Published online by Cambridge University Press: 03 June 2009
The present study investigates children's bias when interpreting novel noun–noun compounds (e.g. kig donka) that refer to combinations of novel objects (kig and donka). More specifically, it investigates children's understanding of modifier–head relations of the compounds and their preference for HAS or LOCATED relations (e.g. a donka that HAS a kig or a donka that is LOCATED near a kig) rather than a FOR relation (e.g. a donka that is used FOR kigs). In a forced-choice paradigm, two- and three-year-olds preferred interpretations with HAS/LOCATED relations, while five-year-olds and adults showed no preference for either interpretation. We discuss possible explanations for this preference and its relation to another word learning bias that is based on perceptual features of the referent objects, i.e. the shape bias. We argue that children initially focus on a perceptual stability rather than a pure conceptual stability when interpreting the meaning of nouns.
We express our gratitude to the children, parents and teachers of the following nurseries and primary schools who participated in our study: Busy Little Bees, Grace Mary Primary School, Little Angels Nursery, Little Robins Day Nursery, Russel Day Nursery, Small World, Coogee Nursery and St Margaret Mary Primary School. We also thank Sidra Aslam, Nicola Duggan and Elizabeth Ann Colledge for creating the stimuli and/or conducting the experiments and Sotaro Kita and the reviewers for comments on an earlier version of this manuscript. This research was supported by a British Academy Joint Activities Grant and a British Academy Small Grant (SG-44074), an Undergraduate Research Bursary from the Nuffield Foundation (URB/33283), a Vacation Scholarship from the Wellcome Trust (VS/06/BIR/A5), NSERC Discovery Grant 203054 and NSERC Discovery Grant 245058.