If quantum gravity doesn't entangle with mass superposition, what alternative mechanism could produce the observed repulsion?

Your question asks about repulsion *without* entanglement between the mass and the gravitational field itself. The archive contains exactly this scenario.

The Weak-Value Mechanism

The key insight is that repulsion can arise from quantum interference of gravitational forces even when the mass and gravity don't become entangled. Here's how:

Imagine a source mass in a spatial superposition—say, "here" and "there" simultaneously [AG-2026.02-534]. Classically, a probe mass would feel an attractive pull toward each position. But quantum mechanically, these two gravitational forces interfere with each other, like ripples on a pond crossing and creating patterns of reinforcement and cancellation.

When you perform the right measurement—specifically, a *postselection* that confirms the source mass ended up in a particular location—the interference pattern becomes asymmetric [AG-2026.02-268]. The probe mass ends up pushed *away*, not pulled in. This is not an error or artifact; it's a genuine prediction of quantum mechanics using the formalism of "weak values," which describe measurement outcomes conditioned on specific post-experiment observations.

Why This Matters

This repulsion is a signature of quantumness itself [AG-2026.02-268]: a purely classical gravitational field cannot produce repulsion, because classical forces always add as vectors—if gravity pulls from two directions, the net result is still inward. Only quantum superposition, with its interference effects, enables the reversal.

Importantly, this mechanism works *without* requiring that the gravitational field and matter become entangled during the interaction [AG-2024.05-603]. The gravitational field remains, in some models, a dependent or auxiliary variable rather than an independent quantum system that gets correlated with the mass [AG-2024.05-605].

A Caveat on Feasibility

The practical challenge is that these effects are small and require careful state preparation and postselection. The papers outline proposals using spin-bearing nanocrystals [AG-2026.02-534], but the required masses, separations, and measurement sensitivities remain at the edge of what is experimentally possible.