{"id":"04aa755a-bc28-4196-a4c8-9c4192ea984a","arxiv_id":"2510.20240","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Li-Yorke and distributional chaos transfer from a dynamical system to its hyperspace and fuzzy-set extensions, while a new Cantor-dense variant transfers back under completeness and linearity assumptions.","lead":"This paper proves that Li-Yorke chaos and its many variants pass from a dynamical system to its hyperspace and fuzzy-set extensions, but not generally back. It introduces a stronger \"Cantor-dense\" chaos that does transfer back under completeness and linearity conditions.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Reverse-transfer rests on [26, Prop 3.5] without proof; if that equivalence fails or is restricted to group/linear maps, Theorems 5.2 and 5.5(b) collapse.","rationale":"The reader's weakest_assumption identifies the external equivalence [26, Prop 3.5] as the key unproved support for the strengthened reverse-transfer results. My concern is a more specific version of the same: the equivalence may not even apply to the spaces used, since [26] appears to target group endomorphisms and linear operators, while K(X) and F(X) are not groups or vector spaces in the relevant sense. This is the single most load-bearing point because Theorem 5.2 is the paper's headline reverse transfer; if the equivalence is false or inapplicable, the conclusion does not follow. The forward-transfer theorems and counterexamples are internally well-argued and do not suffer from this issue. Therefore the paper remains a solid extension conditional on the external result, matching the reader's CONDITIONAL verdict. I see no additional internal error that would change the verdict, so I recommend no change.","tokens_in":37471,"tokens_out":23073,"duration_ms":172977,"concrete_test":"Obtain [26] and read the statement and proof of Proposition 3.5. Verify (i) that the proposition is stated for arbitrary continuous maps on completely metrizable spaces, not only for endomorphisms of groups or linear operators; (ii) that the proof does not use group multiplication, translation-invariance of the metric, or linearity in an essential way (e.g., to move open sets or to build Cantor sets). If the proof exploits group structure, then Theorem 5.2 is unsupported and the verdict should be REJECT or CONDITIONAL pending a direct proof. As a cross-check, attempt to prove or disprove the equivalence by constructing a sensitive Polish system with dense proximal relation but no Cantor-dense scrambled set; such a counterexample would falsify the external proposition and hence Theorem 5.2.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's strongest advertised result — that Cantor-dense Li-Yorke chaos transfers from (F(X),f_hat) back to (K(X),f_bar) — is not proved directly. Theorem 5.2 (and the CD-LYC parts of Theorem 5.5) invoke [26, Proposition 3.5] to pass from dense proximal relation + sensitivity to CD-LYC. This step is load-bearing: without it, all that has been shown is that the fuzzy system has dense Prox and sensitivity, which by itself does not imply the existence of a Cantor-dense scrambled set (and the paper gives no Mycielski construction). The cited proposition is quoted as valid for 'a continuous map f acting on a completely metrizable space,' yet reference [26] is titled 'Chaos for endomorphisms of completely metrizable groups and linear operators on Fréchet spaces'; the hyperspace K(X) and the fuzzy spaces F∞(X),F0(X) carry no group or linear structure. If Proposition 3.5 was proved only for group endomorphisms (or only for linear operators), then applying it to (K(X),f_bar) is invalid. Even if the statement is true generally, the paper does not establish it, and the strong 'Cantor-dense' requirement (a Cantor set inside every prescribed open set) is exactly where a subtle Baire-category argument would be needed. Thus the central reverse-transfer claim is only as secure as an external, unverified equivalence.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies transfer of Li-Yorke-type chaos between a metric dynamical system (X,f), its hyperspace extension (K(X), f_bar) with the Hausdorff metric, and the fuzzy extension (F(X), f_hat) with the supremum, Skorokhod, sendograph, and endograph metrics. The main forward results (Theorems 3.1 and 3.2) show that LY, mean LY, and distributional chaos of types DC1–DC3 transfer from (X,f) to (K(X),f_bar) and from (K(X),f_bar) to (F(X),f_hat), with the endograph metric requiring boundedness or diameter conditions. Three counterexamples answer Question 1.1 positively and show optimality of boundedness assumptions. The paper then introduces Cantor-dense Li-Yorke chaos (CD-LYC) and, using proximality and sensitivity results from Section 