The results of the present study provide evidence favoring the existence of RD between auditory events that are phonologically similar but do not share identity at any level. Response biases and/or simple confusion were dismissed as the possible cause of RD in this study by applying conservative data correction procedures. However, the use of distinct data correction procedures did not affect the evaluation of the hypothesis under consideration and statistical analyses always supported the same explanation. Lexical effects were also discarded as a possible account of the RD observed here on several grounds; (a) the materials used (letter names and meaningless syllables), (b) the rate of presentation (too fast to produce lexical effects, as assessed in previous literature), and (c) the use of the same materials repeatedly and predictably throughout the experiment. This result challenges previous hypotheses arguing that RD and RB between similar targets were produced by lexical competition (e.g., Chialant and Caramazza, 1997; Miller & MacKay, 1994). Indeed, the hypothesis supported by the present data argues that the representations involved at tSe level of processing where RD and RB occur are subserved by multiple codes (Bavelier, 1994, 1999) not by a single unitary representational node.
Gallés et al. (Tue,) studied this question.