When Kalton and Peck met Fourier (Q6073541)

This paper is pure jaywalking through Kalton's ``Centralizer Avenue''. To fix ideas, and cheating just a bit, given a Banach algebra \(A\) and two Banach \(A\)-modules \(X,Y\), an \(A\)-centralizer \(\Omega: X \curvearrowright Y\) is a homogeneous map such that, for all \(a\in A\), all \(x\in X\), and some \(C>0\), \[ \|\Omega(ax) -a\Omega (x)\|\leq C \|a\|\|x\|. \] It was the beautiful discovery of \textit{N.~J. Kalton} [Nonlinear commutators in interpolation theory. Providence, RI: American Mathematical Society (AMS) (1988; Zbl 0658.46059); Trans. Am. Math. Soc. 333, No. 2, 479--529 (1992; Zbl 0776.46033)], based on the paper of \textit{N. J. Kalton} and \textit{N. T. Peck} [Trans. Am. Math. Soc. 255, 1--30 (1979; Zbl 0424.46004)] that \(L_\infty\)-centralizers are at the core of Banach space theory, firmly stuck at the intersection between interpolation theory, homological Banach space theory and complex variable theory. A good reason why this is so is because many classical Banach spaces carry with them an \(L_\infty\)-module structure (for different choices of the algebra \(L_\infty(\mu)\)): spaces with unconditional basis or Köthe spaces among them. At the center of it all, one always encounters the first and most important \(L_\infty\)-centralizer: the Kalton-Peck map \[ \mathsf{KP}(x) = x \log \frac{|x|}{\|x\|} \] whose demeanor, as well as the basic facts about centralizers, can be followed in [\textit{F.~Cabello Sánchez} and \textit{J.~M.~F. Castillo}, Homological methods in Banach space theory. Cambridge: Cambridge University Press (2023; Zbl 1519.46001)]. The map \(\mathsf{KP}\) is an \(\ell_\infty\)-centralizer on spaces with unconditional basis, an \(L_\infty\)-centralizer on Köthe spaces, and can be worse things in other spaces (see below). Bounded maps are trivially centralizers, so one is interested in unbounded centralizers. The present work deals with the so far unfashionable \(L_1\)-centralizers, uncovering a few mindblowing spots: hidden Diagon alleys, Escher crossings and Spiderwick gardens to which one can only arrive through a backdoor in the wardrobe. In what follows, \(G\) is a compact abelian group, \(\Gamma\) is its dual group of characters, and \(\Omega\) an \(\ell_\infty(\Gamma)\)-centralizer on \(\ell_2(\Gamma)\) (or an \(\ell_\infty(S)\)-centralizer in the third assertion below). A carefree description of some results of the authors could be: \begin{itemize} \item If \(\mathscr F\) denotes the Fourier transform, \(\mathscr F^{-1} \Omega \mathscr F\) is an \(L_1\)-centralizer. \item \(\mathscr F^{-1} \mathsf{KP} \mathscr F\) is an unbounded \(L_1\)-centralizer. \item If \(S\subset \Gamma\) is a Sidon set, then \(\Sigma \Omega \mathscr F |_{S}\) is an unbounded \(L_1\)-centralizer. \item \(L_\infty\)-centralizers that commute with translations are \(L_1\)-centralizers. \end{itemize} But, of course, the awe is in the details. None of those assertions exhausts the subtleties of the proofs in the paper. And I have not touched the applications to the existence of nontrivial exact sequences also presented in the paper, among which my favourite is the explicit construction of a nontrivial sequence \[ 0\longrightarrow \ell_1\longrightarrow Z\longrightarrow c_0\longrightarrow 0 \,. \] Readers interested in these wonderful objects should read [loc. cit. Zbl 1519.46001; Proposition~5.2.20 and Section~7.5.2] or the comments in Section~5.2 of this paper; those interested in exact sequences of Banach spaces with additional \(G\)-module structure could go to [\textit{J.~M.~F. Castillo} and \textit{V.~Ferenczi}, Bull. Malays. Math. Sci. Soc. (2) 46, No.~4, Paper No.~135, 47~p. (2023; Zbl 1528.46059)]. Those who want to witness the work of Cabello about the secret life of centralizers could just read my earlier zbMATH review of the paper [\textit{F.~Cabello Sánchez}, Rev. Mat. Iberoam. 37, No.~6, 2309--2346 (2021; Zbl 1482.46095)].