The study of literacy has hitherto been confined to the acquisition of the skills of reading and writing per se, without any significant reference to the wider issues which condition the extent and content of literacy, as well as those issues which, in a circular way, are in turn conditioned by the extent, content, and the very nature of literacy in a given community. Seemingly identical issues are differently emphasized in literate and non-literate communities; abilities and objectives vary between the two (as, indeed, between any number of sub-divisions of each). Types of emphasis put on communal issues, social objectives and aspirations determine the content of literacy, and the motivations for or constraints on its spread. Linguistic communities are variously stratified: there can be, and often are, socially marked strata, with or without overtones of prestige or social gradation; there can be occupationally determined diversification; where education is widely prevalent, there can be unifying influences, based on models of excellence and on modes of imposition that are characteristic in educational systems, which may open new roads to standardization and its consequences. The values attached to different forms of linguistic behaviour are not the same in every community: it cannot be said, for instance, that the speech habits of, say, the artisans is regarded as sub-standard or inferior (or standard or superior, for that matter) the world over. Likewise, not every community automatically looks upon its literates (if there are any literates) for models of excellence: in some societies it is not easy to distinguish between a literate and an illiterate individual by means of their form of speech alone; the crystallization of such notions as 'educated speech', etc. is the product of long periods of literacy and education and of educated leadership.

Even in the same community, values attached to speech types are not always static; they change along with changes in other political and social values. Let me give an example. In the case of Tamil, the speech of the Brahmin community had enjoyed some regard and esteem until very recently. The changes in the political, and concomitant cultural, values have relegated the Brahmins to a less prestigious position in the Tamil society. As a consequence of this political and cultural change, the high or formal variety of Tamil has been stripped of Brahman (=Sanskrit) features, and the 'high' feature slots thus vacated, have been filled with seemingly 'pure' Tamil characteristics discovered in the classical Tamil usage. I have already said that similarity in social stratification does not in itself grant the same status to comparable speech styles. That is to say, the fact that the high caste speech is prestigious in one community does not in itself predict that in all caste-based societies the high caste forms will be regarded as prestigious ipso facto. For instance, even during the time when the high caste speech was regarded with respect in Tamilnadu, Sinhalese speech habits were never stratified for prestige on a caste-scale. The criteria a community may choose for defining Linguistic prestige thus depends on the community's own sets of values which may change and be moved up or down along the parameter of prestige. While linguistic stratification is symptomatic of social stratification, and linguistic prestige is symptomatic of the entire complex embodied in this sociolinguistic activity. The study of literacy in the context of its relation to sociolinguistic complexities is still virgin territory.

A society's attitude to literacy is dependent upon a number of factors. The most significant of these, for the present purpose, are the following: the presence or absence of a sense of prestige with which certain forms of linguistic behaviour are held in the society; the associations people may find necessary to make between prestigious speech behaviour and good written language; the presence or absence of any strong movements dedicated to puristic endeavours; the social desire to either delimit the domains of literacy for the creation of an elitist minority or propagate literary activity beyond such limits to produce universal literacy; the extent of freedom and influence the writers and the written word may enjoy in a given community; the types of pressures brought to bear upon the learner as well as the learner's capacity to cope with such pressures; and, perhaps above all, the society's own thinking as to whether literacy is desirable or not.

All these complexities are well demonstrated in diglossic communities. I have, therefore, made an attempt in this book to interpret the meaning of literacy in diglossic circumstances with reference to three diglossic communities, namely, the Kannada, Sinhalese and Tamil speaking communities of South Asia. As for as I know, this is the first attempt to study problems of literacy or interpret the implications of literacy from this point of view. Owing to this pioneering nature of my survey, therefore, it will be required of me to justify further my choice of diglossia (as opposed to, say, bilingualism) and then why I choose these three particular communities; for it has been suggested (if only by those whose research has not been in the area of in-depth analysis of problems of diglossias) that diglossia is a miniature bilingual situation and that problems of literacy in diglossias cannot be any more intricate than or different from those in bilingual situations, and that the latter are already known. Linguistically, most studies of diglossia that are available in print are largely typological comparisons of the high and low features with little reference to social correlations and motivations; this approach has not helped our understanding of diglossic behaviour as a particular sociolinguistic issue distinct from bilingualism. In short, the theoretical relevance of diglossia has not been studied very much. I shall, for this reason now take a very brief excursion into the nature of diglossia dealing particularly with the functional dissimilarities between diglossia and bilingualism, as well as between diglossia and 'dialects'; as a continuation of this I shall discuss in the next chapter, again briefly, the nature of writing conventions and their potential capacity to foster diglossic behaviour. In these introductory discussions, I shall also examine the beliefs that sustain diglossia and endeavour to relate these beliefs to the notions that are held about literacy in such communities.

