{"id":"b8b6b473-88f8-4c8d-8a21-4bc0c2045363","arxiv_id":"2505.17401","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The alternating sum of parabolic inductions and restrictions of an affine Hecke algebra module equals a module twisted by an explicit sign-and-parameter involution, with a finite-group analogue proved under restrictive assumptions.","lead":"The paper proves a duality formula for Hecke algebras with unequal parameters, extending Kato's equal-parameter result, and gives a relative version for finite Hecke algebras following Howlett-Lehrer. It also sketches a comparison with Aubert-Zelevinsky duality for p-adic groups.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Section 7 asserts compatibility with Aubert–Zelevinsky duality, but (7.3) is written down without a proof that it defines an involution or matches D_G under the Bernstein equivalence.","rationale":"The reader's verdict is CONDITIONAL with moderate confidence, and I agree that Theorems 3.2 and 5.2 are the substantive contributions while the abstract overstates what Section 7 proves. The reader's primary flagged assumption was Assumption 5.1 for Theorem 5.2; my main concern is different: Section 7 does not prove the claimed compatibility with Aubert–Zelevinsky duality at all, since (7.3) is asserted rather than derived as an involution. This is a missing proof in a central advertised claim, not a refutation of the main theorem. The reader did mention the unproved formula (7.3), so my agreement is partial rather than complete. Since the main theorems are not invalidated, the appropriate verdict remains CONDITIONAL: the paper should be accepted only after the Section 7 claim is either proved or explicitly downgraded to a conjecture/program. No fatal flaw was found in the proof of Theorem 3.2; the complex argument there is standard and the unequal-parameter generalization is credible.","tokens_in":35293,"tokens_out":28419,"duration_ms":241161,"concrete_test":"State Section 7 as a theorem with precise hypotheses and prove it: define the putative Hecke-side involution * on generators T_s and T_gamma of End_G(i_G^P Sigma), verify the quadratic and braid relations directly, and show that under the Bernstein equivalence E_G the functor D_H in (7.3) equals E_G o D_G o E_G^{-1} up to the factor (-1)^{|I0|}. In particular, track whether the equivalence from [Roc02] uses normalized parabolic induction: if it does, recompute (7.3) with the corresponding normalized induction on the Hecke side, and check whether the central-character shift cancels in the alternating sum. If the shifted identity fails or the braid relations fail for any unequal-parameter example, the compatibility claim in the abstract is unsupported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The abstract lists compatibility with Aubert–Zelevinsky duality as a proven result, but Section 7 contains no theorem statement. It derives (7.3) by formally translating the group-side expression D_G = sum (-1)^{|I|} i_G^{P_I} o r_G^{P_I} to the Hecke side. What is missing is the verification that (7.3) is the Grothendieck shadow of an algebra involution on End_G(i_G^P Sigma): no involution * is defined on this generalized affine Hecke algebra, no quadratic/braid relations are checked for the images of the generators, and no proof is given that D_H equals [M^*] for modules. The identification of these endomorphism algebras with generalized affine Hecke algebras is cited from [Sol22] without checking its hypotheses for the specific cuspidal pairs and unramified twists used here. There is also no tracking of normalized versus unnormalized induction: the p-adic functors in (6.1) are normalized, while the Hecke-side Ind/Res in (7.3) are written as unnormalized tensor/restriction functors; a mismatch of central characters would change the formula. These are not contradictions with Theorem 3.2 or Theorem 5.2, but the advertised compatibility claim is not established. The paper should either provide a complete proof of (7.3) with explicit hypotheses or remove the compatibility statement from the abstract.