I. Regional and Social Variation

The earliest references to regional variations in Tamil geographical dialects are found in Tolkaappiyam and in the commentaries of early Tamil grammars. They have been discussed in detail by Kamil Zvelebil (1959), T.P.Meenakshisundaram (1964), G.Srinivasa Varma (1978) and S.Sakthivel (1981). These references are only to lexical variations and not to phonological or grammatical differences.

It was Kamil Zvelebil who started the analysis and presentation of regional variations in modern Tamil dialects late in the fifties and early in the sixties, in his papers Dialects of Tamil. 1, II and Ill, (1959, 1960), and Spoken Language of Tamil (1964) presented a regional classification of Tamil dialects. In Dialect of Tamil I, he presented phonetic transcription, segmental and supra-segmental transcription, the materials he collected on tapes for 50 text sentences and also some texts--mostly folk-tales, for eight dialects (1) Madras (2) Madurai (3) Dindugul (4) Erode (5) Tirunelveli (6) Tuticorin (7) Ramnad (8) Trinconamalee (Ceylon). His informants are known by their names and place but no details are available about their caste, age, education and other variables pertinent to the study of Tamil dialects.

Zvelebil’s Dialects of Tamil II is about Madras Brahmin dialect. It is here that for the first time, he recognizes vertical dialects, though Bloch recognized it as early as 1910. He writes.

“Within the limits of local dialects we have naturally to distinguish among different cultural and educational levels. (E.g. an old illiterate woman of Madras suburbs will speak somewhat differently than a school boy of 16 who has passed primary and secondary education, though they may both use the local form of speech.)”

Apart from these differences there are distinct forms of speech spoken by different communities, which are, sometimes, commonly used by these communities irrespective of the territorial divisions (e.g. the type of speech spoken by Brahmins in Madras City, by the Chettiars of Chettinad etc.) (p.572).

Bloch recognizes a regional standard and Standard Spoken Tamil, but is not clear yet in himself. It was M.Shanmugam Pillai in 1972 who, recognized a Regional Standard Tamil (RST) and Standard Spoken Tamil (SST) for the first time as distinct between themselves and both different from written Tamil and tried to define the contexts for their uses, and recognition.

In his paper, Zvelebil gives 25 specimens of Tamil spoken in and around Madras city, Chingelpet, North Arcot, South Arcot. His informants were of different age, education and caste. He did not distinguish the Iyer and Iyengar speech differences.

In the Appendix to his paper he reproduces 50 text sentences (which he used for dialects of Tamil 1) for Tiruchirapally, Tirunelveli and Jaffna dialects.

Dialects of Tamil III, was published in Archiv Orientalni (1960) and on the basis of these data, he wrote his paper on Finite Verb Terminations in Colloquial Tamil in 1963, and Spoken Language of Tamilnadu in 1964. He attempted a regional classification of modern Tamil dialects (1964) as follows: He recognized four dialect regions in Tamilnadu.

RegionSubgroups Centres of Prestige
1. Northern (aruva: vatatalai, aruva:) - a. North East - Madras city, Chingelpet
- b. North West - Velur, Chittoor, Krishnagiri
-c. South Arcot (partly)
2.Eastern (punal, panri) - a. Kaveri Delta - Tiruchi, Thanjavur, Karur
- b. Pudukottai
3. Western - a. West North - Salem, Erode (Bangalore)
- b. West South - Coimbatore, Dindugul
4. Southern (Vatapa:nti, tenpa:nti) - a. Madurai
- b. South West - Tirunelveli, Nagercoil
- c. South East - Ramnad, Tuticorin
5. Ceylon - a. North
- b. colombo?
- c. North East - Triconamalee
- d. South East - Baticaloo

He discusses isoglosses for each of these dialects which is tabulated along with those given by Sakthivel.

Sakthivel (1981), based on the survey carried on by the department of Linguistics, Annamalai University, repeats the four regions of Kamil Zvelebil but gives more isoglosses, but also does not mention some, which Zvelebil gives. His regions are,

	1. Northern Dialect - Madras, Chingelpet and North 	Arcot Districts.
	2. Central Dialect - Tiruchirapally, Thanjavur and South Arcot Districts.
	3. Western Dialect - Salem, Dharmapuri, Coimbatore and Nilgiri Districts.
	4. Southern Dialect - Madurai, Ramnad, Tirunelveli and Kanyakumari Districts.

