Autochthonous language communities and the Race Relations Act
Department of Celtic, University of Edinburgh
Copyright © 1998 Wilson McLeod.
First Published in Web Journal of Current Legal Issues in association with Blackstone Press Ltd.
More than twenty years after its enactment, the scope of the Race Relations Act 1976 remains less than fully clear. It is still unsettled, for example, whether and on what basis the English, Welsh and Scots are entitled to protection under the Act.
The relationship between Britain's autochthonous languages communities -- principally the Welsh and Scottish Gaelic communities -- and the RRA has received little attention, both because these communities rarely register on the horizon of those in Britain's urban core, and because there is no intuitive link between these groups and racial legislation.
Nevertheless, a plausible case can be made that these groups, especially the Gaelic community, should be recognised as distinct ethnic groups under the statute and given appropriate protections.
The issue is particularly significant in the context of employment, as more and more jobs in Wales and Scotland are designated as Welsh or Gaelic-essential, thereby creating the risk of discrimination claims from members of groups who are unlikely to have such language skills. The paper argues that whether or not the autochthonous language communities are formally recognised as distinct ethnic groups, employers' efforts to maintain and develop autochthonous language use should receive substantial deference.
More than twenty years after the enactment of the Race Relations Act 1976, fundamental questions about the scope of its protections remain unresolved. It is still not entirely clear, for example, whether, and on what basis, the peoples of the different national entities that make up Great Britain -- England, Wales and Scotland(1) -- constitute protected groups under the Act. Although this problem has attracted considerable attention of late, a related but in some ways more complex question has been largely ignored: what protection, if any, do Britain's autochthonous language communities enjoy under the Act? How are Gaelic speakers in Scotland and Welsh speakers in Wales to be classified within the framework of the Act, and how are their rights -- if any -- to be measured against those of other protected groups?
These questions have scarcely been considered at all by industrial tribunals or by the courts, and the little treatment they have received has been problematic and unsatisfactory. In Gwynedd County Council v Jones  ICR 833, the Employment Appeal Tribunal refused to recognise any distinction between Welsh-speaking Welsh people and English-monoglot Welsh people, so as to hold incompetent a claim by two English-monoglot Welsh claimants that reserving certain local authority jobs for Welsh speakers violated the claimants' rights under the RRA. The analysis in Jones was superficial, indeed cryptic, and a range of important sociological and legal questions were ignored. These questions have become all the more critical in light of changing policies and increasing attention at the UK and European levels to the rights and needs of autochthonous linguistic groups. If carefully and sensitively construed, the RRA may serve as a mechanism to reinforce these policies and solidify the position of the autochthonous languages.
The RRA is, however, a less than ideal framework for determining rights or resolving claims of this kind; it was enacted with a fairly specific purpose -- the protection of the then relatively new immigrant groups of Asians and Afro-Caribbeans -- but has since been used as a rather blunt instrument to deal with a much wider range of issues, including the rights of gypsies, Rastafarians and Sikhs. In particular, the Act's focus on "racial groups" as the catch-all category for analysis tends to impose unfortunate conceptual barriers. Section 3(1) of the Act provides that the term "`racial group' means a group of persons defined by reference to colour, race, nationality or ethnic or national origins, and references to a person's racial group refer to any racial group into which he falls." All the Act's protections work within this framework of the "racial group," so that to secure the Act's protection any community must necessarily be classified as a "racial group."
In modern social and political thinking, it is awkward and unfamiliar to think of, say, the Scots and the English, or for that matter the English and the French, as different "races," but that is essentially what the Act commands, since the term "racial group" is defined so expansively as to include "nationality", "national origins" and "ethnic origins". Although this conceptual obstacle should properly be ignored, so that, for example, distinct "nationality" or "ethnicity" should be fully sufficient to define a "racial group," it is clear that this terminology has discouraged courts and tribunals from granting recognition to groups that are not readily definable under the more colloquial understanding of a "racial group."
The term "autochthonous language," used throughout this article, may also seem unfamiliar and unwieldy, but is becoming an important term of art. At a European level, autochthonous minority languages have been distinguished from immigrant minority languages and granted special protections, most notably through the Council of Europe's European Charter for Regional or Minority Languages and certain European Community funding programmes set in place for the autochthonous languages of member states (European Commission Budget Line No. B3-1006). Whether this distinction is ultimately a justified one, or simply bespeaks an inappropriate lack of regard for the rights of immigrant language communities, is a significant question in structuring an anti-discrimination policy for Britain and Europe; for present purposes, however, the particular importance assigned to autochthonous languages is relevant for assessing the rights under the RRA of the UK's autochthonous language communities. This article will focus principally on the 500,000 strong Welsh-speaking community and especially on the 65,000-strong Scottish Gaelic community, whose potential claim to recognition as a distinct "ethnic group" -- and thus a protected "racial group" within the meaning of the Act -- is arguably the strongest of the autochthonous language communities.(2)
A number of RRA issues may arise in the context of autochthonous language communities. First, these groups might be recognised as "racial groups" in their own right. Such recognition could be significant in two different ways. Initially, it would mean that a member of the group is in a position to show that she belongs to a protected class and could therefore assert that she has been a victim of unlawful discrimination; and in a more positive sense, it might also give special leeway to measures designed to promote the use of the autochthonous languages.
More generally, and probably more importantly as a practical matter, employers who require autochthonous language skills for their staff might face discrimination claims from members of other protected groups. Even if the autochthonous language communities do not merit recognition as racial groups in their own right, such discrimination claims could readily be defeated, depending on the nature of the job at issue, but justifying autochthonous language requirements may require a complex and sensitive analysis, attuned to various competing legal and social policies.
These different questions are dealt with in turn.
To tackle the question of recognition under the RRA, Welsh speakers in Wales and Gaelic speakers in Scotland should be considered separately, for two reasons. First, the Jones precedent, which has little if any direct relevance to the Gaels, represents a significant if not insuperable obstacle for Welsh speakers. Second, the cultural and sociolinguistic situations of the two communities are very different in crucial respects, so that their classification under the statute need not be identical.
