other in their speech habits. While still maintaining some idiosyncrasies of their own, they achieve what Le Page and Tabouret-Keller (1985) call
“focusing.” Creoles have developed both from individual speakers’ attempts to speak the lexifier and through their mutual accommodations in the contact settings.7For convenience, we simplify this more complex picture of contact by focusing on the communal language level. In reality, however, the primary actions identified here through the notions of competition and selection take place in the minds of individual speakers. Recognizing the reality described in this section enables us to account for variation within the community.
2.2 The development of creoles: what the histories of individual colonies
has generally been downplayed since the 1970s. The legacy of Krapp (1924), Kurath (1928), Johnson (1930), Faine (1937), Hall (1966), and Valkhoff (1966), among others, has been barely noticeable in the context of Atlantic creoles. Exceptions include D’Eloia (1973) and Schneider (1989) regarding African-American vernacular English (AAVE), and Chaudenson (1979, 1989, 1992) in the case of mostly Indian Ocean French creoles.
Overall, superstratists claim that creoles have typically extrapolated, through normal adaptive processes, structural alternatives that were already present in metropolitan and/or colonial varieties of the lexifiers.
The new vernaculars did not innovate much in the sense ofex nihiloor UG- based creations advocated by Bickerton (1984, 1992). Nor did they accept much substrate influence that did not have some model, however partial or statistically limited, in their lexifiers. However, Chaudenson (1989, 1992) and Corne (1999) recognize the role of substrate influence, without which creole vernaculars would be more similar to (ex-)colonial varieties of European languages spoken by descendants of Europeans, such as Québécois and Cajun French. Chaudenson takes creoles’ lexifiers to have been approximations by slaves of colonial European speech of the late homestead phase, i.e., varieties that were not significantly restructured compared to European colonial speech. According to Chaudenson, during the homestead phase, identified as société d’habitation, all those born in the colonies spoke the same colonial varieties of European languages, regard- less of race, because they all lived in the same integrated settings.9(This is an important point against arguments that creoles developed due partly to a break in the transmission of the lexifier. Transmitters of the lexifier need not have been European.)
Since the mid-1980s I have contributed to the “complementary hypothe- sis.”10I argue that the only influences in competition are structures of the lexifier and of the substrate languages. The language bioprogram, which need not be understood to be operating exclusively in children, regulates the selection of structural features from among the options in competition among the language varieties in contact. It can incorporate my ecology- sensitive markedness model, according to which the values of the compet- ing alternatives are determined by diverse factors, such as regularity or invariance of form, frequency, generality, semantic transparency, and per- ceptual salience, among apparently a host of other factors that future research is likely to uncover.
Along with Le Page and Tabouret-Keller (1985) and Chaudenson (1992), I criticize a shortcoming common in most studies, viz., the comparison of creoles’ structural features with those of the standard varieties of their lex- ifiers rather than of their nonstandard varieties. The illusion that the Europeans with whom the non-Europeans interacted on the plantations
spoke the standard varieties of their lexifiers is not consistent with sociohis- torical information available about the beginnings of the colonies. Some seventeenth-century letters addressed to, for instance, the West Indian Company, the Virginia Company, the [Dutch] West India Company, or their other European counterparts reveal that their authors were typically low-ranking employees who had been sent on difficult ground-breaking missions in the colonies. They corroborate studies of Ship English, such as Bailey and Ross (1988:196–7), who argue that “most of the sailors were illiterate, including many of the captains and masters.” According to this literature, the varieties spoken aboard the ships must have been more non- standard than the ships’ logs indicate, especially as the written medium may have skewed the samples in the direction of the standard variety.
These observations are consistent with historical accounts according to which large proportions of the immigrant European populations consisted of defector soldiers and sailors, destitute farmers, indentured laborers, and sometimes convicts. That is, the vast majority of the (early) colonists came from the lower strata of European societies. As much of their correspon- dence indicates (e.g., Eliason 1956), they spoke nonstandard varieties, inherited by the vernaculars of rural and low-income Whites in such loca- tions as the Piedmont, Appalachian and Ozark Mountains in the USA.
The same may be said of French varieties spoken on the Caribbean islands of, e.g., St. Barths and St. Thomas.
It has often been argued against the non-relexificationist version of the substrate hypothesis (represented chiefly by Alleyne and Holm) that the Africans could not influence the structures of the emerging creoles because of their extensively diverse linguistic backgrounds. However, as noted above, the European lexifiers themselves were typically heterogeneous.
