila says there is a natural relation
(svdbhavika sambandha) between an object and its property. Just
itself and) its capacity (to reveal the generic property).

Parthasirathimisra comments on 31b-32: Yatha hy abhivyaktisaktih
svdbhdviki tathd sambandhasydpi svabhdva eva hetuh. Kiscid vyaktayah kincit
sdmdnyam svabhdvdd evdtmand sambadhnantiti. Just as the capacity (of an
individual object) to reveal (a generic property) is natural, the relation (of the
individual directly to the general property) has nature as its cause. Certain
individual entities simply out of their own nature bind a certain general
property to themselves. SV, pp. 390-91.
33A class property is not to be distinguished from the class which is a special
sort of aggregate.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
as things have natural capacities, the nature of the individual th

binds to itself a certain property. The general property is direc
perceptible and immediately evident in an individual obje
Specific features of the object are not the mediate cause of ou
knowing it. On the contrary, in exactly the same way that on
would identify an object as a cow because it is identical to thos
specific features, one can identify it because it is identical to
general property cowness. Therefore, he concludes that th
individual has a natural connection to the general property.34
On the other hand, Kumarila does not regard the
individual and the general property as entirely the same. The
individual object is at once understood to be different from the
general property cowness and identical to it. When one says,
"This is a cow," the word 'this' denoting the individual gives the
cognition of difference, the word 'cow' denoting the general
property gives the cognition of sameness. Yet due to the identity
expressed by the verb 'to be' in the sentence, both words refer to
the same entity as the substratum of these two opposing
cognitions. Because the two cognitions are in a single entity, the
individual and the general property are not entirely different.
That the individual is at once different from and the same
as the general property is not contradictory. One thing differs or
is the same as others in different ways. One determines the unity
and diversity of an object considering different natures (?3.3.2,
note 29).
3.3.4. Class Property: Abstract Generic

Property versus Shape
In his Tantravarttika on the Akrtyadhikarana, Kumari

argues vigorously to show that the class property is not a sh
He comments that Sabara's purpose in defining the class prope
(ak.rti) as a general property (sdmdnya) at the beginning of
Akrtyadhikarana (IA1) is to remove the doubt that it is a
34Akrtivada 45. Sdsnddyekdrthasambandhi gotvam ity upalaksanam; na ca

svasamavdyy eva kevalam cihnam isyate. 46. Sdsnddibhyas tu pindasya bhedo
ndtyantato yadd; sdmdnyasya ca pindebhyas tadd sydd etad uttaram. 47. Kasmdt
sdsnddimatsv eva gotvam? Yasmdt taddtmakam; tdddtmyam asya kasmdc cet,
svabhdvdd iti gamyatdm. 45. The (marks of a cow) dewlap, etc. point out the
(generic property) cowness which belongs to the same objects that the dewlap,
etc. do; nor is it true that only that which inheres in something can be a mark of
it. 46. When the object is not entirely different from the dewlap, etc., nor the
general property from the objects, then the following will suffice to answer: 47.
(Question:) Why is the general property cowness present only in individuals
with the specific features, dewlap, etc? (Answer:) Because it is identical to
them. If (one asks): Why is it identical to them? (I answer:) Understand that it
is identical to them from its own nature. SV, p. 393-94. See also note 32.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MiMAMSA 213
configuration of parts (avayavasamsthdna) in Mimamsa as it is in

Nyaya.35 He argues as follows: First, a configuration cannot be a
class property because each individual is shaped slightly different.
Hence, each individual has its own configuration not common to
other individuals. For the purpose of this argument he accepts the
Nyaya-vaisesika view that each quality is one of many quality
instances of a kind. A general property of those qualities inheres
in each quality instance of the kind. Each quality instance in turn
inheres in a single substance. He considers configuration to be a
quality instance which inheres in a single substance. Second, the
general property of all configurations cannot define the class of
cows, for example, because it belongs to any object which has
shape, not only to cows.
Kumarila denies that there is any middle road. One
cannot specify the configuration in such a way to include only the
intended individuals and exclude all others. One cannot specify
the shape of a cow with enough precision to distinguish it from a
horse yet generally enough to include all cows. The only thing
that identifies all and only cows is the generic property cowness.
Furthermore, he says, class properties belong to
substances, qualities and actions. Only earthly substances
independently have shape, and certainly the substances direction,
time, and the self, as well as qualities and actions, do not have
shape. Therefore, in order that a class property be the sort of
thing which can belong to any of them in the same manner, it
must be simply any general property (sdmdnya) not a
configuration (avayavasamsthana).36
35Satyam evdkrtih purvam pratyaksena nirtpitd; samsthdndaankayd tv atra

sdamnydtmd nirupyate. True, he (Sabara) described the class property as directly
perceptible before (1.1.5 S2). But here (IA1), because of the doubt that it may be
the configuration, he describes it as being a general property. MD, part 2, p.
236.
36Evam prapancitam sarvam arthasdmanyam dkrtih; na samsthdnam kuto hy

etad dtmddigunakarmasu? A class property is any general property of objects as
I have just detailed above. It is not a configuration because where is the
configuration in substances like the self, in qualities and in actions?
Sarvesu hi pdrthivesu gavddighatddisu samsthdnam bhavet. Agnitoyavdyv-
dksdindndm tu pdrthivadravyaparigrahavaidd dkdrdnuvrttih kalpyeta na
svdtantryena, dikkaldtmamanasdm tu gunakarmandm ca na kathamcid api
samsthanam sambhavati. Sadmnyamdtram tu sarvasvavisesdnuvrttirupam
sambhdvyate. Tasmdd etad evdbhidheydkrtilaksanam navayavaracana-
samsthdndkrtitvam avakalpate. There may be configuration in all material
objects such as cows, etc. and pots, etc. Although one may suppose that fire,
water, air and space have the shape which they take on because of being
included in a material substance, they do not have shape independently.
However, shape is in no way possible for (the substances) direction, time, the
self and mind, and qualities and actions. But an abstract general property in the
form of that which recurs in each of the individual objects it qualifies is
possible in them all. Therefore, just that is the character of a class property fit to
be denoted. If a class property were a configuration constructed of parts it

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Kumarila's arguments that shape cannot be a class

property are not entirely sound. He notes that there is general
agreement that a word denotes the common element which is the
basis of the same cognition with respect to different things
whatever that element may be. Parthasarathimisra comments that
the disagreement lies in the nature of that element.37 It seems
plausible that Sabara intends shape itself to be a common property
defining a class of individual substances.
wouldn't serve the purpose.

Samsthdnasya ca nfsitvdt prativyakti ca bhedatah; sdmdnyavyavahdritvam
nakrtitve 'vakalpate. Atha samsthdnasdmdnyam dkrtitvena grhyate; asvddisv
api tulyatvdd bhaved dkrtisamkarah. Na cdvdntarasamsthdnam sarvagopinda-
vrtti yat; asvddibhyo nivrttam ca gosabddlambanam bhavet. Because
configuration perishes and differs for each individual, (if being an dkrti meant
being a configuration) if something were an dkrti it would be impossible for it
to be the general property denoted by a word. If the configuration general
property were taken as class property, there would be confusion of class
properties because it is as much in horses, etc. (as it is in cows). Nor can a more
specific configuration which is present in all the bodies of cows and absent
from all horses be the object of the word 'cow'. [Note: Samsthdnasdmdnya is the
general property in all configurations. Configuration is a conjunction quality
instance which belongs to a single individual. The general property
conjunctionhood is present in all conjunctions, that is, in any shape
whatsoever. Hence, the class property could not distinguish between apples and
oranges, or as Kumarila exemplifies objects which are obviously different,
between cows and horses.]
Tasya hy upalaksanam alocyamdnam na jdter anyal labhyate tatas ca jdtir eva
sdmdnyam iti nyayendpadyate. Tena prathamapdde rucakasvastika-
vardhamdnakoddharandt samsthdndkrtyabhidhdndsankaksapddadarsanavad dsid
ity etdbhydm prasnottardbhydm vydvartyate. atas ca dravyagunakarmandm
ydvatkimcit prdg vyaktibhyah sadmnyam tat sarvam dkrtir evety abhipretya
mdtrasabdah prayuktah. Because, search as one may, one cannot find any
identifying mark of it other than the generic property. Therefore, it follows
logically that the general property (sdmdnya) is just the generic property (jdti).
Because of the example of (the three golden objects) the necklace, the
auspicious symbol (svastika) and the cup in the first pdda (1.1.5. S2), the doubt
arose that the word 'class property (akrti)' means configuration as it does in
Gautama's view (Nyaya). This question ("What is a class property?") and its
answer ("any general property of substances, qualities or actions" (IA1)) remove
that doubt. Hence Sabara uses the word 'any (-mdtra)' intending that a class
property is any general property whatsoever of substances, qualities or actions
up to but excluding individuals. MD, part 2, p. 237.
37Akrtivada 4. Tannimittam ca yat kincit sadmnyam sabdagocaram; sarva
evecchatfty evam avirodho 'tra vddindm. Everyone wants whatever the common
element is which is the basis of (the same cognition of different objects) to be
the range of a word so the theorists agree on this matter. Parthasarathimisra:
Ekabuddhinimittam yat sdmdnyam jdtirupam sdrupyarupam apoharupam, tasya
sabdavdcyatvam sarvavddindm avivddam, tatsvarupe tu vivdda iti. All the
theorists agree that the general property which is the basis of a single cognition
(with respect to many objects), whether it has the form of a generic property, the
similarity (of the objects to each other) or the removal (of other objects from
knowledge), is the object denoted by a word, but disagree as to its nature. SV, p.
385.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MIMAMSA 215
Although Kumarila's remark that not all entities which one

identifies as having general properties have shape is a propos,
discussions concerning class properties, in the first instance,
concern ordinary material substances which do have shape. The
discussions concerning the meaning of a word in Nyaya and in
the Mahabhasya certainly choose such objects as examples. The
concept easily extends from that of general shape to that of any
general property whatsoever. Hence the question is the
following: In the case of ordinary material objects such as cows,
falcons, ornaments, and pots, what does Sabara, and even
Kumarila himself, consider a class property to be, a shape or an
undefined abstraction?
It has been amply described how a general property is the
basis of having the same recurrent cognition with respect to
distinct objects.38 I have also described how the shape of an
object may be conceived of either as a quality belonging to that
object alone or as one held in common.39 If it is conceived of as
common then it is a general property. Alternatively, if it is
conceived of as the quality conjunction inhering in a single
substance then a specific type of conjunction, more specific than
the general property of all configurations, constitutes a property
common to each of similar shapes and is the basis of the same
cognition with respect to the objects. In either case, a shape
general property is capable of defining a class.
While arguing that configuration cannot be a class
property (akrti), Kumarila adopts the Nyaya-vaisesika view that
each quality is one of many quality instances manifesting a
quality general property and inheres in a single substance (?3.3.1
note 22). However, while arguing that speech (sabda) is eternal
he rejects this view in favor of the view that the quality itself is
general. He says that there is no need to accept any general
property other than this. Individual differences in the appearance
of the quality can be accounted for by the differences of the
individual substance instances in which it inheres. There is no
need to posit a different quality instance for each differen
appearance of the quality. For example, there is one gener
quality white present in any number of individual white
substances and which is one of the several colors in which the
general property color (ripata) inheres.40 By analogy, there is a
single general shape characterized by a dewlap, tail, hump, head,
horns, and four legs present in each individual cow and which is
38See ?1.3.1.5, ?2.2, Gautama 2.2.69, and Uddyotakara's debate with the
Buddhist (Gautama 2.2.64 U4-U6), ?3.3.3.
39See ?1.2.3, ?2.2, Gautama 2.2.64 U4.
40Sabdanityatadhikarana 41 b. Etayaiva disd vdcyd sukldder api nityatd. 412.
Samsargamdtrabhedena syat tatrapi bhedadhfh; svari7pam tu tadeveti ko jdtim
kathayisyati. SV, p. 591. See Halbfass (1980: 233).

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
one of the many configurations in which the general property

all configurations inheres.
Immediately following his argument that configuration
cannot be a class property (dkrti), Kumarila argues against t
Vaisesika notion that an exclusive specific property qualifies ea
individual eternal entity. He says that properties, each of which
general, can form an exclusive set capable of uniquely specifyi
each individual. Hence, it is unnecessary to presume that there
a unique specific property for each individual (?3.3.2, note 2
The basis for presuming that a property exists over and above a
individual substance in which it is present is that it is present
more than one individual substance as the basis for recognizing
to be the same as another individual substance. If the property
were present in only one substance there would be no basis for
presuming that it is different from its substratum. This reasoni
applies to all individuals, not just to eternal substances.
Analogous reasoning applies to qualities. If a quality
were present in just one individual substance, there would be n
basis for presuming it to exist over and above the substance itse
The fact that qualities are the basis for recognizing individ
substances to be the same is the only testimony to their existen
Therefore, by fiat qualities are general in nature. Due to the f
that one perceives various individual material substances to be t
same, one accepts as the basis of that cognition some gener
element, call it the class property. The question is whether Saba
and Kumarila consider the class property of a class of individu
material objects, which is the basis for the same cognition one h
with respect to each of those individual objects, to be just
abstraction, or a general shape.
In Nyaya, the configuration, among other features, of an
object such as a cow is an inferential mark of the generic proper
(Gautama 2.2.68 V1, U1). In the case of other substances like
clay, gold, and silver, Vacaspatimisra suggests what Uddyotakara
implies, that color is an inferential mark of the generic property
(Gautama 2.2.68 U2, ?2.2 note 15).41 Sabara and Kumarila
consider the class property to be directly perceptible; one need
not infer it. Kumarila states in his Akrtivada that one does not
require a second general property in the form of an inferential
mark in order to make the abstract general property known
(?3.3.3 note 32). Hence one does not require the general shape
including such features as a dewlap, hump, tail, etc. to indicate the
cowness of a certain animal; one directly perceives it. Nor does
one require the shape of a falcon with crooked wings and fanned
tail to indicate falconhood. Hence there is no direct relation
between the abstract general property, which is the basis of
41Kumarila himself expresses this view in the Slokavarttika, Vanavada

SV, pp. 439-40.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MIMAMSA 217
general cognition, and a shape. The only relation between them

is that they both belong to the same individuals (?3.3.3 note 34).
There is also no logical reason to have them both. The
class property is directly perceptible in the sense that one
immediately sees the object as what it is. But asserting that the
generic property is an abstract entity other than the shape of a
cow or falcon, and other than the color of clay, gold, or silve
requires some evidence. Leaning on the assertion that the clas
property is directly perceptible and the fact that perception is
inherently positive, Kumarila gives no evidence. We directly
perceive shape, color, and other features. We do not directly
perceive a generic property. Presumption could establish that th
generic property exists as the basis of our general cognition
However, it cannot establish that it exists in addition to shap
because shape serves to identify the object. One can only
presume that without which the general cognition would not
occur. If shape is coextensive with the general cognition, then it
accounts for the general cognition and there is no ground for
presuming another basis.
I have just mentioned Sabara's statement that the word
'cow' denotes the class property qualified by the dewlap, etc. If
the general shape itself is not the,class property; but rather, an
abstract general property is, in what way do these features qualify
it? In Akrtivada 45 Kumarila says that these parts serve to point
out the general property cowness from among the other general
properties present in a cow (?3.3.3 note 34). If the general
property does not require an inferential mark to make it known,
why does Sabara use one in this instance? If the general property
is directly perceptible, why doesn't Sabara just mention it directly?
Why mention other features common to cows which are directly
perceptible. Isn't it rather that Sabara mentions these features to
describe the shape general property which is itself the class
property?
Kumarila blames Sabara's example of the golden
ornaments for giving rise to the doubt that a class property might
be a configuration (?3.3.4 note 36). The cause of doubt is that
the example does not necessarily show that a generic property as
distinguished from an individual is the meaning of a word. It
leaves open the possibility of concluding that shape as
distinguished from the substance is the meaning of a word. The
substance in all three is the same, gold. Yet one uses a different
word for each of the three objects, a necklace, a svastika, and a
cup. Concomitant presence and absence determine what the word
means. Therefore, the word does not denote the substance; it
denotes what differs from one object to the next. The shape
differs. Hence, a word denotes shape. One identifies the object as
what it is by its shape. One shape is a necklace, another is a
svastika, a third is a cup.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Kumarila's objection implies that Sabara chose a bad

example. However, it seems to me that he chose the example of
ornaments deliberately. Choosing three gold ornaments intend
their common material substance to be brought to the fore. Th
point is to show that there is no difference in material but only
shape; only shape can be the basis of the use of a different wo
not substance.42 Because he requires the word-meaning to
eternal, however, he makes it clear that he does not mean shape
it inheres in a single individual. If Sabara had not meant sha
itself to be the class property which a word denotes, he would n
have used the example of golden ornaments. He would have
stuck with the example of the cow with which he had introduce
the section. By using the example of golden ornaments, he show
that he means shape as distinguished from the substance to be t
meaning of a word.
Sabara cites the Vedic injunction to build a falcon altar a
evidence that a word denotes the class property and not an
individual. The injunction is carried out by piling up bricks
the figure of a falcon. Sabara argues that the injunction is
possible if the word 'falcon' denotes the class property bu
impossible if it denotes an individual because it is impossible to
create a falcon by piling up bricks. It is obvious to all that it i
also impossible to create a class property by piling up bricks. A
class property is eternal. Creating the class property mean
manifesting it, constructing an individual object to which the cl
property belongs and which is therefore a member of the class.
it is impossible to build an individual, it is equally impossible t
manifest the class property.
Kumarila astutely observes this and concludes that
whether the word 'falcon' denotes the class property or an
individual the injunction means that one should construct
something similar to the object denoted by the word 'falcon'. He
says that it is possible to make an altar similar to the class
property. Regarding the impossibility of making one similar to
an individual, he continues as follows:
And if one selected one individual (to which to make the altar
similar) one would not be able to make it similar to any other
individual. Therefore, the action of building a falcon altar is not
possible if the object denoted by a word is an individual (mA note
12).
It is certainly true that a good copy of one individual differs from

other individuals because one individual differs from another in
some respects. However, each individual is similar to every other
42The example is very similar to Patafijali's example of gold shaped into

different objects in the Paspasahnika. See ?1.2.3.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MIMAMSA 219
in the class. It is precisely this similarity which the class property

captures. Hence, it is not true that by being similar to one
individual the falcon altar would fail to be similar to any other.
Kumarila's point, however, is that the altar should not be more
similar to one individual than to another. It should be similar just
to that element which is common to all the individuals. If it is,
then it is similar to the generic property not an individual.
Jayantabhatta, in the Nyayamafijari, states the following
objection to Kumarila's argument:
Syenavyaktya cet sadrsyam istakaku.tasya ndsti, vyaktyantarena

vyabhicdrdt; nitardm amurttayd jdtyd sadrsiyam dkaseneva na
tasydvakalpate.
If the pile of bricks is not similar to an individual falcon because it
would fail to be similar to another individual, then it is no more fit to
be similar to the insubstantial generic property than it is (fit to be
similar) to space.43
If the class property is merely an abstract generic property, there

is simply no similarity between it and a shape which is not
mediated by an individual. Kumarila himself recognizes that the
dewlap and other marks are able to serve to point out the generic
property cowness because they and the generic property belong
to the same individual object (?3.3.3 note 34). Hence, the
individual mediates the relation. The generic property belongs to
a real individual of the class not to a model. The shape in which
one lays the bricks is the figure of any individual which has that
generic property, say for instance, the first one one sees or the
first remembered.
One may argue that Kumarila's statement, "if one selected
one individual (to which to make the altar similar) one would fail
to make it similar to any other individual," intends to point out the
following facts: An individual by its very nature is unique; it is
not common to the class. The altar has to represent any falcon
whatsoever not a particular falcon. Hence the altar must have a
property common to all and only falcons. The fact of being a
property common to all and only falcons makes the general shape
coextensive with the generic property. This coextension is the
basis of the shape's greater similarity to the generic property than
to any one individual.
This argument has two difficulties: First, Kumarila has
argued that configuration inheres in a single individual, or if it is
considered to be general, the configuration of a cow is no
different from that of a horse. If he sticks by these arguments, he
cannot determine a general shape which belongs to all and only
individuals with a certain generic property. If the shape is not
43Nyayamafijari, part. 2, p. 37, lines 13-14.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
coextensive with the generic property, then there is no basis f

the shape's greater similarity to the generic property.
Alternatively, if the shape is coextensive with the generic
property, there is no reason to presume that the two are distinct
entities.
Second, the following position is obviously contradictory:
The shape is coextensive with the generic property but the altar
with that shape is not a real falcon so it can't have the generic
property. Since it has the shape but not the generic property, the
shape and generic property are distinct entities. Yet they are
similar because they are coextensive. The contradiction is
between the following two sets of statements: 1) The generic
property does not belong to a model; the shape does. 2) The
generic property is coextensive with the shape. By definition,
coextension is just the very fact of belonging to the same entities.
Now the first statement is the ground for distinguishing the
generic property from shape. The second statement is the ground
for similarity between the generic property and shape. Two
things cannot be both the same and different in the same respect
at once.
Therefore, if the class property is an abstract ge

property and not a general shape, Sabara's example of the
altar fails to constitute evidence that a word denotes the class
property and not an individual. On the other hand, however,
shape itself is the class property it succeeds in a most
straightforward manner even without resorting to similarity.
Whatever has the shape of a falcon is a falcon. Therefore, a pile
of bricks in the shape of a falcon is a falcon too. The word
'falcon' in the injunction to build a falcon altar denotes shape
which is the class property of the class of falcons.
There are two drawbacks to these arguments which
indicate that Sabara recognizes a class property to be more than
just shape. First of all, the dichotomy of shape and substance,
where the latter is conceived of as the material from which an
object is made, such as brick, or sinews, etc., does not hold up i
Sabara's arguments concerning the meaning of a generic term. I
has been discussed how Patafijali assumes that it is the dravy
which is associated with dual and plural number rather than the
akrti, in the arguments under Panini 1.2.64 and 1.2.58. The
implication of this, namely that he must conceive a dravya as an
individual not as a material, has also been pointed out (?1.2.8).
Similarly, Sabara's presentation of the argument for the view that
a generic term denotes an individual includes a reason which
applies to an individual but not to a material. Under Jaimini
1.3.31, Sabara says, "If the animal brought for a Vedic
performance passes away, one should offer another one of the
same color and age." The argument states that being other makes
sense for a dravya but not for an akrti. "Even another individual

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MIMAMSA 221
animal (pasu-dravya) has the same akrti (IIE)." The argument

adduces this as a reason for considering that an individual
(vyakti) is denoted. Hence Sabara uses both the words 'dravya'
and 'vyakti' to mean an individual, not a material.
This has profound implications against the argument that
Sabara considers shape to be the class property. If shape is the
class property, then in the case of the falcon altar, the class
property belongs to the pile of bricks assembled in the shape of a
falcon. In this case, the pile of bricks is a member of the class of
falcons and hence is a falcon. If the pile of bricks is a falcon,
then it is an individual falcon too. Yet Sabara himself says that in
building the altar one does not succeed in constructing an
individual falcon. One is no more able to build a shape without
constructing an individual than one is able to manifest a class
property without creating an individual. Therefore, in either case,
whether the class property be an abstract generic property or a
general shape, the example of the falcon altar fails to show a word
used for a class property without an individual. The difficulties
of Sabara's argument concerning the falcon altar will be examined
in more detail shortly.
Sabara's argument to prove that a word is eternal and that
each enunciation does not constitute a new word (?3.2.2, note 13)
shows that his concept of a class property is clearly broader than
just the shape of a material object. If a word were new each time,
it could not denote the class property because every object has
many class properties. New each time, the word could not form a
relation with the appropriate one among them. This argument is
not valid on the view that what is meant by a class property (dkrti)
is the shape of material objects.
Sabara explains the reason that the word, if it were new
each time uttered, could not denote the class property in the
following manner:
For the creator of the word-meaning relation would relate the word to
the class property by pointing at the class property. But because there
are many class properties in any particular cow, by what means other
than the word itself would he indicate the general property cowness
singled out from all the other properties in the cow as the one denoted
by the word 'cow'? (?3.2.2 note 13)
If he considered a class property to be just a shape, there would

be no reason that the creator of the word-meaning relation could
not relate the word to shape by pointing at it, even if there are
many shapes present in the cow. Because what could many
shapes be but the shapes of various parts? Each part has its own
location and so can be pointed to.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Sabara may consider other common elements which are

not abstractions besides shape to be class properties.44 For
example, if the word 'cow' occurs once while someone points to
an object of a certain shape, endowed with breath and motion and
having a certain size, color, etc., one would be uncertain whethe
the word referred to the shape, breath, motion, size, color, or
some other property of that object. However, if the same word
'cow' occurs again and again while someone indicates the animal
of that shape even when the other properties, color, age, size, etc
vary, then one may understand that the word does not refer to
those other properties; rather, it refers to the object with the
recurrent property or properties. The point of the argument is
that the procedure to determine meaning, concomitant presence
and absence, requires repetition of the same word with an abiding
object. This concomitant presence and absence could succeed i
shape is the object it denotes.
On this interpretation, the term 'dkrti' then has two senses:
1) It is that which is the common element among a number of
objects, no matter whether it be a shape, a quality, or an action.
As such it is an abiding entity. 2) It is a producible individua
shape as in a falcon altar which one creates.45
It seems more likely that Sabara uses the term 'dkrti' in the
two senses mentioned at the start of the discussion: 1) a particula
shape, and 2) an abstract generic property. These are the senses
of the term familiar in both grammar and Nyaya. Hence in the
argument for the eternity of speech sounds, it is the abstrac
generic property which is meant. The other generic properties,
such as substancehood and existence, accompany the generic
property cowness in the object. The shape clearly serves as an
obvious example of a property which recurs and hence of an
abstract generic property. In Sabara's arguments neither the
conception of a class property as just shape, nor the conception of
it as an abstract generic property are consistently adhered to. H
may conceive of a class property abstractly but deliberately
confound it with shape in order to prove that a class property i
denoted by a generic term, or he may in fact not clearly
distinguish the senses of the term 'dkrti' himself.
44Just as he conceives shape as general unlike Nyaya and Vaisesika, he may

conceive qualities and parts as general just as he does actions. The use of the
word 'general (samdnya)' in Vaisesika sutra 1.1.18, Dravyagunakarmandm
dravyam kdranam sdmdnyam, suggests that it is also possible to consider
substance as a common element. VS, p. 30. However, Sabara would never apply
the concept of class property (dkrti) to a single substance as the common cause
of two or more objects. For example, gold in three golden ornaments is not a
class property (dkrti) but a single substance (dravya).
45D'Sa (1990: 91) seems to have these two senses in mind. He (p. 87) certainly
interprets the generic dkrti in the first sense.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MiMAMSA 223
Kumarila clearly conceives

of generic terms as one concer
property versus individual.
arguments to free the concept
shape, the fact that he assume
some similarity to the gene
conceives the generic proper
than shape what similarity i
intentions, in trying to defen
concerning the falcon altar, sh
property.
3.3.5. The Bearing of the Nature of a Class

Property and an Individual on the Denotation
of a Generic Term
The class property is eternal; the individual substanc

which has that property is not. If the meaning is the clas
property, the word-meaning relation will be eternal and n
otherwise. Hence Mimamsa milst argue that the meaning of th
word is the class property and not the individual substance.46 T
Akrtyadhikarana presents initial arguments for the view that
generic term denotes an individual alone, then rejecting th
presents conclusive arguments for the view that a generic ter
denotes only the class property.
3.4. Analysis of Arguments in the

Akrtyadhikarana
A detailed outline of the argument for the concl

the Akrtyadhikarana, that a generic term denotes o
property, appears at the beginning of the translation
this chapter. Sabara discusses four views of what the m
a generic term is: 1) the individual alone, 2) the
individual or individual as being of such and such a ty
qualified class property or class property as belon
individual and 4) the class property alone. First he pre
view that the individual alone is the meaning then
presses his objector to accept first that the class proper
46Kumarila states this succinctly in Akrtivada 1. A krtivya

sambandho nityatdsya ca; na sidhyetdm iti jnatva tad vacyatv
Knowing that if the word-meaning relation were with respect to an
than the class property, neither the relation itself nor its eternity
in Mimamsa we say that the class property is denoted. SV, p. 385.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
included in the meaning, second that the individual must b

excluded.
He rejects the first view, that the individual alone is the
meaning, because the following two facts make it untenable: 1)
The general property must be an element in the resulting
cognition in order to account for the knowledge of an individual
as the same as others of the class (IIIC3aii, 3b). 2) The general
property must identify the individuals the word denotes in order
to account for the use of the word with respect to all and only
individuals of the class, specifically members unknown
previously (IIIC4).
Number one implies that the cognition of an individual
following the use of a generic term does not have the individual
alone as its object. Its object is the individual qualified by the
class property. Sabara says that no one can apprehend a qualified
object without knowing its qualifier. If the general property
serves to mark the individuals the word denotes, then "one would
apprehend an individual which is qualified. And if qualified, one
would understand the qualifier first. No one can apprehend
something qualified unless the qualifier is apprehended (IIIC4b)."
Kumarila states the principle as follows:
Visistagrahanam nestam agrhFtavisesanam; ...

Agrhltavisesane tadanuraktd visesyabuddhir nopajdyata iti
tadgrahanam drdatavyam.
It is not accepted that there is knowledge of the qualified without
knowledge of the qualifier; ...
(Because of the principle) "Without knowing the qualifier, the
knowledge of the qualified, which is based on it, does not arise," one
must respect the knowledge of the (qualifier).47
For the general property to identify an individual as an

object the word may denote, one must know the general property
to be present in the individual before applying the word to the
individual. However, that does not necessarily imply that the role
of giving knowledge of the general property must fall to the
word. Explaining this limitation of the principle, Kumarila
continues as follows:
47TV, vol. IV, p. 341. Kumarila refers to this principle twice in the
Akrtyadhikarana. His first reference is by way of paraphrase. He says, "Because
without denoting a qualifier it is impossible to denote a qualified object it
qualifies.... (Visesanam anabhidhdya tadvisistavisesydbhidhdndsambhavdd....)"
(MD, part 2, p. 241, line 12.) In his second reference, he says, "Because of the
principle, 'Unless the qualifier is known,' etc. (nadgrhitavisesaneti nydydd...)."
(MD, part 2, p.248, line 4.) The principle appears in the Mimamsanyayaprakasa
(p. 7, line 12.) as follows: Ndgrhitavisesana buddhir visesya upajdyate. Unless
the qualifier is known, knowledge of the qualified object does not arise. The
Mimamsanyayaprakasa quotes the verse line which I quote from Kumarila.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MIMAMSA 225
Na tv anabhihitavisesanatvena kimcid dusyati. Yadi namadnabhi-

hitam api visesanam kenacid anyena pratydyyamdnam visista-
pratyayotpddandya parydptam bhavati, kim tatrdbhidhdndnurodhena.
But no fault arises if the word doesn't denote it. If even without
being denoted something else conveys the cognition of the qualifier
so that it is competent to bring about the cognition of the qualified
object, then why does it have to be denoted?
In the case of a generic term, there is nothing else which

consistently gives knowledge of the general property. If the word
denoted only an individual and did not supply the knowledge of
the general property, something else would have to. One can't
rely on the individual object itself to give knowledge of the
general property because knowledge of the individual results
from the word which doesn't give knowledge of the individual
until the general property indicates the individual as an object to
denote. Knowledge of the general property would depend on
knowledge of the individual and knowledge of the individual
through the word would depend on knowledge of the general
property. Hence, it would be circular to say that the general
property indicates the individual for the word to denote and the
individual gives knowledge of the general property. Therefore,
the word does not denote the individual alone.
The proponent of the view that the word denotes only the
individual agrees that the general property must identify the
individuals the word denotes (IIIC4), and that the general
property is an element in the resulting cognition (IIC). However,
he denies that it must be denoted itself in order to do so, claiming
that one knows the general property from the individuals of the
class. He claims that the individual gives knowledge of the
general property, knowledge of the general property brings about
the word's relation to individuals, and the word gives knowledge
of the individual; yet denies that this is circular.
He argues something like the following: Even without the
use of the word one knows individual objects by direct perception
and other means of knowledge. The individual objects known in
this way give knowledge of the general property entirely
independent of the use of the word. The general property so
known brings about the word-meaning relation between the word
and any individual which has that property. Thereafter whenever
the word is used, the word generates cognition of that individual.
Then the cognition of the individual generates cognition of the
general property. In this way one arrives at the cognition of the
individual as being of such and such a type. It is not necessary
that the cognition of the qualifier precede the cognition of the
qualified.
Sabara points out that such a view is not cogent.
Knowing the relation between a word and its meaning depends

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
upon concomitant presence and absence; hence, it requires

knowing each of the two independently yet concurrently more
than once (?3.2.2, notes 10,13). Concomitant presence and
absence can establish the relation between a word and an
individual object provided that one knows the object
word each independently first. In this case the relation
them would succeed directly, independent of the class
and there would be no reason to presume the mediat
class property there.
However, a generic term is used for members of t
hitherto unknown, for example, a newly born calf (
order to achieve the use of the word for such an object o
to the mediation of the class property because one has
concomitant presence and absence of the word and
property with regard to other individuals (IIIC4). N
neighbor rushes in and announces the birth of a calf. I
not seen the calf yet and his use of the word does not
class property to mind how will cognition of either
property or the new individual arise? The only answer
that it won't arise. By merely existing the class proper
lead one to know that which has that property. It can
by being an object of cognition.48 Therefore, the word
gives knowledge of the class property.
The objector's explanation could account for the u
word for individual objects of which direct perception
means of knowledge have given cognition previously
even there it is cumbersome because it involves the un
mediation of the class property) but it cannot account f
of a word for a member of the class not known previo
knowing the class property as belonging to known mem
the class, one does not know that it belongs to the n
Therefore, the class property can neither bring about
relation with the new object nor be a part of the re
cognition. Hence Sabara correctly requires that the clas
serve as a mark for all the individuals and correctly con
as such a mark it is both an element in the denoted me
the logically prior element (IIIC4).
The views that the word denotes either the qualified
individual or the qualified class property with the other element
taking a subordinate position in the meaning as the qualifier are
moderate views. These are the views Patafijali considers in his
commentary on 1.2.64 vdrttika 53 (53c). Sabara rejects these
48As Kumarila says in Akrtivada 14. Bhaven nirvisayd buddhir yadi saktir na
grhyate; na hi sadbhdvamdtrena visayah kascid isyate. Cognition would have
no object if one did not know the capacity (of a general property to produce the
same cognition). It is unacceptable that something be an object (of cognition)
just because it exists. SV, p. 387.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MiMAMSA 227
moderate views for the following reasons: 1) If the individual

were involved in the meaning at all, the meaning would be
individual, not general (IIIF). One would require the unqualified
class property as a truly general property to identify the qualified
meaning, just as the view that the individual alone is the meaning
required the general property to identify the objects which the
word denotes. 2) Since the individual and the class property are
inseparably connected, once one knows one of them that one will
make the other known. Hence one cannot presume that the word
denotes more than one (IIIE1). 3) Concomitant presence and
absence show that the word denotes the class property and not the
individual (E3).
The second argument, that only one of the two
inseparable elements can be denoted because knowledge of that
one is sufficient to supply knowledge of the other, is equally valid
on the view that the individual alone is the meaning of the word
and on the view that only the class property is the meaning of the
word. The objector mentions that both views are equivalent in so
far as action is impossible on either view yet both views can
account for the action by having the denoted element indicate it
(IIIB). The argument that the word denotes the individual points
out that actions such as washing rice are impossible on the class
property (IIIBla). Sabara will respond that the word denotes the
class property riceness, and the class property indicates the
substance to wash (IIIB2a, IVB1). The argument that the word
denotes the class property points out that actions, such as building
a falcon with bricks, are impossible on an individual (IIIBlb).
The objector could respond that the word denotes a real
individual falcon and the individual falcon portrays its class
property as the shape in which to lay the bricks (IIIB2b). Both
arguments rely on the fact that the individual and the class
property are inseparably connected in order to eliminate the other
from being a part of the meaning (IIB, IIIE1).
The arguments seek to uphold an economy of word-
meaning relations. The objector states the principle of economy
as follows: It is unreasonable for a word to denote more than one
meaning (anyayas cdnekdrthatvam) (IIB). The word 'ekdrthatvam
(that there be one meaning)' glosses the word 'arthaikatvam' in
sutra 1.3.30 interpreted as a statement of the proponent in the
Lokavedadhikarana:
Jaimini 1.3.30. Because (Vedic texts) enjoin performance, each word

has just one meaning because there is no division.
As stated, the sutra applies specifically to Vedic context versus

ordinary context. It asserts that we don't see any differenc
between a word used in the Veda and in ordinary use so the wo
is the same. In addition, if the meaning of the word used in th

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Veda were different from the meaning of the word in ord

usage, we would be unable to carry out the actions the
enjoins. However, the sutra may be generalized to apply to
two contexts. As such it says that we don't see any diff
between a word in one use and that word in another use. If the
meaning in one use were different from the meaning in another
use, we would not understand which one was intended.
Kumarila comments that if more than one meaning per
word were accepted, the word-meaning relation would not b
fixed, one would have to posit more than one signifying capacity,
and the cognition of each meaning would be optional. In the case
of connected meanings, by deriving one meaning from the other,
one can avoid these faults; hence, just one should be accepted as
the denoted meaning (IIB note 4).
The principle that each word have only one meaning
complements the principle Jaimini states just four sutras earlier in
the Vyakaranadhikarana that each meaning have only one word:
Jaimini 1.3.26. Anydyas canekasabdatvam.