Analytically, it is best to begin by considering the problem of discrimination against Welsh or Gaelic speakers -- refusal to hire a Welsh or Gaelic speaker for a job, or imposition of some condition that brings about indirect discrimination against Welsh or Gaelic speakers. Discrimination of this kind is unlikely to prove the most serious source of actual litigation under the Act -- although such problems are certainly not imaginary, given the disturbing vitality of anti-Gaelic prejudice in Scotland(3) -- but it makes sense to start with this situation because it commands a focus on the significant question of whether Welsh or Gaelic speakers properly constitute a "racial group" under the Act. Answering this question will assist in analysis of other potential problems relating to the RRA, including the legality of employment restrictions designed to promote the welfare of these autochthonous language communities and the rights of members of other racial groups to challenge employment requirements involving autochthonous language skills. The legal position of Welsh and Gaelic speakers under the RRA should be the same whether membership in these communities is used as a sword or as a shield; that is to say, if Welsh or Gaelic speakers are given recognition and protection in a case where a member of the group asserts that she has suffered discrimination, such recognition should apply equally when a member of another racial group claims discrimination in favour of Welsh or Gaelic speakers.
Although the RRA uses the generic term "racial group," several distinct kinds of groups are actually subsumed into this general category. Five different terms are used in section 3(1) -- "colour," "race," "nationality," "ethnic origins" and "national origins" -- and the use of such verbal distinctions tends to suggest that these should be understood as discrete categories, "separate and alternative" from each other (Boyce v British Airways plc EAT/385/97 (quoted in Northern Joint Police Board v Power  IRLR 610, 613)). Although the English, Scots and Welsh could conceivably be distinguished on the basis of separate "nationality", "ethnic origin" or "national origin" -- depending on whether or not "nationality" is restricted to the meaning of citizenship in a sovereign state -- the critical question with respect to the Gaels and to Welsh speakers must be whether they properly constitute an "ethnic group"-- "a group of persons defined by reference to . . . ethnic origins" -- since it would seem clear that none of the other enumerated categories applies.
In the lead case, Mandla v Dowell Lee  2 AC 548, Lord Fraser set out the factors to be considered in assessing whether a protected "ethnic group" exists,
"For a group to constitute an ethnic group in the sense of the Act of 1976, it must, in my opinion, regard itself, and be regarded by others, as a distinct community by virtue of certain characteristics. Some of these characteristics are essential; others are not essential but one or more of them will commonly be found and will help to distinguish the group from the surrounding community. The conditions which appear to me to be essential are these, (1) a long shared history, of which the group is conscious as distinguishing it from other groups, and the memory of which it keeps alive; (2) a cultural tradition of its own, including family and social customs and manners, often but not necessarily associated with religious observance. In addition to these two essential characteristics the following characteristics are, in my opinion, relevant, (3) either a common geographical origin, or descent from a number of common ancestors; (4) a common language, not necessarily peculiar to the group; (5) a common literature peculiar to the group; (6) a common religion different from that of neighbouring groups or from the general community surrounding it; (7) being a minority or being an oppressed or dominant group within a larger community, for example (say, the inhabitants of England shortly after the Norman conquest and their conquerors might both be ethnic groups.) A group defined by reference to enough of these characteristics would be capable of including converts, for example, persons who marry into the group, and of excluding apostates. Provided a person who joins the group feels himself or herself to be a member of it, and is accepted by other members, then he is, for the purposes of the Act, a member. . . . In my opinion, it is possible for a person to fall into a particular racial group either by birth or adherence, and it makes no difference, so far as the Act of 1976 is concerned, by which route he finds his way into the group." ( 2 AC at 562)
Applying these criteria, Mandla held that Sikhs constituted an "ethnic group"; in subsequent cases Gypsies have also been held to be an "ethnic group" (CRE v Dutton  IRLR 8 (CA)), while Rastafarians have not (Dawkins v Department of the Environment  IRLR 284,  ICR 517 (CA)).
In making these assessments, and in considering the position of Britain's constituent national groups under the Act, courts and tribunals seem to have overemphasised the role of "race" in the analysis. Although Lord Fraser's speech in Mandla emphasised the need to construe the word "ethnic" "relatively widely, in a broad, cultural/historic sense," his initial observation that "the word `ethnic' still retains a racial flavour" ( 2 AC at 562) has received undue emphasis in subsequent decisions -- an approach that has attracted academic criticism (MacEwen 1997, p 93). This comment was seized upon as the basis to deny ethnic group status to Rastafarians in Dawkins; similarly, although Jones had previously asserted that it was "obvious" that the Welsh constituted an ethnic group ( ICR at 836), the EAT has recently relied on the "racial flavour" principle to rule in Boyce v British Airways plc EAT 385/97 (quoted in Northern Joint Police Board v Power  IRLR 610, 613) that the English, Welsh and Scots are not entitled to RRA protection on the basis of distinct "ethnic origins".
The "racial flavour" required for "ethnic group" status is not, however, coterminous with the broader term "race". The inquiry appears to be somewhat more specific. Lord Fraser amplified his reference to "racial flavour" by noting that "the word `ethnic'... is used nowadays in an extended sense to include other characteristics which may be commonly thought of as being associated with common racial origin" (Mandla  2 AC at 562) (emphasis added). Lord Fraser cited with approval the decision of the New Zealand Court of Appeal in King-Ansell v Police  2 NZLR 531 that "a group is identifiable in terms of its ethnic origins if it is a segment of the population distinguished from others by a sufficient combination of shared customs, beliefs, traditions and characteristics derived from a common or presumed common past even if not drawn from what in biological terms is a common racial stock" ( 2 NZLR at 543). This nuanced approach to the concept of "racial flavour" means that traits based on a relatively recent shared tradition can be sufficient even in the absence of a common racial stock; Lord Fraser concluded in Mandla that Sikhs constitute an ethnic group for RRA purposes "although they are not biologically distinguishable from the other peoples living in the Punjab" ( 2 AC at 565). It should be noted, however, that Lord Templeman's speech in Mandla appeared to emphasise the racial dimension to a rather greater degree, arguing that "a group of persons defined by reference to ethnic origins must possess some of the characteristics of a race, namely group descent, a group of geographical origin and a group history" ( 2 AC at 569).