I. Spoken Tamil

Early in this century, a few western scholars wrote about the spoken forms of Tamil and very few Tamil scholars were interested in them. Jules Bloch (1990) reported the differences between the Brahmin, Non-Brahmin and untouchable Tamil. L.V. Ramaswamy Aiyar reported differences between Brahmin Tulu and folk Tulu. Beschi wrote a grammar for spoken Tamil – Kottuntamil (1728, reprint 1971). R.P. Sethu Pillai (1939), for the first time, made an attempt to present the differences between written and colloquial Tamil. Since then, till the sixties, there was almost a complete silence in the study of social dialects or in colloquial Tamil.

Not many native scholars took interest in the spoken form of Tamil till the sixties. Tamil scholarship was identified with the mastery of ancient and medieval classical literature and grammar and it is these scholars who also were interested in language study-philology, as it was then called, and for them spoken Tamil was untouchable, impure, and polluting, with all its implications, though they live with it all the time. The political and social philosophy of the Dravida Kazakam (DK) and Dravida Munneerra Kazakam (DMK) patronized the glorification of ancient Tamil literature and the culture enshrined in them. Another important factor was the dominance of Caldwell’s Comparative grammar attitude. Philologists of these decades were interested more in the problems—Sanskrit verses Dravidian—identified with Tamil, approach to establish the independence or interdependence of one upon the other, initiated by Caldwell. Even, scholars, who were interested in the scholarly elaboration of Caldwell theories, could not resist this temptation. Perhaps this was a political and social necessity, because of the dominance of the theory that everything came and originated from Sanskrit. Moreover, the studies by these philologists, K.V.Subbayya, C.P.Venkatarama Aiyar, L.V. Ramaswamy Aiyar, S.Anavaratanayagam Pillai, Venkatarajulu Reddiar, P.S.Subrahmanya Sastri, S.K.Chatterji etc., on Tamil and Dravidian, were based on the written form of the languages. It was the western scholars again M.B.Emeneau and T.Burrow, who resorted to the analysis of spoken forms of Dravidian languages, and used those materials for studies in Comparative Dravidian. Dialect studies, whether regional or social, were not thought of, till the end of the fifties. Referring to Jules Bloch’s paper in 1910, Susan S.Bean says, that for nearly fifty years his contribution stood virtually alone (1974).

The sixties of this century is a landmark in the development of linguistic studies in Tamilnadu after Caldwell and Bloch. This is also true of South Asian languages in general. Some of the American structural linguists and their Indian counterparts in the late fifties and early sixties got interested in caste dialects, as it was then called, and in the evolution and evaluation of standard forms of Tamil. As for Tamil the papers by William Bright (1960), William Bright and A.K. Ramanujan (1964), Ramanujan (1968), M. Shanmugam Pillai (1965 a, 1965 b, 1968 reproduced 1981) started the discussion of caste dialects. A hypothesis was proposed by William Bright and A.K.Ramanujan concerning the brahmin (B) and non-brahmin (NB) speech. “In general, the brahmin dialect seems to show great innovation on the more conscious levels of linguistic change – those of borrowing and semantic extension—while the non brahmin dialect shows greater innovation in less conscious type of change—those involving phonemic and morphological replacements”. Bright explains these tendencies in terms of literacy. He claims that “a possible hypothesis is that literacy, most common among Brahmins has acted as a brake on change in their dialects-that the ‘frozen’ phonology and grammar of the literary language have served to retard change in Brahmin speech”. Another version of this hypothesis (1964) is “on the part of B toward greater use of foreign vocabulary and semantic shifts; on the parts of NS towards shifts in native phonology and in morphology”. They conclude that “Upper and lower class dialects innovate independently of one another and in two ways, here labeled conscious and unconscious. They claim that “literacy, where ever it is present in human societies, acts as a brake on the process of linguistic change. Susan S.Bean (1974) by a comparative study of the Sociolinguistic studies so far accomplished for South Asian Languages formulates some interesting hypothesis and new directions for further research. Bright and Ramanujan’s hypothesis has been elaborated and challenged by a few like M. Shanmugam Pillai (1968), D.P.Pattanayak (1975), V. Balasubramaniam (1980), G. Sankaranarayanan (1980).

