indigenous languages, colonization, European languages, language endangerment, African languages, diversification, Romance languages, language attrition, Mufwene, vernaculars, Oxford University Press, European colonization, influence, European language, populations, language evolution, endangerment, substrate influence, genetic linguistics, Native American languages, Cambridge University Press, Bantu languages, Brenzinger, Celtic languages, classical languages, linguistic diversification, West Germanic languages, Khoisan languages, Western European languages, England, Western European countries, Middle English, Nettle & Romaine, native speakers, Grenoble & Whaley, Editions Odile Jacob, European populations, white collar, language shift, minority languages, indigenous African languages, indigenous populations, economic system, lingua franca, lingua francas, indigenous language, endangered languages, ancestral language, Mouton de Gruyter, Native Americans, M. Paul Lewis, Romance linguistics, References Brenzinger, socio-economic system, Magnus Huber, Lindsay J. Whaley, Northern Arizona University, Indo-European
LANGUAGE ENDANGERMENT: WHAT HAVE PRIDE AND PRESTIGE GOT TO DO WITH IT? Salikoko S. Mufwene University of Chicago 1. Introduction A lot has been written on language endangerment since the 1990s, especially since Krauss (1992). Recently the literature has been augmented by books such as Crystal (2000), Dixon (1998), Grenoble & Whaley (1998), Hagиge (2000), and Nettle & Romaine (2000), as well more regionally focused publications such as Cantoni (1997), Brenzinger (1998), and Reyhner et al. (1999). Generally this literature has expressed apprehension, primarily deploring the fact that linguistic diversity is being lost at a dramatic speed, and occasionally arguing that the native speakers of these languages are losing their cultural heritage. In Nettle & Romaine, the latter concern is highly melodramatized. The reader may easily forget to think beyond European colonization as an ecological reason that has accelerated language endangerment around the world. They may also fail to note that the speed and gravity of the process have not been uniform everywhere. Yet it is useful to figure out why. With the exception of the contributions in Brenzinger (1998), which deal mostly with the expansion of some African languages
at the expense of other indigenous languages, most of these publications decry the powerlessness of the colonized populations and the fact that colonization has left them no choice but allegedly to lose pride in, and shift from, their ancestral languages. The analyses of the process have tended to be uniform, revealing little variation from one part of the world to another, even in such a well documented book as Nettle & Romaine (2000), which, for example, highlights accurately the fact that the extinction of indigenous
2 languages has been more severe in Australia than in Papua New Guinea
. Although both countries have been colonized by United Kingdom
, one was a settlement colony and the other an exploitation colony. As I show below, this difference is generally correlated with variation in colonization styles, which fostered different kinds of relations between the colonizers and the colonized populations and affected the indigenous life styles much less in exploitation than in settlement colonies (Mufwene 2001). The literature has said very little about costs and benefits to the affected populations, especially from the point of view of how they have adapted to changing socio-economic ecologies. One exception to this trend is Ladefoged (1992). He asks whether linguists are justified in condemning the fact that some people have found it more advantageous for them to shift from their ancestral language to another which they find more useful. The dominant focus on the geographical expansion of Western European languages both as vernaculars in former settlement colonies and as lingua francas in former exploitation colonies has in fact left little room for examining the world order in the coexistence of the indigenous languages.1 Hence, competition and selection among languages sharing the same socio-economic niche have been interpreted mostly in terms of economic power relations, to which Dorian (1988) responds aptly with the phrase "ideology of contempt." It is therefore not surprising that loss of pride and lack of prestige, both due to being ranked ethnographicaly at the bottom of the repertoire of language choices, have routinely been invoked as explanations for why speakers of indigenous languages in former colonies have given them up. Examination of language 1 Brenzinger 1998 is exceptional, naturally because this coexistence is almost the book's focus.
3 endangerment under relatively egalitarian situations, such as why an indigenous vernacular is losing ground to another, or why French and Spanish are losing some of their imperial lingua franca functions to English, could have suggested different explanations. Ladefoged's position could have been givenmore serious consideration thanit has. Because it focuses onindividual speakers, and proprietors, of a language as the main agents of its evolutionary trajectory, including giving it up if ecologically necessary.2 I argue that this indifference is in part associated with a general misunderstanding of ecological factors that lead to language attrition and death. Linguists have seldom interpreted the processes of language attrition and death first as the results of adaptive responses of speakers to changing political and socio-economic conditions around them. Perhaps guided by a static notion of culture, they have capitalized on the price that the affected populations had to pay in losing their ancestral heritage. They have given little attention to what the populations have gained, or just hoped to accomplish, in the changing socio-economic ecologies they experienced. Grenoble and Whaley (1998) are an interesting exception. Observing that "speakers abandon their native tongue in adaptation to an environment where use of that language is no longer advantageous to them" (22), they try to determine how different ecological factors are weighted relative to each other and point out the greater significance of socio-economic factors. I focus precisely on this ecological aspect 2 I use the term ecology here in reference to what is identified in Mufwene (2001) as "external ecology," referring to the sociohistorical setting in which a language is spoken, one that is largely determined by socio-economic factors. As for the term evolution and its derivative evolutionary, which recur in this essay, I use it not only for change in structure and in pragmatic constraints associate with a language but also in relation to whether it will thrive, lose vitality, or die.
