Pramanas And Language Dispute Between Dinnaga Dharmakirti And Akalanka

Author(s): Piotr Balcerowicz
Publisher: Piotr Balcerowicz
Catalog link: https://jainqq.org/explore/269198/1

Page #1

Journal of Indian Philosophy (2005) 33: 343-400 DOI 10.1007/s10781-004-6673-6
© Springer 2005
PIOTR BALCEROWICZ
PRAMANAS AND LANGUAGE: A DISPUTE BETWEEN
DINNAGA, DHARMAKĪRTI AND AKALANKA*
In Three Simple Chapters (Laghiyas-traya, LT), a concise though seminal work on epistemology written by a Digambara Akalanka Bhatta (c. 720-780), we come across a verse that contributed to some controversy and to a variety of interpretations. It reads (LT 10cdllab):
jñānam adyam matih samjñā cintă câbhinibodhikam/ prän näma-yojanāc chesam śrutaṁ sabdânuyojanāt //
Its contents is rather lucid and straightforward, it would seem. At first attempt, we could render it as follows:
Int. I 'The first [kind of) cognition (comprises) sensuous cognition (here: memory), recognitive cognition, and inductive thinking (and) determined cognition, (and occurs) before the association with name (speech). The remaining (kind of cognition) is testimonial (mediated) cognition on account of the application of speech.'
The interpretation of the verse should not be very problematic. In it Akalanka continues his exposition of his epistemological ideas. Here he iterates the idea that there are only two cognitive criteria (pramāna),' divided into perception (pratyakşa) and indirect cognition (parokşa). The first variety comprises four subtypes: (1) sensuous cognition (matijñāna), which embraces all cognitive processes that are directly based on the activity of sense organs and that culminate in the act of retention (dhāraṇā), or saving the directly perceived image of an object to the residue memory; (2) recognitive cognition, the idea
* The present paper is a part of my research subsidised by the Alexander von Humboldt-Stiftung.
For the reasons why I consistently render pramāna as 'cognitive criterion' see: Balcerowicz (2001a: 140-145, n. 4).

Page #2

344
of which is the determination and identification of things through comparing them with their image preserved in memory; (3) inductive thinking, or association, that consists in generalisation on the basis of a series of single similar events; in the course of it, general ideas, images, principles and rules are derived that may later serve as a background for universal generalisations, indispensable to any inference; and finally (4) determined cognition which completes sense-related cognitive processes with a definite conclusion; for that reason the final stage is also called 'inference' (anumāna).2 The second variety includes all remaining kinds of cognitive process, which in the above verse are given a joint name of testimonial (mediated) cognition (śruta), the common feature of which is that they are all mediated cognitions, as opposed to the first variety of cognitions that are acquired directly by the cogniser. The very possibility of cognition to be mediated, i.e., to be conveyed to other people, is due to language, which is the vehicle of information and the means of expressing the contents of one's own knowledge. For this reason, Akalanka seems to relate the first variety of cognition, viz. perception, with non-verbal sphere, and the second variety, viz. indirect cognition, with the use of language and application of verbal symbols.
The continuator of Akalanka's ideas, a commentator of a number of works and an author of independent treatises, Vidyananda Pātrakesarisvāmin (first half of 9th century) quotes the above verse of Laghiyas-traya and offers his own paraphrase of its contents, which entirely agrees with Interpretation 1:
PIOTR BALCEROWICZ
Int. 2 The first [kind of] cognition [comprises] memory, recognitive cognition, inductive thinking and determined cognition; it arises before the association with name (speech). The remaining [kind of cognition] is testimonial (mediated) cognition [that arises] on account of the application of speech.
As for this idea, Akalanka said the following: "The first [kind of] cognition [comprises] sensuous cognition (here: memory), recognitive cognition, and inductive
2 Clearly, the Sanskrit term anumana has a much wider connotation than 'inference'. In a number of contexts it refers to all kinds of reasoning which is based on symbols (including speech elements) of any system of coding; it proceeds from one symbol (word or idea), that has already been comprehended, to another, new one, of which we know it is related to the former one. In this way the term 'anumana' is applicable to typical cases of inference, but also to the use of language in which words as symbols connote ideas (universals), which are further related with each other or with their referents. This aspect of anumana is clearly visible, for instance, in Dinnaga's system. This also explains why Akalanka takes abhinibodha and anumana as synonyms.

Page #3

A DISPUTE BETWEEN DINNAGA, DHARMAKĪRTI AND AKALANKA 345
thinking (and) determined cognition, (and occurs before the association with name (speech). The remaining [kind of cognition) is testimonial (mediated) cognition on account of the application of speech".13
3 TÁVA, 239: jñānam ädyam smrtih samjñā cintā câbhinibodhikam prāg-nama-sarvrtar sesam śrutam sabdânuyojanät // atrákalanka-devāh prāhuh - jñānam adyam smrtih samjna cintā câbhinibodhikam/ pran nama-yojanac chesam śrutam sabdânuyojanāt // These two verses are followed by Vidyānanda's comment, which is additionally interesting in term of the qualitative usage of the particle eva (TSVA, p. 239-240): tatrêdam vicaryate - mati-jñānād ādyad ābhinibodhika-paryāntāc chesam śrutam śabdânuyojanād evêty avadhāranam, śrutam eva sabdânuyojanād iti vā? yadi śrutam eva sabdanuyojanād iti pūrva-niyamah tada na kaścid virodhah, śabda-samsrstajñānasya aśruta-jnanat va-vyavacchedāt. atha sabdanuyojanād eva śrutam iti niyamah tadā śrotra-mati-pūrvakam eva śrutam na cakşur-ādi-mati-pūrvakam iti siddhantavirodhah syāt. sārvyavahārikam šābdam jnanam śrutam ity apeksayā tathā-niyame tu nêsta-bädhästi cakṣur-ādi-mati-pūrvakasyāpi śrutasya paramarthato 'bhyupagamat sva-samaya-pratipatteh. - As regards the idea, it is (now) examined. The remaining [kind of cognition different] from sensuous cognition, which is the first [kind of cognition), (with its varieties) up to determined cognition, is testimonial cognition. Is it the restriction of the following kind: [1] "Testimonial cognition farises) exclusively on account of the application of (articulated] speech" or (2] "Exclusively testimonial cognition (arises) on account of the application of [articulated] speech"? [Ad 2) If it is the foregoing restriction: "Exclusively testimonial cognition (arises) on account of the application of [articulated] speech", then there is no contradiction, because [in this way) it is excluded that the cognition connected with (articulated] speech is non-testimonial cognition. (Ad 1] If the restriction is: “Testimonial cognition (arises) exclusively on account of the application of (articulated) speech", then testimonial cognition is exclusively (such a cognition) that is preceded by sensuous cognition derived from the organ of hearing, it is not (a cognition) preceded by sensuous cognition derived from the organ of seeing and other (senses]; [and] that would yield a contradiction with the doctrine. But if (we assume] such a restriction in dependence on the rule that "testimonial cognition is conventional verbal cognition", then there is no accepted (sc. legitimate) subversion, because testimonial cognition is (then) ultimately accepted - in accordance with the contention of sour own doctrine -as also (a cognition) preceded by sensuous cognition (derived from all sense-organs such as the eye, etc.' For Vidyānanda's further comments that follow this passage see n. 85. To explicate Akalanka's idea, Vidyānanda Pātrakesarisvāmin adds in the same spirit in TSVA, p. 243: śabdânuyojanāt tv eşā śrutam asty akşa-vittivat / sambhavabhāva-sarvittir arthâpattis tathânumā // nāmâsamsrsta-rūpā hi matir eşā prakirtitā / nâtah kaścid virodho 'sti syād-vādâmrta-bhoginām // -
- "But if this (sensuous cognition were to arise) on account of the application of speech, then let it be testimonial cognition (sc. indirect cognition), like cognition derived from sense-organs, and other cognitive criteria such as) equivalence, apprehension of absence (as negative proof) as well as inference (which arise on account of the application of speech). However, this sensuous cognition has been declared to have the form which is disconnected from name (speech). Therefore, there is no contradiction for those who take delight in the nectar of the doctrine of modal description.'

Page #4

346
PIOTR BALCEROWICZ
The only difference between Vidyānanda and Akalanka is the use of the term 'mati' (sensuous cognition), which is replaced by Vidyānanda with an unequivocal term “smrti' (memory). This should not surprise us because Akalanka himself uses these two terms interchangeably several times (cf. LT 25, LTV 61). This apparent synonymity of such seemingly divergent concepts as 'sensuous cognition' (which approximates the idea of perception) and memory' is in itself a bit eccentric, and will necessitate further explanation.
In his commentary on Siddhasena Divākara's Sammati-tarkaprakarana, another Jaina thinker Abhayadeva-sūri (11th century) extensively commented on the issue of the division of cognitive processes and their relation to language, raised by Akalanka in the above verse:
Such a cognition (which occurs) before the association with speech is sensuous cognition; it becomes the remaining [kind of cognition), viz. testimonial (mediated) cognition, which is of various subdivisions, which is produced owing to the application of speech and which is unclear cognition."
'Here, the division is as follows: such cognition (which occurs) before the association with speech, which is divided into memory etc. and which operates as being capable of activating infallible practical action is sensuous cognition; however, (such a cognition) which is manifested as a result of the association with speech is in its entirety testimonial (mediated) cognition.”
Abhayadeva confirms the initial analysis of the contents of Akalanka's verse. Interestingly, we find a new element in Abhayadeva's exposition, which was absent in Akalanka's: the reference to the practical side of cognition which is capable of activating infallible practical action' (avisaṁvādi vyavahāra-nirvartana-kşamaṁ) - a clear influence of the well-know idea of Dharmakirti, viz. the efficacy to execute causally efficient action (artha-kriyā-sāmarthya).
If we suppose that what Vidyānanda and Abhayadeva say, which is unanimous with Interpretation 1, accurately depicts what Akalanka had in mind, we still face a serious difficulty: how is it possible to reconcile such processes as memory, recognition, recollection,
4 TBV, p. 553.13-14: prāk sabda-yojanāt mati-jñānam etat sesam aneka-prabhedam sabda-yojanād upajāyamanam avisadam jñānam śrutam.
5 TBV, p. 553.20–21: atra ca yat sabda-samyojanāt prāk smrty-adikam avisamvādivyavahāra-nirvartana-kşamam pravartate tan matih, sabda-samyojanāt prādur-bhūtar tu sarvam śrutam iti vibhāgah.

Page #5

A DISPUTE BETWEEN DINNĀGA, DHARMAKĪRTI AND AKALANKA 347
inductive thinking (association) or inference with the idea of direct cognition, or perception? And that is precisely what all these interpretations imply. In addition, Akalanka understandably relates inference (anumāna) to indirect cognition (parokşa) at another place (LT 4). Did he not notice the contradiction to which his statement of LT 10cd-11ab leads to, namely that inference, said to be synonymous with determined cognition (ābhinibodhika = mati-jñāna), merges in character with direct cognition, or perception? Was this difficulty not noticed by his commentators?
The premonition that the initial Interpretation 1 may not portray Akalanka's intentions adequately finds its confirmation with Prabhācandra-sūri (11th century), the author of the Nyāya-kumudacandra (NKC), the main commentary on Akalanka's Laghiyas-traya. His gloss on the verse accentuates entirely different aspects and the overall interpretation conveys ideas that radically oppose Interpretation 1:
Int. 3 'The first [kind of cognition is sensuous cognition. Memory, recognitive cognition, as well as inductive thinking and determined cognition are the remaining [kinds of cognition, different from sensuous cognition,) viz. [they make up] testimonial (mediated) cognition (which occurs before the association with name (speech) and on account of the application of speech."
6 The above translation follows Prabhācandra's commentary, the complete text of which runs as follows (NKC, p. 404.4-405.9): yat prathama-kārikāyām 'seșam' avisadam jñānam ity uktam tat kim? śrutam avispasta-tarkanam 'śrutam avispastatarkanam ity-abhidhänāt. kim yat nāma-yojanāj jāyate 'visada jñānam tad eva śrutam, utânyad api? ity āha - pran nama-yojanāt. namnah abhidhānasya yojanāt pūrvam upajāyate yad aspastam jñānam tac chrutam nāma-yojanâjanitàrthâspasta"jnana-sādharmyād ity-abhiprāyah. cintă ca ity atra ca-śabdo bhinna-prakramah sabdanuyojanät ity asyânantaram drastavyah. tena na kevalam nama-yojanāt pūrvam yad aspastam jñānam upajāyate tad eva śrutar kim tu sabdânuyojanāc ca yad upaJayate tad api śrutam ity sangrhitam bhavati. kim tad?ity āha - samjna ity ādi. cintă ca ity ayam ca-sabdaḥ punar bhinna-prakramah matih ity asyânantaram smrti-samuccayârtho drastavyah. tena smrty-ādy avisadam jñānamśrutam ity uktam bhavati. indriya-prabhavam mati-jnanam tu deśato vaisadya-sambhavāt sāṁvyavahārikam prat yakşam ity uktam, tasya śrutasya kim karanam? ity āha - jnanam adyam karanam. kim-nāma? ity āha - matih iti. na câgama-virodhaḥ - 'mati-pūrvam śrutam" ity-abhidhānāt. - 'What is this "remaining" (cognition of which it was said in the first verse (sc. LT 3) that it is unclear cognition? [It is) testimonial cognition, which is indistinct suppositional reasoning, (thus) on account of the explicit statement: "Testimonial cognition is indistinct suppositional reasoning'. Is testimonial cognition only that which arises through the association with name (speech), or also something else? To answer this, (Akalanka) says: before the association with name (speech). That nonlucid cognition which arises prior to the association with name, i.e., verbal statement, is testimonial cognition, because it is similar to the non-lucid cognition of an object which is not produced by the association with name - such is (Akalanka's) intention. In the following (phrase): and inductive thinking, the speech element and occurs in the irregular position (sc. is displaced) and should be interpreted immediately

Page #6

348
PIOTR BALCEROWICZ
To understand Prabhācandra's commentary it is crucial to notice the following equivocation: he applies the same term 'mati' twice in two different meanings. First, the term is taken in the sense of 'sensuous cognition' in the phrase 'The first [kind of] cognition is sensuous cognition' (jñānam ādyam matiḥ = NKC: jnanam ādyam kāranam. kimnāma? ity āha – matiḥ iti). Secondly, Prabhācandra uses it again in the sense 'memory;' (mati = smrti) in the phrase: “Memory, recognitive cognition, as well as inductive thinking and determined cognition are the remaining skinds of cognition) viz. testimonial cognition' (matih samjñā cintā câbhinibodhikam...seșaṁ śrutam). He tries to justify his interpretational ‘manoeuvre' by way of explaining various positions of the copula 'ca', that is said to be placed wrongly, and suggesting that the actual sequence of words in the verse is slightly different. However, Prabhācandra is justified in taking the term 'mati' to mean 'memory (smrti), apart from its usual meaning, by the Jaina tradition itself, e.g., by Umāsvāmin's remark that 'sensuous cognition, memory, recognitive cognition, inductive thinking [and] determined cognition are synonymous expressions' (TS 1.13: matiḥ smrtih samjñācintâbhinibodha ity anarthāntaram), a statement that also Akalanka refers to at another occasion (SViV 2.17, p. 115.15-18).
The implication of Prabhācandra's interpretation is that such cognitive processes as memory, recognitive cognition, inductive
(Foonote 6 Continued). after the phrase: on account of application of speech. Therefore, not only is testimonial cognition that (cognition) which arises prior to the association with name (speech), but also is testimonial cognition that (cognition) which arises on account of the application of speech - thus has [the idea] been recapitulated. What is this [testimonial cognition)? To answer this, (Akalanka] says: recognitive cognition etc. This speech element "and" [in the phrase) and inductive thinking, occurs again in the irregular position (displaced) and should be interpreted immediately after the phrase: sensuous cognition, as having the meaning of conjunction with memory'. Thus, what this amounts to is the following: memory, etc. are unclear cognition, i.e., testimonial cognition. However, sensuous cognition, which has its origin in senses, has been described as conventional perception (in LTV 4), in so far as it can occur as partially [having] clarity. What is the cause of this testimonial cognition? To answer this, [Akalanka] says: the first [kind of] cognition is the cause (of this testimonial cognition). What is its name? To answer this, (Akalanka] says: sensuous cognition. There is no contradiction with the scriptural testimony [of TS 1.13), because of the explicit statement: “Testimonial cognition is preceded by sensuous cognition" [TS 1.20).' a Cf. TÁVA, p. 237: śrutam aspastam tarkanam.
Sr: yojanāj janitârthâspasta. CA, Śr: atra śabdo.
A, Śr: avisada-jñānam.
B: -prabhava-mati-o. FTS 1.20.

Page #7

A DISPUTE BETWEEN DINNĀGA, DHARMAKIRTI AND AKALANKA 349
thinking or determined cognition are separated from sensuous cognition (mati-jñāna). Such an opinion of Prabhācandra, in turn, contradicts the gist of Umāsvāmin's aphorism of TS 1.13, quoted above. Granted Prabhācandra's exposition be correct, Akalanka would have to renounce another vital tenet of Jaina tradition, which he did accept on other occasions explicitly, and thereby he would inevitably run into contradiction again. The problematic nature of such an interpretation was evident to Prabhācandra, who was quick to add in the concluding section of the gloss (NKC, p. 405.7-9) that: 'There is no contradiction with the scriptural testimony [of TS 1.13], because of the explicit statement: "Testimonial cognition is preceded by sensuous cognition” [TS 1.20).' This artifice of interpreting the sequence of cognitive acts such as memory, etc. in terms of their causes does not explain the difficulty away: these cognitive acts (sc. memory etc.) may be caused by perception but it does not mean that they all share with it the same nature.
To understand the contents of Akalanka's enigmatic verse and its far-reaching consequences, one should take a closer look at the historical context in which Akalanka formulated it and a range of ideas which are at its intellectual background, namely the centuries-old discussion among Indian philosophers what cognitive criterion (pramāna) is and what its true nature consists in. It is also important to take into account some important shifts in its understanding among the Jainas themselves.
The very notion of cognitive criterion' (pramāna) was a foreign body in the system of the early evolving Jaina theory of cognition. It was adopted gradually under the influence of the methodologicalanalytical stream of Anvikșiki and the Nyāya school. The influence is undeniably attested by the choice of terminology and the typology of cognitive criteria which we find in Jaina Āgamas, which follows that of Anvikșiki and Nyāya. The appropriation of the idea of 'cognitive criterion' by Jaina thinkers must have taken place probably after 1st or 2nd century CE, i.e., when the term and its divisions had already become well-established in the Anvikșiki and Nyāya traditions. Initially, in Jaina sources, the term pramāna referred exclusively to classifications known from these two traditions, and comprised: perception (pratyaksa), inference (anumāna), analogy-based reasoning

Page #8

350
PIOTR BALCEROWICZ
(aupamya) and scriptural testimony (āgama).' Interestingly, the notion of cognitive criteria' was introduced into Jainism as an idea (along with its classification), not as a separate term. Before the Jainas started using the term “pramāna', the current term was 'hetu' (Prakrit heū), or ‘source of cognition' or 'cognitive tool. That was also in perfect harmony with the usage we find in the Caraka-saṁhitā.
Further, the earliest fragments of Jaina texts of epistemological relevance (probably after about 3rd or 2nd century BCE) testify that Jainas applied two schemes of classification of cognition that may have coexisted as complementaries.
The first scheme distinguished two basic kinds of cognition: immediate, direct cognition (pratyakşa) and mediated, indirect cognition (paroksa). In the initial phase of the development of Jaina epistemology the borderline between them was regulated by the interpretation of what ‘aksa', or 'perceiving organ', is. Until 6th - 7th centuries, for the Jainas the eye and other sense organs were merely physical or physiological mediators in acquiring cognition. The exclusive perceiving organ' was the cognitive subject (ātman) himself, or "living element (jiva), viz. the soul. All cognition which occurs in the cognitive subject (the soul) without the assistance of any additional factors, such as sense-organs or the mind, was taken to be immediate and direct (pratyakşa), whereas any cognitive act which could not possibly take place without the support of senses, etc. was considered indirect (paroksa).
Such an approach conspicuously differed from the mainstream of Indian epistemology and general usage with which basically most Indian thinkers complied: direct cognition, or perception, originates in senses, whereas any cognition that follows the activities of the mind is an indirect variety of cognition, and these varieties were named and classified differently. No wonder that Jaina terminological usage must have seemed awkward to many and may have led to unnecessary misunderstanding and complications.
The second typology comprised five kinds of cognition: sensuous cognition (mati-jñāna), testimonial cognition (śruta-jñāna), clair
7 Viy 5.4.26[3] (vol. 1, p. 201.1-2): pamāne cauvvihe pannatte, tam jahā - paccakkhe, anumāne, ovamme, āgame.
8 Thān 336 (p. 149): ahavā heū cauvvihe pannatte, tam jahā - paccakkhe aņumāne ovamme âgame.
9 E.g. Cars 3.8.33: atha hetuh - hetur nāmôpalabdhi-kāranam. tat pratyakşam anumānam aitihyam aupamyam iti.

