Yonatan Tesfaye Fessha
LLB (Addis Ababa), LLM (Human Rights and Democratisation in Africa) (Pretoria), PhD (Western Cape)
Postdoctoral Research Fellow, Community Law Centre, University of the Western Cape, South Africa
Edition: AHRLJ Volume 9 No 2 2009
Pages: 501 - 523
Citation: (2009) 2 AHRLJ 501-523
Download article in PDF
Special thanks go to Professor Nico Steytler for the incisive comments he gave on earlier versions of this article.
The success of a federal arrangement in accommodating ethnic diversity cannot be measured solely on the basis of its language rights regime. However, it is generally agreed that a well-designed language rights regime goes a long way in contributing either to the effective reconciliation, unity and diversity or to the eventual polarisation of cultural communities. This article focuses on the challenges of adopting an inclusive language policy in multi-lingual states. Using two case studies, South Africa and Ethiopia, it examines the different policy alternatives for accommodating linguistic communities.
As language is often one of the key expressions of ethnic identity, language rights in a federal state are 'invested with a symbolism of its own'.  It represents the recognition (or the lack thereof) of the linguistic identities of the state's constituent units. As a result, a language policy often correlates with visions of uniformity or visions of diversity.
Of course, language policies go beyond the symbolic realm of recognition. The recognition and protection of language rights and linguistic identity form an important part of human rights. This assertion may, of course, appear at odds with the individual liberal position that provides the theoretical foundation for most international human rights instruments, including the United Nations (UN) Charter  and the Universal Declaration of Human Rights (Universal Declaration).  Major international documents, other than providing for a right against discrimination (based on language), do not mention group-specific rights.  The widely-recognised language rights in the context of 'fair trail and the due process of law' are similarly informed by this same individualist philosophy. 
Notwithstanding this individualistic orientation, the human rights instruments that impact language rights, either as part of a non-discrimination clause or through other specific rights, affect the power of the state in areas of language policy. In other words, the protection of many of the rights recognised in international human rights instruments has implications for states' language policies. This is because a language policy has the capacity to affect the enjoyment of other rights. A language policy that only promotes a single language group can have the effect of discrimination as it can create barriers to the exercise of voting, education and other rights.  As noted by De Varennes: 
If public authorities prevent the use of any language, including minority language, in private activities, this could potentially, depending on the type of intervention, breach a number of rights such as the right to private and family life, freedom of expression, non-discrimination, or the right of persons belonging to a linguistic minority to use their language with other members of their group.
A more controversial category of rights are language rights that mandate the use of different languages by public authorities. These language rights, which are also the major focus of this contribution, provide individuals what the principle of non-discrimination cannot provide, namely, the right to obtain government services through the medium of one's language. It is also these language rights that are relevant to the linguistic identity of minorities. Except for the widely-recognised right to use a minority language when required for the purposes of a fair trail and due process, however, most human rights treaties say very little about the use of language by state authorities. Even article 27 of the International Covenant on Civil and Political Rights (ICCPR), which attempts to address language-related claims, makes reference only to persons belonging to 'such minorities' and not to the groups themselves:
In those states in which ethnic or linguistic minorities exist, persons belonging to such minorities shall not be denied the right, in community with other members of their group, to enjoy their own culture, to profess and practise their own religion, or to use their own language.
This means that the right recognised by article 27 is an individual right that, in so far as language rights are concerned, only protects the use of a language as between private individuals.  The only two instruments that clearly impact on the use of a minority language by state authorities are the Framework Convention on the Protection of National Minorities and the European Charter on Regional and Minority Languages, both of which are regional instruments.  These two instruments provide for the so-called 'sliding-scale approach', which requires that 
[w]here public authorities at the national, regional or local level face a sufficiently large number of individuals who use a minority language, authorities must provide an appropriate level of services in this language.
The absence of general human rights treaties that contain provisions on the use of language by state authorities begs the issue of states' response to the use of language for the purposes of government. That is exactly what this contribution intends to address. It focuses on the challenges of adopting an inclusive language policy in multi-lingual states. Using two case studies, those of South Africa and Ethiopia, it examines the different policy alternatives for accommodating linguistic communities. It does this in four separate but interrelated parts. First, it discusses the two major language policies that are adopted by countries that are charcterised by linguistic diversity. It then proceeds to examine the South African constitutional approach to language rights. This is followed by an examination of the Ethiopian approach. The article concludes by bringing together the discussion on the experience of the two multi-linguistic states and identifying some of the key lessons, which may also assist other multi-linguistic states that are struggling with the similar challenges of adopting an inclusive language policy.
This contribution is decidedly narrow in its approach as it deals with the specific subject of language use by state authorities. This choice is informed by the purpose of the article, namely examining the relevance of language rights regimes to a state that seeks to respond to the exigencies of linguistic diversity. Issues of discrimination based on language and other aspects of language rights are not the focus of the article. These and other issues, like language in education policy, are discussed only to the extent that they are relevant to make a point to the main thesis of this article.
2. Brief comparative overview of language rights regimes
Multi-ethnic states adopt different language policies. Some adopt a policy modelled on the individualistic approach, or commonly known as the personality principle, to the whole issue of language rights, while others opt for a territorial model of language planning. Under the territorial approach, the rights to exercise language rights are confined within the defined territory of the minority language. Under the personality principle, by contrast, an individual speaker can, by and large, exercise language rights irrespective of his or her geographical location. As noted by Reaume: 
Neither model [of language policy] dictates very precisely a concrete language policy on language, and indeed the literature displays a distinct lack of precision in the use of the two models to illuminate the concrete polices adopted in various jurisdictions.
The aim of this brief comparative overview is, however, simply to place the discussion on language polices in Ethiopia and South Africa in a broader context, by indicating the general trend in countries facing similar challenges. For the purpose of this article, we will stick to the dichotomy of the personality and territorial principle as commonly understood.
2.1 Personality model
According to the personality principle, individuals are entitled to use their mother tongue in every part of the country with few territorial restrictions. This vests individuals with a right in their personal capacity, regardless of where they live. In other words, the linguistic preferences of speakers, wherever they reside, are given a central place in the regulation of language use. An important element of this principle is, however, the criterion 'where numbers warrant', which implies that 'language rights may be granted only when there are [a] sufficient number of particular language speakers to warrant language protection'.
