{"id":"7ed6ad35-5a9c-4ea6-8e93-c4ab5e03bd55","arxiv_id":"2411.15345","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Every compact Nil 3-manifold occurs as a cusp cross-section either in all arithmetic commensurability classes or in exactly one, and also in some non-arithmetic complex hyperbolic or H^2_C-manifold.","lead":"This paper classifies which arithmetic complex hyperbolic 2-manifolds can have a given compact Nil 3-manifold as a cusp cross-section, and proves every such shape appears as a cusp in some non-arithmetic complex hyperbolic or H^2_C-manifold. The classification splits the seven Nil families into those that appear in every Picard modular commensurability class and those that appear in exactly one.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The proof of Theorem 4 runs the cusp-separability promotion in SU(2,1), but families (3) and (4) have antiholomorphic cusp groups; as written, the non-arithmetic H^2_C case is not established.","rationale":"The paper's principal new classification, Theorems 1-3, is supported by explicit representations and a clean eigenvalue argument, and I see no reason to dispute those conclusions. The conditional verdict is driven by Theorem 4's non-arithmetic promotion step. For families (3) and (4), the cusp group necessarily contains an antiholomorphic reflection, so the separability argument cannot be run inside SU(2,1). The proof as written makes the stronger and false claim that these cusps can be embedded in a complex hyperbolic orbifold, contradicting Theorem 3. The final 'intersection with the first factor' step is also imprecise, but it has an evident repair by pulling back Lambda_0 through phi, so it is not the decisive obstruction. The genuinely load-bearing question is whether Bergeron's separability lemma extends to the full isometry group and the R-reflection lattices used in Corollary 2. A check of the hypotheses of that lemma settles whether the missing step is a routine extension or a real gap in the non-arithmetic existence theorem. Since the arithmetic classification is unaffected and the non-arithmetic gap is repairable in principle, the reader's CONDITIONAL verdict remains appropriate.","tokens_in":18921,"tokens_out":15202,"duration_ms":142493,"concrete_test":"Check whether Bergeron's 'lemme principal', as quoted in Proposition 3.8 of [M1], applies to G = Isom(H^2_C) modeled as the real algebraic group PU(2,1) ⋊ <sigma> with H = Stab_G(q_infinity), and to the R-reflection lattices ~Gamma(6,1/6), ~S(6,sigma_1), ~S(6,sigma_4), ~S(4,sigma_1) from Corollary 2. If G is linear algebraic and H is an algebraic subgroup, then the same proof gives separability of cusp subgroup G_infinity(O_d) in these lattices and Theorem 4 can be repaired by replacing 'G = SU(2,1)' with this full isometry group. If the hypotheses fail, no argument in the paper establishes the embedding step for families (3) and (4), and the non-arithmetic claim for those families would require a new separability proof.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The arithmetic classification in Theorems 1-3 is well supported by explicit holonomy representations and by the eigenvalue obstruction in Theorem 2, and I found no defect there. The load-bearing weak point is the proof of Theorem 4, specifically the promotion of an immersed Nil cusp to an embedded cusp cross-section in a non-arithmetic quotient. For families (3) and (4), horizontal and double half-twist, the cusp group necessarily contains antiholomorphic isometries: Theorem 3 proves this, and Lemma 5 realizes these groups inside G_infinity(O_d) = G^0_infinity(O_d) union G^0_infinity(O_d) sigma. Step (2) of the proof of Theorem 4 then invokes Bergeron's separability lemma with G = SU(2,1) and H = Stab_G(q_infinity), and concludes: 'We now have a non-arithmetic lattice Gamma_1 in SU(2,1) such that N is diffeomorphic to a cusp cross-section of the orbifold H^2_C/Gamma_1.' For families (3) and (4), this contradicts Theorem 3: no quotient of H^2_C