A function is invertible when every output traces back to exactly one input. In formal terms, this means the function must be both one-to-one (injective) and onto (surjective), a combination mathematicians call bijective. If either property is missing, the function can’t be reliably reversed, because the inverse would either lose information or leave gaps.
The Two Requirements: One-to-One and Onto
A function pairs each input with an output. For the reverse pairing to also be a valid function, two conditions must hold.
One-to-one (injective): No two different inputs produce the same output. If f(x₁) = f(x₂), then x₁ must equal x₂. This guarantees that when you try to reverse the function, each output points back to a single input rather than forcing you to choose between two or more possibilities.
Onto (surjective): Every element in the codomain (the set of possible outputs) actually gets hit by some input. The range of the function equals its entire codomain. This ensures the inverse has something to say about every value you might feed it, with no orphaned outputs that lack a corresponding input.
When both conditions are satisfied, the function is bijective, and a theorem in mathematics confirms the relationship directly: a function has an inverse if and only if it is bijective. One condition without the other isn’t enough. A function that’s one-to-one but not onto will have an inverse that’s undefined for some values. A function that’s onto but not one-to-one will have an inverse that tries to assign multiple values to a single input, which breaks the definition of a function.
What an Inverse Actually Does
If f is an invertible function, its inverse (written f⁻¹) reverses the mapping: f⁻¹(y) = x precisely when f(x) = y. The two functions undo each other completely, which produces the cancellation formulas:
- f⁻¹(f(x)) = x for every x in the domain of f
- f(f⁻¹(x)) = x for every x in the domain of f⁻¹
Composing a function with its inverse in either order gives the identity function, the function that simply returns its input unchanged. This is the defining property of an inverse: it’s the function that, when composed with the original, produces the identity.
One common source of confusion: the notation f⁻¹(x) does not mean 1/f(x). The superscript here indicates the inverse function, not a reciprocal.
The Horizontal Line Test
If you’re looking at a graph and want to know whether the function is invertible, the horizontal line test gives you a quick visual answer. Imagine drawing horizontal lines across the graph at every possible height. If every horizontal line crosses the graph at most once, the function is one-to-one and its inverse will be a function. If any horizontal line crosses the graph in two or more places, the function fails the test, meaning two different inputs share the same output, and the function isn’t invertible as written.
This is the graphical equivalent of checking injectivity. A parabola like y = x², for instance, fails immediately: a horizontal line at y = 4 crosses the curve at both x = 2 and x = −2. The function can’t be inverted because the output 4 doesn’t point back to a unique input.
Graphing the Inverse: Symmetry About y = x
When an inverse does exist, its graph has a clean geometric relationship to the original. Every point (a, b) on the graph of f becomes the point (b, a) on the graph of f⁻¹. You’re swapping the x and y coordinates. The result is that the graph of a function and its inverse are mirror images of each other across the line y = x. If you folded the coordinate plane along that diagonal, the two curves would land on top of each other.
Restricting the Domain to Force Invertibility
Many useful functions aren’t naturally invertible over their full domain, but you can make them invertible by narrowing the domain so the function becomes one-to-one. This is one of the most practical tools in algebra and calculus.
Take y = x² + 1. Over all real numbers, it’s a parabola that fails the horizontal line test. But if you restrict the domain to x ≤ 0 (the left half of the parabola), every output now comes from exactly one input, and the restricted function passes the horizontal line test. You can then find its inverse normally. The same idea works for x ≥ 0, giving you the right half instead.
This is exactly how inverse trigonometric functions are defined. The sine function oscillates forever and is wildly non-invertible over all real numbers. But restricted to the interval [−π/2, π/2], it’s strictly increasing and one-to-one, so arcsin can be defined as its inverse on that interval. Every “inverse” of a periodic or symmetric function you encounter in math relies on this domain restriction technique.
Even more complex quadratics can be handled. A function like f(x) = x² − 3x + 2, restricted to x ≤ 1.5, keeps only the left half of the parabola (everything at or to the left of the vertex). That half is strictly decreasing, passes the horizontal line test, and is therefore invertible.
Monotonic Functions and Calculus
In calculus, there’s a powerful shortcut for identifying invertible functions. A function that is strictly increasing over an interval (always going up, never flat or dipping) is automatically one-to-one on that interval. The same is true for a function that is strictly decreasing. These are called strictly monotonic functions, and they always pass the horizontal line test because they never revisit the same output value.
If a function is differentiable, you can check this by looking at its derivative. A function with a positive derivative throughout an interval is strictly increasing there. A function with a negative derivative throughout is strictly decreasing. Either way, it’s invertible on that interval. This connection between derivatives and invertibility is one reason finding where f'(x) = 0 matters so much: those are the points where the function changes direction and potentially stops being one-to-one.
Putting It Together
The core idea is straightforward: a function is invertible when it creates a perfect one-to-one pairing between its domain and codomain, with nothing doubled up and nothing left out. You can verify this algebraically (show that f(x₁) = f(x₂) forces x₁ = x₂ and that every element of the codomain is reached), graphically (horizontal line test), or through calculus (check that the derivative never changes sign). When a function isn’t invertible on its natural domain, restricting the domain to a monotonic piece is the standard fix.