Antarvyapti Interpreted in Jainism

Author(s): Atsushi Uno
Publisher: Z_Nirgranth_Aetihasik_Lekh_Samucchay_Part_1_002105.pdf and Nirgranth_Aetihasik_Lekh_Samucchay_Part_2
Catalog link: https://jainqq.org/explore/269047/1

Page #1
'Antarvyapti' Interpreted in Jainism Atsushi Uno The term "antarvyapti" used in the syllogism of Jaina logic is, together with the opposite term "bahirvyapti", found in the Jaina and the Buddhist texts alone. Among scholars, there is a great divergence of opinions on the first user of the term. To the best of my understanding, the Nyayavatara of Siddhasen. Divakara (the 8th century)* was the first to use the term. The Nyayavatara (abbr. Nya), a short manual of Jaina logic cast in 32 verses, mentions the antarvyapti immediately after two verses treating two kinds of drstantas. antarvyaptyaiva sadhyasya siddher bahir-udahrtih i vyartha syat tad-asadbhave'py evam nyaya-vido viduh 1120|| [Running rendering] When the establishment of probandum (sadhya) is secured exclusively by internal variable concomitance (antarvyapti), the citation of external example (bahir-udahrti) will be useless. And such will also be the case even when internal concomitance is absent (or unknown). Thus say those who are conversant with logic. The NyA has a sole commentary Vivsti by Siddharsi (c. A. D. 906) and it has a sub-commentary Tippana by Devabhadra (the latter half of the 12th century) scarcely referring to the verse under consideration. Among the subsequent Jaina works which developed the theory of antarvyapti on the basis of the NyA and the Vivrti are : the Pramananayatattvaloka (abbr. PNT) by Devasuri (1080-1169), its auto-commentary the Syadvadaratnakara (abbr. SyR), its abridged commentary the Ratnakaravatarika (abbr. RaA; * Siddhasena Divakara, as recent researches show, was not the author of the Nyayavatara but of Nayavatara. He was either Siddharsi alias Siddhasena (late 9th-early 10th cent.) or may be some other pre-medieval Siddhasena. -- Editors.

Page #2
'Antarvyapti' Interpreted in Jainism 311 by Ratnaprabha), and the Pramanamimamsa (abbr. PrM) by Hemacandra (1089-1172). The description of antarvyapti in the NyA has been taken over in the PNT almost ad verbatim, yet it may hardly be possible to grasp the meaning of the verse without the help of commentaries. It is due to the ambiguity of the expression that came to allow different interpretat about the verse. Though the idea of antarvyapti took its rise in Jainism, it is a wonder that the term itself disappeared in later Jaina works. However, Ratnakarasanti of the later Buddhist logic employed this term antarvyapti. His way of using the antarvyapti is somewhat different from that of Jainism and it is, as it were, a modified "kevalanvayi" inference of the NyayaVaisesika school. This thesis aims at having a general look at the Viviti and later Jaina works in order to clarify the purport of Siddhasena's antarvyapti. Before entering the main issue, I would like to give an outline of the syllogistic form of Indian logic and my opinion of antarvyapti. Technical terms used for the definition of antarvyapti will be understood in the following way. Antah (internal example) indicating paksa or the subject of argument, and vyapti residing in the antah is antarvyapti. Opposite term for "antarvyapti" is "bahirvyapti" (or bahyavyapti : external invariable concomitance), residing in the bahih. The bahih (external example; the basis other than paksa) refers to the twofold example, which is to be cited as the third member of the syllogism. When one accepts the function of the antarvyapti, the mention of bahir-drstanta (the external example) as well as of bahirvyapti (external variable concomitance) is cumbersome. That is, the external example (bahih) is useless. Inference is divided into two : inference for oneself (svarthanumana) and inference for others (pararthanumana). The former, being useful for oneself and primarily non-verbal, functions in one's consciousness only. That is, inferential knowledge of probandum (sadhya), secured by the cognition of probans (hetu; smoke) and of invariable concomitant + He was the disciple of Devasuri of Brhad-gaccha. The date of Ratnakaravatarika is c. A. D. 1165. -- Editors.

