Universality of Magic in Local Quantum Field Theory

Summary

A short, sharp hep-th paper (Valentin Benedetti, Atish Dabholkar, Marcello Dalmonte; ICTP Trieste + INFN + Bologna; 27 pp, 17 Jul 2026) proving a clean structural fact: no stabilizer state on a lattice can flow, in the continuum limit, to the vacuum of a local QFT — or to any state that resembles the vacuum at short distances. Since a flat entanglement spectrum is a necessary condition for a state to be a stabilizer state, and physical QFT states provably cannot have a flat spectrum, every physically relevant QFT state carries non-zero magic (nonstabilizerness). Simulating QFT states on a quantum computer therefore necessarily requires resources beyond stabilizer states and Clifford gates — the Gottesman–Knill “classically simulable” sector can never reach the continuum vacuum.

The engine of the argument is exactly the algebraic machinery this wiki is built around. The Reeh–Schlieder theorem makes the vacuum (and low-energy states near it) cyclic and separating for every local algebra; the Bisognano–Wichmann theorem fixes the vacuum modular operator on a wedge to ${\Delta }_{\Omega }=e^{-2\pi K}$ with $K$ the boost generator, whose spectrum is the continuous real line; and the type III₁ nature of local algebras (type-iii-von-neumann-algebras) forces the Connes invariant $S(\mathcal{A})=\{0\}\cup {\mathbb{R}}^{+}$ to be continuous. A flat entanglement spectrum corresponds to a modular operator proportional to the identity (trivial modular flow) — a type-I / discrete feature that a faithful normal state on a type III₁ algebra simply cannot have. Magic in QFT is, in one sentence, a shadow of type III₁.

This makes the paper an unusually direct outside contact with the group’s program: it takes the same objects — cyclic-separating states, the vacuum modular Hamiltonian, Rindler-wedge reduced structure, free-field Bogoliubov coefficients — that drive the Bell-CHSH and Araki–Uhlmann relative entropy lines, and reads a new quantum-information resource off them. Notably, Atish Dabholkar is also the contributor whose Strings 2026 question asks for a modular-invariant entanglement entropy in gravity — the same modular-theory frontier from the string-theory side.

Key Results

1. Flat entanglement spectrum ⟺ stabilizer (the QI input). For a pure stabilizer state on a qudit chain, the reduced density matrix on any spatial region is proportional to a projector, so all Rényi entropies coincide: $S_n(A)=\log {\lambda }_A$, independent of $n$ (§2.2). This flatness is preserved under Clifford operations and even holds for disconnected regions, hence extends to the Rényi mutual information $I_n(A_1,A_2)$ — a UV-finite, regulator-independent quantity whose flatness must survive the continuum limit. Non-flatness is a witness of magic (Tirrito et al. 2024); the $n=2$ anti-flatness $F_2({\rho }_A)$ lower-bounds the distance to the stabilizer set, $F_2({\rho }_A)/8\le {\min }_{\sigma \in \mathrm{STAB}}\Vert \psi -\sigma \Vert$ (Cao et al. 2024).

2. The QFT vacuum is magical (Bisognano–Wichmann route). A regularized-lattice computation of the wedge modular Hamiltonian $K_{\Omega ,A}={\int }_A{d}^{D-1}x{x}^{1}H(x)$ shows that demanding a flat spectrum forces every non-diagonal piece of the energy density to vanish, collapsing the Hamiltonian to a multiple of the identity. A Lorentz-invariant QFT with a non-trivial Hamiltonian therefore cannot have a flat vacuum spectrum (§3.1). A parallel replica-trick argument: flatness would require $Z[{\mathcal{M}}_{n,A}]=(Z[\mathcal{M}]{)}^n/{\lambda }_A^{n-1}$, impossible when local degrees of freedom feel the conical defect.

3. III₁ factors as the root of magic (the algebraic core, §3.2). Every cyclic-separating vector defines a faithful normal state; by Reeh–Schlieder the vacuum and states approximating it at short distances (k-particle states, conformal primaries and descendants) are all cyclic-separating; by the continuous Connes spectrum of type III₁ their modular operators have continuous spectrum; hence their entanglement spectra cannot be flat. Run backwards: any lattice stabilizer string that survives the continuum limit must converge to the identity, because the only local operator annihilating a cyclic-separating vector is zero.

