TODAY

Special Seminar:

Tuesday, June 30, 1:00pm in 46-3310

Exact units of quantification in language acquisition
David Barner, UCSD

For adults, expressions like "many pies" or "three pies" are evaluated by first identifying the pies in the context, and then determining whether the words "many" and "three" are true of these individuals, as a set. However, when a novel word is used - e.g., "many blickets" / "three blickets" - the adult must infer what counts as a single unit of blicket. Only then can it be determined whether there are many, three, or none. The first set of studies that I describe investigates how children solve this problem, given their small vocabularies in early acquisition. How can quantifiers and number words emerge in speech at around 2 years of age, when children typically know only a few hundred words?

A second problem is that contexts which contain "many pies" often also contain "some pies" "most pies" and even "all pies". Similarly, when there are 6 pies, there are also, 5, 4, 3, 2, and 1 pies. The second part of the talk will explore how children eliminate these alternative interpretations of quantifiers and number words, by starting with "weak" lower bounded number word meanings, and strengthening them by negating stronger alternatives. I argue that previous studies vastly underestimate children's ability to use inference to strengthen word meanings in acquisition. As a result, these studies may ascribe overly rich innate mechanisms to account for integer acquisition.

Kathleen V. Dickey

Development Coordinator

Massachusetts Institute of Technology

Dept. of Brain & Cognitive Sciences
phone 617.324.5399
fax     617.258.9216
kvdickey@...