Di Bacco, Federica, Tieu, Lyn, Moscati, Vincenzo, Sevdali, Christina, Folli, Raffaella and Romoli, Jacopo
Proceedings of WCCFL34, Salt Lake City
Publication year: 2016


Children and adults have been reported to differ in their interpretation of scopally ambiguous sen- tences such as Every horse didn’t jump over the fence (Musolino 1998; Gualmini 2004; Gualmini et al. 2008; Musolino & Lidz 2006; see also Lidz & Musolino 2002; Musolino et al. 2000; Musolino & Lidz 2006; Kra ̈mer 2000; Moscati & Crain 2014; Moscati et al. 2016, among many others). A recent approach in the literature treats this difference as fully pragmatic in nature. In particular, Gualmini et al. (2008) have proposed an explanation based on what they call the Question-Answer Requirement (QAR), which locates the source of the difference in the understood Question Under Discussion (QUD) in the context. The main idea behind the QAR is that any sentence is to be understood as an answer to a QUD. As a consequence, in the case of scopally ambiguous sentences, a given reading of the sentence is accessible (to adults and children) only if it constitutes a possible answer to the contextual QUD. Children and adults are then claimed to differ only in how they handle and accommodate QUDs. In particular, if the reading that would answer the salient QUD is false in the context, adults, but not children, are able to accommodate a new QUD in order to access the true interpretation of the ambiguous sentence.
Given that it locates the difference only in the way QUDs are managed, the QAR approach predicts that if the effects of the QUD are controlled for, children and adults should perform alike. In this study, we indirectly probed the role of the QUD in scope ambiguity resolution by moving to a linguistic context where there is no sentence that constitutes an answer to the relevant QUD. To do so, we tested children’s and adults’ comprehension of scopally ambiguous questions, rather than declaratives. The main finding of our experiment is that 4–6-year-old children and adults display the same rates of access to the inverse scope readings of such questions. This finding is consistent with a general assumption underlying the QAR approach, which is that there is no grammatical difference in children’s and adults’ ability to access inverse scope readings. Indeed, once the QUD factor is controlled for, the two groups perform alike.
We further discuss the compatibility of our results with two alternative hypotheses concerning the QAR and the interpretation of questions. The first possibility is that the QUD is simply not a factor that affects the interpretation of questions. This would explain why we found no difference between children and adults. In this scenario, the QAR approach to scope interpretation should not be extended to questions. The second possibility is that the QAR approach should be formulated in such a way that it can be extended to explain performance on questions, in particular invoking a notion of super- question and sub-question (Roberts, 1996). In this latter scenario, the lack of a difference between children and adults would be due to the fact that our experimental contexts made salient a super-question that facilitated inverse scope interpretations. We end by discussing a possible formulation of such an extended QAR approach to questions, and sketch a follow-up experiment that could distinguish between the two scenarios just mentioned.