Published online by Cambridge University Press: 26 January 2009
Objective consequentialism is often criticized because it is impossible to know which of our actions will have the best consequences. Why exactly does this undermine objective consequentialism? I offer a new link between the claim that our knowledge of the future is limited and the rejection of objective consequentialism: that ‘ought’ implies ‘can’ and we cannot produce the best consequences available to us. I support this apparently paradoxical contention by way of an analogy. I cannot beat Karpov at chess in spite of the fact that I can make each of many series of moves, at least one of which would beat him. I then respond to a series of objections. In the process I develop an account of the ‘can’ of ability. I conclude with some remarks about the bearing this attack has on subjective consequentialism.
1 If ours is a genuinely indeterministic universe, that is, if some future states of affairs have a genuinely objective probability of more than 0 and less than 1, then it may be appropriate for objective consequentialism to allow that the right action is the one with the highest ‘expected value’, where expected value is arrived at by taking the value of each possible outcome of the act, multiplying this value by the objective probability of that outcome's resulting from the act, and then taking the numbers thus obtained for each possible outcome and adding them together. If objective consequentialism is so qualified, itis at least as difficult to determine which action the view requires of one as it is on the more straightforward approach. So the difficulty I raise applies with equal force whether one includes objective probabilities or not. For this reason, and because it is considerably easier to formulate and to understand if one does not, I shall ignore the complication created by objective probabilities in the rest of the paper. On the other hand, if one allows subjective probabilities, that is, probabilities relative to what the agent knows, believes or could find out, a role in the definition of consequentialism, one has shifted to subjective consequentialism. I am grateful to an anonymous referee for Utilitas for help in formulating this footnote.
2 For a statement of these objections, see Quinton, Anthony, Utilitarian Ethics, London, 1973CrossRefGoogle Scholar; Gibbard, Allan, Wise Choices, Apt Feelings, Cambridge, p. 43Google Scholar; and Hudson, James, ‘Subjectivization in Ethics’, American Philosophical Quarterly, xxvi (1989)Google Scholar.
3 In an excellent pair of articles, ‘On the Logic of Ability,’ Journal of Philosophical Logic, xvii (1988)Google Scholar, and ‘Action and Ability’, Journal of Philosophical Logic, xiv (1990)Google Scholar, Mark A. Brown develops a logic for the ‘can’ of ability and argues forcefully for the conclusion that the following two inferences are not valid when the ‘diamond’ is interpreted as the ‘can’ of ability:
(C〈〉) |- 〈〉 (AvB) → (〈〉A v 〈〉B),
(T〈〉) |- A → 〈〉A.
4 Suppose my opponent remains unconvinced. Suppose she insists that, in the case in which I do win, I can win, although she grants that I cannot in the other cases. Suppose she also endorses the principle that ‘ought’ implies ‘can’. Since it is about as unlikely that one will accurately guess which actions would satisfy objective consequentialism (in the long haul) as that I will beat Karpov or win the billion dollars, I am prepared to accept the implication that objective consequentialism has about a one in a billion chance of being true.
5 We may need to weaken the point still further. Sampras can ace most tennis players even though his choosing to do so does not always guarantee that he will. Perhaps it is enough that his choosing to do so makes it very likely that he will.
6 For detailed discussion and bibliography, see Bennett, Jonathan, Events and Their Names, Oxford, 1988, pp. 188–231Google Scholar.
7 An anonymous referee for Utilitas and Andrew Cortens raised this point.
8 It may be that the agent (for example, the agent who has been convinced by the argument of this paper) will not believe of any action that it has the best consequences, and will not have good reason to believe of any action that it has the best consequences. In other words, the difficulty for objective consequentialism may spill over into difficulties for subjective consequentialism. This does not count as a reason to disregard the difficulty I have outlined for objective conseqentialism, since objective consequentialism and subjective consequentialism are not the only moral theories on the market.
9 Andrew Cortens raised the following objection to this argument:
1. If your argument against objective consequentialism is sound, then so is this argument:
i. It is sometimes true that we do not know how to do what we ought, because we do not know and cannot find out what the correct moral theory is.
ii. If we do not know how to do what we ought, then we are unable to do what we ought.
iii. ‘Ought’ implies ‘can’.
iv. So, it is not true that we ought to do what we ought.
2. But (iv) is obviously false.
3. So, your argument must be unsound.
I believe that philosophers exaggerate the difficulty that we have in discovering moral truths. In most cases, it is pretty clear what we ought to do (even if it is not clear which is the correct theory underlying this fact). Moreover, with a little bit of reflection, empathy, requests for advice from moral authorities and so on, we can usually figure out what we ought to do even if we are initially puzzled. This suggests that the cases in which we are completely unable to know what to do are rare. Secondly, much of the difficulty with figuring out what we ought to do could well be due to our failing various moral or epistemic duties, e.g., we have allowed ourselves to become callous or self deceptive. Ignorance which is due to such culpability does not excuse, just as (other kinds of) inability which are due to our own failures do not excuse. Finally, we should note that the earnest attempt to do the right thing almost always raises the probability that we will do the right thing, even if it does not guarantee it. I suggested above that being able to do something does not imply that one would do it if one tried, but only that one's trying to do it would significantly increase the likelihood that one would do it. It follows, even by the standards developed here, that we are (at least, almost always) able to do what we ought. If there are rare cases in which we are unable to know what we ought to do, and if ‘ought’ implies ‘can’, then in those rare cases, it will not be true that there is anything that we ought to do. Such situations would perhaps be analogous to cases in which an agent is physically completely incapacitated and it is probably appropriate to say that there is nothing that she ought to do. As long as such cases are sufficiently rare, their existence need not undermine my position.
10 I am grateful to Roger Crisp, Jonathan Bennett, Daniel Howard-Snyder, Andrew Cortens, Tom Downing, Hud Hudson, Phillip Montague, Richard Purtill and an anonymous referee for Utilitas for comments on earlier drafts of this paper.