3
$\begingroup$

Show that $P=X^4+X+2$ is reducible over $\mathbb{F}_9$, but irreducible over $\mathbb{F}_{27}$.

I would appreciate any hints. (I know $P$ is irreducible over $\mathbb{F}_3$ by brute force, and that it factors completely in $\mathbb{F}_{81}$ by noting $\overline{X}, \overline{X^3}, \overline{X^9}, \overline{X^{27}}$ are distinct roots in the isomorphic field $\mathbb{F}_3[X]/(P)$ (is there a nicer way of showing this, other than looking for roots in this quotient?))

Thanks in advance.

$\endgroup$
  • 1
    $\begingroup$ Compare with this question. $\endgroup$ – Dietrich Burde Dec 12 '15 at 16:50
  • 2
    $\begingroup$ A hint: If $\alpha,\alpha^3,\alpha^9,\alpha^{27}$ are the zeros of $P$ in $\Bbb{F}_{81}$, then can you show that the coefficients of the factors $$P_1(x)=(x-\alpha)(x-\alpha^9)$$ and $$P_2(x)=(x-\alpha^3)(x-\alpha^{27})$$ are in $\Bbb{F}_9$. Therefore $P=P_1P_2$ in $\Bbb{F}_9[x]$. $\endgroup$ – Jyrki Lahtonen Dec 12 '15 at 17:02
2
$\begingroup$

The excelent hint of @Jyrki Lahtonen: takes care of the $\mathbb{F}_9$ case.

HINT: ( for $\mathbb{F}_{27}$)

The following result is easy to prove: Let $P$ an irreducible polynomial over $K$ and $K \subset L$ a Galois extension. Then the group $Gal(L/F)$ acts transitively on the irreducible $P_1$, $\ldots$ , $P_m$ factors of $P$ over $L$. In particular, $P$ decomposes over $L$ in $m$ irreducible polynomials of equal degree and $m \mid [L:K]$.

As a consequence: if $P(X)$ of degree $n$ is irreducible over $K$ and $L/K$ is Galois of degree relatively prime to $n$ then $P(X)$ remains irreducible over $L$.

${\bf Added:}$ In fact, the same result takes care of the $\mathbb{F}_9$ case. We know that $P$ decomposes (completely) over $\mathbb{F}_{81}$ into $4$ factors (of degree $1$). If $P$ were irreducible over $\mathbb{F}_9$ the Galois group $Gal({F}_{81}/ \mathbb{F}_9)$ ( of order $2$) could not act transitively on the factors. Now the group $Gal(\mathbb{F}_{9}/ \mathbb{F}_3)$ acts transitively on the factors of $P$ over $ \mathbb{F}_9$, and there are more than one, so there are two, of equal degree.

$\endgroup$
2
$\begingroup$

The facts that $\mathbb F_{q'}$ is a field extension of $ \mathbb F_{q}$ if and only if $q'$ is a power of $q$, that there is a 'unique' extension with given cardinality or degree, and therefore that all extensions are normal (irreducible poly with a root in the extension factors completely) make this problem, as asked, pretty straight-forward almost without calculation, though I get less than you did with calculation.

By calculation (!), $P$ has no roots over ${\mathbb F}_3$. Hence it cannot have an irreducible cubic factor, which means that the polynomial cannot factor over $\mathbb F_{27}$, as otherwise it would have to do completely.

Hence $P$ must factor completely over $\mathbb F_{81}$ because either a root of $P$ generates the 'unique' extension of degree 4 over $\mathbb F_3$, or $P$ factored over $\mathbb F_3$ as quadratic polynomials, and so factored completely in $\mathbb F_9$ (which you ruled you by calculation - but...)

Either way $P$ has to factor in $\mathbb F_9$, because we know that it factors completely in $\mathbb F_{81}$, which is the 'unique' extension of degree 2 over $\mathbb F_9$.

$\endgroup$
0
$\begingroup$

It has no root in base field. Therefore roots of $f$ cannot be in $\mathbb F_{27}$. Since it is a polynomial over $\mathbb F_3$, all roots are in $\mathbb F_{9}$ or $\mathbb F_{81}$. Therefore it is reducible over $\mathbb F_9$.

$\endgroup$

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.