Epistemological Point Of View Of Bhartrhari

Author(s): Ashok Aklujkar
Publisher: Ashok Aklujkar
Catalog link: https://jainqq.org/explore/269402/1

Page #1

The Epistemological Point
of View of Bharthari
ASHOK AKLUJKAR
Professor of Sanskrit and related subjects, Department of Asian Studies, University of British Columbia, Vancouver, Canada. B.A. and M.A. degrees in Sanskrit and Pali from the University of Poona; Ph.D. in Sanskrit and Indian Studies, Harvard University, 1970. Head of the Department of Asian Studies, University of British Columbia, 198085. Research Interests: Paninian grammatical tradition, Sanskrit non-religious philosophy, Sanskrit belles lettres in general. Has published research mainly on Bharthari, the Grammarian-philosopher. Has also published numerous articles and reviews in professional journals, and in 1992 published Sanskrit: An Easy Introduction to an Enchanting Language in three volumes, along with 5 ninety-minute audio cassettes.
$1.1 There are several reasons why Bharthari's Trikāndi (respectively, 'B' and 'TK' in abbreviation) or Vakyapadiya' constitutes a very important watershed in the history of Indian philosophy. I have indicated them in some of my earlier papers (e.g., Aklujkar 1993:$$3.26) and stated them explicitly and somewhat extensively in the Introduction to my forthcoming publication Excursions into Pre-Bharthari Thought. One of the reasons why the TK is a very significant landmark in the history of Indian philosophy—in fact, in the history of philosophy in general is the epistemological viewpoint implicit in it. This viewpoint is, indeed, as I hope to demonstrate presently, quite unique
$1.2 There are some relatively superficial and partial parallels to parts of B's viewpoint in the Samkhya, Yoga and Mimarnsă traditions. For example, the accepted means of cognition are pratyaksa, anumana and Sabda," and, correspondingly, the implicit kinds of cognition are pratyaksa, anumiti and sabda in the TK, as they are in the Yoga-stra and the Samkhya-kärika. The buddhi, roughly speaking, intellect-will or 'mind,' is viewed as essentially consisting of linguistic units in both the TK and the Yoga-bháşya (YB in abbreviation). Thought or

Page #2

2
Concepts of Knowledge
The Epistemological Point of View of Bharthari O
3
epistemological concern in the form of jnana is associated with (a) language or linguistic concern in the form of Sabda and (b) reality or ontological concern in the form of artha in the Sabara-bhasya and the YB, as it is in the TK." However, such similarities are more a sign of the age to which B's work belongs and do not cover the core of his epistemological theory or his original contribution.
31.3 Among the surviving texts on semantics in the tradition of Pāņinian grammar, the TK is the earliest. It is also unique in that it has many statements of epistemological import, corresponding to which the later artha-granthas such as the Vaiyakarana-siddhanta-karika by Bhattoji-diksita, the Vaiyakaraṇabhusaņas by Kaunda-bhata, the Vyakarana-siddhanta-sudha-nidhi by Parvatiya Visvešvara-sári, and the Vaiydkarana-siddhanta-mañjuşās by Nägesa have very few. In fact, as far as epistemology goes, except for the theses of sentence primacy and sphofa in verbal communication, the traditions of Kashmir Saivism and Poetics (alarkara-sastra, kavya-śāstra) may be said to have continued the line of B's epistemology more than the tradition of the Paniniyas. It is not that the later Paniniyas have rejected B's epistemology. They have simply not expressed much interest in restating, extending or modifying it. Being more interested in the grammatical or linguistic semantic (as distinct from the philosophical semantic) side of their topic, they have rarely gone beyond quoting relevant pronouncements of B.
$2.1 The TK, as would be clear to anyone who reads all of its three books, was not primarily composed as a work in what we would call philosophy. As Karikas 1:24-26 state, it pertains to eight topics which are primarily topics we would include in the theory of language or theory of grammar.' Both these theories could be called 'philosophy' in a secondary sense ('philosophy as identification of abstract underlying principles, philosophy as a product of theorizing') in our time, but, as is implied by the overall context of this paper, such an extension of the term 'philosophy is not what I have in mind. That the theories give the remaining philosophy proper a 'linguistic turn' and make B the earliest philosopher known to us who approaches problems of epistemology and ontology from a linguistic point of view is, from B's perspective, an incidental but, from our perspective a very significant achievement
(Aklujkar 1999, lecture 2). I have not read another work in Sanskrit philosophy, except the Yoga-sútra, which achieves so much in so few words. For this reason, we should not expect to find in the TK statements that put epistemology or ontology at the forefront. These branches of philosophy, rather, appear as raising their heads within the frame of a linguistic-grammatical theory. Thus, most of the epistemological observations figure in as analogies or occur as asserted parallels, although, for the purpose of the present essay, I shall look upon them as if they are made with the intention of speaking specifically about epistemology.
$2.2 Even within the limits I have set for myself, some of you will immediately notice that my interpretation of B's philosophy differs significantly from that of Gaurinath Sastri, K.A. Subramania Iyer, Bishnupada Bhattacharya, Madeleine Biardeau and Raghunath Sharmato name only some of the scholars who have published extensive expositions of B's thought utilizing the original Sanskrit texts. This has not been done just for the sake of carving a niche (or erecting a tomb) for myself. Bis considerably earlier than Sankara. Much misunderstanding of him has occurred at the hands of scholars who could not shake off, while thinking of B, what they read in or about Sankara's works. Secondly, none of the scholars whom I have named above has put together or was in a position to put together, because of the earlier bad editing of B's text, all the pieces of evidence which should be taken into account to determine B's philosophy. I undertook this exercise in my unpublished 1970 Harvard University dissertation, The Philosophy of Bharthari's Trikandi. There I have given all the evidence I could gather from his own words and from the words of his ancient commentators for every single view I have attributed to him. I have done my best to read B internally (that is, from within), comprehensively and consistently from a philosophical point of view only. I have also specified the few areas in which he and his ancient commentators may be suspected to differ.
As my objective here is the limited one of determining the general features of B's epistemological viewpoint and as my dissertation will probably be published in a revised form in the next few years, I have not given here the evidence that would support my determination of the general features through the supporting details. If the evidence I have

