CC Problem and ER=EPR in dS
First the cosmological constant problem: why is the gravitational effect of the vacuum so extraordinarily small compared with the energy scales suggested by quantum field theory?
Second the ER=EPR question: what is the relationship between quantum entanglement and the connectivity of spacetime?
The first concerns the value of a parameter in the gravitational field equations. The second concerns the structure of quantum correlations and the emergence of geometry. Yet in de Sitter space, where there is no spatial asymptotic boundary and physical descriptions are naturally tied to observers, these questions may intersect in an interesting way.
The central idea explored here is that the algebra of observables accessible to an observer may provide a principled description of the causal region, information and entropy relevant to cosmological selection. The cosmological constant would then enter not only through the curvature of spacetime, but also through the observer’s horizon, thermal scale and accessible algebra.
This is not a proposed solution of the cosmological constant problem. It is an attempt to identify a mathematical framework in which the observer, quantum information and cosmological geometry can be studied together. The distinction between established results and conjectures will be important throughout.
1. The cosmological constant problem
Einstein’s field equations can be written as
$$ G_{\mu\nu}+\Lambda g_{\mu\nu} = 8\pi G_N T_{\mu\nu}. $$The cosmological constant can equivalently be moved to the matter side of the equation and interpreted as a vacuum-energy density:
$$ \rho_\Lambda = \frac{\Lambda}{8\pi G_N}. $$The problem is not simply that this quantity is small. The difficulty is that quantum field theory permits large contributions to vacuum energy, including contributions associated with massive fields and vacuum phase transitions. Once gravity is included, these contributions must be reconciled with the extraordinarily small observed effective vacuum density.1
A schematic expression is
$$ \rho_{\Lambda,\mathrm{eff}} = \rho_{\Lambda,\mathrm{bare}} + \rho_{\mathrm{loops}} + \rho_{\mathrm{phase\ transitions}} +\cdots. $$A counterterm can be chosen so that the sum agrees with observation. The difficulty is explaining why this cancellation is so precise and why it remains stable under quantum corrections.
There are several logically distinct questions:
- Why is the effective vacuum energy so small?
- Why is it positive rather than exactly zero?
- Why is its value stable under radiative corrections?
- Why does it become dynamically important around the present cosmological epoch?
These are related, but a proposal addressing one does not automatically solve the others.1 The distinction will matter below. A theory might provide a probability distribution for different cosmological constants without explaining radiative stability. Conversely, a theory might make vacuum-energy contributions insensitive to gravity without selecting the observed positive value.
2. From entanglement to algebraic connectivity
The original ER=EPR proposal associates quantum entanglement with geometric connectivity. Its familiar example is the thermofield double state of two quantum systems, whose gravitational description can be a two-sided black hole connected by an Einstein–Rosen bridge.2 However, entanglement magnitude alone cannot characterize connectivity. A system can possess parametrically large entanglement without its semiclassical bulk being connected.
Engelhardt and Liu proposed an algebraic refinement of ER=EPR. Rather than classifying connectivity only through entanglement entropy, they use the operator-algebraic structure of the systems.3
A von Neumann algebra is classified, in part, by its type:
- Type I algebras admit ordinary density-matrix descriptions and behave like familiar matrix algebras.
- Type II algebras admit faithful normal semifinite traces and have a continuous dimension theory.
- Type III algebras have no faithful normal semifinite trace.
The algebraic ER=EPR proposal associates these structures with bulk connectivity in a specific large-\$N\$, semiclassical holographic setting. In particular, type I algebras are associated with disconnected systems, while type III\$_1\$ algebras play a central role in the description of classical wormhole connectivity. The proposal also distinguishes classical wormholes from quantum wormholes, where there is no stable classical geometry.3
The important qualification is that this is not a universal theorem saying that the abstract type of any two von Neumann algebras determines whether an arbitrary spacetime is connected. The holographic setting, the physical representation and the classicality assumptions are essential.
Nevertheless, the conceptual shift is significant:
$$ \text{Entanglement magnitude} \quad\longrightarrow\quad \text{Operator-algebraic structure}. $$The question becomes not merely how much two systems are entangled, but how their observable algebras are represented, embedded and related to one another. That distinction is especially relevant to de Sitter space.
