In this talk, I argue that (1) linguistic typology should tackle discourse functions more directly, and (2) linguistic typology should tackle minor, low-frequency, constructions more directly. I argue the first point by presenting work focusing on discourse structuring in the languages of the Oregon Coast (and a bit beyond), and surveying work by myself and by others on the expression of discourse functions, especially the main event-line of narratives. I use these to illustrate the richness of the morphosyntactic expression of discourse functions, present different axes of cross-linguistic variation in discourse domains, and show how we may tackle this richness to better understand cross-linguistic diversity.
In the second part of the talk, I argue that there may be many construction types that are considered cross-linguistically rare, but are actually cross-linguistically common. Specifically, they can be identified in many languages, but tend to have low frequency because of their function. Their low frequency in individual languages often leaves them out of published grammatical descriptions which leads to the inaccurate belief that they are cross-linguistically rare. I illustrate this by the so-called Possessive-Like Attributive Constructions (PLACs; also referred to as Dependency Reversal in Noun-Attribute constructions). In these constructions, modifiers expressing property-terms have isomorphic structural coding means to semantic heads modified by entity-terms (e.g., a beauty of a construction). I show that such constructions are attested in the languages of the Oregon Coast as well as in languages from many other parts of the world. I suggest that the function of these constructions has to do with the discourse status of their semantic head and propose several axes of typological variation in PLACs based on patterns of their actual usage and the systems of modification constructions they are a part of.