Abstract
An important characteristic of the word superiority effect (WSE) observed in normal subjects is that it extends to words displayed in mixed-case letters, e.g. fAdE is better identified than gAdE (McClelland, 1976). Because upper- and lower-case letters are treated as functionally equivalent within the orthographic system (e.g. “A”/“a” map onto abstract letter identities; Coltheart, 1981), it is often argued that orthographic codes mediate the WSE. In the present paper, we report an intact WSE in a letter-by-letter surface alexic (IH) when words and “word-like” pseudowords were displayed quickly and then masked. Consistent with the claim that orthographic codes mediated these effects, the WSE extended to words typed in mixed-case letters, and IH failed to show a WSE for the same class of words for which he has impaired orthographic knowledge; namely, low frequency words. Based on these results, we argue that IH gains better access to orthographic knowledge than current theories of letter-by-letter reading would predict.