Higher Categorical Coherence Breakdown, Self-Measurement, and Finite Entanglement Entropy in Quantum Field Theory
Full text
Higher Categorical Coherence Breakdown, Self-Measurement, and Finite Entanglement Entropy in Quantum Field Theory Andrei T. Patrascu 1 1 FAST Foundation, Destin FL, 32541, USA email: andrei.patr[email protected] Two long–standing puzzles in fundamental physics are usually treated independently: the measurement problem in quantum mechanics and the divergence of entanglement entropy in algebraic quantum field theory (AQFT). The former arises because linear, unitary evolution offers no intrinsic mechanism for outcome selection, while the latter reflects the fact that local algebras in AQFT are type III1 factors, for which no density matrix exists and the vacuum restricted to a region exhibits an area–law divergent entropy. In this paper we argue that both puzzles stem from an over–abundance of coherence, and we show that they admit a unified resolution within a categorical framework. We introduce the categorical density matrix (CDM), a natural transformation assigning states to spacetime regions enriched with Tomita–Takesaki modular data. We then allow controlled failures of higher categorical coherence (pentagon/hexagon identities), valued in local automorphisms of the net. Passing these defects through the CDM induces local, completely positive Lindblad–type dephasing terms together with a nonlinear modular feedback term in the evolution. Physically, this corresponds to self–measurement: ultraviolet modes act as an intrinsic meter, continuously monitoring near–boundary energy densities, while modular feedback stabilizes pointer–like structures. We prove that the Reeh–Schlieder properties (cyclicity and separating) are structurally preserved, but the short–distance two–point singularities responsible for the entanglement divergence are dynamically softened. This can be summarized by an effective dephasing exponent δ > 0, so that correlators behave as |x−y|−(d−1)+δ at small separation. Collar–free entanglement diagnostics such as mutual information between adjacent regions and shape–variation first laws become finite once δ≥d− 2. Thus the same mechanism that embeds measurement into the dynamics also removes the sharp–limit entanglement divergence, without invoking a cutoff or minimal length. Operationally, the dynamics produces an approximately central pointer subalgebra in boundary collars, creating an effective type-I window within a structurally type-III theory. In unifying the measurement problem with the entropy problem, this work suggests that higher categorical structure is not only a language for quantum foundations but also a tool to reformulate the ultraviolet behavior of quantum fields. We conclude by discussing implications for black hole entropy, modular chaos, and possible observational signatures of intrinsic self–measurement. I. INTRODUCTION Two long–standing puzzles placed side by side Quantum theory, in both nonrelativistic and relativistic guises, delivers extraordinary predictive power while leaving open two structural puzzles that, although usually treated separately, are deeply connected. (i) The measurement problem. In the textbook (von Neumann) account, unitary evolution and the projection postulate coexist uneasily: the theory supplies no intrinsic mechanism that selects an outcome or even a preferred set of macroscopic variables (“pointer states”) in closed systems. Decoherence theory [ 23 ] explains the emergence of classical records for open subsystems but, by design, does not replace the postulate. Competing moves—Everett’s relative-state interpretation [ 24 ], objective-collapse models (GRW/CSL) [ 25 , 26 ], and proposals for nonlinear quantum mechanics (together with their well-known pitfalls [27, 28])—each address parts of the problem with different costs. (ii) The entropy problem in algebraic QFT. Within the algebraic (Haag–Kastler/Wightman) framework [ 1 , 3 , 102 ], local observable algebras A ( O )attached to bounded regions are, under very general conditions, type III factors (indeed typically type III1 ) [ 7 – 9 ]. As a result, there is no trace and no bona fide density matrix for A ( O ): the von Neumann entropy of the vacuum restricted to O is neither defined nor finite. When one regulates (lattice, momentum cutoff, split property) one finds an area law divergence S(O;)∼κd Area(∂O) d−2+ (subleading), diagnosed already in seminal works [ 12 – 15 ] and thoroughly reviewed in [ 16 , 17 ]. The form of the divergence (proportional to boundary area) is universal; the coefficient κd is scheme dependent. Modular theory and the Bisognano–Wichmann theorem [ 5 , 98 ] sharpen this picture by identifying local modular “temperatures” responsible for near-boundary entanglement.
2 These two puzzles are not orthogonal. The Reeh–Schlieder theorem [ 4 ]—which asserts that the vacuum is cyclic and (for regions with nonempty causal complement) separating for every local algebra—encodes a vacuum state that is maximally coherent across all scales and cuts. That very coherence powers the area divergence. A unifying perspective: higher categorical coherence, self-measurement, and entanglement The present work develops a framework in which both puzzles are addressed together by slightly enlarging the kinematics of quantum theory and encoding it in a functorial, region-wise state assignment we call the categorical density matrix (CDM). The two key ideas are: (A) Categorical density matrix (CDM). Instead of representing a global state by a single trace-class operator (which does not exist on type III algebras), we treat the state as a natural transformation ρ: Loc =⇒Stat ◦A,O 7→ ρO∈ A(O)+ ∗, compatible with inclusions and symmetries. Tomita–Takesaki modular data ( JO, ∆ O )and regional modular flows σO t are thereby part of the structure [ 5 , 102 ] and can be manipulated without ever invoking traces (cf. Araki relative entropy [20]). (B) Higher categorical coherence breakdown ⇒intrinsic self-measurement. In standard categorical encodings of quantum theory, commutative triangles enforce linear superposition; higher coherence (pentagon/hexagon) ties together composition of embeddings/quantum processes. We posit controlled defects in higher coherence (beyond triangles) valued in local automorphisms of the net. Passing these defects through ρ induces local, completely positive (CP) evolution terms plus a nonlinear modular feedback term in the CDM dynamics. Physically, UV modes become an internal meter that continuously and weakly “monitors” near-boundary energy densities of IR modes (self-measurement), while modular feedback stabilizes the corresponding pointer structures. Concretely, on each region Owe consider a local, covariant CP evolution of the form ∂tρO=−i[HO, ρO]−ZO ddx γ(x) [T00(x),[T00(x), ρO]] −λ[Kρ,O,[Kρ,O, ρO]] ,(1) where HO is the local Hamiltonian, T00 the energy density, and Kρ,O := −ln ∆ O the (state-dependent) modular generator for ( A ( O ) , ρO ). The double commutator with T00 realizes a local GKSL/Lindblad dephasing channel [ 21 , 22 ]; the second, nonlinear double commutator represents higher-coherence-induced feedback in the modular pointer basis. Both terms are local and CP by construction; covariance and positive energy are preserved. Intuitively: the theory contains its own weak, continuous measurement channel—no external apparatus is required. UV modes continually read out coarse energy-density patterns in a thin collar of the entangling surface; the nonlinear modular term enacts a (state-dependent) stabilization of those patterns. This is the “system measuring itself.” What this buys: from measurement to finite entanglement The same mechanism that installs an embedded measurement process dynamically alters short-distance correlations responsible for the QFT entropy problem. In ordinary (linear) AQFT, the two-point/OPE singularity that controls near-boundary correlations behaves like hO ( x ) O ( y ) i∼|x−y|−(d−1) as |x−y| → 0. Under the evolution (1) , high-modular-frequency components near the cut are damped; one can summarize the net effect by an effective UV-dephasing exponent δ > 0, hO(x)O(y)iρt∼ |x−y|−(d−1)+δ(t)(|x−y| → 0), reflecting the chaotic/scrambling character of the intrinsic self-measurement at very short scales. Without introducing any kinematic cutoff or minimal length, this dynamical softening of the correlator changes the scaling of collar-free entanglement diagnostics: • The mutual information of two parallel half-spaces separated by s behaves as I ( A : B ; t ) ∼ Area ×s−(d−2−δ(t)); it is finite as s→0whenever δ≥d−2.
3 • The first-law variation for shape deformations, δS = δhKi , picks up the same softened kernel and becomes finite for planar deformations as soon as δ≥d−2. Thus a sufficiently strong self-measurement (in the precise sense of the modular damping it induces) removes the sharp-limit divergence of QFT entanglement without spoiling locality or positive energy. In the operator-algebraic sense the local algebras remain type III1, but operationally the dynamics creates an effective type-I window on a near-central pointer subalgebra (smeared T00 in a boundary collar), so that entropies at any finite resolution exist and are physically meaningful. Historical context and relation to prior work Our construction builds on and integrates several strands: 1. Algebraic QFT and modular theory. The Haag–Kastler net and Wightman frameworks [ 1 , 3 , 102 ], Reeh–Schlieder [ 4 ], and Tomita–Takesaki modular theory [ 5 ] (with the Bisognano–Wichmann identification for wedges [ 98 ]) provide the kinematical backbone. The type III nature of local algebras and the split property [ 7 , 10 , 103 ] explain why naive entropy fails and why regulators introduce area terms. 2. Entanglement entropy in QFT. The area law and its universal subleading pieces (e.g. log terms tied to anomalies, finite constants in odd spatial dimension for spheres/balls) have been established across free and interacting settings, with field-theoretic and holographic tools [12–19]. 3. Open-system dynamics and constraints. Completely positive, Markovian semigroups (GKSL/Lindblad) [ 21 , 22 ] capture continuous weak measurement and decoherence; we adapt this structure locally at the level of nets to preserve microcausality. We emphasize compatibility with no-signaling: nonlinearity enters via modular functionals of ρ that are local in the algebraic sense. 4. Foundations and nonlinear QM. We are mindful of no-go results (e.g. Gisin and Polchinski [ 27 , 28 ]) that show how unconstrained nonlinear dynamics leads to superluminal signaling in the presence of entanglement. Our construction avoids those pitfalls by keeping evolution CP and local on each A(O)and confining state-dependence to modular objects built from ρOitself. Contributions of this paper 1. We formulate a categorical density matrix (CDM) for AQFT as a natural, modularly enriched assignment of states to regions, suitable for type III local algebras. 2. We introduce a controlled failure of higher categorical coherence (pentagon/hexagon defects) valued in local automorphisms of the net and show how, when transmitted through the CDM, it yields alocal, CP evolution supplemented by a nonlinear modular feedback term, Eq. (1) , implementing intrinsic self-measurement. 3. We prove that Reeh–Schlieder (cyclicity and separating) is structurally stable under this evolution, while the near-boundary UV kernel controlling entanglement is dynamically softened. We extract a quantitative dephasing exponent δfrom the high-frequency tail of the modular spectral density. 4. We show that collar-free entanglement diagnostics (mutual information of adjacent regions; shape derivatives via the first law) become finite as soon as δ≥d− 2, providing a regulator-free criterion for the removal of the sharp-limit divergence. We highlight the universal pieces (anomaly-controlled logs/finite terms) that remain untouched. 5. We discuss operational consequences: an emergent pointer subalgebra (smeared T00 in a boundary collar) becomes approximately central under the dynamics, creating an effective type-I window and making entropies at finite resolution well-defined and measurable in principle.
4 Organization of the paper Section 2 reviews AQFT, Reeh–Schlieder, modular theory, and the entropy problem. Section 3 defines the CDM and its modular enrichment. Section 4 introduces higher coherence defects and derives the local CP + nonlinear modular evolution. Section 5 analyzes Reeh–Schlieder under the deformed dynamics. Section 6 develops the entanglement diagnostics and the δ -criterion. Section 7 presents examples and compares with regulated computations (split property, lattice). Section 8 discusses conceptual implications (measurement as self-dynamics, black-hole parallels), limitations, and open problems. Appendices collect modular-analytic proofs and technical lemmas. Notation. We work on ( d+ 1)-dimensional Minkowski space. Regions O are double cones or wedges as specified; A ( O )denotes the associated von Neumann algebra; ρO the regional CDM state; σO t the modular flow; Kρ,O the modular generator. For disjoint regions A, B , I ( A : B )denotes mutual information (defined algebraically via relative entropy [20]). II. ALGEBRAIC QFT, REEH–SCHLIEDER, MODULAR THEORY, AND THE ENTROPY PROBLEM A. Conventions and notational choices Throughout this section we work on D -dimensional Minkowski spacetime M1,D−1 with metric signature (+,−,...,−). We write D≡spacetime dimension, d ≡D−1 (spatial dimension). For a spacetime region O ⊂ M1,D−1 , A ( O ) ⊂B ( H )denotes the von Neumann algebra of (bounded) local observables associated to O . The causal complement is O0 , and the double complement O00 ; a region is causally complete if O = O00 . We refer to the net of local algebras as a Haag–Kastler net, and we freely use Wightman-field language when useful [1, 3, 102]. B. Haag–Kastler axioms, spectrum condition, and covariance A (vacuum) QFT in the algebraic sense is the data O 7−→ A(O)⊂B(H),Ω∈ H, U :Poincaré → U(H) satisfying: 1. Isotony: if O1⊂ O2then A(O1)⊂ A(O2). 2. Locality (microcausality): if O1⊂ O0 2then [A(O1),A(O2)] = 0. 3. Poincaré covariance: U(g)A(O)U(g)−1=A(gO)for all Poincaré g. 4. Spectrum condition: the joint spectrum of the generators of the translation subgroup in U lies in the closed forward light cone V+(positive energy). 5. Vacuum: there is a unique (up to phase) unit vector Ω ∈ H with U ( a )Ω = Ω for all translations a ; Ωis cyclic for the quasi-local algebra A ≡ SOA(O)k·k. In the Wightman setting, local algebras are generated by smeared fields φ ( f )with supp f⊂ O ; the spectrum condition implies strong analyticity properties (“tube analyticity”) of vacuum correlation functions. C. Reeh–Schlieder theorem: statement, proof skeleton, and physical content Theorem II.1 (Reeh–Schlieder [ 4 ]) . For any nonempty open region O , the vacuum Ωis cyclic for A ( O ): A(O)Ω = H. If O06=∅, then Ωis also separating for A(O):AΩ=0with A∈ A(O)implies A= 0.