4 below and argue that most of the factors traditionally invoked to account for language attrition and death are not as significant as they are made to be, for instance, the roles of education, of the media, or of the power and prestige associated with particular speakers in promoting and spreading the dominant population's language variety. 2. Does Literacy Prevent Language Endangerment? Much of the literature has invoked prestige of the prevailing language and the affected population's lack of pride in its linguistic heritage among the most important reasons for the decreasing number of speakers of various languages, hence their endangerment. Literacy in the prevailing language has also been invoked as a factor in its favor, over its competitors. Thus, some of the efforts to preserve (indeed not to maintain nor to revitalize!) some of the moribund languages, have focused on developing a writing system and literacy in it (e.g., Hinton 1995). The adequacy of the latter approach to language endangerment is questioned by the fact that the most celebrated dead languages, e.g., Hittite, Sanskrit, ancient Greek, Latin,3 and Gaelic, have bequeathed us rich literary traditions. It is obvious that writing systems and literacy among their speakers did not prevent their death. More evidence against the significance of literacy in the maintenance of languages can be found the history of European colonization since the 16th century. Several written European languages lost the competition to another European language, for reasons that were largely economic or political. For instance, in the United States Dutch, French, German, Italian, and a host of others were given up as vernaculars even before some native American language
s were. English speakers were 3 I return to the question of whether Latin is really a dead language below.
5 not necessarily more literate; as a matter of fact a large proportion of them before the Revolution were not. The increasing usage of English in the colonial economic system, aided by political factors, disfavored the other languages. It promised opportunities the others did not. It appears that the development of writing systems for, and literacy in, some endangered languages guarantees not their revitalization but their (lifeless) preservation like preserves in a jar. The above observation is indeed true of, for instance, Classical Latin and Ancient Greek, which have been well preserved through the written record. One can in fact infer their grammatical systems through these texts. However, the finiteness of the texts is evidence of absence of vitality in these languages. Part of the evidence that these classical languages are no longer alive lies in the fact that no texts have been produced in themsince theyceased being languages of scholarship, as much as by the fact that the scholars who used them as lingua francas did not feel free to innovate in their structures.4 Seldom have living languages preserved such frozen systems, with their users so constrained from innovating. The scholars who used Latin and Greek to disseminate their scholarly findings did not use them as productively as Cicero and Homer, for instance, did. As I show below, the development of Romance languages and Modern Greek probably owe nothing to deviations by the post-GrecoRoman authors
from the original norms. Instead, these authors contributed conservative 4 A reviewer of the manuscript of this book correctly observed that the Latin and Greek used by modern scholars were medieval, not the classical varieties. As important as the distinction must be to classicists, it remains that both the classical and medieval varieties are generally considered dead, though the case of Latin is less clear, as I show below. What bears on this essay is that, if one chooses to argue that Latin never died, because it turned into Romance languages, that particular Latin is Vulgar Latin, the counterpart of the nonstandard vernaculars today.
6 prescriptivism to the grammars of modern European languages. As observed in note 4, those who indeed believe that Latin and Greek are not dead because they have continued in forms of respectively the Romance languages and Modern Greek should consider the following observations. The Romance languages developed from Vulgar Latin, the nonstandard and perhaps unwritten varieties spoken by the lay people and Roman soldiers, not from Classical Latin. Modern Greek developed from the Koinй variety, the spoken, "leveled" and impoverished one compared to the Ancient Greek of classical texts.5 Suchevolutionaryevidence suggests that it takes more thanprestige and writing systemfor a language variety to thrive. The gradual shift of the Irish from Gaelic, in all its spoken and written functions, to English is a reminder of the significance of other apparently more critical factors in determining whether or not a language will be eroded by competition from another language in the same ethnographic setting. Dorian (2001) explains in passing how economic pressures on Sutherland Scots fishermen, rather than loss of pride in their heritage, led them to give it up in favor of English, despite their isolation from the mainstream of the British population. As discussed in section 6, the fact that, unlike Native American languages, several languages in sub-Saharan Africa have not been endangered by the European languages brought over by the former colonizers calls for a more adequate understanding of variation in the coexistence and competition of languages for communicative space. 5 The question of whether or not Latin is dead arises also when one considers the fact that Latin is spoken as a lingua franca at the Vatican. Since it has no native speakers in this function, it is similar to pidgins around the world, and to indigenized varieties of European languages in especially Africa and Asia, where they are transmitted from nonnative to nonnative speakers. It is definitely not as dead as Hittite. For in interesting discussion of different ways in which languages can be considered dead, see Hagиge (2000, Ch. 5).
7 3. A Question of Time Depth Needless to say that such classical languages have seldom been mentioned in the recent academic discourse on language endangerment. The omission reflects in part the traditional way genetic linguists have accounted for language evolution. Their cladograms speak loud and clear, as "older" languages speciate into "younger" ones and so forth, never really dying but increasing in diversity. In a way, the term "dead language" contradicts the practice. The same genetic linguistics tradition also hardly mentions the casualties of linguistic diversification. For instance, studies of the development of Old and Middle English
say almost nothing about the concurrent loss of Celtic languages in England up to the 17th century. Gaelic and its influence on Irish English do not come into the picture until the 18th century. Even genetic Romance linguistics, which produced the termsubstratum, says almost nothing about the loss of Celtic and other languages in the Western European countries
where French, Spanish, and Portuguese in particular are now spoken.6 Among other things, such an omission makes it difficult to seek to understand, for example, what has enabled Breton to survive the expansion and prestige of French so much longer than its Celtic neighbors and why Basque may thrive for another while. Mufwene (2001, Ch. 7) shows that one of the things that make it difficult to understand the diversification of the Bantu languages into different groups is the fact that there is little 6 This observation is not about Celtic substrate influence, which is well acknowledged (e.g., Posner 1996). It is about the fact that the literature barely explains that the development of Romance languages is a concomitant of a protracted shift of the Italic, Gaulish, and Iberian populations from Celtic and other indigenous languages (Mufwene 2001). It takes reading sources such as Martinet (1986) and to some extent (Hagиge 2000) to develop this perspective.
