Iteration 17 synthesis — the (E_O) assault

Status: Clock-(1)-δ-triggered analytical iteration — the first dedicated, direct attack on the residual hypothesis (E_O) (iteration 16 located it but did not attack it). Five finder lenses (E1–E5), each targeting (E_O) from a distinct technology; three independent binding adversarial referees (default stance REFUTE); a watch-mode sweep #2 run in the same pass. Every load-bearing 2020–2026 citation web-verified live this session (author of the integration independently re-verified the four most load-bearing: arXiv:2209.04681, 2312.08525, 2012.00565 v3-comment, and the wedge/Marrakchi–Vaes facts). Outcome: Verdict UNCHANGED — 16th consecutive confirmation (PARTIAL coherence / Program A encodes-not-generates / not-yet-physics / not forecastable). The principal hedge STAYS at R7 — (E_O) was neither proved nor disproved. Unlike iteration 16 (which advanced R6 → R7), no hedge movement: no R-advance, no R-retreat, no numeric ID consumed, carrier-convergence count stays FIVE. What is genuinely new and refereed: (1) the free-field case of (E_O) is sharply reduced to a concrete one-particle spectral question, and the naive "no-eigenvalue-0 ⟹ ergodic" reading is refuted; (2) all three structure-theory routes to (E_O) for general nets provably fail; (3) the adversarial disproof fails on eight candidate mechanisms — no carrier, the 16-iteration streak is intact. Last updated: 2026-07-07 Iteration: 17

1. Why this iteration was legitimate (the clock rule)

The ROADMAP terminal-saturation rule forbids new analytical iterations "unless one of those clocks delivers a genuinely new input." Iteration 16 pinned the entire remaining no-go to a single named open hypothesis and designated it the principal watch item (δ): (E_O) — vacuum ergodicity on massive double-cone algebras. It never attacked (E_O) as a mathematical object; the iter-16 synthesis §7 named "a proof of (E_O)" as "the honest expected outcome" of the next legitimate step. Iteration 17 is that step: a dedicated, five-lens, direct assault on (E_O). This is new work, not a re-run of a closed lever. (Watch-sweep #2, run in the same pass, also re-surfaced the (E_O)-adjacent Marrakchi arXiv:2606.23636 — see watch-sweep #2.)

2. The exact object

Hypothesis (E_O) [OPEN]. For the vacuum ω on the double-cone algebra M(O) of a massive theory in the standing class (H) (isotony, locality, Haag duality, split, Poincaré covariance with unique vacuum, Reeh–Schlieder, Bisognano–Wichmann for wedges, local algebras hyperfinite type III₁), the modular flow σ^ω_t = Ad Δ_O^{it} is ergodic: M(O)_ω := M(O) ∩ {Δ_O^{it}}′ = ℂ1 for every double cone O.

The wedge analogue M(W)_ω = ℂ1 is a theorem (Borchers scaling: the boost is the modular flow, the two lightlike translation semigroups are scaled to infinity, a boost-fixed vector is translation-fixed by a Cauchy–Schwarz equality hence = cΩ, and Reeh–Schlieder promotes this to the algebra). That engine has no analogue for a bounded double cone — no translation semigroup maps O into itself — which is exactly why (E_O) is the residual. The massless/conformal double cone is geometric (Hislop–Longo: Δ_O is a dilation), so (E_O) is expected to close there; the open case is specifically massive double cones, where Δ_O is non-geometric.

