Across many domains of cognition, information processing is more effective when it has been performed recently than when it has not. This effect - termed repetition priming - is ubiquitous; however, its causes are not well understood. Here, I draw attention to the concept of antipriming and argue that it is crucial for understanding priming. Antipriming is a measurable impairment in processing information owing to recent processing of other information when the representations of information overlap and compete. Strengthening one representation after its usage causes priming for that item but also antipriming for some other, non-repeated items. Recent evidence demonstrates priming and antipriming within visual object identification systems. These findings might reflect a form of maintenance relearning of superimposed knowledge representations.