What Is The Rule For Inverse Functions And Differentiation?
Inverse Functions and Differentiation is the calculus rule that gives the derivative of an inverse function directly from the derivative of the original. You never have to differentiate the inverse from scratch. State it precisely:
$$\text{If } f \text{ is one-to-one and differentiable, } b = f(a), \text{ and } f'(a) \neq 0, \text{ then } \quad (f^{-1})'(b) = \frac{1}{f'(a)}.$$
Written as a function of its own input, with $a = f^{-1}(x)$, the same rule reads:
$$(f^{-1})'(x) = \frac{1}{f'\big(f^{-1}(x)\big)}.$$
The one idea to hold onto: you evaluate $f'$ at the original-side point $a$, not at $b$. In Leibniz notation the rule looks like flipping a fraction:
$$\frac{dx}{dy} = \frac{1}{;dy/dx;}.$$
Here $y = f(x)$ measures the output as the input changes, and $x = f^{-1}(y)$ measures the input as the output changes. Those two rates are reciprocals of each other. This rule is a cornerstone of calculus and the fastest route to the derivative of logarithms and inverse trigonometric functions.
How Do You Derive The Inverse Function Derivative Rule?
The derivation is one line of the chain rule. Start from the defining identity of an inverse, true for every $x$ in the domain:
$$f\big(f^{-1}(x)\big) = x.$$
Differentiate both sides with respect to $x$. The left side is a composition, so the chain rule applies; the right side is just $x$, whose derivative is $1$:
$$f'\big(f^{-1}(x)\big) \cdot (f^{-1})'(x) = 1.$$
Now solve for the unknown $(f^{-1})'(x)$, which is allowed exactly when the factor $f'(f^{-1}(x))$ is not zero:
$$(f^{-1})'(x) = \frac{1}{f'\big(f^{-1}(x)\big)}.$$
That is the whole proof. The non-zero condition is not a technicality bolted on afterwards; it is the step where we divide, and dividing by zero is what the hypothesis $f'(a) \neq 0$ forbids.
Why Does The Inverse Function Derivative Rule Work?
The algebra above is airtight, but the picture is what makes it stick. The graph of $f^{-1}$ is the graph of $f$ reflected across the line $y = x$, and that reflection is doing something specific to every tangent line.
Reflection sends each point $(a, b)$ on the graph of $f$ to the point $(b, a)$ on the graph of $f^{-1}$. The coordinates simply trade places.
Reflecting across $y = x$ swaps the horizontal and vertical directions. A rise of $\Delta y$ over a run of $\Delta x$ becomes a rise of $\Delta x$ over a run of $\Delta y$, so the slope flips to its reciprocal.
A tangent line of slope $m$ at $(a, b)$ therefore becomes a tangent line of slope $\tfrac{1}{m}$ at $(b, a)$. That is exactly $(f^{-1})'(b) = \tfrac{1}{f'(a)}$.
If $f'(a) = 0$, the tangent at $(a, b)$ is horizontal. Its reflection is a vertical line, which has no finite slope, so the inverse is not differentiable there. This is the geometric meaning of the $f'(a) \neq 0$ hypothesis.
The rule also explains why $f$ must be one-to-one to have an inverse at all. A one-to-one function is strictly monotonic on the interval, so its graph passes the horizontal-line test, and the reflection is again the graph of a genuine function rather than a sideways curve that fails the vertical-line test.
How Do You Find The Derivative Of ln x From e^x?
The natural logarithm is the inverse of the exponential, so its derivative falls straight out of the rule. Let $f(x) = e^x$, which is one-to-one and differentiable with $f'(x) = e^x$. Its inverse is $f^{-1}(x) = \ln x$, defined for $x > 0$.
$$(\ln x)' = \frac{1}{f'\big(f^{-1}(x)\big)} = \frac{1}{e^{\ln x}}.$$
Since $e^{\ln x} = x$ for every $x > 0$:
$$(\ln x)' = \frac{1}{x}.$$
Final answer: if $g(x) = \ln x$, then $g'(x) = \dfrac{1}{x}$ for $x > 0$.
Notice $f'(x) = e^x$ is never zero, so the non-zero-derivative condition holds everywhere, which is why $\ln x$ is differentiable on its whole domain. The same argument turns any logarithmic function into a clean derivative.
How Do You Differentiate arcsin x Using Its Inverse?
Take $f(x) = \sin x$ restricted to $\left[-\tfrac{\pi}{2}, \tfrac{\pi}{2}\right]$, where it is one-to-one and increasing. Then $f^{-1}(x) = \arcsin x$ and $f'(x) = \cos x$. Apply the rule:
$$(\arcsin x)' = \frac{1}{\cos\big(\arcsin x\big)}.$$
Let $\theta = \arcsin x$, so $\sin \theta = x$ and $\theta$ lies in $\left[-\tfrac{\pi}{2}, \tfrac{\pi}{2}\right]$, where cosine is non-negative. Using $\cos\theta = \sqrt{1 - \sin^2\theta}$:
$$\cos\big(\arcsin x\big) = \sqrt{1 - x^2}, \qquad \text{so} \qquad (\arcsin x)' = \frac{1}{\sqrt{1 - x^2}}.$$
Final answer: $(\arcsin x)' = \dfrac{1}{\sqrt{1 - x^2}}$ for $-1 < x < 1$.