8 reference to the languages spoken by the other populations with which the Bantu came in contact during their colonization of Central and Southern Africa. Yet the history of the region shows that the Pygmies and Khoisans, whose demographic proportions have dwindled drastically, used to inhabit much of the territories now occupied by the Bantu populations. The presence of clicks, which are more typical of Khoisan languages, in some of the more southerly Bantu languages (such as Zulu and Xhosa) is in fact evidence of substrate influence from the non-Bantu populations that have been decimated, absorbed, or pushed to the harsher frontiers. In this particular case, literacy does not matter. It is another story whether pride and prestige, rather than economic and military power, do. A similar comment can be made about the dispersal of Indo-European languages. The presence of Basque, Finnish, Lappish, and Hungarian in what can now be described as IndoEuropean territory is comparable to that of pockets of Pygmy languages spoken in Cameroon or Khoisan languages spoken in Namibia, South Africa
, and Tanzania. They are among the rare more indigenous languages of Europe which survived the dispersal of Indo-European populations and their languages (Martinet 1986, Hagиge 2000), suggesting that the displaced languages must have contributed to the speciation of Indo-European through substrate influence.7 The spread and speciation of Latin into today's Romance languages, or of some West Germanic languages into Old English varieties in England and today's modern English varieties around the world (including English pidgins and creoles), are continuations of the expansion of Indo-European. This is a long evolutionary process that has repeatedly produced 7 Hagиge (2000:69f) reports of Cicero complaining in his book Brutus about non-native influence in some Latin of his time that he considered "corrupted."
9 new and more varieties from those that prevailed over their competitors. In settlement colonies, the expansion and diversification are consequences of shifts from other indigenous and non-indigenous languages, and these processes entailed the deaths of several of them. The typical linguistic maps of the spread of the languages spoken by the Jutes, Angles, and Saxons into England, or of the spread of Latin into the European Romance countries, have given the false impression that no languages were spoken there before their arrival. Thus, they have led us to the false assumption that, for instance, English, French, Spanish, and Portuguese as they are spoken in Europe diversified only, or primarily, by internally-motivated change. The same inaccuracy has continued in studies of the later stages of the same languages. Despite our knowledge of the history of the European colonization over the past four centuries, the spread of the same European languages in the colonies among other Europeans from different ethnolinguistic backgrounds (for instance the appropriation of English by Europeans of Dutch and German descents in North America) has been discussed as if no language contact, hence language shift and substrate influence, had been involved. Only in the case of non-Europeans and recent immigrants have these factors been given the attention they deserve.8 In the case of the Americas, comparisons of equally well-informed studies of both the early and later evolutionary stages of the relevant European languages will reveal several similarities confirming Posner's (1996) observation that even creoles lexified by Romance languages and neo-Romance varieties. Colonization and language contact and shift have 8 To be sure variation in the ecologies of contacts, includingpatterns of inter-group interactions and the extent of structural kinship between the target language
s and those it has come in contact with, plays an important role in determining the structural features of the emergent varieties (Mufwene 2001). This matter need not concern us here.
10 repeatedly played a central role in changing the linguistic landscape of our world, in which some languages have prevailed at the expense of others. In the most dramatic literature, language endangerment has typically been presented as a recent development, without time depth in perspective. More recent publications such as Hagиge (2000) and Nettle & Romaine (2000) provide a little of this background but focus more on the present and recent past (especially Nettle & Romaine), because language extinction is taking place at a faster rate. What continues to be lacking is mention of the fact that in the United States, for instance, the current loss of Native American languages is undoubtedly a continuation of the same process that led earlier, or concurrently, to the death of European languages other than English, e.g., Dutch in New Netherland (New Jersey and New York) or French in Maine, and of African languages. This is an evolution that was caused by a socio-economic system that during the colonial period
was controlled primarily by England. With English establishing itself as the language of the economic machinery and of the colonial and post-colonial administrative structure, everybody else that functioned or was involved within the evolving system (including the African slaves) had to learn it. Gradually the prevalence of English as a lingua franca and ultimately as a vernacular was at the expense of those who were integrated, willfully or not, in the system. I return to this below.9 The current academic discourse on language endangerment has also ignored the fact that a concomitant of language spread has been diversification (mentioned above as the focus of 9 Unlike French in Louisiana, which has experienced protracted death, Quйbйcois French has been rescued not so much by the power of ideological commitment to French culture but mostly by its adoption as the language of any level of the economic system. Ideology alone has, for instance, not helped Esperanto gain momentum, nor has it revitalized Irish in Ireland.
11 genetic linguistics), a process similar to speciation in population genetics, according to which members of a species which settle in different geographical areas mayspecialize into different subspecies in response to the divergent ecological conditions of their existence.10 Although nobody will deny the fact that languages seem to have been disappearing at a faster pace over this past century, discussions of the subject matter
have typically lopsided the scenario in the opposite direction of the genetic linguistics bias and have capitalized on loss of languages. However, re-examining language diversification since, e.g., Proto-Indo-European or Proto- Bantu from the point of view of population movements and contacts is likely to reveal that no speciation was simply internally motivated, and what we know of recent history is probably informative about genetic linguistic developments of distant pasts. Population movements and contacts have undoubtedly been the primary catalysts to the linguistic adaptations which have led to speciation. From an evolutionary point of view, the present linguistic landscape of almost any territory is an interesting balance sheet of births and deaths in the history of its languages. A better understanding of the global picture will help us interpret more adequately what is going on today and what role pride, if the label is justified at all, and lack of literacy have played in the current rapid endangerment of languages. 4. Why Some Non-Prestigious Vernaculars Are Not Endangered but Others Are 10 Mufwene (2000, 2001) argues that the analog of a language in biology is a parasitic species, not an organism. Among the many reasons given are the fact that a language varies internally in ways that the species but not the organism trope can handle, it does not evolve uniformly among its speakers, and we constantly face the question of how the evolutionary paths of individual idiolects translate into (variable) evolutionary paths of a language. This is similar to the following question in population genetics: how do selections of genes at the level of individuals translate into (variable) collective selections at the level of a species? There are of course important differences between the biological and linguistic species, but these need not concern us here; they are irrelevant to the present essay.
