{"id":"669f237e-9cdd-4299-a321-cf67eebc2d4a","arxiv_id":"2607.07150","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"Four counterexamples to converses of geometric-finiteness implications in round convex projective geometry are constructed using 4-dimensional domains invariant under the irreducible representation of SL₂(ℝ).","lead":"The paper constructs explicit 4-dimensional convex projective domains that serve as counterexamples, proving that four implications between notions of geometric finiteness in round convex projective geometry cannot be reversed. This settles the sharpness of a diagram of implications that the authors established in a prior erratum.","discovery_kind":"unclear","skeptic_critique":{"model":"glm-5.2","headline":"The smoothing lemma (A.2) is the weakest link, but the core geometric arguments are independently verifiable and the concern is narrower than it first appears.","rationale":"The reader's identification of Lemma A.2 as the weakest link is correct and well-targeted. The concern is real but narrow: it does not affect the core geometric arguments (height function estimates in §4–5, volume computation in §7, Gromov-hyperbolicity failure in §6), which are explicit and independently checkable. The smoothing appendix is only needed to upgrade the constructed convex domain from 'convex with C¹ boundary' to 'round' (strictly convex with C¹ boundary). Even if Lemma A.2 had a gap, the main geometric content — that the height function can be made small enough for infinite volume and non-hyperbolicity — would survive; only the 'round' qualifier of the counterexample would be at risk. The proof of Lemma A.2 is substantially self-contained (Facts A.3–A.4 are reproved), and the compactness hypothesis is a standard condition in this setting. The concern is that the verification for the specific application is asserted rather than demonstrated, which is a legitimate gap but one that is very likely fillable. The CONDITIONAL verdict is appropriate: the paper's central claims are sound modulo this verification step, and the concern is precisely scoped enough that a targeted check would resolve it. I do not think the verdict should change from CONDITIONAL, but I note that the correctness risk is lower than 'unknown' — the geometric arguments are verifiable and the only uncertainty is in a standard smoothing technique applied to a well-controlled geometric setting.","tokens_in":29108,"tokens_out":885,"duration_ms":879015,"concrete_test":"Verify that the locally finite cover {U_F/G_F} of ∂C∖Λ/Γ constructed in the proof of Lemma A.2 satisfies the compactness condition on totally geodesic boundary pieces for the specific case of Proposition 5.4. Concretely: check that for the Γ-invariant domain Ω₁ constructed in Proposition 5.4 (convex hull of the graph of v, with balls B_x in O_R), every maximal closed face F of ∂Ω₁∖Λ has compact totally geodesic pieces in its boundary. If any face has non-compact totally geodesic pieces, the inductive smoothing in Lemma A.2 may not converge, and roundness of Ω is unverified.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The reader correctly identifies Lemma A.2 (Appendix A) as the least secure point. The authors acknowledge [CLT18, Prop. 8.3] is false in general (tetrahedron with two vertices removed) and rely on unpublished email confirmation from Cooper that compactness of totally geodesic boundary pieces suffices. However, examining the proof of Lemma A.2 more carefully, the argument is largely self-contained: it uses Fact A.1 (convex hull of C¹ sets has C¹ boundary — elementary), Corollary A.5 (local Hessian-convex smoothing via the M_ε trick from [CLT18, §8], reproved here as Facts A.3–A.4), and an inductive patching argument over a locally finite cover. The compactness hypothesis enters specifically to ensure the inductive limit converges and the cover is locally finite. For the paper's main application (Proposition 5.4), the domain is Γ-invariant with Γ a noncocompact lattice of SL₂(ℝ), and the relevant boundary structure is controlled by the SL₂(ℝ)-orbit decomposition from Lemma 2.1. The real question is whether the local-to-global patching in the proof of Lemma A.2 actually needs the full strength of the unpublished compactness claim, or whether the specific structure here (transitive SL₂(ℝ)-action on ∂_ni C, explicit fundamental domain I) makes the argument