Towards a Network-based Approach to French-Creole Diachrony

International audience The similarities between the French-based Creoles of the Americas (particularly Haiti and Louisiana) and those of the Indian Ocean (especially Mauritius), have led some, notably including Goodman (1964), Hull (1979) and McWhorter (2000), to claim that they descend from a common protolanguage originating in West Africa. Chaudenson (1974, 1981, 2001) rejects this notion, arguing that cross-regional similarities can be explained by parallel divergence from nonstandard forms of early modern French under similar conditions. However, he also argues that within the Indian Ocean, the direct transplantation of a contact variety known as Bourbonnais from Réunion to Mauritius during the initial settlement of the latter played a definitive role in shaping the Indian Ocean group. Yet others, such as Goodman (1992), Parkvall (1995, 1999), and especially Hull (1992) posit more nuanced scenarios involving partial diffusion of a transoceanic contact variety that incorporated influences from different colonies with passage of time. Here, we present digitally-compiled textual evidence to show that by the early 19th century, a transoceanic continuum existed, bounded by normative French on the one hand, and basilectal Creole varieties on the other. In the middle lie various forms of “Para-Creole”, ranging fromostensibly simplified French with few contact features, to genuine attempts by French speakers to reproduce contact varieties with varying levels of fidelity. While some of these varieties have been studied previously on an individual basis, such as the (in)famous Français Tirailleur (Skirgård, 2013), approaching them simultaneously allows one to identify several non-trivial parallels between Creoles and other contact-shaped Gallo-Romance varieties. Of particular note, we find evidence that well into the 19th century, features which now demarcate one language or region were at least marginally present in others. To begin, we first consider third-person plural pronouns and the pre-nominal plural marker ban(de) which Chaudenson considers key evidence for his Bourbonnais theory. We show that into the early 19th century, li, now a now-singular third-person pronoun, was used as third-person plural pronoun and an article in the Atlantic as well as the Indian Oceans, and that ban(de) is attested with expanded plural senses in Louisiana. Then, we consider the possibility of a relationship between the imperfective marker ka of the Lesser Antilles and the idiosyncratic Réunionese past imperfect construction (le)te (k)i. More specifically, we provide evidence from Walloon for a European origin of the latter (Lemacle, 1960; Saratxaga, 2012), a view which Chaudenson had posited on the basis of evidence from St. Barths. We further show that the former was perceived by French speakers as an overgeneralization of relative clauses, and reproduced as a past marker, a tendency which has parallels in Français Tirailleur. Overall, while such 19th (and 18th) century data does not completely explain the origin of each feature, it lends credence to the idea of feature-level transmission with modification through the networks of transoceanic institutions like the Catholic Church, the French East India Company and the military, as well as through popular culture.