Cofactorial automorphism-invariance is not transitive

From Groupprops
Revision as of 00:14, 28 March 2016 by Vipul (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

This article gives the statement, and possibly proof, of a subgroup property (i.e., cofactorial automorphism-invariant subgroup) not satisfying a subgroup metaproperty (i.e., transitive subgroup property).
View all subgroup metaproperty dissatisfactions | View all subgroup metaproperty satisfactions|Get help on looking up metaproperty (dis)satisfactions for subgroup properties
Get more facts about cofactorial automorphism-invariant subgroup|Get more facts about transitive subgroup property|

Statement

It is possible to have groups HKG such that H is a cofactorial automorphism-invariant subgroup of K and K is a cofactorial automorphism-invariant subgroup of G, but H is not a cofactorial automorphism-invariant subgroup of G.

The key observation here is that the set of prime divisors of the order of K may be a proper subset of the set of prime divisors of the order of G. Shrinkage in the set of prime divisors is necessary to construct a counterexample.

Facts used

  1. p-automorphism-invariant not implies characteristic

Proof

By Fact (1), we can find a prime number p and p-groups HK such that H is p-automorphism-invariant in K but not characteristic in K. In particular, we can find an automorphism of K of order p relatively prime to p such that H is not invariant under that automorphism. Define G as the semidirect product of K by the cyclic group generated by that automorphism. Then:

  • H is cofactorial automorphism-invariant in K: p is the only prime dividing the order of K, and H is invariant under p-automorphisms.
  • K is cofactorial automorphism-invariant in G: In fact, K is a normal Sylow subgroup of G and hence a characteristic subgroup of G. So it is invariant under all automorphisms of G.
  • H is not cofactorial automorphism-invariant in G: It is not invariant under the conjugation operation by the generator of the cyclic p-group.

Concretely, using the minimal example provided in the proof of Fact (1), we obtain that for any odd prime, we can construct H,K,G to have orders p3,p6,2p6 respectively.