A while back I came across Ernest Gellner's Words and Things (unrevised ed., 1963). It is jam-packed with insights. Here is an example:
Linguistic Philosophy [O. L. philosophy] absolutely requires and presupposes [Logical] Positivism, for without it as a tacit premiss, there is nothing to exclude any metaphysical interpretation of the usages that are to be found, and allegedly "taken as they are," in the world. (p. 86)
Exactly right. For if the anti-metaphysics of logical positivism is not presupposed, how can the O.L. philosopher rule out as meaningless metaphysical ways of talking? People talk in all sorts of ways, not all of them mundane. People talk metaphysics for example. I do it all the time, and it certainly seems to me and some of my interlocutors that I am talking sense. For example, I say things like, 'Existence is a necessary condition of property-possession: nothing has properties unless it exists' and there are people who understand me.
Suppose out in the desert there is a little commune of Bradleyans. Their form of life involves playing a language game in which words like 'internal relation,' 'external relation,' 'Absolute,' 'appearance,' and others have well-defined functions. If meaning is use, these words have meaning because they have a use in this community. How can it be said that language has gone on holiday in a case like this? How distinguish holiday from workaday uses of language?
To make the distinction one has to just assume something like the positivistic stricture on metaphysics. On has to just assume that some language games are meaningless. But there is no basis for this distinction if one takes the uses of words as the source of their meaning.