REVIEW 1 major objections 7 minor 25 references
On a Ramsey--Tur\'{a}n variant of Roth's theorem
T0 review · 1 major / 7 minor · reviewed 2026-08-06 · deepseek-v4-flash
Pith's one-line read For a homogeneous linear equation over F_p, solution-free sets with sublinear Cayley independence are small exactly when a nonempty subset of the coefficients sums to zero, completing the Ramsey–Turán analogue of Roth's theorem.
desk verdict A clean, genuinely new Ramsey–Turán variant of Roth's theorem over F_p with a complete classification; the proof is solid and the small blemishes are typo-level. 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 engine is a lemma about directed graph systems. A restricted digraph system consists of properly edge-colored digraphs on a common vertex set together with a function assigning each vertex a small set of forbidden colors; a proper rainbow directed path is a path whose $i$-th edge lies in the $i$-th digraph, whose edge colors are all distinct, and whose colors avoid the forbidden sets of the two endpoints. The lemma shows that when each digraph has independence number at most $|V|/(100^{k'}\ell^2)$, such a path of length $k'$ must exist. In the degenerate case the Cayley digraphs generated by coefficient dilations $c_iA$ are the digraphs, colors are the elements of $A$, and the zero-sum subset of coefficients provides a disjoint family of solutions of a smaller subequation; the rainbow path then threads these solutions together into a full solution of $L$.
What would settle it
Take the Schur equation $x+y-z=0$ and fix a small $\varepsilon>0$; Theorem 3.1 says $\limsup_{p\to\infty}D(L,\varepsilon,p)/p\le 100^{4}\cdot27\,\varepsilon$, while Theorem 1.3 says the true order is $\Theta(\varepsilon)$. Any infinite family of primes whose normalized maximum size exceeds that upper bound, or any degenerate equation whose normalized maximum size decays more slowly than linearly in $\varepsilon$, would contradict the quantitative claims; for the non-degenerate direction, checking the paper's explicit construction for $x+y+z=0$ on moderate primes is a finite verification that its independence is $O(p/\log p)$.
Extended reading notes
Core claim
Let $L:c_1x_1+\cdots+c_kx_k=0$ be a homogeneous linear equation with $k\ge 3$ nonzero integer coefficients, and call $A\subseteq\mathbb{F}_p$ solution-free when no $k$-tuple of distinct elements of $A$ solves $L$. The theorem states that every solution-free $A$ with $\alpha(\mathrm{Cay}_{\mathbb{F}_p}(A))=o(p)$ has $|A|=o(p)$ exactly when $L$ is degenerate, meaning some nonempty subset of its coefficients sums to zero. For non-degenerate equations the paper constructs solution-free sets of linear size whose Cayley graphs still have independence $O(p/\log p)$, so sublinear independence does not force smallness. For degenerate equations the proof yields $d(L,\varepsilon)\le 100^{k+1}k^3\varepsilon$, where $d(L,\varepsilon)$ is the asymptotic maximum density of solution-free sets with independence at most $\varepsilon p$.
Load-bearing premise
The degenerate direction leans on the classical density theorem that every constant-positive-density subset of $\mathbb{F}_p$ contains a solution to any zero-sum equation in at least three variables with distinct entries; if that theorem or its distinct-entry convention failed, the classification would not follow.
Editorial extensions
If this is right
- For any degenerate equation, the maximum density of a solution-free set with $\alpha(\mathrm{Cay}_{\mathbb{F}_p}(A))\le\varepsilon p$ is at most $100^{k+1}k^3\varepsilon$, so the density tends to zero with $\varepsilon$.
- For any non-degenerate equation, there is a fixed $\beta>0$ and a solution-free set of size at least $\beta p$ with $\alpha=O(p/\log p)$, which is $o(p)$; hence sublinear independence does not suffice to force smallness unless a zero-sum coefficient subset exists.
- For the Schur equation $x+y-z=0$, the quantitative rate is $d(L,\varepsilon)=\Theta(\varepsilon)$, so the general linear upper bound cannot be improved in the exponent.
- For every equation whose coefficients have nonzero total sum but some zero-sum subset, the lower bound is $\varepsilon^{\Theta(1)}$, matching the qualitative classification.
