This article presents a systematic review of the literature examining the relationship between self-talk and performance. "Second-generation questions" regarding potential mediators and moderators of the self-talk-performance relationship were also examined. A total of 47 studies were analyzed. Results indicated beneficial effects of positive, instructional, and motivational self-talk for performance. Somewhat surprisingly, two evidence-based challenges to popular current viewpoints on self-talk emerged. First, negative self-talk did not impede performance. Second, there was inconsistent evidence for the differential effects of instructional and motivational self-talk based on task characteristics. Results from the mediation-based analysis indicate that cognitive and behavioral factors had the most consistent relationships with self-talk. The findings are discussed in the context of recent theoretical advances, and the article includes recommendations for future research (e.g., the use of designs allowing the testing of meditational hypotheses) and for current applied practice (e.g., avoiding the use of thought-stopping techniques).