0
$\begingroup$

In my alg 2 class today my teacher explained properties of polynomials and how to find their roots. One thing he emphasized was that if you know one complex root, you instantly know the other, because the other is just the conjugate (assuming all real coefficients). This made sense in terms of quadratic equations, as when you solve them you can clearly see that you have -b plus the discriminant or minus the discriminant. But I am having trouble seeing why complex roots always come in pairs for higher degree polynomials I found a couple proofs online (which all say the same thing, pretty much) but do not really get them because they rely on properties of conjugates that I do not understand. In fact, we have barely even discussed conjugates aside from the fact that the conjugate of a+bi is a-bi. Apparently this also applies to irrational roots (if you know one, the conjugate is also a solution). And also this does not work for functions with coefficients that are not real. I have researched this and I still am really lost on why this is. Is this question simply too advanced for the type of math I am taking right now? Or is there an easier way to understand this concept (or the various properties of conjugates)? Thanks!

Here is also an example of a video with the proof: https://www.youtube.com/watch?v=XkWz76dxkkI

$\endgroup$
3
  • 2
    $\begingroup$ This is true only if the coefficients are real numbers. It is proved by just taking complex conjugate in the equation $p(x)=0$. $\endgroup$ Commented Oct 23 at 0:11
  • 3
    $\begingroup$ See Complex conjugate root theorem, and the multiple posts on this site regarding this theorem, such as the current $341$ results from searching complex conjugate root theorem. $\endgroup$ Commented Oct 23 at 0:21
  • 1
    $\begingroup$ Please always search for answers before posting questions. $\ \ $ $\endgroup$ Commented Oct 23 at 4:29

2 Answers 2

2
$\begingroup$

As noted, this is only true if the coefficients are real numbers. To prove it, it's enough to prove two lemmas: $\bar {z_1} + \bar {z_2} = \overline {z_1+z_2}$, and $\bar {z_1} \cdot \bar {z_2} = \overline {z_1 z_2}$. You can prove these facts by expressing $z_n$ in the form $x_n + i y_n$ and then doing the arithmetic by brute force.

Once you know these facts are true, you can prove (by induction on the degree of the polynomial, if necessary) that for any $z \in \Bbb C$ and any $p(x) \in \Bbb R[x]$ we have $p(\bar z) = \overline {p(z)}$, so $p(z) = 0 = \bar 0 = \overline {p(z)}=p(\bar z)$, which means that whenever $z$ is a root of $p(z)$, so too is $\bar z$.

To test your understanding, do you see why the proof fails if $p$ has one or more non-real coefficients?

$\endgroup$
1
  • 1
    $\begingroup$ Please strive not to post more (dupe) answers to dupes of FAQs. This is enforced site policy, see here. $\endgroup$ Commented Oct 23 at 4:25
1
$\begingroup$

Here's a solution that only uses some simple algebra:

Imagine we have some polynomial with real coefficients (the only way this works, as you mentioned):

$P(x) = \alpha_0 + \alpha_1x + \alpha_2x^2$

Here, $\alpha_0, \alpha_1,$ and $\alpha_2$ are all real numbers. Any quadratic equation has two roots, and we already know one is the complex root $a + bi$, with $b \neq 0$. Let's call the other root $c$, and try to find out what it is.

By factoring, we have:

$P(x) = (x-(a+bi))(x-c) = (x-a-bi)(x-c)$

Expanding,

$P(x) = x^2 - ax - bix -cx + ac + bci$

And grouping like terms:

$P(x) = x^2 - (a+bi+c)x + c(a+bi)$

Now it's more managable. First of all, we know $a + bi + c = \alpha_1$, which we defined to be a real number. This gives us some new info about $c$. We now know $c$ must at least include some term $-bi$ in order to cancel out the imaginary part $+bi$.

We still don't know the real part of $c$, so we can let $c = d - bi$, and try to find $d$. Using the last term of our equation: because $\alpha_0$ is real, $c(a+bi)$ must be real. We can plug in our new definition of $c$ and expand:

$\alpha_0 = c(a+bi) = (d-bi)(a+bi) = ad - abi +bdi-bi^2 = ad - abi + bdi + b^2$

Like before, we know that $\alpha_0$ is real, so the imaginary parts must cancel out. Using our variables, this means:

$-abi + bdi = 0 \Rightarrow bdi = adi$

Dividing by $bi$ on both sides (note: you can't divide by 0, hence why we set $b \neq 0$ at the start) yields $d = a$. We can finally plug this back in to $c$, the other root of our polynomial $P(x)$, to get:

$c = a - bi$

which is the conjugate of $a + bi$.

$\endgroup$
2
  • 1
    $\begingroup$ Welcome to Mathematics Stack Exchange. OP was asking about higher degree polynomials $\endgroup$ Commented Oct 23 at 0:50
  • $\begingroup$ ...though I didn't (down)vote $\endgroup$ Commented Oct 23 at 1:29

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.