It is unreasonable to have more than one word denoting one
meaning.49
Sabara has the objector state the principle that it is unreasonable to

have more than one meaning per word parallel to Jaimini 1.3.26.
Both principles uphold an economy of word-meaning
relations. Jaimini 1.3.26 states that one can only presume tha
there be the minimum number of word-meaning relations which
account for the usage. One cannot presume that a vernacular
word similar to a Sanskrit word for the same meaning bears an
independent relation to the meaning because one can account for
the fact that it conveys the meaning by its similarity to and
derivation from the Sanskrit word. Sabara points out a
qualification: unless there is justification for assuming more than
one. A second word is accepted in that meaning if the Veda
actually states one.
Conversely, there is no reason for a word to denote more
than one meaning, unless there is some justification for presuming
an additional meaning. One cannot presume that one cognition
bears an independent relation to a word if it can consistently be
accounted for via a second cognition following the use of that
word. If the first cognition cannot be accounted for by deriving
it from the second then by all means it must be accepted as a
second meaning. In addition, the second cognition must be one
which cannot be accounted for by deriving it from the first.
Hence, the cognition which does not depend on other meanings
yet consistently arises upon the use of the word is the meaning of
49MD, vol. 2, p. 213.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MiMAMSA 229
the word. That is, one presumes that a word bears a signifying
capacity with respect to a meaning if one cannot account for the
cognition of that meaning other than by the word denoting it.
3.5. Criteria of Denotation
Although words and meanings are directly percept

word-meaning relation is imperceptible. Therefore, neithe
perception (pratyaksa) nor the other means of knowledg
on it (anumdna, upamdna) can give knowledge of it (?3.1.
4). The Veda does not directly state which words matc
meanings. Although speech in the form of dictionarie
statements of knowledgeable people do give such kno
they are not an ultimate source of knowledge. Their kno
depends on other means of knowledge. Hence, based on
observation and evaluation of usage by any of the other means of
knowledge, presumption (arthdpatti) is the final means by which
Mimamsa determines a word-meaning relation. Presumption is
based on not being able to account for the facts otherwise
(ananyathdsiddhi).
Whenever a word is used, many meanings appear with
it.50 Two criteria, according to Mimamsa, determine which
among these is the denoted meaning: 1) concomitant presence
and absence (anvaya and vyatireka), and 2) inability to account
for the cognition otherwise (ananyalabhyatva). 1) Concomitant
presence and absence of a word w and a meaning x is the fact that
whenever word w is present meaning x is present and whenever
word w is absent meaning x is absent (w _ x). 2) Inability to
account for the cognition otherwise is directly based on the
character of presumption. One may presume the cognition of x
to be caused by a word if there is no other way to account for the
cognition. Therefore, only if a meaning is coextensive with a
word, and in addition one cannot account for the presence of the
meaning otherwise, should the meaning be considered the
denoted meaning of the word.
The procedure of concomitant presence and absence has
been discussed in relation to Patafijali's discussion of the
denotation of generic terms (?1.3.1.1). Kumarila emphasizes the
role that the means of knowledge known as presumption ought to
play in determining the denoted meaning. He states outright that
50Kumarila enumerates seven meanings which occur with the use of a class
term: Gaur ity uccarite sapta vastuni pratibhdnti nah: Jdtir vyaktis ca
sambandhah samuho lingakdrake; Samkhyd ca saptamT. When the word 'cow' is
uttered seven objects appear to us: The generic property, an individual, their
relation, the aggregate (of those three), gender and participant in the action, and
number is the seventh. MD, part 2, p. 256, lines 18-20.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
although two or more meanings appear with the use of a word

the word denotes only the one some other means of knowledg
does not supply:
Sarvatraiva tdvac chabdaid uccaritdd aneko 'rtho gamyate. Na ca tdvatd

sarva evdbhidheyo bhavati, tadbhdvabhdvitvdnyathdnupapatty-
adhinatvdd vdcakasaktikalpandydih. Sa ndmartho 'bhidhtyate, yatr
sabdasya vdcakasaktih kalpyate. Kva cdsdv ad4std sati kalpyate?
Yatra tatkalpandm antarena pratttiprayogau ndvakalpete. Yatra t
anyavisayayaiva saktyd kalpitayd tathdnyatrdpi tau sidhyatas,
tatrdnupalabdhyd saktyabhdvah pratlyate.
One always understands many meanings from an uttered speech
form. That doesn't imply that they are all its denoted meanings
because positing the signifying capacity depends on not being able
otherwise to account for the occurrence of the meaning when the
speech form is used. The denoted meaning is the (meaning) in which
one posits the speech form's signifying capacity. In which (meaning
does one posit (the speech form's signifying capacity), since we can
see it? (In the meaning) in which cognition and usage would not
succeed without positing it. But one knows that a word does no
have the capacity to denote a certain meaning, by means of non
perception (anupalabdhi), if cognition and usage succeed in tha
meaning, and in the other meanings as well, by positing that th
word has the capacity to denote just another meaning.51
Three statements, shown symbolically in table 1, outline

the procedure Kumarila describes to determine a word-meaning
relation: In statement 1, "One always understands many meanin
from an uttered speech form," he recognizes that the cognition
many objects arises regularly with the use of a certain word. B
recognizing that the word does not necessarily denote all these
objects saying, "That doesn't imply that they are all its denote
meanings," he introduces two statements concerning how t
determine, by means of presumption and non-perception, whi
among these objects is denoted.
In statement 2, he states that one can presume that the
word denotes an object if one cannot account for the cognition
that object otherwise, "The denoted meaning is the (meaning) i
which ... cognition and usage would not succeed without positin
(the speech form's signifying capacity there)." Note that Kumar
says, "without positing," not "positing that it is not." The phras
"without positing," tells one not to do something; it does n
outline any positive procedure to follow. Kumarila intends that
there are a limited number of possible causes of the cognition.
One of them must cause the cognition. If the one in questi
does not, one of the others must. Conversely, if none of t
possible causes other than the one in question succeeds in
51MD, part 4, p. 338.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MIMAMSA 231
Table 1
Determining Denotation
1. (Ax)[Ksx -> (Em1...mn)(Kmlx &...& Kmnx)]
2. (Ai)[-Dsmi -- -i(Ax)Kmix] -> Dsmi
3. (Ai, Aj: i?j)[Dsmj -> (Ax)(Kmix & Kmjx)] -> -nDsmi
1. For all objects x, if one knows a speech form s for x then there are many
meanings m1, ..., mn such that one knows m1,..., and one knows mn.
2. For all meanings mi of an object x, if it is true that if it were not the case
that the speech form denoted mi then it would not be the case that for all x one
knows mi of x, then s denotes mi.
3. For any mi and mj, if it is the case that if the speech form denotes mj then
for all x one knows mi of x and one knows mj of x, then it is not the case that
the speech form denotes mi.
bringing about cognition, the one in question must. Hence the

procedure is the following: Assume that it is not the case that s
denotes mi. If it is not the case that for all x one knows mi of x,
then the assumption is false. One concludes that s denotes mi.
Kumarila considers non-perception of an expected object
(anupalabdhi) to be the means of knowledge (pramdna) whereby
one determines the absence of that object. One can presume that
a word bears a capacity to denote only those meanings of which
cognition would not arise unless they were denoted. Non-
perception testifies that any meanings whose cognition can be
accounted for otherwise are not denoted. If one of the other
possible causes succeeds in accounting for the cognition o
meaning then one cannot presume that the word does. In
statement 3, he says, "But one knows that a word does not have
the capacity to denote a certain meaning, by means of non-
perception (anupalabdhi), if cognition and usage succeed in that
meaning, and in the other meanings as well, by positing that the
word has the capacity to denote just another meaning." The
procedure is to assume that s denotes mj. If it follows that for all
x one knows mi and one knows mj, then it is not the case that s
denotes mi.
Table two clarifies the relationships between hearing a
generic term and knowing, by any means whatsoever, the class
property and individuals to which the generic term refers. It
shows the eight possible combinations of the three cognitions,
knowledge of generic term, knowledge of the class property, and
knowledge of an individual.
Rows 1 and 8 show the normal situations, row 1 where a
word is used and we perceive both the class property and an
individual, row 8 where a word is not used and we don't perceive
either. If only the situations represented in these two rows in the
table occurred, concomitant presence and absence would establish
that the word denoted both the class property and the individual.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Table 2
Cognition Data
Kwx Kcx Kix Circumstance
1. T T T Normal use.
2. F T T See individual and recognize its class property.
3. T F T First hear word and see unknown object.
4. F F T One asks, "What is this?"
5. T T F The example of the falcon and other images.
6. F T F See image.
7. T F F Don't know or forget what the word means.
8. F F F No knowledge
'Kyx' represents the general statement: One knows y related t
just three instances of y: the generic term w, the class p
individual i. 'T' in a column represents the truth, and 'F' th
statement at the head of the column: He knows the word w
knows the class property c of x (Kcx), or he knows an individ
i of x (Kix). A separate variable x serves as an index var
association of the three related objects, the word, the class
individual. It stands for the class of objects associated
Although Sabara does not mention any such thing as a class
class property, I speak of a class for convenience in the pres
example, if we speak of the falcon class, f (falcon) instantia
statements Kwx, Kcx, and Kix to yield the following three s
1) One knows the word 'falcon' of the class of falcons (K
2) One knows the class property falconhood of the cl
(Kcf), and
3) One knows an individual falcon of the class of falcons (Kif).
The right hand column of the table describes a circumstance exemplifying the
case the first three columns represent.
Row 2 shows the presence of cognition of the class

property and individual in the absence of the word (IIIE3a). Row
7 shows the absence of cognition of the class property and the
individual in the presence of the word (IIIE3b). The cases
represented in these rows show the lack of concomitant presence
and absence between the word and either the class property or the
individual. However, the knowledge of the class property and
individual in row 2 derives from perception, and row 7 shows the
presence (not the knowledge) of the word.
Row 3 shows the presence of the word and the individual
in the absence of cognition of the class property. This occurs
only when one has not grasped the word-meaning relation and
has not recognized the general type of the object. An elder may
point to a cow and say, "This is a cow," to a child. The
knowledge of the individual is directly perceived, and the child is
not a full-fledged speaker of the language because he does not
know the word-meaning relation. Row 6, showing the absence of
hearing the word and the absence of the cognition of the
individual on the cognition of the class property, occurs only

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MiMAMSA 233
when one knows a class property through perception or some

means of knowledge based on perception.
Row 4 shows the absence of the word and the class
property in the presence of the cognition of the individual. R
5 shows the presence of the word and the class property in t
absence of the cognition of the individual. The situation in row
occurs only to one who directly perceives an individual a
neither knows what type of thing it is nor what it is called. Ro
represents cases like the example of the falcon. Sabara gives t
example in order to establish the concomitant presence o
knowing the word and knowledge of the class property and a
counterexample disproving the concomitant presence of know
the word and the verbal cognition of the individual (IIIG4a).
The discussions concerning word and meaning with
which we are dealing assume a number of limitations. Rat
than complicate the relational statements by including th
assumptions in them explicitly, we will state the assumptions
advance as background and restrict our discussions to the dom
they frame. First of all, although knowledge is properly a relat
between two objects, a knower and a known, we use the predic
K with a variable representing the known in relation to an in
variable x. We do not include a variable for the knower because
the knower is constant in our discussion. He is any competent
speaker of the language in which w is a word so that he actively
knows w and the word-meaning relation. The result is that
knowing a word w does not mean just hearing it but rather
knowing w and the word-meaning relation which is located in w
and has w as one of its relata.52 In some instances Sabara and
Kumarila state that something is present or absent, particularly
52According to later refinements, the word-meaning relation participates in

bringing about all verbal knowledge. First one lears the relation which is
stored inactively in memory. Hearing the word arouses it or causes one to
remember it. Then knowledge of the word together with the word-meaning
relation brings about the cognition of the object. For example, Nagesa in his
chapter describing the primary word-meaning relation (Saktinirupana) says:
Taddharmavacchinnavisayakas'bdabuddhitvdvacchinnam (jndnam) prati
taddharmdvacchinnanirupitavrttivijisitajdnaam hetuh. Ata eva ndgrhftavrtti-
kasya sibdabodhah. (PLM, p. 17.) Knowledge of the word qualified by the
word-meaning relation described with respect to the object qualified by a
certain property x is the cause of verbal knowledge of an object qualified by x.
Hence, one who has not comprehended the word-meaning relation does not have
verbal cognition. He adds: Jiidne vrttivaisistyam ca svavisayakodbuddha-
samskarasdmdnddhikaranyasvdsrayapadavisayakatvobhayasambandhena
bodhyam. Ato nagrhitavyrttikasya ndpi tatpadam ajdnato... bodhah. (PLM, p.
22.) The knowledge (of the word) is qualified by the word-meaning relation in
two ways: 1) it occurs in the same (human) substratum as the aroused memory of
the word-meaning relation, 2) it has as its object the word which is the
substratum of the word-meaning relation. Hence (verbal) cognition does not
arise for one who has not comprehended the word-meaning relation or for one
who does not perceive the word....

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
when discussing concomitant presence and absence, meaning th

a competent speaker knows it as we have just described. In such
cases, we restate their statements directly in terms of knowledge of
the thing.53
Except for the cognition of a word, the knowledge under
discussion rarely derives from direct perception. It derives either
directly from words or indirectly from words via the word
meanings. Sabara's discussion assumes that this is understood.
Hence, we assume the statement, "For all y, if it is not the case that
y is a word then it is not the case that one knows y by means of
direct perception, or by means of inference or comparison based
on direct perception {(Ay)(--Wy -> -,Py)}." In the rare case that
it is otherwise, this will be explicitly stated. Because denotation is
a direct causal relation between knowledge of the word and
knowledge of the denoted meaning, the statement that the word
denotes the meaning implies that if one knows the word then one
knows the meaning. In other words if w denotes a meaning m
then, for all x, if one knows w of x then one knows m of x
{(Am)[Dwm -> (Ax)(Kwx -> Kmx)]}. This is stated as an
assumption in tables 4-7, 10 and 11, line 1. Finally, Sabara and
Kumarila's discussions are not vacuous. They speak of ordinary
words of a language known to competent speakers of the
language. Hence we assume the statement, "For all x, one knows
the word for x {(Ax)(Kwx)}." This is stated as an assumption in
tables 4-7, 10 and 11, line 2.
Table 3
Verbal Cognition Data
Kwx Kcx Kix Circumstance
1. T T T Normal use.
2. F F F Kcx & Kix assume Pc & -,Pi.
3. F F F Kwx assumes knowing relation; Kix assumes -,Pi.
4. F F F Kix assumes -Pi.
5. T T F The example of the falcon and other images.
6. F F F Kcx assumes --Pc.
7. F F F Kwx assumes knowing word-meaning relation.
8. F F F No knowledge
Table 3 shows the truth or falsity of the three

shown in table 2 given the assumptions we have just s
1, 5 and 8 remain the same. Our discussions assume that
knowledge of a word includes active knowledge of the word
meaning-relation located in it and having it as one of its relata. In
the circumstances shown in rows 3 and 7 of table 2, while it is
53Indian philosophy recognizes two types of causes: that which causes simply
by being (satsatkdrana) and that which causes by being known (jhanasatkdrana).
We are dealing strictly with the latter.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MiMAMSA 235
true that one hears a word, one does not know it. Hence those
rows under column 1 of table 3 bear an 'F' showing that the
statement, "He knows a word," is false. In addition, we have
assumed that knowledge of objects other than words does not
derive from perception unless stated so. In the circumstances
shown in rows 2-4 and 6 of table 2, although one may know the
class property or individual by direct perception or some means
based on it, one cannot know them independently of perception;
it is not the case that all of the knowledge in the situation derives
either directly or indirectly from speech. Hence the 'T's in rows 2
and 6 under column 2, and in rows 2-4 under column 3 showing
that the statement, "He knows x," is true have been changed to 'F's
in table 3 showing that the statement is false given the assumption.
Surveying the observed facts depicted in table 3, we see
two types of situations arising upon a competent speaker's
cognition of a generic term. There are the ordinary cases such as
the example of rice where knowledge of the word 'rice' is
followed by knowledge of the class property riceness and
knowledge of individual grains of rice {Kwr & (Kcr & Kir)}.
There are also cases such as the example of the falcon where
knowledge of the word 'falcon' is followed by knowledge of the
class property falconhood, but not by knowledge of an individual
falcon {Kwf & (Kcf & -,Kif)}. There are no cases which are not
of either one of these types. In order to account for these facts,
one wishes to explain the correspondence in terms of a cause and
effect relation among these three elements: the word, the class
property and an individual. Assuming that the other two, as
meanings, are causally derivative from the knowledge of the
word, given the assumption of knowledge of each word one must
account for knowledge of each class property and in most cases
knowledge of an individual.
Excepting the cases like the falcon example, the
correspondences are uniform: knowledge of the word is followed
by knowledge of the class property and knowledge of an
individual {(Ax)[Kwx & (Kcx & Kix)]}. Given the assumption
of knowledge of any word {(Ax)Kwx} one must account for
knowledge of the class property and knowledge of an individual
{(Ax)(Kcx & Kix)}. The simplest solution is to assume that the
word denotes both the class property and the individual {(Dwc &
Dwi)}. Given that one knows any word and that to denote
implies that if one knows the word then one knows the denoted
meaning, if the word denotes both the class property and the
individual, it follows that one knows both the class property and
the individual. Table 4 shows this argument as a symbolic logic
derivation. The final conclusions in grammar and Nyaya concur
with this solution.
In his argument to determine whether a word denotes a
class property or an individual, Sabara acknowledges that

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Table 4
A generic term denotes
both the class property and the individual.
1. (Am)[Dwm -> (Ax)(Kwx -> Kmx)]
2. (Ax)Kwx
3. Dwc & Dwi
4. Dwc & Dwi (R, 3)
5. Dwc (&E, 4)
6. Dwi (&E, 4)
7. (Am)[Dwm -> (Ax)(Kwx -> Kmx)] (R, 1)
8. Dwc - (Ax)(Kwx - Kcx) (AE, 7)
9. Dwi -> (Ax)(Kwx - Kix) (AE, 7)
10. (Ax)(Kwx - Kcx) (->E, 5,8)
11. (Ax)(Kwx -> Kix) (->E, 6,9)
12. Kwa -> Kca (AE, 10)
13. Kwa -Kia (AE, 11)
14. (Ax)Kwx (R, 2)
15. Kwa (AE, 14)
16. Kca (->E, 12,15)
17. Kia (->E, 13,15)
18. Kca&Kia (&I, 16,17)
19. (Ax)(Kcx & Kix) (AI, 18)
20. (Dwc & Dwi) -> (Ax)(Kcx & Kix) (-I, 3-19)
cognition of both follows the use of a word. At the begi

the Akrtyadhikarana, asked why the question arises, the
proponent answers, "Because when the word 'cow' is spoken, one
understands a general property (IB1)." While arguing against the
view that a word denotes both the class property and an
individual, the proponent says, "It is self-evident that one
apprehends the individual when a word is uttered (IIIE2a)."
Although Kumarila acknowledges that cognition of other
meanings occurs as well, the current discussion does not consider
them. The present argument considers only two meanings, the
class property and an individual, and assumes that a generic term
denotes at least one of the two. We state this as the following
assumption in table 7, line 4 and table 11, line 5: w denotes c or w
denotes i {Dwc V Dwi}.
Sabara argues that the word does not denote the
individual because concomitant presence and absence proves that
one can account for the cognition of the individual from the class
property. Immediately following the last statement quoted he
continues as follows:
But the distinction, whether one apprehends it from the word or from
the class property, is not evident. That is determined by means of
concomitant presence and absence (IIE2b-3).

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MiMAMSA 237
Then he makes the following two statements intending to establish

that knowledge of the class property rather than of the word is the
direct cause of knowledge of an individual:
a) Even without the use of the word, whoever apprehends a class

property, certainly apprehends an individual (IIIE3a). b) But even
upon hearing the word, whoever should fail to apprehend the class
property at some time because his mind is dull, would certainly not
understand that individual (IIIE3b).
These statements assert the following: a) Whether one has

cognition of the word or not, if one knows the class property then
one knows an individual. {(Ax)[(Kwx V -,Kwx) -> (Kcx ->
Kix)]}. b) Whether one has cognition of the word or not, if one
does not know the class property then one does not know an
individual. { (Ax)[(Kwx V -Kwx) -> (-Kcx -> -,Kix)] }. By the
law of the excluded middle (p V -ip), the statements imply the
concomitant presence and absence of knowledge of the class
property and knowledge of an individual {(Ax)[(Kcx -> Kix) &
(-,Kcx --> Kix)] }.
The phrases, "even without the use of the word" and "even
upon hearing the word," show that in this case Sabara puts aside
the assumption that the cognition of the object is not caused by
perception or a means of knowledge dependent on perception.
He states that there is concomitant presence and absence of
knowledge of the class property and knowledge of an individual,
regardless of the means of that knowledge (See table 2 rows 2 and
7; the present statement ignores rows 3-6 in which these two
cognitions are not concurrent). The phrase, "because his mind is
dull," in statement b sets aside the assumption that knowledge of
the word includes active knowledge of the word-meaning
relation. Therefore, although the statements assert the presence of
the word with the absence of both meanings, and the absence of
the word with the presence of both meanings, they do not
contradict the concomitant presence (and absence) of knowledge
of the word and knowledge of its meanings. Knowledge of a
word and knowledge of its meanings are only (and need only be)
coextensive insofar as the knowledge of the meanings arises from
speech. Because they set aside the assumptions, Sabara's two
statements (IIIE3a and IIIE3b) do not deny the coextension of
knowledge of the word, as including active knowledge of its
signifying capacity, and verbal knowledge of its meanings.
Sabara's introduction to his statement of the concomitant
presence and absence of the class property and an individual
show that he intends the known class property to be the direct
cause of knowledge of an individual. He intends the statements to
show that one knows the individual "from the class property
(IIIE2b)." From his preceding statement it appears that he intends

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
the direct causation to be mutual due to the inseparable relation

the class property and an individual. He says, "...the class
property is inseparably related to the individual, and once one
knows that relatum one knows the other relatum (IIIE1)."
Kumarila interprets the objector's argument that the word denotes
only the individual to rely on this mutual causation, and interprets
Jaimini's word 'because there is no division (avibhdgat)' as
implying this mutual causation. Kumarila comments as follows:
To answer the objection, "Then how does one have the cognition of
the general property?" (Jaimini) says, "Because there is no division."
Because an individual and its class property are eternally inseparable,
the individual being denoted can convey the understanding of the
class property. (IIC note 5)
When the objector states that the view that a word denotes
the individual and the view that a word denotes the class property
are equivalent, he shows that he understands the inseparability of
the class property and individual to imply that knowledge of
either one causes knowledge of the other. He states that, on the
view that the word denotes only the class property, the class
property serves to make known the individual and, on the view
that the word denotes only the individual, the individual serves to
make known the class property (IIIB2). These statements
interpret the concomitant presence and absence of knowledge of
the class property and knowledge of an individual to be
equivalent to their biconditional {(Ax)(Kix - Kcx)}.54 This
statement appears as an assumption in tables 5 and 6, line 3.
The aim of Sabara's argument is to show that, because one
can account for the knowledge of the individual from the
knowledge of the class property, one cannot presume that the
word denotes the individual. This is exactly a specific form of the
general statement of presumption included in Kumarila's
procedure discussed above (table 1, statement 3). This statement
appears as an assumption on line 4 of table 5 which shows
Sabara's argument as a derivation in symbolic logic.55 The
paragraph below the table summarizes the argument.
54Modern deductive logic holds the biconditional to be equivalent to the

conjunction of the conditional and negative conditional {[(p -> q) & (-ip ->
-q)] = (p - q)}. However, this equivalence relies on the intelim rules of negative
introduction and elimination which are equivalent to the Mimmmsa means of
knowledge by which the relation of cause and effect is established,
presumption. Presumption is the weakest means of knowledge.
55Read: If it is true that, by assuming that the word denotes only the class
property c (and not an individual i), for all classes x one knows c of x and i of x,
then it is not the case that the word denotes i.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MIMAMSA 239
Table 5
One cannot presume that a word denotes an individual.
2. (Ax)Kwx
3. (Ax)(Kix - Kcx)
4. JDwc -> (Ax)(Kcx & Kix)] -> -Dwi
5. Dwc
6. (Am)[Dwm -> (Ax)(Kwx -> Kmx)] (R, 1)
7. Dwc -> (Ax)(Kwx -> Kcx) (AE, 6)
8. Dwc (R, 5)
9. (Ax)(Kwx -> Kcx) (->E, 7,8)
10. Kwa ->Kca (AE, 9)
11. (Ax)Kwx (R, 2)
12. Kwa (AE, 11)
13. Kca (->E, 10,12)
14. (Ax)(Kix Kcx) (R, 3)
15. Kia -Kca (AE, 14)
16. Kia (-E, 13,15)
17. Kca&Kia (&I, 13,16)
18. (Ax)(Kcx & Kix) (AI, 17)
19. Dwc - (Ax)(Kcx & Kix) (->I, 5-18)
20. [Dwc -> (Ax)(Kcx & Kix)] - -,Dwi (R, 4)
21. -,Dwi (->E, 19,20)
A word conveys cognition of b
need denote only one of the tw
class property the cognition of
still arise, then we cannot presu
one knows the class property, o
of the word. Therefore, if the w
will cause cognition of the cla
cognition of an individual. Cog
arises. Therefore, we cannot p
Since no other means of knowl
we conclude the word does not denote an individual.
The proponent of the view that the word denotes only the
individual uses the same general procedure to show that one
cannot presume that the word denotes the class property (IIB,
IIIE1). He makes the following assumption corresponding to the
assumption on line 4 of table 5: If it is true that if the word
denotes only an individual i (and not the class property c) then
for all classes x one knows c of x and i of x, then it is not the case
that the word denotes c. This statement appears as an assumption
on line 4 of table 6 which shows his argument as a derivation in
symbolic logic.
Given exactly the same assumptions (tables 5 and 6, lines
1-3, and line 4 which, in each of tables 5 and 6, is an instance of

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Table 1, line 3), by assuming that a word denotes a class propert

(table 5, line 5) or by assuming that a word denotes an individu
(table 6, line 5), one can account for the cognition of both t
class property and the individual (tables 5 and 6 line 18). Given
just the assumptions by which Sabara concludes that a word do
not denote an individual (table 5 line 21), the objector conclude
that a word does not denote a class property (table 6 line 2
Hence Sabara's second argument against the view that the word
denotes both the class property and the individual could just as
well establish that a word denotes an individual as it can establish
that a word denotes only the class property.
In order to show that the view that a word denotes only a
class property surpasses the view that a word denotes only an
individual, Sabara explains the significance of the example of the
falcon. The objector had argued that the class property can mark
the individuals which the word denotes without being denoted
itself just as a staff marks a staff-bearer without being denoted by
the word 'staff (IIC). Generic terms can denote only the
individuals having the class property just as the term 'staff-bearer'
denotes only the person bearing the staff. Sabara refutes this as
follows:
...it is not the case that just as the word 'staff-bearer' is not used for
the staff, the word 'cow' is not (used) for the class property. For that
very reason we showed the example of the word 'falcon' which
denotes only a class property. Because the word 'falcon' is uttered
when there is no connection with an individual falcon, by
concomitant presence and absence one knows that it denotes a class
property (IIIG4a).
The argument approximately follows the second part of the

general procedure Kumarila describes. Kumarila says that the
denoted meaning is the meaning of which cognition would not
occur without positing that the word denote it (table 1, line 2).
Table 7, headed by an instance of table 1 line 2, shows the
argument symbolically; the paragraph below summarizes it. The
example of the falcon altar is central to this argument. It is an
instance in which a generic term does not give knowledge of an
individual (assumption on line 3). Hence it serves as a
counterexample to the statemeiit, "If one knows the word w for
then one knows an individual i of x," disproving the concomitant
presence and absence of knowledge of w and knowledge of i.
Since one knows the generic term while not knowing an
individual, due to the lack of concomitant presence and absence,
it cannot be the case that a generic term denotes an individual.
If the individual is not known from the word, the
inseparable connection of the individual and its class property
cannot serve as a way to knowledge of the class property. In fact,

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MiMAMSA 241
Table 6
One cannot presume that a word denotes the class property
2. (Ax)Kwx
3. (Ax)(Kix - Kcx)
4. [Dwi - (Ax)(Kcx & Kix)] - --Dw
5. Dwi
6. (Am)[Dwm -> (Ax)(Kwx -> Kmx)] (R, 1)
7. Dwi -> (Ax)(Kwx -> Kix) (AE, 6)
8. Dwi (R, 5)
9. (Ax)(Kwx -> Kix) (->E, 7,8)
10. Kwa - Kia (AE, 9)
11. (Ax)Kwx (R, 2)
12. Kwa (AE, 11)
13. Kia (-->E, 10,12)
14. (Ax)(Kix - Kcx) (R, 3)
15. Kia - Kca (AE, 14)
16. Kca (=E, 13,15)
17. Kca & Kia (&I, 13,16)
18. (Ax)(Kcx & Kix) (AI, 17)
19. Dwi -> (Ax)(Kcx & Kix) (-I, 5-18)
20. [Dwi -> (Ax)(Kcx & Kix)] - -Dwc (R, 4)
21. -,Dwc (--E, 19,20)
the inability of the word to give knowledge of the individual

contradicts the assumption that a word denotes an individual. It
directly contradicts the assumption because the fact that a word
denotes an object implies concomitant presence and absence of
knowledge of the word and knowledge of the object. Therefore,
the assumption must be rejected. It cannot be the case that a word
denotes only an individual and not the class property; it must
denote the class property.
Sabara's arguments attempt to establish logical precedence
by means of the concomitant presence and absence of different
elements. In order to show that knowledge of the word directly
causes the cognition of the class property, and cognition of the
class property directly causes knowledge of an individual; Sabara
must show that knowledge of the word and class property are
coextensive, knowledge of the class property and knowledge of
the individual are coextensive independent of the use of a word,
and knowledge of the word and knowledge of an individual are
not coextensive. However, his counterexample to the coextension
of knowledge of the word and knowledge of an individual is also
a counterexample to the coextension of knowledge of the class
property and knowledge of an individual.
The example of the falcon is absolutely crucial to showing
that a generic term denotes only the class property. However, by

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
citing the example of the falcon, Sabara shows an instance

which one knows the class property without knowing any
individual. This fact is inconsistent with the assumption he ma
to show that, if a word denotes only a class property, one c
account for knowledge of both the class property and an
individual. That assumption was the following: if one knows a
class property then one knows an individual.
One proves that statements are mutually inconsistent by
deriving a contradiction from them. Table 8 derives a
contradiction from the following two assumptions: 1) If one
knows the class property of a class x then one knows an
individual of x. 2) One knows the falcon class property and it is
not the case that one knows an individual falcon.
By setting forth the example of the falcon (IIIA) and
stating that the reason the example was set forth was to show that
the word denotes the class property even when there is no
cognition of an individual (IIIG4a), Sabara implies that it is not
true that knowledge of either an individual or the class property
regularly gives knowledge of the other. It contradicts the
statement that whoever apprehends the class property certainly
apprehends an individual (IIIE3a). If the word 'falcon' refers
only to the class property without any individual, the statement
that if one knows the class property then one knows an individual
is not true, the knowledge of one and the knowledge of the other
are not inseparably connected, and one cannot count on knowing
an individual merely from knowing the class property (or vice
versa). Hence, if Sabara's statements concerning the concomitant
presence and absence of knowledge of the class property and
knowledge of an individual meant that the knowledge of one
mutually causes knowledge of the other, his third argument
would contradict his second. The example of the falcon
establishing coextension of cognition of a word and of the class
property and refuting coextension of cognition of the word and
of the individual would contradict the coextension of knowledge
of an individual and knowledge of its class property.
However, the inseparability of an individual and a class
property does not imply that knowledge of a class property
always provides knowledge of an individual and vice versa. It
means only that it is the nature of a property to reside in an
individual. And as a matter of fact the explanations which Sabara
himself gives as to how one arrives at the understanding of the
qualified individual in the examples of washing rice, number, and
replacement under Mimamsasutra 1.3.35 do not require that
knowledge of either a class property or an individual be a direct
cause of knowledge of the other. They require only that the class
property and the individual be connected.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MiMAMSA 243
Table 7
(--Dwc -> -_(Ax)Kcx) -> Dwc
\ I . - i
(AE. table 1, stmt. 2)
. . -

2. (Ax)Kwx
3. -,Kif
4. Dwc V Dwi
5. -Dwc
6. Dwc V Dwi (R, 4)
7. Dwc
8.
9. -Dwc (R, 5)
10. Dwc (R, 7)
11. -,-,Dwi (--, 8-(9,10))
12. Dwi (-E, 11)
13. Dwi
14. Dwi (R, 13)
15. Dwi (VE, 6,7-12,13-14)
16. (Am)[Dwm -> (Ax)(Kwx -> Kmx)] (R, 1)
17. Dwi -> (Ax)(Kwx -> Kix) (AE, 16)
18. (Ax)(Kwx -> Kix) (->E, 15,17)
19. Kwf -> Kif (AE, 18)
20. (Ax)Kwx (R, 2)
21. Kwf (AE, 20)
22. Kif (->E, 19,21)
23. -,Kif (R, 3)
24. ---Dwc (-1, 5-(22,23))
25. Dwc --. - - x '-,) -
(--E. 24)
-/~~~~~~~~~~~~~~
Suppose that a word did not denote the class property. Then it would denote
an individual. But it can't denote an individual because if it did it could not
avoid giving knowledge of an individual whenever the word was used. For
instance, whenever the word 'falcon' was used it would give knowledge of an
individual falcon. Yet in the example of the falcon altar the word 'falcon' is
used and knowledge of an individual falcon does not arise. Hence, a word does
not denote an individual, and the supposition that it is not the case that a word
denotes a class property is false. Therefore, a word denotes a class property.
Table 8
Inconsistency of inseparability and falcon example
whether inseparability be (p - q) or [(p -> q) & (--p -> --q)]
1. (Ax)(Kcx -> Kix)
2. Kcf & -Kif
3. (Ax)(Kcx - Kix) (R, 1)
4. Kcf - Kif (AE, 3)
5. Kcf & -Kif (R, 2)
6. Kcf (&E, 5)
7. Kif (->E, 4,6)
8. -,Kif (&E, 5)

