Theorems on Continuous Functions

Take f(x) = x^3 - x - 1. Compute f(1) = 1 - 1 - 1 = -1 and f(2) = 8 - 2 - 1 = 5. So at x = 1 the function is negative, and at x = 2 it is positive. Somewhere in between, the function must have crossed zero — it must have a root between 1 and 2.

Why "must"? Because f is a polynomial, and polynomials are continuous. A continuous function cannot jump from -1 to 5 without passing through 0. It has to cross. There is no way around it.

That seems obvious when you picture a smooth curve — of course a line going from below zero to above zero has to cross the axis. But making this obvious idea precise, and then using it to prove things about equations you cannot solve by hand, is one of the most powerful moves in analysis. The theorems in this article tell you what continuous functions on closed intervals are guaranteed to do — not what they might do, but what they must.

The Intermediate Value Theorem

The most important theorem in this article is this one.

Read this carefully. The theorem says: if the function starts at f(a) and ends at f(b), and k is any height between those two values, the function must hit that height at least once somewhere between a and b. The function cannot skip over any intermediate value — it has to pass through all of them.

The word at least matters. The theorem does not say there is exactly one c — there might be several. All it promises is that at least one exists.

Why does the IVT hold? A proof sketch

The idea behind the proof uses the completeness of the real numbers — the fact that the real line has no gaps. Here is the argument, stripped down but honest.

Suppose f(a) < k < f(b) (the case f(a) > k > f(b) is symmetric). Define the set

This set is non-empty because a \in S (since f(a) < k), and it is bounded above by b. By the completeness property of the real numbers, S has a least upper bound — call it c.

The claim is that f(c) = k. The proof works by ruling out the other two possibilities.

Case 1: f(c) < k. If f(c) were strictly less than k, then by continuity there is a small interval around c where f stays less than k. That means points slightly to the right of c are still in S — which contradicts c being the least upper bound of S. So this case is impossible.

Case 2: f(c) > k. If f(c) were strictly greater than k, then by continuity there is a small interval around c where f stays greater than k. That means no point in this interval belongs to S, so the least upper bound of S would be strictly less than c — again a contradiction. So this case is also impossible.

The only remaining possibility is f(c) = k. And since f(a) < k < f(b), the point c cannot be a or b, so c \in (a, b).

Notice what made this work: the completeness of \mathbb{R} (guaranteeing c exists) and the continuity of f (ruling out the two bad cases). Remove either one and the theorem can fail. On the rationals, for instance, the function f(x) = x^2 - 2 is continuous and satisfies f(1) = -1 < 0 and f(2) = 2 > 0, but there is no rational number c with f(c) = 0, because \sqrt{2} is irrational.

Bolzano's theorem — the root-finding corollary

The most common application of the IVT is a special case with its own name.

This is the IVT with k = 0. If f(a) < 0 < f(b) (or f(a) > 0 > f(b)), then 0 is between f(a) and f(b), so the function must hit 0 somewhere in between.

Bolzano's theorem is the rigorous version of the opening observation: f(1) = -1 and f(2) = 5 have opposite signs, so f(x) = x^3 - x - 1 has a root between 1 and 2.

Narrowing the root: bisection

Bolzano's theorem does more than tell you a root exists — it gives you a method to find it to any desired accuracy. This is the bisection method.

You know f(1) < 0 and f(2) > 0. Check the midpoint: f(1.5) = 3.375 - 1.5 - 1 = 0.875 > 0. So the root is between 1 and 1.5 (since f(1) < 0 and f(1.5) > 0). Check the new midpoint: f(1.25) = 1.953125 - 1.25 - 1 = -0.296875 < 0. So the root is between 1.25 and 1.5. Each step halves the interval. After 10 steps, the interval has width 1/2^{10} < 0.001 — you have the root to three decimal places.

This is a genuine numerical algorithm, used in practice when no closed-form solution exists. Every step applies Bolzano's theorem to the new, smaller interval. The algorithm converges because each interval is nested inside the previous one and the widths shrink to zero.

The Extreme Value Theorem

The second major theorem about continuous functions on closed intervals concerns maxima and minima.

In words: a continuous function on a closed interval always has a highest point and a lowest point. It cannot escape to infinity, and it cannot approach a value without reaching it.

Both hypotheses — continuous and closed interval — are essential. Drop either one and the conclusion can fail.

• Open interval, continuous function. Take f(x) = 1/x on (0, 1]. The function is continuous on this interval, but as x \to 0^+, f(x) \to \infty. There is no maximum. The problem is that the interval is not closed — the endpoint 0 is missing, and the function blows up as it approaches the missing endpoint.

• Closed interval, discontinuous function. Define f(x) = x for x \in [0, 1) and f(1) = 0. This function is discontinuous at x = 1. On [0, 1] the values of f get arbitrarily close to 1 (take x = 0.999, 0.9999, \ldots) but never actually reach 1. There is no maximum.

The EVT says that if you have both conditions — continuity and a closed interval — these pathologies cannot happen. The function is trapped on a closed, bounded set, and continuity prevents it from "almost reaching" a value without hitting it.

Why the EVT matters

The EVT is the theoretical backbone of optimisation. When a problem says "find the maximum value of f(x) on [a, b]," it is the EVT that guarantees that maximum actually exists. Without this guarantee, you could spend forever looking for something that isn't there.

