{"id":"73206d16-0065-4cca-86c4-b105547e070b","arxiv_id":"2505.04492","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For adjoint Shimura varieties in the superrigid regime, the canonical overconvergent F-isocrystal has Frobenius conjugacy classes with rational, ℓ-independent characteristic polynomials matching the canonical ℓ-adic local systems.","lead":"This paper proves that on certain Shimura varieties, the canonical p-adic Frobenius structure (an F-isocrystal) has the same rational Frobenius conjugacy classes as the canonical ℓ-adic local systems for every prime ℓ. The result completes a compatibility statement for non-abelian, superrigid Shimura varieties.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 5.6's p-adic special-point compatibility rests on an explicit but unproved crystalline-period calculation; if the formula is wrong, Theorem 4.5's conclusion no longer follows.","rationale":"The reader's weakest-assumption analysis correctly points to the imported crystallinity result PST+24, Theorem 7.1 and to Lemma 5.1. Those are genuine preconditions, and I do not dispute them. However, the most load-bearing gap visible inside the manuscript is the self-flagged omission in Lemma 5.6: the p-adic Frobenius computation at special points is asserted rather than proved. This is not a mere exposition gap because the compatibility at special points is exactly what converts the outer automorphism from Lemma 5.5(1) into the inner automorphism needed for Theorem 4.5. The reader's verdict of CONDITIONAL is therefore appropriate and should remain. My concern supplements rather than replaces the reader's list of conditions: both the imported p-adic Hodge theory and the internal crystalline-period calculation should be independently verified before the theorem is regarded as fully established.","tokens_in":23016,"tokens_out":9826,"duration_ms":103151,"concrete_test":"Write out the proof of Lemma 5.6's displayed Frobenius formula. For a special point s and a finite unramified extension F_v = E(s_{K0})_v, compute D_cris(xi composed with rho_p,s) directly: decompose xi_K composed with inn(a_p)(rho_p,s|Gamma_F_v) into crystalline characters, calculate the action of the linearized Frobenius on each rank-one D_cris using Fontaine's periods and Lubin-Tate theory for unramified representations, and verify that phi^{[kappa(v):F_p]} has characteristic polynomial xi(q^{-1}_{Frob_v}); in particular, check the normalization rec_v(variation_pi_v) = geometric Frobenius against [Con11, Proposition B.4]. If the formula is confirmed, the step stands; if not, identify precisely where Lemma 5.5(2) fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In Section 5.3, Theorem 4.5 is reduced by Lemma 5.5 to checking that the companion rho_pi_rightsquigarrow_lambda,v and the canonical rho_lambda,v agree at special points. For lambda different from p this is supplied by KP24, but on the p-adic side Lemma 5.6 is the only place where the Frobenius of the F-isocrystal at a special point is computed. There the authors assert that a 'calculation ultimately rests on Lubin-Tate theory ([Con11, Proposition B.4]), to which strictly speaking we should add the corresponding calculation of crystalline periods for unramified representations' yields the formula phi^{[kappa(v):F_p]} = direct sum of psi_i(rec_v(variation_pi_v)) chi_i(variation_pi_v)^{-1}, and that this has characteristic polynomial xi(q^{-1}_{Frob_v}). No derivation or reference for the added crystalline-period calculation is given. Because the equality of semisimple classes at special points is the input that upgrades the outer automorphism tau from Lemma 5.5(1) to an inner one, a sign or reciprocity error here would invalidate the conclusion of Theorem 4.5, not just a side remark. The paper flags the omission honestly, but it remains an unproved load-bearing step.