The language policy adopted by Canada represents this approach. Under the Official Languages Act, passed in 1969, both French and English were granted official status for all purposes of the federal government.  As a result of these language rights, which were later included in the 1982 constitutionally-entrenched Charter of Rights and Freedoms, Canadian citizens are entitled to federal government services in either official language where there is a significant demand. Furthermore, the Constitution provides all Canadian citizens, where numbers warrant, the right to have their children educated at elementary and secondary levels in either official language. This policy entailed a symmetrical application from coast to coast, whereby French-speaking minorities outside Quebec and the English-speaking minority inside Quebec receive equal constitutional protection. 
The language policy adopted by Canada, which is designed as a means to achieve a nation-building objective, aims to separate linguistic differences from the collectivities, territories and institutions which constitute them. Although the predominantly French-speaking Quebec province, like the rest of the provinces in Canada, can adopt its own language policy, the application of the federal official languages policy, which is based on the personality principle, means that Quebec has achieved only limited territorialism.  In fact, as Rubio-Marin indicates 
'[t]he Charter has de-territorialized language rights by attaching them to speakers of both French and English rather than to provinces, recognizing thereby minority status to both the Francophone minorities outside Quebec and the Anglophone minority inside Quebec.
In this regard, the personality approach represents a reluctance to recognise the Francophone community as a territorialised linguistic community.  The 'de-territorialisation' role of language rights in promoting a monist vision of the Canadian state and rejecting the idea of a territorialised nation within the state of Canada is also evident in the following remarks of Pierre Trudeau, one of Canada's most influential Prime Ministers, also known for his view of pan-Canadianism: 
If minority language rights are entrenched throughout Canada, then the French-Canadian nation would stretch from Maillardville in BC to the Acadia community on the Atlantic Coast … Quebec cannot say it alone speaks for French-Canadians.
A common criticism directed against the personality approach is that it has the tendency to perpetuate the dominant position that a historically privileged language group enjoys in the state. It is likely to have the effect of strengthening the pressures for assimilation to the dominant group.  This is illustrated by (the fact that the secondary status of French has not changed with the adoption of the policy of official bilingualism) the secondary status of French in Quebec even long after the adoption of the policy of official bilingualism. In Quebec, the cultural division of labour was such that 'capital spoke English and labour spoke French', thus resulting in English occupying a disproportionate place in Quebec in relation to French,  that in a way encouraged people from the other group to assimilate to that language group, thus resulting in providing a disproportionate place to the historically dominant group.  Thus, despite the application of the federal policy of bilingualism, English continued to remain as the majority status language with French relegated to a secondary level. That prompted Quebec to embark upon what is often called 'the language normalisation process', the adoption of a series of language policies that are aimed at elevating the status of French within the province by reversing the disproportionate place it occupies in its own province.  As noted by Gagnon, 
[t]he decision on the part of the Quebec government to implement a series of laws to redress past practices created bad feelings outside Quebec, since it was felt that the federal policy of bilingualism had responded to Quebec's demands.
As the foregoing discussion shows, the personality approach to language planning follows a non-exclusive approach and allows individuals to use the language of their preference across the country. In short, the personality approach emphasises an individualistic orientation of the right. In such a vision, linguistic differences are individual attributes, protected from 'coast to coast' by a central state. The major characteristic of this model of language policy is that it, with its integral coast-to-coast bilingualism, forms part of Canada's failure to extend a symbolical recognition of its multi-national character, presenting a monist vision of the state in the symbolic realm.  It regards the state as one bi/multi-lingual nation. Although the personality principle may be feasible in a country like Canada where there are few linguistic communities, the capacity of this approach to give practical effect to the act of recognition in a multi-ethnic context is questionable. As indicated earlier, the tendency of this particular language policy to perpetuate the dominant position of a historically dominant language is also something that must be looked into. Without a deliberate intervention from the state or the constituent government, it is likely that the historically dominant language group will continue to remain as the majority status language with the languages of other groups relegated to a secondary level.
2.2 Territorial model
Other states have responded to the language problem by adopting the territorial model of language planning. Under such systems, the official language would often be that of the majority of the locality. Individuals have a right to services in that language only, regardless of what their mother tongue is. This often has the effect of promoting unilingualism although 'the connection between the territorial dimension of language policy and unilingualism is not a logical one'.
A good example of the territorial model is Belgium, where both French-speaking Flanders and Dutch-speaking Walloon endorse unilingualism with Brussels being the only region that has adopted official bilingualism. Individuals moving into the other parts of Belgium must assimilate. French-speaking Belgians moving to a Flanders territory will have to send their children to Flanders schools and vice versa. There is thus a strict policy of both territorial and individual monolingualism, implying that '[t]here are no all Belgian language rights'.  Similarly, in Switzerland, language guarantees are provided on the basis of the principle of territoriality. German-speaking Swiss moving to a French-speaking canton has to leave behind any prior claim to language protection. 
The territorial model of language policy represents a recognition of the linguistic identities of the constituent units. It also provides ample room for a community to develop its language and culture. Of course, the territorial model of language planning is not without problems. One of the palpable consequences of the territorial model in Switzerland is that 'many Swiss do not actually become multilingual'.  There is also a concern that this particular approach risks developing isolated communities and scores low in the promotion of inter-group solidarity. That might explain the description of Switzerland as a country composed of three groups that 'stand with their backs to each other'.  It must, however, be admitted that, under certain circumstances, such a language policy might be the only way to hold the state together. In the case of Belgium, for example, it is argued that Belgium would not have existed as one state today had such an arrangement not been made.  In Switzerland, too, the territoriality principle to language is considered to be instrumental in guaranteeing peace among the different language groups. 
A variant of the territorial model of language policy is adopted in Spain. The Castilian language has been the dominant language in Spain for centuries while other national languages were suppressed.  The 1978 Constitution made Castilian the official language of the central government, statewide, while making languages of the autonomous communities co-officials in their respective communities. It further imposes a duty on all citizens of Spain to learn Castilian and the right to use it. This limits the use of local languages to the activities of the autonomous communities thus, unlike the Canadian Constitution, denying any official status in relation to central institutions such as the government, congress, the administration and the courts. Although the sanction of the Castilian language as co-officials of the autonomous communities have played an important role in promoting national unity, it also has the adverse effect of further entrenching the disproportionate status the local language holds in its own area. As noted by Agranoff, 'the language issue has created the single most protracted policy conflict' in Spain. 
3. Comparing language rights regimes in South Africa and Ethiopia
Against the backdrop of the brief comparative observations provided above, the article proceeds to examine the language rights regime both in South Africa and Ethiopia. It commences to do so by looking at South Africa.
