The Interaction of Contrastive Stress and Grammatical Context in Child English Speakers’ Interpretations of Existential Quantifiers
Loading...
Date
2009-06
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
The Ohio State University
Abstract
Scalar implicatures, such as the “some, but not all” implicature associated with the existential quantifier “some”, are systematically canceled in downward entailing (DE) environments, such as the antecedent clause of a conditional sentence (cf. Ladusaw 1979). We observe, however, that this no-implicature “some” must be prosodically de-accented. This contrasts with the SOME used with implicatures which must carry a L+H* pitch accent. Further, pitch-accented SOME is more associated with implicatures than is the de-accented version of some. The vowel-reduced sm (cf. Postal 1964, Milsark 1977) is more associated with the pure existential, no-implicature interpretation than is the prosodically de-accented, but full-vowel version of some. Further, pitch-accented SOME appears rarely without an implicature, and vowel-reduced, de-accented sm may not associate with an implicature.
Given the apparent importance of vowel duration and pitch accent for implicature generation, we investigate child English speakers’ awareness of them. Guasti et al (2005) shows that 7 year-old Italian-speaking children generate implicatures with the Italian version of some. Similarly, Chierchia et al (1998) show that child English and child Italian speakers both compute and cancel implicatures associated with English and Italian some. Papafragou & Tantalou (2004) also show that child Greek speakers compute the implicature associated with Greek some, which must be accompanied by contrastive stress. However, the importance of prosodic and segmental properties for implicature generation and cancellation in child English have not been addressed.
We predict 1) that if pitch accent is crucial to computing an implicature, children should generate an implicature with pitch-accented SOME, not only in non-DE environments, but also in DE environments, 2) if vowel reduction is crucial to implicature cancellation, then sm should allow more implicature cancellation than deaccented some and 3) if the presence/absence of pitch accent is crucial to implicature cancellation, then deaccented some should allow more implicature cancellation than pitch-accented SOME.
Using a between-subjects design with six groups (3 groups of adults, n=51; and 3 groups of children, n=40, age range=3;8-5;8 , mean age=4;5), we tested our 3 predictions using a video-recorded Truth Value Judgment Task. Regarding prediction 1, adults generated implicatures with pitch-accented SOME in both DE and non-DE environments - more than with either sm (χ2=4.37, p=.037) or some (χ2=11.6, p=.001). Children generated implicatures with pitch-accented SOME in DE contexts more than with deaccented some (χ2=7.17, p=.007). This was not true in non-DE contexts. For prediction 2, adults allowed significantly more implicature cancellation with sm than with deaccented some in non-DE contexts (χ2=4.37, p=.037), but not in DE contexts, confirming that DE environments cancel implicatures in the absence of pitch accent. The same results were significant for children in both non-DE contexts (χ2=5.6, p=.018) and DE contexts (χ2=16.9, p<.0001). For prediction 3, adults allowed more implicature cancellation with deaccented some than with pitch accented SOME in both non-DE (χ2=11.6, p<.001) and DE contexts (χ2=18.5, p<.0001), confirming the importance of pitch accent in implicature cancellation. The same was true for children, but only in DE contexts (χ2=7.17, p=.007).
Description
Keywords
quantifier, implicature, children, contrastive stress