The dynamics of linguistic diversity: egalitarian multilingualism and power imbalance among northern Vanuatu languages ALEXANDRE FRANÇOIS Abstract The Torres and Banks Islands, two small archipelagos of northern Vanuatu, are home to 9400 inhabitants and to 17 distinct languages. With an average of 550 speakers per language, this region constitutes an extreme case of the linguistic fragmentation which is typically observed throughout Melanesia. This study presents the linguistic diversity of that area, examines its social underpinnings and outlines its historical dynamics. These islands form an integrated network where a variety of social forces interact, sometimes in conlicting ways. A long lasting bias toward cultural differentiation of local communities has led historically to the linguistic mosaic observable today. This traditional fostering of diversity was correlated with a principle of egalitarian multilingualism. But while these ancient social attitudes have survived to this day, the linguistic diversity of northern Vanuatu has already begun to erode, due to various recent social changes. These changes have reshaped the language ecology of the region and already resulted in the partial loss of earlier linguistic diversity. While northern Vanuatu is still linguistically diverse today, the increased imbalance of power among languages potentially makes the weaker varieties vulnerable in the decades to come. Keywords: Vanuatu; Torres Islands; Banks Islands; linguistic ecology. 1. Multilingualism in Melanesia: two trends in conlict Social networks in the world are potentially subject to two kinds of pressures: centripetal forces, which result in more social integration and more homogeneous social practices — vs. centrifugal forces, whereby individuals or groups emphasize their difference and tend to diverge from each other. The interplay of such conlicting pressures may affect cultural practices as well as linguistic 0165–2516/12/0214–0085 © Walter de Gruyter Int’l. J. Soc. Lang. 214 (2012), pp. 85 – 110 DOI 10.1515/ijsl-2012-0022 86 A. François behavior. Across the world, cultural areas differ in the way they balance homogeneity vs. heterogeneity, depending on an intricate set of geographical, historical and social factors. Thus modern European societies have developed a marked tendency for producing cultural and linguistic homogeneity across vast human networks, thereby erasing earlier forms of heterogeneity. At the other end of the spectrum, a number of small-scale societies in the world show greater tolerance for diversity among local groups: Small-scale societies . . . are economically self-suficient, and proudly form the center of their own social universe without needing to defer unduly to more powerful outside groups. Their constructive fostering of variegation — which holds social groupings to a small and manageable size, and keeps outsiders at a suitable distance — is not offset by the need to align their language with large numbers of other people in the world. (Evans 2010: 14) Such a description would it well the various cultural areas traditionally grouped under the term “Melanesia”. This macro-region consists typically of small-scale, egalitarian societies among which cultural and linguistic diversity is the norm — in contrast, for example, with the politically more integrated societies of Polynesia (Sahlins 1963; Laycock 1982; Pawley 1981, 2007). To take an example from island Melanesia, as many as 106 distinct languages are spoken in Vanuatu (Tryon 1976; Lynch & Crowley 2001) for a population of about 234,000 (VNSO 2009). With only about 2200 speakers per language, this country has the world’s highest linguistic density per capita (Crowley 2000: 50). This paper aims speciically at observing the dynamics of linguistic diversity in the Torres and Banks Islands, a socially coherent cluster of islands located in the northern parts of Vanuatu.1 As many as 17 distinct languages are spoken in this area, for a small population of 9400 inhabitants. The average igure of 550 speakers per language reveals an even higher linguistic density than what is found for Vanuatu as a whole. This extreme language diversity may be approached in two different ways. One possible approach could take it as a given and observe its various linguistic manifestations: thus, in earlier publications, I have endeavored to describe the diversity of these languages’ phonological systems, of their morphosyntactic structures, their lexicons and so on.2 Conversely, rather than taking it for granted, this paper aims at observing this diversity for its own sake, and exploring the social dynamics which underlie it. Based on various ieldtrips carried out in northern Vanuatu since 1997, I here propose to describe the various parameters which shape up this region’s language ecology (Haugen 1972; Mühlhäusler 1996). I will show that the languages of northern Vanuatu are subject to two conlicting tendencies. On the one hand, traditional social practices — in The dynamics of linguistic diversity 87 particular, the spatial anchoring of social groups, as well as egalitarian multilingualism — tend to foster cultural and linguistic diversiication. But on the other hand, contrary forces are also at play, which result in power imbalance between languages, and in the potential erosion of existing diversity. While some of these leveling forces may have always been present in the region, others have been accentuated under modern circumstances: local migrations and community mergers; expansion of the Bislama pidgin; asymmetrical bilingualism, especially in the context of modern formal education. Overall, the mutual balance between competing forces thus appears to change in history. While pre-modern societies in northern Vanuatu would typically give in to pressures of diversiication, more recent changes in their social organization lean towards the erosion, in the long term, of the linguistic heterogeneity that has survived until today. 2. Traditional forces of linguistic diversity This study will begin with a brief presentation of the language situation in the Banks and Torres Islands. I will then illustrate how the linguistic heterogeneity observable today can be explained mostly by longstanding social attitudes which are common in this part of Melanesia. 2.1. The Torres and Banks Islands, a mosaic of languages The Torres and Banks Islands of northern Vanuatu form a relatively small area, with a total land surface of 882 km2. Its modest population of about 94003 is distributed across 10 islands and approximately 50 villages. Even though these communities form together a relatively coherent social network, they speak 17 different languages. These languages are all closely related (see Section 2.2), yet have historically diverged so much that they have lost mutual intelligibility. Figure 1 provides a map of the area, together with an approximate number of speakers for each language. 2.1.1. Varying degrees of vitality. As the statistics on Figure 1 suggest, these 17 languages nowadays enjoy varying degrees of vitality. Four of them are clearly moribund: they are not transmitted to younger generations any more and are only remembered by a handful of elderly speakers. Mwesen, with about 10 speakers, is giving way to the locally dominant language Vurës. The three speakers of Olrat4 have shifted to Lakon, the major language on the west coast of Gaua. Lemerig had ive speakers in 2003 and has now gone down to just two individuals, who live in an area settled by Mwotlap speakers. As for Volow, it is only remembered by the son of its last speaker, who died in 1986.5 88 A. François Figure 1. Map of northern Vanuatu islands, showing languages names with numbers of speakers At the other end of the spectrum, Mwotlap is currently thriving, with as many as 2100 speakers of all ages — among whom 1650 live on Motalava island. By local standards, this is a large community, perhaps even larger than what it was in the 19th century, before the demographic downturn of the 1900s (see Section 3.1.1). Between these two extremes, the other Torres and Banks languages tend to number in the hundreds — from 200 for Lehali, to 2000 for Vurës. The average number of speakers per language is 550 (or 720 if one removes the four moribund languages from the count). In this part of island Melanesia, it appears that just a few hundred speakers may form a viable speech community. Despite igures which seem low by world standards, most of these languages are still healthy today, because — except for the four moribund ones — they are still transmitted to children. In this regard, they are safe from immediate endangerment (see Crowley 1995, 2000). Many of these languages are spoken by a single village or two; the maximum is six villages for one language, and the average is three (50 villages for 17 languages). These villages are usually located on coastal areas, where ishing and inland resources can easily be combined. While language territories all The dynamics of linguistic diversity 89 Table 1. Surviving vs extinct communalects reported in the oral tradition of the Torres and Banks Islands