Bhartrhari on linguistic understanding.

Dash, Achyutananda

Linguistic communication is the very foundation of a civilized society (vacam eva prasadena loka-yatra pravarttate). Speech behaviour or linguistic communication is significant to us, not because we communicate more effectively through it, but our social, spiritual, moral and intellectual life is fully dependant upon "speech behaviour"; and thus speech is part of our identity. Therefore, speech has always enjoyed the supreme status at par with the Supreme Being in the Indian intellectual tradition. Though we understand the significance and importance of speech behaviour, sometime we are not very conscious about the mechanism of speech behaviour and linguistic understanding (LU). In other words, hardly, we think on "why and how someone understands exactly the same thing we want to communicate through our linguistic communication". It is not difficult to understand why sometime we fail to communicate our thoughts properly to the listener, and thus there is no elaborate scientific investigation into this issue. On the contrary there is a vast literature explicating the theories and techniques on "linguistic understanding" found in Indian tradition. The ancient Indian Linguists were very serious not only on speech mechanism but also on the theory of "linguistic understanding". Understanding means knowledge or experience. The first-person experience is the very foundation of Indian thinking / philosophy. And first-person experience is gained through contemplation and/or observation. Observation is the most common means of knowledge acquisition, which is common in the cognitive science and material science as well. Indian schools of philosophy however, treat knowledge as "divine" and the knowledge is the only means of knowing the material world. It is said:

Samvid eva hi bhagavati vastupagame nah saranam (1)

i.e., "Goddess Consciousness" alone is our resort for gaining the knowledge of an entity.

There are controversies on the issue whether knowledge is a substance, a quality, or an action. There are also controversies on whether knowledge / consciousness is eternal or transient. However, irrespective of their affinity, to one or the other schools of thought, almost all thinkers in Indian Philosophy accept that the knowledge is objective (savisayaka) and it is a product. The Naiyayikas almost aphorizing a statement say "visayanirupyam hi jnanam" i.e, the knowledge indeed is described by its object; which to some extent holds true to most of the schools of Indian Epistemology. The Vedantins' though think that the knowledge in its ultimate state refers the Brahman, the Supreme Being (2), still they do agree with the view that the knowledge of mundane reality is object-reflexive and it is a product. Its productivity is possibly conceived not in a very empirical sense, but in the sense of theory building exercise in the Nyaya-Vaisesika system, which of course is a matter of debate. This however is not a matter of our concern in present context.

Indian epistemologists are very particular and very sincere about two aspects of cognitive enterprise: (1) the causal mechanism of the knowledge and (2) its cognitive structure that is being presented to the mind through the causal mechanism. Knowledge is a product and it is produced through a well-defined mechanism (karana-vyapara). The knowledge, thereafter, is presented through the mind (the internal sense organ = antah-karana) to the Self in a well-defined structure. The self, according to the Naiyayikas, is the abode of knowledge. The knowledge, according to them, is a quality (guna) of the Self and it remains in the Self, and thus the Atman (3) is defined in terms of the knowledge. That shows the significance of knowledge not only in the system of Nyaya epistemology but also in the Vaisesika system of ontology as well.

The knowledge is caused by the Cognitive Mechanism (CM) known as Pramanas, defined very logically as prama-karanan (valid means of cognition), that is acceptable in almost all the Indian Systems of philosophies without any reservation. No branch of Indian Philosophy denies the validity and dependability of the pramanas. Since everything, starting from the mundane reality to the high intellectual exercise is dependant on the knowledge, the causal mechanism of knowledge (pramana) has gained prominence and remained indispensable in the sastric traditions in India. The pramanas (means of valid cognition or causal mechanism of knowledge) are four only according to the Nyaya school of thought, and they are pratyaksa (perception), anumana (inference), upamana (analogy) and sabda (speech behavior or linguistic communication).

COGNITIVE MECHANISM & COGNITIVE STRUCTURE

For present demonstration we shall call pramana (pramana-karanam = the causal means of knowledge) as Cognitive Mechanism (CM), and the formal presentation of prama (yatharthanubhavah), as the Cognitive Structure (CS). It won't be exaggerated to say that the Sastrakaras, i.e., the exponents of the Indian intellectual tradition, try to present the 'cognitive event' in such a picturesque or photographic manner that one can actually see its contents. This is the unique achievement of cognitive enterprise in the Indian intellectual traditions, which is not seen elsewhere in the world.

The CM has a minimum essential schematic structure that can be presented in the following formulae:

(Ic + F) [right arrow] R

where, Ic stands for "Instrumental cause (karanam)", F stands for "Function or Auxiliary action" (vyapara) and R stands for "Result" or the "end product"(phalam). (5)

This model is applicable to explain the origin of any mental event, especially the cognitive event, that we may call a knowledge episode. The Indian intellectual tradition emphasizes on the fact that the cognitive events vary depending upon their cognitive mechanisms. There are differences of opinions on the classes of different knowledge episodes. We can, for present purpose, accept the fourfold classification of cognition viz. (a) perceptual cognition (Pratyaksa), (b) inferential cognition (Anumiti), (c) analogous cognition (apamiti) and (d) linguistic understanding (sabdabodha) as has been presented in the schools of Nyaya.

Now having said this if we will apply the above mentioned schematic formula in the case of perceptual cognition (pratyaksajnanam), the sensory faculty (indriya) will be the Ic = "instrumental cause" (karanam, i.e., pratyaksa-pramana), the sense-object-contact (indriyartha-sannikarsa) will be the F = "function" (vyapara) and the "end-product" will be the R, i.e., the "result" (pratyaksa-prama).

