On Polynomial Progressions Inside Sets of Large Dimension

Ben Krause

On Polynomial Progressions Inside Sets of Large Dimension

T0 review · 3 major / 3 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Sets in [0,1] with dimension close to 1 and Hausdorff content not too small must contain non-trivial four-point polynomial progressions.

desk verdict An abstract-only note with a plausible new continuous analogue of Peluse's polynomial progressions theorem, but the key transfer from Sobolev decay to arbitrary high-dimensional sets is exactly where the stress-test concern lands. read the letter →

arxiv 2508.04680 v2 pith:JZAQKCZN submitted 2025-08-06 math.CA math.CO

Ben Krause This is my paper · ORCID

classification math.CAmath.CO MSC 28A7842B2511B30

keywords polynomialprogressionsHausdorffdimensioncontentSobolevestimatesaveragesFourierfractalsetsgeneralizedarithmetic

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper claims that dimension alone — almost full Hausdorff dimension plus a lower bound on Hausdorff content — forces fractal subsets of the unit interval to contain polynomial progressions with four points. For any triple of polynomials $P_1,P_2,P_3$ vanishing at the origin at different rates, every set $E\subset[0,1]$ with $1-c(\mathcal{P})<\dim_H(E)<1$ and Hausdorff content sufficiently large for its dimension contains a non-trivial configuration $\{x, x-P_1(t), x-P_2(t), x-P_3(t)\}$ with $t\neq 0$. The route runs through a Sobolev, or frequency-decay, estimate for polynomial-averaging operators: when one of the averaged functions has no low frequencies, the $L^1$ norm of the average decays exponentially in the frequency scale, and this functional decay transfers to characteristic functions of sets, converting Fourier vanishing into additive structure. The paper also proves that large Hausdorff dimension plus Fourier dimension above $1/2$ forces a generalized three-term arithmetic progression with rational step ratios. The result is a continuous analogue of a discrete polynomial-progressions theorem, tying Sobolev decay to the geometry of fractals.

What carries the argument

The load-bearing object is a polynomial-average Sobolev estimate of the form $$\left\|\$int_0^{1}$ \prod_{k=1}^m f_k(x - t^k)\,dt\right\|_1 \leq \mathrm{Const}\cdot $2^{{-\mathrm{const}}$\cdot l} \prod_{k=1}^m \|f_k\|_m$$ whenever some $f_i$ vanishes on $\{|\xi|\leq 2^l\}$: exponential decay in the dyadic frequency scale $l$ for averages along polynomial curves. The paper's main theorem relies on a deep version of this estimate, valid for the full class of polynomial triples vanishing at different rates, together with a transfer argument that converts the functional bound into a statement about characteristic functions of fractal sets. The dimension and content hypotheses are exactly what the transf

What would settle it

A concrete test: construct a Cantor-type set $E\subset[0,1]$ with $1-c(\mathcal{P})<\dim_H(E)<1$ and Hausdorff content bounded below whose digit restrictions prevent $x$, $x-P_1(t)$, $x-P_2(t)$, $x-P_3(t)$ from being simultaneously present for any $t\neq 0$; a single such set for one allowed triple would refute the main theorem. Short of that, numerically evaluate the displayed $l$-decay estimate for a monomial triple such as $(t,t^2,t^3)$ and check whether the decay constant stays uniform as $l$ grows, since degeneracy there would break the transfer argument.

Watch

Extended reading notes

Core claim

The central claim is that non-trivial polynomial progressions are guaranteed inside subsets of $[0,1]$ of near-full Hausdorff dimension. Given any triple $\{P_1,P_2,P_3\}$ of polynomials vanishing at $0$ at different rates, every $E\subset[0,1]$ with $1-c(\mathcal{P})<\dim_H(E)<1$ and Hausdorff content bounded below in terms of the dimension contains $x\in E$ and $t\neq 0$ with $\{x, x-P_1(t), x-P_2(t), x-P_3(t)\}\subset E$. The proof converts a uniform Sobolev estimate — exponential decay in $l$ of the $L^1$ norm of a polynomial-averaging operator when one input has Fourier support outside $\{|\xi|\leq 2^l\}$ — into a statement about characteristic functions of sets. A second result forces

Load-bearing premise

The load-bearing premise is that the Sobolev $l$-decay estimate holds uniformly, with constants depending only on the polynomial triple, so that the exponential decay survives the passage from functions to characteristic functions of sets and can be absorbed by the 'sufficiently large' thresholds on $\dim_H(E)$ and Hausdorff content.

