A History of Indian Philosophy Volume 4

Indian Pluralism

by Surendranath Dasgupta | 1949 | 186,278 words | ISBN-13: 9788120804081

This page describes the philosophy of inference (anumana): a concept having historical value dating from ancient India. This is the second part in the series called the “madhva logic”, originally composed by Surendranath Dasgupta in the early 20th century.

Part 2 - Inference (anumāna)

The cause of inference is a faultless reason (through which by virtue of its association anything can be ascertained). The nature of this association or concomitance is described by Jaya-tīrtha as being inseparable concomitance (avinābhāva). Vyāsa-tīrtha urges in the Tarka-tāṇḍava that this inseparable concomitance ought really to mean contradiction of experience leading to inadmissible assumption or implication (anupapatti). When anything experienced in a particular space-time relation must be invalid except on the assumption of some other thing, in some other space-time relation, it must be admitted that such a particular relation subsisting between the two is a relation of concomitance (vyāpti), leading to the inference of the latter through the former[1].

Vyāsa-tīrtha urges that this view of. inference has also been supported by Madhva in his Pramāṇa-lakṣaṇa, where he says that the residual method (pariśeṣa) is the essential method in all cases of valid inference[2]. Reduction to absurdity in regard to any valid experience is what necessitates the supposition in an act of inference.[3] Jaya-tīrtha in his Pramāṇa-paddhati has indeed defined concomitance (vyāpti) as inseparability (avinā-bhāva); this inseparable concomitance cannot be described as being in all cases agreement in absence, i.e., the absence of the reason, hetu, in all cases of the absence of the probandum (sādhya), or the inferred entity; for there are cases where, in spite of the absence of such negative instances, inference is possible, e.g., sound is expressible on account of its being an object of knowledge; now here no such negative instance is available where there would be no expression; hence in such cases of impossible-negative (kevalānvayi) inferences the above definition of concomitance, which requires the existence of negative instances for the ascertainment of concomitance, would not apply. Also no kind of spatial association of the reason and consequence (sādhya) can be urged as being an indispensable condition of concomitance: for there can be the inference of rain in the upper part of a country from perceiving a rise of water in the river in the lower part, and there is no spatial contiguity between the reason and consequence. So the main point in concomitance determining inference is the reduction of an incontrovertible experience into an impossibility, which necessitates the assumption of the inferred entity. It is this which has also been described as the law of unconditional and invariable association (sāhacarya-niyama). In the well-known example of fire and smoke what is described as the unconditional and invariable coexistence of the absence of smoke in all cases of the absence of fire is also a case of reductio ad absurdum (anupapatti). It would apply with equal force in the cases of impossibfe-negatives (kevalānvayi) ; for there also the impossible absence of the consequence would render the reason absurd; and hence the assumption of the consequence is necessary.

Vyāsa-tīrtha refutes at great length the definition of inference given by Gaṅgeśa in his Tattva-cintāmaṇi, where he explains concomitance as the coexistence of consequence and reason as qualified by the fact of the absence of the latter in each case of the absence of the former. Had it not been for the fact that in inferences of the type of impossible-negatives (kevalānvayi) no negative instances are available where we might have been acquainted with cases of absence of the consequence being also cases of absence of the reason (sādhyābhāvavad-avṛttitvam), Gaṅgeśa would have been glad to define concomitance (vyāpti) as unconditional and invariable non-existence of the reason in all cases of the non-existence of the consequence (sādhyābhāvavad-avṛttitvam). But owing to the above difficulty Gaṅgeśa was forced to define concomitance as coexistence (sāmānādhikaraṇya) of the consequence and reason where the reason is also qualified as the repository of the negation of all possible conditions which could invalidate its unconditional and invariable relation to the consequence (sādhya)[4]. The insight of Gaṅgeśa in formulating such a definition consists in this, that he thinks that universal existence of the reason in case of the consequence is alone sufficient for an inference of the latter from the former, provided that the reason is pure and unmixed by the presence of any other entity. It is the presence of other entities mixed with the reason that may invalidate its universal coexistence with the consequence; so, if that could be eliminated, then mere universal existence of the reason in cases of the consequence would be sufficient to establish a relation of concomitance between the former and the latter.

Vyāsa-tīrtha, however, points out that the existence of the reason in cases of the consequence is not universally valid in all cases of inference. Thus in the inference of rain in the upper regions from perceiving a rise of water in the river in the lower regions there is no spatial coexistence of the reason in the consequence; so also in the inference that the constellation Rohiṇī will shortly rise in the east because the constellation Kṛttikā has already risen. In all such cases and in all cases of inference the view of reductio ad absurdum (anupapatti) can always define concomitance in the best possible way and therefore can also serve as the best ground for all kinds of inference, including the class known as impossible-negatives (kevalānvayi). For in the example given of that class, “this is expressible because it is an object of knowledge”, we can argue that the denial of non-expressibility is a necessary postulate for the validity of the incontrovertible experience of its being an object of knowledge[5]. An objection may be raised that, non-expressibility being as fictitious an entity as a round square, there would be no meaning in further denying it. To this Vyāsa-tīrtha’s reply is that negation may apply even to the fictitious and the non-existent (aprāmāṇika)[6].

