REVIEW 3 major objections 5 minor 34 references
Lower bounds on concentration through Borel transforms and quantitative singularity of spectral measures near the arithmetic transition
T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash
Pith's one-line read For the almost Mathieu operator, upper bounds on packing and multifractal dimensions of spectral measures now capture the sharp arithmetic transition, tending to zero as ln λ approaches β(α) from below.
desk verdict New Borel-transform criterion is solid and independently useful; the AMO application is important but hinges on an asserted extension of [20] that a referee should check. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The central object is the m-Borel transform $J_{\mu,m}(x,\varepsilon)=\varepsilon^m\int_{\mathbb{R}} d\mu(y)/(|x-y|^m+\varepsilon^m)$, a one-parameter deformation of the usual Borel transform that recovers $\varepsilon\,\Im\int d\mu/(x-y-i\varepsilon)$ when $m=2$. The key mechanism is the criterion that a positive $\liminf$ of $\varepsilon^{-\varsigma}J_{\mu,m}(x,\varepsilon)$ implies a lower bound on $\mu([x-\varepsilon,x+\varepsilon])$ and hence an upper bound on the local upper exponent $\gamma^+_\mu(x)$; this turns boundary growth of m-functions into packing and multifractal dimension bounds. For the almost Mathieu operator, the machinery is completed by partial localization, which gives exponential decay of generalized eigenfunctions on scales between $q_n^{t_1}$ and $q_n^{t_2}$, and by Lemma 8.2, which converts that decay into lower bounds on products of norms of solutions and ultimately into $\Im M_j(E+i\varepsilon)\ge\varepsilon^{-t}$ for $t<\ln\lambda/(2\beta-\ln\lambda)$. These m-function lower bounds are exactly the boundary behavior the general criterion needs, and applying the criterion with $m=2$ yields Theorems 1.1 and 1.2.
What would settle it
Read the proof of [20, Theorem 3.5] and mark each inequality that uses the original hypothesis $\lambda>e^{\beta(\alpha)}$; if any of them requires $\ln\lambda-\beta(\alpha)>0$ rather than the positivity of $\ln\lambda+8\ln(sq_{n-n_0}/q_{n-n_0+1})/q_{n-n_0}$ stated in Remark 6.5, then Lemma 7.3, Lemma 8.2, and the m-function lower bound Lemma 9.2 all fail, and with them Theorems 1.1 and 1.2. A successful check would instead confirm that only the displayed positivity is used, which is exactly what the paper asserts.
Extended reading notes
Core claim
The central discovery is that the arithmetic transition of the almost Mathieu operator is quantitative rather than just a yes/no threshold. For every $\alpha$-Diophantine phase $\theta$ and every $\phi\in\ell^2(\mathbb{Z})$, Theorem 1.1 gives $\dim_P^+(\mu_\phi)=0$ for $\lambda\ge e^{\beta(\alpha)}$ and $\dim_P^+(\mu_\phi)\le 2(1-\ln\lambda/\beta(\alpha))$ for $\lambda<e^{\beta(\alpha)}$, while Theorem 1.2 gives $D^+_{\mu_\phi}(q)\le (2\beta(\alpha)-2\ln\lambda)/(2\beta(\alpha)-\ln\lambda)$ for $q\ge 3/2$. The bounds are transition-capturing: both right-hand sides tend to $0$ as $\ln\lambda\uparrow\beta(\alpha)$, unlike earlier quantitative estimates that stayed bounded away from zero. The proof establishes two things the authors call firsts: partial localization, meaning exponential decay of generalized eigenfunctions on intermediate scales in the singular continuous regime, and a general mechanism by which lower bounds on concentrations of a Borel measure follow from boundary behavior of its m-Borel transforms. In the application, the m-function lower bounds of Lemma 9.2 combine with the general criterion at $m=2$ to force the dimension bounds.
