부산대학교 도서관

This paper surveys applicative constructions in the northwest Amazon and adjacent Andean slopes. Previous research suggests that this area may feature contactinduced diffusion of valency markers, specifically applicatives, across family boundaries. On closer scrutiny, however, there seems to be no firm basis for this conclusion. First of all, applicatives do not seem to be overly common in the area: many languages do not have applicatives, and those that do often have no more than one (with a few notable exceptions). Second, although some commonalities can be observed between the applicative constructions across the area, they involve common features of applicative constructions anywhere, like suffixed applicative markers, and a preponderance of benefactive or sometimes malefactive semantics. To a lesser extent, comitative-related semantics are found. There is some overlap in form, but these seem to be largely coincidental. A possible exception is a connection between Arabela (Zaparoan) and Yagua (Peba-Yaguan), which do show signs of non-accidental similarities. In addition, there are some intriguing but inconclusive functional similarities between Shiwilu (Kawapanan) and some of the Arawakan languages of central Peru, outside the northwest Amazon.