Transference for amenable actions

From MaRDI portal





Suppose that \(G\) acts amenably on a measure space \(X\) with quasi-invariant \(\sigma\)-finite measure \(m\). Let \(\sigma\) be an isometric representation of \(G\) on \(L^p(X,dm)\), \(1\leq p<\infty\), and \(\mu\) a finite Radon measure on \(G\). In this paper, the authors prove that the operator \(\sigma_\mu f(x)=\int_G\sigma(g)f(x) \,d\mu(g)\) for \(x\in X\) has \(L^p(X,dm)\)-operator norm not exceeding the \(L^p(G)\)-operator norm of the convolution operator \(\lambda_\mu f(h)=\int_G f(g^{-1}h)d\mu(g)\) for \(h\in G\). The authors also prove an analogous result for the maximal function associated to a countable family of Radon measures \(\{\mu_n\}_{n=1}^\infty\). The basic transference theory goes back to the earliest works of R. R. Coifman and G. Weiss around 1970, which has many different applications and has been generalized to different contexts; the central role is played by an amenable group.











This page was built for publication: Transference for amenable actions

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4654123)