Tongue-tied in Singapore: A language policy for Tamil?

By Harold F. Schiffman


(July 02, New Delhi, Sri Lanka Guardian) The Tamil situation in Singapore is one that lends itself ideally to the study of minority language maintenance. The Tamil community is small and its history and demographics are well known. The Singapore educational system supports a well-developed and comprehensive bilingual education program for its three major linguistic communities on an egalitarian basis, so Tamil is a sort of “test-case” for how well a small language community can survive in a multilingual society where larger groups are doing well. But Tamil is acknowledged by many to be facing a number of crises; Tamil as a home language is not being maintained by the better-educated, and Indian education in Singapore is also not living up to the expectations many people have for it. Educated people who love Tamil are upset that Tamil is becoming thought of as a “coolie language” and regret this very much. Since Tamil is a language characterized by extreme diglossia, there is the additional pedagogical problem of trying to maintain a language with two variants, but with a strong cultural bias on the part of the educational establishment for maintaining the literary dialect to the detriment of the spoken one. This paper examines these attempts to maintain a highly-diglossic language in emigration and concludes that the well-meaning bilingual educational system actually produces a situation of “subtractive” bilingualism.


The Tamil situation in Singapore is one that lends itself ideally to the study of minority language maintenance. The Tamil community is small—Indians constitute around 7% of the population, but Tamils are only 4%—and its history and demographics are well known. The Singapore educational system supports a well-developed and comprehensive bilingual education program for its three major linguistic communities (Gopinathan, 1994; Gopinathan & Mani, 1983) on an egalitarian basis, so Tamil is a sort of “test-case” for how well a small language community can survive in a multilingual society where larger groups are doing well. But Tamil is acknowledged by many to be facing a number of crises; Tamil as a home language is not being maintained by the better-educated, and Indian education in Singapore is also not living up to the expectations many people have for it (SINDA report, 1991).

I will not address the second issue in this paper, but the first one, failure to maintain Tamil, is one that affects Singapore's ethnicity policy and is therefore cause for concern in many quarters. It is a fact that educated Singapore Tamils are not keeping Tamil as a home language, and have difficulty getting their children to see the value of it. Uneducated Tamils in Singapore continue to speak Tamil, but are not becoming very literate in it: but since they leave school before having to pass lots of tests, are not troubled by this. Educated people who love Tamil are upset that Tamil is becoming thought of as a “coolie language” and regret this very much.

Since Tamil is a language characterized by extreme diglossia (Ferguson, 1958; Schiffman, 1993), there is the additional pedagogical problem of trying to maintain a language with two variants, but with a strong cultural bias on the part of the educational establishment for maintaining the literary dialect to the detriment of the spoken one. This bias is so strong among Tamil speakers that one may almost be forced to conclude that the Tamil teachers, in some cases, would not be upset if spoken Tamil were somehow to be replaced totally by Literary Tamil (LT), an outcome that has never happened in any other Tamil-speaking country. In fact educated “Tamils" in Singapore are often able to speak only LT, but their mother tongue is in many cases actually English. As Gopinathan and Mani (1983) put it, this is a case of subtractive bilingualism—English replaces Tamil, and Tamil is subtracted from the equation.

Tamil language maintenance in Singapore, as in any country, is a societal issue. Schooling is usually thought of as the mainstay, the sine qua non of language maintenance efforts, so the fact that this policy is failing, even with (now) compulsory “mother tongue" instruction for Chinese, Malays and Tamils, should be cause for general alarm in Singapore, not just among Tamils. In my study of this situation, which involved a three-month long ethnographic participant-observation study of language use in Singapore, I attempted to talk to many elements of the Tamil community to try to see what factors are contributing to the subtractive-bilingualism that has already been identified.

The conclusions of this paper will be an attempt to delineate what I see as the root causes. Each segment of the community (Tamil teachers, parents, students, the Curriculum Development Institute, the Ministry of Education, the intelligentsia, the press) has opinions, sometimes strong ones, and I will deal with them as I proceed. I should also point out that I have been a student of Tamil for 40 years, a teacher of it for 30 of those years, and have concentrated my research on this language—its structure, its socio-cultural context, its grammar—for most of my academic career. I therefore had certain expectations about the Singapore situation before undertaking this study, and some strong opinions about what constitutes successful language maintenance derived from my experience with Tamil and other languages that are undergoing, or have undergone language shift.

Review of the Literature

The Tamil language in Singapore has been the subject of a number of studies, and the work of S. Gopinathan and his associates (Gopinathan, 1994; Gopinathan & Mani, 1983; Gopinathan, Pakir, Kam, & Saravanan, 1998) lays out the situation, especially as it has to do with education, quite comprehensively. The issue of societal diglossia has been examined by various authors (Britto, 1986; Ferguson, 1959), and by the author (Schiffman, 1997) as it applies to Tamil. The work of Mitchell (1985, 1986) is useful for a comparison of Tamil diglossia with Arabic diglossia. Fishman (1967/1981) provides the theoretical background for the kind of non-genetic diglossic situation that Tamil finds itself in in Singapore. LaPonce (1987) is a helpful study of the issue of territoriality and language, especially as it affects minority languages such as French in North America. Mertz (1982) provides useful background for “folk-Whorfian” ideas (or “ideologies”) about language. My own work on linguistic culture and language policy provides a theoretical background for issues of how language policy works at a grass-roots level. In particular, I saw the situation in Singapore as in some ways parallel to the German-American case that I describe in Schiffman (1987), where a diglossic language in emigration is in a Fishman-type diglossia with yet another language.

Language Maintenance and the Sociology of Language

Sociologists of language like to view language maintenance as falling into a number of different types. One kind is known as corpus planning or “treatment”; I prefer the terms corpus and status management. My original goal for my time in Singapore on a Fulbright fellowship was to establish how the Tamils of Singapore were maintaining their language in the face of a national language policy that emphasizes English education, but benignly promotes the maintenance of three other languages. Since the Tamils are known for their intense language loyalty in India and Sri Lanka, I was expecting to find that their love of the language and intense language maintenance efforts—manifested in India and Sri Lanka with strong opposition to Hindi, Sanskrit and English[1]—would result in effective language maintenance within the Singapore context. The approach taken by Tamils in India and Sri Lanka is the corpus “planning” strategy noted above. It is perceived by Tamils to be the most important kind of language maintenance, but in this day and age, and in Singapore, it may in fact have little relevance. The failure to perceive this lack of relevance may be the crux of the issue here, and this in fact the main conclusion of this paper.