Island Surviving communalects Extinct communalects Hiw Ureparapara Motalava N Vanua Lava S Vanua Lava S Gaua SW Gaua W Gaua Vonqō Lehali Mwotlap (†)Lemerig Vera’a, Vurës, (†)Mwesen Koro, Dorig (†)Olrat Vurē (= Lakon) †Vëqöyö, †Vësëv †Nto †Volow, †Dagmel †Päk, †Tolap †Wetamōt †Viar †Togla, †Qätärew vary in their size, they typically correspond to the distance which an individual can encompass within half-a-day’s walk. 2.1.2. The memory of dialect fragmentation. The linguistic fragmentation of the area is even higher if one considers that some of these languages consist of more than one dialect. Thus, Lo-Toga encompasses two close varieties, spoken respectively on Lo and on Toga. Local populations have also kept the memory of a number of local speech varieties which have gone extinct in the last few generations. Sometimes, a few distinctive words or phonetic characteristics are still remembered today — but usually too little to evaluate whether these varieties were indeed dialects or separate languages. In actual fact, people usually do not draw any distinction between dialect and language; instead, they view all spatially anchored linguistic peculiarities as characteristic of a local “language” tied to a given place.6 In this section, I will use the term communalect as a neutral term for any speech tradition tied to a speciic community. Table 1 lists those extinct (†) communalects of the Torres and Banks Islands, whose existence is still remembered in the oral tradition. Overall, the Torres and Banks islanders have kept the memory of 28 distinct communalects. Of these, 17 are still spoken to this day,7 while 11 are merely remembered for their existence. Besides oral tradition, another valuable source of information in this respect is the detailed survey published by Codrington in 1885. The early missionary linguist describes there the linguistic fragmentation of his time, which appears to be even higher than what is remembered today. Here is how, for example, he describes the island of Vanua Lava. (In this citation, I underline the names of communalects which are still alive today, and italicize those whose existence is still remembered. Note the correspondences of language names: Motlav = Mwotlap, Vatrat = Vera’a, Mosina = Mwesen, etc.). 90 A. François Two small inhabited islets lie close to the eastern side; on one, Raveña, the language of Motlav is spoken, on the other, Qakea, that of Mota. On the island itself each of the districts or groups of villages has its own dialect, viz. Pak, Lusa, Sasar, Leon, Vatrat, Vuras (Avreas), Mosina, Lomrig, Nawono, Alo Teqel, Qatpe, Tolav, and Qe’i. Some of these are, no doubt, very much alike, but the natives themselves thought them different; and between, for example, Pak and Mosina the difference is considerable. The dialect of Nawono, Port Patteson, is lost, the labour trade having destroyed the population, at one time considerable. (Codrington 1885: 331) The last sentence of this quote already foreshadows the discussion I will propose later (see Section 3.1) on possible factors for the erosion of earlier linguistic diversity. But at this point of this study, the crucial observation is that the total number of distinct communalects reported for the Torres and Banks Islands had reached, during the 19th century, a total of at least 35 — of which half have survived to this day. This indicates the extreme degree of linguistic heterogeneity which three millennia of diversiication were able to produce within such a small territory. 2.2. Dynamic processes of convergence and divergence An overview of the linguistic diversity found in northern Vanuatu can be obtained by observing the way in which a random sentence would be translated into the 17 surviving Torres and Banks languages, seen in Table 2.8 Table 2. Linguistic diversity among Torres and Banks languages: an example sentence Hiw Lo-Toga Lehali Löyöp Volow Mwotlap Lemerig Vera’a Vurës Mwesen Mota Nume Dorig Koro Olrat Lakon Mwerlap sisə tati jɵjməgʟen wugʟɔɣ kwe i nihə tat lolmərɛn ʉrβɛ kwɛ e kɛj tɛtnɛ ɣlal ɣalsɛ kwɒ ki͡ɛj tɛ ɣilal ʧøjmat ʧɛk͡pwɛ ŋgɪj ɛt ɣilal ɣalsi tɛŋg͡bwɛ kɪj ɛt ɪɣlal ɣalsi k͡pwɛtɛ tær ɪ ɣɒlɒl ʔørmaʔ ʔæ.kiʔis ndir ɪʔ lamai ɛntɛɣ ʔɪn nɪr ɣɪtɪɣilal warɛɣ tɛn nɪr ɛtɛ lɪlɪ maŋtɛ βɪs nra ɣate ɣlala mantaɣ tk͡pwe nir βitis ɣil liŋliŋi mi nɪr sɔwsɛ βrɪɣɪl taβul tɛ nɪr tɪ rɔŋ taβul wʊs.mɛlɛ nɪj tɪ rɔŋ βɪlɪː wʊs.mɛlɛ ɣɪː atɪ rɔŋ kɛrɛ aβʊh.malɛ kɛr ti βalɣɛ͡ar mɪnmɪn tɪkwɪtɛ͡a 3pl not.yet1 know properly not.yet2 [obl] ‘They don’t know our language very well yet.’ nə nə nnnnɔnɪn ɔ ɔ o u na ɔ məŋa βəɣəβaɣə βap βaβap ɣatɣat hɔhɔlɛ tɛktɛk tɪktɪk k͡pwak͡pw ɣatlɛ βaβae luwluw lŋa βalβalaw ususraː ɛlŋa nɞ- liŋɪ art speech = ta məʈə munɣɛn ŋ͡mwɔni͡ɛn njɔnɣɪn nɔnɔnɣɪn mʊɣʊt mundɪ namøɣynɪn mɔɣɔnin naŋ͡mwonina namɣin -ɣɪn namɪɣɪn mʊʧ -nɣɪʧ -ɣɛ͡an poss:1incl.pl The dynamics of linguistic diversity 91 As Table 2 shows, the coniguration that typically obtains is that a single line of interlinear glossing corresponds to a wide variety of word forms. Both dimensions of this paradox — parallel structures, divergent word forms — call for an explanation (see François 2011). The structural parallelism often found across the 17 languages can be explained by the sustained relations of social contact in which the communities of this archipelago have always been engaged. Relations of trade, exchange, alliances, have deined a social network in which cultural and linguistic contact was the norm (Huffmann 1996, Bedford and Spriggs 2008). Entrenched practices of interisland marriages, whereby women — sometimes men — build a family in a language community different from their own (see Section 2.3), result in several languages being spoken in the same village. Adults are often multilingual, and raise their children in more than one language. Another consequence of this exogamous tendency is that kinship networks extend from island to island, across the entire archipelago. These various factors of language contact have resulted in a strong degree of structural and semantic convergence among the languages of northern Vanuatu, in a way similar to numerous other cases of language contact reported for other parts of the world.9 Considering the degree of contact among these village communities, the real paradox of these languages is therefore not so much their structural similarity, but rather the high degree of diversity found in the forms of their words. This heterogeneity is all the more conspicuous when one knows that the 17 languages all share a common ancestor — Proto Oceanic, the language spoken by the bearers of the “Lapita” cultural complex who irst settled the islands of Vanuatu about 3200 –3000 BP (Kirch 1997; Pawley 1999; Bedford 2006). Both archaeological and linguistic evidence shows that the modern diversity found among Vanuatu languages results neither from any early genealogical diversity, nor from external inputs, but rather from a gradual process of internal linguistic diversiication: The rapid spread of Lapita from the Bismarcks to West Polynesia between 3200 and 2900 BP had a linguistic correlate. The speech of the Lapita colonists in the different island groups must have been relatively homogeneous, little differentiated from Proto Oceanic. . . . After the irst phase of colonisation, the archaeological and linguistic record indicates that in the Southern Melanesian archipelagos, a sequence of demographic and cultural changes occurred which led to weakening or loss of communication between distant sister communities. . . . Most linguistic innovations spread only short distances and the speech traditions of distant communities diverged. (Pawley 2007) What was initially a homogeneous language community turned into a loose dialect network, within which the accumulation of local linguistic innovations gradually increased the divergence between dialects. The linguistic aspects of these processes of diversiication, which involve mostly lexical replacement 92 A. François and sound change, are not problematic per se, and relect universal tendencies in language evolution. What is more problematic is the extreme density of these changes within such a reduced space (François 2011). This density is best explained by a social bias. Deeply entrenched in this part of the world, is the high indulgence for cultural differentiation from one local community to the other. The emergence of diversity did not merely result from geographical isolation and separate development of languages. A key component in the historical process of cultural and linguistic heterogenization, is this ideological bias towards the active differentiation among local communities. It is no accident that each language in this region is named after a speciic place — whether a whole island (Hiw, Mota, Mwotlap, Mwerlap); a bay or coastal area (Löyöp, Vurës, Lakon); a village (Dorig, Vera’a, Mwesen, Volow); or an ancient hamlet which is now abandoned (Lemerig, Nume, Olrat). In these parts of the Paciic, constant reference is made to the precise anchoring of things and people within the social and geographic space. The canvas of toponyms constitutes a chart against