{"id":"a8f6babf-6a1d-421d-8646-753de67a86ab","arxiv_id":"2507.02588","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The author proves an equality of gerbe-twisted p-adic integrals on SLn and PGLn Higgs bundle moduli spaces for arbitrary rank and degree, generalizing the coprime-case result of Groechenig, Wyss, and Ziegler.","lead":"This mathematics paper proves an exact match between two large families of spaces built from a curve, in cases that earlier results did not cover. It matters because such a match is a concrete test of mirror symmetry, a prediction from string theory that different-looking spaces can describe the same physics.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Sign inconsistency in the Tate-pairing phase of Lemma 4.2: the displayed derivation uses the opposite argument order from Theorem 4.4; for the alternating local Tate pairing this changes the phase of the claimed identity.","rationale":"The reader's weakest-assumption analysis correctly identifies Proposition 3.3 as a delicate geometric point, and I do not dispute that assessment. However, when stress-testing the central claim Theorem 4.4, the most immediately checkable weak point is the Tate-pairing phase computation in §4.2. The paper contains an apparent inconsistency in the order of arguments of the Tate pairing: the derivation before Lemma 4.2 writes ⟨δ(L0), β(y)⟩_Γ, while Lemma 4.2(2), equation (4.4), and Theorem 4.4 all use ⟨β(y), δ(L0)⟩_Γ. For Γ = Pic^0(C)[n] with the standard Cartier self-duality, H^1(F, Γ) carries an alternating local Tate pairing, so the two orderings differ by a sign. If the displayed equation is authoritative, the phase factor in Theorem 4.4 has the wrong sign; if Lemma 4.2 and Theorem 4.4 are authoritative, the displayed equation must be corrected. Either way, the proof as printed does not resolve the discrepancy. This is not a manufactured concern: the phase factor is an essential part of the main theorem, distinguishing it from the coprime case and carrying the degree-shift information. The proposed check—comparing the two occurrences against the explicit Hilbert-symbol case and the maps in diagram (4.2)—would settle in a few lines whether the claimed equality is correct as stated or needs a sign correction. Because the fix is likely local and does not undermine the overall strategy, I would not reject the paper, but I would make acceptance conditional on reconciling the sign convention and correcting whichever occurrence is wrong.","tokens_in":30385,"tokens_out":34321,"duration_ms":410648,"concrete_test":"Fix a p-adic field F and take the model case Γ = μ_n, where H^1(F, μ_n) ≅ F^×/F^{×n} and the local Tate pairing is the Hilbert symbol (a,b)_n, which is alternating. Compute both ⟨δ(L0), β(y)⟩ and ⟨β(y), δ(L0)⟩ in this model, then trace the derivation of the displayed equation before Lemma 4.2 using the maps δ, β in diagram (4.2). If the pairing is alternating, one of the two lines has a sign error; the integrand in Theorem 4.4 must then be matched to the sign-corrected line. If the intended convention is symmetric, the paper should state this explicitly, since it contradicts the standard convention for the local Tate pairing on a self-dual finite group scheme.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central equality, Theorem 4.4, has on its right-hand side the phase factor exp(2πi ⟨β_Le(−), δ(L0)⟩_Γ). This is also the form stated in Lemma 4.2(2) and used in (4.4). However, the derivation of Lemma 4.2 is written with the opposite order: in the displayed equation after (4.3) the term appears as ⟨δ(L0), β(y)⟩_Γ. For Γ = Pic^0(C)[n] with the self-duality coming from the principal polarization, the local Tate pairing on H^1(F, Γ) is alternating: ⟨a,b⟩ = −⟨b,a⟩. If this convention is in force, the two expressions differ by a sign, and the phase factor in Theorem 4.4 is not the one obtained from (4.1)–(4.3), unless the displayed line before Lemma 4.2 is a typo and (4.3) is the intended substitution. The paper never states its pairing convention explicitly, so the reader cannot determine from the text which occurrence is authoritative. This is load-bearing because a sign error here changes the equality of p-adic integrals themselves, not merely a corollary or an interpretation step. The issue is independent of the delicacy of Proposition 3.3: even granting Proposition 3.3, the Tate-pairing computation converting the SYZ isomorphism into a gerbe-function equality must be sign-consistent.