Language maintenance in Tamil Nadu[2], and in contested Sri Lanka, also involves “status” planning (or management), and this involves various measures undertaken to restrict the domains of Hindi, Sanskrit and English (in Tamil Nadu), and Sinhala (in Sri Lanka) so that Tamil can recapture the domains of elementary and secondary education, the media, and so forth. This has been more successful in terms of keeping back Hindi and Sanskrit, but in the case of Sinhala in Sri Lanka, of course, a misguided attempt to reduce the domains of Sinhala by attempting to secede from Sri Lanka has resulted, of course, in a dreadful and intractable civil war. In the case of English, which is perceived in some ways as a buffer against Hindi (and Sinhala) efforts are ambivalent, and many of those who decry angilak kalappu use English and even send their children to English-medium schools. The result is that English is still the main language of higher education in Tamil Nadu; in Sri Lanka the battle to replace English with Sinhala, even in higher education, has been much more intense. In India, of course, the central government has no control over local educational policies, so no attempt to impose Hindi as a medium of instruction in Tamil Nadu universities and colleges has ever been (or will ever be) attempted.

In Singapore (and in Malaysia) language status policy is not under the control of the Tamil community, and Tamils are therefore in the position that Telugu or Kannada speakers are in Tamil Nadu: they are a tiny minority, have no say in overall policy formulation, and are forced to work within the provisions set by the larger society.[3]

Being denied the ability to affect status management, then, Singapore Tamils have been thrown back on the only area left to them, corpus management, that is, the care and feeding of the body of the language, for example, its vocabulary, its grammar, and its literature. Since this is what Tamils excel at back in the homeland, this seems a culturally appropriate thing to do, and has resulted in excessive attention to linguistic purism as a strategy for corpus management. At one meeting of a Tamil literary group that I attended, a visiting writer from India castigated the assembled group on their overzealous purism; his generation of Tamil writers in India has now tired of an earlier generation's zeal for this enterprise, and is now more interested in writing in the idiom of the people.

The local people, however, loudly disagreed; they held that the reason for the decline and fall of Tamil in Singapore is angilak kalappu, or mixing English (and Malay) words in with pure Tamil; eliminate this and all would be fine. At one point in the proceedings I was asked for my reactions to this, to which I replied that as a student of language maintenance I saw at least two other problems: one is that Tamil does not have a homeland or territory (LaPonce, 1987) in Singapore,[4] and secondly, Singapore Tamil speakers tend to not speak Tamil with their children, arguing that if people do not speak Tamil, the language will most certainly die out in Singapore. People argued back with me, saying that even when they do speak with their children, the children do not reply in Tamil, and do not learn it. They had no answer for the territory problem, which is also a domain problem. Tamil has lost almost all the domains it used to have (Sobrielo, 1985, quoted in Ramaiah 1987, p. 13), and is reduced to the domains of home and family, and then only for the uneducated.

What Kind of Tamil Do People Want?

Another issue that it is difficult to get people to understand is that language maintenance for exiled or “overseas” (emigrant) peoples cannot be restricted to the literary or educated variety, but must also be geared to efforts on behalf of the spoken language. Tamil, being extremely diglossic, has a literary variety (LT) that is not comprehensible to the uneducated, and vice versa. LT is based on a thirteenth century norm, the nannuul written by the grammarian Pavanandi. Many linguistic changes that have occurred in the spoken language (ST) since then have not been reflected in LT, so the situation is comparable to what would have happened in English if we still wrote English as Chaucer did, and maintained a formal pronunciation style of Middle English separate from the way we pronounce modern English.

In my own research on the maintenance of German in the United States (Schiffman, 1987), I attribute part of the blame for the loss of the German language, despite excellent German schools operated by different religious bodies, to the failure to recognize that German was diglossic and that the home and community domain had switched to English. This left German as a school language exclusively, and then only in religious instructions and German as a subject. (Of course German was also used in church, but my focus was on what schools attempted to do.) The parallels with Tamil in Singapore are quite striking, in my view.

In diglossic languages (Ferguson, 1959) as extreme as Tamil, the linguistic culture almost always values the older (literary, formal) variety more, viewing it as the “real” language, and the spoken variety as corrupted and “ungrammatical.” So diglossic linguistic cultures are most concerned with the corpus management of the literary language; in the late 19th and early 20th century this became a battleground in Tamil Nadu, and led to the rise of a number of political parties, one of which, the Dravida Munneetra Kazaham, gained power in Madras State in 1967.

The revolutionary zeal, the evangelical, rhetorical and oratorical skills of the DMK mesmerized a generation of Singapore Tamils as well, but the power of that rhetoric has now waned in India, as people have seen successive DMK governments fail to deliver the promised goods. In Singapore, with its small population of Tamils, corpus management is hyperpuristic, but vainly so, since the “enemy,” if we can speak of one, is somewhere else. The kind of Tamil that is taught is unfortunately a kind that has no communicative value for younger people, so as they grow up in Singapore, they cannot see any use for the language. And they are right—this variety is truly without any practical utility, and that is the fault of the Tamil establishment, which has chosen a norm that is useless for communication among live speakers.

Ironically, many meetings I attended were conducted in three varieties: LT, ST, and English, and not just for my benefit (though I, like some others, actually used two of the three). What Tamil corpus managers fail to see is that Tamil speakers, whether in Singapore or elsewhere, actually have a repertoire of languages available to them; they use LT for certain purposes, ST for many purposes, and English for yet others. The three are in a kind of symbiotic relationship; trying to deny any of them their place is detrimental to the others.

