RUNHETC-2025-36 Finite Entropy Implies Finite Dimension in Quantum Gravity

Tom Banks · 2025

Read RUNHETC-2025-36 Finite Entropy Implies Finite Dimension in Quantum Gravity by Tom Banks (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-17856.txt.

Introduction

T Banks argues that a finite heat-like count in quantum gravity means a finite count of states. His 2025 paper builds on work by Jacobson, Fischler, Susskind, and Bousso. Black hole facts hint that a patch of space-time holds just so much information.

Banks says string models of de Sitter space, with endless states, miss the key point. The hard part is to define how such a space decays and what far-off measures show. For times short next to its life, he argues, a finite-state model can describe the space well. The claim is a stance in a live debate.

Catalog record

CREATORTom Banks
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE20 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-17856

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

RUNHETC-2025-36 Finite Entropy Implies Finite Dimension in Quantum Gravity Tom Banks Abstract Quantum Field Theory (QFT) introduced us to the notion that a causal diamond in space-time corresponded to a subsystem of a quantum mechanical system defined on the global space-time. Work by Jacobson [ 1 ] , Fischler and Susskind [ 2 ] and particularly Bousso [ 3 ] suggested that in the quantum theory of gravity this subsystem should have a density matrix of finite entropy. These authors formalized older intuitive arguments based on black hole physics. Although mathematically, Type II von Neumann algebras admit finite entropy density matrices, the black hole arguments suggest that the number of physical states in these subsystems is finite. The conjecture that de Sitter (dS) space has a finite number of physical states was first made in [ 4 ] [ 5 ] . “String theory constructions” of “meta-stable dS states” always involve a model with an infinite number of states. The real challenge of such models is to define the asymptotic correlation functions of the space-time to which the dS space decays, and show how to isolate properties of the dS “resonance” from those correlators. We argue that the theory of the resonance itself is adequately described, on time scales shorter than its lifetime, by a model with a finite number of states. 1 Causal Diamonds and Entropy Since the advent of local QFT, we have become used to associating causal diamonds in space-time with subsystems of the QFT Hilbert space. This is true in the sense that they possess a von Neumann sub-algebra with non-trivial commutant in the full operator algebra, but it is well known that these algebras are Type III_{1} and have no density matrices. The ideas of Bekenstein-Hawking-Jacobson-Fischler-Susskind and Bousso, which associate finite entropy to these regions in a hypothetical theory of quantum gravity show us that something drastically different must happen in such a theory. A very old hand waving argument 1 1 1 I first heard this from G. ’t Hooft in the 1980s, but I’m sure it pre-dates that. shows that in the semi-classical approximation most of the states that even a cut-off QFT assigns to a diamond back-react on the geometry to form a black hole larger than the diamond. Eliminating such states leaves over a QFT entropy that scales like (A/4G_{N})^{\frac{d-1}{d}} 2 2 2 Cohen, Kaplan and Nelson [ 11 ] showed that eliminating these states was entirely compatible with the experimental success of QFT. . This is not even enough to account for the entropy that Bekenstein and Hawking assign to a black hole and Jacobson (effectively) assigns to any causal diamond. Indeed, ’t Hooft’s proposal to model black hole entropy in terms of near horizon QFT states ran into trouble [ 12 ] with back reaction long before the firewall problem proved it could not work [ 13 ] . A lot of recent work, following the beautiful papers of [ 17 ] has been based on the proposal [ 18 ] that one could somehow systematically correct those results by using the crossed product construction of a Type II algebra from the Type III QFT algebra. In [ 18 ] , the crossed product was taken with the action of the boundary Hamiltonian, which [ 17 ] had shown did not act on the LL algebra at leading order in 1/N . Indeed, that lack of action was the reason the subalgebra of a causal diamond has a non-trivial commutant. There are a few disturbing things about this proposal, whose purpose was to construct an algebra with a trace so that entropy formulas made sense. The most striking is that the crossed product algebra contains the original Type III algebra as a subalgebra, whereas we expect the effect of quantum gravity is to reduce the number of observables in a finite region. The original idea of the crossed product came from the existence of the modular group of automorphisms of a Type III_{1} von Neumann algebra \Delta^{it} and the fact that the modular operator \Delta=e^{-K} had many of the properties of a density matrix, but was not an operator in the algebra itself. Physically, the reason that diamond subalgebras in QFT do not have density matrices is because of the infinite entanglement of the diamond with states exterior to the diamond but arbitrarily close to the diamond boundary. This has nothing to do with gauge invariance. It is an issue in theories with no gauge invariance or in the algebra of local gauge invariant operators in Yang-Mills theory. If we look at the sector of the gauge theory not invariant under local gauge transformations, we have an additional issue of entanglement with gauge degrees of freedom on the boundary. The crossed product construction adds enough elements to the algebra such that density matrices like e^{-K} become elements of the algebra. In a gauge theory this should include the non-local fields that implement the entanglement with the boundary. The authors of [ 23 ] have proposed a method of doing this systematically, in a particular gauge, to all orders in the 1/N expansion. The proposal of [ 18 ] appears to mix this up with the UV entanglement of the boundary field theory. In a correct model of quantum gravity that UV entanglement should simply disappear. The work of [ 12 ] [ 13 ] shows that even a simple cutoff on the near horizon states in field theory is not the right way to treat it. Indeed, we have known for some time [ 14 ] [ 15 ] that the proper way to treat entanglement of the bulk with the boundary in AdS/CFT is through the tensor network/Error Correcting Code formalism. This inevitably introduces a UV cutoff length of order the AdS radius on the bulk field theory. The interpretation of this formalism in terms of causal diamonds appeared in [ 16 ] and we are beginning to understand its connection to the work of [ 17 ] [ 25 ] 3 3 3 The arguments of [ 17 ] take place in the strict N=\infty limit. However, to avoid confusion with the Type III operator algebras in the boundary CFT for finite N , the authors argue that they are working with a finite ultraviolet cutoff. If one chooses that cutoff to be a lattice model of the sort used in tensor network constructions then on small enough lattices the arguments of [ 17 ] , which invoke the infinite N bulk generalized free field algebras of [ 22 ] [ 23 ] are not even approximately correct. In [ 25 ] we speculated that the correct framework for the work of [ 17 ] was a double scaling limit where the tensor network shell is taken to the boundary as N is taken to infinity. . In the current note, we want to present a simple argument based on semi-classical black hole physics that the number of states in a black hole or any other finite area causal diamond in Minkowski or AdS space, is finite. To make this argument we want to start from a model of quantum gravity that everyone believes in, and which

Read the complete text