ECMI Panel

Tove Skutnabb-Kangas



This is NOT an abstract, but a paper (drawing on my recent paper in Fishman (ed.) 1999). Since I would like as much discussion as possible, I present here definitions and evaluations of the various models and relate them to a certain extent to Roma education. I would like conference participants to relate them already in advance to models they know, and to think of the relative merits of each model. This is especially important in relation to the concept of 'segregation' which for many seems to be a ghost and which needs careful analysis. Is it physical or mental? Is it a means or a goal? Is it initial and temporary or permanent?. Likewise, I would like participants to reflect on my claim that it is only the strong models that should be promoted. Using transitional early-exit models (which many of those who are afraid of the segregation ghost, even some (many?) Roma, seem to favour, at least in some countries) seems to be a defensive strategy, with some possible initial instrumental gains but long-term losses which amount to linguistic and cultural genocide according to Article II(e) in the UN Genocide Convention 'forcibly transferring children of the group to another group'. Is this a price the Roma are willing to pay? Do they know about the long-term consequences? More radical solutions are needed - but are they possible in the present situation of increasing racism?


'If social integration is taken to be a psychological state characterized by positive self/ingroup identity along with positive other/outgroup identification, then bilingualism, both at the individual and at the social levels, seems to promote social integration.' 

(Ajit Mohanty)


A good educational programme leads to the following goals from a language(s), identity, labour market and life chances point of view:

  1. high levels of multilingualism
  2. a fair chance of achieving academically at school,
  3. strong, positive multilingual and multicultural identity and positive attitudes towards self and others, and
  4. a fair chance of awareness and competence building as prerequisites for working for a more equitable world, for oneself and one's own group as well as others, locally and globally.

I see these goals as positive for ALL children, both minority and majority. Of course the education of ethnic minorities also has to fulfil further demands that can be made on any good education. My paper will look at some of the prototypes for education from the point of view of what role language(s), especially the medium of education, plays for the achievement of the three goals above. I shall also discuss relative merits of various programmes for Roma children at a general level, but it is up to the Roma themselves to decide which goals are the most urgent ones in various situations, countries and times, and how they see the chances of implementation.

The prototypes can be discussed in terms of non-models, weak models and strong models of bilingual/multilingual education.

The non-models do not reach the three goals but lead often to monolingualism or very strong dominance in the majority language and a negation of goals 2 and 3.

The weak models, even when assimilationist, are not quite as harsh for the child. They may often lead to somewhat better chances of school achievement. But in general they do not reach the goals either, especially not goals 1 (where they may reach limited bilingualism) and 3.

Strong models are the only ones which may reach the goals at a group level. Their linguistic aim is to promote multilingualism (or, minimally, bilingualism) and multiliteracy.

The non-models and weak models -- models which are insufficient in reaching the goals, which violate linguistic and cultural human rights and participate in committing linguistic genocide -- are, regrettably, still the most common models for educating indigenous and minority children.

At a group level the bulk of minority children still 'fail' in school. Many are pushed out early, and the school achievement of many is below that of majority children as a group. Later on, they are overrepresented in unemployment and youth criminality statistics and other statistics showing results of an unequal society.

This is the general picture and it is especially true of the Roma. However, there are both individuals and groups who are an exception and are managing well, sometimes even better than majority children. Mostly they do it despite the way their education is organised, not because of it.

Deficiency-based theorizing and assimilation

In the West, the ideology of the nation state has to a large extent prevailed until now. In this linguistically, culturally and socially homogenous community of an integrated population, national and linguistic identities are supposed to coincide. Most states in the world are of course NOT nation-states in this sense. Since there 7-10,000 oral languages (and probably equally many sign languages) but only some 200 states, there are necessarily many 'nations' and speakers of many languages in all but a few states. Therefore this (assimilatory) 'integration' and homogenisation has to be achieved through social engineering and state-initiated reforms. Formal education has always played a decisive part in trying to achieve it. The nationalistic, racist (and classist and sexist) tendencies in this ideology were directed towards all those who had to be forcefully 'uplifted' from their 'otherness': linguistic minorities, the working class, women, etc.

This static and ethnocentric view still prevails in many countries: the whole burden of 'integration' is on the dominated groups and individuals alone, they are the ones who have to change. The dominant group is presented as non-ethnic. Its values are presented as The Norm, or as Standard, and as somehow 'shared' and 'universal', rather than particularistic and changing, like all values are. When the majority population is presented in this way as an integrated mainstream, homogenously sharing universal cultural values, this characteristic legitimates its access to most of the power and resources. These are, of course, shared unevenly on a class and gender basis within the majority population, but this is often not mentioned when integration is discussed.

For a few national minorities (e.g. Swedish-speakers in Finland, English- and Afrikaans-speakers in South-Africa), their right to exist, to define independently who they are (to endo-categorize), and to reproduce themselves as minorities and, accordingly, to have mother tongue medium education, have been more or less self-evident.

For most minorities who have (some of) these rights today, achieving them has been a result of a long struggle. Most minorities do not have these rights. Most minorities are still exo-categorized, i.e. defined by others. The problems they (may) face in the educational system are (mis)diagnosed by representatives of the dominant group(s).