M. Shanmugam Pillai in his paper Fisherman Tamil of Kanyakumari (1968 reproduced 1981) compares the Fishermen Tamil (KF) with the Harijan (H), Nadar (N) and Vellala (V) Tamil dialects of the same district, at the phonological, morphological and lexical level, Sorting out the common and distinctive features for these dialects and comparing them, he concludes that innovations in dialects can only be comparative and it is always doubtful whether any dialect can enjoy monopoly otherwise. Out of 83 items discussed, 38 are innovations for KF when compared with other dialects. In a tracheotomy of Non-Harijan Tamil dialects as (1) Brahmin (2) Higher Non-Brahmin (3) Lower –Non-Brahmin, KF is found to be nearer to Lower-Non-Brahmin Tamil dialect. He also suggests in his paper that dialect switching is due to the hierarchy of the caste structure, coupled with its prestige and social politics and not economics or education. The economically and educationally backward low Non-Brahmins (and also the Harijans who are still far below) and the economically and educationally forward Brahmins, both switch on to a dialect of the upper Non-Brahmin castes. Brahmins switch on to this because of social and political necessity.

V. Balasubramaniam (1980) testing the hypothesis of Bright and Ramanujan (1964) in his own way, at lexical, morphological and phonological levels, reports that the exclusive features attributed to B are found in NB dialects also particularly in the Vellala and Mudaliyar dialects of Tanjore and South Arcot District. There was a period, during previous generation before the Non-Brahmin movement gained momentum, when the Vellalas and Mudaliyars of these districts, considered the B Tamil prestigious and imitated it. They even criticised their children if they will not speak like Brahmins, and spoke ND Tamil. How can one be sure that this has not resulted in the diffusion of these features into their dialect? Unless this is established otherwise, Balasubramaniam’s argument cannot be valid. To test the corollary of their hypothesis, he has picked up the Dravidia Smartha Iyer of Mysore, whose ancestors have migrated to Mysore nearly two hundred years before, from Tanjore district—he calls their speech Mysore Iyer Tamil—MIT. He lists a number of changes in their dialect, at phonological, morphological and vocabulary levels. There could not be borrowings from Kannada, according to him. Here B Tamil, MIT, has a number of innovations, which according to Bright-Ramanujam hypothesis could not be. But, how can a Tamil dialect transplanted in midst of an entirely new linguistic surrounding for nearly two centuries, be expected to be similar to B Tamil of Tamilnadu and help to prove or disprove the hypothesis? MIT is entirely a different problem which will not affect their hypothesis in any way, which is concerned with the dialects in Tamilnadu subjected to sociolinguistic stains there in. However, his third point whether literacy is a brake in language change has to be investigated.

G.Sankaranarayanan(1980) citizens the methodology of the early studies on sociolinguistics. Their data was not collected by two individuals sampling across sections of speakers. To represent two caste dialects, one for each in Mysore, by Bright, is questioned. Then he says that among the Brahmins, there are two groups, Iyer and Iyengar, who differ in speech and culture, and whose speech disproves the hypothesis that ‘it is B which has innovated by introducing the loan words. B frequently preserves non-native phonology, which NB assimilates to native pattern’ (Bright and Ramanujan 1964: 160, 161). The Author who questioned the methodology of these linguists should have followed the method he is advocating, to disprove their hypothesis. He has not done so and unfortunately, he also follows Bright-Ramanujan method.

Bright-Ramanujan hypothesis may be right or wrong. But in the anxiety to explore the hypothesis, both of them have not followed Labov’s or any other scientific method, which they advocate rightly. On the other hand, they have been following the same method, questionable in the choice of dialects.

Sociolinguists who are allergic to the expression caste, led by Pattanayak, themselves admit that caste is a dominant variable among the many variables. Who among the caste oriented sociolinguists, has claimed that caste is the only variable? Nor did they claim that their studies represent a homogenous stereotype, Brahmins from any part of Tamilnadu. It has several stereotype caste markers, some of which have diffused to NB also. It cannot be claimed this for other Tamil castes, though one could easily locate the region. The younger generation both B and NB, which is switching on to a standard spoken form, will not help in this game. The anti-caste, or may be inter caste sociolinguistics, by the way, is a good political jargon to hide realities which is typical of Indian politics. Pattanayak, like a politician says “In fact, the talk about caste dialects may even have negatively reinforced caste identities and feelings”. But the impression is that the talk about caste dialects in the last fifteen years have contributed much to the evolution of a standard spoken form for Tamil, slowly wiping out the caste markers, though it has never been anyone’s aim. When the caste markers are overtly demonstrated by the linguists they become very much aware of them, and hence avoid using them and they drop the stereo forms.

