On one view of language use, language is principally a means of communication. A speaker uses language principally to communicate the contents of his mental states, and especially the contents of his beliefs. One proponent of this view is D. H. Mellor (1990).
Mellor prefers to practise philosophy than to preach about how to do it:
In the sense in which astronomers are interested not in astronomy but in the stars, I am interested not in philosophy but in the various philosophical topics dealt with in this book [Mellor 1991a] - topics on which I find discussions of what philosophy is and how to do it shed very little light. I think the proof of our methods lies rather in the results of our applying them, and my case for my method, such as it is, rests on the contents of the ensuing chapters.
(Mellor 1991a: xv)
Distinguish two questions. (Q1) What is Mellor's method? (Q2) Is it a good method? Mellor answers (Q2) by inviting us to assess the results of the application of his method. These results are his theories of various philosophical topics, and the degree of their success. Fair enough, but this does not answer (Q1). If Mellor's method has good results, it is a good method. But we still need to know what his method is. So (Q1) deserves an answer. Here is an outline of such an answer. In making explicit what is largely implicit, the following occasionally goes beyond what Mellor has said in print. 1
Take a monadic term 'F'. (The following outline carries over to polyadic terms.) A concept C is suitably associated with 'F' so that C provides the meaning of 'F'. Given what 'F' means (the universal closure of), the open sentence 'Fx' entails certain propositions. These are the connotations of 'F'. They constrain what F-ness, the property which 'F' expresses, is. Some connotations are informative, others not. Some are obvious, others not. Philosophical