Load-bearing premise
The load-bearing premise is that a regularity theorem for nonresonant sites, proved in an earlier paper under the strict condition $\lambda>e^{\beta(\alpha)}$, remains valid when $1<\lambda\le e^{\beta(\alpha)}$ because only the positivity of a certain displayed quantity is used in that proof; the current paper does not reproduce the original proof, so this extension must be checked in [20, Theorem 3.5]. If the strict inequality is actually needed anywhere, the partial-localization step, the m-function lower bounds, and both main theorems lose their support.
Editorial extensions
If this is right
- At $\lambda\ge e^{\beta(\alpha)}$, every spectral measure of the almost Mathieu operator with Diophantine phase has upper packing dimension zero; the new case is $\lambda=e^{\beta(\alpha)}$, where singular continuous spectrum can still occur.
- For $1<\lambda<e^{\beta(\alpha)}$, the upper packing dimension is at most $2(1-\ln\lambda/\beta(\alpha))$, so the singular continuous spectrum becomes quantitatively more singular as the coupling approaches the transition.
- For every $q\ge 3/2$, $D^+_{\mu_\phi}(q)\le (2\beta(\alpha)-2\ln\lambda)/(2\beta(\alpha)-\ln\lambda)$, giving a multifractal analogue of the vanishing bound.
- The general criterion (Theorem 1.3 and its corollaries) lets one bound the packing, Hausdorff, and multifractal dimensions of any Borel measure from the boundary behavior of its m-Borel transforms, independent of the almost Mathieu operator.
- The authors expect the same formulation, with $\ln\lambda$ replaced by $\min_E L(E)$, to apply to other models with sharp arithmetic transitions, using the same partial-localization and m-function input.
Reading between the lines
- The m-Borel transform criterion is stated for arbitrary Borel measures, so it could be applied outside quasiperiodic operators, for example to measures whose m-functions are known to grow like a power law, as a general tool for bounding packing dimension from boundary behavior.
- The factor 2 in Theorem 1.1 is likely not sharp: the bound becomes vacuous when $\beta(\alpha)\gg\ln\lambda$, so a sharper subordinacy estimate might replace it by $1-\ln\lambda/\beta(\alpha)$ and change the predicted dimension near the transition.
- The threshold $q\ge 3/2$ in Theorem 1.2 comes from applying the general criterion with $m=2$; varying $m$ could lower this threshold or improve the multifractal bound, an extension that the paper's mechanism makes available.
- If the same partial localization works for other models with sharp arithmetic transitions, the paper's expectation that $\ln\lambda$ be replaced by $\min_E L(E)$ would turn the packing-dimension vanishing near the transition into a universal phenomenon; that is a testable program but not yet a theorem.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper develops a new general criterion (Theorem 1.3) showing that a pointwise lower bound on the m-Borel transform J_{\mu,m}(x,\epsilon) of order \epsilon^\varsigma implies an upper bound on the upper concentration exponent \gamma_\mu^+(x), with corollaries for packing and multifractal dimensions. This criterion is then applied to the almost Mathieu operator with \alpha-Diophantine phase in the hyperbolic regime 1<\lambda\le e^{\beta(\alpha)}. The main results, Theorems 1.1 and 1.2, give upper bounds \dim_P^+(\mu_\phi)\le 2(1-\ln\lambda/\beta(\alpha)) and D_{\mu_\phi}^+(q)\le (2\beta(\alpha)-2\ln\lambda)/(2\beta(\alpha)-\ln\lambda) for q\ge 3/2. These bounds tend to zero as \ln\lambda approaches \beta(\alpha) from below, which the authors present as the first quantitative capture of the arithmetic transition from the singular-continuous side. The proof combines partial localization of generalized eigenfunctions (Theorem 7.1) with power-law subordinacy estimates on m-functions, building on the hierarchical machinery of the authors' earlier work [20].