Labov’s Survey of Martha Vineyard and New Yark City, and Trudgill’s Norwich English, are begining to have their impact on sociolinguistic research in Tamil, but not very intensive or extensive. Yet K.Karunakaran in his A Note on Linguistic Change: a Sociolinguistic Appraisal (1978) reports that phonetic changes i > e, i > o in terms like tira, tera, tora, and i > u in items like piti, puti are socially conditioned. i > e is found consistently in the speech of educated speakers in their informal speech behaviour and they are also consistent in retaining i in formal speech. Contrary to this change, i > o is consistent in the speech of uneducated speakers. In some cases the change i > e and retention of i are conditioned by the social status of the speakers, upper class versus lower class, high occupational status with low occupational status etc. Some speakers at Annamalainagar who begin using u and o for i, quickly realizing the formal situation, switched over to u > a (aru > ara) and o > i. He also gives a list of such changes in consonants, but does not describe the conditioning social variables. The degree of assimilation of the borrowings, he says, is conditioned by social parameters, education etc., of the speakers. We may perhaps start with a hypothesis that in phonology, if there is more than one phonological situation, i.e., linguistic environments the differences may be conditioned by variables other than linguistics. The same hypothesis can be extended to grammatical and lexical levels.

In another paper, A Study of Social Dialects in Tamil--Methods and Practices (I978), Karunakaran presents a model to present the grammar of a language. Keeping the caste constant, Chidambaram Vellala, introduces many variables, economic status-higher income, middle income and lower income groups, education--higher, secondary and primary, location—P1 and P2. The number of segmental phonemes, clusters and some grammatical features presented are conditioned by these variables. He also presents model rules with these variables operating in specific linguistic contexts.

As an extension of this model, in Study of Social Dialects of Tamil (1981), he demonstrates how rules could be framed to describe the distribution of the linguistic features- the variants, conditioned by sociological and other variables, and, he demonstrates it with particular reference to Tamil at phonological and grammatical levels.

And thus, the eighties begin with the impact of Labov and Trudgill in sociolinguistic research in Tamil.

2. Standard Tamil

Early in this century, till early fifties, Tamil philologists used to say that modern written Tamil, perhaps as pure as it could be, is the Standard Tamil. The question that arose in the late fifties is “standard for what, to speak or to write?”. If it is to speak, it is aiming at a prescriptive norm, as was natural with the grammarians then, who were more prescriptive than descriptive. If it is for writing, the style differs very much between authors. Pandits at one extreme and short story writers and novelists at the other extreme, and in between, there are a wide range of variations, about which they were not concerned.

Kamil Zvelebil, in 1959 in Archiev Orientalni, for the first time, argued for the existence of a Standard Spoken Tamil. He said that the colloquial Tamil is distinct from modern written Tamil. He writes,

“Between the cultivated speech and the folk speech there is another level to be found which may be called common speech and which is, though it may vary slightly from place to place and from community to community, fairly homogeneous for the whole linguistic area. This is usually called ‘colloquial Tamil’” (p.573).

In 1963, he again discussed this problem in Tamil Culture (Vo.X.No.3). Standard Spoken Tamil is being evolved out of the Tamil spoken by the educated middle class Pillai’s of the cities of the districts of Madurai, Tiruchirapally, Thanjavur and South Arcot. But he did not support his hypothesis with any data.

Again in 1964, in Archiev Orientalni, he suggested that Standard Spoken Tamil evolves out of the middle class speakers, primarily non-brahmin castes, who are involved in mercantile and professional activities in central Tamilnadu i.e. Tiruchi, Thanjavur, Karur and South Arcot regions.

Shanmugam Pillai in his Caste Isoglosses in Kinship Terms (1965, reproduced 1981) suggested that the data presented in that paper may be a small piece of evidence in support of Zvelebil’s theory. But later in his paper Fishermen Tamil of Kanyakumari (1965, reproduced 1978) he reported that many of the peculiarities listed among the older generation are being replaced by other forms in the speech of educated younger generation in this dialect. Dialect switching, he reported, is due to the hierarchy of the caste structure, coupled with its prestige and politics, and not due to any economic and educational factors. The educationally and economically forward higher caste-the Brahmins, and economically and educationally backward lower non-brahmins and scheduled castes, switch on to the dialect of the higher non-brahmin castes. The rising fisherman generation switch on to a dialect of the Vellala Community in that district. In 1972, in his paper Tamil Today he argued for the existence of a regional standard apart from SST. In another paper Tamil Literacy and colloquial (I972, a) he referred to the fact that the speech of the younger generation is different from that of the older generation, and SST evolves out of the speech of the educated younger generation, where children of different castes and regions mix together, live together and communicate between themselves, in schools and colleges.

