Among the many groups of scholars whose work now illuminates science, technology and medicine (STM), 1 historians, it seems to me, have a key responsibility not just to elucidate change but to establish and explain variety. One of the big pictures we need is a model of the varieties of STM over time; one which does not presume the timeless existence of disciplines, or the distinctions between science, technology and medicine; a model which is both synchronic and diachronic, and both cognitive and social. To that end, this brief paper presents a historical typology of STM from about 1700 to the present by focusing on four ' ideal' socio-cognitive types -four knowledge structures which correspond to four sets of social relations. To some extent these are period specific, but they do not have to be -hence, one may hope, the flexibility and usefulness of the model.In the quest for models, the historiography of medicine proves particularly helpful. First, because of the breadth of the term 'medicine', which spans laboratories and state-policies, folk medicine and multi-national pharmaceutical companies. Secondly, because modern historiography of medicine, from its foundations in the 1930s, has been more closely related to social history than has most historiography of science. Thirdly, because of the French connection: from Georges Canguilhem to Michel Foucault and to Bruno Latour, the most notable of recent French analysts have been particularly concerned with biological sciences and medicine. Indeed the formulation with which I here begin, published in 1976 by Nicholas Jewson, was partly French in its inspiration.Historians of medicine will doubtless know Jewson's paper on 'The disappearance of the sick man' and perhaps his more detailed paper on patronage in eighteenth-century