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs an involution * on extended affine Hecke algebras with unequal parameters by setting T*_w = (-1)^{ℓ(w_fin)} q(w) T^{-1}_{w^{-1}}, and proves that the Alvis-Curtis-type operator D[M] = Σ_{I⊆S} (-1)^{|I|} [Ind_I Res_I M] equals [M*] for finite-dimensional modules (Theorem 3.2). It also proves a relative version for the finite endomorphism algebras arising in Howlett-Lehrer theory under the assumption that the component group C(Λ) is trivial (Theorem 5.2), and it claims compatibility with Aubert-Zelevinsky duality for Bernstein blocks (Section 7). The paper includes a long appendix on homology representations and a detailed proof of Howlett-Lehrer's character identity.","tokens_in":35542,"tokens_out":5934,"duration_ms":44844,"significance":"Theorem 3.2 is a genuine extension of Kato's duality theorem to unequal parameters, and the proof is detailed enough to be checked; it also corrects an error in Kato's Lemma 2. Theorem 5.2 provides a finite-group-theoretic counterpart in the Howlett-Lehrer setting, although only under Assumption 5.1. The advertised compatibility with Aubert-Zelevinsky duality is not established in Section 7, so the paper's lasting contribution rests mainly on Theorems 3.2 and 5.2.","major_comments":[{"comment":"The compatibility with Aubert-Zelevinsky duality is asserted but not proved. Section 7 contains no theorem statement: formula (7.3) is introduced with 'the involution is' and there is no verification that it defines an involution on End_G(i_G^P Σ), no check that it matches D_G under the Bernstein equivalence, and no verification that the hypotheses of [Sol22] are satisfied for the specific cuspidal pairs and unramified twists used. Moreover, (6.1) defines D_G using normalized functors i_G^{P_I} and r_G^{P_I}, whereas (7.3) is written with unnormalized tensor/restriction functors Ind/Res inherited from the Roche diagrams; the paper does not discuss how normalization or central characters are matched. Since the abstract advertises this compatibility as a proved result, the author must either supply a complete proof with explicit hypotheses or remove the compatibility claim from the abstract.","section":"Section 7 and Abstract"},{"comment":"The proof of Theorem 5.2 is too compressed at two load-bearing points. The complex (5.7) is asserted to have cohomology only in one degree by analogy with [Sol66]/[DL82], but the boundary map d_i is described only informally, and the acyclicity argument is not adapted to the present indexing by WI\\C_{I0}(I)/W(Λ) with W(Λ)=R(Λ). The final sentence 'this is done using Lemma 3.5 for the finite case' does not supply the needed verification that χ_{I0}^K intertwines the whole EG(Λ)-action with the twisted action, especially for products T_w with w ∈ W(Λ). Please expand this proof or state explicitly which steps are being quoted from the cited sources.","section":"Section 5.1, proof of Theorem 5.2"},{"comment":"The relative theorem is proved only under Assumption 5.1, i.e. under the condition that the component group C(Λ) is trivial. This is stated in the body, but the introduction and abstract present the result as a general 'relative version for finite Hecke algebras' without emphasizing that the main Howlett-Lehrer cases with nontrivial C(Λ) are not covered. The paper should qualify the scope of Theorem 5.2 in the abstract and in Section 1.2.2, or prove the general case.","section":"Section 1.2.2 and Assumption 5.1"}],"minor_comments":[{"comment":"Corollary 6.4 says 'i_G^Q is equivalent to i_G^Q'; presumably one of the two occurrences should be i_G^{\\bar Q} (or a similar opposite parabolic), otherwise the statement is tautological.","section":"Section 6.2, Corollary 6.4"},{"comment":"In item (1) the text says 'The involution on the Hecke algebra side becomes (3.1) in the Section 3'; equation (3.1) is the definition of D[M], not the involution. The reference should be to Theorem 3.2 or equation (3.2)/(3.3).","section":"Section 7, item (1)"},{"comment":"There are numerous typographical issues: 'courterparts', 'The author also own a lot', inconsistent use of 'S-I. Kato' vs 'S.-I. Kato', and the line 'Ω˙ HpWaff , qsq– HpWpRq, qsq' where the dot should be a semidirect product symbol. A careful proofreading pass is needed.","section":"Throughout"},{"comment":"The notation E^1_I is used in Theorem 5.2 before being defined; the definition appears only in the theorem statement itself. It would help to define it in Section 5.1 or just before the theorem.","section":"Section 5, Theorem 5.2"}],"recommendation":"major_revision","confidential_remarks":"The paper contains a solid and checkable proof of Theorem 3.2, but the Section 7 claim is substantially over-advertised. I recommend asking the author to either prove the compatibility statement or remove it from the abstract. The manuscript would also benefit from a statement of Assumption 5.1 in the introduction's main theorem."