Śrīharṣa on Knowledge and Justification – 2

Śrīharṣa on Knowledge and Justification – 2

Śrīharṣa on the Scope of Reasoning

With regard to the establishment of categories, a generally agreed-upon maxim is that ‘validity of something can be established [P1] not by a mere proposition but by [P2a] providing definition and [P2b] evidence in its confirmation.’[1] I am analyzing this maxim in two sections, with the second having two clauses. The first one negates the possibility of establishing a category by mere conviction or a simple proposition. In a dialectical practice, a thesis cannot confirm itself, and one cannot present another thesis (P2) in affirmation of the first thesis (P1), as long as P2 is not confirmed. Śrīharṣa is explicit in saying that ‘the establishment of a category is contingent upon its definition.’[2] This maxim instantaneously leads to circularity and absurdity: definition itself needs to be further defined. A definition can either rest on already defined categories (which cannot be the case) or on undefined categories, and in either case, an unacceptable consequence results.

The issue of justification leads to the same circularity. Something unjustified cannot justify something else, and something cannot justify itself, as in that case, justification is irrelevant. Now the issue is, most philosophers are open to questioning the categories but are not open in the same way to question the system of justification. How is a dialectical approach possible, one can argue, if the assumption is that a dialectical approach does not result in establishing the truth? Also a virtue argument ensues: what could possibly be the virtue in demolishing the system that justifies our truth claims? At this point, the argument sounds alarmingly similar to the one we often hear: are the anarchists subject to the same justice system? Unwilling to be labeled a polemical anarchist, Śrīharṣa initiates his discourse with virtue argumentation, first laying out that it is not upon him to justify the system to dialectically engage with his opponents.

Virtue argumentation shifts the focus from epistemology to the argumentation theory itself, and what becomes important is what can and cannot be questioned in a dialectical setting.[3] Śrīharṣa identifies this dialectical process as a ‘conversation’ (kathā) and maintains that this process can be meaningful only with a ‘contract of agreements’ (samayabandha). His focus here is on the agreed-upon rules prior to engaging in a debate. If what is agreed is that everything is arguable, is it virtuous to argue on the very rules that are supposedly agreeable? If the rules are the most basic and shall never be challenged, a conversation with a skeptic is not feasible. On the other hand, if the rules are themselves questionable, how can they grant the unquestionable truth with a total certainty? The Nyāya opponents of Śrīharṣa may ponder why anyone would want to destabilize the ‘system,’ and not want to have a thesis. However, the question stands, if all questions are fair game, why would a question regarding the validity of the system itself be wrong? While the Nyāya school looks for a dialectical closure, Śrīharṣa unravels those conclusions: after all, every conclusion is a thesis on its own, open to a new set of questions. Classical Hindu and Buddhist discourse on the nature and scope of debate, highlighted in texts such as Vādanyāya, outline the limits of this dialectical approach, providing sufficient materials to read virtue argumentation. The dogmatic school of Nyāya has attempted to define the parameters of rules since its earliest texts, such as the Nyāyasūtra of Gotama. The challenge for Śrīharṣa is to establish himself in the dialectic circle without utilizing the basic virtue argumentation.[4] It would be a shame to be one of the best bowlers and be disqualified for the game. He therefore rejects the thesis that the means of justification are intrinsically confirmed (siddha) and therefore are not subject to investigation.

The issue is, why shall the interlocutor accept the hypothesis that the means of justification are not subject to questioning? Śrīharṣa gives four possibilities and rejects all the options.[5] In what follows, the four arguments in sequence demonstrate four types of relationships, generally used in classical Hindu-Buddhist debates. These arguments of concomitance are based on:

  1. Concomitance due to pervasion (vyāpya-vyāpaka-bhāva): If x is a mango tree, x is a tree. This can be introduced in the form of categorical syllogism (all x are y, no x is y, some x are y, some x are not y).
  2. Concomitance due to causal relation (kārya-kāraṇa-bhāva): If x is an effect, it must have a cause. The argument Śrīharṣa provides can be introduced both in the forms of modes ponens (p à q, p –| q) and modes tollens (p or q, ~p –| q).
  3. Concomitance due to the existence of R indicating the existence of S (gamya-gamaka-bhāva): If a person with a shaved head who is wearing ochre garments is an indicator of a monk and if Caitra fits the description, he is a monk.
  4. Reductio ad absurdum (atiprasaṅga): if cause and effect are identical, everything will be eternally nascent; if they are different, anything can cause anything.

