Seselwa Serialization and its Significance

Abstract

Hitherto it has been believed that serial verb constructions were not found in the French Creoles of the Indian Ocean. In fact, there exists a wide variety of such constructions in both Morisyen and Seselwa; extensive examples from Seselwa are provided. However, for less basilectal speakers, such constructions must carry on each verb a tense marker which is identical to that on the matrix verb. This fact, together with the misanalysis of Seselwa i as a reprise pronoun, served to disguise the nature of Seselwa serials. The data surveyed has significant consequences for future studies of creoles and for theories of creole origins.

Serial Verb Constructions in Indian Ocean French Creoles (IOCs)

“…the French creoles of the Indian Ocean, Seychellois (SC) and Mauritian Creole (MC) in particular) have serial verb constructions (SVCs). Bickerton (1989) identified a set of constructions, particularly in SC, with properties characteristic of serialization, and claimed that SVCs do indeed exist in this language, contrary to what had previously been assumed (see, for instance Jansen, Koopman, & Muysken 1978). 1 He also argued that their presence in SC and MC lends support to his hypothesis that SVCs in creoles are the result of creolization and the Language Bioprogram Hypothesis (Bickerton 1989(Bickerton , 1990(Bickerton , 1994(Bickerton , 1996 on the grounds that the substrates of MC and SC were non-serializing Bantu languages and could not therefore have been the source of their SVCs.…”

“…Bickerton (1989) identified a set of constructions, particularly in SC, with properties characteristic of serialization, and claimed that SVCs do indeed exist in this language, contrary to what had previously been assumed (see, for instance Jansen, Koopman, & Muysken 1978). 1 He also argued that their presence in SC and MC lends support to his hypothesis that SVCs in creoles are the result of creolization and the Language Bioprogram Hypothesis (Bickerton 1989(Bickerton , 1990(Bickerton , 1994(Bickerton , 1996 on the grounds that the substrates of MC and SC were non-serializing Bantu languages and could not therefore have been the source of their SVCs. However, both Seuren (1990Seuren ( , 1991Seuren ( , 1995 and Corne et al (1996) have disputed Bickerton's claim and argued that the constructions he identifies as SVCs are in fact covert (asyndetic) coordinate constructions.…”

“…It argues on the basis of relevant data, particularly from MC, that there is in fact evidence of serialization in the IOCs. 2 However, unlike Bickerton (1984Bickerton ( , 1989Bickerton ( , 1996, it attributes the existence of SVCs in these languages not to creolization of an impoverished input by children, as claimed by Bickerton, but instead to an independent internal development in which consecutive imperatives in the input were reanalysed as SVCs.…”

“…The proposal has implications for the development of SVCs in other creoles, particularly those with serialization substrates, for it raises the question as to whether their presence in these creoles could also be seen as a consequence of an independent development, with possible convergence with one or more substrates, rather than direct transfer (as claimed by Lefebvre 1998 andMcWhorter 1992, amongst others) or creolization and the language bioprogram (as claimed by Bickerton 1984Bickerton , 1989Bickerton , 1996. Importantly, it also suggests a more substantial role for input and adults in the development of serialization than does either the substrate hypothesis or the universalist hypothesis.…”

“…For a more recent source, see Muysken & Veenstra (1995: 291), although it is conceded there that SC has serial constructions but of the 'TAKE' type. But Bickerton (1989) lists other types of SVCs in SC.…”