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 tarka (ratiocination): a concept having historical value dating from ancient India. This is the third part in the series called the “madhva logic”, originally composed by Surendranath Dasgupta in the early 20th century.

Part 3 - Tarka (ratiocination)

The determining oscillation constituent in a mental process leading to inference is called tarka or ūha[1]. Gautama, in his Nyāya-sūtra, describes it as being ratiocination with a view to knowledge of truth, involving attempt at determination of any fact as possessing a particular character, based on a proper enquiry regarding the cause of such a determination. Thus there is a desire to know the truth about the nature of selves as knowers. Are they produced or are they uncreated? If they were created, they would suffer destruction, like all created things, and would not suffer or enjoy the fruits of their own deeds. If they are uncreated, they may very well continue to exist for ever to suffer or enjoy the fruits of their deeds and undergo rebirth. So the self which undergoes rebirth and enjoys or suffers the fruits of all its deeds must necessarily be uncreated[2]. Vātsyāyana says that tarka is neither included within the accepted pramāṇas nor is it a separate pramāṇa, but is a process which helps the pramāṇas to the determination of true knowledge[3]. KeśavaMiśra, in his Tarka-bhāṣya, is inclined to include it under doubt[4]. But Annam bhaṭṭa, in his Tarka-dīpikā, says that, though tarka should properly be counted under false knowledge (viparyaya), yet, since it helps the pramāṇas, it should be separately counted[5]. The usefulness of tarka in inference consists in assuring the mind of the absence of any cases of failure of existence of the reason in the consequence and thereby helping the formation of the notation of the concomitance of the reason and the consequence[6]. Viśvanātha says that tarka clears away the doubts regarding the possible cases of failure (vyabhicāra) of the reason (e.g., if smoke existed in any instance where there was no fire, then fire would not be the cause of smoke), and thereby renders the knowledge of concomitance infallible and so helps the work of inference not in a direct, but in an indirect way (pāramparayā)[7]. Viśvanātha further adds that such a tarka is of five kinds, namely consideration of the fallacy of self-dependence (ātmāśraya, e.g., if the knowledge of this jug is produced by the knowledge of this jug, then it should be different from it), mutual dependence (anyonyāśraya, e.g., if this jug is the object of the knowledge as produced by the knowledge, then it should be different from this jug), circle (cakraka, if this jug is produced by something else produced by this jug, then it should be different from anything produced by something else produced by this jug), vicious infinite (anavasthā, e.g., if the class concept “jug” refers to all jugs, it cannot refer to things produced by the jug), contradictory experience (pramāṇa-bādhitārthaka-prasaṅga, e.g., if smoke exists where there is no fire, then it could not be produced by fire, or if there was no fire in the hill, there would be no smoke in it)[8].

Mathurānātha, in explaining the function of tarka in the formation of the notion of concomitance (vyāpti), says that, even when through noticing the existence of smoke in all known cases of fire and the absence of smoke in all those places where there is no fire, one decides that smoke is produced by fire or not, it is there that tarka helps to remove all legitimate doubts. As Gaṅgeśa shows, such a tarka would proceed thus: Either smoke is produced by fire or it is not produced there. So, if smoke is produced neither by fire nor by not-fire, it is not produced at all. If, however, there are the doubts whether smoke is from not-fire, or whether it can sometimes be where there is no fire, or whether it is produced without any cause (ahetuka), then none of us can have the notion of inseparable existence of fire in all cases of smoke so as to lead us to action (sarvatva sva-kriyā-vyāghātaḥ)[9]. A course of thought such as is called tarka is helpful to the formation of the notion of concomitance only when a large number of positive and negative cases has been actually perceived and a provisional certainty has been reached. Even when the provisional certainty is reached, so long as the mind is not cleared by the above tarka the series of doubts (sarkśaya-dhārā) might continue to rise[10]. It cannot be urged, says Gaṅgeśa, that, even when by the above method the notion of concomitance has been formed, there might still arise doubts whether fire might not be the cause of smoke or whether smoke might be without any cause; for, had it been so, you would not always (niyata) make fire when you wanted smoke, or eat when you wanted to satisfy your hunger, or use words to carry your ideas to others. Such regular attempts themselves show that in such cases there are no doubts (śaṅkā); for, had there been doubts, these attempts would not be so invariable. It is not possible that you would be in doubt whether fire is the cause of smoke and yet always kindle fire when you try to get smoke. The existence of doubt in such cases would contradict your invariable attempt to kindle fire whenever you wanted smoke; doubts can be admitted only so long as one’s actions do not contradict (sva-kriyā-vyāghāta) them[11].

