Note: Argument updated!
Note 2: For the RCist, "concept" means "rational concept," throughout, in order to distinguish epistemic from "poetic" or "mystical" considerations.
Note 3: "concept" can without loss of meaning be replaced with "experience."
Traditional "metaphysical realist" theories of knowledge (MR) assume that truth is a correspondence relation: When a concept (C) corresponds to the way the world (W) is, it is true (and false otherwise).
C ---corresponds ---->W = true
C---fails to correspond--->W = false
In order for me to know whether any given concept corresponds to some aspect of the world, and assuming that I have a clear understanding of correspondence itself, it seems that I need independent access to two things: (1) the concept and (2) the relevant portion of the world. Only then will I be in a position to judge whether the two correspond. Obviously, I have access to my own concepts through simple reflection. Do I also have independent, nonconceptual access to the world? (It must be nonconceptual, otherwise I will be simply be comparing concepts -- an activity more in line with the coherence theory of truth). Here's my reconstruction of the antirealist's argument:
1. All concepts are in the mind
2. All concepts are conceptualized
3. All concepts are concepts-of-some-x
4. All concepts-of-some-x are in the mind (from 1)
5. There can be no unconceptualized concepts-of-some-x (from 2)
6. All concepts of apples are concepts-of-some-x (from 3)
7. There can be no unconceptualized concepts of apples (from 5 and 6)
8. There can be no concept of an apple that is not a concept (truism)
9. An “unconceptualized apple” is not a concept
10. There can be no concepts of unconceptualized apples (from 8 and 9) (Ernst von G.: “…we cannot possibly [rationally, non-poetically] conceive of an unexperienced world” p. 1, Introduction to RC)
11. Metaphysical realism (MR) involves concepts of unconceptualized apples (by definition)
12. MR is wrong (from 10 and 11)