Quarry School

College Algebra / Inverse functions

Inverse functions

Subject: Mathematics

All 7 lessons, in order.

The unit at a glance

  1. Undoing a function, and what f^-1 means

    Think of a coat check.

  2. Checking that two functions are inverses

    Think of a round trip.

  3. One-to-one functions and the horizontal line test

    Think of a coat check.

  4. Domain and range trade places

    Picture a coat check.

  5. Reading an inverse from a table and from a graph

    Picture the sign-in sheet at a gym.

  6. Restricting a domain so an inverse exists

    Picture a town on a long straight road.

  7. Finding the inverse of a formula by swapping and solving

    Think of wrapping a gift.

Rules (7)
  • Undoing a function, and what f^-1 means

    When f has an inverse function and f(a) = b, then f−1(b) = a. The inverse function f−1 takes each output of f back to the input it came from. The −1 is part of the name, not a power: f−1(x) does not mean 1f(x).

  • Checking that two functions are inverses

    Two functions f and g are inverses of each other exactly when both of these are true: f(g(x)) = x for every input x that g accepts, and g(f(x)) = x for every input x that f accepts. When both hold, g is f−1 and f is g−1. If either one fails, even for a single input, they are not inverses.

  • One-to-one functions and the horizontal line test

    A function f is a one-to-one function when different inputs always give different outputs: f(a) = f(b) happens only when a = b. A function has an inverse function exactly when it is one-to-one. Horizontal line test: a function is one-to-one exactly when no horizontal line meets its graph at more than one point.

  • Domain and range trade places

    If f is a one-to-one function, then: domain of f−1 = range of f, and range of f−1 = domain of f.

  • Reading an inverse from a table and from a graph

    If f is a one-to-one function, then: (1) the table of f−1 is the table of f with its two rows traded; (2) the point (a, b) is on the graph of f exactly when the point (b, a) is on the graph of f−1; (3) the graph of f−1 is the graph of f reflected (flipped, as in a mirror) across the line y = x.

  • Restricting a domain so an inverse exists

    A restricted domain is a smaller set of inputs than the full domain of a function. If f fails the horizontal line test, choose a restricted domain on which f is one-to-one. On that piece f has an inverse function. The usual choice also keeps every output of f, so the inverse accepts all of them. For f(x) = x2 the usual restricted domain is x ≥ 0, written [0, ∞). The domain of f−1 is then the range of f, and the range of f−1 is the restricted domain.

  • Finding the inverse of a formula by swapping and solving

    For a one-to-one function f: write y = f(x), swap x and y to get x = f(y), solve for y, and name the result f−1(x). If solving produces ±, the domain of f picks the sign, because the outputs of f−1 are the inputs of f. The domain of f−1 is the range of f.

Recipes (7)
  • Undoing a function, and what f^-1 means
    1. Read f−1(b) as a question in plain words: b came out of f. Which input went in?
    2. Rewrite the question as f(?) = b. The number inside f−1( ) is an output of f.
    3. Find that input. In a table, find b in the output row and read the input above it. With a formula, set the formula equal to b and solve for x.
    4. Write the answer as f−1(b) = the input you found.
    5. Check by going forward: put your answer into f and confirm that b comes out.
  • Checking that two functions are inverses
    1. Write down both rules, f(x) and g(x), side by side.
    2. First order, f(g(x)): copy the rule for f, but wherever it has x, write an empty pair of parentheses. Then put the whole formula for g(x) inside the parentheses.
    3. Simplify. The result must be exactly x. Not x + 2, not 3x, not −x.
    4. Second order, g(f(x)): copy the rule for g with empty parentheses in place of x, and put the whole formula for f(x) inside.
    5. Simplify. Again the result must be exactly x.
    6. Decide. If both orders gave x, for every input the inner function accepts, f and g are inverses. If either one gave anything else, they are not, and you can stop at the first failure.
    7. Spot check with one number: pick a number f accepts, run it through f and then g, and confirm it comes back. Then pick a number g accepts and run it through g and then f.
  • One-to-one functions and the horizontal line test
    1. Say the question in plain words: do any two different inputs share an output?
    2. From a table or a list of pairs: read along the outputs. A value that sits under two different inputs means the function is not one-to-one. No repeated output means it is one-to-one.
    3. From a graph: hold a horizontal line at the bottom of the picture and slide it to the top. If at any height it meets the graph at two or more points, the function is not one-to-one. If it meets the graph at most once at every height, the function is one-to-one.
    4. From a formula: look at its graph and use the horizontal line test, or write f(a) = f(b) and see whether that forces a = b.
    5. To say a function is not one-to-one, write down the proof: two different inputs and the output they share, such as f(−2) = 4 and f(2) = 4.
    6. State the conclusion about undoing: a one-to-one function has an inverse function, and a function that is not one-to-one does not.
  • Domain and range trade places
    1. Make sure f is a one-to-one function, so that f−1 exists: no output repeated in a table, or a graph that passes the horizontal line test.
    2. Find the domain of f: every input f allows.
    3. Find the range of f: every output f actually produces. List only outputs that really come out, not every number you can imagine.
    4. Trade the two lists. Domain of f−1 = range of f. Range of f−1 = domain of f.
    5. Check with one pair. Pick an allowed input a and compute b = f(a). Then f−1(b) = a, so b must sit in your new domain and a must sit in your new range.
  • Reading an inverse from a table and from a graph
    1. Check that f is a one-to-one function. In a table, no number may appear twice in the output row. On a graph, use the horizontal line test: no horizontal line may meet the graph twice.
    2. To find one value, such as f−1(7), say what it asks: 'f gave the output 7; which input was it?' Find 7 in the output row of f, or at height 7 on the y-axis, and read back to the input.
    3. To get the whole table of f−1, trade the two rows. The old output row becomes the new input row.
    4. If you want the new inputs in order from smallest to largest, slide whole columns. Never separate an input from its own output.
    5. To get the graph of f−1, pick several clear points on the graph of f, write each one as (a, b), and swap it to (b, a).
    6. Plot the swapped points and join them with the same kind of shape the original had. Draw the line y = x lightly: each point and its swap must sit straight across that line from each other, the same distance away.
    7. Check one value both ways: if you wrote f−1(b) = a, the table or graph of f must show f(a) = b.
  • Restricting a domain so an inverse exists
    1. Run the horizontal line test on the whole graph. If no horizontal line meets it twice, stop: the function already has an inverse and needs no restriction.
    2. If the test fails and the graph turns around once, as a parabola or an absolute value graph does, find that turn. For a squared expression, the turn is where the part being squared equals 0. For x2 that is x = 0. For (x − 2)2 it is x = 2.
    3. For a graph with that one turn, keep the turning point and everything on one side of it. Unless the problem names a side, keep the right side: x ≥ the turning value.
    4. Run the horizontal line test again on the kept piece only. Every horizontal line must now meet it at most once.
    5. Check whether any output was lost. The usual choice keeps the range of the whole function; a piece with a smaller range still has an inverse, but that inverse accepts fewer numbers.
    6. Write the restricted domain next to the formula, for example f(x) = x2, x ≥ 0. It is now part of the function.
    7. Trade places: the domain of f−1 is the range of f, and the range of f−1 is the restricted domain. Every answer the inverse gives must lie in the restricted domain.
  • Finding the inverse of a formula by swapping and solving
    1. Make sure f is a one-to-one function on its domain. Use the horizontal line test, or use the restricted domain you are given. If two inputs share an output, stop: there is no inverse function yet.
    2. Write y in place of f(x). If the function comes with a domain condition such as x ≥ 0, write it beside the equation.
    3. Swap the letters. Every x becomes y and every y becomes x, in the equation and in the domain condition. So x ≥ 0 becomes y ≥ 0.
    4. Solve the swapped equation for y. One y: undo the steps in reverse order. Two y's in a fraction: clear the fraction, gather the y terms on one side, factor y out, divide. A squared y: take the square root of both sides and write ±.
    5. If ± appeared, keep the one sign that fits the condition on y.
    6. Write f−1(x) = your result, and state its domain, which is the range of f.
    7. Check. Send one number through f, then send the result through f−1. You should land on the number you started with.
Traps (7)
  • Undoing a function, and what f^-1 means

    Reading f−1(x) as 1f(x). On the name of a function, −1 means inverse, not 1 over. To avoid the mistake, say f inverse out loud and turn f−1(b) into the question f(?) = b before you compute anything. When 1 over f(x) is meant, it is written 1f(x)−1 or [f(x)], with the −1 outside brackets.

  • Checking that two functions are inverses

    Stopping after one order. Getting f(g(x)) = x feels like the finish, but it is half of the test. With f(x) = x2 and g(x) = x, the first order gives (x)2 = x and passes, while the second order gives g(f(−3)) = 9 = 3, which is not −3. Always write both lines, f(g(x)) and g(f(x)), before you answer yes.

  • One-to-one functions and the horizontal line test

    Using the wrong line. A vertical line tests whether a graph is a function at all. A horizontal line tests whether that function is one-to-one. The graph of y = x2 passes the first test and fails the second. Say the question out loud before you draw the line: "does one input have two outputs?" calls for a vertical line, and "do two inputs share one output?" calls for a horizontal line.

  • Domain and range trade places

    Writing the domain of f again as the domain of f−1. The two functions trade their lists, and the two lists are often different. For f(x) = x−2 + 1 the tempting answer is 'domain of f−1: [2, ∞)', but [2, ∞) is what f−1 gives back, not what it takes in. To avoid this, write two labeled lines before you answer, 'goes into f−1' and 'comes out of f−1', and fill the first line from the outputs of f.

