Stokes Phenomena between AdS/CFT and dS/CFT

arXiv:2609.10677 · hep-th, gr-qc · Submitted 2026-09-09 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: High Energy Physics: Theory. The day in high-energy theory: gravity, strings, fields and the questions left open.

Kai: We're looking at the paper: "Stokes Phenomena between AdS/CFT and dS/CFT".

Title and authors: Kai: So we're looking at how these two holographic ideas connect, using this paper "Stokes Phenomena between AdS/CFT and dS/CFT". We've seen it’s about saddle points changing when you analytically continue things.

Mira: It’s trying to establish a holographic consistency check by comparing the bulk path integral to the Liouville zero-mode integral. That suggests the dS contour is homologous to just one thimble, but the AdS contour becomes an infinite sum of those thimbles under continuation.

Lev: That difference in saddle content, that shift from one single saddle to an infinite family, that’s where things get tricky for building real systems. It sounds like a serious mathematical hurdle to cross before we can even think about applying this to hardware constraints.

Kai: Right. The paper is exploring this specific problem in the context of the analytic continuation between AdS/CFT and dS/CFT correspondences, using three-dimensional pure gravity and two-dimensional Liouville theory.

Mira: And they test this condition by comparing the bulk and boundary theories in their spherically symmetric sectors. They define the dS3 minisuperspace path integral from a positive Lorentzian lapse, determine its Lefschetz thimble decomposition, and then define the AdS3 contour by analytically continuing the radius l of AdS and a bulk cutoff parameter.

Lev: So they're setting up these two separate paths to see if they end up in compatible places when you change those parameters. It sounds like a lot of bookkeeping just to get the setup right.

Kai: And on the boundary, they reduce Liouville theory to its constant mode, and then define the AdS and dS integrals by analytically continuing the central charge c and the boundary cutoff parameter epsilon. Both calculations independently lead to Gamma functions exhibiting that same Stokes phenomenon under analytic continuation between dS and AdS.

The paper's summary: Mira: So, to summarize what they found in "Stokes Phenomena between AdS/CFT and dS/CFT", they showed that both the bulk calculation and the boundary calculation of Liouville theory independently exhibit the same Stokes phenomenon when analytically continued between dS and AdS.

Kai: That means if holography is supposed to hold, then the bulk path integrals should select compatible integration cycles, but here we see one side selects a single thimble while the other selects an infinite sum.

Lev: That sounds like a potential problem for running anything on real hardware if we can't control which saddle point dominates. It suggests that the correspondence isn't just about matching the leading terms; it’s about matching the entire structure of saddles across different regimes.

Mira: Exactly, and they found that this agreement is confirmed at leading semiclassical order when you look at the zero-mode partition function, Z0dS, which they found to be homologous to a single Lefschetz thimble.

Kai: That’s a big piece of evidence because it matches the bulk calculation at that same leading semiclassical order; it's like getting two different maps pointing to the same spot initially.

Lev: But Mira, if Z0dS is only one saddle, what about the rest of the integral? Does that single thimble capture all the physics we actually care about when we try to build something?

Mira: Well, they point out that this agreement in leading semiclassical order suggests that holographic dictionaries should actually be modified beyond just that leading order. They find a eleven over twelve difference in the Gamma function arguments, which is where things get interesting <ref:2609.10677#pg1>.

The paper's improvements: Kai: So, the paper suggests two key improvements stemming from this consistency check. First, because of that eleven over twelve difference in the Gamma function arguments, they suggest a modification to the holographic dictionary <ref:2609.10677#pg1>.

Mira: They propose that this modification should look like three times l AdS squared G N equals c minus eleven over two plus some small correction term, which is just O of c minus one. This means how bulk and boundary parameters map onto each other needs adjustment at the quantum level.

Lev: That sounds complicated to implement on a real chip. It implies that even if we get the leading order right, we have to account for these subleading terms when designing error correction codes or whatever physical system we're modeling.

Kai: And they also clarified something about why there is a difference in saddle content. They found it’s not necessarily a contradiction but rather due to truncations in their calculations.

Mira: They suggest that this hints at the possibility that additional Hartle–Hawking contributions might be present, perhaps carried by nonzero Liouville modes or full theory integration cycles, which we haven't accounted for yet.

Lev: So they’re saying the discrepancy isn't a failure of the physics itself but more about our current level of approximation when we truncate the theory or use simplified models. That gives us a direction for where to look next in our error correction research.

Conclusion: Kai: So, to wrap up this paper, "Stokes Phenomena between AdS/CFT and dS/CFT" confirms that the bulk and boundary calculations independently show the same Stokes phenomenon when analytically continued between dS and AdS.

Mira: The main implication is that holographic dictionaries need modification beyond the leading order to account for corrections like three times l AdS squared G N equals c minus eleven over two plus O of c minus one.

Lev: It suggests that we might need to look at those subleading terms when we try to run these ideas on real hardware, because the agreement isn't perfect at the semiclassical level.

Kai: Yeah, and they also point out that the difference in saddle content is likely due to truncations rather than a fundamental contradiction between AdS/CFT and dS/CFT.

Mira: That opens up possibilities for incorporating things like nonzero Liouville modes into our theory to see if we can resolve those differences through full theory integration cycles.

Lev: It’s good to know that the math is consistent at this level, but it also tells us exactly where the next layer of complexity lies for actual physics implementation.

Kai: That's all on this paper for now. We'll be looking at what comes next in these holographic explorations.

