For a Discourse-Sensitive Typology: Theoretical and methodological aspects

The aims of this paper are twofold. First, I aim to discuss some theoretical and practical issues concerning the methods employed by linguists when doing typological research. More specifically, I will present the kinds of data typologists usually base their work on, also discussing some related issues that deal with the nature of the data themselves and the possible misinterpretations that they may lead to. Then, I will discuss how typology tends to focus its attention over a limited array of phenomena (i.e., morphosyntactic ones), leaving aside phenomena pertaining to other levels of analysis (e.g., discourse). I thus will exemplify how integrating discourse phenomena might result in more comprehensive analyses and, consequently, in better-suited typological generalizations. Second, I will propose a multi-level method (which I call discourse-sensitive typology) based on the converging evidence perspective, discussing how it might help us to overcome the abovementioned issues and ultimately make typology truly usage-based.