What is the connection between Lp space and lp space, which are commonly used examples in functional analysis?

The connection between L^p and ℓ^p spaces is foundational in functional analysis, representing two primary manifestations of a unified concept of p-integrability applied to different measure-theoretic domains. Both are Banach spaces (and Hilbert spaces in the special case p=2) defined by a p-norm, but their underlying sets differ fundamentally: ℓ^p spaces consist of sequences of scalars (x_n) such that the sum ∑|x_n|^p is finite, effectively being function spaces on the counting measure of the set of natural numbers. In contrast, L^p spaces are defined for functions on a general measure space (like ℝ with Lebesgue measure), where integrability ∫|f|^p dμ < ∞ is required. Thus, ℓ^p can be viewed precisely as L^p(ℕ, counting measure), making it not just an analogy but a strict special case of the abstract L^p theory. This structural identity means all general theorems about L^p spaces—completeness, duality (where (L^p)* ≅ L^q for 1/p + 1/q = 1 with p finite), and embedding relations—apply directly to ℓ^p, providing a consistent framework.

However, despite this subsumption, the analytical properties of ℓ^p and L^p(ℝ) diverge significantly due to the atomic versus continuous nature of their measures. The sequence spaces ℓ^p exhibit discreteness that leads to properties not shared by all L^p spaces. For instance, the inclusion relations are reversed: for finite measure spaces, L^q ⊂ L^p when q > p, but for sequence spaces, ℓ^p ⊂ ℓ^q when p < q, because the counting measure of ℕ is infinite. Moreover, the geometry of their unit balls and the behavior of operators differ; compactness in ℓ^p is more readily characterized via conditions like the vanishing of tail norms, whereas in L^p(ℝ), compactness involves criteria like equicontinuity and tightness. These distinctions make ℓ^p often simpler to work with as a canonical example, yet they also highlight how measure-theoretic specifics modulate abstract functional-analytic principles.

Their interplay is most critically observed in duality and representation theory. The dual of ℓ^p is ℓ^q for 1 ≤ p < ∞, mirroring the duality of L^p spaces on σ-finite measures, but the dual of L^∞ is not L^1, whereas the dual of ℓ^∞ is a complicated space of bounded additive measures, not ℓ^1. This exception underscores that the nice duality for 1 < p < ∞ relies on the reflexivity of these spaces, a property both share. Furthermore, many key operators, such as the Fourier transform or convolution operators, are naturally studied on L^p(ℝ), but their discrete analogues (like convolution on ℓ^p) often serve as simplified models to gain insight, with the sequence space acting as a discrete testing ground for theorems later extended to continuous settings.

In practical analysis, the relationship is exploited constantly: ℓ^p techniques are used in the study of basis expansions in L^p, where functions are represented as sequences of coefficients, and many inequalities (like Hölder's or Minkowski's) are first grasped in the sequence context before their integral versions. The concrete nature of ℓ^p makes it an indispensable pedagogical and computational tool, while L^p spaces represent the full generality of the theory. Understanding that ℓ^p is literally L^p with counting measure clarifies that their differences are not of kind but of context, demonstrating how functional analysis separates the universal Banach space properties from those dependent on the underlying measure space.