12 Perhaps linguists should also ask themselves whether the fact that some varieties are ranked low ethnographically and are said to have no, or little, prestige has not been confused with whether their speakers take no, or little, pride in them. Interestingly, there are highly stigmatized language varieties such as Appalachian English, African-American English, and several creoles around the world which do notseemto be particularly endangered bythe more prestigious varieties with which they have coexisted and in which their speakers acquire literacy. An answer to this apparent puzzle may lie in the fact that, despite linguists' common claim that creoles are separate languages relative to their lexifiers, speakers of all these stigmatized vernaculars think that they speak the same language as the prestigious variety in which they are provided literacy. There is between their vernaculars and the standard variety a division of labor
that creates no competition of the sort that would lead to the attrition or loss of the nonstandard and less prestigious ones. While there are socio-economic functions that require the prestigious variety, they socialize in the non-prestigious ones, with which they identify themselves and whose usage is typically considered more intimate and/or personable. In the case of African-American vernacular English (AAVE) and other nonstandard American English vernaculars, their speakers socialize so little across ethnic or other social boundaries
that none of them is endangered by the other. This suggests that geographical or social coexistence of varieties, even one involving a prestigious and one or more non-prestigious varieties, is not a sufficient condition for endangerment if the varieties are not competing for the same communicative functions. The above observations question the role of schooling as an ecological factor that
13 promotes language endangerment. Such an explanation has indeed been invoked to account for the endangerment of Native American languages, with reference to boarding schools of the 19th century in which Native American children were forced to speak only English and punished for using their ancestral languages (The Voegelins & Schutz 1967). It has also been invoked to account for the spread of English in Ireland at the expense of Gaelic. However, Odlin (1997, this volume) argues that Irish English developed and spread primarily through migrant workers
who found the nonstandard English to which they were exposed useful and learned it naturalistically. (This is contrary to the earlier introduction of English in Ireland through the administrative elite and through the school system. In the latter case it was taught like Latin, with more emphasis on reciting paradigms and translation than on oral skills.) The migrants helped it feel useful to the rest of the population, a significant proportion of which learned it the same way, naturalistically, outside the classroom setting. Likewise, we may argue that English has spread among Native Americans and endangered their ancestral languages not necessarily because of school systems which have dispensed knowledge in English but because of a socio-economic system in which it has been increasingly necessary to command English in order to function in the work place and interact with the larger population. It has thus become more and more practical for those wishing to find jobs not available to them on the reservations to speak English. Like in Ireland, it must have become more and more onerous to Native Americans to have to speak both English and an ancestral language that offers them no, or few, economic advantages when they can do well with just one language. As much as such adaptations have
14 come at an expensive cost of losing knowledge in their ancestral heritage, it is hardly justified to argue that lack of pride in such heritage is the reason for the shift in linguistic vernacular. Practicality and the principle of least effort seem to be the explanation. Once we put things in perspective, it seems that Native Americans have only belatedly been absorbed by the same process in which African and some European languages died in the United States. Indeed, Dutch, German, and Swedish can no longer be counted among American vernaculars, and French in Louisiana can be declared moribund. 5. A Tale of Two Communities The above observations make it relevant to relate a tale of two nonprestigious American communities. The first is the African-American meta-community on the coast of South Carolina and Georgia, which speaks a nonstandard English dialect that linguists prefer to call Gullah or Sea Island Creole.11 The second is the Ocracoke community, on the coast of North Carolina, which is White and speaks a stigmatized nonstandard dialect identified as Ocracoke Brogue (Wolfram and Schilling-Estes 1995). Both communities have been invaded by outsiders, who are generally Whites and are better off socio-economically. The physical looks of these Island communities have evidently also been transformed. There are signs of urbanization on both islands. Because of the typically segregated residential practice in the United States, the South Carolina and Georgia Sea Islands communities are more segregated than Ocracoke. On both the Georgia and South Carolina islands, the new settlers reside where there are very few 11 I am using the term meta-community here on the model of the term meta-population in ecology, in reference to what is really a set of several communities interconnected by their histories and by dispersing individuals that commute among them.
15 African American
s. There is hardly any cross-race socialization on them. Visual encounters at the supermarket can hardly count, as they often involve no verbal exchange (except at the cash register). These are not the kinds of interaction that may have any significant consequences in terms of language evolution. On Ocracoke, however, the traditional Whites and the newcomers are relatively integrated and there have been marriages among them. As much as linguists have feared that Gullah would soon disappear with the influx of outsiders into the coast of South Carolina and Georgia, Gullah is still viable, albeit as a vernacular which, to be sure, is not spoken of necessity by every coastal African American, but there was also never a time when the linguistic situation was different (Mufwene 1994, 1997). The main danger to Gullah is continuous exodus of its speakers from the coastal African-American communities. On the other hand, Ocracoke brogue is endangered (Wolfram and Shilling-Estes 1995), threatened indeed not by standard English but by the vernacular of the newcomers, mostof whomare fromneighboring mainland communities of North Carolina. Perhaps the relative prestige of a language variety does indeed help it prevail in a particular community. However, it looks like the prestige itself follows from another more fundamental factor thatindependentlyhelps it prevail, viz., the fact thatit offers its speakers a means to function adaptively in a specific socio-economic ecology. More than prestige, usefulness bears on whether or not a language variety will be adopted by another group. A prerequisite in such cases is social, or socio-economic, integration of the communities in contact. The absence of integration fosters separate senses of identity and certainly explains why in general African-American English varieties (including Gullah) need not be considered endangered.
