Given the nature of the exercise, I think the point is not using the Fundamental Theorem of Algebra to show there must be $n$ roots. Anyway, the polynomial is the following: $$x^n + x + 1$$

So far I've tried to reach an absurd by proposing a root that is both root of the polynomial and its derivative, but I'm stuck there. Any hints for this kind of exercise?

  • $\begingroup$ Yes, you should consider the case that the polynomial has a common root with its derivative. That means that the polynomial and its derivative have a common divisor. You can find the greatest common divisor (gcd) using the Euclidean Algorithm. That is a hint, with that you will be able to solve the problem. $\endgroup$ Nov 19, 2017 at 3:01
  • $\begingroup$ Im entering a loop here. I get (x^n +x+1:nx^(n-1) + 1) = (nx^(n-1) + 1:(n-1)x/n + 1) and when I start dividing this I get stuck in a loop $\endgroup$ Nov 19, 2017 at 3:07

1 Answer 1


We may safely assume $n\geq 2$. If $f(x)=x^n+x+1$ we have $f'(x)=n x^{n-1}+1$ and $$ \gcd(f(x),f'(x)) = \gcd(f'(x), n f(x)-x f'(x))=\gcd(nx^{n-1}+1,(n-1)x+n). $$ On the other hand $(n-1)x+n$ only vanishes at $x=\frac{n}{1-n}$, while $$ n\left(\frac{n}{1-n}\right)^{n-1}+1 $$ is positive for any odd $n$ and strictly negative for any even $n$, since $n^n>(n-1)^{n-1}$.
It follows that $f(x)$ and $f'(x)$ cannot have common roots, hence all the roots of $f(x)$ are simple.

  • $\begingroup$ Why this is true? $$ \gcd(f(x),f'(x)) = \gcd(f'(x), n f(x)-x f'(x)) $$ $\endgroup$
    – IrbidMath
    Nov 19, 2017 at 4:33
  • 1
    $\begingroup$ @Ameryr: any polynomial which simultaneously divides both $f(x)$ and $f'(x)$ also simultaneously divides both $f'(x)$ and $nf(x)-x\,f'(x)$. In particular the greatest-common-divisor of $f(x)$ and $f'(x)$ is also the greatest-common-divisor of $f'(x)$ and $n f(x)-x f'(x)$. It is just the main principle of the Euclidean algorithm. $\endgroup$ Nov 19, 2017 at 4:37
  • $\begingroup$ We can prove it by induction on the degree i guess? $\endgroup$
    – IrbidMath
    Nov 19, 2017 at 4:40
  • 1
    $\begingroup$ @Ameryr: no induction is needed. In a Euclidean ring, if $d$ is a divisor of both $a$ and $b$, $d$ is a divisor of $a\cdot\text{whatever}\pm b\cdot\text{whatever}$. $\endgroup$ Nov 19, 2017 at 4:44
  • $\begingroup$ Oh. Yeah thats true any linear combination $\endgroup$
    – IrbidMath
    Nov 19, 2017 at 4:47

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.