Document Type

Thesis

Degree Name

Master of Science (MSc)

Department

Psychology

Faculty/School

Faculty of Science

First Advisor

Todd Ferretti

Advisor Role

Thesis Supervisor

Abstract

Recent research examining differences in the way the left (LH) and right (RH) hemispheres of the brain process language have used the visual half-field (VHF) paradigm to examine whether each hemisphere can independently process information from sentences. The current study expanded upon such work by using event related brain potential (ERP) measures to examine how the comprehension of thematic role knowledge, a process essential to successful sentence comprehension (MacDonald, Pearlmutter, & Seidenbeerg, 1994), is undertaken in each hemisphere. During language comprehension, agents (entities that initiate action in an event) depicted by nouns (e.g., cop) have been shown to be associated with verbs presented in active voice (e.g., present participle—cop was arresting), whereas patients (entities that have action imposed upon them—e.g., crook for the event arrest) are associated with verbs presented in passive voice (e.g., crook was arrested—Ferretti, McRae, Elman, & Ramshaw, 2005). Thus, the currently study examined how the cerebral hemispheres conjointly (Experiment 1) and independently (Experiment 2) process thematic fit (whether the noun and verb are thematically related) in conjunction with morphosyntactic information indicating whether a verb is in active (arresting), or passive (arrested) voice in order to further understand each hemisphere’s sentence processing abilities. In Experiment 1, ERP responses to verbs (presented centrally) preceded by a common agent noun and presented in related-active (e.g., The teacher was lecturing the man), unrelated active (e.g., The shepherd was lecturing the man), related-passive (e.g., The teacher was lectured by the man) and unrelated passive (e.g., The shepherd was lectured by the man), sentences were analyzed in order to examine the effects of relatedness and voice on the processing of the critical verb. Experiment 1 results showed that brainwave amplitudes demonstrated greater semantic facilitation for related active than for unrelated active verbs. However, this effect did not occur as strongly for passive verbs, indicating that the congruency between the agent and verb form significantly affected processing. In Experiment 2, the exact same experimental design was employed with the procedural variation that the critical verbs were presented 2⁰ to either the left or right of visual fixation, so only one hemisphere initially receives the information. Results from this experiment showed that with LH presentation, as with central presentation, related active verbs were processed more easily than unrelated active verbs and the magnitude of this effect was larger than that for analogous differences found with passive verbs (although sensitivity to relatedness for passive verbs was slightly larger than in Experiment 1). However, in the RH, differences in relatedness were found only for passive items indicating that sensitivity to relatedness differences in this hemisphere may be mediated more by frequency differences between verb forms than with LH processing (or central presentation). These results provide evidence against the theoretical contentions that the RH is only sensitive to simple word-level semantic relations (e.g., Faust, 1998), and are supportive of an ‘expectancy’ account of LH language processing (e.g., Coulson et al., 2005; Federmeier & Kutas, 1999).

Convocation Year

2007

Convocation Season

Fall

Download

Share

COinS