{"id":"618765f7-6ed7-48a9-85e5-b544e30d27db","arxiv_id":"2607.13950","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A new combination theorem for Chabauty paths yields a continuous path through Kleinian groups whose groups are pairwise non-isomorphic, and a decomposition theorem (chromatography) splits convex cocompact paths into finitely many isomorphism-bump paths.","lead":"This paper studies continuous paths through the space of hyperbolic 3-manifolds viewed as discrete groups, proving new tools to combine and decompose such paths. It constructs a path along which every group is a different isomorphism type, and proves a 'chromatography' decomposition theorem for convex cocompact paths.","discovery_kind":"new_method","skeptic_critique":null,"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies continuous paths in the Chabauty space D_n of torsion-free discrete subgroups of Isom^+(H^n). It proves a combination theorem for paths (Theorem B), uses it to construct a 'Binary Path' Φ:[0,1]→D_3 along which the isomorphism type is injective as a function of t, with the binary expansion of t encoded in the torsion of the abelianization of Φ(t) (Theorem A), and introduces 'chromatography', a decomposition theorem for convex cocompact paths in D_3 (Theorem C). A final example shows that not every convex cocompact path freely decomposes into isomorphism bump paths.","tokens_in":27939,"tokens_out":31482,"duration_ms":294799,"significance":"If the stated results are correct, Theorem A is a striking contribution to the study of the Chabauty topology: it gives a continuous path through pairwise non-isomorphic Kleinian groups, with an explicit and checkable invariant. Theorem B is a natural and potentially useful combination tool for paths, and Theorem C introduces an appealing structural framework for decomposing convex cocompact paths. The paper is ambitious, carefully written, and the constructions are concrete. However, two load-bearing gaps—one in the proof of Theorem A and one in the dependence on the author's unpublished preprint—need to be addressed before the paper can be accepted.","major_comments":[{"comment":"The sentence 'Since each Γ_{p_n} has non-empty domain of discontinuity, for any m∈Z there is a PSL_2C conjugate Γ_{n,m}^{p_n} of Γ_{p_n} such that ψ(H^3−A_{n,m})⊂A_{n,m} for all ψ≠id' is unproved. This is exactly hypothesis (3) of Theorem B, and without it the paths B_{n,m} cannot be shown to satisfy the hypotheses of the Combination Theorem; the continuity of Φ is therefore not established. The statement is plausible—one can choose a point in the domain of discontinuity with trivial stabilizer and a sufficiently small boundary disk avoiding its orbit, then conjugate—but the manuscript needs a proof or a precise reference. As written this is a genuine gap in the proof of the paper's central theorem.","section":"Section 3.2, proof of Theorem A"},{"comment":"The proof of Theorem C relies on the tracking maps J_{s,t} and their properties, imported as Proposition 4.2 and Lemma 4.3 from the author's unpublished preprint [34] (see also Lemma 4.5's use of '[34, Prop. 4.3]'). These are the core machinery for identifying free factors along paths. Without a proof or a published, refereed version of [34], the decomposition theorem cannot be verified from this manuscript alone. Please include the needed statements and proofs, or provide a complete reference to a refereed source.","section":"Sections 4.1–4.2"},{"comment":"The proof of Proposition 4.9 uses a corrected version of Bowditch's theorem stated only in a footnote: CC(M)⊂M^{≤η}∪W∪N_η(C). The footnote says this follows by 'filling 2-spheres...', but no proof is supplied. Since this inclusion is used to show that an embedded thick convex core gives a free factor, it is load-bearing. Please provide a detailed proof of the modified assertion or a reference where it is proved.","section":"Section 4.2, Proposition 4.9 (footnote 2)"}],"minor_comments":[{"comment":"The path B:R→D_3 is only defined for t∈[0,1] (with B(0)=B(1)={id}); outside [0,1] it should be set to {id} to match the statement.","section":"Example 3.3"},{"comment":"The displayed formula Ab_tor(Φ(t)) ∼= ∏ Ab_tor(B_{n,m}(t)) is misleading: the abelianization of an infinite free product is the direct sum of the abelianizations, not the Cartesian product. The subsequent counting arguments should be phrased for the direct sum (or 'restricted direct product').","section":"Theorem A proof"},{"comment":"The proof chooses a point q∈H^n∖∪ A_P and uses that this complement is non-empty and open. The hypotheses should explicitly include that the complement of ∪ A_P is non-empty, or that the union is a proper closed subset.","section":"Theorem B statement"},{"comment":"The proof appeals to 'standard techniques (see [11])' to show that {A,B,C} is a classical Schottky generating set. Please provide enough detail to make the lemma checkable, or indicate the exact statement in [11] that yields it.","section":"Lemma 4.17"}],"recommendation":"major_revision","confidential_remarks":"The paper is promising, but the two gaps in the central proofs are real. The heavy reliance on the author's own unpublished preprint [34] should be addressed in any revision, possibly by adding an appendix with the tracking results. The half-space conjugacy assertion in Theorem A is likely true but needs a proof. I would not recommend rejection; the results are plausible and potentially significant."