Do cry over spilt milk: possibly you can change the past

Monist, The, July, 2005 by Peter B.M. Vranas

There is widespread agreement, even among those who accept the possibility of backward causation, that it is impossible to change the past. I argue that this agreement corresponds to a relatively uninteresting understanding of what changing the past amounts to. In one sense it is indeed impossible to change the past: in no possible world is an action performed which makes the past in that world different from the past in that world. In another sense, however, it may be possible to change the past: maybe in some possible world an action is performed which makes the past in that world different from the actual past. I argue that those who accept the possibility of backward causation are committed to accepting the possibility that the past changes in the latter sense.

1. Introduction

On the 24th of July 1987 I did something I had never done before: I revealed to a woman that I was in love with her. I wish I had not done so: the timing was bad. I was about to go abroad for a number of years, and I suspect that the woman's knowledge of this fact weighed heavily in her refusal to form a relationship with me. I wish I had spoken only after coming back from abroad. But it's no use crying over spilt milk--or is it?

Can I now, in 2005, make it the case that I had not spoken to that woman in 19877 The suggestion sounds absurd. It is widely agreed that it is logically impossible to change the past. Even love is bound by the laws of logic. But what if time travel to the past were possible, as some physicists think it may well be? (1) Then, according to many philosophers, it would be possible to affect the past--to have a causal effect on it--but it would still be impossible to change the past. (2) To appreciate the distinction between affecting and changing the past, suppose the following scenario is possible. In 2005, being 40 years old, I enter a time machine. After two minutes, as measured by my watch, I exit the machine on the 24th of July 1987, still being 40 years old. Then I go to a payphone, I call my 22-year-old younger self, and I try to talk him out of speaking to the woman. It turns out, however, that he had not even considered doing so; perversely, it is my very phone call which makes him decide to speak. In this scenario I have a causal effect on the past: in 2005 I initiate a causal sequence which results in the occurrence of a declaration of love in 1987. So I affect the past. Still, I don't change or transform the past: in the scenario it is not the case that there is a first 1987 without the phone call and a second 1987 with the phone call. There is rather a single 1987, which includes the phone call. It seems then that even time travel would not give me what I want, namely an ability to transform the past.

But is this really what I want? Do I really want it to be the case that there is a first 1987 in which the declaration of love occurs and a second 1987 in which the declaration does not occur? No. I rather want it to be the case that the declaration never happened; I want it to be the case that there is a single 1987 in which, as a result of something I do in 2005, the declaration does not occur. To use a label, I want to replace the actual past: I want to bring about a nonactual past. If so, then what I want may be possible after all: if the scenario in the last paragraph (in which the phone call does not prevent but rather causes the declaration of love) is possible, then the alternative scenario is also possible in which the phone call does prevent the declaration of love. Granted, this alternative scenario is not actual, since in the actual world the declaration did occur in 1987. But the question is whether I can bring about this alternative scenario, whether I can replace the actual past. And my point is that if I can bring about the original scenario then I can also bring about the alternative scenario: I can now enter a time machine, exit in 1987, and make a phone call which prevents my younger self from speaking to the woman. Admittedly I will not do so, but it does not follow that 1 cannot.

Let me recapitulate what I did in the last two paragraphs. First, I went over a well-known distinction between affecting and changing (in the sense of transforming) the past. Second, I introduced a--to my knowledge novel (though cf. Mavrodes 1984: 143f)--distinction between two senses of 'changing the past': transforming and replacing the past. Third, I suggested that changing the past is more interestingly understood as replacing than as transforming the past; this is my first main thesis in the present paper. Fourth, I suggested that if affecting the past is possible (alternatively: feasible) then so is replacing the past; this is my second main thesis in the paper. I did not defend the possibility of affecting the past: this would be a large project, because it would require defending the possibility of backward causation. But many philosophers already accept that affecting the past is possible; if my two main theses are correct, then these philosophers should further accept that changing the past in the more interesting sense--namely replacing it--is also possible.