In the same manner in case of linguistic understanding (sabdabodha, the cognitive event / knowledge episode caused by speech behavior), the minimum causal mechanism required is:

{(Ic = knowledge of word-elements + F = knowledge of their meanings) [right arrow] R = (hearer's) cognitive-episode from sabda}

It is stated by Visvanatha Tarka-pancanana in brief in his Karikavali-Muktarali in the following karika:

Pada jnanam tu karanam dvaram tatra padarthadhih| Sabdabodhah phalam tatra saktidhih sahakarini[parallel] (6)

i.e., the knowledge of the words is the instrumental cause (Ic--in our formulation), the knowledge of the word-meaning (recollection of word-meaning) is the function (F--in our formalization), the linguistic understanding is the result (sabdabodha, i.e. R--in our formalization), and the knowledge of the denotation function (sakti jnana) is the auxiliary cause (sahakari-karana).

The knowledge episode or the cognitive structure (CS) in our formulation being caused so by the operation of the causal mechanism (CM) is distinct from each other due to differences of CMs. For instance, if we take into account a piece of information / cognition like--"This is a car", it may not create any botheration to a common man whether it is a piece of perceptual cognition or a piece of linguistic understanding. On the contrary, if a Naiyayika is asked to analyse the piece of information, he may end tip asking a counter question like what are the CMs? In other words, the Naiyayikas would like to know if the given information i.e. "this is a car" is the R, then what are Ics and Fs? If the Ic is the visual sense organ (eye) and if the F is the eye-object-contact, then it is a piece of perceptual knowledge episode. On the contrary, if the knowledge episode is caused by the sentence, "this is a car", that consists four words viz. "this" "is" "a" and "car', then these words are the Ics and the recollection of their meanings are the Fs causing the linguistic understanding. According to Naiyayikas, the former is different from the latter. There is a standard principle (7) that is stated as karanabhedat karya-bhedah i.e., the distinct nature of the product depends upon the distinction of their causes. Therefore, the CS being caused by a different means of cognition or CM is considered different, even though they refer to one and the same mundane reality. Let us call the mundane reality precisely as "Object Structure" (OS) for the sake of convenience. Now we can say that one and the same object structure i.e. OS may figure differently in different CSs caused by different CMs. In other words, different Object Structures correspond to different Cognitive Structures (8) depending upon their causal mechanism. The Sastrakaras again almost aphorizing another principle say: "visayanirupyam hi jnanam", the cognitive episode is described by the object structure. Bhatrhari, the doyen of Indian Semantics also says:

jneyena na vina jnanam vyavahare 'vatisthate (9)

i.e., No knowledge that does not corresponds to the object structure can be employed in the practical life.

Without delving more into the details on this issue we can very well propose that the object structure (OS = visaya) carry a treat significance in defining or deciding the nature of the cognitive structure (CS) and that is the very foundation of the knowledge episode. There is another issue that is quite relevant in this context and the attention of the readers may he drown to that is that: whether an OS (object structure) corresponding to the respective CS (cognitive stricture) reflects uniformity irrespective of it causal mechanisms (CMs) or not? Moreover, whether different CMs (causal mechanisms) can be employed for the causation of different cognitive episodes where one and the same object structure corresponds to it or not?

These questions lead to a serious epistemic issue known as pramana-samplava vis a vis pramana-vyavastha in the Indian cognitive enterprise, hotly being debated upon by the Buddhists and Naiyayikas. The Naiyayikas accept pramana-samplava theory in contrast to the pramana-vyavastha theory of Buddhists. The pramana-samplava theory proposes that one and the same OS can be represented with the help of more than one CS. The Buddhists however do not agree with this theory. They uphold the pramana-vyavastha theory that defines the definitive nature of CMs for a definite set of CS. The Buddhists' theory is based on the nature of the OS accepted in their ontological framework, where they accept only two types of padarthas viz. svalaksana-visaya (exclusive nature of things, distinguished from rest of the world) and samanyalaksana-visaya (commonality of things, that remains intact even after the demise of the thing). The exclusive nature of the thing i.e. the exclusive OS (object structure) binds the Buddhists with a definite framework of CM (cognitive mechanism) for the causation of the cognitive episode. The Hindu philosophers in general and the Naiyayikas in particular cannot tolerate / entertain such a thesis and they out-rightly reject the Buddhists' views and stick to their pramana-samplava theory. Without going into the further intricacies of these theories we may take up the following instances for making our point clear in the school of Nyaya.

The perceptual cognition of "nila-ghatah" is different from that of the linguistic understanding generated by the sentence "nilah ghatah", according to the principles of CM i.e. karana-bhedat karyabhedah. Before examining why and how the two so called CSs being caused by different CMs differ let us formalize the cognitive structure.

COGNITIVE STRUCTURE

Each and every knowledge episode caused by the CM (except that of the indeterminate cognition, (nirvikalpaka-pratyaksa)) is presented to the mind in a qualifier-qualificand-structure, and that can be represented schematically in a model that we will call: PSV-Model: where P stands for Prakara, S stands for Samsarga and V stands for Visesya.

The PSV-model of the CS is of two types: (a) Simple PSV-structure & (b) Complex PSV-structure.