Typically the minorities themselves have been and are blamed for any failures. 'Reasons' for problems in school have been said to stem from several deficiencies or handicaps in the minorities themselves. The minority child, her parents and her group have been seen as causing the problems. The most often used 'diagnoses' for these problems have been, according to Stacy Churchill, linguistic (related to either L2 or L1), social or cultural (L = language; L2 = second language; L1 = first language, mother tongue). The child is seen as suffering from handicaps which are

  • L2-related: the minority child (and her parents) do not know the dominant language, e.g. English, well enough;
  • social: the children's parents represent the lowest social groups, with little formal education, high rates of unemployment, etc; the children do not get enough school-related support at home;
  • cultural: the minority culture -- family patterns, gender roles, relations between the generations, etc -- is different from the dominant group's culture; there is a cultural clash; this prevents the child from achieving; and, in a later phase,
  • L1-related: the child does not know her mother tongue well enough and is therefore left without a solid basis for L2-learning too.

In deficiency theories, the minorities themselves, their characteristics (including bilingualism), are seen as The Problem. Measures to cure the problems have typically included more L2-teaching, social support, some forms of multicultural or intercultural education, and some L1-teaching, respectively. Quick assimilation, linguistically and culturally, and acceptance of the dominant group's linguistic, social and cultural norms, have been official or unofficial goals. Roma have both been excluded from formal education and themselves resisted the assimilatory formal education. All four deficiency-related explanations have been and are being used to explain the non-participation, early push-out and lack of success in schools for Roma children.

The educational models used in these early phases of deficiency theorizing can be called non-models and, at the most, weak models of bilingual education. I do not know of any strong models of Roma education - yet.

Non-models and weak models for the education of ethnic minority (and majority) children.

A submersion or sink-or-swim programme is a programme where linguistic minority children with a low-status mother tongue are forced to accept instruction through the medium of a foreign majority language with high status. They are placed in classes where some children are native speakers of the language of instruction. Usually the teacher does not understand the mother tongue of the minority children. The majority language constitutes a threat to the minority children's mother tongue. It runs the risk of being displaced or replaced -- a SUBTRACTIVE language learning situation. The mother tongue is not being learned (properly); it is 'forgotten'; it does not develop because the children are forbidden to use it or are made to feel ashamed of it. This is the most common -- and most disastrous -- method in the present world for educating minority children.

In another variant of a submersion programme powerless majority children (or groups of minority children in a country with no decisive numerical and/or power majorities) are forced to accept instruction through the medium of a foreign (often former

colonial) high-status language (because mother tongue medium education does not exist). This happens in mixed mother tongue classes, mostly without native speakers of the language of instruction. But it also happens in linguistically homogenous classes, sometimes because MT-medium education does not exist and sometimes because the school or the teachers hesitate to implement a MT-medium programme. The teacher may or may not understand the mother tongue of (some of) the children. The foreign language of instruction is not learned properly. At the same time, the children's mother tongues are being displaced and not learned (properly) either in formal domains (for instance literacy is not achieved). Often the children are made to feel ashamed of their mother tongues, or at least to believe in the superiority of the language of instruction. Many African, Latin-American and Asian countries use these programmes.

A transitional programme is a programme where linguistic minority children with a low-status mother tongue are initially instructed through the medium of their mother tongue for a few years. But their mother tongue has no intrinsic value, only an instrumental value. Teaching through the medium of the mother tongue is not seen as a right that the child is entitled to. The mother tongue is seen as useful only so far as its auxiliary use enhances the knowledge of the dominant language. Using the mother tongue also gives the children some subject matter knowledge while they are learning the majority language.

As soon as they can function, at least to some extent, in the majority language orally (early-exit), or at the latest around grade 6 (late-exit), they are transferred to a majority language medium programme.

A transitional programme is a more sophisticated version of submersion programmes, a more 'humane' way of assimilating. Transitional bilingual education encourages shift to first dominance and later on monolingualism in the majority language. These programmes are common in the education of migrant children in some of the more progressive settings (some programmes in Sweden, Holland, USA etc). They are also used in parts of 'Anglophone' Africa.

A segregation programme is a programme where linguistic minority children with a low-status mother tongue are forced to accept instruction through the medium of their own mother tongue. Or the national language of their country of origin: e.g. Kurdish children from Turkey in Bavaria, Germany, are taught through the medium of Turkish not Kurdish. The children are in classes with minority children with the same mother tongue only. The teacher may be monolingual or bilingual but is often poorly trained. Often the class/school has poorer facilities and fewer resources than classes/schools for dominant group children. The teaching of the dominant language as a second/foreign language is mostly poor or non-existent. Many present education models for Roma children can also be characterised as segregation, either by direct educational design or as a result of demographic and economic circumstances.

But majority children also have non-models: mainstream monolingual programmes, maybe with some foreign language teaching. In the worst case (as in many North American or Russian classrooms), no foreign languages are taught at all, leading to what we call monolingual stupidity or monolingual reductionism. In other countries, one or several foreign languages are taught as subjects for a few hours per week. This is still the preferred mode of foreign language instruction in the world. The best foreign-language-as-a-subject teaching, for instance in The Netherlands or in the Nordic countries, can give a solid basis for bi- or multilingualism, if it is combined with travel or using the language in daily intercourse later.

