Decoding Letters: Eye-Tracking Evidence that Bigram Frequencies Influence Reading and Naming

Language learning, word recognition, and reading rely on the continuous extraction of statistical regularities within a language. The current study examined how the orthographic structure of words in the English language influences letter recognition, a precursor to reading and writing. Twenty-five native English speakers named lists of letters while their eye movements and vocal responses were recorded. Letters in each list were arranged in high-probability or low-probability sequences based on the bigram frequency between two adjacent letters. Eye movement data showed that participants fixated longer on lists of letters arranged in high-probability than low-probability sequences, likely because letter pairs with higher bigram frequencies activated a larger network of words than letter pairs with lower bigram frequencies. Higher next bigram frequency was associated with longer fixations on each letter, whereas higher preceding bigram frequency, current letter frequency, and next bigram frequency were each associated with faster letter naming fluency. We conclude that statistical regularities in language build expectations for upcoming letters, with implications for reading, pattern extraction, speech perception, and language learning and processing more broadly.