The Mean Value Theorem

18 min read

Textbook Reference

Both OpenStax sources are free and openly licensed.

Key idea

The Mean Value Theorem says something you already believe about driving: if your average speed over a trip was $90$ kilometers per hour, then at some exact instant your speedometer read exactly $90$.

Here is the precise version. Take a function that is smooth (no breaks, no corners) over an interval $[a, b]$. Draw the straight line connecting its two endpoints; the slope of that line is the average rate of change. The theorem promises there is at least one point inside the interval where the tangent line, the instantaneous rate of change, is parallel to that connecting line. The instantaneous rate equals the average rate, somewhere.

The simplest case to picture first is when the two endpoints sit at the same height. Then the connecting line is flat, and the theorem says the function must level off (have a horizontal tangent) somewhere in between. That special case has its own name, Rolle’s Theorem, and the general theorem is just Rolle’s picture with the whole graph tilted.

Why does a calculus course care about this? Because it is the bridge from the derivative back to the function. The derivative is built from the function; the Mean Value Theorem lets you run the logic backward and deduce facts about the function from facts about its derivative. Almost every later result, including “a positive derivative means the function is increasing” and “two functions with the same derivative differ by a constant,” is the Mean Value Theorem in disguise.

Prerequisite Check

Before this lesson, make sure you can do all of the following:

• State what it means for a function to be continuous on a closed interval (review Continuity at a Point)
• State what it means for a function to be differentiable on an open interval (review Differentiability)
• Compute the slope of a secant line through two points: $\dfrac{f(b) - f(a)}{b - a}$
• Compute a derivative of a polynomial and evaluate it at a point
• Solve a simple equation for an unknown $c$, including discarding solutions outside an interval

If “continuous on the closed interval but differentiable on the open interval” sounds like hair-splitting, slow down on it. The hypotheses are exactly where this theorem is misapplied.

Quick Reference

Rolle’s Theorem. If $f$ is continuous on $[a, b]$, differentiable on $(a, b)$, and $f(a) = f(b)$, then there is a number $c$ in $(a, b)$ with \[ f'(c) = 0. \]

The Mean Value Theorem. If $f$ is continuous on $[a, b]$ and differentiable on $(a, b)$, then there is a number $c$ in $(a, b)$ with \[ f'(c) = \frac{f(b) - f(a)}{b - a}, \qquad \text{equivalently} \qquad f(b) - f(a) = f'(c)(b - a). \]

Plain reading. The instantaneous rate of change at some interior point equals the average rate of change over the whole interval. Geometrically, some tangent line is parallel to the secant line through the endpoints.

The two hypotheses, stated carefully. Continuous on the closed interval $[a, b]$ (endpoints included); differentiable on the open interval $(a, b)$ (endpoints not required). If either fails, the conclusion can fail.

The headline consequences.

• If $f'(x) = 0$ on an interval, then $f$ is constant there.
• If $f'(x) = g'(x)$ on an interval, then $f$ and $g$ differ by a constant.
• The sign of $f'$ tells you whether $f$ is increasing or decreasing (the next section).

Key Concepts

1. Rolle’s Theorem: The Flat Case

Start with the easiest picture. Suppose a smooth function has equal values at the two ends of an interval. Then it must turn around somewhere in between, and at the turning point the tangent is horizontal.

Rolle’s Theorem. If $f$ is continuous on $[a, b]$, differentiable on $(a, b)$, and $f(a) = f(b)$, then there exists $c$ in $(a, b)$ with $f'(c) = 0$. (Plain gloss: a smooth curve that returns to its starting height must have a horizontal tangent somewhere between.)

The reason is short. If $f$ is constant, every point has $f'(c) = 0$. If $f$ rises above the endpoint height somewhere, it attains a maximum at some interior point $c$ (this is guaranteed by the Extreme Value Theorem, since $f$ is continuous on a closed interval), and at an interior maximum of a differentiable function the derivative is zero. The symmetric argument handles the case where $f$ dips below. Either way, some interior $c$ has $f'(c) = 0$.

2. Rolle’s Theorem at Work: Counting Roots

Rolle’s Theorem is the standard tool for proving a function has exactly one root: existence comes from one theorem, uniqueness from Rolle.

Example 1. Prove that $x^3 + x - 1 = 0$ has exactly one real solution. (This is Stewart 3.2, Example 2.)