3.1 The South African experience
Section 6 of the South African Constitution regulates the use of language. It determines the official language of the Republic as well as the provinces and municipalities. Like the interim Constitution, section 6 of the Constitution recognises 11 languages as the official languages of the Republic. The conferring of official status on all 11 languages sends the message that all linguistic groups are regarded equal by the South African Constitution. Symbolically, it reinforces the normative guide set by the Preamble that 'South Africa belongs to all who live in it'. 
Section 6 of the Constitution departs from its counterpart in the interim Constitution in important ways. First, the stipulation of the interim Constitution which mandated the government to 'create conditions for the equal use and enjoyment' of all official languages is, to some extent, qualified in the 1996 Constitution by the introduction of a preferential treatment clause that applies in relation to some of the official languages. The Constitution, under section 6(2), emphasises 'the historically diminished use and status of the indigenous languages' and mandates the government to 'take practical and positive measures to elevate the status and advance the use of these languages'. The special treatment afforded to previously marginalised indigenous languages is further entrenched in the Constitution as it, under section 6(4)(2), subjects the enjoyment of 'parity of esteem and equitable treatment' of all official languages to 'the state's obligation with regard to indigenous languages'. Obviously, 'equitable treatment' in this context does not mean equal treatment. As Currie aptly notes, it is 'a treatment that is just and fair in the circumstances'.  This, at least, means that a language policy has to take into account the structural inequalities of the different languages spoken in the country, entailing a special treatment of some of the official languages.
The special treatment of previously marginalised indigenous languages is even symbolically expressed in the manner in which the 11 official languages are listed in section 6(1). Unlike in the interim Constitution, in which the official languages are listed alphabetically, the 1996 Constitution, under section 6(1), lists the 11 official languages, starting with the language that lacks widespread usage and ending with the one that enjoys extensive usage.  In other words, usage informs the listing order of the 11 official languages. This symbolic expression can, in fact, serve as a guide for interpretation. Strydom argued that '[t]he purpose of the structure is to change the order preference in a deliberate attempt to give textual prominence to languages lacking widespread usage'. 
Second, the language clause of the 1996 Constitution introduces a plethora of considerations that were not included in the interim Constitution and which must now be taken into account when the different spheres of government decide their official languages.  The newly added considerations are usage, practicality, expense, regional circumstances, and balancing the needs and preferences of the population. A third point of departure is that the interim Constitution stipulation, which prohibited the downgrading of rights relating to languages and the status of languages existing at the commencement of the interim Constitution, is omitted. That specific clause was of special concern to the Afrikaner community who feared the marginalisation of their language in post-apartheid South Africa.
3.1.1 For the purposes of government
The use of language for the purposes of government is a major manifestation of the officialisation of a language.  According to section 6(3), which outlines the use of language for the purposes of government, the national and provincial governments can select any of the official languages for the purposes of their administration. It mandates both national and provincial governments to consider the factors listed in section 6(3) and to adopt at least two official languages for the purposes of government.  Their decision to use any of the official languages must be based on 'usage, practicality, expense, regional circumstances and the balances of the needs and preferences of the population as a whole or in the province concerned'.  In this regard, local governments are spared from the complexities of these considerations as they are only required to take into account language usage and the preferences of their residents. The subsection injects a minimum condition by enjoining the national government and each provincial government to use at least two official languages.  In South Africa, where the different ethnic groups are relatively geographically concentrated, the regional preference to language usage provides ample opportunity to promote regional languages and to facilitate the promotion of self-management of ethnic communities.
The 1996 Constitution also maintained the Pan-South African Language Board which was established by the interim Constitution. As stated in the interim Constitution, the Board is mandated to promote and create conditions for the development and use of all official languages.  The list of languages which the Board is mandated to promote and develop is, however, amended to include the Khoi, Nama and San languages, as well as sign languages.  In addition, the Board is entrusted with the additional task of promoting and ensuring respect for all languages commonly used by communities in South Africa, including German, Greek, Gujarati, Hindi, Portuguese, Tamil, Telegu and Urdu.  This duty of the Board is also extended to languages used for religious purposes in South Africa, including Arabic, Hebrew, Sanskrit and other languages.  The 1999 South African Language Board Amendments Act added the responsibility of preparing a dictionary for all 11 official languages.
The conferring of official status to all 11 languages is criticised by some as counterproductive. They argue that this policy is not practically realisable and may eventually result in unilingualism; they consider the policy as 'impractical egalitarianism'.  This fear is compounded by the fact that the Constitution, as mentioned earlier, subjects the equal treatment and use of all 11 languages to a plethora of practical considerations. Although scholars like Alexander concede to the unavoidability of the use of such 'safety clauses', he strongly warns that clauses which are 'allegedly based on technical and economic grounds, are more usually the perfect loopholes for reducing the principle of equal treatment to mere lip service'. 
The official recognition of all 11 languages as equal at the national level could be considered as a reflection of a state that emphasises national unity by assuring that all language groups have a place in South Africa, while the regional preferences in language usage represent a recognition of the need to provide for regional languages. Theoretically, the Constitution introduces a variant of the territorial model of language planning at a provincial level. Unlike the traditional territorial model, however, it does not simply grant an official status to the language of the majority of the locality and limit the use of other local languages. It rather allows provincial governments to consider the factors listed in section 6(3) from a provincial context and to adopt at least two official languages for the purposes of provincial government.  The province of the Western Cape, which is predominantly inhabited by Afrikaans and Xhosa speakers, has, for example, adopted three official languages: Afrikaans, isiXhosa and English. The decision not to advance one particular language but to, at least, recognise two official languages at the provincial level is an important recognition of intra-provincial diversities. Yet, this same aspect of the provincial language policy portrays a state that discourages the identification of a single language with a particular territory and promotes social cohesion and national unity through its language policy. The language clause of the Constitution has thus the dual role of promoting national unity and accommodating ethnic diversity.
It is, however, important to note that a reading of section 6(3) does not reveal which of the considerations listed therein should be given paramount importance when determining an official language. In the absence of such clear guidance, it all depends on the level of importance that policy makers attach to these determining factors. For those that stress practical and economic considerations, the official multilingualism policy serves no purpose beyond 'a symbolic gesture'. A recommendation by the Institute of Chartered Accountants in June 2000 suggested that Afrikaans be abolished and English becomes the sole medium of communication, training and examination. The Institute's annual expenditure of R600 000 on translation, reproduction and printing is cited as the main reason behind the recommendation.  From the perspective of ethnic accommodation, on the other hand, practical considerations should not be used as an excuse to trample on the constitutionally-sanctioned official multi-lingualism. This perspective underemphasises the considerations of 'practicality … and expense' in the use of official languages. It rather stresses the importance of the constitutionally declared official multi-lingualism and the normative guidelines that declare the enjoyment of 'parity of esteem and equitable treatment of all official languages'. 
