Mean value theorems
Rolle’s theorem, Lagrange’s Mean Value Theorem, and Cauchy’s Mean Value Theorem: statements, assumptions, proofs, geometric meaning, and standard applications.
Drafted with Aria, reviewed by the AiCanCode.org team · Updated . How we write and check our notes. Spotted an error? Use Give Feedback at the bottom of the page.
Why it matters
Engineers use mean value theorems to justify approximations, bounds, and error estimates whenever a quantity changes smoothly with another variable (time, temperature, stress, voltage, etc.). They are the bridge between average change over an interval and instantaneous rate at a specific point, which underlies sensor calibration, control-system tuning, and numerical methods.
Key ideas
- Differentiability on (a, b): f is differentiable on the open interval (a, b) if f′(x) exists for every x in (a, b). A derivative at x measures the instantaneous rate of change or slope of the tangent at x.
- Secant line: The straight line joining the points A(a, f(a)) and B(b, f(b)) on the graph of f. Its slope is (f(b) − f(a)) / (b − a).
- Tangent line: The line touching the curve at one point without cutting it nearby. Its slope at x = c is f′(c).
Exact statements (standard university form) and derivations:
- Rolle’s Theorem
- Statement: Let f be a real-valued function on [a, b]. If (i) f is continuous on [a, b], (ii) f is differentiable on (a, b), and (iii) f(a) = f(b), then there exists at least one c in (a, b) such that f′(c) = 0.
- Geometric interpretation: If a smooth curve starts and ends at the same height, there is some interior point where the tangent is horizontal.
- Proof (standard):
- By continuity on [a, b], f attains its maximum and minimum at some points (Weierstrass Extreme Value Theorem). Let x_max and x_min be where the maximum and minimum occur.
- If f is constant, then f′(x) = 0 for all x in (a, b). Any c works.
- If not constant, at least one extremum occurs in the open interval (a, b) because f(a) = f(b). At an interior extremum of a differentiable function, by Fermat’s theorem, the derivative is zero. Hence there exists c in (a, b) with f′(c) = 0.
- Diagram to draw and label: Draw axes. Plot a smooth curve starting at A(a, y0) and ending at B(b, y0) at the same height. Mark a top point C between them. Draw a tangent at C; show it is horizontal (slope 0). Label A, B, C, and indicate f(a) = f(b).
- Lagrange’s Mean Value Theorem (LMVT)
- Statement: Let f be a real-valued function on [a, b]. If (i) f is continuous on [a, b] and (ii) differentiable on (a, b), then there exists at least one c in (a, b) such that f′(c) = (f(b) − f(a)) / (b − a).
- Geometric interpretation: There is at least one point where the tangent to the curve is parallel to the chord AB.
- Derivation from Rolle’s theorem (standard construction):
- Define the straight line ℓ(x) passing through (a, f(a)) and (b, f(b)): ℓ(x) = f(a) + [(f(b) − f(a)) / (b − a)] (x − a).
- Consider g(x) = f(x) − ℓ(x). Then g is continuous on [a, b] and differentiable on (a, b). Also g(a) = g(b) = 0.
- By Rolle’s theorem, there exists c in (a, b) with g′(c) = 0.
- Compute g′(x) = f′(x) − [(f(b) − f(a)) / (b − a)]. Hence g′(c) = 0 implies f′(c) = (f(b) − f(a)) / (b − a).
- Diagram to draw and label: Draw axes. Plot f between A(a, f(a)) and B(b, f(b)). Draw the chord AB. Mark a point C where the tangent is parallel to AB. Label slopes: slope(chord) = [f(b) − f(a)]/(b − a) and slope(tangent at C) = f′(c).
- Cauchy’s Mean Value Theorem (CMVT)
- Statement: Let f and g be real-valued functions on [a, b]. If (i) f and g are continuous on [a, b] and (ii) differentiable on (a, b), then there exists some c in (a, b) such that (g(b) − g(a)) f′(c) = (f(b) − f(a)) g′(c). If additionally g′(x) ≠ 0 on (a, b) and g(b) ≠ g(a), this can be written in ratio form as [f(b) − f(a)] / [g(b) − g(a)] = f′(c) / g′(c).
- Derivation using Rolle’s theorem (standard auxiliary function):
- Define h(x) = [f(b) − f(a)] g(x) − [g(b) − g(a)] f(x).
- h is continuous on [a, b] and differentiable on (a, b).
- Compute h(a) = [f(b) − f(a)] g(a) − [g(b) − g(a)] f(a) and h(b) = [f(b) − f(a)] g(b) − [g(b) − g(a)] f(b). Then h(b) − h(a) = 0, so h(a) = h(b).
- By Rolle’s theorem, there exists c in (a, b) such that h′(c) = 0.