which every cultural practice and every social group will be located and identiied. People are often heard commenting, with conspicuous pleasure and excitement, on particular manners of cooking, mat-weaving, dancing or singing, which may differ, sometimes quite subtly, between two local groups. This ideological bias, which is widespread in Melanesia (Thurston 1987, 1989; Dutton 1995), tends to foster linguistic diversity. Should some innovation emerge within a group of a few individuals, it will often spread quickly to an entire village or set of adjacent villages, via individual events of microdiffusion (Labov 1963, 2001) — a phenomenon aptly described as linguistic epidemiology by Enield (2003, 2008). At some point, the spread of said innovation will meet the limits of a speciic dialect or language community, as it is perceived by individuals. Of course, due to areal contact and multilingualism, some of these innovations may spread further to other communities — indeed we just saw this typically happens with syntactic, phraseological and semantic patterns. However, when innovations speciically affect the phonological substance of words (via sound change or lexical replacement), they are typically assigned an emblematic role — namely, that of a linguistic shibboleth that enhances the difference between a particular local community and its neighbors. A typical consequence is that each community will end up having its own word form for a given meaning, often highly divergent from its neighbors. Thus, as Table 2 shows, the words for ‘properly’ or ‘speech’ have diverged to such an extent that each local community has its own distinctive phonological form. This active process of linguistic heterogenization explains how the original unity of the irst Vanuatu settlers eventually fragmented into a mosaic of distinct languages, with different phonologies and vocabularies, as illustrated in Table 2. The dynamics of linguistic diversity 93 2.3. The power of egalitarian multilingualism In a way, this high degree of divergence is paradoxical, considering the amount of contact-induced convergence which also takes place among the same region. However, the two phenomena — socially emblematic differentiation vs. widespread contact — should really be viewed as two sides of the same coin. The reason why Melanesian communities could afford such linguistic diversity is precisely their constant willingness to learn the tongues of their neighbors. Within such a uniied social network as the Torres and Banks archipelago, the indulgence towards language fragmentation is only sustainable as long as the social norm is to preserve egalitarian multilingualism. While linguistic diversity is arguably triggered by the desire for social emblematicity, it needs egalitarian multilingualism to be maintained over generations. These two keywords — “egalitarian” and “multilingualism” — refer to important social attitudes in the region. These small-scale societies are egalitarian in two ways. First, each village community is essentially acephalous, and political power is distributed horizontally across families, with little village-internal hierarchy. Second, the relationship between local communities is one of mutual respect and peaceful alliance, with virtually no relationship of dominance or prestige of one community over the others. This egalitarianism between social groups is mirrored in the balance of power between languages. No language in this region is ever represented as more prestigious, useful, or important than another. In the traditional world, no local community would undergo the pressure to align its language to that of another one. On the contrary, the social and geographical diversity of spatially-anchored groups is expected to be relected in the diversity of their linguistic practices; and all languages of the region are deemed equal in this respect. The second important notion here is multilingualism. I have already mentioned (see Section 2.2) the tendency to marry outside one’s own language community. Marrying outside one’s community is not a rule, and indeed, in large language communities such as Mwotlap or Vurës, it is common to marry someone speaking the same language, and thereby found a monolingual family. However, between 20% and 30% of unions involve spouses from a different island, with a distinct language background.10 As two speakers of distinct languages found a family, they become familiar with each other’s language. Usually, the dominant language in the household will be the one of the village where the couple has chosen to live. In about 61% of cases, the woman relocates to her husband’s village (Vienne 1984: 240), and becomes luent in his language; in 39% of cases, it is the reverse situation. In these mixed couples, the children would normally be raised bilingually (but see Section 3.4 for more recent trends). A corollary of such exogamous tendencies is that any village will include, at any point in time, an immigrant 94 A. François population of men and women originating from a number of different language communities. Thus, while the main language spoken in the village of Lahlap (Motalava I.) is Mwotlap, it is not rare to hear conversations in Hiw, Lehali, Vera’a, Vurës or Mwerlap, as expatriates meet in the lanes of their host village. In the case of Lahlap, the multiplicity of languages does not affect much the native population of the island, who is mostly monolingual in Mwotlap (see Section 3.2). But there are other places in the Banks Islands where multilingualism is much more widespread among the native population. Thus, in the village of Jōlap (500 inhab.) on the west coast of Gaua I., four distinct languages are heard in public every day: Lakon, Olrat (and its variety Viar), Dorig — and occasionally Bislama (see Section 3.4). There is enough multilingualism among the small population of this village for everybody to understand, and occasionally speak, each other’s languages. In such areas, multilingualism can sometimes be observed even at the micro scale of a single household, or for single individuals. Just in the Jōlap family with whom I was staying, I was able to hear four distinct languages (Lakon, Olrat, Mwerlap, Bislama) spoken on a daily basis, plus others (Mwotlap, Dorig, English) which the same people could also speak. Similarly, as I was recording oral literature in various parts of this archipelago, a single individual would sometimes offer to tell me stories in three or four distinct languages. This general propensity to learn other people’s languages, and the general multilingualism which prevails in parts of the archipelago, is arguably an important factor in the preservation of linguistic diversity. Even languages whose irst speakers are few in number can thrive in such an environment, as the wider community can be expected to learn it as a second language. In such a situation, for example, the very few remaining speakers of Olrat hardly feel any pressure to adopt the main language of the village, and they can still go by with their own language in their daily interactions, knowing that they will be understood. Although this moribund language will eventually have to give in to the main language Lakon, the practice of egalitarian multilingualism allows the language shift to be a slow process, spanning over several generations (see Section 3.1). To sum up, the extreme diversity of languages in this part of Melanesia is best explained by the interplay of two complementary social biases: • The key to the emergence of linguistic diversity is the exploitation of language in its emblematic function, as it seals each community’s anchoring in (social and geographic) space. • The key to the maintenance of linguistic diversity is the generalized practice of egalitarian multilingualism, whereby local communities are willing to learn each other’s languages. The dynamics of linguistic diversity 95 The interplay of these two social tendencies has made it possible for a relatively small population — a few thousand individuals — to develop a mosaic of 35 distinct speech traditions, among which half have survived to this day. 3. Recent trends towards the erosion of linguistic diversity The previous section reviewed the various forces at play in the emergence and maintenance of linguistic diversity. Combined together, these factors have fostered a high degree of divergence between locally anchored languages. Yet crucially, such a general balance of forces characterizes the linguistic ecology of this region under what could be labeled “traditional” circumstances — that is, those social practices which prevailed before the 20th century, plus those contemporary practices which continue them today. A number of contrary forces exist, which tend to weaken linguistic diversity. Arguably, some of these forces may have existed in pre-modern times, yet were never powerful enough to counter the opposite tendency towards linguistic fragmentation. However, more recent circumstances, which arose at the end of 19th century, have begun to erode the inherited diversity of languages. The present section will focus on three major social factors which disfavor linguistic diversity: post-contact migrations, and the subsequent reshaping of communities (Section 3.1); asymmetrical bilingualism (Section 3.2), especially in the context of schools (Section 3.3); and the growing inluence of Bislama as a lingua franca, even among close languages (Section 3.4). Combined together, these new conditions tend to weaken the two pillars of linguistic diversity — namely, the emblematic use of language with relation to space, and the tradition of egalitarian multilingualism. This results in a decrease in the number of distinct languages. 