Why do quantum superpositions of the gravitational field produce repulsion through interference rather than attraction?

This is a subtle question that hinges on quantum interference—and the preprints reveal that repulsion doesn't come from superposition *per se*, but from how we measure it.

The interference mechanism

The key insight is that repulsion arises only under conditional measurement (postselection). Here's the physics: when a source mass is prepared in a spatial superposition and then *measured in a specific location*, the probe mass responds as if it had interacted with a peculiar gravitational configuration [AG-2026.02-268]. The two possible paths the source could have taken—which would each pull the probe inward—can interfere destructively, producing a net repulsive force.

This is mediated by what physicists call a "negative weak value" [AG-2026.02-534]. Weak values are statistical artifacts that emerge when you ask: "Given that I measured the source here, what was the average force the probe felt?" The answer can be anomalously large and point in an unexpected direction—even opposite to the classical expectation. It's not that gravity has flipped; rather, the conditional average over a carefully selected subset of outcomes produces repulsion.

Why this matters for quantum gravity

The crucial point is that classical gravity cannot do this. A classical gravitational field, no matter how it's distributed, always pulls [AG-2026.02-268]. Repulsion is therefore a signature of quantum interference of distinct gravitational configurations—evidence that spacetime itself exists in a superposition [AG-2026.02-534]. This makes the effect a potential experimental witness to gravity's quantum nature.

There is also a different mechanism at play in linearized quantum gravity: certain quantum processes—specifically cross-coupling between gravitoelectric and gravitomagnetic polarizabilities—can produce genuine repulsive interactions at short range [AG-2025.01-366]. Here, quantum vacuum fluctuations of the gravitational field induce opposing dipole moments that repel.

A note on feasibility

The postselection-based repulsion requires exquisite control: the experiments are proposed with spin-bearing nanocrystals but remain extremely challenging [AG-2026.02-534]. Whether such experiments can be realized without overwhelming decoherence is still open [AG-2024.05-603].