A two weight inequality for the Hilbert transform assuming an energy hypothesis (Q442181)

Let \(\omega\) and \(\sigma\) be two weight functions. For an interval \(I\), set \[ E(I,\omega):=\left\{\mathbb{E}_I^{\omega(dx)}\left[\mathbb{E}_I^{\omega(dy)} \frac{x-y}{|I|}\right]^2\right\}^{1/2}, \] where \[ \mathbb{E}_I^{\omega(dx)}f:=\frac1{\omega(I)}\int_If(x)\omega(dx) \] with \(\omega(I):=\int_I\omega(dx)\) and \(P(I,\omega):=\int_\mathbb{R}\frac{|I|}{(|I|+\mathrm{dist}(x,I))^2}\omega(dx)\). For all intervals \(I\subset\mathbb{R}\) and measurable sets \(J\subset \mathbb{R}\), define the functional \(\Phi(I,J):=\omega(I)E(I,\omega)^2P(I,\chi_J\sigma)^2\), and for fixed \(\gamma,\;\varepsilon\in(0,\infty)\), define the functional \(\Psi_{\gamma,\varepsilon}\) by \[ \Psi_{\gamma,\varepsilon}(I,J):=\sup_{I\supset\bigcup_{s\geq1}I_s} \left[\inf_{s\geq1}\left(\frac{|I|}{|I_s|}\right)^\gamma\right] \sum_{s\geq1}\Phi(I_s,J), \] where the supremum is taken over all \(\varepsilon\)-good subpartitions \(\{I_s\}_{s\geq1}\) of \(I\), which means that \(\{I_s\}_{s\geq1}\) is a subpartition of \(I\) satisfying that \(\mathop{\mathrm{dist}}(I_s,e(I))>\frac12|I_s|^\varepsilon|I|^{1-\varepsilon}\), where \(e([a,b]):=\{a,b,\frac{a+b}2\}\). Weights \(\omega\) and \(\sigma\) are said to satisfy the \(A_2\) condition if \[ \left[\int_I\left(\frac{|I|}{|I|+|x-x_I|}\right)^2\,\omega(dx) \int_I\left(\frac{|I|}{|I|+|x-x_I|}\right)^2\,\sigma(dx)\right]^{1/2} \leq C|I|, \] where \(x_I\) is the center of the interval \(I\). For a signed measure \(\omega\) on \(\mathbb{R}\), define the Hilbert transform \(H\omega\) of \(\omega\) by, for all \(x\in\mathbb{R}\), \(H\omega(x):=\text{p.\,v.}\int_\mathbb{R}\frac 1{x-y}\,\omega(dy)\) with usual principal value integral. Weights \(\omega\) and \(\sigma\) are said to satisfy the two testing conditions if for all intervals \(I\) in \(\mathbb{R}\), \(\int_I [H(\chi_I\sigma)(x)]^2\,d\omega(x)\leq C\int_I\,d\sigma(x)\) and \(\int_I [H(\chi_I\omega)(x)]^2\,d\sigma(x)\leq C\int_I\,d\omega(x)\). In this paper, the authors provethe following: Let \(\omega\) and \(\sigma\) be locally finite positive Borel measures on the real line having no point masses in common, namely, \(\omega(\{x\})\sigma(\{x\})=0\) for all \(x\in\mathbb{R}\). Assume that there exist \(\gamma\in(0,\infty)\) and \(\varepsilon\in(0,1)\) such that \(\omega\) and \(\sigma\) satisfy the following Energy Hypothesis, \[ \sum_{s\geq1}\Psi_{\gamma,\varepsilon}(I_s,I)\leq[\mathcal{F}_{\gamma,\varepsilon}]^2 \sigma(I)\quad \mathrm{and}\quad \sum_{s\geq1}\Psi_{\gamma,\varepsilon}(I_s,I) \leq[\mathcal{F}_{\gamma,\varepsilon}^{\ast}]^2 \omega(I) \] with \(\mathcal{F}_{\gamma,\varepsilon}<\infty\) and \(\mathcal{F}_{\gamma,\varepsilon}^\ast<\infty\), for all \(\varepsilon\)-good subpartitions \(\{I_s\}_{s\geq 1}\) of \(I\). Then the two weight inequality \[ \|H(\sigma f)\|_{L^2(\omega)}\leq C\|f\|_{L^2(\sigma)} \] holds if and only if the pair \((\omega,\sigma)\) of weights satisfies the \(A_2\) condition and the testing conditions. Moreover, the authors exhibit a pair of weights which satisfy the two weight inequality, yet they do not satisfy the Pivotal Condition of Nazarov-Treil-Volberg, which implies that the Pivotal Condition is not necessary for the two weight inequality.