3.1. Post-contact social changes and language loss 3.1.1. Demographic changes at a large scale. The second half of the 19th century saw the development of contact with the Western world, in the form of traders, missionaries, or labor recruitment ships. This period of contact resulted in a sudden demographic collapse in the decades around 1900. Vienne (1984: 400) shows that the Banks islands went from about 7,000 inhabitants in 1880, down to 2,000 in 1935: this amounts to a loss of 70% of the population in just a couple of generations. Among the direct causes for this demographic downturn was a series of epidemics, which affected the northern area as much as other islands of Vanuatu 96 A. François (Crowley 1997). Besides its human cost, a side effect of this sudden depopulation was the loss or weakening of several languages. The very low number of speakers for each communalect — often just a few dozens — obviously made them vulnerable to such dramatic demographic change. Roughly at the same time, another cause for massive depopulation was the development of labor trade, or “blackbirding” (1860 –1904), during which a large number of individuals were recruited to work on the colonial plantations of Queensland and Fiji (Fox 1958; Gundert-Hock 1991). As early as 1885, Codrington reported that “the dialect of Nawono, Port Patteson [east Vanua Lava], is lost, the labour trade having destroyed the population, at one time considerable” (see the quotation in Section 2.1.2). These catastrophic events were limited to the few decades around 1900. However, they were later followed by slower demographic processes which took place throughout the 20th century, whereby rural families would seek work in the wealthier cities of Vanuatu, namely Luganville (Espiritu Santo) and Port Vila. For some individuals, the city experience only lasted a few years, before they went back to their home island; but for others, the migration to the city was to become permanent. In these families, the vernacular language may still be spoken for one or two generations — thus, strong Mwotlapspeaking communities can be found in the Mango area of Luganville, and in Port Vila. However, it is common for urbanized children to lose their vernacular in favor of Bislama (see Section 3.4). This being said, the process of rural depopulation, while common in many parts of the world, is still relatively limited in Vanuatu; the 2009 census shows that 76 percent of the population still lives in rural areas (VNSO 2009). Whether they involve the departure of individuals to foreign countries or to urban areas, these forms of emigration constitute a factor in the demographic weakening of small communities, and the potential erosion of linguistic diversity. But interestingly, the modern era has also brought about another form of migration: population movements which took place within the Torres and Banks area, and sometimes within a single island. While these local migrations have not triggered an immediate loss of people or languages, they have resulted in the reshaping of the social landscape of the northern archipelago; as a consequence, they have redeined the power relations among local languages. This social process is the object of the next section. 3.1.2. Local migrations and the reshaping of communities. Vanuatu’s traditional economy combines the horticultural exploitation of land resources with practices of ishing and gathering in coastal areas. The habitat takes the form sometimes of villages on the coast, and sometimes of smaller hamlets in mountainous areas. Part of the linguistic fragmentation which developed in northern Vanuatu — especially as relected by the extinct dialects in Table 1 — relects The dynamics of linguistic diversity 97 this demographic pattern of scattered habitat, as each separate village or hamlet had developed its own language variety. The Torres and Banks Islands still count a number of small hamlets to this day, with sometimes just four or ive households, and no more than 25 or 30 inhabitants. However, according to oral tradition, such dwelling practices were even more common in the past, and have tended to decrease during the course of the 20th century. Several hamlets — especially those located in isolated mountainous areas — became depopulated as their inhabitants relocated to coastal villages (see Vienne 1984: 23). This migratory process often resulted in the merger of two or more formerly distinct communities into one, and in the progressive extinction of some vulnerable varieties under the pressure of a locally dominant language. The motivations for such migrations were multiple. Even in pre-modern times, communities would relocate so as to adapt to the evolution of their land or sea resources, or with the hope to conquer new ground over the wilderness of unexplored areas of their island. This is how, for example, settlers from Motalava began to colonize the northeastern coast of Vanua Lava I., or how Mwerlap speakers settled in eastern Gaua (see Figure 1). Sometimes, a population needed to escape from an area which had become unsafe. For example, the eruption of the Gaua volcano in 2008 forced the relocation of all west-coast villagers to the Nume-speaking area, for two years. Likewise, a cyclone in the atoll of Roua (east of Ureparapara) forced its small population to relocate permanently to the east coast of Ureparapara in the 1950s (Vienne 1984: 39). In this case, the community’s language — now Löyöp — was kept alive, but resulted in the extinction of †Nto, the native dialect of Ureparapara’s east coast. Even though social groups always had reasons for relocating their villages, it appears that the last few generations have seen even more of these local migrations. As we saw above, the series of epidemics around 1900 depopulated villages, in part, due to the actual death of many of its villagers; but it also pushed the survivors to leave their moribund inland hamlets, and join the larger villages on the coast. While contact with the Western world had been the cause of these epidemics, it was also a further incentive for inlanders to settle on the coast, as new commodities (iron tools, medicine, trading opportunities) were coming from the sea. The local migrations thus had an economic component. Sometimes, this attraction towards the Western world took the form of labor migration to foreign countries (see Section 3.1.1), but other individuals chose to work on the local coconut plantations for copra, which had been developed in the coastal areas of the archipelago’s larger islands, Vanua Lava and Gaua. The second half of the 19th century saw the Christianization of northern Vanuatu by the Melanesian Mission, an Anglican missionary organization 98 A. François founded by Bishop George Selwyn (Fox 1958; Hilliard 1978). The Mission chose Mota (Banks Islands) as the language of Christianization, for its biblical teachings and translations. For about three generations, this choice resulted in an increased use of this language across northern Vanuatu islands — at least in church contexts. However, the inluence of Mota was arguably marginal and short-lived; except for a handful of Mota loanwords being adopted here and there (e.g. tataro ‘pray’, wolowolo ‘a cross’, totogale ‘image’), it did not cause any major change in the linguistic practices of the populations. Beyond the use of Mota, the Christianization of the area during the period 1860 –1940 had other indirect effects upon the linguistic landscape of the area — particularly through the way it reshaped the relations between local communities. Some Torres islanders recall how their islands used to be divided by tribal conlicts and ights, until they were paciied by missionaries; the result of this paciication was that formerly hostile communities eventually merged into uniied villages.11 Also, missionaries preferred to build churches on coastal villages, which were easily accessible to ships. Inlanders were encouraged to leave their hamlets and join the new communities (Hilliard 1978), where they could have easy access not only to religious education and celebrations, but also to health facilities. The combination of these various historical events explains why, for example, the mountainous areas of northern Vanua Lava were slowly depopulated during the 20th century. The many inland hamlets of the island, each of which used to be characterized by its own communalect (see the quotation of Codrington [1885] given in Section 2.1.2), merged into the larger villages of Vētubōsō (in