Page #9

A DISPUTE BETWEEN DINNAGA, DHARMAKĪRTI AND AKALANKA 351
voyance (avadhi-jñāna), mind-reading (manah-paryāya-jñānal') and absolute knowledge (kevala-jñāna)." These two models were quickly combined into one consistent scheme which became the standard for centuries to come. It also comprised subdivisions, or stages, of sensuous cognition: sensation (avagraha), cogitation (inā), perceptual judgement (avāya") and retention (dhāraṇā). The scheme found its classical formulation with Umāsvāmin in the Tattvârtha-sūtra
(TS):13
cognitive criteria (pramāņa)
indirect cognition (parokşa)
direct cognition (pratyakşa)
sensuous cognition (mati-jñāna) scriptural cognition (śruta-jñāna)
clairvoyance (avadhi-jñāna)
mind-reading (manaḥ-paryāya-jñāna) - absolute knowledge (kevala-jñāna)
- sensation (avagraha)
cogitation (īhā) - perceptual judgement (apāya) - retention (dhāraņā)
Model 1
10 The Digambaras, including Akalanka (RVār), preferred the variant: manahparyaya-jñāna.
"Viy 8.2.22–23 (p. 336): ...pamcavihe nāne pannatte, tam jahā - ābhinibohiya-nāne suya-nāne ohi-nāne manapajjava-nāne kevala-nāne. ... ābhinibohiya-nāne catuvvihe pannatte, tam jahā - uggaho ihā avāo dhāraṇā. NamS 6[28–29] i 8 (pp. 6, 9): ...nānadarsana-gunānam... nānam pamcaviham pannattam. tam jahā - abhinibohiyanānam suya-nānam ohi-nānam manapajjava-nānam kevala-nāņam.
12 The Svetämbaras prefer the term apāya.
13 TS 1.9-12, 15. Umāsvāmin's actual classification, based on classifications found also in the Agamas, was far more complex and contained the blanket idea of cognitive application (upayoga): (1) jñāna: (1) abhinibodhika-jñāna with four stages: (a) avagraha, (b) ihā, (c) apāya, (d) dhāranā, (2) śruta-jñāna, (3) avadhi-jñāna, (4) manahparyāya-jñāna, (5) kevala-jñāna, (II) darśana: (1) cakşur-darśana, (2) acaksurdarśana, (4) avadhi-darśana, (5) kevala-darśana. Comp. Viy 2.10.9[2] (p. 115): ...jive nam anamtānam ābhinibohiya-nāna-pajjavānam evaṁ suta-nāna-pajjavānam ohi-nāna-pajjavānam manapajjava-nāna-pajjavāna kevala-nana-pajjavanaṁ mati-annāna-pajjavānam suta-annāna-pajjavānam vibhamga-nana-pajjavānam cakkhu-daṁsanapajjavānam acakkhu-damsaņa-pajjavānam ohi-damsana-pajjavāna kevala-damsanapajjavānam uvaogam gacchati, upayoga-lakkane nam jive... See also STP 2.3, 5, 6, 8, 16, 23, 27.

Page #10

352
PIOTR BALCEROWICZ
Model 1 underwent various modifications. One of them was to introduce the division into cognitions that are distinct, having definite contents (sākāra = jñāna), and indistinct, not having any definite contents (anākāra = darśana, 'vision').4
All these historical remarks are necessary to understand Akalanka's intention underlying the problematic verse. It will also be vital to bear in mind that Akalanka penned the Royal commentary (RVār) on Umāsvāmin's Tattvārtha-sūtra, which represented the above Model 1. As the author to the most fundamental Jaina text, which has been considered the paramount exposition of Jaina doctrine until today, Akalanka not only knew the above scheme but also accepted it, at least declaratively. What it practically meant was that he had to avoid any open contradictions with the tradition he glossed.
Around the 7th and 8th centuries some radical changes were introduced into Jaina systematics, which were subsequently endorsed by Akalanka. The 'culprits' were Pātrasvāmin, the author of the lost Trilaksana-kadarthana, Siddhasena Mahāmati, the author of the Nyāyâvatāra (NA),"s and the fundamental shift was directly compelled by certain Buddhist ideas and new epistemological concepts introduced by Dinnāga and Dharmakirti.
14 Cf. the following Canonical model (AnD): pramāna: (1) jñāna: (1) pratyakşa: (a) indriya-pratyakșa (śrotra, cakşus, ghrāna, jihvā, sparsa), (b) no-indriya-pratyaksa: (avadhi, manah-paryāya, kevala), (2) anumāna: (a) pūrvavat, (b) sesavat, (c) sādharmya, (3) aupamya (...), (4) agama [...], (II) darśana: (1) cakşur-darśana, (2) acaksurdarśana, avadhi-darśana, kevala-darśana; see AņD 435-471 (p. 173-179): [435] se kim tam jiva-guna-ppamāne? ... tam jahā - nāna-guna-ppamāne damsana-guna-ppamāne caritta-guna-ppamāne. [436] se kim tam nāna-guna-ppamāne? ... tam jahā - paccakkhe anumāne ovamme āgame. [437) se kim tam paccakkhe? ... tam jahā - imdiya-paccakkhe, noimdiya-paccakkhe ya. [438] se kim tam imdiya-paccakkhe? ... tam jahā – soiṁdiya-paccakkhe cakkhur-imdiya-paccakkhe ghānimdiya-paccakkhe jibbhimdiyapaccakkhe phāsimdiya-paccakkhe. se tam imdiya-paccakkhe. [439] se kim tam noiṁdiya-paccakkhe? ... tam jahā - ohi-nāna-paccakkhe manapajjava-nāna-paccakkhe kevala-nāna-paccakkhe. se tam noimdiya-paccakkhe. se tam paccakkhe. [440] se kim tam anumāne? ... tam jaha - puvvam sesavam dittha-sāhammavam...[471] se kim tam daṁsana-guna-ppamāņe? ... tam jahā - cakkhu-damsana-guna-ppamāne acakkhudaṁsana-guna-ppamāne ohi-darsana-guna-ppamāne kevala daṁsana-guna-ppamāne ceva. cakkhu-darsane cakkhu-damsanissa ghada-pada-kada-raghādievsu davvesu, acakkhu-daṁsane acakkhu-damsanissa āya-bhāve, ohidamsanam-daṁsanissa davvadavvehim savva-paijavehi ya. se tam daṁsana-guna-ppamāne.
15 Since the author of the Nyāyāvatāra is different from Siddhasena Divākara, the author of the Sammati-tarka-prakarana, it is advisable to refer to the former as Siddhasena Mahāmati, see: Balcerowicz (2001b).

Page #11

A DISPUTE BETWEEN DINNĀGA, DHARMAKĪRTI AND AKALANKA 353
One of the two sources of new Jaina classification, which found its earliest and fullest expression in the Nyāyâvatāra, was Dinnāga's wellknown statement of PS and PSV 1.2:
The two cognitive criteria are perception and inference, [because the cognoscible object has two characteristic... For there is no cognoscible other than the individually marked (unique particular) and the generally marked (universal thing), because we shall further demonstrate that perception has as its datum the individually marked (unique particular), whereas inference has as its datum the generally marked (universal thing)."
The passage introduced a new ground for distinguishing two cognitive criteria: two divergent aspects of reality, which can be known either in terms of absolutely unique sensation, or a sense impression which points to itself only, or in terms of concepts that actually convey a synthetised image, embedded in a set of relations that allow the cogniser to group items, construct hierarchies of objects, draw inferences etc., and, finally, to make use of speech, as the medium of all concepts. It is precisely its application that draws the line of distinction between the two spheres:
Perception is free from conceptual construction, (which, in its turn,) is connected with name, class, etc."
Perceptual data that constitute the contents of perception, such as a particular patch of colour, a unique touch sensed in a particular moment etc., carry the information which does not extend over their actuality: the unique particular communicates nothing but its self, or its presence. The contents of ‘inference', which actually stands for any kind of valid speech- and concept-related piece of cognition, is artificially constructed, generated on the basis of a range of perceptual
16 PS, PSV 1.2: pratyakşam anumānam ca pramāne laksana-dvayam / prameyam... na hi sva-sämānya-lak saņābhyām aparam prameyam asti. sva-laksana-vişayam hi pratyakşam, sāmānya-laksana-visayam anumānam iti pratipadayisyāmah. Quoted in Prajñākaragupta, PVA, p. 169.3, 213.6 and by Simha-sūri in DNCV, p. 88.3-89.1; On this quotation see HATTORI (1968: 76-79, notes 1.11, 1.13, 1.14) and E. Steinkellner's notes 44 (p. 28) and 51 (p. 29–30) PVin 112. Cf. also NM, p. 50: As regards one's understanding there are only two pramānas, I mean: inference and direct perception (pratyak sa and anumāna) since (the other pramäņas admitted by different schools) such as tradition (sabda), analogy (upamāna) etc. are include in these two. Thus there are only pramanas, by which can apprehend the thing in itself (svalak sana) and its generality (sāmānyalaksana). There is no other knowable besides these two, which can be apprehended by a pramāņa different from those [already referred to).'
17 PS, 1.3cd: pratyakşam kalpanapodham nāma-jāty-ādi-yojanā. Cf. Hattori (1968: 83, 85).

Page #12

354
PIOTR BALCEROWICZ
data, memory and conceptual framework imposed on the perceptual data by the language; such conceptualisation conveys a generalised object, be it an arrangement of various perceptual data that have been independently apprehended by different sense organs and subsequently correlated with each other into one 'whole', or a set-up of abstracted features, the essential nature of which are its relations with other conceptual entities." It is precisely what facilitates acquisition, or expansion of knowledge of a conceptual object B; of which we have no direct knowledge, on the basis of an object A, which has already become the contents of our cognition and of which we know that it is related to the object B by a certain relation R. Dinnāga maintains, that the rules governing acquisition of knowledge and co-ordination of its contents are the same wherever any system of symbols is involved. All cognitive processes based on symbols, including verbal communication, can be therefore classified as inference, and for this reason
Verbal cognition is not an additional cognitive criterion, different from inference, because it names its object through (the procedure known as] "exclusion of the other" in the same way as (the inference: "x is impermanent, because it is produced", determines its object to have the quality of "impermanence" on the basis of the already known quality of] "being produced" etc.20
For Dinnāga all our knowledge can be organised in two disconnected compartments, the latter of which is defined by two coextensive and inseparable ideas: conceptualisations and speech. Whether we reach the conclusion that there is fire on a mountain by relying on fire's specific mark, or symbol, viz. smoke which we directly perceive, or by relying on the term 'fire', which is verbally communicated to us, both these procedures are essentially not distinct from each other: a symbol triggers the knowledge of its bearer to which it is 'attached'. Both reliable statements referring to things invisible to us and phenomena perceived by us, of which we know are attributes of other objects, serve the same purpose in precisely the same manner: 'Since
18 Cf. Herzberger (1986: 106-144), Hayes (1988: 133-144). As regard the historical background which influenced the development of the idea of the unique particular (svalaksana) as the object of perception see: Singh (1984: 117-135).
19 Cf. e.g., Hayes (1988: 173-219).
20 PS 5.1: na pramāņântaram sābdam anumānāt tathā hi tat/ krtakat vadivat svärtham anyâpohena bhāşate // The original acc. to TSaP 1513-1514. Cf. the translation in: Hayes (1988: 300) and Herzberger (1986: 145-146).

Page #13

A DISPUTE BETWEEN DINNĀGA, DHARMAKĪRTI AND AKALANKA 355
there is similarity in terms of infallibility (as regards) the utterance of an authoritative person, it has the status of inference. What we eventually learn via these two procedures is likewise not different from each other: the cognitive contents is always a general, verbally expressible concept, not a particular individual. In this sense we acquire a vague idea of the inferred or verbally communicated thing, .e., of its general features only that would not, however, allow us to identify any particular individual on its basis, or to distinguish between similar individuals covered by the same concept.
What Dinnāga actually does in his analysis is to search for one clear criterion that help us reduce all sound cognitive phenomena to a minimum number of subtypes of cognitions.
As it is well known, this general conceptual framework was taken over by Dharmakirti who introduced some crucial innovations (PV, 2.1-3):
[U] Cognitive criteria are two, depending on two kinds of data (which, in their turn, are also two] because of their potential in the case of perception) or lack of potential (in the case of inference) to execute causally efficient action. [For instance), a hair-net is not a really existing) object, because it does not (warrant) the fulfilment of [a causally efficient action that concludes with the appropriation of] the object. [2][We recognise that there are only two cognitive criteria) also because of similarity (in the case of inference) and lack of similarity (in the case of perception), and also because of the capability of the data to become the contents in the case of inference) or not (to become) the contents (in the case of perception) of speech, and also because (the object's) image is either present or absent when other causal factors are present. [3] What is capable of causally efficient action is (called] here the ultimately existent; what is different is called the conventionally real. These two are (respectively) the individually marked (unique particular) and the generally marked (universal thing).22
21 PS 2.5: āpta-vādāvisamvāda-sāmānyād anumānatā / As quoted in PVSV4 (p. 108.1) and incorporated into PV4 3.216 / PVSV4 (p. 109.5-11) = PV3 3.216, PVSV3 (p. 72.21-26).
22 PV 2.1-3: mānam dvividham visaya-dvaividhyāc chakty-aśaktitah / artha-kriyāyām keśâdir nartho 'narthâdhimokşatah // sadrśâsadrśatvāc ca visayâvişayatvatah / śabdasyanya-nimittanam bhāve dhi-sad-asattvatah // artha-kriyā-samartham yat tad atra paramartha-sat / anyat saṁvrti-sat proktam te sva-sāmänya-lakşane // Cf. also NB 1.12–17: [12] tasya vişayah sva-laksanam. [13] yasyârthasya samnidhānâsamnidhānābhyām jñāna-pratibhāsa-bhedas tat sva-laksanam. [14] tad eva paramârtha-sat. [15] artha-kriya-samarthya-laksanatvād vastunah. [16] anyat sāmānya-laksanam. [17] so 'numānasya vişayah. - 'The datum for this (perception) is the individually marked (unique particular). An object in the case of which (we experience] the change in (its) representation in cognition depending on [its) proximity or distance is what is individually marked. Nothing else but this (unique particular) is the ultimately existent, because any real thing is characterised by the efficacy to execute causally efficient action. What is different [from it] is the generally marked (universal thing). It is the datum for inference.'
marthya-lak. The datum the case of Wis) prox

Page #14

356
PIOTR BALCEROWICZ
In these three verses Dharmakirti takes over Dinnāga's twofold classification of pramānas in accordance with two kinds of cognoscible objects. In addition, however, the datum for perception fulfils an additional condition: it is endowed with the capability, or potential, to execute causally efficient action (artha-kriya-samarthya). In other words, what is ultimately existent is also causally efficient and triggers perception; when it is subsequently manifested in the cognition as an image (pratibhāsa), or verbally expressible concept, and no longer stands for what it really is but for a whole class or bundle of features, it is neither causally efficient nor ultimately real.25
Inference, on the other hand, is inseparable from conceptualization and mentally abstracted ideas which are based on similarities which - and that is also Dharmakirti's innovation, that goes back to Bhartshari" - do not exist in real individuals but are superimposed on the phenomenal world. The existence of common features in really existing things is a product of the mind, and it is the fundament for concepts and symbols, including linguistic symbols. Similarity, or generic features, cannot be observed directly in perception: it is the mind that correlates all the perceived data in classes in accordance with the framework imposed by language, and the correlatum of the framework is conventional reality (samvrti-satya). Accordingly, the contents of inference, viz. the objects indicated by words, symbols and ideas, are conventions dictated by the language. Ultimately real things, absolutely unique, void of any likeness to other objects and not amenable to verbal means of expression, never enter conceptual thinking or mental images (dhi). Dharmakirti avails himself of the nomenclature coined by Dinnāga (sva-lakṣaṇa and sāmānya-lakṣaṇa), who took recourse to Vasubandhu's statements, 25 but gives it a new
23 Cf. PV 2.53d-54: ...meyam tv ekaṁ sva-laksanam // tasmād artha-kriyā-siddheḥ sad-asattā-vicāraṇāt / tasya sva-para-rūpabhyāṁ gater meya-dvayam matam //
24 Cf. Herzberger (1986: 81-83, 102-103).
28 AKBh2 4.14cd (p. 349.11-13): kāyam sva-sāmānya-laksanābhyām pariksate, vedanāṁ cittam dharmāṁś ca. svabhāva evaişām sva-laksanam. sāmānya-laksanam tv anityatā saņskrtānāṁ duḥkhatā sāsravānāṁ śünyatânātmate sarva-dharmānām. - *The body is (now being] examined, as well as sensation, consciousness and elementary constituents of reality, by way of (their) particular and general characteristic. The particular characteristic is their essential nature alone. The general characteristic, on the other hand, is the impermanence in case of (elementary constituents) that are contingent, the condition of suffering in case of (elementary constituents) that are (suffused) with the [four] defilements and both emptiness and insubstantiality in case of all elementary constituents.

Page #15

A DISPUTE BETWEEN DINNAGA, DHARMAKIRTI AND AKALANKA 357
tinge. Further, in Pramāṇa-vārttika26 and Nyāya-bindu,27 he augments Dinnaga's definition of perception with an explicit idea of nonerroneousness: 'Perception is free from conceptual construction and non-erroneous', a position tacitly accepted by Dinnaga.28 In contradistinction to perception, which pertains to ultimately existent entities that are effectively causal, inference cannot but be erroneous (bhrānta), inasmuch as its actual referents are fictitious.
Another distinction, also introduced by Dinnaga, that will subsequently prove crucial in the context of Jaina epistemological theories, is that of inference for oneself (svârthânumāna), as an internal process of association and drawing general conclusions, and inference for others (parârthânumāna), as a formalised proce
26 PV 2.123-125:
pratyakṣam kalpanapodham pratyakşenâiva sidhyati | praty-atma-vedyah sarveṣām vikalpo nama-samśrayaḥ || samhrtya sarvatas cintam stimitenântarâtmanā | sthito 'pi cakṣuṣā rūpam ikṣate säkṣajā matih || punar vikalpayan kiñcid äsin me kalpanêdrsi iti vetti na pūrvôktavasthāyām indriyad gatau ||
27 NB 1.4: tatra pratyakṣam kalpanapodham abhrāntam.
28 Cf. PSV 1.17 Hattori (1968: 36-37, esp.: Section 3,1 Bb). The idea is also implicit in PS, 1.7cd-8ab, where various kinds of cognition are dismissed as perception proper:
bhrānti-samvṛti-saj-jñānam anumänânumānikam || smartâbhilaṣikam cêti pratyakṣâbham sataimiram |
- (1) The cognition in the form of illusion, (2) [the cognition of] the conventionally real, (3) inference, (4) inferential cognition [as a result of inference], (5) recollection, (6) mental condition tainted by desire along with (7) cognition affected by ophthalmological disorder - these are [varieties of] fallacious perception (sc. are not perception proper).' This idea is further explicated in PSV 1.8ab: tatra bhrāntijnanam mrga-trsnikâdiṣu toyâdi-kalpana-pravṛttatvät pratyakṣâbhāsam, etc. ('Among these [varieties of fallacious perception] the cognition in the form of illusion, in cases of a mirage etc., is a fallacious perception because it is activated by a conceptually constructed idea of water, etc., ... What is unclear in the verse is, firstly how we construe sataimiram. An alternative interpretation of the verse, if we took it as a qualifier of pratyakṣabham, would be: (1) The cognition in the form of illusion, (2) [the cognition of] the empirical reality (the conventionally real), (3) inference, (4) inferential cognition [as a result of inference], (5) recollection or (6) mental condition tainted by desire - these are [varieties of] fallacious perception, which is accompanied by confusion.' The latter interpretation is that which is preferred by Hattori (1968: 28), whereas we find the former (adopted by me) in Bijlert (1989). Another question is which of these cognitions, above enumerated, are genuine cases of fallacious perception. Hattori (1968) concludes that Dinnaga distinguished three kinds of pratyakṣabha (nos. 1, 2, 3), which is contended by Wayman (1978), according to whom Dinnaga distinguished four kids of pratyakṣābha (nos. 1, 2, 3, 7); a compromise solution is offered Franco (1986), who relies on direct suggestions of Lambert Schmithausen (Franco (1986: 85)).

