Recent research has shown that pseudowords created by transposing letters are very effective for activating the lexical representation of their base words (e.g., relovution activates REVOLUTION). Furthermore, pseudoword transpositions of consonants are more similar to their corresponding base words than the transposition of vowels. We report one experiment using pseudowords created by the transposition of two consonants, two vowels, and their corresponding control conditions (i.e., the replacement of two consonants or two vowels) in a lexical decision task while Event Related Potentials (ERPs) were recorded. The results showed a modulation of the amplitude of the N400 component as a function of the type of pseudoword (transposed-letter versus replacement letter pseudowords), and this modulation was different for transposed consonants and vowels. These results suggest that consonants and vowels play a different role during word processing. © 2007 Elsevier Ireland Ltd. All rights reserved.Keywords: Visual-word recognition; Transposed-letters; ERPs; Consonants and vowels When we read it is common to misread words like causal and casual. A growing number of studies have shown that, during the recognition of any given word, not only is the representation of the word itself activated, but also the representations of similarly spelled words ("neighbors"). Although the majority of "neighborhood" experiments have focused on one specific type of neighbor: one-letter different neighbors (e.g., trail and train), recent research has shown that transposed-letter neighbors may be even more perceptually similar to the target stimulus than oneletter different neighbors (trail and trial; [28,31]). Most notably, the presence of transposed-letter effects has critical implications for the choice of an input coding-scheme in visual-word recognition. Most current computational models of visual-word recognition [7,13,20] assume that each letter is encoded in a different "letter-channel", and hence they cannot accommodate the presence of transposed-letter effects.To overcome the limitations of a channel-specific codingscheme, a number of input coding-schemes have been proposed (SERIOL model [34] [14]; overlap model [12]). Although the basic mechanisms of how letter position is encoded differ across these models, they all predict that transposed-letter neighbors like casual and causal are perceptually very similar. There is one caveat, however: for simplicity's sake, these models assume that consonants and vowels are processed in the same way. As shown in the present paper, this assumption may be an oversimplification. Recent transposed-letter experiments [28,27] have shown that consonants and vowels play a different role in visual-word recognition. In particular, Perea and Lupker [28] obtained masked priming effects for consonant transpositions (relovución-REVOLUCIÓN versus retosución-REVOLUCIÓN), but not for vowel transpositions (reluvoción-REVOLUCIÓN versus relavición-REVOLUCIÓN). Furthermore, these findings have been extended...