Official Language Policies in Multilingual Societies: Evidence from the United States, South Africa, and Canada

An estimate of the total number of the world's “living” languages in 2005 was 6,912. Since there were concurrently 192 members of the United Nations, it is obvious that very few of these states could be homogeneous, linguistically speaking. Indeed, there may be only one: North Korea. Linguistic pluralism, or heterogeneity, among the world's countries is thus close to being universal.1 This paper is concerned with how three particular sovereign states accommodate linguistic pluralism among their peoples in the daily operations of their national, state, and provincial governments. I am especially interested in whether—and, if so, how—two or more languages regularly function as the acknowledged official languages of such governments, or alternately just one language fills this role, or perhaps, legally speaking, none at all do.2 And then further, I will seek to identify which of various possible contextual “factors” (as I shall refer to them) appear associated with each such result.

My three contemporary cases are the United States, Canada, and South Africa. Each is decidedly plural linguistically.3 The US is listed as having 311 distinct languages spoken within its population of now 301 million persons. One hundred and sixty-two of these languages are indigenous; the remainders are then linguistic “imports.” English obviously is predominant in the US, but the second most commonly spoken language there today is Spanish, which at the time of the 2000 census was the home language of 10.7 percent of the US population. French was a far distant third, being in 2000 the home language of just 0.61 percent of all Americans.4 Meanwhile, among 33.3 million Canadians, 145 distinct languages are spoken, 85 of them being indigenous tongues. Here too, home language English speakers predominate, though less so than in the United States. French, the second most commonly spoken language in Canada, is home language for about 23 percent of the total population, as it was in 1996 for just under 82 percent of the Quebec provincial population.5 Finally, in South Africa, with a total population in 2007 estimated at 47.9 million persons, there are 35 different languages spoken, nearly two-thirds of them indigenous. Speakers of Zulu are the largest home-language group, constituting 23.8 percent of all South Africans in 2001. Xhosa-speakers ranked second at 17.6 percent. So, alone among our three cases, South Africa has, linguistically speaking, no single numerically predominating mother tongue community, and indigenous languages have far greater popular currency than is true in either the US or Canada.

In South Africa, therefore, and most especially since racial apartheid was ended there in 1994, the daily operating needs of government have required crafting of a relatively complicated national official language policy, judged by contemporary northern North American standards.6 This South African policy I shall refer to as “legal multilingualism.” It contrasts with the official language policy in Canada that I shall call “French–English binationalism.” The latter is both more straightforward and comprehensive than contemporary South African legal multilingualism because, unlike it, Canadian binationalism rests on the presumption of differences existing in Canada between two whole cultures—nationalities even—not just differences among more sociologically limited ethnicities, as is generally postulated to be the case in South Africa. But Canadian binationalism is also more severe in its implementation than South African legal multilingualism; that is, it is strict and far-reaching, even intrusive, especially in Quebec.

And both these arrangements contrast with official language policy in the United States, which I will term a policy of “soft monolingualism.” This designation is meant to underscore two different features of current American language policy. One is its unquestioning acceptance of the rightful dominance of English in American national life; the second is the policy's tolerance and compassion for limited English proficiency (LEP) residents, while they are individually accommodating themselves to the dominance of English, or presumably trying to.

I have more to say about each of these national policies below, but hopefully it is clear how fundamentally different they are—philosophically and practically, each one from the other two. My analytic task is to show how one can account on the “input” side of the ledger in these three cases for the “dependent variable” variation on the official language policy “output” side, the national differences of country language policy briefly alluded to above.

America's Soft Monolingualism

At the federal level, the United States has never had an official language of government, just a single unofficial one: English, a language which is presently used and understood “very well” or “well” by perhaps 96 percent of the nation's population.7 Still, beginning in 1981, efforts have periodically been made by a few individual US senators and congressmen to enact into law English as the sole official language of the country. But to date none of these efforts in the congress has succeeded, and many federal government services to the public are in fact available in languages other than English.8 The 2000 federal census questionnaire, for example, was published in six languages, and many federal government agencies now routinely respond in Spanish if contacted by individual members of the public in that language.9

Bilingualism exists officially in the US at state level only in Hawaii, where both English and native Hawaiian are recognized; but it is found informally in several other states. In New Mexico, for example, state laws are published in both Spanish and English. De facto bilingualism in New Mexico is scarcely surprising, for here native speakers of Spanish in 2000 were 28.7 percent of the state's population. This is two percentage points higher than Texas, the state with the next highest percentage of Hispanics in 2000. But with 8.9 million Spanish-speaking persons resident in the state, California had in 2000 by far the highest absolute number of such persons.

