The languages of Indian philosophy

One facet of “Indian philosophy” that may be perfectly obvious, but is nevertheless worth talking about at the beginning stages of a blog dedicated to the subject, is that nearly all of the texts we’ve been discussing or will be discussing—almost everything that Matthew Dasti’s overview encompasses—are composed in Sanskrit. (Matthew actually noted “the use of Sanskrit as a lingua franca.”) “Indian philosophy” is strikingly coextensive with “Sanskrit philosophy.”

There are many important exceptions, but many of these end up proving the rule. Buddhism and Jainism, for example, initially rejected Sanskrit both for transmitting canonical texts and for religious and philosophical discussion. But many Buddhist communities had adopted Sanskrit for one or both of these purposes by the mid-1st millennium CE, and some Jains, too, came to use Sanskrit for the purposes of explaining their own canonical texts and arguing with non-Jains.

Sanskrit was, of course, the dominant language of culture-power for a long time in South Asia. But its dominance in philosophical discourse was even more profound than in other areas of culture. What was it about Sanskrit—or philosophy—that philosophical discourse was almost entirely conducted in Sanskrit? Where are the exceptions and what do they mean? And when, if it all, were these strictures on language finally relaxed, so that philosophy could speak in other languages? (For one answer to this last question, see Nalini Bhushan and Jay Garfield’s anthology of Indian Philosophy in English.)

One factor, although perhaps just as much an effect as a cause, is a “Sanskrit language ideology” that was formulated and promoted by Mīmāṃsakas and seems to have been rather widely accepted. (I’ll discuss one dissenter, the Jain scholar Prabhācandra, in a later post.) Pollock, in a recent paper that seeks to address the “question of language” head-on, formulates the basic postulate of this ideology thus (2011: 22): “Correct language is required for the correct communication of reality.” Sanskrit is correct by definition; other languages can only get at reality indirectly, that is, by causing us to infer the Sanskrit words for the phenomenon in question. Correct, in this position, also means “eternal”: the only way that one can talk about eternal truths is to use a language that is itself eternal.

One of the key passages for this position is the section on ritual manuals (Kalpasūtrādhikaraṇa, 1.3.6: see another post on the same section) in Kumārila Bhaṭṭa’s Tantravārttika. The earlier discussion on this topic had focused on what kind of authority inhered in the “digests” of the Vedas that told Brahmin priests exactly how to perform certain sacrifices. One reason why they weren’t considered authoritative is that they were “improperly composed” (nāsanniyamāt, 1.3.12)—they lacked features such as accents that truly authoritative Vedic texts were supposed to have. But Kumārila refocuses the argument on the authority of Buddhist and Jain scriptures. They cannot be considered authoritative because—and here Kumārila reinterprets asanniyamāt—they are not in the correct language. Not only are they not in Sanskrit, but they deviate sharply from standard Prakrit, which in Kumārila’s world was the only alternative to Sanskrit if one wanted to compose a text that people could actually read and understand.

Although Kumārila frames the question primarily in terms of eternality and non-eternality, we might think in terms of a functional stability and perspicuity: Sanskrit offered a standard language, which was intelligible to educated people over a vast area, with crystal-clear rules of derivation. There also seems to have been a widespread consensus about how to map the syntax of Sanskrit onto philosophical argumentation (with the bhāva suffixes, the use of the hetu-pañcamī, etc.). But couldn’t any language have fulfilled these functions? How do we account for Sanskrit’s position in philosophical discourse? How do we relate functionalist explanations to social and historical explanations?

Sheldon Pollock, “The Languages of Science in Early Modern India.” Pp. 19-48 in Sheldon Pollock (ed.), Forms of Knowledge in Early Modern Asia: Explorations in the Intellectual History of India and Tibet, 1500-1800. Durham: Duke University Press, 2011.

