Word segmentation of overlapping ambiguous strings during Chinese reading

Area of Science:

• Cognitive Psychology
• Psycholinguistics
• Computational Linguistics

Background:

• Chinese reading involves segmenting characters into words, which can be ambiguous.
• Ambiguous strings in Chinese allow for multiple word interpretations.
• Understanding word segmentation mechanisms is crucial for reading research.

Purpose of the Study:

• To test three hypotheses explaining the segmentation of overlapping ambiguous strings in Chinese.
• To determine the underlying mechanism of word segmentation in Chinese readers.
• To investigate the role of competition and context in resolving reading ambiguity.

Main Methods:

• Conducted three experiments to evaluate different word segmentation hypotheses.
• Focused on 3-character ambiguous strings where adjacent characters can form words.
• Compared the left-priority, independent processing, and competition hypotheses.

Main Results:

• Results indicated that the competition hypothesis best explains the observed data.
• Evidence suggests that characters activate all possible word candidates simultaneously.
• The winning word is determined by its activation level, not solely by position or context.

Conclusions:

• The findings support a competition-based account for resolving Chinese word segmentation ambiguity.
• Word segmentation is not strictly left-prioritized or context-dependent but relies on activation-based competition.
• This research provides insights into the dynamic process of word recognition in Chinese.