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a genuinely novel paper. It proves a combination theorem for continuous paths in the Chabauty topology, then uses it to construct a path in D_3 along which every group is pairwise non-isomorphic, with the binary expansion of t recoverable from torsion in abelianizations. That's striking. It also proves a decomposition theorem for convex cocompact paths and supplies a counterexample to the stronger decomposition one might hope for. The core ideas are good, and the exposition is mostly careful.\n\nThe Combination Theorem (Theorem B) is the real new tool. The proof uses a displacement inequality (Claim 2.1) and a compactness lemma to pass from a sequence of reduced words to a limit word; I followed the argument and it is coherent. Example 2.5, showing condition (2) is necessary, is a nice touch. The Binary Path construction is clever: it combines many 'bumps' indexed by primes and half-spaces, and the abelianization torsion detects which bumps are present, so the isomorphism type encodes t. That part is sound up to one gap. The proof asserts without argument that for each convex cocompact group with nonempty domain of discontinuity and each half-space A_{n,m}, one can conjugate the group so that every nontrivial element sends the complement of A_{n,m} into A_{n,m} — exactly hypothesis (3) of Theorem B. Continuity of the Binary Path rests on this. It may well be true, but it is not proven and it is not a standard one-liner. This needs fixing.\n\nThe decomposition section is structurally interesting. Theorem C itself is stated carefully: it decomposes a subpath D, not the whole path P. The abstract's use of 'characterizes' is therefore too strong, especially since Example 4.16 shows a convex cocompact path that does not freely decompose into isomorphism bump paths. The bigger issue is that the tracking formalism (Proposition 4.2, Lemma 4.3, 'strong uniqueness') is imported wholesale from the author's own unpublished preprint [34]. A referee will need to verify those lemmas; they carry a lot of weight. I did appreciate the honest treatment of Question 4.15.\n\nVerdict: send it to peer review. The main theorems are new and the work is serious. The referee should chase down the half-space conjugacy claim and the [34] dependencies. With those resolved, this would be a solid paper for AGT or Geometriae Dedicata.","headline":"Genuinely new path-combination theorem plus an exotic pairwise-non-isomorphic path in D_3; the main gap is an unproved ping-pong conjugacy assertion, and the decomposition section leans heavily on the author's unpublished preprint.","tokens_in":28281,"tokens_out":2810,"would_cite":true,"duration_ms":45920,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["30F40","57M50"],"pacs":[],"model":"deepseek-v4-flash","headline":"A continuous path in hyperbolic 3-space exists along which every group is a different isomorphism type.","keywords":["Chabauty topology","Kleinian groups","combination theorem","binary path","isomorphism bump path","chromatography","convex cocompact","free decomposition"],"falsifier":"For one prime p from Lemma 3.2, take the non-cyclic convex cocompact group Γ_p with H_1 = Z^2⊕Z/p and try to find a round disk D such that every nonidentity element maps the complementary disk into D. A concrete computation: choose two noncommuting elements a and b and test the commutator [a,b] on a point far outside D; if [a,b] sends that point back outside D for every candidate D, then the asserted conjugates cannot exist, and Theorem A's continuity argument collapses at that step.","tokens_in":27792,"feed_emoji":"🌀","tokens_out":25554,"duration_ms":239353,"temperature":0.7,"pith_summary":"This paper aims to show that the Chabauty space of torsion-free discrete subgroups of hyperbolic 3-space is far larger than the classical deformation spaces of a fixed group. Its headline result is a continuous path Φ:[0,1]→D_3 along which Φ(t) is isomorphic to Φ(s) if and only if t=s; the binary expansion of t is encoded in the torsion of the abelianization of Φ(t). The proof rests on a new combination theorem for paths, which assembles smaller 'isomorphism bump' paths into a continuous path when the pieces satisfy a half-space ping-pong condition. The paper also introduces 'chromatography,' a decomposition theorem showing that any path of convex cocompact groups in D_3 has a companion path that starts at P(0), stays a free factor at every time, and decomposes into finitely many isomorphism bump paths. A final example shows that not every convex cocompact path itself admits such a free decomposition into bump paths.","feed_headline":"One path through hyperbolic 3-space with no two groups isomorphic","feed_subtitle":"Binary expansion of t is coded in homology torsion, so the isomorphism type identifies the parameter.","key_machinery":"The load-bearing object is the path-level Combination Theorem. Given a locally finite collection of paths P:I→D_n and disjoint closed sets A_P with uniform separation, if