In the CS, where the schematic representation of the cognitive episode would have the minimum model of PSV, having at least one prakara (qualifier), one visesya (qualificand) and only one sa[??]sarga (relation) can be viewed as the first type of the PSV-structure. For instance, the perceptual determinate cognition (savikalpaka-pratyaksa) of a single entity like ghatah (a pot) would have the minimum PSV structure, where ghatatva (the pot-ness) appears as P = prakara (the cognitive qualifier), and ghata (pot) as V = visesya, (the cognitive qualificand), and samavaya (inherence) as S = samsarga, the cognitive relation between the P&V that makes the cognitive structure a qualified cognition (visista-jnana).

COGNITIVE RELATION & FUNCTIONAL RELATION

I would like to make it clear here that the relation in the CM is different from the relation in the CS or OS. In other words, the relation between the causal instrument (kara a) and the object can be defined as 'functional relation', whereas the relation between the P and V in the cognitive structure can be defined as 'cognitive relation'. (10)

For instance in the perceptual cognition of ghata, the eye is the instrument (Ic) and eye-object-contact is the function (F) and the result (R) is the "perceptual cognition of ghata". Here the relation of the Ic with the object (ghata) is samyoga (contact) which is called sannikarsa, the perceptual functional relation Whereas, in the CS the relation between the P&V (ghatatva and ghata) is samavaya (inherence), that is called cognitive relation being presented in the CS by the visesana-visesya-bhava-sannikarsa. Therefore, while describing the CM, we should carefully consider:

(a) What is the Ic in present context?

(b) How does it function (vyaparavad bhavati)? In other words, what is the F in present context?

(c) What is R = result (phalam), the product'?

The R would have a CS, in the PSV-model. Therefore, we must ask further questions like:

(d) What is the P in present cognitive structure?

(e) What is the V in present cognitive structure?

(f) What is the S between them?

(g) We also have to see how each of there namely P, S, and V are being represented to our mind?

And thus finally we must ask:

(h) What is the formal CS being represented after the whole exercise?

In other words, which cognitive element is the chief-qualificand and which elements are its qualifiers and how they are related each other forming the complex cognitive structure?

THE CM & CS OF LU

The Naiyayikas propose that the padarthas (word meanings) represented in the Linguistic Understanding (sabdabodha) are apparent through the pada-vitti (relation between words and their meanings) alone, not otherwise. (11) Pada stands for a morpheme or an atomic phonetic unit that bears some meaning (12), but not necessarily a finished word (13) as held by the old school of Nyaya or Vyakarana. This distinction is required to determine the padarthas (word-meanings) and samsargas (relations) in the sabdabodha that constitute it.

The relation in linguistic understanding (sabdabodha) is always to be aka[??]ksa-bhasya or apparent by samsarga-maryada (the principle of representing the relations in linguistic understanding) and the padarthas would always be either prakara (relational qualifier) or visesya (relational qualificand) to them. Let us take the example: nilah ghatah, to examine the CS of the linguistic understanding (sabdabodha). It is a simple sentence having only two words according to the grammar; nilah is adjective and ghatah is noun. But in the sabdabodha framework of sastric analysis, there are four padas viz. nila, sU, ghala, and sU. Here sU, the nominal ending or case-suffix stands for the singular number. According to Painian grammar it has a meaning and thus, it is a pada as per the definition--"saktam padam". As stated above, the first step towards representing sabdabodha is: each and every pada should represent its meaning through its vitti (signification function). In the next step, the padarthas so presented are being related to each other through the principle of samsarga-maryada giving the cognitive structure of the linguistic understanding (vakyathabodha or sabdabodha) of the sentence in question. Let us apply the said mechanism in the present sentence. The word nila in this context means nilatva-vat, (the substance colored with blue, but not just the "blue colour"), according to the lexical principle: gue sukladayah pu[??]si gu ili[??]gas to tadvati (14), i.e., the colour representative words like sukla (white) etc. mean "white colour' when they are used in masculine gender, but if they are used in the same gender that of the substantive then they would mean "possessive of the same (colour)". Here, nila is a qualifier to the ghata, because, that (nila) agrees with the grammatical gender of the ghata and thus, it means nilatra-vat "blue-possessive". The nominative suffix sU (added to the adjective nila) means abheda or "identity". The nominative base ghata means "pot", and the nominative (sg.) suffix added to it (sU) is added only for grammatical correct-ness (sadhutvartha). According to some it has the same meaning as that of the nominative base as per Paini (15). Now, the meanings so represented are "nilatva-vat", "ghatah" and their mutual relation abheda (identity) is apparent through samsarga-maryada. That gives the sabdabodha: nilatvavad-abhinnah ghatah. This represents a simple PSV structure of the sentential cognitive episode where P is nilatvavad and V is ghatah and S between them is abheda. It may be noted that if and only if a simple relation appears through samsarga-maryada giving rise to a sabdabodha then only it will be called a simple PSV structure of sabdabodha. If there are more than one relations or a complex relation apparent through samsarga-maryada then the cognitive PSV structure will be a complex one. The complexity of the cognitive structure arises on account of many other factors, which we would not like to go into details here. I just wanted to point out here that the whole exercise on CS through CM as has been presented in the school of Nyaya no doubt is scientific and presents a well-defined model of Knowledge Representation Scheme. The Naiyayikas' schemata of linguistic understanding (sabdabodha) are certainly--as far as the model, the methodology and the structure is concerned--superb, and precise. It is very convincing and scientific, that appeals to a logical mind.

