Does the postselection measurement need to happen before or after the gravitational interaction occurs?

Your question hinges on the timing of measurement relative to the gravitational interaction—a subtle but crucial point in quantum gravity experiments.

The answer is: **the postselection must happen *after* the gravitational interaction** [AG-2026.02-534, AG-2026.02-268].

Here's why this matters. In these experiments, a source mass is first prepared in a spatial superposition, then allowed to gravitationally interact with a probe mass while both remain in their quantum states. During this interaction, the probe experiences a superposition of gravitational forces—one from each location of the source. Only *after* this interaction completes do you postselect (measure and condition on) the source mass being in a specific location.

The remarkable result is that this *retroactive* postselection creates an effective repulsive gravitational force on the probe [AG-2026.02-534]. This seems counterintuitive: how can a measurement you make *later* change what happened *earlier*? The key is that you're not actually changing the past; rather, you're conditioning your analysis on a specific measurement outcome. The quantum interference between the two gravitational field configurations (from the two possible source locations) only manifests when you postselect one outcome [AG-2026.02-268]. Before postselection, the probe has experienced an entangled superposition of forces; after postselection, that superposition collapses, and the interference pattern reveals repulsion—which classically would be impossible [AG-2026.02-268].

This is achieved via negative weak values, a technique from quantum measurement theory where conditioning on a specific detector outcome reveals information that seems to violate classical intuitions [AG-2026.02-534].