3. Why de Sitter changes the algebraic question
Anti-de Sitter space has a spatial asymptotic boundary. In AdS/CFT, the boundary theory supplies a relatively clear starting point for identifying the observable algebra. De Sitter space is different. It has no analogous spatial asymptotic boundary on which one can simply place the complete theory.
Instead, an observer naturally describes a static patch bounded by a cosmological horizon. Different observers have different causal patches, and therefore different accessible observables.
This changes the question:
Which algebra should represent the physical system when the observer is itself part of the quantum description?
3.1 The static-patch algebra
In continuum quantum field theory, the local algebra associated with a suitable spacetime region is generically a type III factor, often type III\$_1\$.4 Such an algebra has no ordinary density matrix or local von Neumann entropy in the familiar finite-dimensional sense.
The de Sitter static patch therefore presents an algebraic problem: how do we define a gravitationally meaningful entropy for an observer whose observable algebra is not a type I algebra?
3.2 The observer and the crossed product
Chandrasekaran, Longo, Penington and Witten addressed this problem by including an observer’s clock in the construction of the static-patch algebra. The crossed-product construction produces a type II\$_\infty\$ algebra, and an appropriate positive-energy restriction gives a type II\$_1\$ algebra.
The latter admits a normalized trace and a corresponding entropy. For semiclassical states, this entropy agrees with generalized entropy, up to a state-independent additive constant:
$$ S_{\mathrm{gen}} = \frac{A}{4G_N}+S_{\mathrm{out}}. $$The empty de Sitter state is the maximum-entropy state in their construction.5 The observer is therefore not simply an external device used to describe an otherwise observer-independent algebra. The clock participates in defining the algebra itself.
3.3 Covariant observers and fluctuating regions
A particularly relevant development is the construction of Chen and Xu, An Algebra for Covariant Observers in de Sitter Space.
Their central observation is that a fixed static patch is not preserved by every de Sitter isometry. To implement the full SO(1,d) constraints, they promote the observer’s geodesic to a quantum degree of freedom.
The observer is described using a reference-frame Hilbert space
$$ \mathcal H_{\mathrm{obs}}=L^2(SO(1,d)). $$The observer is consequently in a superposition of geodesics, and the associated static patch fluctuates. The gauge-invariant algebra is an averaged modular crossed-product algebra over these fluctuating patches. The authors construct a faithful normal trace and establish that the resulting algebra is type II. For multiple observers, the algebra has a nontrivial center.6
In the multi-observer construction, the type is more specifically type II\$_\infty\$. This should not be confused with the type II\$_1\$ algebra obtained by the positive-energy restriction in the single-observer CLPW construction.
The center contains projections of the schematic form
$$ P_{a\leftarrow b}, $$which encode whether, and over which parts of a trajectory, observer \$b\$ is visible to observer \$a\$.
These projections provide an algebraic description of relational causal accessibility. They do not, by themselves, establish the existence of an Einstein–Rosen bridge. This is an important distinction. The causal structure of two observer patches and the quantum connectivity of their observable algebras are related questions, but they are not identical.
4. The center, relative information and the measure problem
The nontrivial center of the multi-observer algebra is particularly interesting because it decomposes the algebra into sectors.
Schematically, suppose we have a family of algebras
$$ \mathcal A = \int^\oplus \mathcal A_\lambda\,d\mu_0(\lambda), $$where the parameter \$\lambda\$ will eventually be identified with a cosmological-constant sector.
An operator in the center acts by multiplication on these sectors.
There is a subtlety here: in the Chen–Xu construction, \$\Lambda\$ is not itself a central operator. Their construction sets the de Sitter radius to one. A family of different cosmological-constant sectors would have to be introduced as an extension of their theory.
4.1 Trace normalization is not a probability distribution
Suppose each fibre admits a faithful normal semifinite trace \$\tau_\lambda\$. We can define a new trace on the direct-integral algebra:
$$ \tau_N(A) = \int d\mu_0(\lambda)\, N(\lambda)\tau_\lambda(A_\lambda), $$where \$N(\lambda)>0\$ is a central normalization factor.