In practice, government seems to have given considerable weight to practical considerations. English has become the lingua franca of government administration to the extent that the policy of multi-lingualism adopted by the Constitution has only come to represent a mere symbolic value. English has become the language for internal and external communication in government departments.  Even when members of the public communicate with government in a language other than English, government departments invariably respond in English.  Both in the National Assembly and the National Council of Provinces, the dominance of English is clear. Major policy documents are often produced only in English. The situation is no different in the courts. The constitutional promises relating to the parity of 11 languages are not given effect to. 
The establishment of the Pan-South African Language Board has not, as yet, had any significant impact on ensuring the implementation of the official language clause. On many occasions, the Board, after investigating complaints with regard to the violation of language rights, has found that the language policies and practices of the government and state corporations violate the Constitution. The findings and recommendations of the Board, however, often fall on deaf ears. The government is broadly criticised for failing to give adequate attention to the Board. On one occasion, the Board felt compelled to write an open letter to President Nelson Mandela criticising and expressing its concern about the tendency towards monolingualism at all levels of government. 
Generally, the discussion on the use of language for the purposes of government reveals a trend that reinforces the suspicion of the critics of the officialisation of all 11 languages. The officialisation of the 11 languages might send the symbolic message that all groups are regarded equally in the public sphere. However, this symbolic message has not been given practical effect. Despite the multi-lingual reality that characterises South African society and a Constitution that declares official multi-lingualism, monolingualsim seems to be the emerging trend. 
3.2 Ethiopia's experience
Article 5 of the Ethiopian Constitution, outlining the basic principle of the language policy, declares that all Ethiopian languages shall enjoy equal state recognition. This is further elaborated by article 39 of the Constitution, which states that every ethnic group in Ethiopia 'has the right to speak, to write, and to develop its own languages; to express, to develop and to promote its culture; to preserve its history'. Based on these constitutional principles, the Constitution declares that Amharic shall be the working language of the federal government while allowing the states to determine their respective working language.  The federal offices that operate in the states employ Amharic as the language of communication.
From the outset, it is important to note that the Constitution, faced with an ocean of linguistic diversity, has opted not to adopt an official language or languages. It has rather opted for a 'working language'. Symbolically, this is obviously designed to avoid the impression that a particular language is favoured above any other at the symbolic level. The Ethiopian system adopted Amharic as the language of government (federal) business without conveying the message that the adopted language is dominant over others. The success of the system in overcoming the dilemma that it tries to circumvent is, of course, something that can be debated. As we shall see in the following paragraphs, there are sections of society that regard the continued use of Amharic at the federal level as a continuation of their marginalisation and the perpetuation of past policies that subordinated all other languages to Amharic.
The constitutional stipulation that allows each regional state to adopt its working language opens a room for the application of a territorial model of language planning, in which case the working language of each member of the federation would be that of the majority of the area.  In practice, five of the nine regional states have endorsed unilingualism. This obviously provides ample room for each ethnic community to develop its language and culture. It also represents recognition of the linguistic identities of the constituent units. An important consequence of this policy is that individuals moving into either of these regions must assimilate. That means that Amharic-speaking citizens moving to an Oromifa-speaking region have to leave behind any prior claim to language protection. As we shall see later, this has created a problem in some areas where an important number of minorities are scattered in the midst of regionally-dominant linguistic groups, especially in major urban areas of some of the member states. It is, however, important to note that the ethnically plural regional states have opted to retain Amharic as their working language. To be precise, three of the four multi-ethnic states (ie the SNNPR, Benishangul and Gambela regional states) have decided to retain Amharic as their working language.
3.2.1 Debate on language policy
This language policy has provoked criticism both from centrifugal and centripetal forces. On the one hand, there are sections of society that regard the adoption of Amharic as the working language of the federal government as 'little more than the continued endorsement of the superior position of the language, and the sections of society associated with it, by the Ethiopian state'.  For these sections of society, the policy is a threat to their ethnic self-determination rights.  It also undermines the constitutional principle that all languages are equal.  The Oromo Federal Democratic Movement, for example, has opposed the sole use of Amharic as the working language of the federal government and calls for the adoption of Oromifa as the working language of the federal government. 
The more vocal criticism comes from the proponents of the idea of Ethiopian nationhood who want to use language as a unifying factor. They criticise the position of the Constitution on language as an attempt to create 'the biblical tower of Babel' in Ethiopia. If that was not the intention, they argue, the drafters of the Constitution would have opted to encourage the use of Amharic, ultimately developing it as the national language.  According to this argument, Amharic could serve 'as an important instrument for the eventual creation of greater cohesion among Ethiopians in language and in a sense of common national destiny as one people'.  The case of India is often invoked to support this line of argument. Haile-Selassie remarks that 'the role of English as a common language among the diverse linguistic groups in India has tremendously assisted in the development of a national consciousness in that country'.  The proponents of this view recommend the Russian and Spanish model of language planning, where the Russian and Castilian languages are respectively used along the languages of the constituent units.  It is similarly contended that non-Amharas, owing to the decreased amount of formal education they receive as a result of the language policy, will effectively lose access to the state apparatus, further disconnecting them from a sense of Ethiopian identity. 
The position of the Constitution on the use of language marks a clean break with the past during which Amharic enjoyed a superior position throughout the country. This is a major departure from the historical pattern in which 'the distribution of the political goods of communication, recognition and autonomy has been highly skewed, benefiting native Amharic-speakers disproportionately'.  Of course, the special place of Amharic in the Ethiopian linguistic landscape has not vanished entirely. It is also true that the retention of Amharic as the language of national communication can appear to some as the continuation of the Amharic hegemony. However, even if one cannot deny the symbolic implications of the retention of Amharic as a federal language, its continued use can hardly be associated with deliberate symbolic dominance. The decision to keep Amharic as the federal working language is no more than a reflection of the position that the language has attained as 'an effective means of national communication'.  This is also evident from the fact that it is not labelled as an official but rather as a working language of the federal government.
