Anumana in Indian Philosophy

by Sangita Chakravarty | 2016 | 48,195 words

This page relates ‘Divisions of Anumana (in Samkhya-Yoga Philosophy)’ of the study on the concept of Anumana (inference) in the Vedic schools of Indian Philosophy. Anumana usually represents the most authentic means of valid knowledge. This paper discusses the traditional philosophical systems such as Nyaya, Vaisheshika, Samkhya, Yoga, Mimamsa and Vedanta.

(B). Divisions of Anumāna (in Sāṃkhya-Yoga Philosophy)

In Indian logic, anumāna has been classified into different ways, viz.,

  1. pūrvavat, śeṣavat, and sāmānyatodṛṣṭa,
  2. kevalānvayi, kevalavyatireki and anvayavyatireki,
  3. svārtha and parārtha and
  4. vīta and avīta.

In the Sāṃkhya-Yoga system, the Sāṃkhyakārikā of Iśvarakṛṣṇa divides anumāna into three kinds which according to its commentators refer to: pūrvavat, śeṣavat and sāmānyatodṛṣṭa.[1]

The Yuktidīpikā and Māṭharavṛtti [2] imply the division into anumāna for its own self (svārtha) and that for others (parārtha) in their discussion of anumāna. The Yuktidīpikā[3] and Vācaspati Miśra[4] discuss the division of anumāna into vīta and avīta.

According to Aniruddha, in addition to the above three kinds of anumāna, i.e. , pūrvavat, śeṣavat and sāmānyatodṛṣṭa there are another three kinds of anumāna, viz., kevalānvayi, kevalavyatireki and anvayavyātireki raising the number of kinds of anumāna into six.

(1) Pūrvavat, Śeṣavat and Sāmānyatodṛṣṭa:

The Sāṃkhyakārikā takes the division of anumāna into three as well established which is offered by Gautama in his Nyāyasūtra. Vātsyāyana in his Nyāyabhāṣya offers two alternate explanations of the nature of pūrvavat anumāna. With regard to this classification of anumāna, the Sāṃkhya adopts the Nyāya view.

A pūrvavat anumāna means an effect that is inferred from its cause. A pūrvavat anumāna is that inference which infers an object belonging to the class of objects perceived,e.g.,the inference of fire by means of smoke, for, objects of the class of fire have been perceived before in the kitchen room and other places. A śeṣavat anumāna means the condition in which the cause is inferred from its effect. Again, a sāmānytodṛṣṭa anumāna is illustrated as the perception of something at some other place caused by movement, e.g. the sun is perceived at different places. Therefore, it is inferred that there is movement of the sun, though it is imperceptible.

It has been already stated that the pūrvavat anumāna means either the anumāna in which the effect is inferred from its cause. According to Yuktidīpikā, the term pūrva means cause and pūrvavat means that which has cause as a probans.[5] It means the anumāna in case of which after observing the cause one comes to know the future effect. For example, one infers future rains after observing rise of cloud in the sky.[6] The Yuktidīpikā, however realizes the difficulty involved in the above example. The valid probans by its very nature should necessarily lead to the probandum and failing it the probans ceases to be a probans. But, the above example lacks in the above condition and hence ceases to be a proper example. According to the author of the Yuktidīpikā, the rise of cloud in the sky is not necessarily the cause of rains. We cannot establish invariable relation between rise of clouds and rains, because there is the possibility of obstruction by wind and the rest.[7] The author of the Yuktidīpikā opines that in that case the definition of pūrvavat anumāna should be understood as that–through which observing the causal power seized amongst the assisting powers free from obstructing elements, one comes to know future rise of effect just as after observing the clay possessed by the potter who is active and having the instruments like the iron rod and the rest, one comes to know the future manifestation of a pot.