4, proves reverse transfer from fuzzy systems to the hyperspace under completeness (Theorems 5.2 and 5.5), and equivalences for linear operators on Fréchet spaces.","tokens_in":37776,"tokens_out":7769,"duration_ms":70451,"significance":"If the results hold, the paper gives a systematic and fairly complete picture of how the main Li-Yorke-type chaotic properties behave under hyperextension and fuzzification. The detailed metric estimates (3.4) and (3.5) are a genuine technical contribution, and the counterexamples are structurally coherent and directly address a question left open in [34]. The proximality and sensitivity transfer results in Section 4 are also new and are proved with care, including a corrected proof of a point left unclear in [20]. The CD-LYC reverse-transfer results are potentially significant, but their current form depends on an external characterization from [26] whose exact scope is not verified in the manuscript.","major_comments":[{"comment":"The reverse-transfer claims for CD-LYC rest entirely on [26, Proposition 3.5], quoted as valid for arbitrary continuous maps on completely metrizable spaces. However, the title of [26] indicates a focus on endomorphisms of completely metrizable groups and linear operators on Fréchet spaces, and the spaces K(X), F∞(X), F0(X) carry no group or linear structure in general. The authors should either reproduce the exact statement of [26, Proposition 3.5] and confirm that it applies to arbitrary continuous maps, or provide a direct proof of the dense proximal relation + sensitivity ⇒ CD-LYC implication. Without this, the conclusions of Theorem 5.2 and Theorem 5.5(b)(iii)–(v) do not follow from the material proved in the paper.","section":"§5.1, Theorem 5.2; §5.2, Theorem 5.5(b)"},{"comment":"In each example the authors write 'We leave to the reader checking that (X,d) is a metric space'. The metrics are piecewise-defined using sets A with prescribed density properties, and the triangle inequality is not immediate, particularly for triples of points sharing a first coordinate where one lies in A and another not. Since these examples are load-bearing for the claimed failures of converse transfer and for the sharpness of boundedness assumptions, the verification should be included or at least sketched. This is fixable but is currently a gap in the proof of the counterexamples.","section":"§3.2, Examples 3.4, 3.5, 3.6"},{"comment":"The step '(i)⇒(iii),(iv),(v)' applies [26, Proposition 3.5] to (K(X),T), (F∞(X), f_hat), and (F0(X), f_hat) after establishing dense proximal relation and sensitivity. Besides the external-proposition concern raised above, the completeness of F∞(X) and F0(X) is cited from [25] and [15] without any statement of the results. The authors should state explicitly the reference results used, because the CD-LYC conclusion requires the full force of the Mycielski-based equivalence, not merely the existence of a scrambled Cantor set in one open set.","section":"§5.2, proof of Theorem 5.5(b)"}],"minor_comments":[{"comment":"There is a typo: the endograph metric is defined as 'the Hausdorff distance d_H(end(u),end(u))', which should be d_H(end(u),end(v)).","section":"§2.2, definition of d_E"},{"comment":"The use of ε′:=min{ε,1} is clear for D1 and D2, but the statement lists LY without a parameter; this creates a small notational mismatch. Consider reformulating the LY case separately.","section":"§3.1, Theorem 3.1(c)"},{"comment":"The set X0 is defined using 'some increasing sequence (n_k)', but the statement does not specify whether the same sequence is used for all points. This should be clarified, since the proof of density of the proximal relation uses X0^2 ⊂ Prox(f,d).","section":"§5.2, Theorem 5.5(b)"},{"comment":"The computation of the liminf for the scrambled set S omits some justifications in the inequalities involving [n] = sqrt(log2 n); for example, the equality of sums over A ∩ [1,n−1] with powers of 2 is not fully spelled out. A short explanation would improve readability.","section":"§3.2, Example 3.5"}],"recommendation":"major_revision","confidential_remarks":"The forward-transfer core of the paper is solid and deserves publication. The main risk is the reliance on [26, Proposition 3.5] in the CD-LYC reverse-transfer theorems; if the editor or a second referee can confirm that the cited proposition indeed holds for arbitrary continuous maps on complete metric spaces, the necessary revisions reduce to minor ones. Otherwise, the authors should supply a proof of the dense proximal + sensitivity ⇒ CD-LYC step in the generality they need. The examples' deferred metric verifications should also be supplied."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague — this is a careful, workmanlike paper. It does what it says: extends [34, Proposition 3] to the full range of Li-Yorke variants, covers the Skorokhod, sendograph, and endograph metrics, and answers the open question from [34] with a genuinely constructed counterexample. The metric estimates in Section 3 are the real core — (3.4) and (3.5) are proved in detail, and the counterexamples are structurally sound. The new notion of Cantor-dense Li-Yorke chaos is reasonable, and the proximality/sensitivity results in Section 4 stand on their own.\n\nThe soft spot is Theorem 5.2. The proof is essentially: CD-LYC implies dense proximality and sensitivity; then invoke [26, Proposition 3.5] to go back to CD-LYC on K(X). That external proposition is load-bearing, and the paper does not prove it. Given the title of [26] — endomorphisms of completely metrizable groups and linear operators on Fréchet spaces — one has to check whether the equivalence is stated for general continuous maps on complete metric spaces or only for those settings. If restricted, Theorems 5.2 and 5.5(b) do not follow as written. This is not a rejection-level problem; it is a request to either prove the needed direction or quote the precise statement. A referee should ask for that.\n\nThe counterexamples leave “checking that (X,d) is a metric space” to the reader. Minor, but it would take a page to fill in; a referee might ask for it. The paper shows good faith — Remark 3.3 explicitly notes which arguments do not seem to work for certain chaos types, and the authors re-prove a part of [20] they found unclear.\n\nWho is this for? Anyone working in hyperspace or fuzzy-set dynamics. It completes the Li-Yorke picture and gives a usable toolbox. I would send it to a serious referee. The main transfer theorems are solid, and the advertised reverse-transfer is plausible but conditional on an external result that needs verification.","headline":"Solid systematic extension of [34] with good estimates and counterexamples, but the advertised reverse-transfer rests on an unproved external equivalence [26, Prop 3.5] that the authors should pin down.","tokens_in":38320,"tokens_out":2565,"would_cite":true,"duration_ms":24121,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["37B02","37B05","37B25","54A40","54B20"],"pacs":[],"model":"deepseek-v4-flash","headline":"Li-Yorke chaos passes to fuzzy extensions of a dynamical system, but only a strengthened variant passes back.","keywords":["Li-Yorke chaos","distributional chaos","fuzzy dynamical systems","Zadeh extension","hyperspaces","Cantor-dense Li-Yorke chaos","proximality","sensitivity"],"falsifier":"Find a complete metric space (X,d) and a continuous f such that (F∞(X),f_hat) is Cantor-dense Li-Yorke chaotic but (K(X),f_bar) is not; this would disprove the main reverse-transfer theorem. Alternatively, exhibit a completely metrizable system that is both sensitive and has a dense proximal relation but fails to have scrambled Cantor sets in every open set—this would invalidate the equivalence on which the proof leans.","tokens_in":37337,"feed_emoji":"🌀","tokens_out":4792,"duration_ms":39418,"temperature":0.7,"pith_summary":"This paper asks what happens to Li-Yorke chaos when a dynamical system is extended to its space of compact subsets and then to its space of normal fuzzy sets. It proves that all main variants—Li-Yorke, mean Li-Yorke, and distributional chaos types DC1 through DC3, plus uniform versions—transfer forward along these extensions for the supremum, Skorokhod, and sendograph metrics, and largely for the endograph metric. The converse transfer fails in general, and the paper supplies counterexamples where the fuzzy system is chaotic though the original system has almost no scrambled pairs. To repair the converse, it introduces Cantor-dense Li-Yorke chaos, which does transfer back from the fuzzy system to the hyperspace when the underlying space is complete. For linear operators on Fréchet spaces, several of these properties become equivalent to non-equicontinuity.","feed_headline":"Li-Yorke chaos survives fuzzification—but only one way","feed_subtitle":"All standard chaos variants pass to fuzzy extensions; only Cantor-dense chaos transfers back.","key_machinery":"The argument runs on two embeddings: x ↦ {x} (points to singletons) and K ↦ χ_K (compacta to their characteristic fuzzy sets). These are isometries from (X,d) to (K(X),d_H) and from (K(X),d_H) into the fuzzy spaces with supremum, Skorokhod, and sendograph metrics, so scrambled pairs lift unchanged. For the endograph metric, which only sees distances capped at 1, the paper develops level-set estimates (Lemmas 2.6 and 2.7) that translate sendograph and endograph distance into Hausdorff distance between α-levels. The strengthened converse uses a recent characterization: on completely metrizable spaces, Cantor-dense Li-Yorke chaos is equivalent to having a dense proximal relation together with s","core_discovery":"The central claim is that chaos survives the trip to larger spaces in one direction but not the other. Given a continuous map f on a metric space X, the paper shows that if (X,f) is Li-Yorke chaotic (in any of its standard variants, including mean Li-Yorke and distributional chaos DC1–DC3, or their uniform versions), then so is the induced system on the hyperspace of non-empty compact subsets with the Hausdorff metric, and so is the system on normal fuzzy sets equipped with the supremum, Skorokhod, or sendograph metric. The endograph metric is more delicate and needs boundedness assumptions. Conversely, chaos in the fuzzy system does not in general force chaos in the hyperspace or in X; expl","pith_inferences":["The positive-transfer results suggest a monotonicity principle: Li-Yorke chaos is robust under passing to the 'cloud' of a system—points, subsets, fuzzy sets—while the inverse problem is where the real structure lives; the paper's CD-LYC is a natural candidate for a chaotic core that survives both directions.","The counterexamples rely on carefully crafted metrics where proximal pairs are rare; one could test whether the same separation between (X,f) and (FE(X),f_hat) occurs for simpler metrics on interval maps, where distributional chaos may be generic.","A testable extension: for non-normal fuzzy sets (dropping normality), the transfer results might fail at the first step, since the characteristic-function embedding relies on compact support and a non-empty 1-level; checking the paper's lemmas in that setting would delimit the boundary.","Because the converse transfer leans on an external equivalence, a proof that bypasses that equivalence might weaken the completeness assumptions or extend to non-metrizable uniform spaces."],"forward_implications":["If (X,f) exhibits any of LYC, MLYC, DC1, DC1½, DC2, DC2½, DC3, or their uniform versions, then so do (K(X),f_bar), (F∞(X),f_hat), (F0(X),f_hat), and (FS(X),f_hat).","The converse fails: there are systems whose fuzzy extension is U-DC1 and U-MLYC while every LY- or D3-scrambled subset of X has at most two points, and every such subset of K(X) is finite.","For the endograph metric, transfer holds for LY, DC1, DC1½, DC2, DC2½ and their uniform versions, but MLYC and DC3 require boundedness or diameter≤1 assumptions, and the paper gives examples showing these assumptions cannot be dropped.","Cantor-dense Li-Yorke chaos transfers from (F∞, F0, FS) to (K(X)) when X is complete, giving a negative answer to the open question for this strengthened notion.","For linear operators on Fréchet spaces, under a dense set of vectors with a null sub-orbit, U-LYC and CD-LYC become equivalent to non-equicontinuity across all the relevant extensions."],"fun_headline_variants":["Chaos flows one way: from points to fuzzy sets","Fuzzy chaos doesn't force original chaos, except Cantor-dense","Li-Yorke chaos goes up, not down","Cantor-dense chaos is the sole bridge back from fuzzy","Chaos ascends to fuzzy, descends only for Cantor-dense"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The reverse-transfer theorem rests on a cited equivalence: on completely metrizable spaces, Cantor-dense Li-Yorke chaos is equivalent to having a dense proximal relation together with sensitivity; if that equivalence fails—or if (F∞(X),d∞) and (F0(X),d0) are not complete—the strengthened converse does not follow.","fun_headline_variants_meta":{"raw":{"variants":["Chaos flows one way: from points to fuzzy sets","Fuzzy chaos doesn't force original chaos, except Cantor-dense","Li-Yorke chaos goes up, not down","Cantor-dense chaos is the sole bridge back from fuzzy","Chaos ascends to fuzzy, descends only for Cantor-dense"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000321,"raw_usage":{"total_tokens":1644,"prompt_tokens":746,"completion_tokens":898,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":490,"completion_tokens_details":{"reasoning_tokens":810}},"tokens_in":490,"tokens_out":898,"duration_ms":8797,"temperature":1.0,"reasoning_tokens":810,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T08:29:50.916127+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find a complete metric space (X,d) and a continuous f such that (F∞(X),f_hat) is Cantor-dense Li-Yorke chaotic but (K(X),f_bar) is not; this would disprove the main reverse-transfer theorem. Alternatively, exhibit a completely metrizable system that is both sensitive and has a dense proximal relation but fails to have scrambled Cantor sets in every open set—this would invalidate the equivalence on which the proof leans.","supporting_citations":[],"review_version":1}