In a recent paper in the Journal of Experimental Psychology: Human Perception and Performance, we showed that listeners seem to reinstate musical context. First, listeners hear several melodies, each of which can be heard with one of two contexts--two different meters, two different keys. Then they hear each melody alone--without context--and judge a following probe event, like a chord or a drumbeat of a particular speed. In three experiments (as well as ongoing work in the lab), listeners' probe-event judgments are swayed toward the context that they had heard with the melody before, even though the context is no longer there. This has implications for models of music perception, and for perception of ambiguous events generally. It also suggests that when you hear that beginner-band song, you're may be hearing the out-of-tune beginner-band version of it playing back in your mind.