Transference of \(H^p\)-multipliers and their maximal functions on product domains (Q1763791)

The author establishes analogues of \textit{K. de Leeuw}'s transference theorem [Ann. Math. (2) 81, 364--379 (1965; Zbl 0171.11803)] and maximal versions corresponding to results of \textit{C. E. Kenig} and \textit{P. A. Tomas} [Stud. Math. 68, 79--83 (1980; Zbl 0442.42013)] for multipliers of multiparameter Hardy \(H^p(\mathcal R^n)\) spaces. The latter are defined by finiteness of \(\sup_{t\in \mathbb{R}^{n}_{+} }\| \varphi_t \ast f\| _{L^p}\) for \(f\) a tempered distribution defined on \(\mathbb{R}^n\) (or on the \(n\)-torus \(\mathbb{T}^n\) in the periodic case to define the periodic analogue \(H^p(\mathcal{D}^n)\)). Here \(\varphi(x)=\psi(x_1)\cdots \psi(x_n)\) with \(\int \psi =1\) and \(\hat\psi\) supported in \([-1,1]\). The dilate \(\phi_t\) is then \(\prod \psi(t_i x_i)/t_i\). One denotes by \(M_p(\mathcal{R}^n)\) the space of multipliers of \(H^p(\mathcal{R}^n)\), namely, \(m\in M_p(\mathcal{R}^n)\) if the mapping \(f\mapsto \int_{\mathbb{R}^n} e^{2\pi i x\cdot\xi} \hat f(\xi) m(\xi)\, d\xi\) extends continuously to \(H^p(\mathcal{R}^n)\); one defines \(M_p(\mathcal{D}^n)\) similarly as the space of multiplier sequences on \(H^p(\mathcal{D}^n)\). The basic transference theorem says that if \(m\in M_p(\mathcal{R}^n)\) then its multi-integer sequence \(\{m(k)\}_{k\in \mathbb{Z}^n}\) belongs to \( M_p(\mathcal{R}^n)\). The author goes on to prove maximal analogues: one says that \(m\) is \(p\)-maximal on \(\mathcal{R}^n\) if \(T^\ast: f\mapsto \sup_{t\in \mathbb{R}^{n}_{+}}| T_t f| \) extends continuously from \(H^p(\mathcal{R}^n)\) to \(L^p(\mathbb{R}^n)\). Here \(\widehat{T_t f}(\xi)=m(t_1\xi_1,\dots,t_n\xi_n) \hat f(\xi)\). Similarly, one says that \(m\) is \(p\)-maximal on \(\mathcal{D}^n\) if a corresponding estimate applies to the dilate sequence \(\{m(t_1 k_1),\dots,m(t_n k_n)\}\). The author proves, under a mild assumption of continuity and convergence at infinity, that if \(m\) is \(p\)-maximal on \(\mathcal{R}^n\) then \(m\) is \(p\)-maximal on \(\mathcal{D}^n\). He also shows that if a suitable family of dilates of \(m\) is uniformly bounded on \( H^p(\mathcal{D}^n)\) then \(p\)-maximality of \(m\) on \(\mathcal{R}^n\) and \(\mathcal{D}^n\) are equivalent.