Empirical problems with domain-based notions of "simple"

This chapter addresses the on-going debate about the relative "simplicity" of creole languages. It proposes that an evaluation of simplicity/complexity must consider not only categorical features of a language but also probabilistic ones, because (it argues) there is a good deal of linguistic structure encoded stochastically in creoles. To illustrate this, it explores four case studies: the marking of inalienable possession in Bislama (Vanuatu), subject agreement in Bislama, possessive marking in Tayo (New Caledonia), and the recent emergence of a new complementiser in Bislama. Substrate, lexifier and cognitive constraints contribute to the emerging shape of all four features. The data argues for perspectives on creolisation that include non-deterministic features, and for a view of language structure straddling what are sometimes seen as discrete levels of linguistic structure.