The much quoted definition of diglossia that Charles A. Ferguson provided in his classic paper (Ferguson 1959a) reads as follows:

Diglossia is a relatively stable language situation in which, in addition to the primary dialects of the language (which may include a standard or regional standards), there is a very divergent highly codified (often grammatically more complex) superposed variety, the vehicle of a large and respected body of written literature, either of an earlier period or of another speech community, which is learned largely by formal education and is used for most written and formal spoken purposes but is not used by any sector of the community for ordinary conversation.
(p. 236)

Ferguson refers to this superposed variety as the high variety or H, and to the 'primary dialects' as the low variety or L. For the sake of convenience, I shall use the same labels in this essay to refer to the appropriate varieties. Excepting in certain matters of detail, Ferguson's definition fits the diglossic communities I have chosen to investigate; the minor modifications to this definition which might be required to accommodate my communities will become apparent in due course.

A comparison with the functioning of bilingualism and dialect diversity, as well as with register or style distinctions obtaining in almost all linguistic communities, would reveal that, in its functional status, diglossia presents a phenomenon which has characteristics unique to itself.

There are, undoubtedly, some similarities between the varieties in a diglossia and the two (or more) 'language' complexes obtaining in a bilingual (or multilingual) setting. There are, however, significant functional differences between diglossic situations and bi- or multi-lingual situations. The main difference is that, in a non-diglossic bilingual situation, the individual has a fair degree of freedom of choice between 'language 1' and 'language 2'; if the participants are equally conversant with both languages, either may be used in most situations. In this sense the individual's choice of 'language 1' or 'language 2' is not necessarily socially pre-determined in relation to the situation involved. There is no social restriction, for instance, against the use of either language by a bilingual individual when he talks to equally bilingual members of his family. In diglossia, on the other hand, the functions of the high and low varieties are socially determined, so that the use of the high variety in normal family conversation is disallowed (i.e. "it is not the done thing"). Likewise, the use of the low variety in circumstances for which the high variety alone is socially prescribed is not generally tolerated. Fishman (1967), therefore, concludes that

bilingualism is essentially a characterization of individual linguistic behaviour whereas diglossia is a characterization of linguistic organization at the socio-cultural level. (p. 34)

Rather than being a subsidiary of bilingualism, diglossia can co-exist with it, so that, as in the case of, say, the educated Sinhalese-Tamil bilingual, the individual is required by the social conventions to make appropriate uses of the high and low varieties appropriate to each community as necessitated by situations and circumstances. Fishman explicitly admitted in 1967 that a co-existence of diglossia and bilingualism would be entirely feasible, and in fact demonstrated this in the title of his paper, namely 'Bilingualism with and without diglossia; diglossia with and without bilingualism'.

Stewart (1962) makes a pertinent distinction between bilingualism per se and diglossia; this I quote below as a particularly relevant point for our present purpose:

Situations involving bilingualism can be expected to be fairly unstable and to eventually result in the dominance of one of the language over the other. In contrast, a diglossia situation is one in which the juxtaposed linguistic systems are sufficiently alike in some ways to encourage their structural fusion at certain points. This, in turn, allows for enough mutual identification of the two systems on the part of their users that they may function as situational variants of each other. Such a functional complementation of two linguistic systems is characterized by more stability than is usual in other kinds of bilingualism, so that diglossia situations may endure for considerable stretches of time without any serious encroachment of one of the languages upon the domains of the other. (p. 149)