Significantly, the EAT ruled in Northern Joint Police Board v Power  IRLR 610, a companion case to Boyce, that Britain's constituent national groups -- i.e. English, Welsh and Scots -- are entitled to recognition under the RRA as distinct "national groups". Although Power did not provide a meaningful definition of "national group", relying instead on a series of vague ipso facto propositions,(4) it would seem fairly certain that this head cannot extend to Gaelic or Welsh speakers.
Turning first to the question of the Welsh language community, the Jones decision is of central importance. In concluding that English-monoglot Welsh people did not constitute a protected racial group, the EAT adopted a view of the Welsh as an undifferentiated unit and thus implicitly determined that Welsh speakers were also not a protected group. Noting the criteria identified in Mandla, Jones opined that language, standing alone, is not an "essential" factor in the determination, and that the analytic emphasis should be on the "combination of factors." As such, it reversed the industrial tribunal and ruled "that it was wrong in law to use the language factor alone and in isolation as creating a racial group" ( ICR at 836).
Although its decision may ultimately have been correct in light of the overall position of the Welsh language in Wales and the nature of the Welsh-speaking community, the EAT in Jones clearly failed to consider the question with any serious analysis. Language cannot properly be considered something that stands alone; in particular, it very often tends to create among its speakers "a cultural tradition of [their] own" (Mandla  2 AC at 562), and it is certainly arguable that such a distinct tradition can be discerned among Welsh speakers. It is unfortunate that the status of the Welsh language community was determined in this essentially negative context; a much more vigorous and culturally sensitive case could have been mounted within the Mandla framework if the question affirmatively presented had been the status of the Welsh-speaking minority community, rather than the English-monoglot majority.
The analysis in Jones was also distorted to some extent by the unhelpful terminology of the RRA, with its reliance on the term "racial group" as the unit of analytic currency. Although Mandla took the proper analytic approach and spoke of "ethnic groups" -- the pertinent subset of the "racial group" under the statute -- the EAT's reasoning in Jones seems to have been confused by the "racial group" terminology. The EAT's evident difficulty in seeing Welsh-speakers and English-monoglots as separate "racial groups" in the ordinary lay sense led it to explain its decision with peculiar images,
"We cannot believe that, for example, a Mrs. Jones from Holyhead who speaks Welsh as well as English is to be regarded as belonging to a different racial group from her dear friend, a Mrs. Thomas from Colwyn Bay who speaks only English" ( ICR at 836).
Although Jones can hardly be considered a helpful precedent from the Gaelic standpoint, the position of the Gaels within the framework of Mandla and Jones is not immediately obvious, both because Jones was such a superficial and unedifying decision, and because the position of Gaelic is very likely stronger than that of Welsh under the Mandla analysis. Some factors point in one direction, some in others. Part of the difficulty arises from the fact that the Gaels are, in many respects, a group in transition, part way -- far along the way? -- toward assimilation into a greater Britain and the global village. A sensitive evaluation of the Gaels' position, however, requires attention to the larger historical trajectory, and not some artificial snapshot of the present situation. This is particularly the case because the determination of the Gaels' status as an ethnic group may affect the legality of measures designed to promote and increase the use of the Gaelic language, measures that will tend, to some extent, to be motivated by a sense that the language and its speakers have suffered injustices in the past at the hands of hostile outsiders from other parts of Britain who most definitely did perceive the Gaels and their language as alien (MacKinnon 1991).
Although the principal factor differentiating the Gaels from other Scots is the use of the Gaelic language itself, it can well be argued that the language is actually the medium of a distinct and separate culture, manifested in a variety of ways including deep-rooted traditions of poetry, song and music, and unique forms of religious worship. To some extent at least, this distinctiveness extends to material existence as well, the present-day crofting communities remaining substantially different in their way of life from the highly urbanised Scottish mainstream. The claim of Gaelic speakers to recognition as an ethnic group is also strengthened by the fact that a very high proportion of Gaelic speakers, relative to the UK's other autochthonous language communities, are native speakers born and brought up in Gaelic-speaking communities in the Hebrides and West Highlands. It would be safe to say that at least 90% of Gaelic speakers come from such backgrounds, whereas the Welsh language community contains significant proportions of learners and non-traditional speakers. In the case of Gaelic, then, there is a very significant link between the ability to speak the language and a distinct culture and way of life, and the language is the badge of a community that has long been outside the societal mainstream.
Significantly from a legal standpoint, this "combination of shared customs, beliefs, traditions and characteristics" is largely "derived from a common . . . past", distinct from the social institutions and practices of Lowland Britain (King-Ansell  2 NZLR at 543). Although Gaelic was once the language of almost all Scotland, language shift and cultural divergence began as early as 1100, and already by 1380 the Gaels were identified as an entirely distinct and separate group. Marginalised by geography and outsiders' hostility, the Gaels have without question emerged from a common history and experience not shared by others (MacKinnon 1991).
Nevertheless, although the Gaels would certainly have merited recognition as an "ethnic group" in 1745 or even 1900, the position is complicated considerably by the effect of subsequent cultural changes and of the language shift from Gaelic to English that has been proceeding through the Highlands and Islands for the last two centuries and is now progressing within the Gaelic heartland of the Outer Hebrides. The difference between those Highland areas where Gaelic remains vital and those where it has recently passed from use is not particularly marked. Even in communities where the language is still strong, many families consist of Gaelic speakers in the older generations and English monoglots among the young, yet it is the prevailing understanding that a "Gael" is simply duine aig a bheil Gàidhlig bho dhùthchas, "a person who has Gaelic by inherited tradition" or, more loosely, "a native Gaelic speaker" (Macaulay 1994, p 43). Whatever the realities of the sociolinguistic situation, in legal terms it seems difficult to sustain a classification by which parents can place their children into a different "ethnic group" from themselves simply by speaking one language in the household rather than another.