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves a non-archimedean topological mirror symmetry statement for moduli spaces of SL_n and PGL_n Higgs bundles on a curve, allowing arbitrary degree and thus going beyond the coprime case treated by Groechenig–Wyss–Ziegler. The main result, Theorem 4.4, is an equality of p-adic integrals over the integral points of the good moduli spaces, with a phase factor given by a local Tate pairing between the class of the SL_n determinant line bundle and a torsor class on the PGL_n side. The proof goes through two SYZ-type symmetries (Propositions 3.1 and 3.3), conversion of these symmetries into identities of gerbe functions via Tate duality, and then a fibrewise p-adic integration argument following the strategy of [GWZ20] and [COW24]. The paper also derives the coprime-case stringy twisted E-polynomial equality, relates the p-adic integrals to intersection cohomology for prime n and deg D > 2g-2, and proposes a conjecture linking them to BPS cohomology.","tokens_in":30572,"tokens_out":10969,"duration_ms":116096,"significance":"If the main theorem is correct, it is a substantial advance: it removes the coprimality assumption that was central to previous non-archimedean proofs of the Hausel–Thaddeus conjecture, and it gives a clean conjectural framework for the non-coprime case in both the classical and meromorphic settings. The paper is largely self-contained and does not fit parameters to data; the central equality is derived internally from the SYZ-type symmetries and p-adic integration. The refinements and cohomological consequences, especially the correction of a parity condition in earlier work, are potentially significant. However, the proof contains a sign inconsistency in the Tate-pairing phase and a key torsor isomorphism whose proof is only sketched, both of which need to be resolved before the main claim can be taken as established.","major_comments":[{"comment":"The derivation of Lemma 4.2(2) has a sign inconsistency that propagates to Theorem 4.4. The displayed computation after (4.3) uses the term ⟨δ(L0), β(y)⟩_Γ, whereas (4.3) and Lemma 4.2(2) state ⟨β(y), δ(L0)⟩_Γ. For the self-dual finite group Γ = Pic^0(C)[n], the local Tate pairing on H^1(F,Γ) is skew-symmetric, so the two expressions differ by a sign; the paper never states its pairing convention. Since Theorem 4.4's phase factor exp(2πi⟨β_{Le}(−), δ(L0)⟩_Γ) is taken from Lemma 4.2(2), the displayed proof does not justify the stated equality unless one of the occurrences is corrected. Please state the convention and make the order consistent throughout, then re-check the sign in the fibrewise argument of Theorem 4.4.","section":"§4.2, Lemma 4.2 and the display after (4.3)"},{"comment":"The proof of the second SYZ-type symmetry is too compressed for a result on which the main theorem depends. In particular, the construction of the Γ_N-equivariant structure on L'_{L1}, the identification N' = N via the normalization of the pulled-back Poincaré sheaf at Nm^{-1}(N'), and the final conclusion by torsor rigidity are asserted rather than proved. Since (4.1) uses Proposition 3.3 to identify the SL_n Hitchin fibre with a product of Prym-torsors, the proof should be expanded or a precise reference should be given for each of these steps.","section":"§3.2, Proposition 3.3"},{"comment":"The verification of the [COW24] measure conditions is summarized rather than fully checked. The codimension computation for the strata S_{k,n1,n2} is quoted without derivation, and the existence of O_F-point lifts is only referenced to [Bay+21, Lemma 21.22] and [AHH23, Theorem A8]. Because the canonical measures µ_can are used in the statement of Theorem 4.4, I would like a fuller verification or an