{"id":"07bb6d98-e645-4b83-a8c9-fcb2b2a514b0","arxiv_id":"2411.17469","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The unipotent block of mod l representations of p-adic GL2, for odd l dividing q+1, is equivalent to perfect complexes over a dg Schur algebra.","lead":"For p-adic GL2, the paper shows that the unipotent block of mod l representations, in the non-banal case where l divides q+1, is equivalent to perfect complexes over a differential graded Schur algebra. The result gives a new algebraic handle on a piece of the p-adic Langlands program that had resisted a derived-category description.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 5.5's lifting step, the backbone of Theorem 6.10, rests on the unverified hand computation in Lemmas 4.14–4.15 and on a compressed p-adic convolution argument; an undetected error there would invalidate Corollary 5.6.","rationale":"I agree with the reader's weakest-assumption analysis. The main theorem has a clear architecture: Q generates via finite global dimension of SR(n) and nilpotence of I (Theorems 3.14, 6.8), and then V generates via the finite-block comparison (Lemmas 4.13, 6.9) and the lifting Corollary 5.6. The only step that is both genuinely load-bearing and not machine-checked is the lifting lemma. If Lemma 5.5 fails, Q need not be isomorphic to I^G Qf, and the chain ⟨Q⟩ = ⟨I^G Qf⟩ = ⟨V⟩ breaks, so Theorem 6.10 and Corollary 6.11 fall. The finite computation in Lemmas 4.14–4.15 is long and the notation is dense; a single counting error in the conjugate-class parametrizations or coefficient cancellations would invalidate it. The p-adic part of Lemma 5.5 also compresses the argument that elements of I+, I0, and I- 'commute with Z'; this is not true in H(G), only on the module P after moving through the tensor product H(G) ⊗_{H(K)} Pf, and the reduction to affine Weyl group elements deserves a fuller justification. Because the paper is otherwise well-structured and the cited external results are appropriate, the right disposition is to require independent verification of this calculation before full acceptance. The concern is about verifiability of a dense hand computation, not about the author's competence.","tokens_in":23817,"tokens_out":37535,"duration_ms":318711,"concrete_test":"Verify Lemma 4.14/4.15 computationally for small q: for q=2 (l=3), q=4 (l=5), q=5 (l=3), in F_l[GL_2(F_q)] form Z_{q-1}=-Σ_{C1}g+Σ_{C2}g+(q-1)1, factor it as Σ μ_i c_i x_i z_i + Σ ν_j d_j w_0 y_j, and check Σ μ_i c_i=0, Σ ν_j d_j=0, (e-e1)Z_{q-1}P1=0, and e2Z_{q-1}=γ. Then lift to the Iwahori–Hecke algebra of GL_2(F_q((t))) and, using the Bernstein presentation, check (1-e1)Z 1_{wI}=0 for all w in a finite set of affine Weyl group elements (diag(̟^k,̟^l) and w_0 diag(̟^k,̟^l) with -3≤k,l≤3). A single failure localizes the error in Lemma 5.5.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim D^b(H1(G)_fg) = <V> and the dg equivalence Corollary 6.11 both pass through Corollary 5.6, which asserts Q ≅ I^G_{Gf,K}Qf. Corollary 5.6 is exactly the equality IΓ = I^G_{Gf,K}(If Γ f), and its nontrivial inclusion is Lemma 5.5: H(G) ⊗_{H(K)} If ⊆ I. Lemma 5.5 in turn depends on the explicit finite computation in Lemma 4.14, where Z_{q-1} = -Σ_{C1}g + Σ_{C2}g + (q-1)1 is factorized as Σ μ_i c_i x_i z_i + Σ ν_j d_j w_0 y_j with cancellations Σ μ_i c_i = 0 and Σ ν_j d_j = 0, and on Lemma 4.15 identifying γ = e2 Z_{q-1}. This is a long hand calculation over a finite field of characteristic l dividing q+1; it is not machine-checked, and any sign or counting error in the parametrizations of C1 and C2 or in the multiplicities of the sums over a' and b would break the cancellation and hence Lemma 5.5. Moreover, the p-adic part of Lemma 5.5 is terse: the proof asserts that elements of I+, I0 and I- 'commute with Z' because Z is central in Gf; the lifted Z ∈ H(G) is not central in H(G), and the reduction g = iw requires an argument that iZ and Zi agree on the specific vector 1_{wI} via the tensor product relation H(G) ⊗_{H(K)} Pf, not in H(G). This step is correct for w = 1 but is less immediate for general affine w, and the case analysis k≤l/k>l is load-bearing. Since no independent verification is