, in particular the theory of principal bundles
, one can ask if a
-bundle "comes from" a subgroup
. This is called reduction of the structure group
), and makes sense for any map
, which need not be an inclusion (despite the terminology).
Formally, given a G
and a map H
(which need not be an inclusion),
a reduction of the structure group
) is an H
such that the pushout
is isomorphic to B
Note that these do not always exist, nor if they exist are they unique.
As a concrete example, every even dimensional real vector space is the underlying real space of a complex vector space: it admits a linear complex structure. A real vector bundle admits an almost complex structure if and only if it is the underlying real bundle of a complex vector bundle. This is a reduction along the inclusion GL(n,C) → GL(2n,R)
In terms of transition maps, a G-bundle can be reduced if and only if the transition maps can be taken to have values in H.
Note that the term reduction is misleading: it suggests that H is a subgroup of G, which is often the case, but need not be (for example for spin structures): it's properly called a lifting.
More abstractly, "G-bundles over X" is a functor in G: given a map H → G, one gets a map from H-bundles to G-bundles by inducing (as above). Reduction of the structure group of a G-bundle B is choosing an H-bundle whose image is B.
The inducing map from H-bundles to G-bundles is in general neither onto nor one-to-one, so the structure group cannot always be reduced, and when it can, this reduction need not be unique. For example, not every manifold is orientable, and those that are orientable admit exactly two orientations.
If H is a Lie subgroup of G, then there is a natural one-to-one correspondence between reductions of a G-bundle B to H and global sections of the fiber bundle B/H obtained by quotienting B by the right action of H. Specifically, the fibration B → B/H is a principal H-bundle over B/H. If σ : X → B/H is a section, then the pullback bundle BH = σ-1B is a reduction of B.
Examples for vector bundles
, particularly the tangent bundle
of a manifold
- is an orientation, and this is possible if and only if the bundle is orientable
- is a volume form; since is a deformation retract, a volume form exists if and only if a bundle is orientable
- is a pseudo-volume form, and this is always possible
- is a metric; as is the maximal compact subgroup (so the inclusion is a deformation retract), this is always possible
- is an almost complex structure
- (which is not an inclusion: it's a 2-fold covering space) is a spin structure.
- decomposes a vector bundle as a Whitney sum (direct sum) of sub-bundles of rank k and n − k.
Many geometric structures are stronger than G
-structures; they are G
-structures with an integrability condition
. Thus such a structure requires a reduction of the structure group (and can be obstructed, as below), but this is not sufficient. Examples include complex structure
, symplectic structure
(as opposed to almost complex structures
and almost symplectic structures
Another example is for a foliation, which requires a reduction of the tangent bundle to a block matrix subgroup, together with an integrability condition so that the Frobenius theorem applies.
-bundles are classified by the classifying space BG
, and similarly H
-bundles are classified by the classifying space BH
, and the induced G
-structure on an H
-bundle corresponds to the induced map
. Thus given a G
-bundle with classifying map
, the obstruction to the reduction of the structure group is the class of
as a map to the cofiber
; the structure group can be reduced if and only if the class of
When is a homotopy equivalence, the cofiber is contractible, so there is no obstruction to reducing the structure group, for example for .
Conversely, the cofiber induced by the inclusion of the trivial group is again , so the obstruction to an absolute parallelism (trivialization of the bundle) is the class of the bundle.
Structure over a point
As a simple example, there is no obstruction to reducing the structure group of a
, thinking of a
-space as a
-bundle over a point, as in that case the classifying map is null-homotopic
, as the domain is a point. Thus there is no obstruction to "reducing the structure group" of a vector space: thus every vector space admits an orientation, and so forth.