{"id":"33929dee-5dbc-432f-9437-0d3034dd4fc5","arxiv_id":"2501.16270","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"Affine cactus groups embed into a semidirect product of an affine Gauss diagram group and the symmetric group, yielding linearity, trivial centre, and torsion bounds.","lead":"This paper proves new structural results about affine cactus groups, embedding them into a semidirect product of a right-angled Coxeter group and the symmetric group. It derives consequences such as a solvable word problem, a trivial centre, and a bound on the order of torsion elements.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 3.4's injectivity proof is not rigorous: 'continuing this process' never shows that arbitrary AD_n reductions lift to AJ_n after quasi-commutations, so τ=1 does not formally imply σ=1.","rationale":"The reader's weakest assumption identifies exactly the informal 'continuing this process' step in the proof of Theorem 3.4. My reading of the full manuscript confirms that this is the load-bearing point: Lemma 3.3 provides local lifts of commuting and cancelling moves in AD_n, but the induction that these lifts compose to a global reduction of σ is never formalized. In particular, quasi-commutation changes a generator rather than merely swapping two letters, so the correspondence between σ and the reducing word τ is not automatic. The paper also contains a concrete error in the definition of s_{k,l} for wrapping circular intervals, which makes ϕ ill-defined as written; this strengthens the need for a corrected and fully detailed proof. These issues are substantial enough to prevent acceptance in the current form, but they appear fixable, so the conditional verdict is appropriate. I do not see evidence that the main construction is false for n≥3, and the intended statement is plausible; hence rejection is not warranted based on this concern alone.","tokens_in":15170,"tokens_out":26113,"duration_ms":236298,"concrete_test":"Redo the proof of Theorem 3.4 as an explicit induction on the length of the AD_n reduction. For each elementary move in the reduction of τ, write the lifted AJ_n relation and verify that the new σ-word has image under ϕ equal to the updated τ-word, and that for every later generator the defining circular interval I_r is unchanged. Pay special attention to the nested case of Lemma 3.3(ii): after replacing σ_{i,j}σ_{k,l} (with [i,j]_c ⊂_c [k,l]_c) by σ_{k,l}σ_{s_{k,l}(j),s_{k,l}(i)}, check that the image under ϕ is the swapped τ-word and that all subsequent I_r values are preserved. If this verification succeeds for all cases, the missing induction can be completed; if any case fails, exhibit a concrete word with τ=1 but σ≠1, disproving injectivity.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is the injectivity of ϕ: AJ_n → AD_n ⋊ S_n (Theorem 3.4). The proof starts with σ in the kernel, forms the corresponding word τ in AD_n, and observes that τ=1. In the right-angled Coxeter group AD_n, any word equal to 1 can be reduced to the empty word by commuting adjacent letters and cancelling adjacent equal letters. Lemma 3.3 shows that each such elementary move on an adjacent pair in the τ-word corresponds to a legal move in AJ_n: equal intervals cancel, disjoint intervals commute, and nested intervals quasi-commute, with one generator replaced by its image under s. The gap is the next sentence: 'continuing this process, we deduce that σ=1.' After a quasi-commutation, the σ-word is not simply the same word with two letters swapped; one generator is replaced by a different generator. The proof never establishes an induction invariant showing that the lifted σ-word still corresponds position-by-position to the reduced τ-word and that the S_n-prefix products for all later letters are unchanged. Without this, the implication τ=1 ⇒ σ=1 is not formally justified. This is not a cosmetic omission: the entire chain of consequences—linearity, solvable word problem, trivial centre, torsion bounds—rests on this step. The concern is compounded by a concrete defect in the displayed formula for s_{k,l} in Section 1: for wrapping intervals such as n=4, [4,2]_c, the formula fails to give a permutation (it can output 5 and collide), so as written the morphism ϕ is not even defined on all generators. Both issues are likely repairable, but the proof as written is incomplete.