Vurës-speaking territory), Vera’a and Mwesen. Today, only two elder individuals can still remember Lemerig, the language of their childhood in the mountains. The same story could be told about the hamlets of Olrat, Viar, Vurē or Qätärew, whose inhabitants all moved down to the west coast of Gaua during the 1960s and 1970s, and merged with the population of Jōlap. Likewise, the oral history of Hiw people tells about various processes of village relocations and island-internal migrations across the last century. Five generations ago, Hiw islanders used to be distributed into ten inland hamlets scattered in the heights of the island, and speaking three different dialects (or languages?): †Vëqöyö in the north, Vonqō in the centre, †Vësëv in the south of the island. Around the beginning of the 20th century, for reasons which have been forgotten — but which result probably from the demographic collapse described above — the ten villages merged into just two villages on the coast: Yaqane and Yawe. Today, the 280 inhabitants of Hiw island speak a single language. In all these cases, the accretion of small hamlet groups into larger coastal communities resulted in the loss of linguistic diversity. For a couple of generations, local immigrants may retain the memory of their distinct origin, and The dynamics of linguistic diversity 99 make a point of speaking their legacy language to their children; this is a way for them to keep using their speech as emblematic of their distinct anchoring in geographic and historical space. However, as further generations grow up in the new village, the meaning of these ancient bonds fades away — especially when they involve a former inland community which no longer exists. Eventually, the pressure of the major language overcomes the need for social emblematicity, and the more vulnerable varieties disappear. 3.2. Asymmetrical bilingualism and power imbalance These migratory tendencies, whereby hamlets would merge into wider communities, had most momentum in the irst decades of the 20th century. Nowadays the process seems to have slowed down, and the hamlets which have survived tend to remain stable. However, the erosion of linguistic diversity still continues, following different processes. One process which can be identiied is asymmetrical bilingualism. I call asymmetrical bilingualism the situation whereby a community speaking language A tends to become bilingual in another language B, while the reverse is not true. Because speakers of B tend not to learn language A, this increases the social pressure upon A speakers to eventually shift to language B. I mentioned earlier the principle of egalitarian multilingualism, whereby all languages were traditionally treated equally. This is true in principle, at least as far as social representations go. Contrary to what is common in other parts of the world, here no explicit hierarchy is ever established between vernaculars, whereby one language would be seen as more prestigious, or socially attractive, than the others. In that sense, social representations still maintain an egalitarian view on language diversity. This being said, the reality of language ecology also involves some de facto imbalance, whereby some languages do prove more inluential than others. This can be understood in demographic terms — one language simply has a greater number of speakers than another — or in terms of social dynamics. For example, some communities may prove particularly more successful than others in their economy, the development of their material culture, the relations with the external world, etc. Currently, such a description may it well languages like Mwotlap or Vurës, which are not only thriving within their own area, but tend to gain ground at the expense of weaker languages. Speakers of these two dominant languages often form monolingual communities, who expect other people to learn their language, and seldom learn others. This comes in contrast with smaller language groups, who usually learn to speak the languages that surround them. Once again, it can be admitted that such imbalance between social groups — some expanding while others modestly survive — must have characterized 100 A. François these archipelagoes at all times, even under traditional circumstances. However, today’s sociolinguistic situation appears slightly different. One clear example of asymmetrical bilingualism is the language Mwesen (10 speakers), southeast of Vanua Lava: all its speakers are bilingual in Vurës — the locally dominant language — while the reverse is not true. The pressure is very high for the younger generations of the Mwesen area to shift their language as they interact with their Vurës-speaking peers. Today, in the village of Mwesen, the vernacular which is most often heard spoken among people is Vurës; as for Mwesen, it is only maintained in the conversations of a few elderly people as they meet in the village. This is also how Volow, the language formerly spoken on the east of Motalava Island, surrendered to the inluence of Mwotlap. A similar imbalance — admittedly less acute — is evident between the two languages of the Torres Islands, Hiw (280 speakers) and Lo-Toga (680). LoToga not only has more speakers than Hiw, it is also spoken in three islands rather than one. The recent airstrip — from where trade goods come in and go out — is located in Lo-Toga territory. Located two hours of motorboat away from these central islands of the Torres group, the people of Hiw sometimes feel they are trailing behind a prosperous Lo-Toga community. To this social imbalance, one may add the widespread belief that Hiw (a highly innovative language) is hard to learn for the non-native, in contrast with the “easy” language of Lo-Toga. The combination of these various factors may explain why the population of Hiw tends to be bilingual in the dominant Lo-Toga, whereas Lo-Toga speakers hardly speak any Hiw. This asymmetry — which is accentuated in the context of school, as we will see below — constitutes a form of power imbalance between the two languages. Even though Hiw still endures as the emblematic language of its island, the situation might result, in the long term, in the increased tendency for its speakers to adopt Lo-Toga as their language — in a way parallel to what already happened for Volow or Mwesen. We saw earlier (Table 1, and Section 3.1.2) that the island of Hiw used to have three distinct speech varieties, which it has now reduced to one. Today, the power imbalance with Lo-Toga constitutes a new threat to the survival of Hiw. This example is a measure of how linguistic diversity in this group of islands has begun to erode dramatically in the last few generations. 3.3. Language relations in the school context The power imbalance between languages exerts perhaps most of its effects at an early age, and this can be observed in a particular context: school. Formal education in modern Vanuatu has various impacts on the linguistic landscape of the population of the Banks and Torres Islands. The dynamics of linguistic diversity 101 Table 3. Main language of education in public and private schools in the province Torba ( Torres–Banks), (Vanuatu Ministry of Education 2009) School level Vernacular English French Total Preschool Primary Secondary 11 0 0 25 16 2 0 7 0 36 schools 23 schools 2 schools total 11 schools 43 schools 7 schools The formal languages of education — English and French — might be said to play some role in the erosion of linguistic diversity. Recent oficial statistics (Government of Vanuatu 2009) show that a fair number of schools use English as their main language of education, as early as the preschool level, shown in Table 3. Arguably, the time spent by young children learning English — and later French — diminishes their exposure to their native language, and this could be seen as a potential threat upon the vernaculars, in a way reminiscent of more heavily colonized countries of the Paciic region. However, the impact of formal education is somewhat limited by the fact that preschools, as well as most primary schools, are usually located within villages, at a walking distance from family homes where the vernacular languages are still vividly used. To this, one may add the fact that early school teachers are sometimes adults from the same community, and naturally tend to address the children in classes using the local vernacular, in spite of the oficial language of education. Finally, the Vanuatu Ministry of Education (2010) has in recent years expressed the intention to introduce vernacular languages in the curriculum of early school years, from Kindergarten to Year 3 students; English and French would be introduced progressively after Year 2. This program, which is meant to be fully implemented by 2025, will hopefully reinforce the sustainability of vernacular languages in formal education. Overall, the exposure of young children to the two colonial languages is therefore currently too supericial to affect their linguistic practices in any signiicant way. In the ideal case where children can remain in their village until at least the age of 