Abstract
This article examines the strategies by which the different and variable signs of failing mental powers become known sufficiently for 'dementia' to be made into a stable bio-clinical entity, that can be tested, diagnosed and perhaps one day even treated. Drawing on data from ethnographic observations in memory clinics, together with interviews with associated scientists and clinicians, we document the challenges that clinicians face across the clinical and research domain in making dementia a stable object of their investigation. We illustrate how the pressure for early diagnoses of dementia creates tensions between the scientific representations of early dementia and its diagnosis in the clinic. Our aim is to highlight the extent to which the work of diagnosing dementia involves an intricate process of smoothing out seemingly insurmountable problems, such as the notoriously elusive connections between brain/mind and body/person. Furthermore, we show that a part of this process involves enrolling patients as minded, agentic subjects, the very subjects who are excluded from dementia science research in pursuit of biomarkers for the pre-clinical detection of dementia.