Abstract - Roembke

Word learning in humans and pigeons – What role does pruning play?

Abstract: In this project, we are interested in understanding the mechanisms that underlie word learning. At the broadest level, we are investigating the possibility that word learning derives from basic associative mechanisms, mechanisms that may be shared with non-human animals. Therefore, this comparative project simultaneously examines how humans learn small sets of word/object mappings and how pigeons learn to map objects onto unique response keys in an analogue task. More specifically, we are investigating two possible learning components to understand which may be more important for successful word learning: 1) learning which objects are not the right referents of a word, i.e. pruning incorrect associations, or 2) learning to link a word to the correct referent (strengthening the correct associations). If pruning appears to be characteristic of learning in both species, it seems likely that it is an important and domain general aspect of learning. (Tanja Roembke, collaborating with Bob McMurray and Ed Wasserman).