Journal of Experimental Psychology: Learning, Memory, and Cognition (in press) Alternating-script priming in Japanese: Are Katakana and Hiragana characters interchangeable?

Models of written word recognition in languages using the Roman alphabet assume that a word's visual form is quickly mapped onto abstract units. This proposal is consistent with the finding that masked priming effects are of similar magnitude from lowercase, uppercase and alternating-case primes (e.g., beard-BEARD, BEARD-BEARD and BeArD-BEARD). We examined whether this claim can be readily generalized to the two syllabaries of Japanese Kana (Hiragana and Katakana). The specific rationale was that if the visual form of Kana words is lost early in the lexical access process, alternating-script repetition primes should be as effective as same-script repetition primes at activating a target word. Results showed that alternating-script repetition primes were less effective at activating lexical representations of Katakana words than same-script repetition primes—indeed, they were no more effective than partial primes that contained only the Katakana characters from the alternating-script primes. Thus, the idiosyncrasies of each writing system do appear to shape the pathways to lexical access. Current models of written word recognition in languages using the Roman alphabet assume that information about the visual form of the word's component letters is rapidly lost in the early stages of processing, so that the critical stages in lexical access are based on the & Dufau, 2008). Consistent with this view, previous research has reported similar size priming effects for uppercase word targets that are preceded by a masked lowercase repetition prime (i.e., one in which the letters are nominally, but not physically, the same; e.g., beard-BEARD) versus a masked uppercase repetition prime (i.e., one in which the letters are physically the same; e.g., BEARD-BEARD) in lexical decision experiments for electrophysiological evidence). Furthermore, and most relevant to the present research, alternating-case repetition primes (BeArD-BEARD) have been shown to be as effective as lowercase repetition primes (beard-BEARD) in masked priming lexical found that alternating-case primes were as effective as lowercase primes in a masked associative priming lexical decision experiment (e.g., similar word identification times for eRrOr-MISTAKE and error-MISTAKE). In sum, ALTERNATING-SCRIPT PRIMING 4 4 empirical evidence in languages using the Roman alphabet strongly suggests that, during visual word recognition, there is a fast mapping of the visual form of the words' component letters onto their corresponding abstract letter representations (see Carreiras, Armstrong, Perea, & Frost, 2014, for a review). An important question is how far this conclusion about the abstract nature of letter/character-level representations can be generalized. That is, although the lowercase-uppercase …