From the recent literature on evidentiality it is emerging that evidentials in general and Reportatives in particular do not form a uniform semantic class (McCready (2008), Matthewson (2011)).The goal of this talk is to describe the meaning and uses of the Arrernte ‘quotative’ akwele, and thereby contribute to a typology of Reportatives. While akwele behaves in similar ways to known Reportatives, it also exhibits several interesting differences, notably in that it can be used (i) in imperatives, and with certain deontic interjection particles, and (ii) to attribute a belief to someone else based on other forms of behavior than verbal report – thus pointing out to a more general kind of evidential/modal marker than a plain Reportative.

We will essentially (i) offer a classification of the different uses of akwele and (ii) propose a tentative semantic analysis accounting for those various contextual uses. Our account will depart from existing analyses (Wilkins 1986, 1989, Author) in that it views akwele as a marker of apparent (evidential) mental state – in effect a ‘super-category’ covering genuine apparent beliefs (epistemic states), apparent intentions (deontic states) and shared/conventional knowledge (cognitive states). Pragmatic enrichment inferences will then trigger the various contextual interpretations observed.