Sabdapramana: Word and Knowledge in Indian Philosophy

Sabdapramana or 'Testimony' is a formidable doctrine within Indian philosophy. A though investigation of this thesis is long overdue. What is sabdapramana (word as knowledge)? What is involved in 'hearing' words? Is the understanding derived through hearing utterances direct or indirect? Does this peculiarly linguistic understanding (sabdabodha) amount to knowledge (prama), or does it depend on certain other conditions for its truth? Further, what short of theories of meaning, understanding, and knowledge would be required to ground a successful sabdabodha as prama, need careful attention. It is sometimes said that Indian thinkers had no particularly interesting theory of understanding.

The present work sets out to address these issues – issues that have engaged traditional and modern thinkers alike. Based on the classic text, Advaita Vadanta-paribhasa of Dharmarajadhavarindra (17th century), the analysis and arguments extend to the views of and criticisms from the Nyaya, Purva Mimamsa and the grammarian / linguistic schools within Indian philosophy, with a treatment of similar concerns in Western philosophy. There is a compelling thesis here that should be taken seriously in any philosophy.

Long discarded as a distinct source of knowledge in Western philosophy, Testimony might be lead to mutual dialogue between philosophy and religion, and pave the way for critical metaphysics.

Dr. Purushottama Billimoria, PhD is a Professor of Philosophy and Religion at Deakin University and Senior Fellow at Melbourne University. He has held fellowships and has lectured in universities in the USA, UK, Europe, Canada, and India. Dr. Bilimoria is a visiting professor at State University of New York, Story Brook and Columbia University. He is an editor-in-chief of Sophia, an international journal for metaphysical theology; and co-editor of the two-volume compendia on Indian Ethics, and the Routledge History of Indian Philosophy. He is currently editor of "Studies of Classical India Series."

Dr. Purusottama Bilimoria's book on sabdapramana is an important one, and so is likely to arouse much controversy. I am pleased to be able to write a Foreword to this book, at a stage in my philosophical thinking when my own interests have been turning towards the thesis of sabdapramana as the basis of Hindu religious and philosophical tradition. Dr. Bilimoria offers many novel interpretations of classical Hindu theories about language, meaning, understanding and knowing. These interpretation draw upon the conceptual resources of contemporary analytic and phenomenological philosophies, without sacrificing the authenticity, that can arise only out of philologically grounded scholarship. He raises many issues, and claims to have resolved some of them. Certainly, he advances the overall discussion, and this is the best one could hope for in writing on a topic to which the best minds of antiquity and modern times have applied themselves.

In this Foreword, I wish to focus on one of the issues which I have raised on earlier occasions, and on which Dr. Bilimoria has several important things to say. The issue is: is sabdabodha eo ipso a linguistic knowing, i.e., sabdaprabha, or does sabdabodha amount to knowing only when certain specifiable conditions are satisfied. It the second alternative be accepted, these additional conditions are satisfied. It the second alternative be accepted, these additional conditions could not be the same as the familiar asatti (contiguity), gogyata (semantic fitness), akanksa (expectancy) and tatparya (intention), for these are, on the theory, conditions of sabdabodha is prama, i.e. true, so that a false (a-prama) sabdabodha would be a contradiction in terms. And yet, in an intuitively clear sense, one does understand a false sentence. What then hearing a false sentence being uttered by a speaker taken to be competent generates in the heart could not be knowledge. It can only be an understanding of the meaning of the sentence. From this, one could proceed to the more general thesis: in all cases – whether the sentence heard is true or false – there is a linguistic understanding. In the case of a true sentence, the understanding becomes a component of knowing. In the case of a false sentence, when the falsity is discovered (and so is the incompetency of the speaker), the knowledge claim is revoked, the underlying understanding stands all by itself.