provided and the exposition is compressed, the main theorem rests on this unverified calculation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves, for G=GL_2(F) with F a p-adic field of residual cardinality q and with l an odd prime dividing q+1, that the bounded derived category D^b(B_1(G)_fg) of finitely generated unipotent smooth R-representations is classically generated by an explicitly defined object V, and that it is triangulated equivalent to per(dg-End(V^\\bullet)), where V^\\bullet is a projective resolution of V. V is the parahoric induction of the direct sum of the trivial representation and the projective module P_1 from the finite reductive quotient G_f. The proof combines Vignéras's progenerator Q and Schur-algebra description of the unipotent block, a finite global-dimension result for the Schur algebra obtained from affine cellularity, the explicit block structure of GL_2(F_q) in the given characteristic as described by Ackermann and Paige, and a lifting argument that identifies Q with the parahoric induction of Q_f. The core new input is the explicit description of the annihilator ideal I_f and the verification that its lifting annihilates the p-adic module P, which is carried out by a long finite-field computation.","tokens_in":24223,"tokens_out":45461,"duration_ms":400591,"significance":"If the proof is correct, the result is a substantial step in the modular derived representation theory of p-adic groups: it gives a complete dg-algebra description of the unipotent block in a non-banal case, with a dg algebra whose zeroth cohomology is the Schur algebra S_R(2). The chosen generator V is simple enough that the dg endomorphism algebra may admit an explicit presentation, and the author indicates a plausible route to generalizing the method to cases where the finite unipotent block has cyclic defect groups. The paper is careful in citing external results, and the finite-block calculations in Section 4 are detailed. I have verified the counting and cancellations in Lemma 4.14; they are correct. The main concern is the exposition of the lifting lemma (Lemma 5.5), which is load-bearing for the main theorem and is currently too compressed, with one literally false statement about commutativity.","major_comments":[{"comment":"The proof of this lemma is load-bearing for Corollary 5.6 and Theorem 6.10, but as written it contains a false assertion: the sentence \"I0 commutes with W as the Weyl group normalises the diagonal elements of G\" is not true; for w0=(0 1;1 0) and i=diag(a,d) in I0 one has i w0 ≠ w0 i. The subsequent equality iZ1_{wI}=Z1_{wiI}=Z1_{wI} therefore needs a different justification. One must prove that i w I = w I for every i∈I0 and every w in the extended affine Weyl group, which is true because w^{-1} I0 w ⊆ I, but this requires an explicit check for diagonal translations and for the two families w=diag(π^k,π^l) and w=w0 diag(π^k,π^l). Similarly, the collapse of z_i and x_i in the case k>l uses the inclusions \\bar I+ w ⊆ w I- and I0 w ⊆ wI; these are true, but they should be shown by the conjugation computations w^{-1} z w ∈ I- and w^{-1} x w ∈ I0. Please rewrite this proof with these steps made explicit; the current compressed argument is not sufficient for the central lifting claim.","section":"§5, Lemma 5.5"},{"comment":"In the proof of Lemma 4.13, the element Y is simplified to -Σ_{C1}g + Σ_{C2}g, while Z_λ is defined as Σ_{C1}g + Σ_{C2}g + λ·1. The equality e2Z_λ = e2(Y+λ1) is not automatic: it uses that e2(Σ_{C1}g) acts as zero on P2, which follows from the displayed character-table values (the component of Σ_{C1}g on both π0 and π_i is 0). This observation should be stated explicitly, because it is needed to justify the form Z_{q-1} used in Lemma 4.15 and then in Lemma 5.5. I have checked the parametrizations of C1 and C2 and the two cancellations Σ μ_i c_i = 0 and Σ ν_j d_j = 0 in Lemma 4.14; they are correct, but the exposition would benefit from displaying the multiplicities more clearly.","section":"§4, Lemmas 4.13–4.15"}],"minor_comments":[{"comment":"The sentence following the short exact sequence swaps the summand attributions: the exact sequence 0 → rad(P1) → P1 → /BD → 0 has