Śrīharṣa, however, arguing from the Vedānta point of view, denies the power of tarka to dispel doubt. He urges that, if it is said that tarka necessarily dispels doubts in all cases and helps the formation of any particular notion of concomitance, then this statement must itself depend on some other notion of concomitance, and so on, leading us to a vicious infinite (anavasthā). Moreover, the fact that we know the universal coexistence of fire and smoke, and do not perceive any other element universally abiding in the fire which is equally universally coexistent with fire, does not prove that there is no such element in it which is really the cause of smoke (though apparently fire may appear as its cause). Our perception can certify only the existence or non-existence of all that is visible under the normal conditions of visual perception; it cannot say anything regarding the presence or absence of entities not controlled by these conditions, or we could only say that in the absence of fire there is absence of a specific kind of smoke; we could not say that there would be absence of all kinds of smoke; for it is just possible that there is some other kind of cause producing some special kind of smoke which we have not yet perceived; mere non-perception would not prove that such a special kind of smoke does not exist at all, since perception applies only to entities that are perceptible and is guided by its own conditions, and cannot therefore apply to entities which cannot be brought under those conditions[12]. The tarka which is supposed to dispel doubt by the supposition of contradiction of experience and which would thus support concomitance , not being itself grounded on concomitance, would naturally fail to do its part; for, if such groundless tarka could be supposed to establish concomitance, that would itself be contradiction (vyāghāta). Udayana had said that, if even when no doubt is present you suppose that doubt might arise in the future, that can only be due to inference, so inference is valid. No doubts need be entertained regarding the concomitance underlying tarka, as that would lead to the contradiction of our own actions; for we cannot say that we believe fire to be the cause of smoke and still doubt it. Śrīharṣa had replied to this by saying that, where there is experience of failure of coexistence, that itself makes the supposition of concomitance doubtful; when there is no experience of failure of coexistence, there is no end of indefinite doubts lurking about; for these unknown doubts are only put an end to when a specific failure of coexistence is noticed; so under no circumstances can doubts be dispelled by tarka[13]. The main point of the dispute consists in this, that, while Śrīharṣa is afraid to trust tarka because of the supposed doubts, Udayana thinks that, if we are so pessimistic, then we should have to stop all our actions. None of them, however, discusses the middle course of probability, which may lead us to action and may yet not be considered as proved valid inference. Vardhamāna, however, in commenting on the above verse of Udayana, refers to Gaṅgeśa as holding that tarka does not lead to the formation of the notion of concomitance[14].

Vyāsa-tīrtha, however, in his Tarka-tāṇḍava, urges that tarka is not an indispensable condition of the notion of concomitance; by faith in trusty persons, or from inherited tendencies, as a result of experiences in past life, or through acquiescence in universally accepted views, we may have a notion of concomitance without going through the process of tarka. He seems, however, to be largely in agreement with the view of tarka as held by Gaṅgeśa according to the above statement of Vardhamāna, in holding that tarka does not lead directly to the establishment of concomitance. For he says that tarka does not directly lead us to the establishment of concomitance, since concomitance is directly grasped by a wide experience (bhūyo-darśana) of coexistence, qualified by a knowledge of absence of failure of coexistence[15]. Vācaspati also holds more or less the same view when he says that it is the sense-organ, aided by the memory of wide experience, that grasps this natural relation of concomitance[16]. Vyāsa-tīrtha says that the determination of absence of vitiating conditions (upādhi), which is a function of tarka, becomes necessary only in some kinds of inference; it is not always awaited. If it were always necessary, then tarka being required for all notions of concomitance and concomitance being the basis of tarka, there would be a vicious infinite[17]. If failures of coexistence are not known, then from cases of coexistence the self may immediately form the notion of concomitance[18].