Thus, the early linguists who dealt with this problem, were shifting their stands within a decade, because their hypothesis were based on observations and participation which gave more and more information as time passed on, and not on any systematic survey.

Franklin C.Southworth in his paper Problems of defining Standard Language in India, England and United States (I 972) deals with the socio-economic factors associated with the establishment of ‘Standard Language’. Standard usage is based on the native usage of certain groups. In most cases the standard forms have been shaped by scholars who were more concerned with linguistic purity than with mass communication, and about Tamil he writes.

More recently (mainly since the beginning of the 20th century), as certain high non-brahmin castes gained political and economic power, the Sanskritized form was replaced by what is called ‘pure’ Tamil, whose vocabulary is extensively borrowed from the classical Tamil.

Southworth refers to the modern written form, when he talks about standard language for Tamil. But he explains how the low castes are in a disadvantageous position as before when this standard Tamil is different from their spoken forms and the situation helps only the privileged class to perpetuate their dominance.

The problem was again raised by Vasanthakumari (1976). She putsforth a hypothesis that SST does not evolve out of any particular region or caste or religion, but out of the speech of the privileged class (questioned by E.Annamalai, 1976). Annamalai (1976) raising a number of objections, puts forth an interesting hypothesis that SST is obtained by eliminating the stigmatised stereotyped or marked features of home dialects, which will give a set of neutral items in SST. He further suggests that there should be demonstration for the existence of SST in the following ways:

	(1) The evidence to switch over from home dialect in mixed group. 
	(2) Stylistically correlated variations of speech. 
	(3) Consistency in the language spoken in mass media, like cinema.
	(4) Hyper-Correction (in Labov’s sense 1971) by speakers of non-standard dialects.

V.Gnanasundaram (1981) carried on a small survey to test the first. His findings were that non-stereo forms resist style shift more than stereo forms. There is no consistency in stereo forms undergoing style shift in mixed groups. He also concludes that SST is nearer to written Tamil.

Shanmugam Pillai and M.Murugesan, M.Shanmugam Pillai and S.Radhakrishnan (1983) carried on an investigation at Pillayyarkuppam in Pondicherry state to investigate the age variability in spoken Tamils among the males as well as the females, comparing different age groups from the Padayachi community at the phonological, morphological, syntactic and lexical levels. Three levels of age groups, one near sixty (E2) another near forty (E1) and the third near twenty (Y) were compared between themselves and they with modern written Tamil (WT). El and E2 are illiterates and Y has elementary or secondary school education. It was found that E1 E2 formed one group, distinct from the group WY, each group sharing a number of features at all levels, different from the other group. The data for each of the age group was checked with five educated informants from the city of Pondicherry, selected at random, without any reference to caste, religion, or age. It was found that these five informants shared 93 percent of the lexical items with Y and only 7 percent with, El, and E2. W and Y constitute a single group, distinct from El and E2 which form another group sharing a number of items at these levels. WY has many stereo forms distinct from E1 E2. Thus Y is nearer to W. At the same time Y, El and E2 has a number of stereo forms. Therefore, not only the elimination of the stereo forms, but also the creation of new stereo forms, contributes to the evolution of SST. In the light of these investigations, Annamalai’s hypothesis may have to be rephrased that SST evolves as a result of the replacement as well as creations of stereo forms in the speech of the educated younger generation, irrespective of caste, region, religion or economic status. The percentage of new stereo forms created is smaller than elimination. Since, WT does not have any stereo forms for caste, or region, Y chooses to come closer to it, to evolve a neutral dialect as SST. Perhaps, if Gnanasundaram’s experiment has introduced age and educational variables, the results would have been much better, in favour of the above hypothesis. All his informants were well educated and perhaps not much different in age either, where only regional variables were in operation and it is not as pertinent as age and education in the evolution of SST. That nothing conclusive could be said as a result of his experiment, may itself be a proof that young age and education are the important variables and most of his informants are tending towards SST.

