An isomorphic characterization of \(L^{1}\)-spaces (Q935939)

Isomorphic characterizations of \(L^1\) spaces can be obtained by means of abstract topological properties of vector lattices. In this paper, it is shown that a sequentially \((\tau)\)-complete topological vector lattice \(X_\tau\) is isomorphic to some \(L^1(\mu)\) if and only if the positive cone can be written as \(X_+= \mathbb{R}_+ B\) for some convex, \((\tau)\)-bounded and \((\tau)\)-closed set \(B \subseteq X_+ \setminus \{0\}\). A linear topology \(\tau\) defined on a topological ordered vector space is monotonically complete if and only if every monotonic \((\tau)\)-Cauchy net in \(X\) is \((\tau)\)-convergent; it is monotonically \(\sigma\)-complete if the same is satisfied with sequences instead of nets. A vector lattice has the Riesz decomposition property if \([0,x+y]=[x,0]+[y,0]\) for all \(x,y \in X_+\). The main representation result of the paper is Theorem 14 that provides an isomorphic characterization of \(L^1\)-spaces; it implies in particular, for the isometric version, the classical Kakutani representation theorem for (AL)-spaces. Let \(X_\tau\) be a real Hausdorff topological vector space endowed with a directed linear ordering \(\leq\) such that \(X_+=\mathbb{R}_+ B\) for some convex, \((\tau)\)-bounded, and \((\tau)\)-closed set \(B \subset X_+ \setminus \{0\}\). The result is given in terms of a particular norm \(\| \; \| _f\) that is defined using a functional \(f\) on \(X\) with some special properties (the construction and properties of \(f\) and \(\| \; \| _f\) are too involved to be explained here). Theorem 14. Assume that \(X_\tau\) has the Riesz decomposition property, and that \(\tau\) is monotonically \(\sigma\)-complete. Then \((X, \| \;\| _f) \simeq L^1(\mu)\) as normed lattices for some measure space \((T, {T}, \mu)\). If, in addition, the topology \(\tau\) is locally solid, then \(X_\tau \simeq L^1(\mu)\) as topological vector lattices. Some corollaries follow easily from this result. For example, if \(X_\tau\) is a sequentially \((\tau)\)-complete topological vector lattice, then \(X_\tau \simeq L^1(\mu)\) as a topological vector lattice for some measure \(\mu\) (Corollary 16). Another result is that, if \(X_\tau\) is a vector lattice and a Fréchet space (metrizable and complete), then \(X_\tau \simeq L^1(\mu)\) as topological vector lattices for some measure \(\mu\) (Corollary 17). Some applications of these results are also given. For instance, Theorem 19 states that if \(Y\) is a real normed space and if the ordering induced by the cone \(Y_+=\mathbb{R}_+ \overline{B_Y(u,1)}\) (\(\| u\| >1\)) has the Riesz decomposition property, then \(Y\) has finite dimension. Other applications dealing with a pointwise ordering on a space of multivariate polynomials are also given.