The purpose of this paper is to outline a thesis which in some ways attemts to retrieve the value of meaning as a constituent element of theories. It will be argued that theories possess in common a more or less explicit relational structure, but, at the same time, they possess a meaningful function by means of which they are instruments of knowledge or, rather, by means of which theories have the value ofknowledge in science. In conclusion, it will be suggested (...) that this approach to theories allows those with weak relational structure (social, historical theories, etc.) to be treated with a degree of adequacy by the philosophy of science. Likewise, it will be suggested that the method of semantic analysis linked to the theory of “possible worlds” could be fruitful for this purpose. (shrink)