As Ferguson has himself shown (in the above quotation from him for instance), there is a significant distinction between normal or commoner garden dialect diversities obtaining in linguistic communities and diglossic distinctions. For one thing, dialects, when they are definable, are recognized on social and geographical parameters, so that they contribute to the organization of multimorphous communities rather than unified ones tending towards identical verbal behaviour. In the case of diglossia, however, the use of the high variety is not restricted to any one geographical area or any one social class; every person is required to learn and use it for the purposes for which it has been prescribed by social convention. The education system ensures that the high variety is taught to everyone who goes through it. Here, a similarity might be seen between high varieties and standard usages: both have a unifying influence; both are educationally inculcated; and so forth. The difference between these two (which, for us, is significant) is that while the high variety in a diglossia has socially determined domains of function, standard languages do not necessarily have such delimited functioning. Note also the following comment by Stewart (1962):

In situations involving different geographical or social dialects, each linguistic sub-system, or dialect, is in most cases used by its speakers to the exclusion of other dialects of the same type. That is to say, speakers of different geographical or social dialects do not normally have command of each other's linguistic systems. Dialect differences, far from being part of the productive linguistic repertoire of the members of the wider speech community, are historically imposed upon individuals by their geographical provenance or group membership.
(p. 150)

As Stewart has shown on the same page (f. n. 4),

It is possible for originally different geographical or social dialects to come to be used coterminously in a diglossia relationship. But when this happens, the speech forms cease to be geographical or social dialects as such, and become instead the potentially common property of all members of the speech community.

The phenomenon that is closest to diglossic behaviour is the differentiation of styles or registers in linguistic communities [of course more particularly in complex linguistic communities (Gumperz 1962)]. We have shown that the use of language 1 or language 2 is not necessarily socially pre-determined for the bilingual individual. He may choose one or the other, or switch from one to the other, depending upon the participant's command in the two languages in question. Every individual, he may choose one or the other, or switch from one to the other, depending upon the participant's command in the two languages in question. Every individual, however, keeps the different styles apart most of the time, under normal circumstances; he does not, for example, use in family conversation a style of speech appropriate for public speaking. This mutually exclusiveness in the functions of different styles is closely parallel to the similar distinction obtaining between the high and low varieties in diglossia. There are, however, at least two important differences between style diversity per se and diglossic differentiation. Firstly, style diversity is a characteristic of any dialect irrespective of whether the linguistic community is diglossic or not. Where the community is diglossic the high variety is shared by all irrespective of their dialect differences, and functions, in this sense, as a pervasive 'style' distinguishing itself from other styles. Secondly, and as Ferguson has himself shown, the high variety in Diglossia in grammatically more complex and reflects an allegiance to antiquity or literary usage. Consequently, the high variety in a diglossia is of a different order and distinguishes itself in its make-up both from the formal styles of speech obtaining in non-diglossic communities as well as from the diverse styles that are available in the low varieties in diglossias themselves. It can be argued, as indeed I shall do later in this essay with special reference to literacy problems, that unlike the normal dialects and styles a person may master for his own use, the high varieties in some diglossias never approach the likeness of a component of the user's linguistic competence. In other words, they continue to retain their superposed character.

I have shown, I hope convincingly, that while it does in some ways resemble bilingualism, dialect diversity and mutually exclusive style functions, diglossia is a phenomenon different from all three of them. Diglossia, thus, emerges as a linguistic institution which is worthy of study in its own right. Not all languages are diglossic, and therefore, a number of questions remain to be answered. For example: Why are only some languages diglossic? What motivates the emergence of diglossia? What are the implications of diglossia on linguistic theory as Well as on our understanding of linguistic behaviour? To what extent is diglossia a stable language situation, and, if it ever breaks down in a society, what motivates its dissolution? The beliefs in and ideologies about literacy in diglossic communities (which are by and large used to literacy) must be coloured by the issues that would emerge as answers to these questions. I shall attempt to give brief answers to these questions and examine the views on literacy held in the chosen diglossic situations and the extent to which these views and expectations are compatible with the actual acquisition and maintenance of literacy.

From what I have said so far, the reader will have seen that diglossia is not merely a linguistic matter involving two formally different varieties. This point is very important. It is true that many studies of diglossia which are available in print concentrate solely on the formal linguistic aspects. Nevertheless, such formal linguistic differences are symptomatic of particular values, such as the concept of prestige, held as salient components in the social organization. Ferguson's own original intention was to bring into focus the sociolinguistic aspects of the phenomenon. Notice his subsequent statement.