Goals for Tamil in Singapore

Another issue that is not at all explicitly clear in Singapore is what kind of goals Singaporeans have for Tamil. The stated policy goal is that language maintenance (Chinese, Malay and Tamil) will help keep Singaporeans in touch with their Asian values, the value systems brought from their home countries (and supposedly shared in Singapore) such as devotion to family, respect for elders, etc. Maintenance of these goals will supposedly act as a bulwark against the western value system which will be learned through English, and typified by such “western” social ills as materialism, divorce, domestic violence, alcoholism, drug addiction, illegitimacy, and so on. In line with this, moral education is also taught through the “mother tongue” rather than through English. The fact that English language is deemed incapable of expressing the values of devotion to family, respect for elders, etc., is a good example of what Mertz (1982) calls “folk Whorfianism,” that is, the notion that certain values (some would call them “ideologies”) are encoded in certain languages, or that certain languages are vehicles for certain values.

One analyst expressed to me the idea that this was a topic left over from the former mother-tongue educational system, and was a way to increase the number of hours devoted to mother-tongue instruction, that is, it is covertly a mother-tongue instruction policy. Other analysts have pointed out to me a number of issues that are not being discussed, such as the fact that policies in Singapore are generally elaborated in ways that satisfy first the needs and desire of Chinese Singaporeans, and the Tamils have worked within that solution to decide that hyper-archaic Tamil will be the norm; since the Chinese have successfully effected a switch from spoken Chinese dialects (such as Hokkien and Cantonese) to Putonghua (which they insist on calling Mandarin), Tamil ought to be able to use Pundit Tamil, too. Since this satisfies deeply-felt notions of what constitutes language maintenance for Tamils, the policy seems appropriate. A strategy that mimicked the “Speak Mandarin” strategy, elaborated for Chinese, might work if spoken Tamil were as diverse as the Chinese dialects are, so that Tamils would be forced to speak LT in order to communicate. Instead, Tamils tend to use English if they cannot communicate, as is the case, for example, when Sri Lanka Tamils come in contact with Indian Tamils.

Some Differences Between Tamil Nadu and Singapore

As already mentioned, there are many differences in circumstances between Singapore and Tamil Nadu:

1. Ambient language. The culture of Tamil in Tamil Nadu assures that a “bath” or all-encompassing environment of Tamil will always surround people, and that bath will provide much passive understanding of Tamil, and support for Tamil. In Singapore there is no Tamil “bath”; the ambient language is English, and that is crowding Tamil out of its traditional domains, leaving it only with the home and religious domains.

2. Literacy. In Tamil Nadu, it is not demanded of all adults that they master LT; census figures state that not more than 50% of Tamils are literate, or only functionally literate, for example up to 4th standard level only. The level of Tamil literacy expected of them is not as high as in Singapore; if people in Tamil Nadu wish to study archaic or Sangam Tamil they can go to university, but in Singapore there is no university level Tamil, so it is forced on the A-level people, who have to show proficiency in it to go on to higher education.

3. Registers, varieties, levels. People in Singapore fail to recognize that Tamil is, because of its ancient traditions, actually a language with many more levels than, for example, English. In English, we usually expect all high-school graduates to be able to read Shakespeare, but we do not expect them, or even all college graduates, to prove active mastery of Shakespeare. That is, we do not expect Americans or British or Australian students to be able to write sonnets or plays using Shakespeare's language. We only require a passive understanding of his writings, and even attaining that is not easy. Certainly most literate Americans and British speakers cannot read Middle English (e.g., Chaucer) without a dictionary or a grammar, but Singapore expects its Tamils to do the Tamil equivalent of that. And Old Tamil (dating from the 2nd-3rd century CE) is also treated as if it is part of the expected competence of A-level students. Though we do have high-school seniors read Beowulf, they read it in a modern translation, and we would never expect anyone but historians of English to learn Old English (which only goes back to the eighth century, not the second century, like Sangat tamiR).

In effect, Tamil consists of at least five different styles or varieties of language; we can distinguish at least the following:

1. Spoken Tamil (peeccu tamiR), spoken and understood by all speakers;

2. Modern LT, the language of social novels and short-stories; educated speakers can read this and write it.

3. Older modern LT, the language of Pavanandi (thirteenth century); educated people understand it, and some people can generate accurate specimens of it.

4. Medieval Tamil (the Alvars, TirukkuraL), and

5. Old Tamil ( sangat-tamiR), which cannot be understood actively by anyone without the commentaries and dictionaries.

Some people would add another level, that of the epic literature, that is, silappatikaaram and maNimeekalai, which comes between Medieval and Old Tamil, and since a hyperpuristic style incorporating material from a number of stages of the language, cultivated by pundits and taught in secondary-schools by Tamil teachers also exists, we must also take “Pundit Tamil” into consideration even though it is artificial and hypercorrect; it typically mixes material (lexicon, morphology) from various stages of the language. One must also note that most literary norms of the past are poetic norms, not norms intended for writing prose. The only prose writing that existed before the modern novel came into existence was that used to expatiate and comment on poetry; it was in itself not thought of as literature. Scholars and linguists therefore differentiate these different stages and forms of the language, but as noted, pundits do not, mixing various levels at will, and ignoring only the spoken styles, which they consider corrupt and degenerate.


Pedagogically, we know that children need to develop their literacy by linking it to their mother tongue skills, which in Tamil are the spoken dialects. If this is not done, most children will have problems with literacy. I would suggest that the low literacy rate among Tamils in India is related to the extreme diglossia in Tamil, but since there are other languages in India with low literacy (e.g., Hindi and Oriya) but without such extreme diglossia, this claim cannot be substantiated. But if we compare literacy in Tamil with the literacy rate in Malayalam, which is the highest in India, we must also note that Malayalam is much less diglossic than Tamil. Children need to learn to be confident of their mother tongue, basing their knowledge of the standard language on their oral skills, and gradually acquiring literate norms and distinguishing between literary and colloquial norms. Primary school Tamil pedagogy in Singapore schools has made great strides to improve this kind of teaching, but it seems to taper off as Cambridge O-level and A-level exams approach.