every nonidentity element ψ∈P(s) satisfies ψ(H^n−A_P)⊂A_P, then the union of the groups P(s) generates a discrete group Π(s) for each s, and s↦Π(s) is continuous in the Chabauty topology. The proof works by tracking reduced words: a distance inequality bounds word length uniformly near a limit, so convergent elements in the generated groups must be products of convergent elements in the constituent paths. Around this core, the paper develops tracking maps that carry elements of P(0) through a path until they 'go to infinity","core_discovery":"The central claim is Theorem A (the Binary Path): there is a continuous map Φ:[0,1]→D_3 such that Φ(t)≅Φ(s) iff t=s. The construction takes a sequence of primes p_n and, for each n and a range of integers m, attaches a convex cocompact group whose first homology contains a Z/p_n factor; as t moves, these factors switch on and off in a pattern that reproduces the binary expansion of t. The torsion of the abelianization of Φ(t) therefore determines t uniquely. Continuity is proved by a combination theorem for paths: a locally finite family of paths, each carrying a disjoint closed half-space and each nonidentity element sending the complement of its half-space into that half-space, generates a","pith_inferences":["The binary/torsion coding is a template for other injective paths: any sequence of finite abelian groups that can be realized as homology torsion of convex cocompact Kleinian groups should yield a path with pairwise non-isomorphic groups, as long as the half-space ping-pong condition can be met.","The cardinality contrast with D_2 (only countably many isomorphism types along any path) suggests that the binary path is essentially a 3-dimensional phenomenon; a natural question the author leaves implicit is whether a path with pairwise non-isomorphic groups can be built in D_n without homology coding, using only geometry of the limit sets.","The unproved conjugation assertion in Section 3.2 is the step most worth stress-testing: if some Γ_{p_n} cannot be conjugated to send the complement of a single half-space into that half-space, then the Binary Path's continuity argument would need a new ingredient, and the theorem would survive only in a weaker form."],"forward_implications":["If Theorem A is correct, the Chabauty space D_3 contains a continuous path with pairwise non-isomorphic groups, and the same embedding of H^3 into H^n makes such paths exist in every D_n, n≥3.","The Binary Path shows that isomorphism type is a complete invariant along this particular path: the torsion of the abelianization recovers the parameter t exactly.","The path-level Combination Theorem turns any sufficiently separated family of isomorphism bump paths into a continuous path, so it gives a general construction method for exotic paths in D_n.","Theorem C guarantees that every convex cocompact path P in D_3 has a companion path D starting at P(0) which stays a free factor and splits into finitely many isomorphism bump paths.","Example 4.16 shows the stronger statement — that every convex cocompact path itself freely decomposes into isomorphism bump paths — is false in general."],"fun_headline_variants":["Binary code in homology torsion marks each point on a path","Path in hyperbolic 3-space where every group is unique","Torsion fingerprints: one group per t on a continuous path","Non-isomorphic groups along a continuous path in Kleinian space","Coding t in torsion: a path of distinct hyperbolic groups"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The Binary Path construction assumes without proof that each convex cocompact group Γ_{p_n} can be conjugated so that a single fixed half-space A_{n,m} absorbs the complement under every non-identity element; this half-space ping-pong condition is what makes the Combination Theorem apply, so if such conjugates do not exist the path is not known to be continuous.","fun_headline_variants_meta":{"raw":{"variants":["Binary code in homology torsion marks each point on a path","Path in hyperbolic 3-space where every group is unique","Torsion fingerprints: one group per t on a continuous path","Non-isomorphic groups along a continuous path in Kleinian space","Coding t in torsion: a path of distinct hyperbolic groups"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000177,"raw_usage":{"total_tokens":1066,"prompt_tokens":616,"completion_tokens":450,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":360,"completion_tokens_details":{"reasoning_tokens":365}},"tokens_in":360,"tokens_out":450,"duration_ms":5045,"temperature":1.0,"reasoning_tokens":365,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T03:19:18.811479+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For one prime p from Lemma 3.2, take the non-cyclic convex cocompact group Γ_p with H_1 = Z^2⊕Z/p and try to find a round disk D such that every nonidentity element maps the complementary disk into D. A concrete computation: choose two noncommuting elements a and b and test the commutator [a,b] on a point far outside D; if [a,b] sends that point back outside D for every candidate D, then the asserted conjugates cannot exist, and Theorem A's continuity argument collapses at that step.","supporting_citations":[],"review_version":1}