1

To be published in:Companion to Classical Indian Philosophy.Edited by Matthew Kapstein,Cambridge University Press, [in press]

Siddhasena Mahâmati and Akalaóka Bhaþþa: A revolution in Jaina epistemology

Piotr Balcerowicz

1. Introduction

Two Jaina contemporaries, a Œvetâmbara philosopher Siddhasena Mahâmati (circa 710/720–770/780) and his younger contemporary Digambara thinker Akalaóka Bhaþþa (c. 720–780)1,are regarded by their respective sects as the fathers of Jaina logic. Even though the claimsmight be slightly exaggerated, both of them indeed revolutionised Jaina epistemology, byradically transforming basic epistemological concepts, which had been based on Canonicaltradition.The former is credited with composing the succinct Nyâyâvatâra (Introduction to Logic;hereafter NA), and is generally confused with a pre-Diónâga thinker Siddhasena Divâkara (c.450–500), the author of the Saômati-tarka-prakaraòa (A Treatise on Correct Reasoning).2

Despite his importance, we know basically nothing about Siddhasena Mahâmati, not even hiscorrect full name. His seminal work belonging to the category of works in 32 verses(dvâtriôœikâ) was commented by a few authors. Siddharši-gaòin completed his commentaryNyâyâvatâra-vivåtti on 21 May 906, and Devabhadra-sûri supplied his gloss Nyâyâvatâra-þippana in the second half of the twelfth century. Œânti-sûri composed his versed commentaryNyâyâvatâra-vârttika-våtti in eleventh century.Akalaóka Bhaþþa was one of most prolific writers among the Digambaras, and his oeuvreranges from commentaries, such as the Royal Scholium (Râja-vârttika) on Umâsvâmin’s Tracton Truth (Tattvârtha-sûtra, TS) and the Work in 800 Statements (Ašþa-œatî) onSamantabhadra’s Examination of Authority (Âpta-mîmâôsâ, ÂMî), to independent and highlydemanding works on epistemology, such as Three Simple Chapters (Laghîyas-traya), namelyProlegomena to cognitive criteria (Pramâòa-praveœa), Prolegomena to viewpoints (Naya-praveœa) and Prolegomena to the Doctrine (Pravacana-praveœa), An Enquiry intoEpistemology (Nyâya-viniœcaya, NVi), An Enquiry into Conclusive Proofs (Siddhi-viniœcaya),A Compendium of Cognitive Criteria (Pramâòa-saógraha, PSa), The lustre of the Jewel ofCognitive Criteria (Pramâòa-ratna-pradîpa), A Brief Exposition of Epistemology (Nyâya-cûlikâ)3 and The definition of cognitive criterion (Pramâòa-lakšaòa)4.

The present essay will, in the first section, focus on the nature of the epistemological shiftswhich both thinkers introduced independently of each other and on the question how and towhat degree Siddhasena’s and Akalaóka’s ideas were partly anchored in Buddhist concepts ofthe pram¹òa school of Diónâga and Dharmakîrti. The second section demonstrate howAkalaóka applies the novel epistemological ideas to construct realistic ontology andobjectivity of both the external world and the cognising subject vis-à-vis Buddhist idealism.

1 On the dating of Siddhasena Mahâmati see BALCEROWICZ (2003a) and BALCEROWICZ (2008: i ff.).2 On the relationship between Siddhasena Mahâmati and Siddhasena Divâkara see BALCEROWICZ

2000, 2001, 2003, 2008: i–ii, v–xli.3 No manuscripts of Pramâòa-ratna-pradîpa and Nyâya-cûlikâ have been discovered so far.4 The work has never been published so far although there are some extant manuscripts of it.

2

To understand the significance of the shifts and developments that took place in Jainaepistemology and logic in 7th and 8th century we should first take a closer look at thetraditional framework of the system prior to the age of change.

2. Traditional epistemology

As it was the case with the majority of philosophers and philosophical schools in India, twomost important issues that concerned Jaina thinkers were a profound soteriological interest,the central idea of which was liberation (mokša), and the search for solid foundations of ourknowledge, on which the belief in liberation and a path leading to it rest.It is very unlikely, despite popular claims of its followers, that the founder-cum-reformer ofJainism Mahâvîra Vardhamâna (Prakrit: Mahâvîra Vaddhamâòa / Vaððhamâòa, 5th centuryBCE5) or his predecessor Pârœva / Pârœvanâtha (Prakrit: Pâsa / Pâsanâtha6), who wereprimarily ascetics preoccupied with moral and soteriological issues, developed any significantinterest in epistemological questions. At an early stage, there were rather unsophisticatedcosmological and ontological beliefs, e.g. the belief in six classes of beings (jîva-nikâya) andsimple concept of deeds related to material consequences (that later developed into a full-fledged concept of karman), that formed the theoretical background for ethics, based on theinsistence on one’s withdrawal from any dealings with the surrounding world andabandonment of any action (akarman).It was probably a few centuries later that epistemological questions started to preoccupy theminds of Jaina thinkers and that happened, in the first place, in confrontation with rivalreligious and philosophical traditions.We may safely assume that, against the popular claims of modern Jainas, such theories ofrelativity, widely associated with Jainism, especially a comprehensive doctrine of multiplexityof reality (anekânta-vâda), were developed much later. The roots of this well-known theorylie in the felt need, first, to fine-tune early Jaina ethical theory, predominantly geared to curbany kind of violence (ârambha, vaira, hiôsâ) against any living being, by determining what aliving being is, and in what sense, as well as what constitutes violence against it, and,secondly, to adjust to the demands of soteriology and to the needs monastic discipline andeveryday life.One of the earliest epistemological ‘analytical’ devices of this sort to be applied to ethics, wasthe conviction, borrowed from the Buddhist probably in the period between 3rd and 2ndcenturies CE, that deeds can be performed in a triple way: by means of the mind (manas),speech (vâc) and body (kâya).A gradual development of ontology after 1st century BCE / 1st century CE, especiallydetailed analysis of the ontological distinction into living beings (jîva) and lifeless elements(ajîva) along with elaborate classifications of both the categories necessitated a developmentof a relevant theoretical analytical devices. After this period, still during the so-called‘Canonical Period’ (from 4th/3rd century BCE till 450–480 CE) as reflected in the JainaCanon eventually codified in the second half of 6th century, complex dialectical ways ofanalysis (anuyoga-dvâra) were developed (that culminated in an intricate theory before5th/6th century), and crucial theoretical determinants were the four standpoints (nikšepa,nyâsa)7 as tools of analysis: substance (dravya), place (kšetra), time (kâla) and actual

5 Various branches of the Jainas date his parinîrvâòa variously to 525 BCE (most poplular), 527/526BCE, 510 BCE, whereas Western scholars assigned him traditionally to the period of 540–478 BCE.Resent research provides strong reasons to believe that he dider some time in the period between427/425 and 395 BC.6 According to Jaina legendary accounts he preceded Mahâvîra by 250 years, viz. was born around 877BCE; for various reasons this tradition seems highly unlikely. Most plausibly Pârœva lived in 6thcentury BCE, i.e. approx. 100 years before Mahâvîra.7 See ALSDORF (1973) and BHATT (1974).

3

condition (bhâva) of an entity analysed. Occasionally, other parameters were added incanonical literature, such as a particular quality (guòa), a mode (paryaya), spatial extension(pradeœa), name (nâma), form (rûpa), material representation (sthâpanâ), transformation(pariòâma) etc.8 Also in the post-canonical literature the authors such as Siddhasena Divâkara’sSaômati-tarka-prakaraòa (STP 59–60, c. 450-500 CE) enumerated similar parameters thatserved the same purpose: substance (dravya), place (kšetra), time (kâla), actual condition(bhâva), mode (paryâya), aspect or part (deœa) and combination of elements, or the whole(saôyoga).These helped determine in what sense, for instance, a particular entity can be considered aliving being or a lifeless element, or in what sense or aspect a particular living being can beinjured, in what sense a particular category of beings is infinite in number or innumerable, inwhat sense, aspect or region the universe is inhabited and by what class of living beings, inwhat sense a living being (say, as a soul or as a material body, or as a composite of soul andbody) is eternal or perishable, endowed with perfect cognition or deficient in its cognition,suffering or insentient, imprisoned in the mundane world of transmigration (saôsâra) orliberated, in what sense or at what stage of development a living being is possessed ofsensuous cognition (mati-jñâna) or omniscience (kevala, the ultimate ideal for the Jainas), etc.All these considerations have ethical and soteriological implications. An example of suchdeliberations are the following passage of the canonical Exposition of Instructions (Viyâha-pannatti):

‘[The senior disciple Indrabhûti Gautama asks Mahâvîra:] We know and we seethat living beings that have two senses, three senses, four senses and five sensesbreathe in and our, they inhale and exhale. But we do not know and do not seethat beings whose bodies are made of earth, water, fire, air or plants breathe inand our, they inhale and exhale. O Lord, do they breathe in and our, they inhaleand exhale?9—Oh Gautama, all these living beings do breathe in and our, they doinhale and exhale. …They breathe in and our, they inhale and exhale: from the standpoint ofsubstance, substances of infinite units of space; from the standpoint of place,substances immersed in innumerable units of space; from the standpoint of time,substances that occur in any situation; from the standpoint of actual condition,substances that are endowed with colour, smell, taste and touch…’ (Viy1 2.1.3–7,Viy2, p. 145–147).

‘Oh, Skadaka, since you have doubt in your mind … whether souls are with limitor without limit, this is the state of affairs (artha): from the standpoint ofsubstance, souls are singular and with limit; from the standpoint of place, soulsoccupy innumerable units of space and are immersed in innumerable units ofspace, however they are with limit; from the standpoint of actual time, there wasno time, nor there will be any time that souls would not exist, souls are eternaland have no limit; from the standpoint of actual condition, souls are possessed oflimitless cognition, of limitless conation and of limitless good conduct, arelimitlessly neither heavy nor light (i.e. have no mass) and have no limit.’ (Viy1

2.1.23[2], Viy2, p. 161).

The four standpoints (nikšepa, nyâsa) served for a Jaina dialectician as important parametersthat helped analyse in which way a particular entity is relevant to morality and one’s conduct,

8 Cf. the general catalogue of such parameters in BHATT (1974: 15–32).9 The question implies an important doubt of moral relevance whether such entities as the fourelements and plants, often considered insentient, are alive or not, ergo whether they can meaningfullybe objects of human moral actions that incur sin, i.e. whether they can be harmed or injured.

4

through which one should refrain from any kind of violence, as well as to metaphysicalquestions pertinent to the attainment of liberation and omniscience, as its corollary.The four basic standpoints (substance, place, time and condition) underwent a gradual two-fold development. On the one hand, in the final phase of the canonical period (c. 5th–6thcenturies), were in extenso incorporated into the method of the seven-fold modal description(sapta-bhaógî, syâd-vâda) as a set of parameters under which a truth-value of a sentencecould be assessed.On the other hand, they were significantly remodelled and came to stand for an independentmethod of the four standpoints (nikšepa-vâda, nyâsa-vâda), codified also in extra-canonicalliterature, especially in Umâsvâmin’s Tattvârtha-sûtra (TS, ca. 350–400 CE).

Other developments in the early phase of Jaina epistemology concerned the classification ofcognitions, the nature of cognition as well as the sources and fundaments of our cognition.10

It seems that the earliest development was an attempt to clasify all cognitions into five types:sensuous cognition, testimonial cognition, or knowledge based on testimony, clairvoyance, ortelesthesia, mind-reading, or telepathy, and absolute knowledge, or omniscience (Model I).These varieties may have been accepted in Jainism from a very early stage and theclassification is present at an relatively early canonical stage. The background for theclassification was not epistemological but religious and doctrinal and therefore it lackedconsistence.

The first of these, sensuous cognition (abhinibodha, mati; lit. ‘apprehension’, ‘mentalprocess’), was initially understood as an ordinary perception, common to all living beings. Inthe course of time, and under the influence of the Nyâya school11, but also to differentiate itfrom the second type, the Jainas tried to clearly distinguish it from other kinds of cognitionand eventually defined it as a cognition that originates in the contact of a sense organ, themind and an object. In this way the ordinary, mundane cognition came to be more or lessequated with sensory perception of other schools in a broad sense. It was, however, notrestricted to cognition derived from perceiving any sensory data but also comprised mentalinsight, or inner apprehension of mental images, memories etc. It covered not only sensationsand awareness of such sensations derived from particular sense organs (including the organ ofhearing) but also mental processes triggered by such sensations, including processes of simplereasoning, sensation of doubt, acts of deliberation and cogitation, recognition etc. Basically, itembraced all that was not included in all the remaining four kinds of cognition! Such aunspecific understanding of sensuous cognition, in which sensations, perceptions and mentalprocesses overlap, were the nucleus of later contradictions in Jaina epistemology whichtroubled some philosophers.The second type of testimonial cognition (œruta; lit. ‘the heard’) covered all cognition that wasnot based on direct experience of the cognising subject but on verbal testimony of anotherperson. In the first place, it connoted all cognition derived from scriptures or orallytransmitted scriptural tradition, i.e. imparted by a religious authority. Thus, testimonialcognition was essentially any kind of cognition revealed indirectly by means of speech, and assuch it was later said (TS 1.20) to be preceded by sensuous cognition, inasmuch as the sense

10 For a useful sketch of these developments see SHASTRI (1990: 196–213, ‘Divisions of Knowledge’);he generally follows the historical stages worked out in SUKHLÂL (1942: 5 ff.). See also the typologyin BALCEROWICZ (2003a: 45–52).11 Esp. by the definition of perception expressed in NS 1.1.4: ‘The cognition which is produced by thecontact of an object with a corresponding sense organ, which is non-verbal, which is unerring, thenature of which is the decision regarding the nature of the object, is perception.’ This influence isclearly reflected in Umâsvâmin’s TS 1.14: (‘Sensuous cognition is occasioned by sense organs and bythe mind’) as well as in Umâsvâti’s TBh 1.11 (‘Sensuous cognition is occasioned by the contact ofsense organs with their respective objects’) and TBh 1.14.

5

of hearing and sensory perception (comprised under sensuous cognition), but later also visualdata in the form of script (text) perceived through ocular sensation, were a prerequisite to it.Since testimonial cognition was genetically related to sensuous cognition, but specificallyreferred to acts of cognition based on the use of language (either in verbal or, rarely written,form) by isolating those of potentially religious importance, its difference from sensuouscognition lay not epistemological considerations but in practical application of such aknowledge to moral conduct as a morally uplifting factor.These first two kinds of cognition were initially distinguished mainly because of the role theyplayed in religious practice: it was testimonial cognition, especially in the form of imparteddoctrine and rules of conduct, that had any religious relevance and could provide devotionalincentive or moral motivation for the pious believer. Sensuous cognition was, on the otherhand, rather morally neutral and merely reflected the surrounding world. Thus, the primaryground for the distinction between these two was that of religious relevance. It was only laterthat another important and conspicuous difference surfaced, directly related toepistemological questions, which emphasised what could be compared to ‘knowledge byacquaintance’ and ‘knowledge by description’ respectively, a distinction introduced byBertrand RUSSELL (1912) and (1917). As in the case of the first type, also testimonialcognition was believed to be common to all living beings, but in practice a category of mostprimitive living beings, i.e. those possessed of only one sense, namely of the tactile sensefaculty (ekêndriya), also called immovable beings (sthâvara), were barred from it.A reflection of the logical primacy, with respect to religious practice and soteriological goals,of scriptural cognition over sensuous cognition was the first place in classifications scripturalcognition occasionally enjoyed, whereas sensuous cognition was mentioned second (e.g. Uttar28.4).The third variety, clairvoyance (avadhi; lit. ‘mental infiltration’), opened a realm of thesuprasensory. Through it, it was possible to grasps objects that were considered physicallybeyond reach of ordinary cognition of two varieties (sensuous or testimonial cognition). Therewere some restrictions imposed on it. First, only macroscopic objects at a distance possessedof physical form (rûpin) were amenable to it. Although it was usually not stated explicitly, wemay infer that excluded were objects that were considered invisible because of their subtlety(sûkšma), such as atoms, or because of temporal distance (things past and future). There wastendency later on to include possibly all physical objects classified as invisible (adrœya).12

Secondly, clairvoyance was not accessible to everyone: some beings, such as divine beingsand denizens of hells, were possessed of it by virtue of their birth; others—and here not onlyascetics but even animals were included—were supposed to be able to acquire it throughspecial practices. The category of clairvoyance certainly was, on the one hand, an attempt toaccomodate a popular belief that that some individuals (humans, ghosts, divine beings etc.)can see things ordinary humans cannot. On the other hand, however, this variety was soongiven an additional significance: along with the fourth variety, mind-reading, became a a rungin a ‘ladder’ of more and more sublime congition that culminated in perfection, i.e. absolutecognition.Telepathy, or mind-reading (manaÿ-paryâya or manaÿ-paryaya; lit. ‘penetration of mind’)was the most sophisticated kind of cognition out of all ‘mundane’ varieties. One was believedto directly observe other persons’ thoughts, i.e. all contents of other minds. However, onlymorally advanced humans, because of their spiritual development and virtues, could acquire itas a ‘side effect’ of their moral excellence and advancement on spiritual path to perfection.The background that provided justification to the belief in telepathy was an almost pan-Indianconviction that thoughts and all mental phenomena are somatic:13 mental phenomena are

12 On the categories of what was considered in India invisible and on the idea of (im)perception of theinvisible (adåœyânupalabdhi) see KELLNER (1999), (2003) and BALCEROWICZ (2005a).13 On the concept of somatism in traditional Indian philosophy and psychology see SCHAYER (1936),DANDEKAR (1941) and KUNST (1968).