By questioning all forms of relationships between a dialectical process and the means of justification, Śrīharṣa intends to demonstrate that there really is no logical necessity for one to be confined within the given parameters of the system of justification as a precondition to engage in a dialogue. As a consequence, Śrīharṣa has not just produced four arguments in rejection of the hypothesis that one should accept the system for entering a dialectic circle, but also has provided four different ways of argumentation. Following are his arguments against the parameters for questioning the system:

  1. Is it because the system of justification is intrinsically inseparable from the dialectical practice (or is it because the system of justification permeates the dialectical practice, just like the entities that we can see are always the entities that we can also touch)?
  2. Or is it because a dialogue is an effect of its cause, resulting in the acceptance of the justification system?
  3.  Or is because it is a common practice to accept the system of justification in a dialectical practice?
  4. Or is it because an over-implication (ati-prasaṅga) is a consequence necessary for an interlocutor to win over a debate without accepting the system of justification?

Śrīharṣa rejects the first objection by demonstrating that the dialectical process is adopted even by those who do not accept the system of justification. For instance, Nāgārjuna in his Madhyamaka dialectics or Jayarāśi in his Tattvopaplava each utilizes the dialectic system without any commitment to the system of justification. Srīharṣa retorts, ‘if those conversations {tasya} are not established, even your effort to refute their statements {tat} cannot be founded on reasoning.’[6] With this premise, it can be argued that, if there is no virtue in arguing without the established norms, what virtue could there possibly be in rejecting those arguments [of Nāgārjuna, for instance]?

After demonstrating how the concomitance based on pervasion is not tenable between the methods of judgment and a dialectical practice, Śrīharṣa argues that the concomitance based on cause and effect relation is also not tenable. Although what Śrīharṣa is actually saying is that a dialectical process is an effect and not a cause of the system of justification, this argument extends to the causal relation between justification and truth. This relates to the realists’ contention that pramāṇa stands for an inextricably essential cause for the origins of veridical knowledge (pramāyāḥ karaṇam). The implications of this ‘causality’ of the means of justification are much wider than that can be addressed here, and the examples can be found in both Indian and Western philosophical discourse. For instance, this issue relates to the broader theory of knowledge, and the classical discourse on prāmāṇya is one of its most disputed topics. At the heart of this problem are the issues of Pyrrhonian skepticism, and examples can be found in the writings of Sextus Empiricus to the contemporary fallibilism of Karl Popper and Charles Peirce.

The system of justification raises further questions and leads to the following consequences:

  1. Circularity: P needs to be a pramāṇa in order to ground S, but only by grounding S, does P become a pramāṇa. In Śrīharṣa’s own terms, ‘you establish the means of justification by it being the cause of the speech act of the interlocutor, and due to the existence of the means of justification, the affirmation of what is being affirmed.’[7] There is also circularity between the dialectical process and the system of justification, as an opponent likes to show, so that the interlocutor moves to affirm the system of justification as a hypothesis.
  2. Hypothesis: P does not need an external means of justification to ground S: this is intrinsically justified. This issue is generally addressed as svataḥ prāmāṇya, or that the justification system does not require extrinsic grounding. As Śrīharṣa puts it, ‘one presenting a hypothesis has to follow the means of justification and reasoning. Also the interlocutor has to demonstrate the limits [of reasoning] such as the contradiction with the thesis or contradictory evidence as part of the dialectical process of knowing the truth.’[8]
  3. Infinite regress: P grounds S and O grounds P ad infinitum.

Eventually, this stream of argument evolves around conviction or belief argument: the realists tend to argue that this is a conviction and therefore does not require further justification. And this relates to the third argument that the system of justification is confirmed due to conviction. The problem, however, is whether this conventional affirmation of the system of justification is tantamount of accepting the system as true in the absolute sense or is a mere convention. If someone asks me how much I owe him, if I supposedly borrowed a million dollars and paid only half of it, and I say, another half a million, that does not mean that I am actually going to pay him that sum. Although this issue does not seem trivial, the classical Svātantrika-Prāsaṅgika discourse on negation that evolved as a consequence of reading the philosophy of Nāgārjuna has much to offer over this issue.

