Pragmatic inferences: Neuroimaging of ad-hoc implicatures

Abstract

During conversation, comprehenders often make pragmatic inferences, or implicatures. Our study concerns ad-hoc implicatures, which are quantity-based implicatures. For example, the sentence “I walked Lassie”, where the addressee knows that the speaker has 2 dogs, signals to the addressee that the speaker wanted to convey an enriched meaning (i.e., she walked Lassie, but not the other dog). On some accounts, it is assumed that these implicatures are derived similarly to the well-studied scalar implicatures. Yet, ad-hoc implicatures received little attention. In the current study, we used fMRI to further uncover the mechanisms that support pragmatic inferences, and specifically ad-hoc implicatures. In our judgment task, we first presented a context picture with several objects, then the target sentence (sans picture) which referred to either the subset (implicature condition) or the whole set of objects (no-implicature condition), and finally asked participants to judge whether a final picture matched the target sentence. Comparing the implicature and no-implicature conditions, we observed activations in the rostrolateral prefrontal cortex, which we linked to inference generation, and in the right inferior parietal lobule, which we linked to theory of mind or attention shift. We also performed an ROI analysis, examining activations related to ad-hoc implicatures in regions previously linked to scalar implicatures and to other types of context-based implicatures, showing overlaps and dissimilarities in both cases. Thus, our results are not completely in line with theories that argue for one type of processing in the derivation of pragmatic inferences.