Abstract for Kristen Syrett's talk

Recruiting form-meaning correspondences in adjective learning
Kristen Syrett (Rutgers University)
April 8, 2008 (Tuesday)
6:30 PM - ; Room 7102, The CUNY Graduate Center

There is by now considerable evidence that preschool-age children are aware of the role that the context plays in setting and shifting the standard of comparison for gradable adjectives, such as 'big' and 'tall'. More recently, we have shown that they also recognize that there are limits to the role the context plays for other gradable adjectives, such as 'bumpy' and 'straight', whose standard is instead determined by a minimal or maximal presence of a property. Appealing to differences in scalar structure allows us to capture these differences neatly: the former have a context-dependent standard, whereas the latter have an endpoint-oriented standard. Adverbs highlight such scalar differences: standard-raising modifiers such as 'very' reference a contextually-based standard, whereas proportional modifiers such as 'completely' select for adjectives with maximally closed scales. Two questions naturally arise for language acquisition: are these cues reliably present in the input, and do children recruit them in word learning? In this talk, I will present evidence from a corpus analysis and a set of experiments using the intermodal preferential looking paradigm that, in tandem, answer these questions affirmatively. I will discuss the implications these results have for infants' expectations about scalar structure and briefly touch upon current research extending this work into the domain of quantifier and number word learning.