Goal. Show a root exists, then show there cannot be two.

Existence. Let $f(x) = x^3 + x - 1$. Then $f(0) = -1 < 0$ and $f(1) = 1 > 0$. Since $f$ is continuous, the Intermediate Value Theorem guarantees a root between $0$ and $1$.

Uniqueness. Suppose, for contradiction, there were two roots $a$ and $b$. Then $f(a) = f(b) = 0$, and $f$ is continuous and differentiable everywhere, so Rolle’s Theorem gives a $c$ between them with $f'(c) = 0$. But \[ f'(x) = 3x^2 + 1 \geq 1 \quad \text{for all } x, \] so $f'$ is never zero. That contradiction means there cannot be two roots.

Conclusion: the equation has exactly one real solution.

Recap. The pattern is reusable: the Intermediate Value Theorem supplies a root, and Rolle’s Theorem (via a derivative that never vanishes) forbids a second. A function whose derivative never changes sign can cross any level at most once.

3. The Mean Value Theorem: Tilting Rolle’s Picture

The general theorem drops the requirement $f(a) = f(b)$. Now the connecting secant line is tilted, and the conclusion tilts with it: some tangent line is parallel to that secant.

The Mean Value Theorem. If $f$ is continuous on $[a, b]$ and differentiable on $(a, b)$, then there is a number $c$ in $(a, b)$ with \[ f'(c) = \frac{f(b) - f(a)}{b - a}. \] (Plain gloss: the steepness of the curve at some interior point matches the average steepness across the interval.)

The proof is Rolle’s Theorem applied to a cleverly chosen helper function. Subtract the secant line from $f$: \[ h(x) = f(x) - \left[\,f(a) + \frac{f(b) - f(a)}{b - a}(x - a)\,\right]. \] This $h$ is continuous and differentiable wherever $f$ is, and direct computation gives $h(a) = 0$ and $h(b) = 0$, so $h(a) = h(b)$. Rolle’s Theorem then provides a $c$ with $h'(c) = 0$. Since \[ h'(x) = f'(x) - \frac{f(b) - f(a)}{b - a}, \] setting $h'(c) = 0$ gives exactly $f'(c) = \dfrac{f(b) - f(a)}{b - a}$.

So the Mean Value Theorem is not a new idea; it is Rolle’s Theorem after subtracting off the slant. Seeing that connection once means you never have to memorize the proof separately.

4. Finding the Guaranteed Point

The theorem promises a point $c$ exists. For a concrete function you can find it by solving $f'(c) = \dfrac{f(b)-f(a)}{b-a}$.

Example 2. Find the value of $c$ guaranteed by the Mean Value Theorem for $f(x) = x^3 - x$ on $[0, 2]$. (This is Stewart 3.2, Example 3.)

Goal. Compute the average rate of change, set the derivative equal to it, and solve for $c$ in the interval.

Since $f$ is a polynomial, it is continuous and differentiable everywhere, so the theorem applies. The average rate of change is \[ \frac{f(2) - f(0)}{2 - 0} = \frac{6 - 0}{2} = 3. \] The derivative is $f'(x) = 3x^2 - 1$. Set it equal to $3$: \[ 3c^2 - 1 = 3 \implies 3c^2 = 4 \implies c^2 = \frac{4}{3} \implies c = \pm\frac{2}{\sqrt{3}}. \] The interval is $(0, 2)$, so discard the negative root: \[ c = \frac{2}{\sqrt{3}} \approx 1.155. \]

Boxed answer: $c = \dfrac{2}{\sqrt{3}}$. At this point the tangent line to $y = x^3 - x$ is parallel to the secant line through $(0, 0)$ and $(2, 6)$.

Recap. Two checks save points. First, confirm the hypotheses hold before solving. Second, keep only the values of $c$ that lie strictly inside $(a, b)$; the theorem makes no promise about the endpoints.

5. Using the Theorem to Bound a Function

The deeper use is not finding $c$, but deducing a fact about $f$ from a fact about $f'$. The equivalent form $f(b) - f(a) = f'(c)(b - a)$ turns a bound on the derivative into a bound on the function.

Example 3. Suppose $f(0) = -3$ and $f'(x) \leq 5$ for all $x$. How large can $f(2)$ possibly be? (This is Stewart 3.2, Example 5.)

