Alpha: I've been thinking about Fatalism again.

Beta: Not again! The philosopher's stone: a persuasive argument for Fatalism! Dream on.

Alpha: Just wait. I know there have been lots of bad arguments for Fatalism, but that doesn't mean there might not be a good one.

Beta: I'm listening.

Alpha: Well, let's begin with something we can agree on. No one can change the past. You agree?

Beta: Yes, that is the orthodoxy.

Alpha: And the reason no one can change the past is because it's logically impossible. To change the past would be to make it the case that some event both happened and didn't happen – which is a logical impossibility.

Beta: Only if the past is real, of course, which Presentists would deny.

Alpha: Yes, I am assuming the reality of the past, otherwise the question of whether it can be changed or not wouldn't arise.

Beta: I will grant you the reality of the past – I'm no Presentist.

Alpha: Glad to hear it. Now can one change the future?

Beta: Well, we think we can affect or influence the future. That's just common sense. By acting one way rather than another, I make the future be one way rather than another.

Alpha: But that wasn't what I asked you. I asked you whether the future can be changed in the sense in which the past cannot.

Beta: I see what you're driving at. If it can't be the case that some event both happened and didn't happen, then equally it can't be the case that some event both will happen and will not happen.

Alpha: Right, and so far I have said nothing about whether I can affect or influence the future, just as I have said nothing about whether it's possible to affect or influence the past.

Beta: Ah, so the impossibility of changing the past does not imply the impossibility of affecting or bringing about the past?

Alpha: That's my view. In possible worlds with easily manipulable backwards causal chains, we can cause or bring about the past, but even in those worlds we don't (and can't) change the past, we just help to make the past be what it was. (It's hard to put this grammatically.)

Beta: So where are we now? What happened to Fatalism?

Alpha: Right. So we agree that we can't change the past or the future (whether or not we can affect either)?

Beta: Agreed.

Alpha: So since Hitler died in 1945 I can't go back in time and kill him in 1930, even if this world allowed for backwards causation and time machines.

Beta: Yes, it can't be true both that Hitler died in 1945 and in 1930. Given that he died in 1945, you can't kill him in 1930. But that's not Fatalism, is it? I thought Fatalism had to do with the future?

Alpha: I never said it was Fatalism. But, since you mention it, let's consider the future. I have a Grandson (I think you've met the little chap), and he will die at a certain time in the future, won't he?

Beta: Yes, but we don't know when, do we? We know when Hitler died though.

Alpha: It's true we don't know, but that's not what I asked. I just asked if it's true now that there is a certain date on which he will die.

Beta: Well, I suppose so. But wait a minute, on some views of time – not just Presentism – the future is unreal. It doesn't exist. On these views, it's not true now that Grandson will die at a certain time in the future. There are no contingent truths about the future.

Alpha: Good point. On both Presentism and the Growing Universe View, the future is unreal. In that case, Fatalism is no threat. So let's assume from now on that eternalism is true and that past, present and future are equally real.

Beta: Aha! So you're not arguing for Fatalism; only for Fatalism given eternalism?

Alpha: Guilty as charged. But most eternalists aren't fatalists, so wouldn't that be an interesting result to establish?

Beta: Maybe. Anyway, where's the argument?

Alpha: Well, let me proceed. Grandson will die at a certain time in the future; let's assume that he will die in his bed in 2060.

Beta: But you don't know that!

Alpha: We've been through this. I'm just stipulating; for the sake of argument; we are philosophers after all.

Beta: Fair enough.

Alpha: So the fact that Grandson will die in 2060 is a fact about the future just as much as the fact that Hitler died in 1945 is a fact about the past.

Beta: So we are assuming.

Alpha: Well, surely you can see what's coming?

Beta: No.

Alpha: Right: we agreed I couldn't go back in time and kill Hitler in 1930.

Beta: Yes.

Alpha: And the reason for that was simple: because he wasn't killed then.

Beta: Yes, that sounds right.

Alpha: Well, by the same token, I can't kill Grandson tomorrow, because he will not die then (even though I have the means and opportunity to kill him tomorrow).

Beta: Hold on, that doesn't sound right.

Alpha: Maybe not, but where have we gone wrong? We are simply treating the future as we did the past, and we have agreed that they are ontologically on a par.

Beta: Ontologically?

