LECTURE 5: SMOOTH MAPS

1. Smooth Maps
Recall that a smooth function on a smooth manifold M is a function f : M → R
so that for any chart1 {ϕα , Uα , Vα } of M , the function f ◦ ϕ−1
α is a smooth function on
Vα . More generally, we can define smooth maps between smooth manifolds:
Definition 1.1. Let M, N be smooth manifolds. We say a continuous map f : M → N
is smooth if for any chart {ϕα , Uα , Vα } of M and {ψβ , Xβ , Yβ } of N , the map
ψβ ◦ f ◦ ϕ−1
α : ϕα (Uα ∩ f
−1
(Xβ )) → ψβ (Xβ )
is smooth. [Note: Both ϕα (Uα ∩ f −1 (Xβ )) and ψβ (Xβ ) are Euclidian open sets.]

Remark. In the definition we assume that the map f is already a continuous map. In
general the smoothness of all ψβ ◦ f ◦ ϕ−1
α ’s does not imply the continuity of f . See
Problem Set 2.
Remark. One can check that if f : (M, A) → (N, B) is smooth, A1 is a chart on M
that is compatible with A, and B1 is a chart on N that is compatible with B, then
f : (M, A1 ) → (N, B1 ) is smooth.
Remark. A map f = (f1 , · · · , fn ) : M → Rk is smooth if and only if each fi ∈ C ∞ (M ).

The set of all smooth maps from M to N is denoted by C ∞ (M, N ). We will

leave it as an exercise to prove that if f ∈ C ∞ (M, N ) and g ∈ C ∞ (N, P ), then
g ◦ f ∈ C ∞ (M, P ). As a consequence, any smooth map f : M → N induces a “pull-
back” map
f ∗ : C ∞ (N ) → C ∞ (M ), g 7→ g ◦ f.
The pull-back will play an important role in the future.
1In this course, when we say “any chart of a smooth manifold M ”, we always mean “any chart in
a given atlas A that defines the smooth structure of M ”.

1
2 LECTURE 5: SMOOTH MAPS

Example. The inclusion map ι : S n ,→ Rn+1 is smooth, since

1
ι ◦ ϕ−1 1 n 1 n 2


± (y , · · · , y ) = 2y , · · · , 2y , ±(1 − |y| )
1 + |y|2
are smooth maps from Rn to Rn+1 . Moreover, if g is any smooth function on Rn+1 , the
pull-back function ι∗ g is just the restriction of g to S n .
Example. The projection map π : Rn+1 \ {0} → RPn is smooth, since
 1
xi−1 xi+1 xn+1

1 n+1 x
ϕi ◦ π(x , · · · , x ) = ,··· , i , i ,··· , i
xi x x x
is smooth on π −1 (Ui ) = {(x1 , · · · , xn+1 ) : xi 6= 0} for each i.

As in the Euclidean case, we can define diffeomorphisms between smooth manifolds.

Definition 1.2. Let M, N be smooth manifolds. A map f : M → N is a diffeomor-
phism if it is smooth, bijective, and f −1 is smooth.

If there exists a diffeomorphism f : M → N , then we say M and N are diffeo-

morphic. Sometimes we will denote M ' N . We will regard diffeomorphic smooth
manifolds as the same.
The following properties can be easily deduced from the corresponding properties
in the Euclidean case:
• The identity map Id : M → M is a diffeomorphism.
• If f : M → N and g : N → P are diffeomorphisms, so is g ◦ f .
• If f : M → N is a diffeomorphism, so is f −1 . Moreover, dim M = dim N .
So in particular, for any smooth manifold M ,
Diff(M ) = {f : M → M | f is a diffeomorphism}
is a group, called the diffeomorphism group of M .
Example. If M is a smooth manifold, then any chart (ϕ, U, V ) gives a diffeomorphism
ϕ : U → V from U ⊂ M to V ⊂ Rn .
Example. We have seen that on M = R, the two atlas A = {(ϕ1 (x) = x, R, R)} and
B = {(ϕ2 (x) = x3 , R, R)} define non-equivalent smooth structures. However, the map
ψ : (R, A) → (R, B), ψ(x) = x1/3
is a diffeomorphism. So we still think these two smooth structures are the same,
although they are not equivalent.
Remark. Here are some deep results on the smooth structures:
• (J. Milnor and M. Kervaire) The topological 7-sphere admits exactly 28 different
smooth structures.
• (S. Donaldson and M. Freedman) For any n 6= 4, Rn has a unique smooth struc-
ture up to diffeomorphism; but on R4 there exist uncountable many distinct
pairwise non-diffeomorphic smooth structures.
 LECTURE 5: SMOOTH MAPS 3

