The experiment says that if I split a photon into two coherent entangled photons, one towards a double-slit, one towards a which-way detector, I can control whether

  • the interference pattern appears by choosing either detecting which way the photon came through or
  • destroying the information about the photon, even if the detector/destroyer is further away than the double slit screen.

Does it imply that I can know about whether the which-way information will be detected or not even before it is detected?

Suppose I have a person controlling the detector 10 ly away from the slits, and the double-slit setup immediately in front. Then just by looking at the screen, will I be able to know whether the person 10ly away decides to observe or not 10 years before he even decides it? I am quite uncomfortable with this thought.

  • 1
    $\begingroup$ Does this answer your question? Why doesn't delayed choice quantum eraser experiment imply retrocausality? My answer on that one points out that you only have to deal with the subtleties of retro-causality if you try to make sense of the experiment without quantum mechanics. If you use QM properly, its not all that unexpected of a result, and no soothsaying is required. $\endgroup$
    – Cort Ammon
    Nov 7 '20 at 5:50
  • $\begingroup$ I get that the photons go all paths at once until a measurement is done. However, I am still grappling with what is going on at the double slit detector before any "which-way" information is detected. Is there any concrete example of what i am going to detect while the other entangled supposed "photon" have yet to reach the which-way detector? $\endgroup$ Nov 9 '20 at 7:19
  • $\begingroup$ What you are going to get is a detection which looks completely random, without any useful information in it. The information which makes things interesting only appears when you use the which-way detector results to process your results in a statistical way, carefully selecting a subset of points based on this new which-way information. The Wikipedia page on the delayed choice quantum eraser shows exactly what you would detect in section 2.1. $\endgroup$
    – Cort Ammon
    Nov 9 '20 at 7:22

No, you cannot use this to communicate faster than light, or get information from the future. I tried to think a way to do that. Turns out, you need to analyse both sets of data from both sites to get to see if interference patterns appears or not. At the double-slit site, you get mush (no interference), only by combination, careful analysis of correlation might the interference pattern might emerge.

So when bringing together both sides, it's clearly already beyond the time delay of the choice and slower than light.


Not the answer you're looking for? Browse other questions tagged or ask your own question.