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The real result here is Theorem 3.2: an unequal-parameter affine Hecke algebra involution matching the alternating sum of parabolic induction/restriction in the Grothendieck group. The proof is detailed, the correction to Kato's Lemma 2 is explicit and looks right, and the handling of the finite Weyl group component via Proposition 3.1 is careful. This is a genuine extension of Kato's work and a useful contribution to the Hecke algebra side of p-adic duality.\n\nTheorem 5.2, the relative finite Hecke algebra version, is also new, but it is honestly conditional on Assumption 5.1 (the component group C(Λ) being trivial). That restriction is stated clearly, and the theorem as stated is credible. The appendix proof of the Howlett–Lehrer character identity is a nice service to the literature.\n\nThe soft spot is Section 7 and the abstract. The abstract claims compatibility with Aubert–Zelevinsky duality as a proven result, but Section 7 contains no theorem statement. Formula (7.3) is written down by formal translation from the group side, but there is no verification that it defines an involution on the generalized affine Hecke algebra, no check of quadratic or braid relations, no proof that the Grothendieck shadow matches D_H, and no tracking of normalized versus unnormalized induction. The citation of [Sol22] for the endomorphism algebra identification is plausible but not checked against the specific hypotheses here. These are not contradictions with Theorems 3.2 or 5.2, but the advertised compatibility claim is not established. The author should either complete the proof or remove the compatibility claim from the abstract and frame Section 7 as a program.\n\nOverall, the citation pattern is sound, the main theorem appears solid, and the paper is clearly the work of someone who understands the machinery. It deserves serious refereeing, but the abstract and Section 7 need to be brought in line with what is actually proved before acceptance.","headline":"Solid unequal-parameter Hecke involution with a real proof; Section 7's Aubert–Zelevinsky compatibility is asserted, not proven, so the abstract overreaches.","tokens_in":36090,"tokens_out":1496,"would_cite":true,"duration_ms":18506,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["20C08","22E50"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that Alvis-Curtis duality for Hecke algebras is exactly an explicit module-level involution, with unequal parameters and in a relative finite setting.","keywords":["Hecke algebras","Alvis-Curtis duality","involution","unequal parameters","affine Hecke algebras","ramification groups","Aubert-Zelevinsky duality","Bernstein blocks"],"falsifier":"Compute $\\mathbb{D}[M]$ and $[M^*]$ for an irreducible module $M$ of a rank-two extended affine Hecke algebra in which two simple reflections are not conjugate and the parameters differ (for instance $q_s = 2$, $q_t = 3$). If the alternating sum of induced and restricted modules has a character different from the twisted module, the central claim fails; for the relative theorem, the analogous test is a cuspidal pair whose ramification group has nontrivial component group, where the equality (5.5) would show whether Assumption 5.1 is removable.","tokens_in":35059,"feed_emoji":"🔄","tokens_out":19478,"duration_ms":136730,"temperature":0.7,"pith_summary":"The paper aims to show that the alternating sum of inductions and restrictions used in Alvis-Curtis duality is realized on Hecke-algebra modules by a single explicit involution, even when the Hecke parameters are not equal. In the extended affine Hecke algebra setting, Theorem 3.2 states that for every finite-dimensional module $M$, $\\mathbb{D}[M] = \\sum_{I \\subseteq S} (-1)^{|I|} [\\operatorname{Ind}_I \\operatorname{Res}_I M]$ equals $[M^*]$, where the twist is $T_w^* = (-1)^{\\ell(w_{\\mathrm{fin}})} q(w) T_{w^{-1}}^{-1}$. The paper also proves a relative version for finite Hecke algebras attached to ramification groups, Theorem 5.2, under the assumption that the component group $C(\\Lambda)$ is trivial. A final section transfers