¹ In his discussion Nagel emphasizes that he is concerned with changes of values: ‘It may happen that a person believes at one time that he will at some future time accept general evaluative principles – principles about what things constitute reasons for action – which he now finds pernicious. Moreover he may believe that in the future he will find his present values pernicious’ (Nagel, Thomas, The Possibility of Altruism (Oxford, 1970), p. 74)Google Scholar. When I speak of goals I mean goals which depend on evaluative beliefs. Mere preferences independent of evaluation are excluded, although they provide goals in the ordinary sense. I will use ‘goals’ and ‘values’ virtually interchangeably.

² I will assume that the agent thinks it is appropriate to apply his present goals to a future time at which he no longer endorses them. In Parfit's terminology, his allegiance to his present goals is not ‘conditional on their own persistence’ (Parfit, Derek, Reasons and Persons (Oxford, 1984), p. 151)Google Scholar. He believes at T1 that his life at T2 should match those goals even though he realizes that at T2 he will no longer want his life to be that way.

³ Parfit apparently sees the traditional way of understanding prudence and the requirement of temporal neutrality as leading to this conclusion. ‘On the Self-interest Theory, this young man [Parfit is talking about the agent in an example of Nagel's] must give the same weight to his present and his predicted future values and ideals’ (Parfit, Reasons, p. 155). The Self-interest Theory goes beyond temporal neutrality because it claims that the ultimate object of rational concern is our own interest. But Parfit seems to think that it is the Self-interest Theory's acceptance of temporal neutrality that commits it to this conclusion.

⁴ It might be denied that what it is subjectively rational for an agent to do at a particular time must be explained in terms of the beliefs about reasons for acting that the agent holds at that particular time. Perhaps it can also depend on goals and beliefs about reasons for acting that she holds at other times. In that case subjective rationality might require her to give weight to her future goals as well as her present goals, even if she herself does not now believe that this is what it is rational for her to do. I will not argue against this suggestion, but I think it will be difficult to make a convincing case that the prudential solution is the best account of subjective rationality.

⁵ Nagel, Possibility, p. 74 fn.

⁶ Parfit, Reasons, pp. 154–5.

⁷ For the purposes of his argument Parfit supposes that this example does not involve a change of values (Reasons, p. 157). But if we suppose that there is a change of values the example retains its force.

⁸ For this reason it is unfortunate that discussions of temporal neutrality often use examples of changing goals. We might agree that we should give equal importance to each temporal part of our lives (for example, we might think that a benefit in the present provides no stronger reason to act than a benefit of the same size in the future), and still be puzzled by cases of conflicting goals. Combining the two issues might make the requirement of temporal neutrality seem less persuasive than it actually is.

⁹ Parfit's aim is to make a case for the present-aim theory and show that there might be persuasive answers to the objections against it. However, I think that what he says about changing values specifically (Parfit, Reasons, pp. 153–6) suggests that he sees it as the best response to that problem, whatever its merits as a general account of reasons for action.

¹⁰ Parfit discussed such an example in Parfit, Derek, ‘Later Selves and Moral Principles’, Philosophy and Personal Relations, ed. Montefiore, A. (London, 1973), sect. IVGoogle Scholar.

¹¹ Parfit comments that it might be logically impossible to do what the prudential solution asks us to do (Parfit, Reasons, p. 155). Presumably the thought is that if we were to give equal weight to our present and future values this would be incompatible with the supposition that we now endorse our present values and reject the future values. I am not sure that this is a real problem for the prudential solution, so I will concentrate on the objection that it asks us to do something unreasonable.

¹² As Parfit explains, the present-aim view does not mandate which values and goals a person should have at a given time. If an agent held the appropriate set of goals, the present-aim view would reach the same conclusion as the prudential solution. Consequently in the examples I use against the present-aim view it would be possible to specify a set of goals such that if the agent held those goals the present-aim view would reach what I consider the intuitively plausible conclusion about the example. But this does not mean that the examples are harmless. The present-aim view says that in cases of changing goals the agent should act on her present goals whatever they are, so the view would generate implausible conclusions in other cases.