In the view of Māṭhara and Vijñābhikṣu, pūrvavat anumāna is based upon past experience. They explain pūrvavat as it was observed earlier.[8] For example, one infers rains after observing rise of clouds in the sky. The Jayamaṅgalā, the Sāṃkhyacandrikā and Gauḍapāda follow the Yuktidīpikā. They explain pūrvavat as the anumāna which has cause as the probans. Here, the term vat can be explained in the sense of like or similar to or possessed of. Māṭhara takes it in the former sense. On the other hand, the other commentators take it in the latter sense.

A śeṣavat anumāna is various as understood in three ways, viz.,(i) from effect to cause,(ii) from one part to the rest and (iii) through elimination. According to the Yuktidīpikā, a śeṣavat anumāna is defined as that in which after observing the accomplishment of effect one comes to know the prior existence of its cause. As for example, one comes to know the meeting of the couple after seeing a boy. The Yuktidīpikā however feels that such a reasoning is also not faultless. There is no invariable concomitance between meeting of the couple and birth of a boy. The birth of a boy does not necessarily lead to the knowledge of meeting of the couple. The birth of Droṇa, etc. is heard to be without the meeting of the couple.[9] For this reason the probans is non-conclusive and that is why the example is rejected by the Yuktidīpikā. It records another example as after seeing the sky red, one comes to know the rise of the moon or the sun. This is also not a faultless example. Therefore, the Yuktidīpikā gives a faultless example. As after seeing the leaf one comes to know the root of water lily or after seeing the sprout one comes to know the seed.[10]

The second interpretation of śeṣavat anumāna mentioned above is given by Māṭhara[11] and Gauḍapāda.[12] It is exemplified by them as after finding a drop of water from the sea to be saltish, one infers that the rest of water is also saltish. Vijñānabhikṣu gives the third of the above mentioned interpretations. Vācaspati Miśra followed Vijñānabhikṣu and he gives the third interpretation. According to him, śeṣavat anumāna is the knowledge with reference to the residual after eliminating the undesirably involved objects when there remains no undesirable involvement of something else.[13] Śeṣavat means that which has śeṣa or an object not known before as its subject matter. In other words, it is the inference (sādhya) which does not belong to the class of any known object. For example, the inference of the difference of earth from all other things by means of earthiness.

In the Sāṃkhyasūtra, it is stated that a sāmānyatodṛṣṭa anumāna is the proof of both Puruṣa and Prakṛti. [14] A sāmānyatodṛṣṭa is that inference which is neither pūrvavat nor śeṣavat. It is from where the apprehension of the vyāpti or pervasion, by generalization (sāmānyataḥ) from the cases of object belonging to perceptible classes, etc., an object of a different class, i.e., an imperceptible object etc., is established by the force of the mark of inference being a property of the subject of the inference. The sāmānyatodṛṣṭa anumāna is understood in two ways-(a) based upon analogy and (b) inferring a characteristic in other cases after observing it in one case. Some of the commentators of the Sāṃkhyakārikā give both of these interpretations. But some commentators give one of them. In Māṭharavṛtti, Māṭhara gives the second interpretation. He exemplifies it as–observing the mango tree having flower, one infers the flowers on the other trees as well.[15] Gauḍapāda gives both of the above interpretations in his bhāṣya.

For the first interpretation, he gives the example as follows-

“The moon and stars have movement because they change the place. Whatever changes the place has movement just as Caitra. The moon and stars change the place. Therefore, they have movement”.