21 Replies to “The languages of Indian philosophy”

  1. Andrew, this is a very interesting topic for conversation (and in my case, unabashed, under-informed speculation).

    I wonder what the role of Pāṇini and the vaiyākaraṇas is on this question. By the dawn of the classical period, our thinkers have access to an incredibly sophisticated, technical analysis of language–and not just any language, but Sanskrit, which surely must have facilitated the tasks of conceptual analysis and refinement. (I saw this while working on the agency volume; the specter of Pāṇini and the kāraka analysis tended to hover around most discussions, if only implicitly, and at a distance in some).

    If I understand correctly, by the dawn of the classical period, Sanskrit was typically learned by study and not by mere habituation, in that you learned it the way most of us have, as a second (third, etc.) language. You would memorize rules and study it as a system. This would facilitate Sanskrit’s use as a technical language, as it may then be more insulated from the fluctuations of vernaculars.

    Again, just a few ideas for conversation.

    • Yes, grammar played a major role in making Sanskrit available as a philosophical lingua franca. In fact one of Kumārila’s interpretations of asanniyamāt is “a non-existent regulation,” viz. “the absence of any rules laid down by a grammatical discourse” (vyākaraṇoktaniyamābhāvāt).

  2. Much of what you have said is fallacious.

    1. Nearly all of the ancient Indian texts are not in Sanskrit – they are rather in Old-Indic. This is a very crucial difference.

    2. Buddhism and Jainism did never reject Sanskrit at any point of time, particularly the oldest strains of Buddhism and Jainism (did not). Old-Indic (the language spoken by all the Indo-Aryans of the Buddha’s time) was incredibly complicated in terms of phonology (needless to say in terms of morphology as well). They had managed to exist without using writing until the time of the Buddha. But the Buddhist sangha did not have the wherewithal to memorize a prose collection (i.e the canonical discourses) like the Brahmins, so they wrote them down in a borrowed script (probably greek or aramaic or both or others) where that Old-Indic of the Buddha’s discourses could not be faithfully represented in writing, resulting in an imperfect record. As the script evolved over the centuries to suit the phonology of Old-Indic better, many of the Buddhist schools did move back to representing their texts in old-Indic, however just one of them (the Theravadins) held on to the imperfect script thinking it was a language by itself distinct from Old-Indic. Since the first written scriptures in India were the pali canon, it helped both Buddhism and the imperfectly written old-Indic (Pali) to become more and more popular. That Pali’s linguistic innovations set the standard for dialects that evolved from it, resulting in the emergence of Middle-Indic.

    (ao) I take this opportunity to remind all readers of our comments policy. If you can’t be polite, relevant, and scholarly, then your comment won’t clear moderation in the future, and repeat offenders may be banned.

    Regarding Buddhist (and Jain) rejection of Sanskrit, see the Pali Vinaya II 139, and the discussion of Brough, “Sakāya Niruttiyā: Cauld Kale Het,” pp. 35-42 in Beichert (ed.), Die Sprache der ältesten buddhistischen Überlieferung, Göttingen 1980; Ruegg, “On the Expressions chandaso āropema, āyataka gītassara, sarabhañña, and ārṣa as Applied to the ‘Word of the Buddha’ (buddhavacana),” pp. 283-306 in Tsuchida and Wezler (eds.), Haranandalahari, Reinbek 2000; Levman, “Sakāya Niruttiyā Revisted,” in Bulletin des études indiennes 26-27 (2008-2009): 33-51; Dundas, “Jain Attitudes towards the Sanskrit Language,” pp. 137-156 in Houben (ed.), Ideology and Status of Sanskrit: Contributions to the History of the Sanskrit Language, Leiden and New York 1996.

  3. Gaveshana or Research derives from the word Go meaning light, a connotation brought to light in modern times by Sri Aurobindo in his epochal The Secret of The Veda. This hermeneutical coup has so unsettled the received wisdom that there is unwillingness to look at the later texts in terms of the Vedic Sanskrit as unveiled by Sri Aurobindo. This blog, however, is a sign that such resistance will recede. [TNM55]

    (ao) Please look again at our comments policy.

