Cone characterization of Grothendieck spaces and Banach spaces containing \(c _{0}\) (Q409290)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Cone characterization of Grothendieck spaces and Banach spaces containing \(c _{0}\)
scientific article

    Statements

    Cone characterization of Grothendieck spaces and Banach spaces containing \(c _{0}\) (English)
    0 references
    0 references
    0 references
    0 references
    12 April 2012
    0 references
    A cone is a non-empty convex subset \(P\) of a Banach space such that \(aP\subset P\) for all positive \(a\). Naturally, a cone is said to be a copy of another one if there is a (norm-) bicontinuous positively linear bijection between them. It is known that a Banach space is non-reflexive iff the positive cone \(l_1^+\) of \(l_1\) embeds in it. In Proposition 2, the authors give a characterization when the self-difference \(P-P\) of such an \(l_1^+\)-copy \(P\) is isomorphic to \(l^1\). In contrast to this, Theorem 3 states that, if a Banach space contains a copy of the positive cone of \(c_0\), then the self-difference of this copy is always isomorphic to \(c_0\). In Theorem 6, using Rosenthal's two dichotomy theorems on \(l_1\) and \(c_0\), the authors give a condition in terms of cones under which a copy of \(l_1^+\) gives rise to a copy of \(c_0\). Theorem 15 states that a Banach space is non-Grothendieck (meaning that its dual contains a \(w^*\)-convergent sequence which does not converge weakly) iff its dual admits a well-based cone \(P\) whose predual cone has empty interior but contains a quasi interior point (see the article for the definitions). Since a separable space is non-Grothendieck iff it is non-reflexive, the following characterization, Theorem 18, may be given: a separable Banach space is non-reflexive iff it contains a closed cone with empty interior whose dual cone is a copy of \(l_1^+\). Reviewer's remarks. (1) The proof of Theorem 3 can be reduced to the same triangle-inequality argument which follows Theorem 4. (2) The first part of the proof of Theorem 6 can be reduced to the observation that, on the one hand, a weakly convergent sequence admits not only convex combinations (as stated in the article) but convex blocks which converge in norm and that, on the other hand, such blocks are not possible for the canonical basis of \(l_1\). This is more elementary and does not use [\textit{E. Casini} and \textit{E. Miglierina}, Nonlinear Anal., Theory Methods Appl., Ser. A, Theory Methods 72, No. 5, A, 2356--2366 (2010; Zbl 1194.46015)].
    0 references
    0 references
    cones
    0 references
    bases for cones
    0 references
    conic isomorphisms
    0 references
    Grothendieck spaces
    0 references
    \({c_0^+}\)
    0 references
    \({\ell_1^+}\)
    0 references
    0 references