Splitting chains, tunnels and twisted sums (Q2022803)

A Banach space \(Z\) is a twisted sum of two Banach spaces \(Y\) and \(X\) if it has a closed subspace \(Y_1\) isomorphic to \(Y\) such that the quotient \(Z/Y_1\) is isomorphic to \(X\). The twisted sum is non-trivial if \(Y_1\) is not complemented in \(Z\). A compact space \(K\) has a tunnel if there is a continuous map \(f:K\to L\), where \(L\) is a linearly ordered topological space, such that \(f^{-1}(x)\) is nowhere dense in \(K\) for each \(x\in L\). Each tunnel in \(K\) induces a non-trivial twisted sum of the space of continuous functions \(C(K)\) and the space \(c_0(I)\) for some set \(I\). The main concern of the paper is the existence of tunnels in the compact \(\omega^*=\beta\omega\setminus\omega\), where \(\beta\omega\) is the Čech-Stone compactification of the set \(\omega\) of natural numbers, and the associated twisted sums. Note that, as a Banach space, \(C(\omega^*)\) can be identified with the quotient \(\ell_\infty/c_0\). The main result (Theorem~3.7) shows that if there is a splitting chain of clopen sets in \(\omega^*\), then a non-trivial twisted sum of \(C(\omega^*)\) and \(c_0(I)\) exists, where the cardinality of \(I\) is that of the continuum. It is shown that splitting chains of clopen sets do exist in \(\omega^*\) under some assumptions, e.g., under the Continuum Hypothesis or in the classical Cohen model, but one cannot prove their existence in ZFC. In fact, the authors do not know if twisted sums like those in Theorem~3.7 exist in ZFC.