7
$\begingroup$

Let $f \colon S^n \to S^n$ be a constant map, $n > 0$. I want to show that $\deg f = 0$. I will do it by definition. Let $\sum_k g_k \sigma^n_k$ be a singular chain, $g_k \in \mathbb{Z}$, $\sigma^n_k \colon \Delta^n \to S^n$, where $\Delta^n$ is a standard $n$-simplex. Then the induced map $f_n \colon C_n(S^n) \to C_n(S^n)$ acts by the rule $$ f_n \left( \sum_k g_k \sigma^n_k \right) = \left( \sum_k g_k \right) \sigma^n_0, $$ where $\sigma^n_0 = f \colon \Delta^n \to pt$. Let's notice that $\partial \sigma^n_0 = 0$ if $n$ is odd and $\partial \sigma^n_0 = \sigma^{n-1}_0$ if $n$ is even. Hence if $n$ is even the image of cycles group under the action of $f_n$ is trivial and hence the induced homomorphism $f_n \colon H_n(S^n) \to H_n(S^n)$ in homology groups is trivial. But how to show that if $n$ is odd then the induced homomorphism $f_n$ in homology groups will also be trivial?

$\endgroup$
8
  • $\begingroup$ I think that if $n$ is odd then $\partial\sigma^{n+1}_0=\sigma^n_0$, so $\sigma^n_0$ is a boundary. So it is the other way round. $\endgroup$ – Stefan Hamcke Apr 2 '13 at 21:16
  • $\begingroup$ @StefanH. but there are no $(n+1)$-dimensional nontrivial chains, so all $n$-dimensional boundaries are all trivial, isn't it true? Ah, I speak about simplicial chains, not singular. Singular chains may be nontrivial for larger dimensions than $n$, right? Thank you! $\endgroup$ – Appliqué Apr 2 '13 at 21:19
  • $\begingroup$ Dont't you want to show that $\sigma_0^n$ is a boundary? Then it is trivial in the homology group. $\endgroup$ – Stefan Hamcke Apr 2 '13 at 21:22
  • $\begingroup$ @StefanH. It is clear that $\sigma^n_0$ is a boundary of $\sigma^{n+1}_0$ because $\partial \sigma^{n+1}_0$ is a sum of $n+1$ copies of $\sigma^n_0$ with alternating signs starting from $+$. The problem arised because I thought about simplicial chains when I was thiking about $n$-dimensional boundaries. Now I solved it. $\endgroup$ – Appliqué Apr 2 '13 at 21:23
  • 2
    $\begingroup$ You can see this very easily by factoring $f$ through the inclusion $S^n - \{ x \} \hookrightarrow S^n$, where $x$ is any point not in the image of $f$: the homomorphism $f_*$ factors through $H_n(S^n - \{x\})$ but $S^n - \{x\}$ is contractible. $\endgroup$ – user314 Apr 3 '13 at 4:51
10
$\begingroup$

You can say a little bit more, the following proposition is also proved in Hatcher's Algebraic Topology

Proposition. Let $f:S^n \to S^n$ be a continuous map, if $f$ is not surjective then deg$(f) = 0$

Proof. Choose a point $x_0$ not in the image of $f$, then you can factor $f$ with the inclusion of $S^n\setminus \{x_0\}$ which is homeomorphic to $\mathbb{R}^n$. Applying the functor $H_n(\text{_})$ gives us the required result.

$\endgroup$
1
  • $\begingroup$ This is the more "elegant" answer. It also works for any ordinary homology theory, not just singular homology. $\endgroup$ – polynomial_donut May 11 '17 at 13:25

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.