More subgroups of index 2
by math_explorer, Jan 4, 2013, 1:01 PM
Let
be a finite group,
an automorphism of
. Let
be the set of all
such that
.
(i) If
then prove that
is abelian.
(ii) If
then prove that
has an abelian subgroup of index 2.
(i)
First, we claim that if
then
commutes with
. Simply note that
and then invert both sides.
Now, fix an element
and consider the set
. As
, it's not hard to see that
(in detail, take the complement and use De Morgan's). From the above if
is in this set then
and
commute, hence
is a subset of the center of
. But the center of
is a subgroup of
, so its order has to be a divisor of
, but since its order is more than
it has to be
itself. That is,
commutes with all of
, i.e.
is in the center
of 
Thus
, and again since
is a subgroup of
with an order exceeding
, we have
and
is abelian.
(ii)
First,
cannot be abelian, because if it is then for
we have
and
. In that case it's clear
would be a subgroup, which is impossible as its order is not a divisor of
.
So, let
be an element not in the center of
but in
(such an element certainly exists as
). Similar to (i), consider
, which is as before a subset of
, and
this time. Of course,
is not the whole of
so it's a subgroup with order exactly
, and index
. Then
.
We claim that
is an abelian subgroup. In fact,
. Now for any
we have
, and we can invert both sides as before, hence
and
commute and we're done.
---
Darn group actions on sets are so annoying when the group and the set are the same, or subgroups of another group, or subgroups of subgroups of something else. Sylow qq






(i) If


(ii) If


(i)
First, we claim that if




Now, fix an element



















Thus






(ii)
First,






So, let












We claim that






---
Darn group actions on sets are so annoying when the group and the set are the same, or subgroups of another group, or subgroups of subgroups of something else. Sylow qq