The younger generation speaks SST and most of the members in it, do not speak their home dialect, which is still otherwise spoken by the older, uneducated members in their families, and as such, the question of dialect switching in mixed group does not arise for them. The first generation educated groups are investigated in contrast with the early illiterate, older generations. What will be the picture if the three age groups are literates? May be, it is true as Annamalai (I976) said,

‘No standard dialect is entirely homogeneous and rigid and what is accepted as standard is a range within which variations are permitted. This is perhaps more pronounced in Tamil’

K.Karunakaran (I980) in Language standardization and linguistic Convergences demonstrates how minority language features -- grammatical and lexical, are being standardized in majority languages. He illustrates how Malayalam features- grammatical and lexical, are standardized in the Tamil spoken in the border areas of Tamil and Malayalam, and how Telugu lexical items are standardized in Tiruttani Tamil, bordering Telugu country. These are the results of language contact between these languages, and one can add, that a sizable majority in these regions are bilinguals. But the question is, will these items be accepted as standard in SST? It is a disputable question. One can say that they are regional standards and standardization processes can operate at different- levels.

II. The Study of Variation in Tamil

I. Tamil Dialectology: A Historical Perspective

It was Jules Bloch who in 1910 started a systematic Tamil dialect study. He published a description of Tamil social dialects in 1910. His tripartite classification of the Tamil socialists consisted of three dialects, namely, 1. Brahmin Dialect, 2. Non-Brabmin Dialect and 3. Harijan Dialect. Many of the dialectological studies carried out in Tamil were very limited in their area of investigation. Much care was taken to describe the structure of the language rather than to describe the variations attested. In this way a number of dialectological works came under the captions A Descriptive study of Saiva Vellala Tamil Dialect, Udaiyar Tamil Dialect, Brahmin Tamil Dialect and so on.

It was only after the publication of the sociolinguistic studies carried out by Bright (1960), some attention was paid to area of speech variation. After this period, the trend has changed considerably, and the attention of the scholars who worked in this field was directed towards the intricate and complex nature of the language which they analysed. In studying the socially correlated linguistic variation, scholars made use of different levels of linguistic structure.

All the works hitherto mentioned approved the existence of caste dialects. Pattanayak (1975) disapproving the existence of caste dialects put forward his hypothesis that caste differences in dialects may be a marginally determinant variable only at the rural sub-caste level. He supported his hypothesis by arguing that ‘scholars who investigated language variation in India have taken caste dialects for granted as a priori assumption. The notion of caste dialect is unscientific and unnecessary. Caste may be one of the contexts along with rural: urban, educated: uneducated, peer group: old generation, male: female, etc. but certainly as an approach to, and category in dialectology has much less validity than what has been attributed to it by scholars”.

Scholars have varied opinion regarding the hypothesis put forward by Patanayak. His hypothesis was partly criticised by Tiwari (1975), Gopinathan Nair (1975), Karunakaran (1975) Ray (1975), Somasekharan Nair (1975), Bright (1975), Joshi (1975), Neethivanan (1975), and Upadhaya (1975).

Annamalai (1975) has made certain suggestions regarding the necessity for conducting an empirical study and a thorough ‘sociolinguistic survey’ of the various Tamil speaking areas in order to establish certain ‘sociolinguistic correlations’ and to describe ‘language use’.

Annamalai (1975) has made certain suggestions regarding the necessity for conducting an empirical study and a thorough ‘sociolinguistic survey’ of the various Tamil speaking areas in order to establish certain ‘sociolinguistic correlations’ and to describe ‘language use’.

2. Recent Trends in Social Dialectology

The publication of Pattannayak’s (1975) paper Caste and Language provoked hot discussions with regard to the caste dialects. Gnanam’s (1980) study which is made in a semi urban area (industrialised are) has shown that caste is not an independent variable in conditioning the linguistic variations attested in the speech behaviour. His study has made it clear that education-cum-economic status is an important social variable as caste is conditioning the variations in speech. 47.6% of variations are conditioned by a single social parameter namely, educational-cum-economic status, in this particular caste at least, is equally important social parameter like the case. Irulappan (1979) has correlated the linguistic variations attested in the speech behaviour of rural community with the sociological parameters. Sivashanmugam (1981) and Muthuswami Pillai (1981) have studied the speech behaviour of urban areas and have differentiated the speech communities by correlating their linguistic variations with sociological parameters. The importance of these studies stem from the fact that these were the first of its kind in correlating speech behaviour with various sociological parameters. These studies employ the research methodologies and techniques of linguistics as well as sociology.

Sivashanmugam (1981), Gnanam (1980) and Irulappan (1979) have studied the linguistic variations in urban, semi-urban and rural areas respectively and they have correlated the variations with the sociological parameters pertaining to the societies concerned.


Diglossic situation exists in Tamil speech community. Tamil language has two varieties, namely the Literary (LT) and Spoken (ST). Both the Literary and Spoken varieties have significant differences in their structure and function and the attitudes of Tamil speech community towards them also are different.

