Diglossia is not a purely linguistic issue. Diglossia is a characterization of the community, and, on the strength of their stratification, distribution of values, definition of prestige, and concomitant linguistic symptoms, communities may be called diglossic or non-diglossic. As I have been examining diglossia in this essay in the context of literacy, the linguistic symptoms have, naturally, been my main concern; I have, however, made brief remarks on social aspects of literacy, particularly social motivations for defining literacy in one way or another.

There are several types of diglossic communities. The most stable type includes Swiss diglossia, where the vernacular usage is not regarded as incorrect or inelegant but is, rather, taught, written in and cultivated as a respectable mode of behaviour. Greek is not like Swiss in this respect, but in Greek, too, both Dimotiki and Katharevusa are written, although with functional differentiation. The South Asian diglossias differ form these two in that writing is demanded in the high varieties. In South Asia, too, there are more than one type of diglossia: there are those like Tamil, that fall in line with Arabic, etc., and derive the prestigious high norms on religious and cultural considerations; there are those like Telugu and Sinhalese where such considerations do not apply. Telugu and Sinhalese have gone their separate ways on the basis of purism. All diglossic behaviour in which the linguistic usage of the classical times have been revived for prestigious purposes are instances of purism (which differentiates the Swiss situation from the South Asian situations). Where the choice of the classical usage has not been motivated by overwhelming religious - cultural considerations, the classical form survives merely as an elitist's preference. Puristic efforts of that order, with little or no religious-cultural backing, cause tension and uncertainly as seen in the Sinhalese situation. In Telugu, purism has been, so to speak, nipped in the bud, allowing for a free dialogue. On the advantages or other wise of the classical tradition; it is this non-puristic approach that is responsible for the different situations obtaining today in Sinhalese and Telugu although they had remarkably similar beginnings in the early nativising times.

I have shown that certain writing systems have the potentiality to create or enhance cleavages between the literary and non-literary usages. Although diglossia is not entirely dependent upon the availability of a literature, the diglossia of South Asia are closely associated with literary histories. We do, in fact, see the early origins of South Asian diglossia as a product of notions of literary excellence and, of course, linguistic purity. It is not surprising, therefore, that the literary performance in these communities is governed by puristic ideals. It is in this area that I have attempted to distinguish between necessary literacy and sufficient literacy.

Any insistence on normative standards of correctness is puristic. One cannot think of any other name to refer to attitudes which embody the notion that the language that is spoken by the people in their daily social intercourse is, somehow, incorrect. Purism, I have said, is socially divisive: it divides the community into purists versus non-purists or into different brands of purists. It may also create an atmosphere, sometimes by threat of force, in which people are compelled to declare their allegiances to the normative behaviour despite their faltering individual performances. Such a situation is an unfortunate one, for, fear of being wrong often deters experimentation in creative expression. The teaching system tends to abide by the puristic sentiments and to perpectuate normative teaching irrespective of these consequences. Notice that, as I have said before, I am not interested merely in people's ability to sign their names and fill in a form; my interest is in people's acquisition of literacy in terms of Nora Goddard's definition already quoted. The tensions that diglossia create seem to be damaging in the pursuit of these goals; the damage, however, cannot be quantified without a great deal of intensive research involving an interdisciplinary team of psychologists, sociologists, linguists and educational theoreticians. I have only scratched the surface of the problem.

Diglossia poses many problems to the theoretical linguist. As I have shown, albeit briefly, even such basic assumptions in theoretical linguistic as competence and deep structure take different meanings in these contexts. The implications of the acceptance of these terms as of the same universal validity have been questioned elsewhere; I have made some remarks along the same lines, but with special reference to diglossia. If diglossia is not an institution which is theoretically relevant, there is little point in pursuing with diglossia studies. It seems to be the case, however, that diglossia throws light not only on social motivations of language maintenance and use but also on implications of social norms and related performance on the formulation of linguistic theory.