RIKEN Center for Interdisciplinary Theoretical and Mathematical Sciences · Graduate School of Science and Engineering, Saitama University · Yukawa Institute for Theoretical Physics, Kyoto University

hep-th, gr-qc

Submitted: 2026-09-09

Updated: 2026-11-17

Comments: 8+5 pages, 3 figures; v2: minor revisions

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 89/100

The gist: Stokes Phenomena between AdS/CFT and dS/CFT explores how saddle points in path integrals change discontinuously under analytic continuation…

Key concepts

Stokes Phenomenon
This refers to how saddle points in path integrals change discontinuously when you analytically continue the integration contour. In this context, it reveals a fundamental difference between the dS and AdS descriptions of the theory.
Lefschetz Thimble Decomposition
This is a mathematical technique used to decompose a complex path integral into contributions from specific saddle points, called thimbles. The coefficient of each thimble depends on how the original integration contour intersects these thimbles.
Liouville Zero-Mode Integral
This integral serves as an independent test for the bulk calculations. When analytically continued to both AdS and dS branches, it is found to be homologous to a single Lefschetz thimble, matching the leading semiclassical bulk results.

Terminology

Summary

Stokes Phenomena between AdS/CFT and dS/CFT explores how saddle points in path integrals change discontinuously under analytic continuation, providing a nontrivial holographic consistency check between AdS/CFT and dS/CFT The agreement between the bulk minisuperspace path integral and the Liouville zero-mode integral demonstrates that the dS contour is homologous to a single thimble while the analytically continued AdS contour is homologous to an infinite sum of thimbles, confirming holographic consistency

Contour Selection in Minisuperspace

The study begins with analyzing the dS3 minisuperspace path integral using a positive Lorentzian lapse N ∈ iR+ and a bulk cutoff parameter a1. The path integral is defined as ZAdS or ΨdS, where the upper and lower signs refer to AdS and dS, respectively The contour in (8) is inherited directly from the positive Lorentzian lapse, regulated as N ∈ −ε + iR+ As a1 → +∞ at fixed ε > 0, the lower endpoint moves to −i∞ Taking ε → 0+ then gives CL = lim ε→0+ (−ε + iR) After the cutoff-dependent shift, the positive Lorentzian half-line becomes a full vertical contour

Stokes Phenomenon in Saddle Points

The Lefschetz thimble decomposition is applied to the dS lapse integral, where its saddle points satisfy ∂Nren FdS(Nm) = 0. The coefficient nm of Jm in the thimble decomposition is determined by the intersection number of the original integration contour CL with Km For the Lorentzian contour, we obtain nm = ⟨CL, Km⟩ = δm0, CL ∼ J0 Thus only the saddle N0 contributes to the Lorentzian contour The change from a single dS thimble to this infinite sum is the Stokes phenomenon

Independent Verification via Liouville Zero-Mode

The Liouville zero-mode integral provides an independent test by analytically continuing it to the AdS and dS branches The zero-mode partition function Z0,dS is found to be homologous to a single Lefschetz thimble This result agrees with the bulk calculation at leading semiclassical order, as Z0,dS ∼ e − πldS/4GN

Consistency and Modifications

The agreement between the two calculations suggests that holographic dictionaries should be modified beyond the semiclassical leading order due to a 11/12 difference in Gamma function arguments This implies that the holographic dictionary should be modified as 3lAdS2GN = c − 11/2 + O(c−1) The comparison also clarifies that the difference in saddle content is due to truncations rather than a contradiction, suggesting that additional Hartle–Hawking contributions might be carried by nonzero Liouville modes or full theory integration cycles

Cutoff Dictionary Relations

The final relations show the leading semiclassical relations between the finite-cutoff Gamma-function bases in the bulk and boundary calculations These are given by-4b 2ϵ 2AdS = -4lAdS4GN a 21,AdS 1 + OGNlAdS + Oϵ 2AdS and 4b 2ϵ 2dS = 4ildS4GN a 21,dS 1 + OGNldS + Oϵ 2dS These relations are motivated by the Chern–Simons reduction but do not impose the Lefschetz thimble decompositions found independently in the main text The Chern–Simons reduction motivates these variables and fixes their leading dictionary, but it does not impose the Lefschetz thimble decompositions found independently in the main text The Chern–Simons reduction motivates these variables and fixes their leading dictionary, but it does not impose the Lefschetz thimble decompositions found independently in the main text.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improved AI system can perform holographic consistency checks by independently reproducing Stokes structures at leading semiclassical order in both bulk and boundary calculations. This provides a nontrivial holographic consistency check for AdS/CFT and dS/CFT correspondences.

  2. The system can now resolve the discrepancy between semiclassical leading-order predictions by suggesting that the holographic dictionary should be modified as 3lAdS 2GN = c - 11/2 + O(c-1). This implies a refinement of how bulk and boundary parameters map onto each other at quantum levels.

  3. The AI can better constrain the choice of integration contour by understanding that the choice of contour is part of the holographic prescription. This leads to an improved method for choosing contours in parameter regions where the integral converges before analytically continuing them.

  4. The system can utilize a comparison between matched symmetry truncations and full theory results to clarify saddle content differences, distinguishing between truncations and nonperturbative contour prescriptions. This allows for a more controlled comparison of simplified models against the full theory's potential complexity.

  5. The AI can identify that the agreement in leading semiclassical order is not sufficient, pointing towards corrections arising from O(1) corrections to the bulk–boundary parameter map or/and to the reduced path-integral measures. This directs future research toward calculating these subleading terms for better accuracy.

Sources

Related papers