3. Per-lens verdicts (all refereed; corrections binding)

Lens	Technology	Thesis	Referee
E1 — free-field modular spectrum (reduction)	Second-quantization functor: Δ_O = Γ(δ_O), σ_t = Ad Γ(δ_O^{it}); reduce (E_O) to a one-particle question	E_O_OPEN — reduced. Correctly identifies that σ_t is a gauge-invariant, number-preserving Bogoliubov flow, so M(O)_ω is not controlled by fixed one-particle vectors alone: Γ(u)dΓ(A)Γ(u)* = dΓ(uAu) makes N = dΓ(1) and every dΓ(E) modular-fixed regardless of point spectrum. Hence "no eigenvalue 0 of ln δ_O ⟹ ergodic" is NECESSARY but PROVABLY NOT SUFFICIENT — the bosonic number-type operators survive. The residual is an affiliation* clause: which functions of ln δ_O are affiliated to M(O).	Target A: PARTIAL. Reduction confirmed; the naive slogan refuted; the correct sufficient condition is E2's (below).
E2 — III₁ centralizer / ergodicity structure theory	Connes spectrum Γ(σ)=ℝ; Marrakchi–Vaes genericity; weak-mixing ⟺ trivial centralizer	E_O_OPEN. The clean sufficient condition: weak mixing of σ^ω on M(O) ⟹ M(O)_ω = ℂ1 (elementary Tomita–Takesaki), and for the quasi-free vacuum this reduces to purely continuous one-particle spectrum of ln δ_O (no eigenvectors). But Γ(σ)=ℝ and III₁-ness are outer-structure invariants strictly weaker than centralizer-triviality; Marrakchi–Vaes genericity (dense G_δ) cannot pin the distinguished symmetric vacuum; and modular mixing on a bounded region is not a corollary of spacelike cluster decomposition (the modular orbit carries nothing to spacelike infinity).	Target B: CONFIRMED (route does not settle it; fatal gap identified).
E3 — wedge→double-cone reduction	Half-sided modular inclusion (Wiesbrock/Borchers), modular intersection	E_O_OPEN — clean impossibility for this route. O = W₁ ∩ W₂ and M(O) ⊂ M(W₁), but Δ_O-fixity does NOT imply Δ_W-fixity (different algebras, same Ω, non-commuting modular operators); HSMI reconstructs the ambient boost/translations (scaled by Δ_W, sliding O out of itself), never the non-geometric Δ_O. The route closes (E_O) iff Δ_O is geometric = the conformal case already conceded.	Target B: CONFIRMED. Two demerits (below), conclusion survives.
E4 — adversarial disproof	Construct a non-scalar element of M(O)_ω	E_O_OPEN — disproof FAILS. Four candidate mechanisms (localized conserved flux; SO(d−1) rotation invariants; hidden point spectrum of ln Δ_O; integrable/almost-periodic modular data) plus four the referee added (split-interpolating type-I factors; central sequences; η₀ = sgn(ln Δ_O); (−1)^N parity) all die on the triple wall {bounded ∧ in M(O) ∧ exactly Δ_O-fixed}. The SO(d−1) stabilizer supplies only unbounded, global, non-local generators; the only bounded modular-fixed vector is Ω (⟹ scalars via Reeh–Schlieder). No entry-ticket carrier.	Target C: CONFIRMED, with a binding calibration (below).
E5 — is it already known?	Exhaustive AQFT / operator-algebra literature sweep	E_O_OPEN in the literature. No paper states, proves, or disproves trivial vacuum centralizer on a bounded double cone (massive OR even massless). The massive double-cone modular operator is explicitly characterized (numerically) yet the ergodicity question is never posed. (E_O) is a genuine open problem, not folklore — and, because the modular operator is in hand, an unusually tractable one for the free field.	Sustained (cross-checked against E1/E2).

4. The one thing this iteration actually establishes (the free-field reduction)

The deliverable is a referee-verified reduction of the free-field case to a clean, decidable spectral statement:

Free-field reduction [INFERENCE, high]. For the free massive scalar, M(O) = R(K_O), Ω the Fock vacuum, Δ_O = Γ(δ_O), σ_t = Ad Γ(δ_O^{it}) a gauge-invariant number-preserving flow. If the one-particle modular Hamiltonian ln δ_O has purely continuous spectrum on the one-particle space (equivalently: the one-particle modular flow is weakly mixing / has no eigenvectors), then σ^ω is weakly mixing on M(O), hence M(O)_ω = ℂ1 — (E_O) holds for the free field, and the net-naturality no-go becomes a theorem on the free-massive-scalar subclass (modulo n₁).

Two binding corrections make this the right statement:

The naive reading is refuted. "No eigenvalue 0 of ln δ_O ⟹ ergodic" is only necessary — it kills fixed Weyl operators W(f) but not the number-type dΓ(A), A ∈ {δ_O^{it}}′, which are modular-fixed for any point spectrum. Absolute continuity is not enough by fixed-vector counting; one needs the full weak-mixing / purely-continuous-spectrum condition, which does discharge the dΓ(A) obstruction (weak mixing forces the fixed-point algebra to ℂ1, bypassing any need to characterize affiliation directly). This corrects the too-quick "trivial kernel ⟹ (E_O)" intuition (bosonic Fock always carries the σ-fixed number operators).
The spectral premise is unproven [OPEN]. The massive double-cone one-particle modular Hamiltonian is known only numerically (Bostelmann–Cadamuro–Minz, On the mass dependence of the modular operator for a double cone, arXiv:2209.04681, Ann. Henri Poincaré 2023 — "close to a multiplication operator, but … dependent on mass and angular momentum"; Cadamuro, arXiv:2312.08525) — a near-multiplication diagonal part plus an unresolved non-diagonal kernel "orders of magnitude smaller … cannot confirm whether zero." No closed-form massive double-cone modular Hamiltonian exists: Longo–Morsella, The massless modular Hamiltonian (arXiv:2012.00565), v3 comment verbatim — "the results on the massive modular hamiltonian of a ball contained a gap, and have been removed." The numerics lean toward continuous spectrum (a multiplication operator has purely continuous spectrum) — hence toward (E_O) holding — but this is not a proof.

Net: the reduction is proof-grade; the spectral fact it needs is [OPEN], numerical-only, leaning-true. So even the decidable base case is genuinely open — but now pinned to a single concrete spectral computation rather than an abstract centralizer question.

