Distributions 12 | Finite-Order Distributions [dark version]

Finite-order distributions are singled out by a sharpened version of the basic estimate for distributions: the same derivative order must control the distribution on every compact set at once. Starting from the standard inequality—bounding |T(φ)| using finitely many sup norms of derivatives of a test function φ on a fixed compact K—the discussion replaces the sum over derivative orders by a maximum, absorbing constants so the bound depends on a single “worst” derivative order M. The key move is changing the quantifiers: instead of allowing M to depend on K, it requires one integer M that works uniformly for all compact sets. When such an M exists, T is called a distribution of finite order, and the smallest such M is its order.

This uniformity makes order-zero distributions especially important. If T is induced by a locally integrable function (a regular distribution), then T(φ) can be written as an integral of f(x)φ(x). Taking absolute values and using the sup norm of φ on the compact support yields an estimate with no derivatives of φ—meaning the regular distribution has order zero. The Dirac delta δ also lands in order zero: its action on test functions depends only on the value φ(0), again requiring no derivatives. So order-zero distributions already include both “ordinary” objects (functions) and “point-supported” objects (δ), making them a natural baseline class.

The most consequential fact is an identification between order-zero distributions and complex Radon measures. A complex Radon measure μ assigns complex values to Borel sets and is finite on compact sets. From such a μ, one constructs a distribution T_μ by pairing with test functions via integration: T_μ(φ)=∫ φ(x) dμ(x), with the integral effectively over the compact support of φ. This pairing automatically satisfies the order-zero estimate, so every complex Radon measure produces an order-zero distribution.

Conversely, the relationship is described as a canonical bijection between the set of complex Radon measures and the set of order-zero distributions. The practical takeaway is that, for order zero, working with distributions or with measures is largely a matter of preference.

A concrete example ties the theory to the familiar Dirac measure. The Dirac measure at the origin, often denoted δ_0, assigns 1 to any set containing 0 and 0 otherwise—representing a unit point charge. Plugging this measure into the integration formula gives T_{δ_0}(φ)=φ(0), which is exactly the defining action of the Dirac delta distribution. The conclusion is that the Dirac measure and the Dirac delta distribution are the same object under the order-zero correspondence.

With finite-order distributions defined via uniform derivative control, the discussion sets up the next step: studying operations on distributions while keeping track of how these operations affect the order class.