Early Buddhist Theory of Knowledge c12: Chapter VII: Logic and Truth

Early Buddhist Theory of Knowledge c12

Language: English

Chapter VII: Logic and Truth

c12.2

(560) In this chapter we propose to investigate what is meant by the four forms of predication or the four logical alternatives (*v. supra*, 182) as they appear in the Buddhist texts. This will be followed up by an inquiry into the nature of truth, as understood in the Buddhist texts.

c12.3

(561) There seems to be considerable confusion among scholars as to what was meant by this four-fold schema. Poussin speaks of it as 'a four-branched dilemma' of Buddhist dialectic and believes that it violates the principle of Contradiction. He says, 'Indians do not make a clear distinction between facts and ideas, between ideas and words; they have never clearly recognized the principle of contradiction. Buddhist dialectic has a four-branched dilemma: Nirvāṇa is existence or non-existence or both existence and non-existence or neither existence nor non-existence. We are helpless'.¹ Mrs Rhys Davids in a passage that we have already quoted (*v. supra*, 497), suggested that they are 'Laws of Thought'. Later, referring to this same passage, she says, 'To revert to the Laws of Thought—the way in which Indian logic presented the second and third of these (Contradiction and Excluded Middle) has been mentioned'.² Barua too favours this interpretation and in fact calls them 'the four laws of thought':³ '. . . the happy result of this mode of discussion or "wrangling" among the learned wanderers was that in the time of Buddha the four laws of thought were recognized as a matter of course. These are in their applications to propositions-

c12.4

¹ *The Way to Nirvana*, Cambridge University Press, 1917, p. 111.

c12.5

² *ERE*, Vol 8, p. 133, Article on 'Logic (Buddhist)'.

c12.6

³ *A History of Pre-Buddhistic Indian Philosophy*, University of Calcutta, p 47, 1921.

c12.7

1. (If A is B), A is B.
2. A cannot be both B and not-B.
3. A is either B or not-B.
4. A is neither B nor not-B.' (*loc. cit.*)

c12.8

He adds that ‘these are implied in such interrogative propositions as are met with throughout the Buddhist canonical texts’ (*loc. cit.*).

c12.9

(562) Now neither Poussin’s view that this logic is due to a failure to understand or respect the principle of Contradiction nor the views of Mrs Rhys Davids and Barua that they are laws of thought, bear critical examination. We find as early as Uddālaka an awareness in Indian thought of the importance of consistency and the principle of Contradiction, when it was held that it could not both be true that Being came out of Being and Non-Being (*v. supra*, 25). In the Nikāyas, consistency is regarded as a criterion of truth (*v. infra*, 597) and it is stated that ‘truth is one and not two’, ekam hi saccam na dutīyam atthi, Sn. 884; *v. infra*, 597. The dilemmas present two alternatives one of which is the contradiction of the other (M. I.392, *v. supra*, 351–6). In fact, in one place in the Samyutta Nikāya, we come very close to a formal statement of the principle of Non-Contradiction Nigantha Nātaputta converses with Citta and in the course of the discussion, the former makes the following two observations about the latter:¹

c12.10

¹ S. IV 298, 299

c12.11

(i) passantu yāva ujuko c’āyam Citto gahapati yāva asatho . . . amāyāvī (p), i.e see how upright, honest and sincere Citta, the householder, is.
(ii) passantu yāva anujuko c’āyam Citto gahapati yāva satho . . . māyāvī (~p), i.e see how Citta, the householder, is not upright, honest or sincere.

c12.12

Citta is anxious to show that Nigantha Nātaputta is contradicting himself and says, sace purimam saccam pacchimam te micchā, sace pacchimam saccam purimam te micchā, i e. if your former statement (p) is true, your latter statement (~p) is false and if your latter statement (~p) is true, your former statement (p) is false. In other words, in the above situation when the statements are of the form p and ~p, it cannot be the case that both p and ~p are true (~(p. ~p)), which is the formal statement of the principle of Contradiction.² Barua has likewise drawn our attention to a passage in the Kathāvatthu³ where it is pointed out that it is self-contradictory to assert that something ‘exists’ and ‘does not exist’ in the same sense at the same time.

c12.13

² Stebbing, *op cit*, p 191

c12.14

³ *Op cit.*, p. 49, *v* Kvu 159

c12.15

(563) The contentions of Mrs Rhys Davids and Barua that they are laws of thought are equally fantastic The ‘four laws of thought’ that Barua is speaking of are presumably the law of Identity (see below, 1), the law of Contradiction (see below, 2), the law of Excluded Middle (see below, 3) and 'the law of Double Negation'¹ (see below, 4). In the first place Barua's 'A is neither B nor not-B' is not the principle of Double Negation as understood in Western logic, where it is defined as equivalent to 'A is not not-A'.² 'A is neither B nor not-B' appears in fact to be a violation of the principle of Excluded Middle, which states that 'A is either B or not-B'. Secondly, a study of the four forms of predication, as they stand, is sufficient to show that they are not laws of thought. The four forms of predication are:

c12.16

¹ Barua, *ibid*

c12.17

² Stebbing, *op cit*, p 191.

c12.18

1. S is P, e.g atthi paro loko (there is a next world).
2. S is not P, e g. natthi paro loko (there is no next world).
3. S is and is not P, e.g. atthi ca natthi ca paro loko (there is and is no next world)
4. S neither is nor is not P, e.g. n'ev'atthi na natthi paro loko (there neither is nor is there no next world).

c12.19

(564) It will be seen that (1) is not the law of Identity but a simple affirmative categorical assertion; (2) is not the law of Contradiction but a negative assertion or denial, it being a problem as to whether it is the contradictory or merely the contrary of (1); (3) and (4) are assertions not recognized in Aristotelian logic although statements of this form sometimes occur in everyday parlance even in Western languages.³ Considered as laws of thought or in the light of their conformity to them (3) violates (i.e. is the negation of) the law of Contradiction and (4) violates the law of Excluded Middle. Judged by their form in the light of Aristotelian logic (3) and (4) are necessarily (logically) false propositions No wonder Poussin said 'we are helpless!'