Even in English, of course, there are colloquial, informal varieties that are not used in print (except in comic books, certain kinds of sensational journalism, etc.) A spoken form like gonna, used by millions of people daily, will not replace “going to” in print, and a lexical item like kids, known by all as an informal word for “children,” will not appear in print in America (at least) unless it is a direct quote of someone speaking. But it will not appear in serious writing, that is, not used in the expository or narrative or descriptive modes of the language. Children in most literate cultures gradually learn to distinguish between these; it is what education prepares them for. But if children (especially in diglossic linguistic cultures) are told that all spoken language is corrupt and bad, they will have no confidence in their own native ability, and they will begin to treat it as useless and valueless language. Certainly they will not be able to link their knowledge of peeccu tamiR to LT, and almost certainly their writing will lack creativity, spontaneity, and innovativeness, and will be no more than a lifeless parroting of dead norms.

In Singapore, some students demonstrate facility in the memorization of these literary forms, in the same way that elite 19th century students in Britain could manipulate classical languages easily. But many Singapore Tamils put this language behind them once they get past the hurdles of Cambridge A-level tests. As things currently stand, no Tamil child can pick up a Tamil children's book and spontaneously read it, since all children's literature is written in a style that is not comprehensible to an ordinary speaker.

The irony is that Singapore's national schools have such high standards that even students who do not make it into the top 8% can easily enter university in other English-speaking countries. It is more expensive for them to go abroad, because National University of Singapore (NUS) is cheaper, but if they are excluded from entering NUS or Nanyang Technological University (NTU) because their language scores are not quite adequate, this is no disadvantage in the real world. The world of science and technology, as we know only too well in the United States, has no use whatsoever for any knowledge of “foreign language.” Engineering colleges in the United States and elsewhere do not require their students to know any language other than English, and even that standard is very low.

Thus Singaporeans who do well in other subjects but fail to achieve high standards in their Cambridge A-levels in language can easily, though not inexpensively, go abroad for further education. The evidence is that boys do less well than girls at this level, and since this requirement was instituted, a higher percentage of young women have been able to enter NUS than young men. It is interesting to note that in Malaysia, since seats are reserved on a quota for each ethnic group, and not released if a particular ethnic group does not fill its quota, Chinese and Indian Malaysians are left on the waiting lists if their scores are not high enough, though there may be empty seats.[5] So these groups now prepare actively for the eventuality that they will have to go abroad for their higher education, and the newspapers are full of advertisements for programs that allow people to finish a B.A. degree in Australia, England, Canada, the United States, etc. These students are in fact preparing to permanently expatriate themselves, since their foreign degrees are not valued as highly as Malaysian degrees in many cases. Singaporeans may be following the Malaysian model, in effect—for the eventuality of not getting a high enough pass on language in A-levels, they prepare to permanently leave the country.

Overt and Covert Aspects of Policy

In my work on language policy (Schiffman, 1996) I view language policy as not only consisting of the overt and explicit aspects found enshrined in the laws, constitutions, or administrative codes of a policy, but including implicit and covert factors of policy as if they are all part of a larger and complete picture.[6] I consider all these factors as integral elements of the policy, since otherwise there is a tendency (both among policy-makers and students of language policy alike) to weight the official policy more heavily than the unofficial policy, and then treat the unofficial policy as some kind of “vexatious” or “pesky” resistance to the “real” (i.e., official) policy, or as some unintended consequence that has to be “fixed” by the real (official) policy makers. Under such an analysis, I see the language policy of Singapore, unfortunately, as overtly anti-Tamil. It is anti-Tamil both at the highest levels, that is, at the level where policy is set, and at another level, at the level where policy about which kind of Tamil will be taught is decided. Presumably, government decision-makers (the Chinese-descent oligarchy) are unaware of what is happening internally in Tamil policy. They are perhaps informed that the best kind of Tamil is being taught, and that high standards are being set for examinations, and they are, one can assume, satisfied with this. But a more cynical interpretation of this would be that this policy, covertly and perhaps only unwittingly, will marginalize Tamil speakers by keeping them out of the action. The action is where English-educated technocrats at the top of the pyramid wheel and deal and make decisions for the rest of society; or, another sphere would be where bilingual Chinese and English educated elites make the decisions about things. In either scenario Tamils spin their wheels learning a language that is of no practical value, leaving the terrain free for others (and this is not Malays) to run things.

At the highest level, the policy appears to be pro-Tamil. However there are some other policies, not specifically language policies, that contradict the ostensibly “pro-Tamil” policy. One of these is the housing policy, which seeks to disperse Tamils and Malays throughout Housing and Development Board (HDB) housing “estates,” and within these estates, to disperse them widely on different floors, so that no concentration of Indians or Malays may arise, however innocently or accidentally this may occur. This dispersal policy is reportedly designed to prevent racial antagonisms, or even racial violence, from arising, the idea being that if Malays and/or Indians are concentrated anywhere, they might rise up and commit violence, whereas if they are dispersed, they cannot easily band together to do this. The negative side of this, however, is that Tamils are so dispersed that Tamil ceases to act as an intra-ethnic language in the Tamil community. Tamil ceases to have a community domain, and is only spoken, if there, within the home. For Malays, the proximity of Malaysia and Indonesia is supportive enough of their language that they can easily perceive the utility of maintaining their language. They only need to cross the causeway and they are in Malay language territory. But Tamils do not have a home territory anywhere within their horizons; in the old days, before HDB estates, areas such as Sembawang and Serangoon Road provided at least some home territory for Tamil, but this has ceased to be the case. Thus the housing policy of minority dispersal has the covert, and perhaps unintended effect, of penalizing and marginalizing Tamil.

Another aspect is that because Tamils are such a small minority, some schools do not have enough Tamil students to offer full programs. Tamils must then take buses to other schools where there is a critical mass of Tamils, and this then creates the perception of “outsiders” coming in to get “remedial” programs, rather than being an integral part of each school. This is also unintended, but it is a perception that busing can create, as we know from U.S. programs as well.

