The quest for non-European creoles

Abstract Kukama has been classified firstly, as a Tupi-Guarani language, secondly, as a language that has undergone massive contact-induced change, and thirdly, as a creole. These different positions are surveyed against definitions of mixed languages and creoles, and properties of Tupi-Guarani languages. We focus especially on its possible status as a creole language, as there is a need among creolists to expand the range of creoles with a broad set of lexifiers and substrates, especially after recent research has revealed a typological profile for creoles. With typological, historical and historical-comparative arguments, it is concluded that the original grammatical system of Kukama was strongly reduced, and subsequently innovated and expanded with mostly Tupi-Guarani roots, a process parallel to that gone through by accepted creoles. As is commonly the case in creoles, very little beyond the lexicon is inherited from its lexifier, a language close to Tupinamba. However, the Kukama language does not quite fit two proposals for creole prototypes, perhaps because the processes of loss and innovation took place in a much more remote past than for identified creoles. In this paper, it will be discussed whether Kukama is a language that defies classification within the current typology of contact languages, or whether it is deviant within current family classifications.