Switching on ‘Trilingual Competence’ without learning languages!

Chandre Dharmawardana, an original article*

The proposed “trilingual” vision for Sri Lanka, hopes to make every citizen acquire at least a good working knowledge of English, Sinhala and Tamil. This is the opposite of the “Indian model” where, e.g., in Tamil Nadu, Tamil is the only language ‘recognized’ by the state, even with substantial non-Tamil minorities. One ‘justification’ for trilingualism is the view that ‘language Politics” caused communal strife and Eelam violence in post-independent Sri Lanka. There were other, deep systemic reasons for the strife. However, this essay examines the issue of trilingualism and how it can be cheaply and rapidly implement via a technological solution by accepting the fact that most people will not learn three, or even two languages. We argue the following: (i) A form of trilingualism can be provided rapidly and cheaply via available information technology without everyone learning the other two languages. (ii) Attempts at trilingual competency using an educational system already burdened by ‘tuition’ would not succeed. (iii) Even inCanada, after four decades of effort at bilingualism, 80% of the people are unilingual. (iv) Sri Lanka’s effort should be directed to creating interest in the other linguistic and cultural heritage of the land.  (v) The incorporation of automatic translation at the level of business and social interactions into cell-phone conversations or text-messages is eminently feasible and opens up the ‘language barrier’. 

Thus each person conducts social and business transactions in their most comfortable language, while the technology provides an instant translated text or voice interface.

I. LANGUAGE AND POLITICS: The Presidential Secretariat has invited the public to send their comments and suggestions to be considered for inclusion in the Draft 10-year National Plan for a Trilingual Sri Lanka (2012-2021). This will be launched by President Mahinda Rajapaksa later in the year [1].  We learnt about this from a news item on the 9th of October. It also stated that the closing date for receiving suggestions is 12th October!

Although the “Sinhala Only Bill” is often cited by superficial observers as the “cause” of communal strife in Sri Lanka, historians have been skeptical. For instance, the British historian Dr. Jane Russell has examined communal politics in the Donoughmore Era, and she documents details about the first Sinhala-Tamil riot which took place in 1939. The riot was rapidly squashed by the British administration although it spread from Nawalapitiya to many other towns [2, 3]. The Ilankai Thamil Arasu Kadchi (ITAK) was formed in 1949 and the Manifesto already contained the doctrine of “exclusive Tamil homelands”. The details of the rise of Tamil nationalism and land claims have been discussed by Roberts [4], Gerald Peries [5] and others. The Tamil publications of the ITAK seem to have called for driving out the Sinhalese and Muslims (‘invaders’) from these ‘homelands’ [6], long before the ‘Sinhala Only’ bill of 1956. Thus an extremely militant Tamil Nationalism already existed prior to the 2nd World War and crystallized into the ITAK.

 Some politically perceptive writers, e.g., Sebastian Rasalingam [7], Gunadheera [8], Ladduwahetty [9] and Thomas Johnpulle [10] have discussed the issue of caste and land ownership in this context. It is argued that a deep reason for the Tamil-Nationalist separatist movement, beginning from 1929 onwards, was tied to the interests of the land-owning upper-caste Tamils who lived inColombo. Today they control the Diaspora. They regarded the North as their private fief, sought to maintain their hold on land, and retain the use of the virtually free services of the lower castes that they had traditionally enjoyed. Within this view, ‘Federalism’ (or devolution) is, at best a mere milepost to separation. Thus it is claimed that the upper castes sought to include the caste system into the constitution, opposed universal franchise, opposed modernization of roads, causeways, education, socialism, and all reforms that weakened the hierarchical society of the North. While this may be putting too much weight into one argument, such factors cannot be ignored in understanding the origin of racist strife inSri Lanka, just as one-sided explanations based purely on linguistic chauvinism have no solid basis.

Be that it may, the 1956 “Sinhala Only” bill did provide the spark for the ITAK to launch civil disobedience movements and fire the Tamil imagination. The violent response of influential sections of the Southern Polity and ‘tit for tat politics’ led to a series of riots, escalating into Eelam wars supported by a wealthy Diaspora.  The thesis that ‘if the Sinhalese spoke Tamil, and if the Tamils spoke Sinhala, there would be no strife’ is not supported by history. The Southern Irish (Catholic) and the Northern Irish (Protestants) both speak English, and both groups are Christians. The battle forKashmiris fought between peoples who have common languages.

