Particles traveling backwards in time

cygnet1

According to the writings of Feynman and others, there is a probability amplitude for a particle to travel forward in time, interact with a field, travel backwards in time to a different spatial position, interact with the field again and resume its path forward in time before it reaches some detection device. To an observer traveling forward in time, this would look like a particle traveling forward in time, a particle/antiparticle pair being generated out of the vacuum state, the antiparticle miraculously making a head-on collision with the original particle and anhiliating it; meanwhile the second particle resumes its path forward in time to the detection equipment.

My question is this: Do we ever see direct evidence of this pair creation and subsequent anhiliation in accelerated particle experiments, or is this simply a computation strategy to predict probabilities of events we actually can detect?

mfb

Those particles/antiparticles are called virtual for a good reason: You cannot directly measure them.
You can measure the whole particle propagation - it is influenced by virtual particles, but you cannot measure the individual feynman graphs leading to this propagation.

cygnet1

Thank you, mfb. That's what I suspected. However, we can directly observe antiparticles. What if the second interaction never occurs? Then the particle keeps going backwards in time indefinitely. To an ordinary observer, this would appear as though an antiparticle appears somewhere, and eventually anhiliates with the particle (as every antiparticle does eventually). This antiparticle would no longer be virtual, since it would appear to originate as part of the initial conditions.

Are non-virtual antiparticles really particles traveling backwards in time? (I'm talking about events that lie inside the light cone.) CPT symmetry seems to imply that they are.

mfb

Real antiparticles? Sure, they are real.
Virtual/real has nothing to do with the particle/antiparticle naming.

It has to, otherwise the whole process does not happen.

This is just a question of taste/definition.

Bill_K

Perhaps. But the backwards-in-time description invites the OP's paradox: "how does the antiparticle miraculously find the event where it was created?"

There is no doubt which way causality flows: from past to future. And as you said, mfb, if the antiparticle was not created, the whole process does not happen.

A clearer way of putting it, I think, is: "as if a particle had gone backwards in time". An antiparticle really does travel forward, just like everything else. But as it so happens, the amplitude for absorbing an antiparticle equals the amplitude for creating a backwards-moving particle. And so for purposes of calculation you can replace one by the other. Feynman's description is picturesque but has led to more than a little confusion.
That's the statement I would agree with.

cygnet1

That's the crux of my dilemma. Suppose we're studying the behavior of an electron passing through an EM field. It seems exceedingly unlikely that a another electron/positron is spontaneously created out of the vacuum field, the positron takes perfect aim at the original electron and annihilates it, letting the second electron take the place of the first. This is assuming that pair creation is a relatively rare event, and when it occurs, the pairs are created at random positions in random directions. Yet this is a second-order process, and leads to significant errors if it is neglected.

On the other hand, if we're able to accept that the original electron does somehow travel backwards in time, it is easier to see how the created positron finds its mark.

At a macroscopic level, there is no doubt. But at a quantum level, doesn't CPT symmetry imply that it is impossible to distinguish an electron traveling backwards in time from a positron traveling forward in time in the opposite direction? (Maybe its spin also has to be reversed.)

If the antiparticle was not created when it was, it would have been created at a different time, perhaps much earlier, so that it was observed as a real antiparticle as part of the initial conditions. Isn't that how real antiparticles are created in the lab?

BTW, Apparently, I'm not the only one thinking about this problem. Another poster just brought up the same issue in another thread in the Quantum Physics forum.

Bill_K

Yes, this is a VFAQ!

In the CPT Theorem, one has to apply time reversal to everything, not just one particle. It says, given any state of a system there will be another state possible in which particles are replaced by antiparticles (and vice versa) and which looks like the original state with space and time inverted; and furthermore this new state obeys the same physical laws as the original one.

Take for example a muon decay, μ^- → e^- + ν + ν. In the time reversed process, e^-, ν and ν come together to form μ^-. But that is not what the CPT Theorem deals with. It deals with the decay of the antiparticle: μ⁺ → e⁺ + ν + ν. "Obeys the same physical laws" means among other things that the particle and antiparticle decay at the same rate.
No, they're both equally unlikely. Particles can interact only if they come into extremely close proximity. Going backwards in time doesn't alter that fact.

ImaLooser

I simply do not see how I can determine whether or not a particle is traveling backward in time, so as far as physics goes it seems rather meaningless. I mean, they can trap antihydrogren for a while. Someone could say, "see, this bottle is full of particles going backward in time!" How can I prove or disprove that? I can't. I can't prove or disprove that they are going sideway, upside down, or sunny side up in time either. So it is hard for me to become interested in this question.