|
Powerful Pictures: Priming Planning or Production or Both? Kerstin Hadelich, Matthew W. Crocker and
Christoph Scheepers It was shown by Prat-Sala and Branigan (2000) that, given a linguistic context, the accessibility of an entity and thus the position of its expression in a sentence is dependent on the salience of this entity. It is yet unclear whether this effect could also be accomplished using a visual context. The goal of the present study was to investigate the effects of repeatedly presented visual stimuli on constituent order and onset latencies. In addition, we analysed the effect of a visual prime on eye movement patterns. This was in order to clarify the strength of the supportive effect of fixating an object for word retrieval indicated in recent studies (e.g. Griffin & Bock, 2000; v.d.Meulen, Meyer & Levelt, 2001). The subjects saw a visual prime followed by an action scene in which one of the objects was involved in an action with another, unfamiliar object. The thematic role of the objects was varied and the prime could either be the agent or the patient of the action (e.g. a crocodile or a magician preceding an image of a magician hitting a crocodile). Subjects had to describe the second scene in a sentence whereby they were free in their choice of sentence structure. Eye movement patterns, onset latencies and utterance content were compared to a no-prime baseline condition. The analyses showed that both prime conditions, independent of the agent status of the prime, seem to speed up production processes. Yet, effects are stronger when the prime reappears as the agent of the action scene. Additionally, subjects preferentially produced active sentences in regular SVO order and did not produce any OVS sentences. OVS sentences are grammatically possible in German and, analogous to Prat-Sala and Branigan's results, could have been expected. However, the prime did influence the probability of an entity occurring in sentence initial position: When the human entity was primed, it occurred more often at the beginning of the sentence than when there was no prime. This leads to the assumption that visual priming, too, is able to increase the conceptual accessibility of an entity and thereby affects word order. As for the eye movements, we observed the usual pattern of subjects spending more time fixating the agent (e.g. initial NP) of the scene before utterance onset. Subjects were influenced by the prime preceding the scene, however, fixating the primed object but spending less time on it.
|