self-contained. The proof as written does invoke the compactness condition but does not fully verify it for the application — it is asserted ('This property is indeed verified in [DGK24, §9] and our case below') without an explicit check.","agreement_with_reader":"agree"},"referee_report":{"model":"glm-5.2","summary":"This paper constructs explicit counterexamples showing that four implications in the diagram of geometric finiteness notions for round convex projective orbifolds (established in the authors' erratum [BM] to [CM14]) are sharp. The setting is 4-dimensional: the authors study the irreducible representation ρ: SL₂(ℝ) → SL₅(ℝ), its limit set Λ ⊂ RP⁴, and the convex hull C of Λ. For a noncocompact lattice Γ ⊂ SL₂(ℝ), they prove: (a) all ρ(Γ)-invariant round convex domains are weakly but not strongly geometrically finite; (b) for any ρ(SL₂(ℝ))-invariant round domain O, the convex core C/ρ(Γ) has finite volume and C is Gromov-hyperbolic for the Hilbert metric of O; (c) there exists a ρ(Γ)-invariant round domain Ω (not SL₂(ℝ)-invariant) where C is not Gromov-hyperbolic and the convex core has infinite Hilbert volume. The construction of Ω uses a height function on ∂_ni C/Γ, with lower bounds from convexity (Lemma 4.2, Proposition 4.3) and an upper bound construction (Proposition 5.4) via convex hulls of graphs combined with a smoothing appendix. An appendix provides a Hessian-convex smoothing procedure combining ideas from [DGK24, §9] and [CLT18, §8].","tokens_in":29297,"tokens_out":2162,"duration_ms":183902,"significance":"The paper resolves a natural question left open by the authors' erratum: whether the corrected diagram of geometric finiteness implications is sharp. The counterexamples are concrete and explicit, arising from a well-studied representation (the 5-dimensional irreducible representation of SL₂(ℝ)). The description of all ρ(SL₂(ℝ))-invariant convex domains (Lemma 2.1, Proposition 2.3) and the parametrization via height functions are useful contributions independent of the counterexample application. The infinite-volume construction (Proposition 7.7) involves a clean, parameter-free argument: the height function u_Ω(x) ≤ Ce^{-f(x)/2} suffices, and the volume estimate reduces to a Colbois–Verovic-type bound (Lemma 7.3, Theorem 7.6). The non-hyperbolicity argument (Proposition 6.2) is elegantly simple: unbounded height functions automatically yield non-Gromov-hyperbolic convex cores. The appendix, while not fully self-contained (see major comments), provides a useful synthesis of smoothing techniques from [DGK24] and [CLT18].","major_comments":[{"comment":"Appendix A, Lemma A.2: The smoothing lemma is the load-bearing technical result ensuring the constructed domain Ω in Proposition 5.4 is round (strictly convex with C¹ boundary). The authors acknowledge that the general result [CLT18, Prop. 8.3] 'is not true' (counterexample: tetrahedron with two vertices removed) and state that it 'probably works under the additional assumptions that totally geodesic pieces in the boundary of the universal cover are compact.' They then assert: 'This property is indeed verified in [DGK24, §9] and our case below, and plays an important role.' However, no explicit verification for 'our case below' is provided in the manuscript. The proof of Lemma A.2 uses an inductive patching argument over a locally finite cover (U_{F_k}/G_{F_k})_k, and the compactness condition is needed to ensure local finiteness and convergence of the inductive limit. The authors should","section":null},{"comment":"either (i) add an explicit verification that the compactness hypothesis holds in the application to Proposition 5.4, or (ii) explain why the specific structure here (transitive SL₂(ℝ)-action on ∂_ni C, explicit fundamental domain I from Lemma 2.1, the domain Ω₁ ⊂ O_R) makes the inductive patching argument self-contained without needing the full unpublished compactness claim. As written, the reader must take on faith that the smoothing works, which is unsatisfying for the central construction.","section":null},{"comment":"Proposition 5.4: The proof constructs Ω₁ as the convex hull of the graph of v, then applies Lemma A.2 to obtain Ω ⊂ Ω₁ with Hessian-convex boundary outside Λ. The claim that ∂Ω is differentiable at Λ uses Lemma 4.5, which requires u_{Ω₁}(r(t))e^t → ∞ for rays r(t) → p ∈ Λ. For parabolic p, the argument gives u_{Ω₁}(r(t)) ≥ e^{χ(f(r(t)))} with