We rely on memory to form beliefs, but we also frequently revisit memories in conversation and private reflection. I show experimentally that such rehearsal of past experiences generates systematic belief biases. Participants are given a set of experiences and then randomized to have conversations about a subset of them, either ones that reflect well or poorly on them. Such rehearsal has large effects on which of the original experiences participants can recall a week later. Crucially, participants appear naive about rehearsal effects: they take what they remember at face value when later incentivized to form accurate beliefs about the full set of original experiences. Rehearsal therefore distorts not only future recall but also future beliefs. Participants also make rehearsal choices without regard to their later distortionary effects. Intrinsic preferences for thinking about certain experiences instead drive rehearsal choices and therefore belief biases: in particular, a preference to reflect on positive experiences unintentionally generates a positivity bias in future recall and beliefs. This mechanism provides a new non-strategic channel through which seemingly motivated beliefs arise and generates novel predictions in a range of economic domains.