4. TQFT is the exception. A topological QFT can be a stabilizer state (the toric code is the paradigm): its replica partition function scales as $Z[{\mathcal{M}}_n]=(Z[{\mathcal{M}}_1]{)}^n{D}^{n-1}$ with $D$ the total quantum dimension, and it is not a net of type III₁ factors (no local observables). Thus magic sharply separates local QFT from TQFT, purely by algebraic type — a distinction non-Gaussianity (free vs. interacting) does not see.

5. Explicit examples (§4).

• CFT vacuum: the Rényi mutual information between disjoint intervals has genuinely $n$-dependent coefficients (e.g. free scalar in $D=4$, eq. 4.12–4.13), so the CFT vacuum is “always magical”; the $n\to 0$ limit of $S_n$ for a sphere likewise carries non-trivial $n$-dependence in its universal (cutoff-independent) coefficient.
• Free-boson particle states in Rindler: using the two-mode-squeezed representation of the Minkowski vacuum and Bogoliubov/Unruh operators, the reduced density matrices of Unruh particle states $|N⟩$ follow a Pascal distribution; the anti-flatness $F_n$ peaks at intermediate $q_{\omega }=e^{-\omega /T_U}$ and, strikingly, the spectrum becomes flatter as $N\to \infty$ (though that limit is a singular, non-Fock state).

6. Holography (§5). Fixed-area states (RT / gravitational path integral) have flat spectra and are stabilizer-like tensor-network states — but the boundary CFT is a net of type III₁ algebras, so there are no faithful flat-spectrum states; bulk physical states, being cyclic-separating even under weak gravity, are magical, with non-flat subleading Rényi corrections from the modified cosmic-brane prescription. The Hartle–Hawking (not Boulware) vacuum is the physical near-horizon state and is magical; in string theory the string length acts as the lattice cutoff and the conclusion persists.

Methods

• Algebraic QFT: type III₁ factors, the Connes invariant $S(\mathcal{A})$, Tomita–Takesaki modular theory, Reeh–Schlieder, Bisognano–Wichmann, the scaling-algebra formalism for the short-distance approach to the BW modular operator.
• Entanglement techniques: replica trick / conical-defect partition functions, modular Hamiltonians for wedges and spheres (conformal mapping), Rényi entropies and the $n\to 0$ continuation, Rindler reduced density matrices via Bogoliubov coefficients.
• Quantum-information / resource theory: stabilizer formalism on qudit chains, the generalized Pauli/Clifford groups, entanglement-spectrum flatness and anti-flatness $F_n$ as magic witnesses, the stabilizer-distance bound.
• Condensed-matter bridge: qudit spin chains, string nets, MPS-magic literature (White–Cao–Swingle; Tarabunga–Tirrito–Dalmonte; Haug–Piroli) — the lattice side where these diagnostics were developed.

Relevance

This paper hands the group a new III₁-rooted resource to compute with the tools it already owns, and it strengthens the “type III₁ as the common thread” thesis directly.

• relative-entropy-qft — same modular machinery, new observable. The wedge modular Hamiltonian and the free-field Rindler reduced structure are exactly what the Araki–Uhlmann relative-entropy computations for coherent, squeezed, and cat states already use. The anti-flatness $F_2$ is a drop-in new quantity computable from the same two-point functions — see the follow-up magic-of-coherent-squeezed-cat-states.
• type-iii-algebras-across-areas — a fourth vertex. Magic joins Bell violation, relative entropy, and embezzlement as a resource whose presence is forced by type III₁. The natural conceptual question is whether these are quantitatively related — a magic ↔ relative-entropy inequality on III₁, tracked in relative-entropy-magic-inequality.
• bell-inequalities-qft — shared root. The cyclic-separating/BW structure that guarantees vacuum Bell-CHSH violation is the same structure that guarantees magic; both are readings of the continuous modular spectrum. Cat states, which the group has shown produce qualitatively new Bell features, are a natural place to look for enhanced magic.
• gribov-zwanziger / confinement-duality — a genuinely new diagnostic. The result assumes standard type III₁ + BW + Reeh–Schlieder. Whether the Gribov-horizon restriction preserves them (the RGZ vacuum’s modified propagators and the open status of Reeh–Schlieder there — cf. entanglement-as-confinement-probe) makes magic a candidate confinement probe: magic-and-the-gribov-horizon.
• gauge-gravity-duality — resource theory in holography. Fixed-area (stabilizer) vs. physical (magical) states connect to the group’s holographic resource theory and embezzlement work, where fixed-area states are the flat-spectrum reference; and it answers, concretely, part of the Goto et al. call for a QFT resource theory (resource-theory-monotones-in-qft, free-states-and-operations-for-resource-theory-in-qft).
• condensed-matter-connections. The lattice-magic diagnostics (anti-flatness, stabilizer Rényi entropy) come from the many-body community (Dalmonte is an author); the continuum-limit statement is a clean instance of the condensed-matter-qft-bridge.

