{"id":"b5a705b0-ecf9-4f57-8fab-c1c81a57bfb6","arxiv_id":"1908.11292","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Morse quasi-geodesics behave locally-to-globally in mapping class groups, CAT(0) groups, 3-manifold groups, and relatively hyperbolic groups, giving new combination theorems for stable subgroups.","lead":"This paper defines a local-to-global property for Morse quasi-geodesics and proves that mapping class groups, CAT(0) groups, 3-manifold groups, and certain relatively hyperbolic groups satisfy it. It then uses the property to extend combination theorems for stable subgroups beyond hyperbolic groups.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 5.1 relies on an unproved transfer of the Morse local-to-global property from a peripheral subset to its uniform neighborhood; this is load-bearing for the relatively hyperbolic inheritance proof.","rationale":"The reader identified the same load-bearing gap: the unsupported assertion that a thick neighborhood of a peripheral inherits the Morse local-to-global property. I agree that this is the weakest point in the central Theorem 5.1. I have read the proof of Section 5 carefully. The deep-point machinery (Sections 5.3–5.5) is extensive, and most of the estimates appear coherent: Corollary 5.12, Proposition 5.13, and the linear ordering of relevant peripherals in Proposition 5.30 are plausible and supported by the cited lemmas. However, Corollary 5.23 is the unique step where the peripheral local-to-global property is actually imported. It contains two distinct transfer claims: (a) from P_i to N_{rR}(P_i), and (b) from Morse-ness in N_{rR}(P_i) to Morse-ness in X. The first is nontrivial because induced metrics on neighborhoods are not automatically quasi-isometric to the subspace; the second is nontrivial because quasi-geodesics in X are not constrained to stay in the neighborhood in which the Morse gauge was obtained. Lemma 5.5 gives linear quasiconvexity of peripherals, so a plausible proof of the transfer exists, but the paper does not provide it. Since Lemma 5.33 depends on every α_i being a uniform Morse quasi-geodesic in X, the gap is load-bearing for Theorem 5.1 and for the claimed 3-manifold and relatively hyperbolic consequences. I do not see evidence that the claim is false, and other parts of the paper (CAT(0), HHS, stable-subgroup applications) appear well supported. Therefore a request for clarification is appropriate, but no stronger action is warranted.","tokens_in":48782,"tokens_out":27392,"duration_ms":257293,"concrete_test":"Write out a complete proof of Corollary 5.23: show that for every fixed R, the induced metric on N_R(P_i) is quasi-isometric to P_i with constants depending only on R and the constants of Lemma 5.5, and then verify that this quasi-isometry transfers the Φ–local-to-global property to N_R(P_i). In particular, check the two inequalities for projections: d_{N_R(P_i)}(x,y) ≤ 2R + d_{P_i}(π_{P_i}(x), π_{P_i}(y)) and d_{P_i}(π_{P_i}(x), π_{P_i}(y)) ≤ μ d_{N_R(P_i)}(x,y) + μ, using the fact that geodesics from x to P_i stay in N_{rR}(P_i). If the second inequality cannot be established, or if a relatively hyperbolic space with a peripheral P_i can be found where N_R(P_i) is not quasi-isometric to P_i, then Corollary 5.23 fails and Theorem 5.1's proof is incomplete.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The proof of Theorem 5.1 funnels through Corollary 5.23, which asserts that if a peripheral subset P_i has the Φ–Morse local-to-global property, then a uniformly thick neighborhood N_{rR}(P_i), with the induced metric, has the Ψ–Morse local-to-global property, and that a quasi-geodesic that is Morse in that neighborhood is Morse in X. The text states the first assertion without proof or citation, and the second is only justified by the one-line phrase 'The distance formula (Theorem 5.7) and Lemma 5.5 then imply...'. Neither implication is immediate: a quasi-geodesic in X with endpoints on α_i may leave N_{rR}(P_i), and Lemma 5.5 only forces it into a larger neighborhood N_{r' rR}(P_i), whose local-to-global property has not been established. Corollaries 5.22–5.23 are used in Lemma 5.33 to conclude that every relevant subsegment α_i is a uniform global Morse quasi-geodesic in X, and Lemma 5.33 is the heart of Theorem 5.1. If the neighborhood transfer fails, or if the upgrade from 'Morse in N_{rR}(P_i)' to 'Morse in X' fails, the relatively hyperbolic inheritance proof has no alternative mechanism to control deep subsegments. Weakly supported claims in Section 3 (e.g., the Gitik-style argument in Theorem 3.1) are less central because they do not affect the main theorem's proof.