Stress or Context for the Computation of Scalar Implicatures

Since the work of Chierchia et al. (2001) and Noveck (2001), children’s ability to compute scalar implicatures (SIs) has been widely studied. The results of many studies showed that while children master the prerequisites to compute SIs, they are not as proficient as adults in computing them (a.o., Papafragou and Musolino 2003; Napoleon and Bishop 2011). Other studies (i.e., Meroni and Gualmini 2013), on the other hand, showed that even four-year-old children can compute SIs on the <some, all> scale at adult-like levels when the context is made pragmatically felicitous. Specifically, they show that the implicature is computed when an explicit Question Under Discussion (QUD) is added, as this reading constitutes the only felicitous answer to that question (Gualmini et al. 2008). Under the most natural intonation, when some is presented after a QUD that contains the stronger alternative all it receives stress, but it is destressed when the QUD contains some, which does not trigger a SI. Interestingly, only the former condition led children to compute SIs. As a consequence, it could well be the case that children in the study by Meroni and Gualmini (2013) compute SIs when the scalar item is stressed (Miller et al. 2005), independently of the QUD. In this study we aim to disentangle the relative contribution of stress and the explicit QUD to children’s (un)successful SI computation. To test the relative contribution of stress and the QUD, a Truth Value Judgment task (TVJt) was designed in which the target sentences exhibited the opposite stress pattern than in the original study. We conclude that prosody also plays a role in helping children to recognize the set of contextual alternatives which lead them to SI derivation. This study strengthens the claim that young children use and need both contextual and prosodic information to make pragmatic inferences.