RUNHETC-2026-3 What is a Gravitational Path Integral? or Gravitational Path Integrals as Fluctuating Gravito-Hydrodynamics

Tom Banks · 2026

Read RUNHETC-2026-3 What is a Gravitational Path Integral? or Gravitational Path Integrals as Fluctuating Gravito-Hydrodynamics by Tom Banks (2026) 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-2601-10834.txt.

Introduction

Tom Banks asks a basic question: what is a gravitational path integral? Such integrals are used to compute quantities in quantum gravity, but their meaning is often left vague. Banks shows how formulas in the literature can follow from a few coarse-grained assumptions. The assumptions link spacetime geometry, entropy, and fluctuations of the modular Hamiltonian of causal diamonds.

A causal diamond is the region a single observer can see and affect. Treating it like a hydrodynamic system, with flow and entropy, grounds the path integral in concrete physics. The view may clarify why gravity computations work when they do. It speaks to theorists who want quantum gravity built on stated premises, not formal recipes.

Catalog record

CREATORTom Banks
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE38 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-10834

Browse the full library · About the Tower of Babel

Full text

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

RUNHETC-2026-3 What is a Gravitational Path Integral? or Gravitational Path Integrals as Fluctuating Gravito-Hydrodynamics Tom Banks Abstract We show how Gravitational Path Integral formulae for various quantities that have been computed in the literature, follow from a few coarse grained hydrodynamic assumptions about the relations between space-time geometry, entropy, and fluctuations of the modular Hamiltonian of causal diamonds. These remarks have implications for the way we think about such path integrals in relation to a more fundamental model of quantum gravity, and to questions about which space-time topologies are actually summed over in real models. 1 Gravitation, Hydrodynamics and Path Integrals An enormous amount of recent work, following the seminal paper of Saad, Shenker and Stanford [ 1 ] has been devoted to deriving coarse grained quantum properties of quantum gravitational systems from Euclidean gravitational path integrals. There has, in the present author’s opinion, been a lot of unnecessary confusion about the meaning of the gravitational path integral in much of this literature. The note [ 4 ] was an attempt to clear this up, but it has not had much impact. The purpose of the present note is to try to make those arguments more precise. Path integrals were invented almost simultaneously to solve two very different kinds of physical processes: the Schrodinger equation [ 2 ] and the time evolution of a Markov process [ 3 ] . They’re relevant because in each of these situations some quantity characterizing the probability of getting from state A to state B in time t, is the sum of the same quantity for each of the possible histories that goes from A to B in time t. In quantum mechanics that quantity is the probability amplitude, while for Markov processes it is the probability itself. In many cases, the resulting equations resemble each other after continuing time to imaginary time. The confusion arises because of the remarkable fact, first fully appreciated by Schwinger, that for many quantum systems one can analytically continue time to imaginary values and get expressions that are much easier to handle mathematically, and then get the correct quantum answers by analytically continuing back. So the question arises: “When we do Euclidean path integrals for gravity, are we doing analytically continued quantum mechanics, or are we doing statistical mechanics/hydrodynamics?”. Much of the literature, following both old ideas from the pre string theory era, and the fact that perturbative string theory and AdS/CFT treat gravitons as quantum mechanical particles, assumes the former, although it is clear that all of the non-perturbative results that have been obtained calculate coarse grained statistical quantities. In addition, many of them violate the factorization properties expected from quantum mechanics, but not from hydrodynamic time averaged quantities. An entirely different point of view emerges if one follows Jacobson’s seminal paper [ 5 ] . This paper shows that Einstein’s equations (dotted into an arbitrary null vector in a Lorentzian space-time satisfying them) are the hydrodynamic equations of the entropy law \langle K\rangle=\frac{A_{\diamond}}{4G_{N}}, (1) for every causal diamond in the space-time. Here K is the modular Hamiltonian of the subsystem represented by the diamond, of whatever model of quantum gravity describes the global space-time. A is the maximal d-2 volume of space-like surfaces in a null foliation of the diamond boundary. Implicit in this definition is an assumption that the geometry defines a particular state, which we will call the empty diamond state of that model, and that finite area diamonds define subalgebras of the operator algebra, which admit density matrices. If general relativity is hydrodynamics, then it seems reasonable to interpret Euclidean gravitational path integrals as part of fluctuating hydrodynamics, rather than an approximate form of analytically continued quantum gravity. Note that this in no way contradicts the perturbative use of the Einstein-Hilbert Lagrangian to compute Feynman diagrams. In many condensed matter systems the low energy excitations are phonons obtained by quantizing the equations of linearized hydrodynamics and doing perturbation theory around them. In the rest of this note we will rephrase Jacobson’s results in an intriguing new way and try to use the work of [ 6 ] [ 7 ] [ 8 ] to write down the fluctuating version of gravito-hydrodynamics. Finally we’ll try to relate this to path integrals for gravity and discuss what kind of topologies one is supposed to sum over. 2 Jacobson Revisited: The Covariant Entropy Principle We have stated Jacobson’s result in language he did not originally use, language that’s perhaps closer to his later paper [ 9 ] . From now on we’ll call the ansatz that a causal diamond has \langle K\rangle=\frac{A_{\diamond}}{4G_{N}} the Covariant Entropy Principle (CEP). It’s a strong form of the covariant entropy bound of [ 10 ] . What we know in familiar asymptotic backgrounds is that the bound is saturated when the diamond is entirely filled with a black hole. The CEP says that this entropy is already there when the diamond is empty. We cannot emphasize too strongly that this is the entropy of the diamond subsystem in the state of the system where the diamond is empty. This seems puzzling until we think about it in quantum field theory (QFT) where an empty diamond means a diamond in the QFT vacuum state. It’s been known since the 1960s that the operator algebra of such a diamond is Type III_{1} in the Murray von Neumann classification, because of the infinite entanglement caused by divergent light cone commutators. Sorkin was the first person [ 11 ] to show this led to an infinite entanglement entropy proportional to the area, but it was only in the 1990s that this was associated with the Bekenstein-Hawking entropy of black holes [ 12 ] . That insight was what finally led to Jacobson’s argument. We’ll see that if we follow through on this line of thought, it leads to an interpretation of energy and momentum in the bulk of a diamond as an entropy deficit - a constraint on fundamental holographic q-bits that live on the diamond boundary. This interpretation holds locally, independent of the values of the cosmological constant (c.c.). It has to be modified on scales larger than the radius of curvature defined by a negative c.c. because in that case, the global empty diamond state is pure. The tensor network/error correcting code construction of the CFT ground state, shows us how this pure state can be built out of units where energy is an entropy deficit. Let us now turn to a rederivation of Jacobson’s result from this slightly novel point of view. We consider a point in a Lorentzian space-time and realize that it could be lying on the maximal area surfaces of many different causal diamonds. Choose one, and vary it slightly by changing the time-like geodesic defining it, or the interval along that geodesic. We can use the Raychaudhuri equation to compute the variation in area of the diamond. Since we are sitting at a maximum of the area, the higher order terms in the equation are negligible and only the Ricci curvature contributes. The variation is \delta S=R_{mn}k^{m}k^{n}/4G_{N}, (2) where k^{m} is the null direction in which the surface has moved. Since this is happening everywhere in space-time, we

Read the complete text