Most wars are between sibling nations sharing the lands and language groups. The battles between Parakramabahu in the Rajarata against the Ruhuna Aristocrats (who also spoke Sinhala !) are well known from the Chulavamsa. The JVP uprising was a confrontation between Sinhala speakers. The LTTE is said to have killed more Tamils than the army!

Thus there is little chance that multi-lingualism would remove the likelihood of communal strife. After all, immigrants, even Tamils who refuse to learn Sinhala in Sri Lanka, succeed in working in Finnish, Polish or Norwegian in a very short time! This shows that they would have had no difficulty in working in Sinhala, spoken by nearly 80% of the population. Language was merely the bone of contention within a situation that already existed, even prior to World War II. It is like the proverbial ‘coconut tree on a property boundary’ over which two villagers would litigate and fight all the way to the supreme court, while destroying each other in the process. If the coconut tree (say, language chauvinism) is removed, there is something else unless their common interests converge.  Already in the 1950s, the Buddhist Theosophical-Society schools like Ananda and Nalanda taught Tamil to the Sinhala students in their middle-school stage. This was given up probably when the politics became too hot.  By then, the ITAK went about discouraging Tamils from learning Sinhala, and by the late 1970s this involved the use of threats and force in many instances. Such discouragement was a logical component of separatist or even devolutionist politics based on the Indian model, or the 13th amendment.

Communal strife is diminished when people develop commercial and social links among one another. Thus, the government’s policy of opening up highways in the North and East, and opening the ‘cajan curtain’ of isolation is the technologically necessary step for the unification ofSri Lankaand the elimination of strife. In regard to language, knowing a client’s tongue is essential for success in trade. The 60 million potential clients in Tamil Nadu give the Sinhalese a valuable reason for learning Tamil, while the Tamils need Sinhala for trading with the Sinhalese and living in the south, as they increasingly do. A new language is an additional window to one’s mind, and hence its cultural value is immense.

II. LEARNING LANGUAGES: Leaving politics aside, learning two foster languages in addition to one’s mother tongue is to one’s advantage. In an earlier era  Sri Lankan students objected to learning English, claiming that to be a kaduva, i.e., an offensive weapon against the rural kid. During my period as the President of theVidyodayaUniversity in the mid 1970s, the Marxist student leaders demanded Russian instead of English! Today everyone accepts the need for English. The present Hon. Minister of Higher Education who strongly supports English Education may remember those days..

Hence we can assume that there is indeed a strong, built-in incentive for English. Indeed, many students pay to learn English from private academies.  However, the same cannot be said regarding a Sinhalese learning Tamil, or a Tamil learning Sinhala. Parents have shown no interest in motivating children to study the ‘other language’. In fact, the school system has failed to deal with even the core subjects and private tuition programs thrive. These programs completely saturate the time of students who are whisked away immediately after school into tuition classes. Successfully studying Tamil by Sinhala students, or vice versa under such an ambiance is unlikely.

One can attempt to enforce the study of the `other’ language by making it mandatory at a scholarship examination held at an intermediate level. However, the minority children would excel in learning the majority language (as found in Canada), and there would soon be accusations of favouritism and infiltration. The program would have the very opposite effect to that of improving ethnic goodwill.  Canadais a rich country that has attempted to enforce bilingualism (since the Elliot-Trudeau era) as an answer to separatism. And yet, after two decades of French immersion and bilingualism, the separatists nearly split Canadaduring the Jean Chretien Era. Even after massive bilingual bonuses and study programs to public servants, most English public servants stay at the `bonjour’ stage and consider the policy of bilingualism a painful punishment. Even with varied immersion programs in schools, English-Canadians mostly end up unilingual unless their parents spoke both languages at home. On the other hand, French Canadians have greater competency in English. It is a world language and they acquire it.  Thus the French Canadians, a minority that has become more bilingual than the majority occupy a majority of key positions in the Federal government! This has led to much resentment among English Canadians. 

However, even this bilingualism of French Canadians is nothing much to crow about. Thus, while more than 90% Quebecers speak French, only about 45% speak English.  Furthermore, taken overall, nearly 80% of the Canadians remain unilingual, even after four decades of highly expensive bilingual educational and government incentives! 

 Of course, you easily learn what you enjoy and like.  Some people easily learn all the hit songs in several languages, while being totally uninterested in the languages themselves. Tour guides learn several languages very rapidly, but only at a very limited level as the guides have a need for it to do his or her job. I. A. Richards, the famousOxfordlinguist pointed out that only about a thousand words in a foreign language are needed in practical situations. Nevertheless, the vast majority of people seem to not make the necessary effort, and remain unilingual. 