c12.20

³ *v*. Article by C Lewy on 'Calculuses of Logic and Arithmetic' in *Proceedings of the Aristotelian Society*, Supplementary Vol 20, pp 36, 37.

c12.21

(565) Now what do these propositions really mean? We have already attempted to answer this question in an article entitled 'Some Problems of Translation and Interpretation II',⁴ and we do not propose to repeat this here. We intend merely to summarize briefly the gist of what we have said and make some added observations.

c12.22

⁴ UCR, Vol 8, pp 45–55, 1950

c12.23

(566) Since then there have been a few articles written around this subject but in none of them is there a serious attempt to clarify the problems involved. The first is that of P. T. Raju writing in the ‘Review of Metaphysics’ (Vol. 7, 1953–4, pp 694–713) on ‘The Principle of Four-Cornered Negation in Indian Philosophy’. Raju is mainly concerned in this article with Nāgārjuna’s negations or denials of the truth of propositions, expressed according the above four-fold formula. He compares them with similar negations of Śaṅkara. However, it did not occur to him that before we can understand the nature of these negations, it was necessary to know exactly what he was negating or in other words the nature of these four types of assertions. He says that ‘the principle seems to have been first used by Sañjaya’ (*op. cit.*, p. 694) adding that ‘those philosophers who gave a negative answer to all four questions were called “eel-wrigglers” by the Buddhists’ (*loc. cit*). Without giving an iota of evidence he goes on to say that ‘Gautama the founder of the school of Nyāya . . . called such philosophers *vitaṇḍavādins*’ (*op. cit.*, p. 695). He then gives an account of the seven forms of Jain predication, which is found in most textbooks on Jainism Where he has tried to throw some light, in passing, on the four-fold formula of the Pāli Nikāyas, it is either not backed up with good evidence or it is plainly self-contradictory. Thus he seems to identify the Jain syād avaktavya (may be, it is unpredicable) with the Buddhist assertion of the form ‘S is neither P nor not-P’. But all that he says on this subject is as follows: ‘Now indescribability means that S is neither P nor not-P. P and not-P are opposites and therefore cannot be applied simultaneously to the same subject . . . It is the same as saying that ‘S is neither P nor not-P’ (*op. cit.*, p. 699) In trying further to explain the statement ‘S is neither P nor not-P’ he says that ‘*in mathematics* “śūnya” means “zero” and *in metaphysics* it means “that which is neither positive nor negative”’ (*op. cit.*, p 701), but one page later without further ado, he denies even this: ‘“then zero becomes a quantity of which all the four alternatives are denied, it is neither positive, nor negative, nor both positive and negative *nor neither positive nor negative*” (*op. cit.*, p. 702, italics mine)’.

c12.24

(567) The next article to appear on this subject is by Professor Archie J. Bahm on the intriguing subject of ‘Does Seven-fold Predication Equal Four-Cornered Negation Reversed?’¹ But his attempt is disappointing, since as he admits and as is evident from the article itself, his knowledge of the subject is gleaned from ‘reading Raju’s article’ (*op. cit.*, p. 130) and ‘his conclusion did not result from a study of the sources’ (*loc. cit.*). He says that ‘this study purports to show that an attempt to reverse the principle of four-cornered negation, so that one may have a four (or more)—fold principle of affirmation, results naturally and logically in the seven-fold system of predication’ (*op. cit.*, p. 128). Despite these obvious defects in Bahm’s article, he does make a significant contribution towards solving the riddle (logical) of the four-fold assertions when he says, ‘To affirm “is”, “is not”, “both is and is not” and “neither is nor is not” jointly when “is” and “is not” are interpreted as contradictories and “neither is nor is not” as involving¹ an excluded middle is to assert contradictorily. But if on the other hand, “is” and “is not” (or “a” and “non-a”) are interpreted as opposites rather than as contradictories and “neither is nor is not” (or “neither a nor non-a”) is interpreted not as involving an excluded middle but as presupposing that there is something which is neither the one nor the other (its opposite), then no contradiction is involved’ (*op. cit.*, p. 128). According to this theory we do not fall into logical difficulties if we treat not-P (in the four assertions) as the contrary and not the contradictory of P. This is in principle the solution that we had offered in our own treatment of the subject and we shall further examine this solution below.

c12.25

¹ *Philosophy East and West*, Vol 7, Nos 3 and 4, pp 127–30.

c12.26

¹ What he probably means is ‘violating the principle of Excluded Middle’.

c12.27

(568) The latest article to appear on this subject is by Shosun Miyamoto, entitled ‘The Logic of Relativity as the Common Ground for the Development of the Middle Way’.² Miyamoto takes (2) to (5) of the five-fold assertions (*v. supra*, 185) which he attributes to Sañjaya (but which in fact are either shared by all the Sceptics or were exclusively Buddhist, *v. supra*, 184–90) and says that four of them are equivalent to (I) to (IV)³ of the four-fold formula, viz.

c12.28

² *Buddhism and Culture*, ed. Susumu Yamaguchi, Kyoto, 1960, pp. 67–88.

c12.29

³ We shall use these roman numerals to refer to the four forms of predication.

c12.30

| *The four-fold formula* | | *Sañjaya’s statements* |
| | | 1. I do not think of them in such a manner (evam pi me no). |
| I | = | 2. I do not think of them as being identical (tathā ti pi me no). |
| II | = | 3. I do not think of them as being different (aññathā ti pi me no). |
| III | = | 4. I do not think of denying them (no ti pi me no). |
| IV | = | 5. I do not think of not denying them (no no ti pi me no). |

c12.31

(*op. cit.*, p. 75).