explicit statement of which theorem in [COW24] applies verbatim to the SL_n and PGL_n stacks.","section":"§4.1, Lemma 4.1"}],"minor_comments":[{"comment":"In the case where both fibres have a rational point, the sentence \"the integrands can be expressed in terms of Tate pairing against the dual fibre, exp(2πi⟨−, h^{-1}_{SLn}(a)(F)⟩) (resp. exp(2πi⟨−, h^{-1}_{SLn}(a)(F)⟩))\" repeats h^{-1}_{SLn} twice; the second occurrence should presumably be h^{-1}_{PGLn}(a).","section":"§4.3, proof of Theorem 4.4"},{"comment":"The paper should state explicitly, at the point where the local Tate pairing ⟨−,−⟩_Γ is first used, whether it is defined to be alternating and in which argument order it is taken; this is directly connected to the sign issue raised in the major comments.","section":"§4.2, first use of Tate pairing"},{"comment":"There is a typo in the abstract: \"Thesep-adic volumes\" should read \"These p-adic volumes\".","section":"Abstract"},{"comment":"The description of the Γ-action on the non-rigidified stack is dense; a short example or a diagram displaying the 1-commutativity condition would improve readability.","section":"§2.4.1, Lemma 2.11"}],"recommendation":"major_revision","confidential_remarks":"The paper is a serious and valuable contribution, and I expect the main ideas to be correct. However, the sign inconsistency in the Tate-pairing phase is a genuine obstacle to verifying the central equality, and the proof of Proposition 3.3 is too condensed for a step that the main theorem depends on. I would ask the authors to correct the sign convention and expand the relevant arguments before the paper is accepted. The reader's report's ACCEPT recommendation appears too optimistic given the internal inconsistency."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Briefly: this paper extends the non-archimedean topological mirror symmetry of GWZ20 to arbitrary rank and degree, dropping the coprimality assumption. That is a real result. It also proves the two SYZ-type symmetries over arbitrary fields and catches a parity error in MS21/MS22, which is a useful service.\n\nThe proof follows the established GWZ20/COW24 strategy. The main theorem (4.4) is a fiberwise p-adic integral equality, reducing to Tate pairings on Prym torsors. The two SYZ symmetries (Props 3.1, 3.3) are proven in the text, and the paper is transparent about where it stands: the p-adic-to-intersection-cohomology dictionary is only established for n prime and even D > K, and the BPS link is explicitly conjectural. No fitted parameters, no circularity.\n\nThe soft spots are in proportion. The sign issue flagged by the stress-test is real and needs to be addressed: the displayed equation after (4.3) has ⟨δ(L0), β(y)⟩ whereas Lemma 4.2(2) and Theorem 4.4 use ⟨β(y), δ(L0)⟩. The paper never states the Tate pairing convention. For the standard alternating pairing on H^1(F, Γ), these differ by a sign. It looks like a typo in the display, and the lemma/theorem are consistent with each other, but the author must state the convention and fix the line. A referee should not let this slide.\n\nProposition 3.3 is the other delicate point. The Γ_N-equivariant universal bundle construction is dense, and the compatibility of Γ_N with the Abel-Jacobi normalization is not fully spelled out. A referee should ask for more detail there. Lemma 4.1 (measure conditions) is summarized from COW24; probably fine, but it should be checked.\n\nOverall: this is a serious contribution to a real program. The central argument holds up; the sign issue is a minor-to-moderate revision, not a fatal flaw. This deserves a serious referee. I'd bring it to a reading group and would cite it if I worked in this area.