{"id":"0c7cc317-f100-44af-b4ba-18ddc01a2669","arxiv_id":"2507.07287","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Translation-covariant disordered spin states with finite correlations are shown to admit disordered matrix product decompositions on Banach bundles, and an IID-sampled AKLT state is proven gapless yet exponentially clustering with Tasaki index -1.","lead":"This paper builds a structure theory for disordered quantum spin states that transform covariantly under an ergodic shift, generalizing matrix product state classifications to random environments. It also exhibits a disordered AKLT chain that is gapless almost surely yet has exponentially decaying correlations and a stable time-reversal index.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 3.12 is false: for a translation-invariant cat state, the quotient norm on B^+ is 1 while the n_omega norm is 1/2, so the Banach-bundle construction in Theorem 3.16 is not justified.","rationale":"The reader's weakest assumption (all-omega versus almost-surely in the informal Theorem A) is a real statement-level issue, but I find a more immediate internal inconsistency in the formal proof of the central structure theorem. The proof of Theorem 3.34's (a)=> (b) direction depends on constructing the Banach bundle B_x via Theorem 3.16, whose Fell-Doran input is the continuity of the quotient norm on the fibers. Lemma 3.9 only proves continuity of the auxiliary norm n_omega. Proposition 3.12 is supposed to identify these norms, but its proof relies on the claim that the canonical pairing is an isometric isomorphism between B^- and (B^+)*. That claim does not follow from the one-sided inequality in Remark 3.11. The cat-state example shows the identification actually fails: the quotient norm on the two-dimensional fiber equals 1 for h=diag(1,0), while n_omega(h)=1/2. Hence the proof of Theorem 3.16, and therefore the central theorem, contains a false lemma. This is load-bearing because it invalidates the construction of the transfer apparatus exactly where finite-dimensionality is used. The concern is not that the theorem is necessarily false, but that the written argument is incomplete; a repair using n_omega as the fiber norm is plausible. The reader stated they found no fatal error in the structure theorems, so I disagree on which concern is most load-bearing. The verdict remains CONDITIONAL, since the required fix and, separately, the a.s.-versus-all-omega reconciliation both need to be supplied.","tokens_in":67719,"tokens_out":29357,"duration_ms":338918,"concrete_test":"Take A=M_2 and psi = (1/2)(delta_{0^Z}+delta_{1^Z}) on A_Z. Compute the two quantities for h = diag(1,0) at site 0: (i) n_omega(h) = sup_{||c||<=1} |psi(c tensor h)|, using local projections P_{[-N,-1]} tensor I to show the supremum is 1/2; (ii) ||[h]||_{B^+} = inf_{d in I} ||h+d||, reducing to the 2x2 compression on {Omega_0, Omega_1} to show the infimum is 1. If these numbers are reproduced, Proposition 3.12 is refuted and Theorem 3.16 lacks a valid continuity argument.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The (a)=> (b) direction of Theorem 3.34 passes through Theorem 3.16, which builds the small-correlations bundle via the Fell-Doran theorem. The Fell-Doran input requires a continuous fiber norm on B^+_{x,omega}. Lemma 3.9 proves continuity of the sup-norm n_omega, but Theorem 3.16 needs continuity of the quotient norm and invokes Proposition 3.12 to identify the two. Proposition 3.12's proof asserts that B^-_{x,omega} is canonically isomorphic to (B^+_{x,omega})* and that every unit functional is represented by a class of B^- norm at most 1. However, Remark 3.11 only proves the contractive inequality |eta([c],[a])| <= ||[c]||_{B^-} ||[a]||_{B^+}; the reverse is not established and is false in general. Concrete counterexample: let A=M_2 and let psi be the translation-invariant state on A_Z given by the equal mixture of the all-0 and all-1 product states, psi = (1/2)(delta_{0^Z}+delta_{1^Z}). At any cut, B^+ has dimension 2, detected by the two classical configurations. For h = diag(1,0) at the first site, the quotient norm is ||[h]||_{B^+} = inf_{d in I} ||h+d|| = 1, since any d in I has zero expectation in the two reference configurations and the 2x2 compression on {Omega_0, Omega_1} cannot make the norm of [[1,x],[y,0]] smaller than 1. But n_omega(h) = sup_{||c||<=1} |psi(c tensor h)| = 1/2, achieved by local projections onto the all-zero string on [-N,-1] tensored with identity. Thus Proposition 3.12 is false in the paper's own setting, and the norm-continuity needed for Theorem 3.16 is not supplied. A repair may be possible by using n_omega as the fiber norm, but the proof as written contains a false lemma.