Page #3
312 Atsushi Uno Jambu-jyoti (vyapti). The latter, however, primarily verbal (vacanatmaka) in nature, being useful to convince the opponent (prativadin) of the knowledge of the probandum (sadhya). But each school of Indian logic came to tend to decrease the number of syllogism in later times, though such decrease depends on the faculty of opponents. (i) pratijna (argument) : the mountain has fire, The argument itself is called paksa, and the subject (mountain) is also called paksa. Fire which is predicate of the argument is called probandum (sadhya), and the mountain possessing fire (dharmin; Sp) is sometimes sadhya. (ii) hetu (reason) : because of smoke. Generally "reason," which enables one to obtain the inferential knowledge, is expressed by the ablative case of hetu (e.g. smoke) as a phrase. It is also called sadhana or sadhaka. However, how does the opponent (prativadin) act upon hearing pratijna and heru succeedingly ? Does he not show any effect or response, saying to himself "what about smoke ?" It may be because he does not know invariable concomitance (vyapti) between smoke and fire, or he may come to know the vyapti only after he is shown any example like a kitchen. [Three cases, on the part of the opponent, will be introduced later.) If the opponent is, after hearing the hetu, aware of the fact that there is fire wherever smoke occurs, he is compelled to construct in his mind the idea of vyapti. However, this vyapti is not expressed verbally by either the speaker (vadin; instructor) or the opponent (prativadin). This vyapti is not installed in the syllogism as an independent member, though happening after the two members. Some are of the opinion that this vyspti is included in the third member "udaharana" (or drstanta), but I do not accept this opinion. For drstanta should be literally an instance of basis on which to affirm the vyapti. However, the paksa (subject of the argument; e.g. mountain) eventually comes to be regarded as the locus of the vyapti, and the inference finishes. Thus any school of Indian logic beginning with the Nyaya-Vaisesika generally accepts the drstanta as any locus outside the paksa, but does not consider the paksa to be a drstanta.

Page #4
'Antarvyapti' Interpreted in Jainism 313 There is almost no material referring to the function of drstanta, but it is used as the third member, convincing the [dull-minded] opponent of the locus of vyapti which comes from the preceding two members. The ambiguous character of the drstanta is originally derived from other reasons, which will be referred to later. However, is the drstanta an indispensable member to the opponent of slow understanding in the inference for others ? I shall show some instances of establishing the probandum (inferential knowledge) even without the help of drstanta. (1) In the three-membered syllogism of Western logic, the major premise (equivalent to vyapti; M - P) and the minor premise (roughly equivalent to paksadharmata; S - M) are necessary for acquiring the conclusion (inferential knowledge; S. P), dispensing with any example. (3) (2) In the case of Indian logic, later Buddhism, the Mimamsaka, and the Sankara-Vedanta accept vyapti (hetu sadhya) and paksadharmata (hetu + paksa) only for acquiring inferential knowledge, omitting drstana. [The antecedent here is meant to occur in the consequent.] The Nyaya-Vaisesika accepts paramarsa by combining the abovementioned two elements into one. That is, vyapti-visistapaksadharmata-jnanam (the knowledge of paksadharmata possessed of vyapti). As has been shown above, to accept the existence of the vyapti in the antah (paksa) exclusively is antarvyapti, without approving of the function of bahih and bahirvyapti as the main cause (gamaka) of inferential knowledge. Drstanta (or udaharana), the third member of the syllogism, was originally set forth in the Nyayasutra (abbr. NS) and the Nyayabhasya (abbr. NBh), and was quite different in nature from what generally has been understood in later schools of logic.

