Eye-tracking as a window into assembled phonology in native and non-native reading

The past 30 years of reading research has confirmed the importance of bottom-up processing. Rather than a psycholinguistic guessing game (Goodman, 1967), reading is dependent on rapid, accurate recognition of written forms. In fluent first language (L1) readers, this is seen in the automatic activation of a word’s phonological form, impacting lexical processing (Perfetti & Bell, 1991; Rayner, Sereno, Lesch & Pollatsek, 1995). Although the influence of phonological form is well established, less clear is the extent to which readers are sensitive to the possible pronunciations of a word (Lesch & Pollatsek, 1998), derived from the varying consistency of grapheme-to-phoneme correspondences (GPCs) (e.g., although ‘great’ has only one pronunciation, [ɡɹeɪt], the grapheme within it has multiple possible pronunciations: [i] in [plit] ‘pleat’, [ɛ] in [bɹɛθ] ‘breath’; Parkin, 1982). Further, little is known about non-native readers’ sensitivity to such characteristics. Non-native readers process text differently from L1 readers (Koda & Zehler, 2008; McBride-Chang, Bialystok, Chong & Li, 2004), with implications for understanding L2 reading comprehension (Rayner, Chace, Slattery & Ashby, 2006). The goal of this study was thus to determine whether native and non-native readers are sensitive to the consistency of a word’s component GPCs during lexical processing and to compare this sensitivity among readers from different L1s.