The Theory of Definition in Indian Logic

Author(s): J. F. Staal
Source: Journal of the American Oriental Society, Vol. 81, No. 2 (Apr. - Jun., 1961), pp.
122-126
Published by: American Oriental Society
Stable URL: https://www.jstor.org/stable/595050
Accessed: 08-09-2019 15:03 UTC

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 Oriental Society is collaborating with JSTOR to digitize, preserve and extend
access to Journal of the American Oriental Society

This content downloaded from 45.79.69.213 on Sun, 08 Sep 2019 15:03:11 UTC
All use subject to https://about.jstor.org/terms
 122 KRAMRISCH: Piisan

cycle the Nourisher in his sacred aspect and ing his golden weapon at the crossroad, at the
majesty, glowing in the darkness of their night
depth of his cyclical path, where he finds the
where he arises 75 as Liberator (vimocana), swing-
lIidden King Soma, like a lost animal. And they
bhuvane viKve arprtah ( 6. 58. 2 ) which can be understood see him, having found what had been lost (6. 54.
as ' inserted in the whole universe ' but also as ' inserted
in every creature.' Pusan is here the inner guide. 10), now at the other pole of his circuit, aglow
75 At the moment when his foot holds down the fire-with love, at the zenith of creation, at the wedding
brand of the two-faced evil one. In this moment Pusan
is symbolically one with Aja Ekapad and with the of Surya. Pusan leads the Bride towards Soma,
pillar of Life (10.5.6; note 16). across heaven, in their home, the Sun.

TIIE TIIEORY OF DEFINITION IN INDIAN LOGIC

J. F. STAAL
UNIVERSITY OF PENNSYLVANIA

1. qIE THEORY OF DEFINITION according to was the first discipline in

in India
India which
which developed
developed
the Indian logicians has been repeatedly studied in
a scientific technique which
which in
in turn
turn influenced
influenced or
or
a wider context. Foucher has given a description
foreshadowed other domains
domains of
of Indian
Indian thought.5
thought.5
cf its treatment in the Tarka-samyraha, a well-In the Rkpratisakhya (? fourth
fourth century
century B.C.)
B.C.)
known manual of Indian logic written inlaksana
the refers to a grammatical
grammatical rule.6
rule.6 In
In Katya-
Katya-
seventeenth century, and in some of its commen-
yana's varttikas ( ? third century
century B.
B. C.)
C.) as
as quoted
quoted
taries; 1 Ingalls has studied its use in thebynew
Patanjali in the Mahabhasya,
Mahabhasya, laksana
laksana and
and
school (navya-nyaya) ;2 and Mlle. Biardeau has denote respectively
laksya respectively aa grammatical
grammatical rule
rule and
and
considered its place in the development of Indian
its object: sabdo lak.syah sutram
sutram laksarwam
laksarwam "the
"the
thought in general.3 In the present paper a word
for- is the object of the rule,
rule, the
the sitra
sitra (of
(of
malization will be given of the theory of definition
Panini) is the rule." It is said,
said, then,
then, that
that it
it is
is the
the
according to its exposition in the Tarka-samgraha
task of grammar to deal withwith both
both laksana
laksana and
and
and in some definitions quoted by the Nyaya-kosa
lak.sya.7 Actually, Panini's
Panini's grammar
grammar (fourth
(fourth cen-
cen-
(NK) from various logical works.4 In additiontury B. a.) owes part of itsits fame
fame toto the
the fact
fact that
that
we will consider a related concept of Indian the
logic,
rules in it do not prescribe
prescribe how
how toto speak
speak correct
correct
expressed by the term samkara. Sanskrit, but describe the
the facts
facts of
of the
the language:
language:
The terms used for definiens ( laksana)the and
examples (drstanta) for for the
the rules
rules are
are taken
taken
definiendum (laksya) in logic as well as thefrom
idea everyday usage (loka).
(loka). The
The underlying
underlying
of definition have an interesting prehistory inprinciple
the of empiricism was
was laid
laid down
down explicitly
explicitly
work of the Sanskrit grammarians, to which wemuch later. Eslrasvamin
only Eslrasvamin (twelfth
(twelfth century
century
direct our attention here before consideringA.the
D.) formulates it as: laksyamulam
laksyamulam lak$anans
lak$anans
logical doctrines. Such a grammatical prehistory
"lak$ana has laksya for itsits base."
base." Nagojlbhat.ta
Nagojlbhat.ta
of a logical doctrine is not surprising as grammar
says: laksyanusciri vyakEydnam
vyakEydnam "" interpretation
interpretation

