4.4 Derivatives of Inverse Functions

Key Takeaways

  • The Inverse Function Theorem states that if g = f^(-1) and f'(g(a)) != 0, then g'(a) = 1 / f'(g(a)); it follows from differentiating the identity f(g(x)) = x with the Chain Rule.

  • The evaluation point is the whole difficulty: f' is evaluated at g(a), the input that f maps to a, and never at a itself.

  • Solve inverse-derivative problems in three fixed steps: find c with f(c) = a so that g(a) = c, evaluate f'(c), then take the reciprocal.

  • Because the graph of f^(-1) is the reflection of f across y = x, a point (c, a) on f with tangent slope m corresponds to the point (a, c) on f^(-1) with tangent slope 1/m.

  • Where f has a horizontal tangent (f'(c) = 0), the inverse has a vertical tangent at a = f(c) and g'(a) does not exist — as with f(x) = x^3 and g(x) = x^(1/3) at the origin.

Last updated: August 2026

4.4 Derivatives of Inverse Functions

Quick Answer: The College Board outline lists "Derivative of the inverse of a function (including arcsin⁡x\arcsin x and arctan⁡x\arctan x)" as a single tested bullet. Section 4.3 handled the named inverse trigonometric cases; this section handles the general rule behind them. If g=f−1g = f^{-1}, then g′(x)=1f′(g(x))g'(x) = \frac{1}{f'(g(x))} — the reciprocal of f′f' evaluated at the input of ff, not at the input of gg. That evaluation point is the entire difficulty of the topic, and it is where nearly every wrong answer comes from.


1. The Inverse Function Theorem

Let ff be one-to-one and differentiable on an interval, and let g=f−1g = f^{-1} be its inverse. If f′(g(a))≠0f'(g(a)) \neq 0, then gg is differentiable at aa and

g′(a)=1f′(g(a)),equivalently(f−1)′(a)=1f′(f−1(a))g'(a) = \frac{1}{f'\bigl(g(a)\bigr)}, \qquad \text{equivalently} \qquad \left(f^{-1}\right)'(a) = \frac{1}{f'\bigl(f^{-1}(a)\bigr)}

Derivation from the Chain Rule

The identity that defines an inverse is f(g(x))=xf(g(x)) = x for every xx in the domain of gg. Differentiate both sides with respect to xx and apply the Chain Rule from Section 4.1 to the left side:

ddx[f(g(x))]=ddx[x]  ⟹  f′(g(x))⋅g′(x)=1  ⟹  g′(x)=1f′(g(x))\frac{d}{dx}\bigl[f(g(x))\bigr] = \frac{d}{dx}[x] \implies f'(g(x)) \cdot g'(x) = 1 \implies g'(x) = \frac{1}{f'(g(x))}

That single line is worth reproducing on scratch paper at the start of any inverse-derivative question — it regenerates the formula and simultaneously reminds you which function gets evaluated where.


2. The Three-Step Protocol

Almost every CLEP-style inverse-derivative item gives you a value of ff and a value of f′f' and asks for g′g' at some number aa. Work it in this fixed order:

StepQuestion to askTypical work
1What is g(a)g(a)?Find the input cc with f(c)=af(c) = a. Then g(a)=cg(a) = c.
2What is f′(c)f'(c)?Differentiate ff and evaluate at cc — not at aa.
3Take the reciprocal.g′(a)=1/f′(c)g'(a) = 1 / f'(c).

Step 1 is the step candidates skip. You are handed the output aa and must run ff backwards to recover the input cc.


3. Worked Examples

Example 1: Recovering the input by inspection

Let f(x)=x3+2x+1f(x) = x^3 + 2x + 1, which is increasing on all of R\mathbb{R} (since f′(x)=3x2+2>0f'(x) = 3x^2 + 2 > 0) and therefore invertible. Find g′(4)g'(4) where g=f−1g = f^{-1}.

  1. Find g(4)g(4). Solve f(c)=4f(c) = 4: testing c=1c = 1 gives 1+2+1=41 + 2 + 1 = 4. So g(4)=1g(4) = 1.
  2. Evaluate f′f' at c=1c = 1. f′(x)=3x2+2f'(x) = 3x^2 + 2, so f′(1)=3+2=5f'(1) = 3 + 2 = 5.
  3. Reciprocate. g′(4)=15g'(4) = \dfrac{1}{5}.

Note that f′(4)=50f'(4) = 50 is irrelevant here. A distractor of 150\frac{1}{50} is standard.

Example 2: A transcendental ff that cannot be inverted algebraically

Let f(x)=x+exf(x) = x + e^{x}. Find (f−1)′(1)\left(f^{-1}\right)'(1).

