Transference of \(L_p\) bounds between symmetrized Jacobi expansions and Dunkl transform (Q2226638)

Let \(M_p({\mathbb{R}})\) and \(M_p({\mathbb{Z}})\) denote the sets of all Fourier \(L^p\)-multipliers on \(\mathbb{R}\) and \(\mathbb{Z}\), respectively, where \(\mathbb{Z}\) is discrete subgroup of \(\mathbb{R}\). The well-known theorem of \textit{S. Igari} [Tohoku Math. J. (2) 24, 201--206 (1972; Zbl 0239.42004)] says that if \(m(\varepsilon n) \in M_p({\mathbb{Z}}), 1 < p < \infty\) for all \(\varepsilon > 0\), then \(m \in M_p({\mathbb{R}})\) and \[\Vert m\Vert_{M_p(\mathbb{R})} \leq \liminf\limits_{\varepsilon\to 0^+} \Vert m(\varepsilon n)\Vert_{M_p(\mathbb{Z})}.\] The paper under review contains the analogous result showing connections between multipliers for the symmetrized Jacobi expansions and multipliers for the Dunkl transform on \(\mathbb{R}\). A similar result concerning multipliers of the weak type \((1, 1)\) is also proved. \par The second result of this paper is the transplantation theorem. The symmetric Jacobi functions \(\psi_n^{(\alpha,\beta)}\) form a complete orthonormal system in \(L^2((-\pi, \pi), \mu_{\alpha,\beta})\) with corresponding measure \(\mu_{\alpha,\beta}\). Define the Jacobi transplantation operator by \[\mathcal{T}_{\alpha,\beta}^{\gamma,\delta}(f)=\sum\limits_{n \in {\mathbb{Z}}} \langle f,\psi_n^{(\gamma, \delta)} \rangle \phi_n^{(\alpha, \beta)}.\] The following theorem is proved. \par Let \(1 < p < \infty\), \(\alpha, \beta, \gamma, \delta > 1\), and \(a, b \in {\mathbb{R}}\). If the Jacobi transplantation inequality \[ \left\Vert \mathcal{T}_{\alpha,\beta}^{\gamma,\delta}(f) \right\Vert_{L^p((-\pi,\pi),d\mu_{\alpha,\beta})} \leq C \Vert f\Vert_{L^p((-\pi,\pi), d\mu_{\alpha,\beta})}, \quad f\in C^\infty_c ((-\pi, \pi) \backslash \{0\}), \] holds, then the Dunkl transplantation inequality \[\Vert (\hat{\mathcal{D}}_\alpha \circ \mathcal{D}_\gamma)f \Vert_{L^p({\mathbb{R}}, \vert x\vert ^{2a+1}dx)} \leq C\Vert f\Vert_{L^p({\mathbb{R}}, \vert x\vert ^{2a+1}dx)}, \quad f\in C^\infty_c ({\mathbb{R}} \backslash \{0\}) \] is also satisfied (with the same constant \(C\)), where \(\mathcal{D}_\alpha\) is the modified Dunkl transform and \(\hat{\mathcal{D}}_\alpha f(\lambda) = \mathcal{D}_\alpha f(-\lambda) \).