Page #5
314 Atsushi Uno Jambu-jyoti In the NS and the NBh, the dharmin of pratijna (e.g, mountain) is compared with either homogeneous example (e.g. kitchen) or heterogeneous example (e.g. pond) by virture of sadharmya (e.g. smoke) or vaidharmya (e.g. -smoke) between the two (dharmin and drstanta), one analyses in the dharmin fire. Inference in the NS and the NBH is, strictly speaking, not inference, but analogy in nature, and has nothing to do with vyapti in the process of thinking. NS 1..34 : Whatever establishes the probandum (Sp or p), on the basis of the similarity with the [homogeneous) example (in accordance that probandum possesses similarity with the example) is the probans (hetu; M). NBh 1.1.34: Whatever establishes the probandum i.e., dharma (S or Sp) in accordance with the similarity with the example is probans (hetu, smoke). By perceiving a certain dharma (e.g. smoke) in the probandum (subject; mountain) and the same dharma in the example (e.g. kitchen), what describes that it (smoke) establishes the probandum (Sp or P) is probans (hetu). (Comment) Accoding to the NBh, probandum (sadhya) is used in two ways, by indicating first S, and then Sp or P. NS 1.1.35: In similar manner, whatever establishes the probandum (Sp or P) is the probans (hetu), on the basis of the dissimilarity with the [heterogeneous) example (e.g. pond) (in accordance that probandum (mountain) does not possess similarity with example). NS 1.7.36: On the basis of the similarity with probandum (sadhya; dharmin; S, mountain), whatever things possess the dharmas (e.g. smoke) of the probandum (mountain) is (homogeneous] example. NS 1.1.37 : On the basis of the dissimilarity with probandum (sadhya; dharmin; S, mountain), whatever things possess things other than the dharmas (e.g. smoke) of the probandum (mountain) is [heterogeneous) example.

Page #6
'Antarvyapti Interpreted in Jainism 315 NyA, Vivrti Here a summary of Viviti will be introduced. based on Vivrti : (1) When the opponent remembers the Vyapti, since the probandum (sadhya) is established by antarvyapti (vyapti in the paksa), external example (bahih) is useless to mention. (2) Even when the vyapti is not known to the opponent, since antaryapti is absent (or not known) external example (bahih) is useless to mention. Thus the expert on logic says. (20) (Running rendering] When the opponent forgets the vyapti, example is mentioned for him. Otherwise (if he does not forget the vyapti) it is not the case (example is not mentioned). "Otherwise" means (1) the opponent remembers the vyapti (is versant with the vyapti) or (2) he is ignorant of the vyapti. In the case (1), the example is not mentioned. For when the opponent remembers the hetu (probans) which is avinabhavin of the probandum (inseparable from probandum), he knows the probans (hetu) in the paksa also. He necessarily comes to know the probandum in the paksa. Thus the vyapti in antah (paksa) establishes invariably probandum, and the vyapti outside the paksa (bahirvyapti) is useless to mention, In the case (2), antarvyapti is either absent or unknown and it is useless to mention the "bahih" (external example). For an observation of the coexistence in a certain place cannot entail the establishment of the vyapti (coexistence) in all places, vyabhicara being observable. Thus when the opponent (student; prativadin) is ignorant of the vyapti, the vyapti is first to be grasped by means of tarka (as pramana), and then comes the sadhya-siddhi. [Comment) According to the Vivrti, the opponents are divided into two groups, and example is considered useful or useless in conformity to these divisions : I. One who forgets tho vyapti (vismrta-sambandha) External example is useful to mention.

Page #7
316 Atsushi Uno Jambu-jyoti II. One who does not forget the vyapti (avismrta-sambandha) (a) One who remembers the vyapti (smaryamana-sambandha) External example (or external vyapti) is useless, since the antarvyapti here functions. (b) One who is ignorant of the vyapti (agrhita-sambandha) Pramana "tarka" is first necessary to enable the opponent to know the vyapti. So bahirvyapti is useless, and antarvyapti does not function. Vivrti regards antarvyapti as principal. It rather neglects bahirvyapti, than regards it as subordinate. In accordance with antarvyapti under varied conditions, bahirvyapti is taken cumbrous or impossible and eventually useless (vyartha). Even in the case (1), the Vivsti does not even refer to the dull-minded opponent. Vivrti's view-point may be diagrammed as follows : [0 = obtainable, * = unobtainable] (1) The first "useless' (a) antarvyapti (0) bahirvyapti (0) cumbrous useless (b) antarvyapti (0) bahirvyapti (X) impossible useless (II) The second "useless" (c) antarvyapti (X) bahirvyapti (X) impossible useless Rendering As for the problem under consideration, the NyA does not describe directly remaining members "upanaya", "nigamana", and "corroborative five members", because this treatise aims at brevity, yet they may be conjectured by the sharp-minded on the basis of the described those members. For so far as the form and the number of members are concerned, the description may be divided into three : the lowest (jaghanya; simple), the middle (madhyama; ordinary) and superior (utkrsta; detailed). The lowest is the description of the hetu only, the middle is the description of two members and so forth, and the superior is the description of all the ten members. Here the middle description of members are described directly, and the lowest (simplest) and the superior (detailed) descriptions of members are indirectly suggested, because there is proof.