Assessing the non-models and weak models

Minority struggle often starts when parents can see that their children are not doing well at school, despite trying to do whatever the majority society and school demand. In addition, the parents often feel that they are losing their children who no longer know the mother tongue, who may feel ashamed of their parents, language and culture, and who assimilate rapidly, but without getting the benefits which were promised with assimilation. Many minorities start the struggle with demands of the learning of the minority's own culture and instruction in the mother tongue, first as a voluntary subject and later also as a required medium. In the initial phases of mother tongue medium education, the legitimation has often been that it leads to better L2-competence and helps school achievement because the children do not lose content while learning L2. This is still the only (grudgingly accepted) general legitimation for bilingual education in most Western countries. It leads in the best cases to transitional early-exit programmes. These are still based on seeing the minority child as deficient and education trying to 'compensate' for the deficiencies.

Transitional early-exit programmes are a more humane way to assimilate minority children than direct submersion (i.e. placing the children in majority-language-medium classes). Still, they are language shift programmes. They do not normally lead to high-level bilingualism.

All non-models and weak models for minority children fit the United Nation's definition of linguistic genocide. When the United Nations did preparatory work for what later became the International convention for the prevention and punishment of the crime of genocide (E 793, 1948), linguistic and cultural genocide were discussed alongside physical genocide, and were seen as serious crimes against humanity (see Capotorti 1979).

When the Convention was accepted, Article 3 covering linguistic and cultural genocide was voted down in the General Assembly, and is not part of the final Convention of 1948. What remains, however, is a definition of linguistic genocide, which most states then in the UN were prepared to accept. Linguistic genocide is defined (in Art. 3, 1) as

Prohibiting the use of the language of the group in daily intercourse or in schools, or the printing and circulation of publications in the language of the group.

'Prohibition' can of course be overt and direct (e.g., killing or torture for using the mother tongue, as in Turkey vis-a-vis Kurds) or covert, more indirect, accomplished via ideological and structural means. If the minority language is not used as the main medium of education and childcare, the use of the minority language is indirectly prohibited in daily intercourse or in schools, i.e., it is an issue of linguistic genocide. This is the situation for most immigrant and refugee minority children in all Western European countries and in the US, Canada and Australia, for many 'national' minorities, and for most indigenous first nations.

How should minority education be organized?

Enrichment-based theorizing and real integration

In the next phase in theorising education, enrichment-oriented theorizing may emerge. Bi- or multilingualism is increasingly seen as something positive, not only for minorities (for whom it is necessary) but also majorities. For majorities, immersion programmes are started (see below), initially mainly for instrumental reasons. Minorities start demanding the right to (positive) endo-categorization. This is often a part of a revitalization process, especially for indigenous peoples and old national minorities. Minorities and indigenous peoples who have been linguistically assimilated, also start immersion programmes of various kinds (for instance 'language nests', as with Māori in New Zealand, Hawai'ians, or several Sámi people in the Norwegian and Finnish parts of Sápmi). In addition, they and those (few) immigrant minorities who are arguing from an enrichment-oriented point of view, often use human rights arguments.

These can be individually oriented ('it is a linguistic human right to learn one's mother tongue fully, and also to learn an official language fully'). They can also be more collectively oriented ('it is a human right for a minority to exist, and this presupposes learning both L1 and L2 fully'). A combination of both types may be expressed in terms of arguing that linguistic and cultural diversity are not only necessary for the planet but positive resources in any society.

With more sympathy towards the right of minorities to endo-categorize, initially transitional late-exit programmes may be accepted. This was, for instance, happening in the United States in the early 1990s, with a strong backlash in the last year or two of the millennium. At the same time, the more human rights oriented demands often start exposing the power struggles involved in minority education as a part of the more general polarization which is happening in the world. Some educators and parents with foresight, from both the minorities themselves and from majority populations, start relating the groups to each other. They claim that high levels of bi- and multilingualism are beneficial for everybody; that majorities need to become bilingual too; that for real integration to happen, both majorities and minorities need change; that granting educational language rights may prevent ethnic conflict and disintegration. It does no good to try and change the minority child to fit a majority school. It is not enough to try to give the minority child an emergency kit so that the child can manage in a racist society. It is not enough to enrich the majority child through a bit of exposure to other cultures. Instead, the whole school has to change. Society has to change.

Often people have discussed only the instrumental necessities for minorities (they have to learn L2) or instrumental benefits for majorities (bilinguals get better jobs) that a certain type of education can lead to. Alternatively, people discuss only ethnically/linguistically motivated identity-oriented necessities that a certain type of education can lead to for minorities. But a less naive human rights argumentation is also emerging. It combines both instrumental and affective benefits.

The new discussions have also led to the development of better educational models.

Strong models for the education of ethnic minority and majority children

Immersion programmes for majorities

An immersion programme is a programme where linguistic majority children with a high-status mother tongue voluntarily choose (among existing alternatives) to be instructed through the medium of a foreign (minority) language. The children are placed in classes which only have majority children, all with the same mother tongue (the classical model), or at least in classes where the medium of instruction is a foreign language for everybody. The teacher is bilingual so that the children in the beginning can use their own language even if the teacher replies in the foreign language. The children's mother tongue is in no danger of not developing or of being replaced by the language of instruction -- they are in an ADDITIVE language learning situation. Canada has been the pioneer of immersion programmes (see Lambert 1972, Swain & Lapkin 1982). Most of them are still in French (for mainly English-speakers) but many other languages are also involved. Most European countries and the United States also have immersion programmes in different languages. Increasingly, immersion is also used for revitalisation or reclaiming of languages - this would be one of the possibilities for Roma.