The problem of group entrance is equally thorny. Someone who learns Gaelic -- whether she be English, American, or German in origin -- is fully qualified to take a job requiring Gaelic skills, so that an employer demanding such skills is not in actual fact making a demand about the ethnic background of the employee. To understand the situation properly, however, it must be borne in mind that there are hardly any English, American or German people -- or, for that matter, Scots who have not acquired Gaelic in childhood -- who have learned the language to a substantial degree of competence; the total number of fluent learners is probably no more than a thousand. The link between linguistic ability and "ethnicity" thus remains strong. For those few who have acquired the language, moreover, it is by no means clear that this accomplishment is sufficient to admit them fully into the Gaelic community. Mandla indicated that membership in a protected group need not be confined to those born into it, but that membership by "adherence" was possible, and that it was sufficient for "a person who joins the group [to] feel himself or herself to be a member of it" provided that he or she "is accepted by other members" ( 2 AC at 562). Various degrees of resentment against so-called "new Gaels" are very familiar in the Gaelic world, where the value of dùthchas (inherited tradition) remains paramount. In sum, then, it is very easy to enter the Gaelic community in some respects -- so easy as to call the distinctiveness of the group into question -- but extremely difficult in other respects. Finally, although there is substantial evidence to support the view that the Gaels should be recognised as an ethnic group distinct from Scottish society as a whole, many Gaels, and especially community leaders, would probably tend to reject this interpretation. The Gaelic revival of recent decades has deliberately emphasised the importance of Gaelic for Scotland as a whole, and attempted to relocate Gaelic to the centre of Scottish life -- though, significantly, these efforts are sometimes resisted or rejected by non-Gaels (Cormack 1994).
On balance, it appears very difficult to articulate a viable basis for protection of Welsh speakers, given the constraint of Jones. The position of Gaelic speakers is considerably stronger, but by no means certain. A simplistic analysis, emphasising the "racial" dimension without deep probing, would tend to work against recognition of the Gaels as an ethnic group, while a more complex, culturally informed inquiry could well produce a different result.
The most likely scenario for legal conflict concerning the rights of autochthonous language communities is that of a claim that reserving a certain job for members of an autochthonous language community constitutes actionable discrimination against members of other protected communities. This was the context of the Jones case, and it is reasonable to foresee further conflicts of this kind, particularly if policies of securing and advancing the autochthonous languages are pursued with any vigour. Significantly, the question of RRA protection for autochthonous language speakers is not relevant here; claims by members of other protected groups may arise --and may be successfully defended in appropriate cases -- even if the autochthonous language communities do not themselves merit recognition and protection under the RRA.
The basic questions to be resolved in this scenario are the same as in any RRA case, i.e. those set out in section 1(1)(b) of the Act,(5) Can the claimant show that she is a member of a protected racial group? If so, can she make out a prima facie case of discrimination -- direct or indirect? Can that prima facie case be rebutted by a showing that the discrimination is justifiable, for example by showing that skills in the autochthonous language are necessary to perform the job? Claims challenging autochthonous language requirements are much more likely to be based on the theory of indirect rather than direct discrimination: if such requirements do have a discriminatory impact, it will almost certainly be as a side effect to their intended purpose of promoting the autochthonous language, rather than the result of a deliberate effort to harm members of other groups. To advance a claim of indirect discrimination, a claimant needs to show that she has suffered a detriment because she is unable to satisfy the language requirement, and that the proportion of members of her racial group able to meet the requirement is "considerably smaller than the proportion of persons not of that racial group who can comply with it" (RRA s 1(1)(b)(i)). If the claimant succeeds in making this showing, the employer must then establish that its discriminatory conduct is nevertheless justifiable.
At the outset, it is important to emphasise that claims of indirect discrimination can succeed only when autochthonous language skills are actually required. When an employer advises merely that such skills would be an advantage, or that persons having such skills will be preferred over persons who lack them, there can be no case. This is the consequence of the much-criticised decisions in Perera v Civil Service Commission (No. 2)  IRLR 166,  ICR 428 (CA) and Meer v Tower Hamlets London Borough Council  IRLR 399, under which a claimant cannot establish that an employer has imposed a discriminatory "requirement or condition" within the meaning of section 1(1)(b) except when the criterion in question "is a must -- something which has to be complied with .... [and] not merely something which could have affected the mind of the employer" (Id. at 402 (Balcombe LJ). For employers who seek to promote the use of autochthonous languages in the workplace -- as for those who seek crafty means of practising invidious discrimination -- Perera grants very significant leeway.
The analytically simplest scenario for an indirect discrimination case would involve a claim by a member of one of the racial groups already recognised under the Act -- an Afro-Caribbean, for example. A viable claim of indirect discrimination under section 1(1)(b) could readily be constructed here, because the overwhelming majority of autochthonous language speakers are white, a much greater proportion of white people than black could satisfy an autochthonous language requirement. Attention would then turn to the question of employer justification -- whether the language requirement, with its indirectly discriminatory impact, was nevertheless permissible in the circumstances.(6) However, given the extremely small black populations in the areas of Britain where jobs requiring such language skills are likely to appear -- west Wales and the Scottish Highlands and Islands -- such claims are likely to remain purely hypothetical.
