Mean Value Theorem and Rolle's Theorem

⚡ Key Takeaways

Rolle's Theorem: If a function is continuous on $[a,b]$ , differentiable on $(a,b)$ , and $f(a)=f(b)$ , then there exists at least one point $c$ in $(a,b)$ where $f'(c)=0$ . Mean Value Theorem (MVT): Under the same continuity/differentiability conditions, there exists $c$ where the instantaneous slope equals the average slope: $f'(c)=\frac{f(b)-f(a)}{b-a}$ . Where it applies: smooth curves on closed intervals. Who/when: Rolle (1691); Lagrange/Cauchy formalized the MVT in the late 1700s–1800s.

💡 The Core Insight

Over any interval, the average rate of change (the secant slope) must be achieved exactly by the instantaneous rate of change (a tangent slope) at some interior point. If you drive 60 miles in one hour, at some instant your speedometer read exactly 60 mph. Rolle's Theorem is the special case where the average rate of change is zero: if you return to your starting height, somewhere your motion momentarily flattened out.

DefinitionConditions of Continuity and Differentiability

A function $f$ satisfies the MVT hypotheses on $[a,b]$ if:

$f$ is continuous on the closed interval $[a,b]$ (no breaks, jumps, or holes, including at the endpoints).
$f$ is differentiable on the open interval $(a,b)$ (a well-defined tangent slope at every interior point — no corners or vertical tangents inside).

Note the asymmetry: continuity is required on the closed interval, but differentiability only on the open interval. This lets the endpoints be sharp corners.

TheoremRolle's Theorem

If $f$ is continuous on $[a,b]$ , differentiable on $(a,b)$ , and $f(a)=f(b)$ , then there exists at least one $c \in (a,b)$ such that

1

\quad f'(c)=0.

Derivation hint: An interior extremum of a differentiable function has zero derivative; combine with the Extreme Value Theorem. Here $c$ is the (possibly non-unique) interior point, and $f'(c)$ is the derivative evaluated there.

TheoremMean Value Theorem

If $f$ is continuous on $[a,b]$ and differentiable on $(a,b)$ , then there exists at least one $c \in (a,b)$ such that

2

\quad f'(c)=\frac{f(b)-f(a)}{b-a}.

Derivation hint: Subtract the secant line from $f$ to create a helper function satisfying Rolle's hypotheses, then apply Rolle. The symbols: $a,b$ are the interval endpoints; $f(a),f(b)$ the function values there; the fraction is the secant slope (average rate of change); $f'(c)$ is the tangent slope (instantaneous rate of change) at the guaranteed point $c$ . Rearranged, equation [2] gives the often-used form

3

\quad f(b)-f(a)=f'(c)\,(b-a).

Derivation hint: Multiply both sides of [2] by $(b-a)$ .

ProofRolle's Theorem

Step 1 — Extreme Value Theorem. Since $f$ is continuous on the closed bounded interval $[a,b]$ , it attains a maximum value $M$ and a minimum value $m$ somewhere on $[a,b]$ .

Step 2 — Two cases. Case A: $M=m$ . Then $f$ is constant on $[a,b]$ , so $f'(x)=0$ for every interior $x$ ; choose any such $c$ and we are done. Case B: $M \ne m$ . Since $f(a)=f(b)$ , at least one of the extreme values $M$ or $m$ is not equal to the common endpoint value. That extreme value must therefore be attained at some interior point $c \in (a,b)$ .

Step 3 — Fermat's interior extremum theorem. At an interior maximum or minimum of a differentiable function, the derivative must be zero. (If $f'(c)>0$ the function is rising through $c$ ; if $f'(c)<0$ it is falling — either way $c$ couldn't be a peak or valley.) Hence

4

\quad f'(c)=0. \qquad \blacksquare

Derivation hint: Examine the sign of the one-sided difference quotients at the extremum and take the limit.

∎

ProofMean Value Theorem from Rolle

Step 1 — Build the secant line. The line through $(a,f(a))$ and $(b,f(b))$ is

5

\quad L(x)=f(a)+\frac{f(b)-f(a)}{b-a}\,(x-a).

Derivation hint: Point-slope form using the two endpoints.

Step 2 — Define the gap function. Let

6

\quad g(x)=f(x)-L(x).