What is necessary therefore is to dispel the doubts as to failure of coexistence (vyabhicāra-śaṅkā-nivṛtti-dvāra). But such doubts come only occasionally (kvacitkaiva) and not always; and such occasional doubts require to be dispelled by only an occasional recourse to tarka. It cannot be argued that the possibility of doubts may remain in all cases and hence in all cases there is necessity for the exercise of the tarka; for it may well be asked, do such doubts arise of themselves in our minds or are they raised by others? On the first supposition one may have doubts even as to the perception of one’s hands and feet, or one might even have doubts in regard to one’s doubts, which would render even the doubts invalid. If it is held that doubts arise only when other possible alternatives are suggested, then it has to be agreed that there will be many cases where no such alternatives would be suggested or the probability of one of them might be so strongly suggested that there will be no occasion for doubts. So it must be admitted that in many cases we have a natural belief in certain orders of coexistence, where no doubts arise of themselves (sva-rasika-viśvāsasyāvaśyakatvān na sarvata śaṅkā)[19]; no one is seen going through a never-ending series of doubts all his life (na cāvirala-lagna-śaṅkā-dhārā anubhūyate). On the second supposition also, no one can suggest that doubts may always arise: in the relation of smoke and fire one cannot suggest that there may still be some other entity, different from fire, which causes smoke; for, if this were a sensible entity, it would have been perceived, and, if it were non-sensible, there would be no proof at all that a non-sensible entity existed or could exist. For, if Śrīharṣa should be so doubtful of all things, it might be suggested that in all the proofs in favour of monism (advaita) there may be a thousand faults and in the arguments of the dualists there may be a thousand good points, and so in consequence of these doubts you could not come to any conclusion establishing your doctrine of monism[20].

If a belief in a concomitance arises, the mere indefinite possibility of doubt does not shake one off his natural conviction of the concomitance as valid[21]. If you yourself would eat whenever you had hunger to appease, you cannot say that you have still doubts that eating may not after all be the cause of appeasing of hunger. Moreover, what is gained by urging that possibility of doubts always remains? Is it meant to destroy the validity of all inference or of all notions of concomitance? No one who wishes to admit the usefulness of inference would think of destroying the means—the notion of concomitance—by which it is established. If concomitance is not established, the Vedāntist will find that it is impossible to understand the meanings of those Vedic monistic words by which he wishes to establish monism. Again, if inference is to be valid, that can only be established by inference and not by perception. Without inference the Vedāntist could neither establish anything nor refute any assertions made by his opponents, contradicting his own doctrines. It seems therefore that Śrīharṣa would carry out an inference as if there were no fear of the supposed doubts and yet, merely for the sake of saying it, say that there is a possibility of the existence of doubts in all inferences[22].

The main points that arise from the above discussion are that, while Śrīharṣa would argue that tarka cannot remove the doubts threatening the validity of any notion of concomitance and while the Naiyāyikas would hold that tarka, on account of its function of removing doubts from notions of concomitance, is a necessary factor of all inferential process, Vyāsa-tīrtha argues that, though the power of tarka in removing doubts is admitted, yet, since in many of our inferences no doubts requiring the help of tarka would arise, it is not true that tarka is a necessary factor in all inferences[23]. From what has been said above it will appear that there is some subtle difference of opinion in the Nyāya school regarding the real function of tarka. But the general tendency seems to be to restrict the function of tarka to removing doubts and thereby paving the way for the formation of the notion of concomitance; but it does not directly produce the notion of concomitance (na tu vyāpti-grāhaka) nor does it verify particular inductions by the application of general principles of uniformity of nature[24].