Changing this factor changes absolute entropy. If a state has density \$h\$ relative to the original trace, then its density relative to the rescaled trace is \$h/N\$. Consequently,
$$ S_{\tau_N}(\omega) = S_{\tau}(\omega) + \int d p_\omega(\lambda)\log N(\lambda). $$But relative entropy is different.
For two normal states \$\omega\$ and \$\sigma\$, Araki relative entropy is a property of the pair of states on the algebra. It does not depend on an arbitrary rescaling of the trace.
For a direct-integral decomposition, it has the schematic form
$$ D(\omega\Vert\sigma) = D(p_\omega\Vert p_\sigma) + \int d p_\omega(\lambda)\, D_\lambda(\omega_\lambda\Vert\sigma_\lambda). $$Thus common central trace rescalings cannot be determined by relative entropy alone.7
This is a genuine structural limitation.
It means that imposing consistency of relative modular data between observers will not, by itself, determine the relative normalization of different cosmological sectors.
There is also an important conceptual distinction:
- A central state distribution \$p(\lambda)\$ describes how likely different sectors are in a physical state.
- A trace normalization \$N(\lambda)\$ specifies how absolute entropy is normalized across those sectors.
They are not automatically the same thing.
The algebraic construction does not supply a physical prior over cosmological constants merely because it has a center.
5. The Henneaux–Teitelboim connection:
Cosmological constant and four-volume
There is another place where the cosmological constant enters naturally: unimodular gravity.
In the Henneaux–Teitelboim (HT) formulation, the cosmological constant appears as a constant of motion rather than simply as a fixed parameter in the action.8
Introduce a vector density \$V^\mu\$, and define
$$ \kappa=8\pi G_N. $$A convenient form of the action is
$$ \begin{aligned} S_{\mathrm{HT}} ={}& \frac{1}{16\pi G_N} \int d^4x\sqrt{-g}\,R\\ &+ \frac{1}{\kappa} \int d^4x\, \Lambda \left( \partial_\mu V^\mu-\sqrt{-g} \right) +S_{\mathrm{matter}}. \end{aligned} $$Variation with respect to \$V^\mu\$ gives
$$ \partial_\mu\Lambda=0. $$Thus the cosmological constant is constant on shell.
Variation with respect to \$\Lambda\$ gives
$$ \partial_\mu V^\mu=\sqrt{-g}. $$Define the spatial integral
$$ T_4(t)=\int_{\Sigma_t}d^3x\,V^0. $$For a spacetime region bounded by two hypersurfaces,
$$ \Delta T_4 = \int d^4x\sqrt{-g}. $$The change in \$T_4\$ is the accumulated spacetime four-volume.
The corresponding canonical term is
$$ S_{\mathrm{symp}} = \frac{1}{\kappa} \int dt\,\Lambda\dot T_4. $$Hence
$$ P_{T_4}=\frac{\Lambda}{8\pi G_N}, $$and, formally,
$$ \boxed{\{T_4,\Lambda\}=8\pi G_N.} $$With canonical quantization, the corresponding formal commutator is
$$ [\hat T_4,\hat\Lambda] = i\,8\pi G_N. $$The cosmological constant is conjugate to a global spacetime four-volume variable.
There is an important qualification. In covariant HT gravity, the four-form variable is not automatically ordinary Schrödinger time. Buchmüller and Dragon explain that the cosmic-time interpretation emerges semiclassically; the quantum state still satisfies the Wheeler–DeWitt equation.9
This matters because it prevents us from treating \$T_4\$ as an ordinary local clock without specifying the quantum state, boundary conditions and approximation.
5.1 What happens to vacuum energy?
Add a constant vacuum-energy density:
$$ S_{\mathrm{vac}} = -\rho_{\mathrm{vac}} \int d^4x\sqrt{-g}. $$The metric-dependent cosmological term can be rewritten using
$$ \Lambda_{\mathrm{eff}} = \Lambda+\kappa\rho_{\mathrm{vac}}. $$The full action then satisfies
$$ S_{\mathrm{HT}}+S_{\mathrm{vac}} = S_{\mathrm{HT}}[\Lambda_{\mathrm{eff}}] - \rho_{\mathrm{vac}}\Delta T_4. $$In a Lorentzian path integral, the extra boundary contribution produces a phase
$$ \exp(-i\rho_{\mathrm{vac}}\Delta T_4). $$This shows how a constant vacuum-energy contribution can be represented as a shift of the effective integration constant together with a four-volume phase.
