7
$\begingroup$

We all know that all the irreducible polynomials in $\mathbb{C}[x]$ are linear and in $\mathbb{R}[x]$ they aren't more than 2 degree. However,in $\mathbb{Q}[x]$ we can have an irreducible polynomial in any degree.

So the question is: for any given $n\in\mathbb{Z}^+$ does there exist a field $K$ $(\mathbb{Q}\subseteq K\subseteq\mathbb{C})$ such that all irreducible polynomials in $K[x]$ have degree $\le n$, and there exists an irreducible polynomial of degree $n$?

$\endgroup$
9
  • 4
    $\begingroup$ I guess that for every number field $K$ and $n\ge 1$ there's a rational polynomial of degree $n$ that is irreducible over $K$. $\endgroup$
    – YCor
    Commented Oct 10, 2023 at 6:58
  • 1
    $\begingroup$ @AlekseiKulikov I think the OP means that all irreducible polynomials have degree $\leq n$ and there exists an irreducible polynomial of degree exactly $n$. (I interpret "over $n$ degree" as "of degree $>n$".) $\endgroup$ Commented Oct 10, 2023 at 8:07
  • 7
    $\begingroup$ As YCor writes, all number fields $K$ have irreducible polynomials of all degrees $n$. For example, fix a prime $\mathfrak p$ of $\mathcal O_K$, and take a monic polynomial in $\mathcal O_K[x]$ of degree $n$ whose reduction modulo $\mathfrak p$ is irreducible over the finite field $\mathcal O_K/\mathfrak p$. You can even make the polynomial to have integer coefficients. $\endgroup$ Commented Oct 10, 2023 at 8:08
  • 13
    $\begingroup$ But actually, even for arbitrary fields, the property in the question can only hold for algebraically closed fields ($n=1$) or real-closed fields ($n=2$), not for any larger $n$. First, if $K$ is not perfect, it has irreducible polynomials of arbitrarily large degrees of the form $x^{p^k}-a$. If $K$ is perfect, then all finite-degree extensions are simple by Artin’s primitive element theorem, hence they have degree at most $n$. Thus, the algebraic closure of $K$ has degree $n$ over $K$, which means it is either algebraically closed or real-closed by the Artin–Schreier theorem. $\endgroup$ Commented Oct 10, 2023 at 8:21
  • 4
    $\begingroup$ I think you meant to say "a field $K$ with $\mathbb{Q}\subseteq K\subseteq\mathbb{C}$ " rather than referring to a number field (which has the specific meaning of a finite extension of $\mathbb Q$ and thus does not include $\mathbb R$ and $\mathbb C$) - is this correct? IIn any case Emil Jeřábek has answered your question. $\endgroup$
    – Will Sawin
    Commented Oct 10, 2023 at 19:29

1 Answer 1

20
$\begingroup$

Any number field $K$ (i.e., a finite extension of $\mathbb Q$) has irreducible polynomials of all degrees $n\ge1$. To see this, fix a prime ideal $\mathfrak p$ of the ring of integers $\mathcal O_K$. Since $\mathcal O_K/\mathfrak p$ is a finite field, there exists a monic irreducible polynomial $\overline f\in(\mathcal O_K/\mathfrak p)[x]$ of degree $n$. It lifts to a monic polynomial $f\in\mathcal O_K[x]$, which is also irreducible, as any splitting of $f$ reduces modulo $\mathfrak p$ to a splitting of $\overline f$. By Gauss’s lemma (which works for monic polynomials over any integrally closed domain), $f$ remains irreducible in $K[x]$ as well.

If $K$ is an arbitrary field, then there exists a maximal degree $n$ of irreducible polynomials over $K$ if and only if $K$ is algebraically closed (in which case $n=1$) or real-closed ($n=2$). Indeed, assume the maximal degree is $n$. Then $K$ is perfect: if $K$ is an imperfect field of characteristic $p$, there is an element $a\in K$ without a $p$th root, and then the polynomials $x^{p^k}-a$ are irreducible for all $k$. By Artin’s primitive element theorem, it follows that every finite extension of $K$ is simple, hence it is of degree at most $n$. Thus, if $L$ is an extension of $K$ of degree $n$ (which exists as there is an irreducible polynomial of degree $n$), then $L$ has no proper finite extension; i.e., $L$ is the algebraic closure of $K$. By the Artin–Schreier theorem, the algebraic closure of $K$ is a finite extension of $K$ only if $K$ is algebraically closed or real-closed.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .