Let $\mathbb Z_n$ be a cyclic group of order $n$ and for each integer $a$ define

$$\sigma_a : \mathbb Z_n \mapsto \mathbb Z_n, \qquad \sigma_a(x) = x^a, \quad\forall \; x \in \mathbb Z_n.$$

Prove that $\sigma_a \circ \sigma_b = \sigma_{ab}$. Deduce that the map $\bar{a} \mapsto \sigma_a$ is an isomorphism of ($\mathbb{Z} / n \mathbb{Z})^{\times}$ onto the automorphism group of $\mathbb Z_n$ (so $Aut(\mathbb Z_n)$ is an abelian group of order $\phi(n)$).

The question is section 2.3 question 26 part d page 75 of Dummitt and Foote's Abstract Algebra.

Understand the other parts, but this one has me a bit baffled.


Define a mapping $\Phi : (\mathbb{Z}/n\mathbb{Z})^\times \rightarrow \mathsf{Aut}\ Z_n$ by $\Phi(\overline{a}) = \sigma_a$. Since, $\sigma_a = \sigma_b$ if and only if $a \equiv b$ $mod$ $n$, $\Phi$ is well defined.

Moreover, letting $x \in Z_n$ be arbitrary, we have $(\sigma_a \circ \sigma_b)(x) = \sigma_a(\sigma_b(x)) = \sigma_a(x^b) = (x^b)^a = x^{ab} = \sigma_{ab}(x)$.

Thus, $\sigma_a \circ \sigma_b = \sigma_{ab}$.

In other words, $\Phi(\overline{a} \overline{b}) = \Phi(\overline{a}) \Phi(\overline{b})$.

Thus $\Phi$ is a homomorphism.

Finally, since every automorphism of $Z_n$ is equal to $\sigma_a$ for some $a$, $\phi$ is surjective. Since $\mathsf{Aut}\ Z_n$ and $(\mathbb{Z}/n\mathbb{Z})^\times$ are finite $\phi$ is bijective. Thus $\Phi$ is an isomorphism.


If you understand the other parts of this problem, then you're practically finished. Indeed, in parts (a) - (c) you've essentially already shown that

$\sigma \in \text{Aut}(Z_n) \iff \sigma = \sigma_a$ for some $a$ coprime to $n$

(Note this is $\iff$ not just $\implies$ by combining (a) and (c).)

and also

$\sigma_a = \sigma_b \iff a\equiv b \pmod{n}$

Combining these two statements you can conclude (how?) that $$ \text{Aut}(Z_n) = \{\sigma_a : 1 \leq a \leq n-1, \text{gcd}(a,n) = 1\}. $$ Accordingly, there is a bijection $$ \Phi: (\mathbb{Z}/n\mathbb{Z})^\times \to \text{Aut}(Z_n) $$ given by $\Phi(\overline{a}) = \sigma_a$ for all $a$ coprime to $n$, where for $a \in \mathbb{Z}$ we denote by $\overline{a}$ the equivalence class of $a$ modulo $n$. (This is well-defined by the second property above.) Since the domain of $(\mathbb{Z}/n\mathbb{Z})^\times$ is a multiplicative group, it's natural to ask whether or not $\Phi$ is a homomorphism (and hence an isomorphism). This can be stated more succinctly as checking whether or not $\Phi(\overline{a}\overline{b}) = \Phi(\overline{a})\circ\Phi(\overline{b})$, or (equivalently) $\sigma_{ab} = \sigma_a\circ\sigma_b$ (where composition of functions $\circ$ is used on the right since it's the natural group operation on $\text{Aut}(Z_n)$.

So you need to determine if for each $x \in Z_n$ it's true that $\sigma_{ab}(x) = \sigma_a(\sigma_b(x))$. Do you see why this is true given the definition of $\sigma_c$?

This is (intentionally) more of a guide and some perspective than a proof. Try to work through the result. I'm happy to answer additional questions.


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.