Undoing a function
The function f(x) = 3x + 1 does two things to its input, in order: it multiplies by 3, then it adds 1. For example, f(3) = 3 × 3 + 1 = 10.
The inverse function of f, written and read "f inverse", undoes f. It sends each output of f back to the input it came from: since f(3) = 10, . The −1 is part of the name, not a power: does not mean 1 divided by f(x).
Reverse the order, and reverse each step
Think of putting on socks and then shoes. To undo that, the shoes come off first and then the socks: the last thing done is the first thing undone. A function is undone the same way.
The last step of f was "add 1", so undo it first, by subtracting 1. The step before was "multiply by 3", so undo it next, by dividing by 3. Starting from 10: 10 − 1 = 9, then 9 ÷ 3 = 3, which is the input that f turned into 10. As a formula, .
Both changes are needed. Keeping the original order, dividing by 3 and then subtracting 1, turns 10 into , not 3. Reversing the order without reversing the steps, adding 1 and then multiplying by 3, turns 10 into 33.
undoes f = 3x + 1 backward: it subtracts 1 first, then divides by 3, and takes 10 back to 3.
The same steps, by algebra
The inverse can also be found by solving. Write y = 3x + 1 and make x the subject. Subtract 1 from both sides: y − 1 = 3x. Divide both sides by 3: . The algebra undoes the steps in the same reversed order, subtract 1 and then divide by 3.
This says which input x gives the output y. The inverse is a function in its own right, so write it with x as its input: .
Check by composing. , and ff⁻¹. Each function undoes the other, whichever acts first.
the square root taken first: √12.6 = 3.54, then ÷ π gives 1.13, not 2; the steps were square then × π, so they are undone as ÷ π then square root, the reverse order
Undo the steps in the order that brings r back
The area of a circle comes from its radius in two steps: square r, then multiply by . Drag the radius to change the circle, and choose the order of the undoing steps. first and then lands on a different number. first and then , the reverse order, brings back r for every radius.
Why the graph is a reflection in y = x
Every point of the graph y = f(x) is an input with its output. f(1) = 3 × 1 + 1 = 4, so (1, 4) is on the graph of f. Then , so (4, 1) is on the graph of . The inverse swaps input and output, so every point (a, b) on the graph of f becomes the point (b, a) on the graph of .
Swapping the coordinates of a point reflects it in the line y = x. To reflect a point in a line is to move it straight across the line, at a right angle, to the same distance on the other side. Check that with (1, 4) and (4, 1). The segment joining them has gradient (1 − 4) ÷ (4 − 1) = −3 ÷ 3 = −1, and y = x has gradient 1. Since −1 × 1 = −1, the segment crosses y = x at a right angle. Its midpoint is ((1 + 4) ÷ 2, (4 + 1) ÷ 2) = (2.5, 2.5), which is on y = x. So (4, 1) is the same distance from the line as (1, 4), on the other side.
The same holds for every point, so the whole graph of is the graph of f reflected in the line y = x. A point that lies on y = x is its own reflection, so wherever the graph of f crosses the line y = x, the graph of crosses it at the same point.
The gold line through the origin is y = x. The steep gold line is y = 3x + 1 and the white line is its inverse, . The marked points (1, 4) and (4, 1) are reflections of each other in y = x.
P = (a, a³) reflects in y = x to P′ = (a³, a): swap the coordinates and the mirror point traces the inverse, y = ∛x, which is still a function
On x³, drag P to (2, 8) and read where its mirror lands
The gold curve is and the green dashed curve is its reflection in y = x, which is the graph of the inverse, the cube root. Drag P along to (2, 8): its mirror P' is at (8, 2), because f(2) = 8 means . Then choose : the point at the same height on the other side of the y-axis reflects onto the same vertical line as P', so the reflected curve is not the graph of a function.
Only a one-to-one function has an inverse
Take . Since and , two inputs give the output 9. An inverse would have to send 9 back to 3 and also to −3, which is two outputs for one input, and a function gives each input exactly one output. So , with every real number allowed in, has no inverse.