Still in California and 25 other American states to date (early 2008), English has been constitutionally, or by simple legislative statute, defined as the sole official language of state government. In at least seven of these twenty-six states (plus Alaska, where the resulting “English-only” statute was later held unconstitutional), the above decisions arose from successful voter initiatives,10 adeptly circumventing in these cases any role in this process for the state legislature. State legislatures are bodies where institutionally entrenched Latino opponents of so-called “English-only” proposals have often proven capable of blocking them.11

An analyst of such matters might be tempted to infer from California's 1986 endorsement of “English-only” for its state government that such stipulations, where they are found, could well be a xenophobic backlash in US states with above-average numbers of residents speaking non-English languages at home. But that surmise would apparently be mistaken. In 2000, the average percentage of such persons among the above twenty-six “English-only” states was actually five percentage points lower than it was among the twenty-four remaining states having no such stipulations on their books. Even Mississippi and Kentucky, each with fewer than 4 percent of their citizens speaking non-English languages at home, have adopted “English-only” provisions. For the country as a whole, the counterpart figure for speakers of non-English languages at home in 2000 was 17.9 percent.12 The California case in this regard is thus somewhat anomalous. Indeed, among the ten US states with the highest percentages of residents speaking a language other than English at home, just two (Florida and Arizona) apart from California have “English-only” state governments. The other seven do not, and this is doubtless due in part to the high unpopularity that must unavoidably attach to any such proposals in states with especially large numbers of native speakers of languages other than English among their inhabitants.

From case to case, the specific wordings of the principle of official English at US state level are inevitably varied, with the result that the on-the-ground practical consequences of these declarations, where they are found, range from the purely symbolic to the far-reaching and dire. Overall, then, official language policy in the United States defies simple or concise description, and in many instances requires a lawyer's training even to understand. But to date, the default position in the United States is that there is no full legal equality for Spanish anywhere at state or federal levels, notwithstanding the fact that with 28.1 million Hispanics/Latinos living in the United States in 2000, the US had then already the fifth largest Spanish-speaking population in the world.13

South African Multilingualism

Meanwhile, in South Africa, following eight decades of quite successful English–Dutch/Afrikaans legal equality in the country's national government, albeit an equality primarily intended for the convenience of the white minority, there have been no fewer than eleven languages now declared officially recognized under the first permanent post-apartheid constitution agreed to in 1996.14 These languages are English, Afrikaans, Zulu, Xhosa, and seven other smaller and internationally less well-known African vernacular languages: Sepedi, Tswana, Southern Sotho, Tsonga, Swati, Venda, and Ndebele. In actual practice, the meaning of this constitutional recognition appears to be this: the South African central (i.e., federal) government can, as a matter of legal right, be required by any citizen of the country to deal with that individual in any one of the eleven official languages, for example, in the courts. But in its daily internal workings, and in the publications the national government issues, only two official languages are required to be used. Yet not necessarily in every case the same two. However in usual practice, English is almost always one of the two chosen, together with an African vernacular language.

Concurrently, at least two of the eleven official languages are required to be used for provincial government business in each of the country's now nine provinces. In practice, English appears to be used in provincial governments not quite everywhere, together with one or more African vernacular languages, and, in instances, Afrikaans too. But across the country in daily life, English has become in fact the country's principal working language and lingua franca, both in and outside of government.

Actual language use in the lower house National Assembly of Parliament in Cape Town mirrors this prevailing pattern. An MP in the National Assembly may choose today to address his or her 399 colleagues in any one of the country's eleven official languages (or the South African sign language), all speeches being simultaneously interpreted into each of the other official languages for the benefit of the full membership.15 Over a four day period at the end of May 2002, picked by me at random, 92 percent of all Hansard-recorded discourses on the floor of the National Assembly were accomplished in English; less than 4 percent occurred in an African vernacular language. Zulu was the vernacular language most often used. As of 2004, parliamentary documents and records were apparently available only in English and Afrikaans, as had been the practice before 1994, but a Language Policy for Parliament document, “edited” in October 2003, projected parliamentary website publication of parliamentary “daily papers” in all eleven official languages by 2008, and of all speeches of MPs “in the languages in which they were delivered.”16

Canadian Binationalism

Contrasting with American official monolingualism, and essentially paralleling the official language dualism in South Africa from 1910 to 1994, Canada authorized in its founding document, The Constitution Act, 1867, use of two languages—French and English—in the federal parliament at Ottawa and in the federal courts of the country. A 1982 redrafted national constitution continued these stipulations, while expressly defining (for the first time) English and French as official languages of Canada. The 1982 document also acknowledged the extension (begun in earnest in 1969) of the principle of French–English bilingualism to the entire federal bureaucracy of Canada, excluding only small, local federal government offices, if the demand there for services in both languages is not “significant.”

The foregoing references are to federal Canada; the constitutional requirements for official language use at the level of Canada's ten provinces are somewhat more nuanced. In 1867, bilingualism was constitutionally mandated for the legislature and provincial courts in Quebec, but only there among the (then) four provinces, and these stipulations dropped out of the 1982 redrafted document. Instead, the 1982 constitution acknowledges English and French as official languages at provincial level only in New Brunswick, where this had in fact been the case legally since 1969. This 1982 stipulation was made explicitly relevant for “all institutions of the legislature and government” of New Brunswick, and for educational and cultural bodies established “for the preservation and promotion” of the English and French linguistic communities there.17

But English–French bilingualism (institutional or personal) has, since the mid-1970s, been in substantial public and legal disfavor outside of federal institutions in largely French-speaking Quebec, and apart from New Brunswick and perhaps eastern Ontario, little needed as a practical matter in the English-dominant provinces elsewhere.18 To be sure, Canada has long had eloquent national leadership advocating a revitalized bilingualism and biculturalism for the country as a whole, and throughout Anglophone Canada, so-called “French immersion” programs have, I am told, recently gained wide popularity among the middle and upper-middle classes. But parallel with this, a longer and contrary tradition of there being in Canada “two [distinct, cultural, and non-intersecting] solitudes”—in fact, two distinct Canadian nations, existing essentially side by side—has remained and, at least in Quebec over recent decades, grown stronger and more strident. It is not without meaning that the Quebec provincial legislature in Quebec City is referred to locally (ordinarily in French, of course) as the Assemblée nationale du Québec. In 1995 in that province, a public referendum proposal that Quebec should actually seek to leave Canada altogether came within 1.16 percent of the vote of actually passing.