Using Amharic along with the regional languages as co-official language in all regional states might, as some argue, help to promote the relationship between the different linguistic groups. This view, however, belies the structural imbalance that exists between Amharic and the other languages and the effect that this imbalance may have on the development of the latter. As indicated in previous sections, Amharic, to the exclusion of all other languages, has been the language of government business for decades. This provides Amharic with a unique position in terms of language status, which other languages would be hard-pressed to compete with. Even in states where the speakers of other languages are in the majority, there is no guarantee that a policy of co-official languages will manage to avoid the dominance of the Amharic language. Without, at least, some kind of 'normalisation of language policy', the regional language will in all likelihood be relegated to a secondary status.
Moreover, even if one accepts the instrumentality of the Amharic language in bringing different ethnic communities together, it is not clear if it has to be given an official status both at the federal and state level. As noted above, Amharic is now the working language of the federal government in which all government business is conducted. Obviously, any communication between the federal government and a member state or between two member states will be conducted in Amharic. Furthermore, with the view to promote the language as the language of national communication, Amharic is being provided as a subject in almost all primary schools throughout the country. It is also important to note that almost all ethnically plural regional states, with the sole exception of Harari, have opted for Amharic as their working language. Generally, Amharic is still given precedence over all other languages.  This means Amharic can still serve as a cohesive force by facilitating communication between and among the different ethnic groups.
It is also not clear if those like Haile-Selassie that criticise the present language system based on the Indian model have really grasped the Indian system. Their criticism rather reflects an incorrect appreciation of the Indian system. It is true that the adoption of English as a language of government business (ie associate additional official language) has facilitated communication between the different ethnic groups in India. Underlying the Indian and, as a matter of fact, some African states' decision to adopt English as their official language, is the very fact that English has a unique neutral status compared to that of other local languages. It is this factor that often motivates the use of English and not other local languages as languages of government business. As Schmied observes, this is specifically true in most decolonised states: 
Ethnic languages are normally not accepted as national languages wherever other groups fear 'tribal dominance' and prefer English, which is 'tribally neutral'. Only tribally neutral
lingue franche have any chance of taking over certain functions from English as national languages.
The decision made by India not to adopt Hindu as a 'working language' of the national government was underlined by the fact that the adoption of Hindu would portray the dominance of the Hindu-speaking group and the relegation of others to a secondary status. It is also important to note that the Indian system recognises 22 state languages. Thus, in addition to allowing the constituent units to adopt their own regional languages, the adoption of the Indian model would have resulted in the 'officialisation' of a 'culturally neutral language'. Amharic, obviously, does not enjoy a neutral status among the different ethnic groups in Ethiopia. Yet, despite this fact, a decision has been made to maintain Amharic as the federal working language. Furthermore, as indicated earlier, it is regarded as the language of national communication and, as a result, it is being taught as a subject in primary schools in non-Amharic speaking parts of the country. Generally, if the suggestion for the Ethiopian system is to emulate the Indian model, one can reasonably argue that the present system provides more than what the Indian system has to offer.
Finally, we turn to the argument that non-Amharic speakers will lose access to the state apparatus as a result of the language policy. It is not at all clear how the language policy will have the effect of compromising the capacity of individuals from a non-Amharic-speaking group to access the state, thereby continuing their historical marginalisation. In fact, the reverse seems to be true in present-day Ethiopia. Regional state government as well as administrative units within each regional state are run and staffed by members of the language group that is dominant in the regional or sub-regional government unit. The new dispensation has opened more opportunities for employment 'to sons of the soil'. As we shall see later, this is also the case at the federal level. More than ever, one can easily observe the appointment of individuals from extremely diverse linguistic background in the different institutions of the national government, including the cabinet. In fact, fluency in Amharic does not seem to be an obstacle in assuming higher offices of the federal government. It is not uncommon to come across Ministers for whom Amharic obviously is clearly not their mother tongue.
4. Towards an inclusive language rights regime
A language rights regime that operates within the context of a multi-ethnic federation should represent a recognition of the linguistic identities of the constituent units. This entails the framing of the language rights regime as a concrete expression of the federalist principle and attempting to achieve a delicate balance between unity and diversity.  It involves the adoption of a language policy that enables cultural communities to promote their language and cultural identity while at the same time promoting inter-ethnic solidarity.  In this regard, the recognition of all languages as equal is an imperative element of any state that seeks to recognise ethnic diversity. Beyond that, however, there is no definite answer on determining the official language(s) of the federal as well as subnational governments.
In terms of the federal language, the options are either to promote particular language(s) or, as in the case of South Africa, regard all languages spoken by the different ethnic groups as official languages of the country. The South African option is obviously viable in a country with few linguistic groups. As it is evident from the South African experience, a country with more than at least ten ethnic groups cannot, for example, expect to practically realise the usage of all languages in all or most business of the federal government. Such kind of policy, as again proved by the experience of South Africa, is an 'impractical egalitarianism'.  Despite the multilingual reality that characterises South African society and a Constitution that declares official multi-lingualism, monolingualism is the emerging trend. This shows that the South African approach will result, more often than not, in a situation where a particular language becomes the 'unofficial official language' of the state. In that case, a mere recognition of all the languages spoken in the country as official languages will only have a symbolic value. In addition, the policy is bound to create discontent among some ethnic groups unless the 'unofficially official language' is a culturally neutral language, as English is for most South Africans  and decolonised states. This is also the only situation where a state can adopt a particular language as official language without provoking a hostile reaction from other ethnic groups.
Adopting a national language remains, however, a challenge in states such as Ethiopia that do not have the benefit of a culturally neutral language. One option in this context is to select a particular language, which in most cases would be the historically dominant language, as the language of government business without bestowing it with the status of an official language. The Ethiopian approach which recognises Amharic as the working language represents this option. This approach seemingly contrasts with the South African model that recognises all languages as official languages. Like the South African model, however, it is underlined by the same principle that recognises all linguistic groups as equal. The difference lies in the way the two systems give expression to this same principle.
Under the working language approach, the selected language will become the working languages of the federal government in which all tasks of government are conducted. Of course, as in the case of the use of Amharic in Ethiopia, the symbolic implication of adopting a particular language or retaining a historically dominant language as the working language of the federal government cannot be easily disregarded. The solution lies in convincing the different ethnic groups that the particular language, as the title suggests, is adopted not to reflect the hegemony of the speakers of that language, but because of the special position that the language has attained as an effective means of national communication. This, however, only becomes practicable when the state demonstrates its commitment to the equality of all languages by adopting some form of territorial model of language planning, expressing regional preferences in language usage.
