N400Effects Of Prediction Within Chinese Contexts Of Relative Constructions

When a context of a strongly-constrained sentence structure is formed, N400as the ERP (event-related potential) component of the follow-up word of the structure will be reduced in amplitude. There are currently two accounts to describe this phenomenon:a passive activation account explains the reduction as a passive spread of activation across long-term semantic memory representations and a contextual-based prediction account states the scene as specific predictions about upcoming input. The dispute lies on the extent of how each of these accounts influences the N400element reduction. This issue has been firstly addressed in this thesis under Chinese relative construction context by formatting semantically associated Chinese relative construction prime and word target pairs within an experimental context that encouraged prediction to a greater degree. The proportion of related pairs was used to modulate the predictive validity of the relative construction prime for the target while holding constant semantic association. A semantic category probe detection task was used to encourage test subjects to process semantic meaning of these experimental pairs without their possible preferences towards the content of trials of interest. A larger N400reduction to related targets was observed in the high-proportion block of relatedness than the low condition. This result is consistent with the hypothesis that predictions towards upcoming input significantly results in the N400effect. In summary, the results suggest that predictability modulates N400amplitude in a dominant manner under a constrained context like Chinese relative construction.