For both Nāgārjuna and Śrīharṣa, the dialectical process is possible in a conventional level, but that does not mean that this conventionality bears any truth in the absolute sense. Nāgārjuna proclaims:

“Buddha’s teaching regarding what exists (dharma) rests on two truths: The truth as such and the truth limited to conventions” (MMK 24.8).


“Just as direct perception is empty [of self-nature], for the reason that all the entities are empty [of self-nature], so also are inference, analogy, and testimony empty [of self-nature], for the [same] reason that all the entities are devoid of self-nature.”[9]

The thesis that linguistic transaction can convey meaning even if there is no corresponding reality, pivotal to understanding the philosophy of Nāgārjuna and Śrīharṣa, takes language outside of representation or picture theory to a theory similar to the ‘language game’ of later Wittgenstein. Having a meaningful conversation, accordingly, does not rely on the ability to map the world as it is, but instead, it rests on the ability to use conventionally defined terms, and a sentence becomes meaningful not in it being the picture of reality but in it following the normative conversational conventions. It is the totality of the states of affairs, both actual and possible, that constitute the ‘world’ in Wittgenstein’s ‘language game.’ The convention of Nāgārjuna or Śrīharṣa can be compared with this ‘possible’ world of Wittgenstein. This, however, is not to identify the soteriological positions of Nāgārjuna and Śrīharṣa. The rejection of conventionality in the absolute level for Nāgārjuna is not to demonstrate that there is something factual, a real world beyond convention, something that really exists in the absolute sense, but that, the convention is to accept the reality in the conventional level while in reality, there really is no substantiality, no reality of any sort, that is possible either intrinsically, or extrinsically, or by both, or by any external reason. Śrīharṣa, on the contrary, accepts something to be the foundation, a penetrating reality of the Brahman, that transcends everyday conventionality. Along these lines, although Śrīharṣa initiates the discourse as a loyal student of Nāgārjuna, he deserts ‘Camp Nāgārjuna’ without compromising the epistemic framework of Nāgārjuna or even his philosophy of language.

Exchange of meaningful words, a precondition for a dialectical process, does not necessarily translate into describing reality. The scope of language is the conceptual world, as Bhartṛhari proclaims: ‘There exists no such a concept that stands in isolation of corresponding to a word, as words articulate all that is cognized, as if cognition is penetrated [by words]’ (Vākyapadīya I.115). It is a fact that the conceptual world is much wider than the external reality. A chain of order can be conceived along these lines, as the semantic world permeates the epistemic process and thus also the system of justification, where the categories occupy only a marginal space.  This makes ‘knowable’ (jñeya) a common property shared by all entities, both phenomenally existing and merely conceptual. Accordingly, an entity here also stands for something that does not exist, with ‘absence’ (abhāva) being one of the categories. All these are nonetheless the entities of our conventional transaction. This is what is meant by prapañca, a mere verbal expression without any corresponding reality. The argument based on conventions actually serves Śrīharṣa in demonstrating that there is no justification for the system of justification itself.

This leads to the final option, that it is absurd for the interlocutor to proclaim victory without accepting the validity of the system of justification. The absurdity embedded with this argument is, if the system of justification gives rise to veridical knowledge where the system does not exist or does not have its own independent existence but nonetheless gives rise to the knowledge that exists, it would be similar to a barren woman giving birth to a child. Śrīharṣa counters this argument by saying that “even we who are neutral regarding the existence or lack thereof of the system of justification follow its procedure.”[10] He goes one step further and counter-argues: the consequence is the same even for you, if you consider the consequence in acquiring the results without a prior confirmation of the system.[11] What is embedded with this argument is that someone who accepts the system of justification is accepting it without it actually producing any knowledge. And this is similar to considering someone a mother, without her actually giving birth to a child. The tricky situation here is, these arguments of Śrīharṣa are not to be interpreted that he is actively rejecting, or proposing a rejection of the system of justification. On the contrary, he is only rejecting the positive confirmation of such a system without any external verification. For Śrīharṣa the system of justification is a useful mechanism that is not just helpful in deconstructing the categories but also the very process of deconstruction.

Śrīharṣa’s arguments are founded on two central premises:

  1. He accepts a provisional existence of the system of justification for a mere conventional use,
  2. He can use the system even without affirming the system.