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
In the example of washing rice, for example, Sabara says

that the word 'rice' denotes the class property. The Vedic tex
mentions washing as an action to be performed on a substance;
not on a class property because it is impossible to wash a clas
property. The second triplet termination on the word 'rice
(vrihin)' indicates that the meaning of the word 'rice' is that upon
which the action of washing is to be performed. Hence the verb
'to wash' (proksati, literally 'sprinkle') makes known a substance i
general. This general knowledge is the knowledge of the
substance without knowing what type of substance. The
substance in this case happens to be rice. The general knowledge
of it is knowing it simply as the thing to be washed. The second
triplet termination on the word 'rice (vrihTn)' brings that general
knowledge of a substance into relation to the object denoted by
the nominal base it follows. The rice class property which the
word 'rice' makes known is understood to specify this general
knowledge of a substance. The class property as qualifier
together with the general knowledge of the substance give
specific knowledge of the rice substance. This specific
knowledge has the form of the knowledge of the substance to be
washed as qualified by the class property riceness, or more
simply, the knowledge of the rice as rice.
In general, the verb in a sentence together with the
nominal termination on a generic term give general knowledge of
an individual and bring that general knowledge into relation to
the class property which the nominal base supplies. Other speech
forms such as a compound or taddhita affix, instead of a nominal
termination along with a verb, may supply the general knowledge
of the individual and its relation to the object denoted by the base.
General knowledge of an individual which happens to belong to
class x {Gix} is knowledge of the individual merely as an
individual substance without specification. If one has general
knowledge of an individual which happens to belong to class x,
and one has this knowledge in relation to the knowledge of the
class property of x, then one has specific knowledge of an
individual of class x {(Ax)[(Kcx & Gix) - Kix]}.
In some cases the action which the verb denotes may be
impossible on an individual of class x. In such cases general
knowledge of an individual in relation to x will not arise {(-Gix}.
If one does not have general knowledge of an individual of x
then one does not have specific knowledge of an individual of x
{(Ax)(-Gix -- --Kix)}. For instance, in the example of the falcon
altar, the verb 'to build (Vci)' implies that one use bricks as
building blocks. The compound 'x-altar (x-cit)' indicates that x is
that which is most desired to be brought about by the act of
building. The verb and compound therefore give general
knowledge of a brick individual of x. Now instantiate x with
falcon. The compound 'falcon-altar (syena-cit)' indicates that the

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MIMAMSA 245
meaning of the word 'falcon' is that which is most desired to b

brought about by the act of building. But it is not the case that
brick individual is the meaning of the word 'falcon'. The genera
knowledge of a brick individual does not make sense in relation
to the meaning of the word 'falcon' so there is no general
knowledge of an individual falcon {-,Gif}. Even knowing th
falcon class property, without general knowledge of an individua
falcon, specific knowledge of an individual falcon does not arise
{_-,Gif ->-,Kif } (IIIA).
In addition to contradicting the example of the falcon
(table 8; tables 2 and 3, line 5), interpreting Sabara's statements
concerning the concomitance of knowing a class property an
knowing an individual as stating that knowledge of a class
property and knowledge of an individual of that class each caus
the other contradicts the examples in lines 3, 4 and 6 of table 2
Sabara certainly would not deny that the situations represented
rows 3-6 occur. The frequently encountered fact that people as
"What is this?" proves that one does know an individual withou
knowing its class property (line 4). If one had no knowledge of
at all, one would not be able to ask about it. On the other hand,
the question, "What is this?," asks for more information about an
individual of which one has some knowledge; it asks what the
individual is. Asking what the individual is asks its class property.
Hence if one does not know the class property of an individual
then one does not really know the individual fully.
Sabara has stated this concomitant absence of knowledge
of a class property and specific knowledge of an individual in
connection with the example of the new-born cow while arguing
against the view that the individual alone is the denoted meaning
of a word. He says, "No one can apprehend something qualified
unless the qualifier is apprehended (IIIC4b)." Kumarila expresses
this in the statement quoted above, namely, "Knowledge of the
qualified is unacceptable without knowing the qualifier." In both
statements, the qualifier is the class property. The qualified is the
individual as qualified by the class property. In these statements,
both Sabara and Kumarila express that one does not know the
individual as qualified by its class property unless one knows the
class property. Knowing an individual as qualified by its class
property may be called specific knowledge of the individual.
Hence Sabara and Kumarila's statements may be restated more
formally in this way: For all classes x, if one does not know the
class property of x then one does not have specific knowledge of
an individual of class x ( (Ax)(-,Kcx -> -,Kix)}.
Collecting our evaluations of Sabara's treatment of the
example of washing rice under sutra 1.3.35, his description of the
falcon example and his statement concerning the example of the
new-born cow that one must know the qualifier to know the
qualified, we conclude that Sabara accepts that there are two

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
degrees of knowing an individual: 1) One may know that it is a

thing, an individual of some type or other. 2) One may kno
specifically what it is, that is, know the individual as being of su
and such a type. The former we call general knowledge of t
individual; the latter, specific knowledge or knowledge of
qualified individual. Specific knowledge of an individual
includes both general knowledge of the individual and
knowledge of its class property. Hence when Sabara speaks of
knowledge of an individual we must determine whether he
intends general or specific knowledge. 'Gix' represents general
knowledge of an individual which happens to belong to a class x,
and 'Kix' represents specific knowledge of an individual of class
x, that is, knowledge of an individual as qualified by the class
property defining class x. The statements summarizing our
evaluation of Sabara's discussion of the examples of washing rice,
the falcon altar and the new-born cow are equivalent to the
following statement: One has specific knowledge of an individual
if and only if one knows the class property of an individual of a
class x and has general knowledge of an individual of class x
{((Ax)[(Kcx & Gix) -> Kix] & [(Ax)(-,Gix - -,Kix) &
(Ax)(-,Kcx - ~Kix)]) = (Ax)[Kix (Kcx & Gix)]}.
Now let us reexamine Sabara's statement of the
concomitant presence and absence of knowledge of a class
property and knowledge of an individual. In this statement when
he speaks of knowledge of an individual he intends specific
knowledge of an individual. The first part stating concomitant
presence is clearly contradictory not only to the falcon example
but also to the perception of, say, a model cow (table 2, lines 5-
6). Knowledge of a class property by itself does not give specific
knowledge of an individual of that class; knowledge of the class
property together with general knowledge of an individual does.
Therefore, although Sabara says, "whoever apprehends a class
property, certainly apprehends an individual (IIIE3a)," in order
not to contradict himself he must be understood to mean the
following: if one knows the class property of class x, in additio
to having general knowledge of an individual i which happens t
belong to class x, then one has specific knowledge of an i of
{(Ax)[(Gix & Kcx) -* Kix]}.
On our former interpretation, the second part of Sabara's
statement which states concomitant absence was inconsistent with
the cases exhibited in rows 3 and 4 of table 2. The knowledge of
an individual which one has in these cases is general not specific
One does not know its class property which is one constituen
element of specific knowledge of an individual; one knows a
individual merely as a thing. If Sabara's statement, "whoever
should fail to apprehend the class property...would certainly not
understand that individual (IIIE3b)," concerns specific knowledge
of an individual, it is perfectly consistent with lines 3 and 4 of

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MIMAMSA 247
table 2. Sabara does not say, "If one does not know a class
property then one does not know an individual at all;" rather, he
says, "one does not know that individual." He uses the
demonstrative pronoun 'that (imam)' to specify the individual.
The specification intended is the class property of the individual.
Hence this part of his statement repeats the statement that one does
not know the qualified object without knowing the qualifier.
Since knowing the qualified object consists in knowing both the
qualifier and the object merely as a thing, the statement is a
tautology: If it is not the case that one knows the class property
then it is not the case that one knows the individual generally and
one knows its class property {(Ax)[-,Kcx --> -(Gix & Kcx)] }.
We have reinterpreted Sabara's statement of concomitant
presence and absence, "whoever apprehends a class property,
certainly apprehends an individual, ...but...whoever should fail to
apprehend the class property...would certainly not understand that
individual (IIIE3)," to mean the following: For all classes x, if
one knows the class property of x and has general knowledge of
an individual i which happens to belong to x, then one has
specific knowledge of an i of x; and if it is not the case that one
knows the class property c of a class x, then it is not the case that
one has specific knowledge of an individual i of x {(Ax)([(Gix &
Kcx) -> Kix] & (-,Kcx -> -_Kix))}. Clarifying Sabara's argument
concerning the falcon example in terms of general and specific
knowledge, we formulated it as the statement that in the absence
of general knowledge of an individual falcon specific knowledge
of an individual would not arise {--Gif -> -,Kif}. This is an
instance of the general statement that in the absence of general
knowledge of an individual of class x specific knowledge of a
individual of class x does not arise {(Ax)(-,Gix -> -IKix)}. If we
add this to the previous statement, we get the following
biconditional: For all x, one has specific knowledge of i of x if
and only if one knows c of x and has general knowledge of i of x
{(Ax)[Kix - (Kcx & Gix)] . This biconditional is consistent with
the example of the falcon. One shows that a set of statements is
consistent by showing one truth value assignment on which all the
statements of the set are true. Table 9 shows such an assignment.
Table 9
Consistency of revised inseparability and falcon example
1. Kcf & -Kif
2. Kif = (Kcf & Gif, an instance of (Ax)[Kix = (Kcx & Gix)]
1 2
Gi f K f K if Kcf & -Kif Kcf & Gi Kif (- cf & Gif)

F T F T T F T

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Our previous summary of the facts of usage notes two

types of cognitions arising upon a competent speaker's cognition
of a generic term: knowledge of both the class property and
individual or knowledge only of the class property I(Ax)[(Kcx &
Kix) V (Kcx & -iKix)]}. Without a factor other than knowledg
of the class property to distinguish when knowledge of an
individual arises following a word's use and when it doesn't, any
set of assumptions which accounted for cognition of both th
class property and the individual was unable to account for
knowledge only of the class property and the absence of
knowledge of an individual as in the example of the falcon altar.
However, from our reexamination of Sabara's concept of the
inseparability of class property and individual it is clear that
general knowledge of an individual serves as such a factor.
Assuming cognition of the word w for x, one knows the class
property c of x, and one knows an individual i of x in addition if
and only if one has general knowledge of an individual which
happens to belong to x {(Ax)[(Kcx & (Kix = Gix)] }.
Based on the initial impression of the inseparability of a
class property and an individual, one could formulate an
argument against the presumption that a word denotes the class
property (table 6) corresponding to the argument against the
presumption that it denotes an individual (table 5). With the
revised understanding of inseparability, a corresponding
argument does not succeed. Assuming the revised statement of
inseparability (tables 10 and 11, line 3), we restate Sabara's
arguments to prove that only a class property is denoted. The
first establishes that one cannot presume that a word denotes an
individual; the second, that one must presume that a word denotes
a class property because one cannot account for the cognition that
arises from a word otherwise. Tables 10 and 11 show these
arguments as derivations in symbolic logic and the parag
beneath describe them.
Table 10 Caption
If it is the case that, if the word denoted only the class property, the cognition
of the class property would arise, and, in addition, cognition of an individual
would arise if and only if one has general knowledge of an individual, then we
cannot presume that the individual is also denoted. Now, if the word denotes
only the class property, the word will cause cognition of the class property.
Look at the case in which one does arrive at specific knowledge of an
individual. Specific knowledge of an individual arises if and only if one has
knowledge of the class property and general knowledge of an individual. The
necessary condition to give specific knowledge of an individual is that one
have knowledge of both the class property and general knowledge of an
individual. The word has given knowledge of the class property. Hence the

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MiMAMSA 249
Table 10
One cannot presume that a word denotes an individual
(AE, table 1, stmt. 3)
2. (Ax)Kwx
3. (Ax)[Kix = (Kcx & Gix)]
4. (Dwc --> (A)[(Kcx & (Kix - Gix)]) -> --wi
5. Dwc
6. (Am)[Dwm -> (Ax)(Kwx -- Kmx)] (R, 1)
7. Dwc -> (Ax)(Kwx -> Kcx) (AE, 6)
8. Dwc (R,5)
9. (Ax)(Kwx -> Kcx) (->E, 7,8)
10. Kwa -> Kca (AE, 9)
11. (Ax)Kwx (R, 2)
12. Kwa (AE, 11)
13. Kca (--E, 10,12)
14. Kia
15. (Ax)[Kix -(Kcx & Gix)] (R, 3)
16. Kia (Kca & Gia) (AE, 15)
17. Kia (R, 14)
18. Kca & Gia (-E 16,17)
19. Gia (&E, 18)
20. Gia
21. (Ax)[Kix (Kcx & Gix)] (R, 3)
22. Kia - (Kca & Gia) (AE, 21)
23. Gia (R, 20)
24. Kca & Gia (&I, 13,23)
25. Kia (-E, 22,24)
26. Kia - Gia (=I, 14-19,20-25)
27. Kca & (Kia = Gia) (&I, 13,26)
28. (Ax)[(Kcx & (Kix Gix)] (AI, 27)
29. Dwc -> (Ax)[(Kcx & (Kix - Gix)] (->I, 5-28)
30. (Dwc -> (Ax)[(Kcx & (Kix - Gix)]) ->-,Dwi (R, 4)
31. -,Dwi ....~~~~~~~~~~~~~~~~~~~~~\. -I
(-4E. 29.30)
? ? y /
necessary condition will be fulfilled only if one has general knowledge of

individual. Now suppose that one does have general knowledge of
individual. The word has given knowledge of the class property, so one
knowledge of the class property and general knowledge of an individual. T
is a sufficient condition to generate specific knowledge of an individu
Hence, since the word has given knowledge of the class property, gen
knowledge of an individual is the necessary and sufficient condition to gi
knowledge of an individual. Therefore, if the word denotes only the c
property one will arrive at knowledge of the class property, and, in additi
one will have specific knowledge of an individual if and only if one has gen
knowledge of an individual. Therefore, one cannot presume that the w
denotes an individual.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Table 11
(-Dwc
x
-- -,(Ax)Kcx)
\ t- \.
- Dwc
- --7
(AE. table I..
7...
1. stmt.
....
2)
1. (Am)[Dwm - (Ax)(Kwx - Kmx)]
2. (Ax)Kwx
3. (Ax)[Kix - (Kcx & Gix)]
4. -Gif
5. Dwc V Dwi
6. -Dwc
7. Dwc V Dwi (R, 5)
8. Dwc
9. -wwi
10. -_Dwc (R, 6)
11. Dwc (R, 8)
12. -,-Dwi (-I, 9-(10,11))
13. Dwi (-E, 12)
14. Dwi
15. Dwi (R, 14)
16. Dwi (VE, 7,8-13,14-15)
17. (Am)[Dwm -> (Ax)(Kwx - Kmx)] (R, 1)
18. Dwi -> (Ax)(Kwx -> Kix) (AE, 17)
19. (Ax)(Kwx -- Kix) (--E, 16,18)
20. Kwf - Kif (AE, 19)
21. (Ax)Kwx (R, 2)
22. Kwf (AE, 21)
23. Kif (-<E, 20,22)
24. (Ax)[Kix (Kcx & Gix)] (R, 3)
25. Kif- (Kcf & Gif) (AE, 24)
26. Kcf & Gif (-E, 23,25)
27. Gif (&E, 26)
28. -,Gif (R, 4)
29. ---Dwc (-,I, 6-(27,28))
30. Dwc I
(-,E. 29)
Suppose that a word did not denote the class property. Then it would denote
an individual. But it can't denote an individual because, if it did, it could not
avoid giving specific knowledge of an individual whenever a word was used.
For instance, whenever the word 'falcon' was used it would give specific
knowledge of an individual falcon. Specific knowledge of an individual arises
only if one has knowledge of the class property and general knowledge of an
individual. This is true in the case of specific knowledge of an individual
falcon too. Hence every cognition of the word 'falcon' would necessarily be
accompanied by general knowledge of an individual falcon. Yet in the
example of the falcon altar, cognition of the word 'falcon' is not accompanied
by general knowledge of an individual falcon. Hence, a word does not denote
an individual, and the supposition that it is not the case that a word denotes a
class property is false. Therefore, a word denotes a class property.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MIMAMSA 251
The success of Sabara's ar

generic term denotes both
depends on instances such
Given such instances, a no
meaning relation will hav
cognition following the u
class property and indivi
property {(Ax)[(Kcx & Kix
the falcon, in which one kn
not know a falcon individual {Kcf & --Kif}, does not have a
balancing instance. Sabara states this fact clearly in concluding
his reason for introducing the falcon example. He says, "But one
never sees a word used for particular rice grains without a
connection with the rice class property. Therefore, (the view) that
a word denotes the class property is better (IIIG4b)." Denying
that there is an instance, in which upon use of the word 'rice' one
does not know the rice class property and yet one knows some
individual rice grains {-,[Kwr & (-,Kcr & Kir)] , intends the more
general statement that one never knows an individual and not a
class property upon the use of a generic term. Formally stated the
statement is this: It is not the case that there is a class x such that
one knows the word for x, and one knows an individual of x but
does not know the class property of x {-((Ex)[Kwx & (-,Kcx &
Kix)]}. Hence data are weighted in favor of the class property;
the cognition of the class property is coextensive with the use of
the word but cognition of an individual is not.
The procedure to determine whether a generic term
denotes the class property or an individual relies on presumption.
Presumption is based on not being able to account for the
cognition of the meaning otherwise. Sabara has shown that
knowledge of the class property and general knowledge of an
individual can account for the cognition of an individual, but one
cannot account for the cognition of the class property other than
from the word. Therefore he concludes that the word denotes
only the class property.
3.6. Words Denoting Qualities
According to Mimamsa, a sentence consists of words

denoting their own meanings together with a word denoting
action (?3.1.3, note 6). Sabara explains that once one
understands the word-meanings, the word meanings convey t
sentence-meaning. For example, the word 'white' denotes
quality white; the quality makes known a substance qualified
it, i.e. a white object (?3.1.3 note 7). The cognition of the
qualified object, that is, the sentence-meaning, does not arise in all

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
usages. If the word-meaning (quality) by itself makes sense, th

no additional meaning is added. On the other hand, if the word
meaning by itself serves no purpose and so is pointless there, th
the sentence-meaning (qualified substance) is understood.56 For
example, when the qualifying word 'white' occurs near the wor
'cow' or 'horse' which denote the general property, then on
understands the sentence to mean a white cow or horse.57
One understands a substance from the use of a quality
word in the same manner as from the use of a generic term. A
quality word denotes a quality just as a generic term denote
class property.58 Because a class property is not separate from t
substance in which it resides, the class property specifies t
substance. Similarly, because a quality is not separate from the
substance in which it resides, the quality qualifies the substanc
In both cases the fact that a property (dharma) and its substratu
(asraya) are not separate makes the substance the most close
related thing to the property. When the denoted meaning of t
word, i.e., the property, doesn't make sense in connection with t
action because it is an immaterial object, but a substance qualifie
by it does, one understands the substratum of the property.
Although the Arunadhikarana is not directly to this point
Sabara illustrates the manner in which one knows a substance
from the use of a quality word there. The Arunadhikarana
consists of the following sutra:
56Yatra-kevalah paddrthah prayujyamdnah prayojandbhdvdd anarthaka

sa.mjdyata ity avagatam bhavati, tatra vdkyartho 'pi tdvad bhavatv iti
visi.stdrthatdvagamyate na sarvatra. When one understands that it would
useless to use the word-meaning alone because it has no purpose, then, letti
the sentence-meaning supplement it, one understands that the qualifie
meaning occurs, not always. MD, part 1, p. 117.
57Nanv evam bhavisyati sdmdnyavdcinah padasya gaur iti vdiva iti vd
visesakam sukla iti va krsna iti vd padam antikdd upanipatati yadd, tadd
vdkydrtho 'vagamyate. (Proponent:) Will not the knowledge of the sentenc
meaning arise (by way of the word-meanings) in the following manner? When
qualifying word such as 'white' or 'black' occurs near a word denoting a genera
property such as 'cow' or 'horse', then one understands the sentence-meanin
MD, part 1, p. 114.
58Tad eva hi dravyam arunimna paricchidyamdnam arundiabdabhidhdinyata
labhate. Tad evdnyagunakam ndrundsabdah saknoty abhivaditum. Arunimdn
esa sabdo na vyabhicarati. Vyabhicarati punar dravyam. Avyabhicdri ca
kdranam kdranavatam istam. Ato 'sya gunah svdrtha iti gamyate. A certain
substance circumscribed with the quality tan attains to being called tan. The
word 'tan' cannot refer to the same substance with a different quality. The word
does not deviate from the quality tan, but it does deviate from the substance.
That which does not deviate is accepted as the cause of a certain effect. Hence,
one knows that the quality is its proper meaning. MD, part 4, p. 45.
At the end of his commentary on 1.1.25, Sabara describes how concomitant
presence and absence determine that a quality word denotes a quality. MD, part
1, p. 117. The passage is parallel to the one in which he determines that a word
denotes a generic term by concomitant presence and absence (IIIE3).

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MIMAMSA 253
3.1.12. Arthaikatve dravyagunayor, aikakarmyan niyamah syat.

Where the substance and the quality have one purpose, because they
both participate in the same action, there should be a restriction.59
The point of the Arunadhikarana is that mentioning a quality in

the same participatory relation to the same action as a substance
neither forms a separate statement nor states an alternative.60 The
direct statement of the participatory relation requires that the
quality be a means in bringing about an action. It does so
whether that action be the one named in this sentence or
elsewhere. If the statement were divided so that the listener heard
only the quality word, the expectation to know which action the
quality brings about (and which substance it qualifies) would
remain. If he heard just a word for the substance, the expectation
to know which action the substance brings about would remain.
Hence, the statement is not divided so there is just one statement.
The quality and substance do not constitute alternatives
because they are not mutually exclusive. Even though the same
nominal termination following the nominal bases denoting them
assigns them the same relation to the action which is their
purpose, they serve different roles in relation to it. The substance
serves as direct participant in the action, the quality qualifies or
further specifies the substance. The result is that the Vedic
sentence enjoins an action on an object limited in two respects: it
must be a certain substance and it must have a certain quality.
Only one object with both characteristics can participate in
achieving the action.
The following is a condensed description gleaned from
Sabara's commentary on the Arunadhikarana of how one knows a
59MD, part 4, p. 33.

60Sabara considers the statement, "One buys the soma with a tan hazel-eyed
one-year-old (female calf) (Arunayd piigdksyaikahdyanyd somam krTndti)." In
this statement, 'tan (arunaya)' denotes the quality tanness; 'hazel-eyed
(pingdksyd)' and 'one-year-old (ekahdyanyd)' refer to a substance, that is, a cow;
'buys (krlndti)' denotes the act of purchase, and the third triplet nominal
termination (ta), on each of the words for tanness and the cow, directly states
that the quality tan and the cow are instruments in the act of purchase.
According to Jaimini 2.1.46, because a quality and a substance have one
purpose, and if one uttered a quality word alone, a listener would expect more to
be said, the statement of the quality, substance, and action constitutes a single
meaningful unit:
Jaimini 2.1.46. Arthaikatvdd ekam vdkyam sdkdhkasam ced vibhdge sydt. An
utterance is one statement due to having one purpose (artha), if its parts would
require something to complete their sense if the utterance were divided.
If a quality and a substance served precisely the same function in bringing
about an action, the action would take place alternatively with either the quality
or the substance but not both according to Jaimini 12.3.9:
Jaimini 12.3.9. Ekdrthds tu vikalperan, samuccaye hy dv.rttih sydt pradhdnasya.
Subordinates which serve the same purpose should alternate, for on their
conjunction the principal action would repeat.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
substance from the use of a quality word: The word for a qual
denotes the quality. A verb denotes an action.61 The nom
termination on the quality word denotes the participatory relat
the quality has with the action (sddhanatva, i.e., karmatv
karanatva, etc.) It is impossible for a quality, being immaterial
serve the function the termination names for it other than b
qualifying a substance which does serve that function. Althou
a quality cannot directly serve as a means to bring about
action which its nominal termination assigns it, it can do so
qualifying a substance which is such a means. Therefore, due
the nature of the objects (arthat, not due to the word's denota
(sabddt)) the quality attains to specifying the substance.62
In both cases, that of a quality and that of a generic term
the point is that the word denotes just one meaning, the qualit
the class property; it does not denote the individual. In
Arunadhikarana, Sabara describes how one knows to perform
act of purchase using the substance qualified by the quality t
Under sutra 1.3.35, he describes how one knows to perform t
action of washing on the substance qualified by the class prope
riceness from the sentence, "One washes rice (Vrihtn proksat
The means by which one knows to perform the action on
substance, in the former case qualified by a quality and in
latter by a class property, is the same. In each case, although
sentence contains no word which directly gives specific
knowledge of such a qualified substance, it contains words
denoting a qualifying property and an action. The term in the
sentence which denotes the qualifying property ends in a
termination which denotes a means of bringing about action.
Hence, the termination directly states that the property is for the
61More precisely, the verbal termination (tii) denotes causal activity in general
(bhdvana); the verbal root (dhdtu) denotes the specific process the activity
results in (phala).
62See especially: Sabara: Iha hi gunam arunimdnam amurtam santam kriydydh
karanam iti sabda upadisati, yat karandbhidhdyinyd trtlydvibhaktyd samyujya
nirdisati: arunayeti. Na cdmuirto 'rthah kriydydh sddhanam bhavitum arhati.
Here a verbal statement (sabda) teaches that the quality tanness, though it is
immaterial, is the instrument of buying, in that it refers to it with the word
'arunayd' joining the nominal base ('aruna) with the third triplet ending which
signifies an instrument. And yet an immaterial object is not fit to be a means of
bringing about an action. MD, part 4, p. 40.
Further on he says: Krnadtinaiva sambadhyate. Evam hi siruyate: arunagunena
krayam abhinirvartayed iti. Yathd ca tena nirvartyate, tathd yatitavyam bhavati.
Na cdvisimsan sddhanam gunah kriydm abhinirvartayatity arthdt sddhana-
visesanatdm pratipadyate. It is connected with the verb 'one buys'. It is directly
stated that one should carry out the purchase by the instrumentality of the
quality tan. One has to make an effort to see that it is carried out by that
instrumentality. But without qualifying a means of bringing about the action, a
quality does not carry it out. Hence, from the nature of the objects (arthdt), the
quality attains to being a qualifier of the means of bringing about the action.
MD, part 4, p. 62.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MIMAMSA 255
purpose of the action. The property, being immaterial, cannot

bring about an action other than by qualifying a substance. By
qualifying the substance, which can be a means to bring about the
action, the property participates in bringing it about too.
Therefore, due to the nature of the objects one knows that the
action takes place on the substance qualified by the property.
3.7. Conclusions Drawn from Models in

Mimamsa and Nyaya
There are two types of usages of generic terms which one

must explain. In most cases, such as the use of the word 'rice
(vrihi)' in Sabara's example, "One washes rice (vrihin proksati),"
and the word 'cow (go)' in Vatsyayana's examples, "Wash the cow
(gdm proksaya)," the speaker intends, and the listener
understands, an individual qualified by a certain generic property
and having a certain shape. In a few cases, such as the use of the
word 'falcon (syena)' in Sabara's example, "Build a falcon-altar
(syena-citam cinvlta)," and the use of the word 'cow (go)' in
Uddyotakara's example, "Make cows consisting of flour
(pistakamayyo gdvah kriyantdm)," the speaker intends, and the
listener understands, an individual which is not directly qualified
by the generic property, but rather, qualified by a shape
indicative of the generic property.
Neither Sabara nor Vatsyayana use the term 'vyakti
(individual)' strictly for individuals qualified by a certain generic
property. Sabara makes the falcon class property coextensive
with shape qualifying both real falcons and falcon altars.
Gautama and Vatsyayana limit the generic property cowness to
real cows but call all substances which have the shape of a cow,
both real and model, individuals. The difference in their use of
the term does not necessarily imply that their concepts of generic
property, individual, and shape are different. However, different
conceptions of these entities bear on the complexity of the
account of the use of generic terms.
In Mimamsa, Sabara's argument that a generic term
denotes a class property assumes that a shape is more closely
associated with a class property than with an individual. If the
dichotomy of class property versus individual is understood as
shape versus substance, completely identifying the concept of a
generic property with that of a certain general shape and the
concept of an individual with that of a certain material, Sabara's
argument that a generic term denotes a class property is very
straightforward. In most cases in which a generic term is used,
both the shape and a certain material are present. In the examples
of models, the shape is present but the certain material is absent.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
The use of the term is coextensive with the shape, not with t
certain material; hence the term denotes the shape. Therefore
one interprets Sabara as completely identifying a generic prope
with shape, and an individual with a material, Sabara may
understood to conclude from the occasional use of a word for a
model that the meaning of the word is not limited to real
individuals but extends to encompass the broader range of th
general shape. The difficulty with this interpretation is that
Sabara clearly uses the term 'dravya' in the sense of an individua
object in the Akrtyadhikarana, at least in the argument concernin
a replacement (?3.3.4, IIE).
In Nyaya, Gautama clearly distinguishes generic property
from shape. A model individual has shape but lacks the generic
property; a real individual has the generic property in addition to
shape. Both, being manifest aggregate substances, are called
individuals. Gautama 2.2.64 cites the fact that sometimes a
listener does not understand a model individual from a gen
term, but understands only a real individual. In these cases,
word does not deviate from the generic property; hence
concludes that the meaning of a generic term must include
generic property.
Gautama accounts for usage for an individual qualified
a certain generic property and having a certain shape by st
the general rule that a generic term denotes all three elemen
individual, its generic property, and the shape indicative of
generic property. Vatsyayana explains that the intention of
speaker determines which is principal and which subordinat
the meaning. However, he does not directly address the issu
how the word is used occasionally for a model which lack
generic property. If the word denotes all three elemen
whenever the word is used, even when it is used for a model,
cognition of all three elements cannot be avoided. Because the
word is used for a model, one would expect that cognition of the
generic property would be included in the cognition of the model
which the word conveys. Yet the generic property cowness is
absent in a model cow. The fact that we know it is a model and
not a real cow implies that cognition of the generic property is
absent in cognition of the model too. Is cognition of the generi
property present in the cognition conveyed by the generic term
when used for a model, or not? We find two principal solutions
to this dilemma in Nyaya.
Uddyotakara, citing the example of flour cows, gives the
model as an instance in which the shape is principal in the
meaning and the generic property and individual are subordinat
The subordinate position of the generic property in the cognitio
is a compromise. The generic property is present in cognition a
the property ordinarily indicated by the shape which the mode
has but not as qualifier of the model. Although the word giv

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MIMAMSA 257
knowledge of the generic property, the circumstances which forc

one to understand the model block knowledge of it.
The view which the Dinakari attributes to Navyanyaya
avoids the uncomfortable compromise of knowing and yet no
knowing the generic property (?2.3 note 30). However, in
exchange it accepts the encumbrance that a word does not
consistently denote the same meaning. On this view a generic
term denotes two different meanings. Generally, it denotes an
individual qualified by the generic property. Although such an
individual surely has the shape, shape is not part of this first
denoted meaning of the word. Secondly, the generic term
denotes an individual qualified by a shape and excludes the
generic property. This double denotation is very close to the
secondary usage described under Gautama 2.2.62.
Including use of a generic term for a model under usage
of a word for that which it does not denote is inappropriate
because the word is used for what it does denote. It denotes the
shape, an individual, and a generic property. In the case of th
model, it is used for the shape and an individual; it just so
happens that the word denotes the generic property in addition t
what it is used for. Use of a generic term for a model is not use o
a word for an object which it does not denote; rather, it is use o
the word for less than what it denotes. Besides, if the use of a
generic term for a model is included as secondary usage under
2.2.62, it cannot serve as an example of shape as principal in the
denoted meaning in primary usage; nor do there seem to be any
other examples of shape as principal in the denoted meaning.
Mimamsa accounts for the two types of usages of generic
terms by stating the general rule that a generic term denotes only
a class property. There is a second rule according to which, as far
as possible, the class property qualifies an individual substance
which is the means of bringing about an action. In the ordinary
use of a generic term both rules apply to account for the
cognition of an individual qualified by the class property.63 In
the exceptional cases in which the generic term is used for a
model, the condition for the second rule, namely, that it makes
sense for an individual qualified by the class property to be a
means of bringing about the action, is absent. Hence the general
rule alone applies to give cognition of the class property.
Cognition of a real individual does not arise.
Kumarila insists that Sabara uses the term 'dkrti (class
property)' strictly for an abstract generic property and that he
sharply distinguishes it from shape. The dichotomy of class
63As explained in section 3.5, because a class property is an immaterial entity

and as such cannot directly participate in the action in which the sentence states
that it participates, but can participate by qualifying an individual, one
understands an individual qualified by the class property.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
property versus individual is not understood as shape versus

substance; rather, it is understood as abstract generic property
versus a manifest real individual (?3.3). Making a distinction
between class property and shape leads to complications. If th
word denotes the generic property and not the shape, an
explanation is required of how one knows the shape in the
example of the falcon altar and of how in that example Sabara's
statement that one knows the class property but not the individua
makes sense.
Kumarila claims that one knows the shape in which to
build a falcon altar by the similarity the shape bears to the class
property (?3.3.4, IIIA note 12). He must show not merely that
shape is similar to the class property, but that it is more similar to
the class property than it is to a real individual. Shape must have
a direct relation to the class property not mediated by the
individual because Sabara claims that cognition of an individual
doesn't arise in the case of the falcon altar. In order that one be
able to know a shape from knowing the class property without
cognition of an individual arising, the shape must have a direct
relation to the class property. His proof of this stands on
uncertain ground. Nevertheless, let us attempt to see how
Kumarila's interpretation would work.
According to Sabara, the means by which one arrives at
the cognition of an individual qualified by the class property in
the ordinary use of generic terms consists of two steps of
knowledge which take place according to the two rules just stated.
In the example of the falcon altar, this two step means of
knowledge may serve to lead to cognition of an individual
qualified by a shape but not to cognition of an individual
qualified by a generic property. In every case, the class property,
as an immaterial entity, cannot directly participate in the action in
which the sentence states that it participates. In most cases it can
participate by qualifying an individual so that one understands an
individual qualified by the class property. In the exceptional
cases such as the example of the falcon altar, it can't participate by
qualifying an individual either. The ordinary condition for the
second rule, namely, that it makes sense for an individual
qualified by the class property to be a means of bringing about
the action, is absent.
The instruction to build a falcon altar gives direct verbal
knowledge of the class property falconhood qualifying, as far as
is possible, an altar to be built of bricks. An individual qualified
by the class property is a real individual of the class. It simply
doesn't make sense to build a real individual falcon or to build an
altar using individual falcons (IIIA). Hence a real individual
qualified by the class property is not understood in this case.
What does make sense is for a substance qualified by a shape
associated with the class property to be a means of bringing about

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MIMAMSA 259
the action. Hence, one understands such a substance, not a

substance directly qualified by the class property.
It remains to make sense of Sabara's statement that in the
falcon example one knows the class property but not an
individual. In the process of arriving at the cognition of an
individual qualified by the shape which is associated with the class
property, it is not true that one knows a class property and not an
individual. In fact, one does know an individual, an individual
altar. One knows an individual altar just as, in the example of a
model cow in Nyaya, Gautama says that one knows an individual
model cow. It is not an individual cow, but it is a composite
substance. Where Gautama calls a finite manifest substance an
individual whether it is qualified by the denoted class property o
not, Sabara only calls such a substance qualified by the denot
class property an individual. In the falcon example Sabara's
statement that one does not know an individual means one knows
an individual altar, not an individual falcon. But how does one
not know an individual falcon yet know the falcon class property
in relation to an individual altar?
The answer is similar to the compromise we arrived at in
the case of Nyaya's explanation of the use of a generic term for a
model. If the generic property' falconhood were the qualifier of
the individual falcon altar then that individual could not help but
be an individual falcon. But falconhood is not the qualifier of the
model. Yet it is present in cognition as the property ordinarily
associated with the shape which the altar has. Although Kumarila
does not accept the relation of shape to generic property to be that
of being an inferential marker as Gautama states, he does maintain
that there is some sort of qualifying or indicating relation.
The relation a shape has with a class property is that of
ordinarily having the same substratum (samdnasrayatva). The
substratum is an individual. Hence the relation between a shape
and a class property is mediated by an individual. For Sabara's
argument to succeed under Kumarila's interpretation, the relation
between a generic property and shape must somehow be closer
than the relation between an individual and its shape. If it is
closer, it cannot be mediated by the individual. Hence one must
find a relation between the shape and the class property which is
not mediated by an individual. One may argue that once the
relation of ordinarily having the same substratum is learned, it
becomes independent of an individual. Thus, once learned, this
becomes the relation by which the shape is closer to the class
property than it is to an individual. That an object should have a
closer relation to another object than it has to a third object which
mediates the relation between the first two is certainly tenuous.
Nevertheless, it is arguable that a direct relation between the shape
and generic property not mediated by the qualified individual

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
does exist. If there is a direct relation, Sabara can assert that one
knows.the class property but not an individual.
Kumarila's interpretation of Sabara's argument that a
generic term denotes only a class property introduces enough
complexity that it becomes no more appealing than the view that
a generic term denotes both a class property and an individual.
Its appeal lies in the simplicity that only one meaning is denoted.
However, it involves additional complexity in arriving at
knowledge of the additional element. According to Sabara, one
arrives at knowledge of an individual qualified by the class
property due to the nature of the objects (arthdt). If class
property is identical to general shape, then merely in the majority
of cases, not in all cases, due to the nature of the objects (arthdt)
one will know the object the speaker intends. However, if shape
is about as different from the class property as an individual is
then one does not know the shape immediately from the word;
rather, one arrives at knowledge of it from the nature of the
objects just as one arrives at knowledge of an individual. Hence,
cognition of the intended object of speech always involves two
steps of knowing, one from the word to the denoted class
property, a second from the denoted class property to the
intended object (either an individual qualified by the class
property or one qualified by the shape).
According to Kumarila the means by which one arrives at
knowledge of the intended object from the denoted class
property, which Sabara speaks of as due to the nature of the
objects (arthdt), is presumption (dksepa, arthdpatti).64
Presumption depends on not being able to account for the facts
otherwise (ananyalabhyatva). One determines that the speaker
intends an individual qualified by a class property because one
cannot make sense of the sentence otherwise. It does not make
sense that one wash the rice class property itself because it
immaterial, hence one presumes that one washes the substan
qualified by the class property. Similarly, on the interpretation
that the class property is not identical to shape, because it does n
make sense either to build the class property falconhood or
build a real individual falcon out of bricks, one presumes that on
64Prabhakara holds that the individual and the class property are necessari
elements of the same cognition (tulyavittivedya). One can never know o
without knowing the other. According to him arriving at knowledge of th
second is not a separate step of knowing; rather, one already knows the oth
too if one knows the first. The class property and individual are bound in o
cognition. The class property is the handle which the word causes one to gr
to give knowledge of both. This view is very close to the view that the wo
denotes both the class property and the individual.
Mandanamisra professes the view that the word has a primary significatio
relation with the class property alone. It has a secondary signification relati
with the individual qualified by the class property.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
MiMAMSA 261
builds the shape. In each use of a generic term, the speaker,

intending an object other than the one the word denotes, uses a
generic term to give knowledge of the generic property
qualifying the intended object. Obtaining cognition of the
generic property from the word, it is by presumption that the
listener arrives at cognition of the intended object. Presumption
(arthdpatti) is also the means by which one determines the
denotation of a word (?3.5). One determines that a generic term
denotes a generic property because one cannot account for the
cognition of the generic property otherwise. One determines that
the word does not denote an individual in addition to the class
property because one can account for the cognition of an
individual otherwise. How does one account for the cognition of
an individual otherwise? By presumption. In each use the
listener presumes the intended meaning from the impossibility to
make sense of the denoted meaning in order to account for the
usage of the sentence.
Hence, according to Kumarila, in each and every use of a
generic term, the listener presumes the intended meaning from the
denoted meaning because one cannot account for the facts
otherwise. But one can account for the facts otherwise; one can
presume that the word denotes those intended meanings. The
only obstacle to presuming that the word denotes both the class
property and an additional meaning is that one can presume the
additional meaning without presuming that the word denotes it.
But the consistent presumption of the additional meaning from
the class property is no stronger than the presumption that the
word denotes that meaning. The two presumptions are
equivalent. Therefore, there is no stronger reason for the view
that a generic term denotes only a class property than for the view
that a generic term denotes both the class property and an
additional meaning. Both in Nyaya and in grammar, the
preferred view is that the additional meaning is also denoted.
The difficulty of the Nyaya view that a generic term
denotes the generic property, an individual and a shape is that the
word does not consistently give knowledge of all three. One must
treat the usual case in which one knows an individual qualified by
the generic property differently from the exceptional case in
which one knows an individual qualified by a shape. The
Mimamsa view carries the advantage of a single word-meaning
relation with a single denoted meaning but carries the
disadvantage of two steps of knowing to arrive at the intended
meaning.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Akrtyadhikarana Outline
I Preliminaries
A. Definitions
1. A class property is a general property of individual
substances, qualities or actions.
2. An individual is that which has specific properti
which distinguish it as unique.
B. Whenever one uses a generic term,
1. One understands a class property.
2. One acts upon an individual.
II. The word denotes an individual.