16 There is no particular socio-economic pressure on African Americans
to want to speak like affluent European Americans, beyond what every other American does, viz., emulate standard English in situations that require it but otherwise speak one's own vernacular, which reveals one's identity and loyalty to one's background. Geographical and cultural isolation provides additional reasons why Gullah and Appalachian English, for instance, may not be considered endangered, although those that have relocated have had to adjust to the new local vernaculars.12 There is apparently much more to language endangerment thanpride and prestige, and than literacy and education. We must understand more about the ecologies in which languages are spoken and identify which factors are hospitable or inhospitable to their "healths," so to speak. Space limitations make it difficult to explain fully the assumptions that underlie my arguments here. In a nutshell, I treat a language as a parasitic species extrapolated over the existence of idiolects that interbreed and reproduce successfully (Mufwene 2000, 2001). The life of a language qua species is at the mercy of the ecology in which it is used and it is individual adaptive responses of its speakers that set the patterns of language evolution. Hock & Joseph 12 Some may wonder whether the same kinds of explanations apply to American Southern English. Indeed they do, with the difference that the American South offers its own socio-economic system in which American Southern English is accepted. In-migrants from other parts of the United States are a minority and they have a choice between assimilating to the local norm, which their children typically do, and not assimilating. In-migration to the Southern hinterlands does not have the same socioeconomic significance as in-migration to eastern coastal communities, where the local, island vernacular is not used in the economic system of the nearest mainland city or local vacation resort, or where the in-migrants often become majorities and are more affluent. Social integration works more or less the same way in both mainland and island communities exerting pressure on the minority or less affluent group to accommodate the majority or more affluent. The peculiarity of Georgia and South Carolina Sea Islands is that they are not socio-economically integrated.
17 (1996) lead more or less to this kind of interpretation of things, but my population genetics approach (Mufwene 2000, 2001) prompts the question of when, in terms of what languages they prefer to speak, individual speakers' decisions amount roughly to collective decisions of populations which gradually give up their ancestral vernaculars, and when they do not amount to such collective decisions. 6. Sub-Saharan Africa versus the Americas: Differences in Colonization Styles My tale of two communities can apparently also shed light on the differential impacts of European colonial languages in Africa and the Americas, which I attribute to differences in colonization styles, viz., settlement colonies, more typical of the Americas, versus exploitation colonies, more typical of Africa (Mufwene 2001). European settlement colonies amount to what Crosby (1986) identifies as "Neo-Europes," i.e., places where European colonists dreamed to "be more comfortably European in life styles than at home, not less" (298) and they did everything they could to preserve their European cultural traditions minus their shortcomings. They often wound up with immigrant populations that prevailed numerically over the Indigenous populations
(Dixon 1998) which were originally marginalized from them in the early stages of colonization. Exploitation colonies were typically ruled by a small minority of European colonists who came to serve metropolitan companies and their governments on a short term, after which they returned to the metropole or were transferred to another colony. The scheme was to exploit the colonies economically primarily to profit the metropole. Countries such as South Africa
18 and Zimbabwe developed on a combination of both settlement and exploitation colonizations.13 In between these major different schemes stand the plantation island colonies (for instance in the Caribbean and the Indian Ocean
), where the European colonists came to settle too but wound up being minorities among former non-European slaves and indentured laborers. Such territories are also marked by the absence of indigenous populations, because the islands may have been uninhabited (the case of Mauritius) or because the Natives were killed (the case of the Caribbean) or primarily driven away (the case of South Carolina's and Georgia's Sea Islands and coastal areas). A brief comparison of the physical infrastructures of all the three kinds of European former colonies clearly shows this variation in colonial philosophies, which can account for the differing patterns of language evolution in them. European languages have prevailed as nationwide vernaculars in settlement and plantation island colonies but not in exploitation colonies. Thus, some new variety of English has become the vernacular of virtually all Americans, but such is not the situation in South Africa with either Afrikaans or English, in part because of the exploitation colony part of its history. (I discuss this below.) One of the reasons for the alignment of plantation colonies with non-plantation settlement colonies is the exogeneity of their populations, which are now wholly or predominantly of nonindigenous ancestry. The multilingual backgrounds and mixes of the slave and indentured laborers did not favor the retention of their substrate languages in the face of advantages 13 In this comparison, the colonial condition of Africans in South Africa is more comparable to that of Native Americans than to that of African Americans. The counterparts of the latter are the colored people, whose ancestors are at least partly of non-African descent. Zimbabwe does not have a counterpart to the condition of African Americans.