Language maintenance (language shelter) programmes for minorities

A maintenance programme or language shelter programme is a programme where linguistic minority children with a low-status mother tongue voluntary choose (among existing alternatives) to be instructed through the medium of their own mother tongue. The children are placed in classes with other minority children with the same mother tongue only. The teacher is bilingual. They get good teaching in the majority language as a second/foreign language, and this is also given by a bilingual teacher. These schools are most often organized by an ethnolinguistic minority community itself. Initially, the students' native language is used for most of the content matter education, especially in cognitively demanding, decontextualised subjects, while the majority language is taught as a subject only. Later on, some (but by no means all) maintenance programmes use the majority language as a medium of education for part of the time. But in proper maintenance programmes the minority language continues as a medium of education in several (or most, or all) subjects throughout the school.

For a few national minorities, maintenance programmes are a self-evident, 'normal' way of educating their children, a natural human right. It is indicative that most minorities of this type, e.g. the Swedish-speakers in Finland, Afrikaans- and English-speakers in South Africa, or Russian-speakers in Estonia, Latvia and Lithuania, are either former power minorities or in a transitional phase where they have to accept the fact that they no longer have the power to impose their will on a numerical majority. But they still do have the power to organise their own children's education through the medium of their own language. Of course it should be a fundamental, self-evident linguistic and educational human right for any ethnolinguistic minority to use its own language as the main medium of education. But in fact most minorities in the world do not have this basic right. A few indigenous peoples (who are numerically a minority in most of their own countries) have maintenance programmes, most of them do not. Most immigrant and refugee minority children do not have access to maintenance programmes either, even if it can be shown that such programmes would result in high levels of bi/multilingualism, enhanced school achievement and more societal equity.

The purpose of this type of multilingual programme is to ensure that language minority children continue to maintain and develop their mother tongue up to a native (national minorities, indigenous peoples) or at least near-native (immigrant minorities) level. Likewise, they learn the majority language at a native level, and become biliterate. In a European context, they typically also learn further foreign languages. Many mother tongue medium programmes in African countries (all but the segregation programmes) could also be counted under maintenance programmes. Birgit Brock-Utne observes that 'in many of the African countries the majority language is treated in a way that minority languages are treated in the industrialized world'.

This type of multilingual programme enriches society at large by promoting pluralism and mutual understanding and by ensuring that minorities gain access to linguistic and educational prerequisites for social, economic and political integration.

The results are positive. A recent World Bank report (by Nadine Dutcher) gives a summary of some late-exit and maintenance programmes, including the largest evaluation done in the United States, by David Ramirez and his colleagues. In my own study, Finnish working class youngsters in Sweden who had had nine years of mostly Finnish-medium education, were compared with mostly middle-class Swedish children in parallel classes in the same schools. In addition to doing almost as well as Finnish controls in Finland on a Finnish language test, they performed slightly better than the Swedish controls on a Swedish language test. Their school achievement was somewhat better than that of the Swedish-speakers. They had a positive bilingual, bicultural, 'bicountrial' identity. This is also the model that responds to human rights demands, as they are presented in the Hague Recommendations. The Hague Recommendations are the new educational guidelines issued by The Foundation on Inter-Ethnic Relations for the OSCE. In 1992, OSCE (= Organisation for Security and Cooperation in Europe) created the position of a High Commissioner on National Minorities 'as an instrument of conflict prevention in situations of ethnic tension' (Rothenberger 1997:3) OSCE (= Organisation for Security and Cooperation in Europe) High Commissioner on National Minorities, P.O.Box 20062, 2500 EB, The Hague, The Netherlands, phone: 31-70-312 5500; fax 31-70-363 5910, email <>; web-site: <>. In order to prevent ethnic conflict, the High Commissioner, Max van der Stoel, published authoritative guidelines in October 1996 for minority education for the 55 member states (which include Canada and the United States). The High Commissioner, Max van der Stoel (1997: 153) stated when launching The Hague Recommendations Regarding the Education Rights of National Minorities that the course of my work, it had become more and more obvious to me that education is an extremely important element for the preservation and the deepening of the identity of persons belonging to a national minority. It is of course also clear that education in the language of the minority is of vital importance for such a minority.

These guidelines, The Hague Recommendations, were worked out by a small group of experts on human rights and education (mainly lawyers and including the author of these lines). The guidelines are an interpretation and concretisation of international human rights standards about minority education. Even if the term used is 'national minority', the guidelines also apply to immigrated minorities, and one does NOT need to be a citizen in order to be protected by the guidelines.

In the section 'The spirit of international instruments', bilingualism is seen as a right and responsibility for persons belonging to national minorities (Art. 1), and states are reminded not to interpret their obligations in a restrictive manner (Art. 3). In the section on 'Minority education at primary and secondary levels', mother tongue medium education is recommended at all levels, including bilingual teachers in the dominant language as a second language (Articles 11-13). Teacher training is made a duty of the state (Art. 14).