  • Reading an inverse from a table and from a graph

    Reading in the forward direction. For f−1(5), the 5 is an output of f, so look for it in the output row of the table, or at height 5 on the y-axis of the graph. If you look for 5 in the input row, or at 5 on the x-axis, you find f(5), the answer to the forward question. Before you read anything, say the question aloud: 'f gave 5; what went in?'

  • Restricting a domain so an inverse exists

    The most common mistake is to restrict the domain on paper and then forget it. Once you keep x ≥ 0 for f(x) = x2, the inputs −1, −2, −3 and the rest are gone. So f−1(9) is 3 only. It is not "3 or −3", and it is not −3. To avoid the mistake, write the restricted domain beside the formula every time you write the function, and check that each answer from the inverse lies inside the restricted domain.

  • Finding the inverse of a formula by swapping and solving

    The most common mistake comes with the square root. After x = y2, people write y = x out of habit, or leave y = ±x as the answer. Neither is safe. Write ± first, then look at the condition on y that came from the domain of f, and keep the one sign that fits it. For x ≥ 0 that is x. For x ≤ 0 it is −x. Then test with a number from the domain of f: it must come back to itself.

Lesson 1 of 7

Undoing a function, and what f^-1 means

Subject: MathematicsLeans on: Function notationLeans on: Reading a table of valuesLeans on: A power of −1 on a numberLeans on: Solving a linear equationOpen this lesson
Leans on 1Comes back in 0

1 In plain words

Think of a coat check. You hand over your coat and get a numbered ticket. That is a function: coat in, ticket number out. At the end of the night you run it backward: ticket number in, coat out. The backward process is the inverse function. It takes each output back to the input it came from. If a function f turns 3 into 7, its inverse turns 7 back into 3. The inverse is written f−1 and read as f inverse. The small −1 is part of the name, a tag that says: this is f, run backward. It is not a power, and it does not mean 1 divided by f.

2 Picture it

12347597210function
Each arrow is f going forward.

Reminder

Reminder: Function notation, Reading a table of values, A power of −1 on a number, Solving a linear equation
  • Function notation. f(3) = 7 means the input is 3 and the output is 7. The number inside the parentheses goes in. The number after the equals sign comes out.
  • Reading a table of values. Inputs are in the top row and each output sits directly below its input. If 3 is on top and 7 is under it, then f(3) = 7.
  • A power of −1 on a number. On a number, the power −1 means 1 divided by that number, called its reciprocal: 5−1 = 15 and 2−1 = 12.
  • Solving a linear equation. Undo the steps in reverse order, doing the same thing to both sides: 2x + 3 = 11 becomes 2x = 8 after subtracting 3, then x = 4 after dividing by 2.

3 Why

Undoing means getting back what you started with. If f sends a to b, the only way to return to a is to send b to a. So f(a) = b forces f−1(b) = a. The symbol itself is a convention. People chose −1 because undoing a function is a little like dividing after multiplying: both take you back to where you began. But here the −1 sits on the name f, not on a number, so it does not mean 1 over f. Ignore that and answers break: with f(x) = 2x + 3, f−1(11) = 4, while 1f(11) = 125.

4 The rule

When f has an inverse function and f(a)=b, then f−1(b)=a.

The inverse function f−1 takes each output of f back to the input it came from. The −1 is part of the name, not a power: f−1(x) does not mean 1f(x).

The same idea, five ways

Say it

f inverse of 7 equals 3. Say f inverse, never f to the minus one.

Write it

The inverse function of f takes 7, an output of f, back to 3, the input it came from.

In math
  • f(a) = b says the same thing as f−1(b) = a
  • f(3) = 7, so f−1(7) = 3
  • f−1(x) does not mean 1f(x)
  • −1[f(x)] = 1f(x), a reciprocal, because the −1 sits outside the brackets
Like

A coat check run backward: hand in ticket 7 and get back coat 3, the coat that ticket was given for.

See it
7, an output of ff⁻¹: which input didf turn into 7?inputoutput3
The inverse machine takes in an output of f and gives back the input that made it.

The same idea, other ways

As a picture: walk the arrows backward

Draw the inputs of f in one oval and the outputs in another, with an arrow from each input to its output. f follows the arrows from tail to tip: 1 to 4, 2 to 6, 3 to 8. The inverse function f−1 uses the very same arrows, walked from tip to tail: 4 back to 1, 6 back to 2, 8 back to 3. No new arrows are drawn. So f(2) = 6 and f−1(6) = 2 are one arrow, read in two directions.

123468function
f goes along each arrow from left to right. f−1 goes back along the same arrow from right to left.
As a story: caller ID

A contact list with one number saved under each name is a function: you pick a name and it gives that number. When no two names share a number, caller ID is its inverse function: a number on the list comes in and the phone shows the name. Caller ID does no arithmetic on the phone number. It does not divide 1 by it. It looks up who the number belongs to. That is what f−1 does: it looks up which input an output belongs to.

With small numbers: a round trip

Let g be the function that adds 5. Put in 4 and g gives 9. To undo that, subtract 5, and 9 goes back to 4. So g(4) = 9 and g−1(9) = 4. Try another: g(10) = 15, so g−1(15) = 10. Going through g and then through g−1 brings you back to the number you started with, like walking 5 steps forward and then 5 steps back.

4g: add 5g⁻¹: subtract5firstsecond94
The number 4 goes in, becomes 9, and the inverse brings it back to 4.
A rule of thumb: two questions

f answers the question: this went in, what comes out? f−1 answers the opposite question: this came out, what went in? Whenever you see f−1(7), say to yourself: 7 came out of f, what went in? Then rewrite it as f(?) = 7 and hunt for the input.

f(3) = 7: 3 went in, 7 came out

f−1(7) = 3: 7 came out, 3 went in

One fact, read in two directions

Why the −1 cannot be a power

Test it with numbers. Let g add 5, so g(4) = 9, and undoing gives g−1(9) = 4. If the −1 were a power, g−1(9) would be 1 divided by g(9). But g(9) = 14, so that reading gives 114. The number 114 does not take you back to 4, so it undoes nothing. A power of −1 belongs on a number, as in 5−1 = 15. On the name of a function, −1 means inverse.

WrittenSay itWhat it meansValue, with f(x) = 2x + 3
f(11)f of 11Put 11 into f and read the output.2 · 11 + 3 = 25
f−1(11)f inverse of 1111 came out of f. Find the input that went in.4, because f(4) = 2 · 4 + 3 = 11
1f(11)one over f of 111 divided by the output f(11). This is a reciprocal.125
−1[f(11)]f of 11, all to the power −1Another way to write 1f(11). The brackets make the −1 a power.125
5−15 to the power −1On a number, −1 is a power: 1 divided by the number.15, no function involved

5 How

  1. Read f−1(b) as a question in plain words: b came out of f. Which input went in?
  2. Rewrite the question as f(?) = b. The number inside f−1( ) is an output of f.
  3. Find that input. In a table, find b in the output row and read the input above it. With a formula, set the formula equal to b and solve for x.
  4. Write the answer as f−1(b) = the input you found.
  5. Check by going forward: put your answer into f and confirm that b comes out.

Evaluating f−1(b) at one number

1
−1Is the −1 outside brackets, as in [f(b)], or is the expression written 1f(b)?
YesThat is a reciprocal. Find f(b), then write 1 over it. Stop here.
NoIt is the inverse function f−1(b). Go on.
↓
2
Are you given one fact such as f(a) = b?
YesRead it backward: f−1(b) = a.
NoGo on.
↓
3
Is f given as a table?
YesFind b in the output row. The input above it is f−1(b).
NoGo on.
↓
4
Is f given as a formula?
YesSet the formula equal to b and solve for x. That x is f−1(b).
NoYou need more information about f before you can answer.
  1. Look at where the −1 sits. On the name, as in f−1−1(b), it means the inverse function. Outside brackets, as in [f(b)], it means 1 over f(b).
  2. Say the question in plain words: b came out of f. Which input went in?
  3. Rewrite it as f(?) = b.
  4. Find the input from whatever describes f: one given fact, a table, or a formula.
  5. Write f−1(b) = that input.
  6. Go forward to check: put the input into f and see that b comes out.

6 Worked example

Worked example

Reading f−1(7) from a table, and telling it apart from two look-alikes

The table shows a function f. Find f−1(7). Then show that the answer is not f(7) and not 1f(7). In plain words, f−1(7) asks: 7 came out of f, which input went in?

input xoutput f(x)15293742710
x: 1, 2, 3, 4, 7; f(x): 5, 9, 7, 2, 10.
Show the steps and the answer
input xoutput f(x)15293742710↑ solve: output given, read every input above it
The 7 in the output row is highlighted, and the arrow points up to its input, 3.
  1. Say what is asked. f−1(7) means: 7 came out of f. Find the input that went in.f−1 takes an output of f back to its input, so the 7 here is an output of f, not an input.
  2. Rewrite the question as f(?) =7.f(a) = b and f−1(b) = a are the same fact read in two directions.
  3. Look along the bottom row of the table, the f(x) row, for 7. It appears once, in the third column.The bottom row holds the outputs of f, and 7 is being used as an output.
  4. Read the input directly above that 7. It is 3. So f(3)=7.In the table each output sits under the input that produced it.
  5. Write the answer: f−1(7)=3.The rule says f(a) = b gives f−1(b) = a. Here a = 3 and b = 7.
  6. Compare with f(7). Find 7 in the top row, in the last column. Below it is 10, so f(7)=10.f(7) uses 7 as an input. f−1(7) uses 7 as an output. Those are different questions, and here their answers differ: 3 and 10.
  7. Compare with 1f(7). Since f(7)=10, this is 110.1 over f(7) is 1 divided by an output. It is a reciprocal, not the inverse function.
Answer
  • f−1(7)=3
  • f(7) = 10, a different question
  • 1f(7) = 110, a reciprocal and not an inverse
Double-check Go forward. Put the answer 3 into f: the column under the input 3 shows 7, so f(3) = 7. Since f turns 3 into 7, undoing f turns 7 back into 3. The answer holds.

