Spoken Sanskrit.

A post by Victor Mair at the Log discusses the phenomenon of Sanskrit as a spoken language. I was surprised that this “astonished” Mair, but I have an old friend who spent time in India learning the language that way, so I have long been aware that it has as widespread a spoken tradition as, say, Latin. At any rate, there are extraordinarily interesting contributions by David Shulman, Whitney Cox, Frederick Smith, Shenghai Li, and Devin Patel; I found Smith’s particularly enlightening:

H. V. Nagaraja Rao is one of the most versatile and literate Sanskrit pandits of the last half century. Practically no Sanskrit pandits speak as well as he does. Most of them do not resort to subtle or complex verb forms, but speak almost entirely in passive forms, with special emphasis on the past passive participle, for the simple reason that it’s an easy form in a complex language. It requires the agent to be in the instrumental, and the object to be the grammatical subject, thus in the nominative case. This is much easier than spinning out all sorts of complicated finite verbs. Much of the Sanskrit that I’ve heard spoken over the last 40+ years is either simplified in this and other ways, or just wrong. Often vocabulary is cut and pasted from Hindi or even Dravidian languages, modified with Sanskrit case endings. It’s definitely the case that Sanskrit is spoken in many situations among those so trained, including (a) classes at Sanskrit schools and universities; (b) formal debates among Sanskritists – usually pandits rather than scholars (although they do sometimes overlap), and generally in semi-religious contexts, such as debates between advocates of logic (nyāya, vaiśeṣika), Vedic exegesis (mīmāṃsā), and non-duality (Vedānta), or between adherents of various schools of Vedānta; or (c) within the context of Vedic sacrificial performances in which the priesthood consisting of individuals who have mastered different Vedic texts (Sāmaveda, Ṛgveda, Yajurveda) come from different parts of India. I have often been in the latter situation, where, for example, the Sāmaveda pandits come from Karnataka, the Ṛgvedins from Maharashtra, and the Yajurvedins from Andhra Pradesh or Tamilnadu. The only language they had in common was Sanskrit. I very much enjoyed those situations because what they discussed was crucial to what would happen in five minutes or the next day – it was not fixed or formal, nor was the quality of their Sanskrit evaluated by the others for its elegance – it was a real language of communication.

Sanskrit as a spoken language is still taught in Sanskrit colleges, such as the Maharaja’s Sanskrit College in Mysore, and a few places elsewhere. There has also been a rather soft movement to develop spoken Sanskrit by an organization in Bangalore called Samskrita Bharati, which has dedicated itself to the perpetuation of spoken Sanskrit across India and even among NRI’s (non-resident Indians) in the US, UK, and elsewhere. This is the highly simplified form of Sanskrit, and requires very little knowledge of grammar, and is conveniently taught to Indians because of a core vocabulary that’s familiar to them, and which they can then set into a few sentence paradigms (almost always in the passive). This is a rather stumbling movement, although this organization claims that they have taught everyone in one town (Mettur, in Shimoga district of Karnataka) to speak Sanskrit on a regular basis. I have visited this town and found that it’s a false claim. At best, some of the ordinary people can ask for a kilo of tomatoes or onions in Sanskrit. That’s about it.

I recommend the whole thing, and of course the comments are well worth reading as well.

Addendum.India’s Sanskrit speakers seek to revive ‘dead’ language” (thanks, Kobi!).


  1. Raj Kumar Jha says:

    When you talk about #Sanskrit, talk also about #Mithila, the land of ancient learning. It’s anguage #Maithili is not only akin to Sanskrit, but even contemporaneously Sanskri is used, taught and loved in #Mithila. It has not only a Sanskrit University in #Darbhanga, but has many distinguished Sanskrit Pathshalas.

  2. There are also a couple of recent movies in Sanskrit.

    In the Language Log comments, someone mentions that Maya Angelou has been translated into Sanskrit, which I surprised me more than the Ambedkar biography in verse.

    I was most interested to read about the use of sandhi among scholars getting dropped by people. This is something that I notice in Tamil too — that the less educated someone is, the more casual they are about sandhi.

  3. A Sanskrit-speaking village in Madhya Pradesh:

    Thanks to Sanskrit, Jhiri has re-discovered some lost technologies of irrigation, conservation and agriculture from the old scriptures. A siphon system of water recharging, for instance, resulted in uninterrupted water supply through the year in the fields. Small check-dams, wells and irrigation facilities followed.

    “It is matter of pride for us to retrieve these old techniques from the scriptures. With no help from the government and without using any artificial systems, we’ve reaped great benefits,” says Uday Singh Chauhan, president of the Vidya Gram Vikash Samity, which runs development programmes in the village.

    But Jhiri’s pride stops at Sanskrit. The first doctor, engineer, economist, scientist or linguist is yet to walk out from it. After finishing school, most village youth join a political party.

  4. Is it proven that Classical Sanskrit was a real spoken language at any time in the past?

    I had an impression that it was an artificial language created specifically for the purpose of compiling epic literary works.

    Of course it was based on the language which existed in the past (several centuries before the time when classical Sanskrit grammars were written), however, the Sanskrit grammarians could have had only vague idea about it.

    I tried to learn Sanskrit before and came to conclusion that it’s kind of reverse Esperanto – artificial, very logical language designed to be as complex as humanly possible.

  5. David Marjanović says:

    Classical Sanskrit is a deliberately designed regularization of Vedic Sanskrit. (Some of the regularizations may have been natural developments, but apparently most weren’t.) It was by no means created out of nothing.

    however, the Sanskrit grammarians could have had only vague idea about it.

    Don’t you think they knew the whole Vedas by heart backwards* and forwards?