Page #16

358
PIOTR BALCEROWICZ
dure of persuasion.29 These two concepts were taken over not only by the subsequent Buddhist tradition and other Indian philosophical schools, but also by Jaina thinkers, with certain interesting modifications.
IV
In an approach much the same as Dinnāga's, who drew one clear line dividing all valid cognitive procedures into two main compartments, rather than distinguishing a range of pramānas in conformity with a set of insignificant individual features, the Jainas attempted to redefine earlier models, maximally simplify general typology of cognitions and specify new criteria for it in a manner that would be commensurable with general Indian standard. Conspicuously, their previous classification of cognitive criteria and the understanding of which cognition is direct or indirect, was strikingly at odds with what most Indians understood under pratyaksa and parokşa, and must have led to frequent undesired misinterpretations and unnecessary controversies with other schools. Furthermore, engaging in a philosophical discourse necessitated compliance with generally accepted terminology, rather than applying what would appear to others an arbitrary vernacular.
Siddhasena Mahāmati was probably not only the first Jaina philosopher to define what pramāna is, but also to advance an entirely
29 See PS2 2.1ab (anumānam dvividhā svarthaṁ tri-rūpāl lingato 'rtha-drk / - 'Inference is twofold. The first one, inference) for oneself is the discernment of an object by means of the triple-formed inferential sign.') and PS2 3.lab (parârthânumānam tu sva-drstârtha-prakāśanam / - 'Whereas inference for others is the proclamation (sc. demonstration) [to others) of the object one has experienced oneself.'), and PSV.2, K 109a.2–3 = V 27a.5 (svârthânumāna): tshul gsum paủi rtags las rjes su dpag paḥi don (V: rjes su dpag par bya bahi don) mthon ba gan yin pa de ni ran gi don gyi rjes su dpag paho. See: Randle (1926: 28-9), Hattori (1968: 78, n. 1.11) and PVin 112, n 1, p. 21.
30 Cf. e.g., NB 2.1-2: /1/ anumānam dvividha. /2/ svartham parârtham ca.
31 Or, at least one of the first Jaina philosophers. Indeed, Samantabhadra, in Apta-mimāṁsā 101a, characterises cognitive criterion as 'the cognition of truth / reality' (AMi 101a: tattva-jñānam pramānam), or perhaps better: as 'the cognition of reals / doctrinal categories', for what is actually meant are such categories as living elements, lifeless elements etc. mentioned e.g., Tattvärtha-sūtra (TS 1.4: jivajivásravabandha-samvara-nirjarā-mokşās tattvam). Later Jaina tradition, e.g., Vādirāja-sūri in his Nyāyaviniscaya-vivarana (NVIV 1.3, p. 57.22), contends that Samantabhadra's statement is a full-fledged definition of cognitive criterion. Some moderns scholars tend to be deceived by that seemingly obvious definitional character of the statement. The full verse of ĀMi 101 reads as follows: tattva-jñānam pramānam. tam yugapat sarva-bhāsanam/ krama-bhāvi ca taj-jñānam syād-vāda-naya-sanskrtam // $

Page #17

A DISPUTE BETWEEN DINNAGA, DHARMAKIRTI AND AKALANKA 359
new classification of cognitive criteria, albeit resting on former phraseology. In the opening verse of the Nyayâvatāra he asserts:
(Foonote 31 Continued).
- 'Cognitive criterion is the cognition of reals (doctrinal categories). It [either] is a simultaneous representation of all [entities (which is the case with the absolute knowledge (kevala))] or occurs consecutively (of this kind are remaining varieties of cognitive criteria). The cognition of these [reals] is composed of the method of sevenfold modal description and of [conditionally valid] viewpoints.'
[§ Here, I follow an emended text (see: Balcerowicz (1999: n. 45), for AMi seems faulty: tattva-jñānam pramanam te yugapat sarva-bhāsanam
krama-bhavi ca yaj-jñānam syad-vada-naya-samskṛtam ||]
However, what Samantabhadra really expresses is actually the contents, significantly elaborated, of Umäsvamin's statement in TS 1.6: pramana-nayair adhigamah. - 'the comprehension of these [categories representing reality, mentioned in TS.1.4] is [accomplished] through cognitive criteria and [conditionally valid] viewpoints'. In the context of Tattvartha-sutra, it is clear that adhigama refers to the categories that have been discussed two sutras earlier (TS 1.4: jiva-ajiva-asrava-bandha-samvara-nirjarāmokṣās tattvam). Thereupon Umāsvamin speaks of 'their descriptive standpoints from four various perspectives (TS 1.5: nama-sthāpanā-dravya-bhāvatas tannyasah), and clearly the expression tan-nyasah is to be understood as tattva-nyasaḥ. Analogously, adhigamah in TS 1.6 means tattvâdhigamaḥ, which is confirmed by the Bhāṣya: eṣam ca jivâdinām tattvänäṁ yathôddiṣṭānām nāmâdibhir nyastānām pramana-nayair vistarâdhigamo bhavati. - 'And also a detailed comprehension of these reals such as living elements etc., as they have been enunciated [in TS 1.4], which are described from various standpoints such as appellation, etc. (TS 1.5), is [accomplished] through cognitive criteria and [conditionally valid] viewpoints. It is only much later, when adhigamah in TS 1.6 was connected not only with the reals mentioned in TS 1.4, and thereby with ontological issues, but also with the very first statement of Tattvârthasutra (TS 1.1: samyag-darśana-jñāna-cāritrāni mokṣa-mārgaḥ. - 'Correct knowledge, [correct] conation and [correct] conduct [constitute] the path to liberation."). It was done by Akalanka, much after Samantabhadra, in his commentary Rāja-vārttika (RVār 1.6: pramane ca nayaś ca pramana-nayaḥ, tair adhigamo bhavati samyag-darśanâdinām jivâdinām. - There are two cognitive criteria and [many] [conditionally valid] viewpoints, hence [in the sutra]: "cognitive criteria and [conditionally valid] viewpoints". Through these is [accomplished] the comprehension of [the three constituents of the path to liberation, mentioned in TS 1.1: correct] conation, etc., and of [the reals, mentioned in TS 1.4]: living elements, etc.' Samantabhadra's statement tattva-jñānam is, therefore, the repetition of the Umäsvamin's idea of tattvâdhigama, and the remainder of the verse AMi 101, especially syad-vada-naya-samskṛtam, paraphrases Umāsvamin's pramana-nayair of TS 1.6. For the above reasons, it is highly implausible to interpret AMi 101a (tattvajñānam pramānam) as a genuine definition of cognitive criterion.
Another problem related to the issue of either Siddhasena Mahamati's and Samantabhadra's priority in the context of the very first Jaina definition of pramāņa is their real temporal dependence: there is no doubt that Samantabhadra lived after Dinnaga, by whom he was influenced, but it is still not clear whether Samantabhadra was prior or posterior to Dharmakirti, who without any doubt flourished before Siddhasena Mahämati (see: Balcerowicz (2001b)). Certain passages of AMi seem to betray Dharmakirti's stamp, but that requires further research.

Page #18

360
PIOTR BALCEROWICZ
The cognitive criterion is the cognition revealing itself and something else [different from it, and it is free from subversion; [it is] two-fold: perception as well as indirect cognition, corresponding to the way of] determination of the cognoscible.32
He maintains the traditional principle that perception is a direct cognition that grasps an object not-indirectly' (aparoksatayā).Sub sequent verses imply, however, that he distinguishes two kinds of perception: conventional, sense-dependent perception and absolute, sense-independent perception (kevala; NA 27), probably contingent on what we take to be the perceiving organ' (akşa). 34 For Siddhasena the category of indirect cognition comprises all remaining kinds of cognition - especially verbal cognition (sābda, NA 8,9) and inference (NA 10), with its two subtypes ‘for oneself and 'for others' (svârthapratyakșa and parārtha-pratyakşa, NA 11) - that can all be grouped together as a separate kind of cognition based on verbal symbols, in a manner very much similar to what Dinnāga did. This new classification of cognitive criteria can be represented by the following diagram (on page 361, Model 2).
What is especially important, Siddhasena Mahāmati plainly indicates that the source of the two-fold division of cognitive criteria is actually not - as Dinnāga wanted it - the existence of two different aspects of reality the way we can cognise it, but two different modes of apprehending reality, or two different epistemic procedures. 35
32 NA 1: pramānam sva-parâbhāsi jñānam, badha-vivarjitam / pratyakşam ca parokşam ca dvidhā, meya-viniscayāt //
33 NA 4: aparokşatayârthasya grāhakam jñānam idrśam / pratyakşam, itaraj jñeyam parokşam grahanêkṣayā //
34 Compare Siddharși-ganin's explanations in NAV 1.7: pratyakşam cêty-ādi; tatra siddhanta-prasiddha-pāramārthika-pratyaksāpekṣayakşa-sabdo jiva-paryāyatayā prasiddhaḥ iha tu vyāvahārika-pratyakşa-prastāvād akşa-dhvanir indriya-vacano grhyate. tataś câkşam pratigatam pratyakşam. yad indriyam āśrityôjjihite 'rtha-sākṣāt-kāri jñānaṁ tat pratyakşam ity arthah. - 'As regards the term “perception (pratyakşa)", [it can be etymologically explained as "directed towards (pratio) the perceiving organ (°akşa)"'). "Concerning that, the linguistic unit “perceiving organ" - with regard to ultimately real perception, well-known from the Canon - is well-known as a synonym of "the living element". Here [in this verse], however, the linguistic unit "perceiving organ" is used - since [we) are dealing (here) with conventional perception - as an utterance (denoting) the senses. And, therefore, that which has gone towards (sc. pertains) “the perceiving organ" is perception. The meaning is as follows: such a cognition perceiving directly an object which commences resorting to the senses is perception.'
35 NA Icd: pratyakşam ca parokşam ca dvidhä, meya-viniscayāt // - '[Cognitive criterion is) two-fold: perception as well as indirect cognition, corresponding to the way of determination of the cognoscible.'

Page #19

A DISPUTE BETWEEN DINNĀGA, DHARMAKĪRTI AND AKALANKA 361
Significantly, the general divisionary binary scheme, traditionally accepted, has been preserved by Siddhasena in NA, albeit nominally, for he has invested it with a new contents that was more consistent with current terminology of the day, and pratyakṣa began to mean *perception' the way it was mostly understood.
cognitive criteria (pramāņa)
indirect cognition (parokşa)
direct cognition (pratyakşa)
verbal cognition (sābda) -
inference (anumāna) -
conventional perception (sāṁvyavahārika-pratyakșa) absolute perception (kevala-pratyakşa)
other
inference for oneself (svârthânumāna) inference for others (parârthânumāna)
perception for oneself
(svârtha-pratyakşa) perception for others (parârtha-pratyakșa)
Model 2
Akalanka adopted the overall model proposed by Siddhasena Mahāmati and presented it at the outset of the Laghiyas-traya: Perception is clear cognition, (divided) into primary and conventional, [whereas) indirect cognition (comprises all] remaining (types of] cognition. Thus, by way of summary, there are two cognitive criteria.36
(Foonote 35 Continued). It should be noted that Siddhasena speaks of two kinds of "[ways of] determination of the cognoscible (me ya-viniscayāt), or two procedures of cognising, and not merely of two kinds of the cognoscible' (* me yāt = 'corresponding to the cognoscible'). That is a different approach from Dinnäga's who as a justification for the number of cognitive criteria mentions two cognisable aspects of reality, not two different ways of cognizing these two aspects, see PSV2 1.2: dve eva yasmād laksana-dvayaṁ prameyam --...there are only two (cognitive criteria), because the cognoscible object has two characteristics.
36 LT 3: pratyakşam visadam jñānam mukhya-samvyavahäratah/ parokşam sesa-vijñānam pramane iti sangrahah //

Page #20

362
PIOTR BALCEROWICZ
He expressed the same idea also on other occasions, for instance in the Nyaya-viniścaya:
Perception is correctly [taken to be] lucid [cognition]; the other [type of cognitive criterion] is testimonial cognition (= indirect cognition) which is free of misapprehension, divided into [such varieties as] recognition etc. By way of summary, there are two cognitive criteria,3
but also in the Pramana-sangraha:
Perception is clear cognition, [and it is] divided in three, [whereas] testimonial cognition is free of misapprehension, [and it is] indirect cognition [which includes] recognition etc. Thus, by way of summary, there are two cognitive criteria. 38
Akalanka was seriously concerned, however, with effective criteria that would enable us to differentiate between direct (pratyakṣa) and indirect (paroksa) cognitions, but at the same their cognitively valid nature would be preserved. The proper criterion of differentiation involved two issues: non-erroneousness of all perceptions as valid pramānas and mediacy or lack of mediacy as a defining factor.
Firstly, he had to face the issue of impeded cognitions, e.g., affected by ophthalmological disorder (taimira), and the question why these are not reckoned among valid perceptions. In the commentary on the verse PV, 2.288, in which Dharmakirti elaborates on Dinnaga's typology of fallacious cases of perception, Prajñākaragupta refers to a certain irksome objection in his Pramanavārttikâlankara, quoting Dinnaga's verse:
What is the purpose to mention cognition affected by ophthalmological disorder? Some say that also cognition affected by ophthalmological disorder is conceptual. [Therefore,] in order to exclude such [a view], cognition affected by ophthalmological disorder is mentioned. ... If [cognition affected by ophthalmological disorder] is non-conceptual [cognition], then it would be perception [proper]. [But] not all [cognition that is] free from conceptual construction is perception [proper]; on the contrary, [it is perception] when it is [also] non-erroneous. [Only such a cognition which is both] non-erroneous and free from conceptual construction is perception, not any kind [of cognition].39
37 NVi, 469 NVi2 3.83 (p. 359):
pratyakşam anjasă spaṣṭam anyac chrutam aviplavam / prakirnam pratyabhijñâdau pramane iti sangrahah ||
38 PSa 1.2 (p. 97.3-4):
pratyakşam visada-jñānam tridhā śrutam aviplavam paroksam pratyabhijñâdi pramane iti sangrahaḥ ||
39 PVA, p. 334.23-27: sataimirika-grahanam kim-artham. taimiram api savikalpakam iti kaścit. tad-vyavartanârtham tad-vacanam. ...yady [taimiram] avikalpakam kalpanapoḍhatvāt pratyakṣam prāptam. na sarvaṁ kalpanâpodhaṁ pratyakşam, api tv abhrantatve sati. abhrāntam kalpanapoḍham pratyakṣam na sarvam.

Page #21

A DISPUTE BETWEEN DINNĀGA, DHARMAKĪRTI AND AKALANKA 363
To place the above objection and Prajñākaragupta's reply, we should first recall that Dharmakirti in PV, 2.28840 distinguished four kinds of fallacious perception (either following or improving on Dinnāga's varieties, see n. 28) and grouped them according to their character. First three kinds of fallacious perception are conceptual (since they contaminate the perceived data with memory images or impose conceptualised relations on it), and therefore cannot be classified as genuine perception, which is by nature non-conceptual. However, one variety, namely cognition affected by ophthalmological disorder (taimira), turns out to be non-conceptual: it is fallacious because it is erroneous, inasmuch as 'it originates through the impediment in (sensory] substratum (sense-organ).' To say, however, that cognition affected by ophthalmological disorder (taimira) is, nevertheless, perception proper would seem to be in agreement with Dinnāga's definition of perception, which only mentions its non-conceptual nature as the proper criterion. In other words, the definition of Pramāna-samuccaya would allow the following concomitance: 'every cognition free from conceptual construction is perception proper'. That would yield the valid conclusion: 'cognition affected by ophthalmological disorder (taimira) is also free from conceptual construction, hence it is perception proper.'
That seemed improper not only do Dharmakirti, who improved on the definition by explicitly adding the qualifier'non-erroneous (abhrānta, NB 1.4), but also to Jaina thinkers, including Akalanka. The first definiens 'non-conceptual (both in Dinnāga and Dharmakirti) was inadmissible to the Jainas, for it would exclude a number of perceptual processes (vide infra). Also the second definiens 'non-erroneous' in Dharmakirti's definition appeared to the Jainas completely inconsistent with the notion of validity which the definition of all cognitive criteria presupposes. The way Dharmakirti defined it, 'non-erroneousness' (abhrāntatva) was limited to pratyakşa only, and did not extend to anumāna. The first to criticise Dharmakirti's seeming amelioration was Siddhasena Mahāmati who contended: “This finference is non-erroneous because it is a cognitive criterion, just like perception.°41 For the Jainas, 'non-erroneousness' could safely be adapted to qualify pratyakşa, however, it could not be taken as its definiens which should distinguish it from indirect
40 PV, 2.288 trividham kalpana-jñānam aśrayopaplavôdbhavam / avikalpam ekaṁ ca pratyakşabham caturvidham //
4 NA 5cd: tad ( = anumanaṁ) abhräntam pramănatvāt samaksavat.

Page #22

364
PIOTR BALCEROWICZ
cognitions, for non-erroneousness' was a defining mark of all pramānas. Thus, abhrānta could not be employed in the definition of perception. 42
That is why Akalanka searched for another criterion - apart from the idea of direct/indirect modes of cognising, proposed by Siddhasena Mahāmati – that would consistently distinguish direct cognition (pratyakșa) from indirect cognition (parokşa). As a matter of fact, Akalanka's intention was to replace Siddhasena's criterion of 'sthe way of] determination of the cognoscible' (meya-viniscaya), which both tended to be a bit vague and yielded some more difficulties that surfaced the moment Siddhasena Mahāmati restructured the overall model of valid cognitive procedures.
42 Interestingly enough, there seems to be some influence of Dharmakirti's claim that inference is erroneous on the way Akalanka understood the relationship between absolute non-erroneousness and cognitive validity (prāmānya). Akalanka occasionally admits that no valid cognition is absolutely correct, and even false cognition possess limited validity. See for instance A$S3 (ad AMi 101, p. 88.4-8), especially the comment on inference in the final sentence: buddher anekāntāt yena ākārena tattva-paricchedah tad-apeksayā prāmānyam. tena pratyakşa-tad-ābhāsayor api prāyaśaḥ sankirņa-prāmānyétara-sthitih unnetavyā, prasiddhanupahatêndriyadrster api candrarkâdişu deśa-pratyāsatty-ādy-abhūtākārâvabhāsanāt, tathā upahatakşâder api samkhyâdivisavāde'pi candradi-svabhāva-tatt vôpalambhật. tat-prakarşâpekṣayā vyapadeśavyavasthā gandha-dravyâdivat. tatha anumānâder 'api kathaṁcit mithya-pratibhāse 'pi tattva-pratipattyáiva prāmānyam. - 'Since cognitive awareness is multiplex, cognitive validity concerns [only] that (aspect of cognition) by means of which determination of factuality (or: categories) [is accomplished). Accordingly, it should be admitted as established that, as a rule, cognitive validity and its opposite of both perception and its fallacy respectively, are commingled, [i.e., in valid acts of cognition there is an element of illusion, and in fallacious acts of cognition there is some valid component]. [It is so), (1) because even the act of seeing - with a sense organ of which (we) well know it is not disabled - of the moon, the sun etc., manifests (their) unreal forms such as (their) proximity in space etc., [and also] (2) because, similarly, also in case the eye which is disabled (and we see a double moon etc.,), even though there is fallibility (unreliability) as regards the number (of the moons) etc., there is (correct] apprehension of the factuality of the essential nature of the moon etc. The distinction in designating [them as cognitive criterion and its fallacy) depends on the degree, like in the case of] fragrant substance (which is both a substance and a fragrant quality). Similarly, also inference etc., - even though in a particular respect [it generates) false representation (of the cognoscible object] - has cognitive validity, inasmuch as it does indeed [lead to the correct] comprehension of factuality. This idea was expressed by Akalanka on other occasions as well, e.g., LTV 22 and SVi 1.15 (and SViV ad loc., p. 73): ātma-samvedanam bhrānter abhrāntam bhāti bhedivat/ pratyakşam taimiram cândram kim nânekānta-vidvisām // ... nâpi tat sarvathā abhrāntam eva svayam advayasyâpi dvaya-nirbhāsa-pratiteh .. His reasons for the claim, in short, were dictated by his strictly realist stand: no piece of our cognition is absolutely false, because all the contents of our cognition is a representation of the external world, and thus it must correspond to something real. Ultimate components of cognition (presentations of individual events) are true, only the relations between them that are imposed by conceptual activity of the mind can be false.

