{"id":"a7eba2e0-8d39-4e82-a7ed-c6050a2f74f8","arxiv_id":"2411.15321","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A reducible representation is Anosov precisely when the eigenvalue gaps of its irreducible block factors grow at least linearly according to a unique large eigenvalue configuration.","lead":"This paper characterizes exactly which reducible representations of a non-elementary hyperbolic group are Anosov, using the eigenvalue magnitudes of the irreducible blocks after block diagonalization. It further shows that the Anosov deformations of any block normalization form a bounded convex domain, and that Anosov-only components of character varieties cannot contain reducible representations.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified; the characterization in Theorem 3.4 and the convexity/boundedness results are internally sound.","rationale":"I read the paper in good faith and tried to break the central claims. Theorem 3.4's two directions are consistent: the 'only if' uses non-elementarity to freeze Q, and the 'if' derives proximality from the block gaps. I attempted to construct an unbounded ray in A^U_1(zeta) by putting the maximum weight on the support of Q; the homomorphism property forced the inequality to reverse on inverses, so the ray cannot remain Anosov, confirming boundedness. I found no incorrect statement that affects Theorem A, B, C, or Corollary D. The minor issue with full-block flag components is purely notational. The reader's concern about non-elementarity is accurate but already assumed in the paper and tested by the paper's own counterexample. Verdict unchanged.","tokens_in":30420,"tokens_out":35319,"duration_ms":334349,"concrete_test":"Verify the boundedness step in Theorem C for a nonzero ray psi in hom(Gamma, D_U): take the witnessing pair (i,j,gamma) with psi_i(gamma) != psi_j(gamma); if the inequality has the wrong sign, apply it to gamma^{-1}, where psi(gamma^{-1}) = -psi(gamma), and confirm the sign flips so that the proof's construction of a t where lambda_k <= lambda_{k+1} goes through.","verdict_should_be":"UNCHANGED","load_bearing_attack":"After reviewing the proof of Theorem 3.4 and the deformation-theoretic consequences, I do not find a load-bearing gap. The reader's highlighted hypothesis, non-elementarity of Gamma, is genuinely necessary: the density of the orbit of an attracting fixed point in the Gromov boundary (used in Section 3.2 to make the large eigenvalue configuration Q independent of the chosen hyperbolic element) fails for elementary groups, and the paper explicitly notes the Z counterexample. Since Gamma is assumed non-elementary, the step is valid. I also probed the boundedness proof of Theorem C for a possible counterexample: if the unique block with q_i,k > 0 carried the maximal weight of a ray psi, one might worry the ray stays Anosov. The apparent issue dissolves because psi is a homomorphism to R; if psi_i(gamma) > psi_j(gamma) for the witnessing gamma, then at gamma^{-1} the inequality reverses, giving the required psi_i < psi_j and the closing of the eigenvalue gap for large t. The only minor technical blemish is in Theorem 3.4(3), where xi^{q_i,k}_{eta_i}(z) is used for q_i,k = dim_K(U_i), although the P_{theta_i}-Anosov limit map is only defined for q in Delta_{U_i}; this is harmless and can be fixed by declaring the full-block flag component to be the whole space U_i. No central claim is affected.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies when reducible representations of a non-elementary word hyperbolic group into GL(V) are Anosov. The main result, Theorem A (proved as Theorem 3.4), characterizes P_theta-Anosov block upper triangular representations in terms of the eigenvalue magnitudes of the blocks of their block diagonalization, via a uniquely defined admissible family Q = (q_{j,k}) of dimensions describing how the top k eigenvalues are distributed among the blocks. The paper also proves that the space of block deformations yielding Anosov representations is a bounded convex open set (Theorem C), gives a concrete description of this space when the block normalization is itself Anosov (Theorem B), and derives a corollary about connected components of character varieties containing reducible representations (Corollary D). The proofs are detailed and largely self-contained, building on a generalized-eigenvector analysis of block upper triangular matrices (Proposition 3.1) and the Kassel--Potrie eigenvalue-gap characterization of Anosov representations.","tokens_in":30688,"tokens_out":47573,"duration_ms":426449,"significance":"If the characterization holds, it provides a complete and explicit criterion for Anosov reducible representations, reducing the problem to exponential growth rates of eigenvalue gaps between blocks. This is a natural and useful generalization