{"id":"8e9ab5ea-4c77-4aa9-8e2f-74015ddf3180","arxiv_id":"2506.06932","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"A formal density theorem says that, under a Newton-Hodge irreducibility condition, deformations of a mod p point with large p-adic monodromy are open dense in every p-adic neighborhood.","lead":"The paper proposes a conjecture that Hecke orbits of positive-dimensional subvarieties of Shimura varieties are p-adically nowhere dense, and proves a local version for Hodge-type Shimura varieties: around a mod p point, the lifts whose p-adic Galois representation has large monodromy form an open dense set. The result is a step toward understanding how Hecke symmetries behave in p-adic topology, with applications to Torelli and Zilber-Pink type questions.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1.5's density half is not proved: Proposition 3.7 only yields density after replacing K by a finite extension, so density of S_{K,x}^{max} in S_{K,x} is unsubstantiated.","rationale":"The reader's verdict is CONDITIONAL with high correctness risk, and I agree. However, the single most load-bearing issue is not Proposition 2.7 (though its proof is sketchy) but the mismatch between Proposition 3.7 and Theorem 1.5. Proposition 3.7 explicitly concludes density only after possibly replacing K with a finite extension. The proof of Theorem 1.5 says it follows from Theorem 2.1 and Proposition 3.7, but this does not give density over the original K. A finite extension K'/K has more points; density of the large-monodromy locus among K'-points does not imply density among K-points. The paper's introduction acknowledges a related obstruction: base-changing to wild ramified extensions can shrink monodromy, so the statement genuinely depends on the field. The proof does not show that generic points can be chosen to lie in the image of the K-points of the Rapoport-Zink space, nor that the constructed point descends. This is a gap in the central claim as stated. Proposition 2.7's issues are real but likely repairable: the p-power map on congruence quotients is surjective for m >= 2 by standard p-adic Lie theory, and Lemma 2.5 is plausible though the Breuil-module argument is incomplete. The field-extension issue is more fundamental because it affects the statement itself, not just a lemma. Therefore I keep the reader's CONDITIONAL verdict and propose a concrete test to check whether the density statement can be rescued.","tokens_in":13117,"tokens_out":19594,"duration_ms":195647,"concrete_test":"Work out the period map for a concrete Hodge-type Rapoport-Zink space with nontrivial stabilizer (e.g., the basic locus of GSp_4) and determine whether the image pi(RZ(K)) contains generic points of F^a(K) for a finite extension K of \\breve{E}. If the only generic points in the image require a strictly larger field K', then the density statement for K fails and the theorem must be weakened. Alternatively, check whether the proof of Proposition 3.7 can be modified: show that for every x in S_{K,x} and every neighborhood U, there exists z in U cap S_{K,x} with large monodromy, without enlarging K.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing gap is in the density half of Theorem 1.5. Proposition 3.7 concludes density only 'after possibly replacing K with a finite extension.' The proof finds a generic point z in F^a(K), but the point y0 in the fiber pi^{-1}(z) is obtained after replacing K by a finite extension K' (since the fiber is etale, any point is defined over a finite extension). Thus the produced large-monodromy point lies in S_{K',x}, not S_{K,x}. Theorem 1.5 asserts density in S_{K,x} for the original K. Since K-points are a proper subset of K'-points, density over K' does not imply density over K; a priori the large-monodromy locus over K' could avoid K-points. The proof does not show that the generic point can be chosen so that the fiber has a K-point, nor that the resulting point descends to K. This is not a technicality: the introduction explains that passing to wild ramified extensions shrinks monodromy (text after Corollary 1.6), so base-changing the target field changes the problem. Theorem 1.5 as stated is strictly stronger than what is proven.