- A set with $\alpha(\mathrm{Cay}_{\mathbb{F}_p}(A))=o(p)$ is a difference intersector, meeting $B-B$ for every $B$ of positive linear density; the theorem therefore classifies when every difference-intersecting solution-free set is small.
Reading between the lines
- The same classification is plausibly portable to $\mathbb{Z}$ with a difference-intersector condition; the paper notes the translation is routine, so an exact integer analogue would be a direct check rather than a new theorem.
- The quantitative gap between $O(\varepsilon)$ and $\varepsilon^{\Theta(1)}$ suggests testing intermediate degenerate equations: equations with an isolated zero-sum pair may already force the linear rate, while equations needing larger zero-sum subsets may not.
- If the independence assumption is strengthened to $o(p/\log p)$ or smaller, the paper's non-degenerate constructions fail, so the classification boundary could shift; whether it does is an open question the paper explicitly raises.
- The use of sparse high-girth graphs in the lower bound suggests that progress in Ramsey–Turán graph constructions would translate directly into sharper polynomial exponents for non-degenerate equations.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proves a Ramsey-Turán analogue of Roth's theorem for homogeneous linear equations over F_p. The main result (Theorem 1.2) classifies equations for which every solution-free set A with independence number α(Cay_{F_p}(A))=o(p) must have size o(p): this holds exactly when some nonempty subset of the coefficients sums to zero. The degenerate direction is proved with the quantitative bound d(L,ε) ≤ 100^{k+1} k^3 ε, using a new rainbow directed path lemma combined with Roth's theorem; the non-degenerate direction is witnessed by an explicit construction with |A| ≥ βp and α(Cay(A)) = O_L(p/log p). The paper also establishes d(L,ε)=Θ(ε) for Schur's equation and d(L,ε)=ε^{Θ(1)} for every non-degenerate equation possessing a proper zero-sum subset of coefficients.
Significance. If the results hold, the paper gives a clean classification of density-regular homogeneous equations under the Ramsey-Turán independence condition, a natural structural restriction introduced by Erdős and Sárközy. The proof strategy is original: the degenerate case is reduced to a graph-theoretic rainbow path lemma, while the non-degenerate case is handled by explicit constructions that transfer Ramsey graph lower bounds into the additive setting. The paper is careful about the distinct-elements convention in the definition of solution-free sets, and the main classification proof is detailed and self-contained apart from standard tools (Roth's theorem, Caro-Wei, Ramsey graph lower bounds). The quantitative bounds, though probably not optimal except for Schur's equation, give a clear picture of the ε-dependence and raise interesting open problems.
major comments (1)
- [Section 4, Theorem 4.4] The proof claims that a graph G with q=t^k vertices and α(G)≤t^{k−1} is guaranteed by Theorem 4.3. With n=t^k, Theorem 4.3 only gives α(G)≤C t^{k−1} log t; the extra log factor invalidates the pigeonhole step α(Cay_Fp(X))≤p/t, because q/t is no longer larger than α(G). The theorem is likely repairable by choosing n≈t^k polylog(t) and absorbing the resulting polylog factors into the exponent C, but as written the proof of d(L,ε)≥ε^C is incomplete.
minor comments (7)
- [Section 3.1, Lemma 3.2] In the induction step, the second and fourth bullets refer to 'the color of →v_i v_{i+1} in D_{i+1}' but the edge from v_i to v_{i+1} in the proper rainbow path lies in D_i; the index should be D_i.
- [Section 3.2, Theorem 3.3] In the proof of the solution-free claim, the sentence 'assume y_1 is the largest among the y_i's' should read 'assume z_1 is the largest among the z_i's'.
- [Section 4, Theorem 4.2] The displayed inclusion 'X_i±X_i ⊆ (8p/9,p) ∪ [0,p/9)' is not literally correct: X_i+X_i is contained in [0,2p/9), while X_i−X_i is contained in (8p/9,p) ∪ [0,p/9). The argument only needs both sets to be disjoint from Y=[p/3,4p/9], so the proof is unaffected, but the inclusion should be stated accurately.
- [Section 4, Theorem 4.4] The base-r digit argument is written as if the pairs {a_j,b_j} are distinct, but coefficients c_j can repeat or cancel; the footnote about partial cancellation should be expanded so that the 'each exponent appears in at least two distinct pairs' conclusion is fully justified.
