A new proof of the odd primary homotopy exponent of spheres (Q453352): Difference between revisions

It is an important theorem of Cohen, Moore and Neisendorfer that there exist maps \(\varphi:\Omega^2S^{2n+1}\to S^{2n-1}\) such that \(E^2\circ\varphi\) is homotopic to \(\Omega^2 p:\Omega^2 S^{2n+1}\to\Omega^2S^{2n+1}\). Here \(p\) is an odd prime, and all spaces are implicitly localised at \(p\), and we also use the symbol \(p\) for the map \(S^{2n+1}\to S^{2n+1}\) of degree \(p\). Moreover, \(E^2: S^{2n-1}\to\Omega^2S^{2n+1}\) is the double suspension map. It follows easily from this theorem that the groups \(\pi_k(S^{2n+1})\) have exponent dividing \(p^n\) for all \(k>2n+1\). The original proof is long, and makes extensive use of differential Lie algebras and homotopy decompositions of Moore spaces. The paper under review gives a new construction of a map \(\varphi\) as above, which is shorter and uses a different set of methods that are more homotopical than algebraic. These methods arise from earlier work of the author with Gray on some related questions; they involve a new way of mixing cofibrations with fibrations. Another feature of the new proof is that the case where \(n\) is divisible by \(p\) is treated first, and the general case is then recovered by relating the \(n\)'th case to the \(np\)'th case. The new construction may or may not give the same map \(\varphi\) as the one produced by Cohen, Moore and Neisendorfer, but that question does not seem to be important for the applications. Both the proof and the applications use an extensive network of fibrations and homotopy pullback squares, including {\parindent=6mm \begin{itemize}\item[{\(\bullet\)}] the Hopf fibration \(J_{p-1}S^{2n}\to\Omega S^{2n+1}\to\Omega S^{2np+1}\) \item[{\(\bullet\)}] the Toda fibration \(S^{2n-1}\to\Omega J_{p-1}S^{2n}\to\Omega S^{2np-1}\) \item[{\(\bullet\)}] the double suspension fibration \(S^{2n-1}\to\Omega^2S^{2n+1}\to BW_n\) \item[{\(\bullet\)}] the Anick fibration \(\Omega^2S^{2n+1}\to S^{2n-1}\to T_{2n+1}\). \end{itemize}} Various standard facts about these fibrations were originally proved by quoting the work of Cohen, Moore and Neisendorfer; they are proved again in the present paper by different methods, to eliminate that dependence.