Ten Possible Explanatory Factors

Five of the following ten hypothesized independent variable factors arise from extrapolations from a six-fold typology of “major linguistic patterns” around the world, put forward in 1967 by Dankwart Rostow.19 The remaining five emerged randomly in my mind over time, prompted by my readings in the substantial scholarly literature on ethnic conflict and accommodation, especially the writings of Arend Lijphart and Donald Horowitz.20 All ten presume to identify singular conditions that, when present individually, make it more likely, or alternately less likely, that a given non-dominant language in a multi-language setting will be accorded recognition as an official language, rather than simply being ignored and relegated to the policymaking side-lines.

Factor 1 can be stated as a formal hypothesis:

The perceived importance of a domestic non-dominant language

(for those for whom it is not their mother tongue) is enhanced if that language also predominates internationally across the relevant region.

If true, Factor 1 ought to have favored official recognition of Spanish in the United States, alongside English, of course. But I am not aware that the existence of the vast Spanish-speaking world to America's immediate south has had this effect. That is, other than as a source of continuing Hispanic immigration into the US, or by helping to maintain by that civilization's close proximity, currency in the Spanish language among those in the US whose mother tongue it already is. But perhaps I am just uninformed on this point.

But in Africa, the near equally widespread use of English as a vital lingua franca and second language in eastern and southern Africa has undoubtedly solidified support for continuing English's earlier long-time (i.e., since 1910) official status in the republic, by making this decision seem, cross-nationally, an entirely natural and unremarkable one. This is in addition to (1) English's undoubted value, in the absence of other possibilities, for facilitating communication with the outside world, i.e., beyond the continent (Factor 2), and (2) English's continuing ability to serve as a necessary domestic lingua franca within polyglot South Africa itself (Factor 3).

All three considerations made it incontestable that English would be chosen as an official language in the “new” South Africa, notwithstanding that although English is the home language of 40 percent of the 4.8 million whites living in the country, of nearly one-fifth of the 4 million mixed-race coloureds, and of almost all (93.8 percent) the country's 0.9 million Indians, only a tiny 0.5 percent of the 33.9 million black Africans in South Africa speak English at home. This number is marginally smaller actually than the number of Africans who speak Afrikaans at home.21 Ranked according to numbers of home language speakers in South Africa today, English comes in sixth, although only very narrowly behind Tswana, which is in fifth place.22

But neither French in Canada nor Spanish in the US gains equivalent favor to that of English in South Africa as a nationally helpful major world language because both Canada and the US already have one. In Canada, the official recognition of French does qualify the country to belong to the 53-member La Francophonie, symbolically paralleling Canada's longstanding partnership with much the same number of other English-speaking countries in the (formerly British) Commonwealth of Nations. In principle, formal recognition of Spanish in the US might have a similar salutary effect on America's relations with the large Spanish-speaking world. But such a consequence in both these cases is not the same thing as opening up an easy medium of communication with the outside world (or even just parts of it) that would otherwise be linguistically burdened, if not blocked. Few Americans or (I imagine) Canadians seriously question the fundamental adequacy of the English language in their country's routine dealings with the international community, however convenient it may also be for their country's diplomats to be concurrently fluent in one or more other important world languages—Russian, Chinese, etc.

Factor 4 is the idea Rostow suggests that a language's taint or stigma of being alien, or of having “vocabulary elements … reflecting a prenational, imperial past,” could be a crucial disqualification for its later official recognition, should such associations prove difficult to dislodge.23 Such a “taint,” one imagines, interpreted through the peculiarities of the South African case, might easily have scuttled any official language future for Afrikaans beyond 1994, inasmuch as many blacks long viewed Afrikaans as a linguistic handmaiden of apartheid. That is, this might have been the case were it not for two other considerations. One of these is that in 1990, Afrikaans was the third most commonly spoken home language (of a total of nearly three dozen) in South Africa since it was the home language of not only 2.9 million white Afrikaners, but also of a substantial majority (perhaps as much as 80 percent) of the aforementioned coloureds, the latter also, with, of course, all black South Africans and Asians/Indians too, undoubted victims of apartheid. In these matters, there being a relatively large number of mother tongue speakers of a given language (Factor 5) not surprisingly generally counts in its favor, independent of other issues.

Probably more important still is the following: given, as many Afrikaners themselves see it, the heroic rise of the Afrikaans language from the early beginnings of die taal 24 in the second half of the nineteenth century in Africa, it was surely obvious to all (black) African negotiators, from 1990 to 1994, that a decision to discontinue or appreciably down-grade the long official standing (since 1925) of Afrikaans in the country could be a “deal breaker” for the (white) Afrikaner government, then still formally in power. And this awareness occurred at a time when many far more intractable constitutional issues were still pending, problems requiring for their resolution the cooperation of those very same Afrikaner governors, notwithstanding their clear political “lame duck” status.