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies affine cactus groups AJ_n, which arise as equivariant fundamental groups of real moduli spaces, from a combinatorial viewpoint. It claims that AJ_n is isomorphic to the generalized cactus group over an affine Coxeter group of type ~A_n (Theorem A / Theorem 2.7), that AJ_n embeds into a semidirect product AD_n \\rtimes S_n where AD_n is a newly introduced 'affine Gauss diagram group' (Theorem B / Theorem 3.4), and that this embedding yields linearity, solvability of the word problem, residual nilpotence of pure affine cactus groups, triviality of centres, and torsion bounds. The paper also proves injectivity of natural inclusions of subfamilies AJ_n^{p,q} and derives a semi-direct product decomposition. The arguments are presentation-based and rely on a word-reduction analysis of the embedding.","tokens_in":15493,"tokens_out":7424,"duration_ms":71700,"significance":"If the main embedding were established rigorously, the paper would provide a useful structural framework for affine cactus groups and would unify several known results about cactus groups, with concrete consequences such as linearity, solvable word problem, residually nilpotent pure subgroups, and torsion bounds. The paper also gives explicit conjectures and tools, especially the affine Gauss diagram group, that could be of independent interest. However, the current manuscript contains foundational presentation errors and an incomplete injectivity proof, so the significance of the announced results cannot yet be assessed. No machine-checked proofs or code are included, and the correctness rests entirely on the manuscript's arguments.","major_comments":[{"comment":"For n≥3 the presentation of W(~A_n) is internally inconsistent. The commuting relation ρ_iρ_j=ρ_jρ_i for |i-j|≥2 includes the pair (1,n), since |1-n|≥2 for n≥3, so ρ_1ρ_n=ρ_nρ_1. Substituting this into the wrap-around braid relation ρ_1ρ_nρ_1=ρ_nρ_1ρ_n gives ρ_1=ρ_n, which collapses the group. Thus W(~A_n) is not the affine Weyl group of type ~A_n as defined, and Theorem A cannot hold as stated. This needs to be repaired with a consistent affine-type presentation, and all later uses of W(~A_n) must be re-examined.","section":"Definition 2.1"},{"comment":"The proof of the isomorphism AJ_n ≃ C_{W(~A_n)} does not account for singleton parabolic subgroups. The set P^{ir,f}(S) contains {ρ_i} for each i, so the generalized cactus group has generators σ_{ρ_i} that have no counterpart among the generators σ_{i,j} of AJ_n. The step 'we set i=i_1, i+1=i_2, ..., j=i_k' fails for k=1, and the asserted bijection between generators is therefore not a bijection. The author must either prove that all σ_{ρ_i} are trivial or redundant in C_{W(~A_n)}, or explicitly restrict the family of parabolic subsets used to define the generalized cactus group; otherwise the claimed isomorphism is unsupported.","section":"Theorem 2.7"},{"comment":"The displayed formula for s_{k,l} is not a permutation for wrapping circular intervals. For example, when n=4 and [k,l]_c=[4,2]_c, the formula gives s_{4,2}(1)=5, which is not an element of {1,2,3,4}; hence the map is not a permutation of [1,n]. Since s_{k,l} is used in the defining relation (1.0.3) and in the morphism π of Proposition 3.1, this invalidates the presentation of AJ_n and the verification that π is well defined. A correct modular cyclic reversal formula must be supplied and the proof of Proposition 3.1 redone.","section":"Section 1, Eq. (1.0.3)"},{"comment":"The injectivity proof of ϕ is not complete. After observing that τ=1 and choosing a sequence of commutations and cancellations reducing τ to the empty word, Lemma 3.3 only justifies one adjacent move at a time. Following a quasi-commutation, the σ-word is not the original word with two letters interchanged: one generator is replaced by its image under a permutation, so the correspondence between letters of the σ-word and letters of the τ-word is disturbed, and later moves in the chosen τ-reduction need not lift to legal moves in the new σ-word. The sentence 'continuing this process, we deduce that σ=1' does not supply the required induction invariant. A formal induction on the reduction sequence, with an explicit invariant