{"id":"90a2d951-b51b-4e10-949a-bf6847a5008c","arxiv_id":"2411.09069","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Every Higman-Thompson group V_n (n≥2) is generated by three involutions, improving the known four-involution bound.","lead":"The Higman-Thompson groups V_n, a family of infinite simple groups with rich structure, can each be generated by just three involutions (elements that square to the identity). This settles an open question from 2023 and shows the sharpest involutive generating bound for these groups.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The final step of Theorem 3.14 hinges on Proposition 2.4, an unproved maximality claim for V_n deferred to a V-only result in [BBQS22]; if that maximality fails for some n≥3, the proof collapses.","rationale":"The paper is otherwise well-executed. The construction of S_α, the unique-difference condition, and the derivation of H_α ⊇ M are coherent, and the claimed result is likely true. However, a central theorem in a pure-math paper should not rest on an unproved 'same proof' claim when the cited source is for a different group (V_2) and the generalization is non-obvious. The reader's verdict already flagged this as the weakest assumption; my stress-test confirms it. I recommend conditional acceptance: the authors must provide a self-contained proof of Proposition 2.4 or a citation that covers V_n for all n. The minor undefined notation V_n^1 (used in Lemma 3.5) should also be fixed, but it is not load-bearing.","tokens_in":7472,"tokens_out":25824,"duration_ms":217309,"concrete_test":"Inspect [BBQS22, Proposition 6.3] and its proof, then re-derive Proposition 2.4 for n=3 by transcribing every step of that proof to the ternary Cantor set, replacing the binary tree with the ternary tree and Sym(2) with Sym(3). Check whether any intermediate claim uses the fact that a root-cone's complement is a single cone (which holds only for n=2) or that Sym(2) is abelian. If such a step exists, Proposition 2.4 is not a consequence of [BBQS22]; the paper must then supply a proof, and until it does Theorem 3.14 is unproved. If the transcription succeeds, the maximality assertion is vindicated.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Proposition 2.4 states that M := ⊕_{i=1}^n x_i V_n ⋊ 9Sym(X_n) is maximal in V_n, and it is the sole basis for the last line of Theorem 3.14: H_α contains M and τ∉M, so H_α=V_n. The proof given is: 'The following is a generalization of [BBQS22, Proposition 6.3]. The same proof given there, replacing V with V_n, works here as well.' But [BBQS22] concerns Thompson's group V=V_2, and maximality of the analogous subgroup in V_2 rests on the complement of x_1C being a single cone x_2C; for n≥3 the complement is a union of n−1 cones. If any step in the [BBQS22] argument uses that two-cone structure (for instance, to reduce an arbitrary element outside M to a translation that together with M generates V), then Proposition 2.4 may fail for n≥3 and Theorem 3.14 does not follow. The rest of the construction (Lemmas 3.8–3.13) is internally consistent: the commutator identities in Lemma 3.8 are valid under the stated support hypotheses, and the subdirect-product argument in Lemma 3.10 is correct. The only load-bearing gap is the unsupported maximality assertion.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs, for each n ≥ 2 and each finite sequence α = (α_1, …, α_ℓ) of involutions in a certain derived subgroup V_n^1, a triple of involutions (9σ, τ, s_α) in the Higman–Thompson group V_n. The main theorem (Theorem 1.2) asserts that this triple generates V_n, thereby answering positively the question whether V_n is (2,2,2)-generated for every n. The proof shows that the subgroup H_α generated by the triple contains M := ⊕_{i=1}^n x_i V_n ⋊ 9Sym(X_n), and then uses the asserted maximality of M in V_n (Proposition 2.4) together with τ ∉ M to conclude H_α = V_n.","tokens_in":7733,"tokens_out":24418,"duration_ms":229521,"significance":"If the gap around Proposition 2.4 is repaired, this is a significant and clean result: it gives an explicit positive answer to a natural question of Corson–Hughes–Müller–Varghese and produces a whole family of three-involution generating sets for every Higman–Thompson group V_n. The core of the argument (Lemmas 3.8–3.13) is self-contained and correct: the commutator computations are transparent, the unique-difference property is used effectively to kill cross terms, the subdirect-product lemma is sound, and the construction is explicit rather than existential. The paper is also careful to make the generating involutions concrete and checkable.","major_comments":[{"comment":"The final step of the main proof is load-bearing and rests entirely on Proposition 2.4, which asserts that M := ⊕_{i=1}^n x_i V_n ⋊ 9Sym(X_n) is maximal in V_n. No proof is supplied; the text says only that it is a generalization of [BBQS22, Proposition 6.3] and that “the same proof given there, replacing V with V_n, works here as well.” This transfer is not automatic: [BBQS22] treats Thompson's group V = V_2, and for n ≥ 3 the complement of x_1 C_n is a union of n − 1 cones, whereas in V_2 it is a single cone. If any step in the maximality argument uses the two-cone structure, Proposition 2.4 may fail for n ≥ 3. Because Theorem 3.14 uses maximality as the sole mechanism to promote the inclusion M ≤ H_α to H_α = V_n, the authors must either include a proof of Proposition 2.4 for all n, or cite a source where it is proved for general V_n, and should explain why the multi-cone complement does not affect the argument. Without this, the proof of Theorem 1.2 is incomplete.","section":"§2, Proposition 2.4; §3, Theorem 3.14"}],"minor_comments":[{"comment":"The notation V_n^1 is used throughout but never defined; the reader must infer that it denotes the derived (or the finite-index simple) subgroup of V_n. Please define it explicitly, especially because when n is odd V_n itself is not simple.","section":"§3, Lemma 3.5"},{"comment":"Theorem 2.5, that the abelianization of V_n is generated by the image of 9(x_1,x_2), is stated without proof or citation. It is used in Lemma 3.12 to pass from V_n^1 and 9σ to V_n, so a reference (for example to Higman's book) or a short proof should be added.","section":"§2, Theorem 2.5"},{"comment":"The case analysis in the displayed definition of s_α appears to omit the case k = 0, i = 1. Formula (3.1) makes the intended definition clear, but the displayed definition should be rewritten as a partition of C_n so that the domain is unambiguous.","section":"§3, Definition 3.1"},{"comment":"The reference [CHMV23] is given with a Google Drive link; a stable arXiv identifier or a journal reference would be more appropriate.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The main theorem is plausible and the construction is elegant, but the unproved maximality assertion in Proposition 2.4 is exactly the kind of load-bearing point that should be fixed before publication. I see no circularity: Proposition 2.4 is independent of the main theorem. If the authors supply a proof or a precise reference for the maximality claim for all n, the paper should be accepted."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague, here's my take on Schesler–Skipper–Wu. The headline result is real: they show every V_n is (2,2,2)-generated by three explicit involutions, improving the four-involution bound and answering the first part of CHMV Question 1.1. The construction is fresh: a spinal-type involution s_α plus the two standard elements σ and τ, with a unique-difference condition that makes commutators of s_α with its conjugates produce exactly the generators they need. The commutator algebra in Lemmas 3.8–3.11 is careful and checks out; the subdirect-product argument and use of local finiteness of the volume-preserving subgroup is clean. I found no circularity: they're not assuming the result.\n\nThe soft spot is exactly where the stress-test note lands. Proposition 2.4 — the maximality of ⊕ x_i V_n ⋊ Sym(X_n) in V_n — is the sole basis for the last line of Theorem 3.14, and it is stated as a 'same proof' generalization of [BBQS22, Prop 6.3], which is proved only for V (=V_2). The worry that the V-proof uses the fact that the complement of a first-level cone is a single cone, and that this breaks for n≥3, is a reasonable one, and I don't think a reader should have to take that on faith. This is load-bearing, not cosmetic. There's also a smaller issue: V_n^1 appears in Lemma 3.5 without definition; presumably it's the index-2 simple subgroup for odd n, but it needs to be spelled out.\n\nThat said, the maximality claim itself may well be true — the proof might carry over with minimal changes — but the authors should either provide the proof or cite a version that covers V_n. As written, a skeptical referee could not verify the main theorem without going to the source and doing the generalization themselves.\n\nWho's this for? People studying generating sets of infinite simple groups; it's a concrete, citable advance that also introduces a transferable technique. If the maximality gap is patched, I'd want to cite it. I'd send it to peer review — the result is important enough and the construction solid enough to deserve referee time, and the referee can be sent to chase down Proposition 2.4. Not a desk reject.","headline":"A genuinely new and likely correct construction proving V_n is (2,2,2)-generated, but the final maximality step is an unproved generalization from V to V_n that a referee must check.","tokens_in":8294,"tokens_out":4041,"would_cite":true,"duration_ms":39426,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["20F05","20E32"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that for every n ≥ 2, the Higman–Thompson group V_n is generated by three involutions.","keywords":["Higman–Thompson groups","(2,2,2)-generated","involutions","spinal elements","commutator stable generating sets","unique difference property","maximal subgroups","branch groups"],"falsifier":"Exhibit, for some n ≥ 2, a proper subgroup K of V_n satisfying ⊕_{i=1}^n x_i V_n ⋊ Sym(X_n) < K < V_n; this would invalidate Proposition 2.4 and remove the final step of the proof. In practice, one could compute the normalizer of the natural first-level subgroup in V_n for small n and check whether any element outside it generates a proper overgroup.","tokens_in":7255,"feed_emoji":"♾️","tokens_out":9714,"duration_ms":82713,"temperature":0.7,"pith_summary":"The paper answers an open question by proving that every Higman–Thompson group V_n, for n ≥ 2, can be generated by three involutions. These groups act on the Cantor set of infinite words over an n-letter alphabet, and they are infinite simple groups (up to an index-2 