The statistical reader: The role of orthographic regularities in reading

Area of Science:

• Cognitive Psychology
• Computational Linguistics
• Psycholinguistics

Background:

• Reading research often focuses on vertical linguistic units (orthography, phonology, semantics).
• Statistical learning theories suggest readers leverage regularities in writing systems for text processing.
• Horizontal regularities (predictive relationships between units) are less explored.

Purpose of the Study:

• To introduce an information-theoretic measure for horizontal regularities in writing systems.
• To investigate how these regularities constrain lexical access under realistic constraints.
• To demonstrate how readers use horizontal regularities for efficient uncertainty reduction during reading.

Main Methods:

• Developed a novel information-theoretic measure for horizontal orthographic regularities.
• Conducted single-word reading experiments.
• Analyzed eye movement data during naturalistic reading.

Main Results:

• The proposed measure captures how horizontal regularities impact lexical access.
• Empirical studies confirm readers leverage these regularities for efficient information processing.
• Findings support computational mechanisms applicable across diverse writing systems.

Conclusions:

• Horizontal regularities are crucial for efficient reading and lexical access.
• The findings offer a unified computational framework for understanding reading across languages.
• This work has implications for reading development and skill acquisition.