查看文章

Counterfactual statements are famously difficult to process, and so are negated sentences and infrequent clause orders. Here, we argue that their combination can ease much of the processing cost when these difficult constructions align to clarify what is being referred to, thereby reducing referential uncertainty. In Experiment 1, we tested how affirmative and negative counterfactual statements (e.g., If there had been (no) zebras, there would have been (no) lions) are interpreted using a web-based eye-tracking paradigm. We found that negation facilitates processing, particularly when a Question under Discussion is about the actual state of affairs. In Experiment 2, reversing the clause order resulted in easier comprehension. These results provide support for a model of incremental language processing that puts the construal of semantic representations front and center.