6

spatially located (in our minds as physical organs) and have their extension. That is why wecould be justified in considering telepathy a very subtle, sophisticated kind of perception, thepercepts of which are extremely fine. For all imperfect beings trapped in transmigration it wasalso a closest mundane approximation to the ideal of omniscience. Again, the category oftelepathy should not be considered as a result of a systematically and logically worked-outsystem of epistemology but rather an offshoot of a popular belief that some individuals, e.g.ascetics, adepts of yoga etc., have special mental powers and can read people’s minds.The culmination of the whole epistemological edifice, the apex of all spiritual and cognitivedevelopment and the fulfilment of the dreams of the imperfect man, anguished by dread ofnihility, ceaseless suffering in the saôsâra and unpredictability of the world, was absolute,perfect knowledge, or omniscience (kevala; lit. ‘the singular one’). The Jaina concept ofomniscience was, it seems, most flamboyant and uncompromising of its kind ever conceivedin India. It had not spatial, temporal or structural limitations and nothing was hidden from it.Omniscience, as understood by the Jainas, was not some kind of potentiality available to theadept in a situation in need (as in Buddhism): the omniscient perfected being (kevalin)literally comprehended everything at once, all things past, present and future, distant andsublime, material and mental. There was nothing in the whole universe (loka) which was notcovered by it; moreover, it reaches even beyond confines of the universe and penetratesinfinite ‘non-world’ (aloka), filled only with space (âkâœa). The attainment of this perfect,absolute knowledge was tantamount to liberation, the consummation of mundane existence ofa soul transmigrating from times immemorial, and was conditioned on extremely austereascetic practices, complete annihilation of all activities and spiritual growth. In the course oftime, as the doctrine of karman, ultimately conceived by the Jainas both as ethically boundactions as well as subtle matter that, being triggered by them, envelops the soul, grew into anintricately elaborated system, detailed explanations were given how the destruction of karmicmatter is related to the attainment of omniscience (e.g. TS 10).Albeit its elements, i.e. particular varieties of cognitions, were discussed from the outset, it israther unlikely that the fivefold classification as such was developed much earlier thanapprox. 1st century BCE, i.e. before Indian philosophers began to classify cognitions in asystematic manner. Certainly the belief in omniscience, as opposed to all mundane varieties ofcognition, was a defining characteristic of the earliest Jaina doctrine. These classification inits fivefold divisions shows no external influence as such; however particular solutions anddefinitions of respective cognitions did develop under the influence of other philosophicalsystems and religions, especially that of Nyâya-Vaiœešika. Certainly a clearly palpableinfluence of Nyâya-Vaiœešika and a reflection of the discussion on the nature of perceptionwas a slight modification of Model I by adding a list of four stages of sensuouscognition: sensation (avagraha), i.e. the first contact of a sense organ and the object,cogitation (îhâ), speculation as regards the character of the sensory data and an internalprocess of its analysis, (c) perceptual judgement (avâya—Digambaras, apâya—Œvetâmbaras,apanoda), which removes doubt as regards the character of the sensory data and determinesthe source of the sensation, (d) retention (dhâraòâ) of the judgement, which is ultimately asource for future reminiscence and memory. The result was II.

Independently of this fivefold scheme, and at a slightly later stage, the Jainas developedanother classification of cognitions into two varieties: direct cognition (pratyakša) andindirect cognition (parokša) (Model III). This bipolar picture reveals external influence. In thefirst place it goes back to two aspects of the world as reflected in one’s cognition and acts aswell as in sources of one’s cognition, well-known from the Brâhmaòas: the direct, immediate(pratyakša) and the indirect, distant (parokša). What was directly in front of our eyes wascalled pratyakša, whereas the realm of the divine, beyond our eyes, was parokša. The divisioninto direct (pratyakša) and indirect (parokša) is found in late Vedic literature and is alsoreflected by Yaska (Nir 7.1, p. 715) in the division of the genesis of Vedic hymns (åc).Unsurprisingly, the terms describing the two realms of reality were transferred onto cognitive

7

procedures and cognitions that grasped their respective spheres. This is reflected as early as inthe earliest Upanišads (BÂU 3.4–5) and in the classification of cognitive procedures found inKauþilya’s Artha-œâstra (AŒ 1.9.4–7, 1.15.19) and replicated in the Mahâ-bhârata (MBh12.57.26, 12.59.68). What we observe in Jainism is a radical semantic shift: pratyakša cameto denote direct cognition of the soul, unmediated by any organ, whereas parokša referred tocognitive acts by means of sense organs and/or the mind which served as instruments ofcognition for the soul.One might ask what prompted this reversal of the meanings of pratyakša–parokša in earlyJainism. For Jaina ethics and soteriology the pivotal idea was that of the loving being (jîva),or soul (âtman): all actions were ultimately performed by it and the mind, speech and bodywere treated only as instruments. It seems that it was the transference of the dominant role ofthe soul over the organic instruments of its actions that resulted in the central role inepistemology assigned to the soul as ‘the perceiving organ’. In later times, the soul was evencalled akša (‘the eye’) is such contexts.Thus, for early Jainas their epistemology was determined by ethical system and moralconsiderations. This radical semantic shift was not as arbitrary and against pan-Indiantradition as one might think.Reflecting a tradition that goes back to Vedic times, and which also treated the soul as theultimate cognitive subject, Pâòini (c. 6th–5th c. BCE) tried to explain the meaning of the termindriya (‘sense organ’) as derived from the term indra (also the name of the god Indra) in thesense of the soul / th eself: ‘Optionally, the term “the perceiving faculty” (indriya) denotes“an attribute of the self (indra)14”, “perceived by the self (indra)”, “created by the self(indra)”, “welcomed by the self (indra)” and “given by the self (indra)”’ (A 5.2.93). Patañjali(2nd c. BCE) reiterated that ‘indirect (parokša) is what is beyond the reach of the perceivingorgan’, but added that ‘the perceiving organ (lit. ‘eye’) is that through which something isattained’ (MBhâ ad A 3.2.115, Vol. III, p. 189). This explanation deliberately left room forambiguity: either ‘the perceiving organ’ could be interpreted as a sense organ, i.e. aninstrument by means of which the agent attains the desired object, or the subject of the actionwho ultimately attains it. Such and similar considerations, reflected also in early grammaticalliterature, may have additionally influenced Jaina thinkers to treat the soul as akša and thebasis for the interpretation of what consists pratyakša and parokša.These two models, the fivefold scheme and the twofold classification, began to merge at somepoint (Model IV), probably around 2 century CE.

Another scheme to classify epistemically relevant mental phenomena was a scheme of schemeof two kinds of cognitive application (upayoga), or two kinds of cognitive faculties of thesoul. Traditionally the cognitive applications bifurcated into cognition (jñâna) and perceptualexperience (darœana) (Model V), which already brought the seed of contradiction into theclassification of cognitions. The difference between the two lay in cognition being distinct,i.e. having definite contents (sâkâra), and in perceptual experience being indistinct, nothaving any definite contents (anâkara, nirâkâra). Sometimes (STP 2.11) the division, whichapparently did not seem to completely unambiguous, the former was defined as manifest(vyakta) and the latter as not manifest (avyakta). This brought, however, inconsistence into theclassification models: perceptual experience (darœana), which was said to be directly relatedto senses, would overlap with some initial stages of sensuous cognition (mati-jñâna), whichrepresented both a different (fivefold) model and was the other category of the cognitiveapplications (upayoga).The idea of upayoga understood as cognitive application of the soul was developed at aslightly later stage than the classification of parokša–pratyakša.15 The scheme of cognitiveapplications was initially independent of the (fivefold or twofold) classification of cognitions. 14 On the equivalence of indra and the self (âtman) see VATSYAYAN (1999).15 It is basically impossible to offer any objective dating here, except for relative chronology at thisstage of research.

8

That is why the upayogas were, at an early stage, discussed independently of the standardfivefold classification of cognitions.16 However, the notion was again probably rooted inmuch earlier concepts: interestingly, upayoga is traditionally defined as the nature, or definingcharacteristic (lakšaòa) of the soul (Uttar 28.10–11, TS 2.8). These early descriptions mentionthe upayoga in the context of the path to liberation (mokša-marga), which consisted ofcognition (jñâna), conation (darœana), or belief, i.e. the proper ethical approach to life andworldview, conduct (câritra) and ascetic practice (tapas) (e.g. Uttar 28.28.2). There weresubsequently (e.g. TS 1.1) reduced to the first three elements, by the addition to each of themthe attribute ‘correct’ (samyañc, viz. correct cognition etc.) and by reducing acetic practice toa sub-category of correct conduct. It is probable that at the earliest stage the term upayoga,mentioned in the context of the path to liberation, referred to the practical capability of thesoul to apply all its innate endowments that were relevant to achieve the ultimate goal ofexistence. A canonical reference to upayoga mentions that it equally relates to happiness andsuffering: ‘The soul has as its defining characteristic the application as regards cognition(jñâna), conation / perceptual experience (darœana), happiness (sukha) and suffering(duÿkha)’ (Uttar 28.10). Thus, at the earliest stage, if there was any twofold division ofjñâna–darœana, it rather concerned not the distinction cognition–perceptual experience (as thelater tradition takes it), but rather the distinction cognition–conation (belief, world view).Historically the list of soul the characteristics that define the soul’s nature was extended andother elements were included: ‘Cognition (jñâna), conation / belief (darœana) conduct(câritra), ascetic practice (tapas), innate energy (vîrya) and cognitive application (upayoga)are the characteristics of the soul’ (Uttar 28.11). This inventory shows that upayogaeventually assumes its classical meaning of cognitive application, i.e. twofold cognitivefaculty of the soul.

Gradually, probably around the fourth century the twofold division of cognitive application(upayoga) merged with the fivefold classification of cognitions. The result (Model VI) was alist which also included varieties of perceptual experience, or indistinct cognitive acts, thatroughly corresponded to the varities of cognition, except for any perceptual experience thatwould correspond to testimonial cognition (œruta-jñâna) or to telepathy (manaÿ-paryâya-jñâna). The reason for that was that both œruta-jñâna and manaÿ-paryâya were felt to beclosely related to language and conceptualisation, therefore could not have theircorresponding indistinct counterpart.

After 2nd century CE another concept enters Jaina epistemology under the indisputableinfluence of the school of Ânvîkšikî and Nyâya. It was the idea of a ‘cognitive tool’ or ‘thecause of valid cognition’ (hetu), primarily developed in the tradition of Ânvîkšikî (1 centuryBCE - 1 century CE)17, which was soon adopted by Nyâya and other schools under the nameof cognitive criterion (pramâòa)18. The historical layers in Jaina canonical literature reflect thehistorical stages of the development of the concept and its name. In the course of time, theJainas adopted the well-known classification into four categories of perception (pratyakša),inference (anumâna), analogy (aupamya) and scriptural testimony (âgama), first classyfingthem as the causes of valid cognition (hetu)19, which reflected the Ânvîkšikî influence, andsubsequently as cognitive criteria (pramâòa)20 (Model VIII). It seems that the notion of‘cognitive criteria’ was introduced into Jainism with ‘the typological package’, i.e. alreadywith the fourfold classification, not as a separate term.

16 E.g. Uttar 28.4 discusses cognitions first, and the mention of upayoga is made in a diferent context(Uttar 28.10–11).17 CarS vi 8.3318 NS 1.1.3.19 Þhâò 336 (p. 149).20 Viy 5.4.26[3] (vol.1, p. 201.1–2).

9

This tradition, albeit adpoted in some strata of the canonical literature, was feld as a foreignbody in Jaina epistemology, and was indeed incompatible with all prior philosophicaltradition of the Jainas, and was subsequently obliterated (after 5th century CE), with oneexception though.What remained was the acutely felt need to classifly all cognitions within the framework ofcognitive criteria (pramâòa), a standard Indian typology of valid cognitions and/or theirsources which every serious system of Indian philosophy was supposed to have.The first attempt to adjust Jaina fivefold classification of cognitions (Model I) to newrequirements of Indian standard epistemology was undertaken by Umâsvâmin (ca. 350-400)who first lists the five standard varieties of cognition and only then adds that these fit wellinto the pramâòa scheme.21 The new classification of two cognitive criteria takes up thecanonical division of cognitions into direct and indirect (Models III, IV). The result is ModelIX, which basically corresponds to Model IV with the addition of two cognitive criteria: directand indirect cognition, that merely replace the direct and indirect cognition of Model IV.Traditionally, cognitive criteria concerned the question of the validity of cognitions (jñâna),i.e. merely one aspect of cognitive faculties (upayoga). What was left out was the othervariety of cognitive faculties (upayoga), i.e. perceptual experience. The Jaina authors werequick enough to include both kinds of cognitive faculties into one scheme that partiallyoverlapped with the model of cognitive criteria (Model IX), independently developed by thesame authors (TS 2.8–9, TBh 2.8–9, SSi 2.8–9). Accordingly, the comprehensive model ofcognitive faculties (Model XI) includes all cognitive acts irrespective of their validity. Wefind there three major categories: five valid kinds of cognition, that overlap with the five sub-varieties of cognitive criteria (Model IX), three varieties of erroneous cognition (ajñâna), bothof which (i.e. jñâna and ajñâna) are distinct (sâkâra), as well as four varieties of perceptualexperience (darœana), which is indistinct (anâkara).This Model XI is, in fact, Model VI augmented with the scheme of erroneous cognitions. Thevery fact that Model XI envisaged separate room for ajñâna was the result of introducing theidea of valid cognitions, connoted by pramâòa. It has also as its direct background anotherModel X found in a canonical work Paòòavaòâ-sutta, on which Umâsvâmin elaborated.Conspicuously, the canonical passage dates prior to the moment when the theory of cognitivecriteria was incorporated into Jaina epistemology (i.e. before TS) insofar as it does not refer toit at all;22 instead it elaborates the idea of cognitive faculties and excludes telepathy (manaÿ-paryâya), perfect cognition (kevala-jñâna) and pefect perceptual experience (kevala-darœana)from its scheme.

The development of scheme of Model VI, which distinguished two kinds of cognitiveapplication (upayoga) took also another, slightly different track to eventually merge with theidea of the three salvific elements, popularly known as three jewels (tri-ratna), that lead toliberation (TS 1.1): correct cognition (jñâna), correct conation (darœana) and correct conduct(câritra). The final model found in late canonical strata (Model XII) tried to accommodate theidea of cognitive criterion (pramâòa), or valid cognitive measures, in the tripartite system ofthe three jewels, and accordingly all the three salvific components were named ‘validcognitive measures’, which consituted the qualities of the soul (jîva-guòa). As innerirreducible qualities they came very close to the idea of defining characteristic (lakšaòa) ofthe soul (Uttar 28.10–11, TS 2.8). In this way, the former idea of twofold nature of the soulconsisting in cognitive application (upayoga) was extended to three qualities, the third being‘valid cognitive measure of the soul’s quality in the form of conduct’ (câritra-guòa-pramâòa). Model XII also shows that the second element, or darœana, of the salvific triad nolonger meant belief or conation, but assumed a clearly epistemic dimension of ‘perceptualexperience’, i.e. darœana as one of the cognitive application (upayoga), and not darœana of 21 TS 1.9–12.22 Paòòavaòâ-sutta mentions the division into pratyakša and parokša in passing, but does not menionany division of cognitive criteria, or the term pramâòa per se.