1. Earlier Study in Tamil Diglossia

1.1. Ferguson, who developed the concept of diglossia first, says that Tamil as a diglossic one which “fits the definition exactly.” Further he says, Tamil diglossia seems to go back to many centuries, since the language of early literature contrasts sharply with the language of early inscriptions, which probably reflect the spoken language of that time.”

1.2. M. Shanmugam Pillai is the first linguist who has clearly pointed out the differences between the two varieties of Tamil on all levels.

1.3. R.P. Sethu Pillai, though during his time the concept of diglossia was not developed, studied the differences between the two varieties of Tamil. He has given some conversion rules existing between the literary and spoken varieties.

1.4. Kamil Zvelebil states, “....... the gap between the two (Standard Literary Tamil and standard Colloquial Tamil) may be very wide indeed, may widen so that Colloquial Tamil and Literary Tamil will be considered as the two opposite poles of the same language.” He has pointed out the differences between the literary and spoken varieties. He has studied in detail about the dialects of Tamil.

1.5. Caldwell also noted the differences between the literary and spoken varieties of Tamil. He said that “Classical Tamil, which not only contains all the refinements which the Tamil has received, but also exhibits to some extent the primitive condition of the language, differs more from the colloquial Tamil than the classical dialect of any other Dravidian idiom differs from its ordinary dialect. It differs from colloquial Tamil so considerably that it might almost be considered as a distinct language. He said that the difference between the literary and spoken varieties is a remarkable peculiarity of the Indian languages, especially of the four “cultured Dravidian languages.”

1.6. De Silva has made some study on Tamil diglossic situation. He has analyzed the relation between diglossia and literacy in Tamil Speech Community.

1.7. Andronov, Trudgil, Ramanujan, Bh. Krishnamoorthy, S.V. Shanmugam, Gnana sundaram and Yesudasan are some of the linguists who have studied one or other aspects of Tamil diglossia in their studies.

2. Diglossia in Earlier Tamil

2.1. Tamil diglossia has been maintained for ages. Evidences for the existence of diglossia in earlier time could be found in ancient grammatical works, literature, inscriptions and folk literature.

2.2. Tolkappiyam is the existing earliest Tamil grammatical work which is supposed to be a work written around third century B.C. It has mentioned two varieties of Tamil Ceyyuִl and Valakku. Vaִlakku, according to Tolkappiyam is the spoken variety and Ceyyul is the literary one. Ceyyuִl was accepted as the variety which should be used in literatures. So it is very clear that even in Tolkappiyar’s period, the dichotomy between the literary and spoken varieties had existed.

2.3. Though ancient Tamil literature were written only in the literary variety, one could find some words belonging to the spoken variety. Sethu Pillai says, “Dialectal words, however, have found their way into the vocabulary of the standard language either by virtue of their intrinsic worth or by the authority of the poets who pressed them into service. It would be almost impossible at this distance of time to discover these words in the ancient poems. Even the great epics Cilappatikaram and Cintamani contain several dialectal terms. According to Adiyarkkunallar, the commentator of Cilappatikaram, Cirumi (a young girl) is a dialect of Kudanadu. Pudai (in the sense of cover) is a dialect of Malainadu and so forth. Nacinarkkiniyar, in his commentary of Cintamani has disclosed the dialectal character of the second personal pronouns nim, of the pronominal genitive nan, the interjection idea and so on. Thus the accession of dialectal terms has enriched the vocabulary of the standard language.”

2.4. As Velu Pillai says, the language of inscriptions seems to be very close to the spoken variety than to the literary one. Kamil Zvelebil also has the same opinion. He states, “One thing is beyond doubt, the language of earliest Tamil inscriptions differs in some features from the language of contemporary literature, and the differences are of such nature that we are entitled to suggest that this epigraphical langauge reflects the spoken language of the time.”

2.5. The use of spoken variety in inscriptions might be due to the reason that inscriptions were written mainly to convey something like King’s orders, land tax, or King’s greatness etc., to the common people. But classical literature was written only by educated persons for literate elites only.

3. The Two Varieties of Tamil Diglossia

3.1. Literary Variety

Modern Tamil has two varieties, LT and ST. Shanmugam Pillai has divided the modern LT as follows.

	1.Pandit Tamil
	2. Commonly accepted Literary Tamil

He calls the Pandit Tamil as an “artificial style” and defines it as one which “is replete with archaic words and constructions and use ‘pure’ Tamil Words”. This LT is being used in some journals like Centamil, Centamilccelvi and Tenmoli. The commonly accepted literary Tamil is the variety used in all popular dailies, weeklies, and monthlies like Tinattanti, Kumutam, Kalaikatir and Peecumpatam.

