Tuesday, December 16, 2008

Changing the Past

The Ethical Werewolf asks:
What modality applies to the impossibility of changing the past? Is changing the past logically impossible, metaphysically impossible, or physically impossible?

There are many possible questions here. On one popular reading, to "change" the past (or the future, for that matter) is logically impossible: given that Bob was shot at time t, it is not logically compatible with this to make the world such that Bob was not shot at time t. So we cannot "change" the past in the sense that "first" time t is one way, and "later"(?) that very same moment somehow differs. Change is when one moment differs from another; a moment cannot differ from itself. But this is a trivial kind of 'impossibility'.

At the other extreme, we may refrain from taking any merely contingent facts as 'given', and ask abstractly: is it possible that I perform some action φ, such that Bob isn't (and never was) shot at time t? And that seems possible in almost every sense: there's a physically possible world where I φ and Bob isn't shot. It just isn't this one. (And, more importantly, neither is it accessible in any important sense.)

So the real question must lie somewhere in between. Most plausibly, it depends on some notion of "dynamic" or time-relative physical (causal) possibility. That is, we take the present moment as given, and ask whether we have (sufficient) causal influence over what happens at time t. So the modality of my inability to (now) influence the past is simply a time-relative version of the modality barring me from time travel. 'Physical impossibility given my current circumstances', perhaps?


  1. The question becomes whether backwards causality is possible. I have no idea if it is but I don't have a problem with it if it is since I don't think causality is a fundamental ontological characteristic of the universe.

    There's nothing in physics rejecting the notion although in GR the nature of tachyons are pretty weird and there are good reasons to think they don't exist. In QM it's trickier since within short periods of time clearly there is backwards causality. However some interpret that as mere mathematics and not something real. (Here thinking of say Feynman diagrams of interacting electrons)

  2. I'm not sure that it's logically impossible that the past should be changed. Say that I could go back into the past and change something, like rewinding and recording over a VCR tape. Obviously the moment would change, but it would seem strange to argue that it never had been in its original state.

  3. John - what you just described sounds incoherent to me. You might be able to see this by asking questions like, "At what time t* did moment t change its intrinsic state?" You must think there are times (i.e. prior to t*) at which t has intrinsic state S1, and other times (all those subsequent to t*) at which t has some other intrinsic state S2. But t is a single, mere moment in time; it is not a persisting or temporally extended substance of the sort that could have different properties or states at different times. The truths about t are timeless truths, for the moment t remains intrinsically the same no matter what moment on the timeline we consider it from. So the idea that one "could go back into the past and change something" is simply incoherent. If you do something in the past, then the past was like that all along (so to speak). See my linked post for more detail.

  4. In your linked post you say that if I go back and change the past, then "it was like that all along". Only from the experiencer's perspective. The fact that we are living through this particular time is not proof of the fact that it hasn't been rewritten a few times before, or isn't going to be rewritten. Why should any moment in time "t" be eternally fixed as it is? We can change things at different points in space.

  5. Points of space endure through time, so they can "change" in the sense of being different at different times. The same obviously cannot be said of points of time themselves. Our best understanding of time is provided by the eternalist's picture of a static timeline. If you try to explicate your proposal in those terms (i.e. where 'change' is simply the difference between how something is at two different points on the timeline), you'll see it becomes senseless.

  6. 'Physical impossibility given my current circumstances', perhaps?

    This sounds promising, but I guess the difficulty is in spelling out the circumstances. If I say something like "Eight years after Gore didn't win, it's impossible for me to make Gore win" that just seems like we're dumped back into logical impossibility.

    If I say something like "Eight years after 2000, it's impossible for me to exert causal influence over an election in 2000" that seems like more physical impossibility. The circumstances here are just that I'm after the relevant time.

    I guess the second question is more likely to be the one I'm asking. I'm not interested in generating a contradiction (and maybe destroying the world through explosion, as Bush came close enough to that for my tastes). I'm just interested in exerting causal influence over an event.

  7. In my personal opinion, time travel is logically impossible because if you go back in time, every time you go back you would do the same thing. For example, lets say you go back in time and rescue 2 people who you think die in your timeline. You can only "change" an event if yourself in the time you go to doesn't change his/her perspective until after they travel back in time. Otherwise, you would know what you "changed" before you went. If you did know then you aren't changing anything. Very interesting concept though.

    Sorry for raining on your parade.


Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)

Note: only a member of this blog may post a comment.