2
$\begingroup$

Let $S$ be a finite set of primes and $G_{\mathbb{Q},S}$ the Galois group of maximal extension of $\mathbb{Q}$ which is unramified ouside $S$. Since $\mathbb{Q}$ has no unramified extensions, $G_{\mathbb{Q},S}$ is topologically generated by inertia groups at primes in $S$, i.e. the subgroup generated by inertia groups at primes in $S$ is dense in $G_{\mathbb{Q},S}$. It leads to the following question:

Is the union of conjugates of the inertia groups at primes in $S$ dense in $G_{\mathbb{Q},S}$?

Any comments and suggestions will be appreciated.

$\endgroup$
2
  • $\begingroup$ I don't know if it's an issue of English, but I don't understand the meaning of the question. Is the question whether the union of all conjugates of all inertia groups dense? $\endgroup$
    – YCor
    Commented Oct 14, 2022 at 14:51
  • $\begingroup$ @YCor Yes, I ask whether the union of all conjugates of all inertia groups at primes in $S$ in $G_{\mathbb{Q},S}$ is dense. $\endgroup$
    – Nobody
    Commented Oct 14, 2022 at 14:56

2 Answers 2

7
$\begingroup$

Not unless $S$ has one element.

If $S$ contains two primes, $p$ and $q$, the Galois group of the cyclotomic field $\mathbb Q(\mu_{p^\infty q^\infty})$ is a quotient of $G_{\mathbb Q,S}$ and is isomorphic to $\mathbb Z_p^\times \times \mathbb Z_q^\times$. The images of the inertia groups at $p$, $q$, and any other prime are $\mathbb Z_p^\times \times \{1\}, \{1\}\times \mathbb Z_q^\times$, and $\{(1,1)\}$ respectively, and the union of these is not dense.

$\endgroup$
2
  • 1
    $\begingroup$ Thanks so much for your counterexample! But why is it true when $S$ has only one element? $\endgroup$
    – Nobody
    Commented Oct 14, 2022 at 14:19
  • 3
    $\begingroup$ @Nobody I don't think I said that it was! It's not true for $S =\{253381\}$ since the extension generated by the roots of $x^5+ x+3$ is ramified only at that prime (since its discriminant is $253381$) and the inertia group consists of an element of order $2$ in the Galois group and the trivial element, but not every element in the Galois group has order $1$ or $2$. $\endgroup$
    – Will Sawin
    Commented Oct 14, 2022 at 15:19
5
$\begingroup$

Not if $S$ has one element.

If $p$ is an irregular prime and $H$ is the Hilbert class field of $\mathbf{Q}(\zeta_p)$ then $\mathrm{Gal}(H/\mathbf{Q})$ has order divisible by $p$ but no inertial element has order $p$.

If $p$ is a regular prime, then $p$ does not divide the class number of $\mathbf{Q}(\zeta_{p^n})$ for any $n$. A theorem of Washington implies that the $\ell$-part of the class group of this tower is uniformly bounded for every $\ell \ne p$. From the class number formula the order of the class groups also become arbitrarily large as $n$ increases. Taken together this implies that the class number of $\mathbf{Q}(\zeta_{p^n})$ is divisible by at least one prime $\ell$ with $(\ell,p(p-1)) = 1$ for sufficiently larger $n$. If $H$ is the Hilbert class field of $\mathbf{Q}(\zeta_{p^n})$ for such an $n$, then $\mathrm{Gal}(H/\mathbf{Q})$ will have order divisible by $\ell$ but no inertial element can have order $\ell$.

Since the prime above $p$ in $\mathbf{Q}(\zeta_{p^n})$ is principal, it (the prime above $p$) splits completely in the Hilbert class field. So in these cases the inertia group above $p$ coincides with the decomposition group. Hnce these examples also demonstrate the stronger claim that not even the union of the decomposition groups above $p$ are dense.

$\endgroup$

You must log in to answer this question.

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