Matrix, SYK-Like, and Lower-Dimensional Models

Solvable matrix and SYK-like systems can reproduce selected de Sitter quantities—spectral densities, entropy differences, two-point functions, or observer thermodynamics—without yet defining a unique dual of de Sitter gravity. The decisive question is which observables match under one declared dictionary and whether the match survives changes of ensemble or coupling. This page focuses on double-scaled SYK proposals for a three-dimensional static patch and its two-dimensional dilaton-gravity reduction. These are current, lower-dimensional, proposal-level models rather than established higher-dimensional de Sitter holography as of 2026-08-20.

Required background. Gauge Theory and String-Theory Preparation provides large-$N$ limits, random ensembles, and constrained systems.

Helpful background. Quantum Gravity frames the bulk dictionary; Quantum Information supplies entropy and operator-algebra tests; Causality and Global Structure supplies static patches; and Semiclassical Gravity supplies de Sitter correlators and thermodynamics.

A doubled constrained SYK model

Begin with even $N$ Majorana fermions, an even interaction order $p$, and a random $p$-body Hamiltonian. Here use the source normalization $\{\psi_i,\psi_j\}=2\delta_{ij}$; it differs from the $\delta_{ij}$ convention on the ordinary SYK page and rescales couplings accordingly:

$$H=i^{p/2}\sum_{i_1\lt\cdots\lt i_p} J_{i_1\cdots i_p}\psi_{i_1}\cdots\psi_{i_p}.$$

The double-scaling limit sends $N,p\to\infty$ with

$$\lambda=\frac{2p^2}{N}$$

fixed. The proposed de Sitter model uses two identical copies and restricts physical states by

$$(H_L-H_R)|\Psi_{\mathrm{phys}}\rangle=0.$$

For definiteness the copies use the same realization of independent zero-mean Gaussian couplings, with variance $\overline{J_{i_1\cdots i_p}^{\,2}}=N\mathbb J^2/[2p^2\binom Np]$ in this normalization. Thus every eigenvalue can pair exactly across copies; independent disorder realizations would generally require a finite energy tolerance. Operators must preserve the constraint. Narovlansky and Verlinde 2023, § 2, Eqs. (15)–(24) define the system. Their spectral density peaks near zero energy; “maximal entropy” refers to a microcanonical coarse graining, not the von Neumann entropy of an individual pure eigenstate. Ensemble averaging remains part of the calculable spectral description.

In the semiclassical $\lambda\to0$ limit, constraint-preserving dressed operators have a two-point function matching a massive scalar Green function in $\mathrm{dS}_3$, with the proposed relation

$$m^2L^2=4\Delta(1-\Delta),$$

after fixing time, radius, and operator normalization. For real $0\lt\Delta\lt1$, this map covers $0\lt m^2L^2\leq1$, the complementary-series range and its threshold in three dimensions; heavier principal-series masses require a separate continuation. The checked match concerns the circle-reduced scalar equation and selected observer-worldline two-point functions, including their prescribed analytic continuation Narovlansky and Verlinde 2023, §§ 3.3–4.1, Eqs. (46) and (55)–(62). It does not establish a dictionary for all angular modes or every local bulk field.

Results and limitations through 2026-08-20

The formulation axes are Hamiltonian ensemble, symmetry class, double-scaling limit, state preparation, constraint, operator dictionary, target dimension, and static-patch region.

Evidence or position	Controlled result	Limitation or adverse discriminator	Date
Doubled DSSYK, Narovlansky and Verlinde 2023, §§2–4, pp. 6–19	Exact spectral data and dressed two-point functions reduce at large $N$ and small $\lambda$ to selected $\mathrm{dS}_3$ static-patch scalar quantities	A full bulk operator map, unique ultraviolet gravity theory, and higher-point locality tests remain unresolved	2023
Observer thermodynamics, Tietto and Verlinde 2025, §§2–5, pp. 2–11	Semiclassical $\mathrm{dS}_3$ entropy with an observer is matched to the known DSSYK spectral density, clarifying two temperature notions	The construction assumes a proposed entropy dictionary and conical-source model; it does not establish complete dynamical equivalence	2025
Static-patch trace adverse test, Chen, Stanford, Tang, and Yang 2026, §§2–4, pp. 8–31	Shockwave OTOCs test cyclicity and positivity of a Euclidean-gravity trace proposal	The tested trace interpretation fails those criteria; any SYK-like realization of that dictionary must evade or reproduce the same calculation	2026
Matrix-model completion position	Ensemble averages and exact spectra suggest a nonperturbative completion of selected gravitational sums	No unique member theory, complete observer algebra, or controlled higher-dimensional de Sitter reconstruction follows from ensemble data alone	2026 cutoff