However, this is not automatically a solution of the cosmological constant problem. If the integration constant is fixed by boundary or initial data, the shift still affects the effective cosmological constant.
The physical question becomes how the global zero mode and the boundary conditions are treated.
Unimodular gravity changes the status of \$\Lambda\$, but does not automatically explain its observed value.910
6. De Sitter thermodynamics:
Cosmological constant as an observer’s scale
The HT construction becomes especially interesting when combined with de Sitter observer thermodynamics.
For four-dimensional de Sitter space,
$$ R_{\mathrm{dS}} = \sqrt{\frac{3}{\Lambda}}. $$The Gibbons–Hawking temperature of a static observer is
$$ T_{\mathrm{dS}} = \frac{1}{2\pi R_{\mathrm{dS}}}, $$and the horizon entropy is
$$ S_{\mathrm{dS}} = \frac{A_{\mathrm{dS}}}{4G_N} = \frac{3\pi}{G_N\Lambda}. $$Eliminating the de Sitter radius gives
$$ \boxed{\Lambda=12\pi^2T_{\mathrm{dS}}^2.} $$Thus the cosmological constant determines the thermal scale experienced by a static observer.
If the observer’s state is KMS at inverse temperature \$\beta_{\mathrm{dS}}\$, its modular automorphism group is related to physical time evolution by
$$ \sigma_s^\omega = \alpha_{\beta_{\mathrm{dS}}s}^\omega, $$up to the sign convention.
The modular parameter \$s\$ is dimensionless. Physical proper time is calibrated by the inverse temperature.
This gives us three distinct notions of time:
- Observer proper time, measured along a geodesic.
- Modular time, generated by the state and observable algebra.
- HT four-volume time, conjugate to the cosmological constant.
They should not be identified with one another.
However, a quantum state with different \$\Lambda\$-sectors would also have different modular thermal scales. This makes the observer’s clock a natural way to formulate relational comparisons between cosmological sectors.
The unresolved problem is how to construct the physical Hilbert-space maps between those sectors. The Chen–Xu construction itself treats a fixed de Sitter radius.
7. Horizon corners and the geometry of observer space
There is another potentially useful connection.
In gravitational edge-mode theory, the area of a codimension-two corner is conjugate to a relative Lorentzian boost angle \$\eta\$ between its normal frames.
The corner symplectic form takes the form
$$ \Omega_{\mathrm{corner}} = \frac{1}{8\pi G_N}\, \delta A\wedge\delta\eta. $$The corresponding momentum is
$$ p_\eta=\frac{A}{8\pi G_N}. $$For the de Sitter cosmological horizon,
$$ p_\eta = \frac{A_{\mathrm{dS}}}{8\pi G_N} = \frac{S_{\mathrm{dS}}}{2\pi}. $$This is a striking relation: the same entropy that characterizes the de Sitter horizon is proportional to the momentum conjugate to the normal-frame boost.11
The space of oriented timelike geodesics in four-dimensional de Sitter can be represented as
$$ \mathcal G_{\mathrm{geo}} = \frac{SO_0(1,4)}{SO_0(1,1)\times SO(3)}. $$This is a six-dimensional homogeneous space.
The coadjoint-orbit construction suggests a natural symplectic form on this space, with a possible level
$$ c=\frac{A_{\mathrm{dS}}}{8\pi G_N}. $$This is a promising mathematical connection between the gravitational horizon corner and the quantum reference frame of a de Sitter observer.
But the physical identification is not yet derived from the complete joint gravitational and observer action. We must derive the moving-horizon boundary terms and avoid counting the same gravitational surface charge twice.
There is also a useful negative result. Two oppositely oriented observer orbits, by themselves, can reduce to a single point after imposing the common de Sitter constraint. Matter and clock degrees of freedom are needed for nontrivial physical relative observables.
So the horizon orbit is not, by itself, an entanglement mechanism or a wormhole criterion.
8. The spectral-geometric perspective
Noncommutative geometry offers another way to express the curvature scale.
The Euclidean continuation of four-dimensional de Sitter space is the round four-sphere \$S^4_R\$.