¹³ A fourth view should be mentioned for the sake of completeness. Instead of making a claim about the significance of the temporal distinction between past, present, and future, it makes a claim about the significance of the different temporal distinction between earlier and later (or prior and posterior). The claim is that in cases of conflicting goals the person should act in accordance with the goals that he will hold later in his life. This view would agree with the present-aim view and the simultaneous-values view about past–present cases, and it would agree with the simultaneous values view about the kind of present–future cases that I have discussed. However, we can imagine a different kind of present–future case. At T1 I must make a decision that will affect my life at T1, but I know that later in life at T2 I will hold different goals which I would then retroactively apply to my decision at T1. The simultaneous values view tells me to make the choice at T1 in accordance with my goals at T1, but this view would tell me to make this choice in accordance with the goals I will come to hold at T2. This conclusion is intuitively implausible, but it does not seem to me to be as implausible as the conclusion that I should act in accordance with my past goals in past–present cases.

¹⁴ Bykvist, Krister, ‘The Moral Relevance of Past Preferences’, Time and Ethics: Essays at the Intersection, ed. Dyke, H. (Dordrecht, 2003)Google Scholar.

¹⁵ Bykvist does not say whether my predictable future preferences about the present should have weight if they agree with my present preferences about the present, but the ideas in his view should lead to that conclusion.

¹⁶ Bykvist's rationale spreads to cases that do not involve changes in goals. Suppose that a person must choose between a gain of well-being in the present and an equal-sized gain of well-being at some future time, in a case where there is no relevant change in the person's goals or values. The view about the significance of person stages seems to imply that the person should choose the present gain. Both choices would benefit the continuing person equally but only the first choice benefits the person stage that makes the choice.

¹⁷ Recent work on well-being by Dworkin, Ronald, Sovereign Virtue (Cambridge, Mass., 2000), ch. 6Google Scholar; Sumner, Wayne, Welfare, Happiness, and Ethics (Oxford, 1996)Google Scholar; and Darwall, Stephen, Welfare and Rational Care (Princeton and Oxford, 2002)CrossRefGoogle Scholar seems to accept some version of the positive response condition.

¹⁸ The same cannot be said about the objective value determinant of well-being. My judgment of the objective value of what happens at T2 will be different depending on whether I make this judgment at T1 or T2, so my view about the amount of well-being that I will experience at T2 will also differ depending on whether I am relying on my values at T1 or my values at T2.

¹⁹ On some views of well-being if the change occurs it can affect my well-being at T2 despite my ignorance. Still, this example reduces the impact on well-being.

²⁰ Of the writers I have mentioned I think that Dworkin and Darwall would accept the enhancement claim, although they might express it in different language. Sumner would not accept it.

²¹ Another possibility is that the gain in well-being cannot be assigned to any particular time in the agent's life. However, it seems appropriate to associate the gain with the activities that the agent is responding to.

²² In a different example a person might look back at his life and realize – correctly – that his past activities had less value than he believed when he engaged in them. Does this kind of retrospective disapproval affect his well-being, and is the effect positive or negative? Here I am tempted to appeal to the thought that the positive response should be proportional to the value of what it is a response to. His life is better because he does at last respond in the appropriate way, though it is not as good as it would have been if his past beliefs about the value of those activities had been true rather than false.

²³ Accepting retrospective value enhancement means that in some cases it would be better for me in terms of well-being to perform now more valuable activities that conform to my future values rather than less valuable activities that conform to my present values. But I would contend that this is not an implausible conclusion to reach about those examples.

²⁴ I would like to thank Nils Holtug, Krister Bykvist, Thomas Hurka, and audiences at the University of Calgary and the University of Toronto for helpful comments on earlier versions of this article.