In practice, once you know a maximum exists (by the EVT), you find it by checking the critical points inside (a, b) — points where f'(x) = 0 or f' does not exist — and comparing their values with f(a) and f(b). The largest wins. This is the closed interval method, and it works precisely because the EVT says there is a winner.

Putting the theorems to work

Common confusions

• "If f(a) and f(b) have the same sign, there is no root." Not necessarily. The function might dip below zero (or above) and come back. All you can say is that Bolzano's theorem does not apply — you cannot conclude a root exists by this method. But one might exist anyway.

• "The IVT finds the root." The IVT proves a root exists. It does not give you a formula for the root. To actually find the root, you use the bisection method (or Newton's method, once you know derivatives) — the IVT provides the guarantee; the algorithm provides the approximation.

• "The EVT applies to any continuous function." Only on a closed, bounded interval. On an open interval like (0, 1) or an unbounded domain like [0, \infty), the conclusion can fail. A continuous function on (0, 1) can approach \infty near 0, and a continuous function on [0, \infty) can grow without bound. The closed-and-bounded condition is the safety net.

• "Continuous means the function stays bounded." Only on a closed interval. The function f(x) = 1/x is continuous on (0, 1) but unbounded. Continuity alone does not control size; you need the interval to be closed and bounded to trap the function.

• "If f achieves its maximum at c, then f'(c) = 0." Only if c is in the interior of the interval. If c is an endpoint, the derivative might not be zero there — the function might be climbing right up to the edge. The closed interval method checks endpoints precisely for this reason.

Going deeper

The role of completeness

The IVT proof above used the least upper bound property (also called the completeness axiom): every non-empty subset of \mathbb{R} that is bounded above has a least upper bound (supremum) in \mathbb{R}. This is the defining property that separates \mathbb{R} from \mathbb{Q}.

On the rationals, completeness fails, and so does the IVT. Define f(x) = x^2 - 2 on the rational numbers. Then f(1) = -1 < 0 and f(2) = 2 > 0, and f is continuous (as a function on \mathbb{Q}), but there is no rational c with f(c) = 0. The root \sqrt{2} exists in \mathbb{R} but not in \mathbb{Q}. The "gap" in the rationals at \sqrt{2} is exactly the kind of gap that completeness fills.

The IVT is, in a deep sense, equivalent to completeness. If you assume the IVT as an axiom, you can prove the least upper bound property from it. The two statements are saying the same thing in different languages: the real line has no holes.

Proof sketch of the EVT

The EVT requires two steps. First, show that f is bounded on [a, b] — that is, |f(x)| \leq M for some number M and all x \in [a, b]. Second, show that f actually attains its supremum.

Boundedness. Suppose f is unbounded. Then for each positive integer n, there exists x_n \in [a, b] with |f(x_n)| > n. The sequence (x_n) lives in [a, b], which is closed and bounded, so by the Bolzano-Weierstrass theorem it has a convergent subsequence x_{n_k} \to c \in [a, b]. Since f is continuous, f(x_{n_k}) \to f(c), which is a finite number. But |f(x_{n_k})| > n_k \to \infty, a contradiction. So f is bounded.

Attainment. Let M = \sup\{f(x) : x \in [a, b]\}. For each n, there exists x_n \in [a, b] with f(x_n) > M - 1/n. Again by Bolzano-Weierstrass, a subsequence x_{n_k} \to d \in [a, b]. By continuity, f(d) = \lim f(x_{n_k}) = M. So f attains its supremum at d. An identical argument for the infimum shows that f attains its minimum.

Applications to equations

The IVT is a powerful existence tool. Here are two classical applications that appear routinely in competitive exams.

Fixed-point theorem. If f : [0, 1] \to [0, 1] is continuous, then f has a fixed point — a value c with f(c) = c. To prove this, define g(x) = f(x) - x. Then g(0) = f(0) - 0 = f(0) \geq 0 (since f maps into [0, 1]) and g(1) = f(1) - 1 \leq 0. If either is zero, you are done. Otherwise g(0) > 0 and g(1) < 0, so by the IVT there exists c \in (0, 1) with g(c) = 0, i.e., f(c) = c.

Odd-degree polynomials always have a real root. If p(x) = x^n + a_{n-1}x^{n-1} + \cdots + a_0 with n odd, then p(x) \to +\infty as x \to +\infty and p(x) \to -\infty as x \to -\infty. So for large enough R, p(R) > 0 and p(-R) < 0. By the IVT, p has a root in [-R, R]. This is why every cubic, every quintic, every odd-degree polynomial has at least one real root — but an even-degree polynomial like x^2 + 1 can avoid the axis entirely.

Where this leads next

These theorems are the foundation for several important ideas that follow.

• Differentiability — once you know a function is continuous, the next question is whether it has a well-defined tangent at every point. Differentiability is the stronger property.
• Differentiation — Introduction — derivatives use limits, and the results in this article underpin many of the existence arguments in differential calculus.
• Properties of Continuous Functions — the algebraic rules (sums, products, composites of continuous functions are continuous) that you use to verify continuity before applying the IVT or EVT.
• Real Numbers — Properties — the completeness of the reals, which is the logical engine behind both the IVT and the EVT.
• Types of Discontinuity — what happens when continuity fails and these theorems no longer apply.