For the ordinary spin Dirac operator on this sphere, the eigenvalues are
$$ \lambda_n^\pm = \pm\frac{n+2}{R}, \qquad n=0,1,\ldots. $$The smallest absolute eigenvalue is
$$ \lambda_{\min}=\frac{2}{R}. $$Since \$R^2=3/\Lambda\$,
$$ \boxed{\Lambda=\frac{3}{4}\lambda_{\min}^2.} $$This is an exact relation for the round four-sphere. It demonstrates that the Dirac spectrum encodes the curvature scale of a specified de Sitter geometry. But it does not select that geometry. The spectral action has the schematic form
$$ S_{\mathrm{spec}}[D] = \operatorname{Tr} f\left(\frac{D}{\Lambda_{\mathrm{cut}}}\right), $$where \$\Lambda_{\mathrm{cut}}\$ is a spectral cutoff, not the cosmological constant. In four dimensions its heat-kernel expansion contains terms of the form
$$ S_{\mathrm{spec}} \sim f_4\Lambda_{\mathrm{cut}}^4a_0 + f_2\Lambda_{\mathrm{cut}}^2a_2 + f_0a_4+\cdots. $$The leading coefficient is related to volume, the next contains curvature, and higher terms contain curvature combinations.12
This is useful for reconstructing or studying geometry once a spectral triple is specified.
It does not, by itself, determine which value of \$R\$ or \$\Lambda\$ the universe should occupy.
The distinction is important:
$$ \text{Algebra and representation} \longrightarrow \text{relational structure}, $$ $$ D \longrightarrow \text{spectral geometry}, $$ $$ \text{Quantum state and dynamics} \longrightarrow \text{sector selection}. $$The Dirac operator can measure the curvature of a de Sitter sector. It does not supply the probability of occupying that sector without additional dynamics.
9. The observer-weighting idea:
A bridge to the causal entropic principle
This is perhaps the most promising connection between algebraic ER=EPR and the cosmological constant problem.
The causal entropic principle proposes that cosmological parameters should be weighted by the entropy produced within a causally connected region.
Bousso, Harnik, Kribs and Perez used this principle to construct an observer-weighted probability distribution for the cosmological constant. In their calculation, entropy production from dust heated by stars dominates the weighting.13
The observer’s causal diamond is therefore not merely a geometrical object. It is the region in which the relevant entropy production is counted.
This is precisely the kind of region described by the de Sitter observer algebra.
Let \$\mathcal A_a\$ denote the algebra accessible to observer \$a\$, and let
$$ \mathcal N_{ab}=\mathcal A_a\cap\mathcal A_b $$denote the algebra shared by two observers, where the intersection is defined in a common physical representation.
The shared algebra provides a candidate notion of intersubjective accessibility: the observables that both observers can measure.
For a non-equilibrium state \$\omega_t\$, one might attempt to define an observer entropy production
$$ \Delta S_a(t) = S_{\mathrm{cg}} \left(\omega_t|_{\mathcal A_a}\right) - S_{\mathrm{cg}} \left(\omega_0|_{\mathcal A_a}\right). $$Likewise, the common observer entropy production would be
$$ \Delta S_{ab}(t) = S_{\mathrm{cg}} \left(\omega_t|_{\mathcal N_{ab}}\right) - S_{\mathrm{cg}} \left(\omega_0|_{\mathcal N_{ab}}\right). $$The subscript \$\mathrm{cg}\$ is important. It indicates that a coarse-graining prescription is required.
For type III algebras, ordinary von Neumann entropy is not available without a regulator or additional structure. Relative entropy remains well-defined for suitable normal states:
$$ D_{\mathcal A}(\omega\Vert\sigma) = S_{\mathcal A}(\omega\Vert\sigma). $$It can measure how distinguishable two states are using only the observables in \$\mathcal A\$.
But relative entropy is not automatically thermodynamic entropy production. A dynamical model and a coarse-graining prescription are still required.
9.1 What our toy model actually showed
We tested this idea in a finite-dimensional model consisting of two observers on a ring, with a quantum label describing their relative position.
We constructed their observable algebras and their common algebra, then evolved the system with random two-site unitaries.
The entropy production measured through the shared algebra was smaller than that measured by either observer separately. When the regions had no overlap, the shared entropy production vanished.