It is evident that this view of concomitance is a later development of theory by Vyāsa-tīrtha. For Jaya-tīrtha, in his Pramāṇa-paddhati, describes concomitance as being inseparable existence (avinābhāva), which he explains as invariable coexistence (sāhacarya-niyama) and also as invariable relation (avyabhicaritaḥ saṃbandhaḥ)[7]. Janārdana, however, in his commentary on the Pramāṇa-paddhati, holds that this sāhacarya-niyama of Jaya-tīrtha must be interpreted to mean the reductio ad absurdum of Vyāsa-tīrtha; otherwise it would be evident to all that his view of concomitance has been intended by the above definition of Jaya-tīrtha; and he supports his view by pointing out that both in the Pramāṇa-lakṣaṇa and in his commentary on the Pramāṇa-lakṣaṇa Jaya-tīrtha has included inference by residues (pariśeṣa) and implication (arthāpatti) within inference, as he thought that the methods of these are practically methods of inference itself[8]. But this only proves that pariśeṣa and arthāpatti are also kinds of inference and not that the method of anupapatti involved in them should be regarded as being the only possible form of inference. Had he thought this to be so, he would certainly have mentioned it and would not have limited his definition of concomitance to invariable coexistence (sāhacarya-niyama). Chalari-śeṣācārya, who faithfully follows the footprints of Jaya-tīrtha, often repeating his language also, explains this invariable coexistence of Jaya-tīrtha as “where there is smoke, there is fire”; but he remarks that this invariable coexistence means only the existence of an invariable relation of the reason to the consequence (atra sāhacaryaṃ hetoḥ sādhyena saṃbandha-mātraṃ vivakṣitam), and not merely existence in the same place (sāmānādhikaraṇya). Coexistence therefore is said to mean here unfailing relation to the consequence (avyabhicarita-sādhya-saṃbandho vyāptiḥ), and this is vyāpti [9]. He also refers to Gaṅgeśa’s definition of vyāpti, noted above, and points out that this definition of vyāpti would be inapplicable in those instances of inference where there is no spatial coexistence (e.g., the inference of rain in the upper regions from the rise of water in the river in the lower regions)[10]. He points out on the strength of such instances that concomitance cannot be defined as coexistence (sāmānādhikaraṇya), but is an unfailing relation which may hold between a cause and an effect existing in different places. On the strength of these instances Chalari-śeṣācārya argues in favour of concomitance without coexistence (vyadhikaraṇa-vyāpti) as being possible, and therefore advocates the dropping of the coexistence as a necessary condition of concomitance. Vyāsa-tīrtha seems to have profited by these remarks and, instead of remaining content with “unfailing relation” of Chaḷari-śeṣācārya, explained this “unfailing relation” as being the definite relation of reductio ad absurdum (anupapatti)[11].

Footnotes and references:


yad-deśa-kāla-saṃbaddhasya yasya yad-deśa-kāla-saṃbaddhena yena vinānupapattis tasyiva tena saha vyāptiḥ.
(MS., p. 1).


pariśeṣo’rthāpattir amumānam ity aviśeṣaḥ.

      and Pramāṇa-lakṣaṇa-ṭīkā, p. 27.


anumānam api āvaśyakānupapattyaiva gamakam.
(MS.,p. 2).


pratiyogy-asamānādhikaraṇa-yat-samānādhikaraṇātyantābhāva-pratiyogitā-vacchedakāvacchinnaṃ yan na bhavati tena samaṃ tasya sāmānādhikaraṇyaṃ vyāptiḥ.
Part II, p. 100 (ed. 1888, Bibliotheca Indica).


idaṃ vācyaṃ jñeyatvāt kevalānvayi anumānam.


tatra sādhyābhāvasya asattvād eva sādhyābhāve sati sādhanasya yopapattis tad-abhāva-rūpānupapatteḥ sattvāt; manmate’prāmāṇikasyāpi niṣedha-prati-yogitvāt.
(MS., p. 6).


Pramāṇa-paddhati, p. 30.


anupapatter vyāptitvaṃ ca pramāṇa-lakṣaṇe pariśeṣārthāpattiḥ anumā-viśeṣa ity atrārthāpattir iva anumānam api āvaśyakānupapattyaiva gamakam ity uktatvāt.
(MS., pp. 1-2).
      Also Pramāṇa-lakṣaṇa-ṭīkā, pp. 5-7.


Cf. Gaṅgeśa’s alternative definition of vyāpti in the section on Viśeṣa-vyāpti:

yat-saṃbandhitāvacchedaka-rūpavattvaṃ yasya tasya sā vyāptiḥ.
Part II, p. 156.


na tu samāṇādhikaraṇyam eva.
, p. 8 a.


Pramāṇa-candrikā, pp. 8a, 9.

Let's grow together!

I humbly request your help to keep doing what I do best: provide the world with unbiased sources, definitions and images. Your donation direclty influences the quality and quantity of knowledge, wisdom and spiritual insight the world is exposed to.

Let's make the world a better place together!

Like what you read? Consider supporting this website: