{"id":"e68f4474-c77a-4b63-b944-fb993afae745","arxiv_id":"1908.07382","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For free group and monoid actions, generalized limit sets are characterized by new internal transitivity conditions, with exact equivalences in shifts of finite type and shadowing systems.","lead":"This paper proves that for actions of finitely generated free groups and monoids, several new notions of 'limit set' are exactly captured by generalized versions of chain transitivity, especially in shifts of finite type and systems with the shadowing property. It extends a well-known characterization of omega-limit sets from ordinary iteration to group actions with multiple possible futures.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 5.6's final inference assumes the family W_w is Hausdorff-closed; this is neither proved nor generally true, so the shadowing result only yields CICT ⊆ closure(W_w) as written.","rationale":"The paper's central claim is that limit sets are characterized by internal transitivity notions, and the shadowing theorems are the main route to this characterization outside shift spaces. The reader identified the weakest spot correctly: the proof of Theorem 5.6 passes from Hausdorff approximation by elements of W_w to membership in W_w, which requires W_w to be closed. The manuscript proves closure for CICT and IBT variants but never for W_w or WF_w. This is not a stylistic gap; for classical Z-actions the analogous family of omega-limit sets is not generally closed under shadowing, and the literature's result is exactly that internal chain transitivity is the closure of that family. Therefore the proof as written only supports CICT ⊆ closure(W_w), not equality. The concern is load-bearing because it affects the central claim of Section 5 and the corresponding corollaries. It is likely fixable by a more careful construction or by replacing the conclusion with the appropriate closure statement, but as written the theorem is unsupported. I agree with the reader's assessment and would keep the verdict unchanged: conditional acceptance pending a correct proof of the missing closedness or a modified statement.","tokens_in":19130,"tokens_out":11443,"duration_ms":122758,"concrete_test":"Test the unstated lemma by trying to prove that W_w is closed for every G-action with G-shadowing. A concrete first case is the free monoid on one generator, where the theory reduces to classical N-actions: exhibit or rule out a shadowing map for which the family of omega-limit sets is not Hausdorff-closed, i.e., a sequence (x_n, w_n) with d_H(omega_{w_n}(x_n), Y) < 1/n but with no pair (x, w) satisfying omega_w(x) = Y. If such a sequence exists, the inference in Theorem 5.6 is invalid. If, instead, one can prove a diagonal argument that extracts a common (x, w) from the approximating sequence, the gap is fillable.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The decisive step is the last inference in the proof of Theorem 5.6. For each n, the proof constructs a pseudo-orbit, shadows it by some x_n, and obtains d_H(omega_{w_n}(x_n), Y) < 1/n. It then states: \"As n was arbitrary, Y ∈ W_w.\" This is precisely an assertion that W_w is closed in the Hausdorff topology: Y is concluded to be a member of W_w from being a Hausdorff limit of members of W_w. No proof of this closedness is given. The paper proves in Theorem 4.10 that CICT is closed, but it never states or proves that the family of omega_w-limit sets is closed. For classical Z-actions, shadowing does not imply that the collection of omega-limit sets is closed; the known result of Meddaugh and Raines is that internal chain transitivity equals the closure of the collection of omega-limit sets, not the collection itself. Thus, without an additional argument, the proof establishes only CICT ⊆ closure(W_w), and equality CICT = W_w can fail if W_w is not closed. The same gap appears in Theorem 5.7, where d_H(omega^F_{w_k}(x_k), Y) < 1/k is used to conclude Y ∈ WF_w without a closedness proof for WF_w.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies continuous actions of finitely generated free groups and free monoids on compact metric spaces. It defines several limit-set notions—the undirected ω(x), the direction-constrained ω_w(x), and the relaxed direction-constrained ω_w^F(x)—and introduces corresponding internal transitivity notions ICT, CICT, IBT, IBT*, and IBT^o. The main claims are that, for shifts of finite type, the limit-set collections are exactly characterized by the new transitivity classes (Corollary 4.14 and Corollaries 4.28, 4.30); that a subshift over a free group has the G-shadowing property if and only if it is a shift of finite type (Theorem 5.5); and that, under G-shadowing or a weak asymptotic shadowing hypothesis, these limit sets coincide with the transitivity classes in general systems (Theorems 5.6–5.8 and 5.13–5.15).","tokens_in":19374,"tokens_out":16644,"duration_ms":170671,"significance":"The free-group setting is a nontrivial and not-yet-standard extension of the classical Z-action theory, and the paper gives useful, explicit constructions for the SFT case. Theorem 5.5 is a clean analogue of Walters' theorem for free-group subshifts. The paper is also careful in distinguishing the group and monoid cases, which genuinely change the internal transitivity notions. However, the central shadowing equalities in Theorems 5.6–5.8 are not fully proved as written: they depend on a Hausdorff-closedness property of the limit-set collections that is neither stated nor established. If that gap is repaired, the paper would make a solid contribution; as it stands, the major advertised conclusion 'limit sets are completely characterized by internal transitivity... under shadowing' is not yet justified.","major_comments":[{"comment":"The final inference of the proof is invalid as written. For each n the proof constructs a word w_n and a point x_n with d_H(ω_{w_n}(x_n), Y) < 1/n, and then concludes 'As n was arbitrary, Y ∈ W_w'. This requires the family W_w to be closed in the Hausdorff metric. The paper proves in Theorem 4.10 that CICT is closed, and in Lemma 4.11 that W_w ⊆ CICT, but closedness of a superset does not imply closedness of a subset. Moreover, for Z-actions with shadowing the known result of Meddaugh and Raines is that internal chain transitivity equals the closure of the collection of ω-limit sets, not necessarily the collection itself, so closedness of W_w is not a consequence of shadowing alone. As written, the proof establishes only CICT ⊆ closure(W_w). The same missing-closedness step appears at the end of Theorems 5.7 and 5.8. The authors should either prove the needed closedness of W_w and WF_w under G-shadowing, replace the conclusion with the weaker closure equality, or supply a compactness/diagonal argument that produces a single limit word w and a single limit point x with exact equality Y = ω_w(x).","section":"§5.6, Theorem 5.6"}],"minor_comments":[{"comment":"The function O is initially declared to have codomain Y, but the later definition O(v) = σ_{v'}(z_{n_v}) for non-prefix words v need not lie in Y unless Y is invariant. The proof only seems to need O(v) ∈ X for arbitrary v and O(u) ∈ Y for prefixes of the constructed word w; please clarify and correct the stated codomain.","section":"§4.13, proof of Theorem 4.13"},{"comment":"The closing sentence of the proof of Theorem 5.6, 'We have already shown that Ww ⊆ CICT and CICT is closed; thus Ww ⊆ CICT', is tautological and does not address the missing converse. This appears to be a typographical or logical slip that should be corrected.","section":"§5.6, final line"},{"comment":"The proof of Theorem 5.7 ends with 'Thus Y ∈ IBT*', which is the hypothesis rather than the desired conclusion Y ∈ WF_w; this is presumably a typo.","section":"§5.7, final line"},{"comment":"In the construction of the shadowing point, 'O(uv)(1)' should presumably be 'O(uv)(e)', using the identity element notation introduced earlier.","section":"§5.5, proof of Theorem 5.5"},{"comment":"The sentence 'We claim that x ∈ X and that ω_w(x) = A' should read 'ω_w(x) = Y'.","section":"§4.13, proof of Theorem 4.13"},{"comment":"In Definition 4.9 the symbol x is used both for the point whose indices i(x), t(x) are chosen and for the starting point of the chain; and in the proof of Theorem 4.10 the notation j(x) and t(x) is used interchangeably. Please make the notation uniform.","section":"§4.9–4.10"}],"recommendation":"major_revision","confidential_remarks":"The core SFT construction and the shadowing-SFT equivalence appear original and likely correct, but the main shadowing characterization rests on the unproved closedness of W_w. If the authors can repair Theorem 5.6, the paper will be a solid contribution; if the stronger equality is in fact false under G-shadowing, the theorem statements will need to be weakened to closure equalities."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nRead the Binder\\u2013Meddaugh paper on limit sets for free group actions. The SFT part is the real contribution: the CICT, IBT*, and IBT^o notions are natural, and the constructions in Thms 4.13 and 4.27/4.29 actually produce points and words realizing arbitrary such sets. That part looks correct and is genuinely new. The SFT-iff-shadowing theorem (5.5) is also a clean extension of Walters’ result.\n\nThe soft spot is in the shadowing theorems. In Thm 5.6, for each n the authors get an omega_{w_n}(x_n) within 1/n of Y in Hausdorff distance, and then just say “as n was arbitrary, Y \\in W_w.” That inference requires W_w to be closed in the Hausdorff topology. The paper proves CICT is closed but never shows W_w is closed, and for Z-actions it is known that shadowing only gives ICT = closure(W), not equality. So the last step doesn’t hold as written; the correct conclusion is CICT = closure(W_w), and similarly for IBT* and IBT^o. That would still be a meaningful theorem and matches Meddaugh–Raines, but it isn’t what the abstract claims. The weak-asymptotic-shadowing results (5.13\\u20135.15) dodge this because they use exact asymptotic shadowing.\n\nAlso minor: a couple of places say chains end with i(x) instead of t(x) (e.g., in the proof of 5.6, v'_{l_k} should be t(x_0^{k+1}), not i(x_0^{k+1})). Cute typos, easy to fix.\n\nMy take: this paper deserves a serious referee, but the referee needs to make the authors redo the shadowing sections. The SFT characterizations are probably what people will use. I’d bring it to reading group to talk about the closure issue.\n\nRecommendation: accept for peer review with major revision expected.","headline":"Solid SFT characterizations, but the shadowing theorems overclaim exact equality without proving W_w is closed—a real gap that a referee should catch.","tokens_in":19928,"tokens_out":5310,"would_cite":true,"duration_ms":50958,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["37B50","37B10","37B20","54H20"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper shows that, for free group and monoid actions with a shadowing property, directional limit sets coincide exactly with a new class of consistently internally chain transitive sets.","keywords":["free group actions","omega-limit sets","internal chain transitivity","consistent internal chain transitivity","shadowing property","shifts of finite type","block transitivity","Hausdorff topology"],"falsifier":"Look for a compact metric space with a free group action that has the shadowing property, and a sequence of directional limit sets whose limit is not a directional limit set; finding one refutes the equality, while proving the collection is always closed would confirm it.","tokens_in":18907,"feed_emoji":"🌀","tokens_out":11349,"duration_ms":101510,"temperature":0.7,"pith_summary":"For ordinary dynamical systems, the $\\omega$-limit sets of a point are known to be internally chain transitive, and under shadowing the converse holds up to Hausdorff closure. This paper transplants that program to actions of finitely generated free groups and free monoids on compact metric spaces, where there is no single 'future' and hence many candidate limit sets. The authors define several directional limit sets — the set $\\omega(x)$ of all limit behaviors, the set $\\omega_w(x)$ of behaviors along a fixed infinite reduced word $w$, and the looser set $\\omega_w^F(x)$ — and introduce matching internal transitivity notions, most notably consistent internal chain transitivity ($\\mathrm{CICT}$). Their central result is that, for shifts of finite type and for actions with the shadowing property, the class of directional limit sets $\\omega_w$ is exactly the class of $\\mathrm{CICT}$ sets; analogous equalities hold for $\\omega_w^F$ with a block-transitivity condition. If correct, this gives a purely combinatorial handle on which closed sets can arise as limit behavior along a prescribed direction, mirroring the classical theorem for $\\mathbb{Z}$-actions.","feed_headline":"When shadowing holds, consistent chain transitivity pins down limit sets","feed_subtitle":"Directional limit sets coincide with consistently chain transitive sets under shadowing, with SFTs as the base case.","key_machinery":"The load-bearing mechanism is the consistency condition embedded in $\\mathrm{CICT}$ together with a word-building construction. Given a closed set $Y$ in $\\mathrm{CICT}$, each point carries a pair $(i(x), t(x))$ of generator letters that control how chains enter and leave $x$. The proof covers $Y$ with finitely many points, strings together $\\varepsilon$-chains between successive cover points (respecting the entry/exit letters), and concatenates the indexing words to form an infinite reduced word $w$. Under shadowing, the resulting chain data define a pseudo-orbit that is shadowed by some point $x$; the limits along the prefixes of $w$ then reconstruct exactly the points of $Y$. A separate block-transitivity notion ($IBT^*$, $IBT^\\circ$) plays the same role for the less directed limit sets $\\omega_w^F$, using pseudo-orbits indexed by arbitrary group elements rather than chains.","core_discovery":"On the paper's own terms, the central discovery is two-fold. First, plain internal chain transitivity is too weak in the free group case: the paper exhibits a shift space subset that is internally chain transitive yet is not any $\\omega_w$-limit set (Example 4.8). The missing ingredient is a directional consistency: for each point $x$ in the set one must be able to choose an entry letter $i(x)$ and an exit letter $t(x)$ with $i(x) \\neq t(x)^{-1}$, so that chains between any two points start with the first point's entry letter and end with the second's exit letter. Second, with this 'consistent' version in hand, the full correspondence holds: Corollary 4.14 states that for a shift of finite type over $G$, $W_w = \\mathrm{CICT}$, and Theorem 5.6 states that for any $G$-action with $G$-shadowing, $W_w = \\mathrm{CICT}$. The paper also establishes that a shift space over a free group/monoid has the shadowing property exactly when it is a shift of finite type, and it obtains analogue characterizations of the coarser limit sets $\\omega_w^F$ by internally block transitive sets with a final point (variants $IBT^*$ and $IBT^\\circ$).","pith_inferences":["The entry/exit letter pair $i(x), t(x)$ in $\\mathrm{CICT}$ assigns a direction at each point of the set; this strongly resembles the structure of the boundary of the free group, suggesting that directional limit sets are governed by the ends of the Cayley graph rather than by the full shift structure.","The unproved closedness of $W_w$ in the shadowing theorem points to a concrete research question: determine whether shadowing forces the family of directional limit sets to be Hausdorff closed, or find a shadowing system where it is not. Answering this would either repair or refute the main theorem.","Since the free monoid case avoids inverses and drops the consistency condition entirely, the technically simpler setting may be the better testbed for extending these results to other graph-directed semigroups, such as actions of groups defined by finite presentations.","The word-concatenation construction used here might be adaptable to actions of other groups with a normal form and a notion of reduced word, yielding analogous limit-set characterizations beyond free groups."],"forward_implications":["In any shift of finite type over a free group, a closed