Property for Artin groups (AIM Problem 8.6): the irreducible spherical case
A group has property if every automorphism has infinitely many twisted conjugacy classes, where and are -twisted conjugate if for some . Problem 8.6 of the AIM problem list Geometry and topology of Artin groups asks: which Artin groups have property ? A remark by Ignat Soroko on the list records the property as known for the spherical types and (), and (), for the Euclidean types and and for certain right-angled Artin groups, and as failing for abelian right-angled Artin groups.
- Result
- Proved(see note)
- Status
- Partial result
- AI contribution
- AI-discovered
- Method
- Argument
- Field
- Geometric group theory; twisted conjugacy in Artin groups
- Posed by
- AIM Problem List, Geometry and topology of Artin groups, Problem 8.6
- Year posed
- —
- Years open
- —
- Solved
- 2026-11-10
- Model
- Claude Opus 5.5 (Claude Code); GPT-6 Astra (Codex)
- Vendor
- Anthropic; OpenAI
- Collaborators
- Jason Davies
- Verification
- Unreviewed
- Publication
- Preprint
- Significance
- 12 / 100
- Disclosed cost
- —
- Wikipedia
- No dedicated article
What was actually shown
Theorem A: every irreducible Artin group of spherical type and rank at least two has property . Types , , , (Calvez and Soroko) and , (Soroko and Vaskou) were known, and the paper notes that follows from a 2026 lemma of Paris and Soroko, so the new cases are , , , , and . Theorem B classifies for the crystallographic quotients , the pure subgroup: it holds in rank at least three, and for exactly when , or is odd and . Proposition C: has an automorphism with Reidemeister number 18, so the quotient does not inherit . The proof relies on Cohen and Paris's computer-assisted theorem that is characteristic, which the authors did not re-run. Non-spherical Artin groups are not covered, so Problem 8.6 stays open in general.
What the AI did
Per the paper's AI transparency statement and its Hexagon provenance statement, Claude Opus 5.5 (through Claude Code) and GPT-6-Astra (through Codex) developed the mathematical arguments, the exact Python certificates, the Lean formalisation and the manuscript, under the author's direction. Hexagon labels the paper as primarily AI-generated text with human understanding of no parts, and the provenance statement says no human expert has independently reviewed the work. The author takes responsibility for the content and for the accuracy of the disclosure. The repository's own audit file is described as an AI audit.
Verification
No independent mathematician has checked this yet; Hexagon's label says a human understands no parts of the paper. Unreviewed on this site's ladder, lowered from the submitted Lean-checked, because Lean does not prove the headline outright. Checked here on 19 November 2026: Hexagon paper 2611.00278v1 (introduction, Theorems A and B, Proposition C, Appendix C, the AI declaration) and the archived AIM page with Problem 8.6. The repository jasondavies/spherical-artin-rinf was read at commit ab2165a, the one the paper pins (67 Lean files, about 8,500 lines, Lean 4.35.0). A grep found no sorry, axiom, admit, native_decide, implemented_by, extern or unsafe. Two statements were read. artinQuotient_exists_reidemeister_eighteen states Proposition C for the presented quotient with no hypotheses. artin_E8_R_infty_FH, and its six siblings for , , , , , , concludes that every automorphism of the presented Artin group has infinite Reidemeister number, but takes four cited results as hypotheses: finiteness of the Coxeter group, Cohen and Paris's theorem that the pure subgroup is characteristic, the rank of its abelianisation, and the Franzsen-Howlett decomposition of Coxeter automorphisms ( and also have versions without the last). Theorem B, on which the paper's own proof of Theorem A rests, is not formalised. Not rebuilt here, and the repository has no CI. The Python certificates were not run.
Sources
- PaperHexagon PDF 2611.00278v1Calvez, Soroko: Property R-infinity for some spherical and affine Artin-Tits groupsSoroko, Vaskou: Property R-infinity for new classes of Artin groups (D_n, n at least 6)Paris, Soroko: Automorphisms of the Artin group of type D_5
- Lean proofCode and Lean repository (jasondavies/spherical-artin-rinf, commit ab2165a)Lean: Proposition C for the presented quotient (Dihedral5/Quotient.lean)
- Lean statementLean: conditional E8 theorem artin_E8_R_infty_FH (four cited results as hypotheses)
- OtherHexagonAIM Problem List, Geometry and topology of Artin groups, Problem 8.6 (archived 13 Feb 2025)
Submitted by MellowGander562 on