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces a local-to-global property for Morse quasi-geodesics in metric spaces and proves that a wide range of groups and spaces satisfy it: CAT(0) spaces, mapping class groups and Teichmüller spaces, graph products of hyperbolic groups, virtually solvable groups, and fundamental groups of closed 3-manifolds. The main technical contribution is a relatively hyperbolic inheritance theorem (Theorem 5.1): if a geodesic metric space is hyperbolic relative to peripheral subsets each having the Morse local-to-global property, then the ambient space has the property. To prove this, the authors develop a theory of deep points for local quasi-geodesics, decomposing a local quasi-geodesic into alternating deep subsegments running close to peripherals and shallow subsegments with uniformly bounded projections. The paper also derives consequences of the property, including stable-subgroup combination theorems, discreteness of translation lengths for Morse conjugacy classes with a fixed gauge, and a Cartan–Hadamard-type local criterion for hyperbolicity.","tokens_in":89,"tokens_out":5114,"duration_ms":114503,"significance":"If the relatively hyperbolic proof is completed, this is a substantial contribution. The Morse local-to-global property is a natural and useful strengthening of the behavior of Morse quasi-geodesics, and the paper shows it holds for many of the standard groups and spaces in geometric group theory. The deep-points machinery for local quasi-geodesics in relatively hyperbolic spaces is a genuinely new tool that likely has independent applications. The consequences are clean and broad: the combination theorems generalize results of Gitik from hyperbolic groups to stable subgroups, and the applications to mapping class groups and 3-manifold groups are valuable. The paper is careful in setting up uniform gauges and in separating the different ingredients of the proof, and the claimed results are consistent and plausible.","major_comments":[{"comment":"The proof asserts without proof that if a peripheral subset P_i has the Φ–Morse local-to-global property, then a uniform neighborhood N_{rR}(P_i), equipped with the induced metric, has the Ψ–Morse local-to-global property. This transfer is not a formal consequence of the definition: the induced metric on a neighborhood can be very different from the ambient metric, and a subsegment of α_i that is a local Morse quasi-geodesic in X need not be a local Morse quasi-geodesic in N_{rR}(P_i) with the induced metric. Since Corollary 5.23 is used in Lemma 5.33 to conclude that every deep subsegment α_i is a global Morse quasi-geodesic in X, this is a load-bearing step. The authors should provide a proof or a precise citation establishing the transfer, with constants depending only on Φ, λ, and ε, uniformly over the family of peripherals.","section":"§5.4, Corollary 5.23"},{"comment":"Even granting the transfer just discussed, the assertion that 'The distance formula (Theorem 5.7) and Lemma 5.5 then imply each α_i is an (N;k,c)–Morse quasi-geodesic in X' is not justified by the cited results. If β is a quasi-geodesic in X with endpoints on α_i, Lemma 5.5 only guarantees that β lies in N_{r' rR}(P_i) for some larger radius r' depending on the quasi-geodesic constants of β. But Corollary 5.23 establishes Morse control of α_i only inside N_{rR}(P_i), not inside the larger neighborhood. The distance formula cannot by itself bridge this gap without an additional argument showing that the relevant part of β is in fact controlled by α_i. Since the conclusion that each α_i is a global Morse quasi-geodesic in X is essential for Lemma 5.33 and hence for Theorem 5.1, this needs a complete proof.","section":"§5.4, Corollary 5.23, final sentence"},{"comment":"In the proofs of parts (1) and (2), the verification that the concatenated path γ is a local quasi-geodesic is delegated to Gitik's arguments with the statement that 'this argument only uses' certain facts that remain true in the present setting. This is a substantial import of an external proof into a setting involving the Morse local-to-global property, and the specific quasi-geodesic constants (3, ε) and (6, ε) are claimed without reproducing the counting or minimality steps. The reader cannot easily verify that the adapted argument yields exactly those constants. Since Theorem 3.1 is one of the main advertised applications of the paper, the proof