Page #3

4
Concepts of Knowledge
The Epistemological Point of View of Bhartrhari a
5
gathered in the dissertation for establishing the details is deemed satisfactory, it follows that the general features I see in B's epistemological approach must also be there.
The present paper should also be read as complementing the following papers of mine: 'Pramanya in the philosophy of the Grammarians' (1988), 'The number of pramānas according to Bharthari (1989) and 'Bharthari's concept of the Veda' (1991).
$3.1 The first and most important general feature of B's epistemological-indeed, his entire philosophical approach which we should note is that the approach is linguistic. While it is fair to say that this sort of characterization of his approach is essentially to be found beginning with Hemanta Kumar Ganguli's 1963 book, Philosophy of Logical Reconstruction, neither Ganguli nor his successors in the field such as Bimal Krishna Matilal, as far as I can recall, have explained why this approach should be considered linguistic and in what sense it is linguistic or what its extent as a linguistic approach is. This is what I intend to accomplish in the following.
$3.2 As already intimated in $1.2, we find B speaking of three categories: šabda for vac), artha, and jnana. The first stands for language, the second for things or entities, and the third for cognitions. In other words, B is concerned with (what we would call) a theory of language, with ontology, and with epistemology-with language, reality and thought (note 6). We then find B trying to establish some general theses about language with appropriate argumentation. Once they are established, he simply asserts parallel or logically related conclusions in the sphere of epistemology; he does not try to establish the epistemological theses in their own right. Similarly, as $84.3-4 below will indicate, B merely works out the implications of his theses regarding language for the problems of ontology; he does not discuss ontological problems separately or as independent of considerations of language. One can, therefore, conclude with justification that, in his view, the key to typically philosophical problems, to topics of ontological and epistemological interest, is to be found in reaching certain definite conclusions about language.
$3.3 This evidence furnished by silence, however, is only one piece
among others. That B's approach to epistemology is cssentially linguistic can be argued for also positively. Assigning centrality to language in our awareness of the world, acceptance of the thesis that our view of the world is through language--that we cannot conceive things independently of language, in pure abstraction, is an important feature of the linguistic approach. It is espoused in an absolutely unambiguous way in B's work. According to him, that which is never reflected in language, i.e., is never expressed by a linguistic expression, directly or indirectly, cannot be assumed to exist. An entity or a category of entities may possibly come into and remain in existence independently of language. But it cannot be cognized or determined as existent without language (Aklujkar 1970:36.1). Moreover, B does not stop just with the common assumption of Sanskrit philosophers that forms of cognitions can be judged only from the linguistic forms used to express them. He goes a step further and explicitly states that no cognition either takes place or is noticed without the involvement of linguistic units (see $5.1 below for a possible qualification).
$3.4 As the third piece of evidence, we should note what I have pointed out in my 1988 article on pramanya, namely that the common Indian division of cognitions into pratyaksa 'perception,' anumiti 'inferential cognition,' and sabda 'testimonial cognition' is viewed by B as secondary and non-essential; it is not the case that language is present only in bda and not in the others. In accordance with this, B's notion of sabda as a pramana-agama is his preferred term-is multilevel and is, at its deepest or highest level, very similar to that of conceptual scheme developed in the writings of Quine.
$4.1 Having thus made a case that philosopher B's approach is essentially linguistic, I will now give a few details about how it affects his epistemology. These details will, in turn, bear out the centrality language enjoys in his thinking and thus constitute a fourth piece of evidence in favour of the characterization offered in 83.1.
While developing a theory of cognition, one is required to postulate at least three entities:
(a) sentience, pure consciousness, or life-principle (cit, citi, caitanya),

