Characteristic of convexity of Köthe function spaces (Q1204243)

Let \(E\) be a Köthe function space on \((\Omega,\Sigma,\mu)\) and \(X\) an arbitrary normed space. The space \(E(X)\) is the space of all measurable functions \(x\) from \(\Omega\) to \(X\) such that the function \(\lfloor x\rceil:\Omega\to\mathbb{R}\), \(\lfloor x\rceil(\omega):=\| x(\omega)\|\), belongs to \(E\) endowed with the norm \(\| x\|=\|\lfloor x\rceil\|_ E\). The following statements concerning the characteristic of convexity \(\varepsilon_ 0(E(X))\) of \(E(X)\) are proved: 1. \(\max\{\varepsilon_ 0(E),\varepsilon_ 0(X)\}\leq\varepsilon_ 0(E(X))\leq\varepsilon_ 0(E)+\varepsilon_ 0(X)- {1\over2}\varepsilon_ 0(E)\varepsilon_ 0(X)\). 2. Both inequalities in 1. are equalities iff \(\varepsilon_ 0(E)\in\{0,2\}\) or \(\varepsilon_ 0(X)\in\{0,2\}\). In particular (a) \(E(X)\) is uniformly convex iff both \(E\) and \(X\) are uniformly convex. (b) \(\varepsilon_ 0(E(X))=\varepsilon_ 0(E)\) if \(X\) is uniformly convex. (c) \(\varepsilon_ 0(E(X))=\varepsilon_ 0(X)\) if \(E\) is uniformly convex. (d) \(E(X)\) is uniformly non-square iff both \(E\) and \(X\) are uniformly non-square. 3. For all \(\varepsilon_ 1,\varepsilon_ 2\in(0,2)\) and \(\varepsilon_ 3\in(\max\{\varepsilon_ 1,\varepsilon_ 2\},\varepsilon_ 1+ \varepsilon_ 2-{1\over 2}\varepsilon_ 1\varepsilon_ 2)\), there exists \(E\) with \(\varepsilon_ 0(E)=\varepsilon_ 1\) such that \(\varepsilon_ 0(E(X))=\varepsilon_ 3\) whenever \(\varepsilon_ 0(X)=\varepsilon_ 2\). 4. For each \(\varepsilon_ 1\in(0,2)\) there is \(E\) with \(\varepsilon_ 0(E)=\varepsilon_ 1\) such that \(\varepsilon_ 0(E(X))=\varepsilon_ 1\) whenever \(\varepsilon_ 0(X)\leq {1\over 2}\varepsilon_ 1^ 2\). 5. If \(\varepsilon_ 0(E)\geq 1\) or \(\varepsilon_ 0(X)\geq 1\) then the upper bound in 1. can be attained, namely: (a) For each \(\varepsilon_ 1\geq 1\), there exists an \(E\) with \(\varepsilon_ 0(E)=\varepsilon_ 1\) such that \[ \varepsilon_ 0(E(X))=\varepsilon_ 1+\varepsilon_ 0(X)-{1\over 2}\varepsilon_ 1 \varepsilon_ 2(X)\;\forall X. \] (b) For each \(\varepsilon_ 2\geq 1\) and \(\varepsilon_ 1\in(0,2)\), there exist \(E\) and \(X\) with \(\varepsilon_ 0(E)=\varepsilon_ 1,\varepsilon_ 0(X)=\varepsilon_ 2\) and \(\varepsilon_ 0(E(X))=\varepsilon_ 1+\varepsilon_ 2-{1\over 2}\varepsilon_ 1 \varepsilon_ 2\). 6. If \(0<\varepsilon_ 0(E)<1\) and \(0<\varepsilon_ 0(X)<1\), then \[ \varepsilon_ 0(E(X))<\varepsilon_ 0(E)+\varepsilon_ 0(X)-{1\over 2}\varepsilon_ 0(E)\varepsilon_ 0(X). \]