subgroup when n is odd). The authors build, for each admissible sequence α, a triple {σ̂, τ, s_α}, where σ̂ swaps the first two letters, τ is a fixed involution with support near the left edge of the Cantor set, and s_α is a spinal involution, a homeomorphism acting along one thin branch of the tree of finite words. The proof shows this triple generates all of V_n, settling the (2,2,2)-generation case of that question.","feed_headline":"Three involutions generate each Higman-Thompson group","feed_subtitle":"For every n≥2, the group V_n is generated by three involutions, settling an open question.","key_machinery":"The spinal involution s_α is the central object: it is defined by s_α = $x_1^{{ℓ+1}}$ σ̂ ∏_{k=1}^{ℓ} x_1^k x_2 α_k and it is a genuine involution exactly when each α_k has order at most 2. The argument is carried by two structural properties of the sequence: the unique difference property, which makes the commutators [s_α, $s_α^{{t^k}}$] isolate single factors x_1^j x_2 [α_j, α_i] without overlap, and commutator stability, meaning the commutators [α_i, α_j] already generate $V_n^{1}$, which lets those isolated factors build up the whole derived subgroup. These lemmas construct larger and larger subgroups of V_n until the maximality of M forces equality.","core_discovery":"The central claim is Theorem 1.2: for every integer n ≥ 2, the Higman–Thompson group V_n admits a generating set consisting of three involutions. The discovery is a uniform construction rather than an existence argument alone: given a sequence α = (α_1, …, α_ℓ) of involutions in the derived subgroup $V_n^{1}$ which is commutator stable, with the positions of non-identity entries satisfying the unique difference property, the three involutions σ̂, τ, and s_α generate V_n. The proof establishes that this subgroup contains the natural subgroup M = ⊕_{i=1}^n x_i V_n ⋊ Sym(X_n), then uses maximality of M in V_n to conclude the subgroup is the whole group.","pith_inferences":["The construction is uniform in n, but the admissible sequence α is guaranteed through existence arguments using finite generation and simplicity of V_n^1; for a fixed n, writing down an explicit α and the three involutions may require nontrivial computation.","The paper answers only the (2,2,2) half of Question 1.1; whether V_n is (2,3)-generated remains open, and nothing in this proof either supplies or rules out such a generating pair.","Because s_α is modelled on spinal tree automorphisms, the same three-involution scheme might extend to other groups acting on rooted trees, provided analogues of the maximality and simplicity lemmas hold."],"forward_implications":["For every n ≥ 2, the group V_n is (2,2,2)-generated, answering the first half of Question 1.1 in the affirmative.","Since V_n is not (2,2)-generated by involutions, three is the minimal possible number of involutions in a generating set for V_n.","The generating triples are parametrized by admissible sequences α, giving a family of small generating sets rather than a single isolated example.","The proof connects Higman–Thompson groups to spinal-element techniques from branch group theory, suggesting a shared toolbox for generation questions."],"supporting_citations":[{"why":"Supplies the definition of V_n, the simplicity of V_n^1, and the abelianization statement used to lift generators from V_n^1 to V_n.","marker":"[Hig74]"},{"why":"Provides Proposition 6.3, whose generalization to V_n is Proposition 2.4 and carries the maximality step on which the final conclusion depends.","marker":"[BBQS22]"},{"why":"Poses Question 1.1 that the paper answers and records the earlier result that V_n is generated by four involutions.","marker":"[CHMV23]"}],"fun_headline_variants":["Three involutions suffice for each Higman-Thompson group","V_n always generated by three involutions","Just three involutions generate every Higman-Thompson group","Three involutions do it for all Higman-Thompson groups"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof relies on the claim, imported from the n=2 case without a written proof, that the subgroup consisting of self-similar copies of V_n on the n top-level branches together with all permutations of those branches is a maximal subgroup of V_n.","fun_headline_variants_meta":{"raw":{"variants":["Three involutions suffice for each Higman-Thompson group","V_n always generated by three involutions","Just three involutions generate every Higman-Thompson group","Three involutions do it for all Higman-Thompson groups"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000729,"raw_usage":{"total_tokens":3183,"prompt_tokens":783,"completion_tokens":2400,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":399,"completion_tokens_details":{"reasoning_tokens":2332}},"tokens_in":399,"tokens_out":2400,"duration_ms":17793,"temperature":1.0,"reasoning_tokens":2332,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T21:07:11.693255+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Exhibit, for some n ≥ 2, a proper subgroup K of V_n satisfying ⊕_{i=1}^n x_i V_n ⋊ Sym(X_n) < K < V_n; this would invalidate Proposition 2.4 and remove the final step of the proof. In practice, one could compute the normalizer of the natural first-level subgroup in V_n for small n and check whether any element outside it generates a proper overgroup.","supporting_citations":[],"review_version":1}