Tangent spaces and calculus on

differentiable manifolds

Léonard TODJIHOUNDE

Institut de Mathématiques et de Sciences Physiques (IMSP),

Porto-Novo, Bénin

E-mail: [email protected] , Web: www.imsp-benin.com

Master course, Luanda, Nov. 2020.

1/40
Plan I
1 Tangent vectors
2 Integral curves, flows
3 Lie bracket
4 Cotagent spaces, 1-forms
5 Linear tangent mappings
6 Tensors, differential forms
7 Lie derivative
8 Exterior product of differential forms
9 Interior product of differential forms
10 Exterior differential of forms
11 Integration of differential forms
Tangent vectors

Observation :
Let p ∈ Rn and γ :] − r , r[−→ Rn be a differentiable curve in Rn ,
with γ(0) = p. For t ∈] − r , r[, γ(t) = (x1 (t), x2 (t), · · · , xn (t)) ∈ Rn .
Then γ 0 (t) = (x10 (t), x20 (t), · · · , xn0 (t)) ∈ Rn and for t = 0 we have,
γ 0 (0) = (x10 (0), x20 (0), · · · , xn0 (0)) = u ∈ Rn .
Let us consider a differentiable function f : U ∈ Rn −→ R defined
in a neighborhood U of the point p.
The restriction of f to the curve γ is the composed function
f ◦ γ :] − r , r[−→ R.

3/40
Tangent vectors

The directional derivative of f with respect to the vector u = γ 0 (0) is :

n n n
d(f ◦ γ) ∂f dxi ∂ ∂
|t=0 |t=0 = ( xi0 (0)
X X X
|t=0 = )f = ( ui )f ,
dt i=1
∂xi dt i=1
∂xi i=1
∂xi

where ui = xi0 (0) is the ith-component of the vector u = γ 0 (0).

Therefore the directional derivative of f with respect to the vector u
is an operator acting on f and depending only on u.
Tangent vectors

The precedent observation will be used to define a tangent vector to

a curve when we replace Rn by a differentiable manifold.
Let M be an n-dimensional manifold. p ∈ M and ] − ε , ε[−→ M be
a differentiable curve in M with γ(0) = p, and D the set of all
functions (ie. : with real values) on M that are differentiable at the
point p.

Definition :
The tangent vector to the curve γ at the point p = γ(0) is the
function γ 0 (0) : D −→ R, defined on D and given by :

d(f ◦ γ)
γ 0 (0)f =: |t=0 , ∀ f ∈ D .
dt

5/40
Tangent vectors

A tangent vector to the n-dimensional manifold M at a point p ∈ M

is a tangent vector at the point q to some differentiable curve
σ :] − r , r[−→ M with σ(0) = p.
The set of all tangent vectors to M at a point p ∈ M will be denoted
by Tp M.

6/40
Tangent vectors

Locally if we denote by ((x1 (t), x2 (t), · · · , xn (t))the components of

the σ(t) ∈ M (for t ∈] − r , r[) in a local chart (U, φ) at p ∈ M, then
for all f ∈ D,

d(f ◦ σ) d
σ 0 (0) = |t=0 = f (x1 (t), x2 (t), , · · · , xn (t))|t=0
dt dt
n n
!
X
0 ∂f X
0 ∂
= xi (0) = xi (0)( )0 f .
i=1
∂xi i=1
∂xi

7/40
Tangent vectors

Thus the tangent vector σ 0 (0) ∈ Tp M can be expressed in the local

chart (U, φ) by
n
∂
σ 0 (0) = xi0 (0)(
X
)0 .
i=1
∂xi

It follows that the set Tp with the usual operations (addition and
external multiplication) forms a vector space of dimension n.

8/40
Tangent vectors

Equivalent definition of tangent vectors

Let M be a n-dimensional differentiable manifold and C ∞ (M) be the

set of all functions of class C ∞ on M.
A tangent vector at a point p ∈ M is a R-linear mapping
Xp : C ∞ (M) −→ R satisfying the Leibniz rule :

Xp (fg) = Xp (f )g(p) + f (p)Xp (g) , ∀ f , g ∈ C ∞ (M) .

The set of all tangent vectors at the point p ∈ M is called tangent

space to M at the point p and denoted by Tp M.

