Introduction
This chapter is devoted to the proof of the inverse and implicit
function theorems. The inverse function theorem is proved in
Section 1 by using the contraction mapping principle. Next the
implicit function theorem is deduced from the inverse function
theorem in Section 2. Section 3 is concerned with various
definitions of curves, surfaces and other geo-metric objects.
The relation among these definitions are elucidated by the
inverse/implicit function theorems.
The Inverse Function Theorem
This chapter is concerned with functions between the Euclidean spaces and the
inverse and implicit function theorems. We learned these theorems in advanced
calculus but the proofs were not emphasized. Now we fill out the gap. Adapting the
notations in advanced calculus, a point x = (x1, x2, · · · , xn) ∈ Rn is sometimes
called a vector and we use |x| instead of ‖x‖2 to denote its Euclidean norm. All is
about linearization. Recall that a real-valued function on an open interval I is
differentiable at some x0 ∈ I if there exists some a ∈ R such that
Here ◦(z) denotes a quantity satisfying limz→0 ◦(z)/|z| = 0. The same situation
carries over to a real-valued function f in some open set in Rn. A function f is called
differentiable at p0 in this open set if there exists a vector a = (a1, · · · ,an) such
that
where (aij) = (∂f i/∂xj) is the Jabocian matrix of f . (4.1) shows near p0, that is, when
z is small, the function F is well-approximated by the linear map DF (p0) up to the
constant F (p0) as long as DF (p0) is nonsingular. It suggests that the local
information of a map at a differentiable point could be retrieved from its a linear
map, which is much easier to analyse. This principle, called linearization, is widely
used in analysis. The inverse function theorem is a typical result of linearization. It
asserts that a map is locally invertible if its linearization is invertible. Therefore, local
bijectivity of the map is ensured by the inevitability of its linearization. When DF (p0)
is not invertible, the first term on the right hand side of (4.1) may degenerate in
some or even all direction so that DF (p0)z cannot control the error term ◦(z). In this
case the local behavior of F may be different from its linearization.
Inverse Function Theorem
Let F : U → Rn be a C1-map where U isoprene in Rn and p0 ∈ U . Suppose that DF
(p0) is invertible. There exist open sets V and W containing p0 and F (p0)
respectively such that the restriction of F on V is a bijection onto W with a C1-
inverse. Moreover, the inverse is Ck when F is Ck, 1 ≤ k ≤ ∞, in U .
Example: The inverse function theorem asserts a local inevitability. Even if the
linearization is non-singular everywhere, we cannot assert global inevitability. Let us
consider the switching between the Cartesian and polar coordinates in the plane:
x = r cos θ, y = r sin θ .
The function F :
(0, ∞) × (−∞, ∞) → R2 given by F (r, θ) = (x, y) is a continuously differentiable
function whose Jacobian matrix is non-singular except (0, 0). However, it is clear that
F is not bijective, for instance, all points (r, θ + 2nπ), n ∈ Z, have the same image
under F .
Example 2:. An exceptional case is dimension one where a global result is
available. Indeed, in MATH2060 we learned that if f is continuously differentiable on
(a, b) with non-vanishing f ′, it is either strictly increasing or decreasing so that its
global inverse exists and is again continuously differentiable.
Example 3: Consider the map F : R2 → R2 given by F (x, y) = (x2, y). Its Jacobian
matrix is singular at (0, 0). In fact, for any point (a, b), a > 0, F (±√a, b) = (a, b). We
cannot find any open set, no matter how small is, at (0, 0) so that F is injective. On
the other hand, the map H(x, y) = (x3, y) is bijective with inverse given by J(x, y) =
(x1/3, y). However, as the non-degeneracy condition does not hold at (0, 0) so it is
not differentiable there. In these cases the Jacobian matrix is singular, so the no
degeneracy condition does not hold. We will see that in order the inverse map to be
differentiable, the no degeneracy condition must hold.
A map from some open set in Rn to Rm is Ck, 1 ≤ k ≤ ∞ if all its components belong
to Ck. It is called a C∞-map or a smooth map if its components are C∞. The condition
that DF (p0) is invertible, or equivalently the non-vanishing of the determinant of the
Jacobian matrix, is called the no degeneracy condition. Without this condition, the
map may or may not be local invertible, see the examples below.
Nevertheless, it is necessary for the differentiability of the local inverse. At this
point, let us recall the general chain rule. Let G : Rn → Rm and F : Rm → Rl be C1
and their composition H = F ◦ G : Rn → Rl is also C1. We compute the first partial
derivatives of H in terms of the partial derivatives of F and G. Letting G = (g1, · · · ,
gm), F = (f1, · · · , fl) and H = (h1, · · · , hl). From
where I is the identity map. We conclude that
DF −1(q0) = (DF (p0))−1,
in other words, the matrix of the derivative of the inverse map is precisely the
inverse matric of the derivative of the map. So when the inverse map is C1,DF (p0)
must be invertible
Lemma 4.2. Let L be a linear map from Rn to itself given by
Now we prove Theorem 4.1. We may take p0 = F (p0) = 0, for otherwise we could
look at the new function F (x) = F (x+p0)−F (p0) instead of F (x), after noting DF (0)
= DF (p0). First we would like to show that there is a unique solution for the
equation F (x) = y for y near 0. We will use the contraction mapping principle to
achieve our goal. After a further restriction on the size of U , we may assume that F
is C1 with DF (x) invertible at all x ∈ U . For a fixed y, define the map in U by
where we have used the formula
4.2 The Implicit Function Theorem
Next we deduce the implicit function theorem from the inverse function theorem.
Theorem 4.3 (Implicit Function Theorem). Consider C 1-map F : U ! Rm where
U is an open set in Rn _ Rm. Suppose that (p0; q0) 2 U satis_es F(p0; q0) = 0 and
4.3 Curves and Surfaces
A parametric curve is a C1-map from an interval I to R2 which satis_es j0j 6= 0 on
I. In this de_nition a curve is not a geometric object but a map. The condition j 0j 6= 0
ensures that the curve does not collapse into a point. A parametric curve becomes a
geometric curve only if we look at its image. On the other hand, a non-parametric
curve is a subset in R2 satis_es the following condition: For each p 0 = (x0; y0) 2 ,
there exist an open rectangle R = (x1; x2)_(y1; y2) containing p0 such that either for
some
C1-function f from (x1; x2) to (y1; y2) such that \R = f(x; f(x)) : x 2 (x1; x2)g, or for
some C1-function g from (y1; y2) to (x1; x2) such that \ R = f(g(y); y) : y 2 (y 1; y2)g.
In the following we show the essential equivalence of these two de_nitions.
