Refining a normal series to get a composition series for a group.

119 Views Asked by At

Am trying to answer the following exercise in "A First Graduate Course in Abstract Algebra" by W. J. Wickless, page 25, Fifth Problem Set, Problem 1.7.3 :

"Prove that any normal series for a group $G$ can be refined, by adding terms, to produce a composition series for $G$."

Wickless defines a composition series as a finite normal series in which every successive factor group is a simple group.


This does not seem to hold for infinite groups.

Consider the additive cyclic group $\bf{Z}$ of integers, and let $p$ be a prime number. In this group, the series $${0}\triangleleft{p{\bf{Z}}}\triangleleft{\bf{Z}}$$ is a normal series, but I cannot refine this to a composition series with a finite number of terms because the series $$\cdots\triangleleft{p^n{\bf{Z}}}\triangleleft{p^{n-1}{\bf{Z}}}\triangleleft\cdots\triangleleft{p{\bf{Z}}}\triangleleft{\bf{Z}}$$ as $n$ runs through the positive integers has every factor group being cyclic of order $p$ and therefore being simple. If we replace the prime $p$ by a prime $q\neq{p}$, then an identical series with $q$ in place of $p$ can be formed and these are not equivalent.


Do I need to assume that given group $G$ is finite? Am I missing something in these definitions?

Thanks!