9/40
Tangent vectors

By setting for any Xp , Yp ∈ Tp M and any f ∈ C ∞ (M)

(Xp + Yp )(f ) := Xp (f ) + Yp (f ) and (λXp )(f ) := λ.Xp (f ) ,

we can see that the tangent space Tp M has a natural structure of

vector space of dimension n.
∂
A basis of Tp M is given by the natural tangent vectors ∂xi |p ,
i = 1 · · · n, defined by :

∂ ∂(f ◦ φ−1 )
(f ) = (φ(p)) ,
∂xi |p ∂xi

where φ ≡ (x 1 , · · · , x n ) is a system of local coordinate around the

point p.
10/40
Tangent vectors

A tangent vector Xp ∈ Tp M can then be writen as :

n
X ∂
Xp = X i (p) .
i=1
∂xi |p

The real numbers X i (p) = Xp (x i ) are called the components of the

tangent vector Xp with respect to the local coordinate system
(x 1 , · · · , x n ).
Exercice :
1. Justify that if f ∈ C ∞ (M) is such that df (p) = 0, then Xp (f ) = 0,
∀ Xp ∈ Tp M.
∂
2. Prove that the family ( ∂x )ni=1 is a basis of Tp M.
i|p

11/40
Tangent vectors

The set ∪ Tp M = {(p , Xp ) / p ∈ M , Xp ∈ Tp M} is called the

p∈M
tangent bundle over M and denoted by TM ;

TM = {(p , Xp ) / p ∈ M , Xp ∈ Tp M}
Proposition
Let M be an m-dimensional differentiable manifold. The tangent
bundle TM of M is a differentiable manifold of dimension 2m.

12/40
Tangent vectors

A vector field on M is a mapping X : M −→ TM, of class C ∞ , that

assigns to each point p ∈ M a tangent vector X (p) = Xp ∈ Tp M.
The set of all vector fields on M is denoted by χ(M) or sometimes by
abus TM.
In a local coordinate system (x i ) on M a vector field X ∈ χ(M) can
be writen as
n
X ∂
X= Xi ,
i=1
∂x i

where X i are real functions defined on M by X i = X (x i ).

13/40
Integral curves, flows

Let X be a vector field on M. An integral curve of X is a curve γ in M

whose tangent vector at any point γ(t) coincides with the tangent
vector X (γ(t)) ∈ Tγ(t) M.
In local coordinate we get the following system of differential
equations :
dγ k (t)
= X k (γ(t)) ; k = 1, · · · , n .
dt
Given a vector field X and a point p ∈ M, there exists only one
integral curve γ of X such that γ(0) = p (Cauchy-Lipschitz theorem
for differential equations).
Let us denote by γ(t, p) the point of this integral curve
corresponding to the value t of the parameter.

14/40
Integral curves, flows
We call flow generated by the vector field X the mapping
γ : R × M −→ M that assigns to each pair (t0 , p0 ) ∈ R × M the point
of the integral curve γ of X with γ(0) = p0 corresponding to the
value t0 of the parameter.

The flow φ : R × M −→ M generated by a vector field has the

following property :
φ(t, φ(s, p)) = φ(t + s, p)
with
φ(0, .) = id .
For t ∈ R, the mapping φt : M −→ M defined by φt (p) = φ(t, p), is a
diffeomorphism.
The family (φt )t is called one-parameter group of diffeomorphisms
defined by the flow of X .
15/40
Integral curves, flows

Example
Let us consider on M = R2 the vector field X defined by
∂ ∂
X (x, y) = −y ∂x + x ∂y .
The flow generated by X is defined by :

γ(t, (x, y)) = (x cos t − y sin t, x sin t + y cos t) .

For a fixed point (x, y), it is a circle centered at the origin.

16/40
Lie bracket
Let X and Y be two vector fields on a differentiable manifold M.
The Lie bracket of X and Y is the vector field denoted [X , Y ] and
defined by :
[X , Y ](f ) = X (Y (f )) − Y (X (f )) , ∀ f ∈ C ∞ (M) .
In local coordinate (x i )i on M we get :
n
X ∂Y i j ∂X i j ∂
[X , Y ] = ( X − Y) .
i,j=1
∂xj ∂xj ∂xi

The Lie bracket is R-bilinear, skew-symmetric and satisifies the

follwing properties :
[X , [Y , Z ]] + [Z , [X , Y ]] + [Y , [Z , X ]] = 0 , ∀ X , Y , Z ∈ χ(M) .
[fX , gY ] = fg[X , Y ] + f (X (g))Y − g(Y (f ))X ,
∀ X , Y ∈ χ(M) , ∀ f , g ∈ C ∞ (M).
17/40
Cotagent spaces, 1-forms

We say that two vector fields commute if their Lie bracket vanishes.

