I been trying to find a ring homomorphism but without luck, for example in the complex numbers we have the conjugation but this not work in the quaternions, since $f(xy) = f(y)f(x)$ supposing $f$, is the conjugation, and we don't have commutativity in the quaternions.

So, is there some ring homomorphism distinct from the identity or the zero ones?

  • $\begingroup$ A homomorphism from where to where? For one from $\Bbb H$ to itself, how about $a+bi+cj+dk\mapsto a+bj+ck+di$? $\endgroup$ – Lord Shark the Unknown Sep 21 '17 at 1:59

If you mean a ring homomorphism from $\mathbb H$ to itself, you can take any noncentral element $x$ and use the ring isomorphism $q\mapsto xqx^{-1}$.

Actually, it can be proven that these (with the identity automorphism) give you all possible automorphisms.

To show that, we can refer to the Skolem-Noether theorem on central simple algebras. Since $\mathbb H$ is a central simple $\mathbb R$ algebra, we know already that all $\mathbb R$-linear automorphisms are inner.

But if $\phi$ is any automorphism $\mathbb H\to \mathbb H$, then it's easy to see that $r$ is central in $\mathbb H$ iff $\phi(r)$ is central in $\mathbb H$, so that $\phi$ restricted to $\mathbb R$ is an automorphism of $\mathbb R$. But the only automorphism of $\mathbb R$ is the identity, so $\phi$ is already $\mathbb R$-linear. Thus all automorphisms of $\mathbb H$ have been found.

Distinct $x$'s are not guaranteed to yield distinct maps. Since $G=\mathbb H\setminus\{0\}$ is a group, we can refer to the basic group theory result about the conjugation action of $G$ on itself. It's easy to prove that the kernel of the action is the center of $G$ (which corresponds to $\mathbb R\setminus\{0\}$ in our $G$) so that $x$ and $y$ produce the same inner automorphism on $\mathbb H$ iff $x=ry$ for some $r\in\mathbb R$.

  • 2
    $\begingroup$ Doesn’t that get them all? $\endgroup$ – Lubin Sep 21 '17 at 2:48
  • $\begingroup$ That work but why takes them all? $\endgroup$ – Hector Gurle Sep 21 '17 at 3:19
  • 3
    $\begingroup$ @EdgarM Lubin is right: the Skolem-Noether theorem says all automorphisms of a central simple algebra are inner, so this (along with the identity) catches all automorphisms of $\mathbb H$ $\endgroup$ – rschwieb Sep 21 '17 at 10:39
  • 3
    $\begingroup$ @Kimball But if $\phi$ is any automorphism of $\mathbb H$, it's easy to see that $r$ is central iff $\phi(r)$ is central, and so $\phi$ restricted to $\mathbb R$ is an automorphism of $\mathbb R$. But the identity is the only automorphism of $\mathbb R$, so in fact $\phi$ is $\mathbb R$ linear already. Only possible because of this nice property of $\mathbb R$, of course. $\endgroup$ – rschwieb Sep 22 '17 at 0:24
  • 1
    $\begingroup$ While we’re on the topic, perhaps it should be pointed that $\Bbb Q_p$ has this property too: the only automorphism is the identity. $\endgroup$ – Lubin Sep 22 '17 at 13:00

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.