relating the current σ-word to the current τ-word under the S_n-action, is needed. Since Corollary 4.1, Corollary 3.6, and Theorems 4.10–4.16 all rely on this injectivity, this gap is load-bearing.","section":"Theorem 3.4, proof"}],"minor_comments":[{"comment":"The abstract contains the typo 'world problem'; it should read 'word problem'.","section":"Abstract"},{"comment":"Figure 3 appears to depict an n-vertex path, not the standard Coxeter diagram of affine type ~A_n, which is a cycle on n+1 vertices. If the author intentionally uses a non-standard presentation, this must be explained.","section":"Figure 3"},{"comment":"The notation [i,j]_c uses strict cyclic order, but the membership condition in the formula for s_{k,l} relies on i∈[k,l]_c without clarifying boundary conventions; a precise modular definition would resolve this.","section":"Section 1, notation"},{"comment":"The phrase 'subject to the relations of AJ_n involving only these elements' is ambiguous and should be stated as an explicit presentation, because it is not automatic that an arbitrary subset of relations defines a subgroup.","section":"Definition 4.2"},{"comment":"In the first sentence of the proof, the phrase 'let σ∈J_n' should presumably be 'let σ∈AJ_n'; otherwise the proof is inconsistent with the statement.","section":"Theorem 4.12 proof"}],"recommendation":"major_revision","confidential_remarks":"The paper has several load-bearing presentation errors: Definition 2.1 is internally inconsistent for n≥3, the formula for s_{k,l} fails for wrapping intervals, and the proof of the main embedding theorem is incomplete. These are substantive but appear to be repairable in principle, so I recommend major revision rather than rejection. The author should be asked to correct the affine Coxeter presentation, handle singleton parabolic subgroups explicitly, fix the permutation formula, and supply a complete induction argument for Theorem 3.4."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear X,\n\nThe paper has a worthwhile core: it embeds affine cactus groups into a semidirect product of a right-angled Coxeter group and S_n, and the intended consequences (linearity, solvable word problem, trivial centre, torsion bounds, residual nilpotence of the pure subgroup) all follow if that embedding works. That target is genuinely new and worth having.\n\nBut the text as it stands has problems that need fixing before the results can be trusted. First, Definition 2.1 of W(~A_n) is inconsistent for n≥3: the commuting rule |i-j|≥2 makes ρ1 and ρn commute, and then the braid relation ρ1ρnρ1=ρnρ1ρn forces ρ1=ρn. This looks like a typo — the commuting condition should be on cyclic distance — but as written it breaks Theorem A. Relatedly, Theorem A as stated cannot be right because the generalized cactus group includes generators for singleton parabolic subgroups, while the map from AJ_n only hits non-singleton ones; for n≥3 the full circular interval [1,n]_c corresponds to an infinite parabolic, so the generator σ_{1,n} has no image. Also, the formula for s_{k,l} in Section 1 is wrong for wrapping intervals (e.g. n=4, [4,2]_c), and the 'otherwise k' clause is clearly a typo for 'otherwise i'; as written the morphism to S_n is not a permutation.\n\nThe bigger issue is the proof of Theorem 3.4. The idea is right: reduce a word in AD_n that equals 1 to the empty word, and mirror each cancellation/commutation in AJ_n using Lemma 3.3. But after a quasi-commutation the σ-word is not the same word with two letters swapped; one generator is replaced by its image under a permutation. The proof says 'continuing this process' without giving an induction invariant that keeps the σ-word aligned with the reduced τ-word. Without that, the implication τ=1 ⇒ σ=1 is not formally justified. This is the load-bearing step for the whole paper, so it needs a rigorous treatment, not a hand-wave.\n\nThat said, I think the claims are likely true and repairable. The approach is sound, and the local analysis in Lemma 3.3 is mostly correct. A careful revision that fixes the presentation errors and supplies a proper induction for the injectivity proof would make this a solid paper.