3
$\begingroup$

I am trying to construct the field of complex numbers as the quotient ring of real polynomials.

Suppose that

  1. $\mathbb C, \mathbb R$ are the fields of complex and real numbers respectively.

  2. $\mathbb R [X]$ is the ring of polynomials over $\mathbb R$.

  3. $\left \langle X^{2} + 1 \right \rangle = \left \{p(X^{2}+1) \mid p \in \mathbb R [X] \right \}$ is the ideal generated by $(X^{2}+1)$.

  4. $D = \mathbb R [X] / \left \langle X^{2} + 1 \right \rangle$ is the quotient ring of $\mathbb R [X]$ modulo $\left \langle X^{2} + 1 \right \rangle$.

Then $$(\mathbb C, +, \cdot) \cong (D, +, \cdot)$$

Could you please verify if my attempt contains logical gaps/errors? Any suggestion is greatly appreciated.


My attempt:

Lemma (Long Division of Polynomial): Let $K$ be a field and $p, q \in K[X]$ with $q \neq 0 .$ Then there are unique polynomials $r, s \in K[X]$ such that $$p=s q+r \quad \text {and} \quad \operatorname{deg}(r)<\operatorname{deg}(q) \tag 1$$

Proof:

  1. Existence

Define $\mathcal Q: K[X] \times (K[X] - \{0\}) \to K[X]$ by $$\mathcal Q(p,q) = \begin{cases} \bar p / \bar q X^{\operatorname{deg} (p)-\operatorname{deg} (q)} & \text{if } \operatorname{deg} (p) \ge \operatorname{deg} (q) \\ 0 & \text{otherwise}\end{cases}$$

where $\bar p, \bar q$ are the coefficients corresponding to $\operatorname{deg} (p), \operatorname{deg} (q)$ respectively.

Define $\langle p_n, s_n \rangle_{n \in \mathbb N}$ recursively by

$$\begin{aligned}\langle p_0, s_0 \rangle &= \langle p, \mathcal Q(p,q) \rangle \\ \langle p_{n+1}, s_{n+1} \rangle &= \langle p_n - s_n q, \mathcal Q (p_{n+1},q) \rangle \end{aligned}$$

Let $n' = \min \{ n \in \mathbb N \mid s_n = 0\}$. It is easy to verify that $s = \sum_{i = 0}^{n'} s_n$ and $r = p_{n'}$ satisfy $(1)$.

  1. Uniqueness

Suppose that $s’$ and $r’$ are other polynomials such that $p=s’ q+r’$ and $\operatorname{deg} (r’) < \operatorname{deg}(q)$. Then $(s’-s) q=r-r’$. If $s’-s \neq 0$ then, from $\operatorname{deg} (p q) = \operatorname{deg}(p) + \operatorname{deg}(q)$, we would get

$$\operatorname{deg} (r-r’) = \operatorname{deg} ((s’-s) q) = \operatorname{deg} (s’-s) + \operatorname{deg}(q)>\operatorname{deg}(q)$$

which, because $\operatorname{deg} (r-r’) \leq \max \{\operatorname{deg} (r), \operatorname{deg} (r’)\} <\operatorname{deg} (q)$, is not possible. Thus $s’=s$ and also $r’=r$.

By lemma, each set in $D$ has at least one element of the form $a+bX$. Assume that both $a + b X$ and $a' +b' X$ belong to the same set in $D$. Then $(a + b X) \sim (a' + b' X)$ and thus $(a + b X) - (a' + b' X) = p (X^2 + 1)$ for some $p \in K[X]$. It follows that $a = a'$, $b = b'$, and $p = 0$. As such, each set in $D$ has exactly one element of the form $a+bX$.

Consider $$\phi : D \rightarrow \mathbb{C}, \quad [a+bX] \mapsto a+b i$$

Clearly, $\phi$ is surjective.

Assume $[a+bX], [a'+b'X] \in D$ such that $\phi ([a+bX]) = \phi ([a'+b'X])$. Then $a+b i = a'+b' i$, and thus $a=a'$ and $b = b'$. Hence $\phi$ is injective.

Next we show that $\phi$ is a homomorphism w.r.t $+$ and $\cdot$.

$$\begin{aligned} \phi([a+b X]+[a'+b' X]) &= \phi([(a+a')+(b+b')X ]) \\ &= (a+a')+(b+b') i \\ &= (a+b i)+(a'+b' i) \\ &= \phi([a+b X])+\phi([ a'+b' X])\end{aligned}$$

$$\begin{aligned}\phi([a+b X]\cdot[a'+b' X]) &=\phi([(a+b X)\cdot (a'+b' X)])\\ &=\phi ([aa'+(ab'+ba') X+bb' X^2]) \\ & = \phi ([aa'-bb'+ (ab'+ba') X+bb' (X^2 + 1)]) \\ &=\phi([aa'-bb'+ (ab'+ba') X] )\\ &=(aa'-bb')+ (ab'+ba') i\\ &=(a+b i) \cdot (a'+b' i) \\ &=\phi([a+b X]) \cdot \phi([a'+b' X]) \end{aligned}$$

This completes the proof.

$\endgroup$
3
$\begingroup$

Good job! I would just add to that a proof of the fact that $\phi\bigl([1]\bigr)=1$ (that's trivial, of course).

$\endgroup$
2
  • $\begingroup$ Thank you so much for your verification ;) $\endgroup$
    – Akira
    Jun 29 '19 at 9:15
  • $\begingroup$ I'm glad I could help. $\endgroup$ Jun 29 '19 at 9:35
0
$\begingroup$

Well, the quotient ring $D = {\Bbb R}[x]/\langle x^2+1\rangle$ is a field since $x^2+1$ is irreducible over $\Bbb R$.

Moreover, the quotient ring $D$ contains a zero of $x^2+1$, namely $\bar x = x+\langle x^2+1\rangle$. This zero satisfies $\bar x^2+1=0$. The other zero is $-\bar x$.

This shows that $D = {\Bbb R}[\bar x]$ (ring adjoint) and since $D$ is a field, $D={\Bbb R}(\bar x)$ (field adjoint).

The mapping

${\Bbb C}=\{a+ ib\mid a,b\in{\Bbb R}\}\rightarrow {\Bbb R}(\bar x) = \{a+b\bar x\mid a,b\in{\Bbb R}\}: a+bi \mapsto a+b\bar x$

is an isomorphism of fields.

$\endgroup$

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.