5 Proof skeleton (analytic continuation and unique continuation). Let ψ⊥ A(O)Ω, so hψ, AΩi= 0 for all A∈ A ( O ). In a Wightman realization, take A as polynomials in smeared fields supported in O and use translations A(x) = U(x)AU(x)−1. The matrix elements F(x1, . . . , xn) = hψ, φ1(x1)···φn(xn)Ωi define boundary values of functions holomorphic in forward tubes thanks to the spectrum condition. Vanishing on a nontrivial open set propagates by the edge-of-the-wedge theorem, forcing F≡ 0for all configurations, whence ψ = 0. Separating follows from locality and cyclicity of A ( O0 ): if A Ω = 0 with A∈ A(O)then hΩ, BAΩi= 0 for all B∈ A(O0); density of A(O0)Ω gives A= 0. Physical meaning. “Cyclic” means: by acting within any nonempty region one can approximate any state vector—a precise expression of the vacuum’s extreme, scale-agnostic entanglement. “Separating” means: no nonzero local observable annihilates the vacuum. Together, they encode that the vacuum is both maximally fertile (local operations generate everything) and robust (no local eraser). D. Tomita–Takesaki modular theory and Bisognano–Wichmann identification Given a von Neumann algebra M ⊂ B ( H )and a cyclic and separating vector Ω, define the closable anti-linear Tomita operator SΩ : M Ω → M Ωby SΩA Ω = A† Ω. Its polar decomposition SΩ = J ∆ 1/2 yields the modular conjugation J (antiunitary) and the modular operator ∆(positive self-adjoint). The associated modular automorphism group on Mis σΩ t(A) := ∆itA∆−it, t ∈R, and ( M, σΩ t )satisfies the KMS condition at β = 1 with respect to the vector state ωΩ ( A ) = h Ω , A Ω i [ 5 ]. Bisognano–Wichmann. For a wedge region W = {x1>|x0|} and M = A ( W )in a Wightman QFT with suitable assumptions, the modular group σΩ t is implemented by the one-parameter Lorentz boosts preserving W[98]: ∆it=U(ΛW(2πt)), J =(CPT-like reflection on W). Consequently, the modular Hamiltonian K:= −ln ∆ for Wis local: KW= 2πZx1>0 x1T00(x)ddx, (2) and the restriction of Ωto A ( W )is a KMS (thermal) state at local Unruh temperature T ( x1 ) = 1 2πx1 relative to the boost flow. This localization of “temperature” near the entangling plane x1 = 0 underlies the area scaling of entanglement. E. Local algebras are type III and the split property Under general conditions (locality, covariance, spectrum condition, existence of a stress tensor, etc.), local algebras are factors with trivial center, and in fact are typically of type III1 in Connes’ classification [ 7 – 9 ]. Type III means: there is no faithful normal trace; hence there is no density matrix or von Neumann entropy for A ( O )as such. This is the mathematical expression of the vacuum’s infinite, scale-free entanglement structure. The split property [ 10 , 103 ] states that for O1bO2 (strict inclusion with a finite “collar”), there exists a type I factor Nwith A(O1)⊂ N ⊂ A(O2). Physically, the split property says that widely separated degrees of freedom admit an approximate tensor product structure; mathematically, it supplies a regulator internal to the net: inside N∼ =B ( Heff )traces and density matrices do exist, so entropies can be defined and then the split distance is sent to zero.
6 F. Why “entropy” is ill-defined for sharp regions and how the area law appears Because A ( O )is type III , there is no trace and no density matrix for the restriction of Ωto A ( O ). To speak of entanglement “entropy” one must either: • introduce an external UV regulator (lattice spacing, momentum cutoff, Pauli–Villars, point-splitting, . . . ), or •use the split property to insert a type I factor as an intrinsic regulator. In either case, the regulated entropy S ( O ; )diverges in the sharp limit → 0with a characteristic area law S(O;)∼κD Area(∂O) D−2+(γlog(L/)for certain Dand CFTs, finite (shapeand theory-dependent),(3) see [12–17]. Here Lis a macroscopic length scale set by O(e.g. its radius). Rindler/thermal derivation (physical picture). For a half-space x1> 0, Eq. (2) implies the vacuum restricted to A ( W )is thermal with local temperature T ( x1 )=1 / (2 πx1 ). At high temperature, the entropy density of a relativistic QFT scales as s ( T ) ∼αDTd for some constant αD depending on the field content. Integrating the local entropy density over the wedge yields S Area ∼ZL dx1s1 2πx1∼αDZL dx11 (2πx1)d∼const d−1=const D−2, in agreement with (3) . The divergence comes entirely from the thin “thermal” layer near the entangling plane where T(x1)is large; it is, therefore, boundary-local and insensitive to bulk details. Replica/heat-kernel derivation (universal subleading terms). The replica trick computes S = (1 − α∂α ) ln Zα|α=1 , where Zα is the partition function on the α -fold branched cover with a conical singularity along ∂O. In a heat-kernel expansion, ln Zα∼X n≥0 an(α)n−D, the coefficient of −(D−2) is proportional to the area of the entangling surface (nonuniversal), whereas the logarithmic term (when present) is tied to integrated curvature invariants and conformal anomalies on the defect and is universal [17–19]. G. Scheme dependence of the area coefficient and what is universal The coefficient κD in (3) depends on the regulator: lattice discretization (and the precise stencil), momentum cutoff (its shape and smoothness), Pauli–Villars choices, point-splitting kernels, or the choice of intermediate type I factor in the split property all change κD while preserving the form of the area law. Physically, κD renormalizes local counterterms supported on the entangling surface; it is not an observable by itself. By contrast: • Logarithmic coefficients (when present) in even D are universal and tied to conformal anomaly coefficients in CFTs. • For odd D and special shapes (e.g. balls), finite constants (such as F in D = 3 CFTs) are universal. • Relative quantities (Araki relative entropy, mutual information of disjoint regions) are finite and regulator-independent [16, 20]. H. Entropy without traces: relative entropy, modular Hamiltonians, and first laws In a type III setting, the correct information-theoretic object is the Araki relative entropy. For normal states ω, σ on A(O), S(ωkσ) := −hln ∆ω|σiω∈[0,∞],
7 where ∆ ω|σ is the relative modular operator. It satisfies positivity, monotonicity under completely positive, normal maps, and data processing. It provides a rigorous measure of distinguishability accessible by measurements localized in O[20]. For a one-parameter family of states ωλ perturbatively close to a reference σ , the first law of entanglement holds: d dλλ=0 S(ωλkσ) = d dλλ=0 hKσiωλ−S(ωλ)=δhKσi, with Kσ := −ln ∆ σ the modular Hamiltonian for ( A ( O ) , σ ). In cases where Kσ is local (wedges/halfspaces by Bisognano–Wichmann, balls in CFT by a conformal map [ 18 ]), δhKσi is a local integral of stress-tensor one-point functions, giving powerful links to energy conditions (e.g. QNEC) and shape variations of “entropy” without ever defining an absolute entropy. I. Mutual information and the diagnosis of divergences For regions A, B, the mutual information is defined algebraically by I(A:B)ω:= S(ωABkωA⊗ωB)≥0, which is finite for disjoint A, B and encodes correlations accessible by local measurements. For adjacent or touching regions, I ( A : B )inherits the same short-distance divergence as the regulated entropy and thus serves as a collar-free diagnostic of the UV structure: for two parallel half-spaces separated by distance s , I(A:B)∼Area sD−2(s→0), with possible additional logarithms depending on D and the theory [ 16 ]. This makes I ( A : B )a precise observable to test any proposed mechanism that claims to soften or remove the UV divergence. J. Putting it together: the physical narrative The picture emerging from Reeh–Schlieder, modular theory, and type III structure is coherent: • The vacuum is extremely entangled across any spacelike cut (RS), to the point that any region, however small, has access (in principle) to the entire Hilbert space by local operations—albeit requiring unbounded resources. • The modular Hamiltonian for simple regions is local and acts as a generator of boosts; the vacuum looks thermal with a diverging local temperature near the cut (Bisognano–Wichmann). • Because there is no trace on A ( O ), absolute entropies are not defined; when one regulates, a universal area dependence appears with regulator-dependent coefficient. The divergence is a sharp signal of scale-invariant short-distance correlations. • The right invariants in the type III world are relative (Araki relative entropy, mutual information, shape first laws). These are robust, finite, and encode what is physically measurable about entanglement. Foreshadowing our mechanism. In later sections we will dynamically modify only the UV portion of this story—leaving locality, covariance, and positive energy intact—by embedding an intrinsic, local self-measurement channel and a nonlinear modular feedback into the categorical density matrix evolution. The vacuum remains cyclic and separating (RS structure intact), but the near-boundary modular spectrum is damped, softening short-distance correlators. As a result, the collar-free diagnostics ( I ( A : B )and shape first laws) transition from the scaling above to milder behavior, and when the damping is sufficiently strong, they become finite without any external regulator. This is the precise sense in which a solution to the measurement problem (self-measurement) simultaneously resolves the entropy divergence in QFT. Remark on dimensions. We emphasize that Eq. (3) is written with the D -dependent power D− 2 (spacetime dimension), equivalently d− 1(spatial dimension). In D = 4 (three spatial dimensions) this gives the familiar S∼Area/2, while in D= 3 it gives S∼Perimeter/.
8 III. THE CATEGORICAL DENSITY MATRIX (CDM): DEFINITION, MODULAR ENRICHMENT, AND NATURALITY A. From global states to a regional, functorial notion of state In type III local algebras, a conventional, global density matrix does not exist for sharp regions. Nevertheless, every physical prediction is regional: it is a number ρO ( A )obtained by pairing a state on A ( O )with a local observable A∈ A ( O ). This motivates replacing the single global object by a family of regional states, related functorially by restriction along inclusions. Concretely: Categories involved. Let Loc be the category whose objects are causally convex open regions O ⊂ M1,D−1 and whose morphisms are embeddings f : O,→ O0 that preserve causal structure (e.g. inclusions). Let vN be the category of von Neumann algebras with normal unital ∗ -homomorphisms. The Haag–Kastler net is a covariant functor A: Loc −→ vN,O 7−→ A(O), with A(f)the inclusion A(O),→ A(O0)for f:O,→ O0. Let Stat : vN →Conv send a von Neumann algebra M to its convex set of normal states M+ ∗ , and a normal unital ∗-homomorphism ι:M→Nto the restriction map Stat(ι) : N+ ∗→ M+ ∗,ω7→ ω◦ι. Definition III.1 (Categorical density matrix (CDM)) . Acategorical density matrix for the net A is a natural transformation ρ: Loc =⇒Stat ◦A, i.e. an assignment O 7→ ρO∈ A(O)+ ∗such that for every embedding f:O,→ O0, ρO=ρO0◦A(f) = ρO0A(O).(4) If a symmetry group G acts covariantly on the net via unitaries U ( g ), we require covariance of the CDM: ρgO(A) = ρOU(g)−1AU(g),∀g∈G, A ∈ A(gO).(5) We say ρis faithful if each ρOis faithful as a normal state on A(O). Physics. Equation (4) asserts consistency of marginals: the state assigned to a smaller region is the restriction of the one assigned to any larger region containing it. This makes ρ apresheaf of states over spacetime. Covariance (5) ensures the CDM honors spacetime symmetries (Poincaré invariance in the vacuum case). Faithfulness is the algebraic correlate of the separating property (no nonzero positive operator has zero expectation). B. Standard form and modular enrichment at the regional level A central advantage of the CDM is that it brings Tomita–Takesaki modular objects into the kinematics on a region-by-region basis, even in the absence of traces [5, 102]. Proposition III.2 (Regional modular data).Let ρbe a faithful CDM. For each region O: 1. There exists a standard form ( HO, πO, JO,PO )of A ( O )with a cyclic and separating vector Ω O implementing ρOvia ρO(A) = hΩO, πO(A)ΩOi. 2. The Tomita operator SO , modular operator ∆ O and modular conjugation JO are defined on HO ; the regional modular flow is σO t(A) := ∆it OA∆−it O, A ∈ A(O), t ∈R. 3. (A(O), σO t)is a KMS system at β= 1 for the state ρO. Sketch. Faithfulness of ρO yields a GNS representation with cyclic vector Ω O that is separating; the Haagerup standard form then furnishes ( JO,PO ). Tomita–Takesaki yields ∆ O and JO , and the KMS property for (A(O), σO t, ρO)is standard [5].
9 Physics. σO t is the intrinsic “thermal” flow seen by an observer confined to O in the state ρO . For wedges in a Wightman theory, σO t is generated by boosts (Bisognano–Wichmann), so KO := −ln ∆ O is local, cf. Eq. (2) in Sec. II. C. Naturality along inclusions: conditional expectations and cocycle transport Modular flows are not strictly functorial under inclusions in general. Two powerful structures relate modular data between nested regions: (i) Conditional expectations (Takesaki). Let N ⊂ M be von Neumann algebras and ϕ a faithful normal state on M . If there exists a ϕ -preserving conditional expectation E : M→N (i.e. ϕ◦E = ϕ ), then Takesaki’s theorem implies σϕ t(N) = Nand σϕ tN=σϕN t,∀t∈R.(6) In our context, for an inclusion i : O,→ O0 , if there exists a ρO0 -preserving conditional expectation EO0 O : A ( O0 ) → A ( O ), then the modular flow on A ( O0 )leaves A ( O )globally invariant and its restriction equals the modular flow computed with ρO. (ii) Connes–Radon–Nikodym cocycles. For two faithful normal states φ, ψ on the same algebra M , the cocycle derivative [ Dφ : Dψ ] t is a strongly continuous unitary one-parameter family in M satisfying σφ t(A)=[Dφ :Dψ]tσψ t(A) [Dφ :Dψ]∗ t,[Dφ :Dψ]t+s= [Dφ :Dψ]tσψ t[Dφ :Dψ]s, and the chain rule [ Dφ : Dχ ] t = [ Dφ : Dψ ] t [ Dψ : Dχ ] t [ 5 ]. When N ⊂ M and φ, ψ are restrictions of the same global state via possibly different conditional expectations, cocycles implement modular transport between flows on Ninduced from M. These structures motivate the following CDM requirement. Definition III.3 (Split-compatibility and modular transport) . A faithful CDM ρ is split-compatible if for every ObO0 there exists a normal conditional expectation EO0 O : A ( O0 ) → A ( O )such that ρO0◦EO0 O=ρO. The associated modular transport along i:O,→ O0is the pair EO0 O,[D(ρO0◦EO0 O) : DρO0]t, which relates σO tand the restriction of σO0 tto A(O)via Eq. (6) and a cocycle when needed. Physics. Split-compatibility is the CDM avatar of the split property (Sec. II). It ensures that, at finite collar thickness, the modular flows are consistently comparable and that regional thermality fits together across scales. In later sections, higher categorical coherence defects will precisely amount to deforming this transport by a small, local automorphism-valued cocycle. D. A modular connection and its flatness under exact coherence It is convenient to package modular transport into a “connection” on the poset (nerve) of regions. Definition III.4 (CDM modular connection).For each inclusion i:O,→ O0, fix split-compatible data (EO0 O, ui(t)) with ui(t)≡[D(ρO0◦EO0 O) : DρO0]t∈ A(O0). For a composable pair Oi −→ O0j −→ O00, define the holonomy around the 2-simplex by Holj,i(t) := uj(t)σO00 tui(t)uj◦i(t)∗∈ A(O00). We say the connection is flat if Holj,i(t) = 1for all tand all 2-simplices. Proposition III.5 (Flatness under exact coherence) . If the split expectations compose strictly, EO00 O = EO0 O◦EO00 O0 , and ρ is strictly natural (Eq. (4) ) with the same global faithful reference on A ( O00 ), then the CDM modular connection is flat: Holj,i(t) = 1. Proof. By Takesaki’s theorem, σO00 t leaves A ( O0 )and A ( O )globally invariant and restricts to σO0 t and σO t respectively; the ρ -preserving property implies the relevant cocycles are identities. The Connes chain rule for cocycles on A(O00)then gives uj(t)σO00 tui(t)=uj◦i(t), hence Holj,i(t) = 1.