1 A. Foucher, Le compendium des topiques ( Paris, 5 See especially L. Renou, "" Connexions
Connexions entre
entre le
le rituel
rituel
1949) 7-14. et la grammaire en sanskrit,"
sanskrit," Journal
Journal Asiatique
Asiatique 233
233
2 D. E. E. Ingalls, Materials for the study of Navga- (1941-42) 105-65; id., " Grammaire et Vedanta,"
nyaya logic, EOS 40 (Cambridge, Mass., 1951) especially Journal Asiatique 245 (1957) 121-33.
80-1. 6 Rkpratisakhya 13. 31 quoted by L. Renou, Terrni-
3 M. Biardeau, " La definition dans la pensee indienne," nolopie grammaticale du sanskrit ( Paris, 1957) 483.
Journal Asiatique 245 ( 1957 ) 371-84. 7 Quoted by Renou, Terminologie 261. Cf. D. S. Ruegg,
4 MM. Bhlmacarya Jhalaklkar, Nyaya-kosa ( Poona, Contributions a l'histoire de la philosophie lingq4ist*que
1928 ) s. vv. Iaksana, ativyapti, avgapti, asombhava, indienne ( Paris, 1959) 33; cf. the present author in:
samkara. Philosophy East aqxd West 10 (1960) 54.

This content downloaded from 45.79.69.213 on Sun, 08 Sep 2019 15:03:11 UTC
All use subject to https://about.jstor.org/terms
 STAAL: The Theory of Definition in Indian Logic
123

follows laksya."8 For a grammatical rule to be replaced by its samjnin. This is expressed in the
correct, it has to conform to usage. This idea has rule: ssarn r?pam sabdasyasabdasamjna " a word
been further developed in Indian musical theory, (in a grammatical rule) which is not a samjna
where laksana and laksya come to denote theory denotes its own form." 11
and practice respectively and where it is stressed The technical terms belong to the metalanguage,
that the actual practice of musicians is the only otherwise consisting of the paribhasa rules which
sound base for a theory of music. For instance, are rules of interpretation or rules which indicate
Ramamatmya (fifteenth century A.D.) says in the how the rules of grammar have to be manipulated.
beginning of his musical treatise Svaramelakala- A typical paribhasa rule is a rule which discusses
nidhi: the relative strength of various sutras or enunciates
general principles regarding the order in which
tasmal laksya pradhanatvam kreva vak$ye'sya
s?btras have to be applied. The status of a pari-
laksanam
. .