c12.32

(569) This is different from our identification of equivalence (*v. supra*, 188) but Miyamoto does not give any reasons for his identifications, nor is he consistent for he later equates (IV) with (1) (*op. cit.*, p. 81). He considers that Sañjaya's 'system is quite near to the Buddhist standpoint of indescribable or inexpressible' (avyākata-, avyākṛta-)' (*op. cit.*, p. 71) and holds that 'Sañjaya's thought is not far removed from the logic of Śūnya of the Mādhyamika' (*loc cit.*). On the logic of the four-fold formula itself he sheds little light. 'This type of a logical view of nature could not entertain any preconceptions nor be compromising by maintaining one particular view only. It expresses a desire to view nature in her actual state, a logic in pursuit' (*op. cit*, p. 75). This seems to stress the similarity of this logic with the anekāntavāda- of Jainism But a few pages later he holds without offering any explanation that Jain syādvāda stands midway between the position of Sañjaya and the Buddhist point of view: 'The Jains maintained the indeterminate non-exclusive view of nature (anekāntavāda), a half-way position between the Sañjaya non-committal no-standpoint view and the Buddhist Middle Way view' (*op. cit.*, p. 79). A page later, however, he identifies the Jain scheme with the four membered logic of Buddhism as follows:

c12.33

| *Jain schema* | | *Buddhist schema* |
| 1. syād asti (may be, it is) | = | I |
| 2. syān nāsti (may be, it is not) | = | II |
| 3. syād asti nāsti (may be, it is and it is not) | = | III |
| $.&4. syād avaktavya (may be, it is unpredicable) / &5. syād asti avaktavya (may be, it is and it is / & unpredicable / &6. syān nāsti avaktavya (may be, it is not and / & it is unpredicable / &7. syād asti nāsti avaktavya (may be, it is and / & it is not, and is unpredicable}$ | = | IV |

c12.34

(570) He observes that since the term 'avaktavya-' (the inexpressible) occurs in (4) to (7) '"It may be inexpressible" refers to member four of the Buddhist logic' and adds 'while at the same time it also refers to the first member of the Sañjaya logic', i e. 'I do not understand them (questions) in such a manner' (*op cit.*, p 81), contradicting what he had said earlier (*v. supra*, 568). He sees a similarity between the systems of Sañjaya, Jainism and Buddhism, though he is not clear as to what he means: 'The crucial fact however about the three logical systems compared above is that in each what corresponds to the fourth member expresses "neither is nor is not", "denial", "inexpressible" and the non-causal (ahetu) and that in these systems Sañjaya, Jaina and Buddhism firmly stand on the side against any form of nihilism' (*op. cit.*, p. 81).

c12.35

(571) In determining what is meant by the four forms of predication, we cannot decide *a priori* what is meant by these particular forms of speech. We have to proceed empirically and examine what is meant in the light of historical and contextual considerations. The Nikāyas seem to regard them as 'the four possible positions' or logical alternatives¹ that a proposition can take. Thus, other religious teachers, it is said, in fixing the status of the Tathāgata or the perfect person (*v. supra*, 380) after death do so 'according to these four positions' (imesu catusu thānesu, S. IV.380) and not 'outside these four positions' (aññatra imehi catuhi thānehi, *loc. cit.*) meaning by the four positions statements of the form 'hoti Tathāgato param maraṇā', etc., in the four alternatives. The fact that it was thought that one could not posit the status of the Tathāgata outside these four positions is significant, it shows that these four positions were considered to be mutually exclusive and exhaustive or in other words to be the possible logical alternatives.

c12.36

(572) Of the four alternatives, the first is a simple affirmative categorical assertion The second appears to be the opposite of this, but whether it is the contrary or the contradictory is a problem that needs to be examined. Judged by Aristotelian standards, the third and fourth alternatives appear to be very problematic (*v. supra*, 564).

c12.37

(573) We have seen that the third, which was a double-ca (and) type of assertion ('ca (and) ... ca (and)') was introduced by the Trairāśika Ājīvikas (*v. supra*, 222–7) to denote a class of things which could not be described by propositions of the first or second type. We may also note the fact that the fourth or double-na (not) type of assertion comes down from the earliest times. It first appears in the Nāsadīya hymn (*v. na asad āsīt na u sad āsīd tadānīm*, RV. 10.129.¹) and is made use of by Yājñavalkya (*neti neti*, Brh. 4.5 15; *v. supra*, 44, 392) and the Māṇḍūkya Upaniṣad (*v. supra*, 392). This kind of expression is used to denote the fact that conceptual epithets cannot be predicated of the non-conceptual (the ultimate reality, the highest mystical state, the impersonal state after death as conceived by Yājñavalkya). Both Mrs Rhys Davids¹ and Barua² have suggested that this kind of expression could be employed to denote that no real attributes could be ascribed to what was unreal, but they have not given any examples of such an usage in Indian thought It is however a plausible suggestion and we have ourselves shown that what was meant by the expression, 'N'eva sassato na asassato ... loko', i e. 'the world ... is neither eternal nor not eternal' was that temporal epithets (sassata, asassata) cannot be predicated of a world (loko) which is timeless (*v. supra*, 400) or unreal (*v. supra*, 407). As for the suggestion of the Comy. that the Sceptic holds that 'neither p nor not p' is the case in rejecting both thesis as well as anti-thesis, we found that it was not corroborated by the evidence of the texts (*v. supra*, 391)

c12.38

¹ Mrs Rhys Davids, refers to them as 'logical alternatives' in one place notwithstanding her other pronouncements: '... the Indian mind playing about with its four logical alternatives, A is B, A is not B, A is both B and not B, A is neither B nor not B, *Buddhist Psychology*, Second Edition, London, 1924, p. 248, Keith comes near to this when he says they are 'the four possible modes of framing a proposition' (*Buddhist Philosophy*, p. 137).

c12.39

¹ 'If not, is A *neither B nor not-B* (in other words is A a chimaera)', *v supra*, 497.