Now this last picture is clearly rejected by the theories of sabdabodha in classical Indian philosophy (with the possible exception of the so-called grammarians). These theories, in fact, take (i) sabdabodha to be, by definition, true, (ii) and consequently have no interesting notion of linguistic understanding. A notion of understanding requires not only a concept of meaning, but also one that is different from that of reference. It is not necessary that this theory of meaning be as strongly Platonistic as the Fregean theory of sense; what is necessary is that it not be a purely referential theory. The Indian theories under consideration did subscribe to a purely referential theory. Words directly refer to their objects. A sentence is knowing this object, if the sentence is true i.e. if there is such an object. What happens if the sentence is false?

The Indian theories generally held the view that we do not have a sabdabodha of a false sentence. (Rather, what we have in such a case is a peculiar mental (manasa) state.) Is this view counter-intuitive? It would be so, only if we construe sabdabodha as linguistic understanding. It is counter-intuitive to say that we do not understand the meaning of a false sentence. But the view is not counter-intuitive to say that we do not understand the meaning of a false sentence. But the view is not counter-intuitive, if by sabdabodha is meant linguistic knowledge: we surely cannot be said to know, when the sentence heard is false.

Let me then it as uncontroversial (1) that we do understand false sentences, but (2) do not have a knowledge upon hearing and understanding such a sentence. (1) Entails a theory of understanding which the Indian theories do not have, (2) leaves the basic theory of sabdabodha intact: sabdabodha is intrinsically prama (at least, with regard to utpatti, if not which regard to jhapti – i.e., it true, to begin with, even if is not known to be so).

But this would be an unsatisfactory solution, for (1) and (2) cannot be held together. A theory of understanding as is implied in (1) requires a theory of meaning – of meaning that is grasped in understanding a sentence, no matter if that sentence is true or false. But such a theory of meaning – which cannot be purely referential, for then (1) will have to be rejected – will undermine the claim made in (2). Nothing less than a purely referential theory can save (2), and yet a purely referential theory will undermine (1). That (2) needs a purely referential theory is corroborated by the passage, Dr Bilimoria cites from the late K.C. Bhattacharyya: a world qua sentence refers "directly to the thing, expresses the thing, touches it." (K.C. Bhattacharyya, Studies in philosophy, Vol. I, Ch. III, p. 83, cited in this book, p. 326, fn. 150). With a purely referential theory of meaning, understanding a false sentence becomes a 'mock-understanding' (as Arindam Chakraborty would have it using a Fregean locution).

Let me here turn to Arindam Chakraborty's paper 'Understanding Palsehoods: A Note on the Nyaya concept of Yogyata' (The Journal of the Asiatic Society, XXVIII, 1986, pp. 10-20) in which he directly confronts the problem. Chakraborty recognizes that on the usual understanding of 'yogyata' as (semantic) compatibility or fitness, although sentences such as "Idleness is green" and "He is sprinkling with fire" are ruled out from generating sabdabodha, a false sentence such as "The jar is on the floor" is not, (when there actuality is no jar on the floor referred to in the context of the utterance). Wishing to rule out this possibility, Chakraborty then interprets a suggestion from Sidhantamuktavali as implying that a sentence 'a is F' is characterized by fitness (yogyata) only if a is F – which makes fitness collapse with 'truth'. Only true sentences are fit and so can generate sabdabodha. Not only contradictory sentences such as "The fireless hill has fire", but also consistent but semantically false sentences sabdabodha. They may generate "a mental state" consisting of memories of the referents of the component words, but they cannot generate a relational, qualificative cognition, for there is no such relational entity to know.