Ladder: from easy to exam-hard

Try each rung before you open its steps.

Rung 1 of 6

One fact, read backward

A one-to-one function h has h(2) = 9. Find h−1(9). In plain words: 9 came out of h, which input went in?

Show the steps and the answer
  1. Read the fact h(2)=9 as: input 2, output 9.In function notation the number inside the parentheses is the input and the number after the equals sign is the output.
  2. Turn the fact around: the output 9 goes back to the input 2. So h−1(9)=2.The rule says f(a) = b gives f−1(b) = a. Here a = 2 and b = 9.
Answer
h−1(9)=2
Double-check Go forward: the given fact says h(2) = 9, so 2 is the input that produces 9. The answer is not 19, and it is not h(9), which the problem never gives.

Rung 2 of 6

A table where the same number sits in both rows

The table shows a function p. Find p−1(3) and p−1(0). In plain words: 3 came out of p, which input went in? Then the same question for 0.

input toutput p(t)0310243142
t: 0, 1, 2, 3, 4; p(t): 3, 0, 4, 1, 2.
Show the steps and the answer
input toutput p(t)0310243142↑ solve: output given, read every input above it
The highlighted 3 is the one in the output row, and its arrow points up to the input 0.
  1. For p−1(3), look for 3 in the bottom row, the p(t) row. It is in the first column.The number inside p−1( ) is an output of p, and outputs live in the bottom row. Gotcha: 3 also sits in the top row, but that 3 is an input.
  2. Read the input above it: 0. So p−1(3)=0.p(0) = 3, and the inverse reads that fact backward.
  3. For p−1(0), look for 0 in the bottom row. It is in the second column, under the input 1.Again the number inside p−1( ) is an output, so search the output row.
  4. Write p−1(0)=1.p(1) = 0, read backward.
Answer
  • p−1(3)=0
  • p−1(0)=1
Double-check Go forward: the column under the input 0 shows 3, so p(0) = 3. The column under the input 1 shows 0, so p(1) = 0. Compare with the tempting wrong reading: p(3) = 1, which is not p−1(3) = 0.

Rung 3 of 6

A formula with one step

Let g(x) = x + 5. Find g−1(12). In plain words: 12 came out of g, which input went in?

Show the steps and the answer
12g⁻¹: undo add 5, sosubtract 5inputoutput7
The inverse of adding 5 is subtracting 5, so 12 goes back to 7.
  1. Rewrite the question as g(x)=12, which is x+5=12.g−1(12) is the input that g turns into 12. This equation finds that one input.
  2. Subtract 5 from both sides: x=12−5=7.g added 5, and subtracting 5 undoes adding 5.
  3. Write g−1(12)=7.7 is the input whose output is 12.
Answer
g−1(12)=7
Double-check Go forward: g(7) = 7 + 5 = 12. The output is 12, as required.

Rung 4 of 6

A formula with two steps

Let f(x) = 2x + 3. Find f−1(11). In plain words: 11 came out of f, which input went in?

Show the steps and the answer
=2x + 311do the same thing to both sides
Finding f−1(11) is the same job as solving 2x + 3 = 11 for x.
  1. Rewrite the question as f(x)=11, which is 2x+3=11.f−1(11) is the input that f turns into 11. Solving this equation finds it.
  2. Subtract 3 from both sides: 2x=8.f multiplied by 2 first and added 3 last. Undo the last step first.
  3. Divide both sides by 2: x=4.Dividing by 2 undoes multiplying by 2.
  4. Write f−1(11)=4.4 is the input whose output is 11.
Answer
f−1(11)=4
Double-check Go forward: f(4) = 2 · 4 + 3 = 8 + 3 = 11. The output is 11, as required.

Rung 5 of 6

What an inverse means in a story

A ride costs C(m) = 3m + 4 dollars for a trip of m miles. Say in words what C−1(19) means, then find it.

Show the steps and the answer
  1. Say what C does: miles go in, dollars come out.The input m is miles and the output C(m) is the cost.
  2. Say what C−1 does: dollars go in, miles come out. So C−1(19) is the number of miles that costs 19 dollars.The inverse function takes each output of C, a cost, back to the input it came from, a distance.
  3. Rewrite as C(m)=19, which is 3m+4=19.This equation finds the one distance whose cost is 19 dollars.
  4. Subtract 4 from both sides: 3m=15.C added 4 last, so undo that first.
  5. Divide both sides by 3: m=5.Dividing by 3 undoes multiplying by 3.
Answer
  • C−1(19)=5
  • Meaning: a ride that costs 19 dollars is 5 miles long.
Double-check Go forward: C(5) = 3 · 5 + 4 = 15 + 4 = 19 dollars. A 5 mile ride costs 19 dollars, so 19 dollars goes back to 5 miles.

Rung 6 of 6

The inverse and the reciprocal side by side

Let k(x) = 3x − 2. Find k−1(10) and 1k(10), and show they are different numbers.

Show the steps and the answer
  1. For k−1(10), rewrite the question as k(x)=10, which is 3x−2=10.k−1(10) asks: 10 came out of k, which input went in?
  2. Add 2 to both sides: 3x=12.k subtracted 2 last, and adding 2 undoes that.
  3. Divide both sides by 3: x=4. So k−1(10)=4.Dividing by 3 undoes multiplying by 3.
  4. For 1k(10), first find k(10)=3·10−2=28.Here 10 is an input of k, so it goes into the formula.
  5. Write 1 over that output: 1k(10)=128.A reciprocal is 1 divided by the number.
  6. Compare: 4 is not 128.One answers which input gives 10. The other divides 1 by the output at 10. They are different questions.
Answer
  • k−1(10)=4
  • 1k(10)=128
  • The two are different, so k−1(x) is not 1k(x).
Double-check For the inverse, go forward: k(4) = 3 · 4 − 2 = 12 − 2 = 10. For the reciprocal: 28 × 128 = 1, which is what a reciprocal must do.

Counterexamples

What it is not, and the tempting wrong moves.

✗ Not this: f−1(7) = 1f(7). In the example table f(7) = 10, so f−1(7) = 110.
The −1 on the name of a function is not a power. 110 is not even an input in the table, so it cannot be the input that produced 7.
✓ Instead: Find 7 among the outputs. It sits under the input 3, so f−1(7) = 3.
✗ Not this: To find f−1(7), find 7 in the input row and read the number below it, which gives 10.
That is f(7), the forward direction. In f−1(7) the 7 is an output of f.
✓ Instead: Find 7 in the output row and read the input above it: f−1(7) = 3.
✗ Not this: f(3) = 7, so f−1(3) = 7.
The inverse must be fed the output, 7, not the input, 3. The fact f(3) = 7 says nothing about what f−1 does to 3.
✓ Instead: f(3) = 7 gives f−1(7) = 3. The two numbers trade places.
✗ Not this: The inverse is the opposite, so if g(x) = x + 5 then g−1(9) = −g(9) = −14.
Inverse means undo, not change the sign. g(4) = 9, so the inverse has to send 9 back to 4, and −14 does not do that.
✓ Instead: g adds 5, so g−1 subtracts 5: g−1(9) = 9 − 5 = 4.
✗ Not this: A −1 written up high is never a power, so 5−1 must be some kind of inverse of 5.
The rule in this lesson is about the name of a function. On a number or a variable, −1 is an ordinary power.
✓ Instead: 5−1 = 15 and x−1 = 1x. Only on a function name, as in f−1(x), does −1 mean inverse.

Tips and tricks

  • Say f inverse out loud. Saying f to the minus one invites the 1 over mistake.
  • Translate before you compute: f−1(b) = ? becomes f(?) = b.
  • Memory cue: f asks what comes out. f inverse asks what went in.
  • Check every answer by going forward. Put your answer into f. If the number you started with comes out, you are right.
  • Look at where the −1 sits. Right after the name and before the parentheses, f−1−1(x), it means inverse. Outside brackets, [f(x)], or on a number, it is a power.
  • In a table, the number inside f−1( ) is found in the output row, never the input row.
  • Every function in this lesson has each output coming from only one input, so there is one way back. What happens when two inputs share an output is handled later in this unit.

7 The trap

Reading f−1(x) as 1f(x). On the name of a function, −1 means inverse, not 1 over. To avoid the mistake, say f inverse out loud and turn f−1(b) into the question f(?) = b before you compute anything. When 1 over f(x) is meant, it is written 1f(x)−1 or [f(x)], with the −1 outside brackets.

Flash cards

Say the answer out loud, then flip.

At a coat check, the function is: coat in, ticket number out. What does the inverse function do?
It runs the coat check backward: ticket number in, coat out. It takes each output back to the input it came from.
A function f has an inverse, and f(5) = 12. What is f−1(12), and why?
f−1(12) = 5. Since f sends 5 to 12 and sends no other input to 12, the way back sends 12 to 5.
Does f−1(x) mean 1f(x)?
No. The −1 is part of the name. It is a tag that says: this is f, run backward. It is not a power. When 1 over f(x) is meant, it is written 1f(x)−1 or [f(x)].
With f(x) = 2x + 3, find f−1(9). Say first what the question asks.
  • It asks: 9 came out of f. Which input went in?
  • Set 2x + 3 = 9, so 2x = 6 and x = 3.
  • f−1(9)=3.
  • Check by going forward: f(3) = 2 · 3 + 3 = 9.