- Compute h′(c) = [f(b) − f(a)] g′(c) − [g(b) − g(a)] f′(c) = 0, rearrange to get the CMVT relation.
- Note: Setting g(x) = x reduces CMVT to LMVT.
- Supporting fact often cited in proofs:
- Fermat’s Theorem on stationary points: If f has a local maximum or minimum at c and f is differentiable at c, then f′(c) = 0.
- Consequences and uses (what they let you conclude):
- Existence of a point with derivative equal to average rate of change (LMVT); special case derivative equals zero (Rolle).
- Monotonicity test: If f′(x) = 0 on an interval, then f is constant there (by LMVT applied to any [x1, x2]). If f′(x) > 0 (< 0), then f is strictly increasing (decreasing).
- Inequalities and bounds: Comparing two functions via CMVT helps prove inequalities (e.g., ln(1 + x) < x for x > −1, x ≠ 0).
- Error terms: The remainder term in Taylor’s theorem (covered under Taylor’s and Maclaurin’s series) is derived using a mean value form.
- L’Hôpital’s rule (covered under Indeterminate forms and L’Hôpital’s rule) follows from CMVT under suitable limits.
Assumptions to check before applying:
- Closed interval [a, b] for continuity; open interval (a, b) for differentiability.
- For CMVT, ensure g′(x) ≠ 0 on (a, b) and that g(b) ≠ g(a) if using the ratio form.
- Endpoints included for continuity, not for differentiability.
Formulas
Slope of secant on [a, b] = (f(b) − f(a)) / (b − a) — f is any real function; units: same as (units of f)/(units of x); applies for a ≠ b.
Rolle’s theorem: ∃ c ∈ (a, b) such that f′(c) = 0 — if f is continuous on [a, b], differentiable on (a, b), and f(a) = f(b); units: derivative units as (units of f)/(units of x).
LMVT: ∃ c ∈ (a, b): f′(c) = (f(b) − f(a)) / (b − a) — if f is continuous on [a, b] and differentiable on (a, b); units: as above.
CMVT: ∃ c ∈ (a, b): (g(b) − g(a)) f′(c) = (f(b) − f(a)) g′(c) — if f and g are continuous on [a, b] and differentiable on (a, b); for the ratio form, additionally require g′(x) ≠ 0 on (a, b) and g(b) ≠ g(a); units: both sides have units of (units of f) × (units of g)/(units of x).
Monotonicity from derivative: if f is continuous on [a, b], differentiable on (a, b), and f′(x) ≥ 0 on (a, b), then f is non-decreasing on [a, b] — applies with standard regularity; units: n/a in pure math.
Constancy test: if f is continuous on [a, b], differentiable on (a, b), and f′(x) = 0 on (a, b), then f is constant on [a, b] — consequence of LMVT; units: n/a.
Note on units: In pure mathematics these statements are dimensionless. In engineering applications, x and f(x) may carry units; derivatives have units (units of f)/(units of x).
Worked examples
Example 1: Verifying LMVT and finding c for a specific function
- Problem: Verify LMVT for f(x) = x² on [1, 3] and find a c such that f′(c) equals the average rate of change.
- Steps:
- Check continuity and differentiability: A polynomial is continuous on ℝ and differentiable on ℝ, so conditions hold on [1, 3].
- Compute the average rate of change: Use formula before numbers. f(b) − f(a) = f(3) − f(1) = 9 − 1 = 8 (no units in pure math) b − a = 3 − 1 = 2 Average slope = (f(b) − f(a)) / (b − a) = 8 / 2 = 4.
- Compute derivative: f′(x) = 2x.
- Set f′(c) = 4 → 2c = 4 → c = 2.
- Final answer: c = 2 ∈ (1, 3). LMVT verified and the required point is at x = 2. The tangent slope is 4, equal to the secant slope.
Example 2: Rolle’s theorem for a polynomial and finding stationary point
- Problem: Show that Rolle’s theorem applies to f(x) = x² − 5x + 6 on [2, 3], and find c with f′(c) = 0.
- Steps:
- Check equal endpoints: f(2) = 4 − 10 + 6 = 0; f(3) = 9 − 15 + 6 = 0. So f(2) = f(3).
- Continuity and differentiability: Polynomials are continuous and differentiable everywhere; conditions hold on [2, 3].
- Apply Rolle: There exists c ∈ (2, 3) with f′(c) = 0.
- Compute derivative: f′(x) = 2x − 5. Solve 2c − 5 = 0 → c = 2.5.
- Final answer: c = 2.5 (i.e., c = 5/2) is the interior point with horizontal tangent, as guaranteed by Rolle’s theorem.
Example 3: LMVT on logarithm; explicit c on [1, e]
- Problem: For f(x) = ln x on [1, e], use LMVT to find c with f′(c) = (f(e) − f(1)) / (e − 1).