Editorial extensions

If this is right

Every $E\subset[0,1]$ meeting the dimension and content thresholds must contain a non-trivial configuration $\{x, x-P_1(t), x-P_2(t), x-P_3(t)\}$ for every allowed polynomial triple — no density or recurrence assumptions needed.
The theorem is the continuous analogue of the discrete polynomial-progressions theorem: near-full dimension alone forces polynomial configurations inside sets.
The proof provides a template: any uniform frequency-decay estimate for a polynomial average yields a geometric progression theorem for sets, with the decay rate governing the dimension threshold $c(\mathcal{P})$.
With Fourier dimension $>1/2$ added, generalized three-term arithmetic progressions with rational step ratios are forced, tying mild Fourier decay directly to additive structure.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

My inference: the exponential nature of the Sobolev decay should yield quantitative bounds on how small $t$ can be in terms of the dimension gap $1-\dim_H(E)$; the paper states no such bounds.
My inference: the Fourier-dimension threshold $1/2$ in the secondary theorem is likely not sharp, and the boundary case $\dim_F(E)=1/2$ is a natural test of whether milder decay suffices.
My inference: the function-to-set transfer suggests that higher-order or higher-dimensional polynomial configurations will follow as soon as the corresponding Sobolev estimates exist, effectively reducing additive combinatorics on fractals to harmonic-analysis estimates.
My inference: the 'sufficiently large' Hausdorff-content hypothesis could be probed computationally on self-similar sets with known dimension and Fourier decay to see how the content threshold behaves in practice.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Quick take: this is a real theorem statement, not a restatement of known work. If the proof goes through, it gives a continuous analogue of Peluse's discrete polynomial progressions result for subsets of the unit interval with Hausdorff dimension close to 1. The second result, for generalized three-term progressions under a Fourier dimension hypothesis, is a nice short proof and has the shape of a sanity check.

The paper does well to lean on Hu-Lie's Sobolev estimates rather than hand-wave around them. The framing is honest: the abstract says "in analogy with" Peluse, and the cited tools are external, not self-citation. The theorem's condition that the polynomials vanish at different rates is natural and matches what you'd need for a polynomial progression to be non-trivial.

Now the soft spot, and it is the same one the stress-test flags. The displayed estimate is an L^1 bound on a polynomial average that decays when one factor is frequency-localized above 2^l. To get a statement about indicator functions of arbitrary sets with only Hausdorff dimension and Hausdorff content, you need to Littlewood-Paley decompose, apply the estimate to each piece, and sum the l-decay. The problem is that Frostman's lemma gives upper regularity, not Fourier decay. There are compact sets of dimension arbitrarily close to 1 with positive content and no Fourier decay. So unless the proof has some other mechanism, the transfer from the Sobolev estimate to characteristic functions is not automatic. The abstract says "sufficiently large" a few times but gives no constants, and the phrase "builds off deep work of Hu-Lie" is doing a lot of work in the one line where the transfer should be.

To be fair: this is an abstract-only review. It is entirely possible that the full proof handles the summation by using the fact that the average is over x and wins a little l^m regularity, or that the threshold depends on the set in a harmless way. The second theorem's Fourier dimension assumption does not contradict the first, but it does make the absence of any Fourier hypothesis in the main theorem conspicuous. A referee will need to check exactly one step: how the l-decay sums to a positive measure statement using only dimension and content.

Who is this for? People working in additive combinatorics and harmonic analysis on the real line. It is a within-subfield contribution, not a breakthrough, but it is a legitimate question and the paper deserves a serious referee. I would send it out, with instructions to verify the transfer step. I would not cite it yet, and I would not bring it to reading group until the full text is available.