by a torsion-free subgroup of SU(2,1) can have such a cusp. To justify the theorem's H^2_C-manifold conclusion for these families, one must prove separability of the antiholomorphic cusp subgroup in the full isometry group Isom(H^2_C), or in the R-reflection extensions ~Gamma and ~S from Corollary 2, and the paper does not supply that argument. The final product-lattice step has a separate but repairable slip: 'the intersection with the first factor of Lambda' cannot contain the diagonal image of rho_3(pi_1(N)); the intended construction is Gamma_2 = phi^{-1}(Lambda_0), which is finite-index, torsion-free, and contains rho_3(pi_1(N)). Thus the decisive open question is whether the separability mechanism applies to the antiholomorphic cusp groups needed for families (3) and (4).","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies which compact Nil 3-manifolds occur as cusp cross-sections of complex hyperbolic 2-manifolds and of the more general H^2_C-manifolds, with focus on the arithmetic commensurability classes determined by the Picard modular groups PU(2,1,O_d). The authors give explicit holonomy representations for all seven families of Nil 3-manifolds: families (1)--(4) are embedded into G∞(O_d) for every squarefree d, and families (5)--(7) are shown by an eigenvalue obstruction to occur only for d=1 or d=3. Theorem 3 establishes that families (3) and (4) cannot appear in any holomorphic quotient and must have antiholomorphic cusp holonomy. The paper also claims, in Theorem 4, that every Nil 3-manifold occurs as a cusp cross-section of a non-arithmetic complex hyperbolic or H^2_C-manifold. The arithmetic part is supported by explicit, checkable matrix computations; the non-arithmetic part has a gap in the promotion argument for the antiholomorphic families.","tokens_in":19203,"tokens_out":15859,"duration_ms":143438,"significance":"The arithmetic classification in Theorems 1--3 is a substantial and largely convincing contribution: it answers, for Nil 3-manifolds, the analogue of Sell's classification for flat 3-manifolds, and it supplies corrected explicit holonomy representations, including a repair of an incorrect representation in the appendix of [M1] for the double half-twist family. The eigenvalue obstruction in Theorem 2 is clean and gives a concrete falsifiable criterion. The holomorphic/antiholomorphic dichotomy in Theorem 3 is also useful. If the non-arithmetic claim in Theorem 4 can be repaired, it would be a significant extension of McReynolds' existence theorem; as written, however, the proof does not establish the claimed result for families (3) and (4), and the abstract overstates the theorem.","major_comments":[{"comment":"The abstract states: \"We also show that every compact Nil 3-manifold occurs as the cusp cross-section of some non-arithmetic complex hyperbolic 2-manifold.\" This is inconsistent with Theorem 3, which says that families (3) and (4) are not cusp cross-sections of any complex hyperbolic manifold or orbifold and can occur only in H^2_C-manifolds. The abstract must be corrected to \"complex hyperbolic 2-manifold or H^2_C-manifold\", or the statement of Theorem 4 must be revised so that the announced claim matches what is actually proved.","section":"Abstract"},{"comment":"There is an unresolved tension between Theorem 1 and Theorem 3 for families (3) and (4). Lemma 5 constructs holonomy representations for these families into G∞(O_d) = G^0∞(O_d) ∪ G^0∞(O_d)σ, so the cusp group contains antiholomorphic isometries, and the resulting quotient is an H^2_C-manifold rather than a complex hyperbolic manifold. If \"commensurability class of PU(2,1,O_d)\" means commensurability among torsion-free subgroups of PU(2,1), then Theorem 1 is false for families (3) and (4) by Theorem 3. If it means commensurability in the full isometry group Isom(H^2_C), this must be stated explicitly, and the abstract and Section 2.3's identification of commensurability classes with Picard modular groups must be qualified, since