Work

Precisiekwantumkosmologie

Year: 2026
Type: dissertation
Abstract: This thesis develops precision quantum cosmology further in order to put the framework on firmer theoretical ground. Contributing to this overarching goal, we pursue three main strands: extending a concrete route to dS/CFT beyond the leading two-derivative order, investigating the consequences of refining the no-boundary proposal with the Kontsevich-Segal-Witten (KSW) saddle-point selection criterion, and studying no-boundary wavefunctions beyond the saddle-point level from the path integral perspective in a precise 2D model. In addition, we study real-time dynamics at one loop in quantum mechanics, using methods inspired by quantum cosmology. In the first strand, we extend a route to dS/CFT based on the intertwinement between dS and AdS saddle points in semiclassical quantum gravity. Since there are no local, diffeomorphism-invariant observables in quantum gravity, we study the entropy of the cosmological horizon, related to the sphere partition function by the Gibbons-Hawking proposal. We show that the four-derivative corrections to the Wald entropy agree with those of the sphere partition function. By deforming the Euclidean time contour into the complex plane, we realize the sphere as two copies of Euclidean AdS, glued together by a complex transition region. This transition region regularizes the volume divergences of AdS in precisely the same way as conventional holographic renormalization. Through AdS/CFT, this provides a relation between the dS entropy and the three-sphere free energy of a dual CFT, valid at least to subleading order and, conjecturally, exactly. At subsubleading order, the CFT predicts a logarithmic correction to the entropy that corresponds to a one-loop effect related to zero modes in the bulk. We study this in a particular embedding in M-theory, in which the CFT is the ABJM theory. Quantizing 11D supergravity on the corresponding –S⁴ × S⁷/ℤₖ background, we identify zero modes of the scalar ghosts, which arise from the quantization of the 3-form, and of the superdiffeomorphism ghosts. Owing to the indefinite (4,7) signature of the background, there are, in fact, infinitely many zero modes, which prevents us from matching the CFT prediction. In the second strand, we investigate the consequences of using the KSW criterion on complex metrics to select which saddle points contribute to the no-boundary wavefunction. In particular, we study models of slow-roll inflation and anisotropic models in pure gravity. We find that near-dS saddles are favoured, whereas strong deformations are generically excluded at large volume. As a result, models predicting a tensor-to-scalar ratio above ∼0.08 are effectively excluded when the observable phase of 50-60 e-folds emerges directly from a no-boundary origin. However, this bound can be evaded by extending the potential with a plateau-like region if the total number of e-folds is sufficiently large. We discuss various objections to the KSW criterion, as well as highly nontrivial situations in which it agrees with more microscopic considerations, and speculate on a weaker, topological criterion. In the interlude on quantum mechanics, we present a formalism for studying real-time dynamics at one loop in the semiclassical limit, inspired by quantum cosmology. We apply the saddle-point approximation to the full time-evolved state, which admits a better-behaved semiclassical expansion than the real-time propagator. We demonstrate the formalism by working through a series of examples, reproducing WKB results. We find complex saddle points in real time, real saddle points in complex time, and complex saddle points in complex time. In particular, we find a finite-time, finite-energy analogue of Coleman's bounce, free of zero modes. The discrete multiplicity of multi-bounce solutions causes the probability of remaining in a metastable region to decay exponentially. In the third strand, we study no-boundary wavefunctions in 2D gravity with a positive cosmological constant, coupled to unitary conformal matter. Upon choosing a fiducial background geometry and fixing the Weyl gauge, the theory factorizes into the matter CFT, bc ghosts, and timelike Liouville theory. We study wavefunctions defined by a gauge-invariant insertion on the disk and find Hartle-Hawking-like and Vilenkin-like wavefunctions, depending on the integration contour for the area of the 2D surface. We also comment on a static patch perspective with a timelike feature. The Liouville model shares many features with higher-dimensional Einstein gravity with a positive cosmological constant. In particular, it suffers from the conformal mode problem, which forces the path-integration cycle to be complexified. We discuss the resulting complex phase and how it differs between saddles and ensembles. Taking the semiclassical limit of large matter central charge, we compute the Liouville wavefunctions at one loop and put forward an all-loop conjecture. The result is in tension with a bootstrap analysis of timelike Liouville theory, in which a reflection symmetry is imposed on the vertex operators. We speculate that this is due to an ambiguity in the choice of path-integration cycle. Finally, fixing the trace of the extrinsic curvature, K, instead of the boundary length, we identify a natural K-independent pairing of two disk path integrals. This pairing may serve as the basis for a well-defined inner product in which K provides an intrinsic notion of time.
Source: Lirias
Language: English
Open Access status: green
OpenAlex ID: W7212410597
Cites: 261
Cited by:
Related to: 10