Skip to main content

Section 22.1 Postnikov Towers

The construction is by repeatedly attaching cells: for each generator of \(\pi_{n+1}(X)\text{,}\) attach an \((n+2)\)-cell to kill it; this may introduce new \(\pi_{n+2}\)-elements, which are killed by attaching \((n+3)\)-cells; and so on. The same technique was used in the construction of Eilenberg–MacLane spaces (Chapter 13). The space \(X_{\leq n}\) is called the \(n\)th Postnikov section of \(X\text{,}\) and is unique up to weak equivalence.
The inclusion \(X \hookrightarrow X_{\leq n}\) extends to a map \(X_{\leq n+1} \to X_{\leq n}\) (since \(X_{\leq n+1}\) is obtained from \(X\) by attaching cells of dimension \(\geq n + 3\text{,}\) all of which can be filled in \(X_{\leq n}\) where \(\pi_i = 0\) for \(i > n\)). This yields the Postnikov tower of \(X\text{:}\)
The bottom layer is \(X_{\leq 1} \simeq K(\pi_1 X, 1)\text{.}\) By converting \(X_{\leq n+1} \to X_{\leq n}\) into a fibration, the long exact sequence in homotopy identifies the fibre as \(K(\pi_{n+1} X, n+1)\text{.}\)