16 •Reeh–Schlieder holds structurally; the local algebras remain type III1 , but operationally a near-central pointer subalgebra emerges at finite resolution, making entropies physically finite. In the next section we analyze in detail how (19) modifies short-distance two-point functions and modular spectra in representative QFTs (free scalar, CFT balls), and we compute the induced softening of mutual information and first-law shape variations. We also discuss constraints from energy conditions and illustrate the absence of superluminal signaling. V. QUANTITATIVE CONSEQUENCES: UV SOFTENING, MODULAR SPECTRA, AND ENTANGLEMENT DIAGNOSTICS A. Set-up and strategy We now quantify how the intrinsic self-measurement dynamics derived in Sec. IV modifies (i) shortdistance two-point structures near entangling surfaces, (ii) modular spectral densities, and (iii) collar-free entanglement diagnostics, focusing on mutual information and first-law (modular) shape variations. We analyze two representative cases: (a) a half-space (Rindler wedge) O=W={x1>0}, where the modular Hamiltonian is local KW= 2πZx1>0 x1T00(x)ddx(23) by Bisognano–Wichmann (Sec. II); (b) a ball BRin a conformal field theory, where KBRis also local by a conformal map. The two building blocks of the dynamics are (Sec. IV): (linear CP) LCP(A) = i[HO, A]−ZO dDx γ(x) [T00(x),[T00(x), A]]; (24) (nonlinear modular) N(ρt) mod(A) = −[Kρt,O,[Kρt,O, A]],(25) combined in Eq. (19). We write the (time-integrated) modular damping strength α(t) := Zt 0 ds γ(s) + λ(s)≥0, `(t) := pα(t),(26) where γ denotes a local average of γ in a near-boundary collar. Intuitively, ` ( t )is the dynamical crossover length below which modular high-frequency modes are strongly suppressed. B. Free scalar wedge: explicit kernel and two-point softening Consider a free real scalar field φ with mass m≥ 0in the vacuum. The equal-time Wightman function in Dspacetime dimensions is W(x) := hφ(0,x)φ(0,0)i=Zddk (2π)d 1 2ωk eik·x, ωk=pk2+m2. Near a smooth entangling hyperplane, the divergence of entanglement diagnostics is driven by the transverse momentum k⊥ to the plane; tangential modes kk integrate to finite prefactors. The linear CP generator (24) with X = Rdx1pγ(x1)T00 ( x1,xk )acts as a Gaussian smearing in the conjugate variable to modular time. In the wedge, modular time and the proper distance x1 are related by the local Unruh temperature T ( x1 )=1 / (2 πx1 ); the cumulant expansion (Sec. IV B) together with the Gaussian Kraus form (18) imply that two-point functions pick up a factor e−2α(t)ω2 in modular frequency ω . For translation-invariant equal-time correlators across the plane this corresponds (via the BW map) to a Gaussian in the transverse momentum, Wt(x1;xk)≃Zdd−1kk (2π)d−1Z+∞ −∞ dk⊥ 2π eik⊥x1+ikk·xk 2qk2 k+k2 ⊥+m2e−2α(t)k2 ⊥.(27)
17 This is the precise sense in which the self-measurement dynamics softens the short-distance kernel: modes with |k⊥| `(t)−1are exponentially suppressed. Proposition V.1 (Short-distance softening of two-point singularity) . Fix m≥ 0and d≥ 2. For any fixed tangential separation xk, the damped correlator (27) has the short-distance expansion as x1→0: Wt(x1;xk) = Cd|x1|−(d−1) +O(|x1|−(d−3))if |x1| `(t), C0 d `(t)d−1+O|x1|2 `(t)d+1 if |x1| `(t), for constants Cd, C0 d> 0depending on d, m . In particular, the point-split short-distance singularity saturates to a finite `(t)-controlled value at fixed time t > 0. Proof. Perform the k⊥ -integral asymptotically via Laplace’s method. For |x1| ` , the Gaussian is broad in k⊥ and the integral reduces to the standard power singularity ∼ |x1|−(d−1) . For |x1| ` , the Gaussian limits the effective k⊥ -domain to |k⊥|.`−1 , so the oscillatory factor eik⊥x1≈ 1, and the integral scales as R∼`−1 dk⊥ ( ··· ) ∼`−(d−1) . Tangential momenta produce finite prefactors by standard power counting. Physics. Whereas the linear theory exhibits a scale-free pile-up of pairwise correlations across the plane down to x1→ 0, the self-measurement-dressed theory dynamically saturates the near-boundary correlation density at the crossover ` ( t ). No microscopic cutoff has been introduced; ` ( t )is generated by the macroscopic evolution of the state under a local, CP dynamics. C. Mutual information of adjacent half-spaces: crossover and finiteness Let A = {x1< 0 } and B = {x1> s} be two parallel half-spaces separated by distance s > 0in the vacuum state evolved to time t by (20) . The mutual information I ( A : B ) t is finite for s > 0and diverges as s→ 0in linear AQFT as s−(D−2) . We now show that the modular damping induces a crossover from the area-law divergence to a finite area-proportional limit once s`(t). Theorem V.2 (Half-space mutual information: dynamical crossover) . Suppose the modular spectral density in the wedge is dressed by the Gaussian factor e−2α(t)ω2 as in (21) . Then the mutual information between Aand Bsatisfies, for s→0, I(A:B)t= Area ×(cDs−(D−2) 1 + o(1)if s`(t), ˜cD`(t)−(D−2) 1 + o(1)if s`(t),(28) with positive constants cD,˜cD depending only on the theory and dimension. In particular, for any fixed t > 0,I(A:B)thas a finite s→0limit proportional to the area. Sketch. For Gaussian (quasi-free) states, I ( A : B )can be bounded above and below by integrals of two-point functions [ 29 , 30 ]. The wedge modular representation reduces the calculation to a spectral integral of the form I(A:B)tArea Z∞ 0 dω ωD−3e−2α(t)ω2e−ωs. For s` , the Gaussian is flat over the dominant ω∼ 1 /s , recovering the s−(D−2) scaling by scale invariance. For s` , the integral is dominated by ω. 1 /` , giving a constant ∼`−(D−2) . Rigorous bounds follow by splitting the integral at ω0 = min{ 1 /s, 1 /`} and comparing to incomplete gamma functions. Power-law tails and the δ -criterion. If the modular spectral density exhibits only a power-law softening, J(ω)∼ωD−3−δat large ω(Sec. IV E), the same estimates yield I(A:B)t∼Area ×s−(D−2−δ)(s→0), with the borderline δ = D− 2giving a logarithm and δ > D − 2giving a finite limit. This recovers the qualitative δ-criterion of Sec. IV with a quantitative spectral proof.
18 D. First-law (modular) shape variations: finiteness under damping Consider the planar entangling surface of the wedge and a small normal deformation x1 = ζ ( xk )with ζ compactly supported and kζk 1. The first-law of entanglement (Sec. III) gives the variation δS =δhKWi= 2πZdd−1xkZ∞ 0 dx1δζ(xk)T00(x1,xk), to linear order in ζ for states where KW is local (see also [ 31 , 32 ]). At second order in ζ (the leading UV-sensitive piece), one finds a quadratic functional δ2S=1 2Zdd−1xkdd−1ykζ(xk)K(xk−yk)ζ(yk),(29) with kernel K given by an integral of the connected two-point function hT00 ( x ) T00 ( y ) ic over x1, y1≥ 0with weights 2 πx1, 2 πy1 [ 31 , 34 ]. In linear AQFT, K ( r ) ∼ |r|−(d+1) at small |r| , producing a UV divergence in the local limit. Theorem V.3 (Finiteness of planar shape variation under modular damping) . Assume the selfmeasurement dynamics dresses the wedge TT -correlator by a Gaussian in modular frequency as in (21) . Then the quadratic kernel K in (29) is integrable at r = 0 and δ2S is finite for any ζ∈H1 ( Rd−1 ) (square-integrable with square-integrable gradient). Moreover, there exists Cd>0such that δ2S≤Cd`(t)−(D−2) kζk2 H1.(30) Sketch. In momentum space along the plane, K ( q )is proportional to the modular spectral density of TT restricted to the wedge and weighted by the polynomial arising from the x1, y1 integrals [ 31 ]. The Gaussian factor e−2αω2 implies K ( q ) .R∞ 0dω ωD−1e−2αω2∼α−(D/2) , uniformly in q ; differentiating under the integral bounds the q -growth, leading to K ( q ) .α−(D−2)/2 (1 + q2 ). Hence Rdd−1q|ˆ ζ ( q ) |2K ( q ) . α−(D−2)/2kζk2 H1, proving (30). Physics. The second variation of entanglement under shape deformations probes the same nearboundary UV kernel that underlies the area law. The self-measurement dynamics damps precisely the modular frequencies that produce the |r|−(d+1) singularity, rendering the quadratic form finite and controlled by `(t). E. CFT balls: modular flow, excitations, and softening In a D-dimensional CFT, the modular Hamiltonian for a ball BRof radius Rat t= 0 is KBR= 2πZ|x|<R R2−|x|2 2RT00(0,x)ddx.(31) Consider a small localized excitation with stress-tensor one-point function δhTµν i. The first law gives δS =δhKBRi= 2πZ|x|<R R2−|x|2 2RδhT00iddx, which is finite already in linear theory. The UV sensitivity appears at second order and in mutual information between adjacent balls. The self-measurement dynamics in a CFT is naturally encoded by the same modular Gaussian damping in the conformally-mapped wedge. Thus the conclusions of Thm. V.2 and Thm. V.3 transfer to balls: at separations below ` ( t ), I ( BR : B0 R )saturates to a finite, area-proportional limit, and second-order shape functionals are finite with ` ( t )-controlled bounds. For non-spherical deformations, our result dovetails with the known shape-kernel analyses in CFT [ 33? , 34 ]: the local power-law divergences there are replaced by finite, `-dependent coefficients here.
19 F. Extracting the dephasing exponent δfrom modular tails Beyond the Gaussian regime, suppose the dressed modular spectral density of a local operator O obeys, at large ω, JO(ω)∼ωβO−δ, δ ≥0, with βOthe linear-theory exponent (βT=D−3for TT in wedges by dimensional analysis). Then: Proposition V.4 (General δ -criterion) . Let I ( A : B )be the mutual information of two adjacent half-spaces at separation s, and let δbe as above for the TT sector. Then as s→0, I(A:B)∼Area × s−(D−2−δ), δ < D −2, log(1/s), δ =D−2, finite, δ > D −2. Proof. Same spectral integral as in Thm. V.2 with ωD−3→ωD−3−δ for the tail; Tauberian theorems or direct estimates give the stated asymptotics. Physics. The strength of self-measurement (how rapidly modular high frequencies are damped) is captured by δ . Gaussian damping corresponds to an effective δ≥D− 2when probed at separations s` : all pairwise, Bell-like UV correlations are flattened into many-body scrambled correlations that do not pile up at the cut. G. Operational entropies at finite resolution: pointer window Define a finite-resolution conditional expectation E` : A ( O ) → N` onto the pointer subalgebra N` generated by T00 ( f )with f supported in a collar of thickness ∼` and with tangential smearing scale ∼` . Set ρt,` := ρt,O◦E`. Since N`is (approximately) type I, its entropy S(ρt,`)is well-defined. Proposition V.5 (Area-law at finite resolution; saturation) . There exist constants C1, C2> 0such that for any region with smooth boundary, C1 Area(∂O) `D−2≤S(ρt,`)≤C2 Area(∂O) `D−2. Moreover, if `` ( t )(weak damping), C1,2 approach the linear-theory values; if `` ( t )(strong damping), S ( ρt,` )saturates and ceases to grow as `→ 0(at fixed t ), reflecting the finiteness of the collar-free diagnostics in Secs. V C–V D. Idea. Upper bound: data processing for relative entropy under E` , combined with the wedge KMS picture and the local entropy density s∼Td integrated over the collar of thickness ` . Lower bound: restrict to independent pointer cells of size `D−1 along the boundary and use strong subadditivity. Saturation for ``(t)follows because the dressed modular spectrum has no weight beyond ∼`(t)−1. Physics. At any finite operational resolution ` , the entropy scales with area as usual; the new effect is that the self-measurement dynamics generates a physical saturation scale ` ( t ), beyond which finer resolution does not unearth more pairwise entanglement: the divergence is dynamically quenched. H. Constraints: energy conditions, QNEC, and no signaling Our evolution (19) preserves microcausality and is CP at each instant. It also preserves the informationtheoretic inequalities that underlie modern energy conditions. By the first law, shape deformations along null directions relate δS to null components of the stress tensor and lead to the quantum null energy condition (QNEC) [ 31 , 32 , 36 ]. Because (i) relative entropy is contractive under CP maps and (ii) our modular term is a CP dephasing at each instant (Prop. IV.2), the proofs of QNEC that rely on relative entropy monotonicity extend to the dressed theory: damping reduces modular variances and thus cannot violate QNEC; if anything, it moves the theory deeper into the allowed region of the inequality. No-signaling is guaranteed by locality of the generator (Prop. IV.1 and Thm. IV.3): operations in O1 spacelike to O2 remain commuting under the flow, so reduced dynamics on A ( O2 )is unaffected by actions in O1.