Lesson 2 of 7

Checking that two functions are inverses

Subject: MathematicsLeans on: Composing two functionsLeans on: Substituting a whole expressionLeans on: DistributingLeans on: Multiplying cancels dividingLeans on: A fraction times its flipLeans on: Square rootsLeans on: Squaring a negativeLeans on: What f^{−1} meansOpen this lesson
Leans on 0Comes back in 1

1 In plain words

Think of a round trip. One bus takes you from home to work. A second bus claims to run the same route backwards. To test the claim, ride the first bus and then the second, and see whether you step off where you started. Then test it from the other end: start at work, ride the second bus and then the first. If both round trips end where they began, from every stop, the two routes undo each other. Two functions are tested the same way. Put a number into one, feed the result into the other, and see whether your starting number comes back. Do it in both orders. If you always get your starting number back, each one is the inverse function of the other.

2 Picture it

xgffirstsecondg(x)x
If f and g are inverses, whatever goes into the first machine comes out of the second unchanged.

Reminder

Reminder: Composing two functions, Substituting a whole expression, Distributing, Multiplying cancels dividing, A fraction times its flip, Square roots, Squaring a negative, What f−1 means
  • Composing two functions. f(g(x)) means do g first, then feed its output into f. Work from the inside out: if g(2) = 7 and f(7) = 1, then f(g(2)) = 1.
  • Substituting a whole expression. Replace x by the whole expression, in parentheses: with f(x) = 2x + 6, f(x − 3) = 2(x − 3) + 6, not 2x − 3 + 6.
  • Distributing. The number in front multiplies every term inside: 2(x − 3) = 2x − 6.
  • Multiplying cancels dividing. 3 · x+53 = x + 5, the same way 3 · 123 = 12.
  • A fraction times its flip. 23 · 32 = 66 = 1. Multiply top by top and bottom by bottom.
  • Square roots. 9 = 3 because 3 × 3 = 9. The √ sign always gives the answer that is not negative, so 9 is 3 and never −3.
  • Squaring a negative. (−3)2 = (−3) × (−3) = 9. A negative times a negative is positive.
  • What f−1 means. The −1 in f−1 is part of a name, not a power. f−1(x) is the inverse function, not 1f(x).

3 Why

An inverse function sends each output back to its input. So if f(a) = b, the inverse must give g(b) = a. Chain those two facts: g(f(a)) = g(b) = a. That is the statement g(f(x)) = x, and it says g undoes f. But being inverses is a two-way promise: f must also undo g, which is f(g(x)) = x. One order can pass while the other fails. Squaring sends both 3 and −3 to 9. A square root machine sends 9 back to 3 only, so the trip that starts at −3 never gets home. Checking one order would miss that.

4 The rule

Two functions f and g are inverses of each other exactly when both of these are true: f(g(x))=x for every input x that g accepts, and g(f(x))=x for every input x that f accepts.

When both hold, g is f−1 and f is g−1. If either one fails, even for a single input, they are not inverses.

The same idea, five ways

Say it

f of g of x equals x, and g of f of x equals x.

Write it

Doing g and then f returns every input unchanged, and doing f and then g returns every input unchanged.

In math
  • f(g(x)) = x
  • g(f(x)) = x
  • (f ∘ g)(x) = x and (g ∘ f)(x) = x, where the small circle is another way to write f(g(x)) and g(f(x))
  • g = f−1 and f = g−1
Like

A round trip: out on one bus, back on the other, and you step off where you got on, whichever end you start from.

See it

f(g(x)) = x

g(f(x)) = x

Both must be true.

Then g = f−1.

The same idea, other ways

As a picture

Line up two machines. Drop 4 into f, which multiplies by 3 and subtracts 5: out comes 7. Drop that 7 into g, which adds 5 and divides by 3: out comes 4. The number you put in is the number you get out. Now swap the machines and do it again. If the number always survives, the machines are inverses.

4f: times 3,then minus 5g: plus 5,then divide by3firstsecond74
Follow the 4: it changes to 7 in the middle and returns to 4 at the end.
As a story

A secret code moves every letter 3 places forward in the alphabet. A decoder moves every letter 3 places back. To trust the pair, you test two things. Encode a message and then decode it: you should read the original message. Decode a message and then encode it: you should again read the original. If either test garbles the text, the two are not a matched pair.

With small numbers

Let f add 4 and let g subtract 4. Start with 10. Through f you get 14, then through g you get 10. Now the other order: through g you get 6, then through f you get 10. Any starting number behaves the same way, because (x − 4) + 4 = x and (x + 4) − 4 = x. Both orders return x, so f and g are inverses.

Why one order is not enough

Let f square its input and let g take the square root. Start with 9, g first: 9 = 3, then 32 = 9. Home again. Now start with −3, f first: (−3)2 = 9, then 9 = 3. You left from −3 and arrived at 3. The trouble is that f sends two different inputs, 3 and −3, to the same output 9, and g can send 9 back to only one of them. The first order never notices. The second order does.

3−39function
Two arrows land on 9, so no function can send 9 back to both 3 and −3.
Rule of thumb

Out and back, both ways. If either trip leaves you anywhere except x, the answer is no.

CheckWhat you doWhat it shows
f(g(x)) = xg first, then ff undoes g
g(f(x)) = xf first, then gg undoes f
Both give xBoth round trips come homef and g are inverses

5 How

  1. Write down both rules, f(x) and g(x), side by side.
  2. First order, f(g(x)): copy the rule for f, but wherever it has x, write an empty pair of parentheses. Then put the whole formula for g(x) inside the parentheses.
  3. Simplify. The result must be exactly x. Not x + 2, not 3x, not −x.
  4. Second order, g(f(x)): copy the rule for g with empty parentheses in place of x, and put the whole formula for f(x) inside.
  5. Simplify. Again the result must be exactly x.
  6. Decide. If both orders gave x, for every input the inner function accepts, f and g are inverses. If either one gave anything else, they are not, and you can stop at the first failure.
  7. Spot check with one number: pick a number f accepts, run it through f and then g, and confirm it comes back. Then pick a number g accepts and run it through g and then f.

Testing whether f and g are inverses

1
Pick a number a that f accepts. Does g(f(a)) come back to a?
YesA good sign, but not proof. Go on and test with x.
NoStop. They are not inverses. One failed number is enough.
↓
2
Does f(g(x)) simplify to exactly x?
YesGo on to the other order.
NoStop. They are not inverses.
↓
3
Does g(f(x)) simplify to exactly x?
YesBoth orders passed for every input each inner function accepts: f and g are inverses.
NoThey are not inverses, even though the first order passed.
  1. Quick screen: pick one number a that f accepts, find f(a), then put that result into g. If you do not get a back, stop: they are not inverses.
  2. Write f(g(x)): the rule for f with the whole of g(x), in parentheses, in place of x. Simplify.
  3. If the result is not exactly x, stop: they are not inverses.
  4. Write g(f(x)): the rule for g with the whole of f(x), in parentheses, in place of x. Simplify.
  5. If this result is also exactly x for every input f accepts, f and g are inverses. If not, they are not, even though the first order passed.
  6. Check your algebra with a number in both orders.

6 Worked example

Worked example

A linear pair, checked in both orders

Are f(x) = 3x − 5 and g(x) = x+53 inverses of each other? In plain words: if you run any number through one of them and then through the other, in either order, do you always get that number back?

Show the steps and the answer
4f: times 3,then minus 5g: plus 5,then divide by3firstsecond74
The 4 that goes into f comes back out of g, which is what g(f(x)) = x looks like for one number.
  1. Write the plan: compute f(g(x)) and g(f(x)), and see whether each one simplifies to x.Being inverses is a two-way promise, so there are two things to test.
  2. Start f(g(x)). The rule for f is f( ) = 3( ) − 5. Put g(x) in the parentheses: f(g(x))=3·x+53−5.f(g(x)) means g goes first and its whole output becomes the input of f.
  3. Simplify 3·x+53 to x+5.Multiplying by 3 cancels dividing by 3, the same way 3 · 123 = 12.
  4. Now f(g(x))=x+5−5=x.Adding 5 and then subtracting 5 leaves nothing extra. The first order passes.
  5. Start g(f(x)). The rule for g is g()=()+53. Put f(x) in the parentheses: g(f(x))=(3x−5)+53.g(f(x)) means f goes first and its whole output becomes the input of g.
  6. Simplify the top: 3x−5+5=3x. So g(f(x))=3x3.−5 and +5 add to 0.
  7. 3x3=x.Dividing by 3 cancels multiplying by 3. The second order passes.
  8. Both orders gave exactly x, so f and g are inverses.That is the rule: f(g(x)) = x and g(f(x)) = x together.
Answer
  • f(g(x)) = x
  • g(f(x)) = x
  • Yes, f and g are inverses of each other, so g = f−1.
Double-check A second way, with a real number in place of x. Start with 4. f first: f(4) = 3 · 4 − 5 = 7, then g(7) = 7+53 = 123 = 4. Back to 4. Now g first: g(4) = 4+53 = 93 = 3, then f(3) = 3 · 3 − 5 = 4. Back to 4 again. Both round trips agree with the algebra.

Ladder: from easy to exam-hard

Try each rung before you open its steps.

Rung 1 of 5

Adding and subtracting

Are f(x) = x + 4 and g(x) = x − 4 inverses? In plain words: does running a number through one and then the other, in either order, always return that number?

Show the steps and the answer
  1. f(g(x))=(x−4)+4.f adds 4 to its input, and the input is all of g(x), which is x − 4.
  2. (x−4)+4=x.−4 and +4 add to 0. The first order passes.
  3. g(f(x))=(x+4)−4.g subtracts 4 from its input, and the input is all of f(x), which is x + 4.
  4. (x+4)−4=x.+4 and −4 add to 0. The second order passes.
Answer
  • f(g(x)) = x
  • g(f(x)) = x
  • Yes, they are inverses.
Double-check With the number 10: g(10) = 6 and f(6) = 10. The other order: f(10) = 14 and g(14) = 10. Both trips return 10.