The eventual deal (or inter-elite pact, as some might see it) on official languages in the “new” South Africa (which incredibly receives just one sentence of coverage in an authoritative 427-page “hidden history” of the multi-year constitutional negotiations that ended apartheid rule25) has been stipulated here already. But surely many of the recognized African vernacular languages, perhaps all nine, lacked in 1994 the substantial literary tradition (Factor 6) or the “diversified vocabulary for modern technology and social organization” (Factor 7) that Rostow emphasized are advantages for relatively minor indigenous languages seeking government recognition as official languages of a country, should these characteristics obtain.26 Zulu and Xhosa (two mother-tongue languages accounting together for about 41 percent of the population) were officially recognized because it was politically impossible not to do so, once it was decided that Afrikaans, notwithstanding its association in the public mind with apartheid, would itself secure such recognition. But once Zulu and Xhosa were “in,” so to speak, it would have been politically almost as difficult to exclude other black African vernacular tongues, and in the end, even the smallest of these, Ndebele, which is the mother tongue of just 1.6 percent of the total population, was also included as an official language.

On the other hand, none of the five non-English traditional Indian languages spoken by some among South Africa's 0.9 million Indians was included, nor were the separate languages of the San, Nama, or Khoi peoples, none of which tongues are ordinarily classified with Bantu African languages. Yet all of these, with six other essentially immigrant languages in South Africa (e.g., Greek, Hebrew), and, interestingly, sign language too, were explicitly designated in Article 6 of the 1996 National Constitution as deserving of “respect” and “promotion.” To help secure these goals, a new Pan South African Language Board was concurrently authorized and charged with that responsibility. Clearly, the political leadership of the “new” South Africa chose in 1994 an unusual “big tent” approach to developing its language policy for the country, leaving what must inevitably be a myriad of implementation problems for another time, should the basic structure of the current policy in practice, and perhaps even in law, not be substantially modified first.

Factors 8 and 9 are relatively straightforward. The former holds that if a language group is not only large, but also territorially concentrated, the argument for officially recognizing that group's language is enhanced, but diminished if the opposite is the case. This is especially true if the political opportunity structure (Factor 9) is itself somehow territorially based or organized. The willingness of a country's courts to “weigh in” in disputes about language rights can also be part of a favorable political opportunity structure, if such interventions occur. If this is significant, it presumably is so independent of population distribution figures. Both these variables have clearly worked together during the past fifty years to strengthen the position of French, in Quebec and in Canada more broadly.27 I am confident, too, that each of these variables also identifies reasons for, if not yet the increasing acceptance of the use of Spanish in the national life of the United States, at least a heightened public appreciation of a need to address the issue in Congress and in several individual US states. I deem the more favorable treatment of Spanish in New Mexico as a consequence of there being a higher percentage of Spanish speakers there (as previously noted) than in any other US state.

The courts of South Africa (there is a constitutional court) have not weighed in yet (that is, since 1994) on a matter of language, as far as I know, and election of members to lower house of the new South African Parliament is now based on proportional representation, a non-territorially based electoral system. Yet because African tribal groups in South Africa each typically have their own traditional homeland in a rural area, wherein their culture and language predominate, the country's newly increased number of federal self-governing provinces (the earlier four are now nine) create possible new arenas for the maintenance and development of several African vernacular languages, albeit in the rural countryside. However, in the polyglot cities of the country's interior, such as Johannesburg, Factors 8 and 9 appear to favor no language at all, except, perhaps because of that failure, English indirectly, as a viable lingua franca.

Indigenousness

My tenth and last suggested factor is this:

If the minority-language candidate for official language recognition is the mother tongue, home language, or favored tongue of an immigrant, diaspora population, consisting of late-comers to the state in question, that language's acceptance for official language status will be more difficult than would be the case if the same language were instead the common tongue of people who regard themselves, and are generally regarded by others, as native or indigenous to the land in which they now live.

The latter are sometimes referred to as a “homeland” population, although the word homeland may carry for South Africans today some apartheid-era connotations that are diverting and analytically unhelpful in the present context, as well as irrelevant. This said, clearly the so-called legitimacy-claim on behalf of languages in the first situation named is different and politically less captivating for out-group “others” than the legitimacy-claim based on “prior occupancy” made on behalf of languages spoken by homeland inhabitants.

While apartheid lasted in South Africa, some of its many African opponents attempted to de-legitimize Afrikaners, and at times all whites living in South Africa, by pointedly terming them “settlers.” This was sometimes referred to as an “Africanist” critique of South African whites. It obviously suggested that South Africa, due to the “prior occupancy” of its several indigenous African peoples, was now by rights a historic “blacks-only,” or at least “blacks-dominant,” society. Resident whites (who at one time locally even referred to themselves as Europeans) were dismissed as just dispossessing foreign “occupiers,” the three centuries’ length of that “occupation” in some parts of the country being held as irrelevant to the central issue.28

Not surprisingly, this kind of attack contrasted sharply with the self-image of most white Afrikaners. Many in this same era openly described themselves as “the first African nationalists” while concurrently publicly celebrating Afrikaans itself as a new, non-European and territorially African language. In the ensuing struggle between these two conflicting orientations, the second ultimately secured the upper hand politically, insofar as Afrikaans, as an authentically African tongue, came to be widely accepted.