So far Vyāsa-tīrtha has been using the word tarka in the accepted Nyāya sense and, using it in that sense, he has been showing that the removal of doubts is not indispensable for the formation of the notion of concomitance. Tarka consists according to him, however, in the necessary awakening of the knowledge of absence of the reason owing to absence of the consequence; taken from this point of view, it becomes identical with inference (anumāna). Jaya-tīrtha also says in his Pramāṇa-paddhati that tarka means the necessary assumption of something else (consequence), when a particular character or entity (reason) is perceived or taken for granted (kasyacid dharmasyāṅgīkare’rthāntarasyāpādanaṃ tarkaḥ)[25]. Granted that there is no fire in the hill, it must necessarily be admitted that there is no smoke in it; this is tarka and this is also inference[26]. Tarka is thus the process by which the assumption of one hypothesis naturally forces the conclusion as true. This is therefore a pramāṇa, or valid source of knowledge, and should not be considered as either doubt or false knowledge, as some Nyāya writers did, or, as other Nyāya writers considered it to be, different from both doubt and decision (nirṇaya). Thus according to Vyāsa-tīrtha tarka has a twofold function, one as the dispeller of doubts and a help to other pramāṇas, and the other as inference. The main point that Vyāsa-tīrtha urges against Udayana (who holds the function of tarka to be merely the removal of undesirable assumptions) and against Vardhamāna (who holds that the function of tarka is merely the removal of doubt of the absence of the consequence) is that, if tarka does not take account of the material discrepancy or impossibility of facts involved in the assumption of the absence of the consequence (fire) when the smoke is present, then even the doubts or undesirable assumptions will not be removed; and, if it does take account thereof, then it yields new knowledge, is identical with inference, and is a pramāṇa itself[27]. Tarka may be treated as a negative inference, e.g., “had it been without fire, it would have been without smoke; but it is not so”. Being such a negative inference, it stands as an independent inference, and, as it may also be used to strengthen a positive inference, it may also be considered in that case an additional support to it (pramāṇānām anugrāhaka), just as what is known by perception may again be strengthened by inference[28]. Its function in removing doubts in other cases remains just as it has been shown before; but everywhere the root principle involved in it is necessary supposition rendering other alternatives impossible (anyathānupapatti), which is the principle also in inference[29].

Footnotes and references:


ūhatvaṃ ca mānasatva-vyāpyo jāti-viśeṣaḥtarkayāmi’’ ity anubhava-siddhaḥ.
I, p. 40.

Tarka is used in the sense of ūha by Jayanta also in the Nyāya-mañjarī, p. 586. Jayanta says that its function as ūha consists in weakening the chances of the weak alternative, thereby strengthening the probability of the stronger alternative and so helping the generation of a valid knowledge of the certainty of the latter alternative. The meaning of tarka here must be distinguished from the meaning “inference” (anumāna), which it has in Brahma-sūtra, II. 1. 12 (tarkā-pratiṣṭhānāt...), and also from its use as the science of logic (ānvīkṣikī), one of the fourteen subjects of learning (vidyā-sthāna). Yājñavalkya-smṛti, I. 3; also Nyāya-mañjarī, pp. 3-4. Ūha is with Sāṃkhya a quality of buddhi and with the Mīmāṃsakas it is a process of application of recognized linguistic maxims for the determination of the sense of words or of sentences (yuktyā prayoga-nirūpaṇam ūhaḥ), ibid. p. 588. Here ūha is used practically in the sense of “inference” and is such a pramāṇa. But here in the Nyāya ūha or tarka stands between right knowledge and doubt.

Thus Jayanta says:

tad eṣa mīmāṃsaka-kalpyamāno nohaḥ pramāṇa-vyatirekam eti pramāṇa-sandehadaśāntarālavartī tu tarkaḥ kathito’tra śāstre (p. 590).


Nyāya-sūtra, I. 1. 40 and Vātsyāyana’s Vṛtti on it.


tarko na pramāṇa-saṃgṛhīto na pramāṇāntaram;
pramāṇānām anugrāhakas tattva-jñānāya parikalpyate.
I. 1. 1.


Tarka-bhāṣya, p. 44.   


vyabhicāra-jñānābhāva-saṃpādakatvena tarkasya vyāpti-grahe upayogaḥ.
      Bhavānandi on Dīdhiti, quoted in Nyāya-kośa, footnote, p. 292.


Tarka-dīpikā, p. 88.


tathā ca dhūmo yadi vahni-vyabhicārī syāt vahni-janyo na syāt ity anena vyabhicāra-śaṅkā-nirāse niraṅkuśena vyāpti-jñānena anumitir iti paraṃparayā evāsya upayogaḥ.
Viśvanātha-vṛtti, I. 1. 40.


