{"id":"776afbd2-3854-4e59-bcb8-a21f6487ecf1","arxiv_id":"2501.04676","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"For nonautonomous linear difference systems, the paper introduces a new intermediate dichotomy spectrum, proposes a conjecture for its structure, and gives a counterexample showing the nonuniform exponential dichotomy spectrum is not preserved under weak kinematic similarity.","lead":"This math paper studies linear difference equations whose rules change over time and introduces a new spectrum for systems with a unique stable-unstable splitting. It also provides an explicit example showing that, contrary to recent papers, a standard coordinate change does not preserve the nonuniform dichotomy spectrum.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Counterexample to [12] is computed under the looser NED of Def. 2.2, not [12]'s stricter α+2θ<0 condition; the asserted persistence is unproved.","rationale":"The paper's central advertised contribution is the refutation of [12, Cor. 5.7] via Example 5.8. The example is only a valid counterexample if it operates under the definition of nonuniform exponential dichotomy used in [12]. That definition (Rem. 2.5 here) imposes α+2θ<0. Example 5.8's parameter condition 3a>ω>a makes α+2θ = −ω+5a > 2a > 0 whenever a stable estimate of the form (5.6) is used. Consequently the weighted systems in the claimed spectral intervals never satisfy [12]'s stricter dichotomy condition, and the computed spectra are not the objects appearing in Cor. 5.7. The paper's brief assertion that the result persists 'even considering α+2θ<0' is unsupported by any calculation or parameter choice. This is the most load-bearing gap because it directly determines whether the main negative claim is established. The concern is addressable: a variant with ω>5a likely produces strict spectra [−ω−5a,−ω+5a] for A and [−ω−4a,−ω+6a] for B, preserving non-invariance, but that computation must be supplied. This does not amount to a rejection of the paper's broader framework; the new definitions and conditional spectral theorem remain interesting. The Reader's conditional verdict is appropriate, and the identified gap overlaps with the Reader's rationale, though the Reader's primary 'weakest assumption' focused on omitted proofs of the standard spectral lemmas rather than on the stricter-condition mismatch.","tokens_in":27509,"tokens_out":14563,"duration_ms":133081,"concrete_test":"Set a=1, ω=6 (so ω>5a, and the unweighted system A can satisfy α+2θ<0). Recompute the nonuniform exponential dichotomy spectra of A and B using the strict definition from [12] (α+2θ<0 and β−2ν>0), rather than Def. 2.2. With the estimates (5.6)–(5.7), one expects Σ_A^strict=[−11,−1] and Σ_B^strict=[−10,0]; if instead these intervals coincide, Cor. 5.7 is not refuted. Also verify for the printed parameters (e.g. a=1, ω=2.5) that at every γ in the claimed intervals α+2θ>0, confirming the example lies outside [12]'s framework.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Example 5.8 assumes 3a>ω>a. For system (5.4), the stable estimate (5.6) gives, for the γ-weighted system, α=−ω−γ+a and θ=2a. At the claimed right endpoint γ=−ω+3a, we have α+2θ = (−2a)+4a = 2a > 0; indeed for every γ in the claimed interval [−ω−3a, −ω+3a], the pair (α,θ) fails [12]'s strict condition α+2θ<0. Thus the systems in Example 5.8 do not belong to the class for which [12, Cor. 5.7] asserts spectral invariance. The spectra Σ_NµD(A) and Σ_NµD(B) are computed under Def. 2.2 (with the weaker condition α+θ<0), not under the definition used in [12]. The sentence 'even considering α+2θ<0 ... the result is the same' is asserted without proof and is incompatible with the stated parameter range. As written, the paper therefore does not establish the claimed refutation of Cor. 5.7. The underlying phenomenon may well be genuine (a variant with ω>5a may work), but that computation is absent.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a spectral theory for nonautonomous linear difference systems under uniform, nonuniform, and slow nonuniform µ-dichotomies. It introduces two new properties (UPP and USP), a new 'unique projector slow nonuniform µ-dichotomy spectrum', and a conjecture (USPP) under which a spectral theorem is derived. The final section claims to refute the invariance of the nonuniform exponential dichotomy spectrum under weak kinematic similarity, previously asserted in [12] and used in [31], via a one-dimensional counterexample (Example 5.8).","tokens_in":27808,"tokens_out":5423,"duration_ms":55099,"significance":"If the counterexample were valid, the paper would remove a load-bearing assumption from known reducibility and normal-form results, which is a substantial contribution. The UPP/USP distinction and the intermediate spectrum are also potentially useful conceptual tools, and the explicit estimates in Examples 2.14 and 5.8 are helpful. However, the central refutation is not currently established: the counterexample is computed under a weaker definition than the one used in [12], and the spectrum computation relies on unproved spectral and optimal-ratio results. The significance of the paper therefore depends on completing or correcting these points.","major_comments":[{"comment":"The counterexample is computed under Definition 2.2, whose nonuniformity condition is α+θ<0, whereas [12] imposes the stricter condition α+2θ<0, as acknowledged in Remark 2.5. For the stable estimate (5.6) of the γ-weighted system one has α=−ω−γ+a and θ=2a; over the claimed interval [−ω−3a, −ω+3a], this gives α+2θ = −ω−γ+5a ≥ 2a > 0. Hence none of the γ-weighted systems in the displayed interval satisfies [12]'s condition. The sentence in Example 5.8 asserting that 'even considering α+2θ<0 ... the result is the same' is incompatible with the stated parameter range and is not proved. As written, the example does not refute [12, Cor. 5.7].","section":"§5.2, Example 5.8 and Remark 2.5"},{"comment":"Theorem 3.11 is stated without proof ('we omit the proof it and leave it as an exercise for the reader'), and Lemma 3.9 is also left as an exercise. Remark 4.11, which identifies endpoints of Σ_NµD(A) through zero limits of the optimal ratio maps, is a consequence of this unproved theorem. Example 5.8 uses exactly this endpoint identification to assert Σ_NµD(A)=[−ω−3a, −ω+3a] and Σ_NµD(B)=[−ω−2a, −ω+4a]. The spectrum computation is therefore not established by the manuscript. This is not a merely expository omission: the nonuniform definition changes the standard spectral argument, and the validity of Lemma 3.9 in this setting is part of what must be shown.","section":"§3, Theorem 3.11 and Remark 4.11"},{"comment":"Lemma 5.9 provides a sufficient condition, (5.9), for a resolvent point of system (1.1) to be a resolvent point of a weakly kinematically similar system. The subsequent argument only shows that this sufficient condition can fail for γ close to the spectrum of (1.1); it does not prove that such a γ is outside the resolvent of the second system. Thus the negative conclusion in Remark 5.11 rests entirely on Example 5.8. Since that example is not computed under [12]'s stricter hypotheses, the stated refutation of [12, Theorem 3.11] and its use in [31] is not supported by the present manuscript.","section":"§5.3, Lemma 5.9 and the concluding discussion"},{"comment":"Theorem 3.21 is explicitly conditional on Conjecture 2.15, which is not proved. The paper is transparent about this, but the abstract and introduction present a 'spectral theorem' for the new spectrum as one of the main results. The reader should be warned more prominently that this result has the status of a conditional theorem pending an open conjecture, especially because the conjecture is also used implicitly in the discussion of the new spectrum's properties.","section":"§3.2, Theorem 3.21"}],"minor_comments":[{"comment":"In Definition 3.2(iv), the slow nonuniform µ-resolvent set is written as 'ΣsNµD(A) = R \\ ΣsNµD(A)'; the left-hand side should be ρsNµD(A).","section":"Definition 3.2"},{"comment":"There are several typographical errors, including 'we omit the proof it and leave it as an exercise' before Theorem 3.11 and 'The other contention' in Lemma 2.6, which should be 'containment'.","section":"Section 3 and Lemma 2.6"},{"comment":"Example 3.14 states that the spectra for Σ_sNED and Σ^UPP_sNED are computed in Example 5.8, but Example 5.8 computes only the nonuniform exponential dichotomy spectrum; the claimed slow and UPP slow spectra are not derived there. A direct derivation or a precise pointer is needed.","section":"Example 3.14"},{"comment":"The text refers to 'lemma 2.21' before Definition 4.8, but no Lemma 2.21 appears in the paper; the intended reference is presumably Lemma 4.6 or another numbered result.","section":"Section 4"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The one thing worth knowing: this paper has a genuinely new idea buried in it, but its headline refutation of [12] is not established as written. The UPP/USP distinction and the intermediate spectrum are the real contributions; the counterexample to weak kinematic similarity is computed under the wrong (looser) definition.\n\nWhat is actually new: the paper defines UPP and USP, shows that µD and NµD always have them, and—more interestingly—that slow nonuniform µ-dichotomy does not (Examples 2.10 and 2.12 are clean). That motivates the intermediate spectrum Σ^UPP_sNµD, which sits between Σ_sNµD and Σ_NµD, and the USPP conjecture. The spectral theorem for this new spectrum is honestly conditional on the conjecture. The discrete optimal ratio maps in Section 4 are also new, and the paper fills in some proofs omitted in [18]. These are real contributions, and the exposition is careful.\n\nThe soft spot is Section 5. The claim that Example 5.8 refutes [12, Cor. 5.7] does not hold up. Corollary 5.7 lives in [12]'s setting where the nonuniform dichotomy requires α+2θ<0. The example only verifies α+θ<0 (Def. 2.2). For every γ in the claimed spectrum [−ω−3a, −ω+3a], one gets α+2θ = 2a > 0 at the right endpoint (and larger elsewhere), so the systems are not covered by [12]'s hypothesis. The sentence promising the same result under α+2θ<0 is unproved and incompatible with the stated range 3a>ω>a. A different parameter regime might work, but it is not in the paper.