The sense of Francophones in Canada that they together, and especially Francophones living in Quebec, descend from one of the country's two “founding peoples,” clearly undergirds their long-standing expectation that the French language in Canada should have an acknowledged, rightful, joint primacy in the public affairs of the country, one that, aside from English, other languages spoken in Canada are simply not entitled to enjoy. Or so the familiar argument goes.29 This is a claim, though one occasionally hears of it being challenged by leaders of various of Canada's Indian tribes (themselves sometimes now referred to as Canadian “first nations”), that is very widely accepted in Canada, and indeed in the world; and it is a powerful legitimization at home of even heavy-handed implementation of Quebec's current official “French-above-all other languages” language philosophy for that province.

When Spain ceded Florida to the United States in 1821, Spain had controlled that New World territory almost uninterruptedly for three centuries. Similarly, before the current US states of California, Nevada, Texas, Arizona, Utah, and parts of Wyoming and Colorado became American territories in the middle of the nineteenth century, this vast region had belonged to Mexico for several decades, and before that, parts of it belonged to Spain for as much as 250 years. The oldest, continuously inhabited city in the present day United States, Pensacola, Florida, was settled by the Spanish in 1559. This was almost a half century before the first English settlers to North America arrived at coastal Virginia in 1607. San Juan Pueblo in northern New Mexico (recently renamed Ohkay Owingeh) was first settled by Spaniards in 1598, and later served for a few years as their territorial capital until that was moved to Santa Fe in 1608. In no valid historical sense, then, can it be said that the Spanish language is less native to Florida and the great American Southwest than English is. Indeed, the opposite contention is an easier argument to defend.

Nonetheless, in the contemporary United States, overwhelmingly it is understood that Hispanics in Florida and the Southwest (not to mention American Hispanics living elsewhere) do not constitute together one or more ages-old “homeland” or indigenous populations, on the model of, say, the Lakota of South Dakota, or Francophones in Quebec. They are viewed as rather more prosaic immigrant diaspora populations, similar to French-Canadians in Maine or the Chinese of San Francisco, albeit vastly more numerous than these, but with precisely the same moral claim that those groups have on Americans generally, not more. In support of this latter contention, recent official population figures can be cited showing, for example, that in the decade of the 1990s, Hispanics in the United States increased their numbers by 58 percent, while the US population overall grew by just over 13 percent.30 This would have been impossible without relatively large infusions of new Hispanic immigrants from elsewhere during this period.

I come now to the reinforcing negative belief among many Americans relative to extending Spanish language rights in the United States. This is the view, held by as many as six in every ten Americans, that illegal immigration into the country, primarily from the south, is a “very serious” problem.31 The bare-bones facts of the case are these. In 2005, it is estimated there were as many as 11.1 million unauthorized immigrants living in the US. This was then almost 4 percent of the total US population. And it is a number one-third larger than its counterpart statistic at the time of the US Census 2000, and, if correct, accounts for fully 30 percent of the total foreign-born population in the US in 2005. Of these 11.1 million illegal immigrants in 2005, 56 percent are believed to have come from Mexico, and a further 22 percent from Central Latin America. Hence, at least 78 percent of the 11.1 million “illegals” were Hispanics, representing about 19 percent of all Hispanics then living in the US.32

Not surprisingly, given these figures, the country's immigration policy, and especially the security of the US–Mexican border, has become a very “live” issue in American politics. So long as this heated controversy continues, it seems unavoidable that, far from being accepted as an indigenous language, a contrary taint of alien-ness will instead attach (doubtless unfairly in very many instances, and certainly ahistorically) to speaking Spanish in the United States, seriously burdening any possible future efforts to sanction the bilingual official use of Spanish in the US government.

Conclusions

Table 1 summarizes the comparative empirical evidence presented in this paper.

Table 1. Favorable and unfavorable factors affecting official recognition of foreign and/or non-dominant languages in multilingual societies

		Canada	RSA	USA	Inter-factor weights
Factor no. (from text)	Issue
1	Regional usage	0	+	+	+2
2	Global access	+	++	0	+3
3	Needed lingua franca	0	+++	0	+3
4	Associated stigma	0	− − −	−	−4
5	Number of native speakers	++	+	+	+4
6	Absence of grand literary tradition	0	+	0	+1
7	Technical vocabulary	0	++	0	+2
8	Territorially concentrated native speakers	+++	+	++	+6
9	Political opportunity structure	+++	+	+	+5
10	Language(s) of indigenous people(s)	+++	++	0	+5
Totals		+12	+11	+4
Likelihood of some recognition of non-dominant languages increases with higher total scores. Scoring Scheme: “+++” = Very favorable; “++” = Moderately favorable; “+” = Slightly favorable; “0” = Neither favorable nor unfavorable; “−” = Slightly unfavorable; “−−” = Moderately unfavorable; “−−−” = Very unfavorable.