20 I. Summary of Section V • The self-measurement dynamics damps modular high frequencies by a factor e−2α(t)ω2 (Gaussian regime) or more generally by a power ω−δ. • Two-point functions across a cut are softened: the short-distance singularity saturates at the dynamical scale `(t) = pα(t)(Prop. V.1). • Mutual information of adjacent half-spaces exhibits a crossover from the linear-theory area divergence to a finite area-proportional plateau (Thm. V.2); for power-law tails one obtains the general δ - criterion (Prop. V.4). • First-law shape variations (quadratic kernel) become finite with ` ( t )-controlled bounds (Thm. V.3); the same holds for CFT balls. • Operational entropies at finite resolution obey an area law with coefficients that saturate once ``(t), reflecting the finiteness of collar-free diagnostics. •Energy conditions (QNEC) and no-signaling remain intact; CP and locality are built in. VI. DERIVING THE INTRINSIC DYNAMICS FROM COHERENCE DEFECTS: STOCHASTIC LIMITS, H–THEOREMS, AND CAUSALITY A. Aim and overview The previous sections motivated and used the regional evolution ∂tρO(A) = ρOi[HO, A]−ZO dDx γ(x) [T00(x),[T00(x), A]] −λ[Kρt,O,[Kρt,O, A]],(32) which we interpreted as intrinsic self-measurement (linear CP dephasing of T00 ) plus a nonlinear modular feedback (dephasing in the Kρ,O -eigenbasis). In this section we (i) derive the linear CP part from higher-categorical modular holonomy via a quantum stochastic limit, (ii) justify the nonlinear modular term as a steepest-descent for relative entropy and give a CP Kraus representation at each time slice, (iii) prove an H –theorem for relative entropy, (iv) establish causality/no-signaling and covariance of the flow, and (v) discuss stationary states and structural stability (type and Reeh–Schlieder). Notation: We use explicit O,A for regions/algebras (instead of shorthand macros) to avoid compilation issues. For an operator X,adX(A) := [X, A]and ad2 X(A) := [X, [X, A]]. B. From modular holonomy noise to a GKSL generator Recall from Sec. IV that higher categorical coherence breakdown appears as a local automorphismvalued 2–cochain wj,i ( t ) ∈ U ( A ( O00 )) relating modular transports along inclusions; small holonomies around 2–simplices are w(∆t) = expi√2 ∆t X −∆t Y +o(∆t), X =X†, Y =Y†∈ A(O),(33) with X, Y local (smeared) operators, cf. Eq. (4.3). Iterating n = t/ ∆ t steps along a grid of inclusions generates a product of such unitaries. As in repeated-interaction and quantum stochastic limits [ 40 , 41 , 43, 93], a central-limit scaling produces a conservative, completely positive semigroup. Theorem VI.1 (Holonomy central limit ⇒ GKSL generator) . Let {Uk}n k=1 be i.i.d. random unitaries of the form Uk = exp (i √2∆t Xk− ∆ t Y )with E [ Xk ]=0, E [ XkX` ] = δk`X2 and X = X† , Y = Y† local, bounded on a common invariant core. Define Φ(∆t) t(A) := EU† 1···U† nA Un···U1, n =t/∆t.
21 Then Φ (∆t) t→etL in the strong operator topology on the local C ∗ -algebra as ∆ t→ 0, with GKSL generator L(A) = i[Heff, A]−ad2 X(A)−adY(A), Heff ∈ A(O)†.(34) If Yis absorbed into Heff, the CP part is precisely −ad2 X. Sketch. A second-order cumulant (Kubo) expansion gives, for one step, E ( U†AU ) = A + ∆ t (i[ Heff, A ] − ad2 X ( A ) −adY ( A )) + o (∆ t ), with Heff the coherent drift from Itô correction and commutators of X, Y . The Trotter–Kato product formula then yields Φ (∆t) t→etL [ 44 ]. Complete positivity follows because −ad2 Xis of Lindblad form with selfadjoint Lindblad operator X[21, 22]. Local choice of X .Locality and covariance single out the smeared energy density T00 as in Eq. (4.9): X=ZO dDxpγ(x)T00(x), with γ≥0compactly supported in O. This yields exactly the linear CP part in (32): LCP(A) = i[HO, A]−ZO dDx γ(x) [T00(x),[T00(x), A]].(35) The construction is a quantum Brownian limit of modular holonomy noise [43, 93]. C. Nonlinear modular feedback as steepest KMS descent Let σO be a faithful reference (e.g. vacuum) on A ( O )and consider the Araki relative entropy Φ[ ρO ] := S ( ρOkσO ). The (formal) gradient flow of Φwith respect to the Bogoliubov–Kubo–Mori (BKM) metric on the state manifold induces, on observables, the Laplacian in modular time [46, 95]: N(ρ) mod(A) = −ad2 Kρ,O(A), Kρ,O:= −ln ∆ρ,O.(36) At each time slice t, the nonlinear map Φρt,∆t(A) = exp−λ∆tad2 Kρt,O(A) = 1 √4πλ∆tZR ds e−s2/(4λ∆t)eisKρt,OAe−isKρt,O,(37) is a convex mixture of *-automorphisms (unitary conjugations), hence completely positive and unital for each fixed ρt; expanding gives Φρt,∆t= id + λ∆tN(ρt) mod +o(∆t), as in Prop. 4.3. Physics. Equation (37) dephases in the Kρ,O basis. Because Kρ,O is local (exactly for wedges; approximately for smooth cuts), this is a local pointer selection near the entangling surface. D. An H–theorem for relative entropy We prove that the intrinsic dynamics (32) monotonically reduces relative entropy to a reference σO. Theorem VI.2 ( H –theorem) . Let σO be faithful on A ( O ). Consider the piecewise-constant-in-time evolution that, on each slice [ t, t + ∆ t ), applies the CP map Ψ t,∆t := e∆tLCP ◦ Φ ρt,∆t to ρt (with Φ ρt,∆t from (37)). Then the Araki relative entropy S(ρt,OkσO)is nonincreasing: S(ρt+∆t,OkσO)≤S(ρt,OkσO)for all ∆t > 0. In the continuous limit, d dt S(ρt,OkσO)≤0. Proof. By Uhlmann monotonicity (data processing) [47, 95], SΨt,∆t(ρt,O)kΨt,∆t(σO)≤S(ρt,OkσO), for any CP, normal, unital Ψ t,∆t . Both e∆tLCP and Φ ρt,∆t are CP, unital; their composition is CP, unital. Because e∆tLCP leaves σO invariant when it satisfies detailed balance w.r.t. σO (see below), and Φ ρt,∆t leaves σO invariant if [ Kσ,O,· ]generates σ ’s modular flow (true by definition), we have Ψ t,∆t ( σO ) = σO , hence S ( ρt+∆tkσ ) ≤S ( ρtkσ ). Divide by ∆ t and take the limit using Trotter–Kato to obtain the differential form.
22 Detailed balance for the linear CP part. The generator (35) satisfies KMS detailed balance with respect to a stationary reference σO(e.g. the wedge KMS state) if hA†LCP(B)iσ=hLCP(A)†Biσ, with the KMS inner product hA, Biσ := R1 0ds σ ( A†σs ( B ) σ−s ); this holds when γ is constant along modular orbits and HO is the modular Hamiltonian (or commutes with it) [ 46 , 48 ]. In that case, etLCP leaves σinvariant. E. Causality and no signaling We formalize the statement that the intrinsic dynamics does not enable superluminal signaling. Theorem VI.3 (Microcausality and product structure) . Let O1,O2 be spacelike separated and let Ai := A ( Oi ), i = 1 , 2. Suppose γ ( x )is supported in O1 and Kρt,O1 acts on A1 . Then the Heisenberg generator L1 acting on A1 satisfies [ L1 ( A1 ) , A2 ] = 0 for all Ai∈ Ai . Consequently, the joint flow on A1∨A2 factorizes as et(L1⊕0) = etL1⊗id , and the reduced dynamics on A2 is the identity: operations in O1cannot influence expectation values of observables in O2. Proof. For the linear CP part, LCP,1(A1) = i[HO1, A1]−ZO1 dDx γ(x) [T00(x),[T00(x), A1]], and [ T00 ( x ) , A2 ] = 0 for spacelike separation, hence [ LCP,1 ( A1 ) , A2 ] = 0. For the modular term, Kρt,O1 is affiliated to A1 ; ad2 Kρt,O1 ( A1 )remains in A1 , so it commutes with A2 . The factorization of the flow and the identity action on A2 follow from the uniqueness of solutions to ∂tAt = L1 ( At )with At∈ A1 (and similarly for states). Physics. Linearity is not required for no signaling; locality of the generator suffices. The nonlinear term depends on ρt but only through Kρt,O1 , which is a local functional; it cannot encode instantaneous information about spacelike-separated marginals. F. Stationary states and long-time behavior We state existence and qualitative properties of stationary states for (32). Proposition VI.4 (Stationary points and LaSalle invariance) . Let σO be a faithful stationary state for the linear CP semigroup etLCP (detailed balance), and suppose γ, λ are bounded with R∞ 0 ( γ ( t ) + λ ( t )) dt = ∞ . Then t7→ S ( ρt,OkσO )is nonincreasing and bounded below by 0, hence convergent. Every ω in the ω –limit set satisfies [Kω,O,A(O)] = 0 (i.e. ωis modular-diagonal on A(O)) and is stationary for the full dynamics. If the linear semigroup is primitive (unique faithful fixed point) [46, 49], then ω=σO. Idea. The H –theorem (Thm. VI.2) gives monotonicity and boundedness of S ( ρtkσ ). LaSalle’s invariance principle for nonincreasing Lyapunov functions implies that the ω –limit set lies in the set where the production rate vanishes. For the modular term, the production rate is ∝λhad2 Kρ,Oiρ (Spohn inequality [ 46 ]), which vanishes iff ρ is diagonal in the Kρ,O basis. Primitivity of the linear semigroup forces uniqueness of the limit. Physics. The system self-thermalizes in the modular sense near the entangling surface: modular coherences are erased, leaving a pointer subalgebra. If the CP part already has a unique faithful stationary state, the dynamics flows to it.
23 G. Structural stability: factor type and Reeh–Schlieder The evolution (32) acts by normal, faithful, CP, unital maps on each A ( O )and preserves inclusions. Such maps preserve factoriality and do not change the von Neumann type (heuristically: a type III factor cannot be CP-contracted into a type I algebra without a singular collapse of the net); see [ 50 , Ch. IV] for background on stability under normal maps. Together with Theorem VI.3, this leaves the Haag–Kastler axioms intact. A detailed RS stability argument was outlined in Prop. 4.4; here we note that cyclicity and separating are stable under local CP perturbations that respect the spectrum condition, by the same tube analyticity/edge-of-the-wedge logic as in Sec. II. H. Covariance and detailed balance: wedge KMS case For the wedge W with modular Hamiltonian KW (Bisognano–Wichmann), choose γ constant on boost orbits and HO commuting with the boost generator. Then LCP is covariant under boosts and satisfies detailed balance with respect to the wedge KMS state σW [ 48 , 49 ]. Consequently, σW is stationary for the linear CP part, and the modular feedback leaves σW invariant (because it is a convex mixture of conjugations by eisKW). Thus σWis a fixed point of the full dynamics. I. Physical summary • The linear CP part of the intrinsic dynamics arises from a quantum Brownian limit of small modular holonomies (Theorem VI.1), selecting X∝T00 by locality. • The nonlinear modular feedback is a steepest KMS descent for relative entropy, with an instantaneous CP Kraus representation (Eq. (37)). • Relative entropy to a faithful reference state is an Lyapunov function ( H –theorem, Theorem VI.2); long-time limits are modular-diagonal and, under primitivity, unique (Proposition VI.4). • The dynamics is causal and non-signaling (Theorem VI.3); Poincaré covariance and wedge detailed balance can be imposed naturally. • Structural properties (factor type, RS) are preserved; operationally, a pointer subalgebra emerges at finite resolution, matching the finite, collar-free entanglement diagnostics derived in Sec. V. VII. EXAMPLES, BENCHMARKS, AND COMPARISON WITH CONVENTIONAL REGULATORS A. Scope and goals We now test the intrinsic self-measurement dynamics developed in Sections IV–VI against explicit models and standard regularizations. Our aims are: (i) to compute two-point softening and mutual information in representative free-field settings where the modular Hamiltonian is local (wedges, CFT balls); (ii) to show how the dynamical crossover scale ` ( t ) = pα(t) reproduces and dominates the area divergence at short distances; (iii) to compare with the split property (type I intermediary) and with lattice regulators in Gaussian states, deriving quantitative inequalities; and (iv) to outline a holographic consistency check. Throughout we avoid gauge subtleties (free neutral scalar, or CFTs with well-defined stress tensor); generalizations follow the same pattern with suitable care for constraints. B. Free scalar in a wedge: explicit Fourier analysis Consider a free real scalar in D spacetime dimensions with Wightman two-point function at equal times, W(x) = hφ(0,x)φ(0,0)i=Zddk (2π)d eik·x 2ωk , ωk=pk2+m2, d =D−1.(38)
24 Let x1 be the proper distance to the entangling plane x1 = 0 and xk∈Rd−1 tangential coordinates. The intrinsic dynamics (linear CP + modular feedback) damps high modular frequencies; by the Bisognano–Wichmann map this appears as a Gaussian suppression in the transverse momentum k⊥, Wt(x1,xk) = Zdd−1kk (2π)d−1Z∞ −∞ dk⊥ 2π eik⊥x1+ikk·xk 2qk2 k+k2 ⊥+m2e−2α(t)k2 ⊥.(39) Lemma VII.1 (Gaussian convolution in the transverse coordinate) . Let G` ( z ) = (4 π`2 ) −1/2e−z2/(4`2) with `=pα(t). Then Wt(x1,xk) = ZR dz G`(z)f W(x1−z, xk), where f Wis the partial inverse Fourier transform of Wwith respect to k⊥only. Proof. Fourier inversion in k⊥ gives f W ( ξ, xk ) = Rdk⊥ 2πeik⊥ξF ( k⊥,xk )with F ( k⊥,kk ) = 2qk2 k+k2 ⊥+m2−1. Multiplying by e−2αk2 ⊥corresponds in ξto convolution with G`. Short-distance behavior. Setting xk = 0 and expanding (39) for |x1| → 0yields the softening already stated in Prop. V.1: for |x1| ` , Wt ( x1,0 ) = C0 d`−(d−1) + O ( |x1|2`−(d+1) ), while for |x1| ` one recovers the usual |x1|−(d−1) singularity [51]. C. Mutual information: exact integrals and asymptotics For two parallel half-spaces separated by s > 0, the mutual information I ( A : B ) t can be bounded in Gaussian theories by transverse-momentum integrals of the damped two-point kernel [ 29 , 30 ]. In the wedge basis, a convenient representation is I(A:B)tArea ×Z∞ 0 dω ωD−3e−2α(t)ω2e−ωs,(40) valid up to multiplicative constants depending on field content. The integral (40) is elementary in terms of incomplete gamma functions: Z∞ 0 dω ωνe−aω2−bω =1 2a−(ν+1)/2eb2/(8a)Γ ν+1 2D−(ν+1) b √a, with Dµ a parabolic cylinder function. Its smalls asymptotics gives precisely the crossover stated in Theorem V.2: I(A:B)t= Area ×(cDs−(D−2) [1 + o(1)], s `(t), ˜cD`(t)−(D−2) [1 + o(1)], s `(t).(41) The constants are cD = Γ( D− 2) (up to theory-dependent prefactors) and ˜cD∝ 1from the ω. 1 /` contribution. Two dimensions (1+1 CFT). For a CFT with central charge c , the interval entropy is S ( ` ) = c 3log ( `/ ) + ··· [ 37 ]. Two adjacent intervals of total length L = `1 + `2 and separation s→ 0have I∼c 3log ( L/s )[ 38 ]. In our framework, the self-measurement replaces the short-distance s by an effective seff =ps2+κ `(t)2(a direct consequence of Lemma VII.1 and conformal mapping), giving ICFT(s;t)≃c 3log L ps2+κ `(t)2+··· ,(42) so that I saturates to c 3logL/ ( √κ ` ( t )) as s→ 0. Here κ is a shape-dependent numerical factor (of order unity) fixed by the precise definition of the modular damping map.