A. Action:
1. Vedic texts enjoin actions such as killing, wash
and cutting.
2. These actions are not possible on a class property.
3. They are possible on an individual.
4. Therefore, the texts make sense only if the word
B. Because a class property and an individual are
inseparably connected, once one knows one of the two,
one comes to know the other. Therefore, it is only
necessary that the word give knowledge of one of the
two. Hence, one cannot presume that the word denotes
more than one of them.
C. Cognition of the general property arises because it serves
as a mark to identify the individuals for the word to
denote; yet the general property itself is not denoted.
For example, the staff is the mark of a staff-bearer, yet
the word 'staff-bearer' does not denote the staff.
D. Number:
1. Vedic texts mention various numbers of the obj
of a class.
2. Number bears no relation to the class property
because there is just one class property per class.
3. Number relates to individuals.
E. Difference
1. Vedic texts mention other objects of the same kin
2. There is only one class property for the kind,
other.
3. There are other individuals of the kind.
263

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
III. The word denotes the class property.

A. Action:
1. Vedic texts enjoin actions such as building a falco
altar.
2. One carries out the action with respect to the cla
property; one makes the shape out of bricks.
3. One cannot create a real individual falcon.
denotes the class property.
5. Although one could construct something using
individual falcons as building blocks, the text states
the falcon as the direct object (karman) of the
action in the sentence, not the instrument (karana).
B. (Objection:) The two views are equivalent with respect
to action.
1. Some action is impossible on either view .
a. It is impossible to wash the class property of
rice (IIA2).
b. It is impossible to build an individual falcon by
piling up bricks (IIIA3).
2. Whichever is denoted can indicate the other as the
participant in action.
a. The word 'rice' denotes the class property, the
class property indicates the individual on which
to perform the action of washing (answers
IIIB 1 a).
b. The word 'falcon' denotes the individual, the
individual indicates the class property (shape)
to build (answers IIIBlb).
C. The word does not denote an individual alone to the
exclusion of its class property.
1. If it did denote an individual alone,
a. It wouldn't denote any other individuals, even
those of the same kind.
b. Because it does refer to other individuals of the
class and an individual is by definition
different from that which is common to more
than one entity, it cannot denote an individual.
2. If it denoted other individuals completely
independent of any properties,
a. Then the object the word denotes would be a
general property.
b. But an individual is different from properties; it
is that in which properties reside.
3. If each individual were completely devoid of all
properties, one would be unable to set a criterion to
delimit the class of individuals suitable to be
denoted by a word.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
AKRTYADHIKARANA OUTLINE 265
a. An exhaustive list of current usage could not

be the criterion to delimit the class because:
i. It wouldn't account for the fact that the
word refers to new members of the class.
ii. It wouldn't account for the cognition of
the general property which is the basis of
knowing each individual as the same.
b. The signifying capacity of the word itself
cannot determine that the word denotes certain
individuals and not others because it does not
denote a random set of individuals (IIIC3aii).
4. The criterion to determine the range of a generic
term is the class property.
a. If the word denoted all and only those
individuals which their class property
indicated, then the word would not produce
cognition of the individual alone; rather, it
would produce cognition of the individual as
qualified by its class property.
b. Knowledge of a qualified object presupposes
knowledge of the qualifier. Therefore, a word
cannot denote an individual alone and rely on
the connection between the individual and its
class property to supply cognition of the latter
(IIB).
D. (Objection:) Both views accept both the individual and
the class property as elements in the meaning of a word,
they differ on which is principal and which secondary:
1. On the view that a word denotes an individual, the
individual is the principal meaning, and the class
property is subordinate.
2. On the view that a word denotes the class property,
the class property is the principal meaning, and an
individual is subordinate.
E. A word does not denote an individual as qualified by its
class property. The class property alone is denoted.
1. Because a class property and an individual are
inseparably connected, once one knows the first,
one comes to know the second. Therefore, it is
only necessary that a word give knowledge of one
of the two. Hence, one cannot presume that the
word denote more than one of them (IIB).
2. It is not necessarily the case that all the objects of
the cognition which follows the use of a word are
denoted.
a. There is cognition of both an individual and
the class property.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
b. It is not obvious which the word denotes and

which the denoted object gives knowledge of.
3. Concomitant presence and absence establish that it
is the class property, not the word, which gives
knowledge of the individual.
a. Even without the use of a word, if one knows a
class property then one knows an individual of
the class.
b. Even when a word is spoken, if cognition of
the class property does not arise, one will not
know an individual.
F. The word does not denote the class property as qualified
by (resident in) an individual.
1. Qualified by an individual, the class property loses
its generality.
2. If the individual were included in the meaning even
as the qualifier, the word would be unable to denote
the class property as qualified by (resident in) any
other individual (IIIC1).
G. Accepting that the word denotes just the class property,
one can account for the cognition of the individual as
principal.
1. Granted that one has the cognition of an individual
as qualified by the class property.
2. The word denotes the class property alone no
matter whether it is principal or subordinate in the
cognition.
a. The way the meanings fit together decides what
is principal and what is subordinate in the
cognition.
i. If the word is used to express its own
meaning, the class property is principal.
ii. If it is used to express another object, then
the class property is subordinate and
qualifies the principal object.
b. The word is not engaged in whether its
meaning is principal or subordinate in the
cognition.
3. The denotation of the word 'cow' is not parallel to
that of 'staff-bearer' (IIC).
a. In the example, because something supplies the
knowledge of the qualifier first, the qualified
cognition can arise in accordance with the
principle that knowledge of a qualified object
presupposes knowledge of the qualifier
(IIIC4b).
i. A separable part of the word 'staff-bearer',
i.e. 'staff, supplies cognition of the staff.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
AKRTYADHIKARANA OUTLINE 267
ii. '-bearer' gives cognition of the person.

iii. These two cognitions together give the
cognition of the person qualified by the
staff.
b. There is no part of the word 'cow' to give
knowledge of the qualifier cowness so that the
qualified cognition could arise.
c. Because of the principle that one must know
the qualifier first in order to have a qualified
cognition, the word 'cow' denotes the class
property.
4. Concomitant presence and absence determine that a
generic term denotes the general property
exclusively.
a. One does find a generic term and class
property, in the absence of an individual (e.g.
'falcon') (IIIA).
b. One never finds a generic term and an
individual, in the absence of the class property
(e.g. 'rice').
IV. Response to the objections in II.

A. 1.3.34 restates IIA, IID and IIE
B. 1.3.35 answers them:
1. Because the class property and individual are
inseparably connected, although a word denotes the
class property, it is used to convey the individual.
a. Because an individual substance directly
participates in the Vedic performance, it is
connected to the transcendental result; the class
property is not so connected because it cannot
directly participate in the performance. Hence,
just the mention of the action gives knowledge
of an individual substance.
b. A word denoting the class property alone is
used to qualify the substance upon which the
action is done. It specifies which substance.
(e.g. 'rice')
2. Because number relates to individuals, it gives
knowledge of unspecified individuals. A word
denoting the class property qualifies the individuals.
3. Because a replacement relates to the substance it
replaces by the relation of being other, the word
denoting the class property will qualify the
substance with the class property to provide the
qualified cognition of the kind of substance.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Jaimini 1.1.5 Translation (Selections)
1.1.5. Autpatikas tu sabdasydrthena sambandhas;

tasya jiidnam upadeso 'vyatirekas cdrthe 'nupalabdhe;
tat pramdnamz Bddardyanasydnapeksatvdt. The relation
of a word with its object is innate; knowledge of it is
verbal instruction, and (verbal instruction) does not fail
with respect to an object beyond the range of
perception; it is a valid means to know (even such an
object, and therefore to know dharma), in the opinion of
Badarayana, because it is independent.1
Nanu sarva eva nirdlambanah svapnavat pratyayah....2

S1) (Buddhist:) All cognition has no basis in real objects just as
dreams don't....
Api ca niyatanimittas tantusv evopddTyamdnesu patapratyayah.

Itarathd tantvdddne 'pi kaddcid ghatabuddhir avikalendriyasya
sydt. Na caivam asti. Ato na nirdlambanah pratyayah. Ato na
vyabhicarati pratyaksam.3
(Proponent:) Moreover, the cognition of cloth has a definite
cause, it occurs when and only when threads are present (as
material cause). Otherwise, even given threads, sometimes one
whose senses are perfectly fine would have the the cognition of a
pot. But it isn't so. Hence, cognition is not without a basis (in real
objects). Therefore, direct perception does not err.
.. Atha gaur ity asya sabdasya ko 'rthah?4

S2) (Objector:) .... Now what is the object denoted by the word
cow?
Sasnddivisistdkrtir iti brumah.

(Proponent:) The class property characterized by the dewlap, etc.
1MD, part 1, p. 28. Sabara interprets 'tasya' to refer to dharma so that the second
phrase says, "the means to know (dharma) is verbal instruction. However,
'sambandha' immediately precedes and makes a natural referent. Nagesa seems
to make a direct reference to this sutra interpreted in this way when he says,
"verbal instruction is grasping the relation (between a word and the object it
denotes) (upadesah sambandhagrahah) (Under Panini 1.2.64, MB, p. 70b)."
2MD, part 1, p. 34. For the full argument, see MD, pp. 34-42 or Gaingnatha Jha
(1933: vol. 1, pp. 12-15).
3MD, part 1, p. 42.
4MD, part 1, pp. 59-62.
269

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Nanv dkortih sadhyasti va, na vd?

(Objector:) Is the class property something to be inferred, or not?
Na pratyaksd sati sddhyd bhavitum arhati. Rucakah svastiko

vardhamdnaka iti hi pratyaksam drsyate.
(Proponent:) Being directly perceptible it cannot be something to
be inferred. We directly perceive a necklace, an auspicious
symbol and a cup.5
Vydmoha iti cen,

S3) If (one objects that) it is delusion.
Na. Ndsati pratyayaviparydse: vydmoha iti sakyate vaktum.

(Proponent:) No. You can't say we are deluded unless another
experience gives evidence that our cognition is false.
Asaty api arthdntara evamjdttyako bhavati pratyayah, panktih,

yutham, vanam iti yatheti cet.
(Objector:) Even when there is no other additional object we
have the cognition of such an object; for example, a row, a herd,
a forest.
Na. Asambaddham idam vacanam upanyastam. Kim: asati vane

vanapratyayo bhavattti pratyaksam evdksipyate: vrksd api na
santfti? Yady evam, pratyuktah sa mdhdydnikah paksah. Atha
kim adkrtisadbhdvavddy updlabhyate: siddhdntdntaram te dusyatTti:
vane 'py asati vanapratyayah prdpnottti? Evam api prakrtam
dusayitum asaknuvatas tatsiddhdntantaradusane nigrahasthdnam
dpadyate, asddhakatvdt. Sa hi vaksyati: Dusyatu yadi dusyati.
Kim tena dustenddustena vd prak.rtam tvayd sddhitam bhavati,
madlyo vd pakso dusito bhavattti. Na ca vrksavyatiriktam vanam
yasmin nopalabhyate 'to: vanam ndst[ty avagamyate. Yadi vane
'nyena hetund sadbhdvaviparitah pratyaya utpadyate, mithyaiva
5A vardhamdnaka is a certain type of pot for drinking. The Amarakosa, 2.9.32a

(p. 313a) says, "Ghatah kutanipav astrt sardvo vardhamdnakah," on which the
Ramgarami says, "...dve pdtrabhedasya."
Vaidyanatha Sastri comments: sauvarnapaddrthagatarucakatvddiripdvdntara-
sdmdnydbhiprdyena bodhyam. Traydnugatasuvarnatvarupasdmdnydbhiprdyam
vedam. He intends the specific properties of being a necklace, etc. in golden
entities. He also intends the general property goldenness present in all three.
MD, part 1, p. 61. That is, the different names denote the narrower class property
as opposed to the one that the three objects have in common. Yudhisthira
Mimamsaka (1977-86, vol. 1, p. 39) uses the Sanskrit words minus their
terminations for the three examples in Hindi but adds, "dbhusanavisesa
(specific ornaments)."
Ganiganatha Jha (1933: vol. 1, p. 21) translates the examples, "the necklace, the
road-crossing, the dish, and so forth," as if they are totally unrelated things.
Taking them as golden ornaments, as Vaidyanatha and Yuddhisthira
Mimamsaka do, brings out the relevance of the objectors argument.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
JAIMINI 1.1.5 TRANSLATION 271
vanapratyaya iti. Tato: vanam ndstTty avagacchamah. Na ca

gavddisu pratyayo viparyeti. Ato vaisamyam. Atha vanddisu
naiva viparyeti, na: te na santfti. Tasmdd asambaddhah pankti
vanopanydsah. Ata upapannam jaiminivacanam: dkrtih sabddrtha
iti. Yathd cidkrtih sabddrthas tathoparistan nipunataram
upapddayisydma iti.
S4) (Proponent:) No. The statement you make (that we cognize
a whole even when there is no object other than the parts) is
irrelevant (it has nothing to do whether the perception of a class
property is true). a) Saying, "When there is no forest, one has the
cognition of the forest," do you mean thereby to object to the
direct perception of trees? If so, this Mahayana Buddhist view
(that no external objects exist) has been refuted (1.1.5. Si). b) Or
by saying, "Even when a forest is not present one cognizes a
forest," do you rebuke him who says that the class property exists,
"another one of your conclusions is false?" In this way too,
unable to refute the proposal at hand, by attacking his other
conclusion, you are defeated, because refuting (the conclusion
that cognition is based on real objects) does not prove (that a class
property does not exist independent of substance). For he'll say,
"If my conclusion (that cognition is based on real objects) is
refuted, let it be refuted! By it being refuted or not do you prove
your point in the current debate (that a class property does not
exist) or refute my view (that it does)?" Moreover, just because
we don't perceive the forest to be distinct from the trees doesn't
mean the forest doesn't exist. If for some other reason a cognition
arises which contradicts the existence of the forest, then certainly
the cognition of the forest is false. Then we understand that the
forest does not exist. But (just because a cognition arose to
contradict the existence of the forest) it is not the case that the
cognition with respect to cows (that they are cows, i.e. cowness) is
contradicted. They are two different cases. On the other hand, if
a cognition does not arise to contradict the existence of the forest
and other aggregates (the row and the herd) then it is not true that
they (the forest, the row and the herd) do not exist. Therefore,
your statement about the row and forest is irrelevant. Hence,
Jaimini's statement that the object denoted by a word is the class
property makes sense. We will explain how the object denoted by
a word is the class property more thoroughly later.6
6 The references are to Jaimini 1.3.33, Ak.rtis tu kriydrthatvdt, and to the

Akrtyadhikarana.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Jaimini 1.3.30-35 Translation
Lokavedaidhikarana (Selections)
1.3.30. Prayogacodanabhdvdd, arthaikatvam,

avibhdgdt.1
(Proponent:) Because (Vedic texts) enjoin performance,
(each word) has just one meaning because there is no
division.
... ... ... ... ... ... ... ... ... ... ... ... ... ...... ... .
Ya eva laukikdh sabdds ta eva vaidikds, ta evaisdm arthd iti.
(Proponent:) Words that occur in the Veda are none other than
the words that occur in ordinary usage; their meanings are none
other.
Kutah?
(Objector:) Why?
Prayogacodandbhavat. Evam prayogacodand sarbhavati

yadi ta eva sabdds ta evdrthd; itarathd, sabddnyatve 'rtho na
pratiyeta. Tasmdd ekasabdatvam iti.
(Proponent:) Because Vedic texts enjoin performance. It
is (only) possible to enjoin performance if the words are none
other and their meanings are none other; otherwise, if the words
were different, the meaning would not be understood.
Ucyate: Prayojanam idam. Hetur vyapadisyatdm.

(An objector) says: This is a purpose. State the cause.
Tato hetur ucyate: avibhdgdd iti. Na tesdm esdm ca vibhdgam

upalabhdmahe, ata evaikasabdatvam. Tams ca trnms cdrthdn
avagacchdmah ato ndnyatvam ca vaddmah.
Then (the proponent) states the cause, "because there is no
division." We do not perceive any division between those (Vedic
words) and these (ordinary words), so the words are the same.
We understand the same meanings (from the Vedic words as we
do from ordinary words), so we say that they too are none other.2
1MD, part 2, p. 229.

2MD, part 2, pp. 231-32.
273

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Akrtyadhikarana (Entire)
Yadi laukikds ta evdrthds tadd samdehah: kim dkrtih sabddrtho
'tha vyaktir? iti.3
(Proponent:) I) If the meanings (of words that occur in the Veda)
are none other than the meanings of those words in ordinary
usage, then the question arises, "Is the object denoted by a word
the class property, or an individual?"
Kd punar dk.rtih? Ka punar vyaktir? iti

(Objector:) A) But what is a class property? And what is an
individual?
Dravyagunakarmandm sdmanyamdtram dk.rtih. Asddhdranavisesa

vyaktih.
(Proponent:) 1) A class property is any general property of
substances, qualities or actions. 2) An individual is that in which
there are exclusive distinguishing properties.
Kutah samsayah?
(Objector:) B) Why does the doubt arise (as to whether the object
denoted by a word is the class property or an individual)?
Gaur ity ukte sdmdnyapratyaydd, vyaktau ca kriydsambandhdt.

(Proponent:) 1) Because when the word 'cow' is spoken, one
understands a general property, 2) but action has a relation with
an individual.
Tad ucyate "vyaktih sabdirtha" iti.

II) The (objector) says: An individual is the object denoted by a
word.
Kutah?
(Proponent:) Why?
Prayogacodanabhdvdt. A lambhanaproksanavisasandtnamz
prayogacodand adkrtyarthe na sambhaveyuh.
(Objector:) A) Because (Vedic texts) enjoin performance.
Injunctions to perform killing, washing, cutting, etc. would be
impossible if the object denoted by a word were the class
property.
Yatroccdrancnarthakyam tatra vyaktyarthah. Ato

'nyatrdkrtivacana iti cet,
3MD, part 2, p. 235-68.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
JAIMINI 1.3.30-35 TRANSLATION 275
(Proponent:) Where the utterance would have no purpose (if the

word denoted a class property), it denotes an individual.
Elsewhere it denotes a class property.
Uktam: anydyas cdnekdrthatvam iti.

(Objector:) B) It has been said that it is unreasonable for a word
to denote more than one object.4
Katham sdmanyavagatir iti cet,

C) If (the proponent) asks: (if the word denotes only the
individual then) how does one understand the general property?5
Vyaktipaddrthakasydkrtis cihnabhiitd bhavisyati ya evamakrtikah

sa gaur iti. Yathd yasya dando 'sti sa danditi na ca dandavacano
dandisabdah, evam ihdpi. 30.
4Kumarila: Nyayena hi sthitam ekasya sabdasydrthaikyam anavasthita-

sam bandhdnekddrstavdcakasaktikalpanavikalpadosaprasangdt, sambaddhdndm
carthdndm anyatardbhidhdnenaivetaratra pratipattisiddheh. (Objector:)
Reasoning has established that each word denotes only one object because
otherwise the relation between the word and its object would not be fixed, one
would have to posit more than one imperceptible capacity to denote (sakti) for
each word, there would be an option (that sometimes the word denoted an
individual and at other times the general property), and because just by
denoting one of connected objects one achieves cognition of the others. MD,
part 2, p. 243.
5Kumarila: Katham tarhy dkrtipratyaya iti ced ata dha: avibhdgad iti. Yatas
tu nityam eva vyaktydkrtyor avibhdgas, tasmdd vyaktir abhihitd satl saknoty
evdkrtim gamayitum. Tad uktam tadbhutddhikarane. Yasya tu padasya yo 'rtho
'bhidheyatvens'ritas tasya tat-parihdrdsambhavdd avasyam tatra-
sthasyaivdrthdntaram ldksanikatvendbhyupagantavyam iti. Api cdkrti-
paddrthakasya vyaktisu sddhdrandnaikdntikatvdn nirnayo na sydd. Vyakti-
paddrthakasya punar atyantdntarbhitadkrtiniscayasiddher na sabdavydpdra-
kalpanopapattih. Tasmdd vyaktir eva sabddrtho ndkrtih. To answer the
objection, "Then how does one have the cognition of the general property?"
Jaimini says, "because there is no division." Because an individual
and its class property are eternally inseparable, the individual being denoted
can convey the understanding of the class property. This has been said in the
Tadbhutadhikarana (?3.1.3 note 6). Because it is impossible for the object
which a word denotes to avoid being denoted when the word is used, one must
accept any additional meaning of the word which is fixed in its denoted object,
as being indicated by the denoted object. Moreover, if one maintained that the
object denoted by a word is the class property, one would not be able to
determine which individual the class property indicated because there would be
equal cause for it indicating each. But if one maintains that the object denoted
by a word is an individual, one can determine (that any individual will indicate)
the class property which is entirely included (in any individual which the word
denotes). Hence, it does not make sense to posit that the word conveys the class
property. Therefore, it is an individual which is the object denoted, not a class
property. MD, part 2, p. 243-44. [Note: If the reading is 'kalpandnupapattih',
supply the individual instead of the class property and translate as follows:
Hence it makes perfectly good sense to posit that the word conveys the
individual.]

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
(Objector:) The class property will serve as a mark: whatever has

such and such a class property is called a cow. Just as whoeve
has a staff (dan.da) is called a staff-bearer (dandin), yet the word
staff-bearer does not denote the staff, in the same way her
(whatever has cowness is called a cow, yet the word 'cow' doe
not denote the class property cowness).6
1.3.31. Adravyasabdatvdt.
D) Because there would be no word for a substance.
Dravyasrayasya sabdo dravyasabdah. Na tatra
dravydsrayavacanasabdo bhavet, yady dkrtih sabddrtho bh
Sad deyd dvddasa deyas caturvims'atir deyd iti na hy d
sadddibhih samkhydbhir yujyate. Tasmdn ndkrtivacanah. 3
A word for a substance is a word for a substance as the
substratum (of the general property and a number). The
be no word denoting a substance as the substratum, if t
denoted by a word were the class property. 1-2) The
"one should give six, one should give twelve, one sho
twenty-four" would not make sense, because the class pr
has no connection with the numbers, six, etc.7 4) Theref
word does not denote the class property.8
6The word 'dandin' is formed by adding the suffix ini in the mean
suffix matup. Panini 5.2.115, Ata inithanau: The affixes ini and
after a nominal base ending in the sound a. Panini 5.2.94, Tadasyast
matup, provides the affix matup after a word meaning x to form a wo
that of which, or in which, x is. The Kas'ika (vol. 2, p. 520) gives 'da
as an example in the meaning contact (samsarga), i.e., one in whose co
is a staff, or loosely, one who holds a staff. See Sabara on Jaimini
"Dandi praisdn anvdha." MD, part 5, p. 50-51.
7Satapathabrahmana 2.2.2.3-5 (p. 39): Tau vai sad dadydt.... Dvadas
Caturvimsatim dadydt.... Taittiriyasamhita 5.4.11.4 (p. 225): Dvd
dadhdti... Sabara discusses number under Jaimini 4.1.15. MD, part 5,
8Sabara's compound 'dravya-dsraya' is a karmadharaya. His sta
straightforward and consistent with the objector's position. Ku
interpreted Sabara as taking the compound to be a bahuvrfhi m
number, quality, or general property) whose substratum is a substa
the same time he rejects such an interpretation because it has several
MD, part 2, pp. 2445. It seems to me that this was simply not what Sab
Kumarila's own interpretation is quite reasonable.
Kumarila: Api cadravyasabdatvad vyakter evdbhidheyatd;
sdmdnddhikaranyam hi nadkrtyarthagundrthayoh. Moreover, the individ
alone is the object denoted because otherwise there would be no word f
a substance; a word denoting a class property and a word denoting a qualit
are not co-referential.
Gosabdadlndm dkrtivacanatve gauh suklah, arunaya pinigdksyaikahdyanya gavd
krndtifti, sad gdvo deydh, ekdm gdm daksindm dadydd ity evamddisu prayogesu
sarvesdm jatigunamdtravacanatvdt sdmdnddhikaranyam na prdpnoti. If the
words 'cow', etc. denoted the class property, in each of the usages, "white cow,"
"One buys (soma) with a tan brown-eyed one-year-old cow," "One should give
six cows," "One should give one cow as fee," because all the words would
denote only generic properties and qualities, they would not attain to being co-
referential.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
1.3.32. Anyadarsanac ca.

E) Because (we) see (provision for) another.
Yadi pasur upakrtah paldyeta, anyam tadvarnam tadvayasam
alabheteti yady dkrtivacanah sabdo bhaved, anyasydlambho
nopapadyeta. Anyasyapi pasudravyasya saivdk.rtih. Tasmdd
vyaktivacana iti. 32. (purvapaksah)
1) (A Vedic injunction states) "If the animal brought (for a Vedic
performance) passes away, one should offer another one of the
same color and age."9 2-4) If the word ('animal') denoted the
class property, offering another one would not make sense. Even
another individual animal has the same class property (so one
would not succeed in offering something other. Therefore, the
word denotes an individual.
1.3.33. Akrtis tu kriyarthatvat.

(Proponent:) III) But the class property is (the object
denoted by a word) because it is for the purpose of the
action.
Tusabdah paksam vydvartayati. A krtih sabddrthah.
The word 'but' rejects the previous view. The class property i
object denoted by a word.
Kutah?
(Objector:) Why?
Gotvasya na hi sambandhah suklatvdrunimddibhih; yena sasthy api tdvat sydt

kutas tv ekarthavrttitd. Na hi gotvdk,rtih sukld ndrund ndpi sadgund; vyaktis tv
evamgund tasmdd vyakter evdbhidheyatd. Mama hi vyaktisabdatvdt sidhyaty
ekarthav.rttitd; tava tv adravyasabdatvad bhaved bhedo gavdsvavat.
Cowness simply has no connection with the color white, the color tan, etc., by
which even a sixth triplet termination could occur (on the word cow, e.g., the
cow's white color), so how can the words 'cow' and 'white' refer to the same
object? For the class property cowness is neither white nor tan nor six-fold, but
the individual has such qualities. Therefore, it is the individual which is
denoted. Because on my view there is a word denoting an individual, the words
succeed in referring to the same object, but because on your view there would
be no word for a substance, their reference would be as different as cows
and horses. MD, part 2, pp. 244-45. Taittiriyasamhita 6.1.6.7-8 (pp. 254-55,
mantras 16, 19): "Ekahdyanyd krtndti. (One buys with a one-year-old),"
"Arunayd pingdksyd krfndti. (One buys with a tan hazel-eyed)." See Sabara on
Jaimini 3.1.12. MD, part 4, p. 33.
9"Pasus ced updkrtah paldyate vdyave tam anudisydnyam tadvarnam tad-
vayasam dlabheta." Katyayanasrautasutra 25.9.1 (part 2, p. 528).
The Karka commentary glosses the word 'paldyate' with the word 'nasyet'.
Hence, it means dies as is also clear from Sabara's reference to the relation of the
replacement to the 'deceased (vinastasya)' below (IVB3). The animal (pasu,
literally, domestic animal) in the Vedic performance is a goat (chdga). See
Jaimini 6.8.30. Pasucodandydm aniyamo 'visesat. 31. Chdgo vd mantravarndt.
("Agnaye chdgasya vapdyd medaso 'nubruihi. Invoke Agni with the fat of the
marrow of a goat."). 'Upikrta' means brought near, tied to the post (yupa) and
prepared. To prepare it for offering, one rotates lit darbha grass around the goat
in the manner of drdtrikam (Hindi, dratO.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Kriydrthatvat. Syenacitam cinviteti vacanam dkrtau

sambhavati, yady dkrtyarthah syenasabdah. Vyaktivacane tu: na
cayanena syenavyaktir utpddayitum sakyata ity asakydrtha-
vacandd anarthakah. Tasmdd dkrtivacanah.
A) Because it is for the sake of the action. 1-2)
(Proponent:) The injunction, "One should build a falcon altar
possible with respect to the class property, that is, if the w
'falcon' denotes the class property.10 3-4) But if the word den
the individual, the injunction would be pointless because, du
the fact that it is impossible to create an individual falco
piling up (bricks), it would enjoin the impossible. Therefore
word denotes the class property.
Nanu syenavyaktibhis cayanam anusthdsyate.