Each of the first three has three varieties, according as it refers to knowledge (jñapti), production (utpatti) and existence (sthiti).

Thus the threefold example of ātmāśraya would be

  1. etad-ghaṭa-jñānaṃ yady etat-ghaṭa-janyaṃ syāt etad-ghaṭa-bhiimaṃ syāt,
  2. ghato’yam yady etad-ghaṭa-janakaḥ syāt, etad-ghaṭa-bhinnaḥ syāt,
  3. ayaṃ ghaṭo yady etad-ghaṭa-vṛttiḥ syāt, tathātvena upalabhyeta.

Example of anyonyāśraya in jñapti: ayaṃ ghaṭo yady etad-ghaṭa-jñāna-janya-jñāna-viṣayaḥ syāt etad-ghaṭa-bhinnaḥ syāt.

Example of cakraka in utpatti: ghaṭoyaṃ yady etad-ghaṭa-janya-janya-janyaḥ syāt tadā etad-ghata-janya-janya-bhinnaṃ syāt.

Mādhava, in his Sarva-darśana-saṃgraha, speaking of older Nyāya tradition, adds seven others,

  • vyāghāta (contradiction),
  • pratibandhi-kalpanā (irrelevant thesis),
  • lāghava (minimum postulation),
  • gaurava (too much postulation),
  • utsarga (general rule),
  • apavāda (exception),
  • vaijātya (class-difference).

But Viśvanātha, whose list of these varies somewhat from the above, as he drops vyāghāta and has prathamopasthitatva, and vinigamana-viraha for pratibandhi-kalpanā, apavāda and vaijātya, holds that these are not properly tarka, but are so called only because they help as accessories to pramāṇas

(pramāṇa-sahakāritva-rūpa-sādharmyāt tathā vyavahāraḥ).
I. 1. 40.


Gaṅgeśa on tarka and Mathurānātha’s commentary thereon.
      Tattva-cintāmaṇi, Part 11, pp. 219-28.


Ibid. p. 220; see also Kāmākhyānātha’s note, also p. 228.


tad eva hy āśaṅkyate yasminn āśaṅkyamāne sva-kriyā-vyāghāto na bhavatīti; na hi saṃbhavati svayaṃ vahny-ādikaṃ dhumādi-kāryyārthaṃ niyamata upādatte tat-kāraṇaṃ tan netyāśaṅkyate ca.
p. 232.


tad-adarśanasya āpātato hetv-antara-prayojyāvāntara-jāty-adarśanena ayo-gyatayā avikalpyatvād apy upapatteḥ; yadā tu hetv-antara-prayojyo dhūmasya viśeṣo draksyate tadāsau vikalpiṣyate iti saṃbhāvanāyā durnivāratvāt.
Śrīharṣa’s Khaṇḍana-kkaṇḍa-khādya, p. 680.


Udayana’s verse ran as follows:

śaṅkā ced anumāsty eva na cec chaṅkā tatastarām
vyāghātāvadhir āśaṅkā tarkaḥ śaṅkāvadhir mataḥ.
III. 7.

Śrīharṣa gave his reply to this by slightly changing Udayana’s words as follows:

vyāghāto yadi śaṅkāsti na cec chaṅkā tatastarām
vyāghātāvadhir āśaṅkā tarkaḥ śaṅkāvadhiḥ kutaḥ.
p. 693.

Gaṅgeśa suggests that the word vyāghāta in Śrīharṣa means failure of coexistence (sahānavasthāna-niyama), while in Udayana it means contradiction of one’s own actions (sva-kriyā-vyāghātaḥ). But, as Vyāsa-tīrtha shows, the word may be taken in the latter sense even in Śrīharṣa.
(MS., p. 25).


atrāsmatpitṛcaranāḥ, tarko na vyāpti-grāhakaḥ kintu
vyabhicāra-jñānābhāva-saharkṛtaṃ sahacāra-darśanam.
III, p. 26.


api ca tarko na sākṣād vyāpti-grāhakaḥ bhūyo-darśana-vyabhicārādarśana-sahakṛta-pratyakṣeṇaiva taḍ-grahaṇāt.
(MS., p. 20).


bhūyo-darśana-janita-saṃskāra-sahitam indriyam eva svābhāvika-saṃban-dha-grāḥi.