Significance. If the proof is completed, the paper would provide the first upper bounds on packing and multifractal dimensions of spectral measures that vanish at the arithmetic transition, a genuinely new and important phenomenon. The general m-Borel-transform criterion in Sections 3–5 is original, appears correct, and is likely to find independent use; the proofs there are detailed and self-contained. The application, however, is more fragile: it depends on several imported statements from [20] whose validity in the newly needed parameter range is asserted rather than demonstrated, and one auxiliary lemma (Lemma 9.3) has a proof that does not appear to be valid as written. These issues are load-bearing for the central claims, so the paper needs substantial revision before the main theorems can be considered established.
major comments (3)
- [Remark 6.5 and Theorem 6.4] Theorem 6.4, quoted from [20, Thm 3.5], is originally stated only for \lambda>e^{\beta(\alpha)}. Remark 6.5 asserts that the proof only uses the positivity of \ln\lambda+8\ln(sq_{n-n_0}/q_{n-n_0+1})/q_{n-n_0}, so the result extends to 1<\lambda\le e^{\beta(\alpha)}. This extension is load-bearing: it is used in Lemma 7.3, Theorem 7.1, Lemma 8.2, Lemma 9.2, and Theorem 9.4, and thus indirectly in Theorems 1.1 and 1.2. The paper does not reproduce or verify the proof of [20, Thm 3.5] in the extended range, so a reader cannot check that every step (including large-deviation estimates controlling the exceptional set where |P_k| is small) survives when \lambda\le e^{\beta(\alpha)}. Please provide a detailed proof of the extension, or at least a precise statement of which inequalities in the original proof are used and why each remains valid in the range 1<\lambda\le e^{\beta(\alpha)}.
- [Lemma 9.2 and parameter t1 (eq. (69))] The proof of Lemma 9.2 defines t_1=(\beta-\ln\lambda)/(\beta+\sigma) and uses the exponent g=\frac12 \ln\lambda/(t_1\beta)-\varepsilon from Lemma 8.2. When \lambda=e^{\beta(\alpha)}, we have t_1=0, so the expression for g is undefined, and the proof of Lemma 8.2 cannot be applied as written. This equality case is explicitly claimed as new in Case 1 of Theorem 1.1 (Remark 1.1), so it is not a peripheral concern. The authors should either give a separate argument for \lambda=e^{\beta(\alpha)} (for example, by a limiting argument from \lambda<e^{\beta(\alpha)} if such a passage is justified, or by showing directly that the lower bound in Lemma 8.2 holds with any finite polynomial exponent when t_1=0) or adjust the statement of Theorem 1.1 to exclude this case.
- [Lemma 9.3 and its proof] Lemma 9.3 claims that the exceptional sets S_0=\{E:2x_0(E)\in\mathbb Z\} and S_1=\{E:2x_0(E)\in\frac12+\mathbb Z\} have zero measure with respect to \mu_{\delta_0} and \mu_{\delta_1}, respectively. The proof, however, argues that for E\in S_0 one has \Im M_1(E+i\epsilon)\to 0 and then invokes 'basic spectral theory' to conclude \mu_{\delta_0}(S_0)=0. This implication is not valid for general Borel measures: the vanishing of the imaginary part of the Borel transform at a point E only indicates that E has zero absolutely continuous density and is not an atom; it does not imply that the set of such points has zero measure with respect to the measure. Since the argument is the only reason given for excluding S_0 and S_1, and the lower bounds on \Im M_1 and \Im M_2 in Theorem 9.4 depend on this exclusion, the proof of Lemma 9.3 needs to be substantially reworked or replaced. The authors should provide a correct argument that these exceptional sets are null (or explain why they are null in the context of the almost Mathieu operator with Diophantine phase).
minor comments (5)
- [Equation (82)] The bound in (82) is written as \|u\|_{L,L}\le C(E)L^{1/2\ln L}, which is ambiguous: it could be read as L^{1/2}\ln L or as L^{(1/2)\ln L}. The later use in Lemma 9.2 requires the bound b(L)\le C(E)L^{1+\epsilon}, which is compatible with L^{1/2}\ln L but not with L^{(1/2)\ln L}. Please clarify the notation.
- [Remark 6.5] The displayed inequalities in Remark 6.5 are typeset in a confusing way, for example '8\ln qt2 n qn−n0'. Please replace these with clearer notation such as q_n^{t_2} and q_{n-n_0} to make the argument readable.
- [Section 2.1] The phrases 'Corollaries of 1.9 and 1.4' and 'Corollaries of 1.5 and 1.6' appear in the proof paragraphs; they should be 'Corollaries 1.9 and 1.4' and 'Corollaries 1.5 and 1.6'.