BARTRAHARI--A CHALLFNGF TO THE NAIYAYIKAS' THEORY OF LU

In spite of the high precision of the Naiyayikas' scheme of sabdabodha one is forced to rethink on the validity of the CMs and CSs when someone goes through Bhartrhari's Vakyapadiya. The question obviously comes to mind is: Does the schematum of the Naiyayikas present a coherent, perfect and comprehensive theory of linguistic understanding and structural analysis? If "Yes" a lot of problems will be easily solved for those who work on Linguistic analysis or Semantics in general and in the field of Knowledge Representation in AI (Artificial Intelligence) in Computer Science in particular. The above theory would solve all the problems of Natural language Processing / Understanding (NLP / NLU) and meaning extraction / semantic representation of any given natural language at large. This is what the world is in need of and any meaningful dialogue in this direction would attract the attention of the world of intelligentsia and any convincing solution would be the achievement of the time / millennium.

However, the answer to the above question certainly is an emphatic "No". In that case the scheme of linguistic understanding as has been presented by Nyaya school of thought is to be re-examined again. The scheme as well as the theory is not satisfactory enough. Professor B.K. Matilal says:

 The Nyaya view of language was admittedly unsatisfactory. It did not
 explain why or how meaning is understood from a word or a sentence.
 For, if these meaning-bearing units were simple groups of
 sequentially uttered sound-atoms, they would lack unity, and hence
 a united meaning cannot be conveyed by them. Neither does it explain
 how these units could be simultaneously grasped in our cognition. To
 talk about perception on the last uttered sound aided by the memory
 impressions of the others is only to reveal the poverty of this
 theory's explanation assumes that we have either the word or the
 sentence as two separate, single meaning-bearing units. One of the
 implications of sphota doctrine is to posit a set of new linguistic
 entities as single meaning hearing units. (16)

Professor Matilal raises other questions like what is the nature of a "word": and what is the modes operandi for grasping the "word" through audio-perception. These questions are amplified in the tradition and attempts have been made in giving answers to them. But the real challenge comes from Bhartrhari in the manner elaborated bellow. The striking point that may he underlined here is that how far the atomic meaning units or the morphophonemic meaning units contribute to linguistic understanding? Can we say that the totality of the understanding or in other words the CS is build up step by step by intertwining one after the other atomic meaning units? Does the CS according to the Naiyayikas appear like that of a fakir's fabric or it has an appearance of a compact, complete and singularly unified fabric? These questions did not come to the atomists' mind. The question is "Why"? What appears to be more astonishing in the tradition that most of the Naiyayikas, and I precisely mean here that the bulk of Navya-Naiyayikas, who have come even a millennium years after Bhatrhari, have ignored his model of linguistic understanding. The reason behind this is yet to be thoroughly investigated upon. However, a quick remark on the problem would be that--possibly they had no answer to Bhartrhari's design of LU, and thus, they simply had to maintain silence on the objections raised by Bhartrhari on the atomic-meaning-units-of-sound-atoms-theory or contribution-of-the-morphophonemic-meaning-units-theory in building up the LU. In our understanding a complete, comprehensive and thorough study on these issues should be carried out to ascertain the real motive behind the silence of the Naiyayikas.

THE LINGUISTIC UNDERSTANDING IN BHARTRAHARIAN PERSPECTIVE

Bhartrhari is very much aware of the fact that language has its teleos in linguistic understanding. Dandin, the author of Kavyadarsa, very appropriately says: vacam eva prasadena lokayatra pravarttate, that means the worldly affair goes on only on account of speech behavior. Therefore, starting from the communication point of view till its epistemic bearing, language plays a major role in every sphere of human life. Now looking at the epistemological as well as the semantic significance of the theory of linguistic understanding, let us briefly examine what Bhartrhari says.

Thinking from the First person's experience point of view, Bhartrhari opines that the formal structure of cognition is invariably language-dependant. His often-quoted karika in this context is:

na so 'sti pratyayo loke yah sabdanugamad rte / anuviddham iva jnanam sarvam sabdena bhasate // (17)

i.e., there is no knowledge in the whole world which is invariably disassociated with the words; (and thus), each and every piece of cognitive episode is closely being intertwined with the words and the words illuminate everything. He further remarks:

vagrupata ced utkramed avabodhasya sasvati | na prakasah prakaseta sa hi pratyavamarsini [parallel] (18)

i.e., If the linguistic form of the cognitive structure is set forth for ever, than the knowledge would not illuminate at all; because, the Vak is the only pratyavamarsini, i.e., the cause of recognition and identification.

From the communication point of view Bhatrhari says:

nalabdhakramaya vaca kascid artho'bhidhiyate (19)

i.e., without proper sequential order of the speech no meaning can be communicated.

Therefore either for cognitive experience (= LU) or for speech behavior of language, the sentence is treated as the very foundation. According to Bhartrhari, a sentence is indivisible single unit, and that alone helps us in grasping the unified sentential meaning. We call it linguistic understanding (LU).

The question as has been elaborately explained above may be asked again in this context also that--what is the CM of the CS in the Bhartrharian framework of LU? From the semantic explication point of view, Bhartrhari plainly says that language or a sentence in that matter cannot be split into words or broken into morphemic units:

pade na varna vidyante varnesv avayava na ca | vakyat padanam atyantam pravibhago na kascana [parallel] (20)

i.e., there are no morphemes in the words and no phonemes in the morphological units as well; the sentence is also not very different from the words.

