Concept · Graphing
What Is an Inverse Function? A Plain Explanation
An inverse function undoes another function exactly — feed it the output, and it hands back the input. Not every function qualifies, and the reason why is worth understanding before finding one by hand.
Undoing a function exactly
A function f takes an input and produces an output. Its inverse, written f⁻¹, runs that process backward: feed it the output, and it hands back the original input. If f(3) = 7, then f⁻¹(7) = 3 — no approximation, no rounding, exactly the input that produced that output.
This is captured precisely by composition. Applying a function and then its inverse, in either order, returns exactly what you started with: f⁻¹(f(x)) = x, and f(f⁻¹(x)) = x. Neither statement is true by coincidence — it is the definition of what makes f⁻¹ the inverse of f, and it is exactly what the Inverse Function Calculator on this site checks numerically before showing a result.
Why the notation is not an exponent
f⁻¹(x) looks like it should mean 1/f(x), and that confusion is common enough to be worth stating plainly: it does not. The −1 here means 'inverse function', a different operation from raising something to a power. 1/f(x) is the reciprocal of f; f⁻¹(x) is a function that undoes f. For most functions the two are not even close to each other.
Not every function has one: the one-to-one requirement
An inverse only exists when f is one-to-one — every output comes from exactly one input. If two different inputs produce the same output, there is no way to reverse the process: given that output, which input should f⁻¹ return?
- f(2) = 4 and f(−2) = 4two different inputs, the same output
- f⁻¹(4) = ?should it return 2, or −2? both are valid
- no single answer exists, so f⁻¹ is not a function herex² over all reals fails the one-to-one test
Graphically, this is the horizontal line test: if any horizontal line crosses the graph of f more than once, f is not one-to-one over that domain, and a full inverse does not exist.
The fix: restrict the domain
x² is not invertible over all real numbers, but restricted to x ≥ 0 it is — every non-negative input now gives a distinct output, and √x correctly undoes it on that restricted piece. This is not a workaround; it is the standard convention, and it is why √4 is defined to mean 2, not ±2, even though both square to 4.
The same issue, and the same fix, applies to sin, cos and tan. sin(0) and sin(π) are both 0, so sine repeats endlessly and has no inverse over all real numbers. Restricted to a principal domain — [−π/2, π/2] for sine — it becomes one-to-one, and arcsin is its inverse on exactly that restricted piece. A calculator that returns arcsin(x) is implicitly using this restriction, not claiming sine is invertible everywhere.
What the graph of an inverse looks like
f⁻¹'s graph is the reflection of f's graph across the line y = x. This follows directly from what an inverse does: every point (a, b) on f's graph corresponds to a point (b, a) on f⁻¹'s graph, since if f(a) = b then f⁻¹(b) = a. Swapping the coordinates of every point is exactly a reflection across y = x.
A quick consequence worth knowing: if f passes through (2, 5), then f⁻¹ passes through (5, 2). Where f is increasing, f⁻¹ is increasing too, since a reflection across y = x does not change a curve from rising to falling.
The domain and range swap places
Because every output of f becomes an input of f⁻¹, and every input of f becomes an output of f⁻¹, the domain and range trade places exactly. If f has domain x ≥ 0 and range y ≥ 3, then f⁻¹ has domain x ≥ 3 and range y ≥ 0.
| f | f⁻¹ | |
|---|---|---|
| Domain | the inputs f accepts | the range of f |
| Range | the outputs f produces | the domain of f |
This is a fast way to sanity-check a candidate inverse: if the domain and range do not swap correctly, something in the algebra went wrong.
The identity function is its own inverse
f(x) = x is the one case where undoing the function changes nothing at all, since there is nothing to undo — f⁻¹(x) = x as well. It is a useful edge case to keep in mind: whatever method you use to find an inverse should return x itself when applied to f(x) = x, and that is exactly what the calculator on this site does.
Vocabulary worth having straight
A one-to-one function is one where no output repeats — every value in the range traces back to a single input, which is exactly the property that makes it possible to undo a function in the first place. Composition of functions is what f⁻¹(f(x)) = x actually is: applying f, then applying f⁻¹ to the result, chained together into a single combined operation.
- Swap x and y — the final step of finding f inverse by hand, once the equation has been solved for x.
- Reflection over y = x — how the graph of f⁻¹ relates to the graph of f, since every coordinate pair is flipped.
- Restricted domain — the smaller piece of the input values, like x ≥ 0 for x², on which a function that is not naturally one-to-one becomes invertible.
- Principal branch — the specific restricted domain convention used for sin, cos and tan, chosen so arcsin, arccos and arctan are single, well-defined functions rather than a menu of equally valid answers.
Questions about what is an inverse function
What does f⁻¹(x) mean?
The inverse of f — a function that undoes what f does. If f(a) = b, then f⁻¹(b) = a. It is not the same as 1/f(x), even though the notation looks similar.
Why doesn't every function have an inverse?
An inverse only exists when the function is one-to-one — every output comes from exactly one input. If two inputs share an output, like 2 and −2 both squaring to 4, there is no single input to return.
What is the horizontal line test?
A quick check for whether a function is one-to-one: if any horizontal line crosses the graph more than once, the function is not one-to-one there, and has no full inverse over that domain.
Why does the graph of an inverse look reflected?
Because f⁻¹ swaps every input and output of f. Swapping the coordinates of every point on a graph is exactly a reflection across the line y = x.