Referee Report

3 major / 3 minor

Summary. The paper (arXiv:2508.04680, abstract only) claims a connection between Sobolev-type decay estimates for polynomial averaging operators and the existence of nontrivial polynomial progressions in subsets of [0,1] of sufficiently large Hausdorff dimension. The main stated theorem asserts that if P1,P2,P3 vanish at the origin at different rates and E has dimension greater than 1 - c(P) and sufficiently large Hausdorff content, then E contains a non-trivial progression {x, x-P1(t), x-P2(t), x-P3(t)} with t != 0. The proof is said to build on work of Hu-Lie. A second, weaker theorem gives a short proof that under an additional Fourier dimension > 1/2 assumption, E contains a generalized three-term arithmetic progression with rational coefficients. Since only the abstract was available, no proof, constant specifications, or transfer argument could be inspected.

Significance. If the main theorem is correct, it would be a meaningful continuous analogue of Peluse's discrete polynomial progressions theorem and would demonstrate a new link between Sobolev estimates for polynomial averages and additive combinatorics of fractal sets. The result is cleanly stated and falsifiable, and the secondary Fourier-dimensional result provides a useful baseline. The paper's reliance on external deep work of Hu-Lie is legitimate, not circular. However, the significance cannot be fully assessed from the abstract alone: the core mechanism — how an L^1 Sobolev estimate for functions with frequency gaps transfers to arbitrary high-dimensional sets — is not shown, and the constants that control the thresholds are unspecified.

major comments (3)

[Abstract, displayed Sobolev estimate] The central analytic input is stated as \|∫∏ f_k(x-t^k) dt\|_1 ≤ C 2^{-c l} ∏ \|f_k\|_m whenever some f_i vanishes on |ξ|≤2^l. The constants C and c are left unspecified. The main theorem's hypotheses '1 - const(P) < dim_H(E)' and 'sufficiently large Hausdorff content' depend on these constants and on P. Without quantitative control or monotonicity of these constants, the threshold cannot be verified, and the claimed universality over all polynomial triples with different vanishing rates is not checkable from the abstract.
[Abstract, main theorem (transfer to sets)] The main theorem is asserted for arbitrary E satisfying only Hausdorff dimension and Hausdorff content hypotheses, with no Fourier-dimensional assumption. To conclude about the indicator 1_E from the displayed estimate, one typically needs a Littlewood-Paley decomposition and summation over frequency scales, requiring quantitative Fourier decay or at least a uniform bound on the decomposition that depends only on dimension and content. Hausdorff dimension and positive Hausdorff content imply a Frostman measure with upper regularity μ(B(x,r)) ≤ C r^s, but that does not imply any decay of |μ̂(ξ)|. The abstract does not describe how the transfer avoids this obstruction. The second theorem's explicit Fourier dimension > 1/2 hypothesis makes the absence of such a hypothesis in the main theorem especially striking. This is the load-bearing step that needs a precise proof.
[Abstract, 'unconditional' claim and reliance on Hu-Lie] The phrase 'strongest (unconditional) result builds off deep work of Hu-Lie' is ambiguous. If 'unconditional' means no Fourier dimension assumption, that is consistent with the stated theorem; if it means independent of any conjecture, the description is too terse to verify. The abstract gives no indication of which properties of the Hu-Lie estimates are used, how the estimate's constants depend on the polynomial triple, or whether the required uniformity holds for the full class of P1,P2,P3. Since the entire main theorem rests on this external input, the abstract alone does not provide enough detail to assess correctness.

minor comments (3)

[Abstract, displayed estimate] There is a notation mismatch in the product: the left-hand product uses index k (f_k(x-t^k)), while the right-hand norm product writes \prod_{i=1}^m \| f_k \|_m, mixing i and k. Please correct.
[Abstract, constants] The abstract uses both 'Const' and 'const' in the same display, which is stylistically confusing. Standardize notation for absolute and polynomial-dependent constants.
[General] Because the full manuscript was not available for review, references and context for Hu-Lie and Peluse could not be checked. The authors should ensure the final version clearly states the provenance of each estimate and any prior partial results on continuous polynomial progressions.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found in the abstract; the result rests on external Hu-Lie Sobolev estimates and a transfer argument, not on its own conclusion.