(Objector:) But piling up will be carried out using individ
falcons (as building blocks).
Na sddhakatamah syenasabdarthah, fpsitatamo hy asau

syenasaenana nirdisyate. Atas cayanena syeno nirvartayitavyah,
sa dkrtivacanatve 'vakalpyate.
(Proponent:) 5) The word 'falcon' does not mean that which most
serves to bring about the act; rather, the word 'falcon' refers to
that which one most wishes to obtain.11 Hence, one must produce
10Taittiriyasamhita, 5.4.11.1 (p. 225): Syenacitam cinvita suvargakdmas syeno

vai vayasdm patisthas syena eva bhutvd suvargam lokam patati.... Syenacitam
iti syena-citam, cinvTta, suvargakdma iti suvarga-kdmah, syenah, vai, vayasdm,
patisthah, syenah, eva, bhutvd, suvargam iti suvah-gam, lokam, patati.... One
desirous of heaven (eternal happiness) should build a falcon (on which to put
the fire). It is the falcon which is the best flyer among birds. Becoming a falcon
he flies to heaven (the world of eternal happiness).
Bhattabhaskaramisra comments: Syenacitam iti. Syena iva clyate vakrapakso
vyastapuccha iti. Sa eva kvip. Nitya eva; ndtra kdmasamyogah. Uktam ca:
syenacidagnindm iti. 'Falcon-built' means built like a falcon, with bent wings
and fanned tail. The falcon itself is what is built. Building the falcon is a
requirement; there is no connection with desire. In addition, it is said, "Among
fires, the falcon-built...."
11The proponent states that the word 'falcon' refers to the falcon as the direct
object of the action of piling up not as the instrument. To do so he indirectly
refers to the following grammatical rules:
Panini 1.4.42, Sddhakatamam karanam: That participant in an act (kdraka) which
most serves to bring about the act is the instrument (karana).
Panini 1.4.49, Kartur ipsitatamam karma: That participant in an act (kdraka)
which the agent (kartr) most wishes to obtain is the direct object (karman).
Panini 3.2.92, Karmany agnydkhydydm: The Kasika (vol. 1, p. 232) says: Ceh,
karmani iti vartate. Karmany upapade cinoteh karmany eva kdrake kvip
pratyayo bhavaty agnydkhydydm, dhdttipapadapratyayasamuddyena ced
agnydkhya gamyate. Syena iva ciyate syenacit. Katikacit. Akhydgrahanam
rudhisampratyaydrtham. Agnyartho hi istakdcaya ucyate: syenacid iti. When
there is a co-occurring word denoting the direct object, the suffix kvip occurs
after the verb root V/ci (to build) to denote the direct object itself, if it is a name
for a fire, that is, if one understands the aggregate of the root, subordinate word

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
a falcon by means of piling up. That is possible if the word

denotes the class property.12
and suffix to be a name for a fire. 'Falcon-built' means built like a falcon.
Heron-built. Panini mentions the word 'name' to communicate that the words so
formed are conventional terms. It is the brick structure for the purpose of the
fire (not the fire itself) which is called 'falcon-built'.
The Padamafijari commentary on the Kasika clarifies the final point:
Akhyagrahanam rudhisampratyaydrtham iti. Agnisabdo loke jvalane
rudhah, vede tv agnyartha istakdcaye 'pi pracurah prayogah... iha tv asaty dkhyd-
grahane lokaprasiddhivasena jvalana eva g,rhyeta: Md grdhi, istakdcaya eva
grhyatdm ity dkhydgrahanam ity arthah. The statement that the word 'name' is to
communicate that the words so formed are conventional terms means the
following: The word 'fire' in normal usage conventionally means the flames but
in the Veda it is also widely used to mean the brick structure whose purpose is
the fire.... But here if the word 'name' were absent from the sutra, because normal
usage is so well known one would understand it only in the sense of flames.
The mention of 'name' tells one not to understand it in the normal sense but to
understand it only in the sense of a brick structure. KNP, part 2, p. 609.
The Nyasa clarifies the meaning of the term as follows: Agnyartho ya istakdndm
racandvisesah sa syenacid ity ucyate, tatraivdyam pratyaya isyate.... A certain
arrangement of bricks for the purpose of a fire is called 'falcon-built'. One wants
the suffix kvip only in this meaning. KNP, part 2, p. 609.
12Kumarila: Kim ca naivestakdbhir dkrtisampadanam abhyupagatam yatah
svabhdvdt sndyvddydrabdhadravyasamavdyitvenopdlambhah sydt. Pistapinda-
simhavat sddrsyasampattih paksadvaye 'pi codandrthah. Sd tv dkrtyantara-
vilaksanasyendk.rtya syan na tu vyaktyantaravilaksanasyenaviesend
'sddharanena vd. Nirvikalpatvdt sdmdnyavisesanapeksam vyaktimdtrasddrsyam
avidheyam, sarvadravyesu tulyatvdt. Ekavyaktiparigrahdc ca vyaktyantarena
sddrsyam na kdryam sydt. Tasmdc chyenacitkriyd vyaktau sabddrthe na
sambhavati. It is not accepted that one constructs the class property with bricks
because then an objection would be possible because the class property
naturally inheres in a substance that is made of tendons, etc. (not of bricks). The
meaning of the injunction ("Build a falcon altar") on both views (the view that
the word 'falcon' denotes the generic property falconness and the view that it
denotes an individual falcon) is to construct something similar (to the denoted
object, i.e., either the generic property or the individual) just as a flour lion (is
made similar either to the generic property lion-ness or to an individual lion).
Something could be constructed similar to the falcon class property which is
different from other class properties, but could not be constructed similar to a
specific or exclusive property of falcons which differentiates one or more
falcons from other individual falcons. The injunction cannot command that one
make something similar to the plain individual, irrespective of all general and
specific properties, because the individual as such is predicateless and the same
in all substances. And if one selected one individual to which to make the altar
similar, one would not be able to make it similar to any other individual.
Therefore, the action of building a falcon-altar is not possible if the object
denoted by a word is an individual. MD, part 2, pp. 246.
Further on Kumarila says: ...dkrtih paddrtha iti vijiidyate. Syenacitam cinvfta iti
sravanat. Atra hi syenavyaktim cayanena kurydd dkrtim veti vdkydrthau sydtdm.
Ydvad istakdbhih syenavyakteh kartum asakyatvdt sndyvddibhir apy anirvrtter
istakdbddhdc chyenavyakteh prayojanakalpandc cinoter amukhydrthatvdt
Karmany agnydkhydydm iti sm.rtiparityagaprasangac ca cayanena syena-
vyaktim kurydd ity evam tdvan nopapadyate. Tathdakrter apistakabhih kartum
asakyatvdd daivanirmitatvdt prayojanakalpandc cinoter ayathdrthatvaprasahgdt
smrtibddhdc cdk,rtim kurydd ity api nopapadyate. Parisesdc chyenam iva citam

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Nanu ubhayatra kriyaya asambhava eva vyapadisyate: Ndkrti

sabdarthah. Kutah? Kriyd na sambhaved dkrtau sabdarthe: vrtht
proksatfti. Tathd na vyaktih sabddrthah, kriyaiva na sambhaved
vyakteh sabddrthatve: syenacitam cinvlteti. Yad apy ucyeta
vrthln proksatlti vyaktilaksandrthdkrtir iti. Sakyam anyatrdpi:
syenacitam cinvlteti vaditum: dkrtilaksandrthd vyaktir iti. Kim
punar atra jyya.h?
(Objector:) B1) It has been pointed out that action is impossibl
on both views: a) The class property is not the object denoted by
a word because action would be impossible if the object denoted
by a word were a class property; for instance (in the sentence
"One washes rice" (one cannot wash the class property of rice; one
washes specific individual grains).13 b) Likewise, an individual is
not the object denoted by a word for the same reason, action
would be impossible if the individual were the object denoted by
a word; for instance (in the sentence), "One should build a falcon
altar" (one cannot build a real individual falcon; one builds the
shape). 2a) The proponent (of the view that the class property is
the object denoted by a word) would answer that (in the example)
"One washes rice," the class property (being denoted) serves the
agnisthalam cayanena nirvartayed iti vdkydrthah, tatas ca yayd kaydcic chyena-

vyaktyd sadrsasydgnes cetum asakyatvdt sarvavyaktisadrsydsambhavdd
atftdndgatavyaktisddrs.yanupapattes ca syendkrtisddrsayasaampattisam bhavdc
cckrtih sabddrtha iti nisc[yate. We determine that a class property is the object
denoted by a word because of the Vedic text, "One should build a falcon altar."
Two sentence meanings are possible here: by piling up (bricks) one should
make either a) an individual or b) a class property. a) It is impossible to make
an individual falcon with bricks, and one doesn't create one using tendons, etc.
because the fact that bricks must be used prevents it. Hence, because one would
have to posit a purpose (of the injunction) for an individual falcon, one would
(interpret the verb) 'to build' in a secondary meaning and so one would have to
abandon the (grammatical) rule, Panini 3.2.92 (which states that the verb 'to
build (v/ci)' compounds with a direct object to form a name of an altar).
Therefore, the injunction (to build a falcon altar) does not make sense meaning
that one should make an individual by piling up. b) Likewise, it is not possible
to make a class property with bricks either because it is created by divine power.
Hence, because one would have to posit a purpose (for the injunction), one
would have to interpret (the verb) 'to build (Vci)' in an untrue sense which the
(aforementioned) rule would prevent too. Therefore, the injunction (to build a
falcon altar) doesn't make sense meaning that one should make a class property
either. What remains is the following sentence meaning: One should create an
altar, i.e a place for fire, like a falcon by piling up bricks. It is impossible to
build an altar similar to any one particular individual falcon, similarity to all
falcons is impossible, and similarity to past and future falcons does not make
sense; but it is possible to bring about the similarity (of an altar) to the falcon
class property. Therefore, we conclude that the class property is the object
denoted by a word. MD, part 2, p. 259-60.
13Katyayanasrautasutra 1.9.1 (part 1, pp. 70, 96) identifies rice (vrlhi) as the
oblation (havih), 2.2.9. concerns washing it. This is what the 'washing
(proksana)' mentioned in IIA refers to. The statement, "vrfhfn proksati," is also
quoted by Sabara on 9.1.11 and 9.1.13, and discussed under 9.2.40. MD, part 6,
pp. 81, 85 and 164.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
purpose of making known the individual. b) On the other view

too (in the example) "One should build a falcon," one can say that
the individual (being denoted) serves the purpose of making
known the class property (shape to build). Which is better here?
Akrtih sabddrtha iti. Yadi vyaktih sabddrtho bhaved,

vyaktyantare na prayujyeta. Atha vyaktyantare prayujyate, na
tarhi vyaktih sabddrthah, sarvasdmdnyavisesavinirmukta hi
vyaktir ity ucyate.
(Proponent:) C) That the class property is the object denoted by a
word is better. la) If an individual were the object denoted by a
word, the word would not be used for another individual. b) If it
is used for another individual, then the individual is not the object
denoted by the word because only that which is devoid of all
general and specific properties is called an individual.
Naisa dosah. Vyaktyantare sarvasamanyavisesavinirmukta eva

pravartisyate.
(Objector:) 2) This fault does not occur. The word will denote
another individual which is in fact devoid of all general and
specific properties.
Yadi vyaktyantare sarvasamdnyavisesavinirmukte pravartisyate,

sdmdnyam eva tarhi tat.
(Proponent:) a) If it denotes another individual devoid of all
general and specific properties, then that which it denotes is
precisely a general property.14
Nety ucyate. Yo hy arthah samanyasya visesanam casrayah sd

vyaktih. Vyaktivacanas ca sabdo na samanye na visese vartate,
tesdam tu asrayam evabhidadhdti. Tena vyaktyantare vrttir adosah.
Na hi tat sdmanyam.
b) (The objector) says: No. That object which is the substratum
of the general property and specific properties is the individual.
A word for an individual is used neither for the general property
nor a specific property but denotes just their substratum.
Therefore, it is not a fault (for the view that the word denotes an
individual) that a word is used for another individual. An
individual simply is not a general property.
14Biardeau (1964: 170) takes 'tat' to refer to the word rather than the individual.
Pataijali on 1.2.64 varttika 41 explains how the word 'vrksa', being common to
all individual trees, would be a single entity occurring in many substrata at
once. It is possible that Sabara has this in mind. However, the objector's answer
contains no argument against a word being a general property. Whereas by
distinguishing an individual from a general property it denies that an
individual is a general property.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Yadi vyakyantaresv api bhavati, sarvasamanyavisesa-

vinirmuktdydm asvavyaktau gosabdah kimiti na vartate?
(Proponent:) 3a) If a word is used for other individuals too, why
doesn't the word 'cow' denote an individual horse which is devoid
of all general and specific properties?
Aha yesv eva prayogo drstas, tesu vartisyate, na sarvatra, na

cas'vavyaktau gosabdasya prayogo drstah; tasmdt tatra na
vartisyate.
(The objector) says: Because it denotes only those for which its
usage has been seen, not all (individual objects), and one has not
seen the word 'cow' used for an individual horse; therefore, it will
not denote one.15
Yadi yatra prayogo drstas tatra vrttih, adya jdtdyam gavi

prathamaprayogo na prdpnoti, tatrddrstatvdt. Sdmdnyapratyayas
ca na prapnoti: iyam api gaur iti, iyam api gaur iti. Iyam va gaur
iti, iyam vd gaur iti sydt. Bhavati tu samdnyapratyayo
'drs.tapurvayam api govyaktau. Tasmdn na: pratyaydpekso
gosabdo vyaktivacana iti sakyata dsrayitum.
(Proponent:) i) If the word 'cow' denoted those individual objects
for which it's usage has been seen, it's first usage for a cow born
today would not obtain because it has not been seen.16 ii) Also
the cognition of the general property, "this is a cow and that is a
cow too," would not obtain. Rather, the cognition "either this is a
cow or that is a cow" would arise.17 But the cognition of the
general property does occur even with respect to an individual
cow not previously seen. Therefore, one cannot resort to the
explanation that the word cow denotes those individuals for
which its usage has been seen.
15Sabara has the objector express the view that denotation is natural as
expressed in the Mahabhasya on 1.2.64 vdrttikas 29, 33 and 34. See ?1.3.1.4.
16Kumarila: Gosabdoccdrane hi pirvam evdg,rhitdsu vyaktisu samanyam
pratiyate; taddkdrajninotpatteh pascdd vyaktayah pratlyante 'tas cdkrti-
pratyayasya nimittdntardbhdvdd vyaktipratyaye ca ptrvapratltasdmanya-
nimittatvdd dk.rtih sabddrtha iti vijfnyate. When the word 'cow' is uttered, one
immediately apprehends the general property before one has grasped the
individuals. One apprehends individuals after the knowledge of the general
property has arisen. Hence, because the cognition of the class property does not
have a cause other than the word and because the previously apprehended
general property is the cause of the cognition of an individual, we conclude that
the class property is the object denoted by a word. MD, part 2, p. 261.
17Kumarila: Yadi ca vyaktayo 'bhidheyd bhaveyus tatas tdsdam citrakhanda-
mun.ddivisesasvarupagrahanad vicitrd sabdoccdrane buddhih sydt. Ekdkdrd
tutpadyate. Tendpy dk.rtih sabddrtha iti nisciyate. Furthermore, if individuals
were denoted, then, because one would grasp their specific variously colored,
broken, shaven natures, each use of a word would produce a different cognition.
But each use produces the same cognition. For this reason too we conclude that
the class property is the object denoted by a word. MD, part 2, p. 261.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Evam tarhi sakteh svabhdva esah, yat kasydmcit vyaktau vartat

kasyadmcin na, yathd agnir usnah, udakam sitam, evam etad
bhavisyatiti.
(Objector:) b) Then it is the nature of the signifying capacity of a
word that it denotes one individual and not another. Just as fire is
hot and water cold (because such is its nature), so this will be (a
word denotes the individuals it does because such is its nature).
Naivam sidhyati. Na hy etad gamyate: kasydmcid vyaktau

vartate, kasydmcin neti.
(Proponent:) It doesn't succeed that way. One does not just
understand that the word 'cow' denotes one individual and not
another arbitrarily.
Satyam etat. Gotvam laksanam bhavisyati: yatra gotvam tasydm

vyaktdv iti.
(Objector:) 4) That's true. Cowness will be the identifying mark;
the word 'cow' will denote each individual in which there is
cowness.
Evam tarhi visista vyaktih pratiyeta. Yadi ca vis'i

visesanam avagamyeta. Na hy apratlte visesane
pratyetum arhanttti.
(Proponent:) a) In that case one would apprehe
which is qualified. b) And if qualified, one would
qualifier first. No one can apprehend something
the qualifier is apprehended.
Astu vis'esanatvendkrtim vaksyati, visesyatvena

dkrtipadarthakasya vyaktir na paddrtho, vyakti
nakrtih. Ubhayam ubhayasya paddrthah. Kasyacit kimcit
prddhanyena vivaksitam bhavati. Tendtrakrtir gunabhdvena
vyaktih pradhdnabhavena vivaksyata iti.
(Objector:) D1-2) Let it be that a word will denote the class
property as qualifier, an individual as qualified. Because it is not
the case that an individual is not denoted for him for whom a
class property is denoted, nor that a class property is not for him
for whom an individual is denoted. Both are denoted for both.
For each something is intended as principal.18 So here an
individual is intended as principal and the class property as
subordinate.
18The segment of text, "It is not the case that .... meant as principle (Na hy dkrti-
padarthakasya .... prddhdnyena vivaksitam bhavati)," is a close paraphrase of
Patafijali on Panini 1.2.64, vdrttika 53 (53c).

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Naitad evam. Ubhayor ucyamdnayor gunapradhanabhavah syat.

Yadi cdtrdkrtih pratfyate sabdena tadd: vyaktir api paddrtha iti na
s'akyate vaditum.
(Proponent:) E) It is not so. If both were denoted, one could be
principal and the other subordinate. But if one apprehends the
class property due to the word, then one cannot say that the
individual is also the object denoted by the word.19
Kutah?
(Objector:) Why?
Akrtir hi vyaktyd nityasambaddhd, sambandhinydm ca tasydm

avagatdydm sambandhyantaram avagamyate. Tad etad
dtmapratyaksam yac: chabda uccarite vyaktih prattyata iti. Kimn
sabddd utdkrter iti vibhcgo na pratyaksah. So 'nvaya-
vyatirekdbhydm avagamyate. Antarenadpi sabdam ya dkrtim
avabudhyeta, avabudhyetaivdsau vyaktim. Yas tuccarite 'pi sabde
mdnasdd apacdrdt kaddcid dkrtim nopalabheta, na jdtucid asdv
imam vyaktim avagaccheta.
(Proponent:) 1) Because the class property is inseparably related
to the individual, and once one knows that relatum (the class
property) one knows the other relatum. 2a) It is self-evident that
one apprehends the individual when a word is uttered. b) But the
distinction, whether (one apprehends it) from the word or from
the class property, is not evident. 3) That is determined by means
of concomitant presence and absence. a) Even without the use of
the word, whoever apprehends a class property, certainly
apprehends an individual. b) But even upon hearing the word,
whoever should fail to apprehend the class property at some time
because his mind is distracted, would certainly not understand that
individual.
19Kumarila: Jdtis cet purvam abhidheyatvendbhyupagatdtraiva

sabdasyopaksTnas'aktitvdn na vyaktivacanatd labhyate. Na cdkrtisamvedanend
'pi vyaktipratfter anyathdpy upapannatvdt satydm gatdv anekdrthatd yujyate
'bhyupagantum. Arthdpattyd ca sabdasya vdcakasaktih kalpyate. Sd ca yadd
vyaktipratiter anyathdpy upapannatvdt ksLnd tadd vdcakasaktikalpandydm
pramdnam ndsti. If the generic property is accepted as denoted first, then,
because the word has used up its capacity to denote there, one cannot accept that
it denotes an individual. Nor is it correct to accept that there is more than one
object denoted (the individual in addition to the class property) while there is
another (more economical) way, because the fact that there is cognition of an
individual along with knowledge of the class property does make sense
otherwise. And one posits a word's capacity to denote by means of
presumption. When presumtion fails because the cognition of an individual
makes sense otherwise, then there is no mode of evidence by which to posit the
capacity to denote (the individual too). MD, part 2, p. 263.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Nanu vyaktiviisisdyam akrtau vartate?

(Objector:) F) Can't the word denote the class property as
qualified by the individual?20
Vyaktivisis.tdydm ced varteta, vyaktyantaravisistd na pratiyeta

Tasmac chabda dkrtipratyayasya nimittam, dkrtipratyayo
vyaktipratyayasyeti.
(Proponent:) 1-2) If the word denoted the class property as
qualified by an individual, one would not apprehend it as
qualified by another individual. Therefore, the word is the cause
of the cognition of the class property, the cognition of the clas
property is the cause of the cognition of the individual.
Nanu gunabhutd pratlyata ity uktam.

(Objector:) G1) But we have said that one apprehends the clas
property as subordinate.21
Na gunabhavo 'smatpaksasya bddhakah. Sarvathd tdvat pratlyate

Arthdd gunabhdvah pradhdnabhdvo va. Svdrtham ced uccdryate
pradhanabhuitd. Atha na svdrtham, pardrtham eva, tato guna
bhuitd. Na tatra sabdavydpdro 'sti.
(Proponent:) The fact that one apprehends the class property as
subordinate doesn't refute our view. 2) It doesn't matter how on
apprehends it. a) The object determines what is principal and
what subordinate in the cognition. i) If a word is uttered for th
sake of its own denoted object, then the class property is
principal; ii) if not for the sake of its own object but just for th
sake of another object, then the class property is subordinate. b
The word is not engaged in this.
Nanu ca danditi na tdvad dandisabdena dando 'bhidhlyate atha ca

dandavisisto 'vagamyate. Evam ihdpi na tdvad dkrtir abhidhtyat
atha cdkrtivisistd vyaktir gamyeteti naitat sddhu.
(Objector:) 3) But it is not the case that the word 'staff-bearer'
denotes a staff and yet we understand one who is qualified by a
20The objector retreats to a new position: the individual is subordinate an

qualifies the class property which is the principal meaning.
21See "Tendtrdkrtir gunabhdvena... (So here the class property is meant
subordinate and the individual as principal (IIID))." The objector cit
Patafijali's statement that both the class property and the individual must
included in the meaning which the word denotes in order to establish the vi
that the word denotes the individual as qualified by the class property (D). T
proponent accepted that one cognizes both the class property and the
individual (E2a). However, he expressed at that point already that the cogniti
one has upon the use of the word is not necessarily the denoted meaning (E2
The objector confuses the two again here and the proponent clarifies that whic
element is principal and which subordinate have nothing to do with what th
word denotes.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
staff. Similarly here too it is not the case that the word denotes
the class property and yet one understands the individual
qualified by the class property. Hence this is not correct (that the
word denotes the class property alone, yet the cognition is of the
individual qualified by the class property).
Ucyate: Satyam dandisabdena dando ndbhidhlyate. Na tv

aprattte dande dandipratyayo 'sti. Asti tu dandisabdaikades'abhuto
dandas'abdo, yena dandah pratyayitah. Tasmdt sddhv etad yat
pratlte visesane vis'istah pratlyata iti. Na tu gosabddvayavah
kas'cid dkfrteh pratydyakah, anyo vyakteh; yata ucyate: tata dkrtir
avagatd, na gosabda dkrtivacana iti. Na ca yathd dandis'abdo na
dande prayuktah evam gos'abdo ndkrtau. Tadartham eva
nirdarsitam kevaldkrtyabhidhdnah syenasabda iti. Tad evam
anvayavyatirekdbhydm asati syenavyaktisambandhe syena-
sabdoccdrandd: dkrtivacana iti gamyate. Na tu vrlhydkrti-
sambandham antarena vrihivyaktau sabdasya prayogo drstah.
Tasmdd dkrtivacanah sabda ity etaj jydyah. 33. (krtah
siddhdntah)
The (proponent) answers: a) It is true that the word 'staff-bearer'
does not denote a staff. However, one is not conscious of the
staff-bearer without being conscious of the staff. i) In this case
there is the word 'staff, present as part of the word 'staff-bearer',
which conveys the cognition of the staff. c) So it is correct that
one knows the qualified object only when the qualifier is known.
b) In the other case, however, there is no part of the word 'cow'
which conveys the cognition of the class property, other than
(what conveys the cognition) of the individual; because you
stated that the class property is understood from the individual,
the word 'cow' does not denote the class property. 4a)
Furthermore, it is not the case that just as the word 'staff-bearer' is
not used for the staff, the word 'cow' is not used for the class
property. For that very reason we showed the example of the
word 'falcon' which denotes just a class property. Because the
word 'falcon' is uttered when there is no connection with an
individual falcon, by concomitant presence and absence o
knows that it denotes a class property. b) But one never se
word used for particular rice grains without a connection with
rice class property. Therefore, the view that a word denotes
class property is better.22
22Kumarila: Syenacid ity ddav dkrtau prayogadarsandd anvayavyatirekdbh

ca jdtir eva vdcyety avadhdryate syenacidvdkye kevaldydm dkrtau prayukta
Kvacid api ck.rtyd vind vyaktimdtre prayogddarsandd...tendkrtir eva sabddr
iti. 33. Because we see a word used for a class property in 'falcon-altar' a
on, and concomitant presence and absence determine that a generic prop
alone is the denoted object because in the falcon altar sentence the word, 'fa
is used for the class property alone. And because we never see a word use

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
1.3.34. Na kriya syad iti ced arthdntare vidhanam na

dravyam iti cet.
IVA) (The following objections were raised above: if the word
denoted the class property) action wouldn't take place,
enjoining (action) on a replacement (wouldn't make sense),
there wouldn't be (any word denoting) a substance.
Atha yad uktam, Na kriyd sambhaved: vrthin proksatiti na
dravyas'abdah sydt: Sad deyeti anyavacanam ca na syat: anyam
tadripam iti, tat parihartavyam. 34.
We must respond to the objections which were raised: (If the
word denoted the class property) 1) no action would be possible
(given the implied command to wash rice in the statement) "One
washes rice" (IIA). 2) There would be no word denoting a
substance (in injunctions such as) "One should give six (cows)"
(IID). 3) There would be no mention of another (as there is in
the phrase) "another of the same color" (IIE).
1.3.35. Tadarthatvdt prayogasydvibhdgah.
B) Because (a word) has (a class property) as its meaning,
its usage is not divorced (from an individual qualified by
the class property).
Akrtyarthatvdc chabdasya yasyd vyakter dkrtyd sambandhas tatra
prayogah. Proksanam hi dravyasya kartavyatayd sruyate.
Katamasya? Yad yajatisddhanam apurvayuktatvdt tasya.
Ndkrteh, asakyatvdt. Tatra vrthisabda dkrtivacanah prayujyate
proksanasrayavisesanaya. Sa hy dkrtim pratydyayisyati, akrtih
pratltt satl proksanasrayam viseksyatiti. Tendkrtivacanam na
virudhyata iti. Evamr: Sad deyd gavo daksineti daksinddravye
samkhydydh prayoktavye: gdva ity dkrtivacano visesakah. Tathd:
anyam iti vinastasya pratinidher anyatvasambandhah. Tatra
pasuusabda dkrtivacana dk.rtyd viseksyatlti. Tasmdd: gaur, asva ity
evamddayah sabdad dkrter abhidhdyakd iti siddham. 35.
1) Because the word denotes the class property, it is used for an
individual which has a relation with the class property. "One
washes rice" directly states washing as something which must be
done to a substance. Why to a substance? Because, due to the
fact that it directly participates in the Vedic performance, it is
connected to the transcendental result.23 The statement does not
just an individual without a class property,...; therefore, just the class property
is the object denoted by a word. MD, part 2, p. 266.
23Literally, "To which type of entity? To that type of entity which directly
participates in the Vedic performance, because such an entity will be connected
to the transcendental result. Not to a class property because it can't be
(connected to the transcendental result because it doesn't participate directly in
the performance)." The question "Katamasya?" asks what is so special about a
substance that singles it out from the pool of entities as that to which washing
is done rather than some other entity.
One could also interpret the question, "What kind (Katamasya)?" as showing
that if the sentence were divided it would require something else to complete its

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
mention washing as something to be done to a class property

because it can't be. This being the case, the word 'rice' which
denotes the class property is used to qualify the recipient of
washing.24 For it will make known the class property and the
class property being known will qualify the recipient of washing.
Therefore it is not contradicted that the word denotes the class
property. 2) In the same way, in the injunction "One should give
six cows as fee," when the number must be connected with the
substance which constitutes the fee, the word 'cows', denoting the
class property, qualifies it. 3) Likewise, when the scripture
enjoins getting an 'other' (goat), there is a connection of otherness
between the deceased and its replacement.25 The word 'animal',
denoting the class property, will qualify the other substance with
the class property. Therefore, it has been proven that words such
as 'cow', 'horse', etc. denote class properties.
sense. The sentence, "The word 'rice', which denotes the class property, is used
there to qualify the recipient of washing. (Tatra vrihi...)," shows that the word
denoting the class property fulfils the sense. The point is that if the sentence
were divided the sense of the parts would be incomplete. This, together with the
fact that the usage is for the sake of action, implies that it is one sentence.
Sabara shows how the class property and substance are construed as the same
object in the meaning of that single sentence.
24The substratum (asraya) of an action is that in which the action directly takes
place. It is either the agent or direct object of the action or both depending on
the particular action. In the case of washing, both the agent and the direct object
are substrata but from the use of the word 'rice' it is clear that Sabara is
discussing the direct object, not the agent.
25Remember that the relation of difference requires that the related elements be
substances because there is only one class property but many substances of the
type (IIE).

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
General Conclusion
We have examined the discussions concerning the

denotation of generic terms (common nouns) in Patanijali's
Mahabhasya, Vatsyayana's commentary on the Nyayasutras of
Gautama, and Sabara's commentary on the Mimamsasutras of
Jaimini. In the course of our discussion we have also considered
what the terms 'dkrti', 'jti', and 'dravya' refer to in each of the
three texts. Each of the discussions includes consideration of the
extreme views that a generic term denotes only an individual,
and that it denotes only a generic property. In Nyaya, a third
extreme position, that a generic term denotes only a shape, is
considered too. Each of the discussions also considers a moderate
view that a generic term denotes both a generic property and an
individual. In the case of Nyaya, this moderate view includes the
shape as a third element in the denoted meaning. On the
moderate view, one of the elements in the meaning is principal
and the rest subordinate in accordance with the intention of the
speaker. Most important, we have examined the arguments for
each of these views and the role that certain principles of
procedure and epistemology play in them.
In his discussions under Panini 1.2.64 and 1.2.58,
Pataijali uses the terms 'dk.rti' and jadti' synonymously to mean a
generic property, that is, a property wholly present in each of
many individuals of a class. He also uses both terms to refer to
the class, the aggregate of all the individual objects in which that
generic property resides. He uses the term 'dravya' to refer to an
individual which is part of the class and substratum of the gener
property. Sabara and Vatsyayana occasionally use the term
'dravya' in this sense but generally use the term 'vyakti' instead
In Nyaya, the term jdti' is used strictly for a generic property
considered abstractly as the cause of the same cognition wit
respect to many objects of a kind. It is that by virtue of which a
individual is recognized as being of a certain kind. A generic
property is clearly distinguished from a shape, for which the term
'dkrti' is used. In the case of most material objects, the shape
serves as an inferential mark of the generic property. Patafijali
recognizes the use of the term 'dk.rti' for a shape. He recognizes
the role a shape plays in making known the generic property of
an object too. However, shape does not enter into his seriou
discussions of the denotation of a generic term; nor does he use
the term 'dkrti' in these discussions to mean shape. Because it is
interchanged with the term 'jdti' in different contexts, 'akrti' cannot
be understood to mean anything other than a generic property.
289

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Sabara's use of the term 'dkrti' raises some doubts. He

does not use the term 'jti' and uses the term 'dkrti' for the
property which the individuals of a class have in common. In the
Akrtyadhikarana it is not absolutely clear whether he means the
term to refer to an abstract generic property or to a shape. In
certain contexts he clearly uses the term 'dkrti' for common
properties other than a shape, thus implying that he conceives of
an dkrti generally as an abstract generic property just as Patanijali
does, and just as Vatsyayana conceives a jdti. However, Sabara
uses the term 'dkrti' for a shape in another context central to his
argument to establish that a generic term denotes a generic
property. The success of the argument depends on closely
associating, or entirely identifying, the two senses of shape and
generic property.
Pataiijali considers two extreme views of what a word
denotes: the view that it denotes only an individual and the view
that it denotes only a generic property. The view that a generic
term denotes only an individual has two variations, depending on
whether one accepts the principle that a word denotes a single
object or allows that it may denote more than one object. If a
word can denote no more than one object, a problem arises in the
case of duals, which refer to two objects, and plurals, which refer
to many objects. According to the principle that one word
denotes one object, to denote many objects many words would
arise. But the accepted usage is that just one word is used with a
plural inflectional termination. To account for this, one would
have to state explicitly that just one of homophonous speech
forms remains and the others are deleted. Such an explicit
statement involves prolixity. A simpler explanation, according to
which an additional rule does not have to be stated explicitly,
would be preferred. Speech forms are used in order to give
knowledge of the object they denote. It is preferable to account
for their use by showing how they arise naturally, based on
certain principles and the ordinary conception of the object. One
method by which this can be done is to reject the principle that a
word denotes just one object. The ultimate criterion in
determining what a word denotes is ordinary usage. In ordinary
usage a singular speech form is used to denote one object, a dual
is used to denote two objects, and a plural is used to denote many
object. Hence one can account for the use of the plural to denote
many objects just by saying that it does so naturally.
This explanation is unsatisfactory because it does not
allow the procedure of determining the meaning of a speech form
by concomitant presence and absence to be carried out to its full
extent. One determines that a speech form denotes a meaning
when in the presence of that speech form that meaning is
understood and in the absence of that speech form that meaning is
not understood. When singular, dual, and plural forms of a word

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
GENERAL CONCLUSION 291
are used, the nominal base recurs and so does a certain cognition.
The basis of this cognition is said to be a generic property
common to each of the individual objects for which the generic
term is used. The simplest explanation for the use of the same
nominal base in each usage of a generic term is that the term
denotes a generic property. In this way the usage of a generic
term to refer generally to any individual of a kind rather than just
to a single particular individual is accounted for too.
There are drawbacks to the view that a generic term
denotes only a generic property. It has difficulty accounting for
the occurrence of different inflectional terminations in different
genders and numbers, and difficulty accounting for the syntactic
connection of a generic term with verbs and quality words. It is
impossible for a generic property, which is a single abstract
object, to have differences, to be associated with number greater
than one, or to participate in action. The simplest explanation for
the occurrence of gender and number suffixes is that they arise
after a nominal base denoting the object to which the semantic
conditions for those suffixes belong. The semantic conditions for
gender and number suffixes are the properties of gender and
number. These belong to individual substances, not to a generic
property.
The view that a generic term denotes only a generic
property accepts that the individual substances in which the
generic property resides are the semantic conditions for different
gender and number suffixes and for the use of a generic term in
syntactic connection with verbs and quality words. Yet it
maintains that the individual is not denoted. This is counter-
intuitive. Hence Pataijali adopts the view that both the generic
property and the individual are denoted. The generic property
accounts for the recurrence of the same cognition with respect to
all individuals of a class and is the semantic condition for the use
of a single nominal base even in the plural. Including the
individual as denoted too accounts for the use of a generic term in
syntactic connection with verbs and different quality words, and
for the use of different gender and number suffixes.
Similar arguments are put forth in Nyaya for the inclusion
of an individual and a generic property in the meaning of a
generic term. In addition, Gautama includes the shape which is
the means by which one infers what an individual substance is.
Of particular interest is his argument against accepting that the
shape alone is denoted. He rejects such a view by saying that a
word denotes the individual which has the shape, and which is
directly connected with the generic property, rather than the
shape itself. Suppose the general shape of a cow is the inferential
mark of the generic property cowness. The generic term denotes
an individual which is the substratum of the cow shape only if the

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
individual is also the substratum of the generic property cowness.