3.2. Spoken Variety

Likewise, the spoken variety of Tamil may be divided as follows:

	1. Regional dialects
	2. Social dialects
		(a) Communal/caste dialects
		(b) Rural Vs Urban dialects
		(c) Literate Vs Illiterate dialects
	3. Literate non- Brahmin spoken variety which could be understood in all parts of Tamilnadu.

The present author, for his analysis of Tamil diglossia, takes the commonly accepted Literary Tamil and the third of the above mentioned spoken varieties as LT and ST respectively.

4. Functional Distribution of LT and ST

4.1. One of the most important features of diglossia which distinguishes it from other language situations is the functions between its high and low varieties.

4.2. In diglossic languages, the functional distinction between the high and low varieties is determined by the concerned societies, not by individuals. Where, when and to whom, the high or low variety should be used is not at the discretion of any individual. The norms prescribed by the society regarding the functional allotment between the varieties should be strictly followed by all. If anyone deviates from the social norm, he may be ridiculed or sometimes condemned by others in his society. So “the importance of using the right variety in the right situations can hardly be overestimated.”

If any newcomer in a particular society, where diglossic situation exists, likes to have active participation in all social functions, he has to learn both the varieties and also he should have a sound knowledge of the social norm regarding the functional distribution.

4.3. Generally the high variety of diglossic languages is used for written purposes. In some diglossic languages, it may also do the formal spoken function. For example, in Sinhalese diglossia, the high variety is used only for the written purposes. It is never used to perform any spoken function. But in the diglossic languages like Tamil and Greek, the high variety is used for both written and formal spoken purposes. But in no diglossic language, the high variety is used for ordinary conversations or for any other informal spoken purposes.

4.4. The low variety of diglossic languages does the informal spoken functions. Generally it is not used for written purposes and so it has no written form. But in Greek, one of the diglossic languages, the low variety also has written form.

This low variety is used in folk literature. In Tamil, this variety is used in written literature also, in some contexts, for some specific reasons, which will be discussed later.

4.5. So the two varieties of diglossic languages do have two different social functions. They are in mutually complementary distribution in doing their social functions. They are not used in the situations which are not assigned to them by the society.

4.6. The following table gives the functional distribution the two varieties of Tamil, the literary and spoken.

SituationThe Variety used
1. Conversation with family members, friends, colleagues, servants and other people in public places such as shop, temple, theatre etc.X
2. Speech in public meetings, Seminars, Conferences and in Assembly
(a) Person with formal education X
(b) Person without formal education. X
3. Letters
(a) Official letters X
(b) Personal letters
      (i) Person with formal education X
      (ii) Person without formal education. X
4. Classroom lectures
(a) on Tamil language and literature X
(b) cartoons X X
5. Medium of education X
6. Newspapers, Journals
(a) editorials, news, articles X
(b) cartoons X X
7. Creative literatures like novels, short stories, poetry
(a) Novels, short stories
      (i) author’s narration
      (ii) conversation among characters

(b) Poetry
      (i) traditional poems
      (ii) modern Poetry and free verse

8. Folk literature such as folk songs, drama, proverbs and idioms X
9. All India Radio and Television
(a) news broadcast X
(b) speeches X
(c) interviews with
      (i) persons with formal education X
      (ii) person without formal education X
(d) agricultural news for villagers X
(e) Advertisements
      (i) some advertisements X
      (ii) some advertisements X
      (iii) some advertisements XX
10. Films
(a) Social films X
(b) Historical films X X
(c) Purana films X X
11. Government orders, letters, circulars X

4.7. From the above table of functional distribution, it is clearly seen that for writing purposes, only the literary variety is always being used. The exceptions to this are:

1. The conversations in some novels, short stories

2. Some works of modern poetry

3. The personal letters written by somebody who has no formal education but has a knowledge of Tamil orthography.

The exception (1) may be explained that the writer’s desire to portray their characters in a realistic background is the main reason. That is, they switch over from the literary to the spoken variety as a literary technique.

Similarly for the exception (2) it may be explained that, the attitudes of the poets to the spoken variety is the reason. They consider the spoken variety as the living language and the language of common masses. So they want to write their poetry in the spoken variety.

For the exception (3) the inefficiency of the concerned persons is the reason. Since they have no formal education, they are unable to write in the literary variety.


Copyright CIIL-India Mysore