  4. An idle wonder — and nothing more than idle since I’m in no position to analyze the Indian tradition — is that I’d be curious to know how what is said here could be compared to the medieval European tradition of philosophy done in Latin. The use of a lingua franca seems, to me, to be necessary for their to be any sort of vibrant cross-cultural exchange.

    • I also mention the use of Latin until Kant whenever I explain to a lay audience what Sanskrit means for Indian philosophy.

      Perhaps an important difference may lie in the fact that Sanskrit has been the language of all sort of culture (from scientific literature to poems, from religious literature to inscriptions) until relatively recent times, whereas Latin has been superseded by vernacular languages as the language of belles lettres already in the Middle Ages?

  5. Apologies if my earlier comment appeared confrontational.

    What you have cited does not answer one simple question: Why would every single Brahmin of the late vedic era (i.e. the Buddha’s time) be reported (as per the Pali canon) to be speaking in Pali when there is no remote evidence in their own (Brahmanical) texts of anything like Middle Indic being in existence in that period?

  6. My problem with Pollock’s thesis (as I read it, perhaps discussing it with him would be a different experience) is that it sounds deterministic: Given the Mīmāṃsakas’ purposes (preserving the supremacy of the Veda and of the Brahmin class and the like), they could have only acted the way they did. By contrast, I am inclined to think that the preservation of a Sacred Text could have also be done in the Hebrew way, that is, by keeping its language free from any other use and safeguarding it for only religious purposes.

  7. Pollock in his book “The Language of the Gods in the World of Men” argues, that until the beginning of the New Era Sanskrit was exclusively the language of Ritual and related subjects, whereas non-ritual activities were held in Prakrits. This accounts for the fact that all early epigraphical evidences (not only by Buddhist Ashoka, but also by Hindu Satavahanas) were in Prakrits. ‘Invasion’ of Sanskrit into the non-ritual sphere is believed to be connected with Shakas, who needed to have some lingua franca and being non-Indians had less problems in ignoring the previous tradition of using Sanskrit.

    • If that’s what Pollock says, then he is saying in effect that Pāṇini was seeking to standardize the single standard dialect, which beggars belief. This seems to be what the world at large believes about sanskrit.

      I think on the contrary that Panini was trying to standardize multiple Old-Indic spoken dialects of his time (one of which was Proto-Pali, the language the Buddha spoke).

      I believe Pali was simply a written representation of an Old-Indic dialect that the Buddha spoke, which is why it had no grammar or even a name for nearly a millenium after the Buddha’s demise (and which is why also that the Brahmins who spoke to the Buddha were not, as per the canon, speaking in an obviously different language). Writing was new to India and the script’s unsuitability in representing Old-Indic phonology is what introduced the gemination and many other innovations that we now consider characteristic of Middle Indic.This is also why we will never find a written form of Old-Indic from late Vedic India that pre-dates Pali.

      ‘Sakaya niruttiya’ in the Buddhist canon simply means “(with their) own etymologies”, where two brahmins were complaining that the buddhavacana was being distorted by people who used their own fanciful etymologies. Nirutta or nirukta does not mean “language”. Buddhism therefore never proscribed Old-Indic.

      • You mean, Pali is older than Sanskrit? This evidently contradicts the data of Indo-European comparative linguistics))

        • Pali is the written form of some late-vedic dialect(s) which were the spoken languages that Pāṇini sought to standardize — please note Pali was not a single dialect, it is commonly believed to be a lingua franca, a mix of many similar dialects (which is exactly what late Vedic was in the Buddha’s time). Nobody disputes that the Buddha lived in the late-Vedic era. Nobody disputes that he spoke the common vernacular of the late vedic era. Nobody disputes that Pāṇini standardized the grammar of late Old-Indo-Aryan dialects (otherwise called late Vedic). If the spoken language the Buddha spoke with the Brahmins of his time (that Pali was the written form of) was late-Vedic, then Sanskrit is the grammatical standard of late-Vedic, hence the language must have existed before the grammar. I’m just putting two and two together and have stated my theory, the evidences that fit the theory are already well researched and widely available. Why did all Buddhist traditions (except Theravada) adopt Sanskrit if the Buddha proscribed it? The answer is simple – they didn’t need to adopt sanskrit, they were already speaking in Old-Indic, but with a more evolved script they started writing it phonetically… therefore Buddhist Hybrid Sanskrit!