The generic term 'cow' does not denote a model cow.
Sabara's discussion of the denotation of generic terms in
the Akrtyadhikarana borrows considerably from the Mahabhasya.
Sabara both argues against views presented in the Mahabhasya,
and adopts principles and arguments from it. Occasionally he
closely paraphrases it. Sabara rejects the view that a generic term
denotes only an individual by developing the argument from the
Mahabhasya that if a generic term denoted a particular individual
it would not refer to another individual of the kind. On one view
expressed in the Mahabhasya, accepting ordinary usage as the
ultimate criterion with regard to what a word denotes justifies
setting aside the principle that one word denotes just one meaning
and the procedure of determining meaning by concomitant
presence and absence. In criticism of this position, Sabara rejects
the view that a word denotes what it does according to what is
seen in ordinary usage, because this view does not account for the
use of ageneric term for a new object of the kind.
Sabara rules out the view that a generic property serves to
indicate which individual objects are denoted without itself being
denoted, because he accepts the principle that an object cannot be
known as qualified by so'mething without the qualifier being
known first. Cognition of an individual as qualified by its generic
property cannot arise unless cognition of the generic property
arises first. Because usage of a generic term leads to knowledge
of an individual as belonging to a certain class, cognition of the
class property qualifying the individual must arise independent of
the individual, directly from the generic term. Therefore, it must
be denoted by the generic term.
Patafijali recognizes the principle that one generally does
not use a word to denote an object which another word has
already made known. In grammar the principle is limited in
scope and emphasis. Sabara gives a similar principle central
importance in his discussion on the denotation of a generic term.
In Mimamsa the procedure to determine what a word denotes
includes, in addition to concomitant presence and absence, the
mode of evidence known as presumption. One presumes that a
word denotes an object, if one cannot account for cognition of
that object otherwise whenever a word is used. Conversely, it
cannot be accepted that the capacity of the word to denote is the
direct cause of cognition of an object, if that cognition can be
accounted for otherwise. As a result of this procedure, and in
accordance with the principle that one word denotes just one
meaning, Sabara argues against the view that the individual is
denoted along with the class property. Thus he rejects the final
views of grammar and Nyaya.
Sabara argues that concomitant presence and absence
show that a generic term denotes a class property and not an

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
individual. Whenever an individual is known upon the use of a

generic term, a class property is also known. But he denies the
converse. He claims that there are cases in which a class property
is known and an individual is not. He gives an example of the
generic term 'falcon' referring to a brick altar in the shape of a
falcon. This example is the cornerstone in his argument to
establish that a generic term denotes only a generic property. I
have mentioned that Sabara's use of the term 'akrti' in this
example raises some doubts concerning his use of the term a
his concept of the entity which it denotes.
The soundness of his argument depends on identify
the generic property with shape, or at least associating it wi
shape more closely than a generic property is associated with
own substratum. Because it is difficult to see what relationsh
other than identity is closer than that between a generic prop
and its own substratum, one is forced to concede either that
Sabara accepts shape itself as a generic property or that his
argument succeeds only by the trick of using the single term
'akrti' in two different meanings to achieve the appearance of an
identity where there is none. If the latter is the case, his argument
to establish that a generic term denotes only a generic property
fails. On the other hand, if he accepts that shape itself is a class
property, the clarity of his concepts must be called into question.
A brick altar in the shape of a falcon is not a falcon. But if shape
is a class property, namely, that by virtue of which an individual
is recognized as being of a certain kind, then if the altar has the
falcon shape it is a falcon. If it is a falcon, it is an individual
falcon. If it is an individual falcon, then the example fails to
serve the purpose it is intended to serve in his argument. In order
to prove that a generic term denotes only a generic property and
not an individual by concomitant presence and absence, he must
show a case in which a generic term is used for a generic property
and not for an individual. The falcon altar is adduced as such an
example. But if the falcon altar is an individual falcon, then it is
not true that upon the use of the generic term 'falcon', cognition
of a generic property arises and cognition of an individual does
not arise. Hence the example fails, and the argument from
concomitant presence and absence fails with it.
In order to save Sabara's argument, Kumarila asserts that
an akrti is an abstract generic property and not a shape but that
shape is more similar to the generic property than it is to an
individual. This is a tenuous assertion. The principal relation
between a shape and a generic property discussed by Kumarila
and in Nyaya is that of an inferential mark to the object to be
inferred. Kumarila is equivocal about accepting such a relation at
all. Even if he does accept this relation, or if he accepts some
other relation of qualification, the relation is mediated by the
individual. A shape is related to a generic property by inhering in

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
the same individuals in which the generic property inheres. If th

individual mediates the relation between a generic property and a
shape, it is impossible for knowledge of a generic property t
give knowledge of a shape without giving knowledge of an
individual. If a listener always arrives at knowledge of an
individual in addition to knowledge of a class property, again
concomitant presence and absence fail to show that only a class
property is denoted.
The only evidence for not including an individual in the
denoted meaning is the limitation imposed by virtue of adopting
presumption as part of the means to determine what a word
denotes. That limitation is that if one can account for cognition
of the individual otherwise one cannot presume that the wor
denotes it. However, the manner in which Sabara accounts for
knowledge of the individual otherwise, in each case of the use of
a generic term, depends on the very same limitation, namely, that
one not be able to make sense of the sentence otherwise.
Sabara explains how knowledge of a qualified individual
arises from a generic term as follows: The generic term denoting
a class property conveys the cognition of the class property. If
the word is used in syntactic connection with action, because it is
impossible to perform action on a class property whereas it is
possible to perform it on an individual, one understands that the
class property qualifies the individual. Hence one understands the
qualified individual because the sentence doesn't make sense
otherwise. Now according to Kumarila's interpretation, in the
case of the falcon altar, if a generic term denotes a generic
property, then when the generic term 'falcon' is used knowledge
of the generic property falconhood will arise. It is impossible to
perform building a brick altar on either the generic property
falconhood or on a real individual falcon. One performs it on a
shape because the sentence wouldn't make sense otherwise. As we
see, cognition of the individual is no more absent in the example
than cognition of the class property is. Either both are absent or
both are present, depending upon what is meant by cognition
being present. The example of the falcon was given by Sabara as
one in which cognition of an individual does not arise but that of
the class property does. Hence the example does not serve its
intended purpose. Without an example showing the presence of
cognition of the class property in the absence of cognition of an
individual, the argument from concomitant presence and absence
for the view that a generic term denotes only a generic property
fails.
In every case of the use of a generic term, one arrives at
knowledge of an individual or shape by reasoning that it is
impossible otherwise. This is precisely the reasoning by which
one establishes that a word has the capacity to denote a certain
meaning. One presumes that an object is denoted if it is

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
impossible otherwise to account for the cognition of it when

word is used. Conversely, one cannot presume that a gener
term denotes an individual because it is possible to account f
cognition of an individual otherwise. The manner in which one
accounts for cognition of an individual in every case is that it
impossible to account for the use of the sentence otherwi
Hence Sabara's argument against the view that an individual
denoted by a generic term in addition to a generic propert
depends on weighing one impossibility more than another.
depends on giving more importance to one presumption than t
another.
Although his arguments to eliminate an individual from
being denoted by a generic term are weak, the view that a generic
term denotes only a generic property is not thereby invalidated.
All that can be said of Sabara's conclusion is this: he considers it
more reasonable, that in each use of a generic term one presume
that an individual is the intended object, than that one presume
that the generic term denotes the individual in addition to the
generic property. Patafijali and Vatsyayana, on the other hand,
consider it more reasonable that a generic term denote an
individual as well as a generic property.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Abbreviations
Abbreviations found in the Notes
Kas. Kasika
KNP KaSikdvrtti... with Nydsa and Padam
MB Pataijali's Vydkarana Mahdbhdsya... E
Sarma, .
MB-K The Vydkarana-mahdbhdsya of Patanjali. Ed. Kielhorn
MBPU Mahdbhdsya of Pataijali... Ed. Guruprasada Sastri.
MD Mimdmsddarsana
MPV Mahdbhdsya-pradtpa-vydkhydndni
ND Nydyadarsanam
ND2 Nydyasutras
Nir. Nirukta. See Sarup, Laksman. 1920-27. Part 3. Reprint
ed.
NM Nydyamaijarl
PLM See Kapil Deva Shastri. 1975.
RVS Rgveda-samhitd: with the Commentary of Sdyanacadrya
SV Slokavdrttika
VS Vais'esikasiutra
Abbreviations found in the Bibliography
ASS Anandasrama Sanskrit Series, Pune

ABORI Annals of the Bhandarkar Oriental Research Institute,
Pune
ALB Brahmavidyd, The Adyar Library Bulletin, Adyar
(Madras)
ALS The Adyar Library Series, Adyar (Madras)
BORI Bhandarkar Oriental Research Institute, Pune
BSOAS Bulletin of the School of Oriental and African Studies,
London
BSPS Bombay Sanskrit and Prakrit Series
CSCRS Calcutta Sanskrit College Research Series
DCDS Deccan College Dissertation Series, Pune
FL Foundations of Language, Dordrecht
HOS Harvard Oriental Series, Cambridge
IA Indian Antiquary, Bombay
IHQ Indian Historical Quarterly, Calcutta
297

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
IIJ Indo-lranian Journal, The Hague

IL Indian Linguistics, Pune: Linguistic Society of India.
JAOS American Oriental Society Journal, New Haven
JBBRAS Journal of the Bombay Branch of the Royal Asiatic
Society
JGJRI Journal of the Ganganatha Jha Research Institute,
Allahabad
JIP Journal of Indian Philosophy, Dordrecht and Boston
and London: Klumer Academic Publishers
JOR Journal of Oriental Research, Kuppuswami Sastri
Research Institute, Mylapore (Madras)
JRAS Journal of the Royal Asiatic Society of Great Britain
and Ireland, London
JUB Journal of the University of Bombay
JUP Journal of the University of Poona, Humanities Section
PCASS Publications of the Centre of Advanced Study in
Sanskrit, University of Poona
PEW Philosophy East and West: A Quarterly of Asian and
Comparative Thought, Honolulu: University of Hawaii
Press
PIFI Publications de l'Institut fran9ais d'Indologie,
Pondichery
RSCG Shri Rajasthan Sanskrit College Granthamala, Varanasi
SVUOJ Sri Venkateswara University Oriental Journal, Tirupati
TPS Transactions of the Philological Society, London
UPSPS University of Poona Sanskrit and Prakrit Series, Pune
WZKS Wiener Zeitschrift fur die Kunde Sidasiens, Vienna
ZDMG Zeitschrift der deutschen morgenldndischen
Gesellschaft, Leipzig and Wiesbaden

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Bibliography
Primary Sources
Amarakosa (Ndmalirgdnusasana of Amarasimha): with the

Rdmasrami or Sudhd Commentary of Bhdnujidiksita. Ed.
Sivadattadadhimatha. Bombay: Nirnaya Sagar Press.
[Reprint: Delhi: Caukhamba Samskrta Pratisthana, 1985.]
Atharvavedasamhitd. Ed. Sripad Damodar Satavalekar. Aundh:

Svadhyaya Mandala, 1939. 4th ed.: Pardi: Svadhyaya
Mandala.
Carakasamhitd by Agnivesa: Revised by Caraka and Drdhabala

with the Ayurveda-Dipikd Commentary of Cakrapdnidatta.
Ed. Yadavasarman. Bombay: Nirnaya Sagara Press, 1941.
Reprint: Kashi Sanskrit Series, no. 288. Varanasi:
Chaukhambha Sanskrit Sansthan, 1984.
Kdrikdvall and Nydyamuktdvalf of Visvandthanydyapancdnana-

bhdttacdrya: with the Dinakari and Rdmarudrf
Commentaries. Ed. Atmaramasarman. Bombay: Nirnaya
Sagar Press, n.d. Reprint: Varanasi: Chaukhamba Sanskrit
Series Office.
Kdaikd: A Commentary on Pdnini's Grammar by Vdmana and

Jaydditya. Part 1 (adhyayas 1-4) and part 2 (adhydyas 5-8),
eds. Aryendra Sharma, Khanderao Deshpande, and D. G.
Padhye. Part 3, Indices, ed. B. R. Sastri and V. Sundara
Sarma. Sanskrit Academy Series nos. 17, 20, 23.
Hyderabad: Sanskrit Academy, Osmania University, 1969,
1970, 1976.
Kdaikdvrtti of Jaydditya and Vdmana: Along with Commentaries

Vivaranapaicikd-Nydsa of Jinendrabuddhi and
Padamanjarl of Haradatta Misra. Ed. Srinarayana Misra. 6
Parts. Varanasi: Ratna Publications, 1985.
Kdthakasamhitd. Ed. Sripad Damodar Satavalekar. Aundh:

Svadhyaya Mandala, 1943. 4th ed.: Pardi: Svadhyaya
Mandala.
*o?
299

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Kitydyana`rautasi7tra with the Karkabhdsya. Ed.

PattabhirBmaS'stri. Ldl Bahddur Sdstri Kendriya Samskrta
Vidydpitha Series, no. 42. 2 parts. New Delhi, 1982-83.
Laghumafijiisd. See SabhZpati SarmB Upgdhydya. 1973.
MaitrdyanTsarmhitJ. Ed. Srip5d Damodar Satavalekar. Aundh

Svidhyaya Mandala, 1942. 4th ed.: PZriri: SvZdhydya
Mandala.
Manusmbrti: with the Hindi translation by Rdmesivara BhattJcdrya.

Bombay: Nirnaya Sagar Press, n.d. Reprint: Delhi:
Caukhamba Samskrt PratisthZn, 1985.
MahdbhJsya of Patafijali: with the Commentaries Mahdbhdaya-

Pradipa by Kaiyata and Pradtpoddyota of Sri NJgoji Bhatta.
Ed. Guruprasgda Sastri. 9 volumes. RSCG 30. Benares:
Bhargava Pustakalaya, 1938. Reprint: 9 parts in 8 volumes.
Varanasi: Vanivilas Prakashan, 1987.
MahdbhsyappradTpapraka-s'a by Pravartakopddhydya:
Commentaires sur le MahdbhQ'ya de Patafijali et le Pradipa
de Kaiyata. Ed. M. S. Narasimhacharya. PIFI 73.
Pondichery: Institut franqais d'Indologie, 1986.
MimJmsJdars'ana of Jaimini with Sabara's Bhdaya. Vol. 1,

TarkapJda with VaidyanJthas'a-strT's Prabhd Commentary
(adhydya 1, pdda 1), ed. TirthahalligrZmimbhijana
SubbZi'stri, reedited by KgSindtha Vdsudev Abhyankar and
GaneSaSgstri Ambaddsa Josi, 1976. Vol. 2, with Kumdrila's
Tantravdrtika (1.2 - 2.1). Vol. 3, with Kumdrila's
Tantravdrttika (2.2 - 2.4), 3d ed., edited by Gane'as'stri
Josi, 1980. Vol. 4, with Kumdrila's Tantravdrtika (adhydya
3), new ed., edited by Kdsinntha Visudeva Abhyamkar and
Ganedas'stri Arnb5d5sa Josi, 1984. Vol. 5, with Kumdrila's
Tupttkd (adhydiyas 4-7), new ed., 1984. Vol. 6, with
Kumdrila's Tup!Tkd (adhydyas 8-10), new ed., edited by
Kdfindtha Vasudeva Abhyamkar and Ganesa'astri Jogi,
1984. Vol. 7, with Kumdrila's Tup!tkJ (adhydyas 11-12),
new ed., 1985. ASS, no. 17. Pune.
MtmJrjzsd-nydya-prakdia by Apadeva: Edited with an Original

Commentary by Mahamahopadhyaya Vasudev Shastri
Abhyankar. 2d ed. Pune: BORI, 1972.
Mahdbhdcvya-pradTpa-vydkhydndni: Commentaires sur le

Mahdbhdaya de Patafijali et le Pradipa de Kaiyata. Ed. M. S.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
BIBLIOGRAPHY 301
Narasimhacharya. PIFI 51, 1-10. 10 volumes. Pondichery:

Institut fran9ais d'Indologie, 1973-83.
Nydyadarsanam: with Vdtsydyana's Bha.sya, Uddyotakara's

Vdrttika, Vdcspati Misra's Tdtparyatlka, and Visvandtha's
Vrtti. Eds. Taranatha, Nyaya-Tarkatirtha, and
Amarendramohan Tarkatirtha. Calcutta Sanskrit Series, nos.
18-19. Calcutta: Metropolitan Printing and Publishing,
1936-44. 2d edition in 1 vol.: Delhi: Munshiram
Manoharlal, 1985.
Nydyasuttras of Gautama: with Vdtsydyana's Bhasya and

Visvandtha-bhattdcdrya's V.rtti. 1985. ASS, no. 91. Pune.
Nydyamaijari of Jayantabhatta: with the commentary

Granthibhanga by Cakradhara. Ed. Gaurinatha Sastri. 3
parts. Varanasi: Sampurnanandasamskrtavisvavidyalaya,
1982-84.
Patahjalayogasutra: with the Bhas4ya of Vydsa and Tika of

Vdcaspatimis'ra. ASS, no. 47. Pune.
Pataijali's Vydkarana Mahjbhjsya: with Kaiyata's Pradlpa and

Nagesa's Uddyota. Vol. 1 (navdhnika), eds. Raghunatha
Sarma and Sivadatta Kudala. Sudhi Series 12. Varanasi:
Sudhi Prakashan, 1987. Vol. 2 (adhydyas 1.2 - 2), ed.
Sivadatta D. Kudala. Bombay: Nirnaya-Sagar Press, 1912.
Vol. 3 (adhydya 3), eds. Raghunatha Sarma and Sivadatta
Kudala. Bombay: Nirnaya-Sagar Press, 1937. Vol. 4
(adhydyas 4-5), ed. Bhargavasastri Joshi. Bombay:
Nirnaya-Sagar Press, 1942. Vol. 5 (adhydya 6), ed.
Bhargavasastri Joshi. Bombay: Nirnaya-Sagar Press, 1945.
Pramdnavdrttika of Acdrya Dharmaklrti: with the Commentary

'Vritti' of A cdrya Manorathanandin. Ed. Dvarikadas Shastri.
Varanasi: Bhauddha Bharati, 1984.
Prasastapddabh.sya (Paddrthadharmasanigraha): with Commentary

Nydyakandalt by Srldharabhatta along with Hindi
Translation. Ed. Durgadhara Jha. Varanasi: Sampurnanand
Sanskrit Vishvavidyalaya Research Institute, 1977.
Rgveda-samhitd: with the Commentary of Sdyanacarya. 2d ed.

Eds. N. S. Sontakke, C. G. Kashikar, et al. 5 Volumes.
Pune: Vaidika Samsodhana Mandala, 1972.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Satapathabrahmana: According to the Mddhyandina Recension

with the Veddrthaprakisa commentary of Sdyandcdrya. Part
1. Bombay: Laxmi Venkateshwar Steam Press, 1940.
Slokavdrttika of Kumdrila Bhatta: with the Commentary

Nydyaratndkara of Pdrthasdrathi Misra. Ed. Dvarikadasa-
sastri. Varanasi: Ratna Publications, 1978.
Taittirlyasamhitd. Ed. A. Y. Dhupkar. 1945. 4th ed.: Ed. Sripad

Damodar Satavalekar. Pardi: Svadhyaya Mandala.
Tarkasamgraha. See Kedaranatha Tripathi. 1977.
Ten Principal Upanishads with Sdaikarabhdsya. Works of

Saikaracarya, vol. 1. Delhi: Motilal Banarsidass, 1964.
Vaisesikasutra of Kandda: with the commentary of Candrdnanda.

Ed. Muni Sri Jambuvijayaji. Gaikwad's Oriental Series, no.
136. Baroda: Oriental Institute, 1961.
Vdkyapadfya of Bhartrhari. Eds. Abhyankar, Kashinath Vasudev,

and V. P. Limaye. UPSPS 2. Pune, 1965.
Vdkyapadtya of Bhartrhari with the Commentaries Vrtti and

Paddhati of Vrsabhadeva: Kdnda 1. Ed. K. A. Subramania
Iyer. DCMS 32. Pune: Deccan College, 1966.
Vdkyapadiya of Bhartrhari with the Commentary of Heldraja:

Kdnda 3, part 1 (samuddesas 1-7). Ed. K. A. Subramania
Iyer. DCMS 21. Pune: Deccan College, 1963.
Vdkyapadfya of Bhartrhari with the Praklrnaprakdsa of Heldrdja:

Kanda 3, part 2 (samuddesas 8-14). Ed. K. A. Subramania
Iyer. Pune: Deccan College, 1973.
Vakyapadiya of Bhartrhari (An Ancient Treatise on the

Philosophy of Sanskrit Grammar): Containing the Tkad of
Punyardja and the Ancient Vrtti: With a Forward by Ashok
Aklujkar: Kdnda 2. Ed. K. A. Subramania Iyer. Delhi:
Motilal Banarsidass, 1983.
The Vydkarana-mahdbhdsya of Patanjali. Ed. Lorenz Franz

Kielhorn. 3 vols. BSPS 18-22, 28-30. Bombay:
Government Central Press: Vol. I, 1880; Vol. 2, 1883; Vol.
3, 1885. Second edition: 1892, 1906, 1909. Third edition
revised and furnished with additional readings references
and select critical notes by K. V. Abhyankar. Pune: BORI,
1962, 1965, 1972. Reprint: 1985.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
BIBLIOGRAPHY 303
Secondary Sources
The titles of works written in modern foreign languages are not

translated. For all works with titles in more than one language,
the title in the language in which the work is written appears first
with the second title in brackets. If the work is in a foreign
language, the English title appears in brackets; if the work is
written in English but published in a foreign language journal,
etc., the foreign title appears in brackets.
Abegg, Emil. 1923. "Die Lehre von der Ewigkeit des Wortes bei
Kumarila." In Antidoron fur Jacob Wackernagel, 255-64.
Gottingen.
Abhyankar, Kashinath Vasudev, ed. 1962. The paribhasendu-

sekhara of Nagojibhatta: with the commentary Tattvddarsa
of MM. Vasudev Shastri Abhyankar. Pune: BORI.
.1967. Paribhssasamgraha: A Collection of Original Works

on Vyakarana Paribhadss. BORI Post-graduate and
Research Department Series, no. 7. Pune.
Abhyankar, Kashinath Vasudev and Jayadev Mohanlal Shukla,

eds., trans. 1975. Patanjali's Vyikarana-mahdbhdsya,
Ahnikas 1-3 with English Translation and Notes. BORI
Research Unit Publications 1. Pune.
Abhyankar, Kashinath Vasudev, and V. P. Limaye, eds. 1963-69.

Mahdbhasyadtpikd of Bhartrhari. Supplements to ABORI
43, 44, 45, 46, 47, 50. Reprint: BORI Post Graduate and
Research Series 8. Pune, 1967, 1970.
Agrawal, Dhannalal Javaharlal, ed. 1985. Mmadmsd-uddharana-

kosa (The citations from Sdbara-bh.sya traced to their
original sources). Pune: Vaidika Samsodhana Mandala.
Aklujkar, Ashok Narhar 1970. "Ancient Indian Semantics."

ABORI 51:11-29.
Bandyopadhyay, Nandita. 1982. "The Concept of Similari

Indian Philosophy." JIP 10: 239-75.
Bedekar, D. K. 1958. "Some concepts based on revelatory

Epistemology." ABORI.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Bergdahl, Roland. 1987. "The Sadhanasamuddesa (karikas 1-11)

of Bhartrhari's Vakyapadiya with the commentary Prakasa
of Helaraja: A Translation with an Introductory Essay and
Explanatory Notes." Ph.D. diss., University of
Pennsylvania.
Bhaduri, Sadananda. 1947. Studies in Nycya-Vaisesika

Metaphysics. Bhandarkar Oriental Series, no. 5. Pune:
BORI. Reprint: 1975.
Bhagavat, Vaman Balakrishna and Saroja V. Bhate, eds., trans.

1986, 1990. Mahabhasyadipika of Bhartrhari: Fascicules 3,
7: Ahnika 6, parts 1, 2. BORI Post Graduate and Research
Department Series 26, 34. Pune.
Bhandarkar, Ramkrishna Gopal. 1927-33. Collected Works, ed.

N. B. Utgikar. 4 vols. Pune: BORI.
Bhattacharya, Bishnupada. 1956. "Philosophical data in Patafnjali's

Mahabhasya." Our Heritage 4:51-65. Calcutta.
. 1962. A Study in Language and Meaning: A Critical

Examination of Some aspects of Indian Semantics. Calcutta:
Progressive Publishers.
Biardeau, Madeleine. 1960. Compte-rendu de Contribution a

l'histoire de la philosophie linguistique indienne, by D. S.
Ruegg. IIJ 4, nos. 2-3:170-72.
. 1964. Theorie de la connaissance et philosophie de la

parole dans le brahmanisme classique. Le monde d'outre-
mer passe et present, Premiere serie, etudes 23. Paris; Le
Haye: Mouton.
Bohtlingk, Otto. 1876. "Katyayana oder Patanjali im

Mahabhasya." ZDMG 29:183-90.
. 1887. Panini's Grammatik. Leipzig. Reprint: 2 vols. Olms:

Hildesheim, 1964.
Boudon, Pierre. 1938. "Une application du raisonnement par

l'absurde dans l'interpretation de Panini: les
jnapakasiddhaparibhasa." Journal Asiatique 230:65-121.
Paris.
Bronkhorst, Johannes. 1983. "On the History of Paninian

Grammar in the Early Centuries Following Patafijali." JIP
11, no. 4 (December): 357-412.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
BIBLIOGRAPHY 305
. 1987. Three Problems Pertaining to the Mahdbhasya. Post-

graduate and Research Department Series, no. 30. Pune:
BORI.
_ , ed., trans. 1987. Mahdbhdsyadlpikd of Bhartrhari: Fascicule

4: Ahnika 1. BORI Post Graduate and Research Department
Series 28. Pune.
Brough, John 1951. "Theories of general linguistics in the

Sanskrit grammarians." TPS 1951:27-46. Reprint: Staal
1972:402-14.
1953. "Some Indian Theories of Meaning." TPS 1953:161-

76. Reprint: Staal 1972:414-23.
1973. "I-ching on the Sanskrit Grammarians." BSOAS

36:248-60.
Buhler, George, trans. 1886. The Laws of Manu. Oxford. Fir

AMS edition: New York: AMS Press, 1971.
Cardona, George. 1967-68. "Anvaya and vyatireka in Indian

grammar." ALB 31-32, V. Raghavan Felicitation Volume,
313-52.
. 1975. "A note on the formulation of Panini 6.1.67." Annals

of Oriental Research, Silver Jubilee Volume: 11-20.
University of Madras, Sanskrit section.
. 1976. Panini: A Survey of Research. The Hague; Paris:

Mouton. Delhi: Motilal Banarsidass. 2d In Press.
. 1977-78. "Still Again on the History of the Mahabhasya."

ABORI 58-59, Diamond Jubilee Volume, 79-99.
. 1981. "On reasoning from anvaya and vyatireka in early

Advaita." In Studies in Indian philosophy: A memorial
volume in honour of Pandit Sukhlalji Sanghvi, ed. Dalsukh
Malvania and Nagin J. Shah, 79-104. Ahmedabad: L.D.
Institute of Indology.
. 1983. Linguistic analysis and some Indian traditions. The

Pandit Shripad Sastri Devdhar Memorial Lectures, 1st ser.,
no. 20. Pune: BORI.
1988. Panini: His Work and its Traditions. Vol. 1,

Background and Introduction. Delhi: Motilal Banar
2d In Press.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Chakravarti, Prabhat Chandra. 1930. The Philosophy of Sanskrit

Grammar. University of Calcutta.
__. 1933. The Linguistic Speculations of the Hindus.

University of Calcutta.
Chakravarti, Pulinbihari. 1951. Origin and Development of the

Sdmkhya System of Thought. Calcutta: Metropolitan
Printing and Publishing House. Second Edition: Delhi,
1975.
Charudeva Shastri, trans. 1968. Bhagavatpatanjali viracita

vydkarana-mahdbhdsya (prathama navdhnika) ka hindi
anuvdda tathd vivarana [The Mahdbhdsya of Pataijali (first
nine dhnikas) translated into Hindi with comments]. Delhi:
Motilal Banarsidass.
Chatterji, Kshitish Chandra, trans. 1957. Patanjali's Mahdbhdsya:

Paspasdhnika (introductory chapter). Usha Memorial Series
7. Second edition: Calcutta: A. Mukherjee, 1957. Reprint:
1964, 1972.
Chatterji, S. C. 1939. The Nydya Theory of Knowledge.

University of Calcutta. Fourth edition: 1972.
Chattopadhyaya, Debiprasad and Mrinalkanti Gangopadhyaya.

1967-73. Nydya Philosophy: Literal Translation of
Gautama's Nydyasuitra & Vdtsyayana's Bhcsya along with a
free and abridged translation of the Elucidation by
Mahdmahopddhyaya Phanibhiusana Tarkavdgisa. 5 parts
[part I (first adhyaya), 1967; part 2 (second adhyaya),
1968; part 3, 1972; part 4, 1973; part 5, 1976]. Indian
Studies Past & Present. Calcutta: R. K. Maitra, R. D. Press.
Chattopadhyaya Kshetresachandra. 1937. "References to Buddhist

Philosophy in the Vrttikdragrantha of Sabarabhdsya on the
Mimamsa sutras of Jaimini," In Jha Commemoration
Volume: essays on oriental subjects, 115-27. Poona
Oriental Series 39. Poona: Oriental Book Agency.
Clear, Edeltraud Harzer. 1990. "Isvarakrsna's Two-level

Perception: Propositional and Non-propositional," JIP 18:
305-40.
Clooney, Francis X. 1988. "Devatadhikarana: A Theological

Debate in the Mimamsa-Vedanta Tradition," JIP 16, no. 3
(September): 277-98.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
BIBLIOGRAPHY 307
__. 1990. Thinking Ritually: Rediscovering the Puirva

Mimdmsd of Jaimini. Publications of the De Nobili
Research Library, no. 17. Vienna.
Coward, Harold G. and K. Kunjunni Raja, eds. 1990. The

Philosophy of the Grammarians. Encyclopedia of Indian
Philosophies, gen. ed. Karl H. Potter, Vol. 5. Princeton:
Princeton University Press and Delhi: Motilal Banarsidass.
Danielsson, O. A. 1883. "Die Einleitung des Mahabhasya."

ZDMG 37:20-53. Leipzig.
Dasgupta, S. N. 1920. The Study of Pataiijali. Calcutta.
Dash, Subas Chandra. 1991. "Samaya: The Word Meaning

Relationship in Nyaya-Vaisesika System." JOI (Baroda) 41,
no. 1-2:57-66.
Devasthali, Govind Vinayak. 1942. "On the probable date of

Sabarasvamin." ABORI 23, Silver Jubilee Volume: 84-97.
Pune.
. 1948. "Sabara and the Nyaya-Vaisesika Darsanas." JOR

18:16-24. Madras.
. 1949. "Positive Data for the Date of Sabarasvamin." JGJRI

6:231-40.
. 1951. "Sabara and Patafijali." JUB 20 (part 2): 101-6.
. 1959. Mimdmsd: The Vdkya-Sdstra of Ancient India.

Bombay: Booksellers' Publishing Co.
Devasthali, Govind Vinayak and Gajanan Balakrishna Palsule,

eds., trans. 1989. Mahdbhdsyadtpikd of Bhartrhari:
Fascicule 6: Ahnika 4. BORI Post Graduate and Research
Department Series 32. Pune.
Deshpande, Madhav 1972. "Kaundabhatta on the philosophy o

nominal meanings: The text of Kaundabhatta's
Namarthanirnaya, with an English translation, explanation
and an introductory essay." Ph.D. diss., University of
Pennsylvania.
. 1992. The Meaning of Nouns, Semantic Theory in

Classical and Medieval India: Namarthanirnaya of
Kaundabhatta Translated and Annotated. Studies of
Classical India 13. Dordrecht: Kluwer.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Dravid, Raja Ram. 1972. The Problem of Universals in Indian

Philosophy. Delhi: Motilal Banarsidass.
D'Sa, Francis X. 1980. Sabdaprdmdnyam in Sabara and Kumdrila:

Towards a Study of the Mlmamsd Experience of Language.
Publications of the De Nobili Research Library, no. 7.
Vienna.
Edgerton, Franklin. 1928. "Some linguistic notes on the Mimamsa

system." Language 4:171-77.
Elayath, K. N. Neelakantan. 1989. "Sentence and its Meaning in

Advaita Vedanta." JOI (Baroda) 39, no. 1-2:51-56.
Filliozat, Pierre-Sylvain. 1973. Introduction to Mahdbhdsya-

pradlpa-vydkhydndni, ed. by Narasimhacharya, vol. 1, iii-
xiii.
1975-86. Le Mahdbhasya de Pataijali avec le Pradipa de

Kaiyata et l'Uddyota de Ndgesa. PIFI 54, 1-5. Vol. 1,
Adhyaya 1 Pdda 1 Ahnikas 1-4, 1975. Vol. 2, Adhydya 1
Pdda 1 Ahnikas 5-7, 1976. Vol. 3, Adhydya 1 Pada 1
Ahnikas 8-9, 1978. Vol. 4, Adhyaya 1 Pdda 2, 1980. Vol.
5, Adhydya 1 Pdda 3, 1986. Pondichery: Institut Franqais
d'Indology.
1991. An Introduction to Commentaries on Pataijali's

Mahabhasya. Post-graduate and Research Department
Series, no. 35. Pune: BORI.
Frauwallner, Erich. 1956. Geschichte der Indischen Philosophie.

Band 2, Die Naturphilosophischen Schulen und das
Vaisesika-system, das System der Jaina, der Materialismus.
Salzburg: Otto Muller Verlag.
. 1960a. "Das Eindringen der Sprachtheorie in die indischen

philosophischen Systeme." In Indologen-Tagung 1959:
Verhandlungen der Indologischen Arbeitstagung in Essen-
Bredeney, Villa Hugel, 13.-15. Juli 1959, 239-43.
Gottingen: Vandenhoeck und Ruprecht. Reprint: Kleine
Schriften, 279-83.
1960b. "Sprachtheorie und Philosophie im Mahabhasya des

Patafijali." WZKS 4:92-118. Reprint: Kleine Schriften, 284-
310.
. 1961a. "Landmarks in the History of Indian Logic." WZKS

5:125-48.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
BIBLIOGRAPHY 309
1961b. "Mimamsasutram I, 1, 6-23." WZKS 5:113-24.
. 1974. History of Indian Philosophy. Vol. 2, The Nature-

Philosophical Schools and the Vaisesika System, the System
of the Jaina, the Materialism. Trans. V. M. Bedekar. New
York: Humanities Press.
. 1982. Kleine Schriften. Ed. Gerhard Oberhammer und

Ernst Steinkellner. Wiesbaden: Franz Steiner Verlag.
. 1984. Nachgelassene Werke. Band 1, Aufsdtze, Beitrdge,

Skizzen, ed. Ernst Steinkellner. Wien: Verlag der
Osterreichischen Akademie der Wissenschaften. [Geschichte
der Indischen Philosophie. Band 4, Nachgelassene Skizzen:
3, Sprachtheorie, 93-119. 4, Mimamsa, 120-22. 5,
Sprachtheorie des Vrttikara, 123-24. 6, Kumarila, 125-26.]
Gachter, Othmar. 1983. Hermeneutics and Language in Puirva

Mtmamsa: A Study in Sabara Bhdsya. Delhi: Motilal
Banarsidass.
Gafiganatha Jha. 1942. Pirvamtmdmsd in its sources. Varanasi:

Banaras Hindu University.
, trans. 1900-8. Slokavdrtika: ...with extracts from the

commentaries "Kdaikd" of Sucarita Misra and
"Nydyaratndkara" of Pdrtha Sdrathi Misra, by Kumarila
Bhatta. Calcutta. 2d ed. 1 vol. Delhi: Sri Satguru
Publications, 1983.
, trans. 1903-24. Tantravarttika: A Commentary on Sabara's

Bhdsya on the Piurvamlmamsd Suitras of Jaimini, by
Kumarila Bhatta. 2 vols. Calcutta. 2d ed. Delhi: Sri Satguru
Publications, 1983.
, trans. 1912-19. The Nydya-sutras of Gautama: with the

Bhdsya of Vdtsydyana and the Vdrtika of Uddyotakara.
Indian Thought 4-11. Reprint in 4 vols.: Delhi: Motilal
Banarsidass, 1984.
. 1916. The Puirva Mtmadmsa Siltras of Jaimini: Chapters I-

III: Translated with an Original Commentary. The Sacred
Books of the Hindus, vol. 10. Allahabad: The Panini
Office, Bhuvaneswari Asrama.
1933. Sdbara-bhdsya. 3 vols. GOS, nos. 66, 70, 73.

Baroda: Oriental Institute. Reprint: 1973.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Garge, Damodar Vishnu. 1952. Citations in Sabara-bhasya. DC

8. Pune.
Gaurinath Sastri 1959. The Philosophy of Word and Meaning:

Some Indian approaches with special reference to the
philosophy of Bhartrhari. CSCRS 5. Reprint: Calcutta:
Century Press, 1983.
Gerow, Edwin. 1984. "Language and symbol in Indian

semiotics." PEW 34, no. 3 (July): 245-60.
Goldsticker, Theodor 1860. Pdnini: His Place in Sanskrit

Literature: An investigation of some literary and
chronological questions which may be settled by a study of
his work. London: Triibner. First Indian Edition: Ed.
Surendra Nath Shastri. Varanasi: Chowkhamba, 1965.
Halbfass, Wilhelm 1976. "Zum Begriff der Substanz (dravya

Vaisesika." WZKS 20:141-66.
. 1976. Zur Theorie der Kastenordnung in der indischen

Philosophie. Gottingen: Vandenhoeck and Ruprecht.
. 1980. "The Vaisesika concept of guna and the Problem of

Universals." WZKS 24:225-38.
1983. Studies in Kumarila and Sankara. Studien zur

Indologie und Iranistik, ed. Inge Wezler, no. 9. Re
Verlag fur Orientalistische Fachpublikationen.
1992. On Being and What There Is: Classical Vaisesik

the History of Indian Ontology. Albany: SUNY Press
Hara, Minoru. 1959. "Note on the Sanskrit word n'-tya

79:90-96.
Haradatta Sharma. 1936. "The Meaning of the word upacara

According to Gotama and the Rhetoricians." Poona
Orientalist 1, no. 1:26-33.
Herzberger, Radhika. 1986. Bhartrhari and the Buddhists: An

essay in the Development of Fifth and Sixth Century Indian
Thought. Dordrecht and Boston: D. Reidel.
Hiriyanna, M. 1938. "Vyadi and Vajapyayana." IHQ 14:261-66.
Houben, Jan E. M. 1992a. "Bhartrhari's Samaya / Helaraja's

Samketa." JIP 20:219-42.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
BIBLIOGRAPHY 311
. 1992b. "The Sambandha-sumuddesa (Chapter on Relation

and Bhartrhari's Philosophy of Language: A Study o
Bhartrhari's Sambandha-samuddesa in the context of the
Vakyapadiya with a translation of Helaraja's Commentary
Prakirna-Prakasa on the Sambandha-samuddesa." Ph.D.
diss., University of Utrecht.
Ingalls Daniel H. H. 1951. Materials for the Study of Navyanydya

Logic. HOS 40.
. 1957. "Dharma and Moksa." PEW 7, nos. 1-2 (April-

July):41-48.
Jacob, George Adolphus, Colonel. 1984. Index to Sdbara-bhcsya.