According to Bhartrhari, the sentence as a single and indivisible meaning-bearing unit is the cause of LU. Bhartrhari does not accept the sakti-graha i.e. grasping of the atomic meaning units and the causal mechanism (CM) for sabdabodha as has been presented above by the Naiyavikas. The theory that "the individual words either denote or connote some meaning" is simply inconceivable for Bhartrhari as far as the LU is concerned. In the Bhartrharian framework of LU, the word-meanings can be classified into two levels: one is the lexical meaning and the other is the syntactic meaning. (21) The former serves a technical purpose whereas the latter plays its role in linguistic communication. Therefore, the word meaning i.e. the lexical meaning is unreal in his framework of linguistic understanding. Some people do not understand the real difference between these two types of meanings, and they take them to be identical. The situation in a nutshell has been presented by Bhartrhari in the following karika:

Abhedapurvika bhedah kalpita vakyavadibhih | Bhedapurvan abhedams tu manyante pada-darsinah [parallel] (22)

i.e. the Vakya-vadins assume difference only on the ground of non-difference or identity-in-difference, whereas the pada-vadins do so otherwise.

TWO SCHOOLS OF THOUGHT--VAKYA-VADINS VS PADA-VADINS

According to Bhartrhari there are two schools of thoughts on linguistic understanding: (a) Vakya-vadins and (b) Pada-vadins, where the former takes the sentence as the base of linguistic communication whereas the others take the pada as its base. The Mimamsakas and the Naiyayikas are pada-vadins and they accept that the meaning-bearing unit in a sentence is the pada (individual words) alone. In other words they do not agree with the vakya-vadins, the grammarians, who uphold the theory that the sentence is an indivisible unit and it is possessive of the power of communication. The Mimamsakas and the Naiyayikas believe in the reality of individual words, their denotative function, and their meaningfulness as far as the linguistic communication is concerned. They cite the Vedic evidence in support of their theory as: padaprakrtih samhita. (23) Here if the word padaprakrti is explained etymologically as a bahuvrihi compound (padani prakrtih yashya sa). According to this etymological explanation the statement would mean that 'the sentence (samhita) is that of which the individual words are the sources'. It means that the syntactically well-formed words constitute a sentence and thus the words are the foundation of a sentence, and the), are the bearer of the meanings. Since this theory gives primacy to the words (padas) as far as the expressive potency of the language is concerned, the exponents of this theory are called pada-vadins. The point may he underlined here that 'syntactically well-formed words constitute a sentence', but the question is how to determine the syntactically well-formed-ness of the words? If it is with the help of grammar, then the grammar, especially the Paninian grammar, derives a sentence at a time and not the individual words (padas). In other words, the syntactically well-formed-ness of the words completely depend upon the syntactically well-formed-ness of a sentence, and such a sentence in turn is communicative. Therefore, Bhartrhari explains the term padaprakrti in the above cited Vedic statement taking it as an instance of tatpurusa compound (padanam prakrtih) in which we get the opposite meaning, namely "the sentence is that which is the source of the individual words". This is an interesting play of the word giving two different meanings in two different etymological explications.

To quote Bhartrhari:

padaprakrtibhavas ca vrttibhedena varnyate | padanam samhita yonih samhita va padasraya [parallel] (24)

i.e., the state of being padaprakrti is described by different vrttis (compound-formation rules), the samhita (sentence) is the source of the words or the sentence depends upon the words (in the formation of the samhita = a sentence).

There are innumerable instances given by Bhartrhari in Vakyapadiya to illustrate that the communicative potency of individual words cannot be acceptable since no word can be said with certainty of having a definite meaning. If no definite meaning can be assigned to any particular word, how can they express their "own meanings"? If no definite meaning can be obtained from any particular word then how the word-meanings would be appropriately related (or the relation be achieved through samsarga-maryada) and give a picturesque presentation of linguistic understanding? This is how a "mule kuthara"-type of criticism on the sabdabodha theory of the Naiyayikas and the Mimamsakas given by Bhartrhari, and thus their theory is rejected out rightly by the grammarians. From Bhartrhari's point of view, the padavadins' theory is just naive theory that may enjoy any critical appreciation.

No binding on the words to denote a set of certain meanings

There cannot be binding on the words (padas) to denote certain meanings or certain set of meanings. In other words, words cannot be considered as meaning bearing units due to their uncertainty nature. Bhartrhari gives innumerable account of that. For instance: raja-purusa may mean "Shine! O man!" instead of "the king's officer". To quote him:

raja-sabdena rajartho bhinnarupena gamyate | vrttav akhyatasadrsam padam aryatra yujyate [parallel] (25)

To cite a few more examples from Bhartrhari we may quote the following verse:

vaira-vasistha-girisas tathaikagarikadaya | kaiscit kathancid akhtyata nimittavadhisankaraih [parallel] (26)

The point Bhartrhari tries to emphasize here is that one and the same word may be derived differently with different stems and suffixes or by different etymology or in different system of grammar. He gives the examples of vaira, vasistha, girisa and ekagarika etc. which can be derived differently to denote different meanings. The commentators give the following etymologies for these words, for the sake of curiosity, which may be noted below:

vairam: vira-karma vairam, (Vairam means the heroic deed of a person); viresu bhavam vairam, (or Vairam may mean, he who is born in the families of heros); virasyedam (Vairam may otherwise mean that 'this (belongs) to the hero); viraya va (idam) vahram (Vairam may also mean that 'this is for the hero). However, the lexical meaning of vairam is "enmity".

vasistham: vasisthena drstam proktam va varsistham, (Vasistham means that (the mantras / suktas, which have been seen or taught by Rsi Vasistha); vasistha-sabdo 'smin vartate iti vasistham, (Vasistham may mean that the word "Vasistha" occurs in this); vasisthasyedam vasistham (Vasistham may otherwise mean that which belongs to Rsi Vasistha).

girisah: girau sete iti girisah, (The word girisa means he who rests in a mountain); girau sira (siro yasya) iti va (The word girisa may otherwise mean that whose mind / head is constantly in words / Goddess Sarasvati).

ekagarikah: aikagarikat caure, (The word ekagarika means 'a theif'); ekagaram sodha, (there are six types of ekagaram); ekam asahayam agaram prayojanam asya jighamsor aikagarika iti, (the killer or the terrorist who has a lone motive, and thus he is called aikagarika); mumusisor aikagarika iti va (the desirous of stealing is also otherwise called aikagarika); Aikagarika lexically means just a person who has a lone abode.

