How do path-integral and curved-spacetime QFT approaches differ mechanically in their propagator calculations?

# Path Integrals and Curved-Spacetime QFT: How Propagators Differ

The two approaches—path-integral quantization and curved-spacetime QFT—calculate propagators in fundamentally different ways, and they can actually give *different answers* for the same physical system.

Path-Integral Approach

In the path-integral picture, the propagator is built by summing over *all possible histories* of a particle or field between two spacetime points. You weight each path by a phase factor determined by the classical action [AG-2024.07-1817]. For a scalar field, this is conceptually straightforward in flat space, but when you try to extend it to curved spacetime, you must parametrize paths along the geometry itself—treating geodesics as the "classical paths" that dominate the sum [AG-2026.01-638]. The extended Lagrangian formalism allows one to parametrize evolution along an abstract parameter rather than coordinate time, making the treatment more general [AG-2024.04-1725].

The key practical challenge is that real-time path integrals are "oscillatory and high-dimensional," making them numerically expensive—recent work uses Picard-Lefschetz theory and eikonal approximations to make this tractable [AG-2025.01-648].

Curved-Spacetime QFT Approach

In curved-spacetime QFT, the propagator emerges from solving the quantum field equations (like the Klein-Gordon or Dirac equation) in a given metric background. Here, spacetime curvature enters *explicitly* through covariant derivatives and the metric tensor itself. The propagator becomes spatially dependent: gravitational fields modify Feynman diagrams by warping propagators and interaction vertices in ways that depend on location [AG-2024.11-199].

The Critical Difference: They Don't Always Agree

A striking result: when you compute the propagator for a particle in the far-horizon region of a black hole using curved-spacetime QFT formalism, it *differs* from the propagator obtained via path-integral methods [AG-2025.12-453]. This tension reveals that the two formalisms, while both rigorous, encode different physical assumptions about what "quantization in curved spacetime" means—especially regarding the ambiguity of particle definition in strong-field regimes.

In weak fields, both methods agree and align with low-energy experiments (like quantum interference in Earth's gravity). But they diverge in strong-curvature regions, signaling that neither formalism alone fully captures quantum gravity.

---

Note: The provided preprints address specific calculational techniques and applications, but don't give a detailed side-by-side mechanical comparison of propagator kernels in the two formalisms—that deeper contrast would require references to standard textbooks on curved-spacetime QFT and path-integral quantization.