Stein’s vector Riesz transform weak-type problem

Prototype Program for a Dimension-Free Weak-(1,1) Estimate for the Vector Riesz Transform A clearer and more structured research formulation

1. Research Objective Let ν be a finite positive measure on ℝⁿ, and let Rν denote the vector Riesz transform. The principal objective is to establish a weak-(1,1) estimate of the form λ |{x ∈ ℝⁿ : |Rν(x)| > λ}| ≤ C |ν|(ℝⁿ), where the constant C is independent of the dimension n. The central difficulty is not the singularity of the Riesz kernel itself. The main difficulty is controlling the interaction of many vector-valued contributions without introducing a factor depending on the number of atoms. The prototype analysis below isolates this difficulty and identifies a possible mechanism for eliminating unwanted N-dependence.
2. Prototype I — The Two-Atom Problem Consider the positive two-atom measure ν = aδ₀ + bδ_y, a,b ≥ 0, y ≠ 0. By rotational and scaling normalization, we may take y = e₁. Thus ν = aδ₀ + bδ_{e₁}. The vector Riesz transform is Rν(x) = cₙ [ a x/|x|ⁿ⁺¹ + b (x−e₁)/|x−e₁|ⁿ⁺¹ ]. Define A(x) = x/|x|ⁿ⁺¹, B(x) = (x−e₁)/|x−e₁|ⁿ⁺¹. Rν(x) = cₙ(aA(x) + bB(x)). The desired estimate is λ |{x ∈ ℝⁿ : |Rν(x)| > λ}| ≤ C(a+b), with C independent of n. 2.1 Why the two-atom problem is nontrivial For one atom, the field is radial and its level sets are exactly balls. With two atoms, the vectors A(x) and B(x) point in different directions. Consequently, the magnitude of their sum cannot be understood solely from the scalar magnitudes |A(x)| and |B(x)|. |aA+bB|² = a²|A|² + b²|B|² + 2ab A·B. The cross term is the first genuinely geometric interaction term. The two-atom configuration is therefore the smallest model in which directional alignment and cancellation appear.
3. Prototype II — The N-Atom Problem Consider ν = Σₖ₌₁ᴺ aₖδ_{cₖ}, aₖ ≥ 0. Vₖ(x) = aₖcₙ (x−cₖ)/|x−cₖ|ⁿ⁺¹, Rν(x) = Σₖ₌₁ᴺ Vₖ(x). The elementary estimate |Σₖ Vₖ|² ≤ N Σₖ |Vₖ|² is valid but too crude for a dimension-free, N-free argument. It forgets that all vectors are generated by the same observation point x and by the same geometric kernel. A successful argument must preserve this common geometry.
4. Step One — Separate Large Individual Contributions Fix λ > 0 and choose 0 < α < 1. At each x, divide the atoms into L(x) = {k : |Vₖ(x)| > αλ}, S(x) = {k : |Vₖ(x)| ≤ αλ}. For a fixed atom, the condition |Vₖ(x)| > αλ defines a ball centered at cₖ. Its volume has the scale |{|Vₖ| > αλ}| ≲ aₖ/(αλ). λ |⋃ₖ {|Vₖ| > αλ}| ≲ α⁻¹ Σₖ aₖ = α⁻¹ |ν|. Thus the individually large contributions can be controlled without any factor N.
5. Step Two — Identify the Genuine Difficulty The remaining points satisfy |Vₖ(x)| ≤ αλ for every k, while |Rν(x)| > λ. No individual atom is large, so the exceptional event must be produced collectively. The key phenomenon is therefore directional coherence rather than individual magnitude.
6. Step Three — Separate Magnitude from Direction θₖ(x) = (x−cₖ)/|x−cₖ|, wₖ(x) = aₖ|x−cₖ|⁻ⁿ. Rν(x) = cₙ Σₖ wₖ(x) θₖ(x). The unit vectors θₖ(x) are not arbitrary. They all originate from the same observation point x. This geometric dependence is the structural feature that an N-atom argument must exploit.
7. Step Four — The Resultant-Direction Test Whenever Rν(x) ≠ 0, define e(x) = Rν(x)/|Rν(x)|. |Rν(x)| = e(x)·Rν(x) |Rν(x)| = cₙ Σₖ aₖ [e(x)·(x−cₖ)]/|x−cₖ|ⁿ⁺¹. This converts the vector magnitude into an adaptive scalar projection while retaining the signs supplied by the geometry. It suggests that the correct argument should be global or measure-level rather than a crude pointwise square-function estimate.
