I am attaching the problem statement and proof of Brouwer Fixed Point Theorem (from Algebraic Topology by Hatcher). I am a newbie in Algebraic Topology.

enter image description here

And here is the proof.

enter image description here

I am confused on how $\pi_1(S^1)$ is zero, if $r$ is such a map as defined in the proof. Because clearly, $\pi_1(S^1)$ contains atleast one homotopy class which is defined as $f_t(s)$. Anyway retraction r is nothing but identity map, also $S_1 \subset D^2$. Hence all $rf_t(s)$ is nothing but $f_t(s)$. And this is a valid homotopy and should be a member of $\pi_1(S^1)$. But how the proof claims that $\pi_1(S^1)$ is zero if all the arguments in the proof are true (hence the contradiction). What I am getting wrong here.

  • $\begingroup$ "Anyway retraction r is nothing but identity map"... The retraction is $D^2 \mapsto S^1$ is not the identity map. The domain and range of an identity map must be equal. $\endgroup$ – Lee Mosher May 27 '17 at 14:44
  • $\begingroup$ en.wikipedia.org/wiki/Retract this says it is identity @LeeMosher $\endgroup$ – user3001408 May 27 '17 at 14:49
  • $\begingroup$ No, it says given a space $X$ and a subset $A$, a function $f : X \to A$ is a retraction if the restriction of $f$ to $A$ is the identify function on $A$, in other words the function $f \bigm| A$ is equal to the identify function $\text{Id}_A$ on the set $A$. The definition of a retraction applies to many important situations where $A$ is a proper subset of $X$, for example in your question where $X=D^2$ and $A=S^1$, and in such a situation the restricted function $f \bigm| A = \text{Id}_A$ is not the same as the given function $f : X \to A$. $\endgroup$ – Lee Mosher May 27 '17 at 15:24

You have a composition of maps $$ S^1 \stackrel{\iota}{\longrightarrow} D^2 \stackrel{r}{\longrightarrow} S^1 $$ where $\iota$ is the inclusion of $S^1$ into $D^2$. This composition is the identity map on $S^1$. So, it should give the identity map on $\pi_1(S^1)$. Since $\pi_1(D^2) = 0$, this is impossible.


The idea here can perhaps be made clearer by seperating the proof in to three steps:

step 1: $\pi_1(S^1)=\mathbb Z$

I will omit the proof for this statement.

step 2: no retract theorem

There is no map $r:D^2 \to S^1$ so that $r \circ i=id$, where $i$ is inclusion, and $id$ is identity. Suppose that there were. Then the induced morphisms on the fundamental group would be $i_*:\mathbb Z \to \pi_1 (D^2)$ and $r_*:\pi_1(D^2) \to \mathbb Z$ with the additional property that $r_* \circ i_*: \mathbb Z \to \mathbb Z$ should be identity, but $\pi_1(D^2)=0$, so this composition is impossible. [For clarity's sake, I replaced $\pi_1(S^1)$ with $\mathbb Z$ in the previous statement]

step 3: fixed point theorem:

Using the map defined by hatcher, notice that if there were no fixed point, the map $r$ would be a retract, which is a contradiction.


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.