Move the choice into the flight
Fixed analyzer settings leave time for a hypothetical influence to cross an apparatus before detection. The 1982 experiment used an optical switch in each arm to direct light towards one of two fixed polarizers. Changing the route changed the effective measurement setting without mechanically rotating a polarizer.
The novelty is therefore a relation among events, not simply a faster instrument. Ask where the setting becomes relevant, where the distant outcome is recorded, and whether a light-speed signal could connect them. A long cable or a distant-looking photograph is not a spacetime analysis.
Read the evidence: A. Aspect, J. Dalibard and G. Roger (1982), Experimental Test of Bell’s Inequalities Using Time-Varying Analyzers, Phys. Rev. Lett. 49, 1804–1807
Keep the distance and the delay together
For two events in one inertial frame, the simplified condition below is spacelike separation: distance exceeds the distance light can travel during their time difference. The original paper compared roughly ten-nanosecond switching intervals with a roughly forty-nanosecond light-travel time between the two sides.
A real timing audit needs event definitions, detector delays, path lengths and margins, including uncertainty. The instant when an electrical command starts is not automatically the instant at which a photon encounters the selected analysis. Replacing all these events with one convenient timestamp can reverse the conclusion.
Read the evidence: A. Aspect, J. Dalibard and G. Roger (1982), Experimental Test of Bell’s Inequalities Using Time-Varying Analyzers, Phys. Rev. Lett. 49, 1804–1807
A periodic switch is not a random generator
The switches used acousto-optic diffraction with independent drives at different frequencies. The authors explicitly described the changes as quasiperiodic rather than truly random, and called for more ideal random and complete switching. That qualification belongs beside the achievement.
Predictability matters even when two measured sequences have little ordinary correlation. A rule known in advance can let a model anticipate a future setting. By contrast, a fast unpredictable setting still needs suitable placement in spacetime; randomness cannot repair an outcome that had enough time to receive the distant choice.
Read the evidence: A. Aspect, J. Dalibard and G. Roger (1982), Experimental Test of Bell’s Inequalities Using Time-Varying Analyzers, Phys. Rev. Lett. 49, 1804–1807Aspect (2022), Experimental tests of Bell’s inequalities: a first-hand account
Do not borrow the wrong Bell statistic
The paper’s normalized rate combination had a local upper bound of zero and a reported value 0.101 with uncertainty 0.020. Its symbol should not be compared numerically with a binary CHSH statistic whose upper bound is two. A shared letter does not make two estimators identical.
As a reading exercise, put the outcome alphabet, normalization and auxiliary exposures next to every displayed inequality. Then identify whether losses are outcomes, ignored events or handled through an additional assumption. This one-page ledger prevents a large violation in one convention from being mistaken for a small violation in another.
Read the evidence: A. Aspect, J. Dalibard and G. Roger (1982), Experimental Test of Bell’s Inequalities Using Time-Varying Analyzers, Phys. Rev. Lett. 49, 1804–1807
Counts and corrections remain part of the claim
The coincidence signal had a background component, estimated using delayed events. Such a correction needs its own measurement procedure. Subtraction is not inherently invalid, but a reader should know what was subtracted and how its uncertainty enters the final comparison.
Separate two hypothetical changes. Moving the delayed window can test background estimation; increasing detection efficiency changes which emitted pairs become observable. The first cannot certify the second. Timing, optical throughput and count analysis are connected apparatus properties that answer different inferential questions.
Read the evidence: A. Aspect, J. Dalibard and G. Roger (1982), Experimental Test of Bell’s Inequalities Using Time-Varying Analyzers, Phys. Rev. Lett. 49, 1804–1807
An advance can have explicit unfinished conditions
Aspect’s later account places the switching experiment within a programme of source and analyzer improvements. The historical lesson is cumulative: a new geometry can test a previously accessible communication route while leaving other assumptions for subsequent experiments.
The existing Bell teaching model explores correlations and sampling. It does not contain the 1982 acoustic standing wave, measured switching profile or full event timing. Use the model for its declared mechanism, then return to the paper for the apparatus claim; neither supports a controllable faster-than-light message.
Read the evidence: Aspect (2022), Experimental tests of Bell’s inequalities: a first-hand account