Abstract

Since Gagné and Shoben (1997), evidence has accumulated that humans use statistical semantic preferences associated with compound constituents in the interpretation of compound nouns. Most of this work has focused on the semantic relation between constituents, rather than interpretation of the constituents themselves. Yet due to the ubiquity of polysemy and homonymy, there are typically several readings to choose from for any given constituent, and it is not clear at the outset which one is to be used. This holds both for modifiers and heads, and for readings that are unrelated (plant factory/organism) as well as those that are metaphorically linked (gold material/colour). Using attested novel compounds, I show that diversity of interpretation is much greater than suggested by studies that focus only on semantic relations and that the ambiguity of the constituents, especially the head, plays an important role in determining both difficulty and diversity of interpretation. Current models of conceptual combination focus either on rather underspecified semantic relations or on ‘slot-filling’ in the semantics of the head; in either case, it is unclear how and at what stage disambiguation of the constituents takes place. The results presented here indicate that future incarnations of such models will need to address this challenge if they are to account for the interpretation not only of familiar but also of novel compounds.

Reference Gagné, C. L. and E. J. Shoben (1997). Influence of thematic relations on the comprehension of modifier–noun combinations. Journal of Experimental Psychology: Learning, Memory, and Cognition 23(1), 71–87.