Parallels are often made between the policy of encouraging Chinese Singaporeans to speak Mandarin and forget dialect, and in fact a major language shift has taken place in ethnic Chinese homes—many Chinese now report (En, 1990, p. 163) that Mandarin is their “predominant household language” (but many also report that English is their “predominant household language,” and not Chinese or dialect.) Mandarin is of course the language of instruction in Chinese language classes, and the importation of Cantonese or other-language soap operas and television programs is now discouraged. This insistence on a high standard of Mandarin, however, differs in some details from the Tamil model, since Mandarin is the official language of the People’s Republic of China (PRC), and has great utility. It is in effect a mother tongue for hundreds of millions of Chinese; the kind of hyper-archaic literary Tamil being required for A-levels in Singapore Tamil classes is not anybody's mother tongue, and ceased to be some centuries ago.

It could actually be argued that the kind of Tamil being taught is not in fact hyper-archaic, because it contains many words that did not exist in the past. This claim is in fact to some extent true, as many new words have been coined to serve as replacements for English words. I refer to this kind of “fake” Old Tamil as “archaistic,” that is, it “looks archaic” because it uses old roots but is not really old. This kind of corpus management is known as “purism,” that is, forming words from what is perceived to be “pure” Tamil sources and roots. I wish to argue that purism is, in the case of Tamil at least, counterproductive. The hyperpuristic style was a strategy developed early in this century in Tamil Nadu, beginning first as a hyperarchaic style (which uses only words from older sources, such as illam for “house,” instead of viiDu or some other more natural form), but then was adapted for coining new words, that is, neologisms, for things that never existed in the past. In some ways this might be termed hyper-pseudo-archaic, since there is a certain falseness to this supposed ancient style; another might be simply epicurean, that is, concerned primarily with a certain style. I prefer the term “archaistic” because it encapsulates best the notions found in the term “hyper-pseudo-archaic.” Thus in my HDB building, the English word “lift” (“elevator”) is referred to by one term lipTu on a sign on the outside, telling people in four languages what to do if there is a breakdown, and as a mintuukki on the inside, telling people, again in four languages, that the lift is under “closed-circuit television surveillance” (uL amaippu kaamira plus a verb meaning “surveille.”) I am unaware of what the official term for “lift” might be in Tamil Nadu, since I have never seen a Tamil sign on any lift there—the assumption there seems to be that if one is using the lift, one knows what to do with it, and what it is called. Notice that the lift is given a translation mintuukki which I take to mean “up-lifter” whereas the word kaamira from English “camera” is used without impunity, for reasons I am not able to fathom. If one travels around Singapore and reads the signs, one finds great inconsistency in the terminology, which I would think might be frustrating. This tendency to constantly tinker with terminology is something that is typical of hyperpuristic language planning, but is probably counterproductive for a language—one should settle on one form and eschew all others. The attitude is paternalistic, that is, “the government knows best” but this is also typical of Singapore—it is a kind of arrogance that is counterproductive for language maintenance.

An Example of This From Tamil Nadu.

The English word “television” has five (six, if you count the Singapore word) equivalents in Tamil. That is, Tamil speakers may use any one or several of the following words to refer to “television”:

1. duurdarshan is the term used, even in Tamil, for government television broadcasting originating in New Delhi. When announcers say itu duurdarshan one knows that they are saying “This is the Indian government television broadcasting system.”

2. tolaikaaTci is the term used for television broadcasts originating in Madras. Another term, tolainookki, may also be used, and is usually the term used for television in Singapore.

3. “Television,” pronounced TelaviSan, is used by many educated people, often to refer to the concept of television, even when speaking Tamil, for example, “TeleviSion vandadukku munnaale, reeDiyoo taan irundadu” (“Before there was television, there was only radio”).

4. “TV,” pronounced Tiivii, is also used widely in spoken Tamil. Example: “naan ungaLe Tiiviile paatteen” (“I saw you on TV”).

5. paDapoTTi, used by uneducated people, is derived from paDam “picture” and poTTi “box,” and is probably modeled on another older similar item, paaDa poTTi “song box,” that is, “radio.”

Thus Tamil has at its disposal at least five words. One, from Sanskrit, is never used except to refer specifically to the all-India government television system, as originating in Delhi. The second item is only used in the speech of television broadcasters, and would be the written form as well. Three and four are of course borrowed. But the last item comes from the spoken resources of the language, and though somewhat puristic-sounding, is obvious to any speaker of Tamil as to its meaning. The others would not be clear to people without education.

In any event both one and two are loan-translations or calques from the international terms: tele- is the Greek word for “distance” and both duur- and tolai convey this; vision is from Latin, and darshan and kaaTci/nookki are the Sanskrit and Tamil equivalents; the Tamil also conveys, by the ending -i on nookk-i (and kaaTci) that an instrument or thing is involved. It is not a “look-ing” but a “look er” or a “thing for looking.”

This is also present in mintuukk-i “thing that lifts,” or “thing for lifting.” My point is that Tamil does not need five words for television. If it is going to try to displace English or Sanskrit, it needs at most one and if the language is going to be useful to people, it needs the one that is most obvious and natural and acceptable to the most people in its meaning, even if it does not come from the most ancient source. paDapoTTi is a “pure” Tamil word but it is from a source that is neither archaic nor deliberately puristic, the natural resource of the living spoken language.

One of the aspects of the evolution of new vocabulary in any language that I know of is that the words that succeed and prevail tend to eventually be short and easy to use. Much of new English vocabulary, for example, starts out as a longish term of some sort, for example, “television” and is then reduced to “telly” (in Britain) or “TV” in North America. The word “fan” meaning person avidly interested in sports or entertainment, comes from fanatic and is a reduction or shortening of that term. (Tellingly, the Tamil word for “fan” is a mis-translation of [electric] fan, i.e., viciri). But Tamil purists do not allow Literary Tamil to use resources such as shortening, or acronyms, despite the prevalence of English acronyms in general use in the language: MGR, HDB, MOE, COE, etc. In South India there is a college known by its initials OCPM.

This has in fact evolved into a “word” oociipiyam and even obeys the morphophonemic caariyai rules for words ending in am, that is, oociipiyattule “in/at OCPM (college).” In English (and other languages) the taking of pieces from several words (not just letters) to make new words is known as “blending.” Such forms as “smog” (from smoke and fog), “modem” (from modulator-demodulator) are known as blends, and they are very common in Japanese, Russian, Spanish, Indonesian, and many other languages. Computer terminology and consumer electronics abound with these things: VCR, PDA, HDTV, DOS, WYSIWYG. GIGAFLOPS, etc. In India the word “kilometre” has been borrowed in various languages; in Tamil it is kiiloomiiTTar often abbreviated on signs as kii.mii.