- [Lemma 9.3] The proof of Lemma 9.3 ends with '\mu_{\delta_0}(S_1)=0', but the statement concerns S_0; this appears to be a typo and should be corrected to '\mu_{\delta_0}(S_0)=0'.
- [Proof of Theorem 1.1] The simplification leading to (108), namely that 2\varsigma/(2-\varsigma) with \varsigma=2(\beta-\ln\lambda)/(2\beta-\ln\lambda) equals (2\beta-2\ln\lambda)/\beta, is correct but not immediately transparent; writing the intermediate algebraic step would improve readability.
Circularity Check
No circularity: the new m-Borel-transform criterion is proved from first principles, and the almost Mathieu application uses published structural theorems as external inputs; the Remark 6.5 extension is a non-circular but unverified proof detail.
full rationale
The derivation chain is not circular. Theorems 1.3, 1.7, and 1.8 are proved directly from measure estimates, Fatou's lemma, and Hölder arguments; the constants m and ς are assumptions of the criterion, not quantities fitted to the target dimensions. The application fixes m = 2 and derives ℑM_j(E + iǫ) ≥ ǫ^{-t} from Lemma 9.2, which uses Theorem 2.3 (from Killip–Kiselev–Last) and Lemma 8.2. Lemma 8.2 is proved from Theorem 7.1/Lemma 7.3, which rely on Theorem 6.4 quoted from [20]. That theorem is a published, parameter-free structural result whose assumptions (Diophantine θ, positivity of the displayed decay rate, E a generalized eigenvalue) do not include the packing or multifractal dimensions being bounded; hence it is independent evidence, not a restatement of the conclusions. The only load-bearing caveat is Remark 6.5, where the authors extend [20, Thm 3.5] from λ > e^β to 1 < λ ≤ e^β by asserting that 'only the fact that ln λ + 8 ln(s q_{n−n0}/q_{n−n0+1})/q_{n−n0} > 0 is used in the proof.' This is an unverified proof-reading claim, and if false the lower bounds in Lemmas 7.3, 8.2, and 9.2 would lose support; however, it does not make the argument circular, because it does not assume the dimension upper bounds or the concentration lower bounds that are being proved. No fitted parameter is relabeled as a prediction, and no known result is merely renamed. Accordingly no circular step is exhibited.
Assumptions & free parameters
free parameters (4)
- t1 =
(beta - ln lambda)/(beta + sigma), sigma -> 0
- t2 =
any value in ((9 beta - ln lambda)/(9 beta), 1)
- sigma =
-> 0
- epsilon =
-> 0
assumptions (6)
- standard math Spectral theorem and m-function identities (27)-(32) relating m-functions of half-line problems to spectral measures of delta_0 and delta_1.
- standard math Theorem 8.1 (Last-Simon [25]): for almost every energy with respect to the spectral measure, there exists a generalized eigenfunction with ||u||_{L,L} <= C(E) L^{1/2} ln L.
- standard math Bourgain-Jitomirskaya [8]: for the almost Mathieu operator with lambda > 1, the Lyapunov exponent is constant ln lambda on the spectrum.
- domain assumption Hierarchical eigenfunction estimates of [20] (Theorems 6.4, 7.4 and Lemma 6.3) apply in the regime 1 < lambda <= e^beta; the paper asserts only positivity of ln lambda + 8 ln(s q_{n-n0}/q_{n-n0+1})/q_{n-n0} is used (Remark 6.5).
- domain assumption The phase theta is alpha-Diophantine; this is a hypothesis of the theorems and holds for almost every theta.
- domain assumption alpha is irrational with 0 < beta(alpha) < infinity; this is the nontrivial regime for the application.