Alpha: They both exist. And neither can be changed.

Beta: Yes, we are assuming that. No, something must have gone wrong. I don't like your conclusion. Ah it's coming back to me. Wouldn't David Lewis say that there is a sense in which I can kill Hitler in 1930, even if there is also a sense in which I can't? (Lewis Reference Lewis1976)

Alpha: Indeed he would. His example involved Tim travelling back to 1920 in order to try to kill his Grandfather (who died in 1957). And yes, Lewis said that there was a mundane sense in which Tim can kill Grandfather in 1920 (if we focus only on Tim's means, motive and opportunity), but a more inclusive sense in which he can't (if we add in the fact that Grandfather died in 1957).

Beta: But Lewis thought that Tim was free to kill Grandfather in 1920?

Alpha: Yes, I believe he did. He didn't think that a time traveller's freedom somehow diminishes as he travels back in time. Of course, given that Grandfather died in 1957, Tim will fail in his attempt to kill Grandfather, but failure does not imply lack of freedom, and each failed attempt, according to Lewis, will have a commonplace explanation (no Hand of Fate required).

Beta: Good. I'll just apply Lewis's reasoning to your Grandson case: there is an ordinary sense in which you can kill Grandson tomorrow and another sense in which you can't. Fatalism avoided! Freedom regained!

Alpha: Well, just a minute. We agreed it was logically impossible to change the future.

Beta: Yes.

Alpha: If I kill Grandson tomorrow, the future would be changed. Right?

Beta: Er, I suppose.

Alpha: Then if it's logically impossible to change the future, it's logically impossible for me to kill Grandson tomorrow. But if it's logically impossible for me to kill Grandson tomorrow, surely there is no sense in which I can kill him tomorrow. Here's an analogy: if Goldbach's Conjecture is false, there is no sense in which anyone can prove it (since doing so is mathematically impossible).

Beta: Just a minute. I smell a fallacy here. ‘I kill Grandson tomorrow’ doesn't look like a logical impossibility. Surely what's logically impossible is killing Grandson tomorrow given that he dies at a later time. From this it doesn't follow that it is logically impossible to kill Grandson tomorrow. Here's an analogy for you: it's logically impossible that Bill's a bachelor given that he's married; but it doesn't follow from this that it's logically impossible that Bill's a bachelor!

Alpha: You've been doing some reading I see. So you are happy with the Lewisian response to my Grandson example?

Beta: No, not really. I was happy enough with it in the case of Tim travelling to the past and trying to kill Grandfather. But I find it worrying that there's a sense in which you can't kill Grandson tomorrow, simply in virtue of the fact that he dies in 2060, even if there is another sense in which you can kill him tomorrow (but won't).

Alpha: I agree. The idea that the existence of future facts generates a sense in which we can't accomplish quite ordinary tasks is a fatalist one, yet seems obligatory for anyone who accepts the reality of the future. That's my argument for Fatalism.

Beta: Well, I can think of one response. If we want to avoid Fatalism, but grant parity between past and future cases, then I will have to go back on what I conceded earlier and deny that you can't kill Hitler in 1930, given that he died in 1945. The truth is: you can kill Hitler in 1930 and you can kill Grandson tomorrow (though, in both cases, you won't). The fact that Hitler died in 1945 doesn't imply that you can't kill him in 1930, only that you won't. The fact that Grandson will die in 2060 doesn't imply that you can't kill him tomorrow, only that you won't. Fatalism avoided.

And, what's more, I don't think these results are counter-intuitive. Here is a simple analogy. Suppose that no one will hit the bull's eye at tonight's darts tournament. It doesn't follow from this that I can't hit the bull's eye tonight, only that I won't. No one, except the fatalist, thinks that I can't hit the bull's eye because it won't be hit. Similarly, given that Hitler died in 1945, it doesn't follow that I can't kill him in 1930, only that I won't. If I can't do something that is because I lack the means or opportunity to do it, not because I will not succeed.

We can make the same point in terms of admissible and inadmissible reasons why someone can't do something. Just as ‘because it will not be hit’ is not an admissible reason why I can't hit the bull's eye, so ‘because Hitler wasn't killed in 1930’ is not an admissible reason why I can't kill Hitler in 1930. The fatalist's mistake is to count such pseudo-reasons as admissible.