{"id":"b05badac-66ae-404b-a0ae-a4a19a18a87b","arxiv_id":"1908.08214","paper_version":3,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Irreducible nonsurjective endomorphisms of free groups are fully irreducible and their mapping tori are word-hyperbolic.","lead":"This paper proves that certain self-maps of free groups, called irreducible nonsurjective endomorphisms, always produce 'mapping torus' groups that are word-hyperbolic, a strong form of geometric tameness. The result matters because it unifies existing ideas about when such groups are hyperbolic and may help classify them.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Central claim rests on unproved external criterion (Thm 6.2 of [15]); internal reduction from clean immersion to absence of a^d is sound.","rationale":"The reader's verdict ACCEPT with moderate confidence matches my reading. The internal proof is coherent: Theorem 4.5 supplies clean immersions for irreducible nonsurjective endomorphisms; the forest argument and Proposition 5.4 are intricate but plausible; no internal contradiction emerged. The only genuinely load-bearing unproved input is Theorem 6.2, which is external and not reproduced here. Since the paper explicitly cites [15] for this criterion, the central claim is conditional on that theorem's correctness; this is a dependency, not a demonstrated flaw, and it does not change the verdict. I therefore agree with the reader's weakest_assumption and recommend UNCHANGED.","tokens_in":17899,"tokens_out":17340,"duration_ms":157261,"concrete_test":"Inspect [15] (arXiv:1809.04761), Theorem 6.3, and re-derive the forward direction: assume f:Γ→Γ is an immersion with no a,d,n satisfying φ^n(a)∼a^d, and prove F∗φ hyperbolic using the Bestvina–Feighn hyperbolicity criterion or the equivalent used there. Verify that each step applies to all immersions, and note precisely where the \"no a^d\" hypothesis is used. If the proof invokes cleanliness, irreducibility, or an additional expanding property without proving it from the immersion hypothesis, then Theorem 6.3 requires a new argument and the verdict should be conditional.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The proof of Theorem 6.3 converts hyperbolicity of F∗φ into the non-existence of a nontrivial a with φ^n(a) conjugate to a^d, via Theorem 6.2 from [15]. The present paper establishes (Thm 4.5 plus the argument inside Thm 6.3) that a clean immersion representing an irreducible nonsurjective endomorphism expands every nontrivial conjugacy class, so it has no such a. The single step not proved here is the forward direction of Theorem 6.2: for an arbitrary immersion, absence of such an a is sufficient for F∗φ to be word-hyperbolic. This is a strong \"if and only if\" for all immersions, not just clean or irreducible ones. If Theorem 6.2 has a hidden hypothesis (for instance, that the immersion be clean, or that the endomorphism be atoroidal in a stronger sense), or if its proof in [15] contains a gap, then Theorem 6.3 fails even though the internal reduction is correct. The reader's weakest-assumption field correctly identifies this same dependency.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves several results about irreducible nonsurjective endomorphisms of finitely generated free groups. It gives a new proof that such an endomorphism can be represented by a unique clean immersion (Theorem 4.5), proves a partial converse for clean immersions whose Whitehead graphs have no cut vertices (Theorem 4.6), characterizes invariant finitely generated subgroups (Propositions 5.3 and 5.4), characterizes fully irreducible injective endomorphisms (Theorem 5.5), and derives that irreducible nonsurjective endomorphisms are fully irreducible (Corollary 5.6). The main hyperbolicity result (Theorem 6.3) states that if an endomorphism is represented by a clean immersion then its mapping torus is word-hyperbolic; in particular, the mapping torus of an irreducible nonsurjective endomorphism is word-hyperbolic. The proof relies on a previously stated hyperbolicity criterion for mapping tori of immersions (Theorem 6.2, from the author's earlier work).","tokens_in":18122,"tokens_out":5201,"duration_ms":54311,"significance":"If the results are correct, the paper makes a substantial contribution to the study of nonsurjective endomorphisms of free