[16] The Jayamaṅgalā also gives the same example. For the latter interpretation, Gauḍapāda gives the same example given by Māṭhara. According to Sāṃkhyacandrikā, in case of sāmānyatodṛṣṭa anumāna, we have some factor other than causal relation as leading to the inferential knowledge. The Yuktidīpikā discusses the sāmānyatodṛṣṭa type of anumāna elaborately. According to it, sāmānyatodṛṣṭa is defined as that after observing the invariable concomitance of the two objects one comes to know the invariable association of the objects of the same group at some other place at some other time.[17] For example, some times after observing the relation of smoke and fire, one comes to know at other time the existence of some other fire through some other smoke.[18] The Yuktidīpikā opines that the above factor is common to the other kinds of anumāna also. Therefore, it cannot serve as a distinguishing feature of anumāna based on general observation. That is why the Yuktidīpikā gives another explanation of anumāna based on general observation. After observing the invariable association of some, observing later on one characteristic out of those, there arises the knowledge of some other unobserved characteristic in some dissimilar object. As for example, on observing Devadattas attaining to some different place through movement, the movement is inferred in case of the invisible planets through their attaining to some different place. Similarly, on observing that the length in case of castle etc. is caused by growth, the growth is inferred in case of the medicinal herbs and trees by observing their length. The Yuktidīpikā is aware of the difficulty involved in this explanation also. It can very well be a case of śeṣavat type of anumāna because the movement and growth are the effects of attaining to some other place and length respectively through which they are inferred. The Yuktidīpikā alleviates the above difficulty on the ground that in case of śeṣavat anumāna there is necessity of the knowledge of cause from effect. But this is not a condition for an anumāna based on general observation.

(2) Kevalānvayi, Kevalavyatireki and Anvayavyatireki:

In the Sāṃkhya system, Aniruddha mentions these three kinds of anumāna, viz., kevalānvayi, kevalavyatireki and anvayavyatireki. [19] But he does not discuss them elaborately. Actually, these divisions are introduced by Uddyotakara on the basis of the nature of vyāpti. In the kevalānvayi anumāna, vyāpti is affirmative only and there is no possibility of counter example. In the kevalavyatireki anumāna, vyāpti can be stated in negative forms and there is no possibility of any homogenous example. Again, in the anvayavyatireki anumāna, vyāpti is stated in both positive and negative forms. The kevalānvayi anumāna is exemplified as that sound is non eternal, because it is produced’. The kevalavyatireki anumāna is exemplified as–‘the living body is not devoid of a soul, because then it would be devoid of life’. The third, i.e., anvayavyatireki anumāna is exemplified as the inference of fire through smoke. Here, both the homogeneous and the counter examples are possible.

(3) Svārthānumāna and Parārthānumāna:

In the Sāṃkhya system, the Sāṃkhyakārikā does not refer to such a division of these kinds of anumāna. The Yuktidīpikā and Māṭharavṛtti seem to imply such a division in their discussion of anumāna. In the view of Yuktidīpikā, the five components of anumāna, viz., proposition (pratijñā) reason (hetu) exemplification (udāharaṇa). application (upanaya) and conclusion (nigamana) are meant for making someone else known. This means by implication that these are not necessary for knowing the things oneself. The Yuktidīpikā is not very strict for the use of these components in making the others known. These components are to be used in accordance with the necessity of the party enquiring about the object. According to Māṭhara, one establishes the object for knowledge of others through the use of five components of anumāna. This is called parārthānumāna. Māṭhara considers three components of an anumāna, viz., pakṣa, hetu and dṛṣṭānta. Here, pakṣa is called pratijñā and dṛṣṭānta is called nidarśana.

For example,

  1. Vahnimānayam pradeśaḥ (pakṣa).
  2. Dhūmavatvāt (hetu).
  3. Yathā mahānasam (dṛṣṭānta or nidarśana).[20]

Though Īśvarakṛṣṇa does not clearly state about the components of anumāna, yet his kārikāvalī gives an idea about the pañcāvayava anumāna.

For example,

  1. puruṣo’sti (pratijñā),
  2. saṃghātaparārthatvāt (hetu),
  3. natavat vyavatisthate liṅgam (dṛṣṭānta),
  4. kṣīrasya yathā pravṛttirajñānasya tathā pravṛttiḥ pradhānasya (upanaya),
  5. tasmāt tṛvidham karaṇam dvāri (nigamana).[21]

In the system of Sāṃkhya, Vijñānabhikṣu[22] and Aniruddha[23] record the five components of anumāna in the terminology current in Nyāya system and accepted in its essentials by other systematists too.

