Physicists sometimes talk about particles going backwards in time. Help me make sense of this. I thought things don't "go" in any particular direction in time. They just "are" there at every moment. And you can imagine playing the tape "forwards" or "backwards" but that has no physical meaning.
$\begingroup$
$\endgroup$
2
-
1$\begingroup$ "Physicists sometimes talk about particles going backwards in time." Which physicists? In what context? Do you mean to ask about tachyons? Or about advanced vs retarded propagators? Or what? $\endgroup$– hftCommented Jun 30, 2023 at 0:11
-
$\begingroup$ Possible duplicates: physics.stackexchange.com/q/391/2451 and links therein. $\endgroup$– Qmechanic ♦Commented Jul 1, 2023 at 9:50
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
5
It's unfortunately common for popular science accounts to use the "going backwards in time" when talking about zigzag causal structures in Feynman diagrams and retrocausal quantum models. It's unfortunate, because as you correctly note, that phrase isn't logically coherent; it uses time in two different senses in quick succession. Feynman himself never used that phrase, I'm almost certain, and in my Rev. Mod. Phys. piece of retrocausal accounts of entanglement we certainly did not use anything like that terminology, instead adopting the block-universe perspective of spacetime events. (The set of those events, as you correctly note, can be "played" forward or backwards, and don't "go" in either direction.)
But instead of dismissing such phrases out of hand, it's perhaps worth analyzing why it's so common for people to use this particular phrase in these contexts. The central point that is being (clumsily) attempted is that the causal order is opposite the temporal order. Any careful and precise way to make this point requires the "interventionist" account of causation popularized by Judea Pearl, and frequently used in the zigzag context by Huw Price.
Suppose you had a zigzag structure $\text{/\/}$ in spacetime, as in one of Feynman's examples of an electron which annihilates the positron from an electron-positron pair. A lazy and technically incorrect way to talk about this would be that the electron "goes forward in time" (on the left), then "goes back in time" (as a positron) in the middle, and then "goes forward in time" (on the right). But the meaningful sentiment of these statements is that the causal structure of this event looks like a directed acyclic graph (Pearl's DAGs), with zigzagging arrows pointing up to the right, then down to the right, then up to the right.
The key here is that interventionist causation doesn't quite work in a single block universe, as it requires counterfactuals; you have to imagine how the events would be different given potential interventions at different points. The causal arrows tell you in what direction to expect the effects of those interventions. Those counterfactuals can't all fit on one spacetime diagram – they refer to counterfactual universes – but they do tie into how we innately think about causation, imagining what would be different under different external interventions. With this in mind, those statements might make a bit more sense: the word "going" is meant to be causal, not temporal, not anything which would be meaningful in a single block universe without imagining counterfactuals.
-
$\begingroup$ Thanks for your reference (free version arxiv.org/abs/1906.04313 ), Ken. I was familiar with another piece of yours A New Class of Retrocausal Models and I coincidentally clicked on a story that came out yesterday with your name on it: aeon.co/essays/… . The element I find positive in the causal zigzags is that each individual component can only "influence" at a maximum of +c or -c. In that respect, you could say it respects relativity. It seems we have to abandon classical forward in time causality no matter what. $\endgroup$ Commented Jun 30, 2023 at 20:03
-
$\begingroup$ "the causal order is opposite the temporal order" - that's a succinct way of putting it, but isn't that a contradiction in terms? What basis would there be to ascribe causes as distinct from effects, if one did not precede the other in time? $\endgroup$– SteveCommented Jul 1, 2023 at 9:22
-
1$\begingroup$ DrChinese: "Influence" is a causal notion, so it doesn't have a speed, but I take your point. Perhaps it's clearer to say that the events which mediate the causation on the zigzag all can be Lorentz Covariant, because they all lie on timelike worldlines. And it's generally clearer to talk about respecting Lorentz Covariance, rather than the vaguer "relativity". $\endgroup$ Commented Jul 1, 2023 at 16:13
-
$\begingroup$ Steve: I agree that if you define the distinction between "cause" and "effect" based on temporal order, this doesn't make sense. It rules out retrocausality by definition. But that's not the definition people instinctively use, or else no one would be able to make sense of time travel movies. Also, no one would be convinced that the light switch causes the bulb to turn on, because those events seem to be instantaneous; you can't tell which one happens first. Instead, our instinctive notion of causation is tied to Interventions; see the links in the 2nd paragraph of my answer. $\endgroup$ Commented Jul 1, 2023 at 16:17
-
$\begingroup$ @KenWharton Agreed. What I refer to as an "influence" was in quotes for exactly the reason you say; since it is more like a mutual influence - and temporal order is NOT factor in the observed results. With the advent of entanglement swapping experiments with independent sources and delayed execution of the swap, it becomes harder and harder ignore Interpretations with some kind of retrocausal component (regardless of what you call it). $\endgroup$ Commented Jul 2, 2023 at 17:44