Buddhist Philosophy in India and Ceylon c26.s3: 4. Dharmakīrti's Theory of Inference

Buddhist Philosophy in India and Ceylon c26.s3

Language: English

4. Dharmakīrti's Theory of Inference

c26.s3.1

Inferential knowledge is essentially of generality; it cannot reach to the immediateness of presentation; the mountain inferred as fiery is not the presentation of the fire on the mountain at a definite point of time. Its validity must be verified, if desired, by the usual process of action; we must, for instance, go to the mountain and see the fire, which we have inferred from the smoke cloud. An advance is made on Dignāga in the process of examination of the conditions of a correct middle term; the middle must be present in those things only in which the thing to be inferred exists, and absent in all those things in which it is not found; the rule must be observed fully in either case, or the inference will be doubtful. Cases where there is the necessary invariable connexion are limited to three only, in Dharmakīrti's view, namely, identity of nature or essence (*tādātmya*), effect of a cause, and non-perception or non-existence. It has been ingeniously, but quite legitimately, pointed out¹ that this gives a classification of judgement on the basis of positive and negative, with the division of the affirmative judgements into analytical, relation of identity, or causal, relation of effect to cause. Reasoning, therefore, which is defined as knowledge of the inferable derived through the middle term, is of three main types: ‘This is a tree, because it is a pine,’ relation of identity, the species allowing the inference of the genus; ‘Here there is fire, because there is smoke,’ the cause fire is inferred from the effect smoke; and ‘Here is no smoke, because it is not perceived,’ as smoke would, if existent, be perceived. This third variety, based on non-perception or non-existence, is divided into eleven classes, of a somewhat needlessly varied character. The ground of the inference, it will be observed, is thus always a general proposition, which rests on mental activity, and falls under one of three categories.²

c26.s3.2

¹ Stcherbatskoi, *Museon*, v (1904), p. 144, n. 6.

c26.s3.3

² NB., p. 31; NB., pp. 104 ff. ; TRD., pp. 41 ff, where four classes of non-perception are enumerated: perception of the contrary, of contrary effect; non-perception of cause, of identity.

c26.s3.4

Inference for the sake of another, or syllogism, is defined as the verbal declaration of the middle term—that is, when the reason is set out in words in order to produce a conviction in others.³ Inference is properly a form of knowledge, and words are, therefore, only inference in a secondary sense, namely, as producing knowledge, the name of the cause being derived from the effect. Such a form of inference is two-fold, either direct or homogeneous (*sādharmyavat*) or indirect or heterogeneous (*vaidharmyavat*): either ‘Sound is non-eternal, because it is a product; all products are non-eternal, like a pot’, or ‘Sound is non-eternal, because it is a product; no non-eternal thing is a product, like ether’. This, however, is a needlessly full form of the syllogism; Dharmakīrti is content with the simple form ‘The hill is fiery, because it is smoky’, the express formulation of the general proposition being unnecessary, since it is contained in the reason ‘because it is smoky’. The omission of the example he defends¹ against Dignāga by insisting that the term ‘smoke’ includes the case cited in the example and also the negative example; but he concedes that it has the value of giving particularity to what is pointed out in a general way by the middle term; the complete enunciation of the reasoning with the concrete example is more effective.

c26.s3.5

³ NB., pp. 108 ff. ; NBT., pp. 46 ff.

c26.s3.6

¹ NB., pp. 115 f. ; MSIL., p. 114, n. 2.

c26.s3.7

Fallacies² are classed by Dharmakīrti according to the old division of fallacies of the thesis, i. e. the minor term combined with the major, e. g. the fiery hill, as insisted upon by Dignāga; of the middle; and of the example, divided according to the homogeneous and the heterogeneous example. His classification of fallacies of the middle, the only important class, is on the basis that a fallacy arises if a characteristic of the middle is unproved (*asiddha*), e. g. ‘Sound is eternal, because it is visible,’ visibility being admitted by neither party to the supposed argument; uncertain (*anaikāntika*), e. g. ‘Sound is non-eternal, because it is knowable,’ where ‘knowable’ is too wide since it covers both eternal and non-eternal things; and as contradictory (*viruddha*), e. g. ‘Sound is eternal, because it is a product,’ the middle contradicting the major. He differs from Dignāga in rejecting two of the forms of fallacy allowed by the former, the first³ being reduced very sensibly to a mere form of contradiction, while the latter⁴ is laid aside on the interesting ground that it does not rise from true reasoning but is based on the scriptures of the two disputants, and scripture alone is no authority for Dharmakīrti.

Linked texts (3)

དེ་ཁོ་ན་ཉིད་བསྡུས་པའི་དཀའ་འགྲེལ — Classical Tibetan
རིགས་པའི་ཐིགས་པའི་རྒྱ་ཆེར་འགྲེལ་པ — Classical Tibetan
Nyāyabindu — Dharmakīrti, Sanskrit