Semiclassical Causality, Chronology, and Quantum Horizons

This chapter asks what causality means when quantum matter sources one semiclassical metric and when fluctuations begin to blur operational propagation. It moves from causal response, through chronology horizons and negative-energy wormholes, to stochastic horizon observables and the boundary where one mean geometry no longer closes the question. The central discipline is to match each conclusion to its contour, state, horizon class, smearing, and approximation order.

Helpful background. Quantum Fields in Curved Spacetime supplies states and microlocal regularity; Einstein Equations supplies constraints and response; Black-Hole Geometry and Mechanics supplies horizon generators; Semiclassical Gravity supplies mean and stochastic validity; and Causality and Global Structure supplies domains of dependence and Cauchy horizons.

Three levels of causal statement

At the first level, quantum fields respond on a fixed or self-consistent mean metric. Holding the initial density matrix fixed, the correct initial-value equation uses an in-in closed-time-path contour and a retarded kernel,

$$\delta\langle T_{ab}(x)\rangle =\int dV_{x'}\, \Pi^{\rm ret}_{ab}{}^{cd}(x,x')h_{cd}(x'), \qquad \operatorname{supp}\Pi^{\rm ret}(x,\cdot)\subseteq J^-(x).$$

An independent change of the initial state adds a separate response term. The kernel and its normalization are developed from the influence action in Hu and Verdaguer 2008, §§ 4.1–4.3, PDF.

At the second level, stress fluctuations induce correlations of smeared, gauge-invariant metric observables. A controlled Gaussian approximation or additional higher moments can turn these into an arrival-time or escape-probability distribution. For example, detecting a signal within one year defines a finite-time success event; failure does not establish that the emitter lies beyond a global event horizon. At the third level, different quantum-geometric branches may have different causal orders or topologies. No retarded kernel on one mean manifold can answer that exact microscopic question.

Chronology claims cross the same levels. A divergent fixed-background stress is evidence for backreaction, not a solved backreacted spacetime. A microlocal obstruction at a compactly generated Cauchy-horizon base point is a theorem under exact hypotheses, not a theorem that all closed timelike curves are forbidden. Keeping these logical levels distinct is the chapter’s main safeguard.

Six pages in causal order

The pages lead from a controlled initial-value problem to the limits of a classical causal geometry:

1. Causal Response in Semiclassical Equations derives the closed-time-path equation, defines the retarded polarization convention, and checks Ward identities and support.
2. Chronology Horizons and Stress–Energy Growth explains image-sum blueshift, finite-stress exceptions, Hadamard failure, and the difference between prescribed-background evidence and prevention.
3. Chronology Protection: Theorem and Conjecture states compact generation, base points, F-locality, and the two Kay–Radzikowski–Wald conclusions without turning them into a universal verdict.
4. Traversable Wormholes and Quantum Sources derives throat null-energy violation and compares quantum inequalities with controlled boundary-coupled constructions.
5. Fluctuating and Smeared Horizons replaces a sharp stochastic coordinate radius with relational timing, redshift, expansion, or escape distributions.
6. Semiclassical Causal Limits and Quantum-Gravity Questions identifies topology change, interfering geometries, exact locality, and nonperturbative dressing as questions for Quantum Gravity.

The order matters. One must first establish causal response on a mean solution, then test fluctuations of operational observables, and only then decide whether the one-geometry description has failed.

Recurring hypotheses and boundaries

Question	Data that must be fixed	What the result does not automatically show
Is response causal?	Initial density matrix, CTP contour, retarded prescription, gauge, counterterms	Stability or small fluctuations
Does stress grow near chronology formation?	Field, state, quotient, subtraction, frame, approach curve	A self-consistent avoidance geometry
Does KRW apply?	Compact generation, base point, scalar field, $D(S)$ Hadamard state, extension and F-locality hypotheses	Universal chronology protection
Is a wormhole source controlled?	Global state, sampling, boundary coupling, full stress, cutoff, fluctuations	Topology creation or nonlinear stability
Is a horizon “smeared”?	Relational probe, detector, state, smearing kernel, confidence level	An observer-independent fuzzy membrane

The Kay–Radzikowski–Wald result is the sharpest stop rule in the chronology discussion: an initial Hadamard Klein–Gordon state cannot extend as a locally Hadamard distributional bisolution at base points of a compactly generated horizon, and the usual algebra cannot extend there with F-locality Kay, Radzikowski, and Wald 1997, § 5, Theorems 1–2, PDF. The missing conclusion is equally important: the theorem does not specify the preceding self-consistent geometry.

Reader routes and checks

For a response route, read pages 1 and 5 and verify retarded support, the Ward identity, positive smeared variance, and gauge-invariant endpoints. For a chronology route, read pages 2 and 3 and compare every claim on four axes: statement, hypotheses, logical status, and missing conclusion. For a wormhole route, read page 4 with the volume’s energy-inequality material, ensuring the sampling scale lies below curvature and boundary scales. To study the limits of classical geometry, read pages 5 and 6 and ask whether all relevant branches share one manifold and causal relation.

Useful null checks recur. Removing a quotient should recover the reference vacuum rather than leave an image divergence. Turning off a double-trace boundary coupling should remove the associated traversable window. Taking a narrow-state, weak-fluctuation limit should restore causal propagation on the mean metric. A broad coordinate variance that disappears in a relational observable is a gauge artifact, not new causal physics.

Review the chapter

A proposed time-machine model has a compactly generated $H^+(S)$. A scalar state is Hadamard on $D(S)$, and one tetrad component of its renormalized stress remains finite along a chosen curve toward a base point. May one conclude that the state extends as a regular local QFT across that point? May one conclude that backreaction prevents the horizon?

Solution

No to both. Finiteness of one coincident-limit component along one curve is weaker than bounded difference from a local Hadamard distribution in a neighborhood of $(x,x)$. Under the stated extension hypotheses, KRW Theorem 2 gives a non-Hadamard obstruction at every base point. The theorem and the finite component do not solve the backreacted semiclassical equations, so neither determines whether the horizon is prevented, replaced, or reached at the edge of validity.

What completion enables

A reader can now derive a causal mean-field response, state exactly what chronology theorems establish, test negative-energy geometries without treating a QEI as universal, and replace coordinate horizon fluctuations with measurable distributions. Just as importantly, the reader can recognize when causal order itself depends on quantum geometry and pass a precise last-controlled record to the neighboring theory rather than extrapolate one metric beyond its domain.

References

• Hu, B. L., and Verdaguer, E. (2008). “Stochastic Gravity: Theory and Applications.” Living Reviews in Relativity 11, 3. DOI. Open PDF.
• Kay, B. S., Radzikowski, M. J., and Wald, R. M. (1997). “Quantum Field Theory on Spacetimes with a Compactly Generated Cauchy Horizon.” Communications in Mathematical Physics 183, 533–556. DOI. Open PDF.