EPM Parts II-III

EPM Parts II-III

Part II (in a nutshell): So why don’t we think of talk about "sense data" as a philosopher’s construction (Sellars says "code"), meant to clarify (or to make explicit) logical connections that are already implicit in "looks" talk:

Something looks red to me. ç è I’m being presented with red sense data.

Sellars here cautions us that the translation into "sense data" talk appears to have some logical implications that are not clearly present in the "looks" talk. Namely, it appears to license the inference that there is something (the sense data) that is red. Unless one is already in the grip of something like a sense-datum theory, one does not usually think that such an existence claim could appropriately be inferred from the left hand side of this translation. If so, then the two claims shouldn’t truly be considered be intertranslatable; the right-hand side expands (not clarifies) the meaning of the left-hand side.

Sellars also mentions that once one invokes the language of sense data, it is nearly irresistible to treat such posits as theoretical entities in part of a quasi-scientific theory of perception, and thus, as possessing determinate properties of their own (e.g., specific shades of pinkness).

Part III: Consider: (1) "I see that the book is green."

(2) "No? Well at least it looks green to me"

1. Does (2) in any way justify (1)? The Cartesian foundationalist thinks so. (2) reports a separate fact from (1). (2) is a report, not about an external object, but rather a report about an APPEARANCE, about which we cannot be mistaken. So by merely having appearances, we come to have a certain bit of KNOWLEDGE, which we can then use to justify the rest of our beliefs.

<Note, by the way, that there is no mention of sense data in the previous paragraph. Sellars is beginning to expand his case to other invocations of the given.>

The Cartesian picture makes it sounds as if "looks"-talk is conceptually prior to "is"-talk, as suggested by the following analysis:

X is red ç è X looks red under normal conditions.

If so, then "X looks P to S" would seem to be reporting an independently intelligible fact involving X, P (or Pness), and S, and the (3-place) relation of looking. <or at least a 2-place relation of looking-P>

3. But, according to Sellars, this seems to get the true conceptual priority backwards. There is liitle to be gained by conceiving looks as some sort of independent relation. Enter the parable of John, which tries to illustrate the plausibility that "looks P" talk really is grasped after one has an antecedent understanding of the notion of something’s actually being P.

4. Sellars’ analysis of looks: X looks P to S does not report a minimal (and independently intelligible) fact about either the world or one’s own experience. Rather, the function of "looks" (or "seems") is to withdraw one’s natural inclination to report that "X is P" in the face of contravening experience (or suboptimal conditions).

Virtues of this analysis:

Recall that one really ought to think of the "red" involved in a seeing that something is red as the same "red" in a looking that red. This analysis allows that to be the case, without having to postulate some weird sort of redness actually present in an episode of mere looking red.

This explains why "looks" talk seems so incorrigible. Insofar as it is actually a withdrawal of a claim, then it isn’t really an assertion of much of anything, and so one cannot be mistaken when one uses it.

It can handle "generic" lookings: "It looks many-sided to me."/ "Well, how many, exactly?" If one is viewing some sort of particular, then it ought to have some determinate number of sided, shouldn’t it?

Can accommodate various degrees of endorsement (Brandom, p. 144):

Compare: "I see that X is P."

"It looks to me as if X is P." (X looks P to me.")

"It looks to me as if there’s a P X over there."

Notice the progressive canceling of existential and qualitative commitments.

5. If Sellars is right, then one can still say that the analysis in 2. above is essentially correct. However, that analysis turns out, not to be part of an analysis of red, so much as an analysis of "normal conditions" as those under which things really are as they seem.