TURead: An eye movement dataset of Turkish reading.

Area of Science:

• Psycholinguistics
• Cognitive Science
• Computational Linguistics

Background:

• The Turkish language, an agglutinative language with shallow orthography, is underrepresented in reading research.
• Understanding oculomotor control in reading requires datasets that capture morphological complexity.

Purpose of the Study:

• To introduce TURead, a novel eye movement dataset for Turkish silent and oral sentence reading.
• To investigate the influence of Turkish morphology, specifically suffixation and word length, on oculomotor control.
• To provide empirical data for psycholinguistic research on reading processes in agglutinative languages.

Main Methods:

• A target-word approach was used, manipulating word length and suffix addition.
• Collected eye movement data (fixation durations, saccades) during silent and oral reading tasks.
• Included linguistic features (frequency, predictability) and cognitive measures (working memory, Cloze scores).

Main Results:

• Eye movement patterns, specifically fixation parameters, were analyzed in relation to word characteristics.
• Findings demonstrate consistency with established patterns in reading literature, validating the dataset's utility.
• The study provides insights into how morphological complexity impacts reading behavior.

Conclusions:

• TURead offers a valuable resource for studying the interplay between morphology and oculomotor control in Turkish.
• The dataset supports further research into reading processes across diverse linguistic structures.
• Results confirm existing psycholinguistic models of reading while highlighting the need for cross-linguistic data.