Bernstein: 'Reading Cavell Reading Wittgenstein'
'The distortion is to imagine that knowledge has an "object" outside of the "language games" of which it is a part—that words refer to "transcendental signified" to use an expression from another tradition, rather than being part of a language which itself produces meaning in terms of its grammar, its conventions and its "agreements in judgement". Learning a language is not learning the names of things outside a language, as if it were simply a matter of matching up signifiers with signifieds, as if signifieds already existed and we were just learning new names for them [...] Rather we are initiated by language into a socious, which is for us the world. So that the foundations of knowledge are not so much based on a pre–existing empirical world as on shared conventions and mutual attunement.' (299)
In Boundary 2 (Vol IX, No. 2, 1981) 295–306.