f(r(t)) = t + O(1) and χ(t) + t → ∞, hence u_{Ω₁}(r(t))e^t → ∞. This is correct, but the role of the condition χ ≤ -log 100 is not clearly explained in the proof of Proposition 5.4 — it is used to ensure the graph of u lies in a uniform neighborhood O_R of C (so that the balls B_x ⊂ O_R have differentiable boundaries), but this logic is only implicit. The authors should make explicit where each hypothesis on χ is used.","section":null},{"comment":"Section 2.5, Remark 2.4: The authors note that [CM14, Prop. 10.6 and 10.7] contained an error (claiming the convex hull of ψ(ℝ) is ℝ³ when it is the parabolic cylinder C_P). They state the error 'does not break the end of the argument.' This is reassuring, but given that the present paper's Theorem 1.1(a) cites [CM14, Prop. 10.6] for the weak-but-not-strong geometric finiteness of O_R/ρ(Γ), the authors should briefly indicate which step of the proof of Prop. 10.6 is affected and why the conclusion survives, so the reader can verify the citation is still valid.","section":null}],"minor_comments":[{"comment":"The notation ∂_ni C (nonideal boundary) is introduced in the introduction but the subscript 'ni' is only explained parenthetically. A brief definition at first use would help readers unfamiliar with [DGK24] terminology.","section":null},{"comment":"Lemma 2.1: The statement uses both PSL₂(ℝ) and SL₂(ℝ) without always specifying which group acts. For instance, the first bullet says 'Each PSL₂(ℝ)-orbit' while the third says 'PSL₂(ℝ) acts transitively' but the equivariant homeomorphism φ is stated for SL₂(ℝ). Consistent use (or an explicit note that ρ factors through PSL₂(ℝ)) would improve clarity.","section":null},{"comment":"Figure 2: The labels Ω_min and Ω_max in the figure caption/figure should be O_min and O_max for consistency with the text.","section":null},{"comment":"Proposition 4.3, proof: 'Following one of these rays will lead you back to a point y of the compact part π⁻¹K in time bounded above by f(x) plus an additive error term, exactly the way this works in the hyperbolic surface with cusps H²/Γ, see Fact 4.4 below.' The reference to Fact 4.4 is slightly misleading — Fact 4.4 is about geodesic chords in horoballs, not directly about returning to the compact part. A more precise explanation of the reduction would help.","section":null},{"comment":"Lemma 7.3: The statement uses Ω_ε for the ε-neighborhood of Ω in the affine chart, but this notation is also used in Lemma 7.4. The proof references 'decreasingness of Ω ↦ Vol_Ω(A)' which should be stated as a separate fact or referenced.","section":null},{"comment":"The bibliography lists [Flé] as 'Preprint, arXiv:2512.00197' — this arXiv ID format (2512) suggests a future date and may be a typo.","section":null},{"comment":"Page 1, line 2 of abstract: 'acted on by ρ(Γ)' — should clarify that Γ is a lattice of SL₂(ℝ) (this is stated later in the abstract but the first mention of Γ is unqualified).","section":null},{"comment":"Theorem 1.1, point 2: '(gf)&(Hyp) does not imply (GF)' — the notation (gf)&(Hyp) is not defined in the list at the end of the introduction. It presumably means (gf) AND (Hyp), but this should be stated.","section":null}],"recommendation":"minor_revision","confidential_remarks":"The paper is a natural and well-motivated sequel to the authors' erratum. The core geometric arguments (Sections 2–7) are solid and independently verifiable. The main concern is the smoothing appendix (Lemma A.2), which relies on an unpublished claim attributed to an email exchange with Cooper. However, the skeptic's analysis is correct that the concern is narrower than it first appears: the specific application has enough structure (transitive SL₂(ℝ)-action, explicit fundamental domain) that the compactness verification should be straightforward to add. This is a minor revision, not major: the central claims do not depend on the full generality of Lemma A.2, only on its application to a very structured case. The paper fits well in the journal's scope (geometric topology / geometric group theory)."