2. The differential of a smooth map

We have seen that if U, V are Euclidean open sets, and f : U → V is smooth,
then for any a ∈ U , one naturally gets a linear map dfa : Ta U → Tf (a) V , which can be
think of as a linearization of f near the point a, and is extremely useful in studying
properties of f near a. Now suppose M, N are smooth manifolds, and f : M → N
smooth. We would like to define its differential dfp at p, again as a linear map between
tangent spaces, which serves as a linearization of f near p. But the first question is:
What is the tangent space of a smooth manifold at a point? Since M is an abstract
manifold, we don’t have a simple nice “geometric picture”.
To define tangent vectors on smooth manifolds, let’s first re-interpret (algebraically)
the tangent vectors in the Euclidean case. Recall that for any point a ∈ Rn and any
vector ~v at a, there is a conception of directional derivative at point a in the direction
~v , which send a smooth function f defined on Rn to the quantity
f (a + t~v ) − f (a) d
D~va f = lim = f (a + t~v ).
t→0 t dt t=0

So any ~v gives us an operator D~va : C ∞ (Rn ) → R. In coordinates, if ~v = hv 1 , · · · , v n i,

then
X ∂f
D~va f = vi i ,
∂x
∞ n
in other words, as an operator on C (R ),
X ∂
D~va = vi i .
∂x
Observe that for any f, g ∈ C ∞ (Rn ),
• (linearity) D~va (αf + βg) = αD~va f + βD~va g for any α, β ∈ R.
• (Leibnitz law) D~va (f g) = f (a)D~va g + g(a)D~va f .
In general, we can define
Definition 2.1. Any operator Da : C ∞ (Rn ) → R satisfying these two properties is
called a derivative at a.

So any vector ~v at a defines a derivative D~va at a. It is not hard to see that the
correspondence v D~va preserves linearity, i.e. Dα~
a a
~ = αD~v + βDw
v +β w
a
~ . The next
a
proposition tells us that the correspondence v D~v is one-to-one, so that we can
identify the set (vector space) of tangent vectors at a with the set of derivatives at a:
Proposition 2.2. Any derivative D : C ∞ (Rn ) → R at a is of the form D~va for some
vector ~v at a.

Proof. For any f ∈ C ∞ (Rn ), we have

Z 1 n
d X
f (x) = f (a) + f (a + t(x − a))dt = f (a) + (xi − ai )hi (x),
0 dt i=1
4 LECTURE 5: SMOOTH MAPS

where Z 1
∂f
hi (x) = i
(a + t(x − a))dt.
0 ∂x
Note that the Leibnitz property implies D(1) = 0 since
D(1) = D(1 · 1) = 2D(1).
By linearity, D(c) = 0 for any constant c. So
n n n
X
i
X
i i
X ∂f
D(f ) = 0 + D(x )hi (a) + (a − a )D(hi ) = D(xi ) (a).
i=0 i=0 i=0
∂xi
It follows that as an operator on C ∞ (Rn ),
n
X ∂
D= D(xi ) .
i=1
∂xi x=a

In other words, if we let ~v = hD(x1 ), · · · , D(xn )i, then D = D~va . 

This motivates the following definition:

Definition 2.3. Let M be an n-dimensional smooth manifold. A tangent vector at a
point p ∈ M is a R-linear map Xp : C ∞ (M ) → R satisfying the Leibnitz law
(1) Xp (f g) = f (p)Xp (g) + Xp (f )g(p)
for any f, g ∈ C ∞ (M ).