The point to which Bhartrhari wants to draw our attention is amply clear from these explanations. He simply underlines the uncertainty or indecisive nature of a "word" and in that case the atomic meanings or the lexical meanings of the words cannot help in getting the exact LU from a speech behaviour.

Moreover, in the system of Nyaya and Mimamsa, we split a word into prakrti & Pratyaya (verbal / nominal stem + suffix) and assign a meaning to each elements and both together give the "understanding" being logically / appropriately related by a relation. Even if this model is accepted, it would not yield the result, because the procedure would fail in many cases. For instance, there are words bereft of suffix like ahan, (a + han + 0) and there are words bereft of stems as well, like the word iyan, which has no stem. But we do understand their meanings and in these cases the above model of LU would not work.

MEANING ACCORDING TO BHARTRAHARI

Bhartrhari defines meaning in the following line:

Yasmims tuccarite sabde yada yo rthah pratiyate | tam ahur artham tasyaiva nanyad arthasya laksanam [parallel] (27)

According to Bhatrhari sabda (language) and "artha" (meaning / thought / things meant) are one and undifferentiated in their pre-verbal or potential state. Although in Bhatrhari's system thought and reality are indistinguishable, both being referred to by "artha", he did accommodate the views of others when he says that a word, being uttered, refers to three distinct items: (a) its own form (i.e. the linguistic utterance), the idea in the intellect (buddhi) and the external object (the mundane reality).

According to Bhartrhari there seems to be no reality beyond the "thing meant". Before the utterance the language along with whatever it conveys or means is like the yolk of a peahen's egg. It is seen that all the variegated colours of a full-grown peacock lie dormant in that state in potential form. Later on these colours are actualized. Similarly in the self (in the speaker-hearer environment), whosoever is gifted with linguistic capability, all the variety and differentiation of linguistic items and their meanings exist as potentialities, and that is just realized through speech behaviour. According to Bartrhari language and thought are identical at that stage. Therefore, the splitting of the words into stems and suffixes as suggested by the Naiyayikas or Mimamsakas for getting the individual / morphological meanings and thereafter building the "linguistic understanding" gradually is simply out of question. Bhartrhari says in a straightforward manner that this is not the way to do the LU business.

HOW FAR THE MORPHOLOGICAL MEANINGS CONTRIBUTE TO THE LU?

The morphological meanings do not anyway contribute in getting the sentential meaning in the system of Paninian theory of linguistic understanding. Probal Dasgupta argues almost in the same line of Bhartrhari in his paper: The Word and Not The Morpheme. (28) He examines the issue taking four direction words in English into consideration namely East, West, South and North. He says, and I quote:

 If you really believe in a sound-meaning contract ...... you are
 bound to accept counterintuitive morpheme cuts of the sort no
 linguist has ever accepted. Thus, you must then invent a morpheme
 'th', meaning 'Polar CGD (Cardinal Geographical Directions) and cut
 North and South into 'Nor-th' and 'Sou-th' correspondingly, a
 morpheme 'st' meaning 'Lateral CGD', appearing in 'Ea-st' and
 'We-st', so segmented! (Dasgupta 1996: 84)

This is comparable to Bhartrhari's age old saying:

na kupa-supa-yupanam anvayo 'rthasya drsyate | ato 'rthantaravacitvam samghatasyaiva gamyate [parallel] (29)

Can we cut k-upa, s-upa, y-upa and build a grammar as has been designed by the Mimamsakas or the Naiyayikas? Can we have a morpheme like--upa, and assign some meaning to it and make the words like kupa (a well), supa (sauce / soup), yupa (a post / pillar). The list can be expanded further with the words like r-upa, bh-upa, st-upa, dh-upa, ap-upa so on and so forth. This is not the way to do grammar or grammatical analysis.

Therefore, no morphophonemic unit can be accepted as meaning-bearing unit and thus no padartha can be obtained from the so called padas (saktam padam) and no sabdabodha of the technical nature can be achieved.

AKHANDA-VAKYA-PAKSAH OF BHARTRAHARI

The grammarians accept the indivisibility of the sentence as far as the LU is concerned. The greatest criticism / aksepa on the stand point of the Grammarians is that: Do we really not understand the word-meanings in course of communication? Is the existence of individual word fictitious? In that case, are the grammar, the lexicon and the grammatical operations unreal? Is it not that all goes against the system of Papinian gramamr itself, when we understand it as the best model of derivation? All the objections / arguments of the pada-vadins are met by Bhatrhari by having recourse to his spirit of accommodation. Punyaraja says: sarvaparsadam hi vyakaranan sastram tatra cayam sthitah ity atra punah pada-padarthav angikaroti. What he means is that after the meaning of the sentence is understood the listener does understand some meaning of the individual words. However, those meanings hardly contribute to the linguistic understanding. In other words, the individual meaning components are understood in course of linguistic operation, but not in course of linguistic understanding, as they are unreal. Because, they essentially are not useful for linguistic communication at all. The sentence gives an understanding of compact sort of meaning-unit, which cannot be actualized with individual words. Bhartrhari strongly upholds his doctrine of indivisibility of the sentence and indivisibility of sentence meaning. However, he admits the need of splitting them into parts as padas and further into prakrtis and pratyayas in order to serve practical purpose of grammatical analysis (sabdanusasanam) alone. This, on the contrary, is not required for linguistic understanding.

