Theme: Semantics and Structural types of pronoun. Plan

Procedure—Participants listened to pre-recorded sentences while looking at displays presented on a computer monitor. The participants’ task was to click (using the computer mouse) on the picture mentioned in the sentence they heard. As in the other experiments, participants were familiarized with the characters’ names during a training phrase. The displays were presented on a Macintosh eMac G4 using Psyscope, which was also used to record participants’ responses. The sound files were played over external speakers. A light-weight head-mounted eye-tracker (ISCAN EC-501) was used to record participants’ eye movements.
Data analysis—Using a SONY DSR-30 digital VCR with jog-shuttle control, a trained research assistant listened to the audio portion of the videotape and determined the onset of critical target sentences, marking the frame at which they occurred. The time between the sentence onset and the onset of the anaphor was measured using Praat software (Paul Boersma and David Weenink, Institute if Phonetic Sciences, University of Amsterdam) and this was used to calculate where in the video record the onset of the pronoun or reflexive occurred. The video record was analyzed frame-by-frame (without sound) starting 30 frames (one second) before the onset of the pronoun or reflexive and continuing for 90 frames (three seconds). Coding consisted of recording frame-by-frame whether the participant was looking to the left character, right character, left picture, right picture or elsewhere. Because the audio was turned off, the coder was blind to experimental condition. These data were then used to determine which regions had been fixated over time.1
Predictions
Predictions for picture-choice data—The picture choice results are predicted to be the same as Experiment 2a. We predict that pronouns will show significant sensitivity to a structural anti-subject constraint as well as a semantic perceiver constraint, and that reflexives will be sensitive primarily to a structural subject constraint, with weaker effects of a semantic source constraint.