OP-46 / carrier-problem attack, iteration 8: the net upgrade — net-modular data adds operators, not a localized indefinite form; the parametrization is the geometry

Status: ATTEMPT executed — the carrier problem (HYP-FACTORIZATION-IS-GEOMETRY), the re-compressed verdict-flipper, was attacked at the one input the iteration-7 single-algebra no-go did not cover: a net/family $O\mapsto\mathcal M(O)$ supplying relative modular operators $\Delta_{O'|O}$ , families $\{J_O\}$ , half-sided modular inclusions (HSMI), modular-intersection data (Wiesbrock), and the modular-Berry connection. The verdict does not flip. The net supplies strictly more operators but no new localized indefinite form; the indefinite/localized dichotomy of iteration 7 is closed under the net upgrade, and localization is shown to enter only through the net's index set — which is the geometry. Proposed hedge move: OP-46 residual MEDIUM-HIGH → toward HIGH, conditional on a single flagged net-constructibility ansatz. Last updated: 2026-06-10 Iteration: 8

Grading instrument: 2026-06-08-iter3-encode-vs-generate-criterion.md ( $\Delta_{\rm geo}$ three-valued flag, $\kappa_\Phi$ , smuggle items $g_1$ – $g_7$ ) extended here with three net-level items $n_1$ – $n_3$ (§1). Boundary inherited from 2026-06-10-iter7-indefinite-pairing-attempt.md (the $\eta_0$ result, vacuity theorem, trichotomy lemma + its constructibility ansatz — the no-go the single-algebra case left scoped). Prior closures not relitigated.

Scope honesty: §§1–3 are this session's own mathematics, derived stepwise with every assumption flagged, and verified numerically in the explicit chiral U(1)-current standard pair (fresh RNG). The closure theorem rests on one flagged net-constructibility ansatz (§1, the net analogue of iter-7's single-algebra ansatz). Citations web-verified this session unless marked unverified.

1. Formalizing the carrier and the net smuggle list

Definition (carrier). Over a net $\mathcal N=\{\mathcal M(O)\}_{O\in\mathcal I}$ with vacuum $\Omega$ , a carrier is a measurable field of fibers $\{V_x\}_{x\in X}$ with an $(n,n)$ Hermitian form $\eta_x$ per fiber such that:

(loc) localized: $q(\psi)=\langle\psi,\eta_x\psi\rangle$ depends only on $\psi$ near $x$ — equivalently the form has vanishing cross-coupling between disjoint regions;
(alg) algebra-compatible: $\mathrm{Ad}\,\eta_x$ preserves the local algebra;
(gen) generated: the field is built from net-modular data without smuggle items.

Smuggle list, extended to net level. $g_1$ – $g_7$ as before; new items: $n_1$ the index set $\mathcal I$ already carries a causal/region order (localization template); $n_2$ a positive-energy/spectrum condition installed via a chosen group representation; $n_3$ a chosen Euler element / wedge assignment (a geometry-germ in the Lie algebra).

Net-constructibility ansatz [INFERENCE — the flagged assumption]. A canonical, modular-covariant Hermitian form built from net-modular data has Gram operator a measurable function of the net's modular generators $\{\ln\Delta_O,\ \ln\Delta_{O'|O},\ P_{O',O}\}$ . (This is the net analogue of iteration 7's single-algebra ansatz " $\mathrm{Gram}\in\{\Delta\}''$ .")

Closure Theorem (net-data dichotomy). [INFERENCE — high modulo the ansatz] Under the ansatz, every such form is exactly one of:

(P) positive and local — a function of a positive translation generator $P\ge0$ (the new datum a HSMI supplies, by Borchers) or of the identity ( $\mathrm{Re}\langle\cdot,\cdot\rangle$ );
(S) signature-free — symplectic $\mathrm{Im}\langle\cdot,\cdot\rangle$ , complex-symmetric $\langle J\cdot,\cdot\rangle$ , real $J$ -split (iter-7 prong (ii), by enumeration);
(N) indefinite but non-local — a sign-of- $\Delta$ form $\varepsilon(\Delta_O)$ or $\varepsilon(\Delta_{O'|O})$ , boost-frequency-diagonal, with a slowly-decaying ( $\sim 1/\theta$ ) kernel coupling arbitrarily distant regions.