the pattern to generalized affine Hecke algebras attached to Bernstein blocks, giving the Hecke-algebra counterpart of Aubert-Zelevinsky duality. If correct, the paper gives a uniform module-level explanation of a duality that was previously known mainly as a character identity.","feed_headline":"An explicit involution realizes Alvis-Curtis duality for Hecke modules","feed_subtitle":"The alternating sum of inductions and restrictions equals an explicit module twist, even with unequal parameters.","key_machinery":"The load-bearing object is the involution $*$ on the Hecke algebra, defined in the affine case on the Iwahori-Matsumoto generators by $T_w^* = (-1)^{\\ell(w_{\\mathrm{fin}})} q(w) T_{w^{-1}}^{-1}$, where $q(w)$ is the product of the Hecke parameters in a reduced expression of $w$ and $\\ell(w_{\\mathrm{fin}})$ is the length of the finite part of $w$. This map is an algebra anti-involution, and it is the module twist appearing on the right-hand side of $\\mathbb{D}[M] = [M^*]$. The proof mechanism that carries the argument is the truncation complex: the induced module $\\operatorname{Ind}_I \\operatorname{Res}_I M$ is written as a quotient of $H \\otimes_{H_I} M$ by images of endomorphisms $\\tau_s$, and the alternating sum of these complexes is arranged so that a spherical simplicial complex argument shows cohomology lives only in degree 0. The kernel is then identified with $M^*$ through the intertwining element $\\chi = \\sum_{w} (-1)^{\\ell(w)} T_w \\otimes T_w^{-1}$, using the three intertwining identities for finite simple reflections, the affine reflection, and the length-zero part $\\gamma \\in \\Omega$. In the relative theorem the same mechanism runs with the ramification group $W(\\Lambda) = R(\\Lambda)$ and the length function $\\ell^{I_0^K}$.","core_discovery":"The central discovery is that a duality defined by an alternating sum of parabolic induction and restriction functors is the same operation as twisting modules by an explicit algebra anti-involution. For the extended affine Hecke algebra $H(W(R), q_s)$ with possibly unequal parameters, the paper proves $\\mathbb{D}[M] = [M^*]$ for all finite-dimensional modules $M$, with $T_w^* = (-1)^{\\ell(w_{\\mathrm{fin}})} q(w) T_{w^{-1}}^{-1}$ on the Iwahori-Matsumoto basis. In the relative finite setting, the analogous statement holds for the endomorphism algebra $\\mathcal{E}_G(\\Lambda)$ of a Harish-Chandra induced cuspidal module, with the twist $T_w^* = (-1)^{|I_0|+\\ell^{I_0^K}(w)} p_w T_{w^{-1}}^{-1}$, provided the component group $C(\\Lambda)$ is trivial. The proof constructs an explicit chain complex whose only nonzero cohomology sits in degree 0, so the alternating sum collapses to a single graded term, and then proves by direct intertwining identities that this term is $M^*$. The manuscript further gives a formula, (7.3), for the counterpart of Aubert-Zelevinsky duality on generalized affine Hecke algebras attached to Bernstein blocks, reducing to the earlier theorem when the supercuspidal support is the split torus with trivial character.","pith_inferences":["Editorial extension: If the equality $\\mathbb{D}[M] = [M^*]$ is stable under specialization of parameters, the involution could be used to transport duality through families of Hecke algebras, giving a parameter-uniform statement beyond the current fixed-parameter theorem.","Editorial extension: The relative theorem's dependence on trivial $C(\\Lambda)$ suggests that a nontrivial component group will require either a twisted involution involving characters of $C(\\Lambda)$ or a sum over components; the semidirect decomposition $W(\\Lambda) = C(\\Lambda) \\rtimes R(\\Lambda)$ is the natural place to look for that correction.","Editorial extension: One testable consequence is that the involution should be compatible with the associated graded affine Hecke algebra at the graded level, so the identity $\\mathbb{D} = (\\cdot)^*$ should survive passage to the associated graded objects of the Bernstein block endomorphism algebras."],"forward_implications":["For any extended affine Hecke algebra with unequal parameters, the Alvis-Curtis style alternating sum can be replaced by the single explicit twist $M^*$, so computations of $\\mathbb{D}[M]$ reduce to inverting basis elements.","In the relative finite setting, the theorem gives a module-level involution on the endomorphism algebras of