Page #8
'Antarvyapti' Interpreted in Jainism 317 (Comment] Though three kinds may be based on the form and number, these three kinds may be applicable to opponents' intelligence in the reverse order. "Suddhi" (corroboration; correctness) in the five corroborative members such as "pratijna-suddhi' etc., is referred to by the NBh and the Nijjutti of Bhadrabahu (5th cent. A. D.). Five members being added by five corroborative members, they may have eventually accepted tenmembered syllogism. It is doubtful whether ten members withstood the practical use. III Later Works : PNT (SYR, RaA), PrM After the Vivrti's opinion, I would like to introduce antarvyapti's theory taught in the later Jaina works. The Nya and the PNT are substantially almost the same. In the NYA primary cause of "sadhyasiddhi" (inference) is antarvyapti, while in the PNT it is the probans (hetu) and the sub-cause is regarded as antarvyapti. In the PNT the term "bahirvyapti" appears for the first time. [It is worth to note the next points. In the later Nyaya school the karana of inferential knowledge has undergone change in the following order: hetu - vyapti - vyapti-jnana - paramarsa.] When, on the basis of antarvyapti, the probans (hetu) can or cannot establish the probandum (sadhya), it is useless to mention the bahrvyapti. (antarvyaptya hetuh sadhya-pratyayane saktasaktau ca bahirvyapter udbhavanam vyartham iti //111.371/) If the probans establishes the probandum, based on antarvyapti exclusively, bahirvyapti is useless. (It mentions three examples.) (I) "Here is fire." (asty atragnih) "Where there is fire, there only exists (is intelligible). (saty-evagnau dhumopapattih) (II) "That my son is speaking outdoors" (mat-putro 'yam bahir-avasthito vakti) "Otherwise such voice is unintelligible." (anyathaivam bhuta-svaraupapattih]

Page #9
318 Atsushi Uno Jambu-jyoti When by only using such either (positive vyapti) or (negative vyapti), sharp-minded (vyutpanna-mati) opponent can establish probandum (inference). (III) "This woman's unborn baby is a boy." (etasyah strio garbhatyarn puman) "because of being her baby." (etad-apatyatvat) "Like other babies who are known as boys." (prasiddha-pumstvetatarapatyavat) For in these examples, even if in the presence of bahirvyapti antarvyapti is not obtainable, probans (hetu) is not considered to be the main cause (gamaka) of the probandum (inferential knowledge). In the third example, "saty-eva pumstve etat-apatyatvat" should be valid probans (hetu). So, such valid antarvyapti is doubted. Bahirvyapti produced from such doubt, though determined, cannot produce anything (useless). (Comment] I would like to test three examples one by one. Syllogisms mentioned in the first two examples are right. In the first example, hetu "smoke" is here omitted. The vyapti being reducible to the form chetu sadhya> is antarvyapti, residing in the paksa (here; mountain?). Here antarvyapti being taken as principal, instructor (vadin) didn't dare to mention drstanta. Since the hetu (probans) is gamaka of inference, bahirvyapti is useless so far as the form is concerned. It is a good example to show how antarvyapti or bahirvyapti functions in accordance with the absence of the presence of drstanta, in relation to two syllogisms having two members in common. In the second example, vyapti being shown in the form <-sadhya > -hetu> is (negative vyapti). <-hetu> is "otherwise unintelligibility of such voice" (anyatha evambhuta-svaranupaparti), and <-sadhya> is "if he is not speaking outdoors" (bahir-vacanabhava). Here any heterogeneous example is purposely neglected, and accordingly bahirvyapti is not mentioned. Thus antarvyapti existing in the paksa is functioning, the hetu is gamaka of inference. In the third example, the hetu is apparently "asiddha-hetu" as well as "sopadhika-hetu", and this hetu is not the cause (ghataka) of inference.