Page #4

6 Concepts of Knowledge
(b) a cognizor, determinate consciousness or intellect (buddhi). serving as the agent of specific acts of cognizing, and
(c) a state, act, or event of cognizing, a particular cognition (jñāna)." As can be seen from the following table, in what B accepts as existing/ in the field of cognition, that is, in the ontology of his epistemology, so to speak, there is a correlation of these entities with the levels of language he accepts:
vaikhari 'language/1, speech' madhyama 'language2' pasyanti 'language3' param pasyantyähḥ rūpam
=jñāna 'cognition' -buddhi 'mind, intellect" -cit/citi/caitanya 'sentience, consciousness'
sükşma nityă atîndriyā vāc, vācaḥ uttamaṁ rūpam, sabda-tattva-brahman, *language-principle, language/4".
Not only is there a correlation, B identifies the three entities-cit, buddhi, and jñāna, respectively with language/4 (par pasyanti-rūpa), language3 (pasyanti), and language2 (madhyama). This is not explicitly stated in his extant works. But there are numerous indications in his remarks to the effect that the situation must be as I have outlined."2 His ancient commentators too give or presuppose the equations I have given.
The reasoning behind the identifications probably was this: If we find no reason to hold that the intellectual" counterparts of utterances, the totality of linguistic units, and the language-principle are different, respectively, from the cognitions, the intellect, and the principle which assumes the form of these two, then we must conclude that we have here not six entities, but only three. Otherwise, (a) there will be unnecessary postulates in the theory; the criterion of theoretical economy or simplicity (laghava) will be sacrificed; and (b) we will also have subscribed to the naive realism that every name has a corresponding object that six names must imply actual existence of six different entities.14
$4.2 B thus exhibits a distinctive linguistic approach to epistemology.
The Epistemological Point of View of Bhartṛhari 7
The approach may seem extreme. But it can also be viewed as showing the courage of following an insight to all its logical consequences. It is not an ideal language philosopher's approach. B was familiar, at least to a considerable extent, with an ideal language in the form of Panini's metalanguage. But he viewed it too as inherently incapable of remaining unambiguous or free from philosophical problems. Nor is B's approach that of an ordinary language philosopher. He does deconstruct some ordinary Sanskrit expressions that had assumed philosophical significance such as abhava, samaväya and sakti. But he does not accomplish this so much through an exploration of the usage of these words as through drawing attention to one or more general truths about language that he has established elsewhere. In spirit, therefore, his approach is closest to that of Quine and Chomsky-an approach depending primarily and directly upon determining the general nature of language, although its similarity with Quine's overall philosophical position is probably far greater than with Chomsky's.
§4.3 Now, to the second and last general feature of B's epistemology which I wish to point out. I think we are realizing increasingly that, although, as students of philosophy, we distinguish between ontology and epistemology, the two are not independent of each other in any philosopher's practice. A philosopher's ontological views are shaped by or revised in the light of his epistemological views, and vice versa. The truth of this observation is seen in an unparalleled way in B's philosophy at least as far as the Indian philosophical tradition goes.
We get one piece of evidence for the observation if we ask the question: What type of apparatus does B propose in order to explain the fact (or what is generally accepted as a fact) that we experience the world? There is no discussion in the extant TK about the apparatus that must be presupposed in order to explain our knowledge or experience of the world. This apparatus, as mentioned above in §4.1, should be the buddhi". The buddhi or buddhi-tattva of B, as his commentators point out, is identical with pasyanti, the third level or aspect of language. It is not an evolute, either logical or chronological, from prakṛti as it is in the Samkhya and Yoga systems. Nor is it distinguished from manas, ahamhkāra, etc. It is simply citi or caitanya in its aspect of holding the entire diversity off