25 D. Shape variation for a planar cut: kernel estimates Let ζ ( xk )be a small normal displacement of a planar cut. The quadratic second variation δ2S is a bilinear form with kernel K built from the TT two-point function (Sec. V). Using (39) , one obtains in momentum space along the plane: Kt(q) = CDZ∞ 0 dω ωD−1e−2α(t)ω2ω2 ω2+q2,(43) with CD> 0. The integral is convergent at large ω and bounded uniformly in q by Kt ( q ) ≤C0 D` ( t ) −D . Moreover, differentiating under the integral, |∂qKt(q)| ≤ C00 D`(t)−(D−2), which implies: Proposition VII.2 (Bounded quadratic form).For any ζ∈H1(Rd−1), δ2S(t) = 1 2Zdd−1q (2π)d−1Kt(q)|b ζ(q)|2≤C `(t)−(D−2) kζk2 H1. Proof. Use Kt ( q ) ≤a + bq2 with a∝`−D and b∝`−(D−2) from (43) . Then R ( a + bq2 ) |b ζ|2≤ akζk2 L2+bk∇ζk2 L2. E. Comparison with the split property: quantitative bounds Fix nested regions O1bO2 with collar thickness δ . The split property provides a type I factor N with A ( O1 ) ⊂ N ⊂ A ( O2 )[ 10 , 103 ]. Let E : A ( O2 ) → N be a normal conditional expectation and define the split entropy of a state ρby Ssplit(ρ;δ) := Sρ◦E(well-defined because N ≃ B(Heff)). The intrinsic dynamics provides a physical coarse-graining channel E`(t) onto the pointer subalgebra supported in a collar of thickness ∼` ( t ). The following shows that, whenever ` ( t ) δ , the two procedures agree up to exponentially small errors. Theorem VII.3 (Split vs intrinsic coarse-graining) . Assume δ` ( t )and that γ is supported in a collar of thickness ∼` ( t )inside O2\O1 . Then there exists a normal conditional expectation E : A ( O2 ) → N such that, for all normal states ρ, ρ◦E−ρ◦E`(t)1≤C e−c δ2/`(t)2, with constants C, c > 0independent of ρ. Consequently, Ssplit(ρ;δ)−S(ρ◦E`(t))≤C0e−c δ2/`(t)2. Sketch. The Gaussian transverse damping in Lemma VII.1 implies an approximate Markov property across collars of thickness ` ( t )(exponentially decaying conditional mutual information). One constructs E by averaging over modular translations in the collar and projecting onto the pointer subalgebra; Lieb–Robinson-type locality bounds for relativistic QFT (implemented here by microcausality and the Gaussian kernel) yield the stated estimate on trace distance. The entropy bound follows from Fannes–Audenaert continuity with dimension replaced by an effective local mode count ∝Area δ `−(D−1) [52]. Physics. The split property inserts a type I factor by hand; the intrinsic dynamics produces an effective type I window at finite resolution. When the split collar is wider than ` ( t ), the two descriptions coincide to exponential accuracy. F. Gaussian lattice benchmark: correlation matrix method Discretize the transverse direction with spacing a and infinite extent along the plane. For a free Gaussian state, the reduced state on a set of sites is completely determined by the two-point (covariance) matrix; its entropy is S=X jhνj+1 2log νj+1 2−νj−1 2log νj−1 2i,(44)
32 Physics. (49) exhibits the usual informational backaction (innovation term) alongside the physical dephasing (Lindblad terms). In our framework, both arise from higher-coherence defects: the former is the (fictitious) “readout” of the internal meter; the latter is the objective dephasing. C. Almost-sure collapse and the Born rule The diffusive SME (49) for a nondegenerate pointer yields almost-sure collapse to pointer eigenspaces with Born weights. Theorem IX.3 (Collapse to pointer sectors with Born weights) . Assume Π ∆ has a purely discrete, nondegenerate spectrum {πk} in the relevant energy sector and that [ HO, Π ∆ ]is negligible on the timescale of monitoring. Then for any initial faithful state ρ0,O, 1. along almost every measurement trajectory, bρt,O→Pkbρ0,OPk/tr(ρ0,OPk)as t→ ∞, for some k, 2. P(collapse to k) = tr(ρ0,OPk)(Born rule), where Pkis the spectral projection of Π∆at πk. Sketch. Standard martingale arguments in quantum filtering (Belavkin) show that Mk ( t ) := tr ( bρtPk ) is a bounded martingale with E [ Mk ( t )] = tr ( ρ0Pk ); Doob’s martingale convergence theorem gives a.s. convergence Mk ( ∞ ) ∈ { 0 , 1 } (nondegeneracy + nondemolition). The innovation drives the posterior mean h Π ∆ibρt toward one of the πk , and the dephasing kills off-diagonals (see, e.g., [ 80 – 82 ]). The limit is a rank-one projection state within the measured sector, and the probability of each sector is its initial weight. Physics. This is the intrinsic (no external apparatus) realization of outcome selection: the system’s own UV modes weakly monitor Π ∆ and drive collapse with the standard Born probabilities, while preserving locality and positivity. D. Consistent histories and decay of interference For a sequence of coarse projectors {P(j) αj} in Fixt ( O )at times t1<··· < tn , the decoherence functional is D(α,β) := trP(n) αn···P(1) α1ρ0P(1) β1···P(n) βn, (Heisenberg evolution suppressed). In linear theory, Dis typically not diagonal. Under our dynamics: Proposition IX.4 (Exponential suppression of off-diagonal histories) . Let P(j) αj, P(j) βj be spectral projectors of pointer fields with k Π (j) ∆k uniformly bounded and α6 = β for at least one j . Then, for monitoring rates γ, λ bounded away from zero on [0, T], |D(α,β)| ≤ Cexp−cZT 0 ds [γ(s) + λ(s)] δ2 α,β, where δα,βmeasures the minimal spectral gap between distinct histories (in an appropriate norm). Idea. Insert identity resolutions in terms of Π ∆ ; each interval contributes a Gaussian dephasing factor e−2αω2 in modular frequency (Sec. V). A Trotter product bound in operator norm then gives the exponential suppression of off-diagonal terms. Details mirror standard proofs of decoherence for continuous monitoring [83, 84]. Physics. Histories built from pointer observables become consistent (approximately orthogonal); classical stochastic processes for coarse energy-density patterns emerge.
33 E. WAY-type constraints and compatibility with conservation laws The Wigner–Araki–Yanase (WAY) theorem constrains exact projective measurements of observables that do not commute with additive conserved quantities (e.g. total energy) [ 85 , 86 ]. Our pointer Π ∆ is a localized energy density, so [Π ∆, H ] 6 = 0 in general; thus only approximate measurements at finite resolution are permissible. Theorem IX.5 (WAY-consistent accuracy bound) . Let T > 0be the monitoring time and ∆Π ∆ the RMS error of the posterior estimate from the current Yt. Then there exists C > 0such that ∆Π∆≥C Tk[Π∆, HO]k, independently of the initial state, with equality (up to constants) in the weak-measurement, long-time limit. Sketch. Follow Ozawa’s universally valid WAY-type inequality for continuous measurements [ 87 ] applied to the instrument (48) generated by (49) with η > 0. The error–disturbance tradeoff bound includes a commutator term with H ; for diffusive monitoring over time T , the Fisher information grows as ∝T , yielding the stated scaling. Physics. Self-measurement respects conservation-law constraints: it is as accurate as allowed by WAY-type bounds, which is exactly what one expects of a physical (rather than postulated) measurement channel. F. Intersubjective agreement and objectivity at the pointer level Two observers restricted to spacelike-separated collars C1,C2 (disjoint along the entangling surface) measure commuting pointer fields. Under the intrinsic dynamics, their reduced states evolve according to identical CP+modular flows; hence their posterior distributions on pointer outcomes agree (up to statistical noise) and concentrate on the same classical value in the long-time limit. Proposition IX.6 (Emergent objectivity in the pointer sector) . Let Π (1) ∆ and Π (2) ∆ be two commuting pointer fields with disjoint supports. Consider independent diffusive readouts with identical monitoring rate. Then the joint posterior state bρt becomes close (in trace distance) to a spectrum broadcast structure on the pointer sector: bρt≈X k pkπkihπk⊗σ(1) k⊗σ(2) k⊗ρk,rest, with σ(i) k approximately perfectly distinguishable for different k , and pk the Born weights. The approximation improves as t→ ∞. Idea. Standard arguments of quantum Darwinism/spectrum broadcast structures adapted to continuous measurements (see [ 59 ] and [ 88 ]) show that independent monitoring channels imprint classical labels redundantly; commuting supports ensure no-signaling and factorized records. The intrinsic dephasing provides the decoherence needed for objectivity. Physics. The classical world emerges as redundant records of pointer variables in local collars; the intrinsic dynamics supplies both the decoherence and the (internal) monitoring needed for objectivity. G. Ensemble independence and no superluminal signaling The nonlinear dependence on ρt is local (through Kρt,O ) and the instrument acts within A ( O ); therefore the arguments of Theorem VI.3 apply unchanged. Proposition VIII.3 ensures preparation noncontextuality: different convex decompositions of the same ρtlead to the same next state under the slice map.