- Steps:
- Continuity and differentiability: ln x is continuous and differentiable on (0, ∞); conditions hold on [1, e].
- Compute average slope: f(e) − f(1) = 1 − 0 = 1; b − a = e − 1. So average slope = 1 / (e − 1).
- Derivative: f′(x) = 1/x.
- Solve 1/c = 1 / (e − 1) → c = e − 1.
- Check interval: e − 1 ≈ 1.71828, which lies in (1, e).
- Final answer: c = e − 1 ∈ (1, e). The tangent slope 1/c equals the secant slope 1/(e − 1).
Example 4: Using CMVT to derive an inequality
- Problem: Prove that for x > −1, x ≠ 0, we have ln(1 + x) < x. Also show equality at x = 0 by continuity.
- Steps (CMVT/LMVT approach):
- Consider f(t) = ln(1 + t) and g(t) = t on [0, x] if x > 0; or on [x, 0] if −1 < x < 0. Both continuous and differentiable on the chosen interval. For g, g′(t) = 1 ≠ 0.
- By LMVT (g is identity), ∃ c between the endpoints such that f′(c) = [f(x) − f(0)] / [x − 0] = ln(1 + x) / x.
- Compute derivative: f′(t) = 1 / (1 + t). For c between the endpoints, we have 1 / (1 + c) < 1 if c > 0; and 1 / (1 + c) > 1 if c < 0 (since −1 < c < 0). Carefully interpret the sign of x:
- If x > 0: c ∈ (0, x) ⇒ 1/(1 + c) < 1 ⇒ ln(1 + x) / x < 1 ⇒ ln(1 + x) < x.
- If −1 < x < 0: c ∈ (x, 0) ⇒ 1/(1 + c) > 1, but x < 0. Multiplying both sides by negative x reverses the inequality, again giving ln(1 + x) < x.
- At x = 0, both sides are 0; the inequality extends by continuity.
- Final answer: ln(1 + x) ≤ x for x > −1, with equality only at x = 0; strict inequality holds for x ≠ 0.
Common mistakes
- Misidentifying the interval: Using a and b out of order or setting a = b (secant slope undefined).
- Algebra slips in auxiliary functions: For LMVT/CMVT proofs, errors in constructing ℓ(x) or h(x) and in simplifying derivatives.
- Mixing roles of f and g in CMVT and forgetting the condition g′(x) ≠ 0; or attempting to apply CMVT when g(b) = g(a) while using the ratio form.
- Assuming uniqueness of c: The theorems guarantee existence, not uniqueness. There may be multiple points c.
- Units confusion: In pure math there are no SI units; in applications, derivative units are (output units)/(input units). Keep consistency.
For your exam
- Derivations/Proofs: Derive LMVT from Rolle’s theorem using the auxiliary linear function. Derive CMVT using the constructed h(x).
- Geometric interpretations: Explain and sketch the chord-tangent parallelism for LMVT; horizontal tangent for Rolle’s theorem. Describe exactly what to draw and label.
- Applications: Prove inequalities (e.g., ln(1 + x) < x for x > −1, x ≠ 0), monotonicity/constancy results via LMVT, root location and number of stationary points.
- Numericals: Given f and [a, b], verify conditions, compute average slope, find a c satisfying LMVT; or verify Rolle’s theorem and compute c with f′(c) = 0.
- What a full-marks answer includes:
- Clear, exact statement with all conditions.
- Hypotheses check before application.
- Correct construction of auxiliary function (for proofs) and clean step-by-step derivation.
- Proper algebra and calculus steps, with reasons (e.g., “by Rolle’s theorem”, “by Fermat’s theorem”).
- Diagram description with labeled points and specified slopes when asked for geometric meaning.
- Final conclusion highlighted (e.g., the value(s) of c or the proved inequality), and notes on existence vs. uniqueness.
Quick check
- State the hypotheses of LMVT and the conclusion in one line.
- For f(x) = e^x on [0, 2], find a c such that f′(c) equals the average rate of change on [0, 2].
- Using CMVT for f(x) = sin x and g(x) = x on [a, b], what does the theorem guarantee?
Answers: (1) If f is continuous on [a, b] and differentiable on (a, b), then ∃ c ∈ (a, b) with f′(c) = (f(b) − f(a)) / (b − a). (2) Average slope = (e² − 1)/(2 − 0); f′(x) = e^x; set e^c = (e² − 1)/2 ⇒ c = ln[(e² − 1)/2] ∈ (0, 2). (3) There exists c ∈ (a, b) with (sin b − sin a)/(b − a) = cos c, i.e., a point where the tangent slope equals the secant slope (by LMVT applied to sin x).
Finished this topic? Mark it so your progress, study plan and readiness keep up.
Stuck on something here?