Page #5

8 O
Concepts of Knowledge
The Epistemological Point of View of Bharthari 0
9
amorosi,
linguistic units (phonemes, words and sentences, i.e., varnas, padas and vakyas) of which a person is aware. There is no physical or material distinction between citi and buddhi. Just a functional one or one od the level of realization of truth.
This is the view of B's buddhi as a static entity-when buddhi is at rest. But there is also a dynamic aspect to buddhi. In that aspect, it is a continuum of cognitions reflecting specific linguistic units (Aklujkar 1970:883.4-5)-cognitions here being sentences or functional equivalents of sentences."
Why, how and in exactly which sense B reduces awareness of the external world to linguistic units is part of his ontological exploration The conclusions he reaches as a result of that exploration have a bearing on his epistemology. The buddhias consisting of linguistic units is a point, where his epistemology lies in with his ontology.
84.4 To explain briefly, B's reduction seems to have its roots in his investigation of linguistic meaning (artha). Through scattered but multisided argumentation, that goes much beyond Frege and those who have developed Frege's insight, B establishes that word meanings are not external things or physical objects." He also points out that it would not do to think of word meanings as images of external things, as precepts. Nor would it be theoretically economical or justifiable to admit a secondary, linguistic, reality (upacara-satta)-something. I believe, corresponding to Putnam's internal realism. If not tied to the perceptible empirical reality, this reality would amount to an uncontrolled licence, a carte blanche, and lead to distinction without difference. Thus, ultimately the distinction between a meaning and its signifier must be only in the view we are accustomed to taking. We must be making an artificial distinction between the cognitions in which signifiers are apprehended and the cognitions in which meanings are reflected; arthamust be a vivarta of sabda (Aklujkar 1970: $85.15-17, 28).
$4.5 В seems to be saying that sense experience must precede conceptualization," although it may not be possible to prove the primacy of sense experience. There is no statement in his writings which would allow us to conclude that denial of the independent existence of physical objects out there in the world was his final philosophical position. At a
particular level in his thought, he unites the generalized existence of physical objects (bhava-satta, maha-satta) with the existence implicit in words or conceptualizations. But, except for a person in the state of spiritual liberation, physical objects are not said to cease to exist (Aklujkar 1970:886.6-8). The question of whether they are, per se, derivable, in the final analysis, from $abda-tattva brahman, the ultimate principle in B's philosophy, is left open."
It suffices for B's purpose that (a) the differentiation of the physical world be viewed as coming from the sabda-tattva brahman, the language principle, and (b) the possibility of a person succeeding in wiping out the traces of the world be admitted.
B is aware of a stronger version according to which the physical objects themselves would be evolutes of the Ultimate, and the physical world would be a lower, dispensable reality-the Ultimate alone would be the final truth. But he neither accepts that version as his only philosophical position nor does he deny the possibility of its being true.
Following his general tendency or strategy of going beyond specific conflicting positions to a non-conflicting common factor or metaposition-of achieving theoretical ascent wherever possible, B takes the minimum he needs to develop a coherent view of his own and declines to be further involved in possibilities that probably cannot, in his view, be proved or disproved logically-that become a matter of accepting this scripture or that scripture, with his personal preference going with the Veda as scripture."
A philosopher wishing to determine whether the physical world consisting of distinct objects or our internal conceptual/linguistic world having a (generally corresponding) diversity should be accorded precedence may conclude that, from our perspective as subjects, the latter should be assigned a more fundamental status on the ground that if the latter did not exist the former would not exist for us. However, this does not amount to denying the existence of the former. In fact, it does not amount to anything more than holding that language is our window to the world (a thesis acceptable to B as pointed out earlier, but one which is really epistemological in nature, not ontological).
$4.6 In consonance with the preceding is B's concept of avidya, the

Page #6

10 O
Concepts of Knowledge
The Epistemological Point of View of Bhartrhari O
11
tying point between epistemology and ontology for many schools of Indian philosophy. His avidya does not consist in sccing sva-bhava or essence where it does not exist (the Nāgarjunian version) or in believing in the reality of the perceptible world (the Vijñana-vadiya, Gaudapādian or Sankarite version). While it can possibly accommodate these understandings of avidya, that is, while it does not necessarily conflict with them, its core consists in not being aware of the distortion that language introduces
Sankara's assertions (a) avidyāvad-visayanyeva sastrani and (b) one does not acquire vidya, one simply removes avidya' have their parallels in B's pronouncement śåstresu prakriya-bhedair avidyaivopavarnyate / anagamavikalpă tu svayam vidyopavartate (TK 2.233), but B's reasons for making the pronouncement are significantly different
B's avidyd consists in the inability of linguistic expressions to refer to things precisely or only as they are. Language is said to introduce distinctions where there are none and to obliterate distinctions where they exist.
The introduction of distinctions is two-dimensional. There is a dimension of space and there is a dimension of time: murti-kriya-vivartáv/ avidya-sakti-pravstti-matram (Vitti 1.1).
Further, since avidya is a property, attribute or capacity of the language principle, it and its two basic dimensions (or its two modes of asserting itself) should be beginningless, just as the language-principle should be beginningless for a philosopher: caitanyavat sthita loke dik kala-parikalpaná / prakrtim prāpinām tam hi ko'nyatha sthapayisyati (TK 3.6.18).
Consequently, in B's view, there has never been a situation at the level of ordinary experience in which there was only pure perception, unattended by any conceptualization, of something in the world. The separation between sense experience in itself and conceptualization is either only theoretical (it is to be accepted as a theoretical necessity, as a kind of a priori) or can be entertained as a possibility only in the case of spiritually advanced beings, that is, only in the case of beings who are not like us-and only at the level of extraordinary experience.
There is no evidential or empirical wedge which would prove the primacy of sense experience.
$5.1 In the preceding, I have implicitly made a distinction between B as a philosopher (in our most prevalent contemporary Western sense of the term) and as a religious thinker. What that distinction suggests is that B, as a philosopher, need not be seen as needing the concept of moksa or, to use his expression in Vrtti 1.5, of brahmanah präptih for the other elements of his philosophy to stand.
Another implication of what I have said so far is that there are levels in B's ontology and they are related to his roles as a thinker. As a Grammarian or Vaiyakarana, he accepts as existing everything that words can denote (even 'hare's horn' is deemed to exist from the Grammarian's perspective). As a philosopher, he admits only the physical things and the language principle as truly existing. Everything else is seen as inseparable from either the things or the principle (time and space, as capacities of the latter, are inseparable from it; all other entities such as qualities, universals, etc. have no separate existence from substance). And as a religious thinker, he entertains the possibility that his philosophical ontic world could be superseded by one in which the language principle alone remains. The assumption then is that a person canreach a certain stage in which his mind (= buddhi, paśyanti)is divested of diversity and he becomes' the language principle.
$5.2 The distinction between B's roles as a Grammarian thinker and as a non-Grammarian thinker is conveyed by the remarks of his ancient commentators, particularly the remarks of Helă-răja. The differentiation between a $abda-pramāņaka ontology and a non-éabda-pramanaka (in effect, corresponding to our philosophical) ontology, which Hela-raja makes," has support in B's remarks. That the non-sabda-pramanaka ontology is not explicitly characterized as philosophical or is not further divided into philosophical and religious is due to the absence of distinctive terminology for philosophy and religion in the Indian tradition."
Post-script:
As the preceding text was read toward the end of the seminar, it then contained, at this point, the following additional observations pertaining to the carlier exchanges in the seminar:

