New results on restriction of Fourier multipliers (Q971916)

The authors develop an extension of the transference method due to Coifman and Weiss which applies not only to \(L^p\) spaces but to general quasi-Banach function spaces. They use it to get de Leeuw's type restriction theorems and maximal restriction theorems for certain classes of quasi-Banach function spaces, namely they show that if \((B,C)\) is a couple of rearrangement invariant spaces defined in \(\mathbb R\) such that \[ \liminf h_C(1/N)h_B(N) <\infty \] where \[ h_X(s)=\sup _{f\in X\setminus \{0\}} \frac{\|D_{1/s}f^*\|}{\|f^*\|} \] and \(m\) is a continuous and bounded Fourier multiplier for the couple \((B,C)\) then \(m|_{\mathbb Z}\) defines a Fourier multiplier for the couple \((B_\mathbb T, C_\mathbb T)\) where \(B_\mathbb T\) and \(C_\mathbb T\) consist in functions \(f\) on the torus \(\mathbb T\) with \(f^*\) in the corresponding r.i spaces associated to \(B\) and \(C\) respectively. Also restriction to lower dimensions are given in the same fashion and results on restriction of maximal multipliers are also provided. A number of classical spaces such as Lorentz spaces, Orlicz spaces and spaces coming from interpolation fall under the scope of their theorem. The main tool is an abstract formulation of Coifman-Weiss transference theorem using Wiener amalgams. Namely, if \(F,E\) are function spaces defined on a measure space \(\mathcal M\), \(B\) is a function space defined on a locally compact, \(\sigma\)-compact group \(G\), \(R\) is representation of \(G\) on the operators acting on \(F\) such that for each a non-empty open set \(V\) of \(G\) the map \(x\to \|\chi_V(v)R_vf(x)\|_B\) is measurable for any \(f\in F\), the authors introduce the Wiener amalgam (associated to \(R\)) as the set of functions \(f\in F\) such that \(\big \|\|\chi_V(v)R_vf(x)\|_B\big \|_E<\infty\). They observe that the transference technique allows them to obtain boundedness of the transfer operator \(T_Kf(x)=\int_G K(u)R_{u^{-1}}f(x)d\lambda(u)\) on Wiener amalgams. For positively-preserving representations, they also manage to show the boundedness of its maximal counterpart. Then identifying the amalgams appearing in several cases they get applications to the boundedness of some transferred operators in a lot of classical function spaces.