Ed. Gaurinath Sastri. The Princess of Wales Sarasvati
Bhavana Studies, no. 4. Varanasi: Sampurnanand Sanskr
Vishvavidyalaya.
Jacobi, Hermann 1911. "The Dates of the Philosophical Sutras o

the Brahmans." JAOS 31:1-29.
. 1929. "Mimamsa und Vaisesika." In Studies in Honor of

Charles Rockwell Lanman, 145-65. Cambridge,
Massachusetts: Harvard University Press.
. Kleine Schriften. 1970. Ed. Bernhard Kolver. Glasenapp-

Stiftung 4, nos. 1-2. 2 Parts. Wiesbaden: Franz Steiner
Verlag.
Joshi, Shivaram Dattatray. 1966. "Adjectives and Substantives as a

Single Class in the 'Parts of Speech'." JUP 25:19-30.
(PCASS, class A, no. 9.) Pune.
. 1968. "Word-integrity and syntactic analysis." JUP 27:165-

73. (PCASS, class A, no. 20.) Pune.
, ed., trans. 1968. Patafijali's Vydkarana-Mahdbhdsya

samarthdhnika (P. 2.1.1). PCASS, Class C, No. 3. Pune.
Joshi, Shivaram Dattatray and J. A. F. Roodbergen, eds., trans.

1969. Pataiijali's Vydkarana-Mahdbhd.sya: Avyaylbhdva-
tatpurusdhnika (P. 2.1.2 - 2.1.49). PCASS, class C, no. 5.
Pune.
, eds., trans. 1971. Pataiijali's Vyakarana-Mahdbhasya:

Karmadhdraydhnika (P. 2.1.51 - 2.1.72). PCASS, class C,
no. 6. Pune.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
, eds., trans. 1973. Patanjali's Vydkarana-Mahabhasya:

Tatpurusdhnika (P. 2.2.2 - 2.2.23). PCASS, class C, no. 7.
Pune.
, eds., trans. 1974. Pataijali's Vydkarana-Mahdbhasya:

Bahuvrlhi-dvandvdhnika (P. 2.2.23 - 2.2.38). PCASS, class
C, no. 9. Pune.
, eds., trans. 1975. Patanjali's Vydkarana-mahdbhdsya:

kdrakdhnika (P. 1.4.23 - 1.4.55). PCASS, class C, no. 10.
Pune.

Anabhihitdhnika (P. 2.3.1 - 2.3.17). PCASS, class C, no.
11. Pune.

Vibhaktydhnika (P. 2.3.18 - 2.3.45). PCASS, class C, no.
12. Pune.

Prdtipadikdrthasesahnika (P. 2.3.46 - 2.3.71). PCASS, class
C, no. 14. Pune.

Paspasihnika. PCASS, class C, no. 15. Pune.
Kane, Pandurang Vaman 1922. "Gleanings from the Bhasya of

Sabara and the Tantravarttika." JBBRAS 26:83-98.
. 1945. "The Mahabhasya and the Bhasya of Sabara."

Bhdrattyd Vidyd 6:43-45.
Kapil Deva Shastri, ed., trans., 1975. Vaiydkaranasiddhdnta-

paramalaghumanjusd of Ndgesabha.tta: Text, Translation,
and Study. Kurukshetra University Publications.
Kedaranatha Tripathi, ed. 1977. Tarkasamgraha with the Dipika

(of Annambhatta) and Hindikald (of Kedaranatha Tripathi)
Commentaries. Varanasi: Kasi Hindu Visvavidyalaya.
Delhi: Motilal Banarsidass.
Kielhorn, Lorenz Franz, ed. 1868. The Paribhdsendusakhara of

Nagojibhatta. Part 1, Sanskrit text. BSPS, 2, 7. Bombay:
The Indu Prakash Press. Second edition: K. V. Abhyankar.
Pune: BORI, 1962.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
BIBLIOGRAPHY 313
trans. 1871-74. The Paribhasendusakhara of Nagojibhatta

Part 2, Translation and notes. BSPS 9, 12. Bombay:
Government Central Book Depot. Revised edition: K. V.
Abhyankar. Pune: BORI, 1960.
___. 1875. "Note on Rajatarangini I 176." IA 4:107-8. Reprint:

Kleine Schriften, Teil 1, 156-57.
1876a. Katyayana and Patafijali: Their relation to each

other and to Panini. Bombay: The Educational Society,
1876. Reprint: Varanasi: Indological Book House, 1963;
Osnabriick: Zeller, 1965; Kleine Schriften, 1-64.
1876b. "On the Mahabhasya." IA 5:241-51. Reprint: Kleine

Schriften, Teil 1, 169-79.
. 1886. "Notes on the Mahabhasya." IA 1886 (July): 203.

[Kleine Schriften, 206-14.]
1969. Kleine Schriften: mit einer Auswahl der

epigraphischen Aufsdtze, ed. Wilhelm Rau. 2 vols.
Wiesbaden: Franz Steiner Verlag.
Kiparsky, Paul, and J. F. Staal. 1969. "Syntactic and semantic

relations in Panini." FL 5:83-117.
Kuiper, F. B. J. 1957. "Vacarambhanam." IIJ 1, no. 2:154-59.
. 1958. Vacarambhanam (II)". IIJ 2, no. 4:306-10.
Kunjunni Raja, K. 1955. "Indian Theories of Homophones and

homonyms." ALB 19:193-222.
. 1963. Indian Theories of Meaning. ALS 91. Madras: The

Adyar Library and Research Center. Second edition: 1969.
1965. "Panini's attitude towards laksana." ALB 29:177-87.
Lahiri, Prabhodh Chandra. 1935. Concordance Pdnini-Patanjali

(Mahabhasya). Indische Forschungen 10. Breslau: M. and
H. Marcus.
Larson, Gerald James and Ram Shankar Bhattacharya, eds. 1987.

Sdmkhya: A Dualist Tradition in Indian Philosophy.
Encyclopedia of Indian Philosophies, gen. ed. Karl H.
Potter, Vol. 4. Princeton: Princeton University Press and
Delhi: Motilal Banarsidass.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Leblanc, Hugues, and William A. Wisdom. 1972. Deductive

Logic. 2d ed. Boston: Allyn and Bacon.
Limaye, Vishnu Prabhakar. 1974. Critical Studies on the

Mahdbhdsya. Hoshiarpur: Vishveshvaranand Vedic
Research Institute.
Limaye, Vishnu Prabhakar, G. B. Palsule and V. B. Bhagavat,

eds., trans. 1985. Mahdbhdsyadipika of Bhartrhari:
Matilal, Bimal Krishna. 1966. "Indian theorists on the nature of

the sentence (vdkya)." FL 2:377-93.
. 1968a. The navya-nyaya doctrine of negation: The

semantics and ontology of negative statements in Navya-
nyaya philosophy. HOS 46. Cambridge, Mass.: Harvard
University Press.
. 1968b. "Gangesa on the concept of Universal Property

(Kevaldnvayin)." PEW 18, no. 3 (July): 151-61.
. 1971. Epistemology, logic, and grammar in Indian

philosophical analysis. The Hague and Paris: Mouton.
. 1973. "The Notion of Substance and Quality in Ancient

Indian Grammar." Acta et Communicationes Universitatis,
II, 2:384-407. Tartu.
. 1977. Nydya-vaisesika. A History of Indian Literature, ed.

Jan Gonda, vol. 6, fasc. 2. Wiesbaden: O. Harrassowitz.
. 1983. "Some comments of Patafjali under 1.2.64." In

PCASS E 9, Proceedings of the International Seminar on
Studies in the Astddhyayl of Pdnini (July 1981), 119-26.
Pune.
. 1985. Logic, Language and Reality: An Introduction to

Indian Philosophical Studies. Delhi: Motilal Banarsidass.
. 1986. Perception: An Essay on Classical Indian Theories of

Knowledge. Oxford: Clarendon Press.
. 1988. "Sabdabodha and the Problem of Knowledge-

Representation in Sanskrit." JIP 16, no. 2 (June): 107-22.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
BIBLIOGRAPHY 315
_ . 1990. The Word and the World: India's Contribution to the

Study of Language. Delhi: Oxford University Press.
Matilal, Bimal Krishna, and Arindam Chakrabarti, eds. 1994.

Knowing from Words: Western and Indian Philosophical
Analysis of Understanding and Testimony. Synthese
Library Studies in Epistemology, Logic, Methodology, and
Philosophy of Science, vol. 230. Dordrecht: Kluwer
Academic Publishers.
Matilal, Bimal Krishna, and Robert D. Evans, eds. 1986. Buddhist

Logic and Epistemology: Studies in the Buddhist Analysis
of Inference and language. Dordrecht: D. Reidel.
Matilal, Bimal Krishna, and Jaysankar Lal Shaw, eds. 1985.

Analytic Philosophy in Comparative Perspective:
Exploratory Essays in Current Theories and classical Indian
Theories of Meaning and Reference. Dordrecht and Boston:
D. Reidel.
Mishra, Umesha. 1926. "Physical theory of sound and its origin in

Indian thought." Allahabad University Studies 2:239-90.
Mishra, Vedapati. 1970. Vyakarana varttika: Eka samiksatmaka

adhyayana [A critical study on Vyakarana-varttika].
Varanasi: Prithivi Prakashan.
Mukherji, Gurusankar. 1992. "Some Linguistic Problems:

Sarfupanam ekasesa ekavibhaktau." In Corpus of
Indological Studies: Professor Ramaranjan Mukherji
Felicitation Volume, vol. 2, ed. Anantalal Thakur, 360-68.
Delhi: Sharada Publishing House.
Narayana Murti, M. S. 1971a. "A Note on the iva-samasa."

SVUOJ 14:47-52.
_ . 1971b. "A Note on Ekasesa." SVUOJ 14:83-98.
1973. "Philosophy of Grammar." In SVUOJ 16,

Proceedings of the Seminar on Andhra's contribution to
Indian culture, 37-54.
Nilakantha Sastri, K. A. 1921. "Jaimini and Badarayana". IA

50:167-74.
Oberhammer, Gerhard. 1964. "Gedanken zur historischen

Darstellung indischer Logik." Orientalistische
Literaturzeitung 59:5-19. Leipzig.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Ojihara, Yutaka. 1965. "A la recherche de la motivation ultime du

Panini-sutra 1.1.62." Miscellanea Indologica Kiotiensia 6-
7:69-85. Kyoto.
1967. "Causerie vyakaranique IV: Jati 'genus' et deux

definitions prepatafijaliennes." Journal of Indian and
Buddhist Studies 16:459-51. Tokyo.
1967-68. "Sur 1'enonce panineen astrivisaya (IV.1.63):

deux interpretations et leur rapport avec le Ganapatha."
ALB 31-32 (V. Raghavan Felicitation Volume): 125-43.
Palsule, Gajanan Balakrishna. 1967. "'Samjfinaym' in Panini." JUP

25:31-75. (PCASS, class A, no. 10.)
ed., trans. 1985, 1988. Mahabhdsyadipika of Bhartrhari:

Fascicules 1, 5: Ahnikas 3, 2. BORI Post Graduate and
Research Department Series 22, 31. Pune.
Palsule, Gajanan Balakrishna and Vaman Balakrishna Bhagavat,

eds., trans. 1991. Mahabhdsyadipika of Bhartrhari:
Pandeya, R. C. 1963. The Problem of Meaning in Indian

Philosophy. Delhi: Motilal Banarsidass.
Paranjpe, Vasudeva Gopala. 1922. Le vdrttika de Kdtydyana: une

etude du style, du vocabulaire et des postulats
philosophiques. Heidelberg: Weiss'sche Universitats-
buchhandlung.
Pashupatinath Sastri. 1923. Introduction to Puirva-Mimamsd.

Calcutta. (Ph.D. diss., University of Calcutta, 1924.) Second
Edition: Gaurinath Shastri, ed. Chaukhambha Orientalia,
1980.
Pathak, Shridhar, and Siddheshvar Chitrao. 1927. Word-index to

Patanjali's Vyakarana-mahdbhdsya. Pune: BORI.
. 1935. Word-index to Pdnini-suitra-pttha and Parisistas.

Pune: BORI.
Perry, Bruce M. 1995. "An Introduction to the Nyayacatu

granthika: with English Translations." Ph.D. diss.,
University of Pennsylvania.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
BIBLIOGRAPHY 317
Potter, Karl H., gen. ed. 1970-. Encyclopedia of Indian

Philosophies. Vol. 1, Bibliography, 2d ed., 1983 [1st ed.
1970]. Vol. 2, Indian Metaphysics and Epistemology: The
Tradition of Nyaya-Vaisesika up to Ganigesa, 1977. Vol. 3,
Advaita Veddnta up to Sankara and His Pupils, 1981. Vol.
4, Samkhya: A Dualist Tradition in Indian Philosophy, ed.
Gerald James Larson and Ram Shankar Bhattacharya, 1987.
Vol. 5, The Philosophy of the Grammarians, ed. Harold G.
Coward and K. Kunjunni Raja, 1990. Princeton: Princeton
University Press and Delhi: Motilal Banarsidass.
Raju, Poola Tirupatti. 1942. "Arthapatti: Its Logical Significance."

Proceedings of the All-India Oriental Conference. Varanasi.
1951. "The Universal in the Western and the Indian

Philosophy." In Radhakrishnan: Comparative Studies i
Philosophy Presented in Honour of his Sixtieth Birthday
379-408. New York: Harper and Brothers.
1971. The Philosophical Traditions of India. London:

Allen and Unwin.
. 1985. Structural Depths of Indian Thought. SUNY Series

in Philosophy, ed. Robert C. Neville. Albany: State
University of New York Press.
Ramasvami Shastri, V. A. 1952. "Bhavana: the leading concept of

Verbal Cognition." Bharattya Vidyd 13:169-76.
Rau, Wilhelm. 1985. Die Vedischen Zitate im Vydkarana-

Mahdbhdsya. Mainz: Akademie der Wissenschaften und der
Literatur; Stuttgart and Wiesbaden: Franz Steiner Verlag.
Renou, Louis. 1942. Terminologie grammaticale du sanskrit.

Bibliotheque de l'Ecole des Hautes-Etudes, nos. 280, 281,
282. 3 vols. Paris: Champion, 1942. Reprint in 1 vol.,
1957.
. 1953a. "Les speculations sur le langage." L'lnde Classique

2:79-84.
. 1953b. "Philosophie grammaticale." L'nde Classique 2:93-

94.
Rocher, Rosane. 1966. "Bhdva 'etat' et kriyd 'action' chez Panini."

In Recherches linguistiques en Belgique, ed. Yvan Lebrun,
113-20. Wetteren: Universa.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC:34:56 UTC
1975. "India." Current Trends in Linguistics, ed. Thomas

A. Sebeok, vol. 13, Historiography of Linguistics, 3-67.
The Hague: Mouton.
Ruben, Walter. 1928. Die Nydyasutra's: Text, ibersetzung,

Erliiuterung und Glossar. Abhandlungen ftir die Kunde des
Morgenlandes, vol. 18, no. 2. Leipzig: Deutsche
Morgenlandische Gesellschaft.
Ruegg, David Seyfort. 1958. "On the term buddhiviparindma and

the problem of illusory change." IIJ 2, no. 4:271-83.
1959. Contributions a l'histoire de la philosophie

linguistique indienne. Publications de l'Institut de
Civilisation indienne, serie in-8, no. 7. Paris: De Boccard.
1960. Compte-rendu de The philosophy of word and

meaning, by Gaurinath Sastri. IIJ 4, nos. 2-3:173-79.
Sabhapati Sarma Upadhyaya, ed. 1973.

Vaiydkaranasiddhdntalaghumaijusda (of Nag
'Ratnaprabhd' Commentary (of Sabhapati Sarma
Upadhyaya). Kashi Sanskrit Series 163. Varanasi:
Chowkhamba Sanskrit Series Office.
Sarma, K. Madhava Krishna. 1968. Panini, Kdtydyana and

Patanjali. Delhi: Lal Bahadur Shastri Samskrita
Vidyapeetha.
.1943. "Vak before Bhartrhari." The Poona Orientalist 8,

nos. 1-2:21-36.
Sarup, Laksman, ed., trans. 1920-27. The Nighantu and the

Nirukta: The Oldest Indian Treatise on Etymology,
Philology and Semantics. Reprinted with three parts in one
volume: Delhi: Motilal Banarsidass, 1984.
Scharf, Peter. 1990. "The Denotation of Generic Terms in Ancient

Indian Grammar, Nyaya, and Mimamsa." Ph.D. diss.,
University of Pennsylvania.
. 1993. "The term 'dkrti' and the Concept of a Class Property

in the Mahabhasya." WZKS 36 (1992), supplement (1993):
31-48. (Paper presented at the VIIIth World Sanskrit
Conference, 27 August - 2 September 1990, Vienna.)
. 1993b. "Assessing Sabara's Arguments for the Conclusion

that a Generic Term Denotes Just a Class Property." JIP 21

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
BIBLIOGRAPHY 319
(March): 1-10. (Paper presented at the XIIIth South Asian

Languages Analysis Roundtable, 25-27 May 1991, Urbana-
Champaign.)
. 1995. "Early Indian Grammarians on a speaker's intention,"

JAOS 115.1: 66-76.
Scharfe, Hartmut. 1961. Die Logik im Mahdbhdsya. Deut

Akademie der Wissenschaften zu Berlin Institut fur
Orientforschung 50. Berlin: Akademie-Verlag.
. 1965. "Vacana 'numerus' bei Panini?" (Kuhns) Zeitschrif

fur vergleichende Sprachforschung auf dem Gebiete d
indogermanischen Sprachen 79:239-46. Gottingen.
. 1971. Pdnini's metalanguage. Memoirs of the American

Philosophical Society, vol. 89. Philadelphia.
_ . 1977. Grammatical Literature. A History of Indian

Literature, ed. Jan Gonda, vol. 5, fasc. 2, 77-216.
Wiesbaden: Otto Harrassowitz.
Sen, Sukumar. 1952-53. "The story of Devadatta in the

Mahabhasya." IL 12:9-16.
Sharma, Rama Nath. 1987. The Astddhydyi of Panini. Vol. 1,

Introduction to the Astadhydyt as a Grammatical Device.
New Delhi: Munshiram Manoharlal.
Siderits, Mark. 1982. "More things in Heaven and Earth," Review

of Dravid (1972). JIP 10:187-208.
. 1985. "Word Meaning, Sentence Meaning, and Apoha." JIP

13:133-51.
. 1991. Indian Philosophy of Language: Studies in Selecte

Issues. Studies in Linguistics and Philosophy, vol. 46.
Dordrecht: Kluwer Academic Publishers.
Smet, Robert de. 1960. "Language and Philosophy in India."

Proceedings of the XII International Congress of
Philosophy 10: 47-54. Florence.
Soni, Jayandra. 1991. "Dravya, Guna and Paryaya in Jaina

Thought." JIP 19:75-88.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Spitzer, Morris. 1927. Begriffsuntersuchungen zum

Nydyabh.sya. Leipzig: Otto Harrassowitz. (Ph.D. diss.,
University of Kiel, 1926.)
Sreekrishna Sarma. 1954. "The theories of the ancient Indian

philosophers about word, meaning, their mutual
relationship and syntactical connection. [Die Theorien der
alten indischen Philosophen iiber Wort und Bedeutung, ihre
Wechselbeziehung, sowie uiber syntaktische Verbindung.]"
Ph.D. diss., Philipps-Universitat Marburg/Lahn.
. 1957. "The words akrti and jati in the Mahabhasya." ALB

21, nos. 1-2 (May): 54-65.
. 1959. "Syntactical Meaning - Two Theories." ALB 23:41-

62.
Staal, J.F. 1962. "Philosophy and language." In T.M.P.

Mahadevan Volume, 10-25. Madras
1966. "Indian Semantics, I." JAOS 86:304-11.
. 1972. A reader on the Sanskrit grammarians. Cambridge,

Mass.: MIT Press; Delhi: Motilal Banarsidass.
1976. "Sanskrit philosophy of language." In History of

Linguistic Thought and Contemporary Linguistics, ed. by
H. Parret. Berlin and New York: Walter de Gruyter.
. 1988. Universals: Studies in Indian Logic and Linguistics.

Chicago and London: The University of Chicago Press.
Strauss, Otto. 1927a. "Altindische Spekulationen liber die Sprache

und ihre Probleme." ZDMG 81:99-151. Reprint: Kleine
Schriften, 220-72.
1927b. "Mahabhasya ad Panini 4.1.3 und seine Bedeutung

fur die Geschichte der indischen Logik." Aus Indiens
Kultur: Festgabe fir Richard von Garbe, ed. J. Negelein,
84-94. Erlangen: Palm und Enke. Reprint: Kleine Schriften,
273-83.
. 1932. Die dlteste Philosophie der Karma-Mfmdmsa. Berlin

LUders.
1983. Kleine Schriften, ed. Friedrich Wilhelm. Wiesbaden:

Franz Steiner Verlag.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
BIBLIOGRAPHY 321
Subha Rao, Veluri. 1969. The Philosophy of a Sentence and its

Parts. New Delhi: Munshi Ram Manohar Lal.
Subrahmanya Sastri, P. S. 1944-62. Lectures on the Mahdbh.sya.

Vols. 1 and 2, Annamalai University Sanskrit Series, nos. 9
and 11, 1944 and 1951. Vols. 3-5, Tiruchirapalli, 1955,
1956, 1957. Vol. 6, Thiruvaiyaru, 1962. Annamalainagar:
Annamalai University and the author. Vol. 1, 2d ed., 1960.
Subramania Iyer, K. A. 1942. "The Conception of Guna among

the Vaiyakaranas." New Indian Antiquary 5:121-30.
Bombay.
. 1945. "The Vaiyakarana Conception of 'Gender'." In

Bhdratakaumudi: Studies in Indology in honour of Dr.
Radha Kumud Mukherjee, vol. 1, 291-307. Allahabad: The
Indian Press.
. 1951. "The Point of View of the Vaiyakaranas." JOR

18:84-96. Madras. Reprint: Staal (1972): 393-400.
. 1950-51. "The Conception of Action among the

Vaiyakaranas." JGJRI 8:165-88.
. 1969. Bhartrhari: A study of the Vakyapadiya in the light

of the Ancient Commentaries. Deccan College Building
Centenary and Silver Jubilee Series, no. 68. Pune.
Suryanarayana Sastri, S. S. 1942. "Word and sense" ABORI 23,

Silver Jubilee Volume, 424-30.
Taber, John A. 1989. "The Theory of the Sentence in Purva

Mimamsa and Western Philosophy," JIP 17, no. 4
(December): 407-30.
Tarkatirtha, Visvabandhu. 1992. "The Nyaya on the Meaning of

Some Words," translated with explanatory notes by J. L.
Shaw. JIP 20:41-88.
Thieme, Paul. 1937. "On the identity of the Varttikakara." Indian

Culture 4:189-209. Calcutta: Indian Historical Institute.
. 1956. "Panini and the Paniniyas." JAOS 76:1-23. Reprint:

Kleine Schriften. Teil 2, pp. 573-95.
1971. Kleine Schriften. Ed. G. Buddruss. 2 vols.

Wiesbaden: Franz Steiner Verlag.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Tiwari, Heeraman. 1994. "One and Many: The Early Naiyayikas

and the Problem of Universals." JIP 22:137-70.
Tola, Fernando and Carmen Dragonetti. 1990. "Some Remarks on

Bhartrhari's Concept of Pratibha." JIP 18:95-112.
Trapp, Valentin, trans. 1933. Die ersten fiinf Ahnikas des

Mahdbhdsyam ins Deutsche iibersetzt und erkldrt. Leipzig:
Harrassowitz. (Ph.D. diss., Ludwig-Maximilians-Universitat
zu Munchen.)
Van Buitenen, J. A. B. 1956. "Vacarambhanam" IL 16:157-62.
. 1958. "Vacarambhanam Reconsidered." IIJ 2:295-305.
. 1957. "Dharma and moksa" PEW 7, nos 1-2:33-40.
Van Nooten, Barend A. 1969. "Panini's theory of verbal

meaning." FL 5:242-55.
. 1983. "Vivaksd, or intention to speak, as a linguistic

principle." In PCASS E 9, Proceedings of the International
Seminar on Studies in the Astddhydyl of Pdnini (July
1981), ed. S. D. Joshi and S. D. Laddu, 43-52. Pune.
Varma, Siddhesvara. 1925. "Analysis of Meaning in the Indian

Philosophy of language." JRAS 1925:21-35.
. 1926. "Analysis of Meaning in Indian Semantics." Journal

of the Department of Letters 13:1-38. University of
Calcutta.
Vattanky, John. 1993. "The Referent of Words: Universal or

Individual, the Controversies Between Mimamsakas and
Naiyayikas." JIP 21:51-78.
Vedavrata, ed. 1962-63. Patanjali's Vyakaranamahabhasya with

Kaiyata's Pradipa and Nagojibhatta's Uddyota. 5 vols.
Gurukula Jhajjar (Rohatak): Haryana Sahitya Samsthana.
Verpoorten, Jean-Marie. 1987. Mlmdmsd Literature. A History of

Indian Literature, ed. Jan Gonda, vol. 6, fasc. 5.
Wiesbaden: Otto Harrassowitz.
Wezler, Albrecht. 1982. "Paralipomena zum Sarvasarvatmakatva-

vada (I). Mahabhasya zu Panini 4.3.155 und seine
einheimischen Erklarer." WZKS 26:149-66.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
BIBLIOGRAPHY 323
. 1985. "A Note on Mahabhasya

dravyam: Studies on Mallavadi
In Bukkyr to ishukyl: Kumoi
[Buddhism and Its Relation to
Honour of Dr. Shozen Kumoi on
Kyoto: Heirakuji Shoten.
. 1994. "Once Again on Pataija

(Studies in Patanijali's Mahabha
Yuddhisthira Mimamsaka. 1973. S

itihdsa [History of Sanskrit g
Kapur Trust.
, trans. 1977-86. Sabara's Jaim

Hindi explanation called drsamata-vimarsint. 5 vols.
Bahalagadha (Sonipat-Harayana): Ram Lal Kapur Trust,
1977, 1978, 1980, 1984, 1986.

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
Index
action 21, 38, 40, 43, 51, 54, analytic strings 36-37, 53, 57,
55, 56, 57, 59, 78, 79, 77-78, 89, 132, 145
89, 123-125, 130, 138, anantya 59, 129 n.60 (See
146, 154-155, 158, 161, infinity)
169, 182, 188, 195, 196, anvaya and vyatireka 19, 40,
203-207, 213, 218, 222, 58 n. 57, 80-81, 168 n.
227, 244, 251-255, 257- 30, 184-185, 201 n. 13,
259, 263, 264, 267, 274, 229, 269, 286 (See
277-280, 286-287, 291, concomitant presence and
294 absence)
agent 43-44, 73 n. 75, 109 n. apavarga 151
18, 111 n. 22, 138, 147 dpta 151-152
n. 91, 155, 175 n. 7, 278 Astadhyayi 1, 3
n. 10, 288 n. 24 atidesa 25, 76, 84, 92-93
aggregate of qualities 24-25,
161-63, 207-8, 206 n.24 Badarayana 4, 269
Aitareyabrahmana 123 n. 48 Bhartrhari 3, 10
akrti 4, 11-18, 21, 23, 27-38, Bhattojidiksita 3
40 n. 40, 50, 58 n. 57, 84 Buddhism 5-6, 323; Buddhist
n. 94, 116-117, 121, argument against generic
124-126, 129 n. 60, 130 properties 155, 158, 160,
n. 61, 131 n. 63, 133 n. 162-163, 182-187, 269,
66-67, 134-135, 140- 271
144, 145 n. 88-89, 148 n.
93, 153, 154 n. 15, 156- calf, new-born 226, 245-246,
160, 167 n. 29, 174, 177, 282
180-182, 185, 188-192, Candrananda 155 n. 18
201 n. 13, 205, 212, 213 Carakasamhita 7
n. 35, 214 n. 36, 215, Chandogyopanisad 28 n. 21,
220-222, 257, 269-271, 139 n. 79
274-280, 282 n. 16-17, class 11, 17, 38-39, 59, 78,
283-287, 289-290, 293 86-87, 125, 163, 289
(See also class property, class property, nature of 11-
form, shape) 18, 21, 23, 27-30, 34-39,
Akrtyadhikarana (Jaimini 48-65, 70, 76, 83, 117,
1.3.30-35) 5, 8, 10, 18, 121, 126, 130, 142, 144-
20, 205, 206, 208, 209, 145, 201, 204-205, 209-
212, 223, 235, 256, 263, 223, 255, 257-261, 263,
274, 290, 292; 269-271, 274, 293 (See
translations 8 also akrti)
Amarakosa 62 n. 66, 137-38 clay 27-29, 153, 156, 160,
n. 74, 146 n. 90, 270 n. 5 164-166, 168, 171, 192,
Atharvaveda 115 n. 32-33 216-17 (See model cow)
Ayurveda 152 n. 11
325

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
cognition, 2, 6, 21, 37, 47- dravya 11-17, 21, 23-30, 35-

50, 52, 41-42, 47-50, 52, 38, 40 n. 40-41, 50, 57,
171, 196, 205, 209, 210- 76 n. 81, 80 n. 85-86, 81
211, 269-271; of n. 91, 84 n. 94, 85, 95 n.
qualifier first 161-162, 9, 96, 79, 115 n. 34, 118,
224-225, 245-246, 266- 121-122 n. 46, 123, 124
267, 283, 286, 292 (See n. 51, 125-126, 127 n.
also identical cognition 57, 128 n. 57, n. 59, 129,
and means of knowledge) 130 n. 61, 132 n. 64, 134
concomitant presence and n. 70, 136 n. 72, 137 n.
absence 19, 40-41, 42, 73, 140, 141 n. 82, 142,
43-44, 46, 81, 84, 87, 143 n. 85, 144-145, 146
165, 217, 222, 225-227, n. 91, 154 n. 17, 155 n.
229, 231-233, 236-238, 18, 175, 176 n. 9, 180 n.
240-242, 246-247, 266- 25, 177, 181, 184, 186,
267, 284, 286, 290, 292- 191, 204 n. 18, 206 n.
294 24, 213 n. 36, 214 n. 36,
configuration (See shape) 13, 220-221, 222 n. 44, 252
19, 31, 34-35, 153-154, n. 58, 253, 256, 274,
156, 164, 165, 167, 169, 276-277, 279 n. 12, 286-
170, 171, 181, 192, 213, 287, 289 (See also
215-217, 219 substance)
duty (See dharma)
danda 275-276, 285-286
(See staff) ekasesa 2, 9, 40, 42, 101,
deletion 43-44, 80-81, 103, 105, 106, 108-111, 113
109-111, 173, 286, 270- n. 26, 115 n. 34, 116 n.
271 (See also lopa, luk, 35, 121, 124, 126, 131,
lup) 134-135, 141 n. 82, 147
denotation 1, 5, 21-22, 45, n. 91 (See remainder of
47, 87, 112-115, 229- one)
230, 235, 282-283 eternity, of a speech form 9-
deviation 59 (See also 10, 13, 22, 27, 10 n. 19,
vyabhicara) 173 n. 2, 196-203, 215,
dharma 21-22, 50, 118-120, 221-222; of the object
195-196, 198-199, 203, denoted 12, 16, 22, 27-
269; property or nature 28, 144 n. 87-88, 154,
132 n. 64, 135 n. 70, 139 184, 198, 203-4, 218,
n. 79, 152 n. 17, 162 n. 223; of the word meaning
24, 175, 179 n. 18, 233 relation 22, 27, 198, 223;
n. 52, 252; dharmin 125 of substances in Vaisesika
n. 52 155 n. 18, 154, 207, 216
Dharmakirti 131 n. 63 extension rule (See atidesa)
difference 36-37, 53-55, 57,
77-78, 89, 116-118, 132, falcon altar 18, 157, 164,
145, 176, 189-190, 205, 165, 167, 215-223, 227,
212, 215, 218, 263, 277, 232-235, 240, 242-248,
287-288, 291 250-251, 255, 258-260,

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
INDEX 327
264, 267, 278-80, 286, gender 7, 25, 30-33, 35-37,

293-294 54, 57, 61-79, 81-83, 89,
form 11-14, 16, 18, 34, 152- 103 n. 5, 118 n. 38, 130,
153, 156-160, 163-168, 133-142, 145 n. 89, 229
171, 180, 182, 185-186, n. 50, 291
189, 191-192 (See also general sentences 50-51, 54-
dk~rti and shape) 55, 59, 118-119, 123-
124, 126-128
Gautama 4; siitra 1.1.1 151 generic property 2-3, 7-8, 11,
n. 1; sa7tra 1.1.2 151 n. 13, 15-19, 30-35, 38-39,
2; si7tra 1.1.3 151 n. 3; 42, 46-49, 54, 61-65, 70-
si7tra 1.1.4 162 n. 21; 73, 75-80, 82-89, 95-96,
sa7tra 1.1.7 151 n. 4; 128, 136, 152-171, 174-
sfitra 1.1.8 151 n. 5; 177, 180-193, 196, 201,
si:tra 1.1.14 151 n. 6; 204-205, 210, 212-213,
sfitra 2.1.52 152 n. 1; 216-217, 219-223, 255-
siltra 2.1.55 152 n. 10; 259, 261, 289-295
sfitra 2.1.56 152 n. 10; generic term, definition of
si7tra 2.1.68 152 n. 11- 22-23, 96; examples of
13; stitra 2.2.13 151 n. 29, 41, 45, 49, 51, 53-56,
7; saitra 2.2.18 157 n. 7; 62, 65, 71, 79, 88, 255
sfitra 2.2.40 151 n. 8; gold 153, 154 n. 16, 160,
si7tra 2.2.58-69 4, 5, 8, 164, 166, 171, 192, 216-
18-20, 169-171, 173- 217
193; siitra 2.2.58 4, 152 golden ornaments 27-29,
n. 9, 169, 173; si7tra 211, 214 n. 36, 217-218,
2.2.59 169, 174; siltra 222 n. 44, 270 n. S
2.2.60 117 n. 36, 154, gotra 30-31, 170, 178-179;
166, 168, 169, 175, 190 suffix 94-95 n. 6-7, 101
n. 32; sfitra 2.2.61 165, n. 1, 103-104 n. 5, n.7
170, 177, 181; siitra grammar 21-22, 40, 43-45,
2.2.62 94 n. 6, 104 n. 7, 47, 84, 88, 101 n. 2, 110
167, 170, 178, 257; si7tra n. 19, 115 n. 31
2.2.63 164-165, 170, grammarians' conceptio
180; satra 2.2.64 154- of gender 65-70, 136-
161, 215 n. 38-39, 165- 140
167, 170, 181, 256; si7tra guna 15-16, 21, 23 n. 7, 24
2.2.65 165, 171, 188; n. 10, 25 n. 15, 31, 40 n.
sfitra 2.2.66 4, 163-164, 41, 57, 68 n. 69, 75 n.
166, 171, 189; stitra 80, 76 n. 81, 79, 80, 81,
2.2.67 153-154, 171, 82 n. 92, 133-135, 138,
190; sfitra 2.2.68 12 n. 140, 142, 151, 154, 169
12, 13, 16, 153, 156-158, n. 1, 175, 176 n. 9, 190-
160, 166, 171, 191, 216; 191, 197 n. 7, 205, 206
satra 2.2.69 153, 155- n. 22, 213 n. 36, 214 n.
156. 171. 192. 215 n. 38: 36, 222 n. 44, 252 n. 58,
saitra 3.1.62-63 151 n. 253, 254 n. 62, 274, 276
6, 191 n. 35; n. 8, 277 n. 8; vowel

