Classical summation in commutative and noncommutative \(L_p\)-spaces (Q537675)

The book under review is a beautiful and original exposition on the topic of almost everywhere convergent orthonormal series. It is accessible to a graduate student, but it is not an easy reading. The student or researcher who succeeds in reading this book will be rewarded with a deep understanding of the subject, both in the commutative and noncommutative setting. The starting point is the Rademacher-Menchov theorem stating that, whenever a sequence \((\alpha_k)\) of coefficients satisfies \(\sum_k\alpha_k|\log k|^2<\infty\), the series \(\sum_k\alpha_kx_k\) converges \(\mu\)-almost everywhere for any orthonormal sequence \((x_k)\) in \(L_2(\mu)\). As proved by Kantorovich, there is also a corresponding and (seemingly) stronger result on the maximal function of the series: \(\sup_j\left|\sum_{k=0}^j\alpha_kx_k\right|\in L_2(\mu)\). By a proper choice of an increasing and unbounded sequence of positive scalars (for example, \((\log\log k)\)) and a proper choice of a summation method (for example, the Cesàro one), we obtain a new theorem on the almost sure convergence of a specific matrix transform of the starting orthonormal sequence and a new theorem on the corresponding maximal function. The author shows that all the classical coefficient tests of this kind lead to some maximal theorem. Moreover, such theorems are applicable not only to orthonormal sequences, but also to the much more general weakly \(2\)-summable ones. The second step is the use of so-called maximizing matrices. From the introduction: ``This allows us to transform basically every classical coefficient test for almost everywhere convergence of orthonormal series \(\sum_k\alpha_kx_k\) in \(L_2(\mu)\) to a theorem on almost everywhere summation for unconditionally convergent series \(\sum_kx_k\) in vector-valued Banach function spaces \(E(\mu,X)\).'' As a consequence, we obtain a series of results on the summation of such series in \(L_p\)-spaces. The last part might be the most demanding one from the reader's point of view, since it investigates noncommutative coefficient tests in the context of von Neumann algebras. The cases considered include the case of the algebra itself, where the limit of the series might fall outside of the algebra. Even more interesting are convergence theorems in Haagerup's \(L^p\) spaces, and, in the case that the algebra is semifinite with a faithful normal semifinite trace \(\tau\) on it, in the noncommutative Banach function spaces \(E(M,\phi)\). The noncommutative results are often motivated by the previous work of Gilles Pisier and Marius Junge, especially Junge's noncommutative Doob's inequality, and by the joint work by the author and \textit{M. Junge} [J. Funct. Anal. 206, No. 2, 322--355 (2004; Zbl 1069.46033)]. To sum up, the book should stand on the shelf of anyone seriously interested in functional analysis and/or probability.