},"author_rebuttal":null,"desk_editor":{"model":"glm-5.2","letter":"The paper constructs explicit 4-dimensional counterexamples showing that four implications in the geometric finiteness diagram for round convex projective geometry are sharp — i.e., their converses fail. Two domains suffice: one where the convex core has infinite Hilbert volume and is not Gromov-hyperbolic, and one (the uniform R-neighborhood O_R) where both are finite/nice but strong geometric finiteness still fails. This cleanly closes out the program from the authors' erratum [BM]. The description of all ρ(SL₂(ℝ))-invariant convex domains (Lemma 2.1, Proposition 2.3) is a nice bonus result — they are exactly O_max, O_min, and the O_R — and the height-function parametrization in §4–5 is a natural and effective framework for controlling how close Ω sits to the convex hull C. The volume estimate (Proposition 7.7, via Colbois–Verovic) and the non-hyperbolicity argument (Lemma 6.1, Proposition 6.2) are clean and independently checkable. The authors are upfront about a prior error in [CM14] (Remark 2.4), which is good practice. The one soft spot is Appendix A, specifically Lemma A.2. The authors acknowledge that [CLT18, Prop. 8.3] is false in general and rely on unpublished email confirmation from Cooper that a compactness hypothesis fixes it. They assert this hypothesis holds in their case but don't verify it explicitly. That said, the concern is narrower than it looks: the proof of Lemma A.2 is largely self-contained (the M_ε smoothing trick is reproved as Facts A.3–A.4, the inductive patching is spelled out), and for the main application the boundary structure is simple — ∂_ni C/Γ is a single SL₂(ℝ)-orbit quotient with a compact-plus-cusps decomposition, so local finiteness of the cover is straightforward. The gap is expository, not mathematical. A referee should ask the authors to either verify the compactness condition explicitly for their case or note that the specific structure makes it automatic. This is a solid paper for specialists in convex projective geometry and geometric finiteness. It deserves a serious referee, and I expect the appendix issue to be fixable in revision.","headline":"Solid counterexamples with one expository gap in the smoothing appendix","tokens_in":29814,"tokens_out":2790,"would_cite":true,"duration_ms":110153,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["20F55","20F65","20H10","22E40","51F15","53C50","57M50","57S30"],"pacs":[],"model":"glm-5.2","headline":"Four geometric-finiteness implications fail in dimension 4","keywords":["convex projective geometry","geometric finiteness","Hilbert metric","Gromov-hyperbolicity","lattice representations","convex core","round convex domains"],"falsifier":"If the smoothing lemma (Lemma A.2) fails for the constructed domain—specifically, if the compactness hypothesis on totally geodesic boundary pieces is not actually satisfied—then the domain Ω may not be round, and the counterexamples to implications involving round convex projective orbifolds would not apply.","tokens_in":29284,"feed_emoji":"🎯","tokens_out":1242,"duration_ms":256317,"temperature":0.7,"pith_summary":"The paper constructs explicit 4-dimensional convex projective manifolds that serve as counterexamples to four converses in a previously established diagram of geometric-finiteness implications. The setting is the irreducible 5-dimensional representation of SL₂(ℝ) acting on real projective 4-space, where a lattice Γ in SL₂(ℝ) preserves a family of properly convex domains. The convex hull C of the limit set is fixed, but the ambient domain Ω can be varied. The authors show that for the SL₂(ℝ)-invariant domains O_R (uniform neighborhoods of the minimal domain), the convex core C/Γ has finite Hilbert volume and is Gromov-hyperbolic. They then construct a different Γ-invariant round domain Ω, lying between C and the maximal domain, whose height function decays exponentially in the cusps. This domain is close enough to C that the convex core has infinite Hilbert volume and fails to be Gromov-hyperbolic, yet the action remains weakly geometrically finite. Two counterexample manifolds (using the same group and convex core but different Hilbert metrics) disprove all four converses simultaneously.","feed_headline":"Four geometric-finiteness implications fail in dimension 4","feed_subtitle":"Explicit 4D convex projective counterexamples show that weak geometric finiteness, finite volume, and hyperbolicity of the convex core are独立","key_machinery":"The irreducible representation ρ: SL₂(ℝ) → SL₅(ℝ) on degree-4 homogeneous polynomials; the height function u: ∂_ni C/Γ → (0, 1/2) parametrizing how far the domain boundary lies from the convex hull C; the 1-Lipschitz