The biggest mistake that both philosophers and social scientists make when trying to explain a stable knowledge-producing activity such as science is to assume that the task requires that one explain science-as-successful or knowledgeas-true. Unlike explanations of, say, the capitalist economy, few explanations of science imply criticism of science as it normally is. Rather, a broadly functionalist explanatory logic is invoked to infer that science works as well as could be expected. (See truth, reliability and the ends of knowledge.) As these explanations have come to incorporate more of the actual history of science, the tendency – especially among epistemologists and philosophers of science – is to pursue invisible-hand strategies whereby even prima facie suboptimal day-to-day features of science, such as its ruthlessness, its drudgery and elitism, are said to be necessary for its overarching good features. Included here are the most recent generation of naturalistic (see naturalism) epistemology (e.g. David Hull, Philip Kitcher), as well as sociologists and economists who point to knowledge as a uniquely legitimizing or productive element in the modern world. (See knowledge society.)

Before turning to this functionalist explanatory logic – whose natural home is Talcott Parsons's social systems theory – three critical responses are worth noting. The first simply reinforces the boundary between how science is and ought to be, arguing that most science performs suboptimally most of the time.