8. Step Five — The Desired Measure-Level Escape from the N-Loss The proposed strategy is to seek a decomposition ν = μ + (−Δ)¹ᐟ²u, with 0 ≤ μ ≤ λ, |μ|₁ ≤ |ν|₁, |{u > 0}| ≲ |ν|₁/λ. The existence of such a pair (u, μ) is a research lemma and must not be assumed without proof.
9. Step Six — Why Such a Decomposition Would Solve the Problem R = ∇(−Δ)⁻¹ᐟ². Rν = Rμ + ∇u. If the construction gives ∇u = 0 almost everywhere on {u = 0}, then {|Rν| > λ} ⊆ {u > 0} ∪ {|Rμ| > λ}. The exceptional set is therefore split into a contact contribution and a residual Riesz-transform contribution.
10. Step Seven — Control of the Residual Term ‖Rμ‖₂² = ‖μ‖₂². ‖μ‖₂² = ∫ μ² ≤ λ∫μ = λ|μ|₁. λ |{|Rμ| > λ}| ≤ |μ|₁ ≤ |ν|₁. This estimate contains no dependence on N.
11. Step Eight — Control of the Positivity Region |{u > 0}| ≤ C₀ |ν|₁/λ. λ |{u > 0}| ≤ C₀ |ν|₁. λ |{|Rν| > λ}| ≤ (C₀ + 1)|ν|₁. Thus the number of atoms disappears completely. If C₀ = 1 under the exact normalization and obstacle estimate, the resulting constant would be 2.
12. The Central Research Lemma The entire strategy reduces to the following statement: For every finite positive atomic measure ν and every λ > 0, construct an admissible pair (u, μ) satisfying ν = μ + (−Δ)¹ᐟ²u, 0 ≤ μ ≤ λ, |μ|₁ ≤ |ν|₁, |{u > 0}| ≤ C₀ |ν|₁/λ, together with the precise localization property for ∇u required to obtain the exceptional-set inclusion.
13. Prototype III — The One-Atom Experiment The cleanest rigorous test is ν = aδ₀. R(aδ₀)(x) = acₙ x/|x|ⁿ⁺¹. |R(aδ₀)(x)| = a|cₙ||x|⁻ⁿ. Therefore the level set {|R(aδ₀)| > λ} is exactly a Euclidean ball centered at the origin, and its volume has the scale |{|R(aδ₀)| > λ}| ≍ a/λ, with the precise dimensional factor determined by the normalization of cₙ. This verifies the correct endpoint scale directly.
14. What the One-Atom Experiment Must Actually Prove The explicit one-atom Riesz field does not itself establish the desired obstacle decomposition. A genuine construction must verify all of the following: • Admissibility: u belongs to the required fractional energy or distributional class. • Distributional identity: aδ₀ = μ + (−Δ)¹ᐟ²u. • Threshold constraint: 0 ≤ μ ≤ λ. • Mass control: |μ|₁ ≤ a. • Contact-set estimate: |{u > 0}| ≤ C a/λ. • Gradient localization: ∇u = 0 almost everywhere outside the appropriate contact/positivity region, in the precise sense required by the construction. Only after these properties have been rigorously established may the decomposition be inserted into the weak-(1,1) argument.
15. Why Guessing the Obstacle Function Is Dangerous A radial formula suggested by the fundamental solution of (−Δ)¹ᐟ² may have the correct scaling but still fail one or more essential requirements. In particular, the distributional identity, positivity of μ, the bound μ ≤ λ, mass control, contact-set size, or gradient localization may fail. Construct u → verify the distributional identity → verify every obstacle estimate. This is the correct order of proof.
16. Prototype IV — Returning to Two Atoms ν = a₁δ_{c₁} + a₂δ_{c₂}. After the one-atom case is understood, the next question is whether the obstacle construction is stable under positive superposition. One must not simply assume u = u₁ + u₂, because fractional obstacle problems are nonlinear. The two-atom experiment therefore tests whether the desired mass and contact-set estimates come from a genuinely nonlinear envelope principle rather than from pairwise treatment of the atoms.
17. The Role of the Two-Atom Geometry Rν(x) = cₙ [a x/|x|ⁿ⁺¹ + b(x−e₁)/|x−e₁|ⁿ⁺¹]. |Rν(x)|² = cₙ² [a²|x|⁻²ⁿ + b²|x−e₁|⁻²ⁿ + 2ab x·(x−e₁)/( |x|ⁿ⁺¹ |x−e₁|ⁿ⁺¹ )]. The cross term contains the essential directional information and may be positive or negative depending on x. This shows why a purely magnitude-based estimate loses important geometry.
