2
$\begingroup$

In the theory of group actions on posets one studies the action of a group $G$ on a poset $\mathcal{P}$ (via order-preserving permutations) partly through its derived action on the order complex of the poset, $\Delta(\mathcal{P})$, which is a simplicial complex constructed from linear chains in the poset, and associated linear representations on the (co)homology of (geometric realizations of) $\Delta(\mathcal{P})$. (See R. Stanley's Some aspects of groups acting on finite posets, 1982, for example.)

In arithmetic we have a basic poset: the collection of all number fields, which is not a finite poset, or even locally finite, but does have an initial element, $\mathbb{Q}$. Call this poset $\mathcal{P}(\mathbb{Q})$. We also have analogous "local" posets $\mathcal{P}(\mathbb{Q}_p)$, one for each rational prime $p$. Each of these posets is acted on by the absolute Galois group of its initial element, $G_\mathbb{Q}$ or $G_{\mathbb{Q}_p}$.

Have these posets and associated simplicial/topological/linear Galois actions been studied in some guise, and if so, where, or is there an elementary error I am making in thinking along these lines? I suspect the latter because order complexes of these kinds of posets seem such natural topological gadgets to attach to number fields and their localizations and study Galois actions on, and yet one doesn’t come across them. Of course, it is also possible that an expert sees immediately that there isn’t likely to be any interesting information to be extracted from looking at these rather unwieldy infinitary objects that isn’t already available with simpler means.

$\endgroup$
3
  • 1
    $\begingroup$ I am definitely not an expert in group actions on posets, but I suspect that nothing interest would come out when applying Galois to the poset of subfields. Every extension $L/Q$ (where $Q=\mathbb{Q}$ or $Q=\mathbb{Q}_p$) which is Galois would be fixed by the action, and more generally every extension $L/Q$ could only be sent to a conjugate $L'$. By the Galois correspondance, this is just a group acting on the poset of its subgroups, which doesn't say much on the group itself, I guess (it is here where I might be wrong). In a way, one is simply studying $G_Q$ by looking at its quotients. $\endgroup$ Commented Mar 25, 2018 at 10:05
  • $\begingroup$ @Filippo Alberto Edoardo, thanks for bringing out the implicit question: how much can be learnt about a group from studying its action on the poset of its subgroups, especially when the poset is topologized, thus bringing cohomological methods in play? $\endgroup$
    – user122285
    Commented Mar 27, 2018 at 23:01
  • $\begingroup$ @Filippo Alberto Edoardo, In the case of number fields we, of course, have more to play with since we have an infinite collection of intertwined posets reflecting local-global interactions. $\endgroup$
    – user122285
    Commented Mar 27, 2018 at 23:04

0

You must log in to answer this question.