We can see that just three of our ten hypothesized test factors (Factor 5, 8, and 9) appear to have been relevant in all three of our cases, although variously so. In particular, the uniquely high concentration (90.7 percent) of all 6.35 million Canadian Francophones in just one province (Quebec), within which province they represented numerically at the time of the 1996 census 81.9 percent of its entire population—these these two statistics, coupled with a Canadian political system within which authority over language policy is shared between the federal government in Ottawa and ten powerful provincial governments, have seemingly virtually guaranteed the continuing preeminence (not just bilingual equality!) of the French language in daily life in Quebec. These three conditions have, working together, presumably also assured the strong lobbying campaign, with its origins perhaps as early as 1963, for strict French–English bilingualism in agencies of Canada's federal government outside of Quebec.

Were these considerations not enough, we have the additional factor (Factor 10) further bolstering both outcomes, namely, that Quebec very early in Canada's history was broadly accepted as the patrie (homeland) of the French people in North America. This view still persists. The existence then of a single, highly concentrated, nationally-minded group, empowered by a very favorable political opportunity structure and a sense of its own historical entitlements, appears to be what distinguishes multinational Canada from multilingual South Africa. South Africa has, as can be seen, many of these same attributes too, but, as Table 1 suggests, in markedly lesser strengths than does Canada.

Save for Factor 10, a similar if weaker argument might well be made for officially recognizing Spanish, similarly bilingually, of course, in the US somewhere, most probably in the great American Southwest, essentially Texas to California. I say “weaker” because of the fact of there being only minority proportions of Spanish speakers in every American state, although significant minority proportions in several of them. A social science “puzzle” is why such official recognition in some of these states has not occurred already. If official bilingualism is politically viable in New Brunswick in Canada, where minority Francophones are 30 percent of the total provincial population, why should official bilingualism be unthinkable in the United States apart from Hawaii—that that is, why not in New Mexico or Texas too, where the proportions of Hispanics in 2000 approached in each case 30 percent of the total population?

Of course, a major difference is that federal Canada is itself bilingual, whereas the US federal government is not. And there is the second distinguishing reason mentioned above. This is of course the widespread public belief that a significant number of Spanish-speaking persons have entered the US illegally. Again, it seems self-evident that it is politically impossible most places in the United States today to contemplate extending language rights to Spanish-speaking populations in the country as long as it is felt that a sizeable number of such persons live in the country illegally. But perhaps this need not be a permanent blockage.

As noted above, the Afrikaans language in South Africa has recently been similarly burdened by its own widely perceived racist stigma. But this stigma was, as I’ve said, likely overcome by the necessary negotiating strategy of the ruling regime's leading extra-parliamentary opponents, in the period just prior to agreement being reached on the country's new constitution in 1994 to 1996. But, as noted, this immediately led to the decision to recognize nine African vernacular languages too, and once this was done, per considerations of Factors 1, 2, 3, 6, and 7, the continuation of English as an official language was itself also assured. As is shown in Table 1, none of these factors necessarily figured in decision-making about official language policy in either Canada or the United States. But in a markedly different political climate regarding immigration policy in the US, Factor 1 considerations might facilitate the official language recognition of Spanish in the future in the US, presumably first at the level of some US state, or states—or, at a minimum, stimulate a serious discussion about such a possibility.

I need to underscore here the greatly added complexity of the South African case that results from the inclusion in the mix of official languages of numerous African vernacular tongues. This makes, unsurprisingly, contemporary South Africa very much like scores of other comparatively young states created in tropical Africa since about 1957 (when Ghana became independent), but very unlike either Canada, excluding remote Nunavut (which is not a province, of course, but a self-governing territory), or the US, exclusive of Hawaii.

And now, some final thoughts on the matter of process. Not surprisingly, in each of our three cases, the initial settling of the official language issue occurred at the times these countries were first established. In the US case, the substance of that decision 218 years ago was of course that the new country would have no official language at all. This meant, in practice, America's de facto working language would be English. Over the ensuing decades and centuries (in two of our cases), the resulting arrangements, different as they were, came to be deeply rooted in the respective national political cultures of these societies, and in their legal systems. Always constitutionally-rooted matters in South Africa and Canada, the absence of any references to official language in the US constitution came to be understood similarly by many in the US, not as a simple oversight of the revered founders, but instead their collective deliberate intent.

This view of the matter in the United States clearly influences the national debate on official language policy in the country today. But it does not imply, even for most of those holding it, an explicit, enforceable constitutional warrant for the status quo language policy position. Indeed, it is precisely because of the absence of such a legal grounding that some in the country and in the congress aspire to change the US constitution to establish one, and thereby create the most legally firm sanction possible. And clearly in the present climate, were there to be such a constitutional sanction, it would be one favoring English alone. But legislating constitutional change for any purpose is a high bar to cross in the US. Meanwhile, and perhaps accordingly, others in the congress have attempted to impact language-of-government issues in fundamental ways through the passing of ordinary legislation. This is certainly an easier course to follow, though to date none of these efforts has as yet proved successful either.33

Never in modern Canada or South Africa have simple legislative majorities been sufficient to change fundamentally the official language policy status quo at national levels, though in Canada several times ordinary federal statutes have been used, as in 1969, to enhance the existing national language policy. National bilingualism before 1994 was constitutionally entrenched in South Africa, as it was from 1867 onward in Canada, a provision reiterated and constitutionally strengthened in the latter case in 1982. In Canada some relief at local levels from unwanted consequences of the country's national legislative straightjacket (as I choose to view it) in this matter was possible, taking advantage of the rights Canadian provinces have had to set language policies for themselves within their own jurisdictions, though not of course for federal entities located within their borders. But until the introduction of a federal political structure in South Africa in 1994, even this kind of locally based deviation from the ordained national pattern was impossible there.