\n\nRecommendation: send to peer review. Require the author to fix the sign convention and expand the proof of Prop 3.3.","headline":"A genuine extension of non-archimedean mirror symmetry to arbitrary degree, with a sign-convention typo that needs fixing but doesn't sink the central idea.","tokens_in":31248,"tokens_out":6310,"would_cite":true,"duration_ms":63419,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14D20","14H60","14F20","14G10","11S80"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that gerbe-twisted p-adic integrals on moduli spaces of $SL_n$ and $PGL_n$ Higgs bundles agree for arbitrary degree, removing the coprime assumption from earlier proofs.","keywords":["Higgs bundles","topological mirror symmetry","p-adic integration","gerbes","Hitchin fibration","Prym varieties","intersection cohomology","SYZ duality"],"falsifier":"Compute the two p-adic integrals of Theorem 4.4 explicitly for a non-coprime example, for instance genus 2 with $n=2$, $D=K$, and even degree, by counting $O_F$-points on the two moduli spaces over a p-adic field and comparing with the predicted gerbe-twisted volumes; a mismatch, or a direct check that the torsor isomorphism in Proposition 3.3 fails for a spectral curve with a non-trivial $N$-class, would falsify the claim.","tokens_in":30062,"feed_emoji":"🪞","tokens_out":11080,"duration_ms":107108,"temperature":0.7,"pith_summary":"This paper establishes non-archimedean topological mirror symmetry for $SL_n$ and $PGL_n$ Higgs bundles in arbitrary degree, removing the coprime restriction that appeared in previous proofs. It shows that for $D = K_C$ or $\\deg D > 2g-2$, the gerbe-twisted p-adic integral over the $SL_n$ moduli space equals the gerbe-twisted p-adic integral over the $PGL_n$ moduli space, up to an explicit Tate-pairing phase depending on the determinant line bundle. The result completes the coprime-case topological mirror symmetry for odd-degree twists and, in the meromorphic case, relates the $SL_n$ integral to intersection cohomology traces. If correct, the coprime assumption is not intrinsic to mirror symmetry for Higgs bundles, and the p-adic integrals become a common tool linking the coprime, non-coprime, and conjectural BPS-cohomology formulations.","feed_headline":"Mirror symmetry for Higgs bundles beyond the coprime case","feed_subtitle":"Gerbe-twisted p-adic volumes on the SL_n and PGL_n sides agree for arbitrary rank and degree.","key_machinery":"The key mechanism is the torsor of trivialisations $\\mathrm{Triv}^0$ of a $\\mu_n$-gerbe, upgraded to a $G_m$-gerbe of $n$-torsion, which converts a gerbe on one Hitchin fibre into a Prym torsor on the dual fibre. On the $PGL_n$ side the gerbe $\\alpha_N$ is $\\Gamma$-equivariant, and its trivialisations are controlled by the relative group scheme $\\Gamma_N$ over $C$ built from the Poincar\\'e line bundle and the Abel-Jacobi map. These identifications let the author express gerbe functions through Tate pairings, and the p-adic measure is constructed from translation-invariant volume forms pulled back along the self-dual isogeny $\\mathrm{Prym}^0 \\to \\mathrm{Prym}^0/\\Gamma$, whose cotangent-volume ratio is controlled by [GWZ20, Lemma 6.15].","core_discovery":"The central claim is Theorem 4.4, an equality of p-adic integrals over $O_F$-points of the good moduli spaces. For any degrees $d,e$ and determinant line bundle $L$, the integral of the gerbe function $f^{e'}_\\alpha$ on the $SL_n$ side equals the integral of $f^{d'}_{\\alpha_N}$ times $\\exp(2\\pi i \\langle \\beta_{L_e}(-), \\delta(L_0)\\rangle_\\Gamma)$ on the $PGL_n$ side, with $d' = d + \\deg D$, $e' = e + \\deg D$, $L_0 = L \\otimes D \\otimes N^{-d'}$, and $D = \\bigotimes_{i=0}^{n-1} O_C(-iD)$. The proof reduces to the smooth Hitchin fibres, where the $SL_n$ fibre is a product of Prym torsors dual to the $PGL_n$ fibre, via the two SYZ-type symmetries: $\\mathrm{Triv}^0(\\alpha^{e'} \\mid h^{-1}_{SL_n}(a)) \\cong h^{-1}_{PGL_n}(a)$ and $\\mathrm{Triv}^0(\\alpha_N^{d'} \\mid h^{-1}_{PGL_n}(a)) \\cong \\mathrm{Prym}^{N^{d'}}$. Gerbe functions become Tate pairings on the fibres, and the canonical p-adic volumes of the dual abelian varieties $\\mathrm{Prym}^0$ and $\\mathrm{Prym}^0/\\Gamma$ are equal.","pith_inferences":["A plausible next step, not taken in the paper, is to extend the equality to all effective divisors $D$ by treating the non-reduced spectral locus, since the fibrewise argument covers the smooth locus and the remaining strata have measure zero.","The parity correction to the