The main objective of this paper is to challenge one of the notions which supports that psychoanalysis is not science, i.e., that experimental evidence holds intrinsic, absolute superiority over clinically derived evidence. I review experimental evidence supporting both the efficacy of psychoanalytic treatment (e.g. meta-analyses and randomly controlled trials) and the plausibility of its psychic mechanisms, including their bidirectional interplay with the nervous system. I argue that psychoanalysts clinically engage the unconscious mind by treating patients' associations, affects, relational patterns, dreams, speech, and silences within the analytic dyad as inferential evidence of these mechanisms that are conceived in clinical-theoretical models. To highlight the contextual fragility of any kind of evidence, I draw a parallel to the precariousness of experimental data on pharmacological mechanisms (mechanisms of change) in medicine—data that often lacks the technology for full elucidation and hinges on physicians' limited training in drug actions. I also discuss some ideas about epistemology. A clinical vignette from psychoanalytic practice illustrates this argument.