Harish-Chandra induced cuspidal modules whenever the ramification group has trivial component group, upgrading the known character identity for ramification groups to modules.","When the supercuspidal support is the split torus with trivial character, the Bernstein-block formula (7.3) reduces to Theorem 3.2, showing compatibility between the affine involution and Aubert-Zelevinsky duality in that case.","For real parameters the involution preserves unitarity of modules, so the duality respects the unitary Iwahori-spherical representations of split $p$-adic groups, a direct consequence of the unitarity theorem in Section 3.4 combined with the criterion for unitarity of Iwahori-Hecke modules."],"supporting_citations":[{"why":"Supplies the duality operator on Hecke algebra modules for equal parameters that this paper extends to unequal parameters and to the relative setting.","marker":"[Kat93]"},{"why":"Provides the truncation complex with cohomology concentrated in degree 0, which collapses the alternating sum to a single term.","marker":"[DL82]"},{"why":"Gives the character-level identity for Weyl groups that is the prototype of the module-level involution theorem.","marker":"[Sol66]"},{"why":"Supplies the Bernstein-Lusztig presentation of the affine Hecke algebra and the basis used in the proof of the involution.","marker":"[Lus89]"},{"why":"Gives the Iwahori-Matsumoto presentation and the decomposition theorem for $T_w$ that defines the involution.","marker":"[IM65]"},{"why":"Develops the ramification group $W(\\Lambda)$, its reflection subgroup $R(\\Lambda)$, and the length function used in the relative theorem.","marker":"[How80]"},{"why":"Proves the character-level duality for ramification groups that Theorem 5.2 lifts from characters to modules.","marker":"[HL82]"},{"why":"Establishes the Hecke algebra structure of the endomorphism algebra $\\mathcal{E}_G(\\Lambda)$ and its parabolic subalgebras used in the relative statement.","marker":"[HL83]"},{"why":"Introduces the Aubert-Zelevinsky duality on Grothendieck groups of $p$-adic groups, which Section 7 compares with the Hecke involution.","marker":"[Aub95]"},{"why":"Identifies endomorphism algebras of Bernstein blocks with generalized affine Hecke algebras, the identification on which formula (7.3) rests.","marker":"[Sol22]"}],"fun_headline_variants":["Hecke involution realizes Alvis-Curtis duality","Explicit twist equals alternating sum duality","Unequal parameter Hecke duality via involution","Relative Alvis-Curtis duality as module twist","Involution matches induction-restriction sum"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The relative theorem's load-bearing premise is Assumption 5.1: the component group $C(\\Lambda)$ in the semidirect decomposition $W(\\Lambda) = C(\\Lambda) \\rtimes R(\\Lambda)$ is trivial, so the ramification group is purely a reflection group and the parabolic subgroups of $W(\\Lambda)$ have the direct-product form used in the proof; if $C(\\Lambda)$ is nontrivial, the statement of Theorem 5.2 is not established.","fun_headline_variants_meta":{"raw":{"variants":["Hecke involution realizes Alvis-Curtis duality","Explicit twist equals alternating sum duality","Unequal parameter Hecke duality via involution","Relative Alvis-Curtis duality as module twist","Involution matches induction-restriction sum"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000277,"raw_usage":{"total_tokens":1675,"prompt_tokens":992,"completion_tokens":683,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":608,"completion_tokens_details":{"reasoning_tokens":613}},"tokens_in":608,"tokens_out":683,"duration_ms":6125,"temperature":1.0,"reasoning_tokens":613,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T14:47:30.773465+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute $\\mathbb{D}[M]$ and $[M^*]$ for an irreducible module $M$ of a rank-two extended affine Hecke algebra in which two simple reflections are not conjugate and the parameters differ (for instance $q_s = 2$, $q_t = 3$). If the alternating sum of induced and restricted modules has a character different from the twisted module, the central claim fails; for the relative theorem, the analogous test is a cuspidal pair whose ramification group has nontrivial component group, where the equality (5.5) would show whether Assumption 5.1 is removable.","supporting_citations":[],"review_version":1}