According to what we can callthe State-Given Theory: Whenever certain facts would make it better if we had some belief or desire, these facts give us a reason to have this belief or desire.
Since this is the basic claim about "state-given reasons" they are not provided with their normative force, if they have any, by reference to the object of desire. It is not the "object" of the desire that is at issue here but our state of having this desire in relation to external state's that would provide a conducive environment for this desire to be had. Parfit does, further, make clear in his brief discussion in Chapter 2, that there can be two kinds of "state-given desire", to wit, those states that are "telic and intrinsic" and those that are only instrumental. The first type relate our wanting an event to the event being in someway good in itself whilst the instrumental view relates it rather to the good presumed effects.
A classic case of assumption of a "state-given reason" is provided in Pascal's wager since it is here assumed that the state provided by the belief in God would be beneficial to us though the kind of benefit involved appears (at least on most renditions of it) only to be instrumental. Parfit describes the general claims of "state-given reasons" as value-based (or pragmatic). But Parfit's approach to these alleged reasons is to deny that they would have any importance if we had them. This is due to the way Parfit construes the alleged force of these reasons. In one example Parfit considers the value of weighing less which might, for example, lead one to be happier and healthier (hence would be instrumentally valuable). Assuming the truth of the claim of the value attached to the state in question we have a reason to set out to weigh less and act in such a way that it will become true that we do weigh less. But whilst we have been given reasons to act in these ways for these reasons it doesn't follow, for Parfit at least, that there is a value simply in the state of "weighing less" over and above the reasons for weighing less.
Initially at least Parfit's claim sounds odd since the value attached to the state in question appears to be precisely that attaining it will lead to the results in question so that there is a value attained if we reach the state in question (assuming the truth of the premises concerning the effect of being in this state). In order to dispel this obvious objection Parfit considers other cases. So, for example, it might be better if we were in a different state but not be able to be in this state so it would be better, for example, to be younger but one cannot cause oneself to become so. This doesn't prevent it from being the case that being younger is preferable even though there is no way the state in question can be attained. But this argument is, again, at least inconclusive as all it seems to establish is that some good things are good though unattainable for us and so not all that is good is something we can attain.
The failure of Parfit's arguments in Chapter 2 against the state-given theory seem to me to motivate his much more extended discussion in Appendix A. Here Parfit considers three possible cases that involve state-given reasons and asks for our reflection on how we would respond to them. The cases involve beliefs and aim to show that we don't have state-given reasons to adopt beliefs. The first case suggests that a Despot makes the threat that I shall be tortured unless I can make myself believe something that it is impossible to hold as true (Parfit gives the example of "2+2=1" but presumably any impossible statement would suffice).
In Case 1 there is a basis for saying that we have a reason to adopt a belief even though the belief in question requires one to accept something impossible and the "state-given theory" appears then to require us to adopt the belief in question. Since there are evident object-given reasons not to accept the belief and these reasons provide me with an understanding of what it means to relate beliefs to the world these beliefs will triumph over the state-given beliefs despite the pragmatic value attached to the latter.
Case 1 shows that there is something peculiar about analysing beliefs in relation to the "state-given theory" if this theory's test for the worthwhileness of a belief consists only in the attainment of a pragmatically valuable state. Notably it cuts only against instrumental forms of the theory, not against telic forms if the latter take other forms of purpose than the instrumental one into account.
Case 2 provides a simpler case since in this the Despot will torture me unless I believe that a closed box is empty. This case is easier since, unless I open the box and discover something within it, I have no countervailing pressure to adopting the belief. However, Parfit contends, this is not in itself sufficient for me to adopt the belief that the box is empty. Just as there is no pressure of an object-given kind to prevent me believing the box is empty, so, likewise, there is no pressure of an object-given kind to lead me to believe it is empty and in lieu of pressure from one side or the other I simply have no reason to have a view on the matter.
Of course, in Case 2, it may be that the box is empty and we could open it to verify this. But the risk of doing so is that it may not be empty and arrival at the state of belief would then be decisively checked. This point seems, however, to point clearly to the basis for object-given reasons for beliefs and prevent the state-given theory from taking hold. Again, the problem with the case, however, is that it relates the state-given theory to an instrumental outcome and doesn't test whether there could not exist states that were good in themselves and about which we could have beliefs asserting the essential goodness of the state.