c12.40

² 'The fourth Law ... is applicable to .. the conception of something which is really nothing ..', *op. cit*, p 49

c12.41

(574) Now if we refer to the four forms of predication by the Roman numerals I, II, III, IV respectively, what we intended to show in our article (*v. supra*, 565) was that III and IV, despite their form, were *contingent* propositions, purporting to give us information about some alleged state of affairs Let us take an example, where we are quite certain about the meaning of III and IV:

c12.42

I Antavā ayam loko (parivaṭumo), 1.e this world is finite (and bounded all round)
II Ananto ayam loko (apariyanto), 1 e. this world is infinite (and not bounded all round)
III Antavā ca ayam loko ananto ca, 1 e. this world is both finite and infinite.
IV N'evāyam loko antavā na panānanto, 1 e. this world is neither finite nor infinite.
@H D. I 22, 23.

c12.43

Here (I) states that the world has the characteristic 'finite' *in all respects*; this is clear from the further qualification, parivaṭumo, 'bounded all round'. II states that the world has the opposite or the contrary characteristic of being 'infinite' *in all respects* (*v* apariyanto).

c12.44

Now III *does not state* that the world is both finite in all respects (I) and infinite in all respects (II). This would be self-contradictory, if we mean by loka- the same thing in both cases. It states that the world is both *finite in some respect* and *infinite in another respect*. For it is said that those who hold this theory 'conceive the world to be finite in one dimension and infinite in another' (uddhaṃ adho antasaññī lokasmiṃ viharati tiriyaṃ anantasaññī, (D. I.23). IV is said to be the point of view of a 'reasoner' (takkī, *loc. cit.*), according to whom the epithets 'finite' and 'infinite' cannot be predicated of the world and hence it is said that he disagrees with I, II, and III (*loc. cit*). This too, it may be noted, is not a contradictory statement but a different point of view quite distinct from I, II and III. It resembles the point of view of Kant who showed in his 'Critique of Pure Reason' that one could on the one hand argue that 'the world was finite (in space and time)' and also that 'the world was infinite'. However, since the conclusions contradict each other we are faced with an antinomy, the truth being that spatial (and temporal) attributes (e g. finite, infinite) are subjective.¹ It will be noticed that according to this four-fold logic, a person who denies that 'the world is finite' and asserts the contradictory, namely that 'it is not the case that the world is finite' may be one who holds II, III, or IV. The Aristotelian logic of two alternatives tends to obscure these finer distinctions while the above logic of four alternatives tends to highlight them, showing them as separate logical alternatives.

c12.45

¹ Immanuel Kant's *Critique of Pure Reason*, Tr by N K Smith, London, 1933; for the proofs that the world is finite and that it is infinite, *v.* pp. 396–402; space is an *a priori* form of intuition, not reality, pp. 71 ff, 77, 80–91, 123, 163, 244, 399–400, 440, 449

c12.46

(575) It is, however, necessary for the purpose of this four-fold logic to interpret II, as Bahm pointed out (*v. supra*, 567), as the contrary or opposite and not as the contradictory of (I). For example, if we take the epithet *sukhī-* (experiencing pleasure, happy), we can have the assertion, I, so sukhī, 'he is happy'. Since 'na sukhī' is the contradictory and 'dukkhī' (unhappy) the contrary of sukhī-, i.e. II would be 'so dukkhī'. III would be 'so sukhī ca dukkhī ca' and IV 'so n'eva sukhī na dukkhī'. III and IV are here not self-contradictory statements. III describes the person, who experiences both pleasurable as well as painful sensations and IV the person whose experiences have a neutral hedonic tone, being neither pleasurable nor painful. We find in fact the following four-fold predication about the hedonic experiences of the soul, which are mutually exclusive and together exhaustive of all the possibilities:

c12.47

I Ekanta-sukhī attā hoti, i e. the soul is extremely happy.
II Ekanta-dukkhī attā hoti, i.e the soul is extremely unhappy.
III Sukha-dukkhī attā hoti, i e. the soul (has mixed feelings of) happiness and unhappiness
IV Adukkhamasukhī attā hoti, i e the soul has no (feelings of) happiness or unhappiness.
D I.31.

c12.48

(576) We may cite numerous examples of this type from the Canon. To take an example which is often repeated:

c12.49

I Puggalo attantapo . . ., i.e. the person, who torments himself (e g the ascetic).
II Puggalo parantapo . . ., i.e the person, who torments others (e g. the hunter).
III Puggalo attantapo parantapo ca . . ., i e. the person who torments himself as well as others (e g. the king who initiates and participates in a sacrifice causing the destruction of life with much discomfiture to himself).
IV Puggalo n'evattantapo na parantapo . . ., i.e the person who neither torments himself nor others (e g the arhant).
M 1 341 ff.; Pug. Pañ 55 ff.

c12.50

One of the reasons for adopting this logic seems to be the fact that this four-fold schema gave a better and finer classification of the empirical data (thus preventing much ambiguity in utterances) than that offered by the strictly dichotomous division.

c12.51

(577) While the majority of the four forms of predication are explicable on the above basis, taking (II) as the contrary and not the contradictory of (I), there are some instances where we are faced with an apparent difficulty. This is where (II) appears by virtue of its form, to be the contradictory of (I):

c12.52

e.g. I atthi paro loko, i.e. there is another world.
II natthi paro loko, i.e. there is no other world.
III atthi ca natthi ca paro loko, i e. there is and is no other world.
IV n'ev'atthi na n'atthi paro loko, i.e there neither is nor is there no other world.
D I.27.

c12.53

(578) Are we to regard, natthi paro loko, as the contradictory or the contrary of atthi paro loko? Going purely on the morphology or the word-form of the sentence it would appear to be the contradictory. But this is due to the ambiguity of language. As Stebbing says, ‘A considerable amount of dispute turns up upon the fact that our words have ragged edges. We cannot always be sure whether the propositions we are concerned to maintain *contradicts* our opponent’s assertion or is merely *incompatible* with it. If the two assertions are contradictory, then in disproving our opponent’s assertion we thereby establish our own. But if they are *contraries*, the disproof of his assertions leaves our own still to be established’ (*op. cit.*, p. 62). We maintain that, the proposition, natthi paro loko, should according to its context be treated as the contrary and not the contradictory of, atthi paro loko, despite the linguistic form. The reason for this would be clearer if we proceed from the unambiguous to the ambiguous cases. In the unambiguous case the difference between the contrary and the contradictory is reflected in the linguistic form, viz.