groups. It extends train-track and outer-space methods from automorphisms to irreducible nonsurjective endomorphisms, establishes that irreducibility implies full irreducibility in this setting, and proves hyperbolicity of the corresponding mapping tori, generalizing the free-by-cyclic case. The proof of Theorem 4.5 is a genuine new proof of a result previously announced by Reynolds, and the treatment of invariant subgroups in Section 5 is careful and appears to be new. The paper also gives a quick proof that the Sapir group is word-hyperbolic. The technical machinery is standard and the arguments are mostly detailed. The main caveat is that the final hyperbolicity theorem depends on the author's earlier criterion in Theorem 6.2, which is invoked without a full statement or proof.","major_comments":[{"comment":"The proof of Theorem 6.3 invokes Theorem 6.2 of [15] as a black box. Since this external criterion supplies exactly the implication 'absence of roots implies word-hyperbolicity' that the argument needs, the central conclusion of the paper is conditional on Theorem 6.2. The manuscript should include a complete statement of Theorem 6.2, clarify its publication status, and explicitly verify that all hypotheses of that theorem are satisfied by clean immersions. If Theorem 6.2 has any hidden assumptions, they must be stated so that the reader can check them.","section":"Section 6, Theorem 6.3"}],"minor_comments":[{"comment":"The abstract states several results without theorem numbers; please add references to Theorems 4.5, 4.6, 5.5, and 6.3 to help the reader navigate.","section":"Abstract and Introduction"},{"comment":"The sentence 'Since there are finitely many combinatorially distinct ways to fold a forest, there are fixed j > i and k ≥ 1 such that the composition of folds S_j → S_{j+k} is homotopic to a homeomorphism' is terse; the finiteness argument should be expanded, since the number of folds may depend on the number of edges in S_j.","section":"Theorem 4.5 proof, Section 4"},{"comment":"The final contradiction in the proof of Theorem 4.6 asserts that the induced map on Δ∞ would have disconnected Whitehead graphs or Whitehead graphs with cut vertices; this step is not fully justified and would benefit from a more explicit explanation of which identified vertices produce the cut vertices.","section":"Theorem 4.6 proof, Section 4"},{"comment":"In Example 4.7, the conclusion that φ_A is fully irreducible is derived from word-hyperbolicity of its mapping torus together with the fact that A has rank two; this implication is not immediate and should be explained explicitly.","section":"Example 4.7"},{"comment":"There are several typographical issues: 'eventially' should be 'eventually' in Proposition 5.3, the affiliation line contains an extra space in 'F ayetteville', and 'c.f.' should be 'cf.'.","section":"Throughout"}],"recommendation":"minor_revision","confidential_remarks":"The paper is solid and the internal reduction in Section 6 is convincing. My only substantive concern is the reliance on Theorem 6.2 of [15], which is not stated in full and is listed as 'to appear'; once that dependency is made explicit and verified, I would be happy to see the paper accepted."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The headline result here is genuine: every irreducible nonsurjective endomorphism of a free group has a word-hyperbolic mapping torus. The proof goes through a clean structure theorem—such endomorphisms have a unique clean immersion representative—and gives a nice corollary that irreducibility implies full irreducibility. The paper also supplies a new proof of Reynolds' immersion theorem, which was previously unpublished, and a useful characterization of fully irreducible injective endomorphisms in Theorem 5.5. I was not expecting all of this to be packaged in one paper, and the main results are new and significant.\n\nThe exposition is careful and honest. The action on outer space is treated properly, the limiting tree argument in Theorem 4.5 is spelled out, and the subgroup rigidity result in Section 5 is a real contribution. The examples, especially the Sapir group, are well chosen and actually illustrate the scope of the theorems.