Iterated doubles of the Joker and their realisability (Q721501): Difference between revisions

The cyclic \(\mathcal{A}(1)^{\ast}\)-module \(J:=\mathcal{A}(1)^{\ast}/(\text{Sq}^{3})\), known as the ``Joker'', was first studied by \textit{J. F. Adams} and \textit{S. B. Priddy} [Math. Proc. Camb. Philos. Soc. 80, 475--509 (1976; Zbl 0338.55011)] in connection with the \(\Omega^{\infty}\)-structure on BSO and BU, and has since appeared in many contexts in stable homotopy theory. To explain the name, the picture on p. 341 should be rotated by \(90^{o}\). The question of its realizability requires first equipping it with a full \(\mathcal{A}^{\ast}\)-module structure (which is possible in two ways, dual to each other) and then realizing this module by a spectrum or space. Using the notion of doubling for \(\mathcal{A}\) and \(\mathcal{A}(n)\), one can construct \(\mathcal{A}(n)\)-modules \(\text{Joker}(n)\) for each \(n\geq 2\) by (iterated) doubling of the original \(J\) (resulting in spaced out versions of the original picture), again with two possible \(\mathcal{A}\)-module structures, dual to each other. The author uses some well-known Toda brackets to show that the two \(\mathcal{A}\)-module structures for \(J\) can be realized by finite \(2\)-local CW spectra Spanier-Whitehead dual to each other (these can be used to recover two of the four realizations of \(\mathcal{A}(1)\)). This is also true of \(\text{Joker}(n)\) for \(n=2,3\). However, in Theorem 4.6 he shows that \(\text{Joker}(n)\) cannot be realized for \(n\geq 4\) (which is clear for \(n\geq 5\) by \textit{J. F. Adams} [Ann. Math. (2) 72, 20--104 (1960; Zbl 0096.17404)]). He also discusses the question of unstable realization (by spaces).