Wiener Tauberian theorem

Wiener Tauberian theorem
Name	Wiener Tauberian theorem
Field	Harmonic analysis; Functional analysis; Fourier transform
Introduced	1932
Author	Norbert Wiener

Wiener Tauberian theorem

The Wiener Tauberian theorem is a foundational result in harmonic analysis and functional analysis relating the nonvanishing of the Fourier transform of an integrable function to the density of its translates in an L^1 or C_0 setting. It links ideas from Fourier series, convolution, and Banach algebra theory and has influenced work in ergodic theory, signal processing, and spectral theory.

Statement of the theorem

The classical statement (for the real line or the circle) asserts that if f is an integrable function on the real line whose Fourier transform has no zeros, then the set of all finite linear combinations of translates of f is dense in L^1(R) (or C_0(R) in the compactly supported/continuous variant). This connects nonvanishing of the transform with the generation of the whole Banach algebra by f under convolution; equivalently, f is a generator of the convolution algebra and its ideal equals the whole algebra. Variants replace L^1(R) by L^p(R) for certain p or by algebras of measures on the circle group or locally compact abelian groups.

Historical background and motivation

The theorem was proved by Norbert Wiener in 1932, building on antecedents in the work of G. H. Hardy and J. E. Littlewood on Tauberian theorems and on earlier studies of Fourier series by Bernhard Riemann and Henri Lebesgue. Wiener formulated the result in response to questions arising in signal processing and the analysis of harmonic synthesis, drawing on the algebraic framework later formalized by Israel Gelfand and Marshall Stone. Subsequent developments were influenced by contributions from Arne Beurling, Salomon Bochner, Stefan Banach, and Raphaël Salem, as the interplay between analytic continuation, zeros of transforms, and density problems became central in modern analysis.

Proofs and key techniques

Proofs exploit the structure of commutative Banach algebras, the invertibility criteria given by the Gelfand transform, and properties of the Fourier transform. Wiener's original argument used complex-analytic approximations and harmonic synthesis ideas; later proofs employed the Gelfand–Mazur theorem, spectral radius formulas, and Wiener’s 1/f theorem viewpoint. Techniques draw on the theory of entire functions as developed by Bernard Malgrange and Lars Ahlfors, and use convolution identities, approximate identities, and factorization results similar to those of Helson and Lowdenslager. Extensions to noncommutative settings invoke methods from operator theory as in work by John von Neumann and Marshall Stone.

Generalizations and extensions

The theorem has been generalized to locally compact abelian groups and to algebras of measures on the circle group, with deepening links to spectral synthesis and Tauberian theory of Arne Beurling. L^p versions, including failure and partial analogues for p ≠ 1, were studied by Leopold Kronecker's successors and by H. Helson, while vector-valued and noncommutative analogues invoked tools from C*-algebra theory as in the work of Gelfand and Kaplansky. Multidimensional and discrete variants connect to results by Salem and Zygmund, and modern formulations tie into the Wiener algebra and multiplier theory used by Ludwig Faddeev and Israel Gelfand.

Applications

Applications arise across signal processing, where the theorem justifies deconvolution and filter design using nonvanishing frequency responses; in control theory for stability criteria; in probability theory for limit theorems involving characteristic functions; and in partial differential equations for constructing fundamental solutions. It informs methods in time series analysis, underpins results in spectral theory of convolution operators, and has influenced algorithmic approaches used at institutions such as Bell Labs and in projects linked to MIT and Princeton University research groups in applied analysis.

Examples and counterexamples

Classical positive examples include the sinc function whose Fourier transform (a rectangle function) does not vanish on its support and generates appropriate ideals; compactly supported functions with nonzero transforms provide generators in the convolution algebra on the circle. Counterexamples demonstrate limitations: functions in L^p for p ≠ 1 may fail to generate dense translates despite nonvanishing transforms, constructions due to Beurling and Salem exhibit spectral synthesis failure, and nonabelian group settings produce obstructions related to the absence of a suitable commutative Gelfand transform.

Category:Harmonic analysis