\n\nThe real soft spots are the dense folding arguments in Theorems 4.5 and 4.6. I could not find a gap, but these are hard to check in a quick read. The other structural dependency is Theorem 6.2 from the author's own previous paper [15]: the main hyperbolicity theorem relies on an \"if and only if\" criterion for all immersions that is not reproved here. That is a legitimate reliance on prior work, not a circularity, and the reduction from a clean immersion to the absence of periodic powers is sound. If [15] has a hidden hypothesis, the main theorem would fail, but there is no evidence of that from this paper; it is a normal conditional result.\n\nOne weaker point: Theorem 4.6 gives a criterion for full irreducibility that the author immediately shows is not necessary, which is fine, but it means the title theorem is not derived from that criterion and the paper is slightly less unified than the introduction suggests.\n\nWho should read this? Geometric group theorists working on free group endomorphisms, outer space, or mapping tori of free groups. The paper deserves a serious referee; I would send it to a strong journal and recommend the referee focus on the folding arguments and on verifying the applicability of [15, Thm 6.3]. I would cite this if I worked in the area.","headline":"This paper proves a long-standing conjecture for nonsurjective free group endomorphisms with clean, careful arguments; the one external dependency is explicit and not a hidden flaw.","tokens_in":18621,"tokens_out":2776,"would_cite":true,"duration_ms":29027,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["20F65","20E05","20F67","20E36"],"pacs":[],"model":"deepseek-v4-flash","headline":"Every irreducible nonsurjective endomorphism of a finitely generated free group has a word-hyperbolic mapping torus, because such endomorphisms admit a unique clean immersion representative and cannot carry periodic conjugacy classes.","keywords":["free groups","irreducible endomorphisms","fully irreducible endomorphisms","mapping tori","word-hyperbolic groups","train track maps","outer space","Whitehead graphs"],"falsifier":"Exhibit a clean immersion representing an irreducible nonsurjective endomorphism for which some nontrivial $a\\in F$ and integers $d,n\\ge 1$ satisfy $\\varphi^n(a)=g a^d g^{-1}$; by the paper's own criterion (Theorem 6.2) the mapping torus would then fail to be word-hyperbolic, disproving Theorem 6.3. In a concrete example this is a finite check: if the word length of $\\varphi^n(a)$ grows linearly for some nontrivial $a$, such a periodic class exists, and the predicted hyperbolicity fails.","tokens_in":17726,"feed_emoji":"🍩","tokens_out":15717,"duration_ms":132257,"temperature":0.7,"pith_summary":"This paper proves that every irreducible nonsurjective endomorphism of a finitely generated free group has a word-hyperbolic mapping torus—the ascending HNN extension $F*_\\varphi = \\langle F,t \\mid t^{-1}xt=\\varphi(x)\\rangle$ has a hyperbolic Cayley graph. More generally, the conclusion holds for every endomorphism represented by a clean immersion of a finite core graph. The proof shows that such an endomorphism has a unique irreducible immersion representative, that irreducibility implies full irreducibility in the nonsurjective case, and that any finitely generated invariant subgroup carrying an expanding conjugacy class has finite index in some iterated image. A prior hyperbolicity criterion then reduces the task to excluding periodic conjugacy classes, which the invariant-subgroup result accomplishes. If correct, the theorem turns a broad family of non-invertible free-group maps into a source of hyperbolic groups.","feed_headline":"Irreducible nonsurjective free-group maps have hyperbolic mapping tori","feed_subtitle":"Non-invertible free-group maps can still produce hyperbolic groups; no periodic conjugacy classes can exist.","key_machinery":"The carrying objects are clean immersions and the hyperbolicity criterion from the author's earlier work. A clean immersion is a graph map that is locally injective, whose transition matrix is primitive, and whose Whitehead graphs at every vertex are connected; it represents the endomorphism on the fundamental group. The criterion (Theorem 6.2) states that if an endomorphism is represented by an immersion, then the mapping torus is word-hyperbolic exactly when no nontrivial element $a$ has $\\varphi^n(a)$ conjugate to $a^d$ for some $d,n\\ge 1$. The new work supplies