that identification is only for holomorphic quotients.","section":"§3, Theorem 1 and Lemma 5"},{"comment":"The promotion argument uses Proposition 2 with G = SU(2,1) and H = Stab_G(q∞). This applies only to lattices in the holomorphic isometry group. For families (3) and (4), the image of ρ2(π1(N)) contains antiholomorphic isometries by Theorem 3 and Lemma 5, so the relevant cusp subgroup lies in G∞(O_d), inside an R-reflection-generated lattice in Isom(H^2_C), not in a lattice in SU(2,1). The conclusion \"We now have a non-arithmetic lattice Γ_1 in SU(2,1) such that N is diffeomorphic to a cusp cross-section\" cannot hold for families (3) and (4). The paper needs a separate separability argument for the full cusp stabilizer H = Stab_Isom(H^2_C)(q∞), or for the non-holomorphic R-reflection groups ~Γ and ~S, and none is supplied.","section":"§3.2, proof of Theorem 4, Step (2)"},{"comment":"The final step of the promotion argument is not correctly stated. The diagonal embedding φ(Γ_1) is not a subgroup of the first factor of Λ, so \"the intersection with the first factor of Λ\" cannot contain the diagonal image of ρ3(π1(N)). The intended construction is presumably Γ_2 = φ^{-1}(Λ_0 ∩ φ(Γ_1)) (or an explicit projection followed by intersection with Γ_1), and one must then verify that Γ_2 is torsion-free, has finite index in Γ_1, and still has ρ3(π1(N)) as its full cusp subgroup. As written, this step is incomplete.","section":"§3.2, proof of Theorem 4, final paragraph"}],"minor_comments":[{"comment":"The displayed Heisenberg group law appears to omit the complex conjugation that makes the form alternating; as printed, Im(z1 z2) is symmetric and does not give the standard 2-step nilpotent Heisenberg group. Please check the convention in equation (1) and in Lemma 2(1), since the computations in Section 3 rely on it.","section":"Equation (1) and Lemma 2"},{"comment":"The statement contains a duplicated word: \"diffeomorphic to to a cusp cross-section\" should be \"diffeomorphic to a cusp cross-section\".","section":"Theorem 6 statement"},{"comment":"The word \"anitholomorphic\" should be \"antiholomorphic\".","section":"Theorem 3 statement"},{"comment":"The groups ~Γ and ~S are R-reflection groups in Isom(H^2_C), not lattices in PU(2,1). Corollary 2 and the proof of Theorem 4 should explicitly state this and identify the relevant index-2 holomorphic sublattice, since the current wording sometimes refers to them as lattices in PU(2,1).","section":"Corollary 2 and proof of Theorem 4"}],"recommendation":"major_revision","confidential_remarks":"The arithmetic classification is likely sound and valuable. The main question for the revision is whether the authors can supply a separability argument in the full isometry group Isom(H^2_C), or at least restate Theorem 4 and the abstract so that families (3) and (4) are only claimed for H^2_C-manifolds; without such a fix, the non-arithmetic existence theorem is not established as written."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take on Paupert–Sell. The arithmetic classification is the real content and it holds up. Theorems 1–3, with Lemma 5's explicit holonomy representations and Theorem 2's eigenvalue obstruction, give a clean answer to which Nil 3-manifolds occur in which Picard modular commensurability classes. They also fix some errors in McReynolds's appendix. I did not hand-check every matrix entry, but the level of detail is right for independent verification, and Theorem 3's proof that families (3) and (4) force antiholomorphic holonomy is clear. The citation pattern is honest: the paper leans on [M1], [M2], [Se], [DPP2], [PW], and those debts are stated up front.