Knowing *about* what it is like

Thanks in large part to my recent email correspondence with L. A. Paul (see here and here for some background on that), I've formulated a position which, as far as I know, is novel in the philosophy of mind.  The basic assumption is this:

(1) One knows *about* what it is like to X iff one knows some of the consequences and/or functional correlations associated with the experience of what it is like to X.  

It follows that, 

(2) one can know about what it is like to X without ever directly experiencing X.  

This only requires that phenomenal properties have functional correlations (these can be statistical or accidental, and not intrinsic, of course) and/or that they are not epiphenomenal.  This is consistent with the common intuition that you can only know what it is like to X if you have some experience of X-ing.  (That is the new-knowledge intuition associated with Frank Jackson's Mary thought experiment:  When Mary the color-blind scientist finally sees color for the first time, she will gain new knowledge by virtue of her new experience, regardless of any knowledge she learned ahead of time.)  

A further implication of my assumption is this:  

(3) One can make rational decisions about what it would be like to X without knowing what it would be like to X.  

For example, imagine that before Mary is about the leave her black-and-white room, she has to make a decision: How does she want to first experience colors?  There is a color chamber designed to ease her transition from black-and-white to the color-filled world.  Studies have shown that people like Mary tend to respond well when the transition is slow, moving from soft, dull tones up to brighter hues.  Mary does not know what it is like to see color (assuming the new-knowledge intuition is correct), but she can know about how these different processes are likely to affect her.  She might not know what it is like to be relaxed by a soft green, but she knows what it is like to be relaxed and she knows that soft green is likely to have that effect on her.  In short, she can rationally decide how she wants to first experience colors without trying to guess at the phenomenal contents themselves.

This seems to pose a direct challenge to Professor Paul's argument that the common way to decide whether or not to have children is irrational.  Her argument is that people make assumptions about what it would be like to have children, but since they can't possibly know what it would be like, they are fundamentally wrong about the way they are making the decision.  If the position I have outlined is correct, then they might not be making a mistake.  They might have reasonable, justified and even true beliefs *about* what it would be like without knowing what it would be like to have children . Their decision may therefore be rational.

I am not sure if Paul will accept (1).  If she does, she might still respond that people making The Decision really are under the false impression that they know what it would be like.  In her view, they're not just relying on beliefs *about* what it would be like.  She might even say it is "bizarre" to approach the decision by focusing on knowledge *about* what it would be like.  I don't see anything bizarre about it, though.  I think it is simpler and more intuitively appealing to think that people are not making the mistake Paul says they are making.  I think people commonly go about such decisions much the way Mary would when deciding how to experience colors for the first time.  

This question could perhaps be answered with better sociological data, but that may not be so easy, since we'd need a way of clearly distinguishing between two sorts of assumptions:  on the one hand, assumptions which purport to involve the phenomenal content in question and, on the other hand, assumptions about that phenomenal content.  More importantly, the data that is available, and which Paul draws on in her paper, seems to support my view at least as well as, and I think even more than, hers.

Incidentally, some philosophers--especially those Pete Mandik calls "non-gappy physicalists," like he and Dan Dennett--might say that there is no difference here at all, and that "knowing what it is like" just is knowledge of the correlations and consequences that I am classifying as knowledge about what it would be like.  I am sympathetic with that view (and non-gappy physicalism), but I'm not 100 percent committed to it.  The difficulty in finding data distinguishing these two cases might lend it intuitive support, though.