10

the path to liberation (mokša-marga) and confirms ambiguity of the term mentioned earlier(p. 8). Model XII is significant also for another reason. It is a clear canonical response to thecriticism levelled by other systems against the Jaina ambiguous usage of the term pratyakšaand their peculiar definition of perception. It tries to find the golden mean between the earliercanonical understanding of perception as cognition directly appropriated by soul without thesensory or mental medium and the idea of perception as cognition derived from sense organs.It specifically distinguishes fivefold sensory perception (indriya-pratyakša), based on fivesense organs, and threefold extrasensory perception (no-indriya-pratyakša), comprising thethree traditional kinds: clairvoyance (avadhi-jñâna), telepathy (manaÿ-paryâya-jñâna) andomniscience (kevala-jñâna). The attempt was not bereft of inconistencies, the major beingapparent conflicts: one between the idea of ocular (cakšur-pratyakša), as a variety ofcognition (jñâna), and ocular perceptual experience (cakšur-darœana), the first kind ofperceptual experience (darœana), and the other, similar one between auditory (œrotra),olfactory (ghrâòa), gustatory (jihvâ) and tactile (sparœa) perceptions and non-ocularperceptual experience (acakšur-darœana). It would be quite difficult to tell where thedifference between these sensory varieties of jñâna and darœana lay. This model was notfollowed by Jaina tradition in the centuries to come.A very similar canonical attempt to reconcile conflicting traditions within Jainism andincompatibility of Jaina understanding of perception with pan-Indian philosophical traditionwas attempted in another canonical work Naôdi-sutta. This culmination of the canonicaldevelopment, which was Model XIII, systematised ca. 450-480, was undertaken either prioror at the time of the third Valabhî Council, which under the guidance of Devarddhi-gaòinKšamâ-œramaòa finally codified the corpus of Jaina scriptures. Just like as Model XII, itdistinguishes two kinds of perception: sensory (indriya-pratyakša)and extrasensoryperception (no-indriya-pratyakša) within the traditional scheme of direct cognition(pratyakša) and indirect cognition (parokša).

We can easily see that the early development of Jaina epistemology focused on classificationof cognitive acts and was marred by a tension between an attempt to faithfully preserve earlyCanonical tradition and a tendency to adapt to new ideas developed by philosophers ofÂnvîkšikî, Nyâya, Vaiœešika schools, that exercised strong pressure on Jaina philosophers toupdate their system of epistemology.The post-canonical period was dominated by tree competing and complementary models –Model IX, Model XI and Model XIII, along with their minor variants and slight modifications– that were developed during the canonical period, viz. by the end of 5th century.

Jaina epistemological theories of canonical and early post-canonical period faced a number ofinsoluble difficulties. One of them has already been mentioned a number of times: how toreconcile the specifically Jaina meaning of the term pratyakša as ‘direct cognition’ with themeaning of ‘perception’ or ‘sensation’, universally adopted by Indian philosophers? Anattempt to solve the problem was a late canonical suggestion to distinguish two kinds of directcognition, namely sensory perception and extrasensory perception. This attempt, however,was not widely acknowledged. Another problem was the overlapping of two cognitivefaculties – cognition (jñâna) and perceptual experience (darœana) – at various levels.

3. The change: in search of criterion

Such was, more or less, the state of affairs still at the turn of the sixth and seventh centuries,before two outstanding contemporaneous thinkers (ca 600-660), the Mîmâôsaka KumârilaBhaþþa and the Buddhist Dharmakîrti, entered the philosophical scene. Their exchange of

11

arguments on epistemological issues23 also had an impact on Jaina stance and forced Jainathinkers to revise their epistemology. It was not only a desire to adjust Jaina classification ofcognitive criteria to what most Indians understood under direct and indirect types of cognitionbut also a strongly felt need to comply with generally accepted terminology, a precondition toengage in a philosophical discourse based on a commonly shared philosophical language, thateventually forced terminological and typological changes.

It was around the eighth century when new ideas eventually materialised in Jainaphilosophical treatises and Siddhasena Mahâmati could be credited with advancing an entirelynew classification of cognitive criteria, although his revolutionary ideas were clad intraditional phraseology, as is often the case with transforming traditional principles ofphilosophical systems closely related religious dogmatics that tends to oppose any change. Inthe opening verse of his Introduction to Logic (NA 1) he defines what cognitive criterion isand distinguishes two kinds of it:

‘The cognitive criterion is the cognition revealing itself and something else (sc.something different from it), and it is free from subversion; it is two-fold:perception as well as indirect cognition, corresponding to the way ofdetermination of the cognoscible.’

The reason for the division is a modus operandi on which cognition rests: as Siddhasenaimmediately clarifies,24 perception is therefore a direct cognition ‘that grasps an object not-indirectly’ (aparokšatayâ), which explains what mutatis mutandis indirect cognition shouldbe. This seemingly tautological and uninformative clarification takes recourse to earliertraditional phraseology of the canonical period. Such an approach, which compromised twoseemingly incompatible traditions, had, as we shall see, a hidden potential that made itrevolutionary.Further, as a subsequent verse suggests (NA 27), one could distinguish two kinds ofperception: conventional, sense-dependent perception and absolute, sense-independentperception (kevala), contingent on two understandings of the term ‘perceiving organ’ (akša,cf. p. 7). Accordingly, perception was defined as a cognition directly acquired either through asense organ (akša as ‘the eye’ understood as ‘a sense organ’), in accordance with a generalIndian approach, or through the soul (akša as ‘the soul’), in compliance with earlier Jainacanonical tradition. All kinds of cognition that could not be subsumed under the head ofperception belonged to the category of indirect cognition, and what Siddhasena particularlyhad in mind was verbal cognition (œâbda, NA 8,9) and inference (NA 10). He furtherdistinguished two subtypes of verbal procedures: ‘sentence for oneself’ and ‘sentence forothers’ (svârtha-vâkya and parârtha-vâkya, NA 10), which has usually been interpreted by

reserachers to stand for two varieties of inference (svârthânumâna and parârtha-pratyakša)25,however Siddhasena seems to extend the division to cover all verbalisation, both innermonologue, a variety of which is ‘inference for oneself’ as an internal process of associationand drawing general conclusions, and interpersonal communication, its variety being‘inference for others’ as a formalised persuasion procedure. Conspicuously, Siddhasenaradically does away with cognitive application (upayoga),26 i.e. with two kinds of cognitivefaculties, i.e. with the distinction into cognition (jñâna) and perceptual experience (darœana),which could indeed be confusing for a consistent epistemologist.

23 See STEINKELLNER (1997).24 NA 4: ‘Such a cognition that grasps an object not-indirectly is perception; the other one should beknown as the indirect cognition, as far as the manner of grasping an object is considered.’25 See, e.g., SHAH (1967: 32).26 See BALCEROWICZ (2003a: 44 f.).

12

Venturing such a consistent epistemological stance, Siddhasena further distinguished twosubtypes of conventional perception: ‘for oneself’ (svârtha-pratyakša) and ‘for others’

(parârtha-pratyakša) – apparently, out of the desire for symmetry or harmonic consistency.The reason is a conviction that all cognitions share both efficacy ‘for others’ (pârârthya) andefficacy ‘for oneself’ (svârthya). One can reason internally, ‘for oneself’ (svârtha−°), with thesole purpose to better understand a particular phenomenon, but one could also convey one’sown understanding to other people (parârtha−°) through a proof formula (prayoga) and arguein favour of one’s own understanding of the phenomenon. Similarly, one usually perceivesobjects and he himself is the sole beneficiary of his own perceptive acts. However, it ispossible, argues Siddhasena, that the ultimate beneficiary of one’s own perceptive act isanother person (parârtha−°). That happens when one’s own perception is communicated thisway or another, e.g. by a gesture, by an utterance or otherwise, and triggers a very similarprocess of perception in another person. As he says, ‘Such an utterance that demonstrates anobject recognised through perception is called perception, because it is the external factor forthe representation’ (NA 12). Just a macroscopic object z, being external to one’s cognitiveapparatus and mind, activates a sense organ and gives rise to its representation in the mind ofthe cogniser, similarly a verbal symbol or gesture of a person x, being in the same wayexternal to the cognitive apparatus of person y, can generate a cognitive process the finalcontents of which is that macroscopic object z. In other words, verbal utterances, gestures orother varieties of meaningful expression through symbols can be classified as cases ofperception under special conditions, viz. if they communicate something that is beingperceived and contribute to the generation of a piece of knowledge in the hearer thatcorresponds to the speaker’s cognitive state derived perceptually. Such a process of‘perception for others’ is structurally far more complex than ‘ordinary’ perception, i.e.‘perception for oneself’. But, Siddhasena would argue, also the process of ‘inference forothers’ is far more complex than ‘inference for oneself’; in the case of perception for othersand inference for others the operation of cognitive apparatus of person x is external toperception or inference, respectively, of person y and it only activates person y’s cognitiveprocess.It seems that the criterion for such a division of all cognitive processes is the idea of theultimate beneficiary of the cognitive process. If the beneficiary is thought to ultimately be theprimary cogniser, the cognition is ‘for oneself’ (svârtha-pratyakša, svârtha-vâkya,

svârthânumâna). If the ultimate beneficiary is intended to be another person and if one wishesto convey one’s cognition to another person either directly (through a perceptive act) orindirectly (through a speech act), then we have the case of the cognition ‘for the other’(parârtha-pratyakša, parârtha-vâkya, parârthânumâna). This classification was directlyinspired by Siddhasena’s Buddhist predecessor Diónâga (480–540),27 who was the first tointroduce a distinction of inference for oneself (svârthânumâna) and inference for others

(parârthânumâna)28. The result of such a complex argument aiming at consistency, simplicityand symmetry was Model XIV.

27 On Buddhist influence on Siddhasena see BALCEROWICZ (2008: ix ff.)28 See PS2 2.1ab: ‘Inference is twofold. The first one, inference for oneself, is the discernment of anobject by means of the triple-formed inferential sign’; and PS2 3.1ab: ‘…inference for others is theproclamation (sc. demonstration) to others of the object one has experienced oneself’. See: RANDLE

(1926: 28–9) and HATTORI (1968: 78, n. I.11). We find the influence of Diónâga’s svârtha / svârtha

distinction and conceptualisation-free perception already in Pûjyapâda Devanandin’s Sarvârtha-siddhi(SSi1 1.6, p. 15.1–2), who says: ‘In TS 1.6, cognitive criterion is twofold: for oneself and for the other.Out of these two, the cognitive criterion for oneself is free from testimonial (œruta) component.Further, testimonial cognition is for oneself and for the other. “For oneself” means “to consist incognition”; “for the other” means “to consist in statements”.’

13

Further, Siddhasena Mahâmati was explicit in indicating that the reason for the two-folddivision of cognitive criteria into perception and indirect cognition was two different modesof apprehending reality, or two different epistemic procedures. He considered it of crucialimportance, for he emphatically expressed the idea twice in his succinct work that the division‘corresponds to the way of determination of the cognoscible’ (NA 1d), viz. ‘as far as themanner of grasping an object is considered’ (NA 4d).His emphasis on two different cognitive procedures had a historical background: it was tocounter an earlier claim made by Diónâga, and later reinforced by Dharmakîrti, that there areprimarily two different, irreducible aspects of reality (sva-lakšaòa and sâmânya-lakšaòa, seebelow, p. 13), and that dyadic division influences the way we can cognise it.Without Siddhasena’s Buddhist predecessors it would not be possible to account for theepistemological turning point marked by Siddhasena’s scheme of cognitive criteria, and theinfluence can be seen at different levels.First although the general binary scheme of cognitions, accepted traditionally by the Jainas,was preserved by Siddhasena in Nyâyâvatâra, it was merely a nominal acceptance, inasmuchas Siddhasena gave it a new dimension and contents: the term pratyakša came to mean‘perception’ the way it was generally understood in India. That was a typical strategyfollowed by the majority of Indian thinkers who would introduce changes and modificationsin the systems not by an outright criticism of their predecessors but by launching a newinterpretation (‘rectification of erroneous interpretations’) of earlier statements.Secondly, Siddhasena’s manner of grouping of all cognitions under two heads likewisefollowed the path earlier taken by Diónâga, albeit with some important modifications. Weshould, therefore, first turn to the Buddhist philosopher. In the opening section of hisPramâòa-samuccaya (PS 1.2ac, PSV ad loc.), he famously states that

‘[PS] The two cognitive criteria are perception and inference, because thecognoscible object has two marks… [PSV] For there is no cognoscible otherthan the individually marked (unique particular) and the generally marked(universal thing), and we shall further demonstrate that perception has as itsdatum the individually marked (unique particular), whereas inference has asits datum the generally marked (universal thing).’

The ground-breaking passage distinguishes two cognitive criteria that would correspond totwo different and irreducible aspects of reality. It can be known - in Diónâga’sphenomenological approach - either (1) in terms of absolutely unique sensation, i.e. a senseimpression which points to itself only, unrelated to anything else either concrete or abstract,and can be called a unique particular (i.e. ‘what is marked by itself’, sva-lakšaòa), or (2) interms of concepts - summarily labelled ‘a universal thing’ (i.e. ‘what is marked by similarity’,sâmânya-lakšaòa) - that actually convey a synthetised image, embedded in a set of relationsthat 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.The line of distinction between the two spheres, that of the unique particular and that of theuniversal thing, is drawn along the application of the idea of conceptuality, or conceptual stateof mind (kalpanâ), that has a universal thing as its referent, whereas – as Diónâga emphasises– ‘perception is free from conceptual construction, which, in its turn, is connected with name,class, etc.’What Diónâga means by the neologism29 ‘that which is marked by itself’ (sva-lakšaòa) is thatperceptual datum that constitutes the contents of perception, e.g. a particular patch of colour, aunique taste or touch sensed in a particular moment etc., carry the information which does not

29 Although the term has its earlier Abhidharmic history, the way Diónâga uses it is novel. For thehistorical background see: SINGH (1984: 117–135).

14

extend over its presently perceived actuality. In other words, the unique particularcommunicates nothing but its own self, or its presence, and by definition is bereft from anykind of conceptualisation or conceptually constructed ideas and verbal designation.All the remaining kinds of cognition, which are collectively subsumed under ‘inference’,share a peculiar feature that their contents is based on conceptually constructed ideas.Accordingly, the contents of ‘inference’, which actually comprises all kinds of valid speech-and concept-related pieces of cognition, is artificially constructed, generated on the basis of arange of perceptual data, memory and conceptual framework imposed on the perceptual databy the language. Such a conceptually constructed idea conveys a generalised object, be it asequence of perceptual data that are subconsciously associated together as one and the sameindividual thing extending in time over moments due to their similarity, be it an arrangementof a string of perceptual data that have been independently apprehended by different senseorgans and subsequently correlated with each other into one ‘whole’, that induces animpression of its relative permanence, or be it a set-up of isolated abstractions, the essentialnature of which are their relations to other conceptual entities and which are abstracted from arange of similar features of perceived phenomena.30 Such a ‘second-hand’ cognition isprecisely what facilitates acquisition, or expansion of knowledge of a conceptual object x, ofwhich we have no direct knowledge, on the basis of an object y, which has already becomethe contents of our cognition and of which we know that it is related to the object x by aparticular relation R. It was Diónâga’s claim that the use of any system of symbols, which arenecessarily constructed conceptually, be it in verbal communication or reasoning, involves theselfsame rules that govern the acquisition of knowledge and co-ordination of its contents.What follows from his assumption is that all cognitive processes based on symbols, includingverbal communication, can be classified as some sort of inference:

‘Verbal cognition is not an additional cognitive criterion, different frominference, 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 “beingproduced” etc.’31

In contradistinction to perceptual experience, the contents of which actually never enters theconceptual mind and of which one is never consciously aware in common experience, butwhich constitutes an immediate experience of the real, in the case of all other kinds ofcognition, which are conceptual, there is something that mediates between the objective factand the contents of the mind. In the case of a piece of cognition, derived second-handedly,that ‘there is fire on a mountain’, although we do not see it ourselves, whether we rely on thefire’s specific mark, or its inferential symbol, i.e. smoke, which we directly perceive, orwhether we depend on the term ‘fire’, being communicated to us by a verbal symbol, in bothcases we conclude the same fact and, structurally and procedurally speaking, both these modioperandi are essentially in no way distinct from each other. In both cases a particular abstractsymbol triggers the knowledge of its (ultimately abstract) denotatum to which it is related. Or,in other words, certain phenomena perceived by us, of which we know that they are attributesof other objects and inferentially lead us to them, as well as verbal expressions, of which weknow that they refer to things other then the verbal expressions themselves (i.e. they are notself-referring), serve the same purpose in precisely the same manner: ‘Since there is similarityin terms of infallibility as regards the utterance of an authoritative person, it has the status ofinference’32. What follows is that the gist of indirect cognition, called by Diónâga ‘inference’,is determined by two coextensive and inseparable ideas, or symbols: conceptualisations and 30 Cf. HERZBERGER (1986: 106–144), HAYES (1988: 133–144, 173–219).31 PS 5.1. For other translations see: HAYES (1988: 300) and HERZBERGER (1986: 145–146).32 PS 2.5, quoted in PVSV, p. 108.1.