Rung 2 of 5

Multiplying and dividing

Are f(x) = 5x and g(x) = x5 inverses? The question asks whether both round trips return x.

Show the steps and the answer
  1. f(g(x))=5·x5.f multiplies its input by 5, and the input is g(x) = x5.
  2. 5·x5=x.Multiplying by 5 cancels dividing by 5. The first order passes.
  3. g(f(x))=5x5.g divides its input by 5, and the input is f(x) = 5x.
  4. 5x5=x.Dividing by 5 cancels multiplying by 5. The second order passes.
Answer
  • f(g(x)) = x
  • g(f(x)) = x
  • Yes, they are inverses.
Double-check With the number 10: f(10) = 50 and g(50) = 10. The other order: g(10) = 2 and f(2) = 10. Both trips return 10.

Rung 3 of 5

A pair that looks right and is not

Are f(x) = 2x + 6 and g(x) = x2 − 6 inverses? The question asks whether both round trips return x. One failure is enough to say no.

Show the steps and the answer
  1. f(g(x))=2(x2−6)+6.f doubles its whole input and adds 6. The input is all of g(x), so it goes in parentheses.
  2. Distribute the 2: 2·x2=x and 2·(−6)=−12. So f(g(x))=x−12+6.The 2 multiplies both terms inside the parentheses.
  3. x−12+6=x−6.−12+6=−6.
  4. x−6 is not x, so stop here. They are not inverses.The result must be exactly x. A leftover −6 means the round trip lands 6 short every time.
Answer
  • f(g(x)) = x − 6, which is not x.
  • No, they are not inverses.
Double-check With the number 10: g(10) = 102 − 6 = −1, then f(−1) = 2 · (−1) + 6 = 4. You started at 10 and returned to 4, which is 10 − 6, as the algebra said. Changing the 6 in g to a 3 repairs the pair: with g(x) = x2 − 3, f(g(x)) = 2(x2 − 3) + 6 = x − 6 + 6 = x, and g(f(x)) = 2x+62 − 3 = (x + 3) − 3 = x.

Rung 4 of 5

Fractions in the rules

Are f(x) = 23x − 4 and g(x) = 32(x + 4) inverses? The question asks whether both round trips return x.

Show the steps and the answer
  1. f(g(x))=23·32(x+4)−4.f multiplies its input by 23 and subtracts 4. The input is all of g(x).
  2. 23·32=66=1, so f(g(x))=1·(x+4)−4=x+4−4.A fraction times its flip is 1, and multiplying by 1 changes nothing.
  3. x+4−4=x.+4 and −4 add to 0. The first order passes.
  4. g(f(x))=32((23x−4)+4).g adds 4 to its input and then multiplies by 32. The input is all of f(x).
  5. Inside the outer parentheses, −4+4=0, which leaves 32·23x.Simplify inside parentheses first.
  6. 32·23x=1·x=x.Again a fraction times its flip is 1. The second order passes.
Answer
  • f(g(x)) = x
  • g(f(x)) = x
  • Yes, they are inverses.
Double-check With the number 6: f(6) = 23 · 6 − 4 = 4 − 4 = 0, then g(0) = 32 · (0 + 4) = 6. The other order: g(6) = 32 · 10 = 15, then f(15) = 23 · 15 − 4 = 10 − 4 = 6. Both trips return 6.

Rung 5 of 5

One order passes, the other fails

Are f(x) = x2 and g(x) = x inverses, when f accepts every number? The question asks whether both round trips return x for every input each function accepts.

Show the steps and the answer
−3f: square itg: take thesquare rootfirstsecond93
The trip starts at −3 and ends at 3, so this order does not return the input.
  1. f(g(x))=(x)2.f squares its input, and the input is g(x) = x.
  2. (x)2=x for every x that g accepts, which is 0 and the positive numbers.x is the number whose square is x, so squaring it gives x back. The first order passes.
  3. g(f(x))=x2.g takes the square root of its input, and the input is f(x) = x2.
  4. Test a positive input, x = 5: 52=25=5.A quick number shows what x2 does. Here it returned the 5.
  5. Test a negative input, x = −3: (−3)2=9=3.f accepts negative numbers, so they must be tested too. (−3)2 = 9, and 9 is 3, never −3.
  6. 3 is not −3, so g(f(x)) is not x for every input. The second order fails.One input that does not come back is enough to fail the check.
Answer
  • f(g(x)) = x, so the first order passes.
  • g(f(−3)) = 3, not −3, so the second order fails.
  • No, they are not inverses as given.
Double-check A second way to see the failure: f(3) = 9 and f(−3) = 9, so two different inputs share the output 9. g(9) is one number, 3. No function g could return both 3 and −3 from the single input 9, so no g can undo this f on every number.

Counterexamples

What it is not, and the tempting wrong moves.

✗ Not this: Checking only f(g(x)) and saying yes. For f(x) = x2 and g(x) = x: f(g(x)) = (x)2 = x, so they are inverses.
The other order fails. Take x = −3: f(−3) = 9 and g(9) = 9 = 3. You started at −3 and came back to 3. One order passing does not force the other to pass.
✓ Instead: Compute both f(g(x)) and g(f(x)). Here the second is not always x, so this f and g are not inverses.
✗ Not this: Trying one number and calling it proven. For f(x) = 2x − 1 and g(x) = 3x − 2, take x = 1: g(1) = 1 and f(1) = 1, so f(g(1)) = 1 and g(f(1)) = 1. They must be inverses.
The number 1 happens to work. Take x = 2: g(2) = 4 and f(4) = 7, not 2. With the letter x, f(g(x)) = 2(3x − 2) − 1 = 6x − 5, which is not x.
✓ Instead: A number that fails proves they are not inverses. A number that works proves nothing by itself. When f and g accept endlessly many inputs, the algebra with x is what covers all of them at once.
✗ Not this: Multiplying the two formulas in place of composing them. For f(x) = 3x and g(x) = x3: f(x) · g(x) = 3x · x3 = x2, which is not x, so they are not inverses.
The test is not about multiplying f by g. It is about feeding the output of one into the other.
✓ Instead: Put g(x) inside f: f(g(x)) = 3 · x3 = x. Then put f(x) inside g: g(f(x)) = 3x3 = x. They are inverses.
✗ Not this: Dropping the parentheses. For f(x) = 2x + 6 and g(x) = x2 − 3, writing f(g(x)) = 2 · x2 − 3 + 6 = x + 3 and deciding they are not inverses.
The 2 in f multiplies its whole input. Without parentheses it only reached the x2 and missed the −3.
✓ Instead: f(g(x)) = 2(x2 − 3) + 6 = x − 6 + 6 = x. The other order is g(f(x)) = 2x+62 − 3 = (x + 3) − 3 = x. They are inverses.

Tips and tricks

  • Memory device: out and back, both ways. Out through one function, back through the other, and you must land on x. Then again with the order reversed.
  • The answer to each check must be exactly x. If you get x plus or minus a leftover number, such as x − 6, the two functions are not inverses.
  • Always wrap the inserted formula in parentheses before you simplify. Most wrong answers here come from a lost pair of parentheses, not from the idea.
  • A number can show that two functions are not inverses. One number that works cannot show that they are. Use numbers to catch mistakes, and the letter x to prove.
  • You may stop at the first order that fails. One failure settles the question.
  • The order of the names does not matter for the verdict: if g is the inverse function of f, then f is the inverse function of g.

7 The trap

Stopping after one order. Getting f(g(x)) = x feels like the finish, but it is half of the test. With f(x) = x2 and g(x) = x, the first order gives (x)2 = x and passes, while the second order gives g(f(−3)) = 9 = 3, which is not −3. Always write both lines, f(g(x)) and g(f(x)), before you answer yes.

Flash cards

Say the answer out loud, then flip.

What two things must both be true for f and g to be inverses of each other?
  • f(g(x)) = x for every input that g accepts.
  • g(f(x)) = x for every input that f accepts.
  • Like two bus routes: both round trips must end where they began.
You worked out f(g(x)) and it came to x. Can you say yes, they are inverses?
Not yet. That is half of the test. You rode the round trip from one end. Now work out g(f(x)) as well. One order can pass while the other fails.
You simplify f(g(x)) and get x + 2. What do you conclude?
f and g are not inverses. The result must be exactly x. Not x + 2, not 3x, not −x. You can stop at the first failure.
With f(x) = (x − 1)2 and g(x) = x + 1, the order f(g(x)) hands back x for every x that g accepts. Why are they still not inverses?
  • The other order fails. Start at −2:
  • f(−2)=(−3)2=9
  • g(9) = 9 + 1 = 4, which is not −2.
  • Squaring sends both −2 and 4 to 9, and g can send 9 back to only one of them. The trip that starts at −2 never gets home.

Lesson 3 of 7

One-to-one functions and the horizontal line test

Subject: MathematicsLeans on: What a function isLeans on: What f^{−1} meansLeans on: Horizontal and verticalLeans on: Squaring a negative numberLeans on: Order of operationsLeans on: Solving a linear equationOpen this lesson
Leans on 0Comes back in 0

1 In plain words

Think of a coat check. You hand over a coat, the input, and get a ticket number, the output. Later you hand back the ticket and the clerk walks it to your coat. That return trip is the inverse. It works only if no two coats got the same number. If two coats both carry ticket 7, the clerk holding a 7 cannot know which coat to bring. A function in which every output belongs to only one input is called a one-to-one function. On a graph you can see this with a ruler held flat. Slide it up the page. If it never touches the graph in two places at once, the function is one-to-one. That is the horizontal line test.

2 Picture it

(−2, 4)(2, 4)−22246(−2, 4)(2, 4)−3−2−1123246
The flat line at height 4 meets y = x2 twice.