Page #23

A DISPUTE BETWEEN DINNĀGA, DHARMAKĪRTI AND AKALANKA 365
The second problematic issue turned out to be mediacy or lack of mediacy as the feature that could help distinguish between direct and indirect cognitions, respectively. To maintain, in harmony with older Jaina tradition, that pratyaksa and paroksa are differentiated on the basis of their either immediate or mediated mode of acquisition, while adopting Siddhasena's new scheme, would ultimately lead to the conclusion that such divergent cognitive acts as sensation or inference are merely two different aspects or varieties of one and the same kind of cognition, insofar as both of them are mediated: either by sense organs in the case of sensation, or by the mind, logical reasons, concepts or words in the case of inference. In practical enterprise, in all empirical spheres of life, every individual has to rely on some agency that mediates cognitive processes, be it perception or inference: on sense organs, the mind, verbal expressions, symbols etc. Preserving the criterion of mediacy, the borders between direct and indirect cognitions would be blurred and the whole edifice would collapse.
The outcome of fitting the new contents, especially the new understanding of what perception is (in consonance with current Indian views), into the old Jaina model was that the only ultimately direct cognitions would be: absolute knowledge (kevala), tantamount to omniscience (sarva-jñāna), or two less perfect forms of such supernatural cognitive powers: clairvoyance (avadhi-jñāna) and mind-reading (manah-paryāya-jñāna). All of these three played an important role in soteriology, serving as vital elements in a range of proofs for the existence of soul (jiva) and final liberation (mokşa),43 being contemplated as paramount ideal, rending legitimacy to the authority of the scriptures and being believed in. Yet, these were inaccessible to any contemporaneous mortal, with the exception of ancient personages occurring in myths and hagiographies. Consequently, one would have to acknowledge that an average person has at his or her disposal no direct means of knowing. That would, with certainly, provoke justified and hardly refutable criticism from all other philosophical schools in India, namely, that the Jainas have to entirely rely on unsubstantiated faith in their doctrines, without any prospect to verify them, at least in some part, empirically. Moreover, the immediate conclusion would be that all knowledge of the world a person may have is obtained in an indirect, mediated manner, hence must be by nature liable to doubt and censure. To admit that all
43 A number of them are formulated, e.g. by Kundakunda in the Pravacana-sära and Samaya-sāra, a most of them summarised by Hemacandra-sūri in the Pramānamimimsā.

Page #24

366
PIOTR BALCEROWICZ
judgements we have are derived from second-hand knowledge would, at best, not only lead to scepticism and undermine the rational and realistic outlook Jaina thinkers intended to cultivate, but also subvert doctrinal framework of Jainism. No testimonial statement of Jaina Canonical scriptures and oral tradition could be amenable to any kind of direct verification, not even simple descriptions of worldly matters contained in scriptures that were used – through the reasoning based on analogy and empirical exemplification (drstānta) - as a basis for supplementary corroboration of extra-mundane and metaphysical tenets (following the principle: 'if the Canon accurately describes mundane affairs, its description of spiritual matters must also be accurate'). If any judgement, including all the assertions of the Agamas, were beyond any first-hand validation, and could only be proved indirectly, which would necessitate further, indirect as it were, proof, infinite regress (anavasthā) would result. Accordingly, one would have to concede that doctrinal and epistemological principles of Jainism are faulty, inasmuch as they lead to logical inconsistencies and defects such as infinite regress and ceaseless search for new warrants to sanction previous conclusions
In the worst case, retaining a single criterion of mediacy that distinguishes between direct and indirect cognitions and concluding that in all empirical dealings we rely solely on indirect cognitions would result in rejecting the reality of the world we perceive, and in assuming its ultimately illusory character.
Besides, it is a tautology to say that a direct cognition 'grasps an object not-indirectly' (aparoksatayā), and therefore is bereft of any substantial contents.
All these considerations may have borne on Akalanka's following reflection expressed in the Nyāya-viniscaya:
Two first two [kinds of cognition, viz. sensuous cognition (mati) and testimonial cognition (śruta),] were rightly termed [by Umāsvāmin in TS 1.11] indirect cognition, whereas the remaining [kinds, viz. clairvoyance (avadhi), mind-reading (manahparyāya) and absolute knowledge, were termed) direct cognition. However, it is only in order to be in agreement with people's opinion, that sensuous cognition is incorporated into the definition of (perception].44
44 NVi, 474cd-475ab = NVi, 3.88 (p. 363): ādye parokşam aparam pratyakşam prāhur āñjasam / kevala loka-buddhyiva mater laksana-sangrahah // Akalanka's approach became widespread in Jainism, compare for instance identical arguments of Abhayadeva in his commentary (TBV) on STP 2.1, p. 595.24-25, where he states that mati-jñāna is ultimately paroksa (as it was described in TS), but according to practical convention it is pratyakşa.

Page #25

A DISPUTE BETWEEN DINNĀGA, DHARMAKIRTI AND AKALANKA 367
Clearly, these were also pragmatic considerations that made Akalanka comply with general usage of the term “perception', even though he himself occasionally endorsed the idea that the ultimate *perceiving organ’ is the soul (ātman): [Ultimate) perception, (which is acquired) independently of sense organs and the mind and which is] free of deviation, is the seeing of definite contents.... [In its case, there takes place) the elimination of dependence on other [factors) in the sense: “Iperception is confined to the perceiving organ (sc. the
il)" .45 And it is the soul without which no cognition could be possible: 'If there were no soul, production, etc. of (any) cognition would be impossible, because there would be no agent.:46
There was no controversy that there is an essential difference between perceptual, first-experience or immediate grasp of sensual data that are currently present, on the one hand, and indirect processes of inference or employment of verbal symbols, on the other, which are operative independently of whether the things we infer or speak about are present or not. All philosophical systems tried to accentuate and explain this difference this way or another, by using such terms as “perception', 'direct cognition', 'sensuous cognition', or 'inference', 'indirect cognition', 'analogy', 'presumption', etc. in a sense that differed from traditional Jaina usage.
What made this essential difference for Dinnāga and Dharmakirti was the idea of conceptualisation, which characterised all non-perceptual kinds of cognition, and which connoted two main aspects: (1) the use of a system of verbally expressible symbols (either a natural language or a logical system or relations) and (2) the application of such symbols only to general properties ('universals'), not to unique qualities ('individuals').
A detailed analysis of complex cognitive processes which Akalanka introduced in the contentious verse LT 10cd-11ab will reveal that for him such a contention was not only a simplification, but also an inaccurate exposition. In order to improve on Dinnāga's and Dharmakirtis definitions as well as on Siddhasena Mahāmati's model, Akalanka introduces a new notion of clarity (vaišadya), or lucidity (spastatva), for instance in the verses of NVi, 3.83 and PSa 1.2 quoted above (p. 362, and nn. 37, 38).
49 RVār 1.12 (p. 53.4,11): indriyânidriyânapeksam atita-vyabhicaram sākāragrahanam pratyakşam. ... akşam prati niyatam iti parâpek şânivrttih.
46 RVär 1.12 (p. 45.8): ātmâbhāve jñānasya karanâditvânupapattih kartur abhāvāt.

Page #26

368
PIOTR BALCEROWICZ
For Akalanka, clarity consists in the way reality is presented in cognition:
In contrast to inference etc., the clarity of [this perception), in the case of discernment (becoming aware of the object), is recognised to be the representing of the particular. What is different from this is non-clarity (in the case of discernment].47
As a matter of fact, in the above manner Akalanka modifies and disambiguates Siddhasena Mahāmati's obscure and tautological definiens: the way of] determination of the cognoscible (meya-viniścaya) of NA 1 (p. 360, and n. 32). Each act of perception - either (1) conventional and sensory, (2) internal and mental, or (3) absolute and extra-sensory - grasps the particular thing, or individual aspect, and this ‘unmediated access to the individual' is clarity, or lucidity. In other words, perception leads the cogniser to clearly discernible and uniquely identifiable entities that are amenable, at least in theory, to the cogniser's actions. Inference, verbal cognition etc., on the other hand, do not bring the cognitive agent to clearly distinguishable individual features of a thing. What is represented in such inferential or verbal cognitive acts are ideas or concepts that emerge as the contents of indirect cognition. Ideas themselves, classes that correspond to the ideas or words expressing the ideas cannot be acted upon individually. For this very reason such a cognition based on ideas and words is 'bereft of clarity (avisada, aspașta), inasmuch as it delivers merely a vague and imprecise notion of what actual individuals, amenable in practical actions, are denoted by words or are implied by inferences. Such an unclear (avisada) cognition merely extends over the whole range of individuals and refers to each of them in precisely the same manner, by pointing to their general common feature, and making them all indistinguishable from each other in our cognition. Thus, indirect cognition reveals no particular thing that is the designatum of a term or an actual source of the idea presented in the cognition. On the contrary, on its basis it is not at all possible to make any practically relevant distinction among singular items aggregated into a conceptual bundle'. And this is precisely what is involved in the notion of lack of clarity' (avaiếadya) or 'absence of lucidity' (aspastatva).
As a justification for the need to introduce another criterion for genuine perception, other than its conceptual character, Akalarka
47 LT 4: anumănâdy-atirekena višeșa-pratibhāsanam/ tad-vaiśadya matam buddher avaišadyam atah param //

Page #27

A DISPUTE BETWEEN DINNAGA, DHARMAKIRTI AND AKALANKA 369
cites the case of fallacious perception which may be by nature nonconceptual, and thus fulfils Dinnaga's criterion, but it remain nonetheless erroneous, inasmuch as it is unclear:
Therefore, fallacious perception, which is non-conceptual (indeterminate), is necessarily unclear.48
This is Akalanka's answer to the problem referred to above by Prajñākaragupta: how is it possible that an erroneous act of sensation, even though it is non-conceptual, is not genuine perception?
,49
As one might expect, the concept of clarity (vaiśadya) as the defining feature of perception subsequently underwent changes and reinterpretations, and was also taken up by the Śvetāmbaras. Of special note is the definition offered by Hemacandra-sūri in the Pramana-mimimsa: 'Clarity is either independence of other cognitive criteria or the distinct representation of the form: "this"." The first feature, independence of other cognitive criteria, invokes the idea that the perceptual cognition is undeniable and obvious in its selfexplanatory character, i.e., it requires no additional justification for the accuracy of its contents to be accepted. The second feature, viz. the distinct representation of the form: 'this', which is mentioned as an alternative definiens, is a direct extension of Akalanka's idea that clarity of perception consists in 'the representing of the particular' (viseṣa-pratibhasanam, LT 4).
Vadideva-suri (11th/12th c.) is another example of a Śvetambara thinker who borrowed Akalanka's definition verbatim: 'Lucidity is the manifesting of particular [features] in a greater degree than
inference etc.
,50
VI
A closer examination would now be required to see what is actually clear in a clear cognition? Is clarity simply the way an object is cognised, viz. certainty implied by clear cognition, or is clarity a feature of the object that is presented in the cognition? Unfortunately, on most occasions Akalanka is not very clear on this point.
48 LTV 23: tasmad aviśadam eva avikalpakaṁ pratyakṣâbham.
49 PMI 1.14: pramāṇântarânapekṣêdantaya pratibhāso vā vaiśadyam.
50 PNTAA 2.3: anumanâdy-adhikyena viseṣa-prakāśanaṁ spaṣṭatvam. = LT 4: anumanâdy-atirekena viseṣa-pratibhasanam | tad-vaiśadyam.

Page #28

370
When he says in his auto-commentary to the Laghiyas-traya (LTV 3): jñānasyaîva viśada-nirbhāsinaḥ pratyakṣatvam, itarasya parokṣatā, the statement can equally imply both possibilities: 'Only such a cognition which reveals [its object] clearly (or: as clear) is perception; the other one is indirect cognition.' The same ambiguity holds true in following passage of PSaV 1.2 (p. 97.5-6): pratyakṣaṁ viśada-jñānaṁ tattvajñānam visadam, indriya-pratyakṣaṁ anindriya-pratyakṣam atīndriyapratyakṣam tridhā. - 'Perception is a clear cognition (or: cognition of something clear), [for] the cognition of reals (doctrinal categories) [as a cognitive criterion51] is clear. It is divided in three: sensory perception, non-sensory (mental) perception and extra-sensory perception.' Other formulations where spasta as a synonym of visada occurs are likewise inconclusive: LTV1 61 (p. 21.6-7): indriyârtha-jñānaṁ spaṣṭam hitâhita-prāpti-parihāra-samartham prādesikam pratyakşam.
PIOTR BALCEROWICZ
'Perception which is a cognition of an object [amenable] to sense organ is lucid, capable of [triggering an action of] either appropriation of what is wholesome or avoidance of what is unwholesome, [and it is] spatially limited'; NVi, 157 (p. 50.22-23) = NVi2 160 (vol. 1, p. 524.21-22): akṣa-jñānânujam spaṣṭam tad-anantara-gocaram | pratyakṣaṁ mānasaṁ câha... - 'Mental perception is brought about as a consequence of a cognition [derived] from sense organs, it is lucid, its domain (contents) immediately [follows] this [cognition derived from sense organ].'
However, in a longer section of the Nyāya-viniścaya in which he discusses the possibility of omniscience as the culmination of a series of supernatural perceptions, Akalanka mentions the following idea:
A lucid discernment with respect to a distant 52 [object that] correctly [takes place] does not stand in contradiction, because it depends on the loosening
51
I take it as a reference to Samantabhadra's characterisation of the nature of pramāna in AMi 101a: tattva-jñānam pramānam (see n. 31). In this way, Akalanka expresses the idea that perception (pratyakṣa) is such a cognitive criterion (tattvajñāna stands for pramāna) that is clear (visada).
52 A 'distant' object may connote not only distance in space, but also in time or 'nature' (in the sense of: 'not perceptible with ordinary sense organs'). See for instance: NSā, p. 170: yogi-pratyakṣam tu deśa-kala-svabhava-viprakṛṣṭartha-grahakam ('Supernatural perception is the grasping of object distant in place, time and own nature'), and NBhu, p. 170: deśa-viprakṛṣṭaḥ satya-lokâdayo 'tidurasthā vyavahitāś ca naga-bhuvanâdayaḥ, kala-viprakṛṣṭās tv atitânāgataḥ, svabhāva-viprakṛṣṭāḥ paramāṇy-ākāśâdaya iti ('Distant in place [means] places far away like existent worlds etc., concealed (from sight) [means] hellish abodes etc., distant in time [means] past etc., distant in terms of one's own nature [means] atoms, ether etc.'): These three kinds of 'distance' correspond to Sabarasvamin's four separate kinds of objects, viz. (1) past and future (bhūta, bhaviṣyant), (2) sublte (sūkṣma), (3) concealed from sight

Page #29

A DISPUTE BETWEEN DINNĀGA, DHARMAKĪRTI AND AKALANKA 371
of karmic veils obstructing cognition in a dream, or (cognition) of a fortuneteller etc..
In the verse we can easily distinguish two separate elements: clarity (spasta), which belongs to cognition (vijñāna), and an object of cognition which is spoken of as being distant' (viprakçșța), not as being "clear’. That would imply that clarity or lucidity primarily qualifies the cognition as such, i.e., the cognitive procedures. We come across a similar indication in Pramāna-sangraha:
Cognition (derived) from sense organs is multiplex, inasmuch as another form (sc. one-sided form) is not established, (and it is) lucid, because the object is proximate. 94
Also here lucidity qualifies cognition separately from its object, of which Akalanka merely says that it is proximate and its proximity warrants clear and lucid perception. Therefore, on the basis of two latter quotations, it seems that Akalanka considered clarity (vaišadya) to belong to the way of cognising (to qualify the cognitive process), not necessarily to the object presented in the cognition. Vādideva-sūri in Pramāņa-naya-tattvâlokālankāra seems to support the interpretation that clarity of perception refers to 'clear mode or presenting the object', rather than the 'clear object itself, for he says:
(Foonote 52 Continued). (vyavahita) and (4) distant (viprakrsta), with kind 3 and 4 subsumed under one heading of deśa-viprakrsta, see SBh 1.2.2 (p. 4.7-9): codanā hi bhūtam bhavantam bhavisyantam sūkşma vyavahitam viprakrstam ity evam-jātiyakam artham saknoty avagamayitum, nânyat kiñcanêndriyam ("The injunction enables (people) to know anything of the following sort: past, present, future, subtle, concealed (from sight) and distant'). Frequently, however, the term viprakrsta does not indicate the temporal span, but is restricted to distance in time', as for instance in: VSV ad VS 1.16 (viyukta-pratyak şa): sūkşma-vyavahita-viprakrsteșu artheșu teşām catustaya-sannikarşād api pratyakşam jāyate; PBh 22.12.2.a: [242] viyuktānām punaś catustayasannikarşād yogaja-dharmânugraha-sāmarthyāt sūksma-vyavahita-viprakrsteșu pratyakşam utpadyate; NMa 2 (I, p. 157.7): darśanasya ca paro 'tiśayah sūksma-vyavahita-viprakrsta-bhūta-bhavisyad-ādi-visayatvam; YS 3.25: pravrtty-āloka-nyāsāt sūksma-vyavahita-viprakrsta-jñānam. It seems, however, that Akalanka uses the term also in the sense of distant in time', since as the example he names a female fortuneteller (ik şanikā), glossed by Vădirāja-sūri (NViV) as grāma-dakini, i.e., someone who foretells the fate. Vādirāja-sūri further implies that such cognitions when one's mind is 'absorbed in meditative trance' (samāhita-citta) are also meant.
53 NVi, 407 (P. 84.13-14) = NVi2 21 (vol. 2, p. 291.1-2): vijñānam añjasā spastam viprakrste viruddhyate / na svapnêksanikâder vā jñanavrtti-vivekataḥ //
54 PSa 4ac (p. 97.16-17): akşa-jñānam anekāntam asiddher aparâkrteh / spastam sannikitarthatvāt...

Page #30

372
PIOTR BALCEROWICZ
'Lucidity [of perception] means the presentation of the particular sin consciousness) in greater degree than in the case of inference etc.)
On the other hand, it is the object as it enters the cognitive act that determines whether it is immediate and direct or whether it is mediated, based on conceptualisation and symbols. It would, however, be mistaken to suppose that the dependency between the object and cognition parallels Dinnāga's model, viz. that perception grasps uniquely individual feature whereas inference and other forms of indirect cognition comprehend a common property.
Prabhācandra appears to support another view, namely that 'clarity' belongs both to the object as it is presented in cognition and to the cognition itself, for he says in his commentary to LT 3: 'It is (such a cognition) which reveals sits object] clearly, i.e., it is lucid representation of the intrinsic nature of itself and the other (sc. external object), in so far as it is not oriented towards anything else [than itself] - it is perception, i.e., it is perceptional cognitive criterion. The other one, i.e., that which reveals [its object] not-clearly (or: which reveals an unclear object), is indirect cognition, i.e., indirect cognitive criterion.956 Interestingly, Prabhācandra's interpretation finds support in another statement of Akalanka:
Furthermore, there is no absolute difference in the datum of these two [kinds of] conceptual (determinate) cognitions: clear (perceptual) and its opposite (inferential), because acts of perception of proximate objects and opposite (sc. distant) objects can easily be demonstrated to have as their datum one and the same object.
The external object for direct and indirect cognitions remains the same entity. However, its representation, or its cognitive 'image', in respective cognitions is different. Therefore, it might seem that the apparent ambiguity in a range of above-quoted passaged (LTV 3, PSaV 1.2, LTV 61, NVi, 157 = NVi2 160) was deliberate and also Akalanka may have understood clarity' both ways: as belonging to the process and to the presented object.
55 PNTĀA 2.3: anumānâdy-ādhik yena visesa-prakāśanam spastatvam.
56 NKC, p. 67.12–14: visada-nirbhāsinaḥ para-mukhâprekşitayā sva-parasvarūpayoḥ spasta-pratibhāsasya pratyak satvam pratyakşa-pramāṇatā. itarasya avisada-nirbhāsinaḥ parokşatā parokşa-pramānatā. Bolded typeface reproduces the pratikas of LT.
37 LTV 23: na ca visadêtara-vikalpayoh visaya-bhedaikāntah pratyāsannêtarârthapratyakşāņām ekârtha-vişayatôtpatteh.

