1
$\begingroup$

$V$ is a finite euclidean vectorspace and $\sigma:V->V$ is a motion, this means that $d(\sigma(a_i),\sigma(a_j))=d(a_i,a_j)$ for an affine coordinate system $a_0,...,a_n$

I know the following two things: $\sigma^2=id_V$ and $A:=\{v\in V:\sigma(v)=v\}$

I already proved that A is an affine subspace, but I would like to know why A is non-empty.

My second question is, why $\sigma$ is identical to a reflection $\sigma_A(v):=\sigma_{V_A}(v-a)+a$ where $\sigma_{V_A}$ is an orthogonal reflection on $V_A$ and $a\in A$ is arbitrary.

$\endgroup$
2
$\begingroup$

I assume that $\sigma$ is an affine map, and $V$ is endowed by a scalar product.

In this case, $\sigma(x)=Mx+b$ for some $M:V\to V$ linear map and $b=\sigma(0)\in V$. Since $\sigma(b)=\sigma(\sigma(0))=0$, we have $Mb+b=0$, so the midpoint of $0$ and $b$, $b/2$ will be in $A$. (Check it!)

(Actually, over a field of characteristic $2$, we can have that $A$ is empty, for example by any translation..)

For the second question, reduce it to the case when $0\in A$, ie. $\sigma$ is actually linear. Then, using that it keeps the distances of a basis, you could prove that it keeps all distances (keeps norm), using the following identity: $$||x-y||^2=||x||^2+||y||^2-2\langle x,y\rangle .$$ (By hypothesis, and assuming $a_0=0$ we have that $||\phi(a_i)||=||a_i||$ and $||\phi(a_i-a_j)||=||a_i-a_j||$ for all $i,j$.)

So we also have that $\sigma$ keeps the scalar product: $\sigma\in O(V)$, that is, fixing an orthonormal basis, its matrix will be an orthonormal matrix, satisfying $\sigma^2=id$, hence having $\pm 1$ as eigenvalues, ...

$\endgroup$
  • 1
    $\begingroup$ Thank you very much for your answer, I already thought that affine geometry is not so interesting for most people (Recently I put on an other question relating to affine geomtry but no answer yet: math.stackexchange.com/questions/243807/centre-of-a-quadric) Anyway, I am not sure if I understand the last part. Instead of $d(\sigma(a_i),\sigma(a_j))=d(a_i,a_j)$ you say that $d(\sigma(v_i),\sigma(v_j))=d(v_i,v_j)$ is also true, so $||a_j-a_i||=||v_j-v_i||$ ? Where can I use the parallelogramm identity $||x+y||^2+||x-y||^2=2(||x||^2+||y||^2)$ $\endgroup$ – Montaigne Nov 26 '12 at 23:07
  • $\begingroup$ Sorry, it is not the parallelogramma identity I thought of, but similar, I corrected $\endgroup$ – Berci Nov 28 '12 at 10:52

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.