On a question of G. Glauberman about a replacement theorem for finite \(p\)-groups (Q5946417): Difference between revisions

Let \(S\) be a finite \(p\)-group. Thompson's replacement theorem states that if \(B\) is an Abelian subgroup of \(S\) which does not normalize the Abelian subgroup \(A\) of maximum order that normalizes \(B\), then there exists another Abelian subgroup \(A^*\) of maximum order such that \(A^*\cap B>A\cap B\), \(A^*\leq AB\), \([A^*,A,A]=1\), and, if \(p\neq 2\), the exponent of \(A^*\) divides the exponent of \(A\). \textit{G. Glauberman} asked [in J. Algebra 196, No. 1, 301-338 (1997; Zbl 0917.20016)] if this result could be extended for \(p\neq 2\) by allowing \(B\) to be nilpotent of class 2 instead of necessarily being Abelian. In this paper, the author constructs examples giving a negative answer to the above question for all primes \(p\). On the other hand, he proves the following weaker version of Glauberman's conjecture for \(B\) of class 2, with the condition \(A^*\cap B>A\cap B\) replaced by \(|A^*\cap B|>|A\cap B|\).