Recent evidence suggests rhar awareness of the response-reinforcement contingency is a necessary, though not always sufficient, condition for the modification of behavior in a verbal conditioning paradigm. A related problem is whether or not an actual behavioral change also requires thar S voluntarily intend to make it, and whether or nor such a change can occur without awareness. Forty female undergraduate Ss were presented with a series of 100 pictures of unfamiliar male students and were asked to indicate whether they thought they would or clislike the person in each picture. Responses to the first 20 pictures were unreinforced and were used to determine an operant level. For the remaining pictures, 20 Ss were verbally reinforced for like responses and 20 for dislike responses. A different E administered a post-experimental questionnaire designed not only to assess awareness of the response-reinforcement contingency, but also to determine whether S felt that such awareness had influenced her responses and whether she was aware of any response change that might have occurred. The results indicated that the effect of reinforcement was unrelated to the response class reinforced. Th~rty-four Ss were judged to be partially or fully aware of the responsereinforcement contlngcncy, and only these Ss showed a significant (6 = 4.05, df = 33, p < ,001) increase In the relative frequency of reinforced responses. However, all but two of these aware Ss maintained that in spire of this awareness they had not voluntarily changed their behavior nor were they aware that such a change had occurred. As a matter of fact, many Ss reported that they had consciously tried not to be influenced by the knowledge of what E was doing, and that their responses simply reflected how they felt about the persons in the pictures. The fact that an apparently involuntary change in verbal behavior occurred is contrary to Dulany's (1962) suggestion thar verbal response selection requires char a behavioral intention be present, but is consistent with findings reported by Verplanck (1962, p. 138) rhar dissociation of rule and behavior may be possible. It should be noted, however, thar the behavioral changes observed, though significant, were relatively small. It is possible that the behavior of some Ss was voluntary but that they were unwilling to admit, in retrospect, thar they had been influenced by E. Further research is necessary, particularly on the adequacy of the post-experimental interview as a reflection of processes that took place during the experiment.