should either reproduce the relevant Gitik argument or give a precise reference and state explicitly the lemma being adapted.","section":"§3.1, Theorem 3.1"}],"minor_comments":[{"comment":"The statement says 'if σ0 ˚ α1 ˚ σ1 ˚ ... ˚ αn ˚ σn is the B–relevant decomposition of an pL2;λ,ǫq–quasi-geodesic', but it should read 'pL2;λ,ǫq–local quasi-geodesic'.","section":"§5.3, Corollary 5.22"},{"comment":"In the sentence 'Once we show that γ is an pL;M1;6,cq–local Morse quasi-geodesic', the constant c should presumably be ε, consistent with the rest of the proof.","section":"§3.1, proof of Theorem 3.1(2)"},{"comment":"The reference [Fin] is cited informally as 'Finks the author'; this should be replaced with the full name of the author and a complete bibliographic entry.","section":"§4.1, Example 4.10(1)"},{"comment":"The symbol R is used both for the neighborhood radius in Lemma 5.5 and for the constant R(λ, ε) in Lemma 5.6; although the usage is locally clear, a remark or notational distinction would improve readability.","section":"§5.1"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper introduces the Morse local-to-global property and proves it for many of the usual suspects: CAT(0) spaces, mapping class groups, Teichmuller space, graph products of hyperbolic groups, virtually solvable groups, and closed 3-manifold groups. That alone makes it worth reading. The new concept is natural and the applications are real: Gitik-type combination theorems for stable subgroups, discreteness of translation lengths for fixed Morse gauge, and a local hyperbolicity criterion. The paper also gives examples of groups with Morse geodesics but without the property, which is a useful sanity check.\n\nThe deepest new work is in Section 5, where the authors develop a deep-points decomposition for local quasi-geodesics in relatively hyperbolic spaces. This is a genuine extension of Hruska's theory, and it is the right tool for the inheritance theorem. The structure of the proof is clear, and the lemmas are mostly self-contained.\n\nThe soft spots are real but not fatal. Corollary 5.23 is too terse. It asserts that a uniform neighborhood N_{rR}(P_i) inherits the Morse local-to-global property from P_i, and then that a Morse quasi-geodesic in that neighborhood is Morse in X. The first assertion is true because P_i and its neighborhood are at bounded Hausdorff distance (the projection map gives a quasi-isometry), but the paper says nothing. The second assertion also follows, since any quasi-geodesic with endpoints on N_{rR}(P_i) stays in a slightly larger neighborhood by the linear quasiconvexity of peripherals, and that larger neighborhood is again quasi-isometric to P_i. Both steps deserve at least a sentence of justification. Similarly, Theorem 3.1 compresses Gitik's argument by reference rather than reproducing it; that is acceptable in a research paper, but it leaves a nontrivial verification to the reader.\n\nI do not think the central argument is broken. The reader's worry about the neighborhood transfer is a real gap in exposition, not a load-bearing flaw in the mathematics. The paper is careful elsewhere, and the pieces needed to fill the gap are already in the paper (Lemmas 5.4, 5.5, and the distance formula).\n\nThis is a paper for anyone working on Morse geodesics, stable subgroups, or relatively hyperbolic groups. It deserves a serious referee. I would send it out, with the advice to ask the authors to expand the proof of Corollary 5.23 and add a remark in Theorem 3.1 about where Gitik's argument is being imported.","headline":"A new local-to-global property for Morse quasi-geodesics, proved for a broad class of groups; the relative hyperbolicity proof has one compressed step that is more an exposition gap than a mathematical flaw.","tokens_in":49612,"tokens_out":5333,"would_cite":true,"duration_ms":49435,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["20F65","20F67","57M07"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that Morse quasi-geodesics satisfy a local-to-global principle in a wide class of groups and spaces, including mapping class groups, CAT(0) spaces, and closed 3-manifold groups.","keywords":["Morse quasi-geodesics","local-to-global property","relatively hyperbolic groups","stable subgroups","mapping class group","CAT(0) spaces","3-manifold groups","deep points"],"falsifier":"Check the transfer in Corollary 5.23 directly: take a peripheral subset $P$ with the $\\Phi$-Morse local-to-global property and ask whether every uniform thickening $N_{rR}(P)$ with the