bhclsa corresponds to the status of a metatheorem

'therefore I shall speak about the laksana (of
in modern logic.
music) basing myself upon (its) laksya." 9
These two types of element of the metalanguage,
The Sanskrit grammarians made use of the idea samjna and paribhasa, are combined in two state-
of definition in dealing with the technical terms ments regarding the scope of their applications.
of grammar. A technical term is called samjna, These two statements are themselves paribhasa-
and a sutra of Panini which introduces such a rules.l2 The first is: yathoddesam samjnapari-
term by definition is called a samjnasZutra. The bSclsam vc samjna and paribhasa (apply) where
denotation of the sam jna is called the sam jnin. they are taught." The second: karyakalam
For instance, in the first rule of Panini's gram- samjnaparibhasam " sarnjna and paribhasa (apply)
mar: vrddhir adaic " a, ai and ats are called at the time the operations (to which they apply
vfddhi," the three vowels constitute the samjnin take place ) ." The first enunciation limits the
and the technical term vrddhi introduced to denote applicability of a samjna or paribhasa to its im-
these is the samjna.l° Panini is aware of the fact mediate vicinity, i. e., to the immediately preceding
that the technical terms occupy a special position and following rules. The second, on the other
in his grammar in that they denote something else hand, declares the universal validity of a samjna
(z. e., the samjnin), whereas other words which are or paribhasa, whereby the latter fully acquires the
quoted in rules do not denote the object to which character of a metatheorem. There is no general
they refer in ordinary usage, but denote their pronouncement as to the definite choice between
" own form." lIence a word occurring in a rule these two alternatives, for this decision depends on
cannot generally be replaced by any of its syno- each case individually. After Panini the problem
nyms, but a samjna occurring in a rule has to be has become to find in each particular case of word
formation a paribhasa which enables us to manipu-
8 Quoted by Renou rbtd. at. also L. Renou, La Dqzr- late the lak$ana of Panini's sl4tras in such a way
ghatavrtts de Saranadeva (Paris, 1940) I 1, 136. The
that the result of their application conforms to the
philosopher Eankara ( eighth century A. D. ) utilizes an
analogous terminology when he speaks about the sub- laksya. :Kielhorn discusses examples of this in
ordination of reason to revelation: apamanusarr-tarka connection with the above two paribhasa rules in
" reasoning following scripture " (Brahma-sqztra-bhasya his translation of the ParibS4sendusekhara.13
2. 1. 11 ) . In Advaita, revelation is to logic in the
This rapprochement of laksana to laksya is not ob-
supra-sensible realm what sense-perception is to logic in
the realm of ordinary experience. Inference is not based tained by changing the form of the laksana of
upon experience or revelation, for it is a separate means Panini's permanent rules, but by changing the
of knowledge, but it depends for its application on the interpretation. Similarly, apparently superfluous
data of experience or revelation.
9 Evaramelakalantdhz 2. 17 ed. M. S. Ramaswami 11 Panini 1. 1. 68: translated and lucidly discussed
Aiyar ( Annamalai, 1932 ) 9. by J. Brough, " Theories of general linguistics in the
10In the formulation the general praetiee is that the Sanskrit grammarians," Transactions of the Philolopical
sam jnin ( the predicate of the sentenee ) precedes the Socrety (1951) 28ff. This shows that the Indians dis-
sanjna (the subject). The above rule forms an excep- tinguished clearly between connotation and denotation
tion manyalartham " for the sake of auspieiousness," ( cf. the present author in: Ind o-Iranian Journal 4
for it enables Panini to commence his grammar with an [1960] 72 ) .
auspicious word meaning " growth, increase." 12 Nrs. 2 and 3 in Nagojlbhatta's Partbhasendusekhara.

This content downloaded from 45.79.69.213 on Sun, 08 Sep 2019 15:03:11 UTC
All use subject to https://about.jstor.org/terms
124 STAAL: The Theory of Definttzon in Indiwan Logtc