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes Conjecture 1.3, a p-adic nowhere-density statement for Hecke orbits of positive-dimensional subvarieties of Hodge-type Shimura varieties. Its main theorem (Theorem 1.5) asserts that, under an HN-irreducibility condition, the locus S_{K,x}^{max} of O_K-points in the formal neighborhood of a mod p point x whose p-adic Galois representation has large monodromy is p-adically open dense. The proof has two parts: Theorem 2.1 establishes openness using Breuil's classification of finite flat group schemes and a congruence-subgroup propagation lemma (Proposition 2.7); Proposition 3.7 establishes density using p-adic uniformization, period domains, and generic points in F^a. Corollary 1.6 derives p-adic nowhere density of prime-to-p Hecke orbits.","tokens_in":13311,"tokens_out":15257,"duration_ms":158049,"significance":"If the main theorem were proved as stated, it would be a significant step toward the proposed conjecture and would generalize Maulik-Poonen's formal density result to Hodge-type Shimura varieties by a genuinely different method (Breuil modules plus period-domain geometry). The paper is readable and honest about limitations, explicitly flagging the obstruction to extending the result to C_p-points. It also imports its main external inputs from Chen, Gleason-Lim-Xu, and Gleason-Lourenco rather than relying on circular self-support. However, the current proofs of both the openness and density halves contain load-bearing gaps, so the advertised Theorem 1.5 is conditional on work that is not supplied in the manuscript.","major_comments":[{"comment":"The density half of Theorem 1.5 is not proved for the original field K. In the proof of Proposition 3.7, the generic point z is produced in F^a(K) by Lemma 3.6, but the fiber \\bar\\pi^{-1}(z) is only shown to admit a point after replacing K by a finite extension K': the proof explicitly says 'any connected component of it is a finite extension K' of K. After replacing K with K', we obtain a K-point y0'. Thus the constructed large-monodromy point lies in S_{K',x}, not in S_{K,x}. Since 'large monodromy' in Definition 1.4 is a property of the Gal(\\bar K/K)-representation, a K'-point does not produce an element of S_{K,x}^{max}; and the text after Corollary 1.6 acknowledges that passing to wild ramified extensions can shrink monodromy, so there is no descent argument. Consequently Proposition 3.7 establishes only density after base change, and Theorem 1.5, which drops this caveat, is strictly stronger than what is proved.","section":"Section 3.6, Proposition 3.7 (and Theorem 1.5)"},{"comment":"The openness half rests on the claim that the p-power map H_m/H_{m+1} \\to H_{m+1}/H_{m+2} is surjective for some m \\ge n. The proof computes the analogous map for GL_{2g} and then says the general case follows from naturality of the p-adic exponential map. This is not a complete argument: the exponential/logarithm is only a local bijection between a neighborhood of the identity and a Lie algebra lattice, and it does not directly identify the graded pieces H_m/H_{m+1} and H_{m+1}/H_{m+2} for an arbitrary reductive Z_p-model G^{der}. A precise congruence-subgroup calculation, or a reference, is needed. Without it, the induction step in Proposition 2.7 and hence the openness conclusion of Theorem 2.1 are unsupported.","section":"Section 2.3, Proposition 2.7 (Eq. (2.9))"},{"comment":"Lemma 2.5, which converts p-adic congruence of lifts into isomorphism of p^{n+1}-torsion, is proved only by a sketch. The claim that the Breuil-module invariants (M, Fil^1 M, \\phi_1) are determined by the reduction modulo p^{n+1} is asserted rather than derived: the diagram chase for the divided Frobenius \\phi_1 assumes the embeddings (2.5)-(2.6) and the required compatibility with the isomorphism \\psi, which are essentially the statements needing proof. Since Theorem 2.1 applies Lemma 2.5 with level n+1, this gap is load-bearing for the openness half.","section":"Section 2.2, Lemma 2.5 and its use in Proposition 2.7"}],"minor_comments":[{"comment":"There are numerous typographical artifacts (for example, 'thep-adic' in the abstract, 'OK' for O_K, and missing subscripts in 'Ocris'), and the reference formatting is inconsistent; a careful proofread is needed.","section":"Throughout"},{"comment":"The final sentence says the constructed point 'maps to a OK point z \\in V_x'; it should be an O_{K'}-point, and the reuse of z for both the generic period and the image point is confusing.","section":"Section 3.6, proof of Proposition 3.7"},{"comment":"The definition of L_i as 'H_i/H_{i+1} mod p^{i+1}' conflates a quotient group with a set of matrices; these should be defined as congruence subgroups in