The author analyses implications of the results of research on negative priming for the theory that attributes this phenomenon to inhibitory mechanisms in selective attention. According to an early inhibitory account of negative priming, the representation of the stimulus ignored or selected against in the prime (preceding) display is actively suppressed, which entails a short-term decrease in the representation's activation level below baseline. As a result, when a response to this stimulus is required in the probe (subsequent) display, accessing the stimulus representation may be more difficult. Contrary to this account, it was found that negative priming may persist for quite a long time and that it depends on the conditions in, and on the broader context of, the probe display in which an earlier distractor appears as the target. These results are congenial with noninhibitory accounts of negative priming, which attribute the effect to a conflict or difficulty arising when on the probe trial an episodic representation of the probe target is retrieved that was established when this stimulus served as a distractor. Following the integrative proposal put forward by Tipper (2001), the author considers modifications the inhibitory account seems to require to accommodate the data indicating that both inhibition and memory processes play a role in negative priming. A basic change is a revision of the assumptions concerning the way the effects of inhibition are carried over from the prime trial to the probe trial. A modified inhibitory account assumes that this transfer involves memory coding (on the prime trial) and retrieval (on the test trial) of inhibitory processes or their effects. It is shown that this modification necessitates a revision of other assumptions of the original inhibitory account, especially those concerning the nature of the representations involved in negative priming and the very concept of inhibition. Taking into account the role learning and memory processes play in selective attention puts the mechanisms of the latter in a new theoretical perspective. The transition is from analysing attentional selection in a narrow focus of the question that concerns how the organism solves a current problem of distinguishing task-relevant stimuli from distractors, to analysing it in a broader context of the issue of how the organism in its interactions with a given environment learns to categorise stimuli as relevant or irrelevant, to represent them in the context of an activity as those that should be attended to and those that should be ignored.