Alhama, R. G. 1 , Siegelman, N. 2 , Frost, R. 3, 5 & Armstrong, B. C. 4, 5
1 Max Planck Institute for Psycholinguistics
2 Haskins Laboratories
3 Hebrew University of Jerusalem
4 University of Toronto
5 Basque Center on Cognition, Brain and Language
Proficient reading requires rapid recognition of words in printed text. This ability has been shown to be correlated with visual statistical learning: for example, sensitivity to transitional probabilities predicts reading proficiency in both L1 and L2. To better understand how a broader set of statistical relationships shapes visual word recognition when considering the constraints of the visual system, we adopt an information-theoretic perspective, and examine how the information-content available at different fixation locations in different languages (English vs. Hebrew) drives word recognition. Using a connectionist model that maps visual inputs onto individual words, we account for a number of behavioral phenomena. First, our model predicts a tendency to fixate near the center of a word, slightly closer to word onset. Second, we demonstrate cross-linguistic differences in the likelihood of fixating at other locations than the preferred location due to availability of information-content. The simulations also make the novel prediction, confirmed by behavioral data, that words with an atypical distribution of information-content across letters are better recognized when fixating at an unusual location in a word.
Overall, this research shows how the language processing system is tuned to the perceptually-constrained statistical regularities of the writing systems, thereby driving proficient reading.