G^{der}(Z_p), not as quotients.","section":"Section 2.3, Proposition 2.7"},{"comment":"Lemma 3.6 reuses the symbol K0 from Section 2.1 without redefinition; in Section 3 the reader must infer that K0 denotes the completed maximal unramified extension of Q_p.","section":"Section 3.6, Lemma 3.6"}],"recommendation":"major_revision","confidential_remarks":"The paper is promising but incomplete. The main theorem as stated is not established: the density half is proved only after a finite base change, and the openness half relies on an unproved congruence-quotient surjectivity statement. I recommend major revision rather than rejection, because the central strategy is credible and the gaps may be repairable within the scope of the manuscript. There is no evidence of circularity or data fitting."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the paper proposes a meaningful conjecture, proves a new formal density statement for large monodromy in Hodge-type Shimura varieties, but the proof as written doesn't establish the stated theorem.\n\nThe genuinely new piece is Theorem 1.5: under HN-irreducibility, the large-monodromy locus in a formal O_K-neighborhood of a mod p point is open and dense. This generalizes Maulik-Poonen from PEL to Hodge type by a different route—Breuil classification plus Rapoport-Zink uniformization—and the strategy is coherent. The conjecture and the discussion around it are honest, and the applications to Torelli and Zilber-Pink questions would be substantial if the theorem were solid.\n\nThe proof has two load-bearing gaps. First, Proposition 2.7 (the openness half) relies on surjectivity of the p-power map H_m/H_{m+1} -> H_{m+1}/H_{m+2} for the congruence subgroups of a general reductive group. The proof only works out GL_{2g} and then asserts the general case. This is probably true via a Lie-algebra argument, but it's not written, and the induction over m is sketched in a way that needs more care. Second, and more serious, Proposition 3.7 only proves density after 'possibly replacing K with a finite extension.' The period map is étale, so the generic point z in the period domain may have no K-rational point in its fiber; the produced large-monodromy point lives over K', not K. Theorem 1.5 asserts density over the original K, and the paper gives no descent argument. The stress-test note is correct that this is not cosmetic: the paper itself observes that wild ramified extensions shrink monodromy, so passing to a larger field changes the problem. As stated, Theorem 1.5 is stronger than what is proven.\n\nAlso, Corollary 1.6 is stated without proof. It may follow once Theorem 1.5 is fixed, but as written it's just an assertion.\n\nThe citation pattern is fine: the external ingredients (Chen's generic monodromy, Gleason-Lourenco connectedness, Kim/Howard-Pappas uniformization) are real and used honestly. I don't see circularity or overclaiming relative to what the authors admit. The weaknesses are in the execution, not the architecture.\n\nWho benefits: anyone working on Hecke orbits, p-adic monodromy, or Shimura varieties of Hodge type. The ideas are worth a serious look, but the main theorem should not be accepted in this form. I'd send it to peer review with a clear message to the authors: complete Proposition 2.7, fix the field-extension issue in Proposition 3.7, and either prove Corollary 1.6 or remove it.","headline":"A plausible and genuinely new formal density result for Hodge-type Shimura varieties, but the proof as written does not establish the stated theorem.","tokens_in":13899,"tokens_out":4996,"would_cite":false,"duration_ms":38457,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["11G18","14L05","14G35","11F80"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper establishes that around a mod $p$ point of a Hodge-type Shimura variety, the points of the formal neighborhood with large $p$-adic monodromy form an open dense set, giving a local version of the proposed $p$-adic nowhere density…","keywords":["p-adic monodromy","Hecke orbits","Shimura varieties","Hodge type","p-divisible groups","Breuil modules","formal neighborhoods","nowhere density"],"falsifier":"Exhibit two $p$-divisible groups over $\\mathcal{O}_K$ whose reductions modulo $p^{n+1}$ are isomorphic as group schemes, where the first has Galois image containing a congruence subgroup of $G^{\\mathrm{der}}(\\mathbb{Q}_p)$ and the second has Galois image containing no such subgroup; such a pair would falsify