Task switch costs (i.e., worse performance when a task is executed after a trial featuring a different task than after a trial featuring the same task) are attributed, in part, to proactive interference resulting from previous processing of the current stimulus in a competing task (i.e., the task switched from). Such stimulus-specific proactive interference may also occur, however, when a current stimulus was presented in a previous task context distinct from the tasks administered in the current switching procedure. To investigate the effect of stimulus presentation in a previous task context, participants switched between tasks afforded by distinct sets of stimuli (i.e., “single-affordant” stimuli) and underwent reversal of the stimulus-response assignment in one of the tasks after a first phase of the experimental session. The reversal increased task switch costs in the task subject to the reversal undiminished by lengthened preparation time (although longer preparation time reduced both reaction time in general and the switch cost). These results demonstrate that stimulus-specific proactive interference affects task performance more strongly in switch trials than in repetition trials and is not reduced by task preparation (thus contributing to the residual switch cost). The reversal-induced component of the task switch cost remained substantial during the remainder of the experimental session. By contrast, performance in the other (unaltered) task suffered only temporary (task sequence-unspecific) impairment, suggesting general processing impediment induced by acute relearning demands.