of the author's earlier work on reducible suspensions, and the convexity/boundedness theorems give a clean deformation-theoretic picture. The paper is careful with the non-elementary hypothesis, including a counterexample showing that the uniqueness of the configuration Q fails for Z. The proofs are thorough and, apart from local presentation issues, the central claims are well supported.","major_comments":[],"minor_comments":[{"comment":"Lemma 4.2 states that [Gamma,Gamma] has finite index if and only if hom(Gamma,D_U)={0}. This is false as stated for the trivial decomposition U=(V), for which D_U={0} while [Gamma,Gamma] may have infinite index (e.g., a free group). The proof already uses dim(D_U)>0, so the statement should assume U is non-trivial or at least that dim(D_U)>0. The false direction is not used in Corollary D, so this is a local issue.","section":"Lemma 4.2"},{"comment":"In the formula for the limit map xi^theta_eta(z), the notation xi^{q_{i,k}}_{eta_i}(z) is used for q_{i,k}=dim_K(U_i), but the P_{theta_i}-Anosov limit map is only defined for indices in Delta_{U_i}. This is harmless if one adopts the convention that the full-block component is the whole space U_i, but the convention should be stated explicitly.","section":"Theorem 3.4(3)"},{"comment":"In the proof, the indexing condition for the maximum is written as 'q_{j,k}<dim_K(U'_j cap A^+_k)', which should be 'q_{j,k}<dim_K(U_j)'. The displayed formula in the statement of Corollary 3.3(3) already uses the correct condition, so this is a typo in the proof.","section":"Corollary 3.3, proof of (3)"},{"comment":"The line 'This contradiction implies that psi_i(gamma)<psi_j(gamma) for some j and gamma' is logically too quick. The contradiction only rules out psi_i=psi_j for all j and all gamma. To obtain the strict inequality, one uses that psi_i and psi_j are homomorphisms: if psi_i(gamma)>psi_j(gamma), then at gamma^{-1} the inequality reverses. Adding this observation makes the argument complete.","section":"Theorem C, boundedness proof"},{"comment":"There are several minor typographical errors, including 'representions' in the abstract and an 'upslope' artifact in the proof of Lemma 4.2. These should be corrected in a final pass.","section":"Throughout"}],"recommendation":"minor_revision","confidential_remarks":"The paper is a solid contribution to the Anosov representation literature, and the reader's positive assessment is justified. The central theorems are sound; the issues I found are local and fixable. The overstatement in Lemma 4.2 should be corrected before publication, and the small proof gaps (limit map convention, the psi_i<psi_j step) should be patched. None of these affect the main characterization or the deformation-theoretic corollaries."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know this is a good paper. The main result (Theorem A/3.4) reduces the Anosov condition for block diagonal representations to an eigenvalue-gap condition between blocks, governed by a unique combinatorial object Q, the large eigenvalue configuration. This is genuinely new: it generalizes the author's earlier q-Fuchsian suspension result to all reducible representations and does not reduce to prior work. The proof is long but coherent: a generalized-eigenvector analysis of block upper triangular maps, followed by a clean eigenvalue-gap argument for the converse. I looked at the non-elementarity hypothesis. It is really needed, because the density of attracting fixed point orbits in the Gromov boundary fails for elementary groups, and the paper explicitly gives the Z counterexample. So that is not a flaw.\n\nThe deformation-theoretic part is also solid. Theorem C says the Anosov deformation domains are convex and bounded open subsets; the boundedness proof has a potential pitfall with rays, but it works because if a homomorphism to R is positive on one block at a group element, it is negative on that block at the inverse. Corollary D, about Anosov-only components avoiding reducibles when the commutator has infinite index, is a natural and correct consequence. The paper is careful about the character variety quotient and about eigenvalue magnitude continuity. Citation practice is fine: the author cites their previous work for context, but the central theorems are proved from standard background (KP22, GW12, Gue+17). The only blemish I found is minor: in Theorem 3.4(3), the notation xi^{q_i,k}_{eta_i}(z) is used when q_i,k = dim_K(U_i), just outside the usual range of the P_{theta_i}-Anosov limit map. It is harmless and can be fixed by declaring that flag component to be the whole space. There are a few typos and some implicit standard facts about hyperbolic groups, but nothing that affects correctness.