Proposition 2.7 and remove the openness assertion of Theorem 1.5.","tokens_in":12862,"feed_emoji":"🧮","tokens_out":12221,"duration_ms":104008,"temperature":0.7,"pith_summary":"This paper proposes a conjecture: in the $p$-adic topology on a Shimura variety, the Hecke orbit of any positive-dimensional subvariety should be nowhere dense, rather than dense as it is in the archimedean topology. The main theorem proves a local version of this. Fix a mod $p$ point whose Newton and Hodge polygons only meet at their endpoints, let $K$ be a totally ramified finite extension of the completed maximal unramified extension of the reflex field, and look at the $K$-integral points in the formal neighborhood of that point whose special fiber is the same $p$-divisible group. Within that set, the points whose $p$-adic Galois representation has large monodromy are open and dense. This is the local density input that the full conjecture needs, and it comes with the corollary that prime-to-$p$ Hecke orbits of such points are $p$-adically nowhere dense.","feed_headline":"Large p-adic monodromy is open dense around mod p points","feed_subtitle":"Formal local density on Hodge-type Shimura varieties, one step toward the p-adic nowhere density conjecture.","key_machinery":"The argument runs on a congruence-to-monodromy mechanism for $p$-divisible groups. Lemma 2.5, obtained from the classification of finite flat group schemes by Breuil modules, says that if two $p$-divisible groups agree modulo $p^{n+1}$, their $p^n$-torsion is isomorphic as group schemes. Proposition 2.7 then propagates large monodromy: if one group has Galois image containing the congruence subgroup $H_n = G^{\\mathrm{der}}(\\mathbb{Z}_p) \\cap K_n$ with $K_n = \\{ M \\in \\mathrm{Sp}_{2g}(\\mathbb{Z}_p) : M \\equiv I \\bmod p^n \\}$, any group whose $p^{n+1}$-torsion is isomorphic has image containing some $H_m$; the proof uses the $p$-power map on congruence quotients, whose surjectivity it asserts. This makes the large-monodromy locus open. For density, the paper uses Rapoport-Zink spaces of Hodge type (formal moduli of deformations of a $p$-divisible group with crystalline Tate tensors), the rigid analytic period map to a $p$-adic flag variety, the connectedness of the admissible period domain, and a lemma that generic points of the period domain are $p$-adically close to any given point; generic points are already known from the theory of $p$-adic period domains to have large monodromy. The uniformization theorem identifies formal neighborhoods in the Shimura variety with quotients of these Rapoport-Zink spaces, which is why the local statement transfers to $S_{K,x}$.","core_discovery":"Let $x$ be an $\\mathbb{F}_p$-point of a Hodge-type Shimura variety and let $(G,[b],\\mu)$ be the associated local Hodge-Shimura datum, with $A_x[p^\\infty]$ the $p$-divisible group of the abelian variety at $x$. Assume the Newton and Hodge polygons of $[b]$ and $\\mu$ do not touch outside their endpoints (HN-irreducibility). For a totally ramified finite extension $K$ of the completed maximal unramified extension of the reflex field, let $S_{K,x}$ be the set of $\\mathcal{O}_K$-points of the formal neighborhood of $x$ whose special fiber is isomorphic to $A_x[p^\\infty]$. The paper proves that the subset $S_{K,x}^{\\max}$ of points whose Galois representation $\\rho_y$ has image containing an open subgroup of $G^{\\mathrm{der}}$ is $p$-adically open and dense in $S_{K,x}$. It reaches this by a different route than the earlier PEL case: openness comes from a congruence criterion for monodromy proved with the classification of finite flat groups, and density comes from producing generic points, which necessarily have large monodromy, in every $p$-adic neighborhood of any point. Theorem 1.5 then yields Corollary 1.6: the prime-to-$p$ Hecke orbit of a point with good reduction and HN-irreducible local datum is nowhere dense in the $p$-adic topology on $\\mathrm{Sh}(\\mathbb{C}_p)$.","pith_inferences":["If the propagation lemma can be strengthened to allow bounded wild ramification of the base field, the formal-local statement may extend from finite extensions of $\\breve{E}$ to all $\\mathbb{C}_p$-deformations, which is exactly the gap the paper identifies.","The same congruence-propagation mechanism might prove analogous open-density statements for other classes of $p$-divisible groups or local Shimura varieties, since the input is the classification of finite flat groups by filtered modules