The cotangent space to a differentiable manifold M at a point p ∈ M

is the dual space Tp∗ M of the vector space Tp M.
The elements of Tp∗ M are called covectors at the point p.
An 1-form on M is a mapping ω : M −→ ∪ Tp∗ M that assigns to
p∈M
each point p ∈ M a covector ωp ∈ Tp∗ M.
The set of all 1-forms on M is often denoted by Ω1 (M).
In local coordinate (x i )i on M, by denoting (dx i )i the dual basis of
∂
( ∂x i
)i of TM, an 1-form on M can be locally writen as :

X ∂
ω= ωi dx i with ωi = ω( ).
i
∂xi

18/40
Einstein summation convention

Einstein summation convention :

The Einstein summation convention or Einstein notation is a useful
notation shorthand for manipulating equations about coordinates.
According to this convention, when the index of a variable appears
twice in a term, we mean the summation over all the values that
this index can take. This index is said to be silent. It is shown once in
the upper position, once in the lower position.

19/40
Einstein summation convention
For example in a local coordinate system (x i ) on a m− dimensional
differentiable manifold M a vector field X ∈ χ(M) can be writen as
n
X ∂
X = Xi
i=1
∂xi
∂
X := X i (Einstein notation)
∂xi
∂
and by denoting (dx i )i the dual basis of ( ∂x i
)i of TM, an 1-form on M
can be locally writen as :
X
ω = ωi dx i
i
ω := ωi dx i (Einstein notation)

20/40
Linear tangent mappings
Let F : M −→ N be a mapping between the m-dimensional manifold
M and the n-dimensional manifold N.
For p ∈ M, we consider the mapping F∗p ≡ dFp : Tp M −→ TF (p) N
defined by

dFp (v)(f ) = v(f ◦ F ) , ∀ v ∈ Tp M , ∀ f ∈ C ∞ (M) .

The map F∗p is linear and is called the linear tangent mapping of F
at p.
In local coordinate (x 1 , · · · , x m ) on M and (y 1 , · · · , y n ) on N we have
n
∂ X ∂F k ∂
dFp ( )= ,
∂xi |p k=1
∂xi |p ∂y k |F (p)

where F 1 , · · · , F n are the components functions of F .

21/40
Linear tangent mappings

By considering the tangent bundles over M and N, we get the linear

tangent mapping F∗ : TM −→ TN induced by F which is defined by

F∗ (X )(F (p)) = F∗p (Xp ) , ∀ X ∈ TM , ∀ p ∈ M .

It holds :
F∗ ([X , Y ]) = [F∗ X , F∗ Y ] , ∀ X , Y ∈ TM .
If F : M1 −→ M2 and : M2 −→ M3 are two mappings, then

(G ◦ F )∗ = G∗ ◦ F∗ .

22/40
Linear tangent mappings

The pull-back of F is the mapping F ∗ : Ω1 (N) −→ Ω1 (M) defined by :

F ∗ (ω)(X ) = ω(F∗ X ) , ∀ ω ∈ Ω1 (N) , ∀ X ∈ χ(M) .

The pull-back satisfies the relation :

(G ◦ F )∗ = F ∗ ◦ G∗ .