11) The first years of education are of pivotal importance in a child's development. Educational research suggests that the medium of teaching at pre-school and kindergarten levels should ideally be the child's language. Wherever possible, States should create conditions enabling parents to avail themselves of this option.

12) Research also indicates that in primary school the curriculum should ideally be taught in the minority language. The minority language should be taught as a subject on a regular basis. The State language should also be taught as a subject on a regular basis preferably by bilingual teachers who have a good understanding of the children's cultural and linguistic background. Towards the end of this period, a few practical or non-theoretical subjects should be taught through the medium of the State language. Wherever possible, States should create conditions enabling parents to avail themselves of this option.

13) In secondary school a substantial part of the curriculum should be taught through the medium of the minority language. The minority language should be taught as a subject on a regular basis. The State language should also be taught as a subject on a regular basis preferably by bilingual teachers who have a good understanding of the children's cultural and linguistic background. Throughout this period, the number of subjects taught in the State language, should gradually be increased. Research findings suggest that the more gradual the increase, the better for the child.

14) The maintenance of the primary and secondary levels of minority education depends a great deal on the availability of teachers trained in all disciplines in the mother tongue. Therefore, ensuing from the obligation to provide adequate opportunities for minority language education, States should provide adequate facilities for the appropriate training of teachers and should facilitate access to such training.

Finally, the Explanatory Note states that

[S]ubmersion-type approaches whereby the curriculum is taught exclusively through the medium of the State language and minority children are entirely integrated into classes with children of the majority are not in line with international standards (p. 5).

Knowing that most minority education in the world today is of submersion type, it is useful to know that this type of genocidal education is not in line with international human rights standards. This also means that the children that the Recommendations apply to might be granted some of the central educational linguistic human rights. The questions now are to what extent the 55 OSCE countries will apply the Recommendations and how they will interpret their scope (see the Special Issue volume 4:2, 1996/1997 of International Journal on Minority and Group Rights which is about the Hague Recommendations). Still, the Recommendations are still recommendations, not binding international law, and no sanctions are envisaged if (when) states violate them - as all states today do.

Two-way programmes and the European Union Schools model for minorities AND majorities

A two-way programme is a programme with 50% minority (with the same mother tongue) and 50% majority children are taught together by a fully bilingual teacher. Initially this happens through the medium of the minority language, later through both. Both languages are taught as subjects to both groups. A two-way programme combines a language shelter programme for the minorities and an immersion programme for the majorities. There are two-way programmes in over 200 schools in the USA. The results are positive.

A European Union Schools model is a programme where each language group is taught separately in their own section. Initially the teaching is completely through the medium of their own language. Later they are taught partially together with children from other sections, through the medium of one or two foreign languages. These are first studied as subjects, then used as media of education in concrete contextually embedded subjects. Only after grade 8 are these languages also used as media of instruction in decontextualised, intellectually and linguistically demanding subjects, according to a careful plan. Both the mother tongue and the first foreign language are taught as subjects throughout grades 1-12. The mother tongue continues to be the medium in several subjects throughout the programme. All staff are bi- or multilingual.

Although the strong forms of multilingual education outlined above have different sociolinguistic realities with regard to the linguistic background of the students and the language(s) of the classroom, and different sociopolitical realities with regard to the power relations between the groups attending and the rest of society, they all share an aim of cultural and linguistic pluralism, with the bi/multilingualism and bi/multiliteracy of students as an avowed minimum aim.

Assessing the leading principles for strong models

The experiments described above have reached good results in terms of the goals we mentioned initially: high levels of bi- or multilingualism, a fair chance of success in school achievement, and positive multilingual/multicultural identities and attitudes. The principles which have to a large extent been followed in them can be formulated as 8 recommendations. They form one possible baseline which the reader can relate to, agree or disagree with. Here are the principles.

1. Support (= use as the main medium of education, at least during the first 8 years) that language (of the two that the child is supposed to become bilingual in initially) which is least likely to develop up to a high formal level. This is for all minority children their own mother tongue. For majority children, it should be a minority language. (The European Schools do not follow this principle completely, because they teach also majority children initially through the medium of their mother tongues, e.g. the the Italian-speaking children in the European School in Italy are initially taught through the medium of Italian, instead of a minority language).

2. In most experiments, the children are initially grouped together according to their L1. Mixed groups are not positive initially, and certainly not in cognitively demanding decontextualised subjects. (Spanish-English Two-way programmes in the U.S.A. are an exception: they have mixed in the same class 50% minority, 50% majority children. All are initially taught through the medium of the minority language, later through both. This may be a relevant factor in accounting for the Spanish-speaking children's sometimes relatively less impressive gains in both languages, compared to English-speaking children in the same programmes. The mere presence of majority language children in the same classroom may be too overwhelming for minority children, despite the minority language being the medium of education).

3. All children are to become high level bilinguals, not only minority children. This seems to be especially important in contexts where majority and minority children are in the same classes.

4. All children have to be equalized vis-a-vis the status of their mother tongues and their knowledge of the language of instruction. Nice phrases about the worth of everybody's mother tongue, the value of interculturalism, etc, serve little purpose, unless they are followed up in how the schools are organised.