induced metric has a $\\Psi$-Morse local-to-global property with $\\Psi$ depending only on $\\Phi$, $\\lambda$, and $\\epsilon$. A concrete counterexample, or a local Morse quasi-geodesic in a relatively hyperbolic space whose deep subsegment stays in such a thickening but is not a global Morse quasi-geodesic, would settle that the proof of Theorem 5.1 fails at this step.","tokens_in":48584,"feed_emoji":"📐","tokens_out":7272,"duration_ms":65637,"temperature":0.7,"pith_summary":"The paper establishes a local-to-global principle for Morse quasi-geodesics: if a path in a geodesic metric space has all of its subpaths of some uniform length as Morse quasi-geodesics, then the whole path is a Morse quasi-geodesic. It proves that the principle holds for CAT(0) spaces, mapping class groups, Teichmuller space, graph products of hyperbolic groups, virtually solvable groups, and fundamental groups of closed 3-manifolds. The proof's engine is a relative hyperbolicity inheritance theorem: a space hyperbolic relative to peripheral subsets inherits the property from those subsets. The argument introduces a \"deep points\" decomposition of local quasi-geodesics into pieces that alternately hug a peripheral subset and avoid all peripherals. If right, a suite of hyperbolic-space theorems, including stable subgroup combination theorems, discreteness of Morse translation lengths, and a local criterion for hyperbolicity, transfer to all these groups.","feed_headline":"Local Morse paths become global Morse in many 3-manifold groups","feed_subtitle":"The same local-to-global rule then covers mapping class groups, CAT(0) spaces, and closed 3-manifold groups.","key_machinery":"The load-bearing mechanism is a theory of deep points for local quasi-geodesics in relatively hyperbolic spaces, extending Hruska's deep points for geodesics. A point in the domain is $P$-deep if witnesses on both sides run within a fixed neighborhood of a peripheral $P$ for a nontrivial length of time; deep points for different peripherals cannot overlap, so a local quasi-geodesic decomposes as $\\sigma_0 * \\alpha_1 * \\sigma_1 * \\cdots * \\alpha_n * \\sigma_n$. Each $\\alpha_i$ is a long subsegment running near a single peripheral $P_i$, and each $\\sigma_i$ has uniformly bounded projections to every peripheral, hence is a Morse quasi-geodesic. When each peripheral has the Morse local-to-global property, each deep $\\alpha_i$ is likewise a global Morse quasi-geodesic, via the claimed inheritance of the property by thick neighborhoods. A linear-ordering argument for the relevant peripherals then forces any quasi-geodesic with the same endpoints to pass close to all the $\\alpha_i$, so the whole local path is Morse.","core_discovery":"The central claim is Theorem 5.1: if a geodesic metric space $X$ is hyperbolic relative to a collection $\\mathcal{P}$ of peripheral subsets and every $P\\in\\mathcal{P}$ has the $\\Phi$-Morse local-to-global property, then $X$ has the $\\Psi$-Morse local-to-global property. A local Morse quasi-geodesic is a map whose subsegments of parametrized length at most $L$ are all $M$-Morse quasi-geodesics, and the property says that for every $M,\\lambda,\\epsilon$ there is a scale $L$ such that every such local path is a global Morse quasi-geodesic. The authors claim the inheritance is uniform: the constants for $X$ depend only on the constants for the peripherals. Together with known cases, this yields the broad list of examples in Theorem D: CAT(0) spaces, mapping class groups, Teichmuller space, graph products of hyperbolic groups, virtually solvable groups, and fundamental groups of closed 3-manifolds all have the Morse local-to-global property.","pith_inferences":["One implicit consequence is that any group hyperbolic relative to Morse limited subgroups is Morse local-to-global, so the theorem automatically covers many groups built from unconstricted pieces, widening the class beyond the listed examples.","The deep-point decomposition may give a tool for transferring coarse contracting or divergence properties from peripherals to the ambient space, which could address open questions about whether Morse quasi-geodesics in Morse local-to-global spaces have stronger contracting or divergence behavior.","The unproved thick-neighborhood transfer in Corollary 5.23 is a testable gap: if a peripheral $P$ has the property but a uniform thickening $N_{rR}(P)$ with the induced metric does not, then the inheritance theorem would need a different argument and might fail."],"forward_implications":["In any finitely generated