23/40
Tensors, differential forms

A tensor of type (r, s) at a point p of a differentiable manifold M is a

multilinear mapping of class C ∞ ,
Ap : (Tp∗ M)r × (Tp M)s −→ R.
A tensors field of type (r, s) (or r contravariant and s covariant) on
M is a mapping A, of class C ∞ , that assigns to each point p ∈ M a
tensor Ap of type (r, s).
In local coordinate a tensors field of type (r, s) can be writen as :

A = Aij11 ij22···i j1 js
···jr dx ⊗ · · · ⊗ dx ⊗ ∂i1 ⊗ · · · ⊗ ∂ir .
s

24/40
Tensors, differential forms
A tensors field of type (r, s) acts on r 1-forms and s vector fields as
the follows :
A(ω1 , · · · , ωr , X1 , · · · , Xs ) =
Aij11 ij22···i
···jr
s j1
× dx (X1 ) · · · dx js (Xs ) × ωi1 (∂i1 ) · · · ωr (∂ir ) =
j1
Aji11 ij22···i js i1 ir
···jr X1 · · · Xs ω1 · · · ωr
s
=
A tensors field of type (0, 1) is a 1-form and a vectors field is a
tensors field of type (1, 0).
A differential form of degree q (or q-form) on M is a tensors field of
type (0, q) sur M that is skew-symmetric.
The set of all q-forms on M is denoted by Ωq (M) and
Ω(M) = ∪ Ωq (M).
q∈N
q
For a fixed point p ∈ M the setb Ωp (M) of all q-forms at the point p
is a vector space.
Tensors, differential forms

Let f : M −→ N be a map between two differentiable manifolds M

and N, and A a tensors field of type (0, s) on N.
We can associate to A a tensors field f ∗ T on M of type (0, s) by
puting :
(f ∗ A)(X1 , · · · , Xs ) = A(f∗ X1 , · · · , f∗ Xs ) .
In the case M = N we can extend the pull-back to any tensors field A
of type (r, s) on M by puting :

(f ∗ T )(ω1 , · · · , ωr , X1 , · · · , Xs ) = T (f ∗ ω1 , · · · , f ∗ ωr , f∗ X1 , · · · , f∗ Xs ) .
Lie derivative

Let A be a tensors field and X a vectors field on a differentiable

manifold M.
The Lie derivative of A in the direction of X is the tensors field LX A
of the same type as A and defined by :
1
LX A|p = lim (A|p − φ∗t A|φt (p) ) , ∀ p ∈ M ,
t=0 t

where {φt }t is the one-parameter group of diffeomorphisms

generated by the vectors field X .

27/40
Lie derivative

The Lie derivative is R-linear with respect to both variables,

C ∞ (M)-linear with respect to the direction X of derivation and
satisfies the properties :

LX Y = [X , Y ] and L[X ,Y ] = [LX , LY ] , ∀ X , Y ∈ χ(M) ,

LX f = X (f ) and LX (dω) = d(LX ω) , ∀ X ∈ χ(M) , ∀ ω ∈ Ωr (M) .

Exercice : prove the properties satisfied by the Lie derivative and
given above.

28/40
Exterior product of differential forms

Let α ∈ Ωr (M) and β ∈ Ωs (M) be two differential forms of degree r

and s respectively.
The exterior product of α and β is the differential (r + s)-form α ∧ β
defined by :

α ∧ β(X1 , · · · , Xr+s ) =
1 X
sgn(σ)α(Xσ(1) , · · · , Xσ(r) )β(Xσ(r+1) , · · · , Xσ(r+s) ) ,
r!s! σ

where σ runs over all permutations of {1, · · · , r + s}.

29/40
Exterior product of differential forms

The exterior product is associaative and satisfies the following

properties :
α ∧ α = 0 if α is of odd degreee .
α ∧ β = (−1)rs β ∧ α , ∀ α ∈ Ωr (M) , ∀ β ∈ Ωs (M) .
In local coordinate (x i )i on M (with dim M = n), if (dx 1 , · · · , dx n ) is a
basis of Ω1 (M) = T ∗ M, then a basis of Ωr (M) is determined by the
exterior products of the form dx i1 ∧ · · · ∧ dx ir , with i1 < · · · < ir
belonging to {1, · · · , n}.
It holds that
n!
dim Ωrp (M) =
(n − r)!r!

30/40
Interior product of differential forms

The interior product of a differrential form ω ∈ Ωr (M) of degree r by

a vectors field X ∈ χ(M), is the differential form iX ω of degree
(r − 1) defined by :

iX ω(X1 , · · · , Xr−1 ) = ω(X , X1 , · · · , Xr−1 ) , ∀ X1 , · · · , Xr−1 ∈ χ(M) .

