scientific article; zbMATH DE number 193810

Publication date 5 June 1993

This reissue of Wiener's classic from 1933 contains a foreword by Jean- Pierre Kahane, beginning with ``This is a most inspired and inspiring book and ending with ``... this little book is one of the great mathematical achievements of the century. Since the first printing only got an 18-lines review (see Zbl 0006.05401), it seems justified to give a more detailed description of the contents of Wiener's book, which is ``in substance an elaboration of a course of fifteen lectures... given at the University of Cabridge... 1932.... My impulse to write a book of this type arose from a dissatisfaction with the preponderand rôle of convergence theory in existing textbooks on the subject and from the need for a treatment more in line with the extensive periodical literature. (from Wiener's preface). In the introduction the author gives a short description of Lebesgue's integral and of its properties (convergence theorems, Fubini's theorem, Hölder's and Minkowski's inequality, Riesz-Fischer theorem, etc.), of the (Riemann-)Stieltjes integral, and he deals with orthogonality. In the first chapter Hermite polynomials and Hermite functions are studied; the latter ones constitute a complete orthonormal system in \(L^ 2\) and they are used to obtain the existence of the Fourier transform for functions in \(L^ 2\) (Plancherel's theorem). Chapter two is concerned with a proof of Wiener's general Tauberian theorem: If \(K_ 1\), \(K_ 2\) are in \(L^ 1\), \(f\in L^ 1\) is bounded and if \[ \lim_{x\to \infty}\int^{\infty}_{-\infty}K_ 1(x-y)\cdot f(y)dy=A\cdot \int^{\infty}_{-\infty}K_ 1(x)dx, \] then \[ \lim_{x\to \infty}\int^{\infty}_{-\infty}K_ 2(x-y)\cdot f(y)dy=A\cdot \int^{\infty}_{-\infty}K_ 2(x)dx, \] provided that the Fourier transform \(K_ 1{\hat{\;}}(t)\) does not vanish for any real t. Under the same condition, linear combinations of translates of \(K_ 1\) are dense in \(L^ 1\). Important for the proof is the fact, proved in {\S} 6, that the Fourier series of 1/f is absolutely convergent, if \(f\neq 0\) and if f has an absolutely convergent Fourier series. In Chapter IV the author deduces from his general Tauberian theorem - the Abel-Tauber theorem for power series \(\sum a_ n\cdot x^ n\), satisfying \(a_ n=o(1/n).\) - The corresonding Tauberian theorem under the weaker one-sided Tauberian condition \(n\cdot a_ n>-K.\) - The prime number theorem, using the kernel \[ K_ 1(x)=e^{-x}\cdot (e^{-2e^{-x}}-e^{-e^{-x}}+e^{-x}\cdot e^{-e^{-x}})\cdot (e^{-e^{-x}}-1)^{-2}; \] \(K_ 1{\hat{\;}}(t)\neq 0\) is true, since the Riemann zeta-function does not vanish on Re s\(=1.\) - The Landau-Ikehara Tauberian theorem (and this is applied to prove the prime number theorem again). In the last chapter the author is concerned with the notion of ``spectrum for measurable functions f: \({\mathbb{R}}\to {\mathbb{C}}\), such that \[ \phi (x)=\lim_{T\to \infty}(1/2T)\int^{T}_{-T}f(x+\xi)\cdot \bar f(\xi) d\xi \] exists for any \(x\in R\). The ``spectrum is \[ \sigma (u)=\frac{1}{\sqrt{2\pi}}\{\lim_{A\to \infty}(\int^{A}_{1}+\int^{-1}_{-A})\frac{\phi (t)e^{-iut}}{- it}dt+\int^{1}_{-1}\phi (t)\frac{e^{-iut}-1}{-it}\}. \] Then the ``spectrum is monotonically increasing, and almost everywhere \[ \phi (x)=\frac{1}{\sqrt{2\pi}}\int^{\infty}_{-\infty}e^{iux} d\sigma (\omega), \] and \(\phi\) differs from a continuous function at most on a set of measure zero. Finally, elementary properties of almost-periodic functions are developed, including Parseval's theorem. Unfortunately, a subject index is missing, We conclude with Kahane's words: ``Wiener's book has a special flavour.... It is highly original, with emphasis on a selection of important topics, with unifying views and methods, large perspective and careful attention to details.

Recommendations

Cited in

(31)

This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4040607)