Rationality is central to modern biomedicine: not only are doctors expected to diagnose and prescribe rationally according to evidence-based guidelines, but patients are increasingly expected to behave in rational ways in their encounters with medical practitioners. Patients are offered guidance on how to ask rational questions of their doctors, and how to make reasoned decisions about what treatment options to pursue. Here, we examine the operation of such rationalities and their ‘others’ in two kinds of clinical encounter: those around methadone maintenance treatment, and those around hormone replacement therapy. In these contrasting but related cases, we argue, clients/patients struggle to perform the kinds of rationality figured in contemporary biomedical discourses. The historical freight of non-rationality attached to the medical conditions being treated, namely ‘addiction’ and the menopause, means that clients/patients are enacted as intrinsically non-rational, which renders suspect their capacity to make the kinds of reasoned decisions that contemporary patients are expected to make. Engaging with the work of cultural studies theorist, Eve Sedgwick, we suggest that such enactments can be understood as a form of habit, produced within the complex social space of the clinical encounter. It is only through taking seriously the habits of rationality and non-rationality in which particular kinds of patients are enacted that we can begin to understand the inevitable limitations of the rationalist model of clinical encounter presented as ‘ideal’ in contemporary health discourses.