The cost of educating in three languages, where the two foster languages reach an intermediate standard can be estimated using school-entry and leaving models, together with available cost figures for uni-lingual education as well as supplementary-language education. It is necessary to train teachers for second-language methods, provide texts, language labs, test and grade students etc.  Assuming a currentGDPof about $42 billion, and the current education expenditure of approximately 5-6% of the GDP  on education, an additional outlay of at least $ 0.3 billion per annum at the start would be needed. The program has to run for at least two decades. The costs will increase as our population and service costs increase with time. The program would thus cost upwards of $10 billion. The money will create some jobs, but it will unfortunately not achieve the trilingual vision. Just like inCanada, 80% of the citizens would still remain unilingual!

 Indeed, given the experience of Canada, where French is a substantial minority, much bigger than the size of the Tamil minority in Sri Lanka, there is little chance that even a moderately ‘trilingual Lanka’ will come into being, in spite of the best efforts of those who are working for it. However, introducing Tamil into Sinhala schools, and Sinhala into Tamil schools up to the middle school level may be regarded as a good political concept enabling mutual inter-cultural understanding.

However, we need to do business with each other in the other’s language, and keep tack of each other politics in the respective language. After all, we saw how the ITAK manifestos written in Tamil differed vastly from those written in English for the ‘Federal party’. Similarly, the Sinhala press in isolation could be more insensitive than its sister publication in English. So language barriers need to be opened up.

How do we do that?

 III. THE TECHNOLOGICAL SOLUTION: In this section we argue that the “Language Problem” has an inexpensive and very practical technological solution that does not ruffle any political sensitivities.  Language as needed for business and social transactions is simple and unambiguous. The vast majority of people use language in that practical sense, As I. A. Richards showed, this requires a mastery of a mere thousand words or so in each language, and the sentence structure of that language. We are dealing with Sinhalese and Tamil, and these have essentially the same sentence structure. Further more, they have many common words and idioms.  Thus the machine translation of ordinary sentences in one language to the either is much simpler than going from, say, Sinhala to English. In fact, we argue that even the circuity and memory capacity in a hand-held device like a cellphone (mobile) is sufficient to achieve the needed lingustic capability Consider a Tamil person, Jega, in Vavniya speaking to his Sinhala friend Chula in Colombo. Jega dials Chula on his cellphone and asks in Tamil, ‘Kolompu pooheiradtham eththanei manikku ’? Chula has set his cell phone to render output in Sinhala. Even if the processor in the cell phone were designed to do a word to word replacement 4 from tamil to Sinhala, he would hear ‘Kolomba dhumriya keeyatadha? ’, i.e., in English, ‘At what time is theColombotrain’ ? In fact, it is easy to ascertain that most Sinhala $ Tamil translations are quite adequately handled by simply plugging in the Tamil words for the Sinhala words, and vice versa, as long as we are not trying to translate poetry! Chula replies into his cell phone in Sinhala, ‘Udei dahayata dhumriyak thiyenava ’, and Jega in Vavniya hears on his cellphone ‘Kaalai paththu manikku oru pooheiradtha erukku ’. Only a modest dictionary of about 1000 words in each language is all that is needed for this type of bilingual communication where each person only uses his own language, but instantly understands the other.

We have illustrated in the above examples a very elementary translation process where a dictionary lookup is used by the computer in the cell phone to do the translation. However, modern information technology is much more advanced than that. Such language translations, e.g., the ones available on your laptop or blackberry browser for translation between, say English and French, use computer codes known as “neural networks”.  Such networks in the computer accept inputs in one language, and output them in the client’s desired language (Fig. 1). Most western languages as well as Asian languages like Japanese are now freely available. Similarly, several local languages (e.g., Sinhala, Tamil, Hindi) could be provided on a Sri Lankan cell phone as options that may be selected from a touch pad.

Neural-network based algorithms are self-learning and acquire more and more vocabulary and context as you continue to use the language facility. In fact, such network modules have to be developed by a computer science laboratory and loaded into computers or cell phones as software. A computer science department in any of the universities inSri Lanka, or the Information and Communication Technology Agency (ICTA) inColombowould be eminently competent to develop such codes for use in cell phones, Internet browsers and other applications. Once it is developed, distribution is very easy as any one interested can down load it on-line for a minimal fee, or indeed free of charge. The cost of such a project, where the Tamil-Sinhala language barrier is opened up is negligible, and achieving the objectives in a very short time is well assured. Such algorithms can be developed beyond the level of simple transactions, and then they can be commercialized for corporate applications in the local languages.

