Based on the event segmentation theory (Zacks et al. 2007), the on- and offset of an external context demarcates the beginning and end of a so-called event, temporally structuring the ongoing stream of stimulation in the environment. Following this notion, the present study is concerned with the role of context in structuring sequences of perception-action couplings (so-called stimulus-response or S-R episodes) in memory. If context also structures sequences of S-R episodes, successful retrieval of an individual episode should be impaired when multiple episodes share a context, presumably due to an increased likelihood of confusing jointly retrieved episodes as compared with a condition in which retrieved episodes are associated with different contexts. This hypothesis was tested by manipulating contextual discriminability on S-R binding and retrieval in an auditory negative priming paradigm with sequences of three successive presentations (i.e., pre-prime, prime, and probe) in each trial. In each presentation, participants identified a target sound accompanied by distractor via a keypress. Sine tones (300 Hz/600 Hz) served as context. Context tones either encompassed pre-prime and prime or changed after the response to the pre-prime target, resulting in conditions with low and high discriminability, respectively. No context was presented during the probe. Increased erroneous probe responding with the former prime response in trials with a distractor-to-target repetition was taken as indicator of S-R binding and retrieval. The increase in S-R binding and retrieval was stronger in the high- than in the low-discriminability condition, suggesting the organization of S-R episode sequences by context.