Social epistemology

5. The social construction of knowledge

There are three senses in which it has been claimed that knowledge is socially constructed (see Constructivism). (1) Some sociologists of science (for example, Bruno Latour 1987) have proposed that the facts reported in scientific journals are fabricated by the manipulation of inscriptions involving ‘modalities’ (that is, ways of embedding sentences in other sentences). It has also been proposed (Joseph Rouse 1987) that many scientific laws strictly apply only to artificially created laboratory phenomena and extend to phenomena in nature only by analogy; thus, a law implicitly refers to the social conditions in which laboratory phenomena are created. (2) It has been proposed (again by Rouse) that the conditions of scientific communal knowledge are social. The argument is that the standards of knowledge to which scientific theories and experimental results are held (for example, the degree of confirmation required) depend on the uses to which the theories are put, and these uses are in turn determined by research needs, hence by the social conditions that determine those needs. (3) Finally, it has been claimed, particularly by those associated with the ‘strong programme’ in the sociology of science, that scientific theoretical and observational communal knowledge, or at least beliefs that are esteemed knowledge, are caused by social, economic and political interests. This sociological causal claim contrasts with rationalism about the causes of scientific theory choice, according to which theory choices are caused by rational reasoning on the basis of observations and theoretical desiderata, except where noncognitive factors (such as desires, emotions, peer pressure and the like) interfere with rational choices and pre-empt them.

One argument for the sociological causal claim is a priori: theories are underdetermined by observations and so choice is not caused by observations alone but by social interests (see Underdetermination). Nor can rationalists appeal to theoretical desiderata like simplicity and explanatory power to explain theory choice in the face of underdetermination, since these desiderata are merely rationalizations of theory choice, not rationally supported independently of choice. This objection to rationalism underestimates the theory-independent grounds for desiderata like simplicity – such grounds as that simpler theories carry less information than complex theories and thus accepting them entails less exposure to the risk of error (see Theoretical (epistemic) virtues). These grounds for deciding between theories are sometimes said to be theory-laden, leaving the choice of theory once again unexplained. But even if this is so, the way in which these grounds are selected may depend on metaprinciples relating grounds and theories.

Sociologists have also supported the sociological causal claim by appeal to historical case studies of the role of interests in theory choice (see Gender and science §3; Marxist philosophy of science; Postcolonial philosophy of science; Sociology of knowledge). But the poverty of the historical materials makes it difficult to establish that interests cause, and are not merely correlated with, theory choice in these cases. However, a correlation between interests and theory choice, even if not shown to be causal, may be hard to reconcile with rationalism, since we would not expect interests to be correlated with rational choices.