No canonical form is simultaneously indefinite and local. The engine:

Covariance–locality obstruction. [INFERENCE, high — analytic + numerically demonstrated] A local indefinite form on the chiral one-particle space is a multiplication operator $m(p)$ with a sign change at some $p_0>0$ . The modular flow rescales $p\mapsto e^{-2\pi t}p$ (Borchers), so $\mathrm{Ad}\,\Delta^{it}:m(p)\mapsto m(e^{-2\pi t}p)$ boosts the sign locus $p_0$ away; the only boost-invariant multiplication operators are constants (positive, no sign change). Hence local $\Rightarrow$ non-covariant $\Rightarrow$ non-canonical; covariant $\&$ indefinite $\Rightarrow$ function of $\Delta$ $\Rightarrow$ non-local. The single-algebra prongs (algebra-incompatibility, non-localizability) thus recur at net level and the net adds nothing that escapes them, because every operator a net supplies is built from boost/translation generators.

Corollary (the parametrization IS the geometry). [INFERENCE, high] Localization enters a net only through $\mathcal I$ , and giving $\mathcal I$ a causal/region structure is giving the geometry ( $n_1$ ). This is a framework theorem: by Brunetti–Guido–Longo, positivity of the energy $\Leftrightarrow$ isotony of the net (so localization and the spectrum condition coincide as one input); by Neeb–Olafsson the object that makes $\Delta^{it}$ geometric is an Euler element $h$ ( $\mathrm{ad}\,h$ eigenvalues $\{-1,0,1\}$ ), "the central ingredient … for wedge localization," from which wedges are defined depending on $h$ and the causal structure. So $n_1,n_2,n_3$ are inseparable from the localization: any net-carrier presupposes them.

2. The candidate evader, computed: the HSMI difference subspace

The one structure that might evade §1 is the relative modular data of an HSMI $\mathcal M(O_1)\subset\mathcal M(O_0)$ (shared $\Omega$ ). Worked in the explicit U(1)-current / chiral standard pair: one-particle space $L^2(\mathbb R_+,dp)$ ( $p>0$ , positive energy); $T(a)$ with $P\ge0$ ; half-line modular flow $\Delta_0^{it}=$ dilation; $H_1=T(1)H_0\subset H_0$ . In rapidity $p=e^\theta$ the modular flow is $\theta$ -translation and $\ln\Delta_0=$ multiplication by the rapidity-frequency $w$ . [ESTABLISHED model — Longo standard pairs; Borchers]

Numerical results (fresh RNG, chiral standard pair):

Borchers relation $\Delta^{it}T(a)\Delta^{-it}=T(e^{-2\pi t}a)$ : rel. err. $10^{-4}$ (discretization-limited). The new datum the inclusion supplies is $P\ge0$ .
$P$ -form (the genuine new relative-modular object): signature $(33,0,7)$ — positive and local ( $\langle f,Pf\rangle=\int p|\psi|^2\,dp\ge0$ ; super-exponentially decaying cross-coupling, $10^{-21}\to10^{-85}\to0$ over separations $3,6,12,24$ ).
$\eta_0=\mathrm{sgn}(\ln\Delta_0)$ form: signature $(20,20,0)$ — indefinite $(\infty,\infty)$ , but $\sim 1/\theta$ kernel; couples disjoint rapidity windows at a constant $7.5\times10^{-3}$ independent of separation — non-local (sign-of-boost = Hilbert-transform sibling).
$(1-\Delta_0)$ graph-Gram: sign $=-\mathrm{sgn}(\ln\Delta_0)$ everywhere — same boost-diagonal non-local object.
Covariance: $[\eta_0,\Delta^{it}]=0$ to $5.7\times10^{-16}$ (canonical, non-local); candidate local $m=\mathrm{sgn}(p-p_0)$ fails covariance (locus boosts away).

Verdict on the candidate: the HSMI difference/collar is definite where local ( $P$ ) and indefinite only where non-local ( $\Delta$ ). The relative modular data of two algebras does not produce a localized indefinite algebra-compatible pairing. The candidate evader is closed.

