A within-subjects comparison of the acquisition of quantity-related inferences

Abstract

This study directly compares quantity inferences from scalar implicatures ('Some of the ducks are black') and uniqueness presuppositions in definites ('the duck is black') to exhaustivity inferences in English it-clefts ('It's the duck that's black') for which the theoretical literature disagrees on the source of inference -pragmatic (like scalar implicatures), or semantic (like presuppositions). We investigate whether within-subjects correlations in acquisition can inform us about the source of exhaustivity inferences. Assuming comprehension is achieved once the necessary basis for meaning is acquired, it-clefts should pattern with presupposition judgments if computing a presupposition is involved and should pattern with scalar implicature judgments if computing an implicature is involved. We conduct three experiments to test how closely it-cleft judgments pattern with other quantityrelated inferences, keeping materials maximally similar. The first two experiments test adult participants using a Truth Value Judgment Task and then a 3-point Rating Task; we find that adults' response patterns to under-informative uses of these constructions differ both across individuals and across inference types, with the Rating Task giving more informative results. In the third experiment, we use the 3-point Rating Task with 4-, 5-, and 6-year olds to characterize response patterns across the three inference types for each individual subject. We find that the individual response patterns children exhibit are consistent with the theory that it-cleft exhaustivity shares an underlying cognitive source with the computation of presupposition inferences, but not with scalar implicature inferences.