    * Easier than it sounds. Word order is meaningless or nearly so, so it ends up determined by the meter.

  6. artificial, very logical language designed to be as complex as humanly possible

    Most of the complexity is not by design — morphologically, “artificial” Classical Sanskrit is actually simpler than “natural” Vedic Sanskrit. It does of course have some features that are more or less “constructed”, e.g. the use of massive multi-story compounds or the “logical” use of cases on abstract nouns where dative = purpose, ablative = cause and so on.

  7. Or from another point of view it’s a typical Prakrit gussied up to look like Vedic Sanskrit. Just another version of what our host calls High Fantastickal Speeche, Ye Olde Craparoonie.

  8. I also have some doubts whether Vedic Sanskrit was a real language.

    I mean it’s clearly a higher register of some language which existed in the past, but not the whole language.

    It’s like having the corpus of state of the union speeches by US presidents over two centuries as the only source of English texts.

    Can you haggle over price of new car using that?

  9. Customer: The manufacturer’s suggested retail price of $25,825, including upgraded audio, is a deal which will live in infamy.
    Salesman: Ask not what you should pay for this car; ask rather what this car will do for you, your lovely wife, and your two adorable children.
    Customer: Mr. Salesman, tear down this price tag!

  10. David Marjanović says:

    the “logical” use of cases on abstract nouns where dative = purpose, ablative = cause and so on.

    …How else would you use cases?

    I mean it’s clearly a higher register of some language which existed in the past, but not the whole language.

    Oh, sure – but that should have little bearing on its grammar.

  11. How else would you use cases?

    Idiosyncratically! In Latin, the accusative is not only the normal object of transitive verbs, but also a local case meaning ‘place to which’ (imus Veronam ‘we are going to Verona’) as well as expressing extent of space or duration of time (nec unum diem remoratus est ‘and he did not wait for one day’, tria milia passuum processit ‘he advanced (for) three miles’).

  12. David Marjanović says:

    Oh, you mean Classical Sanskrit cases have nothing but their “logical” uses?

  13. Rather, the idiosyncratic uses are stretched beyond all bearing until they are systematic. Latin (and German and English) have a dative of purpose, but you can’t use it for just any old purpose like you can in Skt.

  14. David Marjanović says:

    I see.

  15. In footnote 5 to a 1995 paper, Rahul Peter Das delightfully skewers Dravidianists, pseudo-Dravidianists, Americans, non-Americans, and John Emerson most delightfully (reparagraphed, some apparatus stripped):

    A side-effect of this ‘Dravidianism’ has been the effort to connect Dravidian with or show the Dravidian origins of such disparate languages as Basque, Japanese, Polynesian languages, Mandingo and the like (not to mention the still unknown language(s) of the Indus culture). These efforts may seem quaint, but they are in fact rather irritating, as much valuable time has to be wasted in convincing the gullible that they have been had.

    Happily, such tendencies are dying out, and are anyway finding ever fewer takers among scholars; however, there are still a few diehards like Clyde Ahmad Winters (Chicago) (whose articles even the otherwise serious Central Asiatic Journal has published), who finds Dravidian loanwords all over the world, and whose latest escapades include the ‘proof’ of Tocharian A and B being not only influenced by, but maybe even originating from Dravidian.

    It is interesting that among the non-Dravidian scholars most vehemently advocating such theories there is a disproportionately large number of North Americans, which seems in keeping with tendencies such as those to be observed e.g. in Joseph H. Greenberg’s ‘world etymologies’, motivated in part seemingly by considerations of domestic politics and cultural conflicts which are very often incomprehensible to non-Americans (on this cf. also Hock 1993b, whence we learn that non-acceptance of Greenberg’s notions on methodological grounds has by certain North American scholars been branded as showing “Eurocentrism”). Attention may also be drawn to the love of models, overarching theories and generalisations, more often than not based, or so it at any rate seems, on scanty primary data, which appears to be, though of course not typical only of North America, particularly wide-spread there.

    I personally have heard many non-American scholars sneer at this as an application to scholarship of ‘quick fix’ and ‘fast buck’ principles of what outsiders (doubtless themselves generalising) perceive to be the American way of life. Though quite unkind, such statements may serve as typical examples showing how such methods, clearly widely accepted, may appear to people with different cultural (and political?) backgrounds.

    But it must be pointed out that such tendencies are not confined to South Asian or Sprachbund Studies, and that non-Americans in general have no call to be complacent, as evinced by the criticism, directed against Dutch scholars of Malay poetics, of Sweeney: “… the insatiable urge to classify, codify, compartmentalize and to impose categories where none may exist often reveals more about the ways of modern scholars than about the tradition.” But irrespective of whom this criticism is directed against, it should be kept in mind by all modern scholars.

  16. David, by “logical” use of cases I meant specifically their use on an abstract noun, often the head of a compound, in ways like the following:

    sandwich-eating-dat. “in order to eat a sandwich”
    sandwich-eating-abl. “because of/after eating a sandwich”
    sandwich-eating-loc. “when/if one eats a sandwich”

    These uses obviously have roots in the basic semantics of the cases, but in later Classical Sanskrit they become, I believe, much more common and more systematized than in the earlier language.

  17. Indeed. I was reflecting on this today, and realizing that you can’t say *He cooked to eating in English to mean He cooked in order to eat.

  18. David Marjanović says:

    sandwich-eating-dat. “in order to eat a sandwich”

    Oh, that works in and around my dialect: zum Butterbrotessen. *evil grin* Be happy we lost the instrumental to phonetic erosion a thousand years ago.

    (The Standard um … zu “(in order) to” construction doesn’t exist in such dialects, and zu + dative is one of the alternatives.)

Speak Your Mind