34 H. Summary of Section IX •We constructed a local pointer instrument for coarse energy density and proved a finite-resolution Born rule (Proposition IX.1). • We introduced a stochastic (trajectory) version of the intrinsic dynamics (SME, Eq. (49) ) consistent with the unconditional master equation (Proposition IX.2). •We proved almost-sure collapse to pointer eigensectors with Born weights (Theorem IX.3). • We established exponential suppression of off-diagonal histories (Proposition IX.4) and compatibility with WAY constraints (Theorem IX.5). • We showed emergent objectivity via spectrum-broadcast structure in the pointer sector (Proposition IX.6). X. RIGOROUS FOUNDATIONS: FORMS, SEMIGROUPS, AND LOCALLY COVARIANT WELL-POSEDNESS A. Scope This section supplies a functional-analytic backbone for the intrinsic dynamics developed in Sections IV– IX. We place all constructions on the local von Neumann algebras A ( O )in standard form, build the linear completely positive (CP) part via (quantum) Dirichlet forms and detailed-balance quantum Markov semigroups (QMS), and justify the nonlinear modular dephasing as a well-defined contraction on the state space. We then prove existence of mild solutions for the full (state-dependent) flow by a Trotter–Kato and minimizing-movements scheme, establish monotonicity properties (Lyapunov structure), and formulate a locally covariant version compatible with the Haag–Kastler functoriality. We conclude with regularity statements (domains/cores) and stability. Throughout, A ( O )denotes the local von Neumann algebra of a causally convex region O ; σO is a fixed faithful normal reference state (e.g. the vacuum or wedge-KMS state) with modular objects (∆ σ,O, Jσ,O ); hAiρ=ρ(A)is the expectation in a normal state ρ;adX(A) := [X, A]. B. Preliminaries: standard form, L2(σ), and KMS inner products Fix M := A ( O )and a faithful normal state σ . By Tomita–Takesaki theory there is a standard representation ( Hσ, πσ, Jσ,Pσ )such that Ω σ∈ Pσ implements σ ( A ) = h Ω σ, πσ ( A )Ω σi . We write A for πσ ( A )when no ambiguity arises. The noncommutative L2 -space L2 ( M, σ )is the completion of M with the KMS inner product hA, Bi2,σ := Z1 0 ds σσs(A†)σ1−s(B),(51) where σs ( · ) := ∆ s σ ( · )∆ −s σ . This inner product is faithful and induces the norm kAk2,σ = phA, Ai2,σ (see e.g. [89, 90, 92]). We will also use the Carré du champ associated with a symmetric generator L: Γσ(A, B) := 1 2L(A†B)−L(A†)B−A†L(B), so that hA, L(A)i2,σ =h1,Γσ(A, A)i2,σ. C. Linear CP part as a Dirichlet form with detailed balance Let T00 ( f )denote the smeared energy density with f∈C∞ c ( O )a real test function. For γ≥ 0locally integrable, define the (formal) derivation δf ( A ) := i[ T00 ( f ) , A ]with common invariant core Aloc (finite polynomials of smeared fields). Consider the quadratic form on L2(M, σ) Eγ(A) := ZO dDx γ(x)kδχx(A)k2 2,σ,(52)
35 where χx is a fixed smooth bump centered at x (a partition of unity makes (52) independent of the choice up to form-equivalence). We write D(Eγ)for its form domain (the closure of Aloc). Proposition X.1 (Closability and Markov property) . Assume f7→ T00 ( f )is essentially selfadjoint on Aloc , and σ is a KMS state for a dynamics that leaves T00 ( f )invariant (e.g. wedge boosts or time translations with appropriate smearing). Then Eγ is a densely defined, closable, Markovian quadratic form on L2 ( M, σ ). Its closure generates a σ -symmetric, conservative quantum Markov semigroup (T t ) t≥0 on Mwith d dt Tt(A)t=0 =LCP(A) := −ZO dDx γ(x) [T00(x),[T00(x), A]],(53) realized in the sense of forms, and satisfying (KMS) detailed balance with respect to σ. Sketch. δχx is a densely defined symmetric derivation valued in L2 ( M, σ ); the integrability of γ and local finiteness of the cover yield a well-defined quadratic form on Aloc , which is closable by the general theory of completely Dirichlet forms on von Neumann algebras (Cipriani–Sauvageot [ 90 ]; see also Goldstein–Lindsay [ 91 ]). Markovianity follows from the Leibniz rule for derivations and the Beurling–Deny criterion in the noncommutative setting. The generator associated to the closed form coincides with (53) on Aloc and extends by form methods to all of D ( Eγ ). KMS symmetry with respect to σ holds because δχx is anti-symmetric under the KMS inner product (51) and γ is a scalar weight, see [ 92 ]. Conservativity (Tt(1) = 1) is automatic for derivation-generated forms. Physics. Form (52) implements the intuitive "weak continuous monitoring of T00 ” as a quadratic penalty on commutators with T00 , in the intrinsic L2 ( σ )geometry. Eq. (53) is the rigorous incarnation of the double-commutator dissipator used in Sections IV–VII. D. QSDE dilation and conservativity The QMS (T t ) t≥0 admits a Hudson–Parthasarathy (HP) dilation on a Bosonic Fock space F with noise fields coupled to T00(x)[93, 94]. Formally, one solves a QSDE dUt=ZdDxpγ(x)T00(x)dA† x−T00(x)dAx−1 2ZdDx γ(x)T00(x)2dtUt, and obtains T t ( A ) = h Ω , U† t ( A⊗1 ) Ut Ω iF . Under the assumptions of Proposition X.1, one can replace T00(x)2by its quadratic form closure on the common core and appeal to conservativity theorems [94]. Physics. The HP dilation fuses smoothly with our “holonomy central limit” picture (Sec. VI): the modular curvature noise becomes a Fock bath coupled to T00 , ensuring complete positivity and providing a canonical dilation. E. Nonlinear modular dephasing: well-defined slices and contractivity Recall the nonlinear map at time step τ > 0, Φρ,τ (A) := exp−λτ ad2 Kρ,O(A) = 1 √4πλτ ZR ds e−s2/(4λτ)eisKρ,OA e−isKρ,O.(54) For fixed ρ ,Φ ρ,τ is a normal, completely positive, unital (CPU) map on M . Define the induced affine map on normal states, Sτ(ρ) := ρ◦Φρ,τ . Proposition X.2 (Well-definedness and monotonicity of Sτ ) . For each τ > 0and normal ρ , Sτ ( ρ )is a normal state. If σis faithful, then SSτ(ρ)kσ≤S(ρkσ), and equality holds iff [Kρ,O,·]acts trivially on the support of ρ(i.e. ρis modular-diagonal).
36 Proof. CPU of Φ ρ,τ implies normality and positivity of Sτ ( ρ ); unitality gives normalization. Data processing for Araki relative entropy under CPU maps [ 95 ] yields S ( ρ◦ Φ ρ,τ kσ◦ Φ ρ,τ ) ≤S ( ρkσ ). Since Φ ρ,τ is a convex mixture of σ -modular automorphisms iff Kρ commutes with ∆ σ (not generally true), we compare S ( ρ◦ Φ ρ,τ kσ )directly by the integral representation of relative entropy (Kosaki’s variational formula [ 96 ]) and convexity of x7→ xlog x along the CPU channel to obtain the stated monotonicity. Equality holds only if all unitary conjugations by eisKρleave ρinvariant in the support sense. Physics. A single nonlinear slice never increases distinguishability from a faithful reference: the modular dephasing is a contraction for relative entropy. This is the H –theorem at the discrete level for the nonlinear part. F. Existence of mild solutions: Lie–Trotter and minimizing movements Define the Lie–Trotter time-step map on states by Tτ(ρ) := ρ◦eτLCP ◦Φρ,τ .(55) Iterate from ρ0:ρ(n) k+1 := Tt/n(ρ(n) k)for k= 0, . . . , n −1and set ρ(n)(t) := ρ(n) n. Theorem X.3 (Global existence of mild solutions) . Let σ be faithful, γ∈L1 loc ( O ), γ≥ 0, and λ≥ 0. Then for every initial normal state ρ0 there exists a family ( ρt ) t≥0 of normal states such that, along any sequence n→ ∞, ρ(n)(t)σ-weak −→ ρtfor all t≥0, and ρtis a mild solution of the intrinsic dynamics: for all A∈Aloc, d dt ρt(A) = ρt i[HO, A]−ZdDx γ(x)ρt [T00(x),[T00(x), A]]−λ ρt [Kρt,[Kρt, A]].(56) Moreover t7→ S(ρtkσ)is nonincreasing and ρtdepends continuously on ρ0in the weak∗topology. Sketch. Compactness: the normal state space is weak ∗ compact and metrizable on bounded sets. Monotonicity: by Proposition X.2 and detailed balance of eτLCP we have S ( ρ(n) k+1kσ ) ≤S ( ρ(n) kkσ ). Thus {ρ(n) ( t ) }n lies in a compact sublevel set of S ( ·kσ ), hence admits convergent subsequences for each t . Consistency across t is obtained by a diagonal argument using the contractivity to build a limiting curve with piecewise-constant approximants (Helly selection). The limit curve ρt satisfies (56) when paired with A∈Aloc via the standard Trotter–Kato and Chernoff product formula adapted to state-dependent perturbations (the nonlinear term contributes its quadratic variation as in (54) ). Lower semicontinuity of relative entropy under weak ∗ limits gives the Lyapunov monotonicity. Continuous dependence follows by nonexpansiveness of both steps in the dual bounded-Lipschitz metric. Remark (minimizing movements). An alternative (variational) construction uses a Jordan–Kinderlehrer–Otto (JKO) scheme on the state space: ρ(n) k+1 ∈argmin ρn1 2τD(ρkρ(n) k) + F(ρ)o, where D is a BKM-type divergence and F ( ρ ) = hHOiρ + λ S ( ρkσ ); the Euler–Lagrange equation yields a discrete gradient flow whose continuous limit is (56) . Existence follows by the direct method of calculus of variations once lower semicontinuity and compactness of sublevel sets are established (see [ 92 ] for the fixed-metric case). G. Regularity: domains, cores, and energy bounds Proposition X.4 (Cores and energy bounds) . Let Aloc be the ∗ -algebra generated by smeared fields in O . Then Aloc is a common invariant core for HO , T00 ( f )and for the generator of etLCP . If supt≤ThHOiρt<∞ for some T > 0, then ρt extends to a normal functional on the form domain of HO for all t≤T.
37 Sketch. Standard results for Wightman fields give essential selfadjointness on Aloc Ω σ . The generator of the closed form has Aloc as a core because δχx maps Aloc into itself, and the Beurling–Deny structure is local. Energy bounds follow from Grönwall-type estimates applied to d dt hHOiρt using that the CP part is dephasing (no net energy input). H. Local covariance and functoriality Let Loc be the category of globally hyperbolic spacetimes with isometric embeddings that preserve orientation and time-orientation, and let vN be the category of von Neumann algebras with normal unital ∗ -monomorphisms. A locally covariant QFT is a functor A : Loc →vN [ 97 ]. Let CP be the category whose objects are pairs ( M, σ )(a von Neumann algebra with a faithful state) and whose morphisms are normal CPU maps intertwining the reference states. Theorem X.5 (Locally covariant intrinsic dynamics) . Suppose: (i) g7→ γg and g7→ λg are natural scalar densities on Loc (e.g. built from curvature scalars and the entangling surface geometry), (ii) the stress-energy density T00 transforms covariantly, and (iii) reference states σO are chosen covariantly (vacuum/KMS). Then the assignment (O ⊂ M)7−→ T(M) t:A(O)→ A(O)t≥0, with T (M) t the intrinsic (CP+modular) evolution constructed above, is a natural transformation with respect to embeddings in Loc: for any embedding ψ:M→Nand region O ⊂ M, A(ψ)◦T(M) t=T(N) t◦A(ψ),∀t≥0. Sketch. For the CP part, covariance follows from the tensorial nature of T00 and the functoriality of A : T(N) 00 ( f◦ψ−1 ) = A ( ψ ) T(M) 00 ( f ) A ( ψ ) −1 . Hence LCP intertwines with A ( ψ ); exponentiating preserves the relation. For the modular slice, Kρ,O is functorial with respect to A ( ψ )when ρ and σ transform covariantly; consequently eisK and the Gaussian average commute with A(ψ), proving naturality. Physics. The intrinsic dynamics is locally covariant: it preserves the causal and geometric structure across spacetimes and regions. This validates its use in curved backgrounds and near horizons (Sec. VIII). I. Stability, contractivity, and convergence to steady states Theorem X.6 (Contractivity and convergence) . Let ρt and ωt be two mild solutions with the same coefficients γ, λ. Then for all t≥0, Sρtkσ≤S(ρ0kσ), Sρtkωt≤S(ρ0kω0), and if the linear CP semigroup is primitive (unique faithful fixed point σ ), then ρt→σ in the weak ∗ topology as t→ ∞. Sketch. First inequality is the Lyapunov monotonicity already proved. For the two-trajectory estimate, use joint convexity of relative entropy and the fact that each step of the Trotterized map is a CPU map applied to both trajectories (data processing). Primitivity implies spectral gap/hypercontractivity for the linear part; combined with the H –theorem, this forces convergence to σ (see [ 92 ] for entropy production and quantum log-Sobolev tools). Physics. The flow is thermodynamically stable: distinguishability from the steady state decays, and for primitive linear dissipation the nonlinear modular dephasing cannot produce limit cycles or new fixed points away from σ. J. Summary of Section X We constructed the linear (CP) part via closable quantum Dirichlet forms satisfying detailed balance, provided an HP dilation ensuring conservativity, and established that the nonlinear modular slices are
38 CPU and contractive for relative entropy. A Lie–Trotter/minimizing-movements scheme yields global mild solutions preserving locality and covariance. Stability and convergence follow from data processing and entropy production. These results underwrite the physical conclusions in Sections V–IX within a rigorous operator-algebraic framework. XI. OPERATIONAL TEST OF INTRINSIC SELF–MEASUREMENT NEAR AN ENTANGLING SURFACE A. Aim and principle The central claim of this work is that a tiny, locally covariant “self–measurement” dynamics—arising from higher–categorical coherence defects and realized as a completely positive (CP) dephasing of modular high–frequency modes in a thin collar around an entangling surface—produces a dynamical crossover length `(t) = pα(t), α(t) := Zt 0 ds Γ(s),(57) and with it three robust consequences: (i) softening of short–distance correlations transverse to the cut; (ii) an area–proportional plateau of mutual information I ( A : B )when two regions are brought within separation s.` ( t ); and (iii) finiteness of second–order shape responses of the entropy. Here Γ is the effective monitoring rate in the collar (defined precisely below). This section gives an explicit, experimentally realizable protocol (superconducting circuits, with a cold–atom alternative) that detects these signatures and interprets them within the AQFT/modular picture developed in the paper. Operational content. We implement the linear CP part of the intrinsic dynamics by continuous weak measurement of a local density/energy proxy supported on a thin collar around the bipartition. Physically, this is a standard QND (quantum non–demolition) monitoring that, at the ensemble level (ignoring the record), generates a local dephasing Lindbladian. Mathematically, in the Heisenberg picture, d dtAt=i[H, At]−X x∈collar Γx[Lx,[Lx, At]],(58) with Lx = L† x a local observable (energy/density proxy) and Γ x≥ 0the calibrated monitoring rate.[ 106 ] Equation (58) is the discrete analogue of the AQFT generator −Rγ(y) ad2 T00(y)confined to a collar. B. Minimal lattice model and mapping to the field–theoretic picture Consider a one–dimensional spin chain (length N) realizing a free–fermion (XX) Hamiltonian H= N−1 X j=1 J(XjXj+1 +YjYj+1),(59) with a bipartition at the bond (j0, j0+1). The collar Cis the set of sites C={j0−d, . . . , j0}∪{j0+1, . . . , j0+d}, with small half–width d = 2 or 3. We continuously and weakly monitor a local Hermitian proxy Lj on each j∈ C, chosen as either the site density Zj (QND in cQED) or a short–range proxy for the bond energy. The unconditional ensemble dynamics generated by (58) is CP, unital, and local. Continuum interpretation. Within the wedge approximation, the entangling cut is locally flat; modular modes are labelled by a “modular frequency” ω . Monitoring Lj in the collar damps high– ω components of modular correlators across the cut by a Gaussian factor exp [ − 2 α ( t ) ω2 ], with α ( t )as in (57) ; this is the lattice avatar of the BW/CFT result discussed in the main text. The emergent length ` ( t )sets the thickness of the operational pointer region near the cut.
