Good-Enough Language Processing: Evidence from Sentence-Video Matching

This paper investigates how detailed a linguistic representation is formed for descriptions of visual events. In two experiments, participants watched captioned videos and decided whether the captions accurately described the videos. In both experiments, videos depicted geometric shapes moving around the screen. In the first experiment, all of the captions were active sentences, and in the second experiment, half of the captions were active and half were passive. Results of these experiments indicate that participants who only encountered active sentences performed less detailed analyses of the sentences than participants who encountered both active and passive sentences, suggesting that the level of linguistic detail encoded reflects the complexity of the task that participants have to perform. These results are consistent with "good enough" models of language processing in which people process sentences heuristically or syntactically depending on the nature of the task they must perform.