One could posit a similar claim by a citizen of one of the other European Union member states. As with the Asian and Afro-Caribbean populations in Britain, each of these nations -- "group[s] of persons defined by reference to ... nationality" -- can be considered a distinct racial group under the statute. Here, a claim of indirect discrimination could be built on the fact that the overwhelming majority of Welsh and Gaelic speakers are citizens of the United Kingdom -- even though the proportion of UK citizens who speak these languages is very small -- and very few citizens of other European Union member states can speak these languages. A claim by a member of another EU member state could also rely on European law: the cornerstone of European employment law is the principle of free movement of workers between member states,(7) and a Welsh or Gaelic requirement not justified by employer necessity could be considered an improper burden to free movement of workers throughout the European Union by erecting a barrier that favoured UK workers -- albeit a tiny minority of them -- at the expense of nationals of other member states. (A claim of this kind was considered in the European Court of Justice's landmark 1989 decision in Groener v Minister of Education (Case 379/87 (1989),  2 ECR 3967), discussed in the context of employer justification, below).
A much more likely scenario -- and a much more politically charged one, given the controversial recent influx of English people into the autochthonous language regions -- would involve a claim by an English person lacking skills in the autochthonous language. Following Power, an English claimant may assert RRA protection based on "national origins", and would thus be in a position to establish a prima facie case, just as the hypothetical black claimant or non-UK European national, in that disproportionately few English people can speak Welsh or Gaelic.
Assuming a discrimination claimant can establish a prima facie case, what of the defence? How and when can requiring autochthonous language skills be justified? The statute provides little guidance: the employer imposing the language requirement will escape RRA liability if it can show the requirement "to be justifiable irrespective of the colour, race, nationality or ethnic or national origins of the person to whom it is applied" (RRA s 1(1)(b)(ii)).
Over the years, the courts have failed to establish a straightforward, workable framework for assessing how discrimination is to be justified under this provision. Despite some occasional clear statements -- such as the Court of Appeal's explanation in Hampson v Department of Education and Science  IRLR 69,  ICR 179 (CA) that "`justifiable' requires an objective balance to be struck between the discriminatory effect of the condition and the reasonable needs of the party who applies the condition" ( IRLR at 191,  ICR at 75) -- there has been "a consistent tendency over time for ... objective standards ... to be weakened by the addition of subjective elements" (Bourn & Whitmore 1996, s 2.60, p 76). Without question, the uncertainty and subjectivity on this point has redounded to the benefit of employers seeking to justify discrimination, and it may fairly be said that the burden of showing justifiability is not an onerous one. While this is perhaps not the ideal posture for British anti-discrimination law in general, it does have the consequence that employers seeking to justify autochthonous language requirements will tend to be in a favourable position vis-à-vis those who might challenge them, so that any reasonable arguments and explanations in support of such requirements should be sufficient to defeat discrimination claims.
Deciding when and in what circumstances a job requirement is "justifiable" is not necessarily straightforward, however, irrespective of the legal test to be applied. As always, there are easy cases and more difficult ones, and deciding where to draw the line is more a matter of making policy decisions than deducing immutable principles.
Consider first the easy cases -- jobs for which the use of the autochthonous language is clearly essential. The most obvious examples in the Gaelic context are teachers in Gaelic-medium school units, and presenters in the Gaelic media. Someone who cannot speak and understand Gaelic is clearly incapable of conducting radio interviews or teaching schoolchildren through the medium of Gaelic, and there can be no doubt that an employer is justified in demanding Gaelic ability for such employees.
Next consider the other extreme -- jobs for which the autochthonous language is required but will in fact never be used at all. Although it might seem at first blush that a requirement of this nature could not possibly stand, it was just such a requirement that the European Court of Justice upheld in Groener. Even though Groener did not implicate the RRA, it is nevertheless a highly relevant precedent, especially given the relative lucidity of its analysis.
Groener held that because of the constitutionally guaranteed official status of the Irish language in the Republic of Ireland, a Dublin college could lawfully impose an Irish-language requirement for an art lecturing post, even though knowledge of Irish was not actually needed to perform the job. The court was careful to confine its ruling to the special, indeed unique, circumstances of the Irish language, which is constitutionally designated as "the national language" and "the first official language" (Bunreacht na hÉireann, Article 8),
"The EEC Treaty does not prohibit the adoption of a policy for the protection and promotion of a language of a Member State which is both the national language and the first official language. However, the implementation of such a policy must not encroach upon a fundamental freedom such as that of the free movement of workers. Therefore, the requirements deriving from measures intended to implement such a policy must not in any circumstances be disproportionate in relation to the aim pursued and the manner in which they are applied must not bring about discrimination against nationals of other Member States.
The importance of education for the implementation of such a policy must be recognised. Teachers have an essential role to play, not only through the teaching which they provide but also by their participation in the daily life of the school and the privileged relationship which they have with their pupils. In those circumstances, it is not unreasonable to require them to have some knowledge of the first official language." ( 2 ECR at 3993)
Commentators have argued that the Groener rule would probably not apply if such a job requirement were imposed for a minority language that has not been granted official status (De Witte 1991, p 170). This is the present position of Gaelic, although legislation to secure its status is currently contemplated (McLeod 1997; Comunn na Gàidhlig 1997). With such languages, in contrast to the situation in Groener, there would be no real state policy at issue to countervail the established state policy of promoting free worker movement.
Indeed, a strong argument can be made that Groener would not apply in all cases where the language involved has been granted official status, but only when that status is comparable in its vigour to the constitutional enshrinement put in place in Ireland. Welsh, for example, has been granted official recognition by the Welsh Language Acts 1967 and 1993, but its position is by no means as secure in legal terms as that of Irish in the Irish Republic. In terms of the balancing test established in Groener, a state policy that establishes a less secure form of official status can be considered a less important governmental aim, so that employment restrictions that interfere with the policy of free movement of workers might more readily be considered disproportionate.
Similarly, it is open to question whether the Groener principle would hold with respect to posts less societally sensitive than that of a teacher (Ó Máille 1990, p 39).(8) For example, office workers who make little contact with the public might be considered to have little relationship to the state's policy of promoting its official language. Again, under the balancing test, the state's interest would be reduced so that the interference with free worker movement could more easily be held to be disproportionate and unjustified.