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
grade 92-93; subordinate 206 n. 22, 227, 273; stitra

140, 204 n. 18, 283, 285 1.3.31 220, 276; smitra
(See quality) 1.3.32 277; smitra
gunas'abda 22, 75 n. 78, 80, 1.3.33 271 n. 6, 277;
142, 81 n. 91, 197 n. 7 smitra 1.3.34 267, 287;
(See quality words) sm7tra 1.3.35 242, 245,
254, 267, 287; smitra
identical cognition 47-49, 52- 2.1.46 253 n. 60; smitra
53, 88-89, 116-118, 153- 3.1.12 75 n. 79, 204 n.
161, 171, 184-188, 192- 18, 253, 277 n. 8; sm7tra
193, 205, 210-212, 214- 4.1.15 276 n. 7; sitra
217, 282, 289, 291 4.2.17-18 276 n. 6;
Indian philosophy of stitra 6.3.11 204 n. 18;
language 5 smitra 6.3.1 2 204, 206;
Indian semantics 3, 5-6, 9 stitra 6.3.31 123 n. 48;
individual, nature of 153, sm7tra 6.8.10 102 n. 2;
171, 190-191, 206-209, sm7tra 6.8.30 123 n. 48,
255-259, 274, 281, 289, 277 n. 9; stitra 10.4.32
293 (See also dravya) 123 n. 48; stitra 12.3.9
inferential mark 19, 153, 253 n. 60
156, 158-160, 164, 171, jati 4, 11-17, 30-36, 38-39,
188, 191-192, 210, 216- 71, 74, 75 n. 78, 81 n.
217, 259, 289, 291, 293 91, 82 n. 92, 84 n. 94,
infinity, of word-meaning 91, 95 n. 9, 96, 97 n. 10,
relations 59 110 n. 20, 116 n. 35-36,
inherence 155 117 n. 37, 118 n. 39, 119
n. 40, 121 n. 45, 124 n.
Jaimini 4-5; si7tra 1.1.1 195 51, 125 n. 52, 126 n. 53,
n. 1; stitra 1.1.2 195 n. 127-128 n. 57, 128, 129
2; stitra 1.1.4 195 n. 4- n. 59-60, 131 n. 63, 133
5; si7tra 1.1.5 4, 7-8, n. 66, 136, 137 n. 73,
119 n. 39, 196, 198, 199 139 n. 79, 141 n. 82, 142
n. 9, 203, 209-211, 213 n. 84, 144 n. 86, 145 n.
n. 35, 214 n. 36, 269- 88, 146 n. 91, 151 n. 1,
271; stitra 1.1.6-23 9, 152, 153 n. 14-15, 154,
198 n. 8; sa7tra 1.1.9 199 155 n. 18, 167 n. 29-30,
n. 11; stitra 1.1.15 60, 174-175, 177-178, 180-
122 n. 46, 200 n. 12; 182, 183, 185, 188-93,
si7tra 1.1.17 199 n. 9; 205, 206 n. 24, 208 n.
si7tra 1.1.18 199 n. 10; 29, 210 n. 32, 214 n. 36-
siitra 1.1.19 8, 201 n. 37, 215 n. 40, 219, 229
13; sfitra 1.1.20 201 n. n. 50, 276 n. 8, 284 n.
14; stitra 1.1.23 202; 19, 286 n. 22, 289-290
stitra 1.1.25 197 n. 6, (See also generic
252 n. 65; stitra 1.3.26 property, class, class
228; si7tra 1.3.30-35 4- property)
5, 8, 273-288; stitra
1.3.30 196, 205 n. 21,

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
INDEX 329
jdtisabda 22, 36, 85, 96, 178, 87, 94; vdrttika 5 94;
190, 128 n. 57 (See also varttika 6 36, 85, 94 n. 5,
generic term) 95; vdrttika 7 4, 9, 16,
Jayaditya 3 36, 79, 85-86, 96;
Jayantabhatta 219 1.2.64 varttika 1 42,
105, 112 n. 24, 140 n.
Kaiyata 2 n. 4, 3, 8, 15 n. 16, 81; vdrttika 2 105;
24 n. 10, 25, 31 n. 27, 33 vJrttika 3 99, 105;
n. 36, 34 n. 37, 39 n. 39, varttika 12 113 n. 27;
52 n. 52, 53, 67, 69 n. varttika 14 104 n. 8;
72, 71 n. 74, 81, 82 n. vdrttika 19 113 n. 26,
92, 93-94 n. 5, 95 n. 9, 147 n. 91; vdrttika 23
101-148 n. 2-93 103 n. 5; varttika 27 36,
kdraka 6-7, 26, 24 n. 11, 61 106; varttika 28 9, 36.
n. 58, 138, 141 n. 82, 109; vdrttika 29 9, 36,
155, 188, 204 n. 18, 229 109, 114 n. 28, 135 n.
n. 50, 278 n. 11, 253, 71, 282 n. 15; vdrttika 30
264 36, 112; vdrttika 31 36,
Kd?ika 3, 33 n. 35, 34 n. 36, 114 n. 30, 112-113;
95 n. 8, 101 n. 1, 104 n. vJrttika 32 36, 99, 114;
7, 105 n. 9, 121 n. 45, vdrttika 33 36, 114-115,
127 n. 56, 276 n. 6, 278 282 n. 15; varttika 34 36,
n. 10, 279 n. 11 114-115, 282 n. 15;
Kathakasamhitd 119 n. 40, varttika 35 2, 4, 36, 48,
123 n. 48, 126 n. 56 56, 65, 116; vdrttika 3 6
Kdtyayana 1-2, 4-5, 7-9, 15- 48-49, 51-52, 116-117 n.
16, 30, 35-38, 99; first 37, 118 n. 38-39, 119 n.
vcYrttika 22, 27, 108 n. 39, 125 n. 52; vdrttika 37
15, 110 n. 19, 174 n. 4, 48-50, 52, 117, 118 n.
198 n. 8; fivasfitra 1 38, 120, 125, n. 52;
vdrttika 12 120 n. 43-44, varttika 38 35, 48, 49,
122 n. 46, 130 n. 62, 200 52, 117, 118-119 n. 39;
n. 11, 201 n. 12; vjrttika varttika 39 48, 50, 51-
13 34, 122 n. 46; 52, 118-120, 122 n. 46;
vdrttika 16 34, 122 n. vdrttika 40a 8, 14, 19,
46; Si'vasaitra 2 vdrttika 1 35, 38-39, 49, 52, 56, 60,
22 n. 6; 1.1.1 vdrttika 9 120-121, 122 n. 46, 201
27 n. 20; 1.1.56 vdrttika n. 12; varttika 40b 8, 14,
19 95 n. 2; 1.2.45 19, 35, 38, 48, 49-50, 52,
varttika 7-8 45 n. 49; 59, 60, 120-122, 144 n.
vdrttika 9 41 n. 43, 58 n. 46; varttika 41 38, 48,
57; 1.2.52 vdrttika 1 50, 52, 60, 121-122, 281
74-75 n. 78; vdrttika 2 n. 14; vdrttika 42 48, 50,
62, 75 n. 78; 1.2.58 52, 58, 121-123; vJrttika
varttika 1 16, 36, 48-49,43 14, 39, 48, 50-52, 58,
84, 96, 91; vdrttika 2 36, 60, 62, 77, 123; vdrttika
92, 93 n. 5; vJrttika 3 44 48, 51, 52, 126, 127
36, 76, 93; vJrttika 4 86- n. 57; vdrttika 45 2, 48,

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
77, 129; vjrttika 46 36, 4.1.63 s'loka-vdrttika 35;

52, 54, 56, 77, 129-130, 4.1.92 vdrttika 6-7 68
143 n. 85; vdrttika 47 39, n. 71, 71 n. 73; 5.1.119
52, 54, 56, 59, 77, 123 n. vdrttika 5 15-16, 23-24,
48, 129-130, 142-143 n. 29 n. 25, 79; 5.2.94
85; vcYrttika 48 8, 35, 38, vdrttika 3 80-81, 82-83
52-53, 56, 60, 77, 120 n. n. 93; vdrttika 4 80-81;
42, 129-130, 144 n. 86, vdrttika 5 76, 81, 82-83
200 n. 11; vdrttika 49 38, n. 93; 6.1.84 vdrttika 4-
52, 55, 56, 59, 77, 129- 5 119 n. 40; 7. 1.74
131, 139 n. 39, 144 n. vdrttika 5 33 n. 35
87; vdrttika 50 37, 52- Kdtydyanas'rautasiitra 277 n.
53, 56, 77, 145 n. 89, 9, 280 n. 13
131-132; vjrttika 51 37, Kaun0abhatta 3, 9
52-53, 56, 77, 131-132, kriya 21, 23 n. 7, 119 n. 40,
145 n. 89; vdrttika 52 36, 121 n. 45, 132 n. 64, 141
52, 55-57, 77, 131-133, n. 82, 143 n. 85, 145-
146-147 n. 91; vdrttika 148, 176 n. 9, 177, 187,
53 4, 9, 19, 35, 56-57, 195 n. 3, 197 n. 6, 204 n.
62, 64-65, 67, 70, 72, 77, 18, 254 n. 62, 271 n. 6,
84 n. 94, 86, 133, 226, 274, 277-278, 279 n. 12,
283 n. 18; vdrttika 54 280, 287 (See also action)
36-37, 56-57, 73-74, 77, kriyakibda 22, 147-148 n. 91
79, 82-84, 142; vdrttika Kumdrila 4, 8, 83, 156, 139
55 37-38, 56, 76-77, 83- n. 79, 154 n. 15, 161,
84, 142, 144; vdrttika 5 6 162 n. 20, 195 n. 6, 197
8, 35, 38, 56, 60, 77, 122 n. 7, 204 n. 18, 205-208,
n. 46, 143; vjrttika 57 210-219, 223-224, 226
56, 77, 144; vdrttika 5 8 n. 48, 228-231, 233-234,
37, 56-57, 77, 145; 236, 238, 240, 245, 257-
varttika 59 36, 39, 56- 261, 275 n. 4-5, 276 n. 8,
57, 77, 145; 1.2.69 279 n. 12, 282 n. 16-17,
vdrttika 2 85 n. 96, 96 n.284 n. 19, 286 n. 22,
10; 1.3.1 vdrttika 6 58 293-294
n. 57; 1.4.1 vdrttika 1
26 n. 17; 1.4.23 vdrttika LaghumainjuSq 25 n. 12
15 26 n. 18; 2.1.1 laksanA 9 (See secondary
vdrttika 21 27 n. 19; reference)
2.2.29 vJrtitika 6 115 n. locus 138
33; vdrttika 14 114 n. 29, lopa 44 n. 47, 80 n. 85-86,
varttika 15 115 n. 31, n. 81 n. 91, 102, 109, 110
33; 2.4.12 vdrttika 1 n. 19, 1 1, 127 n. 66,
113 n. 27; 3.2.1 vjrttika 137 n. 74, 173 (See
5 115 n. 31; 4.1.3 deletion)
varttika 6 74; vdrttika 7 103, 43 n. 45, 44 n.
luk 80,
75 n. 80; ilokavdrttika 47, 80 n. 88, 81 n. 91, 83
66-68, 137; 4.1.21 n. 93, 103 n. 5-6, 104 n.
ilokavdrttika 93 n. 5;

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
INDEX 331
7, 115 n. 33, 173-174 n. number 36-39, 41, 47, 54,

3 57, 61-62, 64-67, 68 n.
lup 74-75 n. 78, 103 n. 7 70, 72-79, 81 n. 91, 82-
83, 84-89, 91-96, 107,
Mahabhasya 1-5, 8-13, 14 126, 130, 133-135, 140-
n. 15, 16-19, 38, 23 n. 7, 142, 148, 169, 175-176,
41 n. 44, 108 n. 15, 122 201, 204 n. 18, 220, 229
n. 46, 198 n. 8, 199 n. n. 50, 242, 263, 267,
11, 201 n. 12, 215, 282 276, 288, 291
n. 15, 289, 292; Nyaya 3-10, 17-20, 151-152,
translation 7-8 155, 157-158, 160, 164-
Mahabhasyadipika 3 166, 174 n. 5, 213, 214
Maitrayanisamhita 123 n. 48, n. 36, 215-216, 219, 222,
126 n. 56 235, 255-257, 259, 261,
Mandanamisra 4, 260 n. 64 289, 291-293; Navya-
Manusmrti 119 n. 40 164, 168, 257
mass nouns 29-30 Nyayabhasya 10, 18, 20
means of knowledge 118-119 translation 8
n. 39, 125 n. 52, 131-132 Nyayavarttika 3, 20
n. 63, 133 n. 67, 151, translation 8
162, 195-196, 202 n. 15, Nyayasftra 3-4, 289
203, 210, 225-226, 229- translation 8
231, 233-235, 237, 238
n. 54, 239, 258, 269-271 one word - one meaning
(See also presumption, 41-48, 87, 105-116, 126
and speech as a means of n. 54, 134 n. 69, 141 n.
knowledge) 82, 228, 290, 292
Mimamsa 3-5, 7-10, 17, 20, one-many relation 47-48, 50,
73 n. 75, 173 n. 2, 195- 52-53, 56, 59-61, 120-
196, 198, 200, 213, 223, 122, 125, 129-130, 143,
229, 238 n. 54, 242, 251, 199-200
255, 257, 261, 289, 292
model cow 157-160, 164- Panini 1-3, 5 sutra 1.1.1 27
168, 170-171, 181-183, n. 20; sutra 1.1.11 115
188, 246, 255-257, 259, n. 33; suttra 1.1.49 102
292 n. 3, 103 n. 5; siutra
1.1.50 102 n. 4; suttra
Nagesa 2 n. 4, 3, 8, 13, 15 n. 1.1.56 92 n. 2, 94 n. 5;
16, 25 n. 12, 33 n. 36, 34 suitra 1.1.62 44 n. 47,
n. 37, 53 n. 53, 64 n. 61, 111 n. 23, 174 n. 3; suitra
68 n. 70, 69 n. 72, 71 n. 1.1.63 44 n. 47; suttra
74, 81 n. 91, 82 n. 92-93, 1.1.64 137 n. 74; suitra
83 n. 93, 94 n. 5, 95 n. 9, 1.1.68 101 n. 2; sutra
96 n. 10, 102 n. 2, 104 n. 1.1.72 103 n. 6; suttra
7, 110-133 notes, 137- 1.2.41 44 n. 48; sutra
148 notes, 233 n. 52, 269 1.2.45 22 n. 6, 41 n. 43,
n. 1 45 n. 49, 58 n. 57; sutra
nihs'reyasa 151, 195 1.2.51 74-75 n. 78, 141

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
332 332 ~DENOTATION OF GENERIC TERMS
n. 82; siltra 1.2.52 35, stitra 1.4.23 26 n. 18;

62, 71, 74-75, 136 n. 72; stitra 1.4.42 278 n. 11;
stitra 1.2.53 68 n. 70; stitra 1.4.49 278 n. 11;
stitra 1.2.56 116 n. 34; stitra 1.4.102 93 n. 5,
stitra 1.2.58 2, 4, 7-9, 140 n. 81; stitra 1.4.103
12, 15-16, 18-19, 35-37, 93 n. 5; stitra 2.1.1 27 n.
39-40, 48, 61 n. 59, 76, 19; stitra 2.1.6 105 n.
79, 84-87, 91, 92 n. 2, 93 10, 121 n. 45; stitra
n. 2-3, n. 5, 94, 95 n. 9, 2.1.30 80; si7tra 2.1.36
1 11 n. 20, 220, 289; si7tra 176 n. 14; stitra 2. 1.5 1
1.2.59 94-95; si7tra 127 n. 56; stitra 2.1.52
1.2.63 94 n. 5; siitra 127 n. 5 6; s i7tra 2.2.6 2 5
1.2.64-73 9, 107 n. 13; n. 1 2; stitra 2.2.24 1 01
stitra 1.2.64 2, 4, 7-9, n. 1, 1 36 n. 7 2; stitra
11-12, 14, 16, 18-19, 21 2.2.28 33 n. 35; stitra
n. 3, 3 0, 3 1 n. 29, 35-37, 2.2.29 43 n. 45, 113 n.
39-40, 42-43, 44 n. 4 7, 28, 114 n. 29, 115 n. 31,
47-48, 50, 55-58, 60, 62- n. 3 3; stitra 2.2.3 6 13 6
63, 64 n. 61, 65, 67, 6 8 n. 72; stitra 2.4.12 113
n. 6 9, 70, 71 n. 74, 7 2, n. 27; stitra 2.4.13 27 n.
79, 82 n. 92, 84 n. 94-95, 19; saitra 2.4.62 104 n.
86-87, 94 n. 6, 99, 1 01, 7; stitra 2.4.63 104 n. 7;
103 n. 5, 107 n. 13, 108, stitra 2.4.64 103 n. 5-6,
113 n. 26-27, 116 n. 34- 104 n. 7; stitra 2. 4.7 1
35, 147 n. 91, 173 n. 3, 43 n. 45, 80 n. 87; stitra
174 n. 4, 182 n. 27, 201 2.4.82 173 n. 3; saitra
n. 12, 220, 226, 269 n. 1, 2.4.85 8; stitra 3. 1.3 2
281 n. 14, 282 n. 15, 283 43 n. 45; stitra 3.1.6 7
n. 1 8, 289; stitra 1.2.65 147 n. 9 1; stitra 3. 1.9 1
107 n. 13; stitra 1.2.66- 109 n. 18; stitra 3.1.9 3
68 33 n. 35; stitra 109 n. 18; stitra 3.1.9 7
1.2.67 63, 118 n. 38, 24 n. 1 1; s i7tra 3.2.1 11 5
136 n. 72, 137 n. 73; n. 3 1; stitra 3.2.23 24 n.
stitra 1.2.69 85 n. 96, 11; stitra 3.2.90 109 n.
96 n. 10; stitra 1.2.73 1 8, 1 11 n. 2 3; stitra
136 n. 72; stitra 1.3.1 5 8 3.2.91 109 n. 18, 111 n.
n. 57; stitra 1.4.1 26 n. 23; stitra 3.2.92 278 n.
17; stitra 1.4.7 92 n. 2; 1 1, 280 n. 1 2; stitra
stitra 1.4.14 4, 174 n. 3; 3.3.102 25 n. 14; stitra
si7tra 1.4.18 127 n. 56; 3.4.67 109 n. 18; stitra
stitra 1.4.21 42-43 n. 3.4.70 24 n. 1 1; stitra
45, 54 n. 55, 61 n. 5 8, 3.4.78 109 n. 18, 140 n.
72-73, 84, 140 n. 81, 141 8 1; stitra 4. 1. 2 140 n.
n. 81-82, 93 n. 5, 173 n. 81; stitra 4.1.3 7, 13, 24,
4; stitra 1.4.22 43 n. 45, 66-67, 71 n. 73, 74, 75 n.
54 n. 55, 61 n. 58, 72-73, 80, 140 n. 8 1; stitra
84, 140 n. 81, 141 n. 8 1- 4.1.4 25 n. 14, 33-34 n.
82; stitra 1.4.23-55 7; 36, 66 n. 64; stitra

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
INDEX33 333
4.1.15 137 n. 74; sfitra 43 n. 45; stitra 6. 1.1I0 6

4.1.2 0 3 1; stitra 4. 1.2 1 115 n. 33; stitra 6.3.34
93 n. 5; si7tra 4.1.41 34 32-33, 34 n. 36; stitra
n. 3 6; siitra 4. 1. 48 17 9 6.3.41 32-33; stitra
n. 18; stitra 4.1.63 8, 6.3.85 101 n. 1; stitra
14, 16, 30-35, 58 n. 57, 6.4.1 103 n. 7, 127 n.
62, 67 n. 66, 70, 137 n. 56; stitra 6.4.143 137 n.
73; stitra 4.1.83 74 n. 74; stitra 6.4.148 31 n.
78, 127 n. 56; siuitra 29, 127 n. 56, 137 n. 74;
4.1.88 127 n. 56; siltra stitra 7.1.39 115 n. 33;
4.1.92 68, 71 n. 73, 103 stitra 7.1.2 22 n. 6; stitra
n. 5; sfitra 4.1.101 94 n. 7.1.74 33 n. 35; stitra
6; saitra 4.1.104 103 n. 7.2.117 103-104 n. 7;
5, 104 n. 7; stitra stitra 7.2.118 127 n. 56;
4.1.105 94 n. 6, 103 n. stitra 7.3.50 22 n. 6;
5; stitra 4.1.162 - 167 stitra 7.3. 1 11 92-93 n.
44 n. 48, 94 n. 6, 104 n. 2; stitra 8.1.4 105 n. 1 0;
7, 1 11 n. 23, 174 n. 3, stitra 8.2.8 1 93 n. 4;
179 n. 22; stitra 4. 2. 16 stitra 8.3.59 22 n. 6
127 n. 56; stitra 4.2.24 pada 4, 42 n. 45, 44 n. 47,
127 n. 56; stitra 4.2.33 101 n. 1, 173-174, 199 n.
127 n. 56; siitra 4.2.67 9
74 n. 78; sa7tra 4.2.81 Parthasdrathimis'ra 214, 208
74 n. 78, 141 n. 82; stitra n. 29, 211 n. 32
5.1.115 25 n. 12, 121 Patafijali, about him 1-3;
n. 45, 122 n. 46; si7tra reference to 1-148
5.1.119 8, 15-16, 23,9 passim, 163-164, 173 n.
24 n. 10- 11, 25 n. 14, 2 9 3, 174 n. 4, 179 n. 18,
n. 25, 40 n. 41, 79-81, 198 n. 8, 206 n. 24, 218
206 n. 24; si7tra 5. 2.3 6 n. 42, 220, 226, 229, 28 1
95 n. 8; stitra 5.2.94 76, n. 14, 283 n. 18, 285 n.
80, 81 n. 91, 276 n. 6; 21, 289-292, 295
si7tra 5.2.115 276 n. 6; polysemy 39, 55, 57, 89,
siltra 5.3.28 33 n. 35; 101, 131, 133, 145-147
si7tra 5.3.55 25 n. 1 6, Prabhakara 4-5
31 n. 28, 40 n. 41, 61 n. Pradipa 3, 8
58; stitra 5.3.74 61 n. pramdna 68 n. 70, 110, 115
58; sfitra 5.3.96 103- n. 34, 116 n. 36, 118 n.
104 n. 7, 182 n. 27; stitra 39, 119 n. 40, 151, 162
5.3.99 103-104 n. 7; n. 22, 180 n. 23, 196,
stitra 6.1.37 179 n. 1 8; 202 n. 15, 231, 269, 284
stitra 6.1.45 137 n. 74; n. 20 (See also means of
stitra 6.1.66 137 n. 74; knowledge)
stitra 6.1.67 44 n. 48, Pramanavdrttika 131 n. 63
109 n. 1 8, 111 n. 23; Pras'astapdda
stitra 6.1.71 44 n. 47, pravrttinimitta (See semantic
1 11 n. 23; stitra 6. 1. 84 condition)
119 n. 40; stitra 6.1.102

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
presumption 19, 34 n. 37, Rgveda 115 n. 32-33, 203

195, 197, 216-217, 227-
231, Sabara, about him 4-5;
238-239, 241, 248-
249, 251, 260-261, 263, translations 8-10;
265, reference to 12-13, 17-
284 n. 19, 292, 294-
295 20, 60, 75 n. 79, 83, 102
principal versus subordinate n. 3, 122 n. 46, 123 n.
cognition of denoted 48, 125 n. 52, 156-157,
object 26 n. 16, 77-79, 163-165, 167, 173 n. 2,
85-87, 89, 95 n. 9, 141, 195 n. 2-3, n. 5, 197 n.
162-164, 166-168, 171, 6-7, 198, 199 n. 9-11,
189-190, 256-257, 265- 200 n. 12, 201 n. 13-14,
266, 283-285, 289 202, 206, 208-209, 212,
principal action 121 n. 213 n. 35, 214-218, 220-
45, 141 n. 82, 253 n. 60 228, 232-238, 240-242,
proper name 6, 22, 58 244-248, 251-260, 269
Purvamimamsa (See n. 1, 276 n. 6-8, 277 n.
Mimamsa) 8-9, 280 n. 13-15, 288 n.
Purvamimamsasutra 4 23-24, 289-290, 292-
translation 8 295,
sabda 15, 21-22, 24 n. 10,
quality, nature of 14-15, 17, 27, 36, 41 n. 44, 42, 58
21, 22 n. 6, 23-25, 29 n. n. 57, 68, 71 n. 74, 75 n.
25, 31, 36-38, 66-67, 69, 78, 79, 82 n. 92, 85, 93,
71, 74-76, 78-84, 89, n. 5, 96, 97 n. 10, 101,
116 n. 36, 117 n. 36-37, 102 n. 2, 105-106, 108-
133-135, 138, 140, 142, 116, 117 n. 37, 119 n.
151, 153-155, 157-158, 40, 121, 122 n. 46-47,
160-161, 164, 168-171, 123 n. 48, 124 n. 51, 126
176, 185 n. 28, 190-191, n. 53-54, 127-128 n. 57,
197 n. 7, 205-207, 213- 129 n. 60, 130 n. 61, 132
216, 222, 251-255, 263, n. 63, 133 n. 66, 134 n.
274, 276-277 n. 8; of 69, 135 n. 71, 136 n. 72,
sound 34, 151 137 n. 73, n. 75, 138,
quality words 15, 22, 62, 73- 139 n. 79, 141 n. 82,
76, 79-84, 197 n. 7, 257- 142, 146 n. 90, 147-148
255, 291 n. 91, 148 n. 93, 151,
152 n. 10-11, 154 n. 15,
relation of property and 174-175, 177-178, 180-
substratum 49-50, 56-57, 183, 187-190, 195, 199,
73-77, 79-84, 142-145, 200 n. 12, 201 n. 13, 202
154-155, 206-209, 211- n. 15, 204, 207 n. 25,
212, 227, 238, 252, 284, 214 n. 36-37, 215, 228,
286 230, 252 n. 58, 254, 269,
remainder of one 9, 42-48, 271, 273-282, 284-287;
88, 105-112, 124, 126, as a quality 138, 191 n.
290 (See Panini su7tra 35 (See speech as a means
1.2.64) of knowledge)

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
INDEX 335
Sabdakaustubha 3 n. 11, 201 n. 12; 2 22 n.

samanya 14-16, 34, 39, 41 n. 6
44, 48, 57, 84, 91, 96, Slokavarttika, translation 8,
116 n. 36, 118 n. 39, 122 154 n. 15, 162 n. 20,
n. 47, 127-128 n. 57, 132 208, 216 n. 41
n. 64, 145-148, 154 n. speaker's intention 4, 9, 19,
16-17, 175, 183, 185- 65-73, 78-79, 85-87, 89-
187, 189-190, 192, 205, 91, 95 n. 9, 96, 134-135,
207 n. 25, 210-211 n. 32, 140, 141 n. 82, 171, 189-
212-213, 214 n. 36-37, 190, 255-256, 260, 289
222 n. 44, 252 n. 57, 270 speech, as a means of
n. 5, 274-275, 281, 282 knowledge 6, 9, 21, 151-
n. 16 152, 163, 195-196, 198-
samanyavisesa 25 n. 12, 154, 199, 203 as sound 173,
162, 186-187, 193, 206 199 (See also eternity)
n. 24, 279 n. 12, 281-282 Sridhara 158 n. 19
samavaya 133 n. 67, 146 n. staff 170, 178, 180, 240,
91, 155, 187, 207, 208 n. 263, 266-267, 276, 285-
29 (See inherence) 286
samavayin 26 n. 18, 187, 212 substance, nature of 130,
n. 34, 279 n. 12 142, 153, 176-177, 213,
samavayikarana 176 n. 9 244, 251-255, 274, 276,
Saiikhya 25 n. 12, 139 n. 79, 287, 288, 291 (See also
151 n. 4, 207 dravya and individual)
Safikhyakarika 7 sun analogy 49, 56, 60, 120,
Sanskrit grammar 6 125, 130 n. 62, 143, 200
Satapathabrahmana 276 n. 7
science 1, 10, 22, 45 Tandyabrahmana 123 n. 48
secondary reference 9, 122 n. Tantravarttika, translation 8,
46, 167, 170, 178-80, 206, 208, 212
257 Tatparyatika 3
semantic condition 6, 22, 33, Taittiriyabrahmana 126 n. 56
61 n. 58, 67, 75-78, 85, Taittiriyasamhita 123 n. 48,
88-89, 180, 291 126 n. 56, 276 n. 7, 277
sentence meaning 5-6, 9, n. 8, 278 n. 10
197-198, 212, 244, 251-
255, 257 n. 63, 258, 260- Uddyotakara 3-4, 8, 20,
261, 280-281, 287-288 154-161, 164-167, 215
n. 23, 294-295 n. 38, 216, 255-256
shape 11-18, 23, 27-31, 34- Uttaramimamsasitra 4
35, 38, 153, 157, 165,
174 n. 6, 212-223, 227, Vacaspatimisra 3, 153 n. 14,
255-261, 264, 280-281, 154 n. 15, 157, 160, 162
289-291, 293-294 (See n. 20, 216
also form) Vaidyanatha Sastrin 270 n5
Sivasutra 1 34, 120 n. 43-44, Vaisesika 3 n. 8, 6-7, 9, 12,
122 n. 46, 130 n. 62, 200 17, 25 n. 12, 29 n. 25,
154, 158 n. 19, 160, 162,

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC
176 n. 9, n. 12, 205, 207, Vyadi 2, 9, 11 n. 11, 48, 99,

208 n. 29, 213, 215-216; 129, 130 n. 60, 133 n.
sutra 1.1.5 176 n. 9, n. 67, 198 n. 8
11, n. 14; sutra 1.1.7 vyakarana (See grammar)
154 n. 17; sutra 1.1.14 vyakti 68 n. 70, 74 n. 77,
176 n. 9, n. 13; suitra 124 n. 51, 125 n. 52, 129
1.1.18 222 n. 44; suttra n. 60, 131 n. 63, 141 n.
1.2.3 154 n. 16; suttra 82, 143 n. 85, 153, 155
2.1.8 192 n. 38; sutra n. 18, 167 n. 28-29, 168
7.2.19 152 n. 10; sutra n. 30, 174-175, 177-178,
7.2.24 152 n. 10; sutra 180, 182, 184, 188-191,
7.2.29 155 n. 18; suttra 201 n. 13, 205 n. 19,
8.1.5 160, 187 n. 29; 206-207, 208 n. 29, 210
suitra 8.1.6 154 n. 17; n. 30, n. 32, 211 n. 32,
vaisesika 191 n. 35 214 n. 36, 219, 221, 229
Vajapyayana 2, 4-5, 9, 11 n. n. 50, 255, 274, 275 n. 5,
11, 48, 99, 116 276 n. 8, 277-278, 279 n.
Vakyapadiya 3, 25 n. 16 12, 280-287, 289 (See
Vamana 3 individual)
vartin 15 vyapti 145-147 n. 90-91, 159
Vatsyayana 3-5, 8, 10, 13, Vyasa 25 n. 12
17-20, 94, 117 n. 36, 151
n. 4-5, n. 7-8, 152 n. 10- wholes 209, 211, 256, 269-
13, 153, 155-156, 158- 271 (See also aggregate
160, 162 n. 21, 163-164, of qualities)
166-168, 173 n. 1-2, 174 word-meaning relation 21-
n. 3, n. 5, 179 n. 22, 191 22, 27, 41-42, 46-47, 58-
n. 35-37, 198 n. 8, 255- 59, 168, 198, 201 n. 13,
256, 289-290, 295 204, 221, 223, 225, 227-
Veda 9, 13, 21-22, 123 n. 48, 230, 232-234, 237, 251,
152, 195-199, 203, 218, 261; learning of 49, 58-
227-229, 244, 253, 263- 59, 117
264, 273-274, 277, 279 word types 22
n. 11, 280 n. 12
veddtiga 21 yadrcchasabda 22 (See also
Vedic words and meanings proper name)
196, 227, 273 yastika 178, 180 (See staff)
vivaksd 4, 9, 25 n. 16, 65, Yogasitra 25 n. 12
67, 68 n. 67, 69-70 n. 72,
82 n. 92, 91, 96, 113 n.
26, 115-116 n. 34, 133 n.
67, 134-135, 140, 141 n.
82, 189-190, 283 (See
also speaker's intention)
Vrttikara 4, 209-211
vyabhicdra 59, 129 n. 60,
210 n. 32, 219, 252 n. 58
(See deviation)

111.93.136.226 on Thu, 05 May 2022 05:09:01 UTC

Nyaya Mimansa

Uploaded by

Copyright:

Available Formats

Nyaya Mimansa

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Nyaya Mimansa

Uploaded by

Copyright:

Available Formats

The Denotation of Generic Terms in Ancient Indian Philosophy: Grammar, Nyāya, and

Stable URL: https://www.jstor.org/stable/1006608

This content downloaded from

The Denotation of Generic Terms

American Philosophical Society

This content downloaded from

ISBN:0-87169-863-3 LC: 94-78520

This content downloaded from

Table of Contents .......................................................................iii

1.2. Substance (dravya) and Class Property (akrti) ............23

1.3. The Denotation of Generic Terms ..............................39

This content downloaded from

1.3.1.5. Reasons for Positing

1.4. Equivalence of the Two Views................................

1.5. Solutions on the View that a Class Property is

1.6. Patanjali's Solution ....... ... ........... ........................ 7

Panini 1.2.58 Translation...........................................................91

Gautama 2.2.58-69 Outline .....................................................169

Mimams ............................................. ........................... 95

This content downloaded from

3.2. The Eternal Nature of a Word

3.3. The Nature of an Individual and of a Class

3.4. Analysis of Arguments in the Akrtyadhikarana ........22

Akrtyadhikarana Outline .........................................................263

G eneral Conclusion ...........................................................289

This content downloaded from

This content downloaded from

This work deals with a topic to which philosophy, most

This content downloaded from

Foundation which gave me the opportunity for

This content downloaded from

By the late fifth century B.C.E. Panini had composed the

ISee Cardona (1976: 238).

This content downloaded from

objects to be of the same kind. For example, one recognizes the

This content downloaded from

Paninians throughout history base their philosophical

This content downloaded from

Initially concerned for the most part with means of corr

9See Nilakantha Sastri (1921).

This content downloaded from

Jaimini's conclusion as explained by Sabara with Katyayana

A considerable amount of scholarship has been do

This content downloaded from

logicians approach making the sense-reference distinction.

This content downloaded from

Indian grammatical philosophy and (pp. 433-548) provide a

Danielsson (1883), Trapp (1933), Chatterji (1957),

This content downloaded from

Panini 2.4.85. However, many terms are left untranslated. He

Jacobi (1929), Hiriyanna (1938), Frauwallner (1960

This content downloaded from

Hiriyanna considers the implications of Vyadi's and

This content downloaded from

both the claims of Mimamsa regarding the nature of speech an

10For more on scientific methodology in ancient India see Staal (1988).

This content downloaded from

The Meaning of the Term 'dkrti'

Although the current work is not intended primarily as

This content downloaded from