Proceedings of the Annual Meeting of the Cognitive Science Society

Propositional accounts of organization in memory have dominated theory in compositional semantics, but it is an openquestion whether their adoption has been necessitated by the data. We present data from a narrative comprehensionexperiment, designed to distinguish between a propositional account of semantic representation and an associative accountbased on the Syntagmatic-Paradigmatic (Dennis, 2005; SP) model. We manipulated expected propositional-interferenceby including distractor sentences that shared a verb with a target sentence. We manipulated paradigmatic-interferenceby including two distractor sentences, one of which contained a name from a target sentence. That is, we increased thesecond-order co-occurrence between a name in a target sentence and a distractor. Contrary to the propositional assumption,our results show that subjects are sensitive to second-order co-occurrence, hence favouring the associative account.