Derivation hint: Measure the vertical distance between the curve and its secant line. This $g$ is continuous on $[a,b]$ and differentiable on $(a,b)$ (a difference of such functions).

Step 3 — Check Rolle's third condition. At the endpoints, $f$ and $L$ agree, so $g(a)=g(b)=0$ . Rolle's Theorem applies to $g$ .

Step 4 — Apply Rolle. There exists $c\in(a,b)$ with $g'(c)=0$ . Differentiating [6]:

7

\quad g'(x)=f'(x)-\frac{f(b)-f(a)}{b-a}.

Derivation hint: The derivative of the linear term $L(x)$ is just its constant slope.

Step 5 — Conclude. Setting $g'(c)=0$ gives $f'(c)=\dfrac{f(b)-f(a)}{b-a}$ . $\qquad \blacksquare$

∎

mathematician · Michel Rolle | 1652–1719

A largely self-taught French mathematician, Rolle published the theorem now bearing his name in 1691 within a treatise on solving equations (Méthode pour résoudre les égalitez). Ironically, Rolle was an outspoken critic of the new infinitesimal calculus of Newton and Leibniz, calling it a "collection of ingenious fallacies." His result was originally a purely algebraic statement about the roots of polynomials — that between two roots of a polynomial lies a root of its derivative. Only later was it recast in the analytic, derivative-based language we use today. The name "Rolle's theorem" was popularized by the German mathematician Moritz Wilhelm Drobisch (1834) and the Italian Giusto Bellavitis.

🧠 Intuition

Imagine the graph of $f$ as a hilly road and the secant line as a straight ramp connecting your start and end points. The MVT says: somewhere along the road, the road's slope is exactly the ramp's slope. Visually, slide the secant line upward (or downward) keeping it parallel to itself; the last place it touches the curve before leaving it is a point where the tangent is parallel to the secant — that touching point is $c$ . For Rolle, the ramp is horizontal (because the endpoints are level). The road that returns to its starting height must have a highest or lowest point in between, and at that peak or valley the road is momentarily flat — that's where $f'(c)=0$ . The deepest takeaway: the average behavior of a smooth function is literally realized as its instantaneous behavior somewhere. This is why the MVT is the workhorse that converts derivative facts into function facts.

Figure

The Mean Value Theorem: a tangent (gold) parallel to the secant (dashed)

ExampleApplying the MVT to a Parabola

Given: $f(x)=x^2$ on the interval $[1,3]$ . Find the guaranteed point $c$ .

Step 1 — Verify hypotheses. $f$ is a polynomial, hence continuous on $[1,3]$ and differentiable on $(1,3)$ . ✓

Step 2 — Compute the secant slope.

\frac{f(3)-f(1)}{3-1}=\frac{9-1}{2}=\frac{8}{2}=4.

Step 3 — Set the derivative equal to it. Since $f'(x)=2x$ :

2c=4 \quad\Longrightarrow\quad c=2.

Step 4 — Check $c$ is interior. $2 \in (1,3)$ . ✓

\boxed{c=2}

For a parabola, the MVT point is always the midpoint of the interval — a tidy special property of quadratics.

ExampleAverage Speed and the MVT (Physics)

Given: A car's position is $s(t)=4t^2+3t$ (meters) over $t\in[0,5]$ seconds. Show that at some instant the car's instantaneous velocity equals its average velocity, and find that instant.

Step 1 — Average velocity (secant slope).

v_{\text{avg}}=\frac{s(5)-s(0)}{5-0}=\frac{(100+15)-(0)}{5}=\frac{115}{5}=23\ \text{m/s}.

Step 2 — Instantaneous velocity. $v(t)=s'(t)=8t+3$ m/s.

Step 3 — Apply MVT. Set $v(c)=v_{\text{avg}}$ :

8c+3=23 \quad\Longrightarrow\quad 8c=20 \quad\Longrightarrow\quad c=2.5\ \text{s}.

Step 4 — Check. $2.5 \in (0,5)$ . ✓

\boxed{c=2.5\ \text{s},\quad v(2.5)=23\ \text{m/s}}

The car's speedometer read exactly the average speed at $t=2.5$ seconds.

