1 Definition and basic ideas
A tempered distribution is a generalized function designed to act on Schwartz functions, which are smooth and decrease faster than any polynomial at infinity. This makes tempered distributions especially suitable for analysis on the whole real line or Euclidean space, where ordinary integrability may fail but controlled growth still permits meaningful pairing with test functions. They are central in distribution theory because they enlarge the class of objects that can be differentiated and transformed by Fourier methods.
1.1 Schwartz space
Schwartz space is the collection of smooth functions whose derivatives, after multiplication by any polynomial, remain bounded. These functions decay rapidly at infinity together with all of their derivatives. The space is commonly denoted by a symbol such as 𝒮 and serves as the standard test function space for tempered distributions.
1.2 Continuous linear functionals
A continuous linear functional assigns a number to each Schwartz function in a linear and continuous way. Continuity here is understood with respect to the natural topology on Schwartz space, which reflects control over all derivatives and rates of decay. Such functionals generalize integration against ordinary functions.
1.3 Tempered distributions as dual objects
Tempered distributions are precisely the continuous linear functionals on Schwartz space. In this dual role, they extend classical functions, measures, and singular objects such as point masses and derivatives of these masses. Their defining feature is not pointwise behavior, but how they interact with rapidly decreasing test functions.
1.4 Examples of tempered distributions
Basic examples include ordinary functions that grow at most polynomially, the Dirac delta distribution, and its derivatives. Locally integrable functions with moderate growth define tempered distributions by integration against Schwartz functions. Finite measures with suitable growth behavior also fit naturally into this framework.
2 Construction and embeddings
Tempered distributions arise by embedding several familiar mathematical objects into the dual of Schwartz space. This construction preserves many operations used in classical analysis while extending them beyond the realm of ordinary functions. The resulting space is broad enough to include both smooth and singular phenomena.
2.1 From locally integrable functions
If a function is locally integrable and does not grow faster than a polynomial at infinity, then it defines a tempered distribution. The action on a Schwartz function is given by integrating the product of the two functions over space. Rapid decay of the test function compensates for the moderate growth of the original function.
2.2 Regular distributions
Any sufficiently regular function can be regarded as a distribution by integration against test functions. When such a function belongs to the tempered class, its associated distribution is called regular in this setting. This identification allows analysis to treat functions and distributions within the same formalism.
2.3 Embedding of measures
Measures of finite order or controlled growth can often be embedded as tempered distributions. The distribution acts by integrating the test function against the measure. This includes point masses and many signed or absolutely continuous measures encountered in analysis.
2.4 Inclusion of classical functions of moderate growth
Classical functions with polynomial growth at infinity fit naturally into the theory. Examples include polynomials, trigonometric functions, and many smooth functions arising in physics and applied mathematics. Their inclusion is one reason tempered distributions are so useful in Fourier analysis.
3 Operations on tempered distributions
One of the main advantages of tempered distributions is that many standard operations extend beyond ordinary functions. Differentiation becomes purely algebraic, translation behaves naturally, and convolution can be defined in important cases. These extensions preserve consistency with the classical formulas whenever both sides make sense.
3.1 Differentiation
Derivatives of tempered distributions are defined by moving the derivative onto the test function with a sign change. This procedure always produces another tempered distribution. As a result, objects with no classical derivative can still be differentiated in a generalized sense.
3.2 Multiplication by polynomials
Multiplying a tempered distribution by a polynomial is always allowed. The polynomial acts on the test function before the distribution is evaluated. This is compatible with the growth conditions built into Schwartz space and is essential for many algebraic manipulations.
3.3 Translation
A tempered distribution may be shifted in space by translating the test function in the opposite direction. Translation is continuous and interacts well with differentiation and Fourier transform. This makes tempered distributions suitable for studying problems with spatial symmetry.
3.4 Convolution
Convolution is a fundamental operation, though it is not defined for every pair of tempered distributions. In favorable cases, it yields another tempered distribution and reflects the smoothing and averaging effects familiar from classical analysis. It is particularly important in solving differential equations and filtering signals.
3.4.1 Convolution with Schwartz functions
A tempered distribution can always be convolved with a Schwartz function. The result is a smooth function, often rapidly decaying, obtained by combining the distribution with a translated test function. This provides a powerful regularization method.
3.4.2 Convolution with compactly supported distributions
When one factor has compact support, convolution is well behaved and extends many classical identities. Compact support prevents uncontrolled growth from interfering with the operation. This case is widely used in local analysis and the theory of differential operators.
4 Fourier transform
The Fourier transform is one of the main reasons tempered distributions are so important. It extends naturally from Schwartz functions to tempered distributions and preserves the duality between decay and smoothness. This extension provides a rigorous framework for frequency analysis of generalized functions.
4.1 Definition on Schwartz space
On Schwartz space, the Fourier transform maps rapidly decreasing smooth functions to other rapidly decreasing smooth functions. The transform is defined by an integral formula with an oscillatory exponential kernel. The result remains within the same test function space, which is crucial for duality arguments.
4.2 Extension to tempered distributions
A tempered distribution is transformed by requiring compatibility with the Fourier transform on test functions. The transformed distribution acts on a Schwartz function by applying the original distribution to the Fourier transform of that test function, up to the chosen normalization. This defines a continuous automorphism of the tempered distribution space.