- [Section 4, Theorem 4.4] The bound 'at most r q^r' for the size of X_Σ appears too small; a safe bound such as (2|X|)^r would still lead to a polynomial bound for r' after absorbing constants, so the numerical estimate should be corrected or loosened.
- [Throughout] Several lemmas are referred to as theorems (e.g., 'Theorem 2.1', 'Theorem 2.2', 'Theorem 3.2' for Lemma 2.1, Lemma 2.2, and Lemma 3.2); consistent numbering would avoid confusion.
- [Section 3.1, Theorem 3.1] The repeated application of Roth's theorem to obtain the family S of disjoint solutions is correct because Theorem 1.1 is stated for the distinct-elements notion of solution-free, but a sentence making this contrapositive explicit would improve readability.
Circularity Check
No significant circularity: the Ramsey–Turán classification is derived from independent external results, not from its own conclusion.
full rationale
The paper's central claim, Theorem 1.2, classifies density-regular homogeneous linear equations under a Ramsey–Turán independence condition. The forward direction (b)->(a) (Theorem 3.1) uses Roth's theorem (Theorem 1.1) as an external black box to find many disjoint solutions of a zero-sum subequation, and then uses a greedy proper-rainbow-path lemma (Lemma 3.2). Roth's theorem is an independent classical result, not the theorem being proved, and the rainbow-path lemma is proved by induction from Caro-Wei degree arguments. The reverse direction (a)->(b) (Theorem 3.3) constructs an explicit linear-size solution-free set with independence number O(p/log p), relying only on elementary modular arithmetic and a clique bound in the Cayley graph. The quantitative lower bounds in Section 4 use external Ramsey graph lower bounds due to Kim and Osthus–Taraz; those results are not self-citations and do not assume Theorem 1.2. The quantity d(L,ε) is defined independently of the theorem's zero-sum-subset condition, so the conclusion is not baked into the definition. The only notable subtlety is the repeated application of Roth's theorem to extract solutions with distinct, pairwise-disjoint elements; this is a standard supersaturation consequence and is not equivalent to the paper's conclusion. Even if the manuscript could be more explicit about that supersaturation step, this is a completeness or correctness concern, not circularity. No load-bearing self-citation, fitted-input-renamed-as-prediction, or definitional equivalence was found.
Assumptions & free parameters
assumptions (3)
- standard math Roth's theorem (Theorem 1.1): any subset of F_p of size at least εp contains a solution to any zero-sum homogeneous equation with at least 3 variables, for p large.
- standard math Caro-Wei bound (Lemma 2.1): a graph with average degree d has independence number at least n/(d+1).
- standard math Ramsey graph lower bounds (Theorems 4.1 and 4.3 by Kim and by Osthus-Taraz).
Cite this review
Pith. "Pith review of On a Ramsey--Tur\'{a}n variant of Roth's theorem." pith.science (2026). https://pith.science/paper/3MZYXTLF
@misc{pith2026250722831,
author = {Pith},
title = {Pith review of: On a Ramsey--Tur\'an variant of Roth's theorem},
year = {2026},
howpublished = {\url{https://pith.science/paper/3MZYXTLF}},
note = {Machine review of arXiv:2507.22831}
}
abstract
A classical theorem of Roth states that the maximum size of a solution-free set of a homogeneous linear equation $\mathcal{L}$ in $\mathbb{F}_p$ is $o(p)$ if and only if the sum of the coefficients of $\mathcal{L}$ is $0$. In this paper, we prove a Ramsey--Tur\'{a}n variant of Roth's theorem, with respect to a natural notion of ``structured'' sets introduced by Erd\H{o}s and S\'ark\"ozy in the 1970's. Namely, we show that the following statements are equivalent: $(a)$ Every solution-free set $A$ of $\mathcal{L}$ in $\mathbb{F}_p$ with $\alpha(\mathrm{Cay}_{\mathbb{F}_p}(A)) = o(p)$ has size $o(p)$. $(b)$ There exists a non-empty \emph{subset} of coefficients of $\mathcal{L}$ with zero sum.