coprime case suggests that endoscopic decompositions for odd-degree meromorphic twists should be revisited wherever an equivariant local system is normalized by a basepoint.","The BPS conjecture can be tested directly by computing the left-hand integral for a small-rank example where the BPS cohomology is known, since the formula predicts the Frobenius trace on the untwisted part."],"forward_implications":["Coprime topological mirror symmetry holds for odd-degree $D$ with the corrected isotypical character twist, so the earlier parity assumption was not needed.","For meromorphic $D > K$ and odd prime $n$, the gerbe-twisted $SL_n$ p-adic integral counts intersection cohomology of the moduli space, extending p-adic integration to singular non-coprime spaces.","The equality holds for arbitrary degrees, so any future BPS sheaf should reproduce these integrals as Frobenius traces, as formulated in Conjecture 5.9.","The refinement in Theorem 4.7 gives a Fourier-type relation between $SL_n$ integrals over degree-shifted determinants and single $PGL_n$ integrals over a fixed specialization locus, providing finer isotypic information."],"supporting_citations":[{"why":"introduces the dual Hitchin system and the SYZ-type mirror symmetry framework that this paper generalises.","marker":"[HT03]"},{"why":"establishes the p-adic integration approach and proves the coprime case that Theorem 4.4 extends to arbitrary degree.","marker":"[GWZ20]"},{"why":"supplies the canonical p-adic measure on moduli stacks of Higgs bundles used to define the integrals.","marker":"[COW24]"},{"why":"provides the intersection-cohomology mirror symmetry in the meromorphic non-coprime case used for the cohomological interpretation.","marker":"[MS22]"},{"why":"supplies the $\\chi$-independence isomorphisms for $GL_n$-Higgs bundles used in the intersection-cohomology step.","marker":"[MS23]"},{"why":"provides the Tate duality pairings that express gerbe functions as pairings in the fibrewise computation.","marker":"[Mil06]"}],"fun_headline_variants":["Mirror symmetry for Higgs bundles now includes all degrees","p-adic volumes bridge SL_n and PGL_n in non-coprime cases","Gerbe-twisted p-adic integrals confirm mirror symmetry","Prym torsors and p-adic volumes prove mirror duality","Non-archimedean mirror symmetry extends to any degree"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is the second SYZ-type symmetry (Proposition 3.3): the torsor of $\\Gamma$-equivariant trivialisations of the gerbe $\\alpha_N$ on a $PGL_n$ Hitchin fibre is isomorphic to the $\\mathrm{Prym}^{N^d}$-torsor, a statement that depends on the compatibility of the group scheme $\\Gamma_N$ with the Abel-Jacobi normalisation of the Poincar\\'e bundle; if this isomorphism failed, the product formula for the $SL_n$ fibre and the Tate-pairing computation would break.","fun_headline_variants_meta":{"raw":{"variants":["Mirror symmetry for Higgs bundles now includes all degrees","p-adic volumes bridge SL_n and PGL_n in non-coprime cases","Gerbe-twisted p-adic integrals confirm mirror symmetry","Prym torsors and p-adic volumes prove mirror duality","Non-archimedean mirror symmetry extends to any degree"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000921,"raw_usage":{"total_tokens":4004,"prompt_tokens":1052,"completion_tokens":2952,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":668,"completion_tokens_details":{"reasoning_tokens":2863}},"tokens_in":668,"tokens_out":2952,"duration_ms":25351,"temperature":1.0,"reasoning_tokens":2863,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T20:26:35.262484+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the two p-adic integrals of Theorem 4.4 explicitly for a non-coprime example, for instance genus 2 with $n=2$, $D=K$, and even degree, by counting $O_F$-points on the two moduli spaces over a p-adic field and comparing with the predicted gerbe-twisted volumes; a mismatch, or a direct check that the torsor isomorphism in Proposition 3.3 fails for a spectral curve with a non-trivial $N$-class, would falsify the claim.","supporting_citations":[],"review_version":1}