Creole studies: Phylogenetic approaches ed. by Peter Bakker et al. (review)

Creole studies has a well-deserved reputation, well known even to linguists with no special interest in the area, of being highly contentious. Many are aware that the main disagreement in our time is no longer over whether universals (Bickerton 1981) or substrate contributions (Holm 1988) were most important in creole genesis—a major debate in the 1980s into the 1990s—but over whether creoles constitute a type of language at all.

Influential publications by Mufwene (2001, 2008), DeGraff (2001a), and more recently Aboh (2015) argue that creolization is simply a kind of language hybridization, qualitatively indistinct from that which produced Romanian or Yiddish, especially considering that second language acquisition, while clearly important in creole genesis, is hardly irrelevant to the histories of perhaps most languages. The idea is that the classification of creoles as a kind of language, born from a radical break in transmission such as pidginization, has been an impressionistic reification, informed by racism and colonialism, and unmoored from the procedures of modern science.

The work chronicled in Creole studies: Phylogenetic approaches serves as a counterpoint to this tempting reconception of creole studies and language contact. While other studies have attempted to demonstrate that creoles are unique in combining certain features (McWhorter 2011, 2018, Daval-Markussen 2013), this book evidences a computational phylogenetic approach to the question, in line with an approach increasingly de rigueur in linguistic work overall. Presenting the work of a project overseen by Peter Bakker at Aarhus University, this volume now stands as a key statement in the debate over the nature of creole languages. Claims that creoles are not a kind of language will henceforth qualify as incomplete without engagement with the works in this book.

The phylogenetic approach tabulates a set of features in a range of languages and then represents the data in diagrams generated by programs such as SplitsTree (Huson & Bryant 2006), graphically revealing clusterings between languages. According to the commonly asserted idea that creoles are not a kind of language, creole languages should not cluster together in diagrams like these. This computational approach, presumably, should show creoles distributed randomly amid other languages, with the folly in the supposed essentialization of creoles duly and conclusively revealed.

But this is not the case, and this volume collects a now imposing amount of evidence in illustration. After three introductory chapters summarizing the general project outlined in the book, creole genesis theories past and present, and the phylogenetic approach, Ch. 4 and Ch. 6 neatly communicate, through various experiments published previously, the weight of evidence in support of the claim that creoles do constitute a structural type of language. Bakker, Daval-Markussen, Parkvall, and Plag (2011) began with the data in Holm & Patrick 2007, comparing ninety-seven features from eighteen creoles. They winnowed that set of features down to sixty-nine, eliminating those not shared by two thirds of the creoles. Phylogenetic analysis shows that in these features, regardless of their lexical base, creoles cluster quite clearly compared to other languages—that is, that there are indeed features typical of creole languages (123). It is reasonable to wonder whether the 'creole' features in question are simply those of analytic languages in general. However, in Ch. 4, Peter Bakker, Eeva Sippola, and Finn Borchsenius find the same result when creoles are compared only with analytic languages (72). Parkvall (2008) compared forty-three complexity-gradient structures from The world atlas of language structures (Haspelmath et al. 2005) in order to avoid possible bias in proceeding from a source like Holm & Patrick 2007, which is focused on features common in creoles. Parkvall's study compared, on the basis of these features, 188 creoles and older languages, and the creoles clustered similarly. This supports the claim that creoles display less grammatical complexity than older languages, presumably due to their ancestry in broken transmission and pidgins (123).

Further chapters strengthen the...