Stored Knowledge versus Depicted Events: what guides auditory sentence comprehension?

In a seminal article, Tanenhaus, Spivey-Knowlton, Eberhard & Sedivy (1995) showed that eye-movements to real-world objects reflect a rapid interplay of utterance and visual environments in sentence comprehension. Further, Kamide, Scheepers & Altmann (2003) found that when linguistic and world knowledge constrain the domain of reference in a visual scene, people even anticipate as yet unmentioned arguments/referents. Studies by Knoeferle, Crocker, Scheepers & Pickering (2003) have since revealed that when linguistic and world knowledge did not disambiguate an initial syntactic and role ambiguity, depicted agent-action-patient events permitted anticipation of thematic role-fillers online. This paper opposes linguistic and world knowledge on the one hand, and visual scenes on the other hand in order to determine their relative importance in auditory comprehension. We observed a preferred reliance of auditory sentence comprehension processes on information that had to be extracted from depicted event scenes. Determining the nature and time-course of the interaction between linguistic/world knowledge and visual scenes is a first step towards developing a theory of real-time auditory sentence comprehension in visual environments. Our finding has implications for theories of the language faculty (e.g., Jackendoff, 2002).