Let $g \in \mathbb{C}[x]$ and $\omega \in \mathbb{C}$. Then:

  1. $\overline{g(\omega)} = \overline{g}(\overline{\omega})$

  2. If $g \in \mathbb{R}[x]$, then $\overline{g} = g$.

  3. Let $g \in \mathbb{R}[x]$ and $g(w) = 0$. Then $g(\overline{w}) = 0$.

(2) is quite obvious, since $z \in \mathbb{R} \Rightarrow z = \overline{z}$. However, I don't understand (1) and (3).

$\omega$ is a variable $\in \mathbb{C}$. So we plug some value in $\omega$ and then we find its conjugate and finally we calculate the value $g(\overline{w})$, right?

In general, I want to prove that in $\mathbb{R}[x]$ every polynomial can be expanded in product of polynomials, where degree of each polynomial $\leq 2$.

I know I have to use lemmas (1), (2), (3) and Fundamental theorem of algebra.

There are 3 cases to prove:

  1. $\deg f \leq 2$. Trivial. Proven.
  2. $\deg f > 2$ and $f$ has root $a \in R$. Then by Bezout's theorem $f = (x - a)q(x)$, where $\deg q = \deg f - 1$. Then I can apply induction on q, until it has no more real roots. Then I should use lemmas (1), (2), (3) and Fundamental theorem of algebra to say something about complex roots, but I stuck here.
  3. $\deg f > 2$ and $f$ has no real roots, and again repeat steps from (2) case.

Can you help me to prove the (1) and (3) lemmas and finish the theorem about polynomial expansion?


$1.$ $\overline{a+b}=\overline{a}+\overline{b}$ and $\overline{ab}=\overline{a}\cdot\overline{b}$ for any two complex numbers $a$ and $b$. Therefore this possibility to split the complex conjugation exists for everything that can be composed from addition and multiplication of complex numbers, like the evaluation of a polynomial.

$3.$ follows from 1. and 2.: $$ 0=g(\omega) \Rightarrow 0=\overline 0 = \overline{g(\omega)}=\overline{g}(\overline\omega)=g(\overline\omega) $$

  • $\begingroup$ So, if $f \in \mathbb{R}[x]$ it has root $\omega$, then $\overline{\omega}$ is also a root, then let $q(x) : = f(x) /( (x - \omega)(x - \overline{\omega}) )$ and I have to repeat this cycle for $q(x)$? $\endgroup$ – False Promise Feb 13 '17 at 17:26
  • $\begingroup$ Yes, that's how it works. You have to show that $(x-\omega)(x-\overline\omega)$ is in $\mathbb{R}[x]$, first. $\endgroup$ – Reinhard Meier Feb 13 '17 at 17:30
  • $\begingroup$ $(x-\omega)(x-\overline\omega) = x^2 - x(\omega + \overline{\omega}) + \omega \cdot \overline{\omega}$. $\omega \cdot \overline{\omega} =|\omega|^2 \in \mathbb{R}$. $\omega + \overline{\omega} \in \mathbb{R}$. And since we are in the field $\mathbb{R} \Rightarrow x \in \mathbb{R}$, right? $\endgroup$ – False Promise Feb 13 '17 at 17:41
  • $\begingroup$ Right. And this sentence exists only because stackexchange complains about short answers. $\endgroup$ – Reinhard Meier Feb 13 '17 at 17:43

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.