c12.55

The distinction is less clear, but still noticeable in:

c12.57

The difference is not noticeable but nevertheless has to be maintained in the following, since the contradictory of (I) is ambiguous and could mean II, III or IV.

c12.59

(579) The fact that propositions of the form, ‘S is P’ and ‘S is not P’ are not always mutually contradictory *in actual usage* is clear from the fact that together they do not always exhaust the possibilities as they ought to according to the law of Excluded Middle Lewy¹ in discussing this problem says that in the case of Mr Braithwaite we are inclined to say that ‘Mr Braithwaite is neither bald nor not bald’ without saying that either the proposition ‘Mr Braithwaite is bald’ or the proposition ‘Mr Braithwaite is not bald’ is true. Lewy suggests at the end of this article that one of the ways of avoiding this problem is ‘to contract a logical calculus’ in which the principle of Excluded Middle does not hold. But it is just this kind of logical calculus, as we have shown in our article (*v. supra*, 565), that is being used in the Pāli Canon, treating statements of the form, ‘S is both P and notP’ and ‘S is neither P nor notP’ (giving a different use for each of these two types) as contingent propositions which are empirically meaningful in certain situations in which, considering the facts, we are not justified in using the propositional forms, ‘S is P’ or ‘S is not P’.²

c12.60

¹ *v.* C Lewy, ‘Calculuses of Logic and Arithmetic’, pp 36 ff in *PAS.*, Supplementary Vol 20

c12.61

² For a full discussion of this problem, *v* pp 49–55 in our article

c12.62

(580) Sometimes in adopting this four-fold schema for purposes of classification only three of the possible four alternatives are mentioned, presumably when no instances are found for the other alternative. Thus, we have:

c12.63

I. saññī-vādā, D. I.31.
II. asaññī-vādā, D. I.32.
IV. n’evasaññī-nāsaññī-vādā, D. I.33.

c12.64

" But there is no mention of saññīasaññī-vādā, presumably because there were no instances of theories which asserted that ‘the soul was both conscious and unconscious’ after death. This is added evidence that the list of theories in the Brahmajāla Sutta need not be considered a merely artificial list, giving mere logical possibilities (*v. supra*, 141).

c12.65

(581) Whether the four-fold logical alternatives were adopted by the Buddhists from the Sceptics or not (*v. supra*, 190), the logic of it was the same. The difference lay in the different uses to which they put the schema. The Sceptics used it to show that each of the four alternatives may or may not be true and that the truth cannot be known. The Buddhists considered at least one of the alternatives to be true in any particular case and made use of it for purposes of logical classification, when they did not resort to dichotomous division (based on the two-fold logic, *v. supra*, 497 f.). They also rejected all four alternatives, when they considered the question meaningless (i.e. a thapanīya pañha) and rarely negated all four alternatives when they considered the answer 'yes' to each of the alternatives as misleading (*v. supra*, 481).

c12.66

(582) Let us illustrate all this. When we have four alternatives, then according to the Law of Non-Contradiction,¹ not more than one alternative can be true and according to the Law of Exclusion (for there is no Middle² since we are dealing with four possibilities) at least one alternative must be true. Now in the case of a two-fold logic or a logic with two alternatives (like the Aristotelian) we have four truth-possibilities since each of the possibilities p or not-p may be true or false. We may represent this by a truth-table as follows:

c12.67

¹ *v.* K. N. Jayatilleke, *op. cit.*, p. 52.

c12.68

² We can still speak of the recognition of the Law of the Excluded Middle in the sense that no 'middle' between 'true' and 'false' is recognized but owing to the ambiguity of the phrase 'Excluded Middle' we have dropped it.

c12.69

| | 1 | 2 | 3 | 4 |
| p | T | F | T | F |
| not-p | F | T | T | F |

c12.70

Of the four possibilities (3) and (4) are excluded by the Laws (Non-Contradiction and Excluded Middle) of the system and we are left with (1) and (2) as the logical alternatives or the alternative possibilities. Likewise, in a four-fold logic or a logic of four alternatives, we are left with only four possibilities, the others being ruled out by the Laws (Non-Contradiction and Exclusion) governing the system. If we draw up a truth-table it will be seen that out of sixteen truth-possibilities, only four (*v.* 1-4, below) alternatives are permissible:

c12.71

| | 1 | 2 | 3 | 4 | 5 | 6 | 7 | etc., up to 16 |
| I. p | (T) | F | F | F | T | F | T | |
| II. notp (contrary) | F | (T) | F | F | T | F | T | |
| III. both p and notp | F | F | (T) | F | T | F | T | |
| IV. neither p nor notp | F | F | F | (T) | T | F | F | |

c12.72

(583) Let us take an example from the Nikāyas that confirms this. Supposing I say '*I know* what has been seen, heard, sensed, thought, attained, sought and reflected upon by the class of recluses and brahmins, *then it would be false for me to say, I do not know* what has been seen, heard . . . *it would likewise be false for me to say, I know and* *do not know what has been seen, heard* ... *and false for me to say, I neither know nor do not know what has been seen, heard* ...'. Now this is a literal translation of a passage appearing at A. II.25, viz. yam .. sassamana-brāhmaṇiyā pajāya ... diṭṭham sutam mutam viññātam pattam pariyesitam anuvicaritam manasā *tam aham jānāmi*. . Yam . diṭṭham sutam ... *tam aham na*¹ *jānāmī ti vadeyyam tam mama assa musā, tam aham jānāmi na ca jānāmī ti vadeyyam tam p'assa tādisam eva, tam aham n'eva jānāmi na na janāmī ti vadeyyam tam mama assa kali*. This example illustrates the fact that when one alternative was taken as true, it was assumed that every one of the other alternatives were false. In this case when (I) is true, it is said that each of the alternatives (II), (III) and (IV) were false.