\n\nThe soft spot is Theorem 4. Step (2) of its proof invokes Bergeron's separability lemma with G = SU(2,1) and H = Stab_G(q∞), and concludes there is a non-arithmetic lattice Γ1 in SU(2,1) with N as cusp cross-section. For families (3) and (4), that contradicts the paper's own Theorem 3: those holonomy representations must contain antiholomorphic isometries, so the cusp group is not contained in SU(2,1). Bergeron's lemma, as used, handles only the holomorphic cusp stabilizer. To complete the proof for (3) and (4), you would need separability for the antiholomorphic cusp subgroup in Isom(H^2_C), or in the R-reflection extensions from Corollary 2, and that argument is not in the paper. As written, Theorem 4 is established for the holomorphic families (1), (2), (5), (6), (7), not for (3) and (4). The final step also has a small slip: 'the intersection with the first factor of Λ' is not a subgroup of Γ1; the intended construction is Γ2 = φ^{-1}(Λ0), which is finite-index, torsion-free, and contains the diagonal image. That one is easily repaired. The abstract makes things worse by claiming non-arithmetic complex hyperbolic manifolds for every Nil 3-manifold, which cannot be right for (3) and (4) in light of Theorem 3.\n\nSo: the arithmetic half is solid and citable. The non-arithmetic half needs a new separability argument before it is accepted. I would send this to a serious referee, with the expectation of a conditional accept once Theorem 4 is fixed. The paper deserves referee time; people working on cusp cross-sections, Picard modular groups, and non-arithmetic lattices in PU(2,1) will want Theorems 1–3 and the explicit representations.","headline":"The arithmetic classification in Theorems 1–3 is solid and worth refereeing; the proof of Theorem 4 has a real gap for the antiholomorphic families (3) and (4), and the abstract overstates the non-arithmetic conclusion.","tokens_in":19890,"tokens_out":4439,"would_cite":true,"duration_ms":37675,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["57K20","32Q45","22E40","11F06"],"pacs":[],"model":"deepseek-v4-flash","headline":"The seven families of compact Nil 3-manifolds are completely classified by which commensurability class of complex hyperbolic 2-manifolds they can bound as cusp cross-sections.","keywords":["Nil 3-manifolds","complex hyperbolic surfaces","cusp cross-sections","Picard modular groups","commensurability classes","Heisenberg group","non-arithmetic lattices","holonomy representations"],"falsifier":"The most direct check is to search for a holonomy representation of a 1/4-twist Nil 3-manifold into $G_\\infty(\\mathcal{O}_d)$ for a squarefree $d\\neq 1$ with the required rotational part $\\zeta_4$; the paper's eigenvalue argument says the eigenvalues $1,1,\\zeta_4$ would have to lie in a cubic extension of $\\mathbb{Q}(i\\sqrt{d})$, which is impossible, so any such representation would refute Theorem 2.","tokens_in":18616,"feed_emoji":"🌀","tokens_out":22172,"duration_ms":149327,"temperature":0.7,"pith_summary":"The paper settles, for compact Nil 3-manifolds, which commensurability classes of cusped arithmetic complex hyperbolic 2-manifolds realize a given manifold as a cusp cross-section. Four of the seven Nil-manifold families (Nil-tori, vertical and horizontal half-twists, double half-twists) occur in the commensurability class of every Picard modular group $\\mathrm{PU}(2,1,\\mathcal{O}_d)$, while the remaining three families (1/4-twist, 1/3-twist, 1/6-twist) each occur in exactly one class, pinned to $d=1$ or $d=3$. The paper also proves that every compact Nil 3-manifold occurs as a cusp cross-section of some non-arithmetic complex hyperbolic 2-manifold or $\\mathbb{H}^2_{\\mathbb{C}}$-manifold, and that the horizontal half-twist and double half-twist families occur only as $\\mathbb{H}^2_{\\mathbb{C}}$-manifolds, never as complex hyperbolic manifolds. Together these results give a complete answer to the geometric boundary problem for complex hyperbolic surfaces in real dimension three.","feed_headline":"Every Nil 3-manifold is a non-arithmetic complex hyperbolic cusp","feed_subtitle":"Four Nil families appear everywhere; three are pinned to a single commensurability class.","key_machinery":"The load-bearing object is the 3-dimensional Heisenberg group $\\mathrm{Nil}$, viewed as the punctured