Parfit presents a number of different arguments against what he terms "subjectivist" views of reasons in Chapters 3 and 4 of On What Matters but the first argument, termed by Parfit the "agony argument", is given special stress by him and in this posting I will look at the argument prior to reviewing Michael Smith's critical response to it.
Parfit presents the "agony argument" in response to what he terms "Case One" and in "Case One" he assumes that some future event will cause a period of agony but, even after ideal deliberation, I find I have no desire to avoid it and adds, somewhat implausibly, that there is no other desire or aim whose fulfilment would be prevented either by the agony in question or by my having no desire to avoid it. Once the case is fully set out it is clear that Parfit has here rather stacked the cards since it is hard to conceive of a period of agony that would not prevent some desires or aims having fulfilment.
The result of the way of describing it that Parfit has given is that in the case in point "subjective" theories of reasons would give no reason to avoid the agony in question. The reason why Parfit takes it that the case has plausibility at all is because he refers to the way in which we relate to present and future agony in terms of having what he elsewhere terms a "bias towards the near". This kind of bias is fairly commonly exhibited since, to take a case that does not usually involve "agony" exactly it is not uncommon to drink to excess despite awareness that later there will be suffering to endure as a consequence. Since the suffering is in the future and the enjoyment is taking place now, the suggestion goes, we rationalise a way of going through the later experience of suffering.
It is not as obvious as these remarks might suggest that the reason why we behave in the way indicated in my example is due to a "bias towards the near" and nor does the example, in any event, show that "future agony" will be accepted in quite the same way as the consequence of heavy drinking often appears to be. In any case the argument Parfit is making is that a "subjective" view of reasons provides no reason in the example of "Case One" to avoid the future agony.
An obvious retort that is considered by Parfit is that when the future occasion of agony becomes present I will at that point have clear desires to avoid the agony in question and that this is sufficient to give me a desire in the present to avoid this future agony. However Parfit suggests that "subjective" theorists of reasons cannot come to this conclusion as one committed to such a view cannot say that "facts about our future desires give us reasons". The ground of this claim is clearly indicated to be that commitment to such a "subjective" theory of reasons ensures that such "facts" cannot be taken into account as taking them into account involves treating the "facts" as "objects" and hence adopting an objective theory of reasons.
This claim concerning the difference between the subjective and objective theories has a peculiar structure like much that Parfit writes about this distinction. It appears to arrogate only to the objectivist a concern with the "facts" since only the objectivist can admit that "facts" are "objects" of self-sufficient standing. However the point about agony is simply that it is something of a phenomenological description and that its phenomenological description points sufficiently of itself to its undesirable nature without having to invoke "objective" qualities at all here.
Parfit attempts to resist this kind of move by suggesting that to make it already involves a "mixed" conception in implying an objective theory of reasons has been coupled with a subjective view of well-being. However I see no need for the "subjectivist" to admit to such a "mixed" conception here at all since the agony in question is a "fact" of a phenomenological kind and not of a different sort. This phenomenological kind of "fact" would be sufficient of itself to make the reason for its avoidance also subjective and not objective and Parfit has provided no reason as yet to think otherwise.
The structure of the "subjective" theory that Parfit supposes becomes clearer when he mentions the notion of temporal neutrality as contrastive with the theory being considered. If you were committed to temporal neutrality you would take the period at which agony was suffered to be indifferent since any such point would be regarded in the same way but Parfit takes the subjective theory under consideration to be constituted by its temporal bias towards the near. At this point, however, it becomes unclear why the argument is being advanced since it touches only upon a specially selected type of subjective theory of reasons and within the body of subjective theories as Parfit has described them it can be seen as an especially eccentric type but nothing in the argument as given touches on subjective theories of reasons as such.
The furtherance of the claim that this argument nonetheless tells against "subjective" theories of reasons in a general way occurs by Parfit citing something from Rawls where Rawls stated that we have no means of telling in advance what ends rational people will pursue and Parfit uses this to argue that it is not possible to claim, on subjectivist grounds alone, that anyone who is fully rational would want to avoid future agony. However, as mentioned before, the simple experience of the agony is reason enough and this can be cited by any type of subjectivist and most definitely by Rawls as agony would be the opposite of a "primary good".
Smith, in replying to Parfit's "agony argument" also refers to the Kantian reply that can be made to it mentioning how universalisation can apply to the very ability to make autonomous choices and that periods of agony are periods in which such choices cannot be made. Since this is so, says Smith, there is a plausible simple universalisation argument that can prohibit future agony.
Parfit's argument was intended to show that there are no plausible reasons on a subjective theory of reasons to avoid future agony but his example required us to accept that subjective theories of reasons manifest an untenable bias towards the near which bias, however, is far from being obviously one that all "subjective" theories of reasons have to manifest. Further, Parfit in presenting this argument neglected the point that agony is subjectively undesirable and hence could be ruled out on these grounds on one type of subjective theory that did not manifest bias towards the near. Finally, as has been noted throughout consideration of Parfit's treatment of "subjective" theories of reasons, he has failed to take the measure of Kantian universalisation (or of Rawls' commitment to the notion of 'primary goods').