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a structure theory for translation-covariant disordered states on one-dimensional quantum spin chains. It defines a class of 'small correlations' states through finite-dimensional quotient spaces B^±_{x,ω}, constructs a Banach bundle from these fibers via the Fell–Doran theorem, and proves a factorization theorem (Theorem 3.34) in the spirit of the Fannes–Nachtergaele–Werner classification. The second half of the paper constructs an IID disordered deformation of the AKLT model and proves, almost surely, that the resulting state has a nearest-neighbor parent Hamiltonian, a deterministic but vanishing bulk spectral gap, exponentially decaying correlations, and a Tasaki index equal to −1. The paper also proves weak-* density of small-correlations states among all translation-covariant states.","tokens_in":68150,"tokens_out":14099,"duration_ms":168454,"significance":"If correct, the paper would provide a genuinely ergodic generalization of the FNW theory and a concrete disordered model exhibiting exponential clustering together with a closed bulk gap—an interesting phenomenon the authors call 'quasi-gapped.' The Banach-bundle transfer apparatus and the aromatic-averaging construction are original and potentially useful tools. The paper also ships explicit Wolfram Language computations in Appendix A, which is a helpful reproducible check of the AKLT ground-state intersections, and the IID AKLT example yields falsifiable predictions (gap closing, index −1) that go beyond the deterministic literature. However, the central bundle construction currently rests on a false norm-identification lemma, and several load-bearing results are deferred to unpublished companion works. The significance of the framework can only be assessed after these points are repaired.","major_comments":[{"comment":"Proposition 3.12 is false as stated. Consider the translation-invariant state ψ = (1/2)(δ_{0^Z} + δ_{1^Z}) on A_Z = (M_2)_Z. At the cut x = −1/2, B^+_{x,ω} is two-dimensional, so the hypothesis holds. For h = diag(1,0) at site 0, every d ∈ I_{x,ω} annihilates both |0^R⟩ and |1^R⟩, so (h+d)|0^R⟩ = |0^R⟩ and hence ||h+d|| ≥ 1; therefore ||[h]||_{B^+} = 1. On the other hand, n_ω(h) = sup_{||c||≤1} |ψ(c⊗h)| = 1/2, attained by taking c to be the projection onto the all-zero string on the left half-chain. Thus ||·||_{B^+} ≠ n_ω even in the finite-dimensional setting. The proof requires the reverse of the inequality in Remark 3.11—that norm-attaining functionals on B^+ are represented by classes in B^- of quotient norm at most 1—which fails: the evaluation-at-0 functional has dual norm 1 but is represented only by classes of quotient norm 2 (e.g. [2P_0^L]). Since Lemma 3.15 and Theorem 3.16 use Proposition 3.12 to obtain continuity of the quotient norm, the Fell–Doran construction of the small-correlations bundle and the (a)=>(b) direction of Theorem 3.34 are not justified as written.","section":""},{"comment":"In the (b)=>(a) direction, after defining the spaces V_ω and W_ω, the proof states that by the Minimality Theorem 3.32 these spaces 'can be mapped isometrically into the small correlations bundle.' But Theorem 3.32(b)—the hypothesis under which such an isomorphism is obtained—is the dense-span condition, which is exactly what needs to be proved and has not been verified at that point. The subsequent boundedness estimates for Q^±_ω therefore rely on the desired conclusion. This is a genuine gap in the proof that finite-dimensionality of a transfer apparatus implies finite-dimensionality of the span in part (a).","section":""},{"comment":"The informal Theorem A and the Abstract state the small-correlations condition only almost surely, whereas the formal Theorem 3.34(a) requires finite-dimensionality for every ω ∈ Ω, and its proof uses each fiber separately. No argument is given for passing from a full-measure subset to all of Ω, or for constructing a measurable Banach bundle on a full-measure subset. This matters for the AKLT application, where nondegeneracy conditions such as those in Lemma 4.7 and Theorem 4.11 hold only almost surely. The authors should either prove the almost-sure version of the structure theorem or explicitly restrict the formal statements to an invariant full-measure set with a measurable