The late-time results agreed with the ordinary finite-size Page entropy of the overlap region.
This is a useful negative result. In that model, shared entropy production was simply the entropy production of the common subsystem. It did not reveal additional wormhole structure.
It does not prove that the same result holds in continuum de Sitter quantum gravity. The toy model is type I and does not reproduce the full Chen–Xu crossed-product construction.
Nevertheless, it tells us what a successful algebraic proposal must improve upon: it must identify information shared by observers that cannot be reduced to ordinary geometric overlap.
9.2 A possible observer-weighting functional
A candidate connection to the causal entropic principle would be
$$ W_{\mathrm{obs}}(\Lambda) = \int_{\mathcal D_a(\Lambda)} d^4x\sqrt{-g}\, \sigma_{\mathrm{prod}}(x), $$where \$\mathcal D_a(\Lambda)\$ is the observer’s causal diamond and \$\sigma_{\mathrm{prod}}\$ is a coarse-grained entropy-production density.
The observer algebra could determine which observables are accessible within the diamond, while matter dynamics would determine the entropy production.
A probability distribution could then take the form
$$ p_{\mathrm{obs}}(\Lambda) \propto p_{\mathrm{prior}}(\Lambda) N(\Lambda) W_{\mathrm{obs}}(\Lambda). $$Here:
- \$p_{\mathrm{prior}}\$ is the distribution supplied by the underlying theory.
- \$N(\Lambda)\$ is a relative sector-normalization factor, if a common Hilbert-space or path-integral prescription defines one.
- \$W_{\mathrm{obs}}\$ is the observer weighting.
The algebraic contribution would be to define the observer-accessible coarse-graining more precisely.
It would not automatically determine the prior or the entropy-production dynamics.
This is where the causal entropic principle and algebraic ER=EPR could genuinely complement one another.
10. Why the usual Euclidean weighting does not solve the problem
The Euclidean de Sitter saddle gives a familiar candidate weight.
For a round four-sphere,
$$ I_E(S^4)=-S_{\mathrm{dS}} = -\frac{3\pi}{G_N\Lambda}. $$The formal semiclassical saddle contribution is therefore
$$ e^{-I_E} = e^{S_{\mathrm{dS}}} = \exp\left(\frac{3\pi}{G_N\Lambda}\right). $$If this is interpreted naively as a probability weight over positive cosmological constants, it diverges as \$\Lambda\rightarrow0^+\$.
Indeed,
$$ \int_0^\epsilon d\Lambda\, \exp\left(\frac{3\pi}{G_N\Lambda}\right) $$does not converge.
This is the familiar Baum–Hawking difficulty.
There are deeper issues as well. The gravitational Euclidean path integral has contour and conformal-factor problems, and a saddle contribution cannot simply be identified with a normalized probability distribution.
Thus the factor \$e^{S_{\mathrm{dS}}}\$ is a useful baseline for discussing possible cosmological measures, but it is not a derived physical prior.14
The same caution applies to proposals involving baby universes and \$\alpha\$-parameters. Their interpretation in higher-dimensional quantum gravity remains actively debated, and should not be assumed to provide a unique measure over cosmological constants.15
11. What have we actually established?
It is worth separating the results from the ideas that remain conjectural.
Established ingredients
The algebraic ER=EPR proposal replaces entanglement magnitude with a structural analysis of observable algebras in its specified holographic setting.
A gravitationally dressed de Sitter static-patch algebra can be constructed as a type II$_1$ algebra, with entropy matching generalized entropy up to an additive constant.5
A covariant multi-observer construction in de Sitter produces a fluctuating-region algebra of type II with a nontrivial center encoding causal accessibility.6
In Henneaux–Teitelboim gravity, the cosmological constant is conjugate to a global four-volume variable.8
De Sitter horizon entropy, thermal scale and curvature radius are related by
$$ R^2=\frac{3}{\Lambda}, \qquad T_{\mathrm{dS}}=\frac{1}{2\pi R}, \qquad S_{\mathrm{dS}}=\frac{3\pi}{G_N\Lambda}. $$Relative entropy is insensitive to common central trace rescalings. Consequently, modular consistency alone cannot determine an arbitrary relative normalization of cosmological sectors.