Unless in a bi-ethnic state where the individual model, which allows citizens to use their language in every part of the country, can serve the same purpose, the territorial approach to language, whereby each region adopts its language(s), is the language planning model that seems to provide effective institutional reality to the act of recognition. Under this model, the subnational units are allowed to adopt regional language(s). This does not necessarily mean promoting unilingualism. As is the case in South Africa, sub-national states that are inhabited by more than one ethnic group can, to the extent possible, recognise intra-substate linguistic diversities by recognising more than one language as working languages of the sub-national government. This would not only represent recognition of intra-substate diversities, but also portray a state that promotes social cohesion and national unity by avoiding the association of a particular territory with a single language. This option may not, however, be appropriate in a situation where a sub-national state is composed of, for example, not less than five ethnic groups. As the experience of ethnically plural sub-national units that adopted Amharic as their working language in Ethiopia suggests, adopting a language that is culturally neutral in the context of the relevant sub-national state is often the only way out. Yet, the system can be used to allow each sub-national state to use its institutional and territorial structure to reflect its linguistic diversity.
Related to this is whether sub-national units should adopt the national official/working language as co-official at the sub-national level. Of course, adopting the national language as co-official promotes social cohesion, especially in countries where there is extensive movement of citizens across internal borders. However, subscribing to this view does not necessarily require the co-officialisation of the national language. The idea of using language as a method of social cohesion can be promoted, for example, by ensuring that children, as in the case of Ethiopia, learn the federal language as a subject in their primary education. The argument for co-officialisation of the federal language at the regional level can have currency only in a situation like in Ethiopia where there are a large number of geographically dispersed ethnic migrants, especially in urban areas, who would be disadvantaged when government business is conducted in the language of the regionally empowered group.
The co-official language policy is not, however, without problems. The problem with this policy is that it has the tendency to promote the hegemonic status that a historically privileged language group enjoys. This is clearly the case in Ethiopia where the adoption of Amharic as co-official language at the regional level would have the effect of maintaining the historically dominant position of that language with the regional language in all likelihood occupying a disproportionate place. This is also supported by the experience of Quebec in Canada where the co-official policy perpetuated the dominant position of English, necessitating the province to embark on what is called 'the language normalisation process'. In a country where the designated federal language is not culturally neutral and where the nationally designated language is a historically dominant language, the co-official policy at the regional level is thus likely to perpetuate the dominant position that the latter enjoys in a state. Without state intervention, the historically dominant language will continue to remain as the majority language status relegating the regional language, albeit numerically dominant, to a secondary level.
In a sub-national state where there are large numbers of ethnic migrants, however, the adoption of the co-official policy seems unavoidable if the system is to accommodate ethnic diversity. The dangers that the co-official policy might pose on the status of the regional language can be mitigated by allowing the sub-national state to adopt what is called 'the language normalisation processes'. As the experience of Quebec in Canada shows, the basic aim of these processes is to restore and maintain the majority status that local languages should assume in their localities. As the experience of Ethiopia suggests, in a country like Ethiopia where there are a large number of ethnic migrants, the absence of a co-official policy easily causes strain on inter-ethnic relationships and runs the risk of alienating particular ethnic groups.
In conclusion, it is true that the success of a federal arrangement in accommodating ethnic diversity cannot be measured solely on the basis of its language rights regime. It is, however, generally agreed that a well-designed language policy goes a long way in contributing either to the effective reconciliation of unity and diversity or to the eventual polarisation of cultural communities and further disintegration of the state.
 P Coulombe 'Federalist language policies: The case of Canada and Spain' in J Tully & A Gagnon (eds) Multi-national democracies (2001) 242.
 The Charter makes no reference to group-specific rights. It simply recognises individual rights. This individualist outlook is, in fact, made clear in the opening article, art 1(3), which outlines the purpose of the UN. According to this article, one of the purposes of the UN is to encourage 'respect for human rights and for fundamental freedoms for all without distinction as to race, sex, language or religion'. The reference to 'fundamental freedoms for all' is interpreted to mean fundamental freedom for all individuals and not groups.
 The Universal Declaration declares that 'everyone is entitled to all the rights and freedoms set forth in this declaration, without distinction of any kind, such as … language'.
 In addition to the UN Charter and the Universal Declaration, art 1 of the UNESCO Convention against Discrimination in Education of 1960, arts 2(1) and 26 of the International Covenant on Civil and Political Rights, art 2(2) of the International Covenant on Economic, Social and Cultural Rights, art 1 of the American Convention on Human Rights, and art 2 of the African Charter on Human and Peoples' Rights provide for the right against discrimination on grounds of language.
 According to arts 14(3)(a) & (f) of ICCPR, an accused has the right to be 'informed promptly and in detail in a language which he understands of the nature and cause of the charge against him'; and is to have 'the free assistance of an interpreter if he cannot understand or speak the languages used in court'. Similar rights are provided in arts 5(2) and 6(3)(a) of the European Convention for the Protection of Human Rights and Fundamental Freedoms and art 8 of the American Convention on Human Rights.
 R Dunbar 'Minority language rights in international law' (2001) 50 International and Comparative Law Quarterly 90-120.
 F de Varennes 'Linguistic identity and language rights' in M Weller (ed) Universal minority rights (2007) 258.
 B Bowring 'Multi-cultural citizenship: A more viable framework for minority rights?' in D Fottrell & B Bowring (eds) Minority and group rights in the new millennium (1999) 6-9. See also De Varennes (n 7 above) 298.
 Art 10(2) of the Framework Convention on the Protection of National Minorities provides that '[i]n areas inhabited by persons belonging to national minorities traditionally or in substantial numbers, if those persons so request and where such a request corresponds to a real need, the parties shall endeavour to ensure, as far as possible, the conditions which would make it possible to use the minority language in relations between those persons and the administrative authorities'. Similar rights are provided in art 10 of the European Charter on Regional and Minority Languages.
 De Varennes (n 7 above) 298. Other major international instruments that make reference to language rights, though not always in legally-binding form, include the UN Declaration on the Rights of Persons Belonging to National or Ethnic, Religious and Linguistic Minorities; the UN Declaration on the Rights of Indigenous Peoples; the Vienna Declaration and Programme of Action; the European Charter for Regional and Minority Languages; the Central European Initiative Instrument for the Protection of Minority Rights; and the Framework Convention for the Protection of National Minorities.
 DG Reaume 'Beyond personality: The territorial and personal principles of language policy reconsidered' in W Kymlicka & A Patten (eds) Language rights and political theory (2003) 271.
 GM Balmer 'Does the United States need an official language? The examples of Belgium and Canada' (1992) 2 Indiana International and Comparative Law Review 445.