It is easy to see that the set of all tangent vectors of M at p is a linear space. We
will denote this set by Tp M , and call it the tangent space Tp M to M at p..
As argued above, if f is a constant function, then Xp (f ) = 0. More generally,
Lemma 2.4. If f = c in a neighborhood of p, then Xp (f ) = 0.

Proof. Let ϕ be a smooth function on M that equals 1 near p, and equals 0 at points
where f 6= c. (The existence of such f is guaranteed by partition of unity.) Then
(f − c)ϕ ≡ 0. So
0 = Xp ((f − c)ϕ) = (f (p) − c)Xp (ϕ) + Xp (f )ϕ(p) = Xp (f ).


As a consequence, we see that if f = g in a neighborhood of p, then Xp (f ) = Xp (g).

In other words, Xp (f ) is determined by the values of f near p. So one can replace
C ∞ (M ) in Definition 2.3 by C ∞ (U ), where U is any open set that contains p. In other
words, as linear spaces, Tp M is isomorphic to Tp U .
Finally we are ready to define the differential of a smooth map between smooth
manifolds. Recall that the differential of a smooth map f : U → V between open
sets in Euclidean spaces at a ∈ U is a linear map dfa : Ta U = Rnx → Tf (a) V = Rm y
∂fi
whose matrix is the Jacobian matrix ( ∂x j ) of f at a. To transplant this conception to
smooth maps, we need to take a closer look at the two interpretation of Ta U : We have
 LECTURE 5: SMOOTH MAPS 5

seen that
P we can identify the (geometric )vector ~v at a with the (algebraic) derivative
a i ∂
D~v = v ∂xi . Note that geometrically,
  * +
∂fi X ∂f1 X ∂fn
dfa (~v ) = · ~v = vj , · · · , vj .
∂xj j
∂x j
j
∂x j

The vector in the right hand side is a vector in Rmy . When interpreted as a derivative
∞ m ∞ m
on C (Ry ), it is a map that maps g ∈ C (Ry ) to
X X ∂fi ∂g X ∂
vj j i = v j j (g ◦ f ) = D~va (g ◦ f ).
i j
∂x ∂y j
∂x
In other words, the derivative that corresponds to the vector dfa (~v ) is the derivative at
f (a) that maps g ∈ C ∞ (Rm ) to D~va (g ◦ f ).
Motivated by these computations, we define
Definition 2.5. Let f : M → N be a smooth map. Then for each p ∈ M , the
differential of f is the linear map dfp : Tp M → Tf (p) N defined by
dfp (Xp )(g) = Xp (g ◦ f )
∞
for all Xp ∈ Tp M and g ∈ C (N ).
The chain rule still holds for composition of smooth maps:
Theorem 2.6 (Chain rule). Suppose f : M → N and g : N → P are smooth maps,
then d(g ◦ f )p = dgf (p) ◦ dfp .
Proof. For any Xp ∈ Tp M and h ∈ C ∞ (P ),
d(g ◦ f )p (Xp )(h) = Xp (h ◦ g ◦ f ) = dfp (Xp )(h ◦ g) = dgf (p) (dfp (Xp ))(h).
So the theorem follows. 
Obviously the differential of the identity map is the identity map between tangent
spaces. By repeating the proof of Theorem 1.2 in Lecture 2 we get
Corollary 2.7. If f : M → N is a diffeomorphism, then dfp : Tp M → Tf (p) N is a
linear isomorphism.
In particular, we have
Corollary 2.8. If dim M = n, then Tp M is an n-dimensional linear space.
Proof. Let {ϕ, U, V } be a chart near p. Then ϕ : U → V is a diffeomorphism. It
follows that dim Tp M = dim Tp U = dim Tf (p) V = n. 
In particular, we see that the tangent vectors ∂i := dϕ−1 ( ∂x∂ i ) form a basis of Tp M .
In coordinates, one has the following explicit formula for ∂i :
∂(f ◦ ϕ−1 )
∂i : C ∞ (U ) → R, ∂i (f ) = (ϕ(p)).
∂xi