rather than a special feature of abelian varieties.","A testable consequence beyond the paper's scope: if the Hecke orbit of the Torelli locus is $p$-adically nowhere dense, then generic abelian varieties over $\\mathbb{Q}$ are not quotients of Jacobians; the paper notes this direction without proving it."],"forward_implications":["If Theorem 1.5 holds, the prime-to-$p$ Hecke orbit of any point of $\\mathrm{Sh}(\\mathbb{C}_p)$ with good reduction and HN-irreducible local datum is $p$-adically nowhere dense.","The formal-local statement is a first step toward the full conjecture: one would need the same open-density statement for all $\\mathbb{C}_p$-deformations, not only those over fixed finite extensions of $\\breve{E}$.","If the full conjecture holds, the set of CM points in $\\mathrm{Sh}(\\mathbb{C}_p)$ is closed in the rigid analytic topology, and the Hecke orbit of the Torelli locus is small enough to yield new statements about which abelian varieties are isogenous to Jacobians.","The proof strategy gives an alternative route to the earlier PEL-type result, replacing $p$-adic Lefschetz arguments with the classification of finite flat groups and Rapoport-Zink uniformization."],"supporting_citations":[{"why":"Supplies the classification of finite flat group schemes by Breuil modules, used to prove Lemma 2.5 that congruence modulo $p^{n+1}$ forces isomorphic $p^n$-torsion.","marker":"[Bre00]"},{"why":"Provides the integral-model theory and crystalline tensors that define the local Hodge-Shimura datum attached to a mod $p$ point.","marker":"[KIS10]"},{"why":"Constructs Rapoport-Zink spaces of Hodge type, including the period map and the group action used in the density argument.","marker":"[KIM18a]"},{"why":"Gives the rigid analytic uniformization of Hodge-type Shimura varieties as quotients of Rapoport-Zink spaces, transferring the local statement to $S_{K,x}$.","marker":"[KIM18b]"},{"why":"Proves that generic points of admissible period domains have large monodromy, which is the source of density.","marker":"[Che14]"},{"why":"Proves connectedness of $p$-adic period domains, ensuring the period map has points in every connected component.","marker":"[GL22]"},{"why":"Extends the generic-point large-monodromy theorem to the setting needed for the present proof.","marker":"[GLX23]"},{"why":"Establishes that weakly admissible points are admissible, so each point of the period domain yields a crystalline Galois representation.","marker":"[CF00]"},{"why":"Provides the earlier PEL-type result that Theorem 1.5 generalizes, and the baseline comparison for the strategy.","marker":"[MP12]"}],"fun_headline_variants":["Large p-adic monodromy is open dense near mod p points","Locus with large monodromy is open dense on Shimura varieties","Open dense large monodromy: a Hecke orbit step","Density of large monodromy points in formal neighborhoods"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof rests on the propagation lemma that two $p$-divisible groups with the same reduction modulo $p^{n+1}$ have the same large-monodromy property, and that lemma relies on two technical steps the paper asserts rather than fully proves: surjectivity of the $p$-power map on congruence quotients and the claim that Breuil-module invariants are determined by reduction modulo $p^{n+1}$.","fun_headline_variants_meta":{"raw":{"variants":["Large p-adic monodromy is open dense near mod p points","Locus with large monodromy is open dense on Shimura varieties","Open dense large monodromy: a Hecke orbit step","Density of large monodromy points in formal neighborhoods"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001819,"raw_usage":{"total_tokens":7163,"prompt_tokens":956,"completion_tokens":6207,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":572,"completion_tokens_details":{"reasoning_tokens":6129}},"tokens_in":572,"tokens_out":6207,"duration_ms":43028,"temperature":1.0,"reasoning_tokens":6129,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T05:47:13.490824+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Exhibit two $p$-divisible groups over $\\mathcal{O}_K$ whose reductions modulo $p^{n+1}$ are isomorphic as group schemes, where the first has Galois image containing a congruence subgroup of $G^{\\mathrm{der}}(\\mathbb{Q}_p)$ and the second has Galois image containing no such subgroup; such a pair would falsify Proposition 2.7 and remove the openness assertion of Theorem 1.5.","supporting_citations":[],"review_version":1}