18. Main Methodological Conclusion The objective should not be to prove a pointwise inequality |Σₖ Vₖ|² ≲ Σₖ |Vₖ|² with a constant independent of N. Coherent alignment prevents such an approach from capturing the true structure. Instead, the preferred architecture is atomic field → fractional potential + bounded residual. The potential absorbs the collective geometric excess, while the residual is handled by L² theory. The final estimate depends on |μ|₁ rather than on the number of atoms.
19. Atomic Reduction and Its Logical Role Finite positive atomic measures isolate the essential geometric difficulty. The relevant reduction principle is that controlling this class is the appropriate testing problem for extending the weak-type estimate to more general L¹ data. However, atomic reduction and obstacle decomposition are logically distinct. The reduction identifies the test class; it does not establish the existence of the decomposition ν = μ + (−Δ)¹ᐟ²u.
20. Final Research Target The next major theorem to establish is: For every finite positive atomic measure ν and λ > 0, construct an admissible pair (u, μ) satisfying ν = μ + (−Δ)¹ᐟ²u, 0 ≤ μ ≤ λ, |μ|₁ ≤ |ν|₁, |{u > 0}| ≲ |ν|₁/λ, together with the precise gradient-localization property needed for {|Rν| > λ} ⊆ {u > 0} ∪ {|Rμ| > λ}. λ|{u > 0}| ≲ |ν|₁, λ|{|Rμ| > λ}| ≤ |μ|₁ ≤ |ν|₁. Therefore: λ|{|Rν| > λ}| ≲ |ν|₁. The decisive achievement is the elimination of dependence on both N and n from the final constant.
21. Immediate Next Experiment The next rigorous step should be the one-atom problem ν = aδ₀. Explicitly construct or identify the appropriate fractional obstacle problem and verify, one by one: • u is admissible. • aδ₀ = μ + (−Δ)¹ᐟ²u. • 0 ≤ μ ≤ λ. • |μ|₁ ≤ a. • |{u > 0}| ≲ a/λ. • ∇u = 0 almost everywhere outside the appropriate contact region. Only after the one-atom construction has been completely verified should the analysis proceed to two atoms and then to the general N-atom configuration.
22. Core Research Principle The route to an N-free estimate is not to force N vectors to behave like orthogonal vectors; it is to replace their collective interaction by a single fractional potential and a uniformly bounded residual whose L² energy is controlled by its mass. One atom → Two atoms → N atoms → General positive L¹ data. At every stage, the critical issue is to obtain the exceptional-set estimate without introducing either the number of atoms N or the ambient dimension n into the final constant.

Prototype Research Report: Toward a Dimension-Free Weak-(1,1) Bound for the Vector Riesz Transform Refined Edition — Atomic Measures, Cancellation Geometry, the N-Atom Problem, and the Measure-Level L² Mechanism 19 August 2026 Abstract This refined report consolidates the prototype mathematical development concerning Stein's dimension-free weak-(1,1) problem for the vector Riesz transform. The central strategy is to study finite positive atomic measures, understand the exact cancellation geometry of the Riesz kernel, identify why naive N-atom summation fails, and replace the N-dependent pointwise estimate by a measure-level decomposition. The report distinguishes explicit calculations from conjectural or theorem-dependent steps and develops a candidate route based on a bounded residual μ and a fractional-Laplacian potential u. This mechanism explains how a constant of the form 2 can arise without any dependence on dimension or on the number of atoms. The document is a research prototype and should not be presented as an independently completed proof until every decomposition and approximation step has been rigorously verified.

1. Stein's dimension-free weak-(1,1) problem Let Rf=(R_1f,...,R_nf) be the vector Riesz transform on R^n, with Fourier multiplier −iξ_j/|ξ| in the j-th component. The endpoint question is whether there exists a universal constant C<∞, independent of n, such that for every n≥1, f∈L¹(R^n), and λ>0, λ |{x∈R^n : |Rf(x)|>λ}| ≤ C ||f||₁. The difficulty is specifically at the endpoint p=1. Dimension-free Lᵖ estimates are known for 1<p<∞, but the weak-(1,1) estimate requires a mechanism that preserves cancellation and does not introduce a dimension-dependent or atom-count-dependent loss.