the clean immersion representative and proves that such a periodic conjugacy class would violate Proposition 5.4, making the criterion's obstruction impossible. Bounded cancellation estimates and the clean train-track structure drive the length and fold arguments that produce the representative and control iterated images.","core_discovery":"The central claim is Theorem 6.3: if an injective endomorphism $\\varphi:F\\to F$ is represented by a clean immersion, then its mapping torus $F*_\\varphi$ is word-hyperbolic; in particular, every irreducible nonsurjective endomorphism of a finitely generated free group is hyperbolic in this sense. Supporting results include Theorem 4.5, which produces a unique irreducible immersion representative with connected Whitehead graphs for every irreducible nonsurjective endomorphism; Theorem 4.6, a partial converse using Whitehead graphs with no cut vertices; and Theorem 5.5, which characterizes fully irreducible injective endomorphisms as those with no periodic cyclic free factor, a clean representative, and image not contained in a proper free factor. Corollary 5.6 then identifies irreducible and fully irreducible for nonsurjective endomorphisms. The decisive technical step is Proposition 5.4: if a finitely generated subgroup $H$ satisfies $\\varphi(H)\\le H$ and contains a $\\varphi$-expanding conjugacy class, then $[\\varphi^k(F):\\varphi^k(F)\\cap H]<\\infty$ for some $k\\ge 0$, which rules out the cyclic periodic conjugacy class that would obstruct hyperbolicity.","pith_inferences":["An implication the author leaves implicit is that the mapping tori constructed here form a class of hyperbolic groups that are not free-by-cyclic in the usual sense, so the proof gives a template for showing hyperbolicity of more general ascending HNN extensions of free groups.","The uniqueness of the immersion representative suggests that the expanding graph together with its leading eigenvalue is an invariant of the endomorphism, potentially allowing distinct irreducible nonsurjective endomorphisms to be distinguished by their clean immersion data.","A natural testable extension is to ask whether the conclusion holds for endomorphisms some iterate of which is represented by a clean immersion; the current proof does not immediately cover that 'eventually clean' class.","Since the obstruction to hyperbolicity is exactly a periodic conjugacy class, the theorem predicts that in all explicit examples, the length of $\\varphi^n(a)$ grows exponentially for every nontrivial $a$; this can be checked directly in small-rank examples."],"forward_implications":["Every irreducible nonsurjective endomorphism of a finitely generated free group has a word-hyperbolic mapping torus, so the group $\\langle a,b,t \\mid t^{-1}at=ab,\\ t^{-1}bt=ba\\rangle$ is word-hyperbolic.","Irreducibility and full irreducibility coincide for nonsurjective endomorphisms, so iterates of an irreducible nonsurjective endomorphism remain irreducible.","The characterization in Theorem 5.5 gives a checkable criterion for full irreducibility of an injective endomorphism: no periodic cyclic free factor, a clean representative, and an image not contained in a proper free factor.","An irreducible nonsurjective endomorphism acts on outer space with a unique attracting fixed point, represented by the unique irreducible immersion, giving a canonical graph model for the endomorphism.","The partial converse Theorem 4.6 shows that a clean immersion with connected Whitehead graphs and no cut vertices represents a nonsurjective fully irreducible endomorphism."],"supporting_citations":[{"why":"Supplies the hyperbolicity criterion (Theorem 6.2) that converts the absence of periodic conjugacy classes into word-hyperbolicity of the mapping torus.","marker":"[15]"},{"why":"Provides the train track representative theorem used to obtain the clean representative needed in Theorem 4.5.","marker":"[4]"},{"why":"Supplies the lamination lemma on convergence of the outer space action and the subgroup-support-leaf argument that Proposition 5.3 extends.","marker":"[3]"},{"why":"Provides the bounded cancellation lemma used in the length estimates that produce the immersion representative.","marker":"[8]"},{"why":"Supplies the fold decomposition and subgroup separability facts used to show irreducible endomorphisms are injective