\n\nA second issue: Theorem 3.11 and Lemma 3.9 are stated with proofs omitted (\"exercise for the reader\"). The spectrum computations in Example 5.8 rely on Theorem 3.11 and Remark 4.11, so the example is contingent on unproved standard material. This is fixable, but the authors should either include the proofs or cite a fully proved version.\n\nWho this is for: people working on nonuniform dichotomy spectra and reducibility. The UPP/USP material deserves attention even if the refutation fails. I would send it to a referee, but with a clear request: verify or drop the counterexample claim, and supply the omitted proofs.","headline":"The UPP/USP material and the intermediate spectrum are genuinely new and worth reading, but the central counterexample to [12]'s spectral invariance is computed under a weaker dichotomy definition and does not refute the cited result as written.","tokens_in":28324,"tokens_out":4584,"would_cite":true,"duration_ms":39703,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["37D25","37B55"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that weak kinematic similarity does not preserve the nonuniform exponential dichotomy spectrum, refuting a recent invariance claim, and proposes new dichotomy properties and an intermediate spectrum.","keywords":["nonautonomous difference equations","nonuniform dichotomy","dichotomy spectrum","kinematic similarity","optimal ratio maps","slow nonuniform dichotomy","spectral theorem"],"falsifier":"Recompute the two spectra in Example 5.8 directly from the definition of nonuniform exponential dichotomy, or verify whether the similarity $S(n)=e^{-an}$ meets the same-$\\varepsilon$ condition required by the weak kinematic similarity definition: if the spectra coincide under that verification, the non-invariance claim is refuted.","tokens_in":27309,"feed_emoji":"🔄","tokens_out":4617,"duration_ms":42045,"temperature":0.7,"pith_summary":"The paper develops spectral theory for nonautonomous linear difference equations with $\\mu$-dichotomies, focusing on nonuniform versions. Its central finding is that the nonuniform exponential dichotomy spectrum is not invariant under weak kinematic similarity: a concrete one-dimensional pair of weakly similar systems has different spectra, contradicting a recent claim used in reducibility and normal-form arguments. The paper also isolates two properties, the unique projector property and the unbounded solutions property, which hold for uniform and nonuniform dichotomies but fail for slow nonuniform dichotomies. It defines an intermediate spectrum built on the unique projector property and conjectures a condition under which spectral theorems hold for it.","feed_headline":"Weakly similar systems can have different nonuniform spectra","feed_subtitle":"A one-dimensional counterexample refutes an invariance assumption behind reducibility and normal-form theorems.","key_machinery":"The central tools are the optimal ratio maps $\\mathrm{st}_P$ and $\\mathrm{un}_P$, defined on spectral gaps as the infimum of $\\alpha+\\theta$ and the supremum of $\\beta-\\nu$ over admissible dichotomy constants. These maps are decreasing and continuous and encode when nonuniformity parameters can be rebalanced across the weighted system. The counterexample uses the scalar family $x(n+1)=\\exp(-\\omega+a(n+1)\\cos(n+1)-an\\cos(n)-a\\sin(n+1)+a\\sin(n))\\,x(n)$, with $S(n)=e^{-an}$ providing the weak kinematic equivalence. The spectral theorem for $\\mathrm{N}\\mu\\mathrm{D}$ (Theorem 3.11) and the boundary identification given in Remark 4.11 are the load-bearing results used to read the spectrum off the optimal-ratio limits.","core_discovery":"On its own terms, the paper establishes that two systems related by a weakly nondegenerate change of coordinates, the discrete analogue of nonuniform kinematic similarity, need not share the nonuniform $\\mu$-dichotomy spectrum. Example 5.8 compares scalar systems whose spectra are $[-\\omega-3a,\\;-\\omega+3a]$ and $[-\\omega-2a,\\;-\\omega+4a]$, refuting Corollary 5.7 of [12] and its use in [31]. The mechanism is traced to optimal ratio maps: near the spectrum, the transferred dichotomy constants violate the conditions needed for the resolvent inclusion, so the error parameter of the similarity cannot be absorbed. The paper additionally argues that a spectrum defined by requiring the unique projector property, $\\Sigma^{\\mathrm{UPP}}_{sN\\mu D}(A)$, sits strictly between the slow nonuniform and nonuniform spectra and may be the invariant object.","pith_inferences":["Inference: the same non-invariance likely occurs in continuous time for nonuniform $(\\mu,\\varepsilon)$-kinematic similarity, since the mechanism is parameter rebalancing rather than discreteness.","Inference: computing $\\Sigma^{\\mathrm{UPP}}_{sN\\mu D}$ for the two systems in Example 5.8 would test the paper's conjecture that this spectrum is the invariant object; equality there would give a practical invariant for block diagonalization.","Inference: the optimal ratio maps give a numerical test: scanning $\\gamma$ near a spectral boundary should reveal where the transferred parameters fail the dichotomy inequalities, delimiting where resolvent inclusion breaks."],"forward_implications":["Reducibility theorems that assume spectral invariance under weak kinematic similarity, such as [12, Thm. 3.11] and the normal forms result [31], lose their justification for nonuniform spectra.","The nonuniform dichotomy spectrum is not a reliable invariant for classifying nonuniformly hyperbolic discrete systems under weak coordinate changes.","The new intermediate spectrum $\\Sigma^{\\mathrm{UPP}}_{sN\\mu D}(A)$, built on the unique projector property, is proposed as a candidate invariant for block diagonalization and should be tested on the same examples.","If the USPP conjecture holds, spectral theorems for the intermediate spectrum follow; if it fails, the resolvent set may have infinitely many gaps or an ill-defined dimension map."],"supporting_citations":[{"why":"Supplies the invariance claim (Cor. 5.7) that the paper refutes, plus the definition of weak kinematic similarity and the reducibility theorem.","marker":"[12]"},{"why":"Source of optimal ratio maps and the continuous-time version of the non-invariance phenomenon.","marker":"[18]"},{"why":"Provides the nonuniform $\\mu$-dichotomy spectrum and kinematic similarity framework for general growth rates.","marker":"[30]"},{"why":"Normal forms result that relies on the refuted invariance of the nonuniform dichotomy spectrum.","marker":"[31]"},{"why":"Introduced nonuniform exponential dichotomies, the underlying concept being studied.","marker":"[5]"},{"why":"Siegmund's dichotomy spectrum, the foundational spectral theorem this work extends.","marker":"[27]"},{"why":"Continuous-time spectral theorem for nonuniform exponential dichotomies, providing the standard argument referenced for the discrete case.","marker":"[13]"},{"why":"Continuous-time nonuniform dichotomy spectrum and normal forms, representing another theorem that assumes spectral invariance.","marker":"[32]"}],"fun_headline_variants":["Weak similarity breaks spectrum invariance","Nonuniform spectra differ for weakly similar systems","Counterexample: weak similarity changes spectrum","Spectrum not invariant under weak kinematic similarity"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The counterexample's spectra are read off using a spectral theorem and a boundary remark whose proofs are omitted, so if the theorem fails under the paper's nonuniform definition, the claimed spectra are not established.","fun_headline_variants_meta":{"raw":{"variants":["Weak similarity breaks spectrum invariance","Nonuniform spectra differ for weakly similar systems","Counterexample: weak similarity changes spectrum","Spectrum not invariant under weak kinematic similarity"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000134,"raw_usage":{"total_tokens":1119,"prompt_tokens":906,"completion_tokens":213,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":522,"completion_tokens_details":{"reasoning_tokens":162}},"tokens_in":522,"tokens_out":213,"duration_ms":2555,"temperature":1.0,"reasoning_tokens":162,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T21:27:29.514651+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the two spectra in Example 5.8 directly from the definition of nonuniform exponential dichotomy, or verify whether the similarity $S(n)=e^{-an}$ meets the same-$\\varepsilon$ condition required by the weak kinematic similarity definition: if the spectra coincide under that verification, the non-invariance claim is refuted.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the invariance claim (Cor. 5.7) that the paper refutes, plus the definition of weak kinematic similarity and the reducibility theorem."},{"cited_title":"A.: Spectrum invariance dilemma for nonuniformly kinematically similar sys- tems","cited_arxiv_id":null,"evidence_quote":"Source of optimal ratio maps and the continuous-time version of the non-invariance phenomenon."},{"cited_title":"Journal of Function Spaces (2024), 6656183, 13 p","cited_arxiv_id":null,"evidence_quote":"Normal forms result that relies on the refuted invariance of the nonuniform dichotomy spectrum."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduced nonuniform exponential dichotomies, the underlying concept being studied."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Continuous-time spectral theorem for nonuniform exponential dichotomies, providing the standard argument referenced for the discrete case."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Continuous-time nonuniform dichotomy spectrum and normal forms, representing another theorem that assumes spectral invariance."}],"review_version":1}