Math 341 Lecture #23

§4.5: The Intermediate Value Theorem

We saw last time for a continuous f : A → R, that if K ⊆ A is compact, then f (K) is
compact; and this led to the Extreme Value Theorem.
There is another topological property of subsets of R that is preserved by continuous
functions, which will lead to the Intermediate Value Theorem.
Theorem 4.5.2 (Preservation of Connectedness). For a continuous function
f : A → R, if E ⊆ A is connected, then f (E) is connected as well.
Proof. Suppose that E is connnected.
The set f (E) is connected if and only if whenever f (E) = A ∪ B for disjoint nonempty
sets A and B, one of A or B contains a sequence that converges to a point of the other.
Suppose that f (E) = A ∪ B for disjoint nonempty sets A and B.
Let
C = {x ∈ E : f (x) ∈ A} = f −1 (A)
and
D = {x ∈ E : f (x) ∈ B} = f −1 (B),
which are the preimages of A and B by f respectively.
Since A is a nonempty subset of f (E) the set C is a nonempty subset of E: each z ∈ A
is given by z = f (x) for some x ∈ C.
Similarly D is a nonempty subset of E.
The sets C and D are disjoint because, if there is x ∈ C ∩D, then f (x) ∈ A and f (x) ∈ B,
contradicting the disjointness of A and B.
The sets C and D satisfy E = C ∪ D because each z ∈ f (E) = A ∪ B is given by z = f (x)
for x ∈ C or x ∈ D.
The set E is connected, so that one of C or D contains a sequence that converges to a
point of the other.
WLOG, suppose C contains a sequence (xn ) that converges to x ∈ D.
The continuity of f implies that f (xn ) → f (x) where f (xn ) is a sequence in A that
converges to f (x) ∈ B.
Therefore f (E) is connected. 
Recall that a subset of R is connected if and only if it is a (possibly infinite) interval.
We now use this preservation of connectedness by continuous functions to prove the
Intermediate Value Theorem.
Theorem 4.5.1 (Intermediate Value Theorem). For −∞ < a < b < ∞, if
f : [a, b] → R is continuous, and if L ∈ R satisfies f (a) < L < f (b) or f (a) > L > f (b),
then there exists a point c ∈ (a, b) such that f (c) = L.
Proof. The finite length interval [a, b] is connected.
The continuity of f implies by Theorem 4.5.2 that f ([a, b]) is connected.
Hence the image f ([a, b]) is an interval.
The values f (a) and f (b) both belong to f ([a, b]).
When f (a) < f (b), the open interval (f (a), f (b)) is then a subset of f ([a, b]) because the
latter is connected.
An L satisfying f (a) < L < f (b) belongs to (f (a), f (b)) ⊆ f ([a, b]), and so there is
c ∈ [a, b] such that f (c) = L.
Because f (a) < L = f (c) < f (b) we have that c 6= a and c 6= b, so that c ∈ (a, b).
A similar argument holds when f (a) > f (b). .
A typical way the Intermediate Value Theorem is used is to prove the existence of real
roots of polynomials.
Example. Does the polynomial
p(x) = x5 − 5x4 + 11x3 − 12x2 + 7x − 1

have a real root?

To answer this affirmatively, we find x1 and x2 such that p(x1 ) < 0 < p(x2 ).
The choice of x1 = 0 gives p(x1 ) = −1, and the choice of x2 = 10 gives p(x2 ) = 59869.
Because p(x) is continuous, the Intermediate Value Theorem guarantees there is (at least
one) c ∈ (0, 10) such that p(c) = 0.
We may not get the exact value of c (there is no analog of the quadratic formula for
quintic polynomials).
But we do get a good rational approximation through Newton’s method,

p(xn )
x1 = −1, xn+1 = xn − for n ∈ N,
p0 (xn )

which gives c ≈ 0.1999050056.

Although we do not get the exact value of c, the point here is that this irrational number
c exists, and the sequence (xn ) from Newton’s methods converges to c.
Corollary (Existence of Roots). For −∞ < a < b < ∞, if f : [a, b] → R is
continuous with f (a) < 0 < f (b), then there exists c ∈ (a, b) such that f (c) = 0.
You will be asked to complete the following proofs of this Corollary using the Axiom of
Completeness, and using the Nested Interval Property.
Start of Proof using the Axiom of Completeness. Set

K = {x ∈ [a, b] : f (x) ≤ 0}.

The set K is bounded above by b because f (b) > 0, and the set K is nonempty as it
contains a because f (a) < 0.
Thus c = sup K exists by the Axiom of Completeness.
The value of f (c) falls into three cases: f (c) < 0, f (c) > 0, or f (c) = 0.
Use the properties of the least upper bound to eliminate the first two cases.
Start of Proof using the Nested Interval Property. Let I0 = [a, b], and consider this in-
tervals’ midpoint z = (a + b)/2.
If f (z) ≥ 0, then choose a1 = a and b1 = z.
If f (z) < 0, then choose a1 = z and b1 = b.
The interval I1 = [a1 , b1 ] has the property that f (a1 ) < 0 and f (b1 ) ≥ 0.
Now use induction to construct a sequence of nested closed intervals In = [an , bn ] such
that f (an ) < 0 and f (bn ) ≥ 0.
Show that ∩In consists of one point c for which an → c and bn → c.
Then use continuity of f to show that f (c) = 0.
The Intermediate Value Property. Is the converse of the Intermediate Value The-
orem value true? For a function f : [a, b] → R, does for all x, y ∈ [a, b] with x < y, and
for all L between f (x) and f (y), there exist c ∈ (x, y) such that f (c) = L, imply that f
is continuous on [a, b]?
Example. The function
(
sin(1/x) if x 6= 0,
g(x) =
0 if x = 0,

is not continuous at 0.
On the interval [0, 1], for any x, y ∈ [0, 1] with x < y, and any L between g(x) and g(y)
there exists c ∈ (x, y) such that g(x) = L.
So the converse of the Intermediate Value Theorem is false as this counterexample proves.
Definition 4.5.3. A function f : A → R has the intermediate value property on an
interval [a, b] ⊆ A if for all x, y ∈ [a, b] with x < y, and any L between f (x) and f (y),
there exists c ∈ (x, y) such that f (c) = L.
Although the converse of the Intermediate Value Theorem is false in general, it is true if
we impose monotonicity on the function.
A proof of this is requested in a homework problem (4.5.3).