Cite this review
Pith. "Pith review of Lower bounds on concentration through Borel transforms and quantitative singularity of spectral measures near the arithmetic transition." pith.science (2026). https://pith.science/paper/JQOZGLUL
@misc{pith2026250112153,
author = {Pith},
title = {Pith review of: Lower bounds on concentration through Borel transforms and quantitative singularity of spectral measures near the arithmetic transition},
year = {2026},
howpublished = {\url{https://pith.science/paper/JQOZGLUL}},
note = {Machine review of arXiv:2501.12153}
}
read the original abstract
We develop tools to study arithmetically induced singular continuous spectrum in the neighborhood of the arithmetic transition in the hyperbolic regime. This leads to first transition-capturing upper bounds on packing and multifractal dimensions of spectral measures. We achieve it through the proof of partial localization of generalized eigenfunctions, another first result of its kind in the singular continuous regime. The proof is based also on a general criterion for lower bounds on concentrations of Borel measures as a corollary of boundary behavior of their Borel-type transforms, that may be of wider use and independent interest.
Reference graph
Works this paper leans on
-
[20]
S. Jitomirskaya and W. Liu. Universal hierarchical structure o f quasiperiodic eigenfunctions. Ann. of Math. (2) , 187(3):721–776, 2018
work page 2018
-
[22]
S. Jitomirskaya and S. Zhang. Quantitative continuity of singula r continu- ous spectral measures and arithmetic criteria for quasiperiodic Sc hr¨ odinger operators. J. Eur. Math. Soc. (JEMS) , 24(5):1723–1767, 2022
work page 2022
-
[1]
A. Avila. The absolutely continuous spectrum of the almost Mathie u oper- ator. arXiv preprint arXiv:0810.2965 , 2008
arXiv 2008
-
[2]
A. Avila and S. Jitomirskaya. The Ten Martini Problem. Ann. of Math. (2), 170(1):303–342, 2009
work page 2009
- [3]
- [4]
-
[5]
J.-M. Barbaroux, F. Germinet, and S. Tcheremchantsev. Frac tal dimensions and the phenomenon of intermittency in quantum dynamics. Duke Math. J., 110(1):161–193, 2001
work page 2001
-
[6]
J.-M. Barbaroux, F. Germinet, and S. Tcheremchantsev. Gene ralized fractal dimensions: equivalences and basic properties. J. Math. Pures Appl. (9) , 80(10):977–1012, 2001
work page 2001
Show all 34 references
-
[7]
Bourgain
J. Bourgain. Green ’s function estimates for lattice Schr¨ odinger operators and applications , volume 158 of Annals of Mathematics Studies . Princeton University Press, Princeton, NJ, 2005
2005
-
[8]
Bourgain and S
J. Bourgain and S. Jitomirskaya. Continuity of the Lyapunov exp onent for quasiperiodic operators with analytic potential. volume 108, pages 1 203–
-
[9]
Damanik and S
D. Damanik and S. Tcheremchantsev. Scaling estimates for solut ions and dynamical lower bounds on wavepacket spreading. J. Anal. Math. , 97:103– 131, 2005
2005
-
[10]
del Rio, S
R. del Rio, S. Jitomirskaya, Y. Last, and B. Simon. Operators w ith singular continuous spectrum. IV. Hausdorff dimensions, rank one pertur bations, and localization. J. Anal. Math. , 69:153–200, 1996
1996
-
[11]
Ge and S
L. Ge and S. Jitomirskaya. Hidden subcriticality, symplectic stru cture, and universality of sharp arithmetic spectral results for type I opera tors. arXiv preprint arXiv:2407.08866 , 2024
2024
-
[12]
L. Ge, J. You, and X. Zhao. The arithmetic version of the frequ ency tran- sition conjecture: new proof and generalization. Peking Math. J. , 5(2):349– 364, 2022
2022
-
[13]
Guarneri and H
I. Guarneri and H. Schulz-Baldes. Lower bounds on wave pack et propaga- tion by packing dimensions of spectral measures. Math. Phys. Electron. J. , 5:Paper 1, 16, 1999
1999
-
[14]
Jitomirskaya
S. Jitomirskaya. Dual lyapunov exponents and operators of t ype I. To appear in the Proceedings of the v (W M)2 2022
2022
-
[15]
Jitomirskaya
S. Jitomirskaya. One-dimensional quasiperiodic operators: glo bal theory, duality, and sharp analysis of small denominators. In ICM—International 28 SVETLANA JITOMIRSKAYA, WENCAI LIU, AND SERGUEI TCHEREMC HANTSEV Congress of Mathematicians. Vol. 2. Plenary lectures , pages 109...