Page #7

12 Concepts of Knowledge
(a) The discussions we have had so far in this seminar have done justice, as far as Indian philosophy is concerned, mainly to the Nyaya concept of knowledge and have almost entirely left untouched those branches of Indian philosophy which are not included in the traditional enumeration of six astika darśanas.
(b) The panditic tradition of the last two or three hundred years has been caught in the grooves of certain texts. It needs to bring back into its discussions texts such as Mandana-miśra's Vibhrama-viveka and B's TK. It also needs to learn to discuss texts with a sharper awareness of the historical development of ideas.
(c) There are anticipations in B to the effect (i) that a distinction between belief and knowledge is useful only as a convenient theoretical fiction, (ii) that a distinction between cognition, mental state, attitude or disposition, and feeling is also useful only as a fiction convenient in our philosophical deliberations and (iii) that the Vaiseṣikas did not remain pure realists when they accepted qualities, movements, universals, particulars, inherence and absence, just as, as was pointed in Dr. Wada's paper, they compromised their realism when they accepted numbers beyond 'one/1'.
The Epistemological Point of View of Bhartṛhari 13
Notes and References
1 (a) The extant works of the Grammarian-philosopher Bharthari, 'who may or may not be identical with the poet-philosopher Bhartṛhari, are: (i) Trikändi or Vakyapadiya, (ii) Mahābhāṣya-tika or Tripädi, published under the title Mahabhasya dipika.
(b) The physical structure of the Trikandi, which word roughly means '3 books," and the relationship of that structure with the available ancient commentaries is as follows:
Brahma-kanda or Agama-samuccaya (kärikäs +Vṛtti):
țikä on kärikäs as well as Vrtti by Vrsabha
Vakya-kanda or Vakyapadiya proper (kärikäs Vṛtti): tīka on kärikäs only; said to be by Punya-raja but more likely to be an abridgment of a tika by Hela-raja. Pada kända or Prakirṇaka (kärikäs in 14 samuddeśas): tika by Helä-raja
2 For an indication of why I do not employ here the word 'knowledge,' meaning 'valid cognition' or 'reliable awareness,' see the last remark in Aklujkar 1988:§2.7 and note 22 that goes with it.
3 (a) See Aklujkar 1988:§§2.1-2.
(b) The actual terms used may be different (e.g., agama and anusravika), and they may have connotations revealing the special concerns of the systems. That their core contents agree is alone relevant here.
4 (a) The terms anumiti and sabda themselves may not occur in all the texts mentioned here.
(b) There is a sense in which all cognition would be sabda according to the TK philosophy. This is pointed out in §3.3 below and in Aklujkar 1988 and 1989.
5(a)... sabdo'pi buddhisthaḥ TK 1.47; ... Sabda-bheda-bhavana-(bija)/nugate buddhi-tattve yo'yarh sabdaḥ... Vitti 1.47; ... sabda-sakti-sarhsrstaya sabdanuviddhaya sabdatmikaya buddhya...Vṛtti 1.123;... väg-rüpāyām buddhau ... Vṛtti 1.133. Supporting indications are available in: Vitti 1.53, Vṛtti 1.130.
(b) In the case of the YB, the evidence for the view attributed to it here is in the form of implication of several statements, that is, indirect.
There is a movement away from sabda when the practitioner progresses from savitarka to nirvitarka samadhi (YB 1.42-43).
Similarly, there is a movement away from narrow intentional-referential activity, oriented toward specific individual objects, to increasingly wider or subtler categorial intending-referring in the substages of savicara samadhi. The substages are expected to culminate in such a wide or subtle intending-referencing that, strictly speaking, it no longer remains intending-referencing-the act involved loses all specificity; it becomes nirvicara (YB 1.44-45).
This is from the point of view of the object. From the point of view of the meditating subject, similarly, inward referring is expected to grow gradually less and

