Alston identifies four “looks-concepts” in Sellars and the Myth of the Given. Three he borrows from Chisholm: the comparative and non-comparative concepts, and the epistemic concept. To these Alston adds what he terms a doxastic concept. The epistemic concept can be glossed as “x looks in the way that would prima facie justify the perceiver in taking x to have a certain feature,” and the doxastic concept can be glossed as “x looks in the way that would ordinarily dispose the perceiver to take x to have a certain feature.”
I’m not sure there are four concepts here, but I won’t quibble with the metaphysical issues now. What I find interesting about Alston’s distinctions is the way they clarify certain issues in the theory of perception.
For example, Alston addresses the nature of the noncomparative concept, what he terms the phenomenal concept, in terms of the theory of appearing. But nothing about the theory of appearing addresses the doxastic or epistemic concepts directly, and here Alston employs the concept of a ground of belief, interpreted reliabilistically. On the reliabilist understanding, the doxastic concept can be interpreted in a naturalistic fashion, explicable by appeal to the laws of nature (or whatever appropriate basis our theory of explanation will posit) connecting perception and belief. Then reliabilism comes into play when explicating the epistemic concept: the naturalistic connection is adequate epistemically only when it is appropriately reliable (though there are other necessary conditions as well for Alston).
Two things are noteworthy here. First, the account of the phenomenal concept does not imply the account of the epistemic concept. Second, on Alston’s account of the doxastic and epistemic concepts, it really doesn’t matter much which account of the phenomenal concept is given.
It’s the second aspect that I find especially interesting in thinking about Sellars. I think Sellars was committed to the view that the epistemic concept had to be understood in terms of something having conceptual or propositional content, but he confused this claim with the claim that the phenomenal concept itself would have to be given an account in terms of its conceptual or propositional content (in order for appearance states to play a role in the theory of justification). He confused these two claims because he assumed, it appears, that an account of the epistemic concept had to derive from or depend in some way on an account of the phenomenal concept. (The rest is well-known: his account of the phenomenal concept appealed to the noncomparative concept, thereby rendering it conceptual or propositional.)
What is intriguing here is the category scheme that results from thinking about the connections between Alston’s four concepts. Sellars uses the noncomparative concept as fundamental for the phenomenal concept. He leaves the doxastic and epistemic concepts unexplained, assuming, I expect, a scientific account of the former, and leaving the epistemic concept a mystery. Alston, as already noted, accounts for the doxastic and epistemic concepts in relation to each other, and addresses the phenomenal concept in an unrelated way.
Is the lack of connection here between accounts of the four concepts a problem? It’s hard to say, but if it isn’t, it opens up a nice opportunity for coherentism. In the past, Wayne Riggs, another of our contributors, and I argued that coherentism can appeal to perceptual states or appearance states without abandoning coherentism. Our argument required that perceptual states have content of some sort. Such a requirement assumes that one’s account of the phenomenal concept must fit together in some way with one’s account of the epistemic concept. If Alston’s approach in epistemology can be adequate, that assumption is false. All we would have needed to require is that perceptual states play a role in the account of the epistemic concept to the extent that they have propositional content. That claim is compatible with the claim that a proper account of the phenomenal concept itself will not appeal to any kind of propositional content.