        • Also (written) Pali also contains an unacceptable number of homonyms to be considered a product of natural linguistic development. The Theravada tradition since the earliest times misunderstands many Pali words (which can be the case only if Pali was not a natural language). Case in point is the word Isipatana for which the commentary gives his own fanciful “nirutti” (that it was a town where rishis landed from heaven, hence called so) when it was actually *ṛṣyavṛjana, a synonym of the word mṛgadāya, with which it is almost always found coupled in the Pali canon.

        • Thomas Oberlies’ grammar of the Pali canon shows that Pali contains some archaisms that even Rigvedic (in its currently available form) has lost. Thus, not only does Pali necessarily have to be a vedic (old-Indic) dialect but also Rigvedic, being a language of the 2nd millenium BCE was likely more conservative in the Buddha’s days than a late Vedic dialect of the mid 1st millenium BCE (Pali), and certainly more conservative than the form of Vedic that exists now.

  8. “The Pali language (the liturgical Prakrit language of Theravada Buddhism) tends to be treated as a special exception from the variants of the Ardhamagadhi language, as Classical Sanskrit grammars do not consider it as a Prakrit per se, presumably for sectarian rather than linguistic reasons.”


    Why was Pali not considered a Prakrit by classical sanskrit grammarians? Maybe because it was not Middle-Indic, but rather the earliest written form of Old-Indic?

  9. I’m not quite sure what exactly the classification of Pali has to do with this topic, since my post was about Sanskrit and the use of Sanskrit by nearly everyone who was engaged in “Indian philosophy” from about the 2nd century to the 18th. Pali is of course another important language of Indian philosophy—but it’s a complex subject, and best broached another time.

    Elisa: the alternatives are what make this question so interesting. Other cultures use different languages for religious and intellectual-cultural purposes, but these lines are rather difficult to draw (especially for Mīmāṃsakas—Mīmāṃsā effectively being considered part of the Veda by some people). For Jains, Prakrit was sometimes used for “inward-looking” philosophical discussion, and Sanskrit for “outward-looking” discussion that engaged with other sects.

    • I’m surprised by your question, as to what exactly Pali’s classification has to do with this topic.

      At least to me it appears canonical Pali’s identity as a late vedic dialect has a very foundational bearing on this topic… it means, firstly that there was no real alternative to Old-Indic (I still would not like to call all Old-Indic by the name ‘Sanskrit’), and secondly that the later post-canonical Pali (which was viewed as a distinct language by Buddhaghosa et al) was probably spoken phonetically-as-written mainly or only in Buddhist (or rather just the Theravadin) circles and not across the board, which is precisely why as the evidence suggests there is no non-Buddhist Pali literature.

      Perhaps this partly answers your first question (What was it about Sanskrit—or philosophy—that philosophical discourse was almost entirely conducted in Sanskrit?)

  10. Pingback: The Philosophical Rasika Report: Listings of Ph.D. Programs in Indian Philosophy (Part I: North America)The Indian Philosophy Blog | The Indian Philosophy Blog

  11. Pingback: Linguistic accommodation and philosophical debateThe Indian Philosophy Blog | The Indian Philosophy Blog

  12. Indian philosophy aims at liberation and future No–Births. Eternal world where eternal living with no birth, no death. Hence the philosophy has to be in eternal language, Sanskrit. Not only in eternal language but through eternal source Vedas.

Leave a Reply

Your email address will not be published. Required fields are marked *