D:3F. 4.
elements in Panini's sutras arex=y means:
notevery element ofbut
deleted x is an
element of y and vice versa.
are pronounced jnapaka rules which disclose laksya
DtF. S. x7Ay
material to which the laksana ismeans: not every element
adjusted. Here of X
is an element of y and vice versa.
is a science which is at the same time empiric in
that it respects and depends
Mrith theonhelp oflaksya,
these notations as also
the logical COIt-
dogmatic in that it adheres
tent of to the expressions
the Sanskrit laksana can bewhich
represented.
was once established. In order to obtain a formalization which is close to
A paribha$a is often called nyayanl4
the structure of the Sanskrit the term
expressionsn we shall
which later denotes (thein system of)
addition make use of a logic. Like-
relation A which holds
wise the paribhasa rules between
of the two Mlmamsa,
classes x and y in the following
which
manner:18 A(x,y) means:
are often rules for the interpretation of"there
theisritual
occurrence
(vrttitva, sattra)
sutras,15 are called nydya.16 Renou of x in has
y." Hence A (x, y) im-
rightly
plies:and
stressed that both Mlmamsa x.y7&0, and - A(x,y) "visent
ATedanta is equivalent to
a donner un enseignementA(s,Y)*
paribhasika, c'est-a-dire
a constituer une hermeneutique, portant
3. The Tarka-dbpika defines sur
a lralidles
definiells
portions du Veda auxquelles ils conferent un role
by: laksyatavacchedaka-samanzyatatvam " co-es-
privilegie." 17 The earlytensiveness
use of(ofthe term
L) with nyaya
the limitor of the class
as synonymous with paribhasa may
of 1," lg which maybe be an indica-
interpreted as: L-1.
tion that elements and laws
Anotherof Indian
definition for logic
L is givencan
by thebeTark-
traced back to the discussion on
bhasa in the problems
following of lan-
terms: dqbsa.na-traya-rahit;o
guage and metalanguage dharrnah
in earlier
" a propertysystems and
which is free from the three
especially in grammar. defects." 20 In each of these three defects we may
2. After this prelude letpresuppose
us considerthat L + 1,thefor this constitutes the
theory
of definition in the laterdefinition
logical of defect. The three defects
doctrines. The referred
to here are ativyapti
theory deals with the various relations which may " overpervasion," avyaptt
hold between the definiens" nonpervasion
(laksana, " and asarnbhava
to be "repre- impossibilittr."
They are defined as (laksya,
sented by L) and the definiendum follows: 21 to be
represented by 1. ) These (1) alaksya-vrttitvatn
relationships (scl. Iaksanasya)
may stivya)-
be
formalized with the help tih of
" overpervasion
a Boolean is occurrence (scl. of L) in the
algebrar
complement
where letters denote classes and whereof 1," whichthemay be formalized as:
notions
A(L,1) or:
of null-clGfss, complement, L. 1+ O. Elsewhere
product, equality 22 this definition
and
tnequality are defined as isfollows:
further extended with the expressioTl: Islk$y-
vrttitte sati, "while (L) occurs in 1," which may
DEF. 1. 0, the null-class, denotes the class to
be formalized as: A (Ij, 1 ) or: L. 1 7& O. Hence the
which no elements belong.
DEF. 2. x, the complement
18 Cf. theof x,
present denotes
author's the
sc Correlations between lan-
guage and
class consisting of the logic in Indianwhich
elements thought" Bulletin of the
are not elements ofSchool
x. of Oriental and Afrtcan Stt4dies 23 ( 1960 ) 109-
222 and: " Fortnal structures in Indian logic" Synthese.
D-EF. 3. x.y, the product of x and
An intere?>ational y,foqdenotes
quarterly the loyical and psycho-
the class consisting of
loyical the
study of theelements
fot4ndations of science 12 ( 196() )
which are elements 279-86.
of x and of y.
19 Quoted NK 696 and Foucher op. cit. 10. The terln
avacchedaka is utilized on account of the occurrence of
13 Bombay 1868, 11 S.
the abstract ( laksyata ) . Cf . the present authol 's
14 Rielhorn, op. att. iv, n. 1.
cc Means of formalisation in Indian and Western logic"
15 There were ritual sutras possessing their own pari-
Proceedinys of the :TIIth International Congress of
bEcl,sa sections ( e. g., Apastaqnba-srauta-sutra 24 )
Philosophy (Venice, 1958) . af.
. Cf. also A. Uno, c; The
L. llenou, Les ecoles qpediques etdetermination
la formationof termsdu Veda
in tTavya-nyaya," Journal o f
( ParisJ 1947 ) 172. Indiax awd Buddhist Studies 7 ( 1958) 61-5.
la See e. g., D. V. Garge, (7itations incit.Sabara-bhasya
20 NK loc.
(Poona, 1952) 252-65. af. llenou in Journat ad
21 T¢rka-dlpika Asiatique
Tarka-sarBgraXla 4. 0¢f. also Tcelssl
233 (1941-42) 119 n. 2. kauqnudt quoted SK 99.
17 L. Renou, Etudes sediques et panineertnes VI ( Paris, 22 ef. Foucher op. cit. 11.
1960 ) 54.

