Several studies (e.g., Nissen & Bullemer, 1987) claimed that reaction times to a repeating sequence may improve although subjects are not aware of the repeating sequence. Perruchet and Amorim (1992) pointed out that the measure of awareness involved in these studies was inadequate (e.g., subjects were not even explicitly asked to retrieve the repeating sequence), and they showed that the dissociation in normal subjects no longer held when awareness was assessed by recall or recognition tests. In this reply, we show that Cohen and Curran's (1993) criticisms of the validity of Perruchet and Amorim's tests and of the theoretical implications of their results are either without foundation or unfalsifiable. We also show that the new experiment by Willingham, Greeley, and Bardone (1993) does not demonstrate dissociation. Both comments further illustrate the widespread uncritical acceptance of dissociation, which probably originates from, but is not theoretically justified by, evidence available for dissociation in amnesic patients.