A Tale of Two Hartle-Hawking Wave Functions: Fully… — Plain-Language Explanation

The Big Picture: Two Ways to Take a Photo of the Universe

Imagine you are trying to take a photograph of the entire universe at a specific moment in time. In physics, this "photo" is called the Hartle–Hawking wave function. It's a mathematical recipe that tells us how likely the universe is to look a certain way.

Usually, to take this photo, physicists use a method called a "path integral." Think of this as summing up every possible history the universe could have had to get to that specific moment.

The problem arises when the universe has a boundary (like the edge of a room). In the famous Anti-de Sitter (AdS) universe (a specific type of curved space), the "floor" of our room is open, not closed. This creates a dilemma: Do we fix the walls of the room, or do we let them wiggle?

This paper explores two different ways to handle this, calling them the "Partially Frozen" method and the "Fully Gravitational" method.

Character 1: The "Partially Frozen" Universe (The Strict Architect)

The Setup: Imagine you are building a model of the universe, but you decide to glue the walls of the room down with super-strong tape. You fix the shape and size of the boundary. You don't allow the walls to move or change at all.

How it works: This is the standard way physicists usually work, especially when connecting gravity to quantum mechanics (AdS/CFT). They say, "We will only count the histories where the walls stay exactly where we put them."
The Result: When the authors calculated the "probability" (or norm) of this universe, the math came out clean and positive. It was a nice, real number, just like you'd expect for a probability. No weird surprises.

Character 2: The "Fully Gravitational" Universe (The Wiggle-Room Explorer)

The Setup: Now, imagine you take that tape off. You decide that the walls of the room are made of a flexible, wobbly material. In this scenario, you don't just sum up the histories of the inside of the room; you also sum up every possible way the walls themselves could wiggle, stretch, and change shape.

How it works: This is closer to the original idea of the Hartle–Hawking proposal, where everything is dynamic. Nothing is fixed by hand; even the boundary is part of the gravitational dance.
The Result: When the authors did the math for this wobbly universe, they found something strange. The probability didn't just come out as a positive number. It came out with a weird, imaginary phase factor (mathematically represented as $\pm i$ ).
The Analogy: It's like trying to measure the weight of a balloon, but because the rubber is so stretchy and alive, your scale starts spinning and giving you a result that includes a "ghost" number. It's not "wrong," but it's definitely not the clean, positive number you'd expect for a simple probability.

The "Phase" Problem: Why the Ghost Number Matters

In quantum mechanics, things can have "phases" (like the timing of a wave). Usually, when you calculate the total probability of something happening, these phases should cancel out, leaving you with a nice, real number.

In the "Frozen" Universe: The phases cancel out perfectly. The result is a solid, positive number.
In the "Wiggly" Universe: The phases don't cancel out. They leave behind a "ghost" number (the imaginary $i$ ).

The authors realized this isn't just a quirk of the AdS universe. They looked at the de Sitter (dS) universe (which is more like our actual expanding universe). In dS, the standard calculation also produces this weird "ghost" phase, which has been a headache for physicists for decades because it makes it hard to interpret the universe's probability.

The "Equator" Experiment: Freezing the Middle

To solve the mystery, the authors tried a clever trick on the de Sitter universe. Instead of freezing the entire boundary (like the "Frozen" AdS case), they froze just the equator (the middle line) of the sphere.

The Analogy: Imagine a globe. Instead of freezing the whole surface, you put a rigid ring around the equator. The top and bottom halves can still wiggle, but they are pinned down at the middle.
The Result: When they calculated the probability with this "partially frozen" equator, the weird ghost phase disappeared. The math became clean and positive again.

The Main Conclusion: It's About Control

The paper's big takeaway is that the "ghost phase" problem isn't caused by the universe itself being weird. It's caused by how much freedom you give the boundaries.

If you let the boundary wiggle freely (Fully Gravitational): You get a messy, complex phase. The math is "fully democratic," but the result is hard to interpret as a simple probability.
If you freeze part of the boundary (Partially Frozen): The phase cancels out, and you get a clean, positive probability.