There has to be equality in the demands made on the children's and the teachers' competencies in the different languages involved, so that the same demands are made on everybody. Both minority and majority children and teachers must be or become bi- or multilingual.

There has to be equality in the role that the languages are accorded on the schedules and in higher education, in testing and evaluation, in marks given for the languages, in the physical environment (signs, forms, letters, the school's languages of administration, the languages of meetings, assemblies, etc), in the status and salaries of the teachers, in their working conditions, career patterns, etc.

It is possible to equalize the children vis-a-vis their knowledge of the language of instruction in several different ways:

A. All children know the language of instruction (maintenance programmes, European Schools initially);

B. No children know the language of instruction or everybody is in the process of learning it (immersion programmes, European Schools in certain subjects in a later phase);

C. All children alternate between 'knowing' and 'not knowing' the language of instruction (two-way programmes in a late phase; alternate-days programmes (50% minority and 50% majority children, the medium of education alternates daily).

5. All teachers have to be bi- or multilingual. Thus they can be good models for the children, and support them in language learning, through comparing and contrasting, and being metalinguistically aware. Every child in a school has to be able to talk to an adult with the same native language.

This demand is often experienced as extremely threatening by majority group teachers, many of whom are not bilingual. Of course all minority group teachers are not high level bilinguals either. But it is often less important that the teacher's competence in a majority language is at top level, for instance in relation to pronunciation, because all children have ample opportunities to hear and read native models of a majority language outside the school anyway, whereas many of them do NOT have the same opportunities to hear/read native minority language models. A high level of competence in a minority language is thus more important for a teacher than a high level of competence in a majority language.

6. Foreign languages should be taught through the medium of the children's mother tongue and/or by teachers who know the children's mother tongue. No teaching in foreign languages as subjects should be given through the medium of other foreign languages (for instance, Turkish children in Germany should not be taught English through the medium of German, but via Turkish).

7. All children must study both L1 and L2 as compulsory subjects through grades 1-12. Both languages have to be studied in ways which reflect what they are for the children: mother tongues, or second or foreign languages. Many minority children are forced to study a majority language, their L2, as if it was their L1.

8. Both languages have to be used as media of education in some phase of the children's education, but the progression in how and how much each is used seems to vary for minority and majority children.

For MAJORITY CHILDREN the mother tongue must function as the medium of education at least in some cognitively demanding, decontextualized subjects, at least in grades 8-12, possibly even earlier.

But MAJORITY CHILDREN can be taught through the medium of L2 at least in some (or even all or almost all) cognitively less demanding context-embedded subjects from the very beginning. L2 can also be the medium of education, at least partially, in cognitively demanding decontextualized subjects, at least in grades 8-12.

For MINORITY CHILDREN the mother tongue must function as the medium of education in all subjects initially. At least some subjects must be taught through L1 all the way, up to grade 12, but the choice of subjects may vary. It seems that the following development functions well:

  • transfer from the known to the unknown;
  • transfer from teaching of a language (as a subject) to teaching through the medium of that language;
  • transfer from teaching through the medium of L2 in cognitively less demanding, context-embedded subjects, to teaching through the medium of L2 in cognitively demanding decontextualized subjects.

The progression used for all children in the European Union Schools seems close to ideal for minority children. The progression in relation to the (minority) MOTHER TONGUE is as follows:

1. All subjects are taught through the medium of the mother tongue during the first 2 years.

2. All cognitively demanding decontextualized core subjects are taught through the medium of the mother tongue during the first 7 years.

3. There is less teaching through the medium of the mother tongue in grades 8-10, and again more teaching through the medium of the mother tongue in grades 11-12, especially in the most demanding subjects, in order to ensure that the students have understood, can express and critically evaluate them thoroughly.

4. The mother tongue is taught as a subject throughout schooling, from 1-12.

The progression in relation to the SECOND LANGUAGE is as follows:

  1. The second language is taught as a subject throughout schooling, from 1-12.
  2. The second language becomes a medium of education already in grade 3, but only in cognitively less demanding context-embedded subjects. Teaching can take place in mixed groups, but ideally together with other children for whom the language is also an L2.
  3. Teaching in cognitively demanding decontextualized subjects only starts through the medium of L2 when the children have been taught that language as a subject for 7 years (grades 1-7) and have been taught through the medium of that language in cognitively less demanding context-embedded subjects for 5 years (grades 3-7). Children should not be taught demanding decontextualized subjects through L2 together with children for whom the language of instruction is their L1, before grade 8. In European Union Schools this is mostly not done even in grades 9-12 in compulsory subjects, only in elective courses.

When applying the principles to the strong models discussed above it appears that the European Union Schools model -- which factually achieves the best results -- gets more plus-ratings than any of the other models. Even if many of these schools are elite schools, they seem to succeed because the model is scientifically sound, not because of their elitism.

The link between ethnicity, identity and language in the struggle for the right to good education for minorities AND majorities.

Ethnicity, identity and mother tongue: primordial sources, shaped by social forces

Ethnicity has been proclaimed dead many times during this century, especially after the second World War. Liberal researchers claimed (and many continue to claim) that it was a traditional, romantic characteristic which would disappear with modernisation, urbanisation, global mobility. Ethnic identities would be replaced by other loyalties and identities: professional, social, gender, interest-group, state-related, global, and so on. Marxist researchers claimed that class-related solidarities which even crossed national borders, would replace ethnicity: international proletariat would unite against world capitalism. Post-modernist researchers now pronounce that we have (should have?) no lasting identities, only flexible temporary nomadic ones: all that is solid melts in the air.

