I am trying to prove that given two isometry on the plane that consist of rotations of the plane around distinct points contain a translation in their group (please see Groups containing two rotations.)

I have showed that they need to be in the kernel and everything but cannot show that $fg$ is not equal to $gf$.

enter image description here

I know that $a$ and $p(a)$ have the same length as orthogonal operators preserve lengths. It makes sense that they don’t equal but how do we know for sure? So basically is: $a+b’=b+a’$ where $a$ and $a’$ have the same length same goes for $b$. How to show this is not possible for non trivial vectors.


It's important to also assume that $f$ and $g$ are nontrivial rotations, or else it is indeed true that $fg = gf$.

So, given that $f$ is a nontrivial rotation about the point $p$ and $g$ is a nontrivial rotation about the point $q$, and $p \neq q$, why don't we check what $fg$ and $gf$ do to the points $p$ and $q$ themselves?

Since $g(q) = q$, we have $$ fg(q) = f(g(q)) = f(q) $$

But since $f$ is not centered at $q$, we have that $f(q) \neq q$. Therefore $$ gf(q) = g(f(q)) \neq f(q) $$ using the fact that $g$ is a nontrivial rotation, and hence $g(x) \neq x$ for all $x \neq q$. Thus, we conclude that $$ fg(q) = f(q) \neq gf(q) $$ $$ \implies fg \neq gf $$

  • $\begingroup$ Thank you, this was helpful $\endgroup$
    – Dhdh
    Dec 4 '19 at 20:54

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.