bundle.","section":""},{"comment":"The quantitative Lyapunov bound in equation (4.37) has the wrong direction. From (4.38)–(4.39) the almost-sure Lyapunov exponent is Λ = E log cos(2Θ_0), which is negative. Since ∫_0^{π/4} |log cos(2x)| dx = (π/4) log 2 and f ≤ ||f||_∞, one obtains Λ = −∫ |log cos(2x)| f(x) dx ≥ −||f||_∞ (π/4) log 2, not ≤. The stated inequality is false; for example, if the density is concentrated near 0, Λ is close to 0, which is larger than the claimed negative upper bound. The almost-sure exponential decay and negativity of Λ remain valid, but the quantitative estimate must be corrected.","section":""},{"comment":"Lemma 4.5, which supplies the Γ-map expansions and estimates used in the proof of purity (Theorem 4.11), the parent Hamiltonian analysis, and the index calculation, is deferred to the unpublished work [27] ('In Preparation'). Corollary 4.16 (deterministic spectrum) is likewise deferred to the preprint [72]. These are not standard external facts but load-bearing consequences specific to the present construction. The manuscript should include full proofs of these statements or replace the references with published, verifiable sources before the claims of Theorem C and Theorem D can be assessed.","section":""}],"minor_comments":[{"comment":"Equation (4.4) states that the transfer operators E_{a,•} are defined by (4.3) for a ∈ M_2, but the operators in (4.3) are defined for a ∈ M_3; this is a typo.","section":""},{"comment":"The statement of Lemma 2.8 lists criteria (i) and (ii), but the proof refers to items (i) and (iii); the numbering should be made consistent.","section":""},{"comment":"In inequality (3.29) the notation is inconsistent: the first line writes ϕ_{ϑ^{-n}ω} without the prime, while the second line writes ϕ'_{ϑ^nω}. The prime and the sign convention should be aligned with the definition of the aromatic average in (3.18).","section":""},{"comment":"The Wolfram code line 'In [10}:= V = NullSpace[P]' appears to reference an undefined matrix P (probably L), and the displayed output is garbled; the appendix should be checked for consistency.","section":""}],"recommendation":"major_revision","confidential_remarks":"The manuscript relies on two companion works by the same group ([27] and [72]) for load-bearing steps, and on [53] for the exponential-clustering input; this combination makes independent verification difficult. The false Proposition 3.12 is central, and the all-omega versus almost-surely gap in Theorem A would need a substantive revision even after the norm issue is repaired. I recommend against acceptance in the present form, but the overall framework may be salvageable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. First, the paper is a real step toward a structure theory for disordered MPS: the converse classification, the weak* density theorem, and the IID AKLT example with gapless bulk, exponential clustering, and Tasaki index -1 are all genuinely new relative to [52,53,60]. Second, the main structure theorem has a load-bearing error that needs fixing before the classification half can be trusted.\n\nThe problem is Proposition 3.12, which claims that the n_omega norm equals the quotient norm on finite-dimensional fibers. That is false even in the paper's own setting. For the translation-invariant cat state (equal mixture of all-0 and all-1 product states on M_2), the fiber B^+ is two-dimensional; for h = diag(1,0), the quotient norm is 1, but n_omega(h) = 1/2. So the continuity of the quotient norm used in Theorem 3.16 is not supplied, and the Fell-Doran route to Theorem 3.34(a)=>(b) is broken. A repair may exist by using n_omega directly as the fiber norm, but the paper does not provide it. This is not a cosmetic gap; it is the hinge of the converse theorem.\n\nThe rest of the paper is better than the reader's moderate confidence might suggest. The IID AKLT construction is concrete, the probability argument for gap closure via rare regions is largely self-contained, and the Tasaki index computation is explicit and checkable. The weak* density proof is a clean ergodic adaptation of FNW.\n\nSoft spots, in ascending order of seriousness. Theorem 4.19(e) has a sign error in the quantitative Lyapunov bound: the inequality should be >=, not <=. The informal Theorem A states the finite-dimensional span condition almost surely, while the formal theorem requires it at every omega; the bundle apparatus is only constructed under the stronger assumption. Lemma 4.5 and Corollary 4.16 delegate key arguments to an in-preparation paper [27] and a preprint [72], which makes parts of Section 4 hard to verify now.