In some areas this has been taken as a word, and the blended form kiimii now stands for kiiloomiiTTar at least in the speech of uneducated people. This seems to me a laudable and natural development, and should be encouraged, because shortening and abbreviating is what seems natural in all the other languages I know. Russian does it, Malay/Indonesian does it, French does it, modern Hebrew does it, English does it. If this process is not allowed to happen in Literary Tamil, how will minsaravacadi (for “electricity”) ever replace karaNTu? How will mintuukki (elevator) ever be expected to replace lipTu? How will tuunduvaNDi (bus) manage to replace bassu? Language planning edicts issued from on high will not make the longer terms succeed.

There are those that argue that Tamil should be open to borrowings from other languages, and that it will not survive if it does not borrow. In fact Tamil does borrow, but borrowings are only common in peeccut tamiR. There are other resources, such as those already mentioned, that is, the natural resources, such as those that produced paDapoTTi. Literary Tamil needs to open itself to these resources—acronyms, blends, shortenings, initials—all of which are natural sources for new words in many languages of the world.

The point for Singapore is that Tamil should meet the needs of Singaporeans, not be designed to fit some idea of Tamil brought from India years ago, or to be pleasing to Tamilians everywhere. English is not the same everywhere, Tamil should not try to be either. Unfortunately, it does not seem that Tamil will be given the freedom to adapt and breathe that it needs to survive in Singapore; instead, it seems that the Tamil language-maintenance policy it currently has is actually a Tamil annihilation policy, because if the language is not allowed to live and breathe, its speakers will desert it; this is a process we already see happening.

Who is to Blame?

As I traveled around Singapore and visited schools, watched teachers teach, watched arts contests and speech contests, and talked to all kinds of people involved in the enterprise of keeping Tamil alive, I was struck by the good will and good intentions of everyone. Many people are trying very hard, and much good can be seen. In particular younger teachers trained recently in modern methodology are doing an excellent job: the Language Experience Approach (LEA) being used in many elementary classes focuses on the task of developing writing abilities of children. Based on the observation that children like to talk about what they have done, LEA makes them write about it. First they are asked to talk about it and then they are guided to write, constructing a story; an interested adult helps, and this models appropriate writing strategies and conventions. The teacher writes down their thoughts, expressed in their words; children acquire the necessary writing skills. The teacher writes the finished story (Class Dictated Story, CDS) on the board.

In these and other new approaches, I would not change a thing in their training. Everyone is doing the best job they know how to do, with the best will and the best of intentions. However, when asked what people see as the problem with Tamil, there is a certain amount of blaming that starts up:

Those people are responsible; it’s the Ministry; it's the Curriculum Development people; it's the teachers, they're too lax; it's the parents, they don't speak with their children; it's the children, they don't appreciate Tamil culture.

Some elements of the system can get quite defensive, or just indicate that they are following guidelines set by others. What I see is that there is lack of agreement between policy-makers (MOE (Ministry of Education), CDIS (Curriculum Development Institute of Singapore), NIE (National Institute of Education), teachers) and consumers, that is, users of the language. Parents and children are not happy, but they feel powerless to change anything. Or, some things have changed, but it may be in some cases too late, that is, some children are already English speakers; their needs are not being met, and here I speak with authority, having taught Tamil to English speakers for 30 years.

The most obvious problem to the outsider is that there is no clarity of goals—how much Tamil, what kind of Tamil, Tamil for what purposes? The government policy, which is to encourage the maintenance of traditional [Asian] culture and values (Gopinathan & Mani, 1983) is clear. The interpretation of what “Tamil maintenance” means, is set by the MOE and the CDIS. Unfortunately, the curious position of Tamil and its diglossia as a minority immigrant language, and the special problems inherent in that situation, are not taken into account.

The strategy of maintaining a puristic Tamil within the domain controlled by the educational establishment is thought of as the only appropriate one within this situation. This is despite the serious lack of congruence existing between corpus management and status management. The only recognition that there is a domain problem is that some teachers and others seem to have decided that since the school is the domain of Tamil, and there is only one kind of Tamil worth preserving, then only “pure Tamil” (tuyyat tamiR) will be allowed in school. They have done what even in Tamil Nadu is not done—they have chased out spoken Tamil from the school. Thus it is no surprise that many young Tamils do not speak Tamil; their competence in spoken Tamil is constantly denigrated, and any confidence they might have is systematically destroyed, or was until recently. How then can we expect any other outcome?

Does Anything Need to be Changed?

Though we now see much use of spoken Tamil in primary education, we do not see any active use of it either in teaching materials or in students’ active writing of it. But my suggestion that spontaneity and creativity will not come in Tamil unless it is used as the first language of instruction with the following objection: “There is no standard spoken language for Tamil; even if we wanted to we could not write it."

Since I have dealt with this issue elsewhere, I will not elaborate further here, but refer the reader to my grammar of the spoken language (Schiffman, 1999). But a great deal of ink has been spilled over this issue in the United States and elsewhere, and I will not attempt to recreate the arguments. The one thing I notice is that children's literature in Singapore, even for the earliest grades, is written in the hyperpuristic style, such that one book on the shelves of the NIE library, Curriculum section, begins a story for children, a story about some children who gather some stones and rocks and build a wall to keep out monsoon rains, with the Tamil words karkaL irukkinrana. Neither of these words (meaning “There are (some) stones”) is a spoken form, and neither would be recognizable to a child knowing only spoken Tamil; in ST the form would be kallu irukku.

As mentioned earlier, no Tamil child can simply pick up such a storybook and just read it. It seems that the goals of the Tamil policy, being unstated, are therefore covert goals; a very cynical view (I shall propose a more optimistic set below) can be described as follows:

1. Students: The purpose of maintaining Tamil language in Singapore is to pass the O-level and A-level tests administered by the Ministry of Education, in order to get 5 A-level passes and gain admission to university. The minimum pass required for Tamil is C6, and the amount of time and energy expended to gain a pass will be equivalent to whatever it takes to get a C6. When entrance to university is gained, Tamil no longer has any economic value, and may be discarded; no further use will be made of it.