boundary of complex hyperbolic 2-space, with isometry group $\\mathrm{Isom}(\\mathrm{Nil}) = \\mathrm{Nil} \\rtimes (\\mathrm{U}(1) \\rtimes \\mathbb{Z}/2)$. The paper writes explicit holonomy representations of the seven Nil-manifold group presentations into the cusp subgroups $G_\\infty(\\mathcal{O}_d)$ of the Picard modular groups, checking the presentation relations with the group law of $\\mathrm{Nil}$. The obstruction part uses the rotational part $u$ of the ellipto-parabolic generator $\\alpha$: for families (5)-(7), $u$ is $\\zeta_4$, $\\zeta_3$, or $\\zeta_6$, and because a matrix representative in $\\mathrm{U}(2,1,E_d)$ has eigenvalues $1,1,u$ lying in a cubic extension of the field of definition $E_d$, the only possible $d$ are $1$ or $3$. For the non-arithmetic existence, the paper uses the fact that the known non-arithmetic lattices have cusp groups isomorphic to $G_\\infty(\\mathcal{O}_1)$ or $G_\\infty(\\mathcal{O}_3)$ (from [DPP2] and [DFP]), and then applies the separability lemma in [B] together with the covering arguments of [M1] and [M2] to promote an immersed cusp to an embedded one in a torsion-free finite cover.","core_discovery":"The central discovery is that the seven families of compact Nil 3-manifolds split into universal and rigid cusp shapes for arithmetic complex hyperbolic surfaces. Using explicit holonomy representations into the cusp subgroups $G_\\infty(\\mathcal{O}_d)$ of $\\mathrm{Isom}(\\mathrm{Nil})$, the paper shows in Theorem 1 that families (1)–(4) appear as cusp cross-sections in the commensurability class of $\\mathrm{PU}(2,1,\\mathcal{O}_d)$ for every squarefree $d$, meaning these shapes occur in every arithmetic commensurability class. In Theorem 2, families (5)–(7) are shown to occur in only one commensurability class each: the 1/4-twist only for $d=1$ and the 1/3- and 1/6-twists only for $d=3$, via an eigenvalue argument that forces the rotational part $\\zeta_4$ or $\\zeta_3/\\zeta_6$ to lie in the field of definition of the class. Theorem 3 shows that the horizontal half-twist and double half-twist families can never be cusp cross-sections of complex hyperbolic manifolds because every holonomy representation must contain antiholomorphic isometries. Finally, Theorem 4 combines the explicit representations with the cusp groups of the known non-arithmetic lattices to prove that every compact Nil 3-manifold occurs as a cusp cross-section of some non-arithmetic complex hyperbolic 2-manifold or $\\mathbb{H}^2_{\\mathbb{C}}$-manifold.","pith_inferences":["The eigenvalue obstruction that pins families (5)-(7) to $d=1$ or $d=3$ suggests a general constraint for almost-flat manifolds modelled on generalized Heisenberg groups: the rotational part of a cusp stabilizer must lie in a field compatible with the lattice's field of definition, which could constrain higher-dimensional complex and quaternion hyperbolic cusps.","The non-arithmetic existence proof likely generalizes to any non-arithmetic lattice that contains a cusp group isomorphic to $G_\\infty(\\mathcal{O}_1)$ or $G_\\infty(\\mathcal{O}_3)$ and satisfies the integrality condition used here, so the result may apply beyond the known examples.","A natural next question is which Nil-tori occur as cusp cross-sections of one-cusped complex hyperbolic 2-manifolds; the explicit representations here may help determine the possible Euler numbers, since the one-cusped case is restricted to Nil-tori with Euler number a multiple of 4.","The complete classification for complex hyperbolic surfaces provides a template for the analogous question for quaternion hyperbolic manifolds, where the almost-flat cusp cross-sections are modelled on higher-dimensional Heisenberg groups and no comparable classification is currently known."],"forward_implications":["For every squarefree $d$, the Picard modular group $\\mathrm{PU}(2,1,\\mathcal{O}_d)$ has cusps whose cross-sections are homeomorphic to Nil-tori, vertical half-twists, horizontal half-twists, and double half-twists.","The 