I argue that ''social epistemology'' can be usefully reformulated as a philosophy of science accounting, specifically one that fosters a critical form of instrumental rationality. I begin by observing that philosophical and sociological species of ''science accountancy'' can be compared along two dimensions: constructive versus deconstructive; reflexive versus unreflexive. The social epistemologist proposes a constructive and reflexive accounting for science. This possibility has been obscured, probably because of the persuasiveness of the Frankfurt School's portrayal of ''critical'' and ''instrumental'' rationalities as polar opposites. In challenging this polarity, I present four ''accountability conditions'' that define pure critico-instrumental rationality. These conditions are most likely to be met during a period of ''crisis'' in the future direction of science. The bulk of the paper concerns a case study of such a crisis that enables us to examine contrasting ways of meeting the four conditions. Ernst Mach and Max Planck debated ''the ends of science'' in the first decade of this century, a time marked by both science's increasing involvement in the means of social reproduction and growing concern that science pursued for its own sake was exhibiting diminishing marginal returns on investment. In analyzing Planck's victory, I conclude that its legacy has been an ever-widening gap between the ''content'' and ''function'' of scientific knowledge in society, as measured by the increasing numbers of jobs that are devoted to mediating the production and distribution of knowledge.