P1 and /BD as the direct summands of Vf and P1 and rad(P1) as the direct summands of Qf, not as stated. The intended meaning is clear from the sequence itself, but the wording should be corrected.","section":"Corollary 4.12"},{"comment":"In the proof, the expression \"IfP2 = cP2 = S′\" uses an undefined symbol c; it should be \"IfP2 = γP2 = S′\".","section":"Lemma 4.11"},{"comment":"The introduction refers to the finite-global-dimension result as \"Theorem 3.15\", but in the text the theorem is numbered Theorem 3.14. Please correct the cross-reference.","section":"Abstract and Introduction"},{"comment":"There are several typos that should be fixed, including \"Furtherore\" in Lemma 5.5, \"together wtih\" in Corollary 5.6, and \"every every\" in the proof of Lemma 6.6.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper is competent and the central result is significant. The main risk is the compressed proof of Lemma 5.5, which is load-bearing and contains a false commutativity statement; after my own verification the argument is repairable, so I recommend major revision rather than rejection. The author should also clarify the sign issue in Lemma 4.13 and fix the minor typographical and cross-reference errors."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a real step forward for the non-banal unipotent block of p-adic GL2. The author proves that for n=2, l odd dividing q+1, Db(B1(G)_fg) is classically generated by V (and by Q), hence equivalent to per(dg-End(V•)) with H^0 = SR(2). She also proves SR(n) has finite global dimension for all n and l ≠ p, extending Deng–Yang's q=1 case. Both are new, and the first gives the kind of derived-level description that has been missing in the non-banal setting.\n\nThe paper is well structured. The strategy is clear: use Vignéras's progenerator, reduce to a finite block problem, use Ackermann's cyclic-defect block description, then lift via an explicit annihilator computation. Citations are precise, and the finite group part is detailed enough to check. I read the main chain of arguments and did not find a gap in the logic.\n\nThe soft spot is exactly where the author says it is. Lemma 5.5, which gives the inclusion H(G) ⊗_{H(K)} I_f ⊆ I, depends on Lemma 4.14: a factorization of Z_{q−1} into c x z and d w0 y terms with coefficient cancellations Σ μ_i c_i = 0 and Σ ν_j d_j = 0. That is a long hand computation over F_l, not machine-checked, and a sign or counting error would break the cancellation. The stress-test note about the p-adic part of Lemma 5.5 also lands: the proof asserts that i commutes with the lifted Z because Z is central in Gf, but in H(G) ⊗_{H(K)} H(Gf) that needs more care for general affine w. The case analysis k≤l/k>l is explicit and I think it is correct, but it is compressed. This is load-bearing, so I would want an independent check or a clearer write-up before treating Theorem 6.10 as fully settled.\n\nThere is a minor typo in Corollary 4.12 swapping the summands of Vf and Qf; the intended statement is clear from Lemma 4.11 and Definition 4.8, and it does not affect Lemma 6.9. No circular reasoning that I can see.\n\nThis paper is for specialists in p-adic representation theory and Schur algebras. It deserves peer review, with a referee who can check both the finite block structure and the Hecke algebra side, with specific attention to Lemma 5.5 and the computation in Lemma 4.14. If that checks out, accept.","headline":"Genuine new result in a well-trodden program; the main theorem is credible but rests on a long hand computation and a compressed lifting argument that need independent checking before I'd rely on them.","tokens_in":24815,"tokens_out":5340,"would_cite":true,"duration_ms":47207,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["22E50","18E30","16G10","20C08"],"pacs":[],"model":"deepseek-v4-flash","headline":"For odd $l$ dividing $q+1$, the derived unipotent block of $p$-adic $\\mathrm{GL}_2$ is classically generated by one explicit parahoric induction $V$, and is triangulated equivalent to perfect complexes over the dg endomorphism algebra of…","keywords":["derived categories","unipotent block","p-adic GL(2)","dg algebra","perfect