\n\nWho should read this: people working on disordered spin chains, ergodic MPS, and symmetry-protected indices. They will find the example useful even if the converse theorem needs revision. A serious referee should see it; conditional acceptance with a request to fix the bundle norm issue is the right outcome.","headline":"Novel and worth refereeing, but the central bundle construction currently rests on a false norm-identification lemma, so the converse classification is not yet proven as written.","tokens_in":68728,"tokens_out":4811,"would_cite":false,"duration_ms":57788,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["46L60","46L30","82B20","37A55","82B44"],"pacs":[],"model":"deepseek-v4-flash","headline":"Disordered spin states with small correlations factor exactly through finite-dimensional random transfer maps, and the paper's AKLT example shows such states can be gapless, exponentially clustering, and topologically nontrivial.","keywords":["ergodic matrix product states","translation covariant states","Banach bundles","disordered quantum spin chains","AKLT model","parent Hamiltonian","spectral gap closure","Tasaki index"],"falsifier":"Simulate the IID AKLT transfer products and compare the decay of $|\\nu_\\omega(S^z_x S^z_{x+\\ell})|$ with the claimed Lyapunov exponent $E\\log\\cos(2\\omega)$: agreement confirms the exponential clustering, while systematically slower decay would falsify it. For the classification, construct a translation-covariant state whose correlation span is finite-dimensional for almost every $\\omega$ but infinite-dimensional on a single orbit of $\\vartheta$; Theorem 3.34's pointwise equivalence predicts that no finite-dimensional transfer apparatus can implement such a state, so producing one explicitly would falsify the theorem as stated.","tokens_in":67499,"feed_emoji":"🧲","tokens_out":19639,"duration_ms":173886,"temperature":0.7,"pith_summary":"This paper asks which disordered quantum spin states — states whose expectation values depend on a disorder parameter — still admit a matrix-product description, meaning the state is computed by multiplying finitely many matrices along the chain. It proves an exact classification: a weakly* continuous, translation-covariant disordered state has finite-dimensional correlation span at every disorder value if and only if it is implemented by a finite-dimensional Banach bundle whose transfer maps, distinguished section, and boundary functional all depend continuously on the disorder (Theorem 3.34). This is the disordered, ergodic generalization of the finitely correlated state theory of [31], and the same construction shows that small-correlation states are weakly* dense among all translation-covariant states (Theorem 3.41). To show the structure carries real physics, the paper samples the one-parameter AKLT family independently at each site: the resulting state is almost surely pure with a nearest-neighbor, frustration-free, translation-covariant parent Hamiltonian, its bulk spectral gap is almost surely zero, and yet its correlations decay exponentially almost surely (Theorem 4.19). The state is also time-reversal invariant almost surely, with the $\\mathbb{Z}_2$-valued Tasaki index equal to $-1$ (Theorem 4.24), so the example demonstrates that symmetry-protected index data can survive both disorder and a closed gap.","feed_headline":"Gapless yet exponentially clustering: a disordered spin state","feed_subtitle":"A disordered AKLT-type example shows the bulk gap can close while spin correlations still decay exponentially fast.","key_machinery":"The small-correlation bundle. Fix a cut separating the chain into a left half $\\mathcal{A}^-$ and a right half $\\mathcal{A}^+$; the paper declares two continuous right-half functions equivalent when the state $\\psi_\\omega$ cannot distinguish them against any left-half observable, and the quotient fibers $B^+_{\\omega}$ record exactly the correlations across the cut. When these fibers are finite-dimensional, the Fell–Doran theorem (Theorem 2.11) gives the disjoint union a unique Banach bundle topology, and translation together with the Koopman map $\\vartheta$ produce the transfer operators $E_{a,\\omega}$ that shuttle vectors between fibers; the factorization (1.3) is then read off from iterating these maps. A minimality theorem (Theorem 3.32) identifies this bundle as the unique minimal implementation of the state. For