Reminder

Reminder: What a function is, What f−1 means, Horizontal and vertical, Squaring a negative number, Order of operations, Solving a linear equation
  • What a function is. Each input gets exactly one output. For f(x) = x2, the input 3 gives 9 and nothing else.
  • What f−1 means. The inverse sends each output back to its input: f(a) = b gives f−1(b) = a. If f(2) = 9, then f−1(9) = 2.
  • Horizontal and vertical. Horizontal means flat, like the horizon. Vertical means straight up and down. Every point on the horizontal line y = 4 has height 4.
  • Squaring a negative number. A negative times a negative is positive: (−2)2 = (−2) × (−2) = 4, the same as 22.
  • Order of operations. Work inside the parentheses first, then the power, then the adding: (0 − 1)2 + 2 = (−1)2 + 2 = 1 + 2 = 3.
  • Solving a linear equation. Undo each step in reverse order, doing the same thing to both sides: 2a − 1 = 2b − 1, add 1 to get 2a = 2b, then divide by 2 to get a = b.

3 Why

An inverse function must itself be a function, so it must send each of its inputs to exactly one place. Suppose f(a) = b and f(c) = b with a and c different. Then f−1(b) would have to equal a and also c. That is two outputs for one input, which no function allows. So a shared output blocks the inverse. The picture test says the same thing. Every point on a horizontal line has the same height, which means the same output. If the line meets the graph at two points, those are two different inputs with one output.

4 The rule

A function f is a one-to-one function when different inputs always give different outputs: f(a) = f(b) happens only when a = b.

A function has an inverse function exactly when it is one-to-one. Horizontal line test: a function is one-to-one exactly when no horizontal line meets its graph at more than one point.

The same idea, five ways

Say it

f is one to one. Every output comes from only one input.

Write it

A function is one-to-one when no two different inputs give the same output.

In math
  • f(a) = f(b) only when a = b
  • a ≠ b gives f(a) ≠ f(b)
  • every horizontal line y = c meets the graph at most once
Like

A coat check where every coat gets its own ticket number, so every ticket leads back to exactly one coat.

See it
(1, 1)(−1, −1)−22−4−224
Each flat line meets the graph of y = x3 at one point only, which is what one-to-one looks like.

The same idea, other ways

As a picture

Look at the graph of y = x2. Lay a flat line across it at height 1 and it touches the graph twice, at x = −1 and x = 1. Lay it at height 4 and it touches twice again, at x = −2 and x = 2. Two touches at one height mean two inputs with the same output. One height with two touches is enough to say the function is not one-to-one.

(−1, 1)(1, 1)(−2, 4)(2, 4)−22246
Each flat line meets y = x2 at two points, so two different inputs share each of those outputs.
As a story with arrows

A list sends each person to the month of their birthday. Going forward is fine: each person has one birthday month, so this is a function. Going backward breaks. If someone tells you only "May", you cannot say whether they mean Ana or Cal. Two arrows landing on the same output is what stops you from running the arrows in reverse.

AnaBenCalDeeMayJulyMarchfunction
Every name has one arrow, so this is a function, but two arrows land on May, so it is not one-to-one.
With small numbers

Doubling: 3 goes to 6, and no other number doubles to 6. Ask "which number doubles to 6?" and there is one answer, 3. Doubling can be undone. Squaring: 3 goes to 9, and −3 also goes to 9, because (−3) × (−3) = 9. Ask "which number squares to 9?" and there are two answers. Squaring cannot be undone until you know more.

input xoutput x²−39−24−1100112439↑ solve: output given, read every input above it
The output 9 sits under both −3 and 3, so the question "which input gave 9?" has two answers.
Why it must be true

Undoing a function means answering one question for every output b: which input gave b? An inverse function needs that question to have exactly one answer every time. A one-to-one function is one where it always does. The horizontal line at height b is the picture of that question: it collects every point whose output is b. One meeting point is one answer. Two meeting points are two answers, and the undoing breaks.

A rule of thumb

Vertical asks: is it a function? Horizontal asks: has it an inverse function? Remember it as H for horizontal, H for has an inverse. A graph must pass the vertical test to be a function at all, and then pass the horizontal test for the function to be undone.

TestThe line you slideThe question it answersIt fails when
Vertical line testAn up-and-down line, moved from left to rightIs this graph a function? Does each input have only one output?Some vertical line meets the graph twice: one input, two outputs
Horizontal line testA flat line, moved from bottom to topIs this function one-to-one? Does each output come from only one input?Some horizontal line meets the graph twice: two inputs, one output

5 How

  1. Say the question in plain words: do any two different inputs share an output?
  2. From a table or a list of pairs: read along the outputs. A value that sits under two different inputs means the function is not one-to-one. No repeated output means it is one-to-one.
  3. From a graph: hold a horizontal line at the bottom of the picture and slide it to the top. If at any height it meets the graph at two or more points, the function is not one-to-one. If it meets the graph at most once at every height, the function is one-to-one.
  4. From a formula: look at its graph and use the horizontal line test, or write f(a) = f(b) and see whether that forces a = b.
  5. To say a function is not one-to-one, write down the proof: two different inputs and the output they share, such as f(−2) = 4 and f(2) = 4.
  6. State the conclusion about undoing: a one-to-one function has an inverse function, and a function that is not one-to-one does not.

Deciding whether a function is one-to-one

1
Is the function given as a table or a list of pairs?
YesRead along the outputs. A value under two different inputs means not one-to-one. No repeats means one-to-one.
NoGo on.
↓
2
Is it given as a graph, or can you picture its graph?
YesSlide a horizontal line from the bottom to the top. Two or more meeting points at any height means not one-to-one. At most one at every height means one-to-one.
NoGo on.
↓
3
Is it a formula whose graph you cannot picture?
YesWrite f(a) = f(b) and solve. If the only way out is a = b, it is one-to-one. If two different inputs can tie, it is not.
NoGo on.
↓
4
Did you find two different inputs with the same output?
YesNot one-to-one. Write the pair as proof. There is no inverse function.
NoIf the check covered every input (every output in the table, every height on the graph, or the algebra of f(a) = f(b)) and found no tie: one-to-one. An inverse function exists.
  1. Say what is being asked: do any two different inputs give the same output?
  2. Pick the tool that fits how the function is given: the output row of a table, a sliding horizontal line on a graph, or the equation f(a) = f(b) for a formula.
  3. Hunt for one tied pair: two different inputs with the same output. One pair settles it as not one-to-one.
  4. If you find no tied pair, give the reason none can exist: no repeat in the table, at most one meeting point at every height, or f(a) = f(b) forcing a = b.
  5. Finish with the conclusion about undoing: one-to-one means an inverse function exists, not one-to-one means it does not.

6 Worked example

Worked example

Can f(x) = x2 − 3 be undone?

Decide whether f(x) = x2 − 3 is a one-to-one function, and say whether it has an inverse function. In plain words: do any two different inputs give the same output?

Show the steps and the answer
(−2, 1)(2, 1)−4−224−4−2246
The flat line at height 1 meets the graph twice, above x = −2 and above x = 2.
input xoutput f(x)−21−1−20−31−221↑ solve: output given, read every input above it
The output 1 sits under two different inputs, −2 and 2.
  1. Look at the graph of f: a U shape with its lowest point at (0,−3).x2 is never negative and equals 0 only when x = 0, so the smallest output is 0 − 3 = −3. On either side of 0 the outputs climb.
  2. Lay a horizontal line across the graph at height 1. This is the line y=1.Every point on that line has output 1, so wherever it meets the graph, the function's output is 1. Height 1 is above the lowest point, so the line cuts through both arms of the U.
  3. Read the meeting points: (−2,1) and (2, 1).The line crosses the left arm above x = −2 and the right arm above x = 2.
  4. Confirm with the formula: f(−2)=(−2)2−3=4−3=1, and f(2)=22−3=4−3=1.Reading a graph by eye can be slightly off, and the formula settles it. (−2)2 = (−2) × (−2) = 4, because a negative times a negative is positive.
  5. Conclude that f is not a one-to-one function.Two different inputs, −2 and 2, share the output 1. One tied pair is enough.
  6. Conclude that f has no inverse function.f−1(1) would have to be −2 and also 2, and a function gives only one output for each input.
Answer
  • f(x) = x2 − 3 is not a one-to-one function, because f(−2) = 1 and f(2) = 1.
  • It fails the horizontal line test, so it has no inverse function.
Double-check A second way, without the graph: build a table for x = −2, −1, 0, 1, 2. The outputs are 1, −2, −3, −2, 1. In the output row, 1 appears twice (under −2 and under 2) and −2 appears twice (under −1 and under 1). Repeated outputs under different inputs give the same verdict: not one-to-one.

Ladder: from easy to exam-hard

Try each rung before you open its steps.

Rung 1 of 5

A table with no repeats

The table shows a function f. Is f a one-to-one function? In plain words: does any output appear under two different inputs?

input xoutput f(x)15293247
x: 1, 2, 3, 4; f(x): 5, 9, 2, 7.
Show the steps and the answer
input xoutput f(x)15293247
The four outputs are all different, so each one points back to a single input.
  1. Read the output row of the table: 5, 9, 2, 7.One-to-one is a question about outputs, so the output row is the place to look.
  2. Look for a value that appears twice. There is none.5, 9, 2 and 7 are four different numbers.
  3. Conclude that f is one-to-one.No two inputs share an output.
Answer
Yes, f is a one-to-one function, so it has an inverse function.
Double-check Try to undo it by reading the table from the bottom row up: 5 came from 1, 9 came from 2, 2 came from 3, and 7 came from 4. Each output leads back to exactly one input. For example, f−1(9) = 2, with no second candidate.

