H-J Rheinberger's book Toward a History of Epistemic Things contains a sophisticated account of scientific reference and scientific method worked out in conjunction with a case study of the laboratory synthesis of proteins. This paper offers a detailed critical analysis of Rheinberger's position from the standpoint of the sociology of scientific knowledge. The central thesis is that Rheinberger's account of reference, whether deliberately or unwittingly, assimilates discourse about the natural world to discourse about the social world. The result is an inadequate account of scientific terms, which does not do justice to the independent character of the objects of scientific knowledge. A further feature of Rheinberger's approach is a commitment to an extreme form of methodological pluralism. This is challenged in the paper on the basis of a sociological reading of Carnap's famous identification of a "continuum of inductive methods."