The Metaphor:
Think of the universe as a chaotic jazz band.

Fully Gravitational: Everyone is improvising, including the drummer and the bassist. The music is free, but it's hard to tell if there's a rhythm (the phase problem).
Partially Frozen: You tell the drummer to keep a steady beat (fix the boundary). Suddenly, the whole band locks in, and you can clearly hear the rhythm (the clean probability).

Summary

The authors found that the "phase problem" in quantum gravity is controlled by whether the path integral is fully dynamic or partially frozen.

In AdS (theoretical universe), letting the boundary move creates a phase; fixing it removes the phase.
In dS (our universe), fixing just the equator removes the phase that usually plagues the calculation.

This suggests that to get sensible, physical predictions (like a clear probability for the universe), we might need to "freeze" certain parts of the spacetime boundary, rather than letting everything fluctuate freely.

Technical Summary: A Tale of Two Hartle–Hawking Wave Functions

Problem Statement
The Hartle–Hawking (HH) proposal defines the wave function of the universe via a gravitational path integral (GPI) over compact Euclidean geometries interpolating between "nothing" and a specified spatial slice. While extensively studied in de Sitter (dS) gravity, the proposal faces persistent issues, most notably the "phase problem" in the norm computation. The one-loop corrected sphere partition function in dS gravity acquires a nontrivial phase, $(\mp i)^{D+2}$ , which contradicts the interpretation of the partition function as a positive-definite norm or a state-counting measure.

The paper investigates whether these issues are intrinsic to dS gravity or if they arise from the specific structure of the gravitational path integral. The authors focus on asymptotically Anti-de Sitter (AdS) spacetime, where natural spatial slices are open and possess a boundary. This feature necessitates an additional spacetime boundary component, $B_-$ , leading to two distinct constructions of the HH wave function:

Fully Gravitational HH Wave Function: The metric on the spacetime boundary $B_-$ is integrated over (summed over). This is a natural extension of the original no-boundary proposal to open slices.
Partially Frozen HH Wave Function: The metric on $B_-$ is fixed (Dirichlet boundary condition). This is the standard formulation in AdS/CFT.

The central question is whether the phase problem in the norm computation depends on whether the boundary is dynamical (fully gravitational) or fixed (partially frozen).

Methodology
The authors employ a combination of semiclassical saddle-point analysis and one-loop perturbation theory across different gravity models:

Semiclassical Analysis in AdS $_3$ and AdS $_2$ :
- AdS $_3$ Einstein Gravity: The authors analyze the HH wave function on a hyperbolic disk spatial slice $\Sigma$ with a dynamical boundary $B_-$ . They utilize a minisuperspace approximation parameterized by the area $A$ and boundary length $L$ of $\Sigma$ . They identify real Euclidean saddles for small $L$ and complex saddles (involving analytic continuation to Lorentzian regions) for large $L$ . The selection of the dominant saddle is guided by steepest-descent contour prescriptions requiring positivity of quantities conjugate to the minisuperspace parameters (extrinsic curvature).
- AdS $_2$ Jackiw-Teitelboim (JT) Gravity: A parallel analysis is performed using a minisuperspace ansatz for the dilaton profile and extrinsic curvature. They examine "maximally-symmetric bubbles" and distinguish between $+$ and $-$ saddles based on the sign of the normal derivative of the dilaton, $\partial_n \Phi$ .
One-Loop Norm Computation in AdS $_D$ :
- The authors compute the one-loop correction to the hyperbolic-ball partition function in $D$ -dimensional AdS gravity.
- Fully Gravitational Case: The boundary $B$ is dynamical (Neumann-type boundary conditions arising from the variation of the action). They derive an effective action for the boundary gravitons by solving linearized Einstein equations in the bulk with off-shell boundary conditions.
- Partially Frozen Case: The boundary is fixed (Dirichlet boundary conditions), corresponding to the standard AdS/CFT setup.
Partially Frozen dS Sphere:
- Motivated by the AdS results, the authors investigate a dS sphere partition function where the metric on an equator (a codimension-1 surface) is fixed. They analyze the one-loop correction by gluing two half-spheres with fixed boundary conditions, carefully accounting for the orientation of the Euclidean time direction and the $i\epsilon$ -prescription for Newton's constant.

