1 The Implicit Function Theorem

Suppose that (a, b) is a point on the curve F (x, y) = 0 where and suppose that this equation can
be solved for y as a function of x for all (x, y) sufficiently near (a, b). Then this part of the curve is
the graph of a function y = ϕ(x) on some interval |x − a| < h with ϕ(a) = b. If ϕ0 (x) exists, we can
compute it by differentiating both sides of the equation F (x, ϕ(x)) = 0 with respect to x to get
∂F ∂F
(x, ϕ(x)) + (x, ϕ(x))ϕ0 (x) = 0
∂x ∂y
∂F
providing that the partial derivatives exist. If ∂y (x, ϕ(x)) 6= 0, we can solve for ϕ0 (x) and obtain
the well known formula
∂F
(x, ϕ(x))
ϕ0 (x) = − ∂F
∂x

∂y (x, ϕ(x))
or, more classically,
∂F
dy ∂x
= − ∂F .
dx ∂y
The precise conditions under which the existence of h and ϕ is assured are furnished by the following
theorem, which is the Implicit Function Theorem for functions of two variables.
Theorem 1. If F (a, b) = 0 and F (x, y) is continuously differentiable on some open disk with center
(a, b) then, if ∂F
∂y (a, b) =
6 0, there exists an h > 0 and a unique function ϕ(x) defined for |x − a| < h
such that ϕ(a) = b and F (x, ϕ(x)) = 0 for |x − a| < h. Moreover, on |x − a| < h, the function ϕ(x)
is continuously differentiable and
∂F
∂x (x, ϕ(x))
ϕ0 (x) = − ∂F
∂y (x, ϕ(x))

There is a corresponding theorem for the case where ∂F∂x (a, b) 6= 0. In this case the curve F (x, y)
is the graph of a function of x = ψ(y) near the point (a, b).
Example. Except for the two √ points (±1, 0), the curve x2 + y 2 = 1 consists of the two continuously
differentiable functions y = ± 1 − x2 , −1 < x < 1. Notice that for F (x, y) = x2 + y 2p we have
∂F 2
∂y = 2y which is zero when y = 0. The points (±1, 0) lie on the two branches x = ± 1 − y ,
−1 < y < 1.
The general Implicit Function Theorem gives condition under which a system of equations
F1 (x1 , . . . , xm , y1 , . . . , yn ) = 0
F2 (x1 , . . . , xm , y1 , . . . , yn ) = 0
...
Fn (x1 , . . . , xm , y1 , . . . , yn ) = 0
can be solved for y1 , . . . , yn as functions of x1 , . . . , xm , say yi = ϕi (x1 , . . . , xm ). Differentiating the
equation
Fi (x1 , . . . , xm , ϕ1 (x1 , . . . , xm ), . . . , ϕn (x1 , . . . , xm ) = 0
with respect to xj we get
∂Fi ∂Fi ∂ϕ1 ∂Fi ∂ϕn
= + ... + .
∂xj ∂y1 ∂xj ∂yn ∂xj
Using the partial Jacobians
 ∂F1  ∂F 
∂F1  ∂F1
∂x1 ··· ∂xm ∂y
1
··· ∂yn
 ∂F2 ··· ∂F2   ∂F1 ··· ∂F2 
 ∂x1 ∂xm   ∂x12 ∂yn 
Dx F =  . ..  , Dy F = 
 . .. 
.
 .. .   .. . 
∂Fn ∂Fn ∂Fn ∂Fn
∂x1 ··· ∂xm ∂y1 ··· ∂yn
These equations with 1 ≤ i ≤ n can be written in the matrix form

Dx F + Dy F Dϕ = 0.

If Dy F is invertible, i.e., if |Dy F | 6= 0, we get

Dϕ = −Dy F −1 Dx F.

Let x = (x1 , . . . , xm ), y = (y1 , . . . , yn ), a = (a1 , . . . , am ), b = (b1 , . . . , bm ) and let

F (x, y) = (F1 (x, y), . . . , Fn (x, y)).

Theorem 2 (Implicit Function Theorem). If F (a, b) = 0 and F (x, y) is continuously differen-

tiable on some open disk with center (a, b) then, if |DyF (a, b)| 6= 0, there exists an h > 0 and a unique
function ϕ(x) = (ϕ1 (x), . . . , ϕn (x)) defined for |x − a| < h such that ϕ(a) = b and F (x, ϕ(x)) = 0
for |x − a| < h. Moreover, on |x − a| < h, the function ϕ(x) is continuously differentiable and

Dϕ(x) = −Dy (x, ϕ(x))−1 Dx (x, ϕ(x)).

Example. Consider the equation F (x, y) = 0 where

F1 (x, y) = x21 + 2x2 + y12 + 2y2 − 8 = 0,

F2 (x, y) = x1 − x22 + y1 − y22 + 3 = 0.

If a = (1, 1), b = (1, 2), we have F (a, b) = 0 and

Œ Œ
Œ2 2 ŒŒ
|Dy F (a, b)| = ŒŒ = −10 6= 0.
1 −4Œ

By the Implicit Function Theorem one has, for x = (x1 , x2 ) sufficiently close to (1, 1),

y = (y1 , y2 ) = (ϕ1 (x), ϕ2 (x)) = ϕ(x)

with ϕ(1, 1) = (1, 2) and

” •−1 ” •
2ϕ1 (x) 2 2x1 2
Dϕ(x) = − .
1 −2ϕ2 (x) 1 −2ϕ2 (x)

An immediate consequence of the Implicit Function Theorem is the following theorem, known as
the Inverse Function Theorem.

Theorem 3 (Inverse Function Theorem). Let y = f (x), where y = (y1 , y2 , · · · , yn ) and x =

(x1 , x2 , · · · , xn ). If DF (a) is invertible, then, for y near b = f (a) and x near a, we have x = g(y)
and Dg(b) = Df (a)−1 .