A function has an inverse only when each of its outputs comes from exactly one input. Such a function is called one-to-one. On a graph, this means no horizontal line meets the graph more than once. The horizontal line y = 9 meets twice, at (−3, 9) and (3, 9). Reflected in y = x, it becomes the vertical line x = 9, which would meet the reflected curve twice, at (9, −3) and (9, 3). A vertical line meeting a graph twice means that graph is not a function.
The fix is to cut the domain. With the domain , the function is one-to-one: each output 0 or more comes from exactly one input that is 0 or more. Its inverse is , because means the square root that is not negative. Check: f(3) = 9 and .
The domain and range change places. with has the range , and that is exactly the domain of . In general, the inputs of are the outputs of f, and the outputs of are the inputs of f.
The white curve is with its domain cut to . The gold curve is its inverse, , and the straight gold line is y = x. The points (2, 4) and (4, 2) are reflections of each other: and .
The usual mistakes
Undoing the steps in the original order. For f(x) = 3x + 1, dividing by 3 and then subtracting 1 gives , which sends 10 to , not back to 3. The last step, adding 1, is undone first.
Reversing only one step. 3x − 1 turns the add into a subtract but still multiplies by 3. Every step is replaced by its opposite, so the multiply becomes a divide as well.
Reading as a reciprocal. , not .
Giving an inverse without cutting its domain. Only with , or only with , is one-to-one.
Worked example: Celsius to Fahrenheit and Back: An Inverse Function, and the Temperature That Reads the Same on Both Scales
Question A temperature of x degrees Celsius is f(x) degrees Fahrenheit, where f(x) = 1.8x + 32. (a) Find f−1(x) and use it to change 95°F into degrees Celsius. (b) Find the temperature that is the same number on both scales.
1.The function f multiplies by 1.8 and then adds 32. The inverse undoes these steps in the opposite order: first subtract 32, then divide by 1.8.
f multiplies by 1.8 and then adds 32. The inverse subtracts 32 first and then divides by 1.8. 2.In symbols, let y = 1.8x + 32. Then y − 32 = 1.8x, so x = y − 321.8. Writing the input as x, f−1(x) = x − 321.8.
Make x the subject of y = 1.8x + 32: f−1(x) = x − 321.8. Its graph is the reflection of y = f(x) in the line y = x. 3.(a) f−1(95) = 95 − 321.8 = 631.8 = 35, so 95°F is 35°C. Check: f(35) = 1.8 × 35 + 32 = 63 + 32 = 95.
(a) f−1(95) = 631.8 = 35, so 95°F is 35°C. The point (35, 95) on f is mirrored to (95, 35) on f−1. 4.A temperature that reads the same on both scales is an input that f leaves unchanged: f(x) = x, so 1.8x + 32 = x.
A temperature that reads the same on both scales is a point of y = f(x) on the line y = x: 1.8x + 32 = x. 5.Subtract x and 32 from both sides: 0.8x = −32, so x = −320.8 = −40.
0.8x = −32, so x = −40. 6.(b) The two scales agree at −40 degrees. On the graph this is the point (−40, −40), where y = f(x) meets the line y = x, and the graph of f−1 passes through the same point. Check: 1.8 × (−40) + 32 = −72 + 32 = −40.
(b) The scales agree at −40 degrees. The graph of f, the graph of f−1 and the line y = x all pass through (−40, −40).
Answer: (a) f−1(x) = x − 321.8; 35°C; (b) −40 degrees
Common mistakes
- Undoing the steps in the same order, dividing by 1.8 first and then subtracting 32, which gives 951.8 − 32 ≈ 20.8. The last step of f was to add 32, so that is the first step to undo.
- Writing f−1(x) as 11.8x + 32. The index −1 on a function means the inverse function, which reverses f. It does not mean the reciprocal of f(x).