The endpoints $x = \pm 1$ are excluded because there $\cos\theta = 0$, which is the $f'(a) = 0$ case: $\arcsin$ has vertical tangents at $\pm 1$. The whole family of inverse trigonometric functions is differentiated this same way.
How Do You Compute A Numeric Inverse Derivative Without Finding The Inverse?
Often the inverse has no clean formula, yet the rule still gives an exact number. Let $f(x) = x^3 + x$ and find $(f^{-1})'(2)$.
Step 1 — find the matching original-side point $a$. Solve $f(a) = 2$:
$$a^3 + a = 2 \quad\Longrightarrow\quad a = 1, \quad \text{since } 1^3 + 1 = 2.$$
Step 2 — check the function qualifies. Here $f'(x) = 3x^2 + 1$, which is positive for every $x$, so $f$ is strictly increasing, one-to-one, and $f'(a) \neq 0$. The rule applies.
Step 3 — evaluate $f'$ at $a = 1$ and reciprocate.
$$f'(1) = 3(1)^2 + 1 = 4, \qquad (f^{-1})'(2) = \frac{1}{f'(1)} = \frac{1}{4}.$$
Final answer: $(f^{-1})'(2) = \dfrac{1}{4} = 0.25$.
We never solved for $f^{-1}$ as a formula, which for a cubic-plus-linear function is genuinely messy. The rule only needed the one point $a = 1$ and the derivative there.
What Are The Derivatives Of Common Inverse Functions?
Every entry below comes from the same rule, $(f^{-1})'(x) = 1 / f'(f^{-1}(x))$. Read the table as "this inverse function inverts that original, and here is the derivative you get."
Table: Common inverse functions and the derivatives the inverse-function rule produces.
Inverse function $g(x)$ | Inverts | Derivative $g'(x)$ | Valid for |
|---|---|---|---|
$\ln x$ | $e^{x}$ | $\dfrac{1}{x}$ | $x > 0$ |
$\sqrt[3]{x}$ | $x^{3}$ | $\dfrac{1}{3x^{2/3}}$ | $x \neq 0$ |
$\arcsin x$ | $\sin x$ on $\left[-\tfrac{\pi}{2}, \tfrac{\pi}{2}\right]$ | $\dfrac{1}{\sqrt{1 - x^2}}$ | $-1 < x < 1$ |
$\arccos x$ | $\cos x$ on $[0, \pi]$ | $-\dfrac{1}{\sqrt{1 - x^2}}$ | $-1 < x < 1$ |
$\arctan x$ | $\tan x$ on $\left(-\tfrac{\pi}{2}, \tfrac{\pi}{2}\right)$ | $\dfrac{1}{1 + x^2}$ | all real $x$ |
Two entries are worth a closer look. The derivative of arccos carries a minus sign because $\cos x$ is decreasing on $[0, \pi]$, so its slope, and therefore the reciprocal, is negative. The derivative of tan inverse is defined for every real number because $\tan x$ has no zero-slope point on its principal branch, so the rule never divides by zero.
Who Discovered The Link Between Inverse Functions and Differentiation?
The reciprocal relationship between a rate and its inverse rate is old, but the notation that made it obvious came out of the seventeenth-century birth of calculus, and a famous quarrel.
Later mathematicians made the idea rigorous. Augustin-Louis Cauchy (1789–1857, France) and Karl Weierstrass (1815–1897, Germany) supplied the limit definitions and the epsilon-delta care that pin down exactly when the inverse is differentiable, which is the modern reason we state the $f'(a) \neq 0$ hypothesis so carefully.
Where Is Inverse Functions and Differentiation Used In The Real World?
Any time two quantities depend on each other and you know one rate, this rule hands you the other for free.
Statistics and simulation: drawing a random sample from a distribution uses inverse-transform sampling, where the quantile function is the inverse of the cumulative distribution, and its slope is the reciprocal of the density. The reciprocal-derivative relationship is what makes the sampling arithmetic work.
Engineering and calibration: a sensor outputs a voltage that rises with temperature at a known rate; the calibration curve you actually read runs the other way, temperature per volt, and its slope is the reciprocal of the sensor's.
Economics: if quantity demanded is a known function of price, the inverse-demand curve gives price as a function of quantity, and its marginal slope is the reciprocal of the original, which is what marginal-revenue calculations rely on.