The original definition of diglossia was based almost completely on factors outside pure linguistics. That is, they were social factors, or factors of function rather than structure. And I would stand by this approach to the problem. As soon as we try to define socio-linguistic situations in terms of linguistic structure, we find that the same kind of structure can be used for different purposes in different communities, and vice versa.
(Ferguson 1962a : p. 173)

This brings us conveniently to the issue of the individual in the context social expectations and demands, in relation, of course, to the perpetuating character of diglossias. There is a communal belief in diglossias that the high variety is somehow more correct and more elegant and presents a respectable medium for the proper conduct of formal affairs. In this sense, the high variety is seen to embody some sort of law and order externally imposed upon the individuals in the society. This does not mean, however, that every individual is capable of adhering to the normative rules of prestigious conduct; nor does it mean that all those who profess to adhere to them do so with equal accuracy are the same degree of dexterity. The illiterate sections of these communities would openly declare their ignorance of the high usage and confess to the incorrectness and inelegance of their linguistic use: there could be, even among the public personalities, a small number of individuals, who might enjoy a high degree of popular esteem and respect, but who, owing to their lack of formal education or book-experience, would not attempt to use the high variety but conduct all formal activities in, at best, a somewhat tidied-up version of their vernacular usage. Both these parties are forgiven by the society. A peasant cultivator in India or Sri Lanka, for instance, would never be publicly denounced for his ignorance of the high language. Likewise, a public personality (like, say, the late K. Kamaraj of Tamilnadu) would never suffer a setback as a result of his apparent lack of mastery in the high usage: I cite the instance of Kamaraj because it is well known that he spoke a low variety Tamil in public this degree of tolerance is built into the desire to maintain group standards, for, as Sprott (1958) says:

It is true that groups vary in the 'tightness' of their standards; some are more 'free and easy' than others, and some members may be tolerated by a group even though they behave, from the point of view of the group, very 'oddly'. (p. 13)

Such tolerance would not be shown to the membership in between, who profess to be literate and conversant with the stratified usage. They are the custodians of the prestigious standards. While all individuals are aware of the availability of 'respectable' standards, it is this middle group that advocates their use and profess to be proficient in it; and, ironically, it is their position that is most vulnerable in the society, for their inadvertent 'errors' of conduct are rarely tolerated or forgiven. Special forums are held, for instance, to denounce the occasional violations of literary rules by writers who, naturally, belong to this middle group membership: in Sri Lanka there is an annual convention for this!

Ferguson has shown on a number of occasions the extent to which myths build up about the excellence of the prestigious usage (e.g. 1959a, 1959b, 1962b). The perpetuation of diglossia rests to a very large extent on these myths and beliefs. Another comment on group standards that Sprott (1958) makes may be quoted in this regard:

Because numbers of groups conceive of the standards of their groups as outside of them individually, because they can be put into words and communicated to a stranger or to a new number, and because they can be a matter of reflection and discussion, on easily gets the idea that they really do come somehow or other from outside. The individual may have intentions of his own which conflict with the standards of his group and he feels 'coerced'. The standards may, indeed arouse such reverence that their origin is attributed to some supernatural being … When group standards are thought of as something apart form the interacting of the group members, we tend to think of them as somehow 'imposed' upon them. This gives rise to the notion that man is naturally unsocial, and that law-givers or moralists must come along and rescue him from his nasty brutish ways. (p. 14)

It so happens that even where the normative rules are explicitly laid down in grammar books and these grammar books are recognized for their excellence by the community at large, a fair proportion of writers rarely succeed in performing in full accordance with these rules. Where the high variety is not recommended for spoken purposes but is limited for writing only, the degree of homogeneity in accuracy is even less marked. In an attempt to classify the various ways in which the vernacular interferes with the high usage, I have examined elsewhere (De Silva 1974a) some of the 'errors' made by prominent Sinhalese writers. In spite of the persistent belief that all educated Tamils keep the two varieties apart and follow the literary rules correctly, there is evidence from modern writings and formal spoken usage that the actual practice falls far below the expected standard. (see Shanmugam Pillai 1965, 1972.) The same is true of Kannada. The present situation with regard to Telugu which is the culmination of a long dialogue on the advisability of maintaining a cleavage between literary and spoken usages and which embodies a semi-legislative attitude to the problem (I refer to the Telugu Language Committee Report, 1973, and its acceptance by the Andhra Government of which more later) illustrates an approach towards relaxing the disparity where its observance has more often failed than been successful.