Dynamicity Predicts Inferred Temporal Order in Complex Sentences: Evidence from English, German, and Polish.

To build an accurate mental model of complex situations, people infer temporal order from sometimes underspecified linguistic information. The basis on which these inferences are drawn is an open question. While previous literature has focused on the role of linguistic structure and discourse pragmatic strategies as important contributors to temporal inferences, here we argue that, under uncertainty, people also use the dynamic properties of the described situations to derive temporal order from language. In three pre-registered studies using English, German, and Polish, adult participants used toys to act out complex situations described by main clause-relative clause structures. We consistently find that non-dynamic state descriptions are temporally ordered first, if the other clause describes a dynamic event. This pattern arises independently of whether dynamicity differences are lexically encoded, like in English or German, or grammatically encoded, like in Polish. More generally, our findings address an important gap in the discussion on the role of eventuality type for temporal inference. While there is substantial research on the significance of telicity and durativity, a third, much more overlooked feature is dynamicity, a concept rooted in event perception, not language. Our results therefore provide a crucial thread to closely weave together language comprehension and event cognition.