Necessity of the Carleson condition for Hölder density transformations

Determine whether the Carleson condition for the transformed square function \(\Delta^F_\mu\) can fail for a uniformly \(n\)-rectifiable measure \(\mu\) and a Hölder-continuous function \(F\) that is not Lipschitz on \(J_0=[c_0^{-1},c_0]\).

Background

For an nn-AD-regular measure with infinite total mass, the paper proves that a uniformly nn-rectifiable measure satisfies an L2/αL^{2/\alpha} estimate for ΔμF\Delta^F_\mu when FF is strictly increasing and α\alpha-Hölder on the density interval J0J_0. This does not yield the desired exponent-2 Carleson estimate when α<1\alpha<1, because the known Carleson estimate for Δμ\Delta_\mu does not control the corresponding lower power of Δμ|\Delta_\mu|. The authors explicitly leave unresolved whether the exponent-2 condition can fail in this setting.

References

We do not know whether eq:carlesonF can actually fail for a uniformly n-rectifiable \mu and a H\"older F which is not Lipschitz on J_0.

eq:carlesonF:

0R ⁣ ⁣xB(x0,R)ΔμF(x,r)2dμ(x)drr    cRn.\int_0^R\!\!\int_{x\in B(x_0,R)} |\Delta^F_\mu(x,r)|^2\, d\mu(x)\,\frac{dr}{r} \;\le\; c\,R^n.

Square Functions and Rectifiability under Monotone Transformations of the Density  (2608.30166 - Le, 31 Aug 2026) in Section 4, immediately after Proposition 4.1 (the proposition labeled `prop:degenerate-easy` in the source)