Submanifold Tangent Space -- from Wolfram MathWorld

The submanifold tangent space at a point of a smooth submanifold is identified with a linear subspace of the ambient tangent space . For the inclusion , its tangent map is injective, and its image is this subspace (Meinrenken 2002, pp. 11-12). For a surface in Euclidean space, it is the tangent plane translated to the origin. The elements of the tangent space are called manifold tangent vectors, and they are closed under vector addition and scalar multiplication. In particular, the tangent space is a vector space.

There are two ways of defining a submanifold, and each way gives rise to a different way of defining the tangent space. The first way uses a parameterization, and the second way uses a system of equations.

(1)

where is the unit open ball in , and . At the point , the tangent space is the image of the Jacobian of , as a linear transformation from to . For example, consider the unit sphere

$S^2={(y_1,y_2,y_3):y_1^2+y_2^2+y_3^2=1}$

(2)

in . Then the function (with the domain $U={(x_1,x_2):x_1^2+x_2^2<1}$ )

(3)

(4)

whose image is the tangent space at ,

$TS^2|_((0,0,1))={(a,b,0)}.$

(5)

An alternative description of a submanifold as the set of solutions to a system of equations leads to another description of manifold tangent vectors. Consider a submanifold which is the set of solutions to the system of equations

where and the Jacobian of the smooth map , with , has rank at the solutions to . A manifold tangent vector at a solution is an infinitesimal solution to the above equations (at ). The manifold tangent vector is a solution of the derivative (linearization) of , i.e., it is in the null space of the Jacobian.

Consider this method in recomputing the tangent space of the sphere at the north pole. The sphere is two-dimensional and is described as the solution to a single equation () . Set . We want to compute the tangent space at the solution (at the north pole). The Jacobian at this point is the matrix , and its null space is the tangent space

$TS^2|_((0,0,1))={(a,b,0)}.$

(9)

It appears that the tangent space depends either on the choice of parameterization, or on the choice of system of equations. Because the Jacobian of a function composition obeys the chain rule, the tangent space is well-defined. Note that the Jacobian of a diffeomorphism is an invertible linear map, and these correspond to the ways the equations can be changed. The basic facts from linear algebra used to show that the tangent space is well-defined are the following.

1. If is an invertible linear map, then the image of is the same as the image of .

2. If is an invertible linear map, then the null space of , for , is the inverse image under of the null space of . More precisely, .

These techniques work in any dimension. In addition, they generalize to submanifolds of an abstract manifold, because manifold tangent vectors depend on local properties. In particular, the tangent space can be computed in any coordinate chart, because any change in coordinate chart corresponds to a diffeomorphism in Euclidean space.

The tangent space can give some geometric insight to higher-dimensional phenomena. For example, to compute the tangent space to the torus (donut) in (which is a flat manifold), note that it can be parametrized by