full rationale

The abstract is the only text available, and within it there is no derivation chain that reduces to its own inputs. The main theorem is asserted to follow from an external Sobolev estimate ('builds off deep work of Hu-Lie') and a transfer from analytic estimates to Hausdorff-dimensional sets; the cited support is external, not self-citational, and is not shown to assume the target result. The second theorem adds a Fourier-dimension hypothesis, which is a strengthening, not a circular reuse of the conclusion. Concerns about unspecified constants or the validity of the transfer from L^1 decay to characteristic functions of high-dimensional sets are correctness or quantitativeness issues, not circularity: no equation in the abstract is shown to be equivalent by construction to the theorem it is meant to prove. Under the rule that circularity requires quoting a specific reduction, no such reduction is present, so the honest finding is no significant circularity.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The paper's central proof rests on external Sobolev estimates and on a transfer from analytic estimates to combinatorial statements about sets. No free parameters or fitted values are visible in the abstract. The constants denoted const(P) and the content threshold are structural parts of the theorem, not fitted quantities.

assumptions (3)

domain assumption Hu-Lie Sobolev estimates for polynomial averages are valid for the polynomial families considered in the theorem.
The abstract states that the main theorem builds on deep work of Hu-Lie, but no derivation or statement of these estimates is included in the abstract. If these estimates fail for the required class of polynomials, the proof does not go through.
domain assumption The displayed l-decay estimate, with norm exponent m, can be applied to characteristic functions of sets and transferred to a density statement about polynomial progressions.
The core connection between the Sobolev estimate and the existence of the progression is only asserted as an analogy in the abstract. The transfer from analytic estimates to combinatorial set structure is a load-bearing step that is not shown.
standard math Standard background facts about Hausdorff dimension, Hausdorff content, and Fourier dimension for subsets of [0,1].
These notions are used in the theorem statement without proof, which is normal for this area of analysis and additive combinatorics.

how reviews work

0 comments

Cite this review

Pith. "Pith review of On Polynomial Progressions Inside Sets of Large Dimension." pith.science (2026). https://pith.science/paper/JZAQKCZN

@misc{pith2026250804680,
  author       = {Pith},
  title        = {Pith review of: On Polynomial Progressions Inside Sets of Large Dimension},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JZAQKCZN}},
  note         = {Machine review of arXiv:2508.04680}
}

abstract

In this note we connect Sobolev estimates in the context of polynomial averages e.g. \[ \| \int_0^1 \prod_{k=1}^m f_k(x-t^k) \|_{1} \leq \text{Const} \cdot 2^{-\text{const} \cdot l} \prod_{i=1}^m \| f_k \|_m \] whenever some $f_i$ vanishes on $\{ |\xi| \leq 2^l \}$ to the existence of polynomial progressions inside of sets of sufficiently large Hausdorff dimension, in analogy with work of Peluse in the discrete context. Our strongest (unconditional) result builds off deep work of Hu-Lie and is as follows: suppose that $\mathcal{P} = \{P_1,P_2,P_3\}$ vanish at the origin at different rates, and that $E \subset [0,1]$ has sufficiently large Hausdorff dimension, \[ 1 - \text{const}(\mathcal{P}) < \text{dim}_H(E) < 1 \] and Hausdorff content bounded away from zero, sufficiently large in terms of its dimension. Then $E$ contains a non-trivial polynomial progression of the form \[ \{ x , x - P_1(t), x - P_2(t), x - P_3(t) \} \subset E, \; \; \; t \neq 0. \] We also provide a short proof that whenever $E$ has sufficiently large Hausdorff dimension and Fourier dimension $> 1/2$, it necessarily contains a non-trivial generalized three-term arithmetic progression of the form \[ \{ x, x - \theta_1 t, x- \theta_2 t\} \subset E, \; \; \; \theta_i \in \mathbb{Q},\ t \neq 0.\]

Discussion (0). Continue with ORCID to comment.

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.