properties such as case reversal (e.g., Feustal, Shiffrin, & Salasoo, 1983; Forster & Davis, 1984). Likewise, phonological properties of masked stimulus words have been shown to influence subsequent processing (e.g., Humphreys, Evett, & Taylor, 1982; Perfetti & Bell, 1991). Furthermore, a number of findings have been reported in which subliminally presented words significantly prime semantic associates (e.g., Balota, 1983; Fischler & Goodman, 1978; Fowler, Wolford, Slade, & Tassinary, 1981; Marcel, 1983) and category coordinates (Damian, 1996; Perea & Gotor, 1997). In combination, these studies suggest that masked verbal stimuli might be processed to relatively "deep" representational levels without conscious awareness of their presentation (see McLeod, 1998, for a comprehensive review). A recent study by Dehaene et al. (1998) confirms and extends these claims about the extent of processing of subliminally presented stimuli. The collection of behavioral data was I would like to thank