Although Groener would not be directly applicable in an RRA case, its underlying principle is highly relevant: that an assessment of whether a language requirement is justifiable should take into account any policy of supporting and sustaining the pertinent language community. As such, the justifiability inquiry here should be considered fundamentally different from the assessment of other employer job requirements which do not implicate important social policies. In effect, there is an additional dimension in cases involving autochthonous language requirements, one that does not arise in typical cases, in which the challenged job requirement (stipulating an arguably excessive number of school leaving qualifications, for example) does not implicate social policy to any significant degree.
This principle was articulated with particular clarity by the Advocate General in Groener, who made the case that European enactments as a whole "recognised that it is essential to preserve Europe's cultural richness and to ensure the diversity of its linguistic heritage" ( 2 ECR at 3981). Nathaniel Berman has commented that "[t]he bold nature of this approach is that it destroys any pretence of maintaining the boundaries between a `legal' and `political' jurisprudence when cultural survival is in question" (Berman 1992, p 1566). Although this approach might be unpalatable for British adjudicators in so strong a form, a diluted version of this understanding does seem permissible and appropriate.
This political dimension is important for the most significant and difficult cases -- the intermediate cases in which an autochthonous language requirement is imposed for jobs where those language skills will be needed on an occasional rather than constant basis. Deciding whether such a language requirement is justified calls for a fact-specific inquiry that will necessarily vary from case to case, and that is unlikely to be straightforward. As a general matter, though, it appears that the balance should be weighted somewhat to favour employer leeway in requiring autochthonous language skills. As shown above, the basic legal rule on justification grants substantial deference to an employer's determination that a given requirement is justifiable, even in run-of-the-mill cases where the requirement in question does not implicate public policy. More specifically, government policies favouring the promotion and development of autochthonous languages should work an increase in the basic level of deference. In particular, assessment of whether a language requirement is justifiable should take into account the fact that present government policy toward the autochthonous language communities -- as manifested by, among other things, substantial public expenditure on government, media, and cultural and community development -- is focused on making proactive efforts to reverse the decline in autochthonous language use, and affirmatively to facilitate and promote such use.
Outside the most obvious spheres of autochthonous language use -- the media and educational contexts noted above -- perhaps the next most viable setting for an autochthonous language requirement is in the social care sector -- doctors, nurses, midwives, community care workers, social workers and so on who provide personal services to members of autochthonous language groups. It is clearly desirable that such carers be able to communicate through the autochthonous language; indeed, in its Jones decision, which arose in the context of a Welsh requirement for carers in a residential home for the elderly, the EAT noted in dictum that a language requirement would appear reasonable in such a setting ( ICR at 837). However, the position is complicated somewhat by the fact that, for practical purposes, all the members of the UK's autochthonous language communities are fully able to speak English as well as the autochthonous language. As such -- and in contrast to the many carers serving, for example, monoglot speakers of Asian languages -- it is not absolutely essential to be able to speak the autochthonous language in order to communicate with members of the autochthonous language communities. Nevertheless, it is clearly justifiable to require autochthonous language skills for jobs of this kind, because many members of these communities -- and especially members served by social carers, who necessarily tend to be elderly, frail, ill or otherwise vulnerable -- are more comfortable using the autochthonous language rather than English, and it would undermine the basic goal of providing effective care to force patients and clients to use English when doing so would be a source of discomfort or stress.(9)
These considerations would be relevant, for example, in assessing how much contact with members of autochthonous language communities would be necessary to justify a language requirement. In some areas -- the Outer Hebrides, for example, or much of western Wales -- the overwhelming majority of the population will belong to the autochthonous language group, and requirements for carers to have autochthonous language skills would be particularly well justified. In other places -- the mainland Scottish Highlands, for example, or the cities and large towns of South Wales -- the proportion will be smaller yet still significant.(10) In light of prevailing policies, a substantial degree of deference should be applied in assessing whether such language requirements are justified, but there would inevitably come a point where the indirectly discriminatory impact of such a requirement would outweigh its benefit. It is safe to say, however, that employers have not come at all close to this point in imposing autochthonous language requirements of questionable necessity. Indeed, the prevailing approach has been very much in the opposite direction; in Scotland at least, language tends largely to be ignored as a hiring criterion in this sector, with a possible detrimental impact on patients and clients.
Government employment as a whole is an important field, particularly at the local level. In addition to various positions for which autochthonous language skills are clearly essential -- Welsh interpreters in Welsh courts, or translators of official documents into Gaelic -- there is a range of jobs for which language skills may be required as a matter of policy, especially if that policy is one of treating English and the autochthonous language "on a basis of equality" (Welsh Language Act 1993 c 38 s 5(2)) such that members of the public are enabled to transact any official business they may have through the medium of the autochthonous language. At the most obvious level, putting such policies into effect may involve requiring that receptionists be able to speak the autochthonous language when receiving telephone calls from members of the public who wish to speak the autochthonous language. It may extend to ensuring that secretaries and executive officers are capable of processing written enquiries from autochthonous language users. It may even reach the stage in which all employees need to speak the autochthonous language because that language, and not English, is used as the operative medium of a particular arm of government -- as, for example, the County Court at Blaenau Ffestiniog, which operates through the medium of Welsh except where otherwise requested (Andrews & Henshaw 1994, p 49).
An autochthonous language requirement would be fairly easy to justify in a work setting where the language was in constant use, but more difficult in an office where the language was used only sporadically. Nevertheless, if public policy suggests that citizens should be able to use their autochthonous language in dealing with government, it is necessary, and justifiable, to allow government to employ a reasonable number of persons to meet those citizens' needs. Indeed, it would undermine such a policy if only a token number of employees were available, so that attempts to exercise the right to use the autochthonous language were slow, frustrating or fruitless. Again, then, there should be substantial weighting in favour of employers' rights to stipulate autochthonous language skills.