\n\nThe right outcome is to send it to a serious referee, with the expectation of major revision. I wouldn't cite it in its current form, but I'd be interested to see the corrected version.","headline":"A promising embedding theorem for affine cactus groups, but the current text has several presentation errors and the main injectivity proof is incomplete.","tokens_in":16048,"tokens_out":4295,"would_cite":false,"duration_ms":39674,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["20F55","20F36"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that affine cactus groups embed injectively into a semidirect product of a right-angled Coxeter group and a symmetric group, deriving linearity, a solvable word problem, a trivial centre, and a torsion bound of $2^{n-1}$.","keywords":["affine cactus groups","generalized cactus groups","Coxeter groups","right-angled Coxeter groups","circular intervals","affine Gauss diagram groups","torsion","word problem"],"falsifier":"Enumerate words in the generators $\\sigma_{i,j}$ for $n=4$ whose image under $\\phi$ is the identity in $AD_4\\rtimes S_4$, and test each against the defining relations of $AJ_4$; a single non-trivial word of this kind would disprove Theorem 3.4, while an exhaustive search up to length eight in $AJ_3$ would make the injectivity claim concrete.","tokens_in":14935,"feed_emoji":"🌵","tokens_out":12068,"duration_ms":94904,"temperature":0.7,"pith_summary":"Affine cactus groups are algebraic versions of cacti drawn on a cylinder: each generator flips the order of a block of marked points that winds once around the circle. This paper tries to show that these groups are well behaved, in particular that they embed injectively into the semidirect product of an affine Gauss diagram group $AD_n$ and the symmetric group $S_n$. If the embedding is correct, several structural questions are settled at once: the groups are linear, their word problem is decidable, their centre is trivial, and every torsion element has order at most $2^{n-1}$. The paper also identifies affine cactus groups with generalized cactus groups over the affine Coxeter group of type $\\tilde{A}_n$, which is what lets Coxeter group theory do the heavy lifting.","feed_headline":"Cactus groups on a circle embed into Coxeter-group products","feed_subtitle":"The embedding gives linearity, a solvable word problem, a trivial centre, and a torsion bound of $2^{n-1}$.","key_machinery":"The machinery is the combinatorial algebra of circularity: circular intervals $[i,j]_c$ and circular sets in $\\mathbb{Z}/n\\mathbb{Z}$, the affine Gauss diagram group $AD_n$ generated by $\\tau_I$ with relations $\\tau_I^2=1$ and $\\tau_I\\tau_J=\\tau_J\\tau_I$ whenever $I\\cap J=\\varnothing$ or one circular set is cyclically contained in the other, and the action of $S_n$ on these sets that produces the semidirect product $AD_n\\rtimes S_n$. The embedding theorem is carried by this action together with Lemma 3.3, which states that a cancellation or commutation visible in the $AD_n$-image of a word corresponds to a cancellation, commutation, or quasi-commutation of the original affine cactus letters. That correspondence is what allows the paper to read the word problem and torsion properties of $AJ_n$ off the right-angled Coxeter structure of $AD_n$.","core_discovery":"The central result is Theorem 3.4: for every $n\\ge 2$, the map $\\phi\\colon AJ_n \\to AD_n \\rtimes S_n$ that sends each generator $\\sigma_{i,j}$ to the pair $(\\tau_{[i,j]_c}, s_{i,j})$ is injective, where $AD_n$ is the affine Gauss diagram group generated by involutions $\\tau_I$ indexed by circular sets over $\\mathbb{Z}/n\\mathbb{Z}$, and $s_{i,j}$ reverses the circular interval $[i,j]_c$. The proof starts from the identification of $AJ_n$ with the generalized cactus group over the affine Coxeter group of type $\\tilde{A}_n$. It then shows that any word in $AJ_n$ whose image in $AD_n \\rtimes S_n$ is trivial must already be trivial in $AJ_n$, because cancellations and commutations in the right-angled Coxeter group $AD_n$ lift back to legal moves among the affine cactus generators.","pith_inferences":["The gap between the constructed elements of order $2k$ and the bound $2^{n-1}$ leaves the exact maximal torsion order open; computing the maximum for $n=4,5$ would likely show the bound is not sharp.","If the embedding is as strong as stated, the solvability of the word problem should extend to a normal form or