Goal. Apply the Mean Value Theorem on $[0, 2]$ and use the bound on $f'$.

Since $f$ is differentiable everywhere, it is continuous, so the theorem applies on $[0, 2]$. There is a $c$ with \[ f(2) - f(0) = f'(c)(2 - 0), \quad \text{so} \quad f(2) = -3 + 2f'(c). \] Because $f'(c) \leq 5$, we have $2f'(c) \leq 10$, hence \[ f(2) \leq -3 + 10 = 7. \]

Boxed answer: the largest possible value of $f(2)$ is $7$.

Recap. This is the theorem’s real power: with no formula for $f$, only a starting value and a speed limit on its derivative, the Mean Value Theorem pins down how far $f$ can travel. A bounded rate of change bounds the total change.

6. The Consequences That Power the Rest of Calculus

Two corollaries follow immediately and are used constantly.

Constant-derivative corollary. If $f'(x) = 0$ for all $x$ in an interval, then $f$ is constant on that interval. (Proof: for any two points, the Mean Value Theorem gives $f(x_2) - f(x_1) = f'(c)(x_2 - x_1) = 0$, so all values are equal.)

Equal-derivatives corollary. If $f'(x) = g'(x)$ for all $x$ in an interval, then $f(x) = g(x) + C$ for some constant $C$. (Proof: apply the previous corollary to $f - g$, whose derivative is zero.)

This second corollary is the reason every antiderivative carries a “$+C$”: two functions with the same derivative can differ only by a constant, so the antiderivative of a function is a whole family, $F(x) + C$.

Important: the conclusion needs an interval. The constant-derivative corollary fails if the domain is split into pieces. The function that equals $1$ for $x > 0$ and $-1$ for $x < 0$ has derivative $0$ everywhere it is defined, yet it is not constant. There is no contradiction, because its domain (all $x$ except $0$) is not a single interval. Always check that the domain is one connected interval before concluding “constant.”

Common Errors Summary

Leveled Practice

Level 1 -- Direct Application

Problem 1. Verify that $f(x) = 2x^2 - 4x + 5$ satisfies the hypotheses of Rolle’s Theorem on $[-1, 3]$, then find all $c$ that satisfy the conclusion.

Problem 2. Find all $c$ guaranteed by the Mean Value Theorem for $f(x) = 2x^2 - 3x + 1$ on $[0, 2]$.

Problem 3. A car travels $180$ kilometers in $2$ hours. Explain, using the Mean Value Theorem, why the speedometer must have read exactly $90$ kilometers per hour at some instant.

Level 2 -- Multiple Steps

Problem 4. Find all $c$ guaranteed by the Mean Value Theorem for $f(x) = \sqrt{x}$ on $[0, 4]$.

Problem 5. Let $f(x) = (x - 3)^{-2}$. Show that there is no $c$ in $(1, 4)$ with $f(4) - f(1) = f'(c)(4 - 1)$, and explain why this does not contradict the Mean Value Theorem.

Problem 6. Suppose $f(1) = 10$ and $f'(x) \geq 2$ for all $x$ in $[1, 4]$. How small can $f(4)$ possibly be?

Level 3 -- Deeper Problems

Problem 7. Use the Mean Value Theorem to prove the inequality $|\sin a - \sin b| \leq |a - b|$ for all real numbers $a$ and $b$.

Problem 8. Two runners start a race at the same time and finish in a tie. Prove that at some instant during the race they have exactly the same speed.

Problem 9. Show that the equation $x^3 - 15x + c = 0$ has at most one solution in the interval $[-2, 2]$, for any constant $c$.

Mastery Checklist

You have mastered this skill when you can do all of the following without referring to notes:

• State Rolle’s Theorem and the Mean Value Theorem, including both hypotheses, from memory
• Explain why continuity is required on the closed interval but differentiability only on the open interval
• Find the guaranteed value of $c$ for a given function by solving $f'(c) = \frac{f(b)-f(a)}{b-a}$
• Discard values of $c$ that fall outside the open interval
• Use Rolle’s Theorem with a non-vanishing derivative to prove a function has at most one root
• Use the form $f(b) - f(a) = f'(c)(b-a)$ to bound a function from a bound on its derivative
• State the constant-derivative and equal-derivatives corollaries and explain why each needs an interval
• Recognize that the “$+C$” on every antiderivative comes from the equal-derivatives corollary

Mental Model

Think of the Mean Value Theorem as a promise that the instantaneous catches up to the average.