Several of their diverse dialects were brought into contact with one another in the colonies. As shown in chapter 3 regarding North America, the varie- ties spoken today in such former colonies are also outcomes of restructur- ing under contact conditions, regardless of the extent of their system-reorganizations. Contact-induced restructuring explains why none of the extra-European dialects is an exact match of any metropolitan variety. As shown in chapter 1,koinéization– if such a name is necessary at all – is the same kind of restructuring that produced creoles by competition and selection, except regarding whether the varieties in contact were geneti- cally and/or typologically related, and whether the languages in contact were all European. (The latter is a shameful tacit assumption of creolistics that is hardly ever stated.) Investigating the development of (ex-)colonial dialects of European languages ecologically should be informative about the nature of feature competition and the factors which regulated specific selections in their respective settings. In relation to creoles, factoring in the
2.2 Development of creoles 35
relevant historical, socioeconomic information about the contact settings and speakers of these colonial varieties should also tell us a great deal about the origins of the typically nonstandard options which found their way into creoles’ structures, even if only in modified forms.
2.2.2 The koiné lexifiers of creoles
A factor that has often been overlooked regarding both the European and non-European elements in the new vernaculars is the demographic signifi- cance of diverse ethnolinguistic groups and how it varied from one period to another during the development of these communities. This factor greatly complicates the language contact formula regarding when a partic- ular language variety was likely or unlikely to influence the development of a new vernacular. I return to this in section 2.3.
Hancock (1969) and Dillard (1972, 1985), among others, emphasize the contribution of an antecedent maritime, or nautical, English jargon to the development of the new colonial varieties. Like Le Page (1960) and Le Page and Tabouret-Keller (1985), Dillard (1985, 1992) invokes the high propor- tion of nautical terms to support this position. On the other hand, Buccini (1995) argues that the making of colonial varieties of European languages may have started in Europe. He presents port cities such as Amsterdam and Utrecht as contact settings where speakers of diverse Dutch dialects met before they sailed for the colonies. New varieties putatively developed there, triggering “leveling” processes (reinterpreted in chapter 1 as restructuring) which would continue up to the time of the colonies. It is not yet clear what role such diachronic processes, which must have taken place in other metro- politan port cities too, played in the development of nautical varieties.
It is plausible that these then-emergent port-city koinés and nautical varieties influenced the vernaculars that developed in the colonies. It is from the port cities, where they stayed for a while, that immigrants left for the col- onies; and they traveled by sea. It is just not clear whether the new metro- politan koinés had already normalized before the colonists emigrated, nor is it evident whether they were dominant among the varieties which were brought from Europe. What Buccini observes may simply boil down to this:
precolonial population movements and contacts in Europe itself must have initiated language restructuring, and forms of these restructured varieties were brought over to the colonies, where the process would continue under new ecological conditions.
The above summarizes what is also suggested by the socioeconomic history of colonization, especially as described by Bailyn (1986) and Fischer (1989) in the case of the USA. They show how British emigrations to North America were basically extensions of population movements that
were taking place in the British Isles, originally in the direction of industrial Southern English cities. These were settings of population and dialect contact, where much of the restructuring that would take place in the colo- nies began. This process is well evidenced by the changes undergone by English in the UK since the seventeenth century.
All the above observations show that there was independent ground for feature competition and selection among the European colonists them- selves, as proved by, for instance, Québécois French and North American White English vernaculars. Regarding the development of creoles, the competition-and-selection situation was made more complex by the pres- ence not only of Africans on plantations but also of other Europeans who did not speak the lexifier natively. In the case of the Chesapeake, numerous non-English-speaking Europeans, especially Germans, counted among the early indentured servants (Kulikoff1991a, 1991b; Menard 1991). Based on Dyde (1993) and Beckles (1990), regarding St. Kitts and Barbados, respec- tively, one is led to speculate that the practice of recruiting labor from places other than the metropole of a colony was a common practice.
Moreover, a closer examination of the linguistic histories of the different European metropoles reveals that not all their citizens or subjects spoke natively the languages which lexified the different creoles.
For instance, as pointed out in chapter 1, rural Ireland was just beginning to Anglicize when England engaged in the colonization of the Caribbean and North America. Chaudenson (1992) cites Father Labbat’s observation that many French citizens (identified in French as “les patoisants”) could be learning French from African slaves of the homestead phase in their Caribbean colonies. The observation suggests that indeed several inden- tured servants in French colonies were not (fluent) speakers of French.