15

speech, that are not really different from each other: the cognitive contents is always ageneral, verbally expressible concept, which conveys some general features, not a particularindividual, which remains specifically the scope of perception only.This is why, for Diónâga, all our cognition is organised in two disconnected compartmentsthat correspond to two aspects of reality. These two aspects of reality are ultimately thegrounds to distinguish two cognitive criteria. What Diónâga actually does in his analysis is tosearch for one clear criterion that help us reduce all sound cognitive phenomena to aminimum number of subtypes of cognitions. That criterion is eventually conceptual state ofmind (kalpanâ).Although these were the ideas initiated by Diónâga that eventually influenced Siddhasena’ssearch for the reason for twofold typology of cognitive criteria, they were mediated by andfiltered through the formulations of Dharmakîrti by the time they entered NA, because thatwork which moulded the structure of NA was Dharmakîrti’s Nyâya-bindu33.In this work (NB 1.12–17) Dharmakîrti, who adopted and modified Diónâga’s scheme andintroduced some crucial innovations, was explicit that the reason to distinguish two differentcriteria was the existence of two distinct realms for our cognition:

‘The datum for perception is the individually marked (unique particular). Anobject in the case of which we experience the change in its representation incognition depending on its proximity or distance is what is individuallymarked. Nothing else but the unique particular is the ultimately existent,because any real thing is characterised by the efficacy to execute causallyefficient action. What is different from it is the generally marked (universalthing). It is the datum for inference.’

On an earlier occasion (PV 2.1–3), Dharmakîrti mentioned a number of additional reasons forthe twofold division of the cognitive criteria – i.e. a thing’s uniqueness as the scope ofperception and things’ similarity as the scope of inference, the thing’s inverbalisability (in thecase of perception) or verbalisability (in the case of inference), the thing’s temporal andspatial presence (in the case of perception) or absence (in the case of inference) during thecognitive act – the twofold division of cognitive criteria:

‘[1] Cognitive criteria are two, depending on two kinds of data which, in theirturn, are also two because of their potential (in the case of perception) or lackof potential (in the case of inference) to execute causally efficient action. Forinstance, a hair-net is not a really existing object, because it does not warrantthe fulfilment of a causally efficient action that concludes with theappropriation of the object.[2] We recognise that there are only two cognitive criteria also because ofsimilarity (in the case of inference) and lack of similarity (in the case ofperception), and also because of the capability of the data to become thecontents (in the case of inference) or because of the incapability to becomethe contents (in the case of perception) of speech, and also because theobject’s image is either present (the case of perception) or absent (the case ofinference) when other causal factors are present during the cognitive act.[3] What is capable of causally efficient action is called here the ultimatelyexistent; what is different is called the conventionally real. These two arerespectively the individually marked (unique particular) and the generallymarked (universal thing).’

33 See: BALCEROWICZ (2001a: xii ff.).

16

There is no doubt that the above line of argumentation to distinguish two cognitive criteria ledSiddhasena Mahâmati to devise a consistent scheme that gave the earlier Jaina tradition atouch of more stringent consistency. In contradistinction to Diónâga’s and Dharmakîrti’sconsiderations, for Siddhasena the reason to distinguish cognitive criteria was not ontological,and thus external to epistemic considerations, i.e. two aspects of reality, but exclusivelyepistemic: two different modes of cognition, or modi of its acquisition. Valid cognition isorganised in two disconnected compartments precisely because there are two modes of itsacquisition, whereas the reality remains one and the same.The above three verses also introduced Dharmakîrti’s novel definition of what the real is:what is ultimately existent is also causally efficient and triggers perception, its definingcharacteristic being the capability to execute causally efficient action (artha-kriyâ-sâmarthya). When the real is subsequently manifested in the cognition as an image(pratibhâsa), or verbally expressible concept, its mental representation is not causallyefficient, hence it is not ultimately real. A serious consequence of such a definition of the realhad its impact on Dharmakîrti’s assessment of the character of inference. Since actualreferents of any conceptual cognitive act are fictitious, insofar as they do not comply with thedefinition of an ultimately existent entity and are not effectively causal, inference cannot butbe erroneous (bhrânta).Such a position provoked still another reaction from the side of Siddhasena, who emphasisedthat all cognitive criteria have to be, by definition, veracious, non-erroneous and notcontradicted by other pieces of cognition obtained through other cognitively valid procedures:‘the cognitive criterion is cognition … which is free from subversion’ (NA 1). Anycognitively valid procedure, if it is supposed to yield a reliable piece of cognition, must itselfbe reliable, and the condition holds valid also in the case of inference (NA 5: ‘Inference is …non-erroneous because it is a cognitive criterion, just like perception’), something whichDiónâga’s and Dharmakîrti’s school vehemently denied, because – as they argued – inferencedid not yield knowledge of the only ultimately existent things (sva-lakšaòa), but only ofconceptually constructed ideas, which are – as actual referents – fictitious. However, asSiddhasena countered, we could not possibly rely on a source of cognition which itself is notnon-erroneous, or else all our knowledge and its fundaments would collapse.He further argued, in accordance with the correspondence theory of truth he stronglyadvocated, that representations of external objects, existing in the mind as mental images(pratibhâsa), are both reliable and true (NA 7: ‘… it is incorrect to assume erroneousness ofall representation’), and that is why, conceptual cognition can be true and can have a reallyexisting object outside of it corresponding to the mental image or idea. That explains whyconceptualisation-based cognitive criterion (parokša) is not only reliable (in the Buddhistsense) but also necessarily true and non-erroneous (abhrânta).

4. In search of descriptive definition

And this leads us to perhaps most important contribution of this period to Jaina epistemology,namely to the first formulation of what a cognitive criterion (pramâòa) actually is. The idea ofpramâòa and the technical term for it had been current in India since the beginning of the firstmillennium CE and there had also been formulated various typological definitions ofpramâòa, which demarcated the notion by merely mentioning a range of specimens to whichthe definition applied, but it was not until seventh century when first descriptive definitions ofit were developed, i.e. proper definitions which describe the character and essential aspectsthe definiendum.As the pioneer we could regard here Dharmakîrti34, who apparently was the first to offer adescriptive definition of pramâòa, in two parts, that specified its character and individual

34 See: FRANCO (1997: 59–61), cf. STEINKELLNER–KRASSER (1989: 3–5).

17

features as follows: ‘the cognitive criterion is an unerring cognition, … alternatively thecognitive criterion is a cognition that manifests a previously uncognised object’ (PV1 1.3, 7).35

That was immediately followed by Kumârila who, in the now lost Båhaþ-þîkâ, formulated hisown definition, preserved as a quotation in a Buddhist work Refutation of the proof ofpermanent things (Sthira-siddhi-dûšaòa, RNA 106.9–11) by Ratnakîrti:36 ‘Such a cognition ofa previously uncognised object, which is determined and triggered by causes that are notdefective, is accepted by people as cognitive criterion.’37

Siddhasena Mahâmati followed the suit and can be considered the first Jaina philosopher tooffer a descriptive definition of pramâòa: ‘The cognitive criterion is the cognition revealingitself and something else different from it (sc. object), and is free from subversion’ (NA 1)38.This step can certainly be considered an advancement of Jaina epistemology.In his opinion, what distinguishes a cognitive criterion from any other act of cognition that isneither true nor absolutely dependable is its dissociation from any kind of subversion (bâdha),or sublation: it is both well-founded (or sufficiently corroborated) and not falsified either byobservation or by logic, as well as consistent with one’s body of true beliefs39. That is theveracious aspect of his definition, stated in the second part of the definition. Further, and thisis the first part of the definition, cognitive criterion not only presents the object to the mind inthe form of its representation (pratibhâsa) but also makes one immediately aware of therepresentation and of the fact that one is aware of it: otherwise one would not be aware of themental image of the cognised thing and would simply not know it consciously.40 Such anunderstanding has two important implications. First, cognitive criterion proper must be acognition through which one becomes immediately aware of the object, and it is not a case, asthe schools of Nyâya and Mîmâôsâ maintained, that one becomes aware of the mental imageof the object in a subsequent act of cognition, after the object has left its impression on themind, through a process of inner perception or inner inference. Further, we immediately knowthat we know – this is, again, not an act of inner inference of the sort: from the fact we knowan object we infer that we know that we know it. Another important implication was therealistic component of any true cognition, and it was indeed and expression of thecorrespondence theory of truth: only such a cognition which directly represents an externalobject can be true. It was a direct criticism of all idealistic philosophies. In a subsequentsection41 Siddhasena – as if following the procedure of Dharmakîrti (see above p. 16) –elaborates on the definition, and in its second formulation introduces one more element, i.e.patency. The term ‘patent’ (sphuþa) became an integral part of most definitions formulated bylater Jaina philosophers, one of the first ones to adopt it was Akalaóka who used itinterchangeably with ‘clear’ (viœada)42.

35 The former is sometimes called ‘epistemological’ definition, whereas the ‘pragmatic’ definition, seeKATSURA (1984).36 See KATAOKA (2003: 99).37 Some (KATAOKA (2003: 95 f.)) consider that the definition had its forerunner in Kumârila’sfollowing statement made in his earlier work: ‘A cognition which is resolute, which has actually arisen,and which does not disagree with another later cognition, should be considered a cognitive criterion’(MŒV 5.5.80), whereas others (FRANCO (1997: 59–60)) deny it arguing that the verse merely statesgeneral conditions for a pramâòa to occur.38 See BALCEROWICZ (2001a: xiv–xv) and (2005: n. 31).39 Cf. NA 21.40 The same idea is expressed in NA 7c.41 NA 7: ‘Cognitive criterion, which is patent and which determines itself and something different fromit, proves to be correct with regard to establishing the dyad’, i.e. it correctly reveals itself and trulyrepresents its object.42 E.g. LT 7, 23; see below p. 23 f.

18

Siddhasena may seem to have been anteceded by Samantabhadra43, who characterisescognitive criterion as ‘the cognition of reals / doctrinal categories’ (ÂMî 101a), such as livingelements, lifeless elements etc., enumerated e.g. in TS 1.4. What looks at first like a definitionturn out to be, when read in the context44, merely a restatement of Umâsvâmin’s TS 1.6:45 ‘thecomprehension of these categories representing reality (mentioned in TS.1.4), is accomplishedthrough cognitive criteria (pramâòa) and conditionally valid viewpoints (naya),’ and was notconsidered a definition by Samantabhadra himself. There is still one more phrase in hisSvayambhû-stotra (SvSt 63cd / 13.3cd) which some46 take to be Samantabhadra’s definitionof cognitive criterion: ‘… like a cognitive criterion, which – with respect to earth – manifestsitself and something else (sc. object) and is characterised by a cognitive state.’ To do so, onewould only prove to be mislead by his reading the passage out of context, because what thefirst hemistich verse of SvSt 63 / 13.3, of which the above quote is a continuation, reads: ‘Wecan speak of one consolidated cognition of the universal and the particular jointly well knownthrough the attributes of continuity and singular character of a perceived thing that aremutually dependent, like a cognitive criterion, which – with respect to earth – manifests itselfand the other (sc. object) and is characterised by a cognitive state.’ In the verseSamantabhadra merely refutes the idea that there are two irreducible aspects of reality whichcan never occur in one act of cognition, which remains divided into two impregnablecompartments, as Diónâga claimed (PS/PSV 1.2ac, see above p. 13), but instead weapprehend both the universal and the individual aspects of one and the same reality in one actof ‘consolidated cognition’ just the way numerically one and indivisible cognitive criterionalways reveals two aspects of reality (the cognised object and itself) at one swoop. That is notonly a refutation of Diónâga but also a reply to the arguments formulated by Dharmakîrti47.Further what some consider a definition would hardly fulfil the requirements of a properdefinition of a pramâòa, even in the eyes of Samantabhadra himself, because the phrase ofSvSt 63cd covers most acts of cognition, be they true or fallacious. Incidentally, however,Samantabhadra’s phrase does contain an element that later features in Siddhasena’sdescriptive definition of cognitive criterion, i.e. that ‘it manifests / reveals itself andsomething else’. Besides, we should always distinguish a conscious attempt of an author toformulate a definition of a phenomenon from a casual remark which centuries later happens tobe interpreted as such, which is a classic case of projecting subsequent development of ideason an earlier phase.48 However, if there were a Jaina philosopher to consciously formulate adescriptive definition of pramâòa prior to Siddhasena Mahâmati it was certainly notSamantabhadra but someone posterior to him, of whom we do not know for certain, apotential candidate being Pâtrasvâmin (alias Pâtrakesarin / Pâtrakesarisvâmin, c. 660–720), theauthor of the now lost Tri-lakšaòa-kadarthana, who, it seems, exercised considerableinfluence on Siddhasena Mahâmati.

43 Traditionally held to have lived around 550, however, there are valid reasons to maintain that hemust have been acquainted with some of Dharmakîrti’s works (e.g. certain passages of ÂMî seem tobetray Dharmakîrti’s stamp), see BALCEROWICZ (2008: ii) and (forthcoming).44 The complete verse of ÂMî 101 reads: ‘Cognitive criterion is the cognition of reals (doctrinalcategories). It either is a simultaneous representation of all entities (which is the case with the absoluteknowledge (kevala)) or occurs consecutively (of this kind are remaining varieties of cognitive criteria).The cognition of these reals is composed of the method of seven-fold modal description (whichcorresponds to pramâòa) and of conditionally valid viewpoints (i.e. naya).’45 BALCEROWICZ (2005b: 359, n. 31).46 FUJINAGA (2003: 164–165).47 Esp. PV1 1.5–6 and PV1 2.76 f.48 For instance Samantabhadra’s phrase of ÂMî 101a was apparently taken by Vâdirâja-sûri (first halfof 10th century) in his Nyâya-viniœcaya-vivaraòa (NViV 1.3, p. 57.22) to be a full-fledged definition ofcognitive criterion. On the other hand, Prabhâcandra in his Svayambhû-stotra-þîkâ (SvStÞ, p. 70-71)does not consider SvSt 63cd a definition of pramâòa but merely – and correctly so – an example ofhow two different, seemingly incompatible aspects can exist in one entity.

19

5. In search of economy

These were the developments which took place within Jainism in the period posterior toDharmakîrti which brought significant changes particularly to Jaina epistemology. In thisperiod the Jainas attempted to redefine earlier models, maximally simplify general typologyof cognitions and specify new criteria for it in a manner that would be commensurable withthe general Indian standard. However, there is one more important improvement to note thepotential impact of which was not restricted only to Jainism but concerned Indian logic and itsmethods.Traditionally, following the methodology of the Nyâya-Vaiœešika school, Indian logiciansapplied a five-membered proof formula (pañcâvayava-vâkya), consisting of the thesis(pratijñâ), the logical reason (hetu), the example (dåšþânta) the application (upanayana) andthe conclusion (nigamana). An alternative was, even at an early date, suggested by Jainaphilosophers who, in addition, applied a ten-membered proof formula, variously understood,but mostly as containing supplementary tools, called ‘clearances’ (œuddhi), to lendverificatory support and to provide unshakeable validity to the main five members. Bothmodels strongly relied on the idea of the example (dåšþânta), which was supposed to providean empirical exemplification, both positive and negative, for a logical reason adduced. In thefifth century the indispensability of some of the members of the classical five-memberedproof formula came to be seriously questioned by the Buddhist master Vasubandhu. In histwo works on logic, now lost Vâda-vidhâna and Vâda-vidhi49, he suggests to dispense withthe last two members of the proof formula, i.e. with the application (upanayana) and theconclusion (nigamana), which he considered a repetition of the logical reason (hetu) and thethesis (pratijñâ), respectively. Vasubandhu and his followers Diónâga and Dharmakîrti stillrelied on empirical (both positive and negative) exemplification of the logical reason that wassupposed to demonstrate that the invariable concomitance (vyâpti) which was supposed tonecessarily relate the proving property (logical reason), or probans (sâdhana) and theinferable property, or probandum (sâdhya).50 This made Indian logic intensional in character,directly dependent on the contents of logical terms and on empirical verification.