Page #8

14
Concepts of Knowledge
The Epistemological Point of View of Bharthari O
15
yogya-bhava/yogyatá
sambandha
compatibility relation between signifier and signatum.
pratyayanga (pratyaya) sambandha
relation responsible for cognition.
dharmainga
sambandha
relation responsible for (cognition and) merit.
less specific, ultimately ceasing to take cognizance even of meditational bliss (ananda) and I awareness (asmita)-lo become, as YB 1.17.18 puts it, nirvastuka. artha-súnya.
These statements indicate that the starting point assumed by the YB for the journey from the ordinary states of consciousness to those of samadhi is a mind which has the means of intending and / or referring, namely internalized linguistic expressions.
(c) In the above, I am not suggesting that the notion of buddhi is identical in the TK and the YB. Unlike in the TK, there is evidence of acceptance of the Sankhya notion of buddhi in YB 2.15.2.20: 3.35.
(d) Those statements in both the works in which buddhi is used in the sense 'cognition, cognizing act or event,' as distinct from 'cognizing organ' (e.g.YB 3.17) are, of course, not intended in the present context.
6 (a) Sabara on Jaimini 1.1.5 has a long discussion bounded by the notions of sabda, artha and janá/vijnana.
(b) gaur iti sabdo, gaur ityartho, gaur itinam ityavibhagena vibhaktandm api grahanam drsam. YB 1.42.
(c) Words indicative of the streams of thought in the TK, as well as of B's areas of thinking, are:
Sabda/vie "linguistic expression (regardless of extent), "language language-and-grammar theory
IMAGINE A SEPARATE artha thing, entity.' 'reality' ontology. COLUMN AFTER THE
jnana 'cognition, mental state,' thought' epistemology. C. TK 3.1.103; 3.3.59; also, ctaddhi tat sarvar yad utajnah wie arthasceti Vrsafha 1.122, p.181.
7 Eight topics or subjects as stated in TK 1.24-26 (this list of topics is primarily applicable to the kanda in which it appears, that is, to the first book):
8 (a) Especially clear is the evidence furnished by the remark: sad api vágvyavaharcnánupagrhitam artha rapam asata sulyam. Vrtti 1.129
(b) The point made here has a bearing on B's notions of vivarta and parinama and our understanding of them. 514.4-6 will observe that B's notion of vivartais different from the one commonly associated with post-Sankara philosophers. Consequently, his notion of parin ma is also different. A detailed delineation of these notions, however, can be undertaken only in a separate publication.
9 (a) na so'sti pratyayo loke yah sabdanugamad ste/ anuviddham iva janam sarvash sabdena grhyate TK 1.131. Variant readings for the last word are: bhasate, gamyate, vartate.
(b) B's view given here does not imply that one must be aware of linguistic units at the time of cognition. He argues for his view by concentrating specifically on those cognition situations in which one is generally not aware of the role of language (Aklujkar 1970: $1.19-26).
10 B may further be understood as holding that, ultimately, there is no division between analytical truths such as 2 + 2 - 4' and synthetic truths such as 'Snow is white.' If we acquainted him with our notion of conceptual scheme, he would probably come up with a conceptual scheme that is inter-related and interdependent.
11 (a) It may be objected that some philosophers (e.g., the Buddhist philosophers) have not accepted the need to postulate all three entities. However, note that, in the case of entity (a), I am not saying that it must be a constant, stable, durable or eternal self. All I am saying is that a life force or fulfilment of the condition of being a living being is required. The Buddhists are not known to have rejected this (rather obviously needed) condition of ordinary experience. They have come close also to accepting the buddhi in their acceptance of a vijana-santana, alaya-vijñana, etc.
(b) Whether the Sanskrit terms used to designate the three entities are precisely the ones I have given here is not a major consideration. A terminology like citta, caitta, etc. will also do.
12 For example, sabda-tattvam evedash va manasakhyam avibhagam Vrtti 1.86.
13 I am using the neologism intellectal' to avoid the connotations of friental and "intellectual," but I do not hold that its introduction is absolutely necessary.
14 B indicates awareness of the arguments that are likely to be made to assert a real difference between the linguistic entities, on the one hand, and the epistemological entitics, on the other, but he does not explicitly deal with them. Given the nature of the problem, the onus of proving that there is in fact a difference should be on those who
anvåkhyeya
Sabda
vikya and pada 'sentence and its (ordinarily) meaning bearing units
pratipadaka
sabda
stems, roots, affixes (artificially meaning-bearing)
meaning in the stage of analysis
apoddharapadartha sthita-laksana artha
artha artha
meaning in actual communication, meaning which does not need analysis to come into effect, typically sentence meaning and, at a certain level, word meaning
karya-karanabhava
sambandha
cause and effect relation between signifier and signatum.