 Balmer (n 12 above) 446.
 M Chevrier 'Language policy for a language in exile' in P Larrivee (ed) Linguistic conflict and language laws: Understanding the Quebec question (2003) 155.
 R Rubio-Marin 'Language rights: Exploring the competing rationales' in Kymlicka & Patten (n 11 above) 60.
 Quebec has continuously been demanding to be recognised as a 'distinct society' in Canada. This was the main agenda in the two rounds of constitutional negotiations: the Meech Lake Accord in 1978 and the Charlottetown Accord in 1990. Recognition of Quebec as a 'distinct society' would have affirmed Quebec's dualist vision of the state, though set out only in the unenforceable paragraphs of a constitutional Preamble. The Quebec government's demand to be recognised as a 'distinct society' has not been successful. See S Tierney Constitutional law and national pluralism (2004) 238; see also R Simeon & L Turgeon 'Federalism, nationalism and regionalism in Canada' (2006) 3 Revista d'Estudis Autonomics I Federals 12. Gagnon & Herivault argue that the outcome of the Meech Lake debate has made the 'distinct society' compromise obsolete and no longer good enough to a majority of Quebecoise. For them, the only acceptable form of recognition today would be the inclusion of a clause recognising the 'Quebec nation' in the Preamble of the Canadian Constitution; A Gagnon & J Herivault 'The unresolved recognition of Quebec' paper delivered at a colloquium on Separatism in Canada: Past, present and future', Institute of Commonwealth Studies/Institute for the Study of the Americas, University of London, London, 4 November 2005 http//www.cst.ed.ac.ut/document (accessed 20 September 2009).
 D Karmis & AG Gagnon 'Federalism, federation and collective identities in Canada and Belgium: Different routes, similar fragmentation' in J Tully & A Gagnon (eds) Multi-national democracies (2001) 154-155.
 Karmis & Gagnon (n 17 above) 155.
 See Balmer (n 12 above) 446. See also Coulombe (n 1 above) 248.
 The Commissioner of Official Languages, in his 1995 Report, noted this fact. The Report concluded that French still had a disproportionate place in relation to English, even in the federal administration where progress in the use and status of French is more visible. The Commissioner concluded that French did not achieve 'a fair status as a language of services and work'. The 1996 Census also revealed that 'the historical pattern of assimilation among francophone minorities has not yet been overcome'; Karmis & Gagnon (n 17 above) 155.
 The Charter of the French Language in Quebec, which is famously known as Bill 101, is a good example that illustrates this situation. Adopted by the Parti Quebecois government in 1977, Bill 101, following the territorial model of language planning, sought to promote the use of French and at the same time restrict the use of English. It obliges both immigrants and Canadians moving to Quebec to send their children to a French school and mandated the display of commercial signs in French only. The court decision abrogated part of this legislation. The Supreme Court in 1979 decided that provisions making French the only official language of legislation and justice violate sec 133 of the British North America Act, 1867, which guarantees legislative and judicial bilingualism in Quebec. Part of the law that restricted the rights to education in English was struck down, entitling not only people who had been educated or whose parents had been educated in English in Quebec, but also those who had been educated in English elsewhere in Canada, to have their children receive education in that language. The Court in 1988 also struck down the rule that imposes French as the only language to be used on commercial signs. See generally K Swinton 'Federalism, the Charter and the courts: Rethinking constitutional dialogue in Canada' in K Knopf et al (eds) Rethinking federalism: Citizens, markets and governments in a changing world (1995) 294-315. See generally J Tully Strange multiplicity: Constitutionalism in an age of diversity (1995) 175.
 A-G Gagnon 'Manufacturing antagonisms: The move towards uniform federalism in Canada' in B de Villiers Evaluating federal systems (1994) 127.
 Simeon & Turgeon (n 16 above) 27. See also Karmis & Gagnon (n 17 above) 152.
 M Keating Plurinational democracy: Stateless nations in a post–sovereignty era (2001) 128.
 T Fleiner 'Switzerland: Constitution of the Federal State and the Cantons' in L Basta-Fleiner & T Fleiner (eds) Federalism and multi-ethnic states: The case of Switzerland (2000) 103-145.
 R Watts 'Language policies and education in Switzerland' in R Watts & J Smolicz (eds) Cultural democracy and ethnic pluralism: Multi-cultural and multi-lingual policies in education (1997) 271-30. See also F Grin 'Language policy in multi-lingual Switzerland: Overview and recent developments' paper presented at the Cicle de confèrencies sobre política lingüística Direcció general de política lingüística, Barcelona, Spain, 4 December 1998.
 J Steiner 'Switzerland and the European Union: A puzzle' in M Keating & J McGarry (eds) Minority nationalism and the changing international order (2001) 137-154.
 Balmer (n 12 above) 443.
 Fleiner (n 25 above) 103-145.
 Spain, after almost four decades of Franco's highly-centralised and homogenising regime, adopted a new Constitution in 1978. During the Second Spanish Republic (1931-1938), which was considered by many as a progressive government, Catalonia, the Basque country and Galicia were allowed to enjoy some level of autonomy. This was, however, short-lived. The Franco regime, which came to power in 1936, emphasised unity and condemned all forms of cultural diversity. The regime suppressed all regional political institutions and laws. It also prohibited the use of Catalans and Basques (Euskera) languages and all sorts of symbolic elements (flags, anthems) of the Catalan and Basque identities. See M Guibernau 'Between autonomy and secession: The accommodation of minority nationalism in Catalonia' in A Gagnon et al (eds) The conditions for diversity in multi-national democracies (2004) 122.
 R Agranoff 'Asymmetrical and symmetrical federalism in Spain: An examination of intergovernmental policy' in De Villiers (n 22 above) 73.
 Another language-related clause of the Constitution, though individualistic in its orientation, is sec 9(3) which states that 'the state may not unfairly discriminate directly or indirectly against anyone on one or more grounds, including … language'. In addition, sec 30 provides that '[e]veryone has the right to use the language and participate in the cultural life of their choice, but no one exercising these rights may do so in a manner inconsistent with any provision of the Bill of Rights'.
 I Currie 'Official languages' in M Chaskalson et al (eds) Constitutional law of South Africa (2002) 37.1-37.15.
 HA Strydom 'Minority rights issues in post-apartheid South Africa' (1997) 19 Loyola of Los Angeles International and Comparative Law Journal 873-914.
 Strydom (n 34 above) 898.
 As above.
