Suppose $X$ is a topological space and $x_0 \in X$. Let $$ X' = X \cup e^{n+1} $$ be obtained from adding a $(n+1)$-cell (so $X'$ is the pushout of the map $\partial e^{n+1} \to e^{n+1}$ and the attaching map $\partial e^{n+1} \to X$).

I want to show that the inclusion $X \to X'$ is an $n$-equivalence, i.e. that the induced map $i^\ast: \pi_k(X,x_0) \to \pi_k(X',x_0)$ is bijective for $k<n$ and surjective for $k=n$.

Using cellular approximation I can prove this when $X$ is a CW complex, but I did not succeed in solving the general case.

  • $\begingroup$ This is true in general, but what kind of attaching map is this? I think this follows from long exact sequence of homotopy groups. $\endgroup$ – Bombyx mori Jan 15 '15 at 8:45
  • $\begingroup$ The attaching map can be any continuous map from $\partial e^{n+1}$ to $X$. $\endgroup$ – user133281 Jan 15 '15 at 8:46
  • $\begingroup$ I think you can work with the pair $(A,B)=(X',X)$. I am not entirely sure if this is a good pair since $X$ is just some topological space. $\endgroup$ – Bombyx mori Jan 15 '15 at 8:49

I assume $X$ is Hausdorff and connected. (The connectedness is just because I don't want to care about base points. You can adapt the argument to the non-connected case. The assumption that $X$ is Hausdorff is more important. I'm unsure whether your claim holds in the non-Hausdorff case.)

An element in $\pi_i(X',X)$ is represented by a map of pairs $(D^i,S^{i-1})\rightarrow(X',X)$ which is homotopic as a map of pairs to a cellular map of relative CW-complexes by the cellular approximation theorem for relative CW-complexes, so it is zero in $\pi_i(X',X)$ for $0\le i\le n$. Your claim now follows from the LES of relative homotopy groups.

You'll find the relative cellular approximation theorem in the book of Spanier (Section 7.6, Cor. 18).

| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.