The move to limit the number of necessary members of the proof formula and to simplify thereasoning procedures found its continuation with Jaina thinkers, the first of this new trend – ifwe are to rely on Buddhist sources, specifically Œântarakšita and Kamalaœîla – beingPâtrasvâmin. In his Tri-lakšaòa-kadarthana,51 he claims that ‘It does not matter if both types ofthe example can be adduced or not, because the example is not any basis for inference’ (TSa1368), and thereby he highlights the economical trend in Indian logic to dispense with all thatis not absolutely essential in logic. Since our inferences are only about the subject of inferencealone and its properties, any empirical exemplification or any additional instantiation ofsimilar relations is redundant. Therefore, the proof formula (sâdhanâvayava) may safelyconsist of only the thesis (NA 14) and the logical reason, based on an invariant relation, whichrelates the inferable property and the proving property (logical reason), to substantiate it. Heproposes to replace all kinds of logical reasons that had so far been in use with a single onecalled ‘inexplicability otherwise’ (anyathânupapatti, anyathânupapannatva). Its sole feature

is, in his opinion, the inseparable connection (avinâbhâva) of the proving property (sâdhana)with the inferable property (sâdhya), and that can be successfully known though suppositional

49 FRAUWALLNER (1933) and FRAUWALLNER (1957).50 A classical example of how such thinking was embedded in the structure of the logical proof wasDiónâga’s Drum Circle of reasons (Hetu-cakra-ðamaru).51 The only available fragments of the work are to be found in Tattva-saógraha (TSa 1364–1379).

20

knowledge (tarka, ûha).52 The economic drive was brought to its extreme by SiddhasenaMahâmati who suggested that, under certain conditions and provided one remembers theinvariant relation, one can effectively argue by applying a single-membered proof formula,that consists of the logical reason alone (NA 14–16)!53

Pâtrasvâmin is echoed by Akalaóka in the Laghîyas-traya (LT 12cd–13ab) who characterisesthe basis for inference as follows: ‘Inference is the comprehension of the sign-possessorthrough the inferential sign, which has as its the sole characteristic the determined cognitionof its inseparable connection with the inferable property…’ He further elucidates the idea inhis autocommentary: ‘For it is impossible to know the relation of essential identity andrelation of causality without the suppositional knowledge of “inexplicability otherwise”; andbecause the logical reason is also operative precisely without these two relations of essentialidentity and of causality, this is the proof that the logical reason has a sole characteristic. For atree etc. inferred on the basis of its shadow is neither the essential nature nor the effect of theshadow etc.’ As the illustration of the tree and its shadow should demonstrate, inference relieson a singular relation that corelates two facts x (sign-possessor / inferable property) and y(inferential sign / logical reason) in a unique way such that the occurrence of y is impossiblewithout the occurrence of x, and on the basis of such presumptive knowledge we infer ‘that itis the case that x’. What relates x and y is the inseparable connection (avinâbhâva), a relationadopted by the Jaina logicians from the Buddhists, is the idea that the logical reason (hetu) isnecessarily related to (‘does not deviate from’) the inferable property (sâdhya):

‘there is no logical reason without the inferable property’, orsâdhya-avinâ hetu-abhâva.54

To put it differently:

‘x extrapolates y, in case there is no x without y’, ory-avinâ x-abhâva.

What the claim of the avinâbhâva relation stating that ‘x is necessarily related to y’ meant wasthat the occurrence of y logically follows from the occurrence of x, and non-occurrence of ylogically follows from the non-occurrence of x.Accordingly, what constituted a turning point in the Jaina approach was the idea that the relationbetween x and y was so well determined and obvious that it required not additional positive(anvaya) or negative (vyatireka) instantiations.To answer the question how such reasonings can be founded, Siddhasena Mahâmati takesrecourse to what he calls intrinsic invariable concomitance (antar-vyâpti), which relates thelogical reason and the inferable property in a necessary manner: ‘Since the probandum isproved only by the intrinsic invariable concomitance, an exemplification from outside would bepointless…’ (NA 20). This peculiar relation renders, in his opinion, all additional positive andnegative instantiations both logically redundant and insufficient. The idea was notSiddhasena’s invention though for he explicitly refers to his predecessors in the same verse.Whether that was Samantabhadra is highly questionable55; most probably the one whomSiddhasena Mahâmati had in mind was Pâtrasvâmin. 52 See BALCEROWICZ (2003b). Later, the idea is succinctly stated by Akalaóka in LT 12: ‘…it is notpossible to know the relation of essential identity and relation of causality (i.e. two basic logicalreasons postulated by Dharmakîrti), without the suppositional knowledge of “inexplicabilityotherwise”.’53 BALCEROWICZ (2003b: 362). The idea finds its precedence in Dharmakîrti who maintained that forthose who know the context ‘simply the logical reason alone is to be stated’ (PV 3.27).54 BALCEROWICZ (2003b: 352 f.).55 Such a suggestion is made by FUJINAGA (2003: 165), who maintains that in ÂMî 6, Samantabhadra‘uses syllogism without example or dåšþânta. … This is the first usage of the antarvyâpti in the history of

21

How can we know intrinsic invariable concomitance (antar-vyâpti) with certainty even when wesee two entities only once and it is not a case of a repetitive experience? The certainty for ourreasonings was believed to be safeguarded by still another process known as suppositionalknowledge (tarka), or presumptive knowledge (ûha), which postulated the followingverificatory procedure :

[x and y are related in a necessary manner because:]whenever there is x, there must be y too;only when there is no y, there is no x.

To recapitulate, the suppositional knowledge establishes the postulate of the intrinsicinvariable concomitance that necessarily holds between x and y. On the basis of such a necessaryrelation, which requires no prior empirical verification through additional instances, one appliesthe logical reason of ‘inexplicability otherwise’ (anyathânupapatti):

If there were no y, there would be no x.

And what follows is the conclusion:

but there is x,so there must be y.

In such a reasoning, we first assume a counter-thesis which is contrary to the postulated thesis‘there is no y’, and subsequently, since we observe a particular fact which is a premise (‘thereis x’) contrary to the entailment of the assumed counter-thesis, we concluded that the latter isfalse: ((¬y → ¬x) ∧ x) →y.56 The structure is that of modus tollendo tollens.This inferential strategy was indeed quite a peculiar device that was, as a matter of fact,supposed to get over the problem of induction, solidly embedded in the strongly empiricallyoriented Indian logic: by means of the new procedure one could validly establish the relationof invariable concomitance that holds between the logical reason and the inferable propertyeven on the basis of a single occurrence observed. As Akalaóka (LT 49) remarked:‘Observation in a singular case does not make evident the invariable concomitance of thelogical reason with the inferable property without inductive thinking. This suppositionalknowledge (i.e. inductive thinking) has as its datum not yet comprehended objects in theirtotality (sc. including all cases). … thanks to the invariable concomitance established bysuppositional knowledge, inference possesses cognitive validity with respect to every singularplace housing an object which is covered by this suppositional knowledge…’ Its nature was athorough understanding of the relation between two entities which would allow one toextrapolate the singular observation on all similar cases.For an Indian logician induction-based reasoning – and such was Indian logic in general –involved primarily two stages: (1) one has to established a link between two individualparticulars (viœeša), which typify two classes, and subsequently (2) one has to extrapolate inorder to establish a invariant relation, universally binding, between the two classes (sâmânya)typified by the pair of the particulars.The difficulty such a model involved lay in the fact that the two stages are equally defeasibleand open to possible revision. At first stage, we can never be sure that the link between two

Jaina philosophy.’ The serious problem with such an interpretation – apart from the erroneous usage of theterm ‘syllogism’, which does not have its Indian counterpart at all (Indian logicians applied the structuresimilar to modus ponendo ponens with a substitution of an individual variable) – is that the verse inquestion is not a formulation of any kind of reasoning: it merely expresses a thesis and we find no logicalreason there. It is not entirely impossible that Samantabhadra was aware of the idea of intrinsic invariableconcomitance, but he never mentions it by name nor applies it in his proof formulas.56 See BALCEROWICZ (2003b: 357).

22

individual particulars is indeed invariant, and not incidental, as long as the we proof that thelink is that of an invariant relation variety that always hold; the soundness of the first stagelies in the validity of the second stage. However, at the second stage, even after numerouspositive verifications, we can never reach certainty that the allegedly invariant relation alwayshold as long as we make sure that no counter-example ever and under any circumstancesobtains, which would amount to examining all possible cases, which as such is impossibility.In a way, the soundness of the second stage depends on the legitimacy of the link observed atfirst level. For the Jainas, the two phases merged: by observing a relation between twoparticulars one could instantly extrapolate to whole classes on the basis of that claim thatotherwise the occurrence of the two related facts would be inexplicable (anyathânupapanna)

insofar as some kind of inseparable connection (avinâbhâva) hold between them. No longerhad one to take recourse to a two-stage inductive reasoning: the proper understanding of thelink between the two fact should suffice to establish an invariant relation between them andthe two classes they instantiated and there was no need any more to verify the postulatedrelation as well as there was no fear that the relation could be falsified by a future observationof an counter-example.Of course, the problem of induction remained: it was shifted from the level of impregnabilityvis-à-vis possible future falsification to the sphere of how we can really know that theoccurrence of one fact (proving property) is indeed inexplicable otherwise, i.e. withoutassuming its relatum (inferable property) one postulates on the basis of some kind ofinseparable connection.Irrespective of the futility of their attempt to come round the problem of induction, whatremained their valuable achievement was that the Jaina thinkers of the eighth century presented acompletely new approach to logic by claiming that proof formulas do not have to be empiricallyverified through other instantiations and, if they are at all valid they are so only by virture oftheir logical structure: the reasoning of the ‘inexplicability otherwise’ type, based on the ideaof inseparable connection (avinâbhâva) known by means of suppositional knowledge, is aformula with two variables (sâdhya and sâdhana), that makes an impression of beingdeductive. Such an approach, the motivation for which was an attempt to simplify reasoningprocedures in order to make them universally binding, without any need for further empiricaljustification other than the premises themselves, opened a genuine possibility of developingstrictly formal logic, in the sense of being independent from circumstances: mereunderstanding of logical rules allows us to infer correctly, without taking recourse toempirical verification or falsification of the logical structure, because the validity ofconclusion relies solely on the premises and not on any other factor external to them. Animportant question here is whether the step taken by the Jaina thinkers was prompted by theirgenuine awareness of what analytically valid inference rules – or more generally analytictruths – are or whether it was an expression of their insightful observation that in ourreasonings we synthesize or extrapolate from repeated experience and observation to such asubtle degree that we are ultimately able to relate pure concepts without taking any recourseto external reality. On the other hand, after Willard Van Orman QUINE’s (1951) famouscriticism, the very analytic-synthetic distinction may seem questionable, hence whether weconsider the new Jaina approach in logic a move towards genuine discovery of the idea ofanalytic truths or a way to better understand the nature of processes through which wesynthesize our individual experiences and apply the synthetic thinking abstracted from ourexperiences to logic losses its importance.Would the move towards economising inference procedures by accepting merely two steps(members) in a proof formula, i.e. the thesis and the logical reason, led the Jainas to dispensewith positive and negative examples altogether? Not really. They maintained that theremaining members of a proof formula can be accepted as long as they fulfil a practical ordidactic role of instruction, although in terms of logic they are perfectly dispensable. For

23

purposes of instruction, rhetoric or persuasion the Jainas even advise to apply a ten-memberedproof formula (NAV 13.2).To pay justice to Buddhist tradition, what Pâtrasvâmin and Siddhasena Mahâmati, followed byAkalaóka, did was rather an inspired verbalisation of what had already been somehow latent inDharmakîrti’s works (e.g. PV1 3.27, NB 3.122), although he himself had hold it unthinkableto do dispense with the example entirely: for all practical purposes, he had considered it anintegral part of the logical reason (hetu).

6. In search of the real world

As an important step towards realistic ontology, Akalaóka improves on the idea, introducedby Siddhasena Mahâmati, that the distinguishing factor for the twofold typology of cognitivecriteria is ‘the manner of grasping an object’ (see above p. 13), and draws its consequences forJaina ontology. His suggestion (LT 3–4, 46, NVi 3.83, PSa 1.2, SVi 1.19), as briefly indicatedabove (p. 17), is that it is the defining feature of perception is its clarity (vaiœadya), or lucidity(spašþatva), whereas that of the indirect cognition is its opposite, and that position is upheldby subsequent Jaina tradition, e.g. by Hemacandra-sûri (1088–1172) in his Pramâòa-mîmâôsâ (PMî 1.13–14).An analysis (LT 5) of the nature of perception, especially its first non-conceptual stage,sensation, ‘which immediately follows the conjunction of the datum and the datum-faculty’,brings him to a conclusion that, irrespective of realist or idealist stance, this conjunctionentails the existence – either real, objective, or mental, subjective – of both the datum as theobject and the ‘datum-faculty, i.e. the sense organ. Whether analysed in terms of objectiverelation (Jaina approach), or as a conjunction of elementary constituents of reality (dharma)that jointly make up an ‘objective basis’ (âyatana) for consciousness (Abhidharmicapproach), or as mental representations in the shape of internal sense-data (inphenomenological approach of Diónâga), the experience of perception reveals a complexnature of any act of cognition or of any object of cognition, or of both. Whether thedifferentiation into datum and datum-faculty were real or merely mental which becomes acontents of each and every cognitive act reflecting objective reality, the fact remains that suchmultiplexity becomes an inherent part of it. Similarly, any contents of cognition reveals thesubject-object division which points to the same conclusion, viz. its multiplex character.Further (LT 23), even consciousness which is not oriented towards external objects but is acase of consciousness in a meditative state will have the same multiplex nature: in one breath‘it is manifested as conceptual (determinate)’ (because it has a determinate contents) and ‘itmanifests its contents clearly’, i.e. it is a case of introspective perception that is by definitionperceptual and non-conceptual (because it is introspective self-illuminating cognition thatdoes not require still another cognition in order to be self-conscious).At all levels of analysis, claims Akalaóka, we always find that a resultant cognition ismultiplex (anekânta), at least in the sense of subject-object distinction, as well as intentional,i.e. it points towards its object.The nature of the object which is represented in cognition is likewise multiplex, on theprinciple of the representational idea of truth: a cognition is true if it represents its objectcorrectly, i.e. the relations within the elements of the object-image in cognition correspond toobjective relations. Since all objects as represented in cognition, as well as all contents ofcognitive acts are multiplex, the real objects, independent of cognition, must also bemultiplex. Now, inasmuch as nothing is ever entirely different (bhedÎkânta) or entirely

identical (abhedÎkânta), the object is not merely established, but it is established asmultiplex, i.e. as the opposite of the absolute difference and of the absolute identity.This multiplex structure of every act of cognition, that always has subjective and objectiveaspects, requires an external factor against which one can judge its truth value; otherwise,‘what would be the distinction between the truth and falsehood?’ (LT 28). Since cognitive

