Classifying map of a principal bundle

A map from the base into BG whose pullback of EG reproduces a given principal G-bundle.
Classifying map of a principal bundle

Fix a GG and a chosen π ⁣:EGBG\pi\colon EG\to BG.

Definition (Classifying map)

Let MM be a and let PMP\to M be a . A classifying map for PP is a continuous map

c ⁣:MBG c\colon M \longrightarrow BG

such that there is an isomorphism of principal GG-bundles

Pc(EG). P \cong c^{*}(EG).

If MM is paracompact (in particular, if MM is a smooth manifold), then:

  • a classifying map exists, and
  • its homotopy class in is uniquely determined by PP.

Equivalently, isomorphism classes of principal GG-bundles correspond to homotopy classes of classifying maps (see ).

Examples

  1. Trivial bundle. For P=M×GP=M\times G, a classifying map can be taken to be constant (and hence null-homotopic).
  2. Hopf fibration. The principal U(1)U(1)-bundle S3S2S^3\to S^2 is classified by a map S2BU(1)CPS^2\to BU(1)\simeq \mathbb{C}P^\infty representing the generator of H2(S2;Z)H^2(S^2;\mathbb{Z}).
  3. Frame bundle viewpoint. The oriented orthonormal frame bundle of a Riemannian nn-manifold is a principal SO(n)SO(n)-bundle; its classifying map MBSO(n)M\to BSO(n) encodes characteristic classes such as the Stiefel–Whitney and Pontryagin classes.