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves a compatibility result between canonical ℓ-adic local systems and canonical p-adic F-isocrystals on adjoint Shimura varieties in the superrigid regime (Assumption 4.1). The main theorem (Theorem 4.5) asserts that, after enlarging the level, the canonical overconvergent G-F-isocrystal E†_v constructed in Theorem 4.4 has a λ-adic companion that is conjugate to the canonical Gad(Qℓ)-local system. The proof combines crystallinity of canonical p-adic local systems (from PST+24 and EG25), a Tannakian argument to upgrade convergence to overconvergence (Lemma 5.1), Margulis superrigidity (Lemma 5.5), and a local calculation at special points (Lemma 5.6). The paper also contains a proof of Drinfeld's comparison theorem for pro-semisimple fundamental groups in arbitrary dimension (Section 3.1).","tokens_in":23226,"tokens_out":14571,"duration_ms":137476,"significance":"If the missing calculation in Lemma 5.6 is supplied, the result is a significant advance: it extends the ℓ≠p compatibility of KP24 to the p-adic realization for non-abelian type Shimura varieties, giving a step toward Kottwitz triples and the Langlands–Rapoport conjecture. The structural arguments are mostly clean and carefully presented: the Tannakian diagram in Section 5.1, the use of full faithfulness of overconvergent versus convergent F-isocrystals, and the reduction to special points are well organized. The paper also provides a higher-dimensional form of Drinfeld's theorem (Remark 3.10) and is explicit about the scope of its assumptions and the loss of primes in Remark 4.6, which is commendable.","major_comments":[{"comment":"The proof of Lemma 5.6 contains an admitted gap in the p-adic calculation, and that calculation is load-bearing for Theorem 4.5. In the proof, after decomposing ξ_K∘inn(a_p)(ρ_{p,s}|_{Γ_{E(s_{K0})_v}}) into crystalline characters ψ_i, the authors assert that the linearized Frobenius ϕ^{[κ(v):F_p]} acts on D_cris by ⊕_i ψ_i(rec_v(ϖ_v)) χ_i(ϖ_v)^{-1}, and that this has the same characteristic polynomial as ξ_K(q^{-1}_{Frob_v}). The text states only that this follows from 'a calculation that ultimately rests on Lubin–Tate theory ([Con11, Proposition B.4]), to which strictly speaking we should add the corresponding calculation of crystalline periods for unramified representations.' No derivation or reference for the crystalline-period part is supplied. This formula is the only input that upgrades the outer automorphism τ from Lemma 5.5(1) to an inner automorphism, so a sign or reciprocity error here would invalidate the conclusion of Theorem 4.5. The authors must either prove the formula or cite a reference that contains it before the paper is accepted.","section":"5.3, Lemma 5.6"}],"minor_comments":[{"comment":"The word 'catgeory' in the first sentence of Section 2.1 is a typo for 'category'.","section":"2.1"},{"comment":"In the statement of Theorem 3.7, the codomain of the monodromy representation ρπ is written as G_{Qλ}; it should be G_{Qπ}, since E has coefficients in Qπ.","section":"3, Theorem 3.7"},{"comment":"In Lemma 4.2(3), the notation 'ρ^ad_{K0,sℓ′}' is missing a comma; it should be 'ρ^ad_{K0,s,ℓ′}'.","section":"4.5, Lemma 4.2(3)"},{"comment":"The proof of Lemma 5.1 is quite terse; please expand the justification that the point [¯V(ξ0)_{E_v}] is isolated in M_dR(E_v) via [EG25, Lemma 4.9], and indicate where the finiteness of the set of isolated points used in the F∗-permutation argument is established.","section":"5.1, Lemma 5.1"},{"comment":"The notation 'πt1(Sv)' in the proof of Lemma 5.5 is undefined and confusing; it should be 'π_1(S_v)' or 'π^t_1(S_v)' with a clarification that it is the (tame) fundamental group.","section":"5.3, proof of Lemma 5.5"}],"recommendation":"major_revision","confidential_remarks":"The structural framework of the paper is convincing, and the missing calculation in Lemma 5.6 appears to be a local, checkable statement. I would be inclined to accept after the authors supply the calculation or a precise reference. I also note that [Con11] is cited as a preprint; the authors should verify its status and perhaps replace it with a published reference if available."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This paper does what it says: it extends KP24 to the ℓ=p case by constructing the canonical overconvergent F-isocrystal and proving its Frobenius classes match the canonical ℓ-adic local systems. The new ingredient is the overconvergence argument—via rigidity of the canonical flat bundle, EG25's Frobenius pullback, and Dwork—plus the companion identification. The structure is clear, and the paper is honest about what is imported and what is ad hoc.