set is a directional limit set if and only if it satisfies the finite combinatorial condition $\\mathrm{CICT}$, so the class of limit sets can be decided without constructing the word or point.","For actions with the shadowing property, the equality $\\mathrm{CICT} = W_w$ transfers the classical characterization from $\\mathbb{Z}$-actions to free group actions, so questions about the existence of points with prescribed limit behavior reduce to checking chain data.","The theorem that a shift space over a free group has the shadowing property exactly when it is a shift of finite type gives a way to recognize SFTs among group shift spaces.","Under the weaker hypothesis of asymptotic shadowing, the same equalities hold, so the characterization is stable and covers systems where ordinary shadowing fails.","For free monoid actions, the consistency condition collapses to plain internal chain transitivity, giving an even simpler characterization of directional limit sets."],"supporting_citations":[{"why":"Supplies the Hausdorff-closure characterization of internal chain transitivity and omega-limit sets for classical Z-actions that this paper extends.","marker":"[16]"},{"why":"Provides the characterization of omega-limit sets in shift spaces that motivates the SFT results.","marker":"[4]"},{"why":"Explores variations of shadowing and their effect on the limit-set/internal-transitivity equivalence, the basis for the asymptotic shadowing results.","marker":"[12]"},{"why":"Establishes the classical result that a shift space has the shadowing property exactly when it is of finite type, which Theorem 5.5 generalizes.","marker":"[20]"},{"why":"Develops limit sets and internal transitivity for Z^d actions, the multidimensional predecessor whose complications motivate the free-group choice.","marker":"[17]"},{"why":"Proves that omega-limit sets are internally chain transitive, the foundational fact that the new transitivity notions refine.","marker":"[13]"},{"why":"Defines semigroup limit sets via filter bases, the general framework from which the authors' limit-set definitions are drawn.","marker":"[2]"}],"fun_headline_variants":["Consistent chain transitivity exactly matches limit sets under shadowing","For free group actions, consistent ICT completes the limit set picture","Shadowing plus directional consistency yields limit set equivalence","In free group shifts, consistent chains are the missing link for limit sets","Directional consistency makes internal chain transitivity exact for limit sets"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument assumes that a limit of directional limit sets is again a directional limit set, but the paper never proves that the collection of such sets is closed under taking limits; if this fails, the shadowing theorem only shows containment in the closure.","fun_headline_variants_meta":{"raw":{"variants":["Consistent chain transitivity exactly matches limit sets under shadowing","For free group actions, consistent ICT completes the limit set picture","Shadowing plus directional consistency yields limit set equivalence","In free group shifts, consistent chains are the missing link for limit sets","Directional consistency makes internal chain transitivity exact for limit sets"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000652,"raw_usage":{"total_tokens":2966,"prompt_tokens":896,"completion_tokens":2070,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":512,"completion_tokens_details":{"reasoning_tokens":1995}},"tokens_in":512,"tokens_out":2070,"duration_ms":14681,"temperature":1.0,"reasoning_tokens":1995,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:20:46.582020+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Look for a compact metric space with a free group action that has the shadowing property, and a sequence of directional limit sets whose limit is not a directional limit set; finding one refutes the equality, while proving the collection is always closed would confirm it.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Hausdorff-closure characterization of internal chain transitivity and omega-limit sets for classical Z-actions that this paper extends."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the characterization of omega-limit sets in shift spaces that motivates the SFT results."},{"cited_title":"Orbital shadowing, i nternal chain transitivity and ω -limit sets","cited_arxiv_id":null,"evidence_quote":"Explores variations of shadowing and their effect on the limit-set/internal-transitivity equivalence, the basis for the asymptotic shadowing results."},{"cited_title":"On the pseudo-orbit tracing property an d its relationship to stability","cited_arxiv_id":null,"evidence_quote":"Establishes the classical result that a shift space has the shadowing property exactly when it is of finite type, which Theorem 5.5 generalizes."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Develops limit sets and internal transitivity for Z^d actions, the multidimensional predecessor whose complications motivate the free-group choice."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Proves that omega-limit sets are internally chain transitive, the foundational fact that the new transitivity notions refine."},{"cited_title":"Braga Barros and Josiney A","cited_arxiv_id":null,"evidence_quote":"Defines semigroup limit sets via filter bases, the general framework from which the authors' limit-set definitions are drawn."}],"review_version":1}