How a Later Choice Changes What the Experiment Can Show
In a delayed-choice experiment, a particle such as a photon is sent through an arrangement with two possible paths. The paths are made to overlap at a detector, where an interference pattern can appear: alternating regions of higher and lower detection probability, like the pattern produced when waves combine.
The key question is whether the setup preserves clues about which path the photon took. If the paths leave distinguishable records, even in another particle linked to the photon, the experiment can reveal path information. If the setup instead makes those paths indistinguishable, the results can show interference. The later choice is a choice about which relationship the measurement will test, not a switch that reaches back and alters a stored trajectory.
The Evidence Depends on the Measurement
In some versions, the photon becomes entangled with a partner, meaning the two particles’ measurement results are connected. A later measurement of the partner can sort the photon detections into different groups. In one grouping, the results show interference; in another, they do not. Viewed all together, the detections do not form a usable interference pattern that changes when the later choice changes.
That distinction matters. The experiment does not let anyone send a message into the past or change a result that was already recorded. It shows that quantum experiments do not always support the everyday assumption that a particle carried a definite, independently knowable path all along. What can be said about the path depends on what information the full measurement arrangement makes available.
The surprising point is therefore not that the future rewrites history. It is that, at the quantum scale, the question an experiment can answer is built into how its evidence is collected.