Proposed connections
A family of observer algebras over different cosmological constants could describe a quantum state spread over different de Sitter radii.
The HT four-volume variable could provide a global relational clock shared by multiple observers.
Gravitational horizon-corner symplectic structures may induce a coadjoint-orbit geometry on the space of observer geodesics.
Relative entropy and common observer algebras may provide a quantum definition of the coarse-graining region used in causal-entropic selection.
The algebraic observer framework may help distinguish observer-dependent information from information that is genuinely shared between observers.
These proposals need to be derived from a common constrained quantum-gravitational model. They are not yet consequences of the cited constructions.
12. What has not been solved
Despite the mathematical structure developed above, we have not derived the observed value of the cosmological constant.
We have not shown that:
- The HT measure is normalizable over positive cosmological constants.
- The observer algebra fixes a prior over \$\Lambda\$.
- The horizon-corner orbit selects a preferred value.
- Relative modular consistency determines cross-sector normalization.
- The entropy of a shared algebra generates a new observer-selection mechanism.
- The spectral action selects the observed de Sitter radius.
Several tempting routes have actually produced useful negative results.
The distinction between type and state has ruled out the idea that changing entanglement alone changes the algebraic connectivity class of a finite-dimensional model.
The trace-rescaling analysis showed that relative modular data cannot determine an arbitrary central normalization.
The ring model showed that shared entropy production can reduce to ordinary overlap entropy.
The simple HT-plus-observer models have not produced a normalizable distribution peaked at the observed cosmological constant.
And the entropy-weighted coadjoint orbit remains a candidate geometric structure, not a derived mechanism for selecting \$\Lambda\$.
These negative results narrow the space of plausible directions. They are not solutions, but they help identify which assumptions must be changed.
13. The research direction: a quantum measure over observers
The most promising question emerging from this programme is not whether algebraic ER=EPR directly determines the cosmological constant.
It is whether a relational observer algebra can provide the correct physical framework for defining the measure over cosmological sectors.
The desired structure would contain
$$ \begin{gathered} \text{Global cosmological variable:} \quad (\Lambda,T_4),\\[2mm] \text{Observer algebra:} \quad \mathcal A_a(\Lambda),\\[2mm] \text{Shared observables:} \quad \mathcal N_{ab}(\Lambda) = \mathcal A_a(\Lambda)\cap\mathcal A_b(\Lambda),\\[2mm] \text{Physical state:} \quad \Psi(\Lambda,\text{observers},\text{matter}). \end{gathered} $$The state would determine how the cosmological sectors are populated. The observer algebras would determine what information can be accessed in each sector. The global four-volume variable would relate the cosmological constant to a relational notion of time.
The main task would be to derive the measure, rather than postulate one.
A concrete calculation would begin with a joint HT-plus-observer path integral,
$$ Z = \int \mathcal Dg\, \mathcal DV\, \mathcal D\Lambda\, \mathcal D(\text{observers})\, e^{iS_{\mathrm{HT}}+iS_{\mathrm{observer}}}. $$One would then:
- Impose the full de Sitter gravitational constraints.
- Include the observer clock and its causal-patch algebra.
- Derive the horizon boundary and corner terms without double-counting gravitational charges.
- Integrate over the global four-volume mode.
- Derive the measure over cosmological-constant sectors.
- Determine whether the observer-dependent entropy production changes that measure.
The decisive outcome would be a factor
$$ Z = \int d\Lambda\, p_{\mathrm{prior}}(\Lambda) N(\Lambda) W_{\mathrm{obs}}(\Lambda) Z_{\mathrm{grav}}(\Lambda), $$where every factor follows from a specified physical construction rather than being independently chosen.
Only then would it make sense to ask whether the resulting distribution is normalizable, whether it is stable under quantum corrections, and whether it can peak near the observed positive cosmological constant.
Conclusion
The connection between algebraic ER=EPR and the cosmological constant problem is not yet a solution of either problem.
What it offers is a different way of organizing the questions.
In de Sitter space, the observer is structurally important. Its clock defines a relational reference frame, its causal patch determines which observables are accessible, and its horizon carries a thermodynamic entropy.
Unimodular gravity adds another layer: the cosmological constant is conjugate to a global four-volume variable. The observer’s proper time calibrates how that volume accumulates, while modular theory relates the observer’s state to the thermal structure of the static patch.
