§6 · Agent 6 · Audit: integral models · 2026-10-18T04:02:26Z

Summary 6.9 · Verdict: (A3a)(A3a) Rank-one models over O (Sakellaridis's Axiom 2.2.1): for every Borel orbit Y of maximal rank and every simple root α, the O-scheme YP_α/R(P_α) is one of the four standard PGL2-spaces over O, with his scheme structure on Y, and YP_α is the closure of its generic fibre minus the closure of the generic boundary. holds for \(\mathcal M_\theta\) over every DVR with \(2\) invertible

proved audit: correct ★ key step

Second audit: correct. Its claim-by-claim outcomes agree with my own grading of Strategy 2.1 to Summary 2.10: no step fails, the three repairs (Audit 6.3, Audit 6.5, Audit 6.6) are valid one-line supplements, and the dependence on S4-11's sign convention is real but harmless because the setting fixes type E IV (52 versus 36 confirmed on the E6 Lie algebra). The established statement, Axiom 2.2.1Sakellaridis's Axiom 2.2.1, relabelled (A3a): rank-one models over O. with Sakellaridis's scheme structure and complement clause for E6/F4 over every DVR with 2 invertible, mixed or equal characteristic, p = 3 included, holds, so Theorem 7.4's use of it is justified; the list of unchecked non-local citations is accurate.

No step of A2's proof fails: three steps needed a line of proof (supplied), one input depends on S4-11's sign convention for the Chevalley involution, and the table was reproduced independently. Sakellaridis's Axiom 2.2.1Sakellaridis's Axiom 2.2.1, relabelled (A3a): rank-one models over O., with his scheme structure and the complement clause, holds for \(E_6/F_4\) over every DVR with \(2\) invertible, mixed or equal characteristic, \(p=3\) included.

Summary 6.9.

Outcomes, claim by claim:

Established: for \(\mathcal M_\theta\) and every DVR \(\mathcal O\) with \(2\in\mathcal O^\times\), the \(B_F\)-orbits are the 45 orbits \(Y_{x,F}\), all of maximal rank; Sakellaridis's \(\mathcal O\)-scheme of each is the smooth orbit scheme \(Y_x\simeq B/B_x\); \(YP_\alpha/R(P_\alpha)\) is \(\mathrm{PGL}_2\backslash\mathrm{PGL}_2\) for the 90 pairs with \(\alpha\in\Phi_x\) and \(B_2\backslash\mathrm{PGL}_2\simeq\mathbb P^1_{\mathcal O}\) for the other 180, \(P_\alpha\)-equivariantly; the complement clause holds for all 270 pairs. So (A3a)(A3a) Rank-one models over O (Sakellaridis's Axiom 2.2.1): for every Borel orbit Y of maximal rank and every simple root α, the O-scheme YP_α/R(P_α) is one of the four standard PGL2-spaces over O, with his scheme structure on Y, and YP_α is the closure of its generic fibre minus the closure of the generic boundary. holds over \(\mathbb F_q[[t]]\) and over the ring of integers of every \(p\)-adic field, \(p\) odd.

  1. Method. The table was recomputed from scratch before A2's scripts were opened (Audit 6.1); the \(\mathrm{Spin}_8\) signs, the torus \(T^\theta\) and \(C_W(w_L)\) were recomputed by exact arithmetic (Audit 6.2); every scheme-theoretic step was re-derived (Audit 6.3–Audit 6.7); the Conrad citations were read in the local PDF (Audit 6.8).
  2. What could have failed and did not. (i) Noncompact imaginary roots, which would create types T or N, are excluded by the triviality of \(\theta_0\) on \(\mathrm{Spin}_8\), confirmed for S4-11's convention. (ii) A disconnected or non-smooth \(T^\theta\) at \(p=3\) does not occur. (iii) The special fibre of the flat closure of an orbit could a priori be larger than the closure of the special-fibre orbit; Lemma 2.7 excludes this by the induction over \(P_\alpha\)-saturations, which works because the same closure sets \(\Sigma_x\) govern \(\bar F\) and \(\bar\kappa\). (iv) The complement clause read with complements in \(X\) is unsatisfiable for non-open saturations in any model, so A2's reading is the only meaningful one.
  3. Repairs supplied. Audit 6.3 step 2, Audit 6.5 step 4, Audit 6.6 step 4; each is a short standard argument, none changes a statement.
  4. Remaining caveat. The citations Steinberg, Mem. AMS 80, Thms. 7.5, 8.1; CGP Prop. A.8.10(2) / Edixhoven Prop. 3.4; EGA IV 11.3.10, 11.10.10, 17.9.1, 18.12.6 were not checked against a local source; their statements as used are standard and their hypotheses are met.

Verdict: the proof of (Q2) in Theorem 2.8 stands.

Builds on

  • Audit 6.1 · Independent recomputation of the \(45\times 6\) table of \(E_6/F_4\) (claim A2-6)
  • Audit 6.2 · Audit of the sign data, the torus \(T^\theta\) and \(C_W(w_L)\) (claim A2-2)
  • Audit 6.3 · Audit of the Borel-orbit classification in every characteristic \(\ne2\) (claim A2-3)
  • Audit 6.4 · Audit of smooth stabilizers and orbit immersions over \(\mathbb Z[1/2]\) (claim A2-4)
  • Audit 6.5 · Audit of the rank-one quotients \(YP_\alpha/R(P_\alpha)\) over \(\mathcal O\) (claim A2-5)
  • Audit 6.6 · Audit of the special fibres of orbit closures over \(\mathcal O\) (claim A2-7)
  • Audit 6.7 · Audit of Sakellaridis's scheme \(Y\) and the complement clause (claims A2-1, A2-8)
  • Audit 6.8 · Audit of the citations, the eleven-point checklist and the scope (claims A2-9, A2-10)

Built on by

Sources