ÁñÁ«ÊÓƵ

Here is a sentence from Sheldon Pollock's introduction to The Language of the Gods in the World of Men : "As we acknowledge the normativity of the actual (which often manifests itself in the textualisation of reality), so we need to acknowledge the actuality of the normative (which manifests itself in the realisation of texts)."

Sentences such as this bring two questions to mind. First, simply: "What does it mean?" In the present case, I think the answer is probably something such as: "The written word not only reflects but also influences human behaviour."

The second, more interesting question is: "Why would the author choose to express himself in such a fashion?" A part of the explanation seems likely to be that a book's rhetorical style often serves to define its intended readership: a highly abstract idiom featuring characteristic usages of nouns such as "discourse" and "representation", and verbs such as "construct" and "constitute", acts as a sign reading "theoreticians only".

But, of course, the rhetoric of abstraction also confers real benefits on the author: it empowers him to say more. In place of individual things, he can deal with whole categories; in place of particular ideas, he can deal with generalised concepts; and this makes it possible for him to draw broad conclusions and make broad claims. Thus it is with Pollock, whose book concerns "two great moments of transformation in culture and power in premodern India. The first occurred around the beginning of the Common Era, when Sanskrit, long a sacred language restricted to religious practice, was reinvented as a code for literary and political expression... The second moment occurred around the beginning of the second millennium, when local speech forms were newly dignified as literary languages and began to challenge Sanskrit for the work of both poetry and polity, and in the end replaced it."

Now, it is always exciting to see a scholar bring radically new perceptions to an old subject. Pollock is clear that the first of his two "moments" formed a crucial turning point in early India: it was "one of the most momentous events in the history of culture and power in Asia"; it involved "dramatic changes" and a "discontinuity" that was "vast and total". It is therefore a little surprising that this "moment of rupture" should have gone unnoticed for so long: was earlier scholarship really so uniformly inept that it could miss something so huge and important? Or should we perhaps look for some other explanation?

The unhappy truth is that the evidence does not support Pollock's theory as clearly as he seems to suggest. Some examples may illustrate this. For Pollock, Sanskrit prior to his first "moment" was "a sacred language restricted to religious practice"; he claims that there is "virtually no indubitable evidence for its employment in any domain we would call... this-worldly". "Indubitable" is a hard requirement but evidence there is; however, Pollock tends to minimise it or to cast doubt on it (to render it "dubitable"). Thus he asserts without explanation that the language codified in the grammar of Panini, which probably dates from the middle of the 1st millennium BC, was a "rigorously normative idiolect restricted to (Vedic) pedagogical environments"; Madhav M. Deshpande's very different (and much more widely accepted) account of it as a "real" language is, also without explanation, dismissed in a footnote as "dubious, to say the least". The later grammarian Patanjali cites examples of the kind of literary Sanskrit that Pollock believes could have been composed only "post-moment"; Pollock therefore favours a minority view, dating Patanjali around AD150 rather than the generally assumed 150BC.

Pollock believes that writing in India did not long predate his first "moment": he refers to "an emerging scholarly consensus" that the Ashokan inscriptions of the 3rd century BC are not merely the earliest surviving examples of writing, but that the writing system was actually first created by the emperor's chancery for use in the inscriptions. (One may wonder why the chancery chose to invent two different scripts.) The scholarly literature on this topic is enormous, and one could not expect Pollock to refer to all of it, but it is striking that the "emerging scholarly consensus" appears to be restricted to two references. It is also striking that while the "structure and complexity" of "post-moment" literature is "unthinkable without literacy", the same is apparently not true of Panini's grammar, a work not lacking structure and complexity.

There is more. It seems, for example, that the Mahabharata had no inconvenient oral prehistory in the BC period. I am not sure why Pollock finds it necessary to take this position, since he considers that what is oral is anyway not "relevant" to the development of literature. (The same happy consideration means that non-Sanskritic authors such as Kabir are not anomalies in the context of Pollock's second "moment": such texts are not literature but "song".) Sherlock Holmes seems to have got it right when he observed of the rush to theorise: "Insensibly, one begins to twist facts to suit theories, instead of theories to suit facts." It is a shame that a work of such enormous erudition as this should imperil its chances of persuading by straining to persuade too much.

John D. Smith is reader in Sanskrit, Cambridge University.