This content downloaded from 45.79.69.213 on Sun, 08 Sep 2019 15:03:11 UTC
All use subject to https://about.jstor.org/terms
 STAAL: The Theory of Definition in Indian Logic
125

definition of ativy2pti is: L. 1 + O) A (L. 17& O). If ativy2pti denotes merely L. 1 1 and avyapfi
14xAMPLES: goh srngitvam " cows are horned denotes merely L. 1 L, case (5) would have been
beasts," or: manusyo brahmanah "Brahmans are neglected by the Indian logicians. Actually the
human beings." examples given to illustrate avyapti are of the
(2) laksyaikadesavrttitvam (scl. Iaksa).zasya) type (5). The first example is explained by the
avyapith "non-pervasion is non-occurrence (scl. Nayaya-bodhinz 23 by observing that there are cows
Of L) in a part of 1," which may be formalized as: which are not black beasts. But it could equally
,-;A(L,1) or: A(L,1) or: L.1+0. Thoughthis well be noted that there are black beasts which are
definition is not further explained by an additional not cows. In other words, the example is a case of
expression, it contains itself the term -ekadesa S(L,1). The second example seems at first sight
" part of " which seems to imply that L does occur to constitute a good example of a valid definition.
in another part of 1, i. e., that A(L, 1) or: L. 17& O. But the commentator explains: there are Brah-
Hence the definition of avapti is: (L.1+0) A mans who do not wear a tuft and a thread, e. g.,
(t.lo). sannyasins. In this case also it could be noted
that there are ksatriyas who do wear a tuft and a
BTAMPLB: gor ntlarupavattvam "cows are black
thread, so that this example too illustrates over-
beasts,' or: sikhasutrav2n br2hmanah " Brah-
lapping.
mans are wearing a tuft and a thread."
Cases of overlapping are actually referred to as
(3) taksyamatr2sattvam (scl. Iaksanasya) asarn-
being defects because of both avy2pti and ati-
bhavah " impossibility is non-occurrence (scl. of
vyapti.24 Hence S(L,1) is defined as:
L) in the whole of 1," which may be formalized
as: L.1 O. [(L 1+0) /\ (L.17&0)] A
P:TAMPLEE3: gor ekasapatvam " cows are solid- A [(L.17&0) A (L.17&0)]
or:
hoofed beasts," or: sundadandav2n brahmanah
"Brahmans are wearing an elephant's trunk." (L.1+0) A (L.17&0) A (L.15£0).
The question arises whether this classification is 4. Elsewhere in Indian logic this concept is
exhaustive. In general there are five possibilities actually defined and esemplified. One of the defi-
for the relations which may hold between the nitions of sarekara "overlapping" quoted by the
clates L and 1: SK 25 iS the following: parasparatyant2bh2va-
(1) (2) (3) s2manadhikaranye sati jatyantarencz, sam2nadhi-
karanyam " that which possesses common-locusness

,A
with another generic character, while there is com-
mon-locusness of mutual relational absence." 26
Another definition speaks analogously of proper-
ties: paraspar2tyant2bhava-s2mcinadhikaranayor
dharmayor ekatra sam2vesah " the coming together
in one place of two properties, which possess com-
mon-locusness of mutual relational absence." Both
L = 1 L.1 - 1 L.l = L
definitions mean that of two classes, which mu-
(4) tually occur in each other's absence, the one also
(s)
occurs in the other. Or, further explained: of two

(X
23 Ibid.