Of course all of us have multiple identities. We may identify at the same time as, say, woman, socialist, ecological farmer, world citizen, mother, daughter, wife, researcher, Finnish, Scandinavian, European, witch, theosopher, lover of music and plants, and so on. This can be done without these identities necessarily being in conflict with each other. Of course, some identities will be more or less salient, focussed and emphasized than others at different times and new identities will emerge or be added, with others fading or being rejected over time.

Still, ethnic identities and, especially, linguistically anchored ethnic identities, seem to be remarkably resilient. Ethnicity is not dead at all; quite the opposite.

Both ethnicity and an attachment to one's language or mother tongue(s) as a central cultural core value, seem to draw on primordial, ascribed sources: you are 'born' into a specific ethnic group and this decides what you mother tongue (or mother tongues, if you have two, for instance with parents speaking different languages) will initially be. But what happens later to your ethnicity, your identity, and your language(s) and how they are shaped and actualised, is influenced by (achieved) economic/political concerns, by your social circumstances and later life. This also influences to what extent you are aware of the importance of your ethnicity and your mother tongue(s) and the connection between them.

This also means that I do not agree with those researchers who see both ethnicity and a mother tongue in an instrumentalist way as something you can choose (to have or not have, use or not use), according to your own whims and wishes. Because of the primordial sources, reaching back into infancy and personal history, neither ethnicity nor mother tongue, nor even identities can be treated as things, commodities, which you can choose at will and chuck out like an old coat if that is what you want. On the other hand, this does NOT mean that they are unchangeable givens or impossible to influence or change.

What is important to study, then, is: under which circumstances can their ethnicity and their language(s) become positive forces and strengths, sources of empowerment in people's lives?

Ethnicity, language and medium of education

Education leading to multilingualism has been used (and fought for, and against) with different motivations. Different strategies have been needed in different countries, even by different minorities in the same country at the same time, because the circumstances have differed. If forced segregation, with denial of access to the power language, was the worst enemy (as in apartheid South Africa, or guest-worker Germany - or the Roma?), a different model had to be chosen, as compared to having forced assimilation as the enemy, with the killing of the minority language as a concomitant. Here I shall relate the models described earlier to the roles that ethnicity and language have played for their development.

For majority groups, ethnicity often does not play any conscious role when choosing the medium of education. In most cases, choosing the majority mother tongue as the medium is done by individuals routinely, as something self-evident -- even if the function of it is to reproduce a majority ethnic identity.

When powerful majorities choose another language as a medium, among existing alternatives, instrumental reasons in most cases prevail: they have recognized the economic and career-oriented benefits of making their own children multilingual, thereby improving opportunities of doing business with and getting ahead in an increasingly interdependent world. Immersion programmes, the European Union Schools and International Schools are examples of this approach of enrichment and extra benefits. In most cases, there is no change of identity for students in these schools, and the connection between language and ethnic identity is often hardly discussed.

In addition to these instrumental reasons, both immersion programmes and, especially, two-way programmes may have an element of an 'integrative' motivation. In situations of potential conflict, multilingual education for majority children may represent a means to improve their understanding of other ethnolinguistic groups with which they are in contact. With revitalising minority groups, different types of two-way programmes may become more common -- an optimistic reading.

For many oppressed and often segregated minorities, ethnic identity was initially not seen as important. Rather, they focussed on school achievement, educational opportunity and equality. Many Roma do this today. Submersion programmes initially gave them a hope of learning the dominant language and culture. These were, rightly, seen as possessing instrumental value which their own languages, cultures and ethnicities could not promise. For most, the hope failed and the promised results did not materialise. In addition, many lost their only linguistic resource, their own language (or did not have any chance of developing it to a high formal level), and many children were made to feel ashamed of their ethnic background. When the first mother-tongue-as-a-subject programmes and the transitional early-exit programmes arrived, the legitimation still was -- and is -- instrumental and L2-oriented: the minority mother tongues got some rights because this led to more L2-competence. Schools should not train minority children for fatherland, folk-dances and ethnolore but for jobs, school authorities say. Often the authorities show a poor understanding of what mother tongue and ethnicity are. They also see the world as either-or, rather than both-and: either L1 or L2; either tradition or modernity; either the parents or a job -- all false polarisations. Many early-exit programmes for indigenous peoples may also belong to this type where educational achievement is emphasized. In theory many Latin American bilingual programmes emphasize the language, culture and ethnicity of the indigenous group. But the model chosen (early-exit) does not lead to the proclaimed goals.

Late-exit transitional, two-way and even maintenance programmes often concentrate on mother tongue medium education for minorities who have previously been excluded from equal educational opportunity and, especially, equality of outcome. Even if the mother tongue is seen as important and there is often a lot of emphasis on creating and supporting positive ethnicity, instrumental career-oriented concerns are at least equally important. These are typically both-and-and models, rather than the either-or which is typical of submersion and segregation but also early-exit transitional programmes.