24

acts are differentiated into true and false or into valid (pramâòa) and invalid (apramâòa)cognitive procedures, there must be some objective criterion against which one could judgethe status of an act of cognition and this must lie outside of the cognition itself, for ‘Wherewould the status of the proof and its opposite (sc. defective proof) come from?’ (LT 28).Akalaóka specifies that a cognition (or a statement as a verbalisation of a mental event) is notself-reflexive when it comes to its truth value but depends on something outside of it. In theargument, Akalaóka (LT 28) applies the logical reason called ‘inexplicability otherwise’(anyathânupapatti), also in its negative formulation: ‘For not all intentional attitudesconcerning human goals deviate from objects, because otherwise absolute deviance of wordsand objects would be possible.’ A similarly objective, non-reflexive point of reference isrequired even in the case of the hearer’s proper interpretation of the speaker’s words whichare to convey the speaker’s intention: the meaning is interpreted against the intentional object.Further, even the Buddhists require an objective point of reference ‘to determine whether theBuddha or anyone else is an authoritative person or not or whether a logical formula is well-formed or not.’ Objective, of course, in a relative sense: as external to the cognition of it. Allsuch decisions require a criterion that lies in a meta-level.By appealing to the analysis of the nature of the data and cognition, Akalaóka claims thatBuddhism idealism is impossible because it is logically self-abortive. Suppose (LT 7) that thereality ultimately (paramârthatas) consists in consciousness, which is maintained by theBuddhists to be one and homogenous, free of the subject-object division. That being the case,one can, for practical reasons, distinguish a valid cognitive procedure (pramâòa) and theresultant cognition (phala). If these belonged to one and the same, ultimately homogenousconsciousness the fallacy would follow that one cognises what has already been cognised andsuch a cognition no longer fulfils the criterion – after the school of Mîmâôsâ, recognised alsoby the Buddhists of Diónâga’s and Dharmakîrti’s school and by the Jainas of Siddhasena’sand Akalaóka’s traditions – of a valid cognition (pramâòa) that is supposed to contributesomething cognitively new. ‘Therefore,’ concludes Akalaóka, ‘an object the essence of whichare substance and modes can be both external and internal in accordance with truth’ (LT 7).Thus, to recognise both the objectivity of the external world and the existence of a real subjectcognising it is indispensable for epistemological enterprise.What stood in the background of Akalaóka’s endeavour to construct realistic ontology on thebasis of his analysis of the structure of cognition was the doctrine of multiplexity of reality(anekânta-vâda), perhaps most spurious and stimulating theory the Jainas ever developed. Iteventually led to another revolutionary development in Jaina epistemology: an enquiry intothe multiplexity of language, prompted by an intuition that no statement is ever complete.Since the reality is so complex and multifaceted, no sentence can reflect it on its own and is assuch ambiguous. Every sentence stands in need of ‘semantic completion’. In other words, theusers of language supply the missing information from the context. The philosopher’s tasknow becomes to develop adequate tools that should disambiguate the language and, whatfollows, suitable procedures of reasoning that would be more context-dependent. Theseimprovements are reflected in Jaina theories of the seven-fold modal description (syâd-vâda)and of viewpoints (naya-vâda), as integral components of the theory of multiplexity of reality(anekânta-vâda). The complexity of these theories was the price to pay for an attempt to drawa consistent realist epistemology that would reflect the diversity of phenomena (prapañca)observed in the world, the ultimately objective status of which Buddhists denied.

BIBLIOGRAPHY

Primary Sources

25

A = Pâòini: Ašþâdhyâyî. Sumitra Mangesh Katre: Ašþâdhyâyî of Pâòini in RomanTransliteration. University of Texas Press, Austin 1987 [reprinted: MotilalBanarsidass, Delhi 1989].

AK = Vasubandhu: Abhidharma-koœa. P. Pradhan (ed.): Abhidharm-koshabhâšya ofVasubandhu. Tibetan Sanskrit Works Series 8, Bauddha Bharati, Patna 1967. [2ndrevised edition: A. Haldar: Bauddha Bharati, Patna 1975].

AKBh = Vasubandhu: Abhidharma-koœa-bhâšya. See: AK.ÂMî = Samantabhadra: Âpta-mîmâôsâ. (1) Vaôœîdhar (ed.): Ašþasahasrî

tarkikacakracûðâmaòisyâdvâdavidyâpatinâ œrîvidyânandasvâminâ nirâkåta. Niròaya-sâgara Press, Bombay 1915. (2) Nagin Shah (ed., transl.): Samantabhadra’sÂptamîmâôsâ. Critique of an Authority [Along with English Translation, Introduction,Notes and Akalaóka’s Sanskrit Commentary Ašþaœatî]. Sanskritsanskriti Granthamâlâ 7,Ahmedabad 1999.

AòD = Aòuoga-ddârâiô (Anuyoga-dvârâòi). See: NaôS.AŒ = Kauþilya: Artha-œâstra. R.P. Kangle (ed.): The Kauþilîya Arthaœâstra. Part I: A Critical

Edition with a Glossary. Part II: An English Translation with Critical andExplanatory Notes. Part III: A Study. University of Bombay, Bombay 1960, 1963,1965 [reprinted: Motilal Banarsidass, Delhi 1986–1988].

BÂU = Båhad-âraòyaka-upanišad. In: Patrick Olivelle: The Early Upanišads. Annotated Textand Translation, Oxford University Press, Oxford–New York 1998: 29–165.

CarS = Caraka-saôhitâ [vi = Vimâna-sthâna]. Vaidya Jâdavaji Trikamji Âchârya (ed.): TheCharakasaòhitâ of Agniveœa. Revised by Charaka and Dåidhabala. With theÂyurveda-Dîpikâ Commentary of Chakrapâòidatta. 3rd edition, Niròaya Sagar Press(Niròaya-sâgara Pres), Bombay (Muôbaî) 1941 [reprinted: Munishiram ManoharlalPublishers, New Delhi 1981].

LT = Akalaóka: Laghîyas-traya. Nyâyâcârya Mahendra Paòðita Kumâr Œâstri (ed.): Œrîmad-

Bhaþþâkalaóka-deva-viracitam Akalaóka-grantha-trayam [Svôpajña-vivåti-sahitamLaghîyas-trayam, Nyâya-viniœcayaÿ, Pramâòa-saógrahaœ ca]. Sarasvatî PustakBhaòðâr, Ahmadâbâd (Ahmedabad) 1996 [1. ed.: Ahmedabad–Calcutta 1939].

MBh = Mahâ-bhârata. (1) V.S. Sukthankar et al. (ed.): The Mahâbhârata for the first timecritically edited. Bhandarkar Oriental Research Institute, Poona 1933–1966.

MBhâ = Patañjali: Vyâkaraòa-mahâ-bhâšya. Bhârgava Œâstri Bhikâjî Joœi, Œivadatta Œarman,Raghunâth Kâœînâth Œâstri (ed.): Œrîman-maharši-Patañjali-nirmitam Vyâkaraòa-mahâ-bhâšyam, œrîmad-upâdhyâya-Kaiyaþa-praòîtena Bhâšya-pradîpena œrîman-Nâgeœabhaþþa-viracita-Bhâšyapradîpôddyotôdbhâsitena Samullasitena. 6 Vols.,Chaukhambâ Saôskåta Pratishthâna, Dilli (Delhi) 2000 [reprinted].

MŒV = Kumârila Bhaþþa: Mîmâôsâ-œloka-vârttika. (1) Dvârikâdâsa Œâstrin (ed.):Œlokavârttika of Œrî Kumârila Bhaþþa with the Commentary Nyâya-ratnâkara of ŒrîPârthasârathi Miœra. Prâchyabhârati Series 10, Târâ Publications, Varanasi 1978.(2) S.K. Râmanâtha Œâstrî (ed.): Œlokavârttikavyâkhyâ Tâtparyaþîkâ of UôvekaBhaþþa. Revised by K. Kunjuni Raja & R. Thangaswamy, University of Madras,Madras 1971.

NA = Siddhasena Mahâmati: Nyâyâvatâra. Edited and translated: Piotr Balcerowicz: JainaEpistemology In Historical And Comparative Perspective - A Critical Edition And AnAnnotated Translation Of Siddhasena Mahâmati’s Nyâyâvatâra, Siddharšigaòin’s

Nyâyâvatâra-vivåti And Devabhadrasûri’s Nyâyâvatâra-þippana. Volume I & II.Second revised edition. Motilal Banarsidass, Delhi 2008 [first edition: JainaEpistemology In Historical And Comparative Perspective - A Critical Edition And AnAnnotated Translation Of Siddhasena Divâkara’s Nyâyâvatâra, Siddharšigaòin’s

Nyâyâvatâra-vivåti And Devabhadrasûri’s Nyâyâvatâra-þippana. Volume I & II. Alt-und Neu-Indische Studien 53,1 & 53,2. Franz Steiner Verlag, Stuttgart 2001].

26

NaôS = Naôdi-sutta / Naôdî-sutta [Nandi-sûtra / Nandî-sûtra]. Muni Puòyavijaya, DalsukhMâlvaòiâ, Amritlâl Mohanlâl Bhojak (ed.): Nandi-suttaô and Aòuoga-ddârâiô.Jaina-Âgama-Series No.1, Shri Mahâvîra Jaina Vidyâlaya, Bombay 1968.

NAV = Siddharši-gaòin: Nyâyâvatâra-vivåti. See: NA.NB = Dharmakîrti: Nyâya-bindu. Theodor Stcherbatsky (ed.): Nyâya-bindu-þîkâ of

Dharmottara with Nyâya-bindu of Dharmakîrti. Bibliotheca Buddhica 7, St.Pétersbourg 1918 [reprinted: Biblio Verlag, Osnabrück 1970].

Nir = Yâska: Nirukta. H.M. Bhadkamkar, R.G. Bhadkamkar (eds.): The Nirukta of Yâska. 2Vols., Reprinted, Bhandarkar Oriental Research Institute, Poona 1985 [first edition:1918, 1942].

NS = Akšapâda Gautama: Nyâya-sûtra. (1) Anantalal Thakur (ed.): Gautamîyanyâyadarœanawith Bhâšya of Vâtsyâyana. Indian Council of Philosophical Research, New Delhi1997. (2) Gangopadhyaya, Mrinalkanti; Chattopadhyaya, Debiprasad (tranl.): NyâyaPhilosophy. Literal Translation of Gautama’s Nyâya-sûtra with Vâtsyâyana’sBhâšya. 5 Parts. Indian Studies (Indian Philosophy in its Sources), Firma KLMPrivate Limited, Calcutta 1967, 1968, 1972, 1973, 1976.

NVi = Akalaóka: Nyâya-viniœcaya. See: LT, NViV.NViV = Vâdirâja-sûri: Nyâya-viniœcaya-vivaraòa. Mahendra Kumar Jain (ed.):

Nyâyaviniœcaya-vivaraòa of Œrî Vâdirâja Sûri, the Sanskrit Commentary on BhaþþaAkalaókadeva’s Nyâyaviniœcaya. Vol. 1 & 2, Bhâratîya Jñânapîþha Prakâœana,Vârâòasî 1949, 1955. [2nd edition: Bhâratîya Jñânapîþha Prakâœana, New Delhi 2000].

Paòò = Paòòavaòâ-sutta. Muni Puòyavijaya, Dalsukh Mâlvaòiâ, Amritlâl Mohanlâl Bhojak(eds.): Paòòâvaòâsuttam. Jaina-Âgama-Series No. 9, Parts 1–2, Shri Mahâvîra JainaVidyâlaya, Bombay 1969–1971.

PMî = Hemacandra-sûri: Pramâòa-mîmâôsâ. Satkari Mookerjee and Nathmal Tatia (transl.,ed.): Hemacandra’s Pramâòa-mîmâôsâ. Text and Translation with Critical Notes.Tara Publications, Varanasi 1970.

PS = Diónâga: Pramâòa-samuccaya. (1) Steinkellner, Ernst (ed.): Dignâga’sPramâòasamuccaya, Chapter 1. A hypothetical reconstruction of the Sanskrit textwith the help of the two Tibetan translations on the basis of the hitherto knownSanskrit fragments and the linguistic materials gained from Jinendrabuddhi’s Þîkâ.Österreichische Akademie der Wisssenschaften, Wien 2005 (April):www.oeaw.ac.at/ias/Mat/dignaga_PS_1.pdf (2) HATTORI (1968).

PSa = Akalaóka: Pramâòa-saógraha, See: LT.PV = Dharmakîrti: Pramâòa-vârttika. (1) Râhula Sâókåtyâyana (ed.): Pramâòavârttikam

Âcârya-Manorathanandi-kåtayâ våttyâ saôvalitam [Dharmakîrti’s Pramâòavârttikamwith a commentary by Manorathanandin]. Parts I–II, Appendix to Journal of the Biharand Orissa Research Society 24–26, Patna 1938–1940. (2) Dalasukha Mâlavaòiyâ(ed.): Âcârya Dharmakîrti kåta Svârthânumâna-pariccheda. General Edition: V.S.Agarwala, Hindu Vishvavidyalaya Nepal Rajya Sanskrit Series 2, HinduVishvavidyalaya Sanskrit Publication Board, Varanasi 1959.

PVSV = Dharmakîrti: Pramâòa-vârttika-svavåtti. See: (4) PV2.RNA = Anantalal Thakur (ed.): Ratnakîrti-nibandhâvalî (Buddhist Nyâya Works of

Ratnakîrti). Tibetan Sanskrit Works Series 3, K.P. Jayaswal Research Institute, Patna1975.

RNA = Ratnakîrti-nibandhâvalî. Anantalal Thakur (ed.): Ratnakîrti-nibandhâvalî (BuddhistNyâya Works of Ratnakîrti). Tibetan Sanskrit Works Series 3; K.P. Jayaswal ResearchInstitute, Patna 1975.

SSi = Pûjyapâda Devanandin: Sarvârtha-siddhi. (1) Phoolchandra Shastri (ed.): ÂcâryaPûjyapâda’s Sarvârthasiddhi [The commentary on Âcârya Griddhapiccha’sTattvârtha-sûtra]. Edited and translated [into Hindi]. Varanasi 1934 [reprinted:Jñânapîþha Mûrtidevî Jaina Grantha-mâlâ 13, Bhâratîya Jñânapîþha Prakâœana,Delhi 2000]. (2) Jinadasa Œastri (ed.): Tattvârtha-våttiÿ Sarvârtha-siddhiÿ—

27

Pûjyapâda Devanandin. Œrîsekhârâma Nemicandra Granthâmâlâ 128, DevajîSakhârâm Diœî & Mâòikacandra-digambara-jaina-parîkšâlaya-mantri, Šodaœapur1939. (3) English translation: JAIN (1960).

STP = Siddhasena Divâkara: Saômati-tarka-prakaraòa. (1) Sukhlâl Saóghavi; Becardâs Doœi(ed.): Saômatitarka-prakaraòam by Siddhasena Divâkara with Abhayadevasûri’sTattva-bodha-vidhâyinî. Gujarât-purâ-tattva-mandir-granthâvalî 10, 16, 18, 19, 21,Gujarât-purâ-tattva-mandir, Amdâvâd 1924–1931 [reprinted: 2 Vols., RinsenBuddhist Text Series VI–1, 2; Kyoto 1984].

SVi = Akalaóka Bhaþþa: Siddhi-viniœcaya. Mahendrakumâr Jain (ed.): Siddhi-viniœcaya ofAkalaóka edited with the commentary Siddhi-viniœcaya-þîkâ of Anantavîrya. 2 Vols.Bhâratîya Jñânapîþha Prakâœana, Vârâòasî 1959.

SvSt = Samantabhadra: Svayambhû-stotra. (1) Jugal Kiœor Mukhtâr ‘Yugavîra’ (ed., introd.):Œrîmat-svâmi-Samantabhadrâcârya-viracita catur-viôœati-jina-stavanâtmaka

Svayambhû-stotra (stuti-paraka jainâgama). Vira-sevâ-mandira-grantha-mâlâ 7,Vira-sevâ Mandira, Sarasâvâ Jilâ Sahâranapura 1951. (2) Pannâlâl Jain (ed.): Âcâryasamantabhadra viracita Svayambhû-stotra [Saôskåta-þîkâ œrî Prabhâcandrâcârya]. 2ed., Bhâratavaršîya Anekânta Vidvat Parišad Pušpa Saôkhyâ 31, BhâratavaršîyaAnekânta Vidvat Parišad, Vârâòasî 1995.

SvStÞ = Prabhâcandra: Svayambhû-stotra-þîkâ. See: SvSt2.TBh = Umâsvâti: Tattvârthâdhigama-bhâšya. See: TS1.

Þhâò = Þhâòaôga-sutta [Sthânâóga-sûtra]. Muni Jambuvijaya (ed.): Þhânaôgasuttam andSamavâyaôgasuttam. Jaina-Âgama-Series 3, Shri Mahâvîra Jaina Vidyâlaya,Bombay 1985.

