How Do People Construct Logical Form During Language Comprehension?

How do people interpret ambiguous sentences containing more than one quantifier, such as Every kid climbed a tree? We report four sentence-picture matching experiments that used priming to investigate whether comprehenders construct logical-form representations during processing. Experiment 1 investigated priming in active-voice sentences containing transitive verbs and found priming effects of quantifier-scope relations. Experiment 2 demonstrated priming effects when prime sentences were in the passive voice (e.g., A tree was climbed by every kid) and target sentences were in the active voice. Experiment 3 used prime sentences containing existentially quantified agents and universally quantified patients (e.g., A kid climbed every tree) and found no priming effects. Experiment 4 showed no priming effects when prime sentences contained plural nouns but no quantifiers (e.g., Kids like to climb trees), thus calling into question a visual-priming account of our priming effects. Our findings suggest that people construct logical-form representations, and they do so after constructing meaning-based representations involving quantifiers and thematic-role information.