The translation from Sinhala to English is more difficult because of the difference in sentence structure in the two languages. A neural-network algorithm is essential for such a translation scheme. As for Tamil$ English translations, it is more than likely that such codes are already being developed for Internet interfaces in South Indian institutions. 

IV. CONCLUSION: We have quoted historians and writers to argue that the naive identification of “language Chauvinism” as the cause of the post-independent civil strife inSri Lanka is just superficial. In any case, even if it were true, the highly laudable objective of makingSri Lanka ‘trilingual’ is quixotic and unattainable with even several decades of effort, as seen from the four decades of futile attempts made byCanada to become bilingual. A practical alternative immediately available from technology is the implementation of online Sinhala ↔ Tamil translation capability. Cell-phone communications mostly involve social and business transactions having a simple structure, suitable for automatic translation followed by electronically spoken output in the desired language.

Thus, even a unilingual individual acquires the capability of easy communication with his colleagues who are competent in some other local language.  The required technological know-how is eminently available in Sri Lanka’s computer-science institutions.

* The author Chandre Dharmawardena is a Principal Research Scientist at the National Research Council of Canada, Ottawa, Ontario and an Adjunct Professorof Physics at the Université de Montreal.


[1] The Sunday Observer,Colombo,09-Oct-2011, {http://www.sundayobserver.lk/2011/10/09/new21.asp}

[2] Jane Russell,Communal politics under the Donoughmore constitution 1931-1947, Tisara Publishers, (1982)

[3] The First Sinhala-Tamil Riot in 1939, occurred after a speech by GG Ponnambalam. {http://dhweb.org/place.names/riot1.htm}

[4] Dr. Michel Roberts, Journal of South Asian Studies, Vol XXVII, no. 1, April-issue. (2004) {http://dhweb.org/place.names/posts/rob-ajwilson.pdf}

[5] Gerald Peiris, Twilight of the Tigers,OxfordUniversityPress (2009).

[6] Sebastian Rasalingam, The Rise and fall of the Ceylonese identity in the face of Chauvinistic Politics, Sri Lanka Guardian, July 2011,{http://www.srilankaguardian.org/2008/07/rise-andfall-of-ceylonese-or-sri.html}

[7] Sebastian Rasalingam, Land Reforms in the North and East ofSri Lanka,IslandNewspaper, Oct. 2011{http://www.island.lk/index.php?page cat=articledetails& page=article-details&code title=36740};

[8] Somapala Gunadheera, A bilateral approach to Reconciliation,Islandnewspaper, Sept. 2001,{http://www.island.lk/index.php?page cat=articledetails&page=article-details&code title=37119}

[9] Neville Laduwahetty, The Island Newspaper, Sept. 2011, Tamil Political Ideology as “The pursuit of territory-based political separateness”,{http://www.island.lk/index.php?page cat=articledetails&page=article-details&code title=35291}

[10] Thomas Johnpulle, Caste Discrimination and Tamil separatistPolitics, Sri LankaGuardian, Oct 2011{http://www.srilankaguardian.org/2011/10/tamil-castediscrimination.html}


Filed under cultural transmission, historical interpretation, language policies, politIcal discourse, Sinhala-Tamil Relations, world events & processes

3 responses to “Switching on ‘Trilingual Competence’ without learning languages!

  1. padraigcolman

    I have approached the subject from a less academic and more frivolous angle here:


  2. Jean-Pierre

    I am glad that this writer has recognized that blaming the Tamil-Sinhala conflict on “Sinhala Only” is utterly simplistic. That may have been very useful to Tamil as well as Sinhala Politicians as a “Casus belli”, but we known from many other conflicts, e.g., in Latin America, that Land ownership and control are almost always the underlying causes. This author has hinted in his introduction that upper-Caste land ownership and their wish to control it by separating off from the central government was a more determining factor.

    I am willing to believe it, and I hope social scientists would look into it and document it more carefully. The use of Sinhala in a geographically tiny country with 75% Sinhala people, with the North and East sparsely populated, seems fair enough.

    In the USA we have lots of Hispanics, with California, Texas and new Mexico as the “traditional homelands”, and the percentage of Hispanics in these provinces tend to be quite high; but absolutely no concession to Spanish has been made for the Hispanics by the US government.

  3. Pingback: On experiments in economic governance: to plan or not to plan? | Thuppahi's Blog

Leave a Reply