Nagesa (30) argues that it is simply impossible to establish the correspondence in between each sentence with its linguistic understanding, and since there is no other straightforward and simplest way to explain the mechanism, the sentences are fictitiously being split into words, and the words further split into stems and suffixes (fictitiously), and thereafter, the meaning assignment is done by permutation and combination. This exercise is being done by the exponents of the grammar just to give a technical justification to the linguistic understanding. However, this is not the desired way to explain the linguistic understanding. And therefore, Bhartrhari proposes the sphota theory and that is the only way out.

BHARTRAHARI'S SPHOTA THEORY: THE ONLY WAY OUT

It is because of the above arguments and for giving a viable theory of linguistic-understanding Bhartrhari proposes the sphota theory. The term sphota has been variously explained leading almost to mystification of the concept. Let us examine a few of these definitions:

(a) varnatirikto varnabhivyangyo 'rtha-pratyayako nityah sabdah sphotah. (31) It means, the sphota is that which is different from the atomic sounds (morphemes or sentences as well), and which is cognizable through the atomic sounds / morphemes / sentences, and which is communicative, that eternal word / sentence is called sphota.

According to the Etymology of the term the definition can follow as:

(b) sphutyate vyajyate varnaih iti sphota iti, i.e. that which is burst forth through the atomic sound units is called sphota.

(c) sphutaty artho 'smat iti sphota iti i.e. Sphota is that from which the meaning is understood.

The sphota is described as part-less, indivisible speech unit and devoid of any internal sequence. The theory, of sphota, as seen by some scholars, is certainly not a mysterious entity, even though some have tried to mystified it by wrong interpretation. We may translate sphota as "Trans-speech" or an abstract level of vak, which is beyond the level of the logistic explanation. Some relate sphota to sabda-brahman, the Eternal Verbatim as well.

Professor S. D. Joshi quoting Bhattoji Diksita says that "the linguistic theory of sphota has no direct connection with the metaphysical doctrine of sabda-brahman". Bhattoji Diksita remarkably points out that sabdabrahman was accidentally discovered by the Vaiyakarana who was searching something else. (32)

Though there are eight-fold classification of sphota, still for Bhartrhari the akhanda-vakya-sphota (sphota in the form of an indivisible, sequence-less, part-less sentence) is the real and the only possible speech unit. This alone cause the linguistic understanding in the hearer's mind rather it reveals the meaning for linguistic communication. This leads to the theory of the pratibha-vakyartha theory that is the very, foundation of poetic theory of linguistic understanding.

CONCLUDING REMARK

Looking from Bhartrhari's theory of the linguistic understanding, the question remains that: if this is accepted then will the cognitive mechanism (CMs) as has been proposed by the Naiyayikas still he operative for the sentential cognition at all? It further may be asked--whether the cognitive structure (CSs) or the knowledge episode he valid being presented in the PSV-model or not? How to account and schematically represent pratibha vakyartha theory in a structural manner? What would be the chief qualificand (mukhya-visesya), or what would be the cognitive qualifier (prakara)? How about to bring the samsarga (the cognitive relation)? Apart from all that, the questions remain--how to account for the cognitive enterprise in general and that of the linguistic understanding in particular? Can we accept nirvikalpaka-type cognition of sabdabodha in that case? Can First Person Experience he a sound ground for explaining the linguistic understanding, vakyartha-avabodhah? If these questions are considerable then it can be concluded that the Nyaya-Vaisesika theory on speech behaviour, that presents a naive doctrine of linguistic understanding needs more attention. It is surprising to think that Bhartrhari, who is an authoritative thinker of pre-Sankara era, possibly to whom Adi-Sandakacarya owes a lot for his thesis of Advairta Darsana, who has somehow or other guided though indirectly many branches of philosophical systems like Saivagamas etc., was ignored by the Naiyayikas. The question is WHY? It is unconvincing to think that Bhartrhari's work was not available to any of the Navya Nyaya scholars starting from Gangesa (13th C) to Dharmadatta Jha (20th C). Well, we do not propose that they should have agreed to Bhartrhari. However they should have agreed to disagree with him and that Would have given a serious dialogue on the cognitive enterprise of linguistic understanding. Ignoring Bhartrhari does not show a healthy sign of sastric debate. There are innumerable arguments and innumerable instances those go against the thesis of the sabdabodha in the schools of Nyaya and mimamsa. Without taking note of them how can the Naiyayikas and Mimamsakas come forward of having a sophisticated theory of linguistic understanding, let it be Abhihitanvaya-vada by Bhattas, Anvitabhidhana-vada by Prabhakaras or Samsargamaryadavada by the Naiyayikas. Bhartrhari is still a challenge for everyone whosoever works on linguistic understanding.

(1.) Upaskara on VS 7.2.26, Vaisesika Darsana (With Three Comm.), Gujrati Press, 1913. p.293.

(2.) Satyam jnanam anantam brahma, Taittiriyopanisad, 2.1.1.