and nonsurjectivity forces unbounded growth of iterated images.","marker":"[18]"},{"why":"Supplies the proposition upgrading weakly clean train tracks to clean ones, and the earlier irreducibility/full-irreducibility equivalence for atoroidal automorphisms.","marker":"[9]"},{"why":"Is the earlier formulation of the immersion-representation theorem that Theorem 4.5 reproves and then uses as the bridge from irreducibility to hyperbolicity.","marker":"[17]"}],"fun_headline_variants":["Hyperbolic mapping tori from irreducible nonsurjective maps","Mapping tori of non-invertible free-group maps are hyperbolic","Irreducible nonsurjective free-group maps yield hyperbolic tori","Hyperbolic tori from nonsurjective irreducible free-group maps","Nonsurjective endomorphisms give hyperbolic mapping tori"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is the previously proved hyperbolicity criterion (Theorem 6.2), that for an immersion representative the mapping torus is hyperbolic if and only if no nontrivial element satisfies $\\varphi^n(a)$ conjugate to $a^d$; the present paper does not reprove that criterion, so an error or unstated hypothesis in it would collapse the main conclusion.","fun_headline_variants_meta":{"raw":{"variants":["Hyperbolic mapping tori from irreducible nonsurjective maps","Mapping tori of non-invertible free-group maps are hyperbolic","Irreducible nonsurjective free-group maps yield hyperbolic tori","Hyperbolic tori from nonsurjective irreducible free-group maps","Nonsurjective endomorphisms give hyperbolic mapping tori"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.002055,"raw_usage":{"total_tokens":7972,"prompt_tokens":890,"completion_tokens":7082,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":506,"completion_tokens_details":{"reasoning_tokens":6995}},"tokens_in":506,"tokens_out":7082,"duration_ms":44704,"temperature":1.0,"reasoning_tokens":6995,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:46:24.955980+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Exhibit a clean immersion representing an irreducible nonsurjective endomorphism for which some nontrivial $a\\in F$ and integers $d,n\\ge 1$ satisfy $\\varphi^n(a)=g a^d g^{-1}$; by the paper's own criterion (Theorem 6.2) the mapping torus would then fail to be word-hyperbolic, disproving Theorem 6.3. In a concrete example this is a finite check: if the word length of $\\varphi^n(a)$ grows linearly for some nontrivial $a$, such a periodic class exists, and the predicted hyperbolicity fails.","supporting_citations":[{"cited_title":"Hyperbolic Immersions of Free Groups","cited_arxiv_id":"1809.04761","evidence_quote":"Supplies the hyperbolicity criterion (Theorem 6.2) that converts the absence of periodic conjugacy classes into word-hyperbolicity of the mapping torus."},{"cited_title":"Train tracks and automorphisms of free groups","cited_arxiv_id":null,"evidence_quote":"Provides the train track representative theorem used to obtain the clean representative needed in Theorem 4.5."},{"cited_title":"Laminations, trees, and irreducible automor- phisms of free groups","cited_arxiv_id":null,"evidence_quote":"Supplies the lamination lemma on convergence of the outer space action and the subgroup-support-leaf argument that Proposition 5.3 extends."},{"cited_title":"The group fixed by a family of injective endomorphisms of a free group, volume 195 of Contemporary Mathematics","cited_arxiv_id":null,"evidence_quote":"Provides the bounded cancellation lemma used in the length estimates that produce the immersion representative."},{"cited_title":"Stallings","cited_arxiv_id":null,"evidence_quote":"Supplies the fold decomposition and subgroup separability facts used to show irreducible endomorphisms are injective and nonsurjectivity forces unbounded growth of iterated images."},{"cited_title":"Leininger","cited_arxiv_id":null,"evidence_quote":"Supplies the proposition upgrading weakly clean train tracks to clean ones, and the earlier irreducibility/full-irreducibility equivalence for atoroidal automorphisms."},{"cited_title":"Dynamics of Irreducible Endomorphisms of $F_n$","cited_arxiv_id":"1008.3659","evidence_quote":"Is the earlier formulation of the immersion-representation theorem that Theorem 4.5 reproves and then uses as the bridge from irreducibility to hyperbolicity."}],"review_version":1}