\n\nThis paper is for anyone working on Anosov representations, higher Teichmuller theory, or character varieties of hyperbolic groups. It deserves a serious referee. I would accept it with minor revisions, asking for the technical fix in 3.4(3) and a remark on the finite-order case in the eigenvalue gap formula. Verdict: accept.","headline":"A solid, self-contained characterization of Anosov representations of block diagonal form, with correct deformation-theoretic consequences; the main theorems are new and the proofs hold up.","tokens_in":31221,"tokens_out":1930,"would_cite":true,"duration_ms":20375,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["20F67","22E40"],"pacs":[],"model":"deepseek-v4-flash","headline":"A block-diagonal representation is Anosov exactly when there is a unique integer table Q of block eigenvalue counts that makes the cross-block eigenvalue ratios grow linearly.","keywords":["Anosov representations","reducible representations","eigenvalue gaps","block diagonal","character variety","hyperbolic groups","large eigenvalue configuration","limit maps"],"falsifier":"Take a non-elementary word hyperbolic group such as a free group F_2, a direct sum V = U_1 oplus U_2, and a block diagonal P_theta-Anosov representation with both blocks irreducible; compute the limit flag at two different boundary points and check whether dim(U_j cap xi_k(z)) is constant in z and equals the unique admissible integer q_{j,k}. If the dimension varies with z, or if two different admissible families both satisfy the linear-growth condition for all infinite-order elements, then Theorem A is false.","tokens_in":30200,"feed_emoji":"🧩","tokens_out":5139,"duration_ms":47055,"temperature":0.7,"pith_summary":"This paper gives a complete characterization of when a reducible representation of a non-elementary word hyperbolic group is Anosov, in terms of the eigenvalue magnitudes of its irreducible blocks. It shows that a block diagonal representation is P_theta-Anosov precisely when one can assign, to each block and each flag level, a nonnegative integer q_{j,k} so that the q_{j,k}-th eigenvalue of one block compared with the (q_{j,k}+1)-th eigenvalue of another block grows at least linearly along the group. This reduces the Anosov property to a check of finitely many eigenvalue ratios, and it determines the Anosov limit map completely from the integer configuration and the block-level limit maps. As a consequence, the deformation spaces of reducible Anosov representations are convex, bounded, open sets, and for many hyperbolic groups, connected components of the character variety that consist entirely of Anosov representations cannot contain reducible ones.","feed_headline":"One integer table decides which reducible reps are Anosov","feed_subtitle":"Block-diagonal representations are Anosov exactly when cross-block eigenvalue ratios grow linearly under a unique configuration Q.","key_machinery":"The central object is the large eigenvalue $\\theta$-configuration Q = (q_{j,k}), a family of integers indexed by the block number j and the flag level k, with 0 <= q_{j,k} <= dim_K(U_j) and sum_j q_{j,k} = k. It records how many of the k largest eigenvalue magnitudes of the full representation are contributed by the j-th block. The load-bearing identity is the eigenvalue-gap formula of Corollary 3.3: log( lambda_k(rho(gamma)) / lambda_{k+1}(rho(gamma)) ) = min_{i,j} log( lambda_{q_{i,k}}(eta_i(gamma)) / lambda_{q_{j,k}+1}(eta_j(gamma)) ), which converts the Anosov linear-growth condition into cross-block eigenvalue ratios. The structured-flag spaces W_{Q}^{U,$\\theta$}(V) and S_{Q}^{U,$\\theta$}(V) then encode exactly which flags can occur as limit flags, namely those meeting each block in dimension q_{j,k}.","core_discovery":"The central claim is Theorem 3.4 (and its block-diagonal specialization, Theorem A): if a representation rho is block upper triangular relative to a direct sum decomposition U, with block diagonalization eta = oplus_{j=1}^m eta_j, then rho is P_theta-Anosov if and only if there is a unique (U,$\\theta$)-admissible family Q = (q_{j,k}) of integers 0 <= q_{j,k} <= dim_K(U_j) such that log( lambda_{q_{i,k}}(eta_i(gamma)) / lambda_{q_{j,k}+1}(eta_j(gamma)) ) grows at least linearly in translation length for all admissible i,j,k. In that case Q is the large eigenvalue $\\theta$-configuration: for every boundary point z, q_{j,k} = dim_K( U_j cap xi^k_eta(z) ), each block eta_j is P_{theta_j}-Anosov with limit map obtained by intersecting the full limit flag with U_j, and the full limit flag decomposes as the direct sum of the contributing block limit flags. This reduces the Anosov condition for reducible representations to eigenvalue gaps of the blocks, and it shows that the discrete configuration Q is a genuine invariant of the representation.","pith_inferences":["The uniqueness of Q suggests a natural stratification of the reducible Anosov locus by