Nielsen-Schreier theorem

The Nielsen-Schreier theorem states that every subgroup of a free group is itself free. This is not a triviality: free groups are defined by the absence of relations, and it is entirely possible that imposing constraints on a group (by passing to a subgroup) could introduce relations. The theorem says this never happens. Freedom, once present, is inherited by all substructures.

The theorem was proved by Jakob Nielsen in 1921 for finitely generated subgroups and extended by Otto Schreier in 1927 to arbitrary subgroups. Nielsen's proof used the combinatorics of reduced words and a rewriting process now called Nielsen reduction; Schreier's proof used the action of the subgroup on the Cayley tree. The topological proof is illuminating: a subgroup \(H\) of a free group \(F\) acts on the same tree that \(F\) acts on, and the quotient graph is a covering space of the bouquet of circles representing \(F\). Since covering spaces of graphs are graphs, and graphs deformation-retract to bouquets of circles, \(H\) is free.

The quantitative form of the theorem is equally striking. If \(F\) has rank \(n\) and \(H\) has index \(d\), then \(H\) has rank \(d(n-1) + 1\). This formula reflects the exponential growth of the free group: the subgroup has more generators because it must encode the branching structure of the coset tree.

The Nielsen-Schreier theorem is the group-theoretic expression of a deep principle: in negatively curved spaces, subspaces inherit the geometry of their ambient space. It is not an accident that the theorem has analogues in covering space theory, in the theory of graphs of groups, and in the geometry of hyperbolic groups. Freedom is a geometric property, and geometry propagates downward.