Some scholars arrive at an opinion from such remarks that for the svārthānumāna one uses the three components of anumāna. According to Māṭhara, both the sets of components are to be used in formal anumāna for making the others known. There is no need of formal use of these components in anumāna for knowing the thing for one’s own self (svārthānumāna).

(4) Vīta and Avīta Anumāna:

Following the Sāṃkhya tradition, Vācaspati Miśra classifies the anumāna into two kinds, viz., vīta and avīta. The division into vīta and avīta is attached much importance in the Sāṃkhya texts especially in the Yuktidīpikā which describes the nature of these varieties in detail. The vīta anumāna means that which functions through an affirmation.[24] Again, the avīta anumāna means that which functions through negation.[25] We do not find this division of anumāna discussed in some other texts of Sāṃkhya-Yoga. The division has found an important place in the system of Nyāya also. However, the early texts of Nyāya system like the sūtras of Gautama and the bhāṣya of Vātsyāyana do not discuss it explicitly though Uddyotakara makes an unsuccessful attempt to trace the reference to them in the sūtras of Gautama. Uddyotakara is the first logician in the Nyāya system who discusses the division into vīta and avīta. The vīta anumāna gives rise to the knowledge of an object when employed in its own essential form while the avīta anumāna does so through refuting the other’s stand.[26] It has been already stated above that Vācaspati Miśra divides anumāna into vīta and avīta. According to him, vīta anumāna means which is present in various ways i.e., besides its presence in pakṣa it is present in sapakṣa and is absent in other dissimilar cases i.e., vipakṣa. On the otherhand, the avīta anumāna is different from it. It means that the avīta is not found in sapakṣa. Of these two, avīta is called śeṣavat anumāna. A śeṣavat anumāna is that in which some of the likely properties of an object are denied and eliminated, the likelihood of their belonging to some others being also denied. We have cognition of that which remains. This kind of anumāna is found in establishing śabda as a guṇa. Through certain grounds śabda can be shown as distinct from other objects. When such distinctive grounds are eliminated, there remains a ground by which śabda can be shown as a guṇa. Since we proceed with the anumāna on a ground that remains, after elimination of other grounds, this inferential process is known as śeṣavat.

Vācaspati Miśra further gives the example of avīta anumāna as–

“Cloth is not different from threads, because it is a quality of them, whatever differs from something can not be a quality of that, just as the cow cannot be a quality of the horse, this is however not a case with the cloth, therefore, cloth is not different from threads.[27]

The vīta anumāna is of two kinds, viz., the pūrvavat anumāna and the sāmānyatodṛṣṭa anumāna. Of these, the pūrvavat anumāna has for its object that universal of which a specific individual has been perceived. The term ‘pūrvavat’ means well known. It is that inferential cognition of which such a universal in the object is called pūrvavat. For example–when from the presence of smoke we infer the presence of fire in general in the hill, this in general is one, of which a specific individual in the shape of a particular fire has been previously perceived in the kitchen[28]

The second form of vīta anumāna is sāmānyatodṛṣṭa which is the general cognition having for its object a general instance of which a specific instance has not been perceived.[29] As for example, when we have an inferential cognition of the sense organ, how could we know about the existence of the eye? For every action to take place there should be an instrument. Seeing is an action. Therefore, we have to infer the eye as a means to the action of seeing. This is sāmānyatodṛṣṭa anumāna. In the term, sāmānyatodṛṣṭa, dṛṣṭa stands for darśana, cognition and sāmānyataḥ stands for sāmānyasya of the universal, the affix ‘tasil’ being capable of signifying the sense of all case endings. Thus, the term sāmānyatodṛṣṭa anumāna stands for the anumāna, inferential cognition of that particular ‘universal’ of which a specific individual has not been perceived.