Page #31

A DISPUTE BETWEEN DINNAGA, DHARMAKĪRTI AND AKALANKA 373
The supposition that Akalanka understood clarity or lucidity of perception in a two-fold manner, as belonging both to the cognitive process and to the cognitive contents (the object), may find, at first glance, some more reinforcement in a passage of the Nyaya-viniscaya 1.3:
It has been correctly stated that the defining characteristic of perception is [that it is] clear and [that it has] definite contents, [and that it is] the percipience of the essence of an object [consisting altogether in] substance, modes, the universal character and the particular character. 58
Accordingly, the object of perception is constituted by an aggregate of a number of features, such as substantial (persistent in time), modal (momentarily changing), general and individual aspects, that all belong to one and the same substratum, and the contents of perception - that is the representation of the extrinsic object (extrinsic in the sense that it is independent of the given cognitive act, be it an external physical thing or an internal cognition that is being reflected upon) - is definite and not indistinct (lit. 'with the form'). In the above passage the clarity of the representation of the object in cognition may at first appear to be implied by the qualifier sākāra, which could hardly apply to the clarity of the cognitive process. What would, however, be problematic with such an interpretation is that the property of 'having definite contents' (sākāratva) is introduced by Akalanka as one of two defining features of perception, next to 'lucidity' (spaṣṭatva). If the idea of definite contents of cognition, viz. the clarity that belongs to the object in cognition, were already implied by the term spasta, it would be redundant to mention it separately in the verse.
Furthermore, on another occasion Akalanka suggests that it is possible to conceive of perception (direct cognition) of an object which itself is unclear, or not clearly manifest:
When it comes to the union of the perceiving organ and the object, [first] the seeing of [mere] existence [takes place, and then] a comprehension of the notion (conceptual impression) in the form of the object - [these constitute] sensation..."
59
The passage describes the first phase of perception, viz. sensation (avagraha). Following earlier Jaina tradition which distinguished two kinds of sensation: sensation of momentary manifestation (vyañjanâvagraha), which is non-conceptual, and sensation of the
58 NVi 1.3 (p. 57):
pratyakṣa-lakṣaṇam prahuḥ spaṣṭaṁ sākāram añjasā dravya-paryaya-samanya-viseşârthâtma-vedanam ||
59 LT 5ac: akşâ rtha-yoge sattâloko 'rthâkāra-vikalpa-dhiḥ | avagraho...

Page #32

374
PIOTR BALCEROWICZ
object (arthâvagraha), which is conceptual," Akalanka speaks of the first moment of sensation in which one merely becomes aware of the fact that something is there (sattâloka), and thereupon there emerges a conceptual image the contents of which consists in a definite form of an object (arthâkāra-vikalpa-dhi). Both these stages of sensation are an integral part of perception, hence both must be clear (visada), or lucid (spasta). It would be difficult to conceive of an object of the first sensation of the vyañjanâvagraha kind which is both clear and nonmanifest, of which one only knows that it is present, insofar as it has stimulated a sense organ. This would imply that clarity of perception qualifies rather the process, not the contents or the object that enters cognition.
In another work Akalanka refers to an idea of an illusory image of a firebrand circle which, despite being false, inasmuch it consists of a series of momentary perceptions, is nevertheless perceived as a clear object:
Since one sees such a false appearance of a firebrand circle etc. (as one unity), which is lucid, it is not the case that one perceives individual acts of apprehension of the unique particular as singular and momentary
60 See TS 1.17-18: arthasya. vyañjanasyavagrahah. The sensation of the object (arthâvagraha) is possible with all sense organs, whereas the sensation of momentary, manifestation (vyañjanâvagraha) concerns all sensory data with the exclusion of vision, see RVär 1.18: vyañjanam avyaktam sabdâdi-jātaṁ tasyâvagraho bhavati. -. ‘Momentary manifestation (means) not manifest [data] generated by sound etc.; the sensation of such (momentary manifestation) occurs.' Akalanka subscribes to this division of vyañjanâvagraha and arthâvagraha, inasmuch as respective parts of Pūjyapada Devanandin's Sarvârtha-siddhi that clearly state this idea are incorporated into Akalanka's Tattvârtha-vārttika: SSi 1.17 (p. 115.4); cakşur-ādi-visayo 'rthah ( = p. 65.28) and SSi 1.18: vyanjanam avyak tam śabdâdi-jātam tasyâvagraho bhavati ( = RVār, p. 66.27-28). The nature of this stage is also described by Akalanka in his Asta-sati, a commentary on Samantabhadra's Apta-mimāṁsā 101, i.e., A$S3 101 (A$S3, p. 88.22–23 = A$S3, p. 1-2)): dhvaner akhandaśaḥ śravanād adhigamo 'pi prāthama-kalpikah tattva-nirņitir eva. - 'Even the very first comprehension (which takes place)" on hearing a sound as a whole (sc. as such) is nothing but a (cognitively valid) ascertainment of the factuality (of the sound).' The fact that he speaks of acoustic sensation, instead of ocular perception (which is the most usual reference when he speaks of perception), is significant, for it is acoustic perception in the case of which we can have sensation of momentary manifestation (vyañjanâvagraha), not possible in the case of ocular sensation. a There can be no doubt that what is implied by the expression tattva-nirniti is cognitive validity, for tattva-nirniti corresponds to tattva-jñānam which is said in AMi 101a to be the defining characteristic of pramāna.
01 Svi 11.27 (p. 733): spastasyâlāta-cakrâder vibhramasyâpi vikșanāt / sva-laksanôpalabdhinam naika-kşanika-darśanam //

Page #33

A DISPUTE BETWEEN DINNAGA, DHARMAKIRTI AND AKALANKA 375
In Akalanka's opinion, we can speak of acts of cognition whose contents is a clear image, viz. whose object is characterised by clarity. However, these are not necessarily true, hence cannot be considered genuine kinds of perception. In other words, it is possible that an object that emerges in cognition possesses clarity (spaṣṭatva), nevertheless it remains an illusion. This would additionally support my suggestion that clarity is for Akalanka primarily a defining feature of the cognitive process, not of its object.
The idea of lucidity or clarity, viz. of the way the object is presented to/in cognition, which is for all practical reasons tantamount to certainty, reminds the two principles, viz. clarity and distinctness (clairement et distinctement), employed by Descartes as criteria for validity of our cognitions. For Descartes, they guarantee that what we apprehend correspond to reality, and thus provide certainty to cognitions. For Akalanka the indubitability warranted by lucidity of cognition was a hallmark of perception and attested to the veracity and reliability of it: perceptual act necessarily furnished one with certainty that what was presented in cognition was at the same time out there: 'here and now'.
VII
The verse of the Nyāya-viniścaya 1.3 (quoted on p. 373, and n. 58) reveals still another important feature of the object of perception. Its object is not merely a unique particular, absolutely devoid of any general character, whose essence can neither be conceptually constructed nor verbally expressed, as Dinnaga understood it. On the contrary, for Akalanka the object of perception is essentially endowed with both particular and universal features, as well as with modal (momentary) and substantial (temporarily persistent) aspects. That is the way an object emerges in a clear manner in cognition as definite data of perception. Thus, the object of perception is a really existing thing which already at the perceptual stage is apprehended as a bearer of multiple features. Its individual, unique aspects are only directly perceived, whereas its common, repeatable characteristics can be comprehended both by perception and indirect cognition alike. This finds confirmation in another passage of the Nyāya-viniścaya:
[129] A really existent thing, which is individually marked (sc. unique particular), can be different (unique), in so far as it is represented as different (unique), [and it is also] endowed with conceptual character; it can [also] be non-different (sc. having uni

Page #34

376
PIOTR BALCEROWICZ
versal aspects), in so far as it is represented [in cognition] as non-different (universal). [130] [The same unique individual], which is distinguished from things that are similar to it [and] is endowed with singular [features], is something that is directly perceptible as well as indirectly cognisable, both as external and internal [data], 62 depending on its own aspect."
63
Thus, it is not the case that our cognitive apparatus, at all stages of cognition, takes an active part and contaminates perceived images with some preconceived notions and ideas, by projecting the contents of memory onto new presentations and by bequeathing some conceptual framework onto phenomenal world, as Dharmakirti understood it. On the contrary, things in themselves are endowed with a complex character, possessed of individual and general features and it is the role of our cognitive apparatus to recognise these features.
Furthermore, it can hardly ever happen, as Akalanka argues in the Siddhi-viniścaya-vivrti, that the object in cognition is apprehended as something absolutely unique, detached from its universal features: 'we do not perceive cognitive episodes within awareness that concern unique particulars and universal things that are absolutely separate from each other... In exactly such a way, [viz. by grasping things as many/much, manifold, etc., we arrive at] a proof [that], because of multiplexity [of appearance], on the level of the ultimate truth [unique particulars and universal features coexist in cognition].'
964
That is why the dichotomy of two different modes of cognising the world (directly or indirectly) is not followed by the dichotomy of the world itself. Cognising the world with the help of direct or indirect cognition should not mean that we live in two unrelated worlds and deal with two unconnected sets of entities: with unique individuals,
62 The external contents of perception (bahis-pratyakṣa) are for instance physical, gross objects, whereas the internal contents of perception (antaḥ-pratyakṣa) are individually apprehended thoughts, contents of introspection, etc., that are immediately triggered by individual events. The external contents of indirect cognition (bahisparoksa) are likewise gross objects, etc., which, however, stand for a whole class of which they are merely individual elements, being objects of inference, contents of concepts, denotations of words, etc. The internal contents of indirect cognition (antaḥparoksa) are general concepts, images, universal contents of symbols etc.
63 NVi 1.129-130 (p. 484):
sad bhinna-pratibhäsena syad bhinnam savikalpakam | abhinna-pratibhäsena syad abhinnam sva-lakṣanam || samanârtha-parāvṛttam asamāna-samanvitam pratyakşam bahir antaś ca parokṣam sva-pradeśataḥ ||
64 SVIV
1.27 (p. 115.12-14): sva-lakṣana-sāmānya-lakṣanaikante punah samvedanâkṛtiḥ na paśyamaḥ... tad evam paramarthataḥ siddhiḥ anekantāt. For the complete text of the commentary see n. 95.

Page #35

A DISPUTE BETWEEN DINNAGA, DHARMAKĪRTI AND AKALANKA 377
which we grasp perceptually, and with concepts and universal features, which we comprehend by means of indirect cognitions. In fact, one and the same entity is a bearer of a bundle of attributes that are differently approached by our cognitive apparatus:
The unique particular is unadulterated (sc. singular), similar (sc. having features common with other members of the same class), endowed with conceptual character, capable [of causally efficient action] by means of its own qualities which occur concurrently and consecutively, [and it is numerically] one.
65
67
Such individual entities, being irreducible and non-repeatable, are at the same time conveyors of common characteristics (samānalakṣaṇa) which make it possible to group individuals into classes, that correspond to concepts and words, on the basis of such similarities. Clearly, this echoes Dinnaga's claim that we name things on the basis of similarities inherent in them directly. A similar idea to Akalanka's, namely that one and the same cognoscible entity has two characteristics, is expressed already in Pramana-samuccaya 1.268 in the well-known passage quoted above. The difference between the two thinkers is, however, crucial. For Dinnaga these two aspects (unique and general) can never be apprehended together in one act of cognition; there is a strict cognitive gap between both aspects, which never intersect in cognition, each being a compartment for a respective pramāna. For Akalanka, on the other hand, both aspects appear already in perception, side by side: the entity is being perceived with its transient, momentary attributes as its 'concurrent' facet (saha-vivartin) as well as with its underlying substantial structure (krama-bhāvin) that is responsible for the fact that all the changing qualities can be correlated with each other and be identified as belonging to one and the same substratum. Such an bundle of features becomes effectively an object that is capable of being acted upon.
65 These are the two notions that later came to be known as synchronic homogeneity (tiryak-sāmānya) and diachronic homogeneity (urdhvatā-sāmānya), see: Balcerowicz (1999).
66 NVi 1.122 (p. 453):
sva-lakṣaṇam asankirnaṁ samānam savikalpam | samartham sva-gunair ekam saha-krama-vivartibhiḥ ||
67 See Herzberger (1986: 124-133), Hayes (1988: 196-204). 68 Quoted on p. 353, and n. 16.

Page #36

378
PIOTR BALCEROWICZ
VIII
A important terminological issue needs next to be clarified: what is the actual relationship between testimonial cognition (śruta) and indirect cognition (parokşa), of which Akalanka speaks? When we compare the four verses quoted before (see p. 362 and 366, and notes 36, 37, 38 and 44), in all of which Akalanka presents the rudimentary division of cognitive criteria and in which he uses śrutam and parokşam interchangeably, it becomes obvious that the term paroksa “indirect cognition') is treated as a synonym to śruta ("testimonial cognition'):
LT 3: NVi, 3.83: PSa 1.2: NVi2 3.88:
pratyakşam ... parokşam... pratyakşam...
śrutam pratyakşam... śrutam aparaṁ pratyakşam... adye parokşam
pramāne iti sangrahah //. pramāne iti sangrahah // pramāne iti sangrahah//
Further, in the commentary to the Pramāna-sangraha we read: "testimonial (mediated) cognition is free of misapprehension, [it is) occasioned by perception, inference and scriptural testimony.ro Thus, Akalanka distinguished three subvarieties of indirect cognition, depending on its source, viz. on the mediating agent (senses, inferable properties and words).
Indeed, the equation śruta = parokşa is a direct consequence that follows when sensuous cognition (mati-jñāna), along with its fours stages of sensation (avagraha) etc., is classified as a conventional variety of perception. If we closely examine Model 1 of cognitive criteria (p. 351), we discover that as soon as we move the four-fold sensuous cognition to the category of direct cognition (pratyakşa), what remains under the heading of indirect cognition (parokşa) is just one variety, viz. testimonial cognition (śruta-jñāna). Thus, it should not be surprising that Akalanka makes use of both of these two terms without much distinction: both of them have the same extension, that is to say they are identical as far as the class of cognitive acts they cover is concerned. However, their intension is different, since they
69 PSaV 1.2 (p. 97.6): śrutam aviplavam pratyakşânumānāgama-nimittam. parokşam pratyabhijñādi smarana-pürvakam. The wording (-nimittam) goes back to the wording of TS 1.14: tad (= mati) indriyanindriya-nimittam (albeit TS speaks of mati-jñāna), which has inspired a number of statements in Jaina philosophical literature. As regards the contents of the phrase, it likewise goes back to TS 1.20: śrutam mati-pūrvam ... - "testimonial (mediated) cognition is preceded by sensuous cognition. The latter is reformulated here by Akalanka as parokşam ... smarana-pūrvakam, for paroksa is a synonym of śruta, and mati is a synonym of smrti/smarana.

Page #37

A DISPUTE BETWEEN DINNĀGA, DHARMAKĪRTI AND AKALANKA 379
connote two different phenomena: paroksa, on the one hand, refers to all indirect, mediated acts of valid cognition, that do not grasp their object in a lucid or clear manner; śruta, on the other, indicates such acts of valid cognitions through which a piece of knowledge acquired by one cognitive subject is communicated and transmitted to another cognitive subject.
The verse of the Nyāya-viniścaya quoted already (NVi2 3.88, p. 366, and n. 44) contains one more important clue. In it, Akalanka provides the rationale why sensuous cognition (mati-jñāna), which has traditionally been classified (together with testimonial cognition (śruta) as indirect cognition, can also be described as a kind of perception: ‘it is only in order to be in agreement with people's opinion, that sensuous cognition is incorporated into the definition of (perception).' What is meant here is conventional perception (sāṁvyavahārika-pratyakşa), which is distinguished, in the same verse, from perception proper (mukhya-pratyakşa), called in the verse aparaṁ. Even though Akalanka mentions neither avadhi (clairvoyance) nor manah-paryāya (mindreading) - traditionally grouped together with the perfect, absolute cognition (kevala-jñāna) - by name in any of his epistemological works, it is clear that what is implied by aparaṁ in the verse NVi, 3.88 are precisely all these three varieties. There can be no doubt that he recognised this tripartite classification of avadhi-manah-paryāya-kevala, for he explicitly and extensively commented on them in his Rājavārttika.
To recapitulate, in Akalanka's scheme, the category of direct cognition (pratyakşa) comprises - as its peculiar conventional variety - sensuous cognition (mati-jñāna) along with its four stages: sensation (avagraha), cogitation (ihā), perceptual judgement (avāya) and retention (dhāranā), the latter being the basis for memory, a kind of storage of experienced events that are subsequently reactivated.
IX
An interpretation of the enigmatic verse of LT 10cd-11ab should take into account the defining feature of direct cognition, viz. the criterion of clarity or lucidity, which should help determine how Akalanka classified such cognitions as sensuous cognition, memory, recognition, induction, determined cognition etc. It would seem that all these cognitive acts grasp, in the first place, general features of an object, and have little in common with the notion of clarity/lucidity attributed to perception. Should we then conclude that Prabhācandra

Page #38

380
suri's interpretation (Int. 3, see p. 347, and n. 6), who catalogues them as varieties of indirect cognition, is the right one and all the remaining commentators misunderstood the verse completely, not noticing the patent contradiction in attributing memory, recognition, etc., a direct, perceptual character?
PIOTR BALCEROWICZ
Prabhācandra's interpretation would find its support in Akalanka's verse of the Nyāya-viniścaya 3.83 (quoted in p. 362, and n. 37) that 'the other [type of cognitive criterion] is testimonial cognition ... divided into [such varieties as] recognition etc.', insofar as the expression 'recognition etc.' (pratyabhijñâdi) stands for the whole set of cognitive processes that includes presumptive knowledge/suppositional knowledge (uha/tarka), determined cognition (abhinibodhika), memory (smrti), inference (anumāna) etc. In latter part of this verse (NVi2 3.83bc) Akalanka is, therefore, concerned with varieties of indirect cognition (parokṣa). That is also in full agreement with another verse of the Nyaya-viniscaya in which Akalanka makes it clear that testimonial (mediated) cognition comprises inference and scriptural testimony:
All this cognition [dealt with just before, including] inference as well as scriptural testimony, is testimonial cognition, provided what is cognised is truth (reality) without inconsistency with tradition."
70
The commentator Vādirāja-sūri expresses what is actually understood from the context itself and on the basis of previous verses (NVi2 3.83-86): 'Not only memory etc., are testimonial cognition, but also inference and likewise scriptural testimony.'
,71
At this stage we should determine how, in Akalanka's thought, the notion of clarity/lucidity, that distinguishes perception and indirect cognition, relates to the issue of conceptualisation, which for Dinnaga differentiates perception from inference?
In the Laghiyas-traya Akalanka describes each of the four cognitive stages comprised under sensuous cognition. Interestingly enough, the conceptual component (vikalpa, kalpana), the associating of 'bare sensory data' with mental concepts, is present in the process almost from the very beginning; that clearly distinguishes Akalanka's position from that of Dinnaga and Dharmakirti for whom such
70 NVi2 3.87 (p. 363):
sarvam etac chruta-jñānam anumanam tathagamaḥ | sampradayâvighatena yadi tattvam pratiyate ||
7 NVIV 3.87 (p. 363.14-15): tad aha - anumanam tathagamaḥ, na kevalaṁ smrtyādikam eva śruta-jñānam, api tu anumānaṁ tadvad āgamaś cêti.