3. Literature direction-map (verified)

Lashkari–Leung–Moosa–Ouseph, Modular Intersections, Time Interval Algebras and Emergent AdS $_2$ , arXiv:2412.19882, JHEP 10 (2025) 153 [verified]: a (twisted) modular inclusion/intersection in any quantum system implies a $\widetilde{\mathrm{PSL}(2,\mathbb R)}$ representation — net-modular data $\to$ group. The geometry is an extra input: verbatim, "AdS $_2$ emerges from insisting on a local description." Localization = the demanded input ( $n_1$ ); the modular conjugation/flow are non-geometric off the algebra (a Generalized Hilbert Transform) and only geometrize in the bulk one chooses.
Wiesbrock, Symmetries and Modular Intersections of Von Neumann Algebras, Lett. Math. Phys., DOI 10.1023/A:1007361114049 [verified]; Half-sided modular inclusion and the construction of the Poincaré group, Commun. Math. Phys. (BF02100104) [verified]: generates $\mathrm{SL}(2,\mathbb R)$ /Poincaré given the $\pm$ half-sided + $J$ -commutator relations; half-sidedness $\Leftrightarrow$ a positive Hermitian generator — positivity is built into the inclusion ( $n_2$ ). Installs the group, not the localization.
Morinelli–Neeb–Olafsson, Orthogonal pairs of Euler elements II, arXiv:2603.26390 (2026) [verified]: a geometric Bisognano–Wichmann for nets of standard subspaces — but Euler elements are assumed ("fundamental … for wedge localization"). Their From local nets to Euler elements, arXiv:2312.12182 [verified], still inputs the Lie group with an Euler element, the BW property, and localizability.
Brunetti–Guido–Longo, modular localization, arXiv:math-ph/0203021 [verified]: positivity of energy $\Leftrightarrow$ isotony; localization $\Leftrightarrow$ positivity for Poincaré — carrier and positivity premise coincide.
Modular-Berry connection (Huang–Ma arXiv:2003.12252; PRD 108 066003; arXiv:2505.04682) [existence verified]: lives on kinematic space — a space of regions, i.e. the index set already carries the geometry; curvature probes bulk lengths (encode), it does not generate localization.
Relative Positions of Half-sided Modular Inclusions, arXiv:2503.18036 [title verified; body unverified — PDF unextractable]: the relative position of two HSMI is a modulus, but within an already-localized translation-dilation structure.

Net (scoped, two-session negative search): every 2020–2026 construction that turns net-modular data into geometry installs the localization through the index set / Euler element / "insist on locality" step. No construction derives a localized indefinite carrier from a structureless net. [OPEN — scoped negative search]

4. Consequences proposed for the wiki

OP-46 hedge: MEDIUM-HIGH → toward HIGH, conditional now on a single net-constructibility ansatz (down from the broader single-algebra scope). The richest structurally-distinct input — relative modular data / HSMI / modular intersection / modular-Berry over a net — is closed; the dichotomy (positive-local vs indefinite-nonlocal) is the net analogue of the iteration-7 trichotomy and survives the upgrade.
HYP-FACTORIZATION-IS-GEOMETRY — formalized and graded. The carrier problem is now stated precisely (§1 Definition) and the closure theorem identifies its content: supplying a localization to net-modular data is identically supplying the geometry (the parametrization-is-the-geometry corollary). This is the sixth independent route converging on the carrier/localization step (the fifth was iter-7's $\eta_0$ non-localizability).
HYP-ENCODING-SCREEN and the CONCLUSION §7 no-test claim inherit the hardened grade with the same condition.
Two proof obligations before HIGH is unconditional: (i) replace the net-constructibility ansatz by a naturality theorem; (ii) rule out a modular-Berry-holonomy-natural localized indefinite form not presupposing kinematic space (the residual escape — it imports the index-set geometry, so the obligation is to show it cannot be made carrier-free).

5. Open subquestions

Can the net-constructibility ansatz be replaced by a naturality theorem (every $\mathrm{Aut}(\mathcal N,\omega)$ -natural localized Gram is a function of the net's modular generators)? [OPEN]
The residual escape: is there a modular-Berry-holonomy-natural localized indefinite form that does not presuppose kinematic space (the space of regions)? Modular-Berry curvature is covariant under a non-modular connection, so it is not literally covered by the covariance–locality obstruction — but it lives on an already-geometric index set. [OPEN]
Does any exceptional-point / non-self-adjoint relative modular structure (PT-symmetric-type) on a net evade the dichotomy by being neither boost-covariant nor a multiplication operator? [OPEN]
Does the parametrization-is-geometry corollary admit a converse — a precise sense in which "a net with no causal index structure" provably cannot carry any localized form (a true non-existence, not an ansatz-scoped no-go)? [OPEN]

Proposed registry items

LEM-NET-CARRIER-DICHOTOMY — (new lemma)

Statement. Under the net-constructibility ansatz, every modular-covariant canonical form from net-modular data is positive-and-local (function of $P$ /identity), signature-free, or indefinite-and-non-local (sign of $\Delta$ , boost-diagonal); none is both indefinite and localized. The HSMI relative-modular candidate is definite where local and indefinite where non-local. Tag [INFERENCE] (high modulo the ansatz).

HYP-FACTORIZATION-IS-GEOMETRY-R1 — (formalization + grading)

Statement. The carrier problem is formalized (carrier = localized algebra-compatible $(n,n)$ field generated from net-modular data without $g_1$ – $g_7$ , $n_1$ – $n_3$ ). The net upgrade closes the structurally-richest escape: localization enters only through the index set, which is the geometry (BGL positivity $\Leftrightarrow$ isotony; Neeb–Olafsson Euler element; Lashkari et al. AdS $_2$ -from-locality). Net-level extension of the fifth carrier route (the boost-covariance engine applied to the family input) — the carrier-convergence count stays at FIVE per the binding referee correction; hedge toward HIGH, conditional on the net-constructibility ansatz. Tag [INFERENCE].