12, the pressure from exogenous languages (whether English, French or other vernaculars) upon younger speakers remains limited. But the situation regarding primary schools can be less ideal. For example, many teachers originate from other parts of Vanuatu, and do not know the local vernacular; they communicate with their students either in the formal language of education of their school (English, French), or — more often — in the pidgin Bislama (see Section 3.4). Also, certain small communities cannot afford a primary school in their village; they must send their children to boarding schools, which host children from different language backgrounds. In this 102 A. François case, the languages which most affect the children’s linguistic practices — and thus bear impact on the future of linguistic diversity — are not so much those of formal education, but rather the vernacular languages spoken among their same-aged peers. Thus, the island of Hiw, with only two villages and 44 households (VNSO 2009), has limited school capacities. During their early years, Hiw children attend preschool in their village, and live in their family homes. At the age of 6, some Hiw children attend the small primary school of their island, but others are sent to a larger boarding school on the island of Lo, in the middle of the Torres group further south. During several years — only interrupted by term breaks — these children will live in an environment where the dominant daily language is Lo-Toga: it is the language spoken by most children in the school, as well as the language used by the adults around them. When they come back to Hiw, the children have acquired luency in Lo-Toga, while the reverse situation (Lo-Toga speakers acquiring Hiw) hardly ever happens. This is obviously an important component in the asymmetrical bilingualism observed above between Hiw and Lo-Toga. The pressure towards linguistic homogenization is even more acute when children reach the age of secondary schooling (about 13 years old). The TorresBanks province has only one public secondary school, at Arep, on the eastern coast of Vanua Lava, near the provincial capital Sola. The language of instruction is English, with some presence of French. However, once again, these exogenous languages are currently exerting little pressure upon the students’ linguistic practices; the real competition takes place among the vernacular languages. Unless they have dropped out from school altogether, teenagers from all around the Torres and Banks attend this institution, each with their own linguistic background. They live there sometimes a whole year without returning to their home island. Their exposure to adults speaking their native language suddenly drops, and their only vernacular interlocutors, for long periods, are the children from the same community. As multilingual peer groups form, the natural tendency is for the most dynamic languages to be adopted as a (micro) lingua franca among students. In Arep, the two dominant languages are Mwotlap and Vurës. After a few years there, young speakers of vulnerable languages will have acquired luency in one of these major languages, and dramatically diminished their exposure to their own language. A legitimate question is what precise impact these boarding schools will have upon the linguistic diversity of northern Vanuatu. One unequivocal effect is a considerable amount of language contact at a young age. This, no doubt, participates in the various forms of contact which have always taken place among northern Vanuatu, and have resulted in the convergence of their linguistic structures (see Section 2.2). On the other hand, it is ambiguous whether this melting pot of languages simply results in more multilingualism; or whether The dynamics of linguistic diversity 103 — as may be feared — it contributes to the strengthening of dominant languages, and the potential weakening of vulnerable speech varieties. Perhaps one key observation, in this case, is the asymmetrical form of the process: while young speakers of Lehali (200 speakers) or Koro (250) will tend to learn Mwotlap or Vurës, the reverse does not happen. In the long term, it is likely that such cases of asymmetrical bilingualism may result in the erosion of linguistic diversity as we know it today. 3.4. The growing use of Bislama While the modern era has somewhat modiied the existing balance between vernacular languages, it has also brought in a new language: Bislama. This pidgin has, to a certain extent, reshaped the language ecology of Vanuatu — in particular, by jeopardizing the traditional model of egalitarian multilingualism described in Section 2.3. Bislama is the Vanuatu variety of the English-based pidgins which formed in the South Paciic during the 19th century. These pidgins had particularly developed in the colonial plantations of Queensland and Fiji, where a high number of Vanuatu islanders had sought work during the “blackbirding” period 1860 –1904 (Tryon and Charpentier 2004). After several years of labor, those individuals who made the journey back home often brought with them the pidgin they had used on the plantations. Various regional forms of the pidgin emerged, inluenced by the vernacular substrates (Clark 1979; Siegel 1998); the more or less uniied variety spoken in Vanuatu is known as Bislama. Bislama spread across the whole archipelago of Vanuatu — then the New Hebrides — as a new lingua franca, a process which rose steadily during the irst half of the 20th century. It was useful when communicating with foreigners, either Westerners or other Paciic islanders. The social status of Bislama was reinforced in 1980, when it was chosen as the “national language” of the newly independent Republic of Vanuatu. Nowadays it is the main language used in the media nation-wide — especially on the radio, the only media easily accessible to rural areas. The interplay between Bislama and the Torres-Banks languages is complex, and deserves a detailed sociolinguistic study of its own. A few hundred individuals originating from the Torres and Banks have migrated to Port Vila or Luganville (Espiritu Santo), the two cities of Vanuatu where the pressure of Bislama is intense: in these urban environments, only the irst generation of immigrants (the parents) are luent in their vernaculars; their children tend to use Bislama as their main language, whether within their family or with their peers. This results in language shift in these families, as vernacular languages are progressively abandoned in favor of Bislama. The 1999 and 2009 censuses 104 A. François Table 4. Main language used at home, by regional province: percentages comparing 1999 and 2009 census data Province (N to S) 1999 2009 Local language Bislama other Local language Bislama other torba sanma → incl. Luganville penama malampa shefa → incl. Port Vila tafea 90.6 60.1 23.8 94.1 83.0 50.4 31.2 95.6 8.3 36.2 67.2 5.3 16.0 39.2 52.4 3.6 1.1 3.7 9.0 0.6 1.0 10.4 16.4 0.8 85.6 51.1 14.0 91.8 74.4 39.7 22.4 91.2 13.8 46.5 81.9 7.6 24.8 53.4 67.8 8.0 0.6 2.4 4.1 0.6 0.8 6.9 9.8 0.8 National, rural National, urban 85.3 29.3 13.3 56.4 1.4 14.3 77.1 20.5 21.7 70.9 1.2 8.6 NatioNal 73.1 23.3 3.6 63.2 33.7 3.1 carried out in Vanuatu (VNSO 2000, 2009) provide igures regarding the language of primary use in private households; these are summarized in Table 4. In the two cities Port Vila and Luganville, Bislama has long become a creole — the irst language for a majority of people. In ten years, it has even progressed by 14.5 percent among urban populations, and 10.4 percent nationwide. This expansion was mostly done at the expense of vernacular languages, whose use in the population has receded from 73.1 to 63.2 percent in just one decade. As far as the rural areas are concerned, however, such processes of complete language shift are still limited. As Table 4 shows, the use of vernacular languages is still largely predominant, for example, in the TorBa (Torres-Banks) province. Bislama is still conined there to a role as a pidgin, an auxiliary language which individuals may choose to speak in certain restricted contexts, when the use of the vernacular is not felt appropriate. In a stark contrast with urban areas, it is almost never the primary language of children. In 1998, in the village of Wasag in Vanua Lava, I remember meeting a seven year-old girl who was monolingual in Vurës, and unable to speak Bislama with any conidence. Although such a situation is rare nowadays, it shows that the vernaculars, in rural areas, are still much in use as the default language in households; Bislama is acquired as a second language, typically at school, through interactions with the teacher (see Section 3.3). In sum, Bislama does not constitute a direct threat to the existing linguistic diversity found in northern Vanuatu — at least not in the sense that vernacular languages would be quickly replaced by a new creole (see Crowley 2000: 125). The dynamics of linguistic diversity 105 However, it does show some indirect effects upon the language ecology of the region. This is especially visible if one pays closer attention to the precise