TS = Umâsvâmin: Tattvârtha-sûtra. (1) M.K. Mody (ed.): Tattvârthâdhigama by Umâsvâtibeing in the Original Sanskrit with the Bhâšya by the author himself. BINS 1044,1079, 1118, Calcutta 1903, 1904, 1905. [= Œvetâmbara Recension]. (2) See: SSi.[= Digâmbara Recension]. (3) See: RVâr. [= Digâmbara Recension]. (3) Englishtranslation: JAINI (1920) and TATIA (1994).

TSa = Œântarakšita: Tattva-saógraha. (1) Embar Krishnamacharya (ed.): Tattvasaógraha ofŒântarakšita with the commentary of Kamalaœîla. 2 Vols., Gaekwad’s Oriental Series30–31, Oriental Institute, Baroda 1926 [reprinted: 1984, 1988]. (2) Dvarikadas Shastri(ed.): Tattvasaógraha of Âcârya Shântarakšita with the Commentary ‘Pañjikâ’ of ShriKamalashîla. 2 Vols. Bauddha Bharati Series 1, Bauddha Bharati, Vârâòasî 1981–1982.

Uttar = Uttarajjhayaòa-sutta [Uttarâdhyayana-sûtra]. (1) Muni Puòyavijayaji and Pt.Amritlâl Mohanlâl Bhojak (ed.): Dasaveyâliyasuttaô, Uttarajhayaòâiô andÂvassayasuttaô. Jaina-Âgama-Series 15, Œrî Mahâvîra Jaina Vidyâlaya, Bombay1977. (2) English translation: JACOBI (1895).

Viy = Viyâha-pannatti / Viyâha-paòòatti / Bhagavaî-viyâha-paòòaþþî [Bhagavatî Vyâkhyâ-prajñapti / Bhagavatî-sûtra]. (1) Bechardas J. Doshi; Amritlal Mohanlal Bhojak (ed.):Viyâhapaòòattisuttaô. 3 parts, Jaina-Âgama-Series 4, Œrî Mahâvîra Jaina Vidyâlaya,Bombay 1974–1982. (2) K.C. Lalwani (ed., transl.): Sudharma Svâmî’s BhagavatîSûtra. Vol. I, Œatakas 1–2, Prakrit Text with English Translation and Notes, JainBhawan, Calcutta 1973.

Secondary Sources

ALSDORF 1973 = Alsdorf, Ludwig: ‘Nikšepa – a Jaina Contribution to ScholasticMethodology’, Journal of the Oriental Institute (Baroda) 22 (1973) 455–463[reprinted in: Ludwig Alsdorf: Kleine Schriften. Ed. Albrecht Wezler, Band X, FranzSteiner Verlag, Wiesbaden 1974: 257–265].

28

Balcerowicz Piotr (2000) “On the Date of the Nyâyâvatâra”, in: Piotr Balcerowicz, MarekMejor (ed.): On the Understanding of Other Cultures. Proceedings of theInternational Conference on Sanskrit and Related Studies to Commemorate theCentenary of the Birth of Stanis³aw Schayer (1899–1941). Warsaw University,Poland, October 7–10, 1999. Studia Indologiczne 7 (2000) 17–57 [reprinted in: PiotrBalcerowicz and Marek Mejor (eds.), Essays in Indian Philosophy, Religion andLiterature. Motilal Banarsidass, Delhi 2004: 313–352].

Balcerowicz Piotr (2001) ‘Two Siddhasenas and the Authorship of the Nyâyâvatâra and theSaômati-tarka-prakaraòa’, Journal of Indian Philosophy 29/3 (2001) 351–578.

Balcerowicz Piotr (2003a) “On the Relationship of the Nyâyâvatâra and the Saômati-tarka-prakaraòa,” Indologica Taurinensia [Proceedings of the XIth World SanskritConference (Turin, April, 3rd-8th, 2000)] 29 (2003) 31–83.

Balcerowicz Piotr (2003b) ‘Is “Inexplicability Otherwise” (anyathânupapatti) OtherwiseInexplicable?’, Journal of Indian Philosophy [Kluwer Academic Publishers] 1–3 (2003)343–380.

Balcerowicz Piotr (2005a) ‘Akalaóka und die buddhistische Tradition: Von derNichtwahrnehmung von Unsichtbarem (adåœyânupalabdhi) zu Allwissenheit’, WienerZeitschrift für die Kunde Südasiens 49 (2005) 151–226.

Balcerowicz Piotr (2005b) ‘Pramâòas and language. A Dispute between Diónâga,Dharmakîrti and Akalaóka’, Journal of Indian Philosophy 33/4 (2005) 343–400.

Balcerowicz Piotr (2008) Jaina Epistemology In Historical And Comparative Perspective - ACritical Edition And An Annotated Translation Of Siddhasena Mahâmati’sNyâyâvatâra, Siddharšigaòin’s Nyâyâvatâra-vivåti And Devabhadrasûri’s

Nyâyâvatâra-þippana. Volume I & II. Second revised edition. Motilal Banarsidass,Delhi 2008 [first edition: Jaina Epistemology In Historical And ComparativePerspective - A Critical Edition And An Annotated Translation Of SiddhasenaDivâkara’s Nyâyâvatâra, Siddharšigaòin’s Nyâyâvatâra-vivåti And Devabhadrasûri’s

Nyâyâvatâra-þippana. Volume I & II. Alt- und Neu-Indische Studien 53,1 & 53,2.Franz Steiner Verlag, Stuttgart 2001].

Balcerowicz Piotr (forthcoming) ‘Dharmakîrti’s criticism of the Jaina doctrine of multiplexityof reality (anekânta-vâda), to be published in the Proceedings of the FourthInternational Dharmakirti Conference, 23-27.08.2005, Universität Wien, Vienna,Austria.

BHATT 1974 = Bhatt, Bansidhar: The Canonical Nikšepa. Studies in Jaina Dialectics.Indologia Berolinensis 5, E.J. Brill, Leiden 1978 [reprinted with a foreword by K.Bruhn & H.Haertel: Bharatiya Vidya Prakashan, Delhi–Varanasi 1991].

DANDEKAR 1941 = Dandekar, R.N.: “Somatism of Vedic Psychology”, Indian HistoricalQuarterly 17 (1941) 70–76.

FRANCO 1997 = Franco, Eli: Dharmakîrti on Compassion and Rebirth. Arbeitskreis fürtibetische und buddhistische Studien Universität Wien, Wien 1997.

FRAUWALLNER 1933 = Frauwallner, Erich: ‘Zu den Fragmenten buddhistischer Logiker imNyâyavârttikam’, Wiener Zeitschrift für die Kunde Morgenlandes 40 (1933) 300–304[reprinted: Erich Frauwallner: Kleine Schriften, hrsg. von Gerhard Oberhammer undErnst Steinkellner, Glassenapp-Stiftung 22, Franz Steiner Verlag, Wiesbaden 1982:479–483].

FRAUWALLNER 1957 = Frauwallner, Erich: ‘Vasubandhu’s Vâdaviddhi’. Wiener Zeitschriftfür die Kunde Süd- und Ostasiens (1957) 104–146 [reprinted: Erich Frauwallner:Kleine Schriften, hrsg. von Gerhard Oberhammer und Ernst Steinkellner, Glassenapp-Stiftung 22, Franz Steiner Verlag, Wiesbaden 1982: 716–758].

FUJINAGA 2003 = Fujinaga, Sin: ‘review of: Piotr Balcerowicz, Jaina Epistemology InHistorical And Comparative Perspective - A Critical Edition And An AnnotatedTranslation Of Siddhasena Mahâmati’s Nyâyâvatâra, Siddharšigaòin’s Nyâyâvatâra-

29

vivåti And Devabhadrasûri’s Nyâyâvatâra-þippana. Stuttgart 2001’, Saôbhâšâ –Nagoya Studies in Indian Culture and Buddhism 23 (2003) 163–165.

HATTORI 1968 = Hattori, Masaaki (ed., transl.): Dignâga, On Perception, being thePratyakša-pariccheda of Dignâga’s Pramâòa-samuccaya, Edition of Tibetantranslations and the Sanskrit text as well as the English translation of the Chapter I.Harvard University Press, Cambridge (Massachusetts) 1968.

HAYES 1988 = Hayes, Richard P.: Dignâga on the Interpretation of Signs. Studies ofClassical India 9, Kluwer Academic Press, Dordrecht 1988.

HERZBERGER 1986 = Herzberger, Radhika: Bhartåhari and the Buddhists. Studies of ClassicalIndia 8, D. Reidel Publishing Company (Kluwer Academic Publishers), Dordrecht1986.

JACOBI 1895 = Jacobi, Hermann: Jaina Sûtras, Part II: Uttarâdhyayana Sûtra andSûtrakåtâóga. Sacred Books of the East 45, Oxford University Press, Oxford 1895[reprinted: Motilal Banarsidass, Delhi 1999].

JAIN 1960 = Jain, S.A.: Reality—English Traslation of Sarvârtha-siddhi. Vira Sasana Sangha,Calcutta 1960 [reprinted: Jwalamalini Trust, Madras 1992].

JAINI 1920 = Jaini, J.L.: Tattvarthadhigama Sutra (A Treatise on the Essential Principles ofJainism). The Sacred Books of the Jainas 2, Kumar Kendra Prasada – The CentralJaina Publishing House, Arrah 1920.

KATAOKA 2003 = Kataoka, Kei: ‘The Mîmâôsâ Definition of Pramâòa as a Source of NewInformation’, Journal of Indian Philosophy 31/1–3 (2003) 89–103 [Proceedings ofthe International Seminar ‘Argument and Reason in Indian Logic’ 20–24 June, 2001– Kazimierz Dolny, Poland, edited by Piotr Balcerowicz].

KATSURA 1984 = Shoryu Katsura: ‘Dharmakîrti’s Theory of Truth’, Journal of IndianPhilosophy 12 (1984) 215–235.

KELLNER 1999 = Kellner, Birgit: ‘Levels of (im)perceptibility. Dharmottara on the dåœya indåœyânupalabdhi’, in:Shoryu Katsura (ed.): Dharmakîrti’s Thought and its Impact onIndian and Tibetan Philosophy. Proceedings of the Third International DharmakîrtiConference, Hiroshima, November 4–6, 1997. Verlag der Österreichischen Akademieder Wissenschaften: Beiträge zur Kultur- und Geistesgeschichte Asiens Nr. 32, Wien1999: 193–208.

KELLNER 2003 = Kellner, Birgit: ‘Integrating Negative Knowledge into pramâòa Theory: theDevelopment of the dåœyânupalabdhi in Dharmakîrti’s Earlier Works’, Journal ofIndian Philosophy 31/1–3 (2003) 121–159 [Proceedings of the International Seminar‘Argument and Reason in Indian Logic’ 20–24 June, 2001 – Kazimierz Dolny,Poland, edited by Piotr Balcerowicz].

KUNST 1968 = Kunst, Arnold: “Somatism: A Basic Concept in India’s PhilosophicalSpeculations”, Philosophy East and West 18 (1968) 261–275.

PADMARAJIAH 1986 = Padmarajiah, Y.J.: A Comparative Study of Jaina Theories of Realityand Knowledge. Motilal Banarsidass, Delhi Varanasi –Patna – Madras 1986 [firstedition: Delhi 1963].

Quine Willard Van Orman (1951) ‘Two Dogmas of Empiricism’, The Philosophical Review60 (1951) 20–43 [Reprinted in: Willard Van Orman Quine: From a Logical Point ofView, Harvard University Press, Harvard 1953].

RANDLE 1926 = Randle, H.N.: Fragments from Dignâga. Prize Publication Fund, RoyalAsiatic Society, London 1926. [Reprinted: Motilal Banarsidass, Delhi 1981].

RUSSELL 1912 = Russell, Bertrand: ‘Knowledge by Acquaintance and Knowledge byDescription’, in: Russell, Bertrand: The Problems of Philosophy. Oxford UniversityPress, Oxford 1912: 72–92.

RUSSELL 1917 = Russell, Bertrand: ‘Knowledge by Acquaintance and Knowledge byDescription’, in: Russell, Bertrand: Mysticism and Logic. Unwin Books, London1917: 152–167.

30

SHAH 1967 = Shah, Nagin J.: Akalaóka’s Criticism of Dharmakîrti’s Philosophy—A Study.Lalbhai Dalpatbhai Series 11, Bharatiya Sanskrit Vidyamandir – L.D. Institute ofIndology, Ahmedabad 1967.

SHASTRI 1990 = Shastri, Indra Chandra: Jaina Epistemology, P. V. Research Series No. 50,P.V. Research Institute, Varanasi 1990.

SINGH 1984 = Singh, Amar: The Heart of Buddhist Philosophy—Dióâga and Dharmakîrti.Munshiram Manoharlal, New Delhi 1984.

STEINKELLNER 1997 = Steinkellner, Ernst: ‘Kumârila’s Îœvarasena, and Dharmakîrti inDialogue. A New Interpretation of Pramâòâvârttika I 33’, in: Petra Kieffer-pülz, Jens-Ue Hartmann(eds.): Bauddhavidyâsudhâkâraÿ: Studies in Honour of Heinz Becherton the Occasion of His 65th Birthday. Swisstal-Odendorf 1997: 625-646.SCHAYER 1936 = Schayer, Stanis³aw: “Über den Somatismus der indischenPsychologie”, Bulletin International de l’Academie Polonaise des Sciences et desLettres, Classe de Philologie (Cracow) 7–10 (1936) 159–168. [Reprinted in:Stanis³aw Schayer: O filozofowaniu Hindusów. Artyku³y wybrane. Ed. by MarekMejor. Polska Akademia Nauk, Komitet Nauk Orientalistycznych, PWN, Warszawa1988: 495–504].

STEINKELLNER–KRASSER 1989 = Steinkellner, Ernst; Krasser, Helmut: Dharmottaras Exkurszur Definition gültiger Erkenntnis im Pramâòa-viniœcaya. Beiträge zur Kultur- undGeistesgeschichte Asiens No.2, Österreichische Akademie der Wissenschaften, Wien1989.

SUKHLÂL 1942 = Sukhlâl Saóghavi: ‘Introduction’, in: Sukhlâl Saóghavi; Dalsukh D.Mâlvaòiyâ; H.K. Devi (eds.): Jñâna-bindu-prakaraòa of Upâdhyâya Yaœovijaya.Singhi Jain Series 16, Bharatîya Vidyâ Bhavan, Ahmedabad 1942.

VATSYAYAN 1999 = Vatsyayan, Kapila: ‘Indriya’, in: Advaitavadini Kaul; SukumarChattopadyay (eds.): Kâlâtattvakoœa. A Lexicon of Fundamental Concepts of theIndian Arts. Vol. IV: Manifestation of Nature, Såšþi Vistâra. Indira Gandhi NationalCentre for the Arts and Motilal Banarsidass, Delhi 1999: 1–68.

APPENDIXTypology of Cognitions

Model I:(1) sensuous cognition (abhinibodha, âbhinibodhika-jñâna, mati-jñâna),(2) testimonial cognition (œruta-jñâna),(3) clairvoyance (avadhi-jñâna),(4) telepathy (manaÿ-paryâya-jñâna),(5) omniscience (kevala-jñâna).

Model II:(1) sensuous cognition (abhinibodha, âbhinibodhika-jñâna, mati-jñâna),(a) sensation (avagraha),(b) cogitation (îhâ),(c) perceptual judgement (avâya—Digambaras, apâya—Œvetâmbaras),(d) retention (dhâraòâ).(2) testimonial cognition (œruta-jñâna),(3) clairvoyance (avadhi-jñâna),(4) telepathy (manaÿ-paryâya-jñâna),(5) omniscience (kevala-jñâna).

Model III:

31

(1) direct cognition (pratyakša),(2) indirect cognition (parokša).

Model IV:(1) direct cognition (pratyakša),(a) sensuous cognition (abhinibodha, âbhinibodhika-jñâna, mati-jñâna),(b) testimonial cognition (œruta-jñâna),(2) indirect cognition (parokša).(a) clairvoyance (avadhi-jñâna),(b) telepathy (manaÿ-paryâya-jñâna),(c) omniscience (kevala-jñâna).