Page #10
'Antarvyapti' Interpreted in Jainism 319 If one is forced to say, the valid hetu should be "being boy" and "being her baby." Thus bahirvyapti is useless. (in the beginning of the RaA is mentioned a verse, which is a compromise between the NyA (20) and the PNT (111.37). The main. cause of inferential knowledge is considered to be antarvyapti, and the term "bahya-vyapti" is mentioned instead of "bahirvyapti.) antarvyaptih saddhya-saktau bahyavyapter varnar vandhyam eva antarvyaptih saddhya-asaktau bahyavyapter varnam vandhyam eva |! When on the basis of antarvyapti the establishment of probandum is possible, the mention of bahyavyapti is useless. When on the basis of antarvyapti the establishment of probandum is impossible, the mention of bahyavyapti is useless. (Like the SyR, the RaA mentions two examples.) (1) "That my son is speaking outdoors" (mat-putro 'yam bahir avasthito vakti) "Otherwise such voice is unintelligible." (anyathaivam bhuta-svaraupapattih] In this case though bahirvyapti does not exist, antarvyapti is accepted as gamaka of inferential knowledge. (II) "He (the unborn boy) is black." (sa syamah) "Because of being her baby." (tad-putrarvat) "Like her other boys." (itara-tat-putratvat) In this case, however, even though bahirvyapti exists, it is not regarded as the main cause (ghataka) of inferential knowledge. (Comment] The first example is the same as the second one of the Syr. The second example is almost the same as the third one of the SyR. In the first example, the description seems misleading. Bahirvyapti (vyatireka-vyapti) and heterogeneous example do exist, though they are not described obviously. PNT further describes two kinds of vyapti (111.38, 39).

Page #11
320 Atsushi Uno Jambu-jyoti Invariable concomitance of the hetu (M) with the sadhya (P) existent in the very object which is considered paksa (S) is antarvyapti. On the other hand, invariable concomitance existent elsewhere is bahirvyapti. (paksikrta eva visaye sadhanasya sadhyena vyaptir antarvyaptih; anyatra tu bahirvyaptih) [Comment] Vyapti being synonymous with vyapyata, it is a nature or a relation possessed by a pervaded (vyapya). Therefore I translated "hetu's vyapti with sadhya" etc. For example : (1) [ "Entity (vastu) is made up of many natures." (vastu anekantatmakam) "Because it is sat." (sattvat) This is not expressed.] "Only when the entity is of many natures (sadhya), then sattva (hetu) is intelligible." (tathopapatti) (II) > "This place has fire." (ayam deso'gniman) "Because it has smoke." (dhumavattvat) "Whatever possesses smoke has fire." (sa evam sa evam) "Like a kitchen" (yatha pakasthanam) [Comment] These two examples show antarvyapti and bahirvyapti respectively. Syllogism (1) is lacking in drstanta, because entity (vastu) is highest connotation being synonymous with existent (sat) in Jainism. The sadhya is and the hetu is , and the vyapti holding between such two notions is nothing but antarvyapti, residing in the paksa (= vastu) only and not elsewhere. Syllogism (2) is equipped with three members : pratijna, hetu, drstanta, and positive-vyapti existent in the outer example (bahirdrstanta), i.e., kitchen is bahirvyapti. Antarvyapti has much to do with drstanta. Hemacandra in PrM first defined drstanta in three sutras, discussed it in his own commentary and concluded that it is by no means indispensable to inference, and eventually referred to antaryapti.