The Yuktidīpikā states that anumāna is of two kinds, viz., vīta and avīta. Both of these kinds are meant for explaining something for others. In the view of Yuktidīpikā, the vīta is that when the probans is applied in its very form. Again, the avīta is through elimination when other possibilities are eliminated. The further explanation offered by the Yuktidīpikā brings out the difference between the vīta and avīta more clearly. It states that the essential form of probans can be of two types, viz., generic and particular. In the case of vīta anumāna the probans is employed in its essential form without any reference to the exclusion of other’s stand. In the case of avīta anumāna, the probans does not prove the probandum directly but wards off other possible alternatives. It is employed to prove something through elimination. According to Yuktidīpikā, if the vīta yields particular results without the need of eliminating other possibilities, the vīta alone is sufficient. If, however, there is the need of eliminating other possibilities, the avīta is resorted to for the purpose. That also runs in the way of an argument as, ‘if it is not admitted so, there arises the undesirable contingency of admitting some other undesirably involved object’. If the elimination of undesirably involved object is not intended, there is no need of avīta. For example, past rains are inferred through flood in river as the cause of the latter. But it involves an undesirable contingency of possibility of considering melting of snow, break of bridge and sports of elephants as the cause of flood. The possibility of these are warded off through avīta as these are negated through the probans like space, time etc. For example, through the place as there is no Himalaya in south. Through time is illustrated as it is the rainy season which leads to admit that it should be rainy water. The Yuktidīpikā seems to hold that the avīta is meant for confirming the vīta but not an independent type of probans. Thus the existence of the cause of universe is inferred through vīta type of anumāna, but it is established through avīta that it can be pradhāna only. The Yuktidīpikā explicitly states that probans is of two types, viz., vīta and avīta. The former is subdivided into five.[30] The vīta is again said to be containing ten components, viz., inquisitiveness, doubt, purpose, conjecturing, to throw aside the doubt, proposition, probans, example, application and conclusion. It is, however, not clear as to how many kinds and components an avīta would have.

The illustrations of application of avīta, however, are found in the context of śeṣavat type of anumāna. Hence, it should not be misunderstood that the probans of anumāna based on analogy only are divided into vīta and avīta. This makes it clear that avīta is only an assisting factor for anumāna. According to Yuktidīpikā, the vīta anumāna should be used first, because otherwise the definition of avīta will be contradicted. The nature of avīta is to assist establishing of a particular thing through elimination of other possible alternatives. If the elimination is warded first and then the establishment of the thesis, it would go against the purpose of avīta.

The division of anumāna into seven kinds is also attributed to the Sāṃkhyas. But Vācaspati Miśra attacks the Sāṃkhya view that espouses seven kinds anumāna. [31] Vācaspati Miśra states that the Sāṃkhyas divide anumāna into seven kinds on the basis of sevenfold inferential marks in the relation in vyāpti.

These seven kinds of relation are as follows:

  1. measurement,
  2. cause,
  3. contact,
  4. opposition,
  5. association,
  6. master and servant, and
  7. killer and the killed.[32]

Vācaspati Miśra states against the view of the Sāṃkhya . He opines that it is illogical to postulate these relations in vyāpti. The theory of Sāṃkhya does not account for the temporal factor as for instance, in the case of opposition. The rain is said to be opposed to the contact of cloud with wind. The past rain however is not opposed to the future contact of cloud with wind. On the contrary, the past rain is favourable to the future contact of these two. The future rain is also not opposed to the past contact of cloud with wind. Hence, the opposition cannot serve as an inferential mark or the relation in vyāpti.

In the available texts of Sāṃkhya , the seven inferential marks and their relations in vyāpti are not found mentioned. The Jayamaṅgalā, however, refers to seven kinds of relation in vyāpti. But, these are not same as recorded by Vācaspati Miśra. The relations like those of master and servant, association, opposition and cause and effect are common to both. Instead of measurement, contact and killer and killed, the Jayamaṅgalā mentions the modification and its source, pot and its possessor, and the object and the being for which it is meant. The Jayamaṅgalā and Vācaspati Miśra have recorded these seven kinds of relation in vyāpti from some ancient text which have been lost to us. Vācaspati Miśra enumerates these varieties of relation in vyāpti but does not illustrate them from which it is possible to deduce a few possibilities. Firstly, forms of vyāpti are those which cannot be favourably applied to the basic Sāṃkhya assumption. As for instance, the prakṛtivikāra or the nimitta-naimittika etc.