the disordered AKLT example the machinery is the one-parameter family of isometries $V_\\omega : \\mathbb{C}^2 \\to \\mathbb{C}^3 \\otimes \\mathbb{C}^2$ with transfer maps $E_{a,\\omega} = V_\\omega^*(a \\otimes \\cdot)V_\\omega$; the gaplessness proof uses rare regions where the sampled angle is close to zero to produce long stretches of slowly decaying correlations, contradicting the exponential clustering bound that a nonzero gap would impose, while the index computation analyzes the Affleck–Lieb twist operator through an explicit combinatorial description of the alternating spin configurations.","core_discovery":"The central claim is Theorem 3.34, stated for a weakly* continuous family of states $\\psi_\\omega$ on the quasi-local algebra $\\mathcal{A}_{\\mathbb{Z}}$ with the translation covariance $\\psi_\\omega \\circ \\tau_k = \\psi_{\\vartheta^k \\omega}$. The following are equivalent: (a) for every $\\omega$, the space $\\{\\psi_\\omega(b \\otimes (\\cdot)) : b \\in \\mathcal{A}^-\\}$ inside the dual of the right half-chain is finite-dimensional; (b) there is a finite-dimensional Banach bundle $E = \\bigsqcup_\\omega E_\\omega$ over the disorder space together with a transfer apparatus — maps $E_{a,\\omega} : E_\\omega \\to E_{\\vartheta^{-1}\\omega}$ linear in $a$, a continuous section $e(\\omega)$, and linear functionals $\\varrho_\\omega$ — such that $\\psi_\\omega(a_m \\otimes \\cdots \\otimes a_n) = \\varrho_{\\vartheta^{m-1}\\omega} E_{a_m,\\vartheta^m\\omega} \\cdots E_{a_n,\\vartheta^n\\omega}(e(\\vartheta^n \\omega))$. The proof takes, for each cut, the quotient of continuous right-half-chain functions by those annihilated against all left-half observables in the state, and the Fell–Doran theorem supplies the bundle topology making the transfer operators continuous. A further result, Theorem 3.41, states that the small-correlation states are weakly* dense among all translation-covariant states, uniformly in $\\omega$. For the disordered AKLT model — a spin-1 valence-bond antiferromagnet whose one-parameter isometries $V_\\omega$ are sampled independently at each site — Theorem 4.19 asserts that the bulk state is almost surely pure, is the ground state of a nearest-neighbor projection-valued parent Hamiltonian, is almost surely gapless, and yet has almost surely exponentially decaying correlations; Theorem 4.24 asserts that it is time-reversal invariant and that the Affleck–Lieb twist expectation converges almost surely to $-1$, giving a well-defined Tasaki index.","pith_inferences":["The pointwise-versus-almost-sure gap in the theorem suggests a measurable-field version of the classification: a state with small correlations only almost surely should still admit a measurable transfer apparatus, with continuity of fiber norms the apparent obstruction in proof.","The rare-region mechanism makes a quantitative prediction: the statistics of slowly decaying stretches, and hence the rate of gap closure, should be governed by the tail probability $P(\\omega_0 < \\delta)$, and a distribution bounded away from the degenerate angle may keep the gap open — a question the paper itself leaves open.","If the Tasaki index is genuinely stable against disorder, other symmetry-protected markers such as string order parameters and reflection indices should also be almost surely constant across realizations, offering a disorder-based route to phase labels without gapped uniqueness.","The analogy to mobility-gapped fermion systems invites the test that low-lying excitations of the GNS Hamiltonian localize near the rare regions; a numerical study of the GNS spectrum on finite chains could indicate whether spatial localization accompanies the closed gap."],"forward_implications":["Every translation-covariant disordered state with small correlations can be evaluated exactly from finite matrix data — transfer maps, section, and boundary functional — making correlation functions, purity, and related observables of such disordered states as computable as for ordinary matrix product states.","Because small-correlation states are weakly* dense among translation-covariant states uniformly in $\\omega$, any weak* continuous property that is uniform in the disorder can be established first on the finite-bundle class and then extended by approximation.","For the IID AKLT state, every translation-covariant finite-range Hamiltonian with this state as a ground state must have a vanishing bulk spectral gap, because the rare-region correlation lower bounds decay