contexts where Bislama is being used. These contexts are extremely versatile, and vary depending on such parameters as location, age groups, nature of the conversation or event; furthermore, the tendencies observed today are prone to quick changes, and would deserve to be closely monitored during the next decades. Bislama is the default language in any interaction with foreigners — whether from outside Vanuatu,12 or from other islands of the archipelago. When the interaction involves two individuals from different parts of the Torres-Banks Islands, they might also choose Bislama as their common language — unless their linguistic knowledge is suficient for them to use each other’s vernaculars. The choice of which language to use in the latter case is not obvious, and I have sometimes heard the same individuals hesitate between the two strategies. What results is often some form of code-switching between Bislama and the local languages. Thus, on Motalava I. the doctor is originally from Hiw (Torres), but has lived long enough on Motalava to acquire reasonable luency in Mwotlap. During his private conversations in the village, he would speak Mwotlap with ease; however, in the more formal context of his medical consultations, he would address the very same individuals in Bislama. There is a strong tendency for people to associate public occasions with Bislama. It is heard most often in church13 — except in some villages — and in public announcements during celebrations or community events. The reason sometimes given for this choice of language is that public speeches are addressed to a crowd which might include some external visitor; the choice of Bislama is here justiied by one’s consideration towards the outsiders. However, the association of public speech with the pidgin is so strong that Bislama will often be the language used for just any public speech, even when all the audience speaks and understands the local vernacular. The increasing tendency is thus to use Bislama as a default, for every interaction involving unfamiliar people from different linguistic backgrounds. This is a major change from the traditional habit of egalitarian multilingualism described earlier (see Section 2.3). In reality, both models coexist in contemporary practices and are involved in a subtle competition. For example, in the village of Jōlap in west Gaua, multilingualism is still the norm, because it is intimately woven into the buildup of families and stable personal relationships. Bislama would there be restricted to those occasions when external visitors are present in a temporary visit and could not be expected to know the various local languages. Conversely, in other places in the Banks, I have heard Bislama used even within a single household. At least four mixed couples, with one parent speaking Mwotlap and another one speaking Mwerlap or Vurës (two languages relatively similar to Mwotlap) were systematically addressing their 106 A. François children in Bislama — based on the surprising conception that their children would be unable to handle a multilingual family. Such an attitude is new, and at odds with the tradition of multilingualism which had prevailed until recently. In Section 2.3, I suggested that multilingualism, whereby local communities are willing to learn each other’s languages, was a key to the maintenance of linguistic diversity. The increasing role of Bislama as the vehicular language in interisland communication — even between close languages — potentially jeopardises the fragile balance among the local vernaculars. Traditionally, a dialect would have been learnt by its community as a irst language, but also understood, at least passively, by a wider network of neighbouring communities. Now that Bislama is taking up the role of lingua franca even between close languages, the consequence is for the more modest languages to be known exclusively within their own community, with little opportunity to gain more speakers. The emerging tendency, for mixed couples, to raise their children in Bislama, is still minor in comparison with the tradition, which still prevails, of learning several vernaculars. However, should this habit develop in the next decades, it would constitute another threat to the transmission of the more vulnerable languages. A potential outcome could be the collapse of linguistic diversity as we know it today. 4. Conclusion: the uncertain future of linguistic diversity It would be very dificult to predict what the linguistic situation in northern Vanuatu will be 50 years from now. The high linguistic density observed today might well survive for several generations (cf. Crowley 1995), and small languages coexist with larger communities. In principle, one could even imagine that the existing fragmentation increases, as contemporary dialects could further drift apart, and separate into distinct languages. The healthy demographic increase observed generally in Vanuatu (VNSO 2009) may support such an optimistic view. Alternatively, however, it is likely that the modern trend towards linguistic homogenisation, which can already be spotted nowadays, will become stronger in Vanuatu, as has happened in other parts of the Paciic. In a way similar to the dramatic changes which characterised the irst half of the 20th century, the erosion of linguistic diversity could easily be accelerated, in the future, by major social changes. This would be the case, for example, if small villages continued to merge into larger monolingual communities, if transport or communication technologies went through dramatic improvement, or if new cultural models began to encourage linguistic levelling and unity at the expense of the traditional model fostering spatially-anchored diversity. The The dynamics of linguistic diversity 107 sociolinguistic dynamics of the Torres and Banks Islands — and more generally of Vanuatu, or of the whole Melanesian region — certainly deserve to be closely monitored in the decades to come. Langues et Civilisations à Tradition Orale (CNRS), Paris Australian National University, Canberra Correspondence address: francois@vjf.cnrs.fr Notes 1. Although the label “northern Vanuatu” is sometimes used (e.g. Tryon 1996) to refer to a wider geographical area that also includes islands further south, the present article will consistently use it to refer to just the Torres and Banks (sub-) archipelago. Likewise, the default reference of the word “archipelago” will be the group of Torres and Banks Islands. 2. For a general assessment of the linguistic diversity found in Vanuatu as a whole, see Tryon (1976). For Torres and Banks languages in particular, François (2005) examines the heterogeneity of phonological systems; François (2007) compares the morphosyntax of noun articles; François (2011) discusses the historical processes of divergence and convergence. 3. The 2009 census carried out by the Vanuatu National Statistics Ofice (VNSO 2009) gives a igure of 9359 inhabitants for the province “TorBa” (Torres-Banks). This shows a +20.7% increase from the igure of 7757 inhabitants observed in 1999. 4. The number of Olrat speakers decreased from 4 to 3 during the writing of this study. 5. Volow can be considered extinct now, as it is no longer spoken today. The reason why I still include it here among the 17 languages of the area, is because it is still remembered today by its “last hearers” (Evans 2010: 209), from whom I was able to collect substantial linguistic data in 2003. A valuable recording of the last luent speaker Wanhan, which the anthropologist Bernard Vienne carried out in 1969 and later handed over to me, has been extremely useful in the task of reconstructing — with the help of Wanhan’s children — what spoken Volow used to be like. 6. As we will see, historical processes of linguistic differentiation inevitably begin to turn two dialects into separate languages (see Section 2.2). 7. The total of seventeen here includes the 16 languages still spoken today (i.e. the 17 languages, minus Volow). Additionally, the dialects of Lo and Toga, while similar enough for the linguist observer to lump them as a single language, are considered as two distinct communalects by their speakers. 8. All forms are given in IPA transcription. Languages are arranged geographically, from northwest to southeast. 9. See, inter alia, Gumperz (1971) for northern India; Enield (2003) for southeast Asia; Ross (2001) for contact between Austronesian and Papuan languages of New Guinea. 10. Vienne (1984: 233) conducted detailed statistics on interisland marriage in northern Banks islands, based on data collected in the 1960’s and 1970’s. Out of 455 unions, 88 (= 20%) involved partners from distinct islands. Because the statistics published by Vienne take the island as a unit of observation, they make it dificult to quantify the cases of linguistic exogamy strictly speaking, as single islands typically include several language communities. Thus, a marriage between speakers of Vera’a and Vurës (two languages spoken on Vanua 108 A. François Lava) constitutes a case of linguistic exogamy, even though it does not appear as such in Vienne’s statistics. Should one take the language community — rather than the island — as the pertinent unit, then the rate of exogamy is necessarily higher than 20, and probably closer to 30 percent. 