19 provided by command (however variable in proficiency) of the language of the economically dominant group. It was necessary for the exogenous slaves and indentured servants to shift to, and appropriate, their masters' language in order to survive in the then-emerging socioeconomic systems. Unlike in exploitation colonies (see below), there was no tiers of these economic systems that could support the non-indigenous non-European languages. Thus the African languages that the slaves brought with them died gradually, if not rapidly, in these colonies, faster in those settings where there weren't many speakers of the same language. In settlement colonies, the indigenous languages that survived the colonial period and could still operate in the ancestral socio-economic systems have been endangered generally as a concomitant of the subsequent and gradual absorption of their speakers within the global socio-economic systems of the new supra-nations. The indigenous populations have found it more and more useful to speak the local European language in order to function in the global socio-economic system. This adaptive language shift is definitely not unlike the adaptive responses of several European immigrants who likewise found it useful to speak the politically and economically dominant European language of the colony where they settled. As time went by, they found it less and less necessary to speak their ancestral vernacular, a process thatwas made easier by their integration in the new system. Overall, social integration fosters both the emergence within the minority population of a variety which is structurally close to that of the majority population and the gradual loss of the minority's ancestral language. Such shifts to the colonizer's language have not been the case in black Africa's former exploitation colonies The latter have operated on a more sharply-demarcated two-tiers
20 economic system: 1) a small white-collar system which is supported by a European language; and 2) a dominant blue-collar system which is supported by an indigenous lingua franca, such as Lingala in two Congos, Swahili in East Africa and the eastern part of the Democratic Republic of Congo, and Hausa in Nigeria. Understandably, these are the languages that have gained more and more speakers among the indigenous populations and endangered other indigenous African vernaculars. Increasingly the lingua francas have also become the native tongues, and often the exclusive vernaculars, of children born in the city. Lack of social integration within the European populations which runs most of the economic systems has kept even the African elite who use the European languages from appropriating them as vernaculars. The need to live their non-professional lives in African languages and often the pressure to preserve their ethnic identities have kept them loyal to the lingua francas and their ancestral languages. Few of their children speak the European languages natively(except, I amtold, in Gabon, where the elite teachFrenchto their children). Those children who have acquired European languages natively still feel the need to be proficient in an indigenous vernacular or lingua franca to communicate with the majority of the Natives who either do not know, or abhor speaking, the European language. The lower one's position is on the white collar scale and/or the more business one has to handle with blue collar workers, the less one uses the European language, regardless of level of education. Socialization out of the work place is typically also in an indigenous language, except in situations where the European language is the most indicated lingua franca, for instance when such elite who speak different indigenous lingua francas interact among themselves.
21 Despite the dominant use of European languages in the media, the indigenous African languages maintain such an important role in the socio-economic lives of most black Africans that there is no particular reason to see them in competition now with the European languages. The school system has helped educated Africans add a European language to their multilingual repertoire, and for very specific communicative functions that are novel to African heritage, but it has not in any way created situations in which these would endanger the vitality of the indigenous languages. It will take drastic changes in the present socio-economic ecology before we can see this ethnographic situation change. 7. Arabian versus European Colonizations of Africa Putting things in perspective, it is informative to note thatAfrica has been colonized twice by outsiders over the past two millennia, first by the Arabs (who have now indigenized) and then by the Europeans. The Arabian domination of North Africa, which started in the seventh century, was on the settlement-colony model. Having conquered North Africa militarily, the Arabs settled and instituted an assimilationist colonial system that made it possible for the dominated populations to Arabize by converting to their religion, Islam, and adopting some of their other cultural values
. The assimilation of the Natives entailed shifting to colloquial Arabic, which won only a pyrrhic victory, as it was influenced by the substrate languages and thus speciated into varieties that are now different from those spoken in Asia. The biggest ensuing change in the North African linguistic landscape was the attrition and extinction of Egyptian and other indigenous languages of the Lybico-Berber family which are survived today by Tashelhit, Tarifit, Kabyle, Tuareg, and the like, which are now minority languages.
22 The European colonization proceeded differently. The European contacts with the African mainland before the mid-19th century were typically sporadic and associated with trade, although some trade forts with a handful of permanent residents, lanзados, were built here and there along the West African Coast to launch expeditions, through native grumettos, into the interior. That trade colonization produced a Portuguese Pidgin, which served as a generalized trade language for most European traders up to the 18th century (Magnus 1999), and Nigerian Pidgin English, Kru Pidgin English, and Cameroon Pidgin English. It is not even clear how widely these non-Portuguese pidgins were spoken before the 19th century. True to the conventional definition of the term pidgin, they developed from occasional contacts between speakers of their lexifiers, the European traders, and their African customers, especially the grumettos who served as intermediaries between them and the Africans of the interior. Permanent and rather intimate, though dehumanizing, contacts between the Portuguese and Africans took place on islands such as Cape Verde, Sa#o Tome, and Principe, which became the antecedents of settlement colonies on Caribbean and Indian Ocean islands, where the English and French prevailed. They all produced creoles. These have functioned as vernaculars of these islands, leading to the loss of ancestral African languages among the descendants of Africans and, mostly by genocide, also to the extinction of the indigenous languages on the Caribbean islands. Like on the plantation island colonies of the New World and the Indian Ocean, the African islands were peopled with exogenous populations, including the Africans, who were absorbed in the socio-economic system despite the strongest form of discrimination against them. They were forced by the extensive societal multilingualism of
23 their settings to adopt the colonial language as their vernacular, simply because these languages enabled them both to communicate more widely among themselves and to survive, however precariously, the new socio-economic pressures onthem. Interestingly, schoolinghad nothing to do with the shift. Ecological necessity did. The exploitation colonization of Africa by Europeans started after the Berlin Treaty in 1885, which divided Africa among the European colonial power