 As aptly argued by Strydom, 'officialising a language is meaningless unless that language is used in all or most of the primary tasks of government — legislative, executive and judicial'; HA Strydom International standards for the protection of minorities and the South African Constitution http://www.fwdklerk.org.za/download_docs/02_05_Int_Standard_Minorities_Publ_PDF.pdf (accessed 20 March 2007).
 It is not, however, clear if this variant of territorial model automatically applies to national departments operating in the provinces.
 Sec 6(3) South African Constitution.
 According to one interpretation, the minimum number of languages to be used for purposes of government is three and not two: English and Afrikaans (as the spirit of the Constitution precludes that their status be diminished) plus at least one African language because the state must, in terms of sec 6(2), take practical positive measures to elevate the status and advance the use of these languages. M Kriel 'Approaches to multi-lingualism in language planning and identity politics — A critique' http://general.rau.ac.za/sociology/kriel.pdf (accessed 20 March 2006).
 Sec 6(5) South African Constitution.
 Sec 6(5)(a) South African Constitution.
 Secs 6(5)(b)(i) South African Constitution.
 Sec 6(5)(b)(ii) South African Constitution.
 V Sacks 'Multi-culturalism, constitutionalism and the South African Constitution' (1997) SA Public Law 683.
 N Alexander 'Language and the national question' in W James & G Maharaj (eds) South Africa: Between unity and diversity (1998) 16.
 It is not, however, clear if this variant of territorial model automatically applies to national departments operating in the provinces.
 Kriel (n 40 above).
 LT du Plessis & JL Pretorius 'The structure of the official language clause: A framework for its implementation' (2000) 15 SA Public Law 505-526.
 Very recently, the decision of the Western Cape Provincial Police Department that only English be used for all internal, including radio, communication has faced protest from a group Afrikaans-speaking police and also among the wider Afrikaans-speaking community. Some threatened legal action against the province's police language policy. The authorities said that their aim was to improve communication between the province's different language groups by encouraging the use of English. The new policy was later removed; See News24.Com 'FW: Afrikaans is under threat' http://www.news24.com/News24/South_Africa/News/0,9294,2-7-1442_2060508,00.html (accessed 20 March 2007).
 Strydom (n 34 above).
 After examining the use of African languages in courts, Hlophe (2001:94) concluded that '[t]he courts continue to lay too much emphasis on practical considerations. Practical considerations in effect are convenience to the presiding judicial officer! The noble goal of language parity will remain elusive as long as the courts continue to adopt this approach, and the legacy of English and Afrikaans as the sole court language will continue … In the result, indigenous African languages are undermined.' JM Hlophe 'Official languages and the courts' (2001) 11 South African Law Journal 690-696.
 H Giliomee 'The rise and possible demise of Afrikaans language' PRASEA Occasional Papers 14 (2003).
 The practice of monolingualism with its promotion of English as the sole language of communication has caused an outcry from communities, especially the Afrikaner community. In an open letter addressed to President Thabo Mbeki, 24 prominent speakers of Afrikaans complained that 'the South African government's commitment to a philosophy of multi-lingualism and cultural pluralism was paying lip service only, as was the commitment to the promotion of the African languages, including Afrikaans' Insig 1999:24. English, they claimed, is what is actually being promoted. The dominance of English is, in fact, conceded by the government. The Minister of Arts, Culture, Science and Technology, when establishing a Language Plan Task Group, noted the increasing tendency towards unilingualism despite the multi-lingual reality that characterises South African society and a Constitution that declares official multi-lingualism. See Department of Arts, Culture, Science and Technology 'Towards a national language plan for South Africa' Final Report of the Language Plan Task Group http://www.dacst.gov.za/arts_culture/language/ langplan/contents.htm (1996) (accessed 20 April 2006).
 Art 5 Ethiopian Constitution.
 The state-based federal offices use Amharic for government business.
 G Cohen 'Identity and opportunity: The implications of using local languages in the primary education system of the Southern Nations, Nationalities and People's Regional State (SNNPR), Ethiopia' unpublished doctoral thesis, University of London, School of Oriental and African Studies, 2000 111.
 L Smith 'Language choice, ethnic self-determination and national unity in contemporary Ethiopia' paper presented at a panel discussion on 'Ethnicity, identity and the right to self-determination' 26 May 2007, Addis Ababa, Ethiopia.
 B Haile-Selassie 'Ethiopia: A precarious ethno-federal constitutional order' unpublished doctoral thesis, University of Wisconsin Law School, 2002.
 M Haile 'The new Ethiopian Constitution: Its impact upon unity, human rights and development' (1996) 20 Suffolk Trans-National Law Review 1-84.
 Minase (n 61 above) 37. Ehrlich similarly argues that the language policy has the effect of 'causing a degree of separation between the various groups'. C Ehrlich 'Ethnicity and constitutional reform: The case of Ethiopia' (1999) 6 ILSA Journal of International and Comparative Law 51-73.
 Haile-Selassie (n 59 above) 213.
 As above.
 Ehrlich (n 62 above). Bloor & Tamrat share the same concern. It is often argued that the present language policy in Ethiopia 'will restrict movement across administrative units, thereby disrupting existing patterns of exchange between different areas and contact between different people'. T Bloor & W Tamrat 'Issues in Ethiopian language policy and education' (1996) 17 Journal of Multi-lingual and Multi-cultural Development 321-338. See also Cohen (n 57 above) 112.
 Smith (n 58 above) 5. As pointed out by Fasil, 'state recognition of every Ethiopian language means that efforts for its development — ie the preservation of literature, the provision for a script, where such does not exist; the documentation of its oral literature; and the further study of each language via grammatical, vocabulary and overall publication and enhanced use of the language — will be done with both state blessing and state support to the extent possible'. F Nahum Constitution for a nation of nations (1997) 55.
 Cohen (n 57 above) 111.
 What might even be problematic is the dominance of Amharic in the majority of the ethnically-plural regional states. This, one may argue, works against the constitutional commitment to promote linguistic diversity and especially the use of local languages even though the scheme benefits from the culturally neutral status of Amharic in the context of the regional languages.
 J Schmied English in Africa: An introduction (1991) 27.
 Coulombe (n 1 above) 242.
 Balmer (n 12 above) 447.
 Sacks (n 45 above) 683.
 This, of course, is not true for Afrikaners for whom issues relating to language historically constitute a central place in their resistance against the British cultural hegemony. This partly explains why the Afrikaners, unlike the other ethnic groups in South Africa, feel so strongly about the dominance of English in today's South Africa.