You cannot solve x+ex=1x + e^x = 1 with algebra, but you can guess and verify: f(0)=0+e0=1f(0) = 0 + e^0 = 1, so f−1(1)=0f^{-1}(1) = 0. Then f′(x)=1+exf'(x) = 1 + e^x and f′(0)=1+1=2f'(0) = 1 + 1 = 2, giving

(f−1)′(1)=1f′(0)=12\left(f^{-1}\right)'(1) = \frac{1}{f'(0)} = \frac{1}{2}

This is exactly why the theorem exists: it produces the inverse's derivative without ever writing down a formula for the inverse.

Example 3: The theorem reproduces the arcsine formula

Take f(x)=sin⁡xf(x) = \sin x on [−π2,π2]\left[-\frac{\pi}{2}, \frac{\pi}{2}\right], so g(x)=arcsin⁡xg(x) = \arcsin x. Then f′(x)=cos⁡xf'(x) = \cos x and

g′(x)=1cos⁡(arcsin⁡x)=11−x2g'(x) = \frac{1}{\cos(\arcsin x)} = \frac{1}{\sqrt{1 - x^2}}

using the reference triangle from Section 4.3 (adjacent =1−x2= \sqrt{1 - x^2}, hypotenuse =1= 1). Every formula in the Section 4.3 master table is this one theorem applied to a restricted trigonometric function — memorize the table for speed, but derive from the theorem if your memory fails on exam day.


4. Geometry: Reflection Across y=xy = x

The graph of f−1f^{-1} is the mirror image of the graph of ff across the line y=xy = x. Reflection swaps the roles of run and rise, so it inverts every slope:

  • If (c,a)(c, a) lies on ff with tangent slope mm, then (a,c)(a, c) lies on f−1f^{-1} with tangent slope 1/m1/m.
  • A steep tangent on ff becomes a shallow tangent on f−1f^{-1}, and vice versa.
  • Increasing functions have increasing inverses; the reciprocal of a positive slope is positive.

When the derivative fails to exist

The theorem carries the hypothesis f′(g(a))≠0f'(g(a)) \neq 0 for a reason. If ff has a horizontal tangent at cc, the reflected tangent on f−1f^{-1} is vertical, and the inverse is not differentiable at a=f(c)a = f(c).

The standard illustration is f(x)=x3f(x) = x^3 at c=0c = 0: f′(0)=0f'(0) = 0, so g(x)=x1/3g(x) = x^{1/3} has a vertical tangent at the origin and g′(0)g'(0) does not exist. Confirm it directly: g′(x)=13x−2/3→∞g'(x) = \frac{1}{3}x^{-2/3} \to \infty as x→0x \to 0.


5. CLEP Traps

  • Trap 1 — Evaluating f′f' at the wrong number. g′(a)=1/f′(g(a))g'(a) = 1/f'(g(a)), never 1/f′(a)1/f'(a). If a table lists both f′(a)f'(a) and f′(g(a))f'(g(a)), one of them is a planted distractor.
  • Trap 2 — Forgetting to reciprocate. Answering f′(g(a))f'(g(a)) instead of its reciprocal is the second most common error.
  • Trap 3 — Inverse function vs. reciprocal function. f−1(x)f^{-1}(x) is the inverse; [f(x)]−1=1/f(x)[f(x)]^{-1} = 1/f(x) is the reciprocal. Their derivatives are unrelated: the reciprocal's derivative is −f′(x)/[f(x)]2-f'(x)/[f(x)]^2 by the Quotient Rule.
  • Trap 4 — Assuming an inverse exists. The theorem needs ff to be one-to-one. On the CLEP this is usually established by showing f′>0f' > 0 (or f′<0f' < 0) everywhere, which is worth stating in your reasoning.
Test Your Knowledge

The function f(x) = x^3 + 2x + 1 is increasing on all real numbers and therefore has an inverse g. What is the value of g′(4)?

A

1/5

B

5

C

1/50

D

1/4

Test Your Knowledge

A differentiable, increasing function f satisfies f(2) = 7, f′(2) = 3, f(7) = 2, and f′(7) = 5. If g is the inverse of f, what is g′(7)?

A

1/5

B

1/3

C

3

D

1/7

Test Your Knowledge

Let f(x) = x^3 with inverse g(x) = x^(1/3). Which statement correctly describes g at x = 0?

A

g is differentiable at 0 because f is differentiable everywhere

B

g′(0) = 0, matching the horizontal tangent of f at the origin

C

g′(0) does not exist, because f′(0) = 0 makes the reflected tangent line vertical

D

g′(0) = 1, because f and g intersect at the origin

Test Your Knowledge

The point (3, 5) lies on the graph of a one-to-one differentiable function f, and the tangent line to f at that point has slope 4. What is the equation of the tangent line to y = f^(-1)(x) at the corresponding point?

A

y − 5 = (1/4)(x − 3)

B

y − 3 = 4(x − 5)

C

y − 5 = 4(x − 3)

D

y − 3 = (1/4)(x − 5)

Sections you finish are checked off in the contents.