1
$\begingroup$

I recently asked the following question (Galois elements determined by action on $n$-th roots of rationals?). Somehow I could not log back into that account, and instead have ended up with a new one! Many thanks for the helpful comments on that question; they confirmed what I had thought was the case but wasn't sure, as well as clarified and corrected it.

My motivation in asking the question was certain comments I had come across to the effect that no element of $Gal(\overline{\mathbb{Q} }/\mathbb{Q})$ other than the identity and complex conjugation could be named, i.e., fully characterized because, I assume, in order to do so one would have to specify its action on all algebraic numbers compatibly with all their algebraic relations.

In lieu of full characterization I wondered if it might be possible and useful to characterize collections of Galois elements by properties of their actions on certain distinguished collections of algebraic numbers, e.g., $\mathcal{R} = \{r^{1/n}: r \in \mathbb{Q}, n \in \mathbb{N} \}$. One could, e.g., ask which elements of $Gal(\overline{\mathbb{Q}}/\mathbb{Q})$ fix $\mathcal{R}$, and study their properties further.

Is that sensible, possible and useful?

$\endgroup$
1
  • 2
    $\begingroup$ There is already a name for Galois elements whose behavior has been very fruitful in number theory: Frobenius elements or conjugacy classes associated to prime ideals. They are not as down-to-earth as complex conjugation ("Frobenius at infinity"), but they are dense in Galois groups over $\mathbf Q$, which is very nice. $\endgroup$
    – KConrad
    Commented Apr 27, 2019 at 21:06

0

You must log in to answer this question.

Browse other questions tagged .