⚠️ The Theorem Promises Existence, Not Uniqueness or Location

The MVT and Rolle's Theorem guarantee that at least one $c$ exists — there may be several, and the theorem gives no formula for finding them in general. Equally important: every hypothesis is necessary. Dropping continuity on the closed interval, or differentiability even at a single interior point, can destroy the conclusion. A common error is applying the MVT to functions with corners (like $|x|$ ) or vertical asymptotes inside the interval. Another error: thinking $f'(c)=0$ means $c$ is a maximum or minimum — Rolle only says the tangent is horizontal, which could occur at an inflection point too.

CounterexampleWhy Differentiability Cannot Be Dropped

Consider $f(x)=|x|$ on $[-1,1]$ . The endpoints satisfy $f(-1)=f(1)=1$ , and $f$ is continuous everywhere, so Rolle's would predict some $c\in(-1,1)$ with $f'(c)=0$ . But

8

\quad f'(x)=\begin{cases} -1 & x<0\\ +1 & x>0 \end{cases}

Derivation hint: Differentiate each linear piece; the slopes are never zero. The derivative is $\pm 1$ everywhere it exists and is undefined at $x=0$ (a corner). There is no point where $f'(c)=0$ . Rolle's Theorem fails — precisely because $f$ is not differentiable at the interior point $x=0$ . This shows differentiability on the whole open interval is indispensable. A second counterexample for continuity: $g(x)=x-\lfloor x\rfloor$ on $[0,1]$ resets at the endpoint with a jump; it is not continuous at $x=1$ , and again no horizontal tangent need exist.

connection

Rolle's Theorem and the MVT sit at the hub of single-variable calculus:

They specialize and generalize each other: Rolle is the MVT with $f(a)=f(b)$ ; the MVT is Rolle applied to a tilted "gap function."
The MVT is the engine behind the Increasing and Decreasing Functions test: if $f'>0$ on an interval then $f$ is increasing there, proven via equation [3].
It underwrites the Fundamental Theorem of Calculus, where the MVT (and its integral cousin) connects antiderivatives to accumulated change.
It generalizes to the Cauchy Mean Value Theorem (the "ratio" form), which in turn powers L'Hôpital's Rule for indeterminate limits.
The result $f'\equiv 0 \Rightarrow f$ constant — foundational for Antiderivatives and Indefinite Integrals — follows directly from equation [3].

🔬 Where this appears

⏳ History

- c. 1300s — Indian mathematician Parameshvara of the Kerala school states a form of the mean value theorem in commentaries, anticipating later results by centuries.

- 1691 — Michel Rolle publishes the algebraic root-spacing result for polynomials in Méthode pour résoudre les égalitez.

- 1797 — Joseph-Louis Lagrange states a version of the MVT in Théorie des fonctions analytiques, framing it through power-series remainders.

- 1806 — André-Marie Ampère attempts an early proof of the MVT as a foundational principle of calculus.

- 1823 — Augustin-Louis Cauchy, in his Résumé des leçons, gives a rigorous treatment and introduces the generalized (Cauchy) Mean Value Theorem.

- 1834 — M. W. Drobisch and later Giusto Bellavitis popularize the name "Rolle's theorem."

- 1860s–1870s — Karl Weierstrass and contemporaries sharpen the underlying notions of continuity and the Extreme Value Theorem, putting Rolle/MVT on fully rigorous (

\varepsilon

–

\delta

) footing.

- Late 1800s — The MVT becomes standard in textbooks as the central lemma linking derivatives to function behavior, a role it holds to this day.

open-problem · Beyond the Classical Theorem

The MVT's elegance is single-variable: in higher dimensions the exact equality $f(b)-f(a)=f'(c)(b-a)$ fails for vector-valued functions — only an inequality version (the mean value inequality) survives, because different components may achieve the mean at different points. Determining the precise location of $c$ , even in one variable, is generally unsolvable in closed form, and the study of how $c$ behaves as the interval shrinks (the "mean value point" asymptotics, where $c \to a + \tfrac{1}{2}(b-a)$ for nice functions) remains a topic of refined analysis. There are also open and delicate questions about mean-value-type theorems for divided differences, complex-analytic functions (where the literal MVT can fail), and discontinuous or fractal "rough" functions where classical derivatives do not exist and weaker generalized-derivative formulations are required.

Mean Value Theorem and Rolle's Theorem

Formal Statement

Why It Holds

Building Intuition

Worked Examples

Common Misconceptions

Connections

Real-World Applications

Historical Development

Subtleties & Limits

See Also