A modular approach to evidentiality

Evidentiality is a well-established morphosyntactic category that has also received a lot of attention in the semantics literature over the last 15+ years. However, it has received scant attention in Lexical-Functional Grammar, despite the fact that LFG’s modular Correspondence Architecture is particularly well-suited to illuminating the phenomenon. In particular, the theory makes possible an account of evidentiality that does not merely conflate the semantic category of evidentiality with its morphosyntactic realization, but which also does not create false equivalences between languages that mark evidentiality morphosyntactically and those that do not. In other words, it enables an account in which we can differentiate languages that morphosyntactically mark evidentiality from those that do not. Yet at the same time it also allows us to capture semantic commonalities between morphosyntactically marked evidentials and expression of evidentiality in languages, like English, that do demonstrate semantic evidentiality, but which do not have a dedicated morphosyntactic paradigm of evidentials. In this paper, we will first consider languages with obligatory, fully grammaticalized evidentiality. We then turn to English as an example of a languages that has the means to express evidentiality, but without fully grammaticalized or obligatory marking