31/40
Exterior differential of forms

The exterior differential of forms on M, is the R-linear mapping

d : Ω(M) −→ Ω(M) such that :
(i) d|Ωr (M) : Ωr (M) −→ Ωr+1 (M)
(ii) d 2 = d ◦ d = 0
(iii) d(α ∧ β) = (dα) ∧ β + (−1)|α| α ∧ (dβ) ∀ α, β ∈ Ω(M).
If ω ∈ Ω1 (M), then dω is a 2-form on M given by :

dω(X , Y ) = X (ω(Y )) − Y (ω(X )) − ω([X , Y ]) .

32/40
Exterior differential of forms

More generally if ω ∈ Ωr (M), then dω a (r + 1)-forme on M given by :

r
X
dω(X1 , · · · , Xr+1 ) = (−1)i+1 Xi (ω(X1 , · · · , Xi−1 , X̂i , · · · , Xr+1 ))
i=1
X
+ (−1)i+j ω([Xi , Xj ], X1 , · · · , X̂i , · · · , X̂j , · · · , Xr+1 ) .
i<j

33/40
Exterior differential of forms

A differential form ω ∈ Ωr (M) is said tobe closed if dω = 0 and it

said to be exact if there exists a differentiable form α ∈ Ωr−1 (M)
such that ω = dα.
Any exact form is closed. Therefore the space of exact r-forms =dr−1
is contained in the space of closed r-forms ker dr .
The quotient space H r (M) = ker dr /=r−1 is called the De Rham
group of cohomology of order r.

34/40
Exterior differential of forms

Properties :
We have the following relations :

LX ω = (diX +iX d) ω and LX (iX ω) = iX (LX ω) , ∀ X ∈ χ(M) , ∀ ω ∈ Ω(M) .

i[X ,Y ] ω = X (iX ω) − Y (iX ω) , ∀ X , Y ∈ χ(M) , ∀ ω ∈ Ω(M) .

iX (α∧β) = (iX ω)∧β+(−1)|α| α∧(iX β) , ∀ X ∈ χ(M) , ∀ α, β ∈ Ω(M) .

35/40
Integration of differential forms

Let M be a differentiable orientable manifold of dimension n.

There exists on M a differentiable form ω of degree n that is nonzero
everywhere. For example the n-form called volume element on M
and defined in local coordinate (x i )i by :

ω(p) = h(p)dx 1 ∧ · · · ∧ dx n ,

where h is a real function on M that is either strictly positive or

strictly negative.
We have then two classes of volume elements on M.

36/40
Integration of differential forms

The choice of the represent of a class fix an orientation on M and

will determine the measure with respect to which we will integrate
functions and n-forms on M.
Let (Ui , φi ) be a local chart on M with associates local coordinate
x = (x i )i , ω be a volume element on M given in local coordinate by
ω = hdx 1 ∧ · · · ∧ dx n and f be a real fonction on M. The integral of
the n-form f ω on the open set Ui ⊂ M is defined by :
Z Z
fω = f (φ−1 −1 1 n
i (x))h(φi (x))dx ∧ · · · ∧ dx .
Ui φi (Ui )

37/40
Integration of differential forms

Consider a partition unity (Ui , εi )i subordinated to an atlas on M

and let us put fi = f εi . It follows
X X
f (p) = f (p)εi (p) = fi (p) , ∀ p ∈ M .
i i

The integral of f ω on M is then defined by :

Z XZ
fω = fi ω .
M i Ui

Exercice : prove that the definition of the integral given above does
not depend neither on the atlas nor on the partition of unity
considered.
Integration of differential forms

Example
Let us consider M = S 1 , U1 = S 1 \ {(1, 0)}, U2 = S 1 \ {(−1, 0)},
ε1 (θ) = sin2 ( θ2 ), ε2 (θ) = cos2 ( θ2 ).
(ε1 , ε2 ) is a partition of unity subordinated to the covering (U1 , U2 )
of S 1 .
We have for example : :
Z 2π Z 2π
θ θ
Z
cos θdθ = cos θ sin2 ( )dθ + cos θ cos2 ( )dθ
S1 0 2 0 2
π π
= + =π;
2 2
which is the well-known result.

39/40
Integration of differential forms

Theorem of Stokes :
Let M be a differentiable manifold of dimension n, oriented,
compact and with bord ∂M.
For any differentiable form α of degree (n − 1) on M, it holds :
Z Z
dα = α.
M ∂M

40/40