Similar questions could arise in the private sector if employers take affirmative steps to advance the use of the language. Many private workplaces in Wales, and rather fewer in Scotland, operate through the medium of an autochthonous language; many more have adopted policies of various kinds that involve the use of these languages in one way or another, such as the provision of telephone operators or other personnel who can deal with enquiries in the autochthonous language.
The basic problem here is one of balance, ensuring that the needs of autochthonous language communities and the policies of promoting the use of autochthonous languages do not obtrude on the rights of others. It may safely be said, however, that to date employers have not come close to bringing about such undesirable consequences, and indeed that the primary goals of serving the communities and strengthening the languages have largely been ignored.
In addition to invoking the basic justification defence set out in section 1(1)(b)(ii), employers requiring autochthonous language skills might also be able to insulate their actions by invoking the affirmative protections enshrined in the Act that insulate employers from discrimination claims that would undermine attempts to promote and secure the welfare of racial groups. Even if these provisions were not directly applicable -- they could only come into play if the autochthonous language communities were affirmatively recognised as racial groups under the Act -- the philosophy underpinning them should be taken into account in assessing the questions of justifiability discussed above.
For example, Section 5(2)(d) of the RRA provides that being a member of a particular racial group constitutes a "genuine occupational qualification" -- a qualification upon which an employer may lawfully insist -- where "the holder of the job provides persons of that racial group with personal services promoting their welfare, and those services can most effectively be provided by a person of that racial group." Similarly, section 35 of the RRA provides that "[n]othing in Parts II to IV shall render unlawful any act done in affording persons a particular racial group access to facilities or services to meet the special needs of persons of that group in regard to their education, training or welfare, or any ancillary benefits." These provisions are frequently invoked as a basis to set aside various public and voluntary sector jobs for members of Afro-Caribbean and Asian racial groups (Tottenham Green Under Fives' Centre v Marshall (No. 2)  ICR 320,  IRLR 219 (EAT); Lambeth London Borough Council v Commission for Racial Equality  ICR 768,  IRLR 231 (CA)).
In effect, sections 5(2)(d) and 35 operate formally to recognise a particular justification for what would otherwise constitute unlawful discrimination. If these provisions did not exist, employers would still be able to argue that reserving a certain job for members of a particular racial group was "justifiable" within the meaning of section 1(1)(b)(ii). The general need to promote the welfare of the population served could certainly constitute a sufficient justification under this general provision; formally enshrining this principle in the statute itself, however, demonstrates its special importance and assures that tribunals and courts are bound to give due recognition to what might otherwise be an unfamiliar or unwelcome proposition. In the context of Welsh and Gaelic, either section 5(2)(d) should be directly applicable, or, alternatively, the goal of promoting the welfare of these autochthonous language communities should be credited as an appropriate justification under section 1(1)(b)(ii) for requiring autochthonous language skills for a given position.
Section 5(3) is of considerable assistance in analysis, both by determining the range of section 5(2)(d) and, indirectly, by providing clarifying guidance for the assessment of justification under section 1(1)(b)(ii). Section 5(3) provides that "[s]ubsection (2) applies where some only of the duties of the job fall within paragraph (a), (b), (c) or (d) as well as where all of them do"; as such, it is not necessary that the population served consist exclusively of the racial group in question. Mutatis mutandis, this provision suggests strongly that it should be permissible under section 1(1)(b)(ii) to impose an autochthonous language requirement for a job in which only some of the requisite duties involved the use of the autochthonous language.
In the context of the autochthonous language communities, section 5(2(d) would be helpful, for example, in considering the position of an employee whose work mostly involves using English but who also uses an autochthonous language some portion of the time in working with autochthonous language speakers. This could well be the situation, for example, with a nurse in a Hebridean district, meeting the needs of a patient group consisting of monolingual English speakers, people bilingual in English and Gaelic who prefer to use English in dealing with a health care professional, and bilingual people who prefer to use Gaelic in that context. The relative proportions might well be significant in assessing the question of justification; obviously, it would be more difficult to justify the application of section 5(2)(d) where Gaelic would only be used only very intermittently, say a few times a year. It is submitted, however, that a fairly vigorous approach is appropriate in this regard, and that any non-negligible proportion of Gaelic usage should suffice. The evident purpose of the statutory provision is to meet the needs of individual members of racial and ethnic groups, and it would offend that purpose if the needs of such individuals were not satisfied because of the misfortune of not being numerous enough. In most cases, after all, the groups being protected here are minorities, and it is inevitable that minorities will sometimes form small proportions of the overall population.
Section 5(4) also provides helpful general guidance, by providing that an employer may not rely on section 5(2) when it already has sufficient employees to meet the need to service the welfare of the racial group at issue.(11) As such, for example, it would be improper to invoke section 5(2)(d) to insist that every single nurse at a hospital serving a partially-Gaelic speaking population be able to speak Gaelic; and, in addition, such insistence should not be deemed "justifiable" under section 1(1)(b)(ii). The employer's goal of meeting the needs of the Gaelic-speaking population could be satisfied by applying the section only to a proportion of the nursing staff, who could successfully meet the hospital's need of attending to patients having a Gaelic preference.
The autochthonous language communities are at the margins of British life, typically ignored by politicians, scholars and activists focused on the urban centre. Their relationship to the RRA has rarely been considered, partially because such a connection is not an intuitive one, especially among those for whom the autochthonous language communities do not register on the horizon. Nevertheless, there are significant and interesting issues at stake. An awareness of the RRA can work a range of positive outcomes for the autochthonous language communities, either by granting them formal recognition or by taking their needs into account and considering ways in which language promotion policies can be advanced in the employment field. Because such policies are likely to strengthen in coming years -- at both the domestic and European levels -- these issues may also grow in importance.
Andrews, J & Henshaw, L (1994) 'Welsh Language in the Courts, 1981-1991' 25 Cambrian Law Journal 47.
Berman, N (1992) 'Nationalism Legal and Linguistic: The Teachings of European Jurisprudence' 24 NYU Journal of International Law & Politics 1515.
