A function f is defined to be continuous at x0 if the following three conditions hold:
f(x0) is defined;
limx→x0f(x) exists;
limx→x0f(x)=f(x0).
For example, f(x) = x2 + 1 is continuous at 2, since limx→2f(x)=5=f(2). Condition (i) implies that a function can be continuous only at points of its domain. Thus, f(x)=4−x2 is not continuous at 3 because f(3) is not defined.
Let f be a function that is defined on an interval (a, x0) to the left of x0 and/or on an interval (x0, b) to the right of x0. We say that f is discontinuous at x0 if f is not continuous at x0, that is, if one or more of the conditions (i) to (iii) fails.
The discontinuity at 2 in Example 8.1(b) is said to be removable because if we extended the function f by defining its value at x = 2 to be 4, then the extended function g would be continuous at 2. Note that g(x) = x + 2 for all x. The graphs of f(x)=x2−4x−2 and g(x) = x + 2 are identical except at x = 2, where the former has a "hole." (See Fig. 8-2.) Removing the discontinuity consists simply of filling the "hole."
The discontinuity at 2 in Example 8.1(a) is not removable. Redefining the value of f at 2 cannot change the fact that limx→21x−2 does not exist.
We also call a discontinuity of a function f at x0 removable when f(x0) is defined and changing the value of the function at x0 produces a function that is continuous at x0.
The kind of discontinuity shown in Example 8.3 is called a jump discontinuity. In general, a function f has a jump discontinuity at x0 if limx→x0− f(x) and limx→x0+f(x) both exist and limx→x0−f(x)≠limx→x0+f(x). Such a discontinuity is not removable.
Theorem 8.1: Assume that f and g are continuous at x0. Then:
The constant function h(x) = c for all x is continuous at every x0.
cf is continuous at x0, for any constant c. [Recall that cf has the value c ⋅ f(x) for each argument x.]
f + g is continuous at x0.
f − g is continuous at x0.
fg is continuous at x0.
f/g is continuous at x0 if g(x0) ≠ 0.
fn is continuous at x0 if f(x0)n is defined.
These results follow immediately from Theorems 7.1–7.6. For example, (c) holds because
limx→x0 (f(x)+g(x))=limx→x0f(x)+limx→x0g(x)=f(x0)+g(x0)
Theorem 8.2: The identity function I(x) = x is continuous at every x0.
This follows from the fact that limx→x0x=x0.
We say that a function f is continuous on a set A if f is continuous at every point of A. Moreover, if we just say that f is continuous, we mean that f is continuous at every real number.
The original intuitive idea behind the notion of continuity was that the graph of a continuous function was supposed to be "continuous" in the intuitive sense that one could draw the graph without taking the pencil off the paper. Thus, the graph would not contain any "holes" or "jumps." However, it turns out that our precise definition of continuity goes well beyond that original intuitive notion; there are very complicated continuous functions that could certainly not be drawn on a piece of paper.
Theorem 8.3: Every polynomial function
f(x)=anxn+an−1xn−1+⋅⋅⋅+a1x+a0
is continuous.
This is a consequence of Theorems 8.1(a–e) and 8.2.
Theorem 8.4: Every rational function H(x)=f(x)g(x), where f(x) and g(x) are polynomial functions, is continuous on the set of all points at which g(x) ≠ 0.
This follows from Theorems 8.1(f) and 8.3. As examples, the function H(x)=xx2−1 is continuous at all points except 1 and −1, and the function G(x)=x−7x2+1 is continuous at all points (since x2 + 1 is never 0).
We shall use a special notion of continuity with respect to a closed interval [a, b]. First of all, we say that a function f is continuous on the right at a if f(a) is defined and limx→a+f(x) exists, and limx→a+f(x)=f(a). We say that f is continuous on the left at b if f(b) is defined and limx→b−f(x) exists, and limx→b−f(x)=f(b).
Definition: f is continuous on [a, b] if f is continuous at each point on the open interval (a, b), f is continuous on the right at a, and f is continuous on the left at b.
Note that whether f is continuous on [a, b] does not depend on the values of f, if any, outside of [a, b]. Note also that every continuous function (that is, a function continuous at all real numbers) must be continuous on any closed interval. In particular, every polynomial function is continuous on any closed interval.
We want to discuss certain deep properties of continuous functions that we shall use but whose proofs are beyond the scope of this book.
Theorem 8.5 (Intermediate Value Theorem): If f is continuous on [a, b] and f(a) ≠ f(b), then for any number c between f(a) and f(b), there is at least one number x0 in the open interval (a, b) for which f(x0) = c.
Figure 8-4(a) is an illustration of Theorem 8.5. Fig. 8-5 shows that continuity throughout the interval is essential for the validity of the theorem. The following result is a special case of the Intermediate Value Theorem.
Corollary 8.6: If f is continuous on [a, b] and f(a) and f(b) have opposite signs, then the equation f(x) = 0 has at least one root in the open interval (a, b), and therefore, the graph of f crosses the x-axis at least once between a and b. [See Fig. 8-4(b).]
Theorem 8.7 (Extreme Value Theorem): If f is continuous on [a, b], then f takes on a least value m and a greatest value M on the interval.
As an illustration of the Extreme Value Theorem, look at Fig. 8-6(a), where the minimum value m occurs at x = c and the maximum value M occurs at x = d. In this case, both c and d lie inside the interval. On the other hand, in Fig. 8-6(b), the minimum value m occurs at the endpoint x = a and the maximum value M occurs inside the interval. To see that continuity is necessary for the Extreme Value Theorem to be true, consider the function whose graph is indicated in Fig. 8-6(c). There is a discontinuity at c inside the interval; the function has a minimum value at the left endpoint x = a, but the function has no maximum value.
Another useful property of continuous functions is given by the following result.
Theorem 8.8: If f is continuous at c and f(c) > 0, then there is a positive number δ such that whenever c − δ <>x <>c + δ, then f(x) > 0.
This theorem is illustrated in Fig. 8-7. For a proof, see Problem 3.