Schwartz Distribution Theory

Motivation: A Rigorous Language for “Delta Functions”

Physicists have long used the Dirac “delta function” δ(x): infinite at the point 0, zero everywhere else, with ∫δ(x)dx = 1. Classically, such an object does not exist. Between 1945–50, Schwartz constructed the theory of distributions—making δ(x) a rigorous object. The key property: every distribution is infinitely differentiable! This provides a strict foundation for physical “functions” like δ(r−r₀) (a point charge).

The Space of Test Functions and Distributions

D(Ω) = C₀^∞(Ω): infinitely differentiable functions with compact support. Topology: φₙ → φ—supports lie within a common K, D^α φₙ ⇒ D^α φ uniformly for all α.

Schwartz space S(ℝⁿ): rapidly decaying functions: sup_x |x^β D^α φ| < ∞ for all α,β. D ⊂ S.

Distribution T ∈ D'(Ω): a linear continuous functional on D. ⟨T, φ⟩ = T(φ).

Regular distributions: f ∈ L¹_{loc} → T_f: ⟨T_f, φ⟩ = ∫fφ dx.

Delta function: ⟨δ, φ⟩ = φ(0). Not a function in the conventional sense.

Derivative of δ: ⟨δ', φ⟩ = −⟨δ, φ'⟩ = −φ'(0). Physically—a dipole.

Derivative of a Distribution

Definition: ⟨T', φ⟩ = −⟨T, φ'⟩ for all φ ∈ D.

Every distribution is infinitely differentiable!

Heaviside function: H(x) = 0 for x < 0, 1 for x ≥ 0. ⟨H', φ⟩ = −⟨H, φ'⟩ = −∫₀^∞ φ'(x)dx = −[φ]₀^∞ = φ(0) = ⟨δ, φ⟩. H' = δ ✓.

Fundamental solutions: P(D)E = δ. If P(D)E = δ, then P(D)(E*f) = f. Solution to P(D)u = f: u = E * f.

Numerical Example

Problem: Prove that (|x|)'' = 2δ(x) in the sense of distributions.

Step 1. From the previous article: (|x|)' = sign(x) as a distribution.

Step 2. (|x|)'' = (sign(x))' in the sense of distributions. By definition: ⟨sign', φ⟩ = −⟨sign, φ'⟩ = −∫_{-∞}^∞ sign(x)·φ'(x)dx.

Step 3. −∫₋∞^∞ sign(x)φ'(x)dx = −[∫₋∞⁰ (−1)φ'dx + ∫₀^∞ 1·φ'dx]. = ∫₋∞⁰ φ'dx − ∫₀^∞ φ'dx = [φ]₋∞⁰ − [φ]₀^∞ = (φ(0)−0) − (0−φ(0)) = 2φ(0).

Step 4. ⟨sign', φ⟩ = 2φ(0) = ⟨2δ, φ⟩. Thus (|x|)'' = 2δ(x) ✓.

Step 5. Physical meaning: |x| is the potential of a “linear spring” (V = k|x|). Force = −V' = −k·sign(x). Derivative of the force = −V'' = −2kδ(x)—“point mass” (or “point load”) at x=0. In mechanics: a concentrated load on a beam is described by a δ-function.

Step 6. Fundamental solution of the Laplacian: Δ(1/(4π|x|)) = −δ(x) in ℝ³. The potential of a point charge q: φ = q/(4πε₀|x|)—a direct consequence of distribution theory.

Real-World Applications

Electrodynamics and mechanics: charge of a point particle = ρ(r) = q·δ(r−r₀). Point load on a structure = f(x) = F·δ(x−x₀). Distribution theory allows one to rigorously solve Maxwell, Navier–Stokes, and elasticity equations with point sources—without “physical” reasoning about “infinite” functions.

Additional Aspects

Distributions (generalized functions in the sense of Schwartz) are defined as continuous linear functionals on S(ℝⁿ), the space of rapidly decaying smooth functions. This allows one to differentiate any locally integrable function indefinitely (in the sense of distributions) and gives a rigorous meaning to objects like δ(x), δ'(x), v.p.(1/x). Convolution of distributions and the Fourier transform are correctly extended to the space of tempered distributions S'. The fundamental solution E(x) of a differential operator L satisfies L·E = δ, and the solution to the nonhomogeneous equation Lu = f is found by convolution u = E * f. The Schwartz kernel theorem establishes a bijection between continuous operators and distributions of two variables—a foundation of the theory of pseudodifferential operators and microlocal analysis.

Connection with Other Areas of Mathematics

The theory of distributions has become the standard language for linear differential equations with singular right-hand sides. In the classical monograph by Hörmander, it is shown that any linear differential operator with smooth coefficients acts continuously on the spaces D' and S' and that fundamental solutions are conveniently described as distributions with controlled growth. For elliptic operators, this is linked to regularity theorems: if Lu = f in the sense of distributions and f is smooth, then u is automatically smooth, formalizing the principle of elliptic regularization.

The Fourier transform on tempered distributions S' lies at the heart of the theory of pseudodifferential operators (Calderón, Zygmund, Kohn, Nirenberg). There, the symbols of operators are described through their action on S', and the singularities of solutions are analyzed microlocally. The concept of the wave front of a distribution (Hörmander) connects distribution theory with modern symplectic geometry and the topology of contact manifolds, since the support of the wave front lies in the cotangent bundle.

In functional analysis, the spaces of test functions and their dual spaces correspond to a whole zoo of nuclear spaces in the sense of Gelfand–Shilov. The Schwartz kernel theorem, relating continuous linear operators to distributions of two variables, is a concrete realization of tensor products of nuclear spaces and is used in the theory of operator algebras.

In probability, generalized stochastic processes, such as white noise, are understood as random distributions: these are measures on D' or S'. The Minlos–Bogolyubov construction connects positive-definite functionals on spaces of test functions to probability measures on the space of distributions, which is used in quantum field theory and stochastic analysis.

Numerical methods for equations with singular sources rely on regularization and weak formulations. The finite element method and boundary element method treat right-hand sides as functionals on Sobolev spaces, i.e., as distributions, and many existence theorems and a posteriori estimates are formulated precisely in this language.

Historical Background and Development of the Idea

The prehistory goes back to ideas of generalized derivatives in Storch, Sokhotski, and Hadamard, as well as the use of singular integrals of Cauchy and Poincaré. However, a complete axiomatic structure did not exist. In 1935, Jean Leray introduced the concept of generalized potential in the context of the Navier–Stokes equations, and Laurent Schwartz in 1945–1950 formalized everything into a unified theory of generalized functions. The key works were published in the Comptes Rendus de l’Académie des Sciences and the two-volume "Théorie des distributions" (Hermann, 1950–1951). It is precisely for these results that Schwartz received the Fields Medal in 1950. Motivating problems came from quantum mechanics and electrodynamics: Dirac’s formalism with bra–kets and the “δ-function”, as well as the Poynting theorem and Maxwell’s equations with point sources, required a rigorous language.