A universal reflexive space for the class of uniformly convex Banach spaces (Q2498480)

From MaRDI portal
scientific article
Language Label Description Also known as
English
A universal reflexive space for the class of uniformly convex Banach spaces
scientific article

    Statements

    A universal reflexive space for the class of uniformly convex Banach spaces (English)
    0 references
    0 references
    0 references
    16 August 2006
    0 references
    In [Proc.\ Am.\ Math.\ Soc.\ 79, 241--246 (1980; Zbl 0438.46005)], \textit{J.\,Bourgain} proved that there is no universal space in the class of separable reflexive spaces: every separable Banach space which contains an isomorphic copy of every separable reflexive Banach space contains actually an isomorphic copy of every separable Banach space. He asked if there exists a separable reflexive Banach space which contains an isomorphic copy of every separable superreflexive Banach space (or, equivalently, of every separable uniformly convex Banach space). In the paper under review, the authors answer this question affirmatively. Their proof uses a result of \textit{S.\,Prus} [Bull.\ Pol.\ Acad.\ Sci., Math.\ 31, 281--288 (1983; Zbl 0547.46007)], which is a partial answer to Bourgain's question: there exists a reflexive Banach space \(X_P\) with a basis such that every separable superreflexive space with the approximation property is isomorphic to a complemented subspace of \(X_P\). The crucial role in his proof is played by the notions of \(p\)-lower and \(q\)-upper estimates for disjoint blocks of a given finite-dimensional decomposition (\(FDD\) with \((p,q)\)-estimates); in particular, this space \(X_P\) is universal for the class of Banach spaces having \(FDD\) with \((p,q)\)-estimates. The authors of the present paper introduce the notion of \((p,q)\)-tree estimates, \(1\leq q \leq p \leq \infty\), and show that for a reflexive Banach space \(X\), it is equivalent to say that: (a) \(X\) has \((p,q)\)-tree estimates, or that: (b) \(X\) is isomorphic to a subspace of a reflexive space having an \(FDD\) with \((p,q)\)-estimates. That answers Bourgain's question, since it follows from \textit{V.\,I.\,Gurarij} and \textit{N.\,I.\,Gurarij} [Math.\ USSR, Izv.\ 5 (1971), 220--225 (1972; Zbl 0248.46017); Izv.\ Akad.\ Nauk SSSR, Ser.\ Mat.\ 35, 210--215 (1971; Zbl 0214.12702)] (see also [\textit{R.\,C.\,James}, Pac.\ J.\ Math.\ 41, 409--419 (1972; Zbl 0235.46031)]) that every uniformly convex space has \((p,q)\)-tree estimates. This proof uses blocking techniques and \textit{M.\,Zippin}'s result that every reflexive space can be embedded into a reflexive space with a basis [Trans.\ Am.\ Math.\ Soc.\ 310, No.\,1, 371--379 (1988; Zbl 0706.46015)]. Actually, to show that (a) \(\Leftrightarrow\) (b), the authors go through a third property: (c) \(X\) is isomorphic to a quotient of a reflexive space having an FDD with \((p,q)\)-estimates.
    0 references
    0 references
    0 references
    finite-dimensional decomposition
    0 references
    \((p,q)\)-estimate
    0 references
    \((p,q)\)-tree estimate
    0 references
    reflexive space
    0 references
    0 references
    0 references