Peer review is a ubiquitous feature of science and fulfills three interrelated roles. Firstly, it is a mechanism used to assess quality based on expert judgement (process). Secondly, it is a decision mechanism used to distribute scarce resources, such as publication space, funding, or employment (outcome). And thirdly, it is an instrument for self-governance in science (context). This is poorly reflected in public debates and, more importantly, in theoretical conceptions informing research about peer review. To move beyond such a “deficit model,” we provide two preliminary considerations that lay the foundation for a more encompassing theory of peer review. First, the peer-review process can be divided into at least eight different practices, which can in turn comprise a wide variety of specific peer-review procedures when combined. Second, peer review not only provides evaluative decisions, but, more importantly, also provides the legitimacy for these decisions. Thus, an encompassing view of peer review should integrate process, outcome, and context. We argue that such a view could start by theorizing peer review as a form of government, not unlike democratic government, which has grown historically around concerns for legibility, responsibility, and responsiveness (Rosanvallon) in a similar way to the Mertonian norms.