39 C. Concrete superconducting–circuit implementation Device and preparation A linear array of N = 24 transmons with nearest–neighbor tunable couplers implements (59) with J/ 2 π≃ 1 . 5MHz. We prepare a near–Gaussian ground–like state either by a matchgate circuit (exact free–fermion ground state at half filling) or by a short adiabatic ramp; both are standard and validated by two–point correlators. Collar–localized continuous monitoring Each qubit j couples dispersively to its readout resonator ( χj shift, linewidth κj ). Driving the resonator of a collar qubit with a weak coherent tone of average photon number nj 1induces unconditional dephasing at rate Γj≈8χ2 j κj nj,(60) in the bad–cavity, on–resonance limit (standard cQED).[107] We choose a Gaussian collar profile Γj0±d= Γ0exp−d2 2σ2 c, d = 0,1,2, σc≃1,(61) and apply the tones for a duration t while the XX Hamiltonian remains on. The effective rate entering (57) is Γ = X j∈C wjΓj≡cgeo Γ0,(62) with geometric weights wj determined by the overlap of Lj with modular energy density and cgeo = O (1). In practice we calibrate cgeo by parity–decay (two–qubit Ramsey) across the central bond. Observables and estimators We probe the mutual information between two blocks A and B of equal size (e.g. 6 qubits each) separated by ssites, I(A:B) = S(ρA) + S(ρB)−S(ρAB),(63) with either the von Neumann entropy (via classical shadows) or the Rényi–2entropy S2 ( ρ ) = −log Tr ρ2 (via randomized measurements). For classical shadows we use random local Cliffords; the standard single–qubit Pauli estimator bρ(1)(P, m) = 3 ΠP,m −I2, P ∈ {X, Y, Z}, m ∈ {±1},(64) tensors over the region and averages over shots. A small eigenvalue clipping produces a positive estimate bρRper region R∈ {A, B, AB}, from which S(bρR)is computed. Predictions: plateau and scaling Let sbe the number of buffer sites between Aand B(with s= 0 adjacent). The theory predicts: I(A:B;t)≈(Iplateau(t), s `(t), decaying tail (model–dependent), s `(t),`(t) = pα(t)≈pcgeoΓ0t. (65) The height Iplateau ( t )is finite and governed by the collar scale ` ( t )(for 1D chains, a log–type lattice analogue of the continuum area law), while the extent of the flat region of I vs. s grows like √t and √Γ0 . Plotting Ias a function of s/`(t)collapses datasets at different (t, Γ0)onto a single crossover curve.
40 Why a plateau? (Intuition) Monitoring the collar implements a CP conditional expectation onto apointer subalgebra generated by coarse energy/density modes near the cut. The conditional mutual information I ( A : B|collar )becomes small (exponentially in the ratio collar thickness / correlation length), forcing I ( A : B )to saturate when A and B approach within the screened distance ` ( t ); this is the operational manifestation of “self–measurement at the seam”. Calibration and extraction of `(t) 1. Single–qubit Ramsey: with tone on only qubit j , fit exponential decay to obtain Γ j and thus the map nj7→ Γjvia (60). 2. Parity–decay across the cut: prepare a Bell state across ( j0, j0 +1), turn on the central tone(s), measure the parity–oscillation decay rate Γ; this fixes cgeo in (62). 3. Plateau–onset fit: at each (t, Γ0), fit I(s)to a two–regime form, e.g. I(s)≈I∞−∆Iexph−s scβi,(66) and define `exp(t) := sc. Test `exp ∝√tand ∝√Γ0; compare with qcgeo Γt. Controls and falsification Three discriminating controls isolate the genuine boundary–localized effect: 1. No monitoring (Γ0= 0): I(s)increases monotonically as s↓0(no plateau). 2. Monitoring far from the cut: same total power applied away from the boundary does not produce a local plateau of Ivs. sat the cut. 3. Uniform monitoring: weak dephasing everywhere suppresses entanglement globally, but does not generate a local plateau versus stied to a length `(t). Failure of the plateau to appear at accessible ( t, Γ 0 ), or failure of `exp to scale as √t and √Γ0 , falsifies the prediction. D. Cold–atom alternative (QND Faraday collar) In a 1D optical lattice (fermions or hard–core bosons), a blue–detuned, far–off–resonant light sheet (waist w≃ 2sites) centered on the cut implements a QND density measurement with site–dependent rates Γ j calibrated by local Ramsey of a spin–wave. Evolution times t = 50–500 ms with Γ 0∼ 10–100 s −1 yield ` ( t ) ∼ 1–4sites. Rényi–2 mutual information I2 ( A : B )is obtained from randomized measurements with a quantum–gas microscope. The same plateau and scaling tests apply. E. Mathematical underpinning of the signatures We outline how (58) yields the plateau and scaling claimed in (65). (1) Dirichlet–form suppression of cross–boundary coherences. For any observable O with support straddling the cut, the dissipative part of (58) contributes −X j∈C Γjh[Lj,[Lj, O]]iρ=−2X j∈C Γj(LjO−OLj)†(LjO−OLj)ρ≤0.(67) Thus any matrix element of ρ that fails to commute with the collar pointer algebra is exponentially damped in time with rate at least Pj Γ j weighted by its overlap. In modular frequency space this appears as a Gaussian factor exp[−2α(t)ω2]for wedge/ball geometries, with α(t)as in (57).
41 (2) Mutual information bound via conditional expectations. Let NC be the von Neumann algebra generated by the pointer variables in the collar and let E : B ( H ) → NC be the ρt –preserving conditional expectation induced by the CP semigroup at time t (Takesaki). Then data processing for relative entropy gives, for the tripartition A|C|B, I(A:B)ρt≤I(A:B|C)ρt+ 2 S(ρt◦E)−2S(ρt),(68) while the cross terms entering I ( A : B| C)are controlled by (67) . For a collar of geometric thickness exceeding the correlation length set by `(t)one obtains I(A:B)ρt≤CArea ×`(t)−(D−2) (D≥2),(69) which in the 1D chain (effective D = 2) yields a finite constant controlled by the collar scale; this is the plateau height. Equation (69) is the operational version of the continuum estimates proved earlier in the paper for wedges and CFT balls. (3) √t scaling of the crossover. From (57) and (62) we have ` ( t ) = pcgeoΓ0t . The onset of the plateau when s.` ( t )follows by comparing the buffer separation to the screening length in the transverse (modular) momentum integral; equivalently, a Laplace–method estimate of the wedge spectral integrals shows that the dominant contribution to I ( A : B )transitions at s∼` ( t )from a power–law tail to the finite collar–controlled constant. F. Statistical pipeline and uncertainties For classical shadows on 6–qubit regions, M≃ 2 × 10 4 random settings per ( t, Γ 0, s )suffice to estimate S ( ρR )to . 0 . 05 bits precision; bootstrap resampling yields robust error bars on I . For Rényi–2 with randomized measurements, M≃ (3–8) × 10 2 random settings with 5 × 10 2 shots each per setting are typical in cold atoms. The analysis reports: (i) I ( s )curves with errors; (ii) extracted `exp ( t )with confidence intervals and fits to `∝√tand ∝√Γ0; and (iii) data collapse quality (reduced χ2). G. Systematics and mitigations Potential confounds include readout crosstalk (mitigate by frequency staggering and verifying non–collar dephasing < 5%), residual ZZ interactions (spin–echo during monitoring; show plateau tracks Γ 0 not ZZ drift), heating by readout drives (keep n. 0 . 2; monitor excited–state population), and tomography bias (cross–validation and PSD projection). Control datasets (no/far/uniform monitoring) separate genuine boundary self–measurement from global entanglement suppression. H. Interpretation in the AQFT framework The collar monitoring implements, in the sense of nets of algebras, a local CP map Φ t : A ( O ) → A ( O ) that is identity outside the collar and a conditional expectation onto the pointer subalgebra inside. The pointer algebra becomes approximately abelian as t grows (dephasing), producing redundant classical records localized near the cut. The plateau in I ( A : B )is not a kinematical artifact (no lattice cutoff or inserted type–I factor); it is a dynamical saturation reflecting that modular UV modes have been softened at the seam by intrinsic self–measurement. The factor type (III 1 ) and Reeh–Schlieder remain intact for the full algebra; only operational diagnostics confined to the collar become classical. I. Summary of decisive signatures The proposal yields a clear, falsifiable experimental pattern: 1. A finite, area–proportional plateau of I ( A : B )at small separations, with extent ∝` ( t )and height controlled by `(t). 2. Scaling: `(t)∝√tand `∝√Γ0for fixed geometry/pulse shaping.
48 Appendix H: Gaussian Lattice Free Models: Covariances and Entropy 1. Covariance matrices For a quadratic Hamiltonian on a lattice with canonical variables (qj, pj), the covariance matrix Γ = hqq>i h1 2(qp>+pq>)i h1 2(pq>+qp>)i hpp>i transforms under a Gaussian channel induced by transverse damping as Γ7−→ (G⊕G) Γ (G⊕G)>, G = diag(e−αk2 ⊥)in k⊥-space. Entanglement entropies then follow from the symplectic spectrum of the restricted Γvia the Peschel formula (44) [54]. 2. Area law with saturation Control of the symplectic eigenvalues using the Combes–Thomas estimate for banded positive matrices and the effective transverse correlation length ξ⊥∼`(t)gives Proposition VII.4 (see also [51, 55]). Appendix I: Cohomology of Nets and Modular Curvature Let N ( Loc )be the nerve of the poset of regions ordered by inclusion. A (local) automorphism-valued 2-cochain assigns to each 2-simplex ( O ⊂ O0⊂ O00 )a unitary wj,i ∈ U ( A ( O00 )) obeying the pentagon (3-cocycle) identity on each 3-simplex: wk,j◦iσO000 t(wj,i) = wk◦j,i wk,j. Gauge transformations by inner 1-cochains ui∈ U(A(O0)) act as wj,i 7→ uj◦iwj,i σO00 t(ui)†u† j. Equivalence classes [ w ] ∈H2 ( N ( Loc ) ,Autloc A )(with modular twist) label modular curvature sectors, providing a classification scheme for intrinsic self-measurement deformations alluded to in Section VIII. Appendix J: Notation, Identities, and a Minimal Preamble Common symbols. A ( O ): local von Neumann algebra; H : Hilbert space; T00 ( f ): smeared energy density; Kρ,O = −log ∆ ρ,O : modular Hamiltonian; σρ t : modular automorphisms; S ( ωkσ ): Araki relative entropy; Γσ: Carré du champ; Dα: parabolic cylinder function; `(t) = pα(t). Useful identities. ad2 X ( A ) = [ X, [ X, A ]]; for any selfadjoint X , −ad2 X is the generator of the CPU semigroup A7→ R g t ( s ) eisXAe−isX ds with Gaussian g t . Data processing: S (Φ( ω ) k Φ( σ )) ≤S ( ωkσ )for any normal CPU Φ. First law (relative entropy): d dλ 0S(ωλkσ) = δhKσi. Appendix K: Statements •The author declares no conflict of interest •No new data has been generated for this manuscript [1] R. Haag and D. Kastler, “An Algebraic Approach to Quantum Field Theory,” J. Math. Phys. 5 , 848–861 (1964).