Controlled results include exact quantum-mechanical spectra and specific semiclassical correlator or entropy matches. Competing positions call these models useful laboratories, duals of a two-dimensional reduction, or microscopic theories of a static patch. The strongest interpretation remains conjectural and model-dependent. The unresolved limitation is a complete observable dictionary that survives higher-point, real-time, and nonperturbative tests.

Matching entropy, spectrum, and correlators

A serious comparison should use at least three independent quantities. First, compare the model density $\rho(E)$ with a gravitational entropy difference in the same ensemble. Second, compare time-dependent two-point functions, including their KMS period and normalization. Third, test a higher-point or response observable sensitive to interactions and causal structure.

The 2025 observer model distinguishes the total entropy of spacetime plus observer from the entropy assigned to ambient de Sitter degrees of freedom. In three-dimensional Schwarzschild–de Sitter, the conical deficit and observer energy lead to two temperature notions; the DSSYK spectral density can reproduce the proposed thermodynamic relation. This improves the dictionary but does not remove its assumptions. An additive entropy normalization or identification of energy must not be adjusted independently for every observable.

The small-$\lambda$ limit is another essential axis. If $\lambda$ measures $G/L$ in the proposed dictionary, then corrections in $\lambda$ should correspond to gravitational quantum corrections. That claim can be tested only if both sides compute the same observable beyond leading order. A formula exact in the quantum model is not automatically an exact gravity result when the bulk map was established only asymptotically.

Structural agreement or tuned agreement

Varying the ensemble is a useful adverse test. If replacing Gaussian couplings by another distribution changes only subleading microscopic details while leaving the claimed semiclassical observables universal, the match is structural. If it destroys the entropy or two-point relation, the proposal must explain why the original ensemble is physically selected.

Likewise, vary operator dimension, energy sector, and state preparation. A proposed geometry should reproduce a family of masses and correlators with one radius identification, not a separately fitted value for each operator. The equal-energy constraint is physically motivated as a form of dressing, but other implementations can produce different physical Hilbert spaces. Their equivalence must be demonstrated.

Finally, a $\mathrm{dS}_3$ or dilaton-gravity match does not establish local four-dimensional gravity. Extra angular modes, propagating gravitons, compactification data, and observer algebras must be added and checked. Lower-dimensional control is valuable precisely when its domain is kept explicit.

What could change these conclusions

The conclusions could change if a theorem gives a complete bulk dictionary and unitary observer algebra, if an independent calculation matches higher-point or finite-$\lambda$ observables, or if a counterexample shows that an advertised correlator or entropy relation depends on an unphysical ensemble choice. The conclusions also depend on whether a higher-dimensional embedding, new state-preparation protocol, or adverse shockwave test changes the proposed target theory.

Exercises

Suppose a DSSYK candidate matches the de Sitter entropy at one energy but its two-point function changes qualitatively under a small ensemble deformation. Explain how this evidence compares with a structural match, and propose two further checks.

Solution

The single entropy match is consistent evidence but not structural evidence because one number can be fitted by an energy map or additive normalization. Sensitivity of the correlator to a small ensemble deformation is an adverse discriminator unless the gravitational dictionary independently selects the original ensemble. One should compare a continuous family of operator dimensions with de Sitter masses using one radius and normalization, then compute a higher-point real-time observable such as an OTOC or response function. It is also useful to vary the physical-state constraint. Agreement across these checks would support a universal sector; failure would confine the model to a tuned analogy.

References

• Chen, Yiming, Douglas Stanford, Haifeng Tang, and Zhenbin Yang. “Negative Shocks versus Static Patch Holography.” Preprint (2026). arXiv:2607.14042.
• Narovlansky, Vladimir, and Herman Verlinde. “Double-Scaled SYK and de Sitter Holography.” Preprint (2023). arXiv:2310.16994.
• Tietto, Damiano, and Herman Verlinde. “A Microscopic Model of de Sitter Spacetime with an Observer.” Preprint (2025). arXiv:2502.03869.