Flexible Letter-Position Coding in Chinese-English L2 Bilinguals: Evidence from Eye Movements

Theories suggest that efficient recognition of English words depends on flexible letter-position coding, demonstrated by the fact that transposed-letter primes (e.g., JUGDE-judge) facilitate written word recognition more than substituted-letter primes (e.g., JUFBE-judge). The multiple route model predicts that reading experience should drive more flexible letter-position coding as readers transition from decoding words letter-by-letter to recognising words as wholes (Grainger et al., 2012). This study therefore examined whether letter-position is coded flexibly in second language English sentence reading for native Chinese speakers, and if this is influenced by English proficiency. Eye-movements were measured whilst 54 adult native Chinese speakers read English sentences including either a real word (e.g., cheaply), a transposed-letter nonword (e.g., ‘chepaly’), or a substituted-letter nonword (e.g., ‘chegely’). Flexible letter-position coding was observed in initial and later processing stages— reading times were longer for substituted-letter than transposed-letter nonwords. Additionally, reading times were longer in both initial and later processing stages for transposed-letter nonwords than real words indicating that, despite encoding letter-position flexibly, readers processed letter-position. Although pre-registered frequentist analyses suggested that English proficiency did not predict overall reading times, Bayes Factors indicated that there was evidence for such a relationship. It is therefore likely that this proficiency analysis suffered from low power. Finally, neither frequentist nor Bayes Factor analyses suggested that English proficiency influenced the difference in reading times between different target word types, i.e. the nature of letter-position coding. Overall, these results suggest that highly proficient L2 learners code letter-position flexibly.