c12.73

¹ This na has been erroneously omitted from the text.

c12.74

(584) When the four alternatives happened to be those of a thapanīya pañha or a meaningless question all four alternatives were *rejected* rather than *negated* because the question in each of the alternatives was not considered to be a proper question (kallo pañho) Thus we have the four alternatives'

c12.75

I. Channam phassāyatanānam asesavirāganirodhā atth'aññam kiñcī ti?
i.e. is there anything else after complete detachment from and cessation of the six spheres of experience?
II ... natth'aññam kiñcī ti?
III ... atthi ca n'atthi c'aññam kiñcī ti?
IV. ... n'ev'atthi no n'atth'aññam kiñcī ti?

c12.76

The replies to each of these questions is of the form 'mā h'evam', i.e. do not say so (A. II.161). The response to the other avyākata-questions is also similar since these questions are 'not answered but set aside' (avyākatāni thapitāni, M I 426, *v. supra*, 481) In other words these questions were *not negated but rejected*. Raju (*v. supra*, 566) and Bahm (*v. supra*, 567) have therefore misdescribed their nature as far as the Pāli Canonical position is concerned by calling this doctrine that of 'four-cornered negation' when it ought properly to be called 'four-cornered rejection'.

c12.77

(585) In fact, it is all the more necessary to distinguish the four-cornered rejection from four-cornered negation, which is also rarely met with in the Nikāyas. Consider the answers to the following four alternative forms of a question:

c12.78

I. Kim nu kho ... vijjāya antakaro hotī ti? Is it the case that one attains the goal by means of knowledge?
II. Kim nu kho ... caraṇena antakaro hotī ti? Is it the case that one attains the goal by means of conduct?
III. Kim nu kho ... vijjācaraṇena antakaro hotī ti? Is it the case that one attains the goal by means of both knowledge and conduct?
IV. Kim nu kho ... aññatra vijjācaraṇena antakaro hotī ti? Is it the case that one attains the goal without knowledge and conduct?

c12.79

Taking II as the contrary of (I), the question is in the form of the four logical alternatives. The answer to each of these four alternatives is of the form 'na h'idam', i.e. it is not so (note the difference from the above), the reason being that while 'knowledge' and 'conduct' are necessary conditions for final salvation they are not sufficient conditions. This denial of all the four logical alternatives is in apparent violation of the Law of Exclusion (*v. supra*, 582), but this is not a peculiarity of this logic, since we meet with this paradoxical situation even with Aristotelian logic, e.g. when a non-smoker is confronted with the question 'have you given up smoking?'

c12.80

(586) It is necessary to distinguish the above four-fold logic from the logic of the Jain *syādvāda*, which is radically different from it. We shall refer to the seven forms of predication of Jainism by the Indian numerals 1, 2, 3, 4, 5, 6, 7 and the four logical alternatives by the Roman numerals I, II, III and IV. Now it was the contention of Raju, Bahm and Miyamoto (*v. supra*, 566, 567, 569) that 4 (syādavaktavya-) was the same as IV (S is neither P nor notP), while Bahm and Miyamoto further identified (*v. supra*, 567, 569) 1, 2, 3 of Jainism with I, II, III of Buddhism respectively. This is mistaken and is due to a failure to understand the logic of each system, as a result of which they seem to have been misled by mere superficial similarities. It appears on the face of it that we could identify 1, 2, 3, 4 with I, II, III, IV respectively:

c12.81

c12.82

But if we examine the interpretation and usage of each we may observe the radical difference behind the superficial resemblance. The interpretation given of syād avaktavyam¹ is that it is logically impossible to assert 1 and 2 of an object *simultaneously* (3 asserts them *successively*). It is implied that the object has a mysterious character, which is impredicable. IV has no such implication and asserts on the contrary that there is an instance of S, such that it does not have the characteristic P or its contrary notP. For example, one of the four possible types of individuals is called 'n'ev'attahitāya paṭipanno na parahitāya' (i.e. he does not act in his own welfare nor for the welfare of others, A. II.95), which means that there is an individual S such that he does not have the characteristic 'attahitāya paṭipanno' (P) nor the characteristic 'parahitāya paṭipanno' (notP). It does not mean that S is indescribable in any way nor that there are no individuals, of whom the characteristics P and notP cannot be predicated simultaneously for in fact the person described as 'attahitāya ca paṭipanno parahitāya ca' (*loc cit*) has these characteristics simultaneously.

c12.83

¹ *v.* Radhakrishnan, *Indian Philosophy*, Vol I, p 303, cp Raju, *op cit*, p 700

c12.84

(587) Another significant difference is that I, II, III, and IV are logical alternatives in Buddhism, whereas this is not the case with the seven forms of predication in Jainism. In Buddhism, as in Aristotelian logic, only one alternative could be true, but in Jainism each of the seven alternate forms of description (or any disjunction of them) could be true. According to Buddhism an object correctly describable by I cannot at the same time be correctly described by II, III, or IV, whereas in Jainism an object correctly describable by 1 is also correctly describable by 2–7, which supplement 1 and do not exclude it.

c12.85

(588) We may define the difference between the Jain logic and the Aristotelian (which is similar in this respect to the Buddhist, *v. supra*, 582, 587) by drawing up truth-tables to illustrate the possibilities that each allows for. For Aristotelian logic, as we have already shown, we have four truth-possibilities of which two, i.e. the cases in which p and not-p are both true and both false are ruled out by the laws of the system. But according to Jain logic, if we confine ourselves to the first two forms of predication only (to avoid complexity) all four truth-possibilities are permissible, *viz.*:

c12.86

| | 1 | 2 | 3 | 4 |
| p (syād asti) | T | F | T | F |
| not-p (syān nāsti) | F | T | T | F |