Referee verdict — track A1-carrier-problem (iteration 8, 2026-06-10)

Adversarial referee, default stance REFUTE; stance not sustained — the submission survives with four minor corrections. All mathematics re-derived; the engine numerically reproduced in fresh code; every load-bearing citation audited live.

Outcome: RESOLVED-NEGATIVE — UPHELD. Headline unchanged (7th consecutive). Principal hedge tightens, does not flip.

The track asks whether upgrading the modular input from a single algebra (M,Ω) to a net O↦M(O) evades the iteration-7 carrier no-go. The answer is no, and the argument is correct in its engine and honest about its limits.

Mathematics (the product) — sound. The covariance–locality obstruction is the load-bearing engine: by the Borchers relation Δ^{it}T(a)Δ^{-it}=T(e^{−2πt}a), the modular flow rescales p↦e^{−2πt}p, so a local indefinite multiplication form with a sign locus p₀ is boosted away — local ⟹ non-covariant ⟹ non-canonical; the only boost-invariant multiplication is the positive constant. I reproduced this numerically (η₀=sgn(ln Δ) covariant and non-local with a constant-strength tail; local sign m(p) covariance error ~0.40) and verified it is robust to the locality-representation subtlety: a sign change in the spacetime coordinate x is also dilated to e^s·x₀, so the obstruction holds whether locality is phrased in x or in p — this defeats the strongest natural objection (that "local" should mean multiplication in x, not p). The relative modular operator Δ_{O'|O} is positive with log over a boost spectrum; its sign-form is indefinite-and-nonlocal — the same (N) class as η₀ — so the genuinely-new HSMI datum splits cleanly into a positive-local generator P and indefinite-nonlocal Δ-forms, with no localized indefinite form anywhere.

Citations — all major references verified live. BGL positivity⟺isotony (verbatim ✓); Wiesbrock modular intersection ⟹ PSL(2,R), DOI matches ✓; Lashkari 2412.19882 — the quoted "insisting on a local representation… implies the emergence of AdS2" is essentially verbatim in §1 (Thms 28–29), and the boundary modular data is genuinely non-geometric (GHT) ✓; Morinelli–Neeb–Olafsson 2603.26390 — Euler element is an assumed input, not derived (the paper says so explicitly), which is exactly what the track needs ✓; Morinelli–Neeb 2312.12182 = Adv. Math. 458 (2024) 109960 ✓; Koot 2503.18036 — relative HSMI position via positive-generator spectral subspaces, supporting the "within an already-localized structure" reading ✓; Huang–Ma 2003.12252 modular-Berry on kinematic space ✓. No fabrication.

Corrections (all minor, all narrowing):

The "Closure Theorem" is not a theorem — it rests on the flagged net-constructibility ansatz, correctly tagged INFERENCE, but should not be styled "Theorem"; and its prong-(S) "signature-free" leg is established by separate enumeration of known form types, not by the ansatz (binding inheritance from iter-7's referee correction — must be stated in the claim).
The Lashkari quote "AdS2 emerges from insisting on a local description" is a mild compression of the paper's actual "…a local representation of the boundary algebras that satisfy Haag's duality" — fix the wording inside the quotation marks.
"Sixth independent convergence" overstates independence: the net closure runs on the same boost-covariance engine as iter-7's route 5 (η₀ non-localizability), extended to a richer input — it is the net-level extension of that route, not a mechanistically independent one.
The hedge move is accurately scoped ("toward HIGH, conditional on one net-constructibility ansatz," two named obligations outstanding, verdict unchanged) — no correction beyond #3.

No-go red-team — clean. No conflict with dS-cardinality (the carrier closure fixes no η=1/4G), the iter-7 trichotomy (this is its family-level generalization with identical ansatz status), OP-48c, or OP-49. The two flagged residuals (A-net ansatz; A-cov modular-Berry escape that imports the index-set geometry n₁) are genuine and correctly identify why HIGH is not yet unconditional.

Net. The extraordinary-claim test does not apply adversely: this is a RESOLVED-NEGATIVE strengthening a standing no-go, not a RESOLVED-POSITIVE. The track absorbs the structurally-richest remaining escape (relative modular operators, {J_O}, HSMI, modular intersection, modular-Berry over a net) as a further instance of the same closure geometry — adds operators, no localized indefinite form, localization living only in the index set, which is the geometry. Corrections are minor and corrective; the principal one (independence overcount) points the verdict no differently. Headline (PARTIAL / encodes-not-generates / not-yet-physics) unchanged; hedge hardened, not flipped.