The possible synthesis is therefore
$$ \begin{aligned} \text{Observer algebra} &\longrightarrow \text{accessible information},\\ \text{Modular structure} &\longrightarrow \text{thermal and relational time},\\ \text{HT gravity} &\longrightarrow \text{volume--}\Lambda\text{ conjugacy},\\ \text{Observer dynamics} &\longrightarrow \text{entropy production and selection}. $$The missing ingredient is the physical state or measure over the different cosmological sectors.
My current view is that algebraic ER=EPR is a promising language for understanding the relational structure of de Sitter quantum gravity. Its contribution to the cosmological constant problem may lie in defining what an observer can access and how information is shared between observers, rather than in producing the numerical value of \$\Lambda\$ directly.
The next step is therefore not another attempt to extract the cosmological constant from an algebraic type or a Dirac eigenvalue. It is to derive the quantum measure over observer-accessible cosmological sectors from a common gravitational theory.
That is where the observer, the algebra and the cosmological constant may finally become parts of the same problem.
References
S. Weinberg, “The Cosmological Constant Problem,” Reviews of Modern Physics 61, 1–23 (1989). https://doi.org/10.1103/RevModPhys.61.1 ↩︎ ↩︎
J. Maldacena and L. Susskind, “Cool Horizons for Entangled Black Holes,” Fortschritte der Physik 61, 781–811 (2013). https://arxiv.org/abs/1306.0533 ↩︎
N. Engelhardt and H. Liu, “Algebraic ER=EPR and Complexity Transfer,” JHEP 07 (2024) 013. https://arxiv.org/abs/2311.04281 ↩︎ ↩︎
J. Yngvason, “The Role of Type III Factors in Quantum Field Theory,” Reports on Mathematical Physics 55, 135–147 (2005). https://arxiv.org/abs/math-ph/0411058 ↩︎
V. Chandrasekaran, R. Longo, G. Penington and E. Witten, “An Algebra of Observables for de Sitter Space,” JHEP 02 (2023) 082. https://arxiv.org/abs/2206.10780 ↩︎ ↩︎
B. Chen and J. Xu, “An Algebra for Covariant Observers in de Sitter Space,” JHEP 09 (2026) 240. https://arxiv.org/abs/2511.00622 ↩︎ ↩︎
H. Araki, “Relative Entropy of States of von Neumann Algebras,” Publications of the Research Institute for Mathematical Sciences 11, 809–833 (1976). ↩︎
M. Henneaux and C. Teitelboim, “The Cosmological Constant as a Canonical Variable,” Physics Letters B 143, 415–420 (1984). https://doi.org/10.1016/0370-2693(84)91493-X ↩︎ ↩︎
W. Buchmüller and N. Dragon, “The Cosmological Constant as a Boundary Term,” JHEP 08 (2022) 167. https://arxiv.org/abs/2203.15714 ↩︎ ↩︎
P. Jiroušek, “Unimodular Approaches to the Cosmological Constant Problem,” Universe 9, 131 (2023). https://arxiv.org/abs/2301.01662 ↩︎
L. Freidel, M. Geiller and D. Pranzetti, “Edge Modes of Gravity. Part I. Corner Potentials and Charges,” JHEP 11 (2020) 026. https://arxiv.org/abs/2006.12527 ↩︎
A. H. Chamseddine and A. Connes, “The Uncanny Precision of the Spectral Action,” Communications in Mathematical Physics 293, 867–889 (2010). https://arxiv.org/abs/0812.0165 ↩︎
R. Bousso, R. Harnik, G. D. Kribs and G. Perez, “Predicting the Cosmological Constant from the Causal Entropic Principle,” Physical Review D 76, 043513 (2007). https://arxiv.org/abs/hep-th/0702115 ↩︎
G. W. Gibbons and S. W. Hawking, “Action Integrals and Partition Functions in Quantum Gravity,” Physical Review D 15, 2752 (1977). https://doi.org/10.1103/PhysRevD.15.2752 ↩︎
D. Harlow, “Observers, Alpha-Parameters, and the Hartle-Hawking State,” (2026). https://arxiv.org/abs/2602.03835 ↩︎