2. Teachers: The purpose of the teaching of Tamil in Singapore is to provide jobs for teachers, whose purpose is to enable students to get 5 A-level passes and gain admission to university. Lip service is also paid to maintenance of Asian family values and the teaching of moral education.

3. National Institute of Education Tamil Teacher Training. The purpose of having Tamil taught in Singapore is to make it necessary to train teachers to teach Tamil; at the lowest levels teachers are encouraged to help children to speak Tamil using recognized techniques such as cooperative learning, interactive learning, etc. At the higher levels the focus is on enabling students to get 5 A-1 passes and gain admission to university. Lip service is also paid to maintenance of Asian family values and the teaching of moral education.

4. CIDS and MOE: The purpose of the teaching of Tamil in Singapore is to maintain “pure grammatical” Tamil. The task of the CIDS (Curriculum Development Institute of Singapore) and MOE (Ministry Of Education) is to set curriculum and prepare teaching materials for the use of teachers and learners of Tamil; at all levels the emphasis is on hyperpuristic Tamil; lip service is paid to the notion of encouraging children to speak, but no testing of this ability will ever be required, and no materials to encourage this goal will be provided. At the higher levels the goal is to make it as difficult as possible for students to get 5 A-1 passes and gain admission to university. The teaching of Asian family values and of moral education is didactic and authoritarian; no recognition of situational difficulties or use of interactive teaching techniques is encouraged.

5. Parents: Parents' goal regarding Tamil is to have their children have as much appreciation of Tamil language and culture as they themselves have, but not more (This usually means the children will have less appreciation of Tamil language and culture than their parents). Whenever the goal of attaining 5 A-1 passes and gaining admission to university conflicts with the previous goal, the latter goal will be take precedence, though it is generally recognized to be at cross-purposes with the former.

6. Writers and Intellectuals: The purpose of Tamil language maintenance in Singapore is to provide job opportunities for people who are otherwise unemployable.

A More Optimistic View, and Some Recommendations

I have put the above covert goals as cynically as possible in order to bring out how conflicted the Tamil language maintenance situation really seems; some of these views are in fact held by members of the Tamil community, or are attributed to others. My own recommendations would be:

1. Introduce into the equation the notion that knowledge of a literary language is necessarily connected to the knowledge of a spoken language, and that the literary language will not survive without knowledge of the real mother-tongue (the language first learned by the child, and spoken before schooling begins), that is, Spoken Tamil. Recognize that the teaching of Tamil has become like the teaching of other classical languages such as Latin and Sanskrit. Though they are interesting and beautiful languages, and contain great riches, their connection with the “mother tongue" is lost.

2. Introduce the notion that Spoken Tamil is a rule-governed form of language, and is actually quite regular, uniform, and standardized. It is a powerful resource, the vehicle of much creativity in the Tamil film, radio plays, and television dramas. It is understood wherever Tamil is still a living language, and Singapore ST is a mainstream dialect of it.[7]

3. Literacy in Tamil should begin by taking advantage of the knowledge of their mother tongue the students bring with them to school. It should be treated as a resource instead of a hindrance. Earliest literacy, such as in the new kindergartens being proposed, should be through the medium of Spoken Tamil. Spontaneous ability to read texts, create texts, and manipulate language can only rest on this knowledge. Earliest reading should prefer sentences like kallu irukku “There were some stones” over karkaL irukkinrana.

4. Decisions can be made about the best point to make a transition to the grammatical forms of LT, whether it is in P1 or P4 (primary school grade 1 or 4). But forms like karkaL irukkinrana should never be the first things novice readers are presented with.

5. Teacher trainees may need to be taught an explicit knowledge of the structure and syntax of ST, since their knowledge of it now is just implicit. They need to be taught that it is regular, rule-governed, and systematic (Schiffman, 1999) and some explicit sociolinguistic training about how the two varieties coexist in the Tamil world probably needs to be made part of the curriculum. Typically now, Tamilians are unaware of how much peeccu tamiR is actually used, even as a medium of instruction, or as a language of “explanation” in schools and colleges. A systematic linguistic introduction to the structure of the target language is an integral part of foreign language teacher-training in the United States, and since many Tamil’s are now not mother-tongue speakers of Tamil, the question of how to teach Tamil as a second-language needs to be explictly dealt with.

6. Something needs to be done to stimulate creative writing for Tamils, and a way to start might be to introduce to Singapore Tamils examples of Singapore Tamil writing. The whole language approach to literacy, which emphasizes reading real language rather than the stilted language of textbooks, would utilize Singapore writing as an example. Great efforts need to be made to find examples of Tamil writing (by real writers) that are appropriate for children. My own first exposure to a real (but simple) Tamil story as a novice learner was Pudumaipittan's Teru ViLakku. Surely other examples of simple but good written Tamil prose can be found.

7. Convene a study group of representatives of the NIE, the Tamil Teachers, the Ministry Of Education, the Curriculum Development Institute of Singapore, recent former students, and parents. Agree to work toward consensus on what the goals and purpose of Tamil maintenance might be, without blaming, finger-pointing, and name-calling, and how best the needs of Tamils in Singapore can be met. Choose a moderator who has no axe to grind.

A Final Note: Egalitarianism and Equality

Much is made in Singapore of how policies are egalitarian, especially the ethnicity policy. But an example from the literature on North America, French in Canada, may be apt. In French Canada (and in adjacent parts of New England) the French language is spoken by a minority of about 5 million, surrounded by a sea of 270 million English speakers. In that situation, the Québécois express the feeling (LaPonce, 1987) that legal egalitarianism is simply not enough, and have attempted to legally restrict and diminish the domains of English within the French-speaking territory.

Though this enrages many English Canadians, who feel that egalitarianism is what the law requires, and is inherently fair, the francophones, however, contend that egalitarianism is not equal, because it does not lead to equal outcomes, but in the case of Canada, to English dominance.