rewriting system for $AJ_n$, since right-angled Coxeter groups admit geodesic normal forms.","The same circular-set construction may transfer to affine cactus groups over other affine Coxeter types, where the notion of circular interval would have to respect the different Coxeter diagram.","The injectivity theorems for the subfamilies $AJ^{p,q}_n$ suggest a filtration of $AJ_n$ by the number of strands involved in a crossing, which could support inductive proofs of further structural properties."],"forward_implications":["$AJ_n$ becomes a subgroup of a linear group, so affine cactus groups are linear for all $n\\ge2$.","The word problem in $AJ_n$ is solvable: a word is trivial exactly when its image in $AD_n$ reduces to the empty word by the standard right-angled Coxeter algorithm.","The centre of $AJ_n$ is trivial for $n\\ge2$, and the centre of the pure affine cactus group $PAJ_n$ is trivial for $n\\ge3$.","Torsion is tightly controlled: $PAJ_n$ is torsion-free, every torsion element of $AJ_n$ has even order bounded by $2^{n-1}$, and elements of order $2k$ exist whenever $2k\\le n$.","The usual cactus group $J_n$ embeds into $AJ_n$, and the pure affine cactus group $PAJ_n$ is residually nilpotent."],"supporting_citations":[{"why":"Introduces affine cactus groups and circular intervals, supplying the objects and the diagrammatic model.","marker":"[8]"},{"why":"Establishes linearity of generalized cactus groups, the property the paper recovers from its embedding.","marker":"[12]"},{"why":"Supplies the Coxeter-group facts used throughout, including parabolic subgroups, longest elements, and linearity.","marker":"[6]"},{"why":"Provides the embedding of usual cactus groups into a Gauss diagram group semidirect product, the template for Theorem 3.4.","marker":"[10]"},{"why":"Defines generalized cactus groups over Coxeter systems, the setting used in Theorem A to identify $AJ_n$.","marker":"[5]"}],"fun_headline_variants":["Affine cactus groups embed in Coxeter semi-direct products","Trivial centre and solvable word problem for affine cactus groups","Cactus groups on a circle: linear, centre-free, solvable words","Embedding affine cactus groups via circular interval reversals"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that every reduction of a word in $AD_n$ to the empty word can be mirrored step by step by cancellations and commutations in $AJ_n$, a step the proof asserts by saying 'continuing this process' without formalizing it.","fun_headline_variants_meta":{"raw":{"variants":["Affine cactus groups embed in Coxeter semi-direct products","Trivial centre and solvable word problem for affine cactus groups","Cactus groups on a circle: linear, centre-free, solvable words","Embedding affine cactus groups via circular interval reversals"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000243,"raw_usage":{"total_tokens":1478,"prompt_tokens":842,"completion_tokens":636,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":458,"completion_tokens_details":{"reasoning_tokens":565}},"tokens_in":458,"tokens_out":636,"duration_ms":5922,"temperature":1.0,"reasoning_tokens":565,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T13:37:32.695878+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Enumerate words in the generators $\\sigma_{i,j}$ for $n=4$ whose image under $\\phi$ is the identity in $AD_4\\rtimes S_4$, and test each against the defining relations of $AJ_4$; a single non-trivial word of this kind would disprove Theorem 3.4, while an exhaustive search up to length eight in $AJ_3$ would make the injectivity claim concrete.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes linearity of generalized cactus groups, the property the paper recovers from its embedding."},{"cited_title":"Bourbaki","cited_arxiv_id":null,"evidence_quote":"Supplies the Coxeter-group facts used throughout, including parabolic subgroups, longest elements, and linearity."},{"cited_title":"Mostovoy","cited_arxiv_id":null,"evidence_quote":"Provides the embedding of usual cactus groups into a Gauss diagram group semidirect product, the template for Theorem 3.4."},{"cited_title":"Bonnaf´ e","cited_arxiv_id":null,"evidence_quote":"Defines generalized cactus groups over Coxeter systems, the setting used in Theorem A to identify $AJ_n$."}],"review_version":1}