16. Animacy in Bislama? Using quantitative methods to evaluate transfer of a substrate feature
The source of and, hence, principal factors constraining, several variables in Bislama, an English-lexified Pacific creole, remain the subject of some dispute. This chapter uses quantitative methods to evaluate the strength of claims that variable presence/absence of arguments in Bislama is principally due to the transfer of preferences in the substrate languages. It focuses particularly on the role that the <i>animacy </i>of a referent plays in determining:a. presence/absence of pronominal subject in a clause,<br />b. the form of 3p agreement, and<br />c. the presence/absence of pronominal objects (Crowley 1990, 2002).<br />Other research has claimed discourse salience (not a substrate feature) and/or direct possession relations (substrate feature) are more relevant (Meyerhoff 2000, 2003a). Clauses of spontaneous conversational Bislama recorded on Malo island in the 1990s <i>and </i>a corpus of 10 Tamambo (the Malo vernacular, Jauncey 1997) narratives or process texts are analysed for the same factors. The results show that animacy is a significant constraint on the subject pronominal variable, but it is not strong for the other two variables. The result is an empirical gain and a theoretical gain. First, claims for transfer of substrate features into Bislama are motivated in a more transparent way than they have been before. Second, we see clearly the potential that multivariate analysis offers for resolving outstanding questions and debates relating to language contact and the role of substrate transfer. This is especially true for the Pacific creoles where we continue to be able to gather, and analyse, substrate corpora.