\n\nThe Tannakian framework is well organized. The reduction of Theorem 4.5 to Lemma 5.5 is sound, and Lemma 5.5 itself is a clean consequence of Margulis superrigidity plus KP24's special-point machinery. I could not find circularity: KP24 handles ℓ≠p, companions are external, and the p-adic Hodge inputs (PST+24, EG25, DLLZ23) are prior work.\n\nTwo places need scrutiny. First, Lemma 5.1's proof that the canonical flat bundle is an isolated point of the moduli stack of flat connections, hence admits an F^a-structure after pullback, is largely delegated to EG25. The argument is plausible—H^1(Γ, ξ0)=0 gives cohomological rigidity—but a referee should check the stack-theoretic step carefully. Second, Lemma 5.6 is the only place the p-adic Frobenius at special points is computed, and the key formula is asserted with a parenthetical acknowledging that a crystalline-period calculation \"should\" be added. That formula is load-bearing: it upgrades the outer automorphism in Lemma 5.5(1) to an inner one. I don't think the formula is wrong—it matches the ℓ≠p value q^{-1}_{Frob_v}, and the reciprocity normalization is flagged—but the derivation is missing, not merely terse.\n\nAlso worth noting: the crystalline input from PST+24, Theorem 7.1 is still a preprint; that is a reliance on the community to verify, not a flaw in this paper.\n\nWho is this for? People working on Shimura varieties, p-adic companions, and the Langlands–Rapoport program. It deserves a serious referee. I would send it out; the referee should ask for a full proof (or a precise reference) for the crystalline-period calculation in Lemma 5.6 and a closer write-up of Lemma 5.1.","headline":"Completes the ℓ=p compatibility for canonical coefficient objects on adjoint Shimura varieties; the main theorem is credible and the two soft spots (overconvergence via EG25, and the special-point crystalline calculation in Lemma 5.6) are real but look patchable.","tokens_in":23828,"tokens_out":2118,"would_cite":true,"duration_ms":20815,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["11G18","14F30","14F20","14F35"],"pacs":[],"model":"deepseek-v4-flash","headline":"On adjoint Shimura varieties satisfying a rank condition, the canonical p-adic F-isocrystal now provably has the same Frobenius characteristic polynomials as every canonical ℓ-adic local system, for all ℓ≠p.","keywords":["Shimura varieties","F-isocrystals","ℓ-adic local systems","companions","p-adic Hodge theory","superrigidity","crystalline local systems","overconvergent isocrystals"],"falsifier":"Pick a closed point x of a mod-p fiber of one of these Shimura varieties and a representation ξ of the adjoint group; compute the characteristic polynomial of the linearized Frobenius on the crystalline realization D_crys(ξ∘ρ_{p,v}) and compare it with the characteristic polynomial of (ξ∘$ρ^{{ad}}$_ℓ)(Frob_x). The theorem predicts rational coefficients and equality for all ℓ≠p; any mismatch at one point would refute it. A more structural check is to look for a nontrivial deformation of the canonical flat bundle inside the moduli stack of flat connections, which would invalidate the rigidity step.","tokens_in":22770,"feed_emoji":"🧮","tokens_out":12013,"duration_ms":104177,"temperature":0.7,"pith_summary":"This paper proves that on adjoint Shimura varieties whose adjoint group has no simple factor of real rank below 2, the canonical p-adic coefficient object—an overconvergent $G$-F-isocrystal with Frobenius structure—has the same rational Frobenius characteristic polynomials as the canonical $\\ell$-adic local systems at every closed point of the mod-$p$ fiber, for every $\\ell\\neq p$. This fills the missing $p$-adic leg of the compatible system of canonical coefficient objects, extending the earlier $\\ell\\neq p$ compatibility of [KP24]. The result matters because such Shimura varieties, including the exceptional ones, have no known moduli interpretation, so this compatibility is a concrete test of the motivic expectation that they carry a universal family of motives with $G$-structure. The proof derives the F-isocrystal from the crystallinity of the canonical $p$-adic local system and then uses rigidity and companion constructions to match it with the $\\ell$-adic side.","feed_headline":"F-isocrystals and