In sum, due to their being embedded in constitutional law, official language policies over the years at the national level in two of our three cases, Canada and South Africa, have been hard to modify or adapt, and intentionally so. South Africa was saved in the last decade from the possibility of political paralysis on this matter because language policy changes in that country could for the moment be folded into the revolutionary, if substantially peaceful, regime change that was concurrently going on, and which then naturally absorbed almost all of the country's available political energies and attentions. Still, the resulting new national language policy for South Africa today has hallmarks of a temporary stopgap, which however could prove difficult either to implement or legally modify in the future. Constitutional change of Article 6 language provisions in South Africa today would require a two-thirds supporting vote in the parliamentary lower house National Assembly, plus the support of six of nine provinces in the upper house. Although this may not represent a great obstacle at the moment under prevailing party-political conditions in the county, my hunch is that in the future actual official language practice in South Africa could be allowed to evolve gradually, independent of what the constitution directs to be the case.

But in Canada, in June 1990, because of a more complex national constitution amending process, it proved relatively easy for just a few disgruntled but strategically situated individuals to sabotage or (more euphemistically said) block prospective constitutional changes that would have impacted, among other matters, the symbolic standing of Quebec within the Canadian confederation. The so-called Meech Lake accord was a package of proposals that the national government of Prime Minister Mulroney and all ten of the provincial premiers spent much time together and not a little political capital developing. Admittedly, the failure of the accord to secure ratification in 1990 did nothing objectively to diminish French language rights in Canada, nor would its passage, had that occurred, have done much similarly to enhance them, at least directly. But subjectively, the failure of that part of the Meech Lake accord that would have placed into the Canadian constitution formal recognition of Quebec's status as a “distinct [Francophone] society” was a serious blow to French ethnic pride in Quebec, and perhaps across Canada more broadly. And this is a sentiment that at its core has long been a pride about French language use.

For many Francophones in Quebec, therefore, the defeat of Meech Lake in June 1990 appears to have felt like a retrenchment in the respect given their language and ethnicity in Canada. More than any other single factor, disappointment at that outcome appears substantially to explain the closeness of the later October 1995 Quebec independence referendum, cited earlier. In the event, as noted, this drastic, if perhaps poorly crafted, proposal was of course defeated, but by such a narrow margin as subsequently to undermine to a degree (I presume) the perceived legitimacy of all positions in this ongoing intra-Canadian dispute regarding Quebec's future status.

Meanwhile, lacking its own constitutional “straightjacket” in this matter of languages used in government, a certain amount of informal official language innovation in the United States has, as I have said, been occurring ad hoc, at both national and state levels, notwithstanding the negative or chilling impact of the present immigration crisis. Perhaps incremental, ad hoc innovation in this topic area is not a bad thing just now, and possibly even indefinitely, given the ever-changing ethnicities of the American population. Indeed, I myself believe this. This then leads me to the following proposition about process and language rights in general, a proposition that I believe each of the three countries we have examined presently tends to confirm, albeit in different ways:

Because (a) official language policies are often seen to require constitutional entrenchment, and (b) constitutions are by design politically difficult to change, in many modern or modernizing societies, (c) official language policies appear in danger of becoming progressively less well adapted to meeting the evolving needs of their communities.

Public acceptance of official language policy as a worthy subject for ordinary legislative action, rather than as fundamental political choices necessitating constitutional entrenchment, should usefully help mitigate this danger, where it exists.

Acknowledgements

The author thanks many colleagues, and especially Tom Karis, Jerome Grieder, Fred Fullerton, and Joseph Copeland for their helpful comments on an early draft of this paper.

Notes

¹ Gordon, Ethnologue, Table 5.

² An “official language” is generally understood to be any language explicitly recognized in law for, at a minimum, daily use in the relevant national or provincial/state legislatures, the courts of the political jurisdictions in question, and the official record of the proceedings of these bodies. Where the law is silent on these matters, the language or languages that are actually used for these purposes are not usually described as “official” (although I have seen this) but rather as “national” languages.

³ Gordon, Ethnologue.

⁴ US Census Bureau, Census 2000, Summary File 3, http://www.census.gov/main/www/cen2000.html

⁵ Countrywide, Chinese in 2001 was the third ranking mother tongue in Canada, spoken by about 2.9 percent of the population.

⁶ The missing North American case in this analysis is obviously Mexico, where (to summarize greatly) the predominance of a single language (Spanish) rivals the predominance of English in the United States. Just under 99 percent of the current Mexican population of 109 million persons are reported to speak Spanish, though for a small minority of these individuals (perhaps 6 percent) who are bilingual or multilingual, their mother tongue is one of the country's sixty-two recognized surviving aboriginal languages.

