Contextual modulation of the neural network underlying the processing of compositional nontransparent meaning

This study examines the neurocognitive mechanisms supporting the processing of sentences involving compositional nontransparent meaning, investigating if the neural correlates are modulated by clausal and extra-clausal prior context. We probe through sentences like “The player jumped for 5 s.“, which engenders an iterative meaning (multiple jumping actions) that is morphosyntactically-unsupported yet obtained at the compositional level ([verb + adverbial]). We hypothesize that the non-transparent meaning is computed by comprehenders’ contextual evaluation, which would be more effortful without guiding cues yet could be facilitated by the presence of biasing information in context. This predicts that the comprehension of nontransparent sentences is contextually modulated, eliciting greater cost than their transparent counterparts—particularly when they are processed in a neutral context than in an iterative-biasing context. The reported fMRI experiment showed that computing nontransparent meaning preferentially recruited the left inferior frontal gyrus (L.IFG), the left middle temporal regions, and the right IFG, in contrast to the transparent counterpart. Crucially, the left frontal activation subserving nontransparent sentences was attenuated in a biasing context, as compared to a neutral context. The context effect was corroborated by the results of iterativity judgments that showed differential iterative vs. non-iterative interpretations for the nontransparent sentences as cued by the clausal and extra-clausal context. While the influence of clausal context has been demonstrated, this study provides novel evidence showing that compositional meaning computation is modulated by prior context in addition. The findings reveal a left-lateralized frontal-temporal network for compositional nontransparent meaning that is subject to contextual modulation beyond morphosyntactic computation.