Over any smooth stretch, the average rate of change is a single number, the slope of the line connecting the endpoints. The instantaneous rate of change, the derivative, varies from point to point. The theorem says these two cannot stay apart: at some interior point the instantaneous rate exactly equals the average rate.

Three things follow from the picture:

• It is an existence promise, not a recipe. The theorem guarantees a point $c$ exists; it does not single one out, and there may be several. For a specific function you find $c$ by solving an equation, but the theorem’s job is only to assert that a solution lives inside the interval.
• The hypotheses are load-bearing, not decoration. A corner (no derivative) or a break (no continuity) lets the curve dodge the conclusion. The absolute-value function on $[-1, 1]$ has average rate $0$ but no horizontal tangent, precisely because of its corner at the origin.
• It runs the derivative backward. The derivative is built forward from the function. The Mean Value Theorem is the gear that lets you reason in reverse: from a fact about the slope to a fact about the function. That reversal is why a positive derivative forces an increasing function, why a zero derivative forces a constant, and why antiderivatives are unique only up to a constant.

Connections

Within Applications of Differentiation (Chapter 3)

• Rolle’s Theorem: The flat-secant special case, and the engine of the Mean Value Theorem’s own proof. Rolle handles “same height forces a horizontal tangent”; the Mean Value Theorem tilts that to “any secant has a parallel tangent.”
• The increasing/decreasing test: The very next result. If $f' > 0$ on an interval then $f$ is increasing there, and the proof is the Mean Value Theorem applied to any two points. The Mean Value Theorem is the foundation that test stands on.
• The first derivative test and curve sketching: Knowing where $f'$ is positive or negative, justified by the Mean Value Theorem, tells you where $f$ rises and falls, which locates maxima and minima and shapes the whole graph.

Toward later calculus (MATH161 and beyond)

• Antiderivatives and the constant of integration: The equal-derivatives corollary says two functions with the same derivative differ by a constant. That is exactly why an antiderivative is written $F(x) + C$, a fact you will rely on throughout integration.
• The Fundamental Theorem of Calculus: The proof that the definite integral can be computed from an antiderivative leans on the Mean Value Theorem (and its integral cousin, the Mean Value Theorem for Integrals). The bridge from derivatives to functions you build here is reused to bridge integrals to antiderivatives.
• Error bounds in approximation: Taylor’s theorem with its remainder, and the error estimates for numerical methods, are quantitative descendants of the Mean Value Theorem. The idea that a bounded derivative bounds the change is what makes those error formulas possible.

Audience Notes

For students who find math intimidating: You do not need the proof to use this theorem. Hold onto the driving picture: average speed of $90$ means you hit exactly $90$ at some instant. To solve a problem, write the average rate of change as a fraction, set the derivative equal to it, and solve for $c$, then throw away any answer outside the interval. That procedure handles every routine problem in this section.

For career-focused students: This theorem is the formal guarantee behind a great deal of engineering reasoning: if a sensor reports a total change and you know a bound on the rate, you can certify how the underlying quantity behaved without measuring it continuously. Bounding total change from a rate limit is exactly Example 3, and it underlies safety arguments in control systems.

For gifted and curious students: Explore the Cauchy Mean Value Theorem, which compares two functions’ rates simultaneously and produces $\dfrac{f'(c)}{g'(c)} = \dfrac{f(b)-f(a)}{g(b)-g(a)}$. It generalizes the theorem here and is the key step in proving L’Hopital’s rule, tying this section directly to the limits-at-infinity material.

For PhD-track students: The Mean Value Theorem is genuinely a real-variable phenomenon; its exact analogue fails for complex-valued and vector-valued functions, where only an inequality (the mean value inequality $|f(b) - f(a)| \leq \sup |f'| \cdot |b - a|$) survives. Understanding why the equality breaks down, and what replaces it, is an early lesson in how analysis on $\mathbb{R}$ differs from analysis on $\mathbb{R}^n$ and $\mathbb{C}$.

Back to Applications of Differentiation | Related: Rolle’s Theorem | Next: Increasing and Decreasing Test