跳到主要內容區塊
:::
開啟
2026/5/14 下午 08:25:49
:::
  • default image
  • 演講或講座
  • 語言學研究所
The Interplay of Contextual Prediction and Lexical Retrieval in Natural Reading: Insights from Eye-Movement and Fixation-Related Potentials

2026-05-18 10:00 - 12:00

加入行事曆

The reading systems framework (RSF), developed by Stafura and Perfetti (2014), proposed that reading comprehension depends on the dynamic interplay between bottom-up word identification (indexed by word frequency) and top-down contextual prediction (indexed by word predictability). However, the modular view holds that context only affects the post-lexical semantic integration and postulates independent (additive) effects of word frequency and predictability. The interactive model suggests that context directly influences lexical processing, allowing interactions between frequency and predictability. Prior research has been inconclusive. Most eye-movement (EM) studies support the modular account, reporting additive effects. In contrast, event-related potential (ERP) studies often find interactions on the N400 component. ERP studies use Rapid Serial Visual Presentation (RSVP) to avoid oculomotor contamination, which doesn't reflect natural, self-paced reading. This research used simultaneous recordings of eye movements and fixation-related potentials (EMFRPs) to investigate how these two information streams interact during natural reading of Chinese sentences. The data showed that contextual information is available and processed early (N1), followed by lexical familiarity checks (P200), culminating in a bidirectional interaction during semantic integration (N400). This research underscores the utility of combining corpus-based analyses with EMFRPs. By capturing naturally occurring language variations, the research revealed that the brain uses all available information—both lexical knowledge and contextual cues—in an incremental, dynamic manner to achieve efficient comprehension.

回頂端