Let $G$ be the group of invertible elements of the ring $\mathbb{Z}/8\mathbb{Z}$. We would like to determine the Dirichlet characters modulo 8. Namely we would like to determine Hom$(G, \mathbb{C}^\times)$.

We denote by $[n]$ the residue class modulo $8$ represented by an integer $n$. Then $G = \{1, [3], [5], [7]\}$. $[3]^2 = [5]^2 = [7]^2 = 1, [3][5] = [7], [5][7] = [3], [3][7] = [5]$. Hence $G$ is the direct product of $\{1, [3]\}$ and $\{1, [5]\}$. Let $\phi \in$ Hom$(G, \mathbb{C}^\times)$. Then $\phi([3]) = \pm 1, \phi([5]) = \pm 1$. We define $\phi_1, \phi_2, \phi_3 \in$ Hom$(G, \mathbb{C}^\times)$ as follows.

$\phi_1([3]) = -1, \phi_1([5]) = 1$.

$\phi_2([3]) = -1, \phi_2([5]) = -1$.

$\phi_3([3]) = 1, \phi_3([5]) = -1$.

Then Hom$(G, \mathbb{C}^\times) = \{1, \phi_1, \phi_2, \phi_3\}$.

Clearly $\phi_3 = \phi_1\phi_2$.

Is the following proposition true? If yes, how can we prove it?

Proposition Let $n$ be an odd integer. Then

$\phi_1([n]) = (-1)^{(n-1)/2}$

$\phi_2([n]) = (-1)^{(n^2 - 1)/8}$

  • $\begingroup$ I would like to point out the following policy of StackExchange because it doesn't seem to be well-known and some users seem to dislike a question to which the poster already knows the answer. [It’s also perfectly fine to ask and answer your own question, as long as you pretend you’re on Jeopardy! — phrase it in the form of a question. To be crystal clear, it is not merely OK to ask and answer your own question, it is explicitly encouraged.] blog.stackoverflow.com/2011/07/… $\endgroup$ – Makoto Kato Nov 7 '13 at 0:25

I mean, you can calculate $\text{Hom}(G,\mathbb{C}^\times)$ for any finite abelian group $G$. Namely, factor $G\cong C_{p_1^{e_1}}\times\cdots\times C_{p_m^{e_m}}$. Then you have an isomorphism

$$\text{Hom}(G,\mathbb{C}^\times)\cong \prod_i \text{Hom}(C_{p_i^{e_i}},\mathbb{C}^\times)$$

defined in the obvious way. Lastly, you have an isomorphism

$$\text{Hom}(C_{p_i^{e_i}},\mathbb{C}^\times)\xrightarrow{\approx}C_{p_i^{e_i}}:(1\mapsto e^{\frac{2\pi i \ell}{p_i^{e_i}}})\mapsto \ell$$

Thus, tracking these isomorphisms back, you get explicitly what the elements of $\text{Hom}(G,\mathbb{C}^\times)$ are.

In your case

$$(\mathbb{Z}/8\mathbb{Z})^\times\cong (\mathbb{Z}/2\mathbb{Z})^2$$

via the usual map. This allows you to explicitly construct all of the characters.

I leave it to you to check the details.

  • $\begingroup$ Dear Alex Youcis, I'm afraid you didn't prove the proposition in my question. $\endgroup$ – Makoto Kato Oct 28 '13 at 2:52

We define a map $f:\mathbb{Z} \rightarrow \mathbb{C}^\times$ by $f(n) = (-1)^{(n-1)/2}$. Let $n, m$ be integers such that $n \equiv m$ (mod $8$). Then $n - 1 \equiv m - 1$ (mod $8$). Hence $(n - 1)/2 \equiv (m - 1)/2$ (mod $4$). Hence $f(n) = f(m)$. Hence we can define a map $\psi_1:G \rightarrow \mathbb{C}^\times$ by $\psi_1([n]) = f(n)$. By Lemma 1 of my answer to this question, $f(nm) = f(n)f(m)$ for odd integers $n, m$. Hence $\psi_1([nm]) = \psi_1([n])\psi_1([m])$. Hence $\psi_1 \in$ Hom$(G, \mathbb{C}^\times)$. Clearly $\psi_1([3]) = -1, \psi_1([5]) = 1$. Hence $\phi_1 = \psi_1$.

We define a map $g:\mathbb{Z} \rightarrow \mathbb{C}^\times$ by $g(n) = (-1)^{(n^2-1)/8}$. Let $n, m$ be integers such that $n \equiv m$ (mod $8$). There exists an integer $k$ such that $n = m + 8k$. Since $n^2 = m^2 + 16mk + 64k^2$, $n^2 \equiv m^2$ (mod $16$). Hence $(n^2 - 1)/8 \equiv (m^2 - 1)/8$ (mod $2$). Hence $g(n) = g(m)$. Hence we can define a map $\psi_2:G \rightarrow \mathbb{C}^\times$ by $\psi_2([n]) = g(n)$. By Lemma 2 of my answer to this question, $g(nm) = g(n)g(m)$ for odd integers $n, m$ Hence $\psi_2([nm]) = \psi_2([n])\psi_2([m])$. Hence $\psi_2 \in$ Hom$(G, \mathbb{C}^\times)$. Clearly $\psi_2([3]) = -1, \psi_2([5]) = -1$. Hence $\phi_2 = \psi_2$.


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.