Some of the varieties then identified by Francophone French as “patois”
were probably Celtic languages, just like Breton. Kibbee’s (1999) “French language policy timeline” shows that in places such as the Midi and Brittany, French did not become a common vernacular until the twentieth century. Thus, in the colonies, much of the same kind of restructuring of the lexifier that took place among the Africans was also taking place among European colonists. An important difference regarding the restructuring of the lexifier among the Europeans and among non-Europeans is the follow- ing: after segregation was institutionalized, selective substrate influence must have applied along racial lines, except for common features that had been selected earlier and were deeply entrenched. That is, with the Africans then interacting more among themselves than with non-Africans, some lin- guistic habits were likely to develop among them that were not attested, nor significant, among other speakers, and vice versa. The nature and extent of such influence are yet to be determined.
2.2 Development of creoles 37
Taking into account the following observations on how the colonies developed from homestead to plantation societies, the position that creoles’
lexifiers were not metropolitan varieties is plausible. There must have been some already locally restructured varieties, and/or something in the process of development, spoken by the founder populations, including Europeans and non-Europeans, of the homestead phase (Chaudenson 1979, 1989, 1992; Le Page and Tabouret-Keller 1985:26).
2.2.3 Normal, uninterrupted language transmission and the development of creoles
Most genetic creolists have taken it for granted that New World and the Indian Ocean creoles developed “abruptly,” within a human generation, after an initial phase during which a pidgin was spoken by adult non- Europeans. The position is also related to the stipulation since Polomé (1983), repeated by several other creolists, that there was a break in the transmission of the lexifier.
We must set aside some incorrect assumptions, for instance, that the lex- ifiers of creoles were monolithic and standard varieties. Then we can realize that the socioeconomic histories of the territories where these vernaculars developed do not support the pidgin-to-creole evolutionary scenario rejected in chapter 1. Contacts and communication in the homesteads were regular, not sporadic. To be sure, there must have been a lot of frustration arising from unsuccessful attempts to communicate in the beginnings.
However, nothing in the ecologies of human contacts during the homestead phases of the development of the colonies suggests reliance on stable pidgins for the purposes of communication. Rather, speakers must have proceeded through normal adult interlanguage stages of second-language acquisition in naturalistic settings toward closer approximations of the lex- ifier.
Another incorrect assumption emerging from the literature is that colo- nial plantations developed overnight, so to speak, with all their peak popu- lation aggregates in place apparently since the foundation of the colonies, and with all the relevant non-European languages represented with their features competing concurrently. Instead, several important historical sources examined by, e.g., Baker (1990, 1993), Singler (1993, 1995), Migge (1993), and Mufwene (1992b, 1997c) confirm Chaudenson’s position that the colonies developed gradually into plantation economic systems, although not necessarily at the same speed (see below). They suggest that creoles developed by continual restructuring of their lexifiers, subject to varying ecologies.
Baker (1996) observes correctly that the homestead phase lasted less time
in Mauritius than in Réunion. Mauritian Creole would thus have devel- oped quite early in the history of Mauritius and in a setting where, given the much smaller proportion of a founder population speaking colonial French, features that were congruent with those of non-European lan- guages were favored. Hence the emergent creole diverged more significantly from the structures of its lexifier than Réunionnais.
The same can also be inferred from Le Page and Tabouret-Keller’s (1985) history of Caribbean colonies and the development of their creoles (see also Williams 1985). For instance, Jamaica, which was colonized by the English almost thirty years later than Barbados (1655 as opposed to 1627), adopted the sugar cane plantation system about twenty years faster than Barbados. Its non-European population grew even faster. By 1690, thirty- five years after the English took Jamaica from the Spaniards (1655), the African population in Jamaica had grown to three times that of the European population: 30,000 vs. 10,000 (Williams 1985:31). The same year, Barbados had 50,000 Africans against 18,000 Europeans, at a ratio of less than three to one in sixty-three years (Williams 1985:31). By the mid- eighteenth century the proportion was over ten Africans to one European in Jamaica, whereas it did not exceed two to one in Barbados, even despite the continuous dwindling of the European population. Note also that in Barbados, the African population remained a minority for the first thirty years, whereas in Jamaica it surpassed the European population within the first twenty years. (More on demographic developments below.) Although Barbados diddevelop a basilect comparable to those of other Caribbean territories by the nineteenth century (Rickford and Handler 1994; Fields 1995) – probably later than Jamaica – it seems to have done so on a smaller scale, which accounts for its disappearance.
These historical observations also reveal that Africans from different regions and language families – often coinciding with different typological groups – became critical to different stages of the development of Atlantic creoles. Curtin’s (1969) general demographic estimates in Table 1 (subject to conventional reservations on his figures) show that the proportion of Africans from the Windward Coast (speaking Mande, Kru, and Western Kwa languages) was significant mostly during the homestead phases of some colonies, e.g., South Carolina and Jamaica. However, it became sig- nificant in some other colonies, e.g., Barbados and Suriname, mostly during their plantation phases. Africans from the Gold Coast and the Bight of Benin (also speaking Kwa languages) became demographically significant during the early eighteenth century, when the basilectalization of several Atlantic creoles was underway. Features of the lexifiers which were Kwa- like were likely to gain selective advantage, barring other factors which may have influenced the restructuring differently (Mufwene 1989c, 1991a).
