Lewis and the Hidden Nature of Reality
One of my problems with Lewis is that he published so little on issues where he thought he had nothing new to say. Sometimes it's tricky to figure out what his views on these issues might have been. Knowing people who knew him personally, or having access to some of his communications would probably help. Have there already been efforts to collect his letters, or even to make some of his unpublished writings available somehow? (If this is really Lewis' computer, the data on it definitely should be backed up soon before it completely turns to dust...)
Here is an example of the problem: In a Memorial Resolution for Lewis, Mark Johnston writes:
In the philosophy of science, he was a structural realist, maintaining that science at its best provides an accurate account of a mind-independent world, though at the inevitable cost of revealing little of the world's intrinsic nature.
Maybe I've just missed it, but I don't remember Lewis ever saying anything about our inability to get at the world's intrinsic nature. Of course, it can been argued that his conception of theoretical terms commits him to some such "noumenalism", but the same charge can be made against virtually any position short of idealism and phenomenalism. (See, e.g., Langton, Kantian Humility, §8.6; Blackburn, "Filling in Space"; and the articles by Smith and Stoljar and Pettit in The Monist, Jan 1998.)
On the other hand, it is not clear to me that these arguments for noumenalism in Lewis' philosophy are very convincing. Couldn't he reply that in fact we do know the intrinsic nature of things, albeit only in a slightly indirect way: For example, we know that certain particles instantiate the fundamental, intrinsic property that realises the up-spin role in our world. Perhaps this is all we could reasonably ask for. Perhaps we're not missing out on the "real nature" of that property here. After all, even though it is quite plausible that things have an intrinsic nature, it is much less clear that properties do (even less so if properties are just classes). Hence any lamenting about our supposed lack of knowledge about the real nature of properties might be simply confused.
I think there is something wrong with this reply, and I also think that noumenalism isn't the only problem for Lewis in this area (for example, there is also the problem for physicalism I mentioned in April). Unfortunately, I don't know what he himself thought about all this.
[Update 2003-08-28: Dave Chalmers and Brian Weatherson both let me know that there's a forthcoming paper by Lewis called "Ramseyan Humility" dealing with these issues.]