Numerical methods: Newton's method for solving equations steps along tangent lines, and its convergence is governed by how an output change maps back to an input change, exactly the inverse-derivative quantity.
Physics of motion: when position is a known function of time, and you re-express time as a function of position, the two rates are reciprocals, so "seconds per metre" is one over "metres per second."
One rule, five fields. The moment a relationship can be run backwards, its rate runs backwards as a reciprocal.
What Are The Most Common Mistakes With Inverse Functions and Differentiation?
These three errors account for most lost marks, and they match the confusions students post most often about which point to use and whether to reciprocate.
Evaluating $f'$ at the wrong point.
Where it slips in:
A student is asked for $(f^{-1})'(b)$ and plugs $b$ straight into $f'$, computing $f'(b)$ instead of $f'(a)$.
Don't do this:
Do not evaluate $f'$ at the point you were given on the inverse side. The number $b$ lives on the inverse's input axis, not the original's.
The correct way:
First find the original-side point $a$ with $f(a) = b$, then compute $f'(a)$. For $f(x) = x^3 + x$ and $(f^{-1})'(2)$, that means solving $a^3 + a = 2$ to get $a = 1$, and using $f'(1) = 4$, not $f'(2) = 13$.
Forgetting to reciprocate.
Where it slips in:
A student correctly finds $f'(a)$ and then writes $(f^{-1})'(b) = f'(a)$, dropping the "one over."
Don't do this:
Do not report $f'(a)$ as the answer. Reciprocating is the entire content of the rule, not an optional last step.
The correct way:
Divide: $(f^{-1})'(b) = \dfrac{1}{f'(a)}$. If $f'(1) = 4$, the inverse derivative is $\tfrac{1}{4}$, not $4$.
Applying the rule where $f'(a) = 0$.
Where it slips in:
A student uses the formula at a point where the original function has a horizontal tangent, for example asking for the derivative of $\sqrt[3]{x}$ at $x = 0$, where its parent $f(x) = x^3$ has $f'(0) = 0$.
Don't do this:
Do not divide by $f'(a)$ when it is zero. The rule's hypothesis fails, and the inverse has a vertical tangent, so its derivative does not exist there.
The correct way:
Check $f'(a) \neq 0$ before dividing. At $x = 0$, the cube root has a vertical tangent and no finite derivative, which is the honest answer, not a number.
Practice Problems On Inverse Functions and Differentiation
Verify each function is one-to-one before applying the rule; answers follow every problem.
For $f(x) = x^5 + 2x + 1$, find $(f^{-1})'(4)$.
(Answer: solve $a^5 + 2a + 1 = 4 \Rightarrow a = 1$; $f'(x) = 5x^4 + 2$, $f'(1) = 7$, so $(f^{-1})'(4) = \tfrac{1}{7}$.)For $f(x) = x^3 + 2x + 1$, find $(f^{-1})'(1)$.
(Answer: $a^3 + 2a = 0 \Rightarrow a = 0$; $f'(x) = 3x^2 + 2$, $f'(0) = 2$, so $(f^{-1})'(1) = \tfrac{1}{2}$.)Given $f(3) = 7$ and $f'(3) = 5$, with $f$ one-to-one and differentiable, find $(f^{-1})'(7)$.
(Answer: $\tfrac{1}{f'(3)} = \tfrac{1}{5}$.)Use the rule to find $(\arccos x)'$ from $f(x) = \cos x$ on $[0, \pi]$.
(Answer: $f'(x) = -\sin x$, and $\sin(\arccos x) = \sqrt{1 - x^2}$, so $(\arccos x)' = -\tfrac{1}{\sqrt{1 - x^2}}$.)For $f(x) = 2x + \sin x$, find $(f^{-1})'(0)$.
(Answer: $2a + \sin a = 0 \Rightarrow a = 0$; $f'(x) = 2 + \cos x$, $f'(0) = 3$, so $(f^{-1})'(0) = \tfrac{1}{3}$.)Explain why the rule cannot give $(f^{-1})'(0)$ for $f(x) = x^3$.
(Answer: $f'(0) = 0$, so the hypothesis fails; the inverse $\sqrt[3]{x}$ has a vertical tangent at $0$ and no finite derivative there.)
Where Should You Go Next After Inverse Functions and Differentiation?
Each door below builds directly on the reciprocal-slope idea.
Inverse functions. Nail the algebra of finding and verifying an inverse first, since the derivative rule assumes you can locate the point $a$ with $f(a) = b$.
The derivative. Revisit the limit definition and the chain rule, the single tool the whole derivation rests on.
Inverse trigonometric functions. Apply the rule across the full arcsin, arccos, and arctan family, the most common place it is tested.
If your child is meeting these ideas for the first time, a live Bhanzu trainer teaches Inverse Functions and Differentiation from the reflection picture outward, so the formula is understood rather than memorised, in the Bhanzu math program.
Was this article helpful?
Your feedback helps us write better content