2.2 Development of creoles 39
By the time the Central Africans (speaking Bantu languages) became demographically significant, during the second half of the eighteenth century, most of the creoles must have already developed the greatest and/or more fundamental parts of their structures and norms. By the Founder Principle, since such demographic significance did not obtain overnight, it was generally more cost-effective for subsequent generations of immigrants (free, enslaved, and indentured) to learn the emerging local vernaculars rather than to develop new ones from scratch. Thus a great deal of features which became associated with the creoles in those early stages became more and more entrenched through adoption by newcomers who targeted the current speakers (creole and seasoned slaves) and would later serve as models for future newcomers. The Founder Principle thus also
WEST ATLANTIC MANDE
KRU KWA
VOLTAIC CHADIC UBANGIAN
S. CENTRAL SUDANIC
BANTU
B A N T U
KHOISAN GoldCoast
Slave Coast GuineaCoast Guinea Coast Se
neg ambia Grain
Coast
Cong
o- Angola
Map 1 Africa: some historical regions and major language groups
Table 1The English slave trade,1680–1800,by African region oforigin,expressed in percentages ofvarying samples, from Philip Curtin (1969:129) SierraWindwardGoldBight ofBight ofCentral PeriodSenegambiaLeoneCoastCoastBeninBiafraAfricaOther 1.1680–8512.027.320.915.76.712.05.4 2.168812.038.018.412.35.211.32.8 3.171314.64.210.431.239.6——— 4.17246.410.65.338.321.33.214.9— 5.17527.032.05.240.412.72.6 6.17717.025.416.049.52.1— 7.17716.72.031.013.13.044.2—— 8.17880.94.75.413.516.829.029.7— 9.17986.26.83.038.245.8— 10.17990.39.89.71.044.834.4— Sources:Data from Davies,Royal African Company,pp.225,233,363;Le Page,“Jamaican Creole,”pp.61–65;Donnan,Documents,2:308–9, 454–56,598;Edwards,British West Indies,2:56.See also text,pp.130–32.
favored continual transmission of the gradually basilectalizing local ver- nacular over the often-alleged break in its transmission. Both sociohistori- cal facts and the evolutionary account proposed in this book make it unnecessary to posit an antecedent pidgin to the creoles which developed in European plantation colonies of the Atlantic and Indian Oceans.
We must remember that the mortality rate was very high among the plan- tation laborers, although this population continued to increase until the nineteenth century. This trend of rapid population replacementand growth (as compared to communities growing by birth) also favored continual restructuring of the extant vernaculars among non-Europeans, especially after segregation was institutionalized on the large plantations and the non-Europeans had less and less fresh input from (colonial) European varieties. Even though the most drastic restructuring may have taken place during the initial and critical transition to the plantation phase in every colony, the basilectalization process probably continued up to the abolition of slavery in the nineteenth century or up to the total collapse of the planta- tion industry, which entailed the end of the importation of indentured labor sometimes from the same places which had formerly supplied slaves.
Having proved adaptive several times before (by Wimsatt’s 1999a Generative Entrenchment principle), the features selected in earlier phases of plantation development stood a good chance of being selected for one reason or another during every round of the competition. However, there was always room for new selections to be substituted for, or added as alter- nates to, some older ones.
This is precisely the kind of evolutionary scenario that Le Page (1960:74–5) and Le Page and Tabouret-Keller (1985:47) suggest for Jamaican Creole. They divide the development of the colony in the eight- eenth century into two major periods: the first half, marked by the preva- lence of Kwa-speaking populations, especially Twi and Ewe; and the second half, marked by a significant Bantu presence. If we ignore the fact that Bantu morphosyntax is not exclusively agglutinating (Mufwene 1994b), the early Kwa prevalence would suffice to account for the selective advantage gained by the Kwa-like morphosyntactic features, most of them being (partially) congruent with patterns of some varieties of the lexifier.
These features include the invariance of the verb regarding person and number, the periphrastic marking of tense and nominal , the intro- duction of relative clauses with a complementizer (including ) rather than with a relative pronoun, and the significant presence of serial verb con- structions.
The few exceptions to the putative selective advantage of Kwa-like fea- tures include Palenquero (Maurer 1987), São Tomense, and Principense (Ferraz 1979), whose initial creators included significant proportions of