(3.) Jnanadhikaranam atma, Tarka-Samgraha, by Annambhatta, Ed. by Y.V. Athalye, (II Edn & IV Impression). Bhandarkar Oriental Research Institute, Pune, 1988, p. 12.

(4.) See more on this issue in my paper "Sabdabodha, Cognitive Priority and the Odd Stories on Prakaratavada & Samsargatavada", Journal of Indian Philosophy, Vol. 27, Kluwer Academic Publishers, Netherlands, 1999, pp. 325-376.

(5.) B.K. Matilal gives a similar schematic presentation: "(I + V) [right arrow] R" where I stands for "Instrumental cause (karanam", V stands for intermediate vyapara and R stands for the "end-product". I have changed the schematic presentation for wider application with my schematic presentation of cognitive structure where I use the PSV-model of the Cognitive Structure. The "V" may cause confusion in both these schematic presentation and thus the change was required. Ref. Matilal, B.K.1990: 51 . The Word and the World: India's Contribution to the Study of Language. Oxford University Press. Oxford. New York.

(6.) Karikavali--Nyaya Siddhanta Muktavali (NSM). Karika--81. by Visvanatha Tarkapancanana. Ed. by C. Sankara Ram Shastri with five commentaries. Sri Balamanorama Series 6. Madras, 1923.

(7.) This is a sarvatantra-siddhanta, principle acceptable to all the branches of philosophies.

(8.) J. N. Mohanty. Reason and Tradition in Indian Thought (An Essay on the Nature of India Philosophical Thinking) Clarendon Press, Oxford, 1992. p. 84.

(9.) Vakya-padiya by Bhartrhari (VP.) I. 89

(10.) The tradition does make a distinction between these two types of relations. The former is called sannikarya in perceptual cognition and the latter is called "sannikarsa-bhasya-sa[??]sarga" i.e., the cognitive relation represented by the "functional relation". This means that the sannikarya in perception is the bhasaka of the relation in the cognitive structure, which is bhasya. This bhasya-bhasaka type of relationship is always there between the "functional relation" and "cognitive relation" in all the instances of the cognitive events.

(11.) The Bhattas' school propose that the padarthas otherwise present also can be incorporated in the Sabdabodha where they do not consider the pada or pada-jnana as the karana. Their theory of sabdabodha is known as Abhihitanvaya-vada.

(12.) saktam padam, Saktivada of Gadadhara.

(13.) See "te vibhaktyantah padam" Nyaya Sutra (NS), 2.2.60. also Cf. "suptia[??]antam padam" P.1.4.14.

(14.) Amara-Kosa, 1.5.17, with Com. vyakhya-sudha by Bhanuji Diksita, Ed. by Pt. Shivadatta, The Nirnayasagar Press, Bombay, 1889. p. 101.

(15.) pratipadikatha-li[equivalent to]ga-parima a-vacana-matre prathama, P.2.3.46

(16.) B. K. Matilal, The Sphota Doctrine of the Indian Grammarians, Philosophy of Language, (A Trilingual Volume Vol. 1, Ed. by Marcelo Dascal, Dietfried Gerhardus et al, Walter do Gruyter-Berlin--New York, 1992, p.611.

(17.) VP. I. 115

(18.) VP.I.132

(19.) VP. I. 89

(20.) VP. I.74

(21) See my paper on 'Indian Theory of Communication' with Special Reference to Pada-Vrtti and Sabdabodha".

(22.) VP.II.57.

(23.) Rk-Pratisakhya II.1

(24.) VP. II.58

(25.) VP.II.35

(26.) VP. II. 171

(27.) VP. II. 328

(28.) Studies in Humanities and Social Sciences (SHSS): Epistemology, Meaning and Metaphysics after Matilal. Ed. by Arindam Chakravarti, Gen. Ed. Mrinal Miri, Journal of IUC, IIAS, Shimla. 1996.

(29.) VP. II. 169

(30.) Nagesa summarizing the theory of Akhanda-vakya-paksah says: "tatra prativakye samketagrahasambhavat tad-anvakhyanasya laghupayenasakyatvac ca kalpanaya padani pravibhajya pade prakrti-pratyaya-vibhagakalpanena kalpitabhyam anvaya-vyatirekabhyam tat-tad-arthavibhagam sastra-matra-visayam parikalpayanti smacaryah" Vaiyakara a-Siddhanta-Laghu-Manjusa by Nagesabhatta, with Kala by Balambhatta and Kunjika by Durbalacarya. Edited by Madhava Shastri Bhandari. Chowkhamba Sanskrit Series 44. Banaras, 1925. pp. 5-6

(31.) Sarva-Darsana-Samgraha by Sayana Madhava, Ed. with an original Com. by MM. Vasudev Shastri Abhyankar, Bhandarkar Oriental Research Institute, Pune, (Third Edn.), 1978. p. 300

(32.) "varatikanvesanaya pravrttas cintamanim labdhavan iti vasistharamayanokta-bhanaka-nyayena sabda-vicaraya pravhttah san prasangad advaita aupanisade brahmany api vyutpadyatam ity abhiprayena bhagavan Bhartrharir vivarttavadadikam api prasangad vyudapadayat" Sabdakaustubha, Quoted by, S. D. Joshi, Sphotanirnaya of Kaundahhatta (With Intro. Trans. and Notes), University of Poona, 1967, p.43.

ACHYUTANANDA DASH DEPARTMENT OF SANSKRIT, DR. H.S.GOUR UNIVERSITYI, SAGAR E-MAIL: <ACHYUT_US@YAHOO.COM>

ACHYUTANANDA DASH

Dr. H.S. Gour University, Sagar