Footnotes and references:


trividhamanumānamākhyātam. Sāṃkhyakārikā, k. 5


Māṭharavṛtti, 5


tatra proyogamātrabhedāt dvaividhyam vītaḥ avītaḥ iti. YD on Sāṃkhyakārikā, k. 5


tābadvividhaṃ vītamavītañca. Sāṃkhyatattvakaumudī on Sāṃkhyakārikā, k. 5


tatra pūrvavat yadā kāraṇamavyuditam dṛṣtvā bhaviṣyattvam kāryasya pratipadyate. YD,on Sāṃkhyakārikā,5


tad yathā meghodaye bhaviṣyattvam bṛṣṭeḥ. Ibid.






na hi dayasamāpatipūrvaka eva prāṇabhṛtaṃ prādurbhāvaḥ,droṇādīnāmanyathotpattiviśeṣaśravaṇāt, YD on Sāṃkhyakārikā, k. 5


parṇam dṛṣṭvā śālukam pratipadyate aṅkuram vā dṛṣṭvā vījamiti tadā śeṣavat. Ibid.


samudrodakavindum prāsya śeṣasya lavaṇabhāva anumīyate iti śeṣavat. Māṭharavṛtti on Sāṃkhyakārikā, k. 5


samudrādekam jalavalam lavaṇamāsādya śeṣasyāpyasti. Gauḍapāda Bhāṣya on Sāṃkhyakārikā, k. 5


śiṣyate pariśiṣyate iti śeṣaḥ sa eva viṣayatayā yasyāstyanumānajñānasya tat seṣavat. Sāṃkhyatattvakaumudī,5


sāmānyatodṛṣṭāt ubhayāsiddhiḥ. Sāṃkhya Pravacanasūtra,1. 103


puṣpitāmradarśanāt anyatra puṣpita āmarā iti, Māṭharavṛtti,5


deśāntarāddeśāntaram dṛṣṭam gatimaccandratārakam, caitra, Gauḍapāda Bhāṣya,5


sāmānyatodṛṣṭam yatra arthayorvyabhicāramupalabhya deśāntare kālāntare ca tajjātīyayoravyabhicāram pratipadyate. YD,5




anenānvayi, vyatireki, anvayavyatireki, pūrvavat, śeṣavat, sāmānyatodṛṣṭañca saṃgṛhitam. Sāṃkhyasūtravṛtti, 1. 100


trisādhanam trayavayava pañcāvayavamityapare. pakṣahetudṛṣṭānta iti trayavayavam. Māṭharavṛtti,5


Sāṃkhyakārikā,17, 42, 35


Sāṃkhya Pravacanabhāṣya,5. 27


Sāṃkhyasūtravṛtti,5. 27


anvayamukhena pravartamānam vidhāyakam vītam. Sāṃkhyatattvakaumudī, p. 51


vyatirekamukhena pravartamānam niṣedhakam avītam. Ibid.


Nyāyavārttika, 1. 1. 35


Sāṃkhyatattvakaumudī, 9


yathā dhūmāt vahnitvasāmānyaviśeṣaḥ parvate anumīyate tasya vahnitvasāmānyaviṣeśasya svalakṣṇaṃ vahniviśeṣaḥ dṛṣṭaḥ rasavatyām. Ibid., p. 53


sāmānyatodṛṣṭaṃ adṛṣṭasvalakṣaṇasāmānyaviṣayam. Ibid.


Vide Kumar, Shiv., Sāṃkhya Yoga Epistemology, p. 158


Ibid., p. 146


Ibid., p. 147

Like what you read? Consider supporting this website: