A new dichotomy for Banach spaces (Q2565231): Difference between revisions

During the last years the author provided a number of very deep and interesting results which settled many problems which were open since the beginnings of Banach space theory. Here he announces a theorem which solves a problem posed in Banach's book: The separable Hilbert space is the only Banach space \(X\) such that every infinite-dimensional subspace of \(X\) is isomorphic with \(X\). This is obtained by combining a theorem of \textit{R. A. Komorowski} and \textit{N. Tomczak-Jaegermann} [Isr. J. Math. 89, No. 1-3, 205-226 (1995; Zbl 0830.46008)] with the following new dichotomy: Every infinite-dimensional Banach space \(X\) contains an infinite-dimensional subspace \(Y\) which either has an unconditional basis or is hereditarily indecomposable (the latter property means that no subspace of \(Y\) is isomorphic to a topological direct sum of infinite-dimensional spaces). The proofs of these theorems will appear elsewhere. The paper under review contains a very interesting description of the history of the problems under consideration as well as many informations on what is known about hereditarily indecomposable spaces. Also one finds a sketch of the proofs: They depend on combinatorial constructions which are in the spirit of the Nash-Williams theorem, and a crucial role is played by assertions on winning strategies for a certain game which deals with blocking a prescribed sequence.