Page #9

The Epistemological Point of View of Bharthari O
17
16
Concepts of Knowledge
assert the difference. Their proofs or arguments, it seems, would boil down to a determination of whether cognition without symbols is possible and whether the symbols involved can be essentially different from the symbols that constitute language. B's view, as stated in 99.3, is that cognition without linguistic symbols is not possible. He also holds that language is innate. The view that all so-called nonlinguistic symbols are ultimately based on linguistic symbols can also be attributed to him with great probability Thus we can infer that he would prefer to take the position that cognition without symbols is not possible and that the symbols involved will not, in the final analysis, be nonlinguistic. In other words, we can interpret him as saying that a wider definition of language and an admission of all symbols as ultimately linguistic in nature together constitute a better theory than a theory which would take the apparent separation between linguistic and nonlinguistic symbols as a given not to be given up.
15 in Vrtti 2.31, B states the Sathkhya-Yoga view after what seems to be a statement of the Nyaya-Vaibesika view.
16 B advocates primacy of the sentence. He accepts the possibility of single-word and single-phoneme sentences (TK 2.40, 270-71).
17 The Sanskrit terms are: buddhyartha or sabdartha'sense, meaning': bathyartha or vastvartha 'reference, referent, actual thing."
18 I infer this from the direction reflected in B's remarks made in contexts other than those in which sense experience or conceptualization per se is discussed, that is, from remarks made 'unawares,' as it were, in which his subconscious assumptions are likely to be revealed. One such remark is sad api vig-vyavahaireniupaghitam artha-ropam Asati tulyarn (Vrtti 1.129) quoted in note 8. The stance or assumption of its author is clearly that physical things exist first and then come into the purview of language, implying that, in general, sensing precedes concept formation (although he would still be free to hold that no particular time can be specified at which the intertwining of the physical world and language can be said to have begun; cf. the characterization of avidyd, which is primarily a language-based notion in B's thought (94.6), as beginningless. We can probably detect similar evidence in other statements of B, suggesting that in his view physical things and their experience through senses deserves priority in systematization.
19 I say this despite several current interpretations in which B is represented as deriving both the concepts and physical objects from labda-tattva brahman or as declaring all diversity in the world to be an illusion or a lower level truth superimposed on sabda-tattva-brahman. I do not agree with those modern interpreters of B who, by adopting a later version of the 'vivarta: parinama' distinction declare B to be a parinama vadin or a vivarta-vadin. As I will point out in a future article, the understanding of vivarta and parinama in the early period of Indian philosophy, to which B belongs, was different.
20 Note the citations in the Vrtti of TK 1. 124. The karikaitself reads thus : Sabdasya parinamo'yam itymnayavido viduh/chandobhya eva prathamam etad visvanh vyavartata
21 (a) See Aklujkar 1991 for B's concept of the Veda, according to which even the nastika philosophies have their ultimate source in the Veda.
(b) The remark made here, obviously, has a bearing on the question of B's religion. I view him as a follower of Brahmanism or Vedism who was not anti Buddhist and who was, most probably, not anti-Jaina either. I expect to be able to be more specific in a future publication.
22 The words quoted here are a part of a longer statement in Sankara's bhasya introducing the Brahma-sutra or Vedanta-sútra 1.1.1. Cf. Upadesa sahasri 1.1.40, summarized in Potter 1981:221.
23 (al Since acquiring vidyd would mean acquiring liberation, Sankara's statement in the Brhadaranyaka-bhasya 3.3, summarized in Potter 1981:195.96, to the effect that liberation, truly speaking, is not acquired it is not a production, an attainment, or a modification or purification-implies that vidya, "liberating knowledge,' is, truly speaking, not acquired. Note also Upadesa sahasrl 2.17.22. summarized in Potter 1981:240, where vidya is said to be revealed not produced or obtained.
(b) It follows that, in the following statement in Potter 1981:6, the use of 'manifested' is appropriate, but the use of acquiring could be misleading: 'Since bondage depends on ignorance, liberation is manifested upon the removal of ignorance by acquiring its opposite, namely knowledge (vidyar.
24 Compare Quine's procedure of beginning the investigation of what exists with the position everything.'
25 sabda-praminakandi Miyacchabda Aha tat paramartha-rpam. Held-raja 3.1.11 p. 24.8-9. Sabde vyavahäre niropitasyaiva vastutvát. Hela-raja 3.7.152 p. 351.18. näsmábhir darśana-vivekah prärabdhah. Hela-raja 3.9.58 p. 70.25-28.
26 asamakhyeya-tattvanam arthanath laukikairyatha / vyavahare samakhyanath fat prijño na vikalpayer TK 2.142; Sabdapramadako lokah sa sastrendnugamyate TK 3.7.38; also TK 2.296-97 and Vrtti thereto; sarvatra i prasiddhist värtha - vyavastha karapam. Anavasthitaiva hi tarkigamabhyaish bhinnesu pravidesu vastu gata vyavastha. Vrtti 1. 106.
27 (a) The intention behind the remarks made here is not to suggest that B's interest in what we would call religion or religious philosophy was not sincere. In his own perspective, the mokya or brahma-prapti frame of his thought could, of course, have been meant seriously. The question of the relationship of the Ultimate or First Principle in his philosophy with the material world, however, is a philosophical question (one to be determined on the basis of logic and argumentation that does not rest on testimony, appeal to extraordinary experience of some kind, or depend on admitting an unverifiable possibility). The answer given to that question does have implications for religious thinking (e.g., for the notion of moksato be accepted and for the method to be advocated for moksa attainment), but the answer itself is not decided at the level of religious thinking. The approving references to the moksa possibility that B may be seen as making or the moksa-supporting citations he gives from texts authoritative to him cannot,
wa shart.
1
20.25-2
# = #

