The theoretical framework of active inference proposed by Karl Friston is currently one of the more actively developed research areas in neuroscience. According to this theory the brain adapts its synaptic activity and architecture in such a way that it de facto comes to mirror the causal structure of events that the organism both encounters and actively induces in its environment. In this contribution we show how active inference can provide a useful perspective to better understand the processes engaged by contemplative practices. More specifically, and focusing on the practice of shikantaza (‘just sitting’) in the Japanese Zen Sōtō tradition, we argue that meditation enacts a peculiar policy with high endogenous epistemic value, whereby the practitioner accrues an intimate, but not necessarily explicit, knowledge about herself. This superordinate policy entails the embodied, active suspension of our habitual reward-seeking and punishment-avoidance behavior, an attitude epitomized by the traditional notion of mushotoku (Jap. ‘nothing to be attained’). From this perspective, we also critically examine some popular claims about meditation that we believe are likely to be misconstrued, and may not be without personal, social and even political consequences.