s: Britain, France, Germany (though it would lose its colonies after World War I), Portugal, Belgium, and to some extent Spain and Italy. Schooling was then introduced as an instrument of colonization, with the partial mission of forming literate civil servants among the Natives. These were also trained both in the local lingua francas and in the colonial language, albeit in a scholastic dialect that was different from the nonstandard vernaculars whose contacts with African languages produced pidgins and creoles in the earlier colonial period discussed above. The colonial societies were clearly segregated, with the civil servants usually serving as a physical buffer between the European colonizers and the Natives. They served as interpreters between the Europeans, although several Europeans indeed learned the local lingua francas. There was no assimilation of the Arabian colonization kind, at least not on a large scale, not even in the Portuguese colonies, where cohabitations of Portuguese men and indigenous women where more common than elsewhere. These are among the few cases that produced not only mulatto offspring but also native speakers of European languages, who still fit in the category of exceptions and did not help spread the colonial languages either. Unlike those who became colonial auxiliaries by education, the mulattos' indigenity was sometimes questioned and they
24 may even have felt more pressure to communicate with the Natives in their indigenous languages. This socio-economic setup, which has continued to date in post-independence Africa, has placed the European colonial languages at the top of the ethnographic ranking of languages. However, it has not made them necessary for the majority of Africans to survive. Thus, as noted above, there has been almost no pressure on the vast majority of black Africans to develop proficiencyin Europeanlanguages unless theywanted white collar jobs. The pressure has been instead to develop proficiency in the indigenous African lingua francas, and these are the ones that have been a threat to the indigenous ancestral vernaculars especially in urban centers. Nowhere do literacy and pride seem to be reasons for shifts from one language to another. Instead necessity and practicality seem to be the most important reasons, just the same reasons highlighted by studies in Brenzinger (1998) regarding contacts of pastoralists and hunter-gatherers. The refugees learn the host population's language. 8. Pride, Prestige, and Literacy: What Have They Got to Do with Language Endangerment? The above discussions should lead us to re-examine at least three reasons which, apparently out of what Dorian (1998) calls the Western "ideology of contempt," linguists have invoked to explain why Native Americans have been shifting from their languages to those of the European colonizers: lack of literacy, of pride, and of prestige. The evidence shows clearly that quite often languages or dialects with the most prestige have not prevailed over their competitors. A case in point is that of Ancient Greek and Classical Latin, which lost to their nonstandard counterparts. These are the ones which have continued in restructured forms
25 through Modern Greek and the Romance languages. It is also noteworthy that all over the world standard dialects have hardly become the prevailing vernaculars of their speakers. Standard French has not yet displaced "les franзais populaires." Colloquial French, rather than the most prestigious variety endorsed by the Acadйmie Franзaise, is the most dynamic French variety spoken today. In English, many nonstandard dialects are far from being moribund. Should they come to be obliterated, it will be colloquial rather than standard varieties that will replace them. In the United States, Gullah and AAVE owe their vitality in part to this ethnographic relationship between the standard and nonstandard varieties, especially to the fact that their speakers can function in the blue collar sector of the economy by modifying their vernaculars only minimally. One can develop the following explanation: speakers of the nonstandard vernaculars feel they speak the same language as the prestigious varieties in which they are provided literacy. At least part of the socio-economic system enables them to function in their nonstandard vernaculars. The division of labor in communicative functions between their vernaculars and the standard varieties makes it unnecessary for them to give up the speech ways that are so natural to them, unless they relocate in different geolinguistic areas. The fact that speakers of the nonstandard vernaculars remain geographically and/or socially isolated from speakers of the more prestigious varieties and do not need the latter varieties while interacting among themselves helps maintain the vitality of the vernaculars.14 14 Likewise, Dixon (1998:82, note 15) points out that "Most of the indigenous languages that survived in Amazonia are spoken by groups which live well away from the main rivers." Under contrary conditions, they would be in sustained contact with mainstream communities and would be under pressure from the dominant socio-economic system to learn the dominant language and/or shift
26 Thus, it appears that the Ocracoke brogue, for instance, is endangered not only because business on the Island has developed in the in-migrants' English vernacular but also because its speakers have been largely absorbed by those of the latter. On the other hand, coastal African Americans, have preserved Gullah because they have not been socially integrated by the more affluent newcomers. Those of them who function in the new economic system try to be bidialectal, developing non-replacive, but simply additional, competence in the inmigrants' variety. A sense of loyalty keeps them from abandoning their variety altogether (Mufwene 1997). The greatest danger to Gullah lies in the numerical attrition of its potential speakers, as more and more African Americans leave the Sea Islands in search of jobs at places where they must shift to the local African-American varieties. However, this trend has been countered to some extent by those who return from the city, out of disenchantment, and tend to speak Gullah with vengeance, often in forms closer to the basilect than spoken by those who have resided continuously on the Islands. On the other hand, division of labor in communicative functions does not explain why Native Americans, like several Europeans before them, have shifted to English in North America. They could preserve the Native American languages for socialization among themselves. Their gradual absorption in the dominant, global socio-economic system must be a large component of the explanation. They differ from Gullah-speakers in that the latter claim they speak English and therefore need not shift to another variety, unless they are away from home. Native Americans maysimplyconsider it more onerous for themto be multilingual in polities to it, therefore their languages would be (more immediately) endangered.
27 where the rest of the population is typically monolingual. The fact that their ancestral languages play no role in the current socio-economic systems deprives them of motivation other then ideology to remain or become proficient in them. In sub-Saharan Africa, the fact that most of the socio-economic system is not global and functions in the indigenous African lingua francas has prevented European languages from endangering the ancestral languages. However, many of the lesser ones are endangered by the Africanmajor languages, typicallythe lingua francas. Populationmovements into and fromthe cities, and more interethnic marriages, are accelerating the process. If relative prestige seems to have played a role in such cases of language shift, lack of pride in one's ancestral language hardly seems to be the reason. In the vast majority of cases, language shift typically is not a conscious decision, either at the level of individual speakers or at the community level. As explained in Mufwene (2001), language shift at the communal level is, like other aspects of language evolution, the cumulative result of decisions made by individual speakers at different speech events. These agents of change are unaware of the long term effects of their communicative practices and regret them when they notice that their beloved language, for which they truly have as much pride as for their underprivileged relatives, is moribund. (It is then often too late to revert the evolutionary course.) All in all, it is practicality and necessity that bring about language attrition and loss among the affected populations, just the same economic reasons that often take underprivileged individuals far away from their families, as much as they would prefer not to part from them, hoping to return to them but often never making it back.