⁷ US Census Bureau, Census 2000, Summary File 3, http://www.census.gov/main/www/cen2000.html

⁸ In 1996, the US House of Representatives passed a bill declaring English to be the country's only official language, but it died in the Senate.

⁹ Executive Order No. 13166, signed by President Clinton on 11 August 2000, directed all federal agencies of government to ensure “meaningful access” to their programs and activities for “eligible persons” with limited proficiency in English as a result of their national origin.

¹⁰ Crawford, James, “Issues in US Language Policy: Language Legislation in the USA,” http://ourworld.compuserve.com/homepages/JWCRAWFORD/langleg.htm

¹¹ Santoro, “Conventional Politics Takes Center Stage,” 901.

¹² US Census Bureau, Census 2000, Summary File 3, http://www.census.gov/main/www/cen2000.html

¹³ An intriguing local exception to this prevailing national pattern was provided in August 2007 when El Cenizo, a small Texas border town on the Rio Grande River, adopted Spanish as the only official language of its municipal government, and concurrently declared the town a sanctuary for illegal immigrants in the US. Yahoo News, 16 August 2007. Available at http://www.englishfirst.org/elcenezo/elcenezoreuters81689.htm

¹⁴ Afrikaans replaced Dutch as one of the two official languages of South Africa in 1925. An interim, post-apartheid, transitional constitution was in place for the years 1994 to 1996 while the details of the 1996 “final” constitution were being negotiated.

¹⁵ Parliament of the Republic of South African in Cape Town, National Assembly Guide to Procedure, 2004, 56. Available at http://www.parliament.gov.za/content/GUIDE.pdf

¹⁶ Library of Parliament, Language Policy for Parliament, 4.

¹⁷ Constitution Act, 1982, Part I, Articles 16(2) to 19(2).

¹⁸ Quebec Bill 101, passed in 1977, has since required, among other matters, even private businesses in Quebec to communicate with their employees solely in French. The official language position in Manitoba is legally unclear. Manitoba joined Canada as a bilingual province in 1870, but legislatively abandoned official use of French in 1890. However, in 1979, the Supreme Court of Canada declared that 1890 decision unconstitutional, and six years later the same court ordered the province to translate all its unilingual (English only) laws into French within three years. In 1996, native speakers of French were 2 percent of the Manitoba population.

¹⁹ Rostow, “Language, Modernization, and Nationhood,” 87–107.

²⁰ Particularly influential with me were: Lijphart, Democracy in Plural Societies; and Horowitz, Ethnic Groups in Conflict.

²¹ Encarta, “South Africa Facts and Figures,” http://encarta.msn.com/fact_631504863/South_Africa_Facts_and_Figures.html

²² International Marketing Council of South Africa, “The languages of South Africa,” http://www.southafrica.info/ess_info/sa_glance/demographics/language.htm

²³ Rostow, “Language, Modernization, and Nationhood,” 98.

²⁴ This is a common colloquial reference in Afrikaans, to the Afrikaans' language itself.

²⁵ Spitz, The Politics of Transition, 42.

²⁶ Rostow, “Language, Modernization, and Nationhood,”, 98.

²⁷ Regarding judicial interventions, see MacMillan and Tatalovich, “Judicial Activism vs. Restraint.”

²⁸ In 1996, Coertzee noted in his book entitled Giving Offense (1–3) that, in addition to underscoring the alleged transient nature of whites residing in South Africa, the use of the term “settler” in the 1980s when referring to whites by such groups as the military wing of the Pan Africanist Congress, also asserted implicitly an unprecedented power of black Africans in South Africa to “name” (i.e., to publicly characterize, or define) their adversaries. Previously, according to Coertzee, such “naming” of especially racial groups in the country had been de facto a prerogative of the dominant white authorities. Together, Coertzee argues, these two distinct “messages” of the use by blacks Africans of the “settler” label “struck a particularly sensitive nerve” among white South Africans during apartheid's final decades, especially the frequently heard PAC rallying cry “One Settler, One Bullet.” The concurrent resonance of these two messages among the mass of black Africans in the country is however less clear.

²⁹ A possible new exceptional case is Nunavut, a self-administered Inuit-majority territory of 31,000 persons in the eastern Canadian Artic, created in 1999. Within Nunavut there are now four official languages: French, English, Inuktitut, and Inuinnaqtun.

³⁰ US Census Bureau, Census 1990, http://www.census.gov/main/www/cen1990.html; US Census Bureau, Census 2000, http://www.census.gov/main/www/cen2000.html

³¹ Pew Hispanic Center, “The State of American Public Opinion on Immigration in the Spring of 2006. A Review of Major Surveys,” fact sheet, 17 May 2006, http://pewhispanic.org/factsheets/factsheet.php?FactsheetID=18

³² Passel, Jeffrey S, “The Size and Characteristics of the Unauthorized Migrant Population in the US: Estimates Based on the March 2005 Current Population Survey,” Pew Hispanic Center Research Report, 7 March 2006, http://pewhispanic.org/reports/report.php?ReportID=61

³³ For a recent review, see Congressional Research Service, “CRS Report to the Congress: English as the Official Language of the United States. Legal Background and Analysis of Legislation in the 110th Congress,” updated 25 January 2007, 2–7.