At this juncture, it is clear that Śrīharṣa is not conceding his central premise even when accepting the conventional reality of the system of justification. His arguments thus force his opponents to uphold the position that ‘even to just initiate a dialectical process, conventional existence [of the system] needs to be approved.’[12] The opponent’s claim comes on the ground of action philosophy: ‘what it means to act is to accomplish something, that is, to bring something to existence that did not exist before.’[13] Accordingly, for an act to be an act, it needs to bring about something, something that was not already there. And what the realist expects to achieve of this argument is, if knowledge is a consequence of the means of justification, then there is a causal relation between these two, and hence, the system exists as a cause: a cause is what precedes its effect (kāryaniyatapūrva-katva).

Śrīharṣa rejects this thesis grounded on a causal relation between the system of justification and truth by arguing that what is required to initiate a dialectical process is only a thesis itself, or a concept that may or may not correspond to reality. It is unreasonable to demand that what exists is affirmed before a dialectical closure. He therefore states: “what is part of the dialectical process {tatra} is the knowledge of existence and not the existence itself.”[14]  This argument relates to the earlier one where the scope of language had been examined. It is not necessary for entities to exist for a linguistic transaction to occur. In particular, if the interlocutor is merely interested in negation, it is not the case that the entities of negation have an antecedent existence. We do negate the rabbit-horns or the round-squares, and say, a barren woman, by definition, cannot bear a child. Just like Nāgārjuna’s use of the convention is not to affirm any metaphysical claims, so also is Śrīharṣa’s, and he does not see the need for a justification system that exists outside of the verbal exchange.

On the other side of the game of polemics, it is not possible for realists to conclude that the knowledge of the system of justification demands the existence of what has been justified, as this would mean that kP à P, and there would be no erroneous cognition, as in all those instances, there would also be the existence of an entity, as they are cognized that way. This leads to the question regarding what is meant by existence, and Śrīharṣa responds to both possible options:

  1. A mere acceptance of the categories suffices to initiate a dialectical process. This is tantamount to saying that a conventional acceptance is sufficient for engaging in discourse. This position is not far-reaching for the realists though, as it is not sufficient to affirm reality.
  2. The existence of the categories free from contradiction is a precondition for a dialectical process. Śrīharṣa gives two options to explain this position and categorically rejects both:
  • The absence of contradiction cannot refer to the two parties and a witness not experiencing contradiction to the thesis only at the moment of the dialectical conflict. Even though the persons debating over something may not see the contradiction with the thesis for the moment, but some other person or the same people in another time may encounter contradiction. This argument resembles ‘The Ten Modes’ of Pyrrhonian skepticism (Mates 1996, 94-110). In essence, this absence of contradiction should be absolute and not relative and impossible to demonstrate. Śrīharṣa further argues, if what is confirmed at the time is only the conventional truth, this will be proving something that is well established (siddhasādhanatā). This argument is noteworthy in its premise that can challenge contemporary responses to skepticism, for instance, the brain in a vat argument (Hilary Putnam in Warfield 1999, 77). Śrīharṣa’s arguments are not to question the apparent, but are meant only to affirm its conventionality. On the other hand, if the claim is that this confirmation at the moment of dialogue establishes the absolute truth, Śrīharṣa argues, this contradicts with the fact that even the entities well-established after a thorough examination can turn out to be false in a later examination.[15] One can derive a basis for doubt on the ground of this argument. Śrīharṣa, however, does not make a positive claim regarding doubt.
  • It is not possible to have the knowledge of the absolute lack of contradiction, and the dialectical process therefore stands on a mere assumption that no contradiction is found at the time of inquiry. Bhartṛhari is aware of this dilemma that ‘even the entities that are logically inferred by some skilled logicians are established otherwise by some other more skilled [people]’ (VP I.34). Although where Bhartṛhari wants to lead this discourse, that there are limits to reasoning and therefore we should accept the testimony, is different from the one acceptable to Nāgārjuna or Śrīharṣa, the problem remains the same that some of the well-established facts turn out to be false. It leaves us in a compromised position that we accept something to be the case as long as contradictory evidence to our thesis is not found. This is what conventionality represents.