too slowly to be compatible with the uniform exponential clustering that a gap would force.","The example motivates a new class of 'quasi-gapped' ground states: almost surely exponentially clustering with random onset lengths, yet with no deterministic onset length and no bulk gap.","The almost sure Tasaki index of $-1$ shows that the $\\mathbb{Z}_2$-valued symmetry-protected index survives disorder and gap closure, so index data remain meaningful for states that are not unique gapped ground states."],"supporting_citations":[{"why":"The deterministic classification of finitely correlated states that Theorem 3.34 extends to the disordered, translation-covariant setting.","marker":"[31]"},{"why":"The Fell–Doran Banach bundle construction used to topologize the small-correlation fibers and make the transfer apparatus continuous in the disorder.","marker":"[33]"},{"why":"The AKLT model whose one-parameter deformation is sampled at each site to build the disordered example.","marker":"[2]"},{"why":"The ergodic matrix product state construction and its randomized exponential clustering estimate that the quasi-gapped example must be reconciled with.","marker":"[53]"},{"why":"The exponential clustering theorem whose uniform correlation bound a nonzero gap would impose and that the rare-region lower bounds contradict.","marker":"[59]"},{"why":"Defines the Z_2-valued Tasaki index through the Affleck–Lieb twist, which Theorem 4.24 evaluates to −1 almost surely.","marker":"[83]"},{"why":"The standard parent Hamiltonian construction for matrix product states, adapted here to the site-dependent disordered case.","marker":"[66]"},{"why":"The companion result that covariant disordered Hamiltonians have deterministic spectrum, used to conclude the bulk gap is almost surely zero.","marker":"[72]"}],"fun_headline_variants":["Ergodic dynamics yield structure theorem for disordered MPS","Disordered spin chains: gapless yet exponentially clustering","AKLT in disorder: gap closes, correlations decay fast, Tasaki -1","Topological dynamics drive ergodic generalizations of MPS","From matrix product states to ergodic disorder: a new structure"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the correlation space is finite-dimensional at every disorder value, because the Banach-bundle construction is proved only under that pointwise condition and has not been shown to work when small correlations hold merely almost surely.","fun_headline_variants_meta":{"raw":{"variants":["Ergodic dynamics yield structure theorem for disordered MPS","Disordered spin chains: gapless yet exponentially clustering","AKLT in disorder: gap closes, correlations decay fast, Tasaki -1","Topological dynamics drive ergodic generalizations of MPS","From matrix product states to ergodic disorder: a new structure"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000638,"raw_usage":{"total_tokens":3091,"prompt_tokens":1245,"completion_tokens":1846,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":861,"completion_tokens_details":{"reasoning_tokens":1761}},"tokens_in":861,"tokens_out":1846,"duration_ms":15601,"temperature":1.0,"reasoning_tokens":1761,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T18:48:44.545865+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Simulate the IID AKLT transfer products and compare the decay of $|\\nu_\\omega(S^z_x S^z_{x+\\ell})|$ with the claimed Lyapunov exponent $E\\log\\cos(2\\omega)$: agreement confirms the exponential clustering, while systematically slower decay would falsify it. For the classification, construct a translation-covariant state whose correlation span is finite-dimensional for almost every $\\omega$ but infinite-dimensional on a single orbit of $\\vartheta$; Theorem 3.34's pointwise equivalence predicts that no finite-dimensional transfer apparatus can implement such a state, so producing one explicitly would falsify the theorem as stated.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The Fell–Doran Banach bundle construction used to topologize the small-correlation fibers and make the transfer apparatus continuous in the disorder."},{"cited_title":"Nachtergaele and R","cited_arxiv_id":null,"evidence_quote":"The exponential clustering theorem whose uniform correlation bound a nonzero gap would impose and that the rare-region lower bounds contradict."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the Z_2-valued Tasaki index through the Affleck–Lieb twist, which Theorem 4.24 evaluates to −1 almost surely."}],"review_version":1}