11. François (2009: 106) describes a similar process for the island of Vanikoro, in the nearby Solomon Islands — an area also under the inluence of the Melanesian Mission. There too, earlier tribal conlicts were paciied by missionaries; this eventually led to the merger of three communities into one, and the loss of Lovono and Tanema languages in favour of Teanu. 12. Occasional knowledge of English or French, acquired through school, is sometimes useful when interacting with tourists. 13. There is considerable variation regarding language use in church. First, a number of protestant denominations are found in the region, many of which are run by a minister of foreign origin — whether from Vanuatu or elsewhere; the latter case forces the use of Bislama. The Anglican church, which is strongest in northern Vanuatu, sometimes employs the local vernacular during ofice, but also often resorts to Bislama or even English. The factors involved in the choice include the perceived degree of linguistic homogeneity of the audience, as well as the minister’s own linguistic background. References Bedford, Stuart. 2006. Pieces of the Vanuatu puzzle: archaeology of the north, south and centre (Terra Australis 23). Canberra: Pandanus Press. Bedford, Stuart & Matthew Spriggs. 2008. Northern Vanuatu as a Paciic crossroads: the archaeology of discovery, interaction, and the emergence of the “ethnographic present”. Asian Perspectives 47(1). 95–120. Clark, Ross. 1979. In search of Beach-la-Mar: towards a history of Paciic Pidgin English. Te Reo 22/23. 3– 64. Codrington, Robert H. 1885. The Melanesian languages. Oxford: Clarendon Press. Crowley, Terry. 1995. Melanesian languages: do they have a future? Oceanic Linguistics 34(2). 327–344. Crowley, Terry. 1997. What happened to Erromango’s languages? Journal of the Polynesian Society 106(4). 33– 63. Crowley, Terry. 2000. The language situation in Vanuatu. Current Issues in Language Planning 1(1). 47–132. Dutton, Thomas E. 1995. Language contact and change in Melanesia. In Peter S. Bellwood, James J. Fox & Darrell Tryon (eds.), The Austronesians: historical and comparative perspectives (Comparative Austronesian Project), 207–228. Canberra: Australian National University. Enield, Nicholas. 2003. Linguistic epidemiology: semantics and grammar of language contact in mainland Southeast Asia. London: Routledge-Curzon. Enield, Nicholas. 2008. Transmission biases in linguistic epidemiology. Journal of Language Contact Thema 2. 299–310. Evans, Nicholas. 2010. Dying words: endangered languages and what they have to tell us (The Language Library). Hoboken, NJ: Wiley-Blackwell. Fox, Charles Elliot. 1958. Lord of the Southern Isles: being the story of the Anglican Mission in Melanesia 1849–1949. London: A. R. Mowbray. François, Alexandre. 2005. Unraveling the history of the vowels of seventeen northern Vanuatu languages. Oceanic Linguistics 44(2). 443–504. The dynamics of linguistic diversity 109 François, Alexandre. 2007. Noun articles in Torres and Banks languages: conservation and innovation. In John Lynch, Jeff Siegel & Diana Eades (eds.), Language description, history and development: linguistic indulgence in memory of Terry Crowley (Creole Language Library 30), 313–326. New York: John Benjamins. François, Alexandre. 2009. The languages of Vanikoro: three lexicons and one grammar. In Bethwyn Evans (ed.), Discovering history through language: papers in honour of Malcolm Ross (Paciic Linguistics 605), 103–126. Canberra: Australian National University. François, Alexandre. 2011. Social ecology and language history in the northern Vanuatu linkage: a tale of divergence and convergence. Journal of Historical Linguistics 1(2). 175–246. Gumperz, John J. 1971. Convergence and creolization: a case from the Indo-Aryan/Dravidian border in India. In Language in social groups: essays by John J. Gumperz (Language science and national development 3), 251–273. Stanford, CA: Stanford University Press. Gundert-Hock, Sibylle. 1991. Mission inluence and labour migration: the case of Vanuatu in the last decades of the 19th century. The Journal of Paciic History 26(1). 98–102. Haugen, Einar Ingvald. 1972. The ecology of language. In Anwar S. Dil (ed.), The ecology of language: essays by Einar Haugen (Language science and national development), 325–339. Stanford: Stanford University Press. Hilliard, David. 1978. God’s gentlemen: a history of the Melanesian mission, 1849–1942. St. Lucia, Q.: University of Queensland Press. Huffman, Kirk. 1996. Trading, cultural exchange and copyright: important aspects of Vanuatu arts. In Joël Bonnemaison, Kirk Huffman, Christian Kaufmann & Darrell Tryon (eds.), Arts of Vanuatu, 182–194. Bathurst: Crawford House Press. Kirch, Patrick Vinton. 1997. The Lapita peoples: ancestors of the Oceanic world (The Peoples of South-East Asia and the Paciic). Cambridge: Blackwell. Labov, William. 1963. The social motivation of sound change. Word 19. 273–309. Labov, William. 2001. Principles of linguistic change: social factors. Oxford: Blackwell. Laycock, Donald C. 1982. Linguistic diversity in Melanesia: a tentative explanation. In Rainer Carle, Martina Henschke, Peter W. Pink, Christel Rost & Karen Stadtlender (eds.), Gava’: Studies in Austronesian languages and cultures, dedicated to Hans Kähler, 31–37. Berlin: Dietrich Reimer. Lynch, John & Terry Crowley. 2001. Languages of Vanuatu: a new survey and bibliography (Paciic Linguistics 517). Canberra: Australian National University. Mühlhäusler, Peter. 1996. Linguistic ecology: language change and linguistic imperialism in the Paciic region (Politics of language). London: Routledge. Pawley, Andrew. 1981. Melanesian diversity and Polynesian homogeneity: a uniied explanation for language. In Jim Hollyman & Andrew Pawley (eds.), Studies in Paciic languages and cultures in honour of Bruce Biggs, 259–310. Auckland: Linguistic Society of New Zealand. Pawley, Andrew. 1999. Chasing rainbows: implications of the rapid dispersal of Austronesian languages for subgrouping and reconstruction. In Elizabeth Zeitoun & Paul Jen-Kuei Li (eds.), Selected papers from the eighth international conference on Austronesian linguistics (Symposium Series of the Institute of Linguistics), 95–138. Academia Sinica. Pawley, Andrew. 2007. Why do Polynesian island groups have one language and Melanesian island groups have many? Patterns of interaction and diversiication in the Austronesian colonization of remote Oceania. Paper presented at “Migrations” conference, Porquerolles, France. Ross, Malcolm. 2001. Contact-induced change in Oceanic languages in North-West Melanesia. In Alexandra Aikhenvald & R. M. W. Dixon (eds.), Areal diffusion and genetic inheritance: problems in comparative linguistics, 134 –166. Oxford: Oxford University Press. Sahlins, Marshall D. 1963. Poor man, rich man. big man, chief: political types in Melanesia and Polynesia. Comparative studies in society and history 5(3). 285–303. 110 A. François Siegel, Jeff. 1998. Substrate Reinforcement and Dialectal Differences in Melanesian Pidgin. Journal of Sociolinguistics 2(3). 347–373. Thurston, William R. 1987. Processes of change in the languages of North-Western New Britain (Paciic Linguistics B-99). Canberra: Australian National University. Thurston, William R. 1989. How exoteric languages build a lexicon: esoterogeny in western New Britain. In Ray Harlow & Robin Hooper (eds.), VICAL 1: Oceanic languages, papers from the ifth international conference on Austronesian linguistics, part 2, 555–579. Auckland: Linguistic Society of New Zealand. Tryon, Darrell. 1976. New Hebrides languages: an internal classiication (Paciic Linguistics C-50). Canberra: Australian National University. Tryon, Darrell. 1996. Dialect chaining and the use of geographical space. In Joël Bonnemaison, Kirk Huffman, Christian Kaufmann & Darrell Tryon (eds.), Arts of Vanuatu, 170 –181. Bathurst: Crawford House Press. Tryon, Darrell & Jean-Michel Charpentier. 2004. Paciic pidgins and creoles: origins, growth and development. Berlin & New York: Mouton de Gruyter. VNSO [Vanuatu National Statistics Ofice]. 2000. The 1999 Vanuatu National population and housing census: Main report. Port Vila, Vanuatu. VNSO [Vanuatu National Statistics Ofice]. 2009. 2009 National census of population and housing: Basic Tables report. Port Vila, Vanuatu. http://www.paciicdisaster.net/pdnadmin/data/ original/VUT_2009_Census_ReportV1.pdf (accessed 19 February 2012). Vanuatu Ministry of Education. 2009. Provincial schools: Torba. http://www.governmentofvanuatu. gov.vu/government-ministries/ministry-of-education/provincial-schools/torba-provinceschools.html (accessed 12 October 2010). Vanuatu Ministry of Education. 2010. Education language policy: provincial consultation. Port Vila: Vanuatu Ministry of Education. Vienne, Bernard. 1984. Gens de Motlav. Idéologie et pratique sociale en Mélanésie (Publication de la Société des Océanistes 42). Paris: Société des Océanistes.