c12.87

(589) It will be noticed that (3) and (4) are not ruled out, since p may be true from one standpoint and not-p from another; similarly p may be false from one standpoint while not-p is also false from another standpoint. It is necessary to remember that Jain logic admits of falsity (nayābhāsa). This happens when the particular assertion is not possible from the standpoint from which it is made. For instance, if I assert that a thing (which exists) does not exist (syād nāsti) from the point of view of its own form (svarūpa), substance (svadravya), place (svaksetra) and time (svakāla), the statement would be false. So in drawing up our truth-table to represent the possibilities according to Jain logic, it is necessary to take account of another variable (in addition to the truth-value), namely the standpoint (naya). Let us for the sake of simplicity, take two alternative forms of predication (syād asti and syād nāsti) and two standpoints, x and y. This gives sixteen truth-possibilities in all, since there would be four possibilities for each of the four possibilities in the above table, viz.:

c12.88

| | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 |
| (p) syād asti | Tx | Tx | Ty | Ty | Fx | Fx | Fy | Fy |
| (not-p) syān nāsti | Tx | Ty | Tx | Ty | Fx | Fy | Fx | Fy |

c12.89

| | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 |
| (p) syād asti | Tx | Tx | Ty | Ty | Fx | Fx | Fy | Fy |
| (not-p) syān nāsti | Fx | Fy | Fx | Fy | Tx | Ty | Tx | Ty |

c12.90

(590) Since it would be self-contradictory to assert jointly the truth or the falsity of both alternates from the same standpoint we can rule out possibilities 1, 4, 5 and 8 by this Law of Non-Contradiction, which holds good for this system. Then again, if we know that p can be true only from standpoint x and not-p only from standpoint y, then it would follow (logically) that p is false from standpoint y and not-p false from standpoint x This means that p would have the truth-possibilities Tx and Fy only and not-p the truth-possibilities Ty and Fx only, all other possibilities (i.e. 3, 6, 10, 11, 12, 13, 14 and 15 in addition to the above) being ruled out. This leaves us with four truth-possibilities as follows.

c12.91

| | 1 | 2 | 3 | 4 |
| p (syād asti) | Tx | Fy | Tx | Fy |
| not-p (syād nāsti) | Ty | Fx | Fx | Ty |

c12.92

(591) This shows the radical difference between the logic of *syādvāda* and of the *catuṣkoṭi*, i.e. the Buddhist logic of 'four alternatives'. The former represents seven forms of predication, which supplement each other, each of which may be true simultaneously with the others from its own standpoint. The latter represents four alternatives of which (as understood in the Canon)¹ only one could be true. In this respect, this logic resembles the Aristotelian, except for the fact that the latter is a two-valued logic of two alternatives while the former is a two-valued logic of four alternatives. Until recently it was believed in the Western world that Aristotelian logic was the only logic and that it reflected the structure of reality but, with the discovery of many-valued logics by Lucasiewicz and Lobochevsky, this view is no longer universally held. This means that our choice of a logical system is to some extent arbitrary and dependent on the needs and nature of our discussion. The Buddhist four-fold logic is in this respect no more true or false than the Aristotelian and its merits should be judged by its adequacy for the purposes for which it is used. It is not at all necessary to feel helpless before it (as Poussin did!), if its nature and significance is understood.

c12.93

¹ Nāgārjuna's position is different but this is beyond the scope of our study

c12.94

(592) Before we leave this topic, we may make some observations on an article by Robinson,² in which he makes a few comments on the *catuṣkoṭi*. He calls this the *Tetralemma* (*op. cit.*, p. 301), but since we have treated the four alternatives as propositional functions, following Schayer,³ and since they are basic to a whole system of logic, we prefer to call this 'the four-fold logic', we do not call Aristotelian logic 'the dilemma' because it is a logic of two alternatives. We agree with Robinson when he says that 'the four members' (i e the four alternatives) are 'in a relation of exclusive disjunction' ('one of but not more than one of "a", "b", "c", "d" is true') (*loc. cit.*); this we have actually shown to be the case (*v. supra*, 583). We cannot agree with his observation that 'the tetralemma resembles the four Aristotelian forms in some ways Both sets comprise propositions constructed from two terms and the constants (functors) 'all', 'some', and 'not' (*op. cit.*, p. 303). If we do so, we treat (I) as an universal affirmative proposition, which it is not (at least always) (*v. supra*, 574). For similar reasons, the explanation of II as corresponding to an Aristotelian E-type proposition, of III as 'the conjunction of I and O forms' (*loc. cit.*) and IV as the 'conjunction of E and A forms' (*loc. cit.*) is unsatisfactory. His proposal to interpret the fourth alternative as: 'No x is A and no x is not A' (*op. cit.*, p. 302), which is true 'when x is null' (*loc. cit.*) is contradicted by usage in the Pāli Nikāyas (*v. supra*, 576, IV).

c12.95

² Richard H Robinson, 'Some Logical Aspects of Nāgārjuna's System' in, *Philosophy East and West*, Vol 6, No 4, January 1957, pp 291–308.

c12.96

³ 'Altindische Antizipationen der Aussagenlogik' in *Studien zur indischen Logik*, Extrait du Bulletin de l'Academie Polonaise des Sciences et des Lettres cracovic 1933, p 93, Schayer represents the four alternatives as p, ~p, (p.~p), (~p ~(~p)) We are, however not treating 'p' as the contradictory of 'p' and hence we have written 'notp' instead (*v. supra*, 186)

c12.97

(593) Let us examine the conception of truth as we find it in the Nikāyas. There is no direct inquiry into the nature of truth (in the epistemological sense) in them, but the value placed on truth (in the wider sense) was so great that some observations about the nature of truth (in the above sense) were, perhaps, inevitable.

