According to a common view, the difference between the “right” kind of reasons that support the distinctive rationality of belief, intention, or other attitudes, and the “wrong” kind of reasons that do not, is that the former are “object-given” reasons while the latter are “state-given” reasons. As I shall argue here, this view is false: it is open to some simple counterexamples.
In this post, I shall explain why the reason that explains why it is irrational to believe Moore-paradoxical propositions (like the proposition that you might express by uttering a first-person present-tensed sentence of the form ‘p and I don’t believe that p’) is a state-given reason, even though it is a reason of “the right kind”. (In a later post, I shall explain why our reasons not to have intentions that would frustrate their own realization are similarly “state-given”.)
Derek Parfit distinguishes between “object-given” and “state-given” reasons as follows (“Rationality and Reasons”, pp. 21f.)
Of our reasons to have some desire, some are provided by facts about this desire’s object. These reasons we can call object-given. … Other reasons to want something are provided by facts, not about what we want, but about our having this desire. These reasons we can call state-given.
This distinction appeals to a non-standard notion of a fact’s being “about” something. In the notorious “strike of the demon” case (discussed by Rabinowicz and Rønnow‐Rasmussen), surely it is a fact “about the demon” that if you admire the demon, then your admiration of him will save the world? In general, surely any fact that cannot be stated without referring to x is a fact “about” x.
Let us take a different approach, then, by using the standard notion of a fact’s being “about” something. Then we can give the following definitions:
“State-given reasons” for you to have attitude A towards object o are provided by facts about how things would be if you had attitude A towards object o; “object-given reasons” for you to have this attitude A towards o are provided by facts about o that are constitutively independent of all such facts about how things would be if you had this attitude A towards o.
Now, consider a Moore-paradoxical proposition – that is, a proposition of the sort that you might express by uttering a first-person present-tensed sentence of the form ‘p and I don’t believe p’. It is clear that it cannot rational for you to believe it. But why is it irrational? The reason that explains why it is irrational to believe this proposition will clearly be a reason of the “right” kind. But what is this reason?
The reason is clearly not that there is anything problematic about the proposition itself. The proposition that you might express by saying ‘p and I don’t believe that p’ might be true; and it might be strongly supported by your evidence. (Indeed, it might have arbitrarily high probability given your evidence. Imagine, for example, that you have just acquired irrefutable evidence that p, and a seemingly infallible oracle has told you that you would never believe that p.) So it is hard to see how there can be any object-given reasons not to believe this proposition.
The reason why you should not believe this proposition does not depend on any features of the proposition itself; it depends on what things would be like if you believed the proposition. Specifically, if you were to believe the proposition, then the proposition would not be true – even if in fact (given that you don't believe the proposition) it is true. (Indeed, if you believed the proposition, then you would be in a position to see that it is not true.) But this is a fact about what things would be like if you had this attitude towards this proposition; it is not a fact about the proposition itself that is independent of facts of that kind. Thus, this fact is a state-given reason not to believe it.
But it seems to be this reason that makes it the case that I am rationally required not to hold this belief. So, some facts about what it is rational to believe depend on state-given reasons. Not all reasons of the “right” kind are object-given reasons.