2. Atomic reduction as the prototype framework A central motivation is the reduction associated with the 2020 Spector–Stockdale work, which connects the endpoint problem to finite positive combinations of Dirac masses. We therefore study ν = Σ_{k=1}^N a_k δ_{c_k}, a_k≥0. The desired atomic estimate has the form λ |{x : |Rν(x)|>λ}| ≤ C Σ_k a_k, with C independent of n and N. Dirac masses are measures rather than L¹ functions, so a rigorous passage back to arbitrary L¹ data requires the exact reduction, regularization, or approximation argument. The calculations below therefore constitute a prototype framework.
3. One-atom calculation For ν=aδ₀, the Riesz kernel is K_n(x)=c_n x/|x|^{n+1}. Hence |Rν(x)| = a|c_n||x|^{-n}. The level set |Rν|>λ is a ball of radius (a|c_n|/λ)^{1/n}, so λ |{x:|Rν(x)|>λ}| = ω_n |c_n| a. Under the standard normalization, ω_n|c_n| is an absolute constant. Thus the one-atom model has the required dimension-free scaling.
4. Two atoms: exact cancellation geometry Take ν=aδ₀+bδ_{e₁}. Define A(x)=x/|x|^{n+1}, B(x)=(x−e₁)/|x−e₁|^{n+1}. Then Rν=c_n(aA+bB). The crucial scalar-product identity is A·B = [|x|²+|x−e₁|²−1]/[2|x|^{n+1}|x−e₁|^{n+1}]. Therefore A·B<0 precisely in the ball centered at e₁/2 of radius 1/2. This identifies the cancellation region geometrically: the two source directions form an obtuse angle at x.
5. General pairwise interaction identity For V_k(x)=a_k c_n(x−c_k)/|x−c_k|^{n+1}, Rν(x)=Σ_k V_k(x), and |Rν|²=Σ_k|V_k|²+2Σ_{k<ℓ}V_k·V_ℓ. Moreover, V_k·V_ℓ = (a_k a_ℓ c_n²/2) [|x−c_k|²+|x−c_ℓ|²−|c_k−c_ℓ|²] / [|x−c_k|^{n+1}|x−c_ℓ|^{n+1}]. The sign is controlled by Euclidean geometry rather than by an explicit factor of n. This is one of the most interesting structural observations of the prototype.
6. Three atoms and the Gram-matrix viewpoint For three atoms, pairwise interactions can have mixed signs. The natural object is the Gram matrix G(x)=(V_i·V_j). It is positive semidefinite and satisfies |Σ_iV_i|²=1ᵀG1. This makes clear why estimating pairwise terms by absolute values can destroy precisely the cancellation that the endpoint theorem must exploit.
7. Why the naive N-atom argument fails The elementary inequality |Σ_{k=1}^N V_k|² ≤ N Σ_{k=1}^N |V_k|² treats the Riesz vectors as arbitrary vectors. It therefore introduces a factor N. The corresponding threshold argument reduces λ to approximately λ/N and loses N in the weak estimate. The triangle inequality is no better because it discards the signed pairwise terms. Thus the failure is a failure of the method, not evidence that Stein's desired estimate must contain N.
8. The correct question: eliminate N through geometry or decomposition We should not attempt to prove the pointwise inequality |Σ_kV_k|² ≤ C Σ_k|V_k|² with C independent of N. Such a statement is much stronger than the endpoint theorem and is not the natural target. Instead we seek an exceptional-set estimate λ |{|Rν|>λ}| ≤ C ||ν||₁, where collective cancellation is retained globally.
9. Large and small contributions For a fixed x and threshold λ, choose 0<α<1 and divide the atoms into L(x)={k: |V_k(x)|>αλ}, S(x)={k: |V_k(x)|≤αλ}. The large contributions can be charged to one-atom exceptional balls. Summing their measures costs only a constant multiple of Σ_k a_k/λ. The unresolved regime is where every individual contribution is small but the vector sum is large. In that regime, the exceptional event must arise from collective directional alignment. This is precisely the information lost by the naive N-vector inequality.
10. Resultant-direction formulation At points where Rν(x)≠0, define e(x)=Rν(x)/|Rν(x)|. Then |Rν(x)| = e(x)·Rν(x) = c_n Σ_k a_k [e(x)·(x−c_k)]/|x−c_k|^{n+1}. This adaptive scalar projection retains signs and may provide a route to a global geometric estimate. It is a candidate tool, not yet a proved endpoint lemma.
11. The central N-atom prototype lemma A purely atomic route would seek an absolute C₀ such that λ |{x:|Rν(x)|>λ}| ≤ C₀ Σ_k a_k, for every finite positive atomic measure, with C₀ independent of both n and N. The preceding sections identify why this cannot be obtained by naive vector summation. The required proof must use the special geometry of the kernel or an equivalent global decomposition.
