Measurement problem and many worlds

[ripe90]

Hello, sorry if I created new thread that is already open, but I did not find answer. I would like to ask you about measurement problem (double slit experiment) and many worlds. When interference pattern is created, the dot on screen just show us in which branch or world we are. But if we observe the path of the particle, pattern is changed. Does it means that the way of branching new worlds is also changed?

 

[andrewkirk]

It depends on how and when the path is measured. The 'branching' happens when a measurement occurs. If the path is measured before the dot is created on the screen, then a branching occurs earlier than it would in the alternative experiment in which path is not measured.

For practical reasons, I think that path-measuring will always require a measurement earlier than would occur in a non-path-marking approach, as long as we interpret 'measurement' broadly enough, and that includes delayed erasure setups. For instance in the Kim et al erasure experiment, the path is in a sense 'measured' as soon as the idler photon is split off, which is before the signal photon strikes the screen.

 

[ripe90]

Thank you for your reply! I was just wondering why the interference pattern is destroyed in case of MWI - because MWI says that there is no wave function collapse. The experiment you send is very interesting, travel back in time :)

 

[andrewkirk]

Say there are n photons that have hit the screen (that's an oversimplification, because the number of photons striking the screen is itself a stochastic outcome, but I think we can ignore that for this purpose).

Then there are 2^n different combinations of which slits the photons have passed through. The wavefunction after the experiment is a superposition of 2^n states, in each of which there is a unique one of those 2^n choices of which slits the photons went through. In every one of those states there is a physicist looking at their screen, and seeing no interference pattern because the photons all came through definite slits and hence could not interfere.

Under MWI, there is no collapse, because all 2^n states are still components of the universal wavefunction. I think that, if we had an uber-physicist that could look at the superposition of all 2^n states, they would see interference. But there is no uber-physicist. By looking at the screen, the physicist becomes part of the system and hence each of the 2^n physicists is in one of the 2^n sub-states, and can only see the pattern of that sub-state.