Does requiring stealth solutions to exist actually constrain which coupling functions avoid ghost instabilities?

This is a sophisticated question about the interplay between stealth solutions and ghost stability—let me work through what the archive reveals.

The short answer: Yes, requiring stealth solutions to exist does impose constraints on which coupling functions avoid ghosts, but the relationship is subtle and context-dependent.

What stealth solutions tell us

Stealth black hole solutions are configurations where a nontrivial field (scalar, vector, or otherwise) couples to gravity yet leaves the spacetime metric unaffected [AG-2025.06-406]. Physically, this means the field's stress-energy tensor vanishes—it decouples from Einstein's equations.

The reason this matters for ghost-avoidance: In higher-order Maxwell-Einstein theories, requiring that generic coupling functions *admit* stealth solutions (like the Reissner-Nordström metric) acts as a selection rule. For instance, [AG-2025.06-406] shows that "for several specific choices of the coupling functions" you can recover standard GR solutions as stealth configurations in richer theories. Those specific choices are the ones that avoid propagating extra ghostly degrees of freedom in those particular limits.

The constraint mechanism

A deeper perspective comes from the degenerate-theory literature. [AG-2025.11-518] analyzes quadratic non-metricity theories and finds that "the theory generically exhibits ghost instabilities" *unless* you land in special subclasses—notably the Symmetric Teleparallel Equivalent of General Relativity (STEGR). The stealth-like property of STEGR (it reproduces Einstein equations without extra fields) is intimately tied to why it avoids ghosts in that context.

Conversely, [AG-2024.10-344] flips the logic: by imposing stealth conditions *first* and asking which geometries and field configurations satisfy them, the "reverse stealth construction" shows that "a gravity model can accommodate a stealth field only if its vacuum solution falls within the geometries permissible under stealth conditions." This is a hard constraint on which theories can have stealth solutions at all.

When stealth isn't the whole story

However, stealth isn't the only way to evade ghosts. [AG-2025.09-466] and [AG-2025.04-686] demonstrate that ghost instabilities can be quenched by strong self-interactions, without requiring the field to be a stealth solution. For instance, [AG-2025.09-466] identifies "classical ghostly field theories with polynomial interaction potentials for which small-data global stability apparently extends to large-data global stability," even when the field does couple back to the geometry.

And [AG-2026.04-561] shows that quantization itself can stabilize a ghost: a harmonically trapped ghost coupled to matter can have a well-defined unitary quantum evolution with a bounded vacuum, provided an integral of motion has a discrete positive spectrum.

The synthesis

So the constraint is real but partial: requiring stealth solutions to exist *does* narrow the space of allowed coupling functions—it's a sufficient (often elegant) condition for ghost-avoidance in certain theories. But it's not necessary. Strong enough nonlinearities or quantum effects can stabilize ghosts in non-stealth configurations. The coupling function must ultimately satisfy either stealth conditions, strong-coupling resummation, or a conserved-quantity structure [AG-2024.08-485]—but these are distinct pathways.