Many writers believe there can be cases which satisfy the following description: starting from an initial state of affairs, it is possible to make a series of changes, none of which alters the value of the state of affairs in any way, but such that the final state of affairs that results from the series of changes is worse than the initial state of affairs. I shall call the claim that there can be such cases the "ex nihilo" claim, since in a sense it asserts that the bad effects of the complete series of changes arise ex nihilo. Proponents of the ex nihilo claim - ex nihilists, as I shall call them - usually advance the claim as part of an argument against act-utilitarianism. If there were cases such as the ex nihilist imagines, then it would be possible to construct variants in which act-utilitarianism unequivocally required behavior which in the aggregate produced sub-optimal consequences. We could construct the sort of case I have called (while denying its possibility) an "act-utilitarian prisoners' dilemma". Act-utilitarianism would be, in Derek Parfit's phrase, "directly collectively self-defeating".
Or so the ex nihilist thinks, and so I carelessly thought when I wrote Utilitarianism and Co-operation. But Wlodzimierz Rabinowicz has pointed out that the ex nihilist's case would not in fact be an act-utilitarian prisoners' dilemma. The reason is easiest to see if we have an example. According to the ex nihilist view of the lawn-crossing problem, there are no bad consequences from one person's crossing a lawn, regardless of how many others do so. So, if each walker gains some small benefit from crossing, it is better than one person cross than none, better that two people cross than one, and so on, right up to its being better that 1,000 cross than 999. But, it is supposed, the consequences of 1,000 crossing are worse than if none cross, which is to say, it is better that none cross than 1,000. But now we have a cycle. 1,000 crossings is better than 999, which is better than 998, ... , which is better than 1, which is better than 0, which is better than 1,000. If this is all true, "better than" is intransitive, and as Rabinowicz and Michael Otsuka have pointed out, it becomes impossible to say that any pattern of behavior is either optimal or sub-optimal. So universal satisfaction of act-utilitarianism in such a case cannot entail sub-optimal results. There is no act-utilitarian prisoners' dilemma.
I overlooked this point in Utilitarianism and Co-operation because I was intent on showing that there simply cannot be such a case as the ex nihilist imagines - which I still think is true and important to establish. If the ex nihilist's case did exist, the intransitivity of "better than" would destroy consequentialist reasoning as we know it. It would be small solace that act-utilitarian prisoners' dilemmas disappeared in the wreckage. I shall revisit my earlier argument against the ex nihilo claim, and elaborate on it slightly, in section II below. I shall point out that the main premise of the argument is effectively equivalent to the denial of intransitivity, which is not surprising in view of what we have just seen. This means that I must confront Larry Temkin's very impressive "continuum argument" for intransitivity.
But I want to begin by focusing on a sort of case in which the ex nihilo claim has its greatest plausibility- cases in which the harm lies in the worsening of the subjective experience of some agent or agents. In such cases the ex nihilist can try to exploit the possibility of imperceptible differences.
Publication Information & Recommended Citation
Regan, Donald H. "Perceiving Imperceptible Harms (With Other Thoughts on Transitivity, Cumulative Effects, and Consequentialism)." In Imperceptible Harms and Benefits, edited by M. J. Almeida, 49-73. Library of Ethics and Applied Philosophy, 8. Dordrecht, Netherlands: Kluwer Academic Publishers, 2000.