Pain serves as a signal to elicit care from others. In turn, displaying pain might be attractive because of the benefits it might bring. Additionally, displaying pain is easy, because helpers distinguish poorly between genuine pain and faked pain. Hence, helpers face the problem of distinguishing true sufferers from free riders, while sufferers face the problem of communicating need convincingly. This article will propose solutions to these adaptive problems. Based on theoretical arguments and on empirical insights from lie detection research, it will be argued that the credibility of pain signals cannot be found in features of the signal itself, but in its context. Namely, pain is obviously credible when the context features unforgeable cues, such as an open wound or the enlarged abdomen of a pregnant woman, but also external cues such as the ice water in cold pressor tasks. In absence of such cues, pain can become credible through costly consequences, such as refraining from rewarding behaviors for a significant period. However, these adaptive mechanisms for communicating need may not be shaped for modern circumstances such as experimental settings and therapeutic encounters.