ABSTRACT
The most difficult task currently facing historians of the natural sciences is to find an adequate way of analysing the relation between scientific knowledge and the social environments in which it has been - and still is - constructed. Mary Douglas's theory of grid/group analysis offers a promising heuristic tool for this purpose. It suggests features of both cosmologies and social environments that may be expected to be found in regular conjunction with each other. Earlier statements of the theory (Douglas 1970, 1975) were in some respects confusing and the terminology was inconsistent; but a coherent formulation can be derived from more recent descriptions (Douglas 1978), enabling the theory to be applied to, and tested on, the cosmologies of scientific knowledge.