49 [2] R. Haag, Local Quantum Physics, 2nd ed. (Springer, Berlin, 1996). [3] R. F. Streater and A. S. Wightman, PCT, Spin and Statistics, and All That, (Princeton Univ. Press, 2000). [4] H. Reeh and S. Schlieder, “Bemerkungen zur Unitäräquivalenz von lorentzinvarianten Feldern,” Nuovo Cimento 22, 1051–1068 (1961). [5] M. Takesaki, Theory of Operator Algebras I (Springer, 1979); see also M. Takesaki, Tomita’s Theory of Modular Hilbert Algebras and its Applications (Springer, 1970). [6] J. J. Bisognano and E. H. Wichmann, “On the Duality Condition for a Hermitian Scalar Field,” J. Math. Phys. 16 , 985–1007 (1975); “On the Duality Condition for Quantum Fields,” J. Math. Phys. 17 , 303–321 (1976). [7] W. Driessler, “On the Type of Local Algebras of Quantum Field Theory,” Commun. Math. Phys. 53 , 295–297 (1977). [8] A. Connes, “Une classification des facteurs de type III,” Ann. Sci. Éc. Norm. Sup. 6, 133–252 (1973). [9] R. Longo, “Algebraic and Modular Structure of Von Neumann Algebras of Quantum Fields,” Commun. Math. Phys. 87, 531–570 (1983). [10] S. Doplicher and R. Longo, “Standard and Split Inclusions of von Neumann Algebras,” Invent. Math. 75 , 493–536 (1984). [11] D. Buchholz, C. D’Antoni, and R. Longo, “Nuclear Maps and Modular Nuclearity,” Commun. Math. Phys. 111, 123–135 (1987). [12] L. Bombelli, R. K. Koul, J. Lee, and R. D. Sorkin, “A Quantum Source of Entropy for Black Holes,” Phys. Rev. D 34, 373–383 (1986). [13] M. Srednicki, “Entropy and Area,” Phys. Rev. Lett. 71, 666–669 (1993). [14] C. G. Callan, Jr. and F. Wilczek, “On Geometric Entropy,” Phys. Lett. B 333, 55–61 (1994). [15] C. Holzhey, F. Larsen, and F. Wilczek, “Geometric and Renormalized Entropy in Conformal Field Theory,” Nucl. Phys. B 424, 443–467 (1994). [16] H. Casini and M. Huerta, “Entanglement Entropy in Free Quantum Field Theory,” J. Phys. A 42 , 504007 (2009). [17] S. N. Solodukhin, “Entanglement Entropy of Black Holes,” Living Rev. Relativity 14, 8 (2011). [18] H. Casini, M. Huerta, and R. C. Myers, “Towards a Derivation of Holographic Entanglement Entropy,” JHEP 1105, 036 (2011). [19] T. Faulkner, A. Lewkowycz, and J. Maldacena, “Quantum Corrections to Holographic Entanglement Entropy,” JHEP 1311, 074 (2013). [20] H. Araki, “Relative Entropy of States of a von Neumann Algebra,” Publ. Res. Inst. Math. Sci. (Kyoto) 11 , 809–833 (1976). [21] V. Gorini, A. Kossakowski, and E. C. G. Sudarshan, “Completely Positive Dynamical Semigroups of N-Level Systems,” J. Math. Phys. 17, 821–825 (1976). [22] G. Lindblad, “On the Generators of Quantum Dynamical Semigroups,” Commun. Math. Phys. 48 , 119–130 (1976). [23] W. H. Zurek, “Decoherence, Einselection, and the Quantum Origins of the Classical,” Rev. Mod. Phys. 75 , 715–775 (2003). [24] H. Everett, ““Relative State” Formulation of Quantum Mechanics,” Rev. Mod. Phys. 29, 454–462 (1957). [25] G. C. Ghirardi, A. Rimini, and T. Weber, “Unified Dynamics for Microscopic and Macroscopic Systems,” Phys. Rev. D 34, 470–491 (1986). [26] A. Bassi, K. Lochan, S. Satin, T. P. Singh, and H. Ulbricht, “Models of Wave-Function Collapse, Underlying Theories, and Experimental Tests,” Rev. Mod. Phys. 85, 471–527 (2013). [27] N. Gisin, “Weinberg’s Nonlinear Quantum Mechanics and Superluminal Communications,” Phys. Lett. A 143, 1–2 (1990). [28] J. Polchinski, “Weinberg’s Nonlinear Quantum Mechanics and the Einstein–Podolsky–Rosen Paradox,” Phys. Rev. Lett. 66, 397–400 (1991). [29] M. M. Wolf, F. Verstraete, M. B. Hastings, and J. I. Cirac, “Area Laws in Quantum Systems: Mutual Information and Correlations,” Phys. Rev. Lett. 100, 070502 (2008). [30] J. Cardy, “Some results on the mutual information of disjoint regions in higher dimensions,” J. Phys. A 46 , 285402 (2013). [31] T. Faulkner, R. G. Leigh, O. Parrikar, and H. Wang, “Modular Hamiltonians for Deformed Half-Spaces and the Averaged Null Energy Condition,” JHEP 09 (2016) 038. [32] H. Casini, E. Teste, and G. Torroba, “Modular Hamiltonians on the null plane and the Markov property of the vacuum state,” J. Phys. A 50, 364001 (2017). [33] A. Allais and M. Mezei, “Some results on the shape dependence of entanglement and Rényi entropies,” Phys. Rev. D 91, 046002 (2015). [34] M. Mezei, “Entanglement entropy across a deformed sphere,” Phys. Rev. D 91, 045038 (2015). [35] P. Bueno and R. C. Myers, “Universal corner entanglement from twist operators,” JHEP 12 (2015) 168. [36] R. Bousso, Z. Fisher, J. Koeller, S. Leichenauer, and A. C. Wall, “Proof of the Quantum Null Energy Condition,” Phys. Rev. D 93, 064044 (2016). [37] P. Calabrese and J. Cardy, “Entanglement entropy and quantum field theory,” J. Stat. Mech. (2004) P06002. [38] P. Calabrese, J. Cardy, and E. Tonni, “Entanglement entropy of two disjoint intervals in conformal field theory,” J. Stat. Mech. (2009) P11001. [39] S. Ryu and T. Takayanagi, “Holographic Derivation of Entanglement Entropy from AdS/CFT,” Phys. Rev.
50 Lett. 96, 181602 (2006). [40] E. B. Davies, Quantum Theory of Open Systems (Academic Press, London, 1976). [41] L. Accardi, A. Frigerio, and J. T. Lewis, “Quantum stochastic processes,” Publ. Res. Inst. Math. Sci. 18 , 97–133 (1982). [42] R. L. Hudson and K. R. Parthasarathy, “Quantum Itô’s formula and stochastic evolutions,” Commun. Math. Phys. 93, 301–323 (1984). [43] S. Attal and Y. Pautrat, “From repeated to continuous quantum interactions,” Ann. Henri Poincaré 7 , 59–104 (2006). [44] T. Kato, Perturbation Theory for Linear Operators (Springer, 1995). [45] D. Petz, “Sufficient subalgebras and the relative entropy of states of a von Neumann algebra,” Commun. Math. Phys. 105, 123–131 (1986). [46] H. Spohn, “Entropy production for quantum dynamical semigroups,” J. Math. Phys. 19 , 1227–1230 (1978). [47] A. Uhlmann, “Relative entropy and the Wigner–Yanase–Dyson–Lieb concavity in an interpolation theory,” Commun. Math. Phys. 54, 21–32 (1977). [48] R. Alicki and K. Lendi, Quantum Dynamical Semigroups and Applications, 2nd ed. (Springer, 2007). [49] A. Frigerio, “Stationary states of quantum dynamical semigroups,” Commun. Math. Phys. 63 , 269–276 (1978). [50] O. Bratteli and D. W. Robinson, Operator Algebras and Quantum Statistical Mechanics II (Springer, 1997). [51] J. Eisert, M. Cramer, and M. B. Plenio, “Area laws for the entanglement entropy,” Rev. Mod. Phys. 82 , 277–306 (2010). [52] K. M. R. Audenaert, “A sharp continuity estimate for the von Neumann entropy,” J. Phys. A 40 , 8127–8136 (2007). [53] I. Peschel, “Calculation of reduced density matrices from correlation functions,” J. Phys. A 36 , L205–L208 (2003). [54] I. Peschel and V. Eisler, “Reduced density matrices and entanglement entropy in free lattice models,” J. Phys. A 42, 504003 (2009). [55] M. B. Hastings, “An area law for one-dimensional quantum systems,” J. Stat. Mech. (2007) P08024. [56] N. Engelhardt and A. C. Wall, “Quantum Extremal Surfaces: Holographic Entanglement Entropy beyond the Classical Regime,” JHEP 01 (2015) 073. [57] T. Faulkner, A. Lewkowycz, and J. Maldacena, “Quantum corrections to holographic entanglement entropy,” JHEP 11 (2013) 074. [58] D. L. Jafferis, A. Lewkowycz, J. Maldacena, and S. J. Suh, “Relative entropy equals bulk relative entropy,” JHEP 06 (2016) 004. [59] H. M. Wiseman and G. J. Milburn, Quantum Measurement and Control (Cambridge Univ. Press, 2010). [60] K. Jacobs, Quantum Measurement Theory and its Applications (Cambridge Univ. Press, 2014). [61] L. Susskind and J. Uglum, “Black hole entropy in canonical quantum gravity and superstring theory,” Phys. Rev. D 50, 2700–2711 (1994). [62] T. Jacobson, “Black-hole entropy and induced gravity,” Phys. Rev. D 50, 6031–6038 (1994). [63] G. Penington, “Entanglement Wedge Reconstruction and the Information Paradox,” JHEP 09 (2020) 002. [64] A. Almheiri, N. Engelhardt, D. Marolf, and J. Maxfield, “The entropy of bulk quantum fields and the Page curve of Hawking radiation,” JHEP 12 (2019) 063. [65] H. Casini, M. Huerta, and J. A. Rosabal, “Remarks on entanglement entropy for gauge fields,” Phys. Rev. D89, 085012 (2014). [66] W. Donnelly and A. C. Wall, “Entanglement entropy of electromagnetic edge modes,” Phys. Rev. Lett. 114 , 111603 (2015). [67] G. Ruzzi, “Homotopy of posets, net-cohomology and superselection sectors in globally hyperbolic spacetimes,” Commun. Math. Phys. 246, 449–486 (2004/2005). [68] J. E. Roberts, “Lectures on Algebraic Quantum Field Theory,” in The Algebraic Theory of Superselection Sectors, D. Kastler (ed.), (World Scientific, 1990). [69] J. Maldacena, S. H. Shenker, and D. Stanford, “A bound on chaos,” JHEP 08 (2016) 106. [70] Y. Li, X. Chen, and M. P. A. Fisher, “Quantum Zeno effect and the many-body entanglement transition,” Phys. Rev. B 98, 205136 (2018). [71] B. Skinner, J. Ruhman, and A. Nahum, “Measurement-Induced Phase Transitions in the Dynamics of Entanglement,” Phys. Rev. X 9, 031009 (2019). [72] A. Chan, R. M. Nandkishore, M. P. A. Fisher, and A. Nahum, “Unitary-projective entanglement dynamics,” Phys. Rev. B 99 , 224307 (2019); see also A. Chan, Y. Li, Y. Huang, and M. P. A. Fisher, Phys. Rev. B 101 , 174306 (2020). [73] Y. Bao, S. Choi, and E. Altman, “Theory of the phase transition in random unitary circuits with measurements,” Phys. Rev. B 101, 104301 (2020). [74] R. Brunetti, K. Fredenhagen, and R. Verch, “The Generally Covariant Locality Principle—A New Paradigm for Local Quantum Physics,” Commun. Math. Phys. 237, 31–68 (2003). [75] E. B. Davies and J. T. Lewis, “An operational approach to quantum probability,” Commun. Math. Phys. 17, 239–260 (1970). [76] A. S. Holevo, Probabilistic and Statistical Aspects of Quantum Theory, 2nd ed. (Edizioni della Normale, Pisa, 2011). [77] M. A. Naimark, “Spectral functions of a symmetric operator,” Izv. Akad. Nauk SSSR Ser. Mat. 4 , 277–318
51 (1940). [78] W. F. Stinespring, “Positive functions on C∗-algebras,” Proc. Amer. Math. Soc. 6, 211–216 (1955). [79] K. Kraus, “General state changes in quantum theory,” Ann. Phys. (NY) 64, 311–335 (1971). [80] V. P. Belavkin, “Quantum continual measurements and a posteriori collapse on CCR,” Commun. Math. Phys. 146, 611–635 (1992). [81] A. Barchielli and M. Gregoratti, Quantum Trajectories and Measurements in Continuous Time (Springer LNP 782, 2009). [82] L. Bouten, R. van Handel, and M. R. James, “An introduction to quantum filtering,” SIAM J. Control Optim. 46, 2199–2241 (2007). [83] M. Gell-Mann and J. B. Hartle, “Classical equations for quantum systems,” Phys. Rev. D 47 , 3345–3382 (1993). [84] R. Omnès, “Consistent interpretations of quantum mechanics,” Rev. Mod. Phys. 64, 339–382 (1992). [85] E. P. Wigner, “Die Messung quantenmechanischer Operatoren,” Z. Phys. 131, 101–108 (1952). [86] H. Araki and M. M. Yanase, “Measurement of quantum mechanical operators,” Phys. Rev. 120 , 622–626 (1960). [87] M. Ozawa, “Quantum measuring processes of continuous observables,” J. Math. Phys. 25, 79–87 (1984). [88] F. G. S. L. Brandão, M. Piani, and P. Horodecki, “Generic emergence of classical objectivity in quantum Darwinism,” Nature Communications 6, 7908 (2015). [89] D. Petz, “A survey of certain trace inequalities,” Functional Analysis and Operator Theory, Banach Center Publ. 30, 287–298 (1994). [90] F. Cipriani and J.-L. Sauvageot, “Derivations as square roots of Dirichlet forms,” J. Funct. Anal. 201 , 78–120 (2003). [91] S. Goldstein and J. M. Lindsay, “Markov semigroups KMS-symmetric for a weight,” Math. Ann. 313 , 39–67 (1999); see also “KMS-symmetric Markov semigroups,” Math. Z. 219, 591–608 (1995). [92] E. A. Carlen and J. Maas, “Gradient flow and entropy inequalities for quantum Markov semigroups with detailed balance,” Commun. Math. Phys. 331 , 887–926 (2014); and “Non-commutative calculus, optimal transport and functional inequalities in dissipative quantum systems,” J. Stat. Phys. 178, 319–378 (2020). [93] R. L. Hudson and K. R. Parthasarathy, “Quantum Itô’s formula and stochastic evolutions,” Commun. Math. Phys. 93, 301–323 (1984). [94] A. M. Chebotarev and F. Fagnola, “Sufficient conditions for conservativity of quantum dynamical semigroups,” J. Funct. Anal. 153, 382–404 (1998). [95] D. Petz, “Sufficient subalgebras and the relative entropy of states of a von Neumann algebra,” Commun. Math. Phys. 105, 123–131 (1986). [96] H. Kosaki, “Relative entropy of states: A variational expression,” J. Operator Theory 16, 335–348 (1986). [97] R. Brunetti, K. Fredenhagen, and R. Verch, “The generally covariant locality principle—a new paradigm for local quantum physics,” Commun. Math. Phys. 237, 31–68 (2003). [98] J. J. Bisognano and E. H. Wichmann, “On the duality condition for a Hermitian scalar field,” J. Math. Phys. 16, 985–1007 (1975). [99] J. J. Bisognano and E. H. Wichmann, “On the duality condition for quantum fields,” J. Math. Phys. 17 , 303–321 (1976). [100] M. Takesaki, Theory of Operator Algebras II (Springer, 2003). [101] F. W. J. Olver, A. B. Olde Daalhuis, D. W. Lozier, B. I. Schneider, R. F. Boisvert, C. W. Clark, B. R. Miller, and B. V. Saunders, eds., NIST Digital Library of Mathematical Functions, http://dlmf.nist.gov/ (Release 1.1.10, 2023). [102] R. Haag, Local Quantum Physics, 2nd ed. (Springer, 1996). [103] D. Buchholz, C. D’Antoni, and R. Longo, “Nuclear maps and modular structures. II. Applications to quantum field theory,” Commun. Math. Phys. 129, 115–138 (1990). [104] M. Reed and B. Simon, Methods of Modern Mathematical Physics II: Fourier Analysis, Self-Adjointness (Academic Press, 1975). [105] This is the algebraic (type–III compatible) analog of pointer bases in decoherence theory; see, e.g., [ 23 , 59 , 60 ]. [106] The nonlinear instantaneous–CP “modular slice” from the main text further sharpens the UV damping but is not required to observe the plateau; the linear CP part alone already produces the predicted crossover. [107] Derivation: with dispersive interaction χa†aZ , the qubit sees phase diffusion with spectral density Snn (0) ≈ 2 n/κ , whence Γ φ≃ (2 χ ) 2Snn (0) = 8 χ2n/κ . Off–resonant corrections and finite efficiency affect the record but not the unconditional dephasing rate to leading order.