Consider two pointed CW complexes $(X,x_0),(Y,y_0)$ and their wedge sum $(X \vee Y,a)$ with $a$ the identification of the two base points. I want to give a sufficient condition under which we have

$$\pi_1(X \vee Y, a) = \pi_1(X,x_0) \ast \pi_1(Y,y_0) ~~.$$

Right now it is not clear to me whether we even need any additional assumption or whether this is true for any CW complexes. In case we need more conditions I am also looking for a counterexample to the case where there are no conditions.

One way I am considering is Seifert-Van-Kampen. Then we only need that there exists some neighborhood $U$ of $a \in X \vee Y$ which has trivial fundamental group. Sufficient for that would be:

Both, $X$,$Y$ have a neighborhood of $x_0$, $y_0$ which (stronlgy) deformation retracts to that point.

I am not sure whether this is true for CW complexes, but it could very well be. It is true that CW complexes are locally contractible, but that is not enough:

First off, for a contractible neighborhood $U$ of $x_0$ we do not even have $\pi_1(U,x) = 0$ (see comments). Moreover, there are contractible pointed spaces (even such that strongly deformation retract to a point) s.t. their wedge sum is not contractible. An example can be found here. However, this space is not a CW complex, so I do still have hope that the theorem holds without any further assumptions.


1 Answer 1


Yes, any point $x$ in a CW complex has a neighborhood which deformations retracts onto $x$. This is part of Proposition A.4 in Hatcher's Algebraic Topology for example.

First off, for a contractible neighborhood $U$ of $x_0$ we do not even have $\pi_1(U,x) = 0$.

Actually we do...? A contractible space has trivial homotopy groups. Is the issue with base points? A contractible space is path-connected, and any path between two points can be used to define isomorphisms between the fundamental groups based at these two points. So if a (sub)space $U$ deformation retracts onto some $x_0$, then $\pi_1(U,x) = 0$ for any $x \in U$.

  • $\begingroup$ My bad, I had a wrong definition of a contractible. I thought it meant that $U \hookrightarrow X$ was nullhomotopic, not necessarily $U \rightarrow U$. $\endgroup$ Apr 10, 2020 at 15:22
  • $\begingroup$ @G.Chiusole I see. This property doesn't really have a name that I'm aware of, but I guess a space $X$ such that any point $x$ has a neighborhood $U$ with $U \to X$ nullhomotopic would be called "semi-locally contractible". $\endgroup$ Apr 10, 2020 at 15:23

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.