5. Why the verdict stands a 16th time, and why the hedge does NOT move

No lens produced a carrier, a reverse-Weinberg–Witten witness, a generative construction, or a distinguishing test. (E_O) was neither proved (no spectral input; no criterion shown to hold for the vacuum on a bounded region) nor disproved (no non-scalar affiliated modular-fixed element exhibited). Per the house rule, a hedge advances only when a condition is discharged or added — neither happened, so HYP-CKV-VACUITY stays at R7, grade (conditional HIGH) and condition string {n₁ AND (E_O)} both unchanged. This is the disciplined outcome: the iteration produced real refereed mathematics, but its verdict-level result is confirmation, not movement.

The verdict is nonetheless modestly hardened at framing level (not proof level): (i) the disproof was attempted head-on and failed cleanly, which strengthens — without proving — the case that (E_O) holds; (ii) the no-go is now shown to be un-closable by every reasoning-decidable structure-theory route (weak-mixing-from-clustering, wedge-to-double-cone HSMI, and genericity-upgrade all carry identified fatal gaps), which sharpens "only external input can move it"; (iii) the literature sweep confirms (E_O) has never even been posed for the bounded massive case, so the wiki's isolation of it is a genuine contribution.

6. Standing-record notes (binding; do NOT propagate the errors)

Citation attribution: arXiv:2209.04681 is Bostelmann–Cadamuro–Minz, NOT "Figliolini–Guido" (one finder mislabeled it; the referee caught it). Figliolini–Guido is the separate 1989 Ann. IHP paper The Tomita operator for the free scalar field and the 1994 On the type of second quantization factors. The mislabel is not written into any registry page.
Overstated sub-claim, not propagated: E3's "for massive nets the wedge pair is not a Wiesbrock modular intersection" is overstated and likely wrong — massive wedge modular flows are geometric (Bisognano–Wichmann; Mund's massive BW theorem), and modular intersections of wedges do reconstruct the Poincaré group for massive nets. E3's conclusion survives on the correct reasoning (the reconstructed flow is the ambient boost/translation, not Δ_O), so only the sub-claim is dropped.
Calibration (binding): E4's "net evidence FOR (E_O) / leaning true" is down-graded by the referee to bare OPEN — an adversary's failure to construct a witness from guaranteed symmetries is absence-of-witness, not absence-of-existence, since the massive double-cone modular structure is spectrally uncharted with no closed form.
Unverified sub-claim: the Figliolini–Guido 1989 "1 is not an eigenvalue of (Δ_B+1)(Δ_B−1)⁻¹" point could not be verbatim-confirmed (numdam metadata only) and stays [unverified]; it is not load-bearing.

Addendum (2026-07-10, iteration 19 — tag RETIRED to [ESTABLISHED]): the full FG 1989 paper was downloaded from numdam and read; p. 429, verbatim: "It is clear from (3.6) that 1 ∉ σ_p(δ)…" — FG state and prove the E=0 exclusion via antilocality (Segal–Goodman). Independently re-proved via factoriality (LEM-K0). See 2026-07-10-iter19-prove-the-numerics.md §1.

7. The re-sharpened watch state (clock 1)

Unchanged in kind from iteration 16; sharpened at (δ):

(α) new definitions/theorems for expectation-free relative-bicentralizer-type invariants — STATIC;
(β) bicentralizer triviality — STATIC (Marrakchi arXiv:2606.23636 is with expectation; the double-cone gate is expectation-free, so it does not transfer);
(γ) expectation-free relative T-invariant computations (adjacent conjecture T(M(O) ⊂ M(W)) = {0}) — STATIC;
(δ) — PRINCIPAL, now sharpened: any resolution of (E_O). The sharpest decidable sub-target is now "prove purely continuous spectrum (no eigenvectors) of the massive double-cone one-particle modular Hamiltonian ln δ_O for the free scalar" — if proved, (E_O) holds on the free subclass (via weak mixing) and R7 → R8 would fire there (the no-go a theorem on the free-massive-scalar subclass, modulo n₁); general (H) stays conditional. Adjacent technology: Bostelmann–Cadamuro–Minz / Cadamuro numerics (arXiv:2209.04681, 2312.08525); the second-quantization functor (Figliolini–Guido; Conti–Morsella arXiv:2305.07606).
(ε) the DCNG definability lever — STATIC (needs an operator-algebra / continuous-logic collaborator).

Clock (2) unchanged: 2027 DESI five-year w(z) (Euclid cosmology co-timed mid-2027) is the nearest dated external event.

8. What could move the verdict from here

Unchanged in kind, sharper in address: (i) a proof of purely-continuous massive double-cone one-particle modular spectrum → (E_O) on the free subclass → R7 → R8 there → the no-test corollary unconditional on that subclass; (ii) a proof (or general-net analogue) of (E_O) → the net-naturality no-go a theorem on class (H); (iii) a massive double-cone counterexample fiber surviving all walls (¬(E_O) is only the entry ticket) → the first generative carrier and a verdict flip; (iv) the external clocks (2027 DESI w(z) — target, not wager). The dossier stands as posed, its §3 residual now carrying the iteration-17 free-field reduction.