complexes","Schur algebra","parahoric induction","affine cellular algebras"],"falsifier":"Compute in the group algebra $R[\\mathrm{GL}_2(\\mathbb{F}_2)]$ with $R$ of characteristic $3$: form $Z_{1} = -\\sum_{g \\in C_1} g + \\sum_{g \\in C_2} g + 1$, factor each element of the two conjugacy classes into the normal forms of Lemma 4.14, and check that both coefficient sums vanish; the factorization is the exact assertion on which Lemma 5.5 and hence Theorem 6.10 depend.","tokens_in":23575,"feed_emoji":"🧮","tokens_out":12564,"duration_ms":94008,"temperature":0.7,"pith_summary":"The paper proves that, for the $p$-adic group $G=\\mathrm{GL}_2(F)$ in the non-banal case where an odd prime $l$ divides $q+1$, the bounded derived category $D^b(\\mathcal{B}_1(G)_{fg})$ of finitely generated unipotent representations is generated by a single explicit object $V$, the parahoric induction from the finite quotient of the direct sum of the trivial representation and the projective indecomposable module $P_1$. It then shows that this derived category is triangulated equivalent to the category of perfect complexes over a differential graded algebra, namely the dg endomorphism algebra of a projective resolution of $V$, whose zeroth cohomology is the Schur algebra $S_R(2)$. Because $V$ is explicit, the dg algebra is a tractable object, and an explicit description of it would give an explicit description of the entire derived unipotent block. The path to the equivalence runs through a long, explicit calculation in the global Hecke algebra that lifts the relation between two finite-block generators to the $p$-adic setting.","feed_headline":"One dg algebra captures the derived unipotent block of p-adic GL2","feed_subtitle":"With odd l dividing q+1, the entire derived unipotent block is perfect complexes over one explicit dg algebra.","key_machinery":"The load-bearing mechanism is the lifting step Lemma 5.5, which proves the inclusion $\\mathcal{H}(G) \\otimes_{\\mathcal{H}(K)} I_f \\subseteq I$, where $I_f$ is the annihilator in the finite group algebra of $V_f = \\mathbf{1} \\oplus P_1$ and $I$ is the corresponding $p$-adic ideal. The proof rests on Lemma 4.14, an explicit factorization of the central element $Z_{q-1} = -\\sum_{g \\in C_1} g + \\sum_{g \\in C_2} g + (q-1)\\mathbf{1}$ in the finite group algebra into products of unipotent, diagonal, and opposite-unipotent factors with coefficient sums that cancel exactly; combined with Lemma 4.15 identifying a generator of the annihilator as $\\gamma = e_2 Z_{q-1}$, this yields $Q = I^G_{G_f,K}(Q_f)$. Exactness of parahoric induction then transports the finite short exact sequence $0 \\to \\mathrm{rad}(P_1) \\to P_1 \\to \\mathbf{1} \\to 0$ to the $p$-adic setting, showing that $V$ and $Q$ classically generate the same subcategory.","core_discovery":"The paper establishes that, for $n=2$ and odd $l$ dividing $q+1$ (and not dividing $q$ or $q-1$), the bounded derived category $D^b(\\mathcal{H}_1(G)_{fg})$ is classically generated by the object $V = I^G_{G_f,K}(\\mathbf{1} \\oplus P_1)$ obtained by parahoric induction from the finite quotient $G_f = \\mathrm{GL}_2(k)$. It also proves that the earlier progenerator $Q$ of the subcategory annihilated by the ideal $I$ lies in the subcategory classically generated by $V$, so the two generate the same derived category. With $V^\\bullet$ a projective resolution of $V$ in $\\mathrm{Mod}(G)$, the main corollary is a triangulated equivalence $D^b(\\mathcal{H}_1(G)_{fg}) \\simeq \\mathrm{per}(\\mathrm{dg\\text{-}End}(V^\\bullet))$, where the dg endomorphism algebra has zeroth cohomology $S_R(2)$. Thus the entire derived unipotent block is encoded as perfect complexes over one explicit dg algebra.","pith_inferences":["The finite calculation of Lemma 4.14 can in principle be checked by computer algebra for the smallest residual cardinalities (for example $q=2$, $l=3$), giving an independent test of the lifting step without repeating the hand computation.","The same scheme should work whenever the finite unipotent block has cyclic defect group, since the explicit Loewy structure used here is available in that generality; the paper notes this expectation, and the missing