This has already been pointed out above in dealing with Śrīharṣa’s objections.


adrṣṭe vyabhicāre tu sādhakaṃ tad ati sphuṭaṃ
jñāyate sākṣiṇaivāddhā mānavadho na tad bḥavet.
(MS., p. 21).


Tarka-tāṇḍava, pp. 22-3.


Ibid. p. 24.


na hi grāhya-saṃśaya-mātraṃ niścaya-pratibandhakam; na ca utpannasya vyāpti-niścayasya balavad bādhakam asti yena autsargikaṃ prāmāṇyam apodyeta.
p. 24.


Ibid. pp. 25-31.


It cannot, however, be said that the Nyāya would urge the necessity of tarka in all instances of inference. The older Nyāya writers do not say anything explicitly on the subject; but Viśvanātha, in his Muktāvalī, states that tarka is necessary only in those cases where there are doubts regarding the forming of the notion of concomitance. Where no doubts naturally arise, there is no necessity of tarka (yatra svata eva śaṅkā nāvatarati tatra na tarkāpekṣāpīti). Muktāvalī, 137.

Dinakara, however, in his commentary on the Muktāvalī 137, thinks that there are two kinds of tarka, clearance of doubts and the formation of concomitance (tarkaś ca divividho samśaya-pariśodhako vyāpti-grāhakaś ca). This however is directly opposed to the view of Vardhamāna cited above.


The wording of Dr Seal’s brief references to the subject of tarka in A History of Hindu Chemistry by Dr P. C. Ray (p. 264) is inexact. He says there:

Tarka or Uha, then, is the verification and vindication of particular inductions by the application of the general principles of Uniformity of Nature and of Causality, principles which are themselves based on repeated observation (bhūyo-darśana) and the ascertainment of innumerable particular inductions of Uniformity or Causality (bhūyo-darśana-janita-saṃskāra-sahitam indriyam eva svābhāvika-saṃbandha-grāhi Vācaspati).”

Thus tarka also helps in dispelling doubt (saṅdeha).

On its function in clearing the way to the formation of the notion of concomitance:

mārga-sādhana-dvāreṇa tarkasya tattva-jñānārthatvam iha vivakṣitam.
p. 586.

Mathurānātha also points out that the function of tarka is to supply such grounds that doubts may not arise, but it is not vyāpti-grāhaka (tarkaḥ śaṅkānutpattau prayojakaḥ...). Mathurānātha on Tattva-cintāmaṇi, Part II, p. 240.


Pramāṇa-paddhati, p. 36a. manmate tu aṅgīkṛtena sādhyābhāvena saha anaṅgīkṛtasya sādhanābhāvasya vyāpakatva-pramā vā sādhyābhāvāṅgīkāra-nimittaka-sādhanābhāvasyāṅgīkartavyatva-pramā vā tarkyate’nena iti vyutpattyā tarkaḥ.
(MS., p. 78).


parvato nirdhūmatvenāṅgīkartavyaḥ niragnikatvena aṅgīkṛtatvād hradavat ity anumānam eva tarkaḥ.
p. 84.


kim ca para-mate tarkasya kiṃ viṣaya-pariśodhane upayogaḥ kiṃ Udayana-rītyā aniṣṭa-prasañjanatvamātreṇa upayogaḥ, kiṃ vā Varddhamānādi-rītyā sādhyābhāva-sandeha-nivarttanena.
p. 92.


sādhanānumānaṃ vinaiva yadi niragnikaḥ syāt tarhi nirdhūmaḥ syāt tathā cāyaṃ nirdhūma iti tarka-rūpānumānenaiva agnisiddheḥ.
p. 90.


sākṣād anyathānupapatti-pramāpaka-tarka-viṣaya-kṛta-virodhasya sattvāt.
p. 89.

Help me keep this site Ad-Free

For over a decade, this site has never bothered you with ads. I want to keep it that way. But I humbly request your help to keep doing what I do best: provide the world with unbiased truth, wisdom and knowledge.

Let's make the world a better place together!

Like what you read? Consider supporting this website: