December 7, 16-17:30 in E20.12

Irene Krämer, Nijmegen

Many, most and all: Children's interpretation of quantified NPs.

Abstract: This study presents data on 4 to 7-year old children's interpretation of weakly quantified Many N and compares children's interpretation of Many to Most and, in particular, All. Experimental results will be presented that show that children's interpretations of Many are different than adults', and that children disregard syntactic information that guides adult responses. I will argue that the source of the deviation lies in pragmatics. Furthermore, I will try to relate this data to a seemingly unrelated phenomenon in child language acquisition, which concerns children's interpretation of universal quantifiers.

The main experiment focused on the distinction between the proportional reading of Many (often labeled "strong") and the cardinal or existential reading (often labeled "weak")(cf. Milsark1977, Partee 1984). These different interpretations involve different Standards of Comparison (Westerstahl 1984). Dutch children and adults were presented with pictures+sentences, such as the ones below:

Unlike adults, children's responses displayed:

No sensitivity to syntactic structure.

Virtual absence of the proportional interpretation in favour of cardinal (weak) interpretations.

Context-independent interpretations of Many.

I propose that children's problem originates in the Mutual Knowledge requirements on determining the Standard of Comparison.

I further propose to extend this proposal to a well-known problem in children's interpretation of universal quantifiers known as "quantifier spreading". On the basis of this phenomenon, it has been proposed that children employ a different mapping from syntax to semantics (Philip 1995, Drozd and Van Loosbroek 1998, Geurts 2001). New experimental data, and careful analysis of the known data, show that the facts are better explained by difficulty with the Mutual Knowledge requirements.