2023
-
[16]
Jitomirskaya and I
S. Jitomirskaya and I. Kachkovskiy. Sharp arithmetic delocaliza tion for quasiperiodic operators with potentials of semi-bounded variation. arXiv preprint arXiv:2408.16935 , 2024
2024 arXiv
-
[17]
Jitomirskaya and I
S. Jitomirskaya and I. Kachkovskiy. Sharp arithmetic localizatio n for quasiperiodic operators with monotone potentials. arXiv preprint arXiv:2407.00703, 2024
2024 arXiv
-
[18]
Jitomirskaya and Y
S. Jitomirskaya and Y. Last. Power-law subordinacy and singula r spectra. I. Half-line operators. Acta Math., 183(2):171–189, 1999
1999
-
[19]
Jitomirskaya and W
S. Jitomirskaya and W. Liu. Arithmetic spectral transitions for the Mary- land model. Comm. Pure Appl. Math. , 70(6):1025–1051, 2017
2017
-
[21]
Jitomirskaya and M
S. Jitomirskaya and M. Powell. Logarithmic quantum dynamical bo unds for arithmetically defined ergodic Schr¨ odinger operators with smo oth po- tentials. Analysis at large: dedicated to the life and work of Jean Bour gain, pages 172 – 201, 2022
2022
-
[23]
S. Y. Jitomirskaya and Y. Last. Power law subordinacy and singu lar spectra. II. Line operators. Comm. Math. Phys. , 211(3):643–658, 2000
2000
-
[24]
Killip, A
R. Killip, A. Kiselev, and Y. Last. Dynamical upper bounds on wave packet spreading. Amer. J. Math. , 125(5):1165–1198, 2003
2003
-
[25]
Last and B
Y. Last and B. Simon. Eigenfunctions, transfer matrices, and absolutely con- tinuous spectrum of one-dimensional Schr¨ odinger operators. Invent. Math. , 135(2):329–367, 1999
1999
-
[26]
X. Li, J. You, and Q. Zhou. Exact local distribution of the absolu tely continuous spectral measure. arXiv preprint arXiv:2407.09278 , 2024
2024 arXiv
-
[27]
W. Liu. Small denominators and large numerators of quasiperiod ic Schr¨ odinger operators.Peking Mathematical Journal to appear
-
[28]
W. Liu. Power law logarithmic bounds of moments for long range op erators in arbitrary dimension. J. Math. Phys. , 64(3):Paper No. 033508, 11, 2023
2023
-
[29]
W. Liu, M. Powell, and X. Wang. Quantum dynamical bounds for long-range operators with skew-shift potentials. arXiv preprint arXiv:2411.00176, 2024
2024 arXiv
-
[30]
Shamis and S
M. Shamis and S. Sodin. Upper bounds on quantum dynamics in arb itrary dimension. J. Funct. Anal. , 285(7):Paper No. 110034, 20, 2023
2023
-
[31]
B. Simon. Almost periodic Schr¨ odinger operators. IV. The Maryland model. Ann. Physics , 159(1):157–183, 1985. QUANTITATIVE SINGULARITY 29
1985
-
[32]
B. Simon. Equilibrium measures and capacities in spectral theory . Inverse Probl. Imaging , 1(4):713–772, 2007
2007
-
[33]
G. Teschl. Spectral theory for Jacobi operators . ProQuest LLC, Ann Arbor, MI, 1995. Thesis (Ph.D.)–University of Missouri - Columbia. (Svetlana Jitomirskaya) Department of Mathematics, University of California, Berkeley, California 94720, USA Email address : sjitomi@berkeley....
1995
-
[1218]
Dedicated to David Ruelle and Yasha Sinai on the occasio n of their 65th birthdays
2002. Dedicated to David Ruelle and Yasha Sinai on the occasio n of their 65th birthdays
2002
Reviewed August 10, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.