Page #12
Antarvyapti' Interpreted in Jainism 321 "Drstanta is not the cause (anga) of inference" (1.1.18). "Because inferential knowledge is established by the probans (sadhana) exclusively" (1.11.19). For inferential knowledge is produced from the cause (sadhana; hetu) which lacks drstanta and is inseparable from probandum (sadhya). So drstanta is by no means the cause of inference. For, drstanta is mentioned for the purpose of (1) getting the knowledge of the probandum, or (2) getting the knowledge of the vyapti, or (3) making the opponent remember the vyapti ? It is not the case (1), because the establishment of the probandum is secured by the above-mentioned probans (hetu) exclusively. And it is not the case (2), because the knowledge of the vyapti is produced from the denial (exclusion) of the hetu in the heterogeneous example (vipaksa; vyatireka; what has not the probandum). And any drstanta is individual. How can individual drstanta totally communicate us the general vyapti ? For other cases of vyapti, another vyapti is to be sought. Since that drstanta is also individual in nature, the decision of the whole is impossible, because drstanta being sought one by one may lead to regressus ad infinitum. And it is not the case (3). For those who are familiar with the vyapti, the vyapti is to be remembered at the sight of the probans (hetu). For those who are ignorant with the vyapti, the vyapti cannot be remembered even at the sight of drstanta. For remembrance presupposes previous knowledge (fire) smoke : -fire )-smoke). "Drstanta is the basis on which to show the opponent the vyapti" (1.1.20) [Question) If the drstanta is not the cause of the inferenial knowledge, why do you dare to give it a definiton ? (Answer) For in order to satisfy the opponent's wish in the inference for others, exceptional drstanta will be allowed. And even in the field of inference for oneself, the definiton of drstanta is not entirely inappropriate in view of the fact that there may be a person who is helped to arrive at the knowledge of antarvyapti from the observation of bahirvyapti found in an example. [Comment) it is the NS that showed the general definion of drstanta. However, the drtanta of later product in other schools being quite different in nature, is necessarily connected with the vyapti. The definition of drstanta

Page #13
322 Atsushi Uno Jambu-jyoti was hardly touched by Jaina and Buddhist texts, with an exception of Hemacandra and Devasuri. Frankly speaking, antarvyapti is the vyapti which is accepted directly in the paksa, not through the medium of any drstanta. IV The Naiyayikas accept three lingas, upon which they put forth conventionally three kinds of inference. Among these is "kevalanvayi anumana". It is explained in the Tarkasamgraha (* 48) as follows: The linga which has vyapti only positively (anvayenaiva) is called "kevalanvayi linga". For example, "A pot is nameable (abhidheya)". "Because of knowableness (prameyatva)" "Like cloth (patavat)", In this case, between the probandum (= abhidheyatva) and the probans (prameyatva) there is no negative vyapti, because all things are nameable as well as knowable. In the Naiyayika's tradition, drstanta cannot be paksa, and when the paksa happens to be all things indicated by etc., any drstanta is not obtainable. In the above syllogism, the Naiyayika might have purposely limited to "pot" etc. as the subject instead of or , in order to evade the difficulty of unavailability of drstanta. In the above-mentioned example (PNT 111.39) "sat is made up of many natures" will be considered to be a modified syllogism of the Naiyayika's "kevalanvayi anumana". Ratnakarsanti, a later Buddhist logician, holds antarvyapti's doctrine for interpreting ksana-bhanga. That is, in the argument "Everything is momentary (Vastu ksanikam)" he could not but seek for vyapti in the paksa i.e., antarvyapti. As has been explained, the establishment of probandum is eventually secured by confirmation of the vyapti in the paksa. However, this problem being quite different in nature from the question whether antarvyapti or bahirvyapti, here will not be taken into consideration. To sum up, (1) When vyapti is sought in the paksa, without accepting drstanta as a member, antarvyapti functions. (2) When vyapti is sought in the drstanta, bahirvyapti operates. (in order to make the opponent of slow understanding remember the vyapti.) (3) Even in two similar syllogisms (e.g. of smoke and fire), a) when it is furnished with drstanta bahirvyati functions (PNT 111.39, the second

Page #14
'Antarvyapti' Interpreted in Jainism 323 example), b) when it lacks drstanta antarvyapti functions (SyR the second example). (4) Syllogism is lacking in drstanta, (i) because it is dependent on the prativadin's arbitrary (in accordance with the opponent's faculty), and not necessarily determined physically. (ii) when positive vyapti is unobtainable (SyR the second example, RaA the first example). (iii) when the paksa happens to be 'sarva', 'sat' etc., indicating the highest connotation (PNT 111 39, the first example; Buddhist's antarvapti-theory). 000