Page #10
18 O Concepts of Knowledge The Epistemological Point of View of Bharthari 0 19 therefore, be used as argument to establish that the religious ontology is B's final philosophicalontology as well and that, since in that ontology only his language principle remains, the physical world must actually be somehow derivable from it in his philosophy too. (b) It also should not be said by way of objection that, if B's ontological thinking is restricted the way I have done, he would become a Dvaitin, whereas the tradition holds him to be an Advaitin. The 'Dvaitin: Advaitin' distinction pertains to the level of religious ontology. B would still remain an Advaitin at that level. Bibliography and Abbreviations (Publication details have not been given in the case of those Sanskrit texts which are available in many editions and in whose case access to a particular edition is not likely to make any difference. The publications of well-known and much published Western philosophers have also not been specified.) Aklujkar, Ashok. 1970. The Philosophy of Bhartphari's Trikandi. Ph. D. dissertation, Harvard University. Unpublished. -1988. 'Pramanya in the philosophy of the Grammarians.' Studies in Indology: Professor Rasik Vihari Joshi Felicitation Volume, pp.1528. Eds. Kumar, Avanindra, et al. New Delhi: Shree Publishing House. -1989. "The number of pramanasaccording to Bhartphari.'. Wiener Zeitschrift fur die Kunde Sudasiens 33:151-58. -1991. 'Bhartshari's concept of the Veda.' Panels of the VIIth World Sanskrit Conference. Vol. IV-V. Ed. Bronkhorst, Johannes. pp. 1-18. Leiden: E.J. Brill. -1993. 'An introduction to the study of Bhartphari.' Asiatische Studien/Etudes Asiatiques 47.1: 7-36. -1999 (forthcoming). Excursions into Pre-Bhartshari Thought. Three lectures given at the Bhandarkar Oriental Research Institute, Pune. Expected to be published by the Institute. B = Bhartphari. Trikandior Vakyapadiya. I have reproduced the text of the karikas and the Vstti from my critical edition under preparation. Those wishing to verify my references to the Vstti prior to the publication of my edition should consult the editions by K.A. Subramania Iyer: (a) Vakyapadiya of Bhartshari with the Vitti and the Paddhati of Vrsabha deva. Poona: Deccan College Postgraduate and Research Institute. 1966. Deccan College Monograph Series no. 32. (b) The Vakyapadiya of Bhartshari, Kanda ll with the Commentary of Punya-raja and the Ancient Vitti. Delhi, etc: Motilal Banarsidass, 1983. I have followed the cnumeration of karikas in: Bhartshari's Vakyapadiya, Die Mulakarikas nach den Handschriften herausgegeben und mit einem Pada-Index versehen. Rau, Wilhelm (cd.). Wiesbaden: Kommissionsverlag Franz Steiner GMBH, 1977. Abhandlungen fur die Kunde des Morgenlandes XLII. 4. Hence the numbers in my edition and those in the editions by Subramania lyer do not always match. However, they arc not far removed from cach other. For the third kanda of the TK, which has a commentary by Hela-raja, see: (a) Vakyapadiya of Bhartshari with the Commentary italic of Hcla-raja. Kanda III, Part 1(samuddesas 1-7). Subramania lyer, K.A. (ed). Poona: Deccan College Postgraduate and Research Institute. 1963. Deccan College Monograph Series no. 21. (b) Vakyapadiya of Bhartshari with the Prakirnaka-prakasa of Hela-ra ja Kanda III, Part II italic (samuddesas 8-14). Poona: Deccan College. 1973. (Continuation of Deccan College Monograph Series no. 21). Ganguli, Hemanta Kumar. 1963. The Philosophy of Logical Construction. Calcutta: Sanskrit Pustak Bhandar. : :." Hela-raja: see under B.. ...... . ..... ** Potter, Karl H.'1981. (ed.) Advaita Vedanta up to Sankara and His Pupils. Motilal Banarsidass CDelhi: MB. The Encyclopedia of Indian philosophies, vol. II. TK = Trikandi. see B. . Vrtti: see under B. Vrsabha: see under B.. YB-Yoga-bhasya: sec Vyasa under Patanjali. * Patanjalii Yoga-sutra with the commentaries of Vyasa and Vacaspati-misza. (ed.) Vidyasagara, Jibananda. 3rd eda, Calcutta 1940.