constraint on log u from convexity (Lemma 4.2); the upper bound u(x) ≤ C·max_i(η_i · e^{-d(x,x_i)}) for convex hulls of graphs (Lemma 5.2); and the smoothing procedure of Appendix A ensuring the constructed domain is round.","core_discovery":"The paper proves that the standard hierarchy of geometric-finiteness notions for round convex projective orbifolds is sharp: four implications that were not previously reversible are genuinely irreversible. The mechanism is a separation between the combinatorial dynamics on the limit set (which determines weak vs. strong geometric finiteness and is independent of Ω) and the metric geometry of the convex core (which depends on Ω). By constructing a Γ-invariant round domain whose boundary approaches the convex hull C exponentially fast in the cusps, the authors make the induced Hilbert metric on C large enough to produce both infinite volume and failure of Gromov-hyperbolicity, while the dynam","pith_inferences":["The counterexamples live in dimension 4, which is the lowest even dimension where the irreducible representation of SL₂(ℝ) yields non-hyperbolic round cusps. Whether analogous separations between geometric-finiteness notions occur in odd dimensions or for other Lie group representations remains open.","The smoothing procedure's reliance on an unpublished compactness hypothesis (acknowledged by the authors) suggests that a complete published proof of the general smoothing lemma would strengthen the foundation not only of this paper but of related constructions in convex projective geometry.","The exponential decay rate e^{-f(x)/2} for the height function is the threshold for infinite volume; understanding whether other decay rates produce intermediate phenomena (e.g., finite volume but non-hyperbolic, or infinite volume but hyperbolic) could refine the classification further."],"forward_implications":["The diagram of geometric-finitess implications for round convex projective orbifolds is now complete: every arrow is either an equivalence or has a known counterexample to its converse.","The separation between dynamical geometric finiteness (independent of Ω) and metric geometric finiteness (dependent on Ω) provides a template for constructing similar counterexamples in other settings where a group preserves multiple invariant convex domains.","The explicit description of all ρ(SL₂(ℝ))-invariant convex domains as the family {O_R : R > 0} ∪ {O_min, O_max} classifies the invariant geometry completely in this representation.","The height-function framework gives a quantitative tool for controlling Hilbert volume and Gromov-hyperbolicity of convex cores in terms of boundary behavior, applicable to other convex projective cusp examples."],"fun_headline_variants":["Four geometric-finiteness implications fail in dimension 4","Geometric-finiteness hierarchy is sharp in 4D convex projective geometry","4D convex domains split limit-set dynamics from convex-core geometry","Counterexamples separate weak and strong geometric finiteness in dimension 4"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The smoothing procedure in Appendix A requires that totally geodesic pieces in the boundary of the universal cover are compact; this condition is verified for the specific domains constructed but the general lemma relies on an unpublished argument.","fun_headline_variants_meta":{"raw":{"variants":["Four geometric-finiteness implications fail in dimension 4","Geometric-finiteness hierarchy is sharp in 4D convex projective geometry","4D convex domains split limit-set dynamics from convex-core geometry","Counterexamples separate weak and strong geometric finiteness in dimension 4"]},"model":"glm-5.2","effort":"low","cost_usd":0.0,"raw_usage":{"total_tokens":692,"prompt_tokens":631,"completion_tokens":61,"prompt_tokens_details":null},"tokens_in":631,"tokens_out":61,"duration_ms":39170,"temperature":1.0,"reasoning_tokens":null,"cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-09T18:47:26.828776+00:00","model_set":{"reader":"glm-5.2"},"falsifier":"If the smoothing lemma (Lemma A.2) fails for the constructed domain—specifically, if the compactness hypothesis on totally geodesic boundary pieces is not actually satisfied—then the domain Ω may not be round, and the counterexamples to implications involving round convex projective orbifolds would not apply.","supporting_citations":[],"review_version":1}