Page #39

A DISPUTE BETWEEN DINNĀGA, DHARMAKĪRTI AND AKALANKA 381
conceptual components were a mere fiction, be it a practically useful fiction. To Akalanka, however, the involvement of the mind which processes bare sensations is an integral part of perception. Conceptual impression, called by Akalanka 'comprehension of the notion' (vikalpa-dhi), or conceptual awareness of the object, which is a synonym of kalpanā and vikalpa, emerges in perceptual process as soon as we become aware of the existence of an object:
When it comes to the union of the perceiving organ and the object, (first) (la) the seeing of (mere) existence (takes place, and then) (1b) a comprehension of the notion (conceptual impression) in the form of the objects - [these constitute) sensation. [It is followed by] (2) cogitation, (which) is the wish (to know) the particular (object, and by] (3) perceptual judgement, (which is] conclusive determination (of the object]. (4) Retention is the cause of memory. Thus sensuous cognition is four-fold. 74
Without this conceptual awareness of the object the cognitive subject could not become aware of anything, not even that he perceives anything:
Any conceptual states of mind, of whose rise and cessation every one is aware, [if they were non-conceptual at the moment of self-illumination), even if they exist, would not be observed in singular acts of perception, just like the singular character of the unique nature [is not observed)."
Conceptualisation thus accompanies through the whole process of cognition, right from the moment when the object enters awareness. Conceptual state of mind defines both acts of perception and indirect
72 This initial phase called sensation of the object (arthâvagraha) corresponds to what the Naiyāyika call nirvikalpaka-pratyakşa. See n. 60.
"This and all the remaining phases are roughly equivalent to the Naiyāyika's savikalpa-pratyaksa.
74 LT 5-6ab: akşârtha-yoge sattâloko 'rthâkāra-vikalpa-dhih / avagraho višeşâkānksehāvāyo viniścayah // dhāraṇā smrti-hetus tan mati-jñānam catur-vidham/
75 LT 24: pratisarviditôtpatti-vyayāḥ satyo 'pi kalpanā / pratyakşe na lakşerams tat-svalaksana-bhedavat // Cf. SVIV 27 (p. 115.11-12): na hi samvitter bahu-bahu-vidha-prabhrty-ākrtayah svayam asamviditā evôdayante vyayante vā yatah satyo 'py anupalakṣitāh syuh kalpanāvat tahtêtarâkrtayah. - 'For no cognitive episodes (that represent the character] of being many / much, manifold, etc. do not arise or cease in the awareness, themselves not becoming the object of awareness, because - although being present - they would not be apprehended, just like the conceptual state of mind; in the same way it is the case with the cognitive episodes of their opposites (such as little, simple, etc. (see TS 1.15-16)].'

Page #40

382
PIOTR BALCEROWICZ
cognition, except for the initial stage of mere contact of a sense organ and its object, when no cognition is yet generated.
Akalanka maintains that such conceptually constructed ideas can also occur in a clear and lucid manner. Besides, conceptualisation has two aspects: it may present its object indirectly, as an idea of vague form, but the way it emerges is clear and lucid (sc. direct):
Any aspect [of cognition] amenable to self-illuminating cognition, in case of conceptual (determinate) cognitions, manifests its contents clearly, because it is manifested as conceptual (determinate), [even] when [a person's] all thoughts are withdrawn (pacified).
76
That may lead to the conclusion that the series of cognitive processes, enumerated in the ambiguous verse LT 10cd-11ab (i.e., memory, recognitive cognition, inductive thinking (association) and determined cognition), even though they are conceptual in nature, could be classified as direct cognitions in terms of the manner in which they themselves become comprehended as acts of cognition.
When we analyse the way Akalanka applies the terms referring to the series of cognitive processes, we discover that he makes use of two series of terms that are normally considered synonyms. For instance the term samjñā (recognitive cognition') is a synonym of pratyabhijñā pratyabhijñāna, but both of them are generally used in slightly different contexts. That is the case with other terms occurring in both
76 LT 23:
sva-samvedyam vikalpānāṁ visadar thâvabhāsanam | samhṛtyâseṣa-cintayām savikalpâvabhāsanāt ||
This is a direct critique of Dharmakirti's idea expressed in PV, 2.123-125: pratyakşam kalpanapodham pratyakṣenaiva sidhyati / praty-atma-vedyaḥ sarveṣam vikalpo nama-samśrayaḥ || samhrtya sarvatas cintam stimitenântarâtmanā /
sthito 'pi cakṣuṣā rūpam ikṣate sākṣajā matiḥ ||
punar vikalpayan kiñcid asin me kalpanêdrsi /
iti vetti na pūrvôktâvasthāyām indriyad gatau || -
-[123] Perception is proved to be free from conceptuality by the very perception itself. [Every] conceptual cognition of all [beings] is cognised by itself individually and is correlated with the words. [124] After one has withdrawn one's own thought from everything, even though one abides tranquil inside oneself, the mind incited by the senses sees [non-conceptually] some form with the eye. [125] When, again, one conceptually conceives of that something, [in that very moment] one realises: "Such and such conceptual judgement has occurred to me!", [but] one is not [aware of that] when there is [still] the [non-conceptual] cognitive state [of mind which has arisen] from the sense-organ, [and] the condition of which has just been indicated [with the words: "After one has withdrawn one's own thought from everything"].' For an alternative translation of the passage see: Dreyfus (1997: 350).

Page #41

A DISPUTE BETWEEN DINNĀGA, DHARMAKĪRTI AND AKALANKA 383
series.?' In this way we have two quasi-synonymous series of cognitive acts that are related causally:
1) mati [memory
samjñā recognitive coginition
→ cintā → inductive thinking
(association)
→ abhinibodha → determined cognition]
2) smrti
[memory
+ pratyabhijñāna → recognition
→ üha/tarka + suppositional
knowledge
→ ābhinibodhika anumāna
inference 'for oneself"]
L
au.1
The difference between these two series lies in their either nonverbal or verbal character. The former of this series is rather to be reckoned among direct, perceptual acts of cognition, whereas the latter most frequently is associated with indirect cognition."
77 See, for instance: (1) LTV 10cd-11ab: ...dhāraṇā, smrtih samjñāyāḥ ..., cintāyāh ..., abhinibodhasya..., (2) SVI 2.1 (p. 120.4): smrtyä pratyabhijñānatā, comp. Svi? 2.1 (p. 120.20–21): mateh smrtih, tatah pratyabhijñā, ata ühaḥ, asmäd aśābdânumā śrutam ity uktam bhavati; and (3) SVIV 2.23 (along with Svił, p. 217.27– 218.1): avagraha adir yasyāḥ să câsau matih samāsaḥ: să ca smrtiś ca sarjñā ca cinta ca abhinibodhaś ca te ātmāno yasya tathôktam. (...Sensuous cognition is a compound piece of knowledge: it is both memory, recognitive cognition, thinking and determined cognition alike...'); (4) RVār 1.13 (p. 58.15-16): mananam matih smaranam smrtih samjñānam samjñā cintana namin an cu:
cintā ābhimukhyena niyatam bodhanam abhinibodhah iti = SVIV 1.27 (p. 115.14-15).
2 US 1415) 78 Since mati and smrti (likewise abhinibodha and ābhinibodhika) are used interchangeably, in order to determine which of these quasi-synonymous series of cognitive acts occurs in the context of either perception or indirect cognition, we should look for occurrences of middle terms (samjñā, pratyabhijñāna, cintā, ūha /tarka). (1) Terms of the first series (mati samjna cinta → abhinibodha) occur, as far as I could see consistently, in the context of sensuous cognition (mati-jñāna), or perception (pratyaksa), e.g.: RVār 1.13 (mananam matih smaranam smrtih saṁjñānam saṁjñā cintanam cintā ābhimukhyena niyatam bodhanam abhinibodhaḥ iti), SVIV 1.27 (mananaṁ vā it matih, smaranaṁ smrtiḥ, samjñānam samjñā, cintanaṁ cintā, ābhinibodhanam abhinibodha iti), SVIV 2.23 (...mati-jñāna-prabheda-lakṣaṇam avagrahâdi-matismrti-samjñā-cintâbhịnibodhâtakam ...), LT 10cd-11ab (jñānamādyam smrtih samjñā cintā câbhinibodhikam), LT 25 (aksa-dhi-smrti-samjñabhiś cintayâbhinibodhikaiḥ), LTV, 61 (p. 21.7: anindriya-pratyakşam smrti-samjñācintâbhinibodhâtmakam). (2) Terms of the second series (smrti pratyabhijñāna ūha/tarka → ābhinibodhika/anumāna) are used in the contect of inference (anumāna) or verbal cognition, e.g.: NVi2 3.83 (śrutam ... prakirnam pratyabhijñādau), NV, 188 (sadrsyâtmani sambandhagrahe ... pratyabhijñādinā siddhyet), PSa 1.2 (paroksam pratyabhijñādi), PSaV 1.2 (parokşam pratyabhijñâdi), PSa 3.17 (pratyakşa-nirnayan na rte sämänyânusmrtis tatah / pratyabhijna tatas tarkah tatah sādhya-parigrahah II, PSa 73, p. 122.21-22 (...anvayavyatirekârtha-vişayatvam sabda-grahanam pratyabhijñānâder...). The only case where both series are mixed is SVI 1.27 (akşa-jñānair anusmrtya pratyabhijñāya cintayan / ābhimukhyena tad-bhedān viniścitya pravartate // ), in which only verbal equivalents of the technical terms are used. The context here is perception. Apart

Page #42

384
PIOTR BALCEROWICZ
What ultimately decides whether these four stages are direct or indirect is their relation to language. Their limited equivalence is described by Akalanka in his own commentary to the problematic verse:
Since [retention] is a cause of infallible memory as the [cognitive] result (cognition), retention is the cognitive criterion [of it]; [similarly] memory [is the cognitive criterion] for recognitive cognition, which is the identification; recognitive cognition [is the cognitive criterion] for inductive thinking, which suppositional knowledge; inductive thinking is [the cognitive criterion] for determined cognition, which is inference etc., [still] before the association with speech. {After the association with speech} [it is] the remaining [kind of] cognition [which is] testimonial cognition, [and] which is of many subdivisions.
As a result, language emerges as an additional criterion that distinguishes direct cognitions from indirect ones. The division into direct (clear and lucid) and indirect (unclear and bereft of clarity) cognitive acts overlaps with the segmentation: non-verbal and verbal:
pratyakṣa (perception)
parokṣa (indirect cognition)
visada (clear)
avisada (unclear) śābda (verbal)
asabda (non-verbal)
akalpita (non-conceptual) | kalpita (conceptual) kalpita (conceptual)
(Foonote 78 Continued).
from the above analysis of the occurrences of single terms belonging to these two series, an additional support of my thesis that that the series containing recognition (pratyabhijñāna) as a kind of indirect cognition, is found in a passage of the Asta-sati ad AMi 101 (AsS3, p. 88.27-28 AşŚ4, p. 279.19-20): pratyabhijñānam pramāṇam, vyavasâyātiśayôpapatteḥ, tat-sämarthyâdhinatvāt pramānatva-sthiteḥ. - 'Recognition is a cognitive criterion, because the excellence of the determination (sc. correct cognition) [of an object though it] is possible, for it is established that it is a cognitive criterion, insofar as it depends on the efficacy of this [determination].' The reference to pratyabhijñāna is inserted among various other types of indirect cognitive criterion which Akalanka argues to be also valid cognitions, but are different from anumana. Clearly, pratyabhijñāna is treated by Akalanka as one of varieties of paroksa. Furthermore, in the relevant portion of his Asta-sahasri ad AşŚ, Vidyananda Pātrakesarisvämin confirms that pratyabhijñāna is a kind of indirect cognition (AsS, p. 279.19: pratyabhijñāṇaṁ pramānam, vyavasayātiśayôpapatteḥ, pratyakṣâdivat, repeated on p. 280.3-4: tatḥ pratyabhijñānam tattva-jñānatvāt pramānam, pratyakṣâdivat), which is juxtaposed with perception.
79 LTV 10cd-11ab: avisamvada-smṛteḥ phalasya hetutvāt pramānam dhārana, smrtih samjñāyāḥ pratyavamarśasya, samjña cintayaḥ tarkasya, cinta abhinibodhasya anumanâdeḥ prāk sabda-yojanat. {śabdânuyojanat} sesam śruta-jñānam aneka-prabhedam.
3 Emendation according to my critical edition of LT/ LTV (under preparation). Without my emendation, and with the reading: cinta abhinibodhasya anumanâdeḥ prāk sabda-yojanāt seṣam śruta-jñānam aneka-bhedam, we would have: ...inductive thinking is [the cognitive criterion] for determined cognition, which is inference etc. [What is the first kind of cognition, i.e., sensuous cognition which occurs] before the association with speech, [becomes] the remaining [kind of] cognition, i.e., testimonial cognition, which is of many subdivisions.' Whether we take the suggested emendation into consideration or not, the meaning would not change considerably.

Page #43

A DISPUTE BETWEEN DINNĀGA, DHARMAKĪRTI AND AKALANKA 385
The above division immediately evokes Dinnāga's assertion, which must have influenced Akalanka's classification and definition of cognitive processes:
When there arises (a cognition as a result of the contact] of a sense organ and an object, it is not possible [for the contents of this cognition to have) verbal character etc. (sc. and conceptual character). 80
Dinnāga's above claim is, clearly, a direct consequence of his distinguishing two autonomous kinds of cognition following the criterion of conceptualisation: for him, conceptually constructed ideas were automatically expressible with the help of verbal symbols. Akalanka, however, argued to the contrary: it is possible to have cognition, of which we are aware and which is, at the same time, conceptual, i.e., its contents is already contaminated by the influence of mental activity such as cogitation, the determination of the nature of perceived object etc. Furthermore, most cognitive processes that constitute perception, with the exception of the initial non-conceptual phase of vyañjanâvagraha, have such a conceptual nature. For Akalanka, the validity of such a conceptual perception is warranted by the fact that each and every subsequent step in the perceptual process is causally dependent on the preceding one, from sensation (avagraha) up to retention (dhāraņā), and further, such a cognitively valid retention results in the subsequent steps (memory, recognition, induction, inference) that are likewise causally related.82 Ultimately,
80 PSV 1.17ab: indriyârthôdbhāve nâsti vyapadesyâdi-sambhavah. Cf. the translation in: Hattori (1968: 36, Section 3,k. lab), Sanskrit text: Hattori (1968: 121, n. 3.3). The phrasing of PSV 1.17ab seems to have triggered, for instance, LT 5ab (akşârtha-yoge sattâloko 'rthâkāra-vikalpa-dhih).
8See nn. 60 and 72, pp. 374 and 381.
82 Cf. STPT 2.1, p. 553, 6-9: atra ca pürva-pürvasya pramanatā uttaröttarsya ca phalatêty ekasyâpi mati-jñānasya cāturvidhyam kathamcit pramāņa-phala-bhedas côpapannah. yathā ca pratibhāsa-bhede 'pi grāhya-grāhaka-samvidāṁ yugapad ekatvam tathā krama-bhāvinām avagrahâdināṁ hetu-phala-rūpatayā vyavasthita-svarūpāņām ekatvam kathaṁcid aviruddham anyathā hetu-phala-bhāvābhāva-prasaktir iti pratipāditam anekaśah. - 'And as regards this (four-fold division of the sensuous cognition) each and every previous (act of cognition) is a cognitive criterion, and each and every following (act of cognition) is (their) result; therefore even though the sensuous cognition is one, it is somehow four-fold and what results is the division into the cognitive criterion and [its) result. And just like with regard to a variety, as it were, of representations, acts of awareness (of the form of] something-to-be-grasped and something which-grasps (preserve their simultaneous unity, in the same manner unity of [the four stages of the sensuous cognition such as sensation, etc. - that (with their] determined intrinsic natures exist consecutively, since they have the form of either cause or result - somehow (presents) no contradiction; otherwise [one would face) an absurd consequence that the cause and result would [both) exist and not exist. This has been taught in many ways.'

Page #44

386
PIOTR BALCEROWICZ
this causal series of cognitions leads to testimonial cognition, which is verbal in character (see LTV 10cd-1lab, p. 42, n. 79). In short, the final element in the process, viz. reliable verbal cognition, is preceded by an act of perception.
Although Akalanka disagrees with Dinnāga as regards conceptual character of perceptions, both of them would agree that verbal cognition that is causally depended on previous direct experience, or perception, is infallible (avisamvādin, see PS/PSV 2.5), viz. it is a cognitive criterion (pramāna, see LTV 10cd-1lab). For Akalanka, the causal sequence of cognitive events, each of which serves as a pramāna for the subsequent one (LTV 10cd-1lab), warrants the validity of testimonial cognition: a valid cognition yields another valid cognition. That is reminiscent of Dinnāga's comments on PS 2.5 (quoted on p. 354/5, and n. 21), in which he explains why testimony of an authoritative person, which is as infallible as inference, can be classified as a kind of the latter: 'Since when one has heard an utterance of an authoritative person, on account of the similarity in terms of infallibility it is called inference. And, as it is said, all name-giving is preceded by perception. 83 Also for Dinnāga a valid act of perception is a necessary ground for the validity of testimony.
Incidentally, Akalanka description of perception, as distinct from indirect cognition in terms of its non-verbal character, is a criticism against Bhartshari's claim that all cognition has a verbal character, and every thought must have its verbal equivalent:
There is no cognition in the human world which would not result from speech. All cognition appears on account of speech as if permeated [by it).**
As against Bharthari's opinion,85 Akalanka endeavours to prove that it is possible to formulate thoughts independently of
83 PSV 2.5 (PSV2, p. 455.1-4): [Vasudhararakṣita:) yid ches pa'i tshig ñid bzun nas kyañ mi bslu bar mtshuns pa'i phyir de yan rjes su dpag pa nid du brjod do || de skad du yan min gi las rnams kyi don du mnon sum son ba'i phyir ro zes 'byun no 1, [Kanakavarman:) yid ches pa'i tshig ñid bzun nas don de la mi bslu ba'i phyir dan | mi 'dra ba'i phyir rjes su dpag pa nid du bcad pa yin te de Itar yan min gi las ni mnon sum snon du 'gro ba can zes bya ba yin no
84 VP 1.131: na so 'sti pratyayo loke yaḥ śabdânugamad rte / anuviddham iva jñānam sarvam śabdena bhāsate //
85 That my contention about the implied criticism against Bharthari's theory is not arbitrary is proved by Vidyānanda's following remark, who comments on Akalanka's verse of LT 10cd-1 lab: in TŠVA, p. 240: athavā na so 'sti pratyayo loke yah sabdânugamās rte /

Page #45

A DISPUTE BETWEEN DINNĀGA, DHARMAKĪRTI AND AKALANKA 387
language; moreover, even such complex cognitive processes as inner association, induction and reasoning may proceed without any involvement of verbal symbols. Concepts and ideas are an integral component of every thought without which no thinking process would be possible; however, we can think non-verbally.
The final picture of Akalanka's classification of cognitive process that emerges after above examination is shown in Model 3 below:
cognitive criteria (pramāna)
P indirect cognition (paroksa)
direct cognition (pratyaksa)
testimonial (mediated) cognition (śruta)
clairvoyance (avadhiy primary perception
mind-reading (manah paryaya) (mukhya-pratyaksa)
absoluto knowledge Xeyalar conventional perception (sāmvyavahărika-pratyaksa)
scriptural testimony (agama)
sensory perception non-sensory (mental) perception (indriya-pratyakşa) (anindriya-pratyakşa)
inference ('for others'] ([parártha-J-unumāna)
sensory cognition (indriya-pratyakşa)
occasioned by sensuous cognition (mati-nimitta)
memory (smrti) recognition (prutyabhijñāna) suppositional knowledge (liha / tarka)
inference ('for oneself
(svartha-|-anumāna)
sensation (avagraha)
cogitation (Tha) perceptual judgement
(apāya) retention (dhārand)
memory (mati = smrti) recognitive cognition (samjña) association (cinta) determined cognition (abhinibodha)
verbal cognition (ndma-yojanat)
pre-verbal cognition (pran nama-yojanar)
Model 3
(Foonote 85 Continued). anuviddham ivâbhāti sarvam śabde pratisthitam // * ity ekānte nirākartu prāg nama-yojanād adyam istam na tu tan nāma-saṁsrstam iti vyākhyānam akalankam anusartavyam. -
- Alternatively, in order to revoke the following absolutist stand (of Bharthari): “There is no cognition which (can exist) without following speech in the human world; everything shines as if pierced (with words), being firmly based in speech," it is accepted that the first (cognition, viz. sensuous cognition), (occurs] before the association with name (speech), but it is not connected with name (speech) - such an explanation should be adopted following Akalańka.' See also n. 3.