c12.98

(594) In the Abhayarājakumāra Sutta, we find statements classified according to their truth-value, utility (or disutility) and pleasantness (or unpleasantness). The intention of the classification is to tell us what kinds of propositions the Buddha asserts If propositions could be true (bhūtam, taccham) or false (abhūtam, ataccham), useful (atthasaṃhitam) or useless (anatthasaṃhitaṃ), pleasant (paresam piyā manāpā) or unpleasant (paresam appiyā amanāpā), we get eight possibilities in all as follows:

c12.99

| 1. | True | useful | pleasant |
| 2. | ” | ” | unpleasant |
| 3. | ” | useless | pleasant |
| 4. | ” | ” | unpleasant |
| 5. | False | useful | pleasant |
| 6. | ” | ” | unpleasant |
| 7. | ” | useless | pleasant |
| 8. | ” | ” | unpleasant |

c12.100

The text reads as follows: "The Tathāgata does not assert a statement which he knows to be untrue, false, useless, disagreeable and unpleasant to others (i.e 8). He does not assert a statement which he knows to be true, factual, useless, disagreeable and unpleasant to others (i.e. 4). He would assert at the proper time a statement which he knows to be true, factual, useful, disagreeable and unpleasant to others (i.e. 2). He would not assert a statement which he knows to be untrue, false, useless, agreeable and pleasant to others (i.e. 7). He would not assert a fact there is a next world, if one thinks that there is no next world, that would be a false conception. When in fact there is a next world, one asserts the statement that there is no next world, that would be a false statement' (Santaṁ yeva kho pana paraṁ lokaṁ; natthi paro loko ti'ssa diṭṭhi hoti, sā'ssa hoti micchādiṭṭhi. Santaṁ yeva kho pana paraṁ lokaṁ: natthi paro loko ti saṅkappeti, svāssa hoti micchāsaṅkappo. Santaṁ yeva kho pana paraṁ lokaṁ: natthi paro loko ti vācaṁ bhāsati, sā'ssa hoti micchāvācā, M. I.402). Thus, while false propositions entertained as beliefs or conceptions or expressed as statements are considered false, when they do not correspond with or deny facts, true beliefs, conceptions or statements are said to be those which reflect or correspond with fact. The words used for true beliefs, conceptions or statements are sammādiṭṭhi, sammāsaṅkappo and sammāvācā respectively, which literally mean '*right* belief, etc.', but here 'right' (sammā) being the opposite of 'micchā' (false) is synonymous with 'true': 'When in fact there is a next world, the belief occurs to me that there is a next world, that would be a true belief ...' (Santaṁ yeva kho pana paraṁ lokaṁ: atthi paro loko ti'ssa diṭṭhi hoti, sā'ssa hoti sammā diṭṭhi, M. I.403).

c12.101

(597) Though truth is defined in terms of correspondence with fact, consistency or coherence is also considered a criterion of truth. We have already cited an example from the Nikāyas, where it is clearly shown that, when two statements contradict each other, it cannot be the case that both statements are true for 'if p is true, not-p is false and if not-p is true, p is false' (*v. supra*, 562). In the Suttanipāta referring to numerous theses put forward by various theorists the question is asked 'Claiming to be experts, why do (they) put forward diverse theories—are truths many and various ...?' (kasmā nu saccāni vadanti nānā ... kusalā vadānā: saccāni su tāni bahūni nānā ... Sn. 885) and answered: 'Truths, indeed, are not many and various' (na h'eva saccāni bahūni nānā ... Sn. 886). It is in this context that the statement is made that 'truth is one without a second' (ekaṁ hi saccaṁ na dutīyaṁ atthi, Sn. 884). The Buddha in arguing with his opponents appeals to this principle of consistency by showing that their theories are false because they are contradicting themselves. Thus, in the debate with Saccaka, the Buddha says at a certain stage in the discussion, referring to his opponent's statements that 'his later statement is not compatible with the former nor the former with the later' (na kho te sandhīyati purimeṇa vā pacchimaṁ, pacchimena vā purimaṁ, M. I.232).

c12.103

elephant. They touch various parts of the elephant such as the forehead, ears, tusks, etc. They are then asked to describe the elephant and each reports mistaking the part for the whole that the elephant was like that portion of the elephant which was felt by him. There are ten conflicting accounts in all corresponding in description to the ten parts touched and these are compared to the ten avyākata-theses put forward by the various recluses and brahmins. If we interpret the parable literally one would have to say that their theses too mistakenly describe the part for the whole and in so far as they constitute descriptions of their partial experience, they have an element of truth but are deluded in ascribing to the whole of reality what is true only of the part or in other words what is partially true. Since it were these very avyākata-theses that were called pacceka-saccas it would appear to be not without justification to translate this word as ‘partial truths’. One may in fact even suggest, though with little historical justification, that these theses were ‘indeterminate’ (avyākata-) in the sense of being neither true nor false¹ analogous to the sense in which avyākata- is used to denote what is ‘neutral’ in moral contexts where ‘what is indeterminate (avyākata-) are acts which are neither good nor evil . . .’ (ye ca dhammā kiriyā n’eva kusalā na akusalā . . . ime dhamma avyākatā, DhS., 583, p. 124). If such an interpretation is to be justified, we would have to say that these avyākata-theses were a product of partial descriptions of reality, their error consisting in regarding these partial accounts as descriptions of the whole of reality. They would be the misdescribed experiences of different thinkers like the blind men’s accounts of the elephant. In fact, it is almost suggested in the Brahmajāla Sutta that not merely the avyākata-theses but all the sixty two philosophical theories ‘result from impressions’ (phassapaccayā, D. I 42, 43; cp. te vata aññatra phassā paṭisaṃvedissantī ti n’etam ṭhānam vijjati, i.e. it is impossible that they would entertain (these theories) without the impressions they had, D. I.43, 44)—i.e. perceptive, sensory and extrasensory and cognitive experience—and that these theories were presumably only partial accounts of reality.

c12.104

¹ This is in fact a sense in which ‘indeterminate’ is used in logic, e g ‘. . . if a three-valued logic is proposed in which the law of the excluded middle is replaced by the trichotomy *every proposition is either true or false or indeterminate* . . .’ A Pap, *Semantics and Necessary Truth*, Yale University Press, New Haven, 1958, p 169

First 99 of 156 passages. Read on

Early Buddhist Theory of Knowledge c12: Chapter VII: Logic and Truth

Chapter VII: Logic and Truth

Cites (37)