ingredient would be an analogous explicit factorization.","A natural next step would be to compute $\\mathrm{dg\\text{-}End}(V^\\bullet)$ explicitly in the smallest case and compare its perfect complexes with independently known objects in $D^b(\\mathcal{B}_1(G)_{fg})$, providing a direct test of the equivalence."],"forward_implications":["The classical generator $Q$ and the simpler object $V$ generate the same derived subcategory, so $D^b(\\mathcal{H}_1(G)_{fg})$ is classically generated by a single parahoric induction from the finite quotient.","The triangulated equivalence $D^b(\\mathcal{H}_1(G)_{fg}) \\simeq \\mathrm{per}(\\mathrm{dg\\text{-}End}(V^\\bullet))$ reduces the study of the derived unipotent block to perfect complexes over one explicit dg algebra.","The zeroth cohomology of $\\mathrm{dg\\text{-}End}(V^\\bullet)$ is $S_R(2)$, so the known module-level description of the subcategory annihilated by $I$ is refined to a derived-level description of the whole block.","Because $V$ is defined from a single projective indecomposable module of $\\mathrm{GL}_2(k)$, the dg algebra is small enough that an explicit presentation can be sought, which would give a concrete description of all objects of the derived category."],"supporting_citations":[{"why":"Defines the progenerator Q, the ideal I, and the module-level equivalence of the annihilated subcategory with modules over the Schur algebra S_R(n); the whole paper extends this picture to the derived level.","marker":"[Vig03]"},{"why":"Supplies Noetherianity of smooth representations, used to pass from finite generation to finite projective resolutions and to control the nilpotent filtration by I.","marker":"[Dat09]"},{"why":"Provides the affine cellular structure of the affine Schur algebra that the paper extends to prove finite global dimension of S_R(n).","marker":"[DY16]"},{"why":"Gives the Loewy structure of the projective indecomposable modules of the finite unipotent block used to describe Q_f, V_f, and the ideal I_f.","marker":"[Ack06]"},{"why":"Provides explicit formulae for endomorphisms of the projective P_2, used to identify gamma = e_2 Z_{q-1} in Lemma 4.13.","marker":"[Pai14]"},{"why":"Supplies the general criterion (Theorem 3.8(b)) that converts a classical generator into a triangulated equivalence with perfect complexes over its dg endomorphism algebra.","marker":"[Kel06]"},{"why":"Provides the definition and finiteness theorems for (idempotent) affine cellular algebras, the tool behind Theorem 3.15.","marker":"[KX12]"}],"fun_headline_variants":["dg algebra encodes derived unipotent block of p-adic GL2","Unipotent block of p-adic GL2 is perfect complexes over dg algebra","p-adic GL2 unipotent block: perfect complexes over a dg algebra"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The main theorem depends on a long hand calculation in a finite group algebra: the central element $Z_{q-1}$ must factor with two coefficient sums that cancel exactly, and any mistake in that calculation would break the lifting step and the dg equivalence.","fun_headline_variants_meta":{"raw":{"variants":["dg algebra encodes derived unipotent block of p-adic GL2","Unipotent block of p-adic GL2 is perfect complexes over dg algebra","p-adic GL2 unipotent block: perfect complexes over a dg algebra"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001216,"raw_usage":{"total_tokens":5002,"prompt_tokens":945,"completion_tokens":4057,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":561,"completion_tokens_details":{"reasoning_tokens":3992}},"tokens_in":561,"tokens_out":4057,"duration_ms":23777,"temperature":1.0,"reasoning_tokens":3992,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T12:06:09.217622+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute in the group algebra $R[\\mathrm{GL}_2(\\mathbb{F}_2)]$ with $R$ of characteristic $3$: form $Z_{1} = -\\sum_{g \\in C_1} g + \\sum_{g \\in C_2} g + 1$, factor each element of the two conjugacy classes into the normal forms of Lemma 4.14, and check that both coefficient sums vanish; the factorization is the exact assertion on which Lemma 5.5 and hence Theorem 6.10 depend.","supporting_citations":[],"review_version":1}