Still others, often threatened ethnolinguistic groups, have adopted multilingual education as a prerequisite or means of linguistic survival. Maintenance/language shelter programmes or revitalisation programmes for minorities, e.g. the Frisian schools in the Netherlands or Kōhanga Reo programmes in New Zealand for the Māori, or Hawai'ian immersion programmes, are of this type. Cultural, ethnic identity is often focussed upon in these programmes and the link between language and ethnicity is strong.

The reasons and goals in using two or more languages in the educational system thus vary greatly, ranging then from increased knowledge and economic gain, to increased mutual understanding, to improved educational opportunity and outcome, to ethnolinguistic survival. Many programmes are multipurpose and combine several of the goals. Linguistic human rights and the minority languages and ethnicities as resources not handicaps are central for many of the most advanced programmes.

Whereas earlier biologically argued racism was used to legitimate unequal access to formal education and, later on, to good jobs, housing, etc, today ethnically and linguistically argued racisms, ethnicism and linguicism, are used. These can be defined as ideologies, structures and practices which are used to legitimate, effectuate and reproduce an unequal division of power and (both material and non-material) resources between groups which are defined on the basis of

  • 'race' (biologically argued racism)
  • ethnicity and culture (culturally argued racism: ethnicism)
  • language (linguistically argued racism: linguicism)

(Skutnabb-Kangas 1986)

Linguicism is a major factor in determining whether speakers of particular languages are allowed to enjoy their linguistic human rights. Ethnicism and linguicism are more sophisticated but equally efficient weapons as biological racism in committing ethnocide (ethnic genocide, the destruction of the ethnic socio-cultural identity of a group) and linguistic genocide.

Linguistic and cultural human rights are a prerequisite for preventing ethnic and linguistic genocide. Lack of these rights, for instance the absence of these languages from school time-tables, makes minority languages invisible. Alternatively, minority mother tongues are constructed as non-resources, as handicaps which are believed to prevent minority children from acquiring the majority language (= the only valued linguistic resource), so that minority children should get rid of them in their own interest. At the same time, many minorities, especially minority children, are in fact prevented from fully acquiring majority resources, and especially majority languages, by educational structures in which instruction is organised through the medium of the majority languages in disabling ways which contradict most scientific evidence.

Human rights, especially economic and social rights, must, according to human rights lawyer Katarina Tomaševski, act as correctives to the free market, overruling the law of supply and demand, in other words: they should guarantee that the basics needed for survival and for the sustenance of a dignified life have no price-tags and are outside market forces. The necessities for survival include not only basic food and housing, but also basic civil, political and cultural rights. Education, including basic educational linguistic rights, is one of the necessities from which price-tags should be removed. According to human rights principles, it is the duty of each government to create conditions for people to provide these necessities for themselves and to provide the necessities for those unable to do so themselves.

The fewer speakers a language has, the more necessary it is for the children to become high level multilinguals, in order to be able to get access to the basic necessities needed for survival. The mother tongue is needed for psychological, cognitive and spiritual survival -- cultural rights; all the other languages, including an official language in the state where the children live, for social, economic, political and civil rights. A child must be able to speak to parents, family, relatives, know who she is, acquire skills of how to think, analyse and evaluate. The mother tongue(s) is (are) vital for this. Further education, job prospects, being able to participate in the wiser society, require other languages. Thus high levels of multilingualism must be one of the goals of proper education.

Everybody, not just privileged elites or poor minorities, needs to be fluent and literate in at least two languages, preferably more. Everybody, not just minorities, needs to become aware of and acknowledge the importance of their ethnic and linguistic roots, in order to be able to develop, analyse, criticize, create and reflect. Language rights for all are part of human rights. Language rights are a prerequisite for many other human rights. Linguistic human rights in education are a prerequisite for the maintenance of the diversity on the planet that all of us are responsible for.

General bibliography

Baker, Colin (1993). Foundations of Bilingual Education and Bilingualism. Clevedon & Philadelphia: Multilingual Matters.

Crawford, James (1989). Bilingual Education: History, Politics, Theory and Practice. Trenton, NJ: Crane Publishing Company.

International Journal on Minority and Group Rights. Special Issue on the Education Rights of National Minorities 4:2, 1996/1997.

Cummins, Jim & Corson, David (ed.). (1999). The Encyclopedia of Language and Education. Volume Bilingual Education. Dordrecht: Kluwer.

Fishman, Joshua A. (ed.). Handbook of Language and Ethnic Identity. New York & Oxford: Oxford University Press.

García, Ofelia (ed.) (1991). Bilingual Education: Focusschrift in Honor of Joshua A. Fishman. Volume I. Amsterdam/Philadelphia: John Benjamins.

Pattanayak, D.P. (1981). Multilingualism and Mother-Tongue Education. Delhi: Oxford University Press.

Rubagumya, Casmir M. (ed.) (1990). Language in education in Africa: a Tanzanian perspective. Clevedon: Multilingual Matters.

Skutnabb-Kangas, Tove (Ed.) (1995). Multilingualism for all. Lisse: Swets & Zeitlinger.

Skutnabb-Kangas, Tove (2000). Linguistic genocide in education - or worldwide diversity and human righs? Mahwah, NJ: Lawrence Erlbaum Associates.

Back to ECMI Panel