1/4-twist Nil-manifold can appear as a cusp cross-section only in the commensurability class of $\\mathrm{PU}(2,1,\\mathcal{O}_1)$, and the 1/3- and 1/6-twist manifolds only in the class of $\\mathrm{PU}(2,1,\\mathcal{O}_3)$.","Families (3) and (4) cannot be cusp cross-sections of complex hyperbolic manifolds; any quotient with such a cusp must be an $\\mathbb{H}^2_{\\mathbb{C}}$-manifold whose holonomy contains antiholomorphic isometries.","Every compact Nil 3-manifold is realized as a cusp cross-section of some non-arithmetic complex hyperbolic 2-manifold or $\\mathbb{H}^2_{\\mathbb{C}}$-manifold, extending the arithmetic existence result of [M1] and [M2] to non-arithmetic settings.","The classification is exhaustive: each of the seven Nil-manifold families is either universal (occurring in all arithmetic commensurability classes) or rigid (occurring in exactly one)."],"supporting_citations":[{"why":"Supplies the original explicit holonomy representations of all seven Nil-manifold families and the separability argument for arithmetic lattices.","marker":"[M1]"},{"why":"Establishes that every Nil-manifold arises as a cusp cross-section of an arithmetic complex hyperbolic 2-manifold and provides the cover-refinement theorems used in the final step.","marker":"[M2]"},{"why":"Provides the explicit structure of the cusp subgroup for d=3.","marker":"[FP]"},{"why":"Provides the explicit structure of the cusp subgroup for d=1.","marker":"[FFP]"},{"why":"Describes the cusp subgroups for all other squarefree d, supporting the universal occurrence in Theorem 1.","marker":"[PW]"},{"why":"Classifies the known non-arithmetic complex hyperbolic lattices and identifies their cusp groups as G∞(O_1) or G∞(O_3).","marker":"[DPP2]"},{"why":"Shows the non-arithmetic lattices lie in R-reflection generated groups with the stated cusp groups, covering the antiholomorphic case.","marker":"[DFP]"},{"why":"Provides the separability lemma used to promote an immersed cusp to an embedded cusp cross-section without an arithmetic assumption.","marker":"[B]"}],"fun_headline_variants":["Nil 3-manifolds: all cusps, some universal, some rare","Seven Nil shapes, four universal, three rare cusps","Cusp census: every Nil 3-manifold appears non-arithmetically","Nil cusps classified: universal vs single-class shapes","All Nil 3-manifolds realized as non-arithmetic cusps"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof of the non-arithmetic existence theorem (Theorem 4) assumes that the separability-based promotion of an immersed cusp to an embedded cusp cross-section, known for arithmetic lattices, also works for the specific non-arithmetic lattices used here, including those whose cusp stabilizers contain antiholomorphic isometries.","fun_headline_variants_meta":{"raw":{"variants":["Nil 3-manifolds: all cusps, some universal, some rare","Seven Nil shapes, four universal, three rare cusps","Cusp census: every Nil 3-manifold appears non-arithmetically","Nil cusps classified: universal vs single-class shapes","All Nil 3-manifolds realized as non-arithmetic cusps"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00019,"raw_usage":{"total_tokens":1362,"prompt_tokens":988,"completion_tokens":374,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":604,"completion_tokens_details":{"reasoning_tokens":276}},"tokens_in":604,"tokens_out":374,"duration_ms":3940,"temperature":1.0,"reasoning_tokens":276,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T14:27:19.607886+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"The most direct check is to search for a holonomy representation of a 1/4-twist Nil 3-manifold into $G_\\infty(\\mathcal{O}_d)$ for a squarefree $d\\neq 1$ with the required rotational part $\\zeta_4$; the paper's eigenvalue argument says the eigenvalues $1,1,\\zeta_4$ would have to lie in a cubic extension of $\\mathbb{Q}(i\\sqrt{d})$, which is impossible, so any such representation would refute Theorem 2.","supporting_citations":[],"review_version":1}