I’ve just finished reading Michael Smith’s The Moral Problem, and I have an issue with his account of the fully rational agent (pp. 155-161). He borrows, and then amends, Bernard Williams’s account. On this account, the fully rational agent must meet the following three conditions:
(i) No false beliefs
(ii) All relevant true beliefs
(iii) Deliberate correctly
Smith takes some time to spell out what he has in mind by (iii). But I want to take issue with (i) and (ii). My question is simple: why think that rationality – even full rationality – requires omniscience, or even omniscience about the relevant matters? Some prima facie evidence against this account would be a case like the following:
Misleading evidence. Phife Dawg is supposed to meet Q-Tip at the studio at 10:00 to record a new album. Q-Tip never misses a recording session, and is always on time. As Phife steps into the elevator up to the studio at 10:05, Jarobi, who never lies, steps out and tells him that Q-Tip is upstairs waiting for him. When Phife gets to the door of the studio, he hears the distinctive beep of Q-Tip’s Skypager, which Q-Tip never lets out of his sight. This is all of the relevant evidence that Phife has, or could be expected to have. Phife then finally forms the belief that Q-Tip is in the studio. But as it turns out, all of this evidence is misleading. Q-Tip is not in the studio; he’s at home asleep.
So Phife has a false belief. Thus, by the account of full rationality – in particular, condition (i) – Phife is not fully rational, because he has a false belief. Though we might agree that Phife is not fully rational for some reason, I think it is implausible to think that he is less than fully rational in virtue of his false belief. In fact, forming this false belief seems to be the most rational belief Phife could form, given his evidence. So (i) is implausible. Rationality does not require one to have only true beliefs.
If Q-Tip is in the studio, then Phife should go in to record. If Q-Tip is not in the studio, then Phife should go look for him. So the belief Phife forms about Q-Tip’s location is practically relevant. So Phife lacks a relevant true belief – the belief that Q-Tip is not in the studio. Thus, by (ii), Phife is not fully rational. Though we might agree that Phife is not fully rational for some reason, I think it is implausible to think that he is less than fully rational in virtue of his lack of a relevant true belief. In fact, forming the belief that Q-Tip is not in the studio would seem to be irrational, given Phife’s evidence. So (ii) is implausible. Rationality does not require omniscience, even about relevant matters.
These problems might not seem terribly important, but I think that they actually cause a bigger problem for Smith’s analysis of normative reasons. That analysis is the following:
Reason: S has a normative reason to A in circumstances C iff S’s fully rational self would desire that S A in C.
Smith wants to use (i) and (ii) to block reason-ascriptions in cases like Williams’s gin/gasoline case. In this case, S has a glass in front of him which he believes (for good reason) to contain gin. But in fact, it contains gasoline. So, Smith wants to say (I think), S has no normative reason to drink the stuff in the glass. This is supposed to be the result delivered by the analysis, since S’s fully rational self would not desire that S drink the stuff in the glass, since he would know that it contained gasoline instead of gin (by (i) and (ii)). But if I’m right that (i) and (ii) are implausible, then Reason will not obviously deliver the verdict that Smith wants in this case. (Note: I disagree with Smith that S has no normative reason to drink what’s in the glass, but that’s another issue.)
Disclaimer: I haven’t done a literature search, so I don’t know if I’m the first person to bring any of this up (I doubt it).