Rung 2 of 5

A table with a repeat

The table shows a function h. Is h a one-to-one function? In plain words: does any output appear under two different inputs?

input toutput h(t)0419253941
t: 0, 1, 2, 3, 4; h(t): 4, 9, 5, 9, 1.
Show the steps and the answer
input toutput h(t)0419253941↑ solve: output given, read every input above it
The output 9 appears twice, under the inputs 1 and 3.
  1. Read the output row of the table: 4, 9, 5, 9, 1.A shared output can only show up as a repeat in this row.
  2. Spot the two 9s. One sits under the input 1 and the other under the input 3.The table says h(1) = 9 and h(3) = 9.
  3. Conclude that h is not one-to-one.Two different inputs, 1 and 3, share the output 9.
  4. Notice that h is still a function.Each input in the top row has one output below it. Repeated outputs are allowed in a function. They only stop it from being one-to-one.
Answer
  • No. h(1) = 9 and h(3) = 9, so h is not a one-to-one function.
  • It has no inverse function.
Double-check Try to undo it: ask which input gave the output 9. The table answers 1 and also 3. An inverse function would need h−1(9) to be a single number, so the undoing breaks exactly at the repeat.

Rung 3 of 5

A slanted straight line

Is f(x) = 2x − 1 a one-to-one function? In plain words: can two different inputs ever give the same output?

Show the steps and the answer
(2, 3)−224−4−2246
The flat line at height 3 meets the slanted line at one point only, (2, 3).
  1. Look at the graph: a straight line that climbs steadily from left to right.Each step of 1 to the right adds 2 to the output: f(0) = −1, f(1) = 1, f(2) = 3.
  2. Lay the horizontal line y=3 across it. It meets the graph once, at (2, 3).A climbing line passes through each height one time and never comes back to it.
  3. Slide the horizontal line to any other height. There is still one meeting point.The line keeps climbing to the right and keeps falling to the left, so it reaches every height exactly once.
  4. Conclude that f passes the horizontal line test.No horizontal line meets the graph at more than one point.
Answer
Yes. f(x) = 2x − 1 is a one-to-one function, so it has an inverse function.
Double-check With algebra: suppose two inputs a and b give the same output, so 2a − 1 = 2b − 1. Add 1 to both sides: 2a = 2b. Divide both sides by 2: a = b. The two inputs were the same number all along, so different inputs cannot share an output.

Rung 4 of 5

A curve that still passes

Is f(x) = x3 a one-to-one function? In plain words: do any two different inputs give the same cube? The gotcha: a graph does not have to be straight to pass.

(2, 8)(−2, −8)−22−8−448
The flat line at height 8 meets the curve only at (2, 8), because the point for x = −2 sits down at height −8.
Show the steps and the answer
  1. Test the tempting pair, 2 and −2: f(2)=2×2×2=8 and f(−2)=(−2)×(−2)×(−2)=−8.With squaring, opposite inputs tied. Cubing uses three factors: (−2) × (−2) = 4, then 4 × (−2) = −8, so the negative sign survives. The outputs 8 and −8 are different.
  2. Lay the horizontal line y=8 across the graph. It meets the graph only at (2, 8).The point for x = −2 is far below, at height −8, not at height 8.
  3. Slide the line to other heights. There is one meeting point each time.The graph climbs the whole way from left to right. It flattens near x = 0 but never turns back down, so it never returns to a height it has already passed.
  4. Conclude that f passes the horizontal line test.No horizontal line meets the graph at more than one point.
Answer
Yes. f(x) = x3 is a one-to-one function, so it has an inverse function.
Double-check A table agrees: for x = −2, −1, 0, 1, 2 the outputs are −8, −1, 0, 1, 8. Five different inputs give five different outputs, with no repeat in the output row.

Rung 5 of 5

Opposite inputs are not the only pair to try

Is f(x) = (x − 1)2 + 2 a one-to-one function? In plain words: do any two different inputs give the same output? The gotcha: the tied pair is not 2 and −2 this time.

Show the steps and the answer
(0, 3)(2, 3)−22424681012
The flat line at height 3 meets the U at (0, 3) and (2, 3), one unit on either side of its center at x = 1.
  1. Try 2 and −2: f(2)=(2−1)2+2=12+2=3, and f(−2)=(−2−1)2+2=(−3)2+2=9+2=11.This is the pair that tied for x2. Here the outputs are 3 and 11, which differ. That does not show the function is one-to-one: one pair that differs proves nothing.
  2. Look at the graph: a U shape with its lowest point at (1, 2).(x − 1)2 is 0 when x = 1 and positive for every other x, so the smallest output is 0 + 2 = 2, at x = 1. The U is centered on x = 1, not on x = 0.
  3. Lay the horizontal line y=3 across it. It meets the graph at (0, 3) and (2, 3).Any height above the lowest point cuts both arms of a U. The meeting points sit the same distance on either side of x = 1: one unit to the left and one unit to the right.
  4. Confirm with the formula: f(0)=(0−1)2+2=(−1)2+2=1+2=3, and f(2)=3 from the first step.The formula settles what the eye read from the graph.
  5. Conclude that f is not one-to-one and has no inverse function.Two different inputs, 0 and 2, share the output 3, so f−1(3) would need to be two numbers.
Answer
  • No. f(0) = 3 and f(2) = 3, so f(x) = (x − 1)2 + 2 is not a one-to-one function.
  • It has no inverse function.
Double-check Another height shows the same thing. The inputs −1 and 3 are each two units from 1: f(−1) = (−1 − 1)2 + 2 = (−2)2 + 2 = 6, and f(3) = (3 − 1)2 + 2 = 22 + 2 = 6. That is a second tied pair, so the horizontal line y = 6 also meets the graph twice.

Counterexamples

What it is not, and the tempting wrong moves.

✗ Not this: The graph of y = x2 passes the vertical line test, so it has an inverse function.
The vertical line test only shows that each input has one output. It says nothing about two inputs sharing an output, and that is what blocks an inverse.
✓ Instead: Use the horizontal line test. The line y = 4 meets the graph at x = −2 and x = 2, so y = x2 is a function but not a one-to-one function, and it has no inverse function.
✗ Not this: For f(x) = x2, f(1) = 1 and f(2) = 4 are different, so f is one-to-one.
One-to-one is a claim about every pair of different inputs. One pair that differs proves nothing. One pair that ties is enough to fail.
✓ Instead: f(−2) = 4 and f(2) = 4 is a tied pair, so f is not one-to-one. To pass, a function needs a reason that covers every pair, such as at most one meeting point at every height.
✗ Not this: The table has the output 9 twice, so it is not a function.
A function may repeat outputs. What a function may not do is give one input two different outputs.
✓ Instead: A table with a repeated output under two different inputs is still a function. It is a function that is not one-to-one.
✗ Not this: The graph of y = x3 bends and flattens near 0, so it fails the horizontal line test.
The test is not about being straight. It is about whether the graph ever returns to a height it has already passed. The graph of y = x3 flattens but never turns back.
✓ Instead: Every horizontal line meets y = x3 exactly once, so it is a one-to-one function.

Tips and tricks

  • Memory device: vertical asks "is it a function?", horizontal asks "has it an inverse function?". H for horizontal, H for has an inverse.
  • Use the edge of a sheet of paper as the horizontal line. Hold it flat and slide it up the graph, watching for a height where it touches twice.
  • A horizontal line that misses the graph, or touches it once, is fine. Only two or more meeting points on one line is a fail.
  • To show a function is not one-to-one you need only one tied pair. Write it down as proof, for example f(−2) = 4 and f(2) = 4.
  • In a table, only the output row decides one-to-one. Scan it for a repeated value.
  • A graph in one unbroken piece that only climbs as you read left to right, or only falls, always passes. One that turns around, down then up or up then down, always fails.
  • Opposite inputs such as 2 and −2 are a good first pair to try, but they are not the only pair. For a U shape, try two inputs the same distance on either side of its lowest point.
  • Failing the test is not the end of the story. A later lesson keeps one piece of such a graph so that an inverse function exists.

7 The trap

Using the wrong line. A vertical line tests whether a graph is a function at all. A horizontal line tests whether that function is one-to-one. The graph of y = x2 passes the first test and fails the second. Say the question out loud before you draw the line: "does one input have two outputs?" calls for a vertical line, and "do two inputs share one output?" calls for a horizontal line.

Flash cards

Say the answer out loud, then flip.

What makes a function a one-to-one function?
Different inputs always give different outputs. No output is shared by two inputs, like a coat check where no two coats get the same ticket number.
A horizontal line meets a graph at two points. What does that tell you?
Both points sit at the same height, so two different inputs share one output. The function is not one-to-one, and it has no inverse function.
Vertical line or horizontal line: which one answers which question?
  • Vertical line: does one input have two outputs? It tests whether the graph is a function at all.
  • Horizontal line: do two inputs share one output? It tests whether the function is one-to-one.
How do you prove that f(x) = x2 is not a one-to-one function?
Write down two different inputs and the output they share: f(−2) = 4 and f(2) = 4. One shared output is enough, like two coats both carrying ticket 7.

Lesson 4 of 7

Domain and range trade places

Subject: MathematicsLeans on: What f^{−1} meansLeans on: Domain and rangeLeans on: One-to-one functionLeans on: Interval notationLeans on: Set bracesLeans on: What a square root allowsLeans on: No zero in a bottomLeans on: The union signOpen this lesson
Leans on 2Comes back in 4

1 In plain words

Picture a coat check. You hand over a coat and get a numbered ticket. The coats the counter will take are its domain: everything that is allowed to go in. The tickets it hands out are its range: everything that actually comes out. At the end of the night the counter runs backward. You hand over a ticket and get your coat back. That backward counter is the inverse function. What goes into it? Tickets, and only tickets that were handed out. What comes out? Coats, and only coats that were checked. So the two lists trade places. The inputs of the inverse are the outputs of the original, and the outputs of the inverse are the inputs of the original.