24 I am grateful to Professor Brough for calling my

attention to this.
2S NK 901. An ancient practice in the Samaveda was
the " overlapping " of rks in order to form a saman:

(V)
see the present author's Naxnbudiri Veda Recitation
('s-Gravenhage, 1961) 72-3.
26 For " relational absence " cf. the present author in:
Lol = O Indo-Iranian Journat 4 ( 1960 ) 71. The treatment of
S (L,1)
where S (L, 1) means: the example in the NE seems to show that the term
" L overlaps 1," i. e.: paraspara " mutual " in the above expression does not
(L. l + l) A (L 1 refer to anyonyabhava "mutual absence."
[+L) A (L.lfO)-

This content downloaded from 45.79.69.213 on Sun, 08 Sep 2019 15:03:11 UTC
All use subject to https://about.jstor.org/terms
126 STAAL: The Theory of Definition in Indian Logic

classes X and y, x occurs in the complement of Another example is the overlapping of udbhu-
y, y occurs in the complement of x, and X occurs tatva "perceptibility" and suklatva "whitenessn';
in y. This may be formalized as follows: for while some white things are manifest, there i6
whiteness without perceptibility (e. 9., in the eye)
A(x,y) A A(ynx) A A(x,y)
and also perceptibility without whiteness (e. 9., in
or:
blueness) .28
(X.y7&0) /\ (X.y7&0) A (X.y+O).
The concept of santkarcl is important because of
The example given corroborates this interpreta- a doctrine according to which every entity is de-
tion.27 The relation samkara is said to hold be- termined by at least one general property, which
tween bSutatva " materiality," and murtcltva need not necessarily be a generic character ( jatr):
" limitation, form.') Now bAutatva occurs in for generic characters are either distinct, or one
p7 thivyadi-catustaya " the quadruple earth etc.' includes the other. Overlapping (samkara) is an
(i. e., earth, water, fire and air) and also in gagana impediment to being a jati.29 Hence udbSutatva
or akasa "ether." On the other hand, murtatva "perceptibility" is not a jati because it overlaps
occurs in the same quadruple and also in manas stsklatva " whiteness." " Imperceptibility " (anud-
"mind." II1 other words: materiality and limita- bSutatva), on the other hand, is a jati: for it is
tion overlap. For the quadruple is both limited considered as various (nana), so that each colour-
and material; ether is material but not limited; hood has its own case of imperceptibility.
and mind is limited but not material.
Another impediment to being a jati is (giving
rise to) infinite regress (anavasthiti, anavastha!
(cf. orp*ros avapz7ros).- These doctrines seem to r
flect difficulties which are not dissimilar to those
dealt with in the theory of types in modern logic.
According to Bochenski a point of difference
between Indian and Western logic is the inten-
sional character of Indian logic.30 While this
view, at the present state of our knowledge, re-
mains valid for a large number of cases, the doc-
trines considered in the above show a marked
estensional character.

28 Zarikavat-muktavati 188 and cf. Svaml Mfidha-

27 D«nakortya ad Eartkavali-1nuktavall ad Bhasd- vananda op. crt. 91-2.
pariccheda 8 ( ed. Eari Rama Eukla [Banaras, 1951 ] 29eamkara jatttse badhokah: Ntlokonthr ap. NE loc.
58); cf. NE loc. crt. and Svaml M§dhavananda, Bhaa- ait. Cf. Ingalls op. cit. 42, n. 49.
parsocheda with Siddhanta-rnuktavoli (Almora, 1954) 80 I. M. Bochenski, FortnaIe Lopik (Freiburg/Munchen,
1 1 n. 4. 1956) 517.

A FIFTEENTH CENTURY RUSSIAN TRAVELLER IN INDIA: COMMENTS

IN (:ONNECTION WITH A NEW EDITION OF
AFANASII NISITIN'S JOURNAL

A TJF.XANDER V. RIASANOVS:EIY
UNIVERSITY OF PENNSYLVANIA

IN 21'HE YEAR 146 6, Af anasil, son of Wikita, a was sent as ambassador by Ivan III, Grand Prince
Russian merchant from the city of Tver' (present of Mosconr, to Farrukh Yasar, Shah of Shirvan,
day lialinin) set out on a remarkable journey. Afanasil Nikitin sailed down the Volga River to
Joining VasilI Papin, another native of Tyer', who Astrakhan'. The following year, he crossed the

This content downloaded from 45.79.69.213 on Sun, 08 Sep 2019 15:03:11 UTC
All use subject to https://about.jstor.org/terms