In this situation, the only way to guarantee equal outcomes, that is, that French speakers will be French-dominant bilinguals, is to create a “safe haven,” a reserved space for French, so that the overwhelming dominance of English can be kept at bay.

Similarly in Singapore, the egalitarianism seems to exist only on paper; the outcome of the policy has not led to the strengthening of Tamil, and the housing policy has guaranteed that no territory for Tamil will exist. In the final analysis, egalitarianism is not equality if one group is ten times the size of the other, whether in North America or in Singapore.

Britto, F. (1986). Diglossia: A study of the theory with application to Tamil. Washington, DC:

Georgetown University Press.

En, L. K. (1993). Singapore census of population 1990. Literacy, Languages Spoken and

Education. Statistical Release No. 3 Singapore: SNP Publishers Pte Ltd.

Ferguson, C. (1959). “Diglossia.” Word, 15, 325-337.

Fishman, J. (1981). Societal Bilingualism: Stable and Transitional. In J. Fishman

Sociolinguistics: a Brief Introduction. (pp. 78-89). (Originally Published in 1967 as Bilingualism with and without diglossia; Diglossia with and without bilingualism. Journal of Social Issues, 23(2), 29-38).

Fishman, J. (1981). Sociolinguistics: A brief introduction. Rowley, MA: Newbury House.

Gopinathan, S. & Mani, A. (1983). Changes in Tamil language acquisition and usage in

Singapore: A case of subtractive bilingualism. Southeast Asian Journal of Social Science, 11(1), 104-116.

Gopinathan, S. (1994). Language policy changes 1978-1992: Politics and pedagogy. In S. Gopinathan, A. Pakir, H. W. Kam, & V. Saravanan (Eds.) Language, society, and education in

Singapore: Issues and trends. Singapore: Times Academic Press.

Gopinathan, S., Pakir, A., Kam, H. W., & Saravanan, V. (Eds.) (1998). Language, society, and education in Singapore: Issues and trends. Singapore: Times Academic Press.

Laponce, J. (1987). Languages and their territories (A. Martin-Sperry, Trans.). Toronto, Canada: University of Toronto Press.

Mertz, E. (1982, July). Language and mind: A Whorfian folk theory in United States language law. Sociolinguistics Working Paper, No. 93. Austin, TX: Southwest Educational Development Laboratory.

Mitchell, T. F. (1985). Sociolinguistic and stylistic dimensions of the educated spoken Arabic of Egypt and the Levant." In J.D Woods (Ed.), Language standards and their codification: Process and application. Exeter Linguistic Studies, 9, pp. 42-57. Exeter, United Kingdom: University of Exeter.

Mitchell, T. F. (1986). What is educated spoken Arabic? International Journal of the Sociology of Language 61, 7-32.

Ramaiah, K. (1987). Linguistic difficulties encountered by secondary four Tamil students in their Tamil compositions. Unpublished M.A. thesis, National University of Singapore.

Schiffman, H. F. (1987). Losing the battle for balanced bilingualism: The German-American

case. Language Problems and Language Planning, 11(1), 66-81.

Schiffman, H. F. (1996). Linguistic culture and language policy. New York: Routledge.

Schiffman, H. F. (1997). Diglossia as a sociolinguistic situation. In F. Coulmas (Ed.), The

handbook of sociolinguistics. London: Basil Blackwell.

Schiffman, H. F. (1999). A reference grammar of spoken Tamil. New York: Cambridge

University Press.

SINDA Report. (1991). At the crossroads: Report of the Action Committee on Indian Education.

Singapore: Action Committee on Indian Education.

Sripathy, M. (1998). Language-teaching pedagogies and cultural scripts: The Singapore primary

classroom. In S. Gopinathan, A. Pakir, H.W. Kam,& V. Saravanan (Eds.), Language, society, and education in Singapore: Issues and trends. Singapore: Times Academic Press.

Author Note

Research for this paper was carried out under a Fulbright Research grant from the Council for International Exchange of Scholars/USIA in early 1994. I would like to express my appreciation to the faculty and staff of the Department of Asian Studies, the National Institute of Education, Nanyang Technological University, Republic of Singapore, for their welcoming assistance.



[1] The current antipathy is strongest against Hindi and is known as Hindi etirppu.

[2] With the exception of the word Tamil Nadu, which already has an English spelling, Tamil words will be transliterated with the following conventions: capital letters represent retroflex sounds, and double letters represent long vowels.

[3] One of the great weaknesses of Indian language policy is the toothless provisions for language groups who live in territories where they are in the minority. It is fine to be a Telugu speaker in Andhra Pradesh; it is not so fine to be one in Kerala, Karnataka or Tamil Nadu, and the constitutional provisions to protect such groups are noticeably difficult to enforce. Each linguistic state, having driven out the perceived oppressor and established its own linguistic regime, turns out to be an even more ferocious oppressor of its own linguistic minority groups.

[4] Singapore's housing policy has expressly worked to disperse the Tamil and Malay population as evenly as possible in the Housing Estates all over the island; the result is that the 7% of Indians never end up in any concentration anywhere in Singapore, so the domain of community, usually tied to a region or even part of the city, is gone completely.

[5] The analogy that occurs to me is an airline that assigns seats by ethnic quota, and flies with empty seats rather than fill them with members of other ethnicities!

[6] I refer to this larger picture as the “linguistic culture,” which is the sum totality of ideas, values, beliefs, attitudes, prejudices, myths, religious strictures, and all the other cultural “baggage” (some writers refer to this as “ideology”) that speakers bring to their dealings with language from their culture. Linguistic culture also is concerned with the transmission and codification of language and has bearing also on the culture's notions of the value of literacy and the sanctity of texts.

[7] The notion that spoken languages that are in a diglossic relationship with literary languages, themselves of course quite standardized, might also exhibit a kind of informal standardization is not generally recognized. The Tamil situation is probably most comparable to that of another highly diglossic language, Arabic (Mitchell, 1985, 1986.)

[South Asia Studies University of Pennsylvania in Journal of Language, Identity and Education, 2002, Vol. 2, No. 2, pp. 105-125 ]
-Sri Lanka Guardian