Bourn, C & Whitmore, J (1996) Anti-Discrimination Law in Britain (London, Sweet & Maxwell).
Comunn na Gàidhlig (1997) Inbhe Thèarainte dhan Ghàidhlig: Tagradh as leth coimhearsnachd na Gàidhlig/Secure Status for Gaelic, A submission on behalf of the Gaelic community.
Cormack, M (1994) 'Programming for Cultural Defence: The Expansion of Gaelic Television' 6 Scottish Affairs 114.
De Witte, B (1991) 'The impact of European Community rules on linguistic policies of the Member States' in Coulmas, F (ed) A Language Policy for the European Community, Prospects and Quandaries (Berlin & New York, Mouton de Gruyter), p 163.
Holmes, A E M & Painter, R W (1996) Swot Employment Law, 4th ed (London, Blackstone Press).
MacAulay, D (1994) 'Canons, myths and cannon fodder' 1 Scotlands 35.
MacEwen, M (1997) 'Discrimination on the ground of race' SCOLAG Journal, April/May 1997, 80.
MacKinnon, K (1991) Gaelic: A Past and Future Prospect (Edinburgh, Saltire Society).
McLeod, W (1997) 'Official Status for Gaelic, Prospects and Problems' 21 Scottish Affairs 95.
Ó Máille, Tomás (1990) Stádas na Gaeilge: Dearcadh Dlíthiúil (Dublin, Bord na Gaeilge) (Páipéar Ócáide -- Uimhir 1).
(1) . The RRA does not apply to Northern Ireland, which is governed by the Fair Employment (Northern Ireland) Act 1976 c.25, and The Race Relations (Northern Ireland) Order 1997, 1997 No. 869 (N.I. 6).
(2). Other autochthonous language communities in Britain include speakers of Scots and Cornish; for various reasons, these communities would be in a much weaker position under the RRA than Welsh or Gaelic speakers. Because the RRA does not apply to Northern Ireland or the Isle of Man, the Irish (Gaeilge) and Manx communities in those jurisdictions are not affected.
(3). As one manifestation of such prejudice, Scottish newspapers routinely feature articles that are strikingly derogatory of Gaelic speakers, using rhetoric of a kind no longer acceptable in public discourse when applied to members of ethnic minorities, and qualitatively different in character from the ritualistic anti-English remarks often found in sections of the Scottish media. A recent article in The Sunday Times' Scottish edition, for example, asserted that "everything about Gaelic and those who promote it is loathsome", while a 1995 piece in The Scotsman claimed that "[t]here is nothing in Gaelic that is worth passing on to the rest of mankind" and dismissed the language as "a low-level peasantish sort of debris". Alan Brown, "Restless Native", The Sunday Times, October 19, 1997, Section 12 (Ecosse), p. 2; Peter Clarke, "Who needs the Gaelic?", The Scotsman, March 11, 1995, "Weekend" magazine, p. 24.
(4). In contrast, the EAT did provide a substantial definition of "nationality", which "has a juridical basis pointing to citizenship, which, in turn, points to the existence of a recognised state at the material time". Power  IRLR at 613.
(5). Section 1(1) provides that:
A person discriminates against another in any circumstances relevant for the purposes of this Act if --
. . .
(b) he applies to that other a requirement or condition which he applies or would apply equally to persons not of that same racial group as that other but --
(i) which is such that the proportion of persons of the same racial group as that other who can comply with it is considerably smaller than the proportion of persons not of that racial group who can comply with it; and
(ii) which he cannot show to be justifiable irrespective of the colour, race, nationality or ethnic or national origins of the person to whom it is applied; and
(iii) which is to the detriment of that other because he cannot comply with it.
The appropriate analytic framework for assessing claims of indirect discrimination under section 1(1)(b) is set out in Raval v Department of Health and Social Security  IRLR 370,  ICR 685 (EAT).
(6). Note that an employer seeking to justify an indirectly discriminatory autochthonous languge requirement could not rely on the argument that members of disadvantaged racial groups could simply learn the autochthonous language following their appointment and thereby satisfy the requirement. In Raval  IRLR 370,  ICR 685, the EAT held that the determination of whether an employee "can comply" with the specified requirements for a job, within the meaning of section 1(1)(b)(i) of the RRA, is to be made as of the deadline fixed for applications. As such, only those applicants with a preexisting competence in the autochthonous language could be considered to satisfy the requirement for purposes of the Act.
(7). This principle is stated in Article 48 of the Treaty of Rome and amplified in Regulation 1612/68 of the Council of the European Communities (October 15, 1968). Note that Regulation 1612/68 does "not apply to conditions relating to linguistic knowledge required by reason of the nature of the post to be filled."
(8). The position may not be as simple as Holmes and Painter suggest in their recent handbook, i.e. that a language requirement "certainly would not be justifiable [under Groener] for employment as, for example, a cleaner" (Holmes & Painter 1996: p 97 (emphasis added)). If the ordinary language of the workplace were Gaelic, and Gaelic was the ordinary language used by supervisors and managers to communicate with cleaning staff, a language requirement might well be justified for members of the cleaning staff.
(9). Significantly, the Welsh Language Board is currently advancing a range of initiatives to encourage doctors, dentists, midwives and health visitors to use Welsh with their patients. "Congratulations Mrs Jones, it's a boyo", Independent on Sunday, 28 December 1997, p 4.
(10). Note that in rural communities that have undergone relatively recent language shift, autochthonous language use will tend to be concentrated among older sectors of the population, whose needs for caring services will be greater. As such, the need for carers with autochthonous language skills will also tend to be greater than the general level of autochthonous language use in the community would suggest.
(11). Section 5(4) provides that Section 5(2)
does not apply in relation to the filling of a vacancy at a time when the employer already has employees of the racial group in question --
(a) who are capable of carrying out the duties falling within that paragraph; and
(b) whom it would be reasonable to employ on those duties; and
(c) whose numbers are sufficient to meet the employer's likely requirements in respect of those duties without undue inconvenience.