4.3 Inversion and Plancherel-type properties
The Fourier transform on tempered distributions satisfies an inversion principle analogous to the classical one. Under standard normalization, applying the transform twice yields reflection, and four applications return the original object. For square-integrable functions, the theory aligns with Plancherel’s theorem, linking spatial and frequency energy.
4.4 Fourier transform of standard examples
The Fourier transform of the Dirac delta is constant, while the transform of a constant function is a delta distribution. Derivatives transform into multiplication by polynomial factors, and polynomial multiplication corresponds to differentiation in frequency. These examples illustrate how tempered distributions unify algebraic and analytic viewpoints.
5 Topological and functional-analytic structure
The theory of tempered distributions depends on a carefully chosen topology on the test function space and the corresponding dual topology. These structures ensure that convergence, boundedness, and continuity behave in a controlled and useful manner. The framework is part of the broader setting of locally convex spaces.
5.1 Topology of the Schwartz space
The topology of Schwartz space is defined by a family of seminorms measuring the size of a function and its derivatives after polynomial weighting. A sequence converges when all such seminorms converge to zero. This topology captures both smoothness and rapid decay simultaneously.
5.2 Dual topology on tempered distributions
The dual space inherits a topology determined by how distributions act on bounded sets of Schwartz functions. This makes it possible to discuss convergence of tempered distributions in a meaningful way. The resulting structure is well suited to limit processes in analysis.
5.3 Continuity and boundedness
Continuity of a tempered distribution means its action is controlled by finitely many seminorms on Schwartz space. Boundedness of sets of test functions plays an important role in describing this control. These properties ensure that generalized functions do not behave arbitrarily on rapidly decreasing inputs.
5.4 Seminorms and convergence
Seminorms provide a practical way to measure growth and decay in Schwartz space. They also allow one to characterize convergence of distributions through uniform estimates. This framework supports approximation arguments and the analysis of limiting processes.
6 Singularities and growth conditions
Tempered distributions are broad enough to include singularities, but only under growth restrictions that make them manageable. Their behavior is governed by how they act on rapidly decreasing test functions rather than by pointwise values. This allows the theory to capture both local irregularity and global decay.
6.1 Polynomial growth
A key condition for temperateness is that growth at infinity be no faster than a polynomial. Such growth is mild enough that rapid decay of test functions compensates for it. This condition excludes many highly divergent objects while retaining a large and useful class.
6.2 Local singular behavior
Tempered distributions may have singularities at isolated points, along sets, or even everywhere in a local sense. Examples include derivatives of step functions and point masses. The theory accommodates these singularities by interpreting them through their action on test functions.
6.3 Support and singular support
The support of a distribution describes where it does not vanish, while the singular support locates where it fails to be smooth. These notions help distinguish local regularity from global distributional behavior. They are important in understanding propagation of singularities in analysis.
6.4 Homogeneous distributions
Homogeneous distributions scale in a predictable way under dilation. They often arise in potential theory, harmonic analysis, and the study of fundamental solutions. Their scaling properties make them especially compatible with Fourier methods.
7 Applications
Tempered distributions appear in many branches of analysis and mathematical physics. Their strength lies in allowing one to use distributional derivatives and Fourier transform techniques on objects that are not ordinary functions. This makes them a standard tool in both pure and applied settings.
7.1 Partial differential equations
In partial differential equations, tempered distributions provide a natural setting for weak solutions and fundamental solutions. Differential operators act cleanly on them, and Fourier transform often converts a PDE into an algebraic equation. This simplifies the study of existence, uniqueness, and regularity.
7.2 Harmonic analysis
Harmonic analysis uses tempered distributions to study signals and functions through frequency decomposition. They allow singular objects to be analyzed with the same tools used for smooth functions. Many results on convolution, multipliers, and transform methods are formulated most naturally in this language.
7.3 Quantum mechanics
In quantum mechanics, distributions model idealized states and observables that are not ordinary functions. The Fourier transform relates position and momentum representations, making tempered distributions particularly useful. They also support formal manipulations that arise in spectral and scattering theory.
7.4 Signal processing
Signal processing often relies on convolution, filtering, and frequency analysis, all of which fit well within tempered distribution theory. Idealized impulses and discontinuities can be represented rigorously by distributions. This helps bridge theoretical models and practical transform-based methods.
8 Related classes of distributions
Tempered distributions belong to a larger family of generalized function spaces. Other classes emphasize different features such as locality, compact support, or stronger decay. Comparing these spaces clarifies the special role of tempered distributions in Fourier analysis.
8.1 General distributions
General distributions act on compactly supported smooth test functions rather than on Schwartz functions. They are broader in some local respects but are not automatically compatible with global Fourier analysis. Tempered distributions form a distinguished subspace with stronger global control.
8.2 Rapidly decreasing distributions
Rapidly decreasing distributions are generalized objects with even stronger decay properties in some formulations. They often arise in refined harmonic analysis and spectral theory. Their relation to tempered distributions depends on the exact functional framework being used.
8.3 Compactly supported distributions
Compactly supported distributions vanish outside a bounded region in the distributional sense. Such distributions have especially good convolution properties and are often easier to localize. They are automatically tempered under standard conditions.
8.4 Semyonov-type and other generalized function spaces
More specialized generalized function spaces are designed to capture additional growth, regularity, or asymptotic features. These spaces may extend or refine the tempered distribution framework for particular applications. They are part of the broader effort to adapt distribution theory to different analytical settings.