Case 3 involves another despotic threat to the effect that the painful outcome will arise unless I adopt the belief that the despot is a genius. Hypnotism is now mentioned as a possible way the adoption of this belief could be combined with providing a check on object-given reasons that would otherwise count against the state having the right effects. This case seems to be one that ensures that the state, if tightly enough defined, can be a sufficient good, independently of object-given reasons. However, Parfit denies this also pointing out that what Case 3 establishes is not that it is best to be in the state in question of having adopted the belief stated but rather that it would be best if we could cause ourselves to be in this state. There are still no reasons to hold the beliefs themselves (which are false) only to make circumstances come about that we could hold them (at least for a time).
So Parfit's real point is to the effect that state-given reasons are insufficient ground for holding a belief to be true even though there may be grounds for causing ourselves to act in a way that enabled us to state that they were true. And this shows the basic point he is making against the instrumental state-given reasons theory. Which is that it does not provide us with a basis for believing something to be the case and if it does not do so it therefore does not provide us with an account of reasons.
Causing ourselves to adopt certain kinds of beliefs is something that we might have reason to do and so be rational. In adopting the activity of bringing it about that we believe something we are acting in a certain way (or set of ways). But there is not a passive way of coming to hold beliefs to be true, just persuading ourselves that we should adopt a belief. Since we can't act in this way it is not irrational to fail to do so.
Once the case is put like this Parfit considers what I think is the core question of the state-given theory. On this view there are two types of question that can be meaningfully asked concerning a belief. The first would be whether the belief was warranted in terms of whether it could be sustained by means that relate to the processes of holding things to be true. The second would be whether a belief satisfied, instead, some practical demand. Once the question is put like this we can see it touches on some important questions, including the arguments Kant gives in the Critique of Practical Reason for the immortality of the soul and the existence of God.
Now that the question of practical state-given reasons has been seriously broached Parfit turns to giving more precision to the view that that there are beliefs that have practical value. Restating this view Parfit claims that it is equivalent to the argument that there are beliefs it would be best to adopt or cause oneself to adopt if possible. This then becomes a question, on Parfit's view, not about beliefs themselves but instead about what we have reason to do or to want.
Putting the question this way makes the argument about state-given reasons one not about beliefs but, whilst this clarification turns the discussion away from beliefs themselves, it still opens a question about the grounds for adopting beliefs and implies that these grounds are not all based on the tracking operation that might be thought to sustain theoretical beliefs generally. But Parfit does not address this claim since he rather holds that practical claims are such by virtue of the role they play in carrying out some operation. (He gives the example of re-doing a sum to make sure it has been well carried out as a practical operation.) Hence he does not consider, despite being on the verge of doing so, the claims made in the Critique of Practical Reason.
Rather than considering these claims Parfit instead moves on to a different kind of state-given theory which concerns desires rather than beliefs:
State-Given Theorists also claim that(2) whenever certain facts would make it better if we had some desire, these facts give us a reason to have this desire.
This argument is considered more plausible by Parfit than the first form of state-given theory since it concerns desires, which have no tracking relation of the sort claimed for beliefs. So the view when formulated in relation to desires would be that we have a reason to have a desire whenever our wanting it would be good (which is a kind of Jamesian "will to believe" argument). In replying to this formulation of the state-given theory Parfit considers a new case.
Case 4 is one in which the despot will torture me unless I develop the desire that he kill me, as, if I do develop the said desire, he'll set me free. On these grounds there would be a reason to adopt the desire in question. But Parfit's point here is much simpler than earlier since now he can simply claim that the basis of resistance to the adoption of the belief is already objectively given and will tell against the pragmatic adoption of the belief.
In Case 4 I have decisive reasons not to want to be killed and so causing myself to adopt the belief I do want to be killed would be causing myself to adopt a belief I have decisive reasons not to adopt. The reason, all the same, to cause myself to adopt the belief that I have decisive reasons not to want to adopt, would be to secure the end that the object-oriented belief states and not due to any intrinsic value of the state I would be in. Again, however, this case cuts only against an instrumental understanding of the state-given theory.
Parfit's lengthy consideration of the state-given theory tells against an instrumental understanding of it in relation to both beliefs and desires but not against telic understanding of it in relation to either belief or desire. Further, the specifics that are involved in practical reasons to adopt a belief that cannot be theoretically mandated are avoided as Parfit makes assumptions about practical reasons in relation to theoretical that do not fit the case of Kant's consideration of practical grounds for beliefs in the Critique of Practical Reason.