ℓ-adic systems now agree on Shimura varieties","feed_subtitle":"Closes the missing p-adic comparison for non-abelian Shimura varieties, moving toward mod-p point descriptions.","key_machinery":"The load-bearing object is the canonical overconvergent $G$-F-isocrystal $\\mathcal{E}^\\dagger_v$, upgraded from the convergent F-isocrystal attached to the canonical crystalline $p$-adic local system $\\rho_{p,v}$. The key mechanism is proving that this convergent object is overconvergent: via the algebraic p-adic Riemann–Hilbert correspondence of [DLLZ23], its underlying flat bundle is identified with the canonical flat $G$-bundle on the Shimura variety, which extends to a logarithmic flat bundle with nilpotent residues on a toroidal compactification; then a Frobenius pullback construction for logarithmic flat bundles, combined with rigidity of this bundle as an isolated point in the moduli stack of flat connections, supplies the required $F$-structure. The companion construction (a dictionary matching F-isocrystals to $\\ell$-adic local systems by equality of Frobenius semisimple conjugacy classes at all closed points) then converts the F-isocrystal into $\\ell$-adic local systems, and superrigidity forces any two such local systems to be conjugate, completing the comparison.","core_discovery":"The central claim is Theorem 1.1, proved as Theorem 4.5: after enlarging the level prime set $N$, for a closed point $v$ of residue characteristic $p$, there exists an overconvergent $G$-F-isocrystal $\\mathcal{E}^\\dagger_v$ on the mod-$p$ fiber $S_v$ whose Frobenius semisimple conjugacy classes are rational and agree with those of the canonical $G^{\\mathrm{ad}}(\\mathbb{Q}_\\ell)$-local system for every $\\ell\\neq p$. For every representation $\\xi$ of $G^{\\mathrm{ad}}$ and every closed point $x\\in S_v$, the characteristic polynomial of the linearized Frobenius on $x^*\\mathcal{E}^\\dagger_v(\\xi)$ has coefficients in $\\mathbb{Q}$ and equals the characteristic polynomial of $(\\xi\\circ\\rho^{\\mathrm{ad}}_\\ell)(\\mathrm{Frob}_x)$. The paper further proves that this F-isocrystal extends to a convergent logarithmic F-isocrystal on a toroidal compactification with unipotent monodromy, and that its Tannakian monodromy is the full adjoint group, so the companion formalism applies. Thus the canonical coefficient objects on adjoint Shimura varieties form a compatible system across all primes, with a single rational Frobenius class at each closed point.","pith_inferences":["If the theorem is right, the $p$-adic realization of the conjectural motive with $G$-structure is now pinned down on these Shimura varieties, completing the compatible system in the same sense as the abelian-type results; this suggests the motivic expectation now has full evidence in the superrigid regime.","The overconvergence mechanism appears to be a general principle: a crystalline local system whose underlying flat bundle is an isolated point in its moduli stack of flat connections should admit an overconvergent $D_{\\mathrm{crys}}$. This could be tested on other rigid locally symmetric varieties and would give a modular-curve-free route to overconvergence.","Because the comparison is made at the level of Tannakian monodromy groups, refinements such as comparing Newton polygons of special fibers, or studying the weight filtration on $D_{\\mathrm{crys}}$, are natural next checks that go beyond the characteristic-polynomial statement of the paper.","The exclusion of $p=2$ and ramified primes is likely technical; extending the Frobenius pullback functor used in the proof should remove these restrictions and enlarge the set of primes covered by the theorem."],"forward_implications":["Corollary 1.4: for points of $S(\\mathcal{O}_F[1/N_y])$ and any place $v\\mid p$, the characteristic polynomial of the crystalline Frobenius on $D_{\\mathrm{crys}}(\\xi\\circ\\rho^{\\mathrm{ad}}_{p,y}|_{\\mathrm{Gal}_{F_v}})$ has rational coefficients and equals the characteristic polynomial of $(\\xi\\circ\\rho^{\\mathrm{ad}}_{\\ell,y})(\\mathrm{Frob}_v)$ for every $\\ell\\neq p$.","At every closed point $x$ of the mod-$p$ fiber there is a single rational semisimple conjugacy class $[\\gamma^{\\mathrm{ad}}_{0,x}]$ in $(G^{\\mathrm{ad}}//G^{\\mathrm{ad}})(\\mathbb{Q})$ that