Model V:cognitive application / cognitive faculties (upayoga):(1) cognition (jñâna), distinct cognitive act, having definite contents (sâkâra),(2) perceptual experience (darœana), indistinct cognitive act, not having any definite contents(anâkara, nirâkâra).

Model VI (4th c.):57

cognitive application (upayoga):(I) cognition (jñâna):(1) sensuous cognition (âbhinibodhika-jñâna),(a) sensation (avagraha),(b) cogitation (îhâ),(c) perceptual judgement (avâya),(d) retention (dhâraòâ).(2) testimonial cognition (œruta-jñâna),(3) clairvoyant cognition (avadhi-jñâna),(4) telepathy (manaÿ-paryâya-jñâna),(5) perfect cognition (kevala-jñâna), or omniscience,(II) perceptual experience (darœana):(1) ocular perceptual experience (cakšur-darœana),(2) non-ocular perceptual experience (acakšur-darœana),(3) clairvoyant perceptual experience (avadhi-darœana),(4) perfect perceptual experience (kevala-darœana), or omni-perception.

Model VII (4th c.):58 (I) cognition (jñâna),(1) direct cognition (pratyakša),(a) complete (omniscient) cognition (kevala),(i) possessed by souls still in the cycle of rebirth (bhavastha-kevala),(ii) possessed by liberated souls (siddha-kevala),(b) incomplete (non-omniscient) cognition (no-kevala),(i) clairvoyance (avadhi),(ii) telepathy (manaÿ-paryâya),(2) indirect cognition (parokša),(a) sensuous cognition (âbhinibodhika),(i) based on verbal cognition (œruta-niÿœrita),

57 Viy 8.2.22–23 (p. 336); Viy 2.10.9[2] (p.115); NaôS 6[28–29] and 8 (pp. 6, 9).58 Þhâò 59–60 (p. 14–15).

32

– (conceptual) sensation of the object (arthâvagraha)

– (non-conceptual) sensation of momentary manifestation (vyañjanâvagraha),(ii) independent of verbal cognition (aœruta-niÿœrita),– (conceptual) sensation of the object (arthâvagraha)

– (non-conceptual) sensation of momentary manifestation (vyañjanâvagraha),(b) testimonial cognition (œruta),included in primary canonical texts (aóga-pravišþa),not included in primary canonical texts (aóga-bahya).(II) perceptual experience (darœana),(1) correct perceptual experience (samyag-darœana),(a) inborn (nisarga),(b) acquired (abhigama),(2) erroneous perceptual experience (mithyâ-darœana),(a) persistent (abhigåhîta),(b) temporary (anabhigåhîta).

Model VIII/a (after 300):the cause of valid cognition (hetu):(1) perception (pratyakša),(2) inference (anumâna),(3) analogy (aupamya),(4) scriptural testimony (âgama).59

Model VIII/b (after 300):cognitive criterion (pramâòa):(1) perception (pratyakša),(2) inference (anumâna),(3) analogy (aupamya),(4) scriptural testimony (âgama).60

Model IX:61

cognitive criteria (pramâòa):(I) direct cognition (pratyakša),(1) sensuous cognition (abhinibodha, âbhinibodhika-jñâna, mati-jñâna),(a) sensation (avagraha),(b) cogitation (îhâ),(c) perceptual judgement (avâya—Digambaras, apâya—Œvetâmbaras),(d) retention (dhâraòâ).(2) twofold testimonial cognition (œruta-jñâna),(II) indirect cognition (parokša).(1) clairvoyance (avadhi-jñâna),(2) telepathy (manaÿ-paryâya-jñâna),(3) omniscience (kevala-jñâna).

59 Þhâò 336 (p. 149).60 Viy 5.4.26[3] (vol.1, p. 201.1–2).61 TS, TSBh, SSi 1.9–12.

33

direct cognition (pratyakša)

clairvoyance (avadhi-jñâna)

mind-reading (manaÿ-paryâya-jñâna)

absolute knowledge (kevala-jñâna)

indirect cognition (parokša)

sensuous cognition (mati-jñâna)

scriptural cognition (œruta-jñâna)

cognitive criteria (pramâòa)

sensation (avagraha)

cogitation (îhâ)

perceptual judgement (apâya)

retention (dhâraòâ)

Model X (late 4th c.):62

cognitive application / cognitive faculty (upayoga):(I) distinct cognitive faculties in the form of cognition (jñâna-sâkârôpayoga),[(1) three correct cognitive faculties in the form of cognition:](a) distinct cognitive faculty in the form of sensuous cognition (mati-jñâna-sâkârôpayoga),

(b) distinct cognitive faculty in the form of testimonial cognition (œruta-jñâna-sâkârôpayoga),(c) distinct cognitive faculty in the form of clairvoyant cognition (avadhi-jñâna-sâkârôpayoga),[(2) three indistinct cognitive faculty in the form of erroneous cognition:](d) distinct erroneous sensuous cognition (maty-ajñânôpayoga),

(f) distinct erroneous testimonial cognition (œrutâjñânôpayoga),

(g) distinct distorted supernatural cognition (vibhaóga-jñânôpayoga),

(II) indistinct cognitive faculties in the form of perceptual experience (darœanôpayoga),(a) indistinct cognitive faculties in the form of ocular perceptual experience (cakšur-darœanôpayoga),(b) indistinct cognitive faculties in the form of non-ocular perceptual experience (acakšur-darœanôpayoga),(c) indistinct cognitive faculties in the form of clairvoyant perceptual experience (avadhi-darœanôpayoga).

Model XI (ca. 350-400):63

cognitive application / cognitive faculty (upayoga):(I) cognitive faculties in the form of cognition (jñânôpayoga),[(1) five correct cognitive faculties in the form of correct cognition (samyaj-jñâna):](a) cognitive faculty in the form of sensuous cognition (mati-jñânôpayoga),

(b) cognitive faculty in the form of testimonial cognition (œruta-jñânôpayoga),

(c) cognitive faculty in the form of clairvoyant cognition (avadhi-jñânôpayoga),

(d) cognitive faculty in the form of telepathy (manaÿ-paryâya-jñânôpayoga),

(e) cognitive faculty in the form of perfect cognition (kevala-jñânôpayoga), or omniscience,[(2) three cognitive faculty in the form of erroneous cognition (viparyaya):]

62 Paòò 9.1912–1914.63 Paòò 9.1912–1914; TS, TSBh, SSi 2.8-9, which comprises the ideas of TS 1.9–12, 1.32 (SSi 1.31).

34

(f) erroneous sensuous cognition (maty-ajñânôpayoga),

(g) erroneous testimonial cognition (œrutâjñânôpayoga),

(h) distorted supernatural cognition (vibhaóga-jñânôpayoga),

(II) cognitive faculties in the form of perceptual experience (darœanôpayoga),

(a) ocular perceptual experience (cakšur-darœanôpayoga),

(b) non-ocular perceptual experience (acakšur-darœanôpayoga),

(c) clairvoyant perceptual experience (avadhi-darœanôpayoga),

(d) perfect perceptual experience (kevala-darœanôpayoga), or omni-perception.

perceptual experience(darœana)

cognition(jñâna)

cognitive faculty (upayoga)

erroneous sensuous cognition(maty-ajñâna)

erroneous testimonial cognition(œrutâjñâna)

distorted supernatural cognition(vibhaóga-jñâna)

correct cognition(samyaj-jñâna)

erroneous cognition(viparyaya)

sensuous cognition(mati-jñâna)

scriptural cognition(œruta-jñâna)

clairvoyance(avadhi-jñâna)

mind-reading(manaÿ-paryâya-jñâna)

absolute knowledge(kevala-jñâna)

ocular perceptual experience(cakšur-darœana)

non-ocular perceptual experience(acakšur-darœana)

clairvoyant perceptual experience(avadhi-darœana)

perfect perceptual experience(kevala-darœana)

Model XII (c. 450):64

valid cognitive measures of the soul’s qualities (jîva-guòa-pramâòa):(I) valid cognitive measure [of the soul’s] quality in the form of cognition (jñâna-guòa-pramâòa),(1) perception (pratyakša),(a) fivefold sensory perception (indriya-pratyakša):—auditory (œrotra),—ocular (cakšur),—olfactory (ghrâòa),—gustatory (jihvâ),—tactile (sparœa),(b) threefold extrasensory perception (no-indriya-pratyakša):—clairvoyance (avadhi-jñâna),—telepathy (manaÿ-paryâya-jñâna),—omniscience (kevala-jñâna),(2) threefold inference (anumâna),65

64 AòD 435–472 (p. 173–180).

35

(a) based on the antecedent (pûrvavat),(b) based on the consequent (œešavat): by effect (kârya), by cause (kâraòa), by quality (guòa),by virtue of a part (avayava), by substratum (âœraya),66

(c) based on the correspondence to/between observed facts (dåšþa-sâdharmyavat): factsobserved universally (sâmanya-dåšþa) and facts observed in particular cases (viœeša-dåšþa),(3) analogy (aupamya),67

(a) by similarity (sâdharmya),(b) by dissimilarity (vaidharmya),(4) scriptural testimony (âgama),68

(a) mundane (laukika),(b) supramundane (lokôttara),(II) valid cognitive measure [of the soul’s] quality in the form of perceptual experience(darœana-guòa-pramâòa),(1) ocular perceptual experience (cakšur-darœana-guòa-pramâòa),(2) non-ocular perceptual experience (acakšur-darœana-guòa-pramâòa),(3) clairvoyant perceptual experience (avadhi-darœana-guòa-pramâòa),(4) perfect perceptual experience (kevala-darœana-guòa-pramâòa),(III) valid cognitive measure [of the soul’s] quality in the form of conduct (câritra-guòa-pramâòa),(1) equanimity (sâmâyika-câritra-guòa-pramâòa),(2) reconfirmation (chedôpasthâpanîya-câritra-guòa-pramâòa),(3) purification by shedding off the karman (parihâra-viœuddhika-câritra-guòa-pramâòa),(4) conduct with subtle residue passions (sûkšma-saôparâya-câritra-guòa-pramâòa),(5) ideal conduct (yathâkhyâta-câritra-guòa-pramâòa).

Model XIII (ca. 450-480):69

cognition (jñâna),(I) direct cognition (pratyakša),(1) fivefold sensory perception (indriya-pratyakša):(a) auditory perception (œrotra-pratyakša),(b) ocular perception (cakšur-pratyakša),(c) olfactory perception (ghrâòa-pratyakša),

65 This threfold division may go back to NS 1.1.5 as well as to the tradition of the Šašþi-tantra.66 This division reaveals the same line of thinking as VS(C) 9.18, by which it must have beeninfluenced: ‘Reasoning based on inferential sign (inference) has the form: “x is the result of y”, “x isthe cause of y”, “x is the conjunct of y”, “x inheres in the same thing in which y inheres”, “x is thecontradictory of y”.’ This must also have been influenced by the theory of inference (anumâna) appliedto medicine, esp. medical examination (parîkšâ) based on tenfold procedure, found in Caraka-saôhitâ,e.g. in CarS vi 8.83 ff.67 Likewise, the division is under direct influence of NS 1.1.6: ‘Inference by analogy is the proof of theinferable property on the basis of its similarity to a property well-known’, supplemented by NS 1.1.34,1.1.35, but also by the practical discusison of inferences based on similarity and dissimilarity inVaiœešika-sûtra (vaidharmya: VS(C) 1.1.21, 2.1.22, 2.2.31, 5.2.21, 5.2.23, 9.12, sâdharmya: VS(C)2.2.29).68 This twofold division goes back to NS 1.1.7 which mentions two kinds of testimony: that relateeither to visible or invisible objects.69 NaôS 7–61 (p. 9–27). This classification is insofar inconsistent as it classifies all sensory cognitiveprocesses (sensaitons through 5 senses) both as sensory perception (indriya-pratyakša) and as sensuouscognition (âbhinibodhika) as a subtype of indirect cognition (parokša). This classification does notknow the notion of cognitive application (upayoga): the term upayoga occurs only once in NaôS 47[ga. 66] (p. 21) in a different context, whereas the term darœana occurs only once in NaôS 71[1](p. 28) in the context of divisions of the Canon (daôsaòa-dhara).

36

(d) gustatory perception (rasana-pratyakša),(e) tactile perception (sparœa-pratyakša),(2) threefold extrasensory perception (no-indriya-pratyakša):(a) clairvoyance (avadhi-jñâna),(b) telepathy (manaÿ-paryâya-jñâna),(c) omniscience (kevala-jñâna),—possessed by souls still in the cycle of rebirth (bhavastha-kevala),—possessed by liberated souls (siddha-kevala),(II) indirect cognition (parokša),70

(1) sensuous cognition (âbhinibodhika),(a) based on verbal cognition (œruta-niÿœrita), with four stages:—sensation (avagraha),—cogitation (îhâ),—perceptual judgement (avâya, apâya), —retention (dhâraòâ) of the judgement,(b) independent of verbal cognition (aœruta-niÿœrita), with four stages:—sensation (avagraha), of two kinds: fourfold (non-conceptual) sensation of momentarymanifestation (vyañjanâvagraha) and (conceptual) sensation of the object (arthâvagraha), thelatter having 6 varieties (5 senses and the mind),—cogitation (îhâ) of 6 varieties (5 senses and the mind),—perceptual judgement (avâya, apâya) of 6 varieties (5 senses and the mind), —retention (dhâraòâ) of the judgement of 6 varieties (5 senses and the mind),[alternative classification]:71

(a*) sensuous cognition (mati-jñâna),(b*) erroneous sensuous cognition (maty-ajñâna),(2) fourteenfold indirect cognition in form of testimonial cognition (œruta-jñâna-parokša),72

(a) threefold testimonial cognition based on linguistic symbols (akšara-œruta),73

(b) manifold testimonial cognition independent of linguistic symbols (anakšara-œruta), e.g.sneezing, coughing, blowing nose etc.,(c) testimonial cognition requiring the synthesising operation of the mind (saôjñi-œruta),(d) testimonial cognition not requiring the synthesising operation of the mind (asaôjñi-œruta),(e) correct testimonial cognition (samyak-œruta), comprises 12 primary canonical texts,(f) erroneous testimonial cognition (mithyâ-œruta), comprises 25 ‘heretical’ texts, etc.

70 NaôS 44 mentions that indirect cognition (parokša) is actually homogenous, and its twosubdivisions (âbhinibodhika and œruta) can for various practical reasons be treated as interchangeable.71 Under clear influence of TBh 1.32.72 NaôS 61 ff. (p. 27 ff.) discusses at length the divisions of œruta-jñâna and the Canon. I just mentionsome varieties.73 The third subdivision has 6 varieties, which include the acquisition of linguistic symbols with all fivesense organs and the mind, e.g. the acquisition of such ‘verbal’ cognition through tongue or touch.

37

indirect cognition(parokša)

direct cognition(pratyakša)

cognition (jñâna)

clairvoyance(avadhi-jñâna)

telepathy(manaÿ-paryâya-jñâna)

omniscience(kevala-jñâna)

sensory perception(indriya-pratyakša)

extrasensory perception(no-indriya-pratyakša)

auditory perception(œrotra-pratyakša)

ocular perception(cakšur-pratyakša)

olfactory perception(ghrâòa-pratyakša)

gustatory perception(rasana-pratyakša)

tactile perception(sparœa-pratyakša)

sensuous cognition(âbhinibodhika, mati)

testimonial cognition(œruta-jñâna)

Model XIV (7th / 8th centuries):cognitive criterion (pramâòa):(I) perception (pratyakša):(1) conventional perception (sâôvyavahârika-pratyakša)(a) perception for oneself (svârtha-pratyakša)

(b) perception for others (parârtha-pratyakša)(2) absolute perception (kevala-pratyakša)(II) indirect cognition (parokša):(1) verbal cognition (œâbda)(a) sentence for oneself (svârtha-vâkya)

(b) sentence for others (parârtha-vâkya)(2) inference (anumâna)(a) inference for oneself (svârthânumâna)

(b) inference for others (parârtha-pratyakša)(3) other kinds of cognition

38

perception (pratyakša) indirect cognition (parokša)

verbal cognition (ωbda)

inference (anumâna)

other

cognitive criteria (pramâòa)

inference for oneself(svârthânumâna)

inference for others(parârthânumâna)

conventional perception(sâôvyavahârika-pratyakša)

absolute perception(kevala-pratyakša)

perception for oneself(svârtha-pratyakša)

perception for others(parârtha-pratyakša)