Morse local-to-global group, stable subgroups satisfy Gitik-style combination theorems: under a short-element intersection condition, the subgroup they generate is their amalgamated free product and is itself stable (Theorem G).","In the mapping class group this yields combination theorems for convex cocompact subgroups, since stable subgroups there are exactly the convex cocompact ones (Theorem M).","Every infinite normal subgroup of a Morse local-to-global group that contains a Morse element must itself contain a Morse element (Corollary 3.6).","For each fixed Morse gauge, the set of translation lengths of $M$-Morse conjugacy classes is a discrete set of rational numbers (Theorem 3.12).","A geodesic metric space with the property that all sufficiently large local triangles are uniformly slim is globally hyperbolic (Theorem 3.15); in particular the universal covers of closed 3-manifolds and CAT(0) spaces pass this local check."],"supporting_citations":[{"why":"Supplies the hyperbolic combination theorems for quasiconvex subgroups that Theorem G extends to stable subgroups.","marker":"[Git99]"},{"why":"Identifies stable subgroups of the mapping class group with convex cocompact subgroups, making Theorem M meaningful.","marker":"[DT15]"},{"why":"Introduces deep points for geodesics in relatively hyperbolic spaces, which Section 5 generalizes to local quasi-geodesics.","marker":"[Hru10]"},{"why":"Provides projections onto peripheral subsets, the bounded quasi-geodesic image lemma, and the distance formula used throughout Section 5.","marker":"[Sis13]"},{"why":"Supplies the isolated-peripheral and unparametrized quasi-geodesic facts and the projection lemma used to bound projections of the non-deep segments.","marker":"[Sis]"},{"why":"Characterizes Morse geodesics in CAT(0) spaces via contracting projections, which is what proves the CAT(0) case.","marker":"[CS15]"},{"why":"Shows hierarchically hyperbolic spaces with bounded domain dichotomy are Morse detectable, yielding the mapping class group and Teichmuller cases.","marker":"[ABD21]"},{"why":"Shows groups satisfying a law or with infinite cyclic center are unconstricted, hence Morse limited, covering virtually solvable pieces of 3-manifold groups.","marker":"[DS05]"},{"why":"Provides the local-to-global theorem for quasi-geodesics in hyperbolic spaces and the local hyperbolicity criterion that the paper adapts.","marker":"[Gro87]"}],"fun_headline_variants":["Local Morse paths become global Morse in many 3-manifold groups","Morse local-to-global holds for CAT(0), mapping class, and 3-manifold groups","Combination theorems for stable subgroups via Morse local-to-global","Local Morse quasi-geodesics are global in relatively hyperbolic groups","Morse local-to-global: new combination theorems and hyperbolicity detection"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the Morse local-to-global property automatically passes from each peripheral subset to every uniformly thick neighborhood of it with the induced metric; the proof uses this transfer without proving it.","fun_headline_variants_meta":{"raw":{"variants":["Local Morse paths become global Morse in many 3-manifold groups","Morse local-to-global holds for CAT(0), mapping class, and 3-manifold groups","Combination theorems for stable subgroups via Morse local-to-global","Local Morse quasi-geodesics are global in relatively hyperbolic groups","Morse local-to-global: new combination theorems and hyperbolicity detection"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001274,"raw_usage":{"total_tokens":5196,"prompt_tokens":916,"completion_tokens":4280,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":532,"completion_tokens_details":{"reasoning_tokens":4179}},"tokens_in":532,"tokens_out":4280,"duration_ms":29095,"temperature":1.0,"reasoning_tokens":4179,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:20:29.557654+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Check the transfer in Corollary 5.23 directly: take a peripheral subset $P$ with the $\\Phi$-Morse local-to-global property and ask whether every uniform thickening $N_{rR}(P)$ with the induced metric has a $\\Psi$-Morse local-to-global property with $\\Psi$ depending only on $\\Phi$, $\\lambda$, and $\\epsilon$. A concrete counterexample, or a local Morse quasi-geodesic in a relatively hyperbolic space whose deep subsegment stays in such a thickening but is not a global Morse quasi-geodesic, would settle that the proof of Theorem 5.1 fails at this step.","supporting_citations":[],"review_version":1}