ABSTRACT
To interpret a noun-noun compound such as “chocolate rabbit” people must find a relation to link the modifier and head noun of the compound (the first and second words). How do people find these relations? Several theorists have proposed a taxonomic approach to noun-noun compounding (e.g. Levi, 1978; Gagné & Shoben, 1997; Gagné, 2001). This approach suggests that there is a fixed, relatively small taxonomy of standard relations such as X uses Y and X made of Y, from which people select the most suitable relation when interpreting a noun-noun compound. With this approach to nominal compounds, the same standard relation will necessarily be used in a great many compounds, for example “chocolate rabbit” “glass table” and “computer network” all use the standard X made of Y relation. For these examples however, it is clear that relatively different specific relations are being used in interpreting the compounds. The taxonomic approach assumes that all these relations are fundamentally the same because they are all examples of the same standard relation.
