Peter Trudgill's account of new-dialect formation is uniformitarian, a position I have embraced explicitly since Mufwene 2001. In Mufwene 2006, I show how similar the mechanisms involved are to those that account for the emergence of creoles, the basic difference lying in the composition of the contact setting's feature pool (see also below). The position I defend is even less moderate, as I argue that in the history of humankind language speciation has basically been a consequence of how internal variation within a language has been affected by migrations of its speakers and additionally by the different contacts the relevant populations have had among themselves (Trudgill's position) and with speakers of other languages in their new, colonial ecologies (Mufwene 2005, 2007, 2008). The reader should not be surprised to see in this response comments that are primarily intended to support and complement the position of the target article.
Migrants typically disturb the traditional internal dynamics of variation within their language, causing some variants to become over- or under-represented in the host setting, compared to the metropole. In the kinds of colonial communities Trudgill discusses, the new feature pool often sets in competition variants which did not compete directly or regularly with each other in the motherland, but which are contributed by dialects spoken in different and sometimes non-adjacent metropolitan communities. The colonial feature pool thus provides new competition dynamics that usually result in new varieties, as is obvious from Trudgill 2004, for instance. In my brand of uniformitarianism, I argue that such colonial settings, in which only dialects of the same language came in contact with each other, differ from those associated with the emergence of, especially, creoles and indigenized varieties of European languages in the fact that the feature pools of the latter settings contain also elements contributed by other languages.
On the other hand, the kinds of colonial settings Trudgill has discussed are rare (e.g., the Falkland Islands), as one must factor in the contacts that the colonists have had with immigrants speaking other languages, as in Australia, South Africa, and white North America. Another important ecological difference between the colonial varieties that Trudgill focuses on and “indigenized” varieties in particular lies in the kinds of varieties the new speakers have targeted and in the mode of “transmission” – that is, vernaculars appropriated by the new speakers primarily through naturalistic language “acquisition” versus scholastic varieties appropriated primarily through books and the classroom. On this parameter, creoles also differ from indigenized varieties in being outcomes of naturalistic appropriation of typically nonstandard vernaculars.
contact is such a significant factor because it makes the composition of the feature pool more heterogeneous. As this heterogeneity varies from one geographical setting to another, contact can cause the same language to evolve differentially. Contact with other languages increases the likelihood of speciation, as is particularly evident from the geographical expansion of several western European languages since the 15th century. A number of factors guide the speakers' selections of features from the new feature pool. Where a language prevails over others, this is in itself a factor privileging a number of its features, because it is targeted by all other speakers, although it only wins a Pyrrhic victory, as is evident from structures of creoles and indigenized varieties. Most of these are legacies of the colonial languages under xenolectal influence rather than mere apports from the new speakers' former languages or mere innovations ex nihilo.
Where the populations in contact assume they speak the same language, there are tacit feature negotiations, “mutual accommodations” in Trudgill's discourse (see also Mufwene 2001), which drive the language toward a new norm. This is basically what Trudgill identifies as “dialect mixture,” a term which I hail over koiné(ization), as it can help us bridge the two research paradigms I have sketched so far. My main reason is that “koinéization” has unfortunately been ill defined as “dialect leveling” or “reduction to common denominator,” as if the different dialects simply got rid of all their differentiating features. If this were really the case, the “new dialects” Trudgill discusses here and in Trudgill 2004 would be too impoverished and reveal no or little variation. As Trudgill shows, just the opposite is the case. More adventurous minds like my own may want to explore speech continua in all the relevant colonial varieties, and not just where creoles are spoken.
Trudgill is correct in positing dialect contact and mutual accommodations among speakers as central factors that account for “dialect mixture.” This is apparently the process that in colonial settings produced varieties that differ from those spoken in the motherland. One peculiarity I find particularly useful in this case is the way Trudgill uses “colonial” and, by implication, “colony.” I see in this a conflation of the usage of the terms in population genetics, in which they allude to geographical relocation, with that in the common parlance, which also conjures up the political and economic domination of an indigenous population by the conquerors. This peculiarity of his discourse should lead readers to realize that, from a language evolution perspective, what matters the most is the fact that populations speaking different language varieties tacitly negotiate the linguistic features that enable them to communicate. Sometimes, the competition and selection process leads to the prevalence of one single language over others, although its structures are influenced (to varying extents) by those of the languages it has displaced. At other times, as in the cases discussed by Trudgill, the process leads to the emergence of a variety where it is more difficult to tell that there is a winner; and “dialect mixture” or “language mixture” just sounds like a more adequate characterization.
