Let $R_n = F_q[x]/\langle x^n - 1 \rangle$, where $F_q[x]$ is a finite field.

Consider $\mu_a$ which acts on $R_n$ like so; $f(x) \mu_a \equiv f(x^a) \bmod (x^n - 1)$ for $f(x) \in R_n$.

Is this an automorphism?

I'm not too sure where to start here. I know polynomial substitution is a homomorphism, so can I instantly deduce that this is a homomorphism?

If I show it's injective, then I can deduce that it is surjective since $R_n$ is finite, but I'm not sure where to start with injectivity?

Any help would be appreciated!


The lifted function $x\mapsto x^a$ on $F_q[x] $ is indeed a homomorphism by the universal property of polynomial rings, however this does not automatically imply that it is a homomorphism on $R_n$. You need to show that $\mu_a(x^n-1)=p(x)(x^n-1)$ for some $p$. This is true because $$x^{an} - 1=(x^n-1)(x^{n(a-1)}+x^{n(a-2)}+\cdots+x^n+1)$$ Thus it is a homomorphism on $R_n$. To show it is an automorphism, you then need to prove the kernel is trivial.

Now, $f(x)$ is a multiple of $x^n-1$ if and only if for all $n$th roots of unity $\zeta$ we have $f(\zeta)=0$. If $f(x^a)$ is a multiple of $x^n-1$, then $f(\zeta^a) = 0$ for all $n$th roots of unity. Since $a$ is relatively prime to $n$, every $n$th root of unity occurs as $\zeta^a$ for some other $n$th root of unity. Thus $f(\zeta)=0$ for all $n$th roots of unity $\zeta$, hence $f(x)$ is a multiple of $x^n-1$. Thus the kernel is trivial.

Note that this must be taken over all $n$th roots of unity, including ones that are not contained in $F_q$. We are essentially working in the algebraic closure.

  • $\begingroup$ Why does the fact that $\mu_a (x^n - 1) = p(x)(x^n - 1)$ show it is a homomorphism on $R_n$? $\endgroup$ – the man Feb 10 at 14:39
  • $\begingroup$ @theman You need to show that for the equivalence class $f(x) +\langle x^n-1\rangle$ we have that $\mu_a$ applied to this is $f(x^a) +\langle x^n-1\rangle$. Thus for an element $g(x) (x^n-1)$ of the ideal, we need that $g(x^a) (x^{an} - 1)$ is in the ideal. This is true if and only if it's true for $g(x) =1$. $\endgroup$ – Matt Samuel Feb 10 at 14:55
  • $\begingroup$ Ah ok ,thank you very much. $\endgroup$ – the man Feb 10 at 15:08
  • $\begingroup$ @theman You're welcome. $\endgroup$ – Matt Samuel Feb 10 at 15:12
  • $\begingroup$ For the kernel, I forgot to say that $gcd(a,n) = 1$. I'm trying to find for which $f(x)$, $f(x^a) \in \langle x^n - 1 \rangle$ right? I'm not sure where to continue from here? $\endgroup$ – the man Feb 10 at 15:13

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.