How do self-interaction potentials stabilize ghost fields without requiring stealth solutions?

This is a genuinely sophisticated question about an apparent paradox in field theory: ghost fields (degrees of freedom with backwards kinetic energy) are classically unstable, yet self-interactions seem to suppress that instability without invoking "stealth" mechanisms (where ghosts decouple at short distances).

How self-interactions stabilize ghosts

The key finding is that certain nonlinear potentials can postpone or arrest the ghost instability indefinitely by controlling how energy flows between normal and ghost sectors.

In a two-field system where one field is ghostly (negative kinetic energy), the natural instability works like a runaway coupling: energy cascades from the normal field into the ghost, which radiates it into increasingly high-frequency modes. However, self-interactions can interrupt this cascade [AG-2025.04-686]. For example, a lifted φ⁶ potential creates transient metastable regimes that partially suppress ghost-induced growth by generating oscillon-like bound structures—localized, quasi-periodic solutions that trap energy [AG-2026.04-1423]. The mechanism isn't mysterious: the nonlinear potential acts as a "speed bump" in phase space, forcing the system to spend longer in classically bounded configurations before instability takes over.

More remarkably, sufficiently strong self-interactions can achieve global stability for all initial data, not just small perturbations [AG-2025.09-466]. Numerical scattering simulations in spherical symmetry show that certain polynomial potentials cause the ghost instability to be quenched classically. The physical picture: the nonlinearity is dominant enough that it controls the dynamics before the ghost can fully couple.

Why this works without stealth

Stealth mechanisms typically hide ghosts by making their couplings vanish under specific kinematic conditions (e.g., during inflation). Self-interactions take a different approach: they remain active but reshape the energy landscape so that the ghost sector becomes a temporary energy sink rather than a runaway drain [AG-2026.04-1423]. Spectral analysis reveals the reason: instability is mediated by nonlinear energy transfer, not instantaneous exponential growth. High-frequency modes are naturally more stable than low-frequency modes at fixed amplitude, so the potential can be tuned to keep the ghost in the UV-dominated regime longer [AG-2025.04-686].

At the quantum level, a related stabilization occurs: an integral of motion with positive discrete spectrum can enforce stability for extended interactions by constraining the Hilbert space structure [AG-2026.04-561]. This is distinct from stealth because the ghost remains coupled; unitarity and boundedness emerge from the algebraic structure, not from decoupling.

The catch

Classical stability ≠ quantum stability. While self-interactions can prevent runaway classical instability, quantum ghosts still face issues with unitarity and vacuum definition in general theories [AG-2024.08-594]. However, in special cases (certain superrenormalizable theories or those with protective gauge symmetries), the ghost can be confined into composite normal particles [AG-2024.08-474] or the theory can be reformulated to retain only physical degrees of freedom [AG-2024.10-448], sidestepping the unitarity problem entirely.

The unifying theme: self-interactions don't eliminate the ghost instability—they tame it by making it gradual, nonlinear, and controllable, opening space for quantum mechanisms to complete the stabilization.