12. Measure-level escape from the N-loss A promising mechanism is a fractional-Laplacian obstacle decomposition. Schematically, seek ν = μ + (−Δ)^{1/2}u, with 0≤μ≤λ, ||μ||₁≤||ν||₁, and a contact/positivity region satisfying |{u>0}| ≤ ||ν||₁/λ. Since R=∇(−Δ)^{-1/2}, Rν=Rμ+∇u. If ∇u=0 almost everywhere outside {u>0}, then {|Rν|>λ} ⊂ {u>0} ∪ {|Rμ|>λ}. For the residual term, the L² identity gives ||Rμ||₂²=||μ||₂²≤λ||μ||₁. Hence Chebyshev gives λ |{|Rμ|>λ}| ≤ ||μ||₁ ≤ ||ν||₁. Combining the two pieces yields the prototype endpoint estimate λ |{|Rν|>λ}| ≤ 2||ν||₁. This is the conceptual 1+1=2 mechanism. The difficult theorem is the rigorous construction of the decomposition and verification of the contact-set, bounded-residual, and approximation properties.
13. One-atom obstacle problem: next rigorous experiment The next test should begin with ν=aδ₀ and explicitly construct the candidate fractional obstacle solution. We should verify, rather than assume, the following properties: • ν=μ+(−Δ)^{1/2}u in the appropriate distributional sense. • 0≤μ≤λ. • ||μ||₁≤||ν||₁. • |{u>0}|≤||ν||₁/λ. • ∇u=0 almost everywhere outside the positivity/contact region in the required sense. • The resulting Riesz level-set estimate reproduces the one-atom bound. The one-atom calculation already verifies the correct scaling, but the obstacle decomposition must be derived independently if it is to serve as the foundation of a new proof.
14. Two-atom obstacle test After the one-atom construction, take ν=aδ₀+bδ_{e₁}. The test is whether the obstacle decomposition can absorb the interaction region where the two vectors cancel or reinforce. The two-atom geometry derived earlier provides an exact laboratory for checking every proposed inequality. In particular, a successful construction must remain stable as the atoms approach each other, as they separate, and as one mass becomes much larger than the other.
15. What the prototype has established • The one-atom weak endpoint calculation is dimension-free. • The two-atom interaction has an exact geometric cancellation region. • The general pairwise interaction formula contains no explicit dimension factor in its geometric numerator. • The naive N-atom vector inequality creates an artificial N-loss. • The correct target is an exceptional-set estimate, not a stronger pointwise square-function inequality. • Large individual contributions can be charged to one-atom estimates; collective small contributions require a new mechanism. • The resultant-direction formulation exposes the signed geometry of the collective field. • A bounded residual plus fractional-potential decomposition offers a principled way to make N disappear. • The L² estimate of the residual produces one unit of the endpoint constant, while the contact region produces the second unit.
16. Research-status statement This document is a refined prototype research report. The explicit kernel identities, one-atom calculation, and elementary N-loss analysis are mathematical components of the prototype. The obstacle-based route is a proposed proof architecture unless every cited theorem and construction is independently verified. In particular, the recent 2026 manuscript claiming a dimension-free weak-(1,1) estimate should be treated as a claimed result pending independent verification. A future article should clearly distinguish proved lemmas, standard external theorems, calculations performed here, and conjectural steps.
17. Conclusion The investigation has moved beyond the original N-atom obstruction. The key insight is that the factor N arises because arbitrary-vector inequalities ignore the special geometry of the Riesz kernel. A dimension-free proof need not make those vectors pointwise orthogonal or uniformly summable. Instead, it can seek a global decomposition that separates the field into a bounded residual, controlled by L², and a fractional-potential part, controlled through its contact set. The proposed next stage is therefore concrete: solve and verify the one-atom fractional obstacle problem, test the construction on two atoms using the exact cancellation identity, and only then attempt the general N-atom theorem. This creates a disciplined path from the prototype calculations toward a possible rigorous proof. References / starting points
18. Y. Ouyang, D. Spector, C. B. Stockdale, “A dimension-free weak-type (1,1) bound for the vector Riesz transform on R^n,” arXiv:2608.18068, submitted August 18, 2026.
19. D. Spector and C. B. Stockdale, arXiv:2004.03382 (2020), concerning the reduction of Stein's question to finite combinations of Dirac masses.
20. E. M. Stein, foundational work on singular integrals and dimension-free Lp estimates for Riesz transforms.