equals the Frobenius class of the $\\ell$-adic local system for all $\\ell\\neq p$ and of the F-isocrystal at $p$ (Remark 1.5).","The F-isocrystal $\\mathcal{E}^\\dagger_v$ extends to a convergent logarithmic F-isocrystal on a toroidal compactification with unipotent monodromy, so the compatibility is valid at the boundary as well as on the open fiber (Corollary 5.2).","The $\\ell$-adic companion $\\rho^{\\mathrm{ad}}_{\\pi\\rightsquigarrow\\lambda,v}$ produced by the companion theorem is $G^{\\mathrm{ad}}(\\mathbb{Q}_\\lambda)$-conjugate to the canonical local system $\\rho^{\\mathrm{ad}}_{\\ell,v}$, so the two local systems agree as Tannakian objects, not merely at the level of Frobenius polynomials (Theorem 4.5).","Together with [Pat25], the compatibility upgrades from the adjoint quotient to the full canonical $G(\\mathbb{Q}_\\ell)$-local systems and the $G$-F-isocrystal (Remark 1.2)."],"supporting_citations":[{"why":"Establishes the ℓ≠p compatibility and the superrigidity-based comparison lemmas that Theorem 4.5 invokes.","marker":"[KP24]"},{"why":"Its Theorem 7.1 proves the canonical p-adic local system is log-crystalline, the input that produces the convergent F-isocrystal.","marker":"[PST+24]"},{"why":"Supplies the Frobenius pullback functor and the isolation criterion that give the canonical flat bundle its F-structure.","marker":"[EG25]"},{"why":"Its logarithmic Riemann–Hilbert correspondence identifies the de Rham realization of the p-adic local system with the canonical flat G-bundle.","marker":"[DLLZ23]"},{"why":"Builds the comparison of pro-algebraic fundamental groups on which the crystalline-to-étale companion correspondence rests.","marker":"[Dri18]"},{"why":"Provides the companion theorem converting overconvergent F-isocrystals into ℓ-adic local systems with matching Frobenius classes.","marker":"[Ked22a]"},{"why":"Gives the superrigidity vanishing H^1(Γ, ξ0)=0 that makes the canonical flat bundle an isolated point in its moduli stack.","marker":"[Mar91]"},{"why":"Supplies the crystalline realization construction that attaches a convergent F-isocrystal to a crystalline local system.","marker":"[Fal89]"},{"why":"Modern formulation of crystalline local systems and D_crys used to state and verify the crystallinity condition.","marker":"[GR24]"}],"fun_headline_variants":["F-isocrystals match ℓ-adic systems on Shimura varieties","Adjoint Shimura varieties: one Frobenius class for all primes","Companion construction unifies F-isocrystals and ℓ-adic sheaves","Compatibility proven for F-isocrystals on adjoint Shimura"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The conclusion depends on the fact that the relevant group has no small-rank simple factors, which is what forces the p-adic local system to be crystalline and the flat bundle to be rigid; if that fails, the overconvergent F-isocrystal may not exist.","fun_headline_variants_meta":{"raw":{"variants":["F-isocrystals match ℓ-adic systems on Shimura varieties","Adjoint Shimura varieties: one Frobenius class for all primes","Companion construction unifies F-isocrystals and ℓ-adic sheaves","Compatibility proven for F-isocrystals on adjoint Shimura"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000745,"raw_usage":{"total_tokens":3304,"prompt_tokens":909,"completion_tokens":2395,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":525,"completion_tokens_details":{"reasoning_tokens":2314}},"tokens_in":525,"tokens_out":2395,"duration_ms":16199,"temperature":1.0,"reasoning_tokens":2314,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T10:24:50.913146+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Pick a closed point x of a mod-p fiber of one of these Shimura varieties and a representation ξ of the adjoint group; compute the characteristic polynomial of the linearized Frobenius on the crystalline realization D_crys(ξ∘ρ_{p,v}) and compare it with the characteristic polynomial of (ξ∘$ρ^{{ad}}$_ℓ)(Frob_x). The theorem predicts rational coefficients and equality for all ℓ≠p; any mismatch at one point would refute it. A more structural check is to look for a nontrivial deformation of the canonical flat bundle inside the moduli stack of flat connections, which would invalidate the rigidity step.","supporting_citations":[],"review_version":1}