Proof of the Jordan-Holder Theorem
The Jordan-Holder theorem establishes the uniqueness of composition factors, showing that the "building blocks" of a group are well-defined regardless of how the group is decomposed.
Statement
Let be a group with a composition series. Any two composition series of have the same length, and the composition factors (up to isomorphism and permutation) are the same.
Proof
We prove the theorem by induction on the length of a composition series.
Base case (): , trivially true.
Base case (): is simple. Any composition series is with one factor . Unique.
Inductive step: Let and be two composition series.
Case 1: . Then the two series share the same maximal normal subgroup, and by induction on , the portions below have the same factors. The top factors are both , so the full series are equivalent.
Case 2: . Since both are maximal normal subgroups of (as the quotients are simple), and , the product is a normal subgroup of strictly containing both. By maximality: .
Let . By the second isomorphism theorem:
Now has composition series (being a subgroup of ). Let be one.
Then:
- is a composition series (factors of , then , then ).
- is a composition series (factors of , then , then ).
By induction on : the original series through has the same factors as the series through then . By induction on : similarly for the series through .
Comparing: both original series have the same multiset of composition factors as the "bridge" series through (the factors of , plus and , which simply swap positions).
Corollary
The Jordan-Holder theorem means we can speak of "the composition factors of " without ambiguity. For a finite group , the composition factors (with multiplicities) are a complete invariant of the "simple building blocks" of . However, the composition factors do NOT determine up to isomorphism: many non-isomorphic groups share the same composition factors (e.g., and both have factors ). The extension problem -- classifying groups with given composition factors -- is one of the hardest problems in group theory.