Page #46

388
PIOTR BALCEROWICZ
The field marked a comprises those varieties of valid cognitions which are not mentioned by Akalanka by name in his works on epistemology (Laghiyas-traya, Nyāya-viniscaya, Pramāna-sangraha and Siddhiviniścaya). Nevertheless, there can be no doubt that he followed Jaina tradition in this respect and accepted all three: including clairvoyance and mind-reading. He comments on them and explains their nature for instance in his commentary on the Tattvärtha-sūtra (RVār 1.9-)
In accordance with the results of the above analysis, the initial translation of the enigmatic verse of LT 10cd-11ab should be modified as follows:
Int. 4 The first [kind of] cognition (also 86 comprises) sensuous cognition (hère: memory), recognitive cognition, and inductive thinking (and) determined cognition; (it occurs] before the association with name (speech). [These four become) - on account of application of speech - the remaining kind of cognition, i.e., testimonial cognition.
Thus, the same sequence of mental processes - i.e., memory, recognitive cognition, inductive thinking and determined cognition, etc. - can be either classified as sensuous cognition (sc. conventional perception), or as testimonial cognition, depending on whether the processes are accompanied by the application of language or not. This interpretation is in agreement with a passage found in the Siddhi-viniścaya:
This sensuous cognition, which consists in the cognitive expectancy (to determine its] datum which is an individual [thing) seized through sensation of its own object (sc. a datum relating to a particular sense organ), also leads - after perceptual judgement has been reached to the fact that a subliminal impression (i.e., retention, dhāranā), amounting to determination of the object, is deposited]88. On account of reason based on the contents of presumptive knowledge (obtained] by a person who recognises (an object) due to memory non-verbal inference is produced, which is determined cognition. It can become testimonial cognition (when) connected with speech.89
(Foonote 85 Continued). aVP 1.131, modified in padas cd. Vidyānander's reading is also found in NBhu, p. 580.24, verse 356.
86 Viz. in addition to the four initial steps: sensation (avagraha), cogitation (iha), perceptual judgement (avāya) and retention (dhāranā).
87 Alternatively: svârthâvagraha could mean: 'sensation of itself and of [its external] object', if we follow Svi? ad loc. (p. 120.6): svam ca arthaś ca tayoh grähako 'vagraha.
88 This is again the description of the four phases: sensation (svârthâvagraha), cogitation (inā = visayâkānkşā), perceptual judgement (avāya = bhedāvāyām upetya) and retention (dhāraņā= sarskāratām yāti).
89 Svi 2.1 (p. 120.2-5): svârthâvagraha-nita-bheda-visayâkānkşâtmikêyam matih bhedāvāyām upetya nirnaya-mayam samskāratām yāty api/ smrtyā pratyabhijñāvatôha-visayād dhetor aśābdânumā kalpyā ābhinibodhiki śrutam atah syāt sabda-samyojitam //

Page #47

A DISPUTE BETWEEN DINNĀGA, DHARMAKĪRTI AND AKALANKA 389
In this manner Akalanka accomplishes two seemingly irreconcilable tasks: he introduces a new model of valid cognitive processes and, at the same time, he declaratively remains faithful to older tradition expressed by Umāsvāmin: 'sensuous cognition, memory, recognitive cognition, inductive thinking (and) determined cognition are synonymous expressions (TS 1.13, p. 6).
The idea expressed in the hemistich of LT 10cd (jñānam ādyam matiḥ samjñā cintā câbhinibodhikam) - as I understand it, viz. that memory, recognitive cognition, inductive thinking (association) and determined cognition are integral parts of perception - is further confirmed by the verse of SVI 1.27, which clearly relates these four mental processes to sensuous cognition, i.e., conventional perception derived from senses:
With the help of cognitions (derived) from sense organs, having remembered [and] recognised (perceptible things), one associates (or: reasons inductively) [and] having determined intentionally individual [objects grasped by] these (cognitive processes, a person) undertakes actions (with respect to these individual objects). 91,92
The verse associates the sequence of cognitive events that are causally related both with acts of perception (akşa-jñānaih) and with practical actions undertaken with respect to individual entities, not universal ideas, for one practically acts towards individual entities. Further, in another place, Akalanka describes how individual entities are being grasped through perception, which lends further support that perception, with the exception of the initial step, has generally conceptual character:
[First] one does not distinguish its own [individual data) from other [things, i.e., one does not even know with what sense organ one perceives), (then) one sees the existence alone (sc. that something is there) because one is not yet aware (of the character
90 The expression ābhimukhyena ... viniścitya 'to determine intentionally' is a paraphrase of the term abhinibodha ("determined cognition'), see Akalanka's explanation in RVār 1.13 (p. 58.15-16): mananam matih smaranam smrtiḥ samjñānam samjna cintanam cinta ābhimukhyena niyatam bodhanam abhinibodhaḥ iti.
91 Anantavirya in his commentary makes it clear that all these processes are indispensable steps in perception, and none of them can be omitted; in SviT ad loc., p. 115.19-23, he enumerates six possibilities, each leaving one cognitive step out, which all lead to impeded cognition.
92 Svi 1.27 (p. 115.6-7): akşa-jñānair anusmrtya pratyabhijñāya cintayan / ābhimukhyena tad-bhedān viniścitya pravartate //

Page #48

390
PIOTR BALCEROWICZ
of the data];" (then) one recognises the individual thing etc. and grasps [it] conceptually through (various) stages of sensuous cognition.94
His auto-commentary, in which Akalanka elaborates on the main ideas accommodated in the above verse, is quite 'revealing. At the outset, Akalanka enumerates six sense organs, the mind (manas)
e va yataḥ satyo kante
punah amanam eva vipralain, smaranam smrlari
93 These two initial steps (avikalpya svam anyasmāt sattā-mātràm acetanāt / pasyan) are the sensation of momentary manifestation (vyañjanavagraha), which is not conceptual. Thereupon follows the sensation of the object (arthâvagraha), which is conceptual, and other conceptual steps of perception.
94 SVI 12.4 (p. 742.1-2): avikalpya svam anyasmāt sattā-mātram acetanāt / paśyan vijānāti dravyâdi mati-bhedair vikalpayan / Compare also Anantavirya's relevant comments in SVIȚ ad loc. (p. 742.6-15).
95 SViV1.27 (p. 115.8-18): sparśana-rasana-ghrāņa-caksuh-śrotra-manāmsi indriyāni. taih sva-visaya-grahanam avagrahady-ātmikā matih bahu-bahu-vidhaksiprânih srtânukta-dhruvāņāṁ sêtara-prakārāņām (TS 1.16). ata evânekānta-siddhih. na hi samvitteh bahu-bahu-vidha-prabhrty-ākrtayaḥ svayam asam viditā evôdayante vyayante vā yatah satyo 'py anupalakṣitāḥ syuḥ kalpanāvat, thatêtarâkrtayah. svalaksana-sāmānya-laksanaikānte punah samvedanâkrtih na paśyāmah, tathāvaipaśyantah katham {kalpanayā ātmānam eva vipralabhamahe. tad evam paramarthatah siddhiḥ anekāntāt. manyate (mananam vā it matih, smaranam smrtih, sarjñānam samjñā, cintanam cintā, ābhinibodhanam abhinibodha iti] tathamanati tattvartha-sūtra-kārāḥ-"matih smrtih samjñā cintâbhinibodha ity anarthāntaram" (TS 1.13) iti. mati-smrty-ādayaḥ śabda-yojanām antarena na bhavantity ekānto na yatas tatrântar-bhāvyeran. tad-ekānte punaḥ na kvacid syuh tan-nāma-smrter ayogāt anavasthānâdeh. - 'Touch, taste, smell, vision, hearing and mind are senses. With their help sensuous cognition, which consists in (four stages such as] sensation, etc., is an act of grasping of respective data for each of these (senses), (and it concerns) such divisions as many/much, manifold, quick, unclear, implicit, fixed and their opposites (TS 1.16). Precisely on the basis (of this act of grasping) the multiplexity (of data) is established. For no cognitive episodes (that represent the character) of being many/ much, manifold, etc., arise or cease in the awareness, themselves not becoming the object of awareness, because - although being present - they would not be apprehended, similarly the conceptual state of mind (would not be cognised); in the same way it is the case with the cognitive episodes of their opposites (such as little, simple etc.). On the other hand, we do not perceive cognitive episodes within awareness that concern unique particulars and universal things that are absolutely separate from each other. [The Buddhists are deceived, but] why do we, not perceiving (things) in precisely such a manner (sc. not perceiving that unique particulars and universal features are absolutely separate from each other), allow ourselves to be deceived {by such assumption}" [that we can perceive universal things that are absolutely separate from each other)? Thus, in exactly such a way, (viz. by grasping things as many/ much, manifold, etc., we arrive at) a proof (that), because of multiplexity of appearance], on the level of the ultimate truth (unique particulars and universal features coexist in cognition). It is maintained that sensuous cognition is understanding, memory is recollecting, recognitive cognition is recognising, inductive thinking is associating, determined cognition is determining intentionally. Accordingly, the author of the Tattvartha-sūtra (1.13] considers that: “Sensuous cognition, memory, recognitive cognition, inductive thinking (and) determined cognition are

Page #49

A DISPUTE BETWEEN DINNĀGA, DHARMAKĪRTI AND AKALANKA 391
being included as the sixth sense. Since the mind grasps concepts and thoughts, the character of mental perception has to be conceptual. Furthermore, Akalanka takes recourse to an earlier idea, expressed by Umāsvāmin (TS 1.16), that perception involves also 'such divisions as many / much, manifold, quick, unclear, implicit, fixed and their opposites', which also directly points to the conceptual nature of perception. Without acknowledging conceptual nature of perception, it would not be possible to explain how the cognising subject becomes aware of his own cognitions; as consequence, no cognition, including conceptual states of mind, would be cognised (anupalakṣitāḥ syuḥ kalpanāvat).
Finally, Akalanka examines the supposition that 'sensuous cognition, memory, etc., do not exist without the association with speech' (mati-smrty-ādayah sabda-yojanām antarena na bhavanti), which he immediately dismisses as simplistic (ekānta). If such cognitive processes were verbal, they should be classified as instances of testimonial cognition (na yatas tatrântar-bhāvyera”), and would turn out to be indirect cognitions. To accept verbal character of such cognitive processes as memory, recognition etc., that are considered by Akalanka to be recognitive cognition, would be tantamount to conceding that each perceptual image or each element of memory would necessarily have its own verbal expression. In consequence, without exact
(Foonote 95 Continued). synonymous expressions." (Suppose] sensuous cognition, memory, etc., do not exist without the association with speech - such is absolutist stand; [but it] is not correct, because [in such a case) these (varieties) should be included in this [testimonial cognition). On the other hand, if [one accepted] such an absolutist stand (that these cognitions are absolutely verbal}, then [these varieties) could not occur with respect to any [object], in so far as it would not be consistent to assume that memory (proceeds with the help of the names of these [objects), because of infinite regress (sc. it would be necessary to have as many names as there are individual things) and other [logical flaws).' sa sviV 1.27 (p. 115.13 reads: katham ātmānam. However, the text should be emended in view of Anantavirya's manifest pratika, overlooked by the editor, in Svi? 1.27 (p. 117.7): ...katham kalpanayā svalaksana-sāmānya-lakşanaikānte samvedanâkrtikalpanayā ātmānam eva vipralabhāmahe vañcayāmah.]
yo Precisely the same idea is expressed in LTV 7ab: bahu-bahu-vidhaksiprânisrtânukta-dhruvêtara-vikalpānām avagrahâdeh svabhāva-bhedan na virudhyate. - 'There is no contradiction (as regards the distinction between cognitive criterion and its result] in the case of sensation and other (stages of sensuous cognition that grasp) such divisions as many/much, manifold, quick, unclear, implicit, fixed and their opposites, because their essential nature is differentiated.'
* Compare Anantavirya's paraphrase, SVIT ad loc., p. 117.25–26: nanu matyadikam sarvam abhidhāna-puras-saram eva svârtham pratyeti iti sabda-śruta evântarbhāvo 'sya.

Page #50

392
PIOTR BALCEROWICZ
co-reference between perceptible data of experience and verbal symbols no perception would be possible.
XI
Above remarks about conceptual character of perception in Akalanka's thought hold true in the case of all conventional varieties of perception. A separate, though related question is whether Akalanka consistently ascribed conceptual character to all other kinds of perception as well, especially to the highest form of perception proper (mukhya-pratyakşa), viz. to absolute knowledge (kevala-jñāna), or to two inferior kinds of perception proper (mukhya-pratvaksa). viż. clairvoyance and mind-reading, which corresponded to what other schools would call 'yogin's perception' (yogi-pratyaksa).
As we know, for Dinnāga and Dharmakirti alike, all subcategories of perception did not differ qualitatively in terms of their non-perceptual character. Dinnāga emphasised that also perception which meditation adepts (yogins) could acquire through their spiritual practices was nonconceptual in character:
Yogins have the insight into the object as such which (sc. the insight) is unrelated to the spiritual preceptor's account. Also the seeing of the object as such, in the case of yogins, which (sc. the seeing) is uncontaminated by notions inculcated] by scriptural testimony is perception. 98
Dharmakirti directly referred to this passage in his Pramānavārttika and acknowledged that also such yogic perception, which seems to be ultimate perception in mundane world, is absolutely free of any conceptual inner states and conceptually constructed ideas:
[Perceptual) cognition of yogins was described before [by Dinnāga). This (i.e., perception) of these (i.e., yogins), which arises through meditation (and) in which the net of conceptuality has been shattered off, manifests itself as purely lucid."
98 PS / PSV, 1.6cd: rnal 'byor rnams kyi bla mas bstan ma 'dres pa yi don tsam mthong | rnal 'byor barnams kyis kyang lung las rnam par rtog pa dang ma 'dres pa'i don tsam mthong ba ni mngon sum mo For Sanskrit text see Hattori (1968: 94, nn. 1.48, 1.49), PS3 ( = DKC, p. 106.1), cf. Hayes (1988: 136, 170, n. 18): yogināṁ gurunirdeśâvyatibhinnârtha-mātra-drk // yoginām apy agama-vikalpâvyavakirņam arthamātra-darśanam pratyakşam. PS 1.6cd is quoted by Akalanka in RVār 1.12, p. 54.14.
99 PV 2.281: prāg-uktam yoginām jñānam teşām tad bhāvanā-mayam / vidhūta-kalpanā-jālam spastam evāvabhāsate //

Page #51

A DISPUTE BETWEEN DINNĀGA, DHARMAKĪRTI AND AKALANKA 393
It would, therefore, be highly interesting whether Akalanka considered that, like all conventional perception, also yogic perception is to a certain degree tainted by conceptually constructed forms. That would, however, indicate inconsistencies in his framework in which all perceptions he has discussed so far are conceptual.
Indeed, in his commentary Rāja-vārttika on the Tattvârtha-sūtra, Akalanka, having rejected several definitions of perception (RVār 1.12, p. 53.26 ff.), including that of Dinnāga, declares in the context of yogic perception that: And also the following is not correct: "There are no entities whose essence is conceptual, there are [only entities) whose essence is non-conceptual”; [it is not correct), because there are no means to comprehend these (non-conceptual entities). For it is possible to demonstrate that there is no object (which would be) non-conceptual and there is no cognition of such [object]; [it is possible to demonstrate that), because there are no characteristics (on the basis of which one could either see them or infer them).100 · Subsequently (RVār 1.12, p. 55.5-25), Akalanka directly criticises Dinnāga's claim that 'Perception is free from conceptual construction' (PS 1.3cd), and demonstrates that no direct cognition can exist which is 'free from conceptual construction' either completely or in a particular respect (p. 55.9–10: kim tat sarvathā kalpanâpodham utâho kathaṁcid iti). In a slightly different context, he adds that: “[Ultimate] perception, (which is acquired] independently of sense organs and the mind (and which is] free of deviation, is the seeing of definite contents.:101 This also supports the hypothesis that Akalanka consistently regarded all kinds of perception to be conceptual.
However, Dinnāga's and Dharmakirti's description of yogic perception the validity of which Akalanka explicitly rejected (RVār 1.12) brings to light a highly significant evidence. In his commentary (RVār 1.12, p. 54.14), Akalanka directly quotes PS 1.6cd and makes reference to Dharmakirti's commentary on it (PV 2.281), which features a crucial term: “[Perceptual] cognition of yogins ... manifests itself as purely lucid' (yogināṁ jñānam ... spassam evâvabhāsate). Even though Akalanka repudiates Dharmakirti's claim that also perception of yogins is completely non-conceptual, this passage as well as the verses that follow (e.g., PV 2.284: spasțâbham nirvikalpam ca
100 RVār 1.12 (p. 54.19-21): "parikalpitâtmanā na santi bhāvā nirvikalpâtmanā santi" iti câyuktam, tad-adhigamöpāyābhāvāt. na hi nirvikalpo 'rtho 'sti tad-vişayam jñānaṁ cêti pratipādayitum sakyam laksanabhāvāt.
101 RVār 1.12 (p. 53.4): indriyânidriyânapeksam atita-vyabhicāram sākāragrahanam pratyakşam.

Page #52

394
PIOTR BALCEROWICZ
bhāvanā-bala-nirmitam) - and which were certainly known to him - may have been the inspiration for Akalanka to introduce the criterion of lucidity or clarity into his definition of perception.
XII
We should now examine how far Prabhācandra's 'solution (Interpretation 3, p. 347, n. 6) differs from Akalanka's model, and with what aim he offered his own interpretation of LT 10cd-11ab. He classifies all processes enumerated in TS 1.13 (p. 348), viz. sensuous cognition, memory, recognitive cognition, inductive thinking and determined cognition, as types of indirect cognition. In this, he follows Umāsvāmin's classification (TS 9-12), who indeed reckoned them as types of parokşa, insofar as they were the stages of sensuous cognition (mati-jñāna), which was traditionally also classified as indirect cognition. In this way, Prabhācandra proves to be closer to earlier Jaina tradition.
For Prabhācandra perception (pratyakşa) was a clear and lucid cognition as well as completely pre-verbal, although in part accompanied by mental conceptualisation and related to conceptually constructed ideas. Indirect cognition was, on the other hand, divided into two subcategories: strictly non-verbal cognition and verbal cognition:
pratyaksa (perception) visada (clear) aśābda (non-verbal) akalpita (non-conceptual) | kalpita (conceptual)
paroksa (indirect cognition) avisada (unclear) aśābda (non-verbal) sabda (verbal) kalpita (conceptual)
How Prabhācandra's alterations changed the general classification of cognitively valid procedures, adopted by Akalanka, can be shown in Model 4 below:
The manner in which Prabhācandra reads into Akalanka's statements his own concepts does not essentially diverge from the manner in which Akalanka himself radically modified earlier Jaina tradition. In terms of historical development of Indian philosophy there was nothing surprising neither about Akalanka nor about Prabhācandra. Their manoeuvres were in agreement with general spirit of Indian philosophical tradition: novelties were introduced in disguise of earlier terminology, as faithful expositions of the earlier masters'

Page #53

A DISPUTE BETWEEN DINNĀGA, DHARMAKĪRTI AND AKALANKA 395
thought. How such interpretations may have eventually wiped out the legacy of the preceptor whose works underwent such a thoroughgoing interpretative treatment can best be illustrated by the relation between Dharmakirti and Dinnāga. In contradistinction to that, Prabhācandra's commentary did not override Akalanka's oeuvre.
cognitive criteria (pramana)
indirect cognition (parokşa)
direct cognition (pratyaksa)
testimonial (mediated) cognition(sruta)
clairvoyance (avadhi) primary perception
mind-coading (manah-parydya) (mukhya-pratyaksa) L
absolute knowledge (kevala) conventional perception (sārvyavaharika-pratyaksa)
'scriptural testimony (agama)
non-sensory (mental) perception
(anindriya-pratyaksa)
sensory perception (indriya-pratyaksa)
sensory cognition (indriya-pratyaksa)
memory (smrti) recognition (pratyabhijfāna) suppositional knowledge (aha / tarka)
inference ['for oneself)
(svartha-]-anumana)
memory (mati = smrti) recognitive cognition (sarja) association (cinta) determined cognition (abhinibodha)
sensation (avagraha) cogitation (tha) perceptual judgement (apaya) retention (dhärand)
verbal cognition (nama-yojant)
pre-verbal cognition
(prdn nama-yojanat) Model 4
As we could see from the above, the development of Jaina epistemology closely followed innovations in the Buddhist pramāna tradition. Dharmakirti and Dinnāga seem to be primary adversaries who triggered Akalanka's criticism and provoked him to reformulate several of his concepts. The influence the Buddhist philosophers exercised on Akalanka and other prominent Jaina thinkers was undeniable. That should not however lead one to believe that Akalanka's thought was somehow inferior to the sources of his intellectual inspiration, certainly not less than Dinnāga's or Dharmakirti's systems who ingeniously applied for instance Bharthari's ideas in their original philosophical edifice. Akalanka availed himself of borrowed ideas in a highly creative, sophisticated and independent manner; his categorisations of

Page #54

396
PIOTR BALCEROWICZ
cognitive phenomena were penetrating and his analysis of epistemological questions were very subtle. What is important to realise is that Akalanka - as any genuine philosopher and ingenious thinker - was a part not only of a his own doctrinal tradition, but also, to the same degree, of Indian philosophical heritage which included also Buddhist thought.
REFERENCES
"[= 19821bhadra. si ommentary
AnD = Aņuoga-ddārāim (Anuyoga-dvārāni). Muni Punyavijaya, Dalsukh Mālvaniā,
Amritlāl Mohanlāl Bhojak (ed.): Nandi-suttam and Anuoga-ddārāim. Jaina
Agama-Series No.1, Shri Mahāvira Jaina Vidyālaya, Bombay 1968. ĀMi = Samantabhadra: Apta-mimāṁsā. (1) Pannālāl Jain (ed.): Apta-mimāṁsā of Samantabhadra Svāmi, with two commentaries: Asta-sati of Bhattākalanka and Devāgama-vrtti of Vasunandi. Sanātana Jaina Granthamālā 10(7), Kāsi 1914. (2) Udayachandra Jain (ed., Hindi tr.): Tattvadipika. A Commentary with Introduction etc. on Aptamimansa of Acharya Samantabhadra. Sri Ganes Varni Digambar Jain Samsthān, Kāsi Vir-Sam. 2501 ( = 1982]. (3) Nagin Shah (ed., tr.): Samantabhadra's Apta-mimāṁsā. Critique of an Authority ( Along with English Translation, Introduction, Notes and Akalanka's Sanskrit Commentary Astašati). SanskritSanskriti Grantha-mālā 7, Ahmedabad 1999. (4) Saratchandra Ghoshal (ed., tr.): Apta-mimāṁsā of Achārya Samantabhadra. Edited with Introduction, Translation,
Notes and An Original Commentary in English. Bharatiya Jnanpith, Delhi 2002. AsS = Akalanka: Asta-sati. See: (1) AM, (3) AM; and (4) AşS. AşS = Vidyānanda Pātrakesarisvāmin: Asta-sahasri. Vamsidhar (ed.): Asta-sahasri
tarkika-cakra-cūdā-mani-syādvāda-vidyāpatină śri-Vidyānanda-svāminā nirākrta.
Nirnayasāgara Press, Bombay 1915. Balcerowicz, P. (1999). "How could a cow be both synchronically and diachronically homogenous, or on the Jaina notions of tiryak-sāmānya and urdhvatā-samanya',
in: WAGLE-QVARNSTRÖM (1999: 211-237). Balcerowicz, P. (2001a). Jaina Epistemology in Historical and Comparative Perspec
tive. Critical Edition and English Translation of Logical-Epistemological Treatises: Nyāyâvatāra, Nyāyâvatāra-vivrti and Nyāyavatāra-tippana with Introduction and Notes. 1-2 Vols., Alt- und Neu-Indische Studien 53, 1-2, Franz Steiner Verlag,
Stuttgart 2001. Balcerowicz, P. (2001b). 'Two Siddhasenas and the Authorship of the Nyāyāvatāra
and the Sammati-tarka-prakarana', Journal of Indian Philosophy 29/3 (2001) 351
378.
Balcerowicz, P. (2003). Dżinizm. Starożytna religia Indii. Historia, rytual, literatura.
Wydawnictwo Akademickie DIALOG, Warszawa 2003. van Bijlert, V. A. (1989). Epistemology and Spiritual Authority. The Development of
Epistemology and Logic in the Old Nyāyā and the Buddhist School of Epistemology with an Annotated Translation of Dharmakirti's Pramāņavārttika II (Pramāna
siddhi) vv. 1-7. WSTBK 20, Wien 1989. CarS = Caraka-samhitā. Vaidya Jādavaji Trikamji Achārya (ed.): The Charakasanhitā of Agniveśa. Revised by Charaka and Dridhabala. With the AyurvedaDipikā Commentary of Chakrapāņidatta. 3rd ed., Nirnaya-sāgara Pres, Bombay 1941 (reprinted: Munishiram Manoharlal Publishers, New Delhi 1981).

Page #55

A DISPUTE BETWEEN DINNĀGA, DHARMAKĪRTI AND AKALANKA 397
text as well as the English caya, Edition of Tibeing the Pratyakşa-pari
DNC = Mallavādin Ksamāśramana: Dvādasara-naya-cakra. Dvādaśāram Nayaca
kraṁ of Ācārya Sri Mallavādi Kşamāśramana. With the commentary Nyāyāgamānusārin of ācārya śri Simhasūri Gani Vādi Kşamāśramana. Ed. with critical notes by Muni Jambūvijayaji. Pt. I (1-4 Aras): Bhavnagar 1966 (reprinted: Bhāvnagar 2000); Pt. II (5-8 Aras): Bhavnagar 1976; Pt. III (9-12 Aras):
Bhavnagar 1988. DNCV = Simha-sūri: Nyāyāgamânusārini Dvādaśara-naya-cakra-vrtti. See: DNC. Dreyfus, Georges B.J. (1997). Recognizing Reality. Dharmakirti's Philosophy and Its
Tibetan Interpretations. State University of New York Press, Albany 1997. Franco, E. (1986). "Once again on Dharmakirti's Deviation from Dignāga on
Pratyakşābhāsa', Journal of Indian Philosophy 14 (1986) 79-97. Frauwallner, E. (1959). 'Dignāga, sein Werk und seine Entwicklung.' Wiener Zeit
schrift für die Kunde Süd- und Ostasiens 3 (1959) 83-164. Hattori, M. (1968). Dignāga, On Perception, being the Pratyakşa-pariccheda of
Dignāga's Pramāņa-samuccaya, Edition of Tibetan translations and the Sanskrit text as well as the English translation of the Chapter 1. Harvard University Press,
Cambridge, Massachusetts 1968. Hayes, R. P. (1988). Dignāga on the Interpretation of Signs. Studies of Classical India
9, Kluwer Academic Press, Dordrecht 1988. Herzberger, R. (1986). Bhartrhari and the Buddhists. Studies of Classical India 8, D.
Reidel Publishing Company (Kluwer Academic Publishers), Dordrecht 1986. LT = Akalanka: Laghiyas-traya. (1) Mahendra Kumār Nyāya Sastri (ed.): Nyāyakumuda-candra of Srimat Prabhācandrācārya. A Commentary on Bhattākalankadeva's Laghiyastraya. 2 Vols., Sri Garib Dass Oriental Series 121, Delhi 1991. (1. ed.: Bombay 1938–1942]. (2) Nyāyâcarya Mahendra Pandita Kumār Šāstri (ed.): Srimad-Bhattâkalanka-deva-viracitam Akalanka-grantha-trayam (Svôpajña-vivrti-sahitam Laghiyas-trayam, Nyāya-viniscayaḥ, Pramāņasangrahaś ca). Sarasvati Pustak Bhandār, Ahmadābād 1996. (1. ed.: Ahmedabad
Calcutta 1939). LTV = Akalanka: Laghiyas-traya-vivrti. See: LT. NA = Siddhasena Mahāmati: Nyāyāvatāra. See: BalCEROWICZ (2001a). NamS = Nardi-sutta / Namdi-sutta (Nandi-sūtra / Nandi-sūtra). Muni Punyavijaya,
Dalsukh Mālvaniā, Amritlal Mohanlal Bhojak (ed.): Nandi-sutta and Anuogaddarāim. Jaina-Agama-Series No.1, Shri Mahāvira Jaina Vidyālaya, Bombay
1968. NAV = Siddharși-ganin: Nyāyāvatāra-vivrti. See: Balcerowicz (2001a). NB = Dharmakirti: Nyāya-bindu. See: DhPr. NSā = Bhāsarvajña: Nyāya-sāra. See: NBhū. NBhū = Bhāsarvajña: Nyāya-bhūşaņa. Svāmi Yogindrānanda (ed.): Srimad-ācārya
Bhāsarvajña-pranitasya Nyāya-sārasya svopajñam vyākhyānam Nyāya-bhūşanam.
Sad-darśana-prakāśana-grantha-mālā 1, Vārāṇasi 1968. NM = Dignāga: Nyaya-mukha. Giuseppe Tucci: The Nyāya-mukha of Dignāga. The
oldest Buddhist Text on Logic, after Chinese and Tibetan Materials. Heidelberg
1930. NMa = Jayanta-bhatta: Nyāya-mañjari. Gaurinath Sastri (ed.): Nyāyamañjari of
Jayanta Bhatta with the Commentary of Granthi-bhanga by Cakradhara. Part 1-3, M.M. Sivakumāraśāstri-Granthamalā 5: Sampūrņānand Saṁskrta
Viśvavidyālaya, Vārāṇasi 1982, 1983, 1984. NVi = Akalanka Bhatta: Nyāya-viniscaya. (1) Nyāyâcarya Mahendra Pandita
Kumār Sastri (ed.): Srimad-Bhattâkalanka-deva-viracitam Akalanka-grantha-trayam (Svôpajña-vivrti-sahitam Laghiyas-trayam, Nyāya-viniscayah, Pramāṇasangrahaś ca] Sarasvati Pustak Bhandār, Ahmadābād 1996. [1. ed.: Ahmedabad-Calcutta 1939). (2) Mahendra Kumar Jain (ed.): Nyāyaviniscaya

Page #56

398
PIOTR BALCEROWICZ
nstitute, Patacandra. See LT" Sanghavi, Mahen
vivarana of Sri Vadirāja Sūri, the Sanskrit Commentary on Bhatta Akalankadeva's Nyāyaviniścaya. Vol. 1 & 2, Bhāratiya Jñānapitha Prakāśana, Vārānasi 1949, 1955
[2nd edition: Bhāratiya Jñānapiha Prakāśana, Delhi 2000). NViV = Vädirāja-sūri: Nyāya-viniscaya-vivarana. See: NViz. PBh = Prasastapäda: Praśastapäda-bhāsya (Padârtha-dharma-sangraha). Johannes
Bronkhorst, Yves Ramseier (ed.): Word index to the Prasastapādabhāşya: a complete word index to the printed editions of the Praśastapadabhāsya. Motilal
Banarsidass, Delhi 1994. RANDLE 1926 = Randle, H. N.: Fragments from Dignāga. Prize Publication Fund,
Royal Asiatic Society, London 1926 [reprinted: Motilal Banarsidass, Delhi 1981). DhPr = Durveka Mišra: Dharmottara-pradipa. Pt. Dalsukhbhai Malvania (ed.):
Pandita Durveka Miśra's Dharmottara-pradipa [being a subcommentary on Dharmottara's Nyāya-bindu-ţikā, a commentary on Dharmakirti's Nyaya-bindu). Kashi
Prasad Jayaswal Research Institute, Patna 1955 (reprinted: 1971). NKC = Prabhācandra-sūri: Nyāya-kumuda-candra. See: LT. PMi = Hemacandra-sūri: Pramāņa-mimmsā. (1) Sukhlalji Sanghavi, Mahendrà
Kumar and Dalsukh Malvaniya (ed.): Kavikālasarvajna-Sri-Hemacandrâcārya-viracitā svopajña-vrtti-sahitā Pramāna Mimāṁsā (with Bhāṣā Tippana of Pandita Sukhlalji Sanghvi]. SPBh 1998. (2) Satkari Mookerjee and Nathmal Tatia (ed.): Hemacandra's Pramāna-mimāṁsā. Text and Translation with Critical Notes. Tara
Publications, Varanasi 1970. PS = Dinnāga: Pramāna-samuccaya. (1) See: Hattori (1968). (2) Hidenori Kitaga
wa (ed.): Indo koten ronrigaku no kenkyū: Jinna (Dignāga) no taikei. Suzuki Gakujutsu Zaidan, Tokyo 1965. (3) In: DNC (Muni Jambūvijayaji), Pt. I, pp. 97
134. PSa = Akalanka: Pramāna-sangraha. See: LT2 (pp. 97-127). PSaV = Akalanka: Pramāna-sangraha-vivrti. See: LT2 (pp. 97–127). PSV = Dinnāga: Pramāna-samuccaya-vrtti. See: PS. PV = Dharmakirti: Pramāna-vārttika. Rāhula Sankrtyāyana (ed.): Pramāņavārtti
kam Acārya-Manorathanandi-krtayā vrttyā samvalitam [Dharmakirti's Pramāņavārttikam with a commentary by Manorathanandin). Parts I-II, Appendix
to Journal of the Bihar and Orissa Research Society 24-26, Patna 1938-1940. PVA = Prajñākaragupta: Pramāna-vārttikalankāra. Rāhula Sankrtyāyana (ed.):
Pramānavārttikabhäsyam or Vārtikālańkārah of Prajnakaragupta. (Being a commentary on Dharmakirti's Pramānavārtikam). Kashi Prasad Jayaswal Research
Institute, Patna 1953. PVin II = Dharmakirti: Pramāna-viniscaya, chapter II (svârthânumāna). Ernst
Steinkellner: Dharmakirti's Pramānaviniscayah. 2. Kapitel: Svarthanumānam. (1) Teil 1. Tibetischer Text und Sanskrittexte. Veröffentlichungen der Kommission für Sprachen und Kulturen Südasiens, Osterreichische Akademie der Wissenschaften, Heft 12, Wien 1973. (2) Teil II. Ubersetzt von E. Steinkellner. Veröffentlichungen der Kommission für Sprachen und Kulturen Südasiens, Osterreichische Akademie
der Wissenschaften, Heft 15, Wien 1979. SBh = Sabarasvāmin: Mimāṁsā-śābara-bhāsya. Yudhisthira Mimāmsaka (ed.): Mimāṁsā-sabara-bhāsyam. Vols. 1-6, Sri Santisvarüp Kapür, Rāmalāl Kapūr Trust Press, Bahālgadh Sonipat-Haryāṇā l: 1987, II: 1990, III: 1980, IV: 1985, V: 1986, VI: 1990. Singh, A. (1984). The Heart of Buddhist Philosophy-Dinnāga and Dharmakirti.
Munishiram Manoharlal Publishers, New Delhi 1984. PNTAA = Vädideva-sūri: Pramāna-naya-tatt välokalankāra. See: SVR. RVār = Akalanka: Tattvartha-vārttika (Rāja-vārttika). Mahendra Kumar Jain (ed.):
Tattvārtha-vārttika (Raja-vārttika] of Sri Akalankadeva. Edited with Hindi Translation, Introduction, appendices, variant readings, comparative notes etc.
.
am "ramana-vârtiarch Society andin). Parts

Page #57

A DISPUTE BETWEEN DINNĀGA, DHARMAKĪRTI AND AKALANKA 399
Parts I-II. First edition, Jñānapitha Mūrtidevi Jaina Grantha-mālā 10, 20 [Sanskrit Grantha), Bhāratiya Jñānapiha Prakāśana, Delhi 1953-1957. [2. ed.: Delhi 1982). SSi = Pūjyapāda Devanandin: Sarvârtha-siddhi. Phoolchandra Shastri (ed.): ācārya
Pujyapāda's Sarvarthasiddhi (The commentary on Acārya Griddhapiccha's Tattvārtha-sūtra). Edited and translated into Hindi]. Varanasi 1934 (reprinted: Jñānapitha Mūrtidevi Jaina Grantha-mālā 13, Bharatiya Jñanapiha Prakaana, Delhi 2000). SVR = Vādideva-sūri: Syād-vāda-ratnâkara. Motilal Lālāji (ed.): Srimad-Vadide
vasūri-viracitah Pramāna-naya-tattvalokālankārah tad-vyākhyā ca Syād-vādaratnâkarah. 5 Vols., Poona 1926–1930 (reprinted: 2 Vols., Bhāratiya Buk Kārporeśan, Dilli (Delhi) 1988). SVi = Akalanka Bhatta: Siddhi-viniścaya. Mahendrakumār Jain (ed.): Siddhi-vi
niscaya of Akalanka edited with the commentary Siddhi-viniscaya-ţikā of Anan
tavirya. 2 Vols. BhJP, V 1959. SVIT = Anantavirya: Siddhi-viniścaya-tikā. See: SVi. SVIV = Akalanka Bhatta: Siddhi-viniscaya-vivrti. See: SVI. STP = Siddhasena Diväkara: Sammati-tarka-prakarana. (1) Sukhlal Sanghavi;
Becardās Dośi (ed.): Sammatitarka-prakaranam by Siddhasena Divākara with Abhayadevasüri's Tattva-bodha-vidhāyini. Gujarat-purā-tattva-mandir-granthāvali 10, 16, 18, 19, 21, Gujarāt-purā-tattva-mandir, Amdāvād 1924-1931 [reprinted: 2
Vols., Rinsen Buddhist Text Series VI-1, 2; Kyoto 1984). TBh = Umāsvāti: Tattvârthâdhigama-bhāsya. See: TS and TT. TBV = Abhayadeva-sūri: Tattva-bodha-vidhāyini. See: STP. Thān = Thānamga-sutta (Sthānanga-sūtra). Muni Jambuvijaya (ed.): Thānaṁgasu
ttam and Samavāyargasuttam. Jaina-Agama-Series 3, Shri Mahāvira Jaina
Vidyalaya, Bombay 1985. TS = Umāsvāmin: Tattvartha-sūtra. M.K. Mody (ed.): Tattvārthādhigama by
Umāsvāti being in the Original Sanskrit with the Bhasya by the author himself.
Bibliotheca Indica 1044, 1079, 1118, Calcutta 1903, 1904, 1905. TŚVA = Vidyānanda Pātrakesarisvāmin: Tattvārtha-śloka-vārtikālařkara. Edited by Manoharalāl, Ramacandra Nātha Rangaji, Nirņāya Sāgara Press, Bombay
1918. TT = Siddhasena-ganin: Tattvarthâdhigama-tikā. Hirālāl Rasikdās Kapadia (ed.):
Tattvarthādhigamasūtra ( A Treatise on the Fundamental Principles of Jainism) by His Holiness Sri Umāsvāti Vāchaka, together with His connective verses commented upon by śri Devaguptasūri & Sri Siddhasenagani and His own gloss elucidated by śri Siddhasenagani, edited with Introduction in English and Sanskrit. 2 Vols., Sheth Devchand Lalbhai Jain Pustakoddhar Fund Series Nos. 67 and 76, Bombay 1926
(Part I: Chapters I-V), 1930 (Part II: Chapters VI-X). Vetter, T. (1964). Erkenntnisprobleme bei Dharmakirti, Veröffentlichungen der Kommission für Sprachen und Kulturen Süd- und Ostasiens Heft 1, Osterreichi
sche Akademie der Wissenschaften, Wien 1964. Viy = Bechardas J. Doshi; Amritlal Mohanlal Bhojak (ed.): Viyāhapannattisuttam. 3
parts, Jaina-Agama-Series No. 4 Part 1/II, Shri Mahāvira Jaina Vidyālaya,
Bombay 1974–1982. VP = Bharthari: Vāk ya-padiya. Wilhelm Rau: Bhartrharis Vākyapadiya. Die
Mülakārikās nach den Handschriften herausgegeben und mit einem Pāda-Index versehen. Abhandlungen für die Kunde des Morgenlandes XLII, 4. Franz Steiner Verlag, Wiesbaden 1977.

Page #58
400 PIOTR BALCEROWICZ VSV = Candrananda: Vaisesika-sutra-vrtti. Muni Jambuvijaya (ed.): Vaisesikasutra of Kanada with the Commentary of Candrananda. With the Introduction by Anantalal Thakur. Gaekwad's Oriental Series 136, Oriental Institute, Baroda 1961. Wagle, N.K.; Qvarnstrom, O. (1999). (ed.): Approaches to Jaina Studies: Philosophy, Logic, Rituals and Symbols. [Proceedings of the International Conference on Approaches to Jaina Studies: Philosophy, Logic, Rituals and Symbols, 31.03 2.04.1995 Torontol, Toronto 1999. Wayman, A. (1978). "A Reconsideration of Dharmakirti's "Deviation from Dignaga on Pratyaksabhasa,' Annals of the Bhandarkar Oriental Research Insti tute, Diamond Jubilee Volume (1978) 387-396. YS = Patanjali: Yoga-sutra. Polakam Sri Rama Sastri and S.R. Krishnamurthi Sastri (ed.): Patanjala-yogasutra-bhasya Vivaranam of Sankara-bhagavatpada. Government Orientla Manuscripts Library, Madras 1952. Rothenbaumchaussee 34 D-20148 Hamburg Germany