Questions Raised

1. What is the magic (anti-flatness $F_2$) of the group’s coherent, squeezed, and cat states in the Rindler wedge, computed from the same Bogoliubov data as their relative-entropy results — and do cat states maximize magic where large-$N$ particle states flatten it? → magic-of-coherent-squeezed-cat-states
2. Is there a quantitative inequality relating magic, Araki–Uhlmann relative entropy, and the Bell-CHSH value for a state on a type III₁ algebra — making the “common thread” precise? → relative-entropy-magic-inequality, and Gap 3 / Project 1 of type-iii-algebras-across-areas.
3. Does the Gribov-horizon restriction change the magic of the confining vacuum — i.e. can magic serve as an order parameter for confinement / a diagnostic that the RGZ vacuum departs from the standard type III₁ + BW structure? → magic-and-the-gribov-horizon
4. How does the magic of the two-sided (TFD / Hartle–Hawking) state change when a traversable wormhole is opened by a GJW deformation (a Connes cocycle) — alongside the group’s Bell-CHSH and relative-entropy analyses of the same operation? → magic-across-traversable-wormhole
5. Does the onset of magic track the confinement / higher-form-symmetry transition between a local gauge theory and its deep-IR TQFT — i.e. is magic a confinement order parameter (with the anti-flatness witness promoted to an intrinsic measure via Clifford orbits, flagged by the authors as undeveloped in a QFT context)? → magic-and-confinement-phases, cf. entanglement-as-confinement-probe.
6. Is the holographic embezzlement cost / capacity related to magic — the departure from the flat-spectrum fixed-area (stabilizer-like) states that both frameworks privilege? → holographic-fixed-area-magic-embezzlement

Related Papers

• 2026-goto-rethinking-qi-gravity-fields — the QI-in-gravity agenda; this paper concretely answers part of its “resource theory in QFT” theme with a fully-worked resource (magic).
• 2025-liu-lectures-entanglement-vna — the type III₁ / modular / crossed-product background the argument rests on.
• strings-2026-open-questions — Dabholkar’s Strings 2026 question on modular-invariant entanglement entropy in gravity (same author, same modular frontier).
• Tirrito, Tarabunga, Lami et al. (2024), Quantifying nonstabilizerness through entanglement spectrum flatness, PRA 109 L040401 [2304.01175] — the flatness/anti-flatness witness lifted here to QFT.
• Cao, Cheng, Hamma, Leone, Munizzi, Oliviero (2025), Gravitational Backreaction is Magical, PRX Quantum 6 040375 [2403.07056] — the $F_2$ stabilizer-distance bound and the holographic magic story.
• White, Cao, Swingle (2021), Conformal field theories are magical, PRB 103 075145 [2007.01303] — magic in CFT on the lattice.
• Chemissany, Gesteau, Jahn, Murphy, Shaposhnik (2025), On infinite tensor networks, complementary recovery and type II factors [2504.00096] — the type II / tensor-network side of the holographic discussion (Gesteau again; cf. strings-2026-open-questions).

See Also

• magic-nonstabilizerness — the concept page.
• type-iii-algebras-across-areas — updated to carry magic as a fourth III₁-rooted resource.
• bisognano-wichmann-theorem, reeh-schlieder-theorem, type-iii-von-neumann-algebras — the three structural inputs.