Language mixture can apply to Lingala, in central Africa, as it has been difficult to identify its lexifier unequivocally, owing particularly to the fact that the languages out of whose contact it emerged are genetically and typologically related. It can also apply to Old English, as it is not evident that the language varieties spoken by the Germanic colonizers of England were either mutually intelligible or dialects of the same language, probably no more than modern Danish, Dutch, and German are. On the other hand, the etymological association of English with the term Angles, the name of one of the colonizing populations, suggests a situation similar to that of creoles, where one language variety won a Pyrrhic victory. These are all interesting research avenues that one should not discount a priori. I would not call the process creolization, as, to my knowledge, there are no structural features nor any global language restructuring process that can be identified by the term. However, I maintain that research on the development of creoles, at least in the way that I approach the subject matter (with much intellectual debt to Chaudenson 2001), suggests that we not assume a priori that creoles have evolved in their own exceptional way (see also DeGraff 2003).
Trudgill is certainly correct is refuting the position that colonial identity drove the structural divergence of “new dialects” both from their metropolitan kin and from each other. This would be tantamount to claiming that evolution is goal-oriented and the colonists had really planned to be different linguistically. They could have done a better job in selecting a language variety that simply had no genetic connection to their heritage language. On the other hand, the invocation of the Founder Principle and of mutual accommodations to account for their emergence is only part of the story. The speakers' mutual accommodations are certainly the social aspect of the mechanisms by which selection from among the competing variants (and language varieties) proceeds. The Founder Principle also suggests that the composition of the feature pool is likely to vary at different stages of the development of a colony, as every wave of immigrants is likely to contribute its share of variants, native or xenolectal, to the colonial feature pool.
There must also be some weighting mechanism that determines which variants are to be preferred in the relevant contact setting. This is when some metric, in the form of ecology-specific markedness or ranking (depending on one's theoretical approach), becomes relevant. Frequency is just one of the factors that determine markedness values. Factors such as simplicity, perceptual salience, semantic transparency, regularity, and more familiarity to particular speakers, among a host of others (discussed in Trudgill 1986 and Mufwene 2001), bear on the selection process. Variation is typically not eliminated because the factors can conflict with each other and/or operate differently from one segment of the relevant population to another.
While accommodation leaves open the possibility of discussing the tacit negotiations among speakers within one time period (as a synchronic process), the Founder Principle imposes a diachronic dimension where one must also factor in the age of the learners. The agency of children in selecting some variants into and out of their idiolects is certainly very significant; it extends the accommodation process over generations. I am not here suggesting the kind of overruling power Bickerton 1981 has attributed to children in the development of creoles since his Roots of language, but rather the kind of position DeGraff 1999 hypothesizes. Some of the substrate features introduced into the colonial language variety by the adult xenolectal speakers would not be transmitted to future generations of learners if the younger learners/speakers (i.e., children) did not integrate them in their idiolects. Conversely, the children act as an effective filter, as the features they avoid are less likely to become entrenched in the new vernacular or available to future learners.
The restructuring equation in the evolution of a language is thus a complex one. There are certainly innovators, including those who introduce xenolectal features or initiate deviations from the current norm; but equally important are also the propagators, who certainly contribute significantly to the selection process, even if this amounts only to reducing the statistical significance of some variants within a population. While “talk[ing] like the others talk” is part of the factors that bring about change, one must remember that, by definition, mutual accommodations do not proceed unidirectionally. Speakers influence, and are in turn influenced by, different speakers in regard to different features. This multidirectional process makes more intriguing the question of how norms as population-wide convergences emerge. The “invisible hand” invoked by Keller 1994 bears as much on language evolution as in economic developments and other forms of cultural evolution.
I am grateful to Trudgill for demonstrating that “common identity” is not part of the complex of processes that produce “new dialects”; it is a consequence or byproduct of it. If identity has a role to play, it must be in resisting influence from outside one's community, which Labov 2001 illustrates with the way the Northern City Vowel Shift in the United States has proceeded, involving white but not African American speakers.