Now, this ingenuous interpretation of 'yogyata' is indistinguishable from 'truth', in fact, it is the same as that. It smacks of ruling out the possibility of understanding false sentences by stipulating by a fiat, as it were, that truth is a condition of sentence-understanding. Even if this anxiety is in some way alleviated, consider this interpretation of yogyata together with the position that what is required for sabdabodha is not yogyata itself but awareness of yogyata, then a certain determination of truth is being built into the theory – which is in conflict with the thesis of paratah-pramanya. The only way of obviating this is to weaken the sense of 'awareness of yogyata (= truth)', to mean 'absence of firm disbelief in absence of yogyata (i.e. in untruth)', but in that case when one believes a sentence (in fact false) to be true (and regards the speaker to be competent etc.) one must be understanding has to find a place in a theory of meanings of the component expressions but no integrated sentence meaning won't do for reasons: to base this claim on the fact that there is no integrated 'featured individual' of the sort is precisely to beg the issue about what one grasps in understanding (as contrasted with knowing); secondly, one cannot distinguish phenomenologically the two experiences (i.e. understanding a false sentence when believed to be true and understanding a true sentence believed to be true); the matter gets – in the third place – more complicated, to the disadvantage of the proposal under consideration, when one considers true sentences taken to be false; and, finally, the way the distinction is sought to be drawn (namely, that in the case of true sentences there is merely memory of 'unrelated' substantial meanings) is compatible with the Prabhakara theory of a-khyati rather than with the Nyaya theory of anyathakhyati.

Let me now return to the sentence quoted from K.C. Bhattacharyya: 'The word directly refers to the things, express the thing, touches it in a sense." Likewise, "The sentence at once refers to an objective relation." There are two things about K.C. Bhattacharyya's thesis (or interpretation of sabdapramana), which deserve attention. For one thing, in the sentence just preceding the first-quoted one, he writes that even in the case of a word remembering is not understanding the meaning. This contrasts with the account we examined earlier: in the case of a false sentence, we do understand the meanings of words by way of remembering them, but that does not constitute self-understanding. K.C. Bhattacharyya is, in fact, distinguishing between (let us not forget, he is interpreting the Vedanta, and not the Nyaya theory) the name, the concept not only requires the name for he goes on to write: "The same determination of the self gives the name and the concept an identical object-reference." I take "the free concept" to mean "meaning" (as distinguished from objective reference). By "the determination of the self', I take it, following the editor's footnote at that place, to mean "the assertive function of the self". Read this together with what follows in the next paragraph (and this is the second thing to be attended to): "The primordial objective reference of a judgement is a provisional belief, a belief, it may be, with a certain general cautiousness induced by experience: if it is only thought, it is at any rate continuous with knowledge." And we are far removed from the Nyaya theory commented upon earlier. It is a theory of direct reference of sentence, mediated by provisional, uncontradicted belief, and also a theory of direct reference of words, again mediated by a concept that is not distinguished from the name. let us remember that the locutions "direct" and "indirect" are ambiguous: one can ask, how direct? You have a concept of mediation by a sense-content (conceived as transparent, objectified only in reflection), which is compatible with a theory of direct reference. This is what I have been intending these critical remarks to point to.

At the end, I will raise another question about the idea of sabda as a means of knowing, as a pramana, in what sense, one wonders, is linguistic knowing (sabdajnana), a grasping (grahana) of a relational ontological structure (or of a properly qualified individual)? Since one may also perceive the same individual (exactly as qualified as in the case of linguistic knowing), how is the purely linguistic grasping of the object different, qua grasping, from a perceptual grasping of it – especially from the sort of perceptual grasping that, according to both Nyaya and Mimamsa, is short through (anuviddha) with linguisticality, i.e. from the so-called savikalpaka percention? Of course, one may want to explain the cognitive difference by means of the different way each is caused, but what I am at present interested in is, how is the one cognition, as a grasping, different from the other then both are graspings of the same object. Perhaps, the distinction between 'direct' and 'indirect' ('aparoksa' and 'paroksa') will help, but this distinction is hopelessly ambiguous as already pointed out. Perhaps the locution of 'grasping' is misleading, when extended from perception to linguistic knowing. It may be that the Jaina distinction between 'visada' and 'avisada' ('clear' and 'not-clear') will help.