Article · Wikipedia archive · Last revised Jun 10, 2026

Parallelization (mathematics)

In mathematics, a parallelization of a manifold of dimension n is a set of n global smooth linearly independent vector fields.

Last revised
Jun 10, 2026
Read time
≈ 2 min
Length
403 w
Citations
2
Source

In mathematics, a parallelization1 of a manifold M {\displaystyle M\,} of dimension n is a set of n global smooth linearly independent vector fields.

Formal definition

Given a manifold M {\displaystyle M\,} of dimension n, a parallelization of M {\displaystyle M\,} is a set { X 1 , , X n } {\displaystyle \{X_{1},\dots ,X_{n}\}} of n smooth vector fields defined on all of M {\displaystyle M\,} such that for every p M {\displaystyle p\in M\,} the set { X 1 ( p ) , , X n ( p ) } {\displaystyle \{X_{1}(p),\dots ,X_{n}(p)\}} is a basis of T p M {\displaystyle T_{p}M\,} , where T p M {\displaystyle T_{p}M\,} denotes the fiber over p {\displaystyle p\,} of the tangent vector bundle T M {\displaystyle TM\,} .

A manifold is called parallelizable whenever it admits a parallelization.

Examples

Properties

Proposition. A manifold M {\displaystyle M\,} is parallelizable iff there is a diffeomorphism ϕ : T M M × R n {\displaystyle \phi \colon TM\longrightarrow M\times {\mathbb {R} ^{n}}\,} such that the first projection of ϕ {\displaystyle \phi \,} is τ M : T M M {\displaystyle \tau _{M}\colon TM\longrightarrow M\,} and for each p M {\displaystyle p\in M\,} the second factor—restricted to T p M {\displaystyle T_{p}M\,} —is a linear map ϕ p : T p M R n {\displaystyle \phi _{p}\colon T_{p}M\rightarrow {\mathbb {R} ^{n}}\,} .

In other words, M {\displaystyle M\,} is parallelizable if and only if τ M : T M M {\displaystyle \tau _{M}\colon TM\longrightarrow M\,} is a trivial bundle. For example, suppose that M {\displaystyle M\,} is an open subset of R n {\displaystyle {\mathbb {R} ^{n}}\,} , i.e., an open submanifold of R n {\displaystyle {\mathbb {R} ^{n}}\,} . Then T M {\displaystyle TM\,} is equal to M × R n {\displaystyle M\times {\mathbb {R} ^{n}}\,} , and M {\displaystyle M\,} is clearly parallelizable.2

See also

See also

Notes

Notes

References

References