The question of how dreams can offer personal guidance and solutions has ignited the interest of humankind for thousands of years. Nonetheless, dreams have often been considered epistemically irrelevant or even deceptive, becoming the target of sceptical debates. Burgeoning research on the function of dreaming for memory integration as well as the development of dream engineering techniques aimed at harnessing dreams for creativity and learning call for a reassessment of the role of dreaming in epistemology. This work addresses unresolved questions concerning the potential use of sleep experiences for knowledge generation. By focusing on the concept of insight and how it is employed in the scientific literature on dreaming to implicitly uphold epistemic goals, I argue that at least a subset of dream uses, when adequately constrained, disclose opportunities for epistemic exploration and expansion. I will proceed by examining three cases and the conditions under which dreams can give rise to knowledge: lucid dreaming, creative ideation at sleep onset, and waking insight following dream discussion. My goal is to show how empirical work on dream consciousness carries tacit assumptions that have far-reaching implications for the epistemology of dreaming and the use of dreams for therapeutic purposes, thus warranting conceptual analysis.