28 The explanation for why neither the Americans of European non-Anglophone descent nor Native Americans have maintained bilingualism may have to be sought in the majority trend. It may appear to be unnecessarily onerous to speakers of such minority languages to be bilingual when the overall society is typically monolingual and the globalizing socio-economic system promotes uniformity. Geographical heterogeneity among languages which are ethnographically equal may be the reason why individual multilingualism is still fashionable in Africa. The tradition of building national unity around linguistic unity in the Western world may be another reason why the Native American languages are endangered now. We should also remember that loss of linguistic diversity in settlement colonies is a concomitant of socio-economic integration, which typically affected Europeans and enslaved populations first, and thenothers. Native Americans are being affected last, because they were last to be integrated in the global economy. On the other hand, the speed of the endangerment and extinction of indigenous languages is faster in North America than in South America
(Nettle & Romaine 2000). The reason is that European settlement colonization is more advanced and pervasive in North America than in South America, where parts of the Amazon forest are still being penetrated. The economies of these two continents are not equally global either, nor are attitudes to indigenous cultures exactly the same. The factors that have produced language endangerment are certainly much more complex thanlinguists have assumed to date. It is certainly dubious that lack of writing systemis a valid explanation at all, especially when we know that languages are primarily spoken. As noted above, some languages with very prestigious literary traditions have not survived today. Also
29 those of the Native American languages which succumbed first to European domination are those that were associated with the Aztec and Inca civilizations (Calvet 1998). If anything writing systems help preserve special forms of languages, such as Old English, Classical Latin, Ancient Greek, Ancient Egyptian, etc. in frozen forms; they do not revitalize them. It is perhaps important too that we in linguistics learn the distinction between preserving a language (like a museum piece), maintaining it in usage, and revitalizing it (by restoring vitality to it). Realistically, we have more control over preservation than over maintenance and revitalization. References Brenzinger, Matthias, ed. 1998. Endangered languages in Africa. Cologne: Rьdiger Kцper. Calvet, Louis-Jean. 1998. Language wars and linguistic politics. Oxford: Oxford University Press. Cantoni, Gina, ed. 1997. Stabilizing indigenous languages. Flagstaff, AZ: Center for Excellence in Education. Northern Arizona University. Crosby, Alfred W. 1986. Ecological imperialism: The biological expansion of Europe, 900- 1900. Cambridge: Cambridge University Press. Crystal, David. 2000. Language death. Cambridge: Cambridge University Press. Dixon, R.W.M. 1998. The rise and fall of languages. Cambridge: Cambridge University Press. Dorian, Nancy. 1998. Western language ideologies and small-language prospects. Endangered languages: Current issues and future prospects, ed. by Lenore A. Grenoble & Lindsay J. Whaley, 3-21. Cambridge: Cambridge University Press.
30 Grenoble,LenoreA.&LindsayJ. Whaley. 1998. Toward a typology of language endangerment.Endangered languages: Current issues and future prospects, ed. by Lenore A. Grenoble &LindsayJ. Whaley,22-54.Cambridge:CambridgeUniversity Press. Hagиge, Claude. 2000. Halte а la mort des langues. Paris: Editions Odile Jacob. Hinton, Leanne. 1995. Current issues affecting language loss and language survival In California. Southwest Journal of Linguistics 14.29-42. Hock, Hans Henrich and Brian D. Joseph. 1996. Language history, language change, and language relationship. Berlin: Mouton de Gruyter. Huber, Magnus. 1999. Atlantic English creoles and the Lower Guinea Coast: A case against Afrogenesis. In Spreading the word: The issue of diffusion among Atlantic creoles, ed. by Magnus Huber & Mikael Parkvall, 81-110. London: University of Westminster Press. Krauss, Michael. 1992. The world's languages In crisis. Language 68.4-10. Ladefoged, Peter. 1992. Another view of endangered languages. Language 68.809-811. Martinet, Andrй. 1986. Des steppes aux ocйans: l'indo-europйens et les "Indo-Europйens." Paris: Payot. Mufwene, Salikoko S. 1994. On decreolization: The case of Gullah. Language and the social construction of identity in creole situations, ed. by Marcyliena Morgan, 63-99. Los Angeles
: UCLA Center for African-American Studies. Mufwene, Salikoko S. 1997. The ecology of Gullah's survival. American Speech 72.69-83. Mufwene, Salikoko S. 2000. Language contact, evolution, and death: How ecology rolls the dice. Assessing ethnolinguistic vitality, ed. by Gloria E. Kindell & M. Paul Lewis. Dallas, TX: Summer Institute
31 Mufwene, Salikoko S. 2001. The ecology of language evolution. Cambridge: Cambridge University Press. Nettle, Daniel&SuzanneRomaine.2000.Vanishing voices: The extinction of the world's languages. Oxford: Oxford University Press. Odlin, Terence. 1997. Hiberno-English: Pidgin, creole, or neither? Occasional Paper
No. 49,Center for Language and Communication Studies, Trinity College, Dublin. Odlin, Terence. this volume. Language ecology and the Columbian exchange. Posner, Rebecca. 1996. The Romance languages. Cambridge: Cambridge University Press. Reyhner, Jon, Gina Cantoni, Robert N. St. Clair, & Evangeline Parsons Yazzie, eds. 1999. Revitalizing indigenous languages. Flagstaff, AZ: Center for Excellence in Education. Northern Arizona University. Voegelin, C. F., F. M. Voegelin, and Noel W. Schutz, Jr. 1967. The language situation in Arizona aspart of the Southwestculture area. Studies in Southwestern ethnolinguistics: Meaning and history in the languages of the American Southwest
, ed. by Dell Hymes and William E. Bittle, 40351. The Hague: Mouton. Wolfram, Walt and Natalie Schilling Estes. 1995. Moribund dialects and the endangerment canon: The case of the Ocracoke Brogue. Language 71.696-721.