the discrete data Q, which could lead to a cell decomposition of the corresponding character variety components; the paper does not develop this.","The Z counterexample shows the non-elementary hypothesis is not merely technical: for elementary groups the large-eigenvalue configuration can depend on the chosen hyperbolic element, so extending the result would require a different invariant.","Theorem B's supremum over infinitely many group elements may be computable by testing only primitive elements in the torsion-free abelianization, as the paper notes, and the paper's Question 4.3 asks whether the deformation domains are finite-sided polytopes; this could be tested on explicit free-group or surface-group examples.","One could in principle use the characterization as an algorithm: enumerate admissible Q, check the cross-block eigenvalue ratios on a generating set, and thereby decide whether a given reducible representation is Anosov for hyperbolic groups with manageable geometry."],"forward_implications":["For any reducible representation, checking the Anosov property is reduced to checking linear growth of finitely many cross-block eigenvalue ratios, once the integer table Q is known.","The Anosov limit map of a block diagonal representation is explicitly computable from the block limit maps via the direct-sum formula xi_eta(z) = ( oplus_{q_{i,k}>0} xi^{q_{i,k}}_{eta_i}(z) )_k, so the configuration Q determines the entire asymptotic flag.","The space A^U_theta(zeta) of block deformations that preserve the P_theta-Anosov property is always convex and bounded (Theorem C), giving a constrained deformation theory for reducible Anosov representations.","If the commutator subgroup [Gamma,Gamma] has infinite index in Gamma, then any connected component of the character variety consisting entirely of Anosov representations contains no reducible representation (Corollary D).","Each block eta_j with theta_j nonempty is itself P_{theta_j}-Anosov, so the Anosov property propagates down the block decomposition to the irreducible factors."],"supporting_citations":[{"why":"Supplies the eigenvalue-gap characterization of Anosov representations (Corollary 4.6) that the paper adapts to block upper triangular representations.","marker":"[KP22]"},{"why":"Gives the prior reducible-suspension framework whose convex deformation domains are generalized here to all reducible representations.","marker":"[Lah24]"},{"why":"Provides the stability theorem (their Theorem 1.2) for openness of Anosov representations, used in proving Theorem C.","marker":"[GW12]"},{"why":"Originally introduced Anosov representations in PSL(2,K) and higher rank, the notion the paper extends.","marker":"[Lab06]"},{"why":"Proved that a representation is Anosov if and only if its semisimplification is, which underlies reducing to the block diagonal case.","marker":"[Gué+17]"},{"why":"Introduced linear u-deformations that the block deformation parameterization generalizes.","marker":"[Bar10]"}],"fun_headline_variants":["Integer table Q decides Anosov for reducible reps","Reducible reps Anosov iff unique integer table Q exists","Eigenvalue gaps and Q: the Anosov test for reducibles","Block diagonal? Check Q-table for Anosov property"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument needs the group's Gromov boundary to be rich enough that the orbit of any hyperbolic element's attracting point is dense; for elementary groups like Z the configuration Q can depend on the element, so the characterization fails without the non-elementary hypothesis.","fun_headline_variants_meta":{"raw":{"variants":["Integer table Q decides Anosov for reducible reps","Reducible reps Anosov iff unique integer table Q exists","Eigenvalue gaps and Q: the Anosov test for reducibles","Block diagonal? Check Q-table for Anosov property"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000242,"raw_usage":{"total_tokens":1488,"prompt_tokens":868,"completion_tokens":620,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":484,"completion_tokens_details":{"reasoning_tokens":545}},"tokens_in":484,"tokens_out":620,"duration_ms":5698,"temperature":1.0,"reasoning_tokens":545,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T14:29:19.197884+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a non-elementary word hyperbolic group such as a free group F_2, a direct sum V = U_1 oplus U_2, and a block diagonal P_theta-Anosov representation with both blocks irreducible; compute the limit flag at two different boundary points and check whether dim(U_j cap xi_k(z)) is constant in z and equals the unique admissible integer q_{j,k}. If the dimension varies with z, or if two different admissible families both satisfy the linear-growth condition for all infinite-order elements, then Theorem A is false.","supporting_citations":[],"review_version":1}