Śrīharṣa brings this conversation to a closure with the final statement that the ‘dialectical process starts by accepting the conventional existence of the system of justification, [the categories being examined,] and so on.’[16] In this entire conversation, Śrīharṣa is hardly interested in making any metaphysical claims regarding the being or non-being of entities, or even in affirming or negating the system of justification. His sole concern is to justify his engagement in a dialectical process without a prior acceptance that the system is outside of the scope of justification. His is thus a conversation upon virtue argumentation, and along these lines, he is raising only the transcendental questions that surround the dialectical process. It is not accepting the system as a precondition for a dialogue, Śrīharṣa argues, against the parameters in which the opponents have made virtue argument. On the contrary, it is about not rigging the system or not transgressing the rules of debate that have been agreed upon by both the sides.[17]

[1] lakṣaṇapramāṇābhyāṃ vastusiddhir [na hi pratijñāmātreṇa] | The first part of this maxim is cited in the Mīmāṃsākoṣa (p. 3339). While this text should have a much earlier reference, the one easily available is in Jaiminīyanyāyamālā –Vistara of Sāyaṇa-Mādhava. The exact text there is: mānādhīnā meyasiddhir mānasiddhiś ca lakṣaṇāt | Jaiminīyanyāyamālāvistara I.1.37, commentary thereon.

[2] lakṣaṇādhīnā tāvallakṣavyavasthitiḥ | KhKh 126: 2 (in the Yogindrananda edition)

[3] For some contemporary readings on virtue argumentation, see Annas 2011 and Cohen 2013.

[4] For a study of the classical rules of debate, see Chinchore 1988.

[5] Ram-Prasad (2002, 138-57) has extensively addressed these possibilities within the context of analyzing being and existence.

[6] Tasyaiva vāniṣpattau bhavatas tannirāsaprayāsānupapatteḥ | KhKh 7:4.

[7] kathāyāṃ kathakavāgvyavahāraṃ prati hetutvāt pramāṇādīnāṃ sattvaṃ satvāc cābhyupagamo bhavatā prasādhyaḥ | KhKh 15:2-3.

[8] pramāṇena tarkeṇa ca vyavahartavyaṃ vādinā | prativādināpi kathāṅgatattvajñānaviparyayaliṅga-pratijñāhānyādy anyatamanigrahasthānaṃ tasya darśanīyam | KhKh 15:7-16:1.

[9] yathā hi pratyakṣaṃ pramāṇaṃ śūnyaṃ sarvabhāvānāṃ śūnyatvād evam anumānopamānāgamā api śūnyāḥ sarvabhāvānāṃ śūnyatvāt. Vigrahavyāvartanī 46.22-47.1.

[10] tasyaiva pramāṇādisattvāsattvānusaraṇodāsīnair asmābhir apy avalambanāt |    KhKh 18:9.

[11] Tasya yadi māṃ prati phalātiprasañjakatvaṃ tadā tvāṃ praty api samānaḥ prasaṅgaḥ | KhKh 19:1.

[12] Kathāṃ pravartayatāpi vyavahārasattābhyupagantavyā | KhKh 19:2-3.

[13] kriyā hi niṣpādanā, asataḥ sadrūpatāprāpaṇam iti yāvat | KhKh 19:3-4.

[14] yataḥ sattājñānasya tatrāṅgatvaṃ na tu sattāyāḥ | KhKh 20:5-6.

[15] prātītikasattvasādhane siddhasādhanam, arthasattvasādhane bādhavirodha ity arthaḥ | Vidyāsāgarī on KhKh 21:18-19.

[16] vyāvahārikīṃ pramāṇādisattām ādāya vicārārambha iti | KhKh 22:4.

[17] tasmād yādṛgvyavahāraniyamaḥ kṛtas tanmaryādā anena nollaṅghiteti yad vādivāgvyavahāre ma-dhyasthāvagamaḥ sa vijayate, yasya tu vacasi naivaṃ tasyāvagamas tasya parājayaḥ | KhKh 22:4-6.

The paper was first published in The Journal of Indian Philosophy (2016) 45:313–329 2014;7:6–24 and has been republished with author’s permission.

Featured Image: Unsplash

Disclaimer: The opinions expressed within this article are the personal opinions of the author. IndiaFacts does not assume any responsibility or liability for the accuracy, completeness, suitability, or validity of any information in this article.