Key Contributions and Results

Distinction Between Two HH Wave Functions:
The paper explicitly constructs and compares the fully gravitational and partially frozen HH wave functions. In AdS $_3$ and AdS $_2$ , the fully gravitational construction involves summing over the boundary geometry, leading to a specific selection of saddles (e.g., the $s_{small}$ saddle in AdS $_3$ and the $+$ -saddle in AdS $_2$ ) determined by contour prescriptions.
Phase of the Norm in AdS:
- Partially Frozen (Fixed Boundary): The one-loop partition function on the hyperbolic ball with Dirichlet boundary conditions is found to be real and positive. This aligns with the expectation from AdS/CFT, where the partition function is dual to a unitary CFT.
- Fully Gravitational (Dynamical Boundary): The one-loop partition function on the hyperbolic ball with a dynamical boundary develops a nontrivial phase of $(\mp i)^{D+1}$ . This phase arises from the negative modes of the effective action for the boundary gravitons (specifically the trace mode and the conformal Killing vector modes).
Resolution of the Phase Problem in dS via Freezing:
The authors demonstrate that the phase problem in dS gravity is not intrinsic to positive cosmological constant gravity alone. By considering a partially frozen sphere partition function in dS (where the equator is fixed), they show that the one-loop phase cancels nontrivially, resulting in a real and positive partition function.
- The cancellation occurs because the two half-spheres (glued at the equator) require opposite $i\epsilon$ -prescriptions for the analytic continuation of Newton's constant to maintain consistent time orientation. This leads to a product of phases $(-i) \times (i) = +1$ .
Semiclassical Wave Function Behavior:
In both AdS $_3$ and AdS $_2$ , the semiclassical HH wave function peaks at a critical size of the spatial slice ( $L_{crit}$ ). For sizes exceeding this critical value, the saddle points become complex, and the wave function becomes an oscillatory sum of complex saddles, analogous to the behavior in dS gravity.

Significance and Claims
The paper claims that the phase problem associated with the norm of the Hartle–Hawking wave function is controlled by the dynamical nature of the spacetime boundary in the path integral, rather than being an inherent feature of the cosmological constant or the spacetime signature.

Control of the Phase: The presence of a phase is tied to whether the gravitational path integral is "fully gravitational" (summing over boundary configurations) or "partially frozen" (fixing boundary data).
- Fully gravitational setups (AdS with dynamical boundary, dS with no fixed subregion) exhibit nontrivial phases.
- Partially frozen setups (AdS with fixed boundary, dS with a fixed equator) yield positive, real partition functions.
Implication for dS Gravity: The results suggest that the phase issue in dS gravity might be resolved by introducing a "frozen" spacetime subregion (analogous to an observer or a holographic screen) that breaks the time reparameterization invariance. The authors draw a parallel between their fixed-equator construction and Maldacena's observer proposal, noting that both break the Hamiltonian constraint in the gravitational sector, though through different mechanisms.
Holographic Interpretation: The fully gravitational AdS computation is interpreted via brane-world holography, where the dynamical boundary corresponds to a lower-dimensional gravity theory coupled to a holographic CFT. The phase found in the bulk AdS calculation matches the phase expected in the lower-dimensional dS-like theory on the brane.

The authors conclude that physically sensible predictions for the wave function of the universe may require extracting them from a partially frozen GPI, although a canonical choice for the frozen subregion in closed universes (like dS) remains an open question.

A Tale of Two Hartle-Hawking Wave Functions: Fully Gravitational vs Partially Frozen