Reference graph
Works this paper leans on
- [1]
-
[2]
T. F. Bloom and J. Maynard,A new upper bound for sets with no square differences, Compos. Math.158(2022), no. 8, 1777–1798
work page 2022
-
[3]
T. Bohman and P. Keevash,The early evolution of theH-free process, Invent. Math.181 (2010), no. 2, 291–336
work page 2010
- [4]
-
[5]
Caro,New results on the independence number, Tech
Y. Caro,New results on the independence number, Tech. report, Technical Report, Tel-Aviv University, 1979
work page 1979
-
[6]
S. Cho, D. Conlon, J. Lee, J. Skokan, and L. Versteegen,On norming systems of linear equations,arXiv:2411.18389, 2024
arXiv 2024
-
[7]
P. Erd˝ os and A. S´ ark¨ ozy,On differences and sums of integers, ii., Bull. Soc. Math. Gr` ece (NS)18(1977), 204–223
work page 1977
-
[8]
P. Erd˝ os and V. T. S´ os,Some remarks on Ramsey’s and Tur´ an ’s theorem, Combinatorial theory and its applications, I-III (Proc. Colloq., Balatonf¨ ured, 1969), Colloq. Math. Soc. J´ anos Bolyai, vol. 4, North-Holland, Amsterdam-London, 1970, pp. 395–404. 15
work page 1969
Show all 25 references
-
[9]
Erd˝ os and P
P. Erd˝ os and P. Tur´ an,On Some Sequences of Integers, J. London Math. Soc.11(1936), no. 4, 261–264
1936
-
[10]
Fiz Pontiveros, S
G. Fiz Pontiveros, S. Griffiths, and R. Morris,The triangle-free process and the Ramsey numberR(3, k), Mem. Amer. Math. Soc.263(2020), no. 1274, v+125
2020
-
[11]
J. Fox, H. T. Pham, and Y. Zhao,Common and Sidorenko linear equations, Q. J. Math. 72(2021), no. 4, 1223–1234
2021
-
[12]
Kelley and R
Z. Kelley and R. Meka,Strong bounds for 3-progressions, 2023 IEEE 64th Annual Sym- posium on Foundations of Computer Science—FOCS 2023, IEEE Computer Soc., Los Alamitos, CA, [2023]©2023, pp. 933–973
2023
-
[13]
J. H. Kim,The Ramsey numberR(3, t)has order of magnitudet 2/logt, Random Structures Algorithms7(1995), no. 3, 173–207
1995
-
[14]
J. Leng, A. Sah, and M. Sawhney,Improved bounds for Szemer´ edi’s theorem,arXiv:arXiv: 2402.17995, 2024
2024 arXiv
-
[15]
Osthus and A
D. Osthus and A. Taraz,Random maximalH-free graphs, Random Structures Algorithms 18(2001), no. 1, 61–82
2001
-
[16]
K. F. Roth,On certain sets of integers, J. London Math. Soc.28(1953), 104–109
1953
-
[17]
K. F. Roth,On certain sets of integers. II, J. London Math. Soc.29(1954), 20–26
1954
-
[18]
Saad and J
A. Saad and J. Wolf,Ramsey multiplicity of linear patterns in certain finite abelian groups, Q. J. Math.68(2017), no. 1, 125–140
2017
-
[19]
Schur, ¨Uber kongruenz x
I. Schur, ¨Uber kongruenz x ... (mod. p.)., Jahresbericht der Deutschen Mathematiker- Vereinigung25(1917), 114–116
1917
-
[20]
J. B. Shearer,A note on the independence number of triangle-free graphs. II, J. Combin. Theory Ser. B53(1991), no. 2, 300–307
1991
-
[21]
Simonovits and V
M. Simonovits and V. T. S´ os,Ramsey-Tur´ an theory, vol. 229, 2001, Combinatorics, graph theory, algorithms and applications, pp. 293–340
2001
-
[22]
Szemer´ edi,On sets of integers containing nokelements in arithmetic progression, Acta Arith.27(1975), 199–245
E. Szemer´ edi,On sets of integers containing nokelements in arithmetic progression, Acta Arith.27(1975), 199–245
1975
-
[23]
B. L. Van der Waerden,Beweis einer baudetschen vermutung, Nieuw Arch. Wiskunde15 (1927), 212–216
1927
-
[24]
V. K. Wei,A lower bound on the stability number of a simple graph, 1981
1981
-
[25]
Zhao,Graph theory and additive combinatorics—exploring structure and randomness, Cambridge University Press, Cambridge, 2023
Y. Zhao,Graph theory and additive combinatorics—exploring structure and randomness, Cambridge University Press, Cambridge, 2023. 16
2023
Reviewed August 6, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.