Recent years have seen growing sociological interest in the role that objects and non-human actors perform in everyday life. Whether as machines, information technologies, artworks, commodities or architectures, objects today raise issues of complexity and controversy (Pels et al., 2002). Borrowing from actornetwork theory the idea that humans and nonhumans are actively involved in the making of social worlds, there are already those who call for a post-social world and an object-centred sociality (Knorr-Cetina, 1997). But how can non-humans be observed? Sociologists are accustomed to socio-constructionist approaches to the sociology of science, or to analyses of tools and innovations couched in terms of networks of actants; methodologically, however, it seems that ideas about how to proceed methodologically are not very well worked out. On the basis of a four-month ethnography conducted in a hospital that has recently introduced a digital clinical records system, 1 discuss the methodological aspects of shadowing non-humans. In particular, adopting Star's insight of an `ethnography of the infrastructure' (Star, 1999), 1 concentrate on how to account for contexts characterized by multiple and non-homogeneous actors and practices and on the implications of such a perspective for organizational analysis.
