1
$\begingroup$

I am trying to prove the fact that every finite field is perfect. Hence, every irreducible polynomial is separable (does not have a repeated root).

This is easy to prove when in a field of characteristic $p $, the irreducible polynomial under consideration has at least one element $ax^b $ such that $p\not | b $. We just differentiate, and prove that the two cannot have any factors in common.

We now take polynomials of the form $\Sigma a_n x^{kp} $ for some whole number $k $. For an example, I took the field to be $Z_3$ and the irreducible polynomial to be $x^6+x^3-1$. I reasoned that if it has a repeated root, then in the polynomial ring $\frac {Z_3}{(x^6+x^3-1)}[t] $, the polynomial $t^6+t^3-1$ will be divisible by $t-x $, And the quotient too will be divisible by $t-x $. However, on dividing $t^6+t^3-1$ by $t-x $, I am getting $x^6$ as a remainder. This is not equal to zero, as we had earlier assumed $x $ satisfies the irreducible polynomial $x^6+x^3-1$. Where am I going wrong?

$\endgroup$
  • $\begingroup$ In the ring $\Bbb{Z}_3[x]$ you have $$x^6+x^3-1=(x^2+x-1)^3,$$ so it cannot be irreducible. $\endgroup$ – Jyrki Lahtonen Aug 11 '15 at 21:50
2
$\begingroup$

The punchline is the following:

Polynomials of the form $\sum a_n x^{np}$ are never irreducible, since we have $$\sum a_n x^{np} = \left(\sum \sqrt[p]{a_n} x^{n}\right)^p.$$

For $\sqrt[p]{a_n}$ to make sense, we need the Frobenius to be surjective, which is the case in a finite field.

All in all we are left with the case you describe in your first lines. So the proof is done.

$\endgroup$

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.