2 Picture it

123579function
The left oval is the domain of f and the right oval is its range.

Reminder

Reminder: What f−1 means, Domain and range, One-to-one function, Interval notation, Set braces, What a square root allows, No zero in a bottom, The union sign
  • What f−1 means. f(a) = b says the same thing as f−1(b) = a. The −1 is a name, not a power. Example: f(6) = 3 gives f−1(3) = 6.
  • Domain and range. The domain is every input a function allows. The range is every output that actually comes out. For the squaring function, the domain is every number and the range is 0 and up, because a square is never negative.
  • One-to-one function. No two inputs share an output. Only a one-to-one function has an inverse function. On a graph, no horizontal line meets it twice.
  • Interval notation. [2, ∞) means every number from 2 up, with 2 included. A square bracket includes the end, a round bracket leaves it out, and ∞ always gets a round bracket. (−3, 7] leaves out −3 and includes 7.
  • Set braces. {5, 7, 9} means exactly the numbers 5, 7 and 9, and nothing between them.
  • What a square root allows. The inside of a square root must be 0 or more, and the root itself is never negative. x−2 needs x − 2 ≥ 0, so x ≥ 2, and its smallest value is 0 = 0.
  • No zero in a bottom. A fraction with 0 on the bottom has no value. In 1x−3, the input x = 3 is not allowed because it gives 10.
  • The union sign. The sign ∪ joins two pieces into one list. (−∞, 3) ∪ (3, ∞) means every number except 3.

3 Why

Start from what an inverse does: f(a) = b says the same thing as f−1(b) = a. A number b can go into f−1 only if there is some input a that f sent to b. That is exactly what it means for b to be an output of f. So the allowed inputs of f−1 are the outputs of f. In the same way, each answer f−1 gives is some a that f accepted, and every such a is given back from its own b. If you fed f−1 a number f never produced, there would be no input to send it back to.

4 The rule

If f is a one-to-one function, then: domain of f−1= range of f, and range of f−1= domain of f.

The same idea, five ways

Say it

The domain of f inverse is the range of f. The range of f inverse is the domain of f.

Write it

Every number that f−1 accepts is a number that f produces, and every number that f−1 produces is a number that f accepts.

In math
  • domain of f−1 = range of f
  • range of f−1 = domain of f
  • f(a) = b means f−1(b) = a: a is in the domain of f, and b is in the domain of f−1
  • a one-to-one f with domain [2, ∞) and range [1, ∞) gives f−1 with domain [1, ∞) and range [2, ∞)
Like

A coat check takes coats and gives tickets. Run backward, it takes tickets and gives coats.

See it

f takes in its domain and gives out its range

f−1 takes in the range of f

f−1 gives out the domain of f

domain of f−1 = range of f

range of f−1 = domain of f

The same idea, other ways

As two machines in a row

Feed 2 into f, where f doubles a number and adds 3. Out comes 7. Feed that 7 straight into f−1, which subtracts 3 and halves. Out comes 2 again. Look at the 7 in the middle. It is an output of f and an input of f−1 at the same moment. Every number f−1 ever receives is a middle number like this one, so the inputs of f−1 are exactly the outputs of f.

2f: double,then add 3f⁻¹: subtract3, then halvefirstsecond72
The 7 between the machines is an output of f and an input of f−1 at once.
With small numbers

Look at the table. The input row is the domain of f and the output row is its range. The question f−1(7) asks: which input gave the output 7? Find the 7 in the output row and look up to its input, 2. So f−1(7) = 2. You can ask f−1 about any number in the output row, and it answers with a number from the input row. Now ask about 6. There is no 6 in the output row, so f−1(6) has no answer. The output row of f is the full list of what f−1 accepts.

input xoutput f(x)152739↑ solve: output given, read every input above it
The output row of f is the list of numbers f−1 accepts, and the input row is the list it gives back.
As a story

A thermometer function f turns Celsius into Fahrenheit by f(x) = 95x + 32, and you use it only for liquid water. It accepts readings from 0 to 100. It outputs readings from 32 to 212, since f(0) = 32 and f(100) = 180 + 32 = 212. The inverse turns Fahrenheit back into Celsius. It accepts 32 to 212 and returns 0 to 100. Ask it about 68 and it answers 20, because f(20) = 36 + 32 = 68. Ask it about 20 and it has no answer: 20°F is below freezing, so f never produced that number.

As shadows on a graph

Picture the graph of f(x) = x + 2. Press the graph flat onto the x-axis and it covers 0 and everything to the right: the domain of f, [0, ∞). Press it flat onto the y-axis and it covers 2 and everything above: the range of f, [2, ∞). The shadow on the y-axis is the list of numbers f−1 accepts. The shadow on the x-axis is the list of numbers f−1 gives back.

domainrange2468246
The shadow on the y-axis, from 2 up, is what f−1 takes in, and the shadow on the x-axis, from 0 right.
As a rule of thumb

Turn every question about f−1 into a question about f. 'Can f−1 take 7?' becomes 'Does f ever output 7?' 'Can f−1 give back 4?' becomes 'Does f accept 4?' You already know how to answer questions about f, so you never need the inverse itself to find its domain and range.

FunctionGoes in (domain)Comes out (range)
Coat checkcoatstickets
Coat check run backwardticketscoats
f(x)=x−2+1[2, ∞)[1, ∞)
f−1[1, ∞)[2, ∞)

5 How

  1. Make sure f is a one-to-one function, so that f−1 exists: no output repeated in a table, or a graph that passes the horizontal line test.
  2. Find the domain of f: every input f allows.
  3. Find the range of f: every output f actually produces. List only outputs that really come out, not every number you can imagine.
  4. Trade the two lists. Domain of f−1 = range of f. Range of f−1 = domain of f.
  5. Check with one pair. Pick an allowed input a and compute b = f(a). Then f−1(b) = a, so b must sit in your new domain and a must sit in your new range.

Finding the domain and range of f−1 without finding f−1

1
Is f a one-to-one function (no output used by two inputs)?
Yesf−1 exists. Go on.
NoStop. There is no inverse function, so there is no domain or range to report.
↓
2
Is f given as a table or a list of pairs?
YesThe input row is the domain of f and the output row is its range. Trade the two lists.
NoGo on.
↓
3
Is f given as a graph?
YesRead the shadow on the x-axis (the domain of f) and the shadow on the y-axis (the range of f). Trade them.
NoGo on.
↓
4
Is f given as a formula?
YesFind the inputs the formula allows (no zero in a bottom, no negative under a square root), then find which outputs come out. Trade them.
NoUse whatever description you have to list the inputs and the outputs of f, then trade them.
↓
5
Does a test pair agree? Pick an allowed a and compute b = f(a).
Yesb is in the domain of f−1 and a is in its range. You are done.
NoEither a list was left in place instead of traded, or the range of f is wrong. Redo that step.
  1. Confirm that f is a one-to-one function, so an inverse function exists.
  2. List what goes into f: its domain.
  3. List what comes out of f: its range.
  4. Write the range of f as the domain of f−1, and the domain of f as the range of f−1. Keep each bracket with its own list.
  5. Test one pair: compute b = f(a) for an allowed a, then confirm b is in the new domain and a is in the new range.

6 Worked example

Worked example

The domain and range of an inverse, with no formula for the inverse

The function f(x) = x−2 + 1 is a one-to-one function, so it has an inverse function. Without finding a formula for f−1, find the domain and the range of f−1. In plain words: which numbers may go into f−1, and which numbers can come out of it?

Show the steps and the answer
(2, 1)(6, 3)domainrange2468101212345
The shadow on the x-axis starts at 2, the domain of f, and the shadow on the y-axis starts at 1, the range of…
  1. Find the domain of f. The inside of the root must be 0 or more: x−2≥0, so x≥2. Domain of f: [2,∞).A negative number has no real square root. Solving x − 2 ≥ 0 finds the smallest input the root accepts. Plug it in to see: f(2) = 0 + 1 = 1, which works, while f(1) would need −1, which is not a real number.
  2. Find the range of f. The root x−2 is 0 at x=2 and grows without end as x grows. Adding 1 lifts every output by 1, so f(x)≥1. Range of f: [1,∞).A square root never comes out negative, so its smallest value is 0 and the smallest output of f is 0 + 1 = 1. The outputs then climb through every number above 1: f(2) = 1, f(3) = 2, f(6) = 3, f(11) = 4.
  3. Trade the first list. Domain of f−1= range of f=[1,∞).f−1 can only accept numbers that f actually produces, and those are the numbers from 1 up.
  4. Trade the second list. Range of f−1= domain of f=[2,∞).f−1 hands back the inputs of f, and those are the numbers from 2 up.
Answer
  • Domain of f−1: [1, ∞)
  • Range of f−1: [2, ∞)
Double-check Test one pair. f(6) = 6−2 + 1 = 4 + 1 = 2 + 1 = 3, so f−1(3) = 6. The input 3 is in [1, ∞) and the output 6 is in [2, ∞). Now test a number outside. Could f−1(0) exist? That would need x−2 + 1 = 0, so x−2 = −1, and a square root is never negative. So 0 is not allowed into f−1, which agrees with [1, ∞).

Ladder: from easy to exam-hard

Try each rung before you open its steps.

Rung 1 of 5

A table

The table shows a one-to-one function h. Find the domain and the range of h−1. In plain words: list the numbers you may feed to h−1, and list the numbers it can give back.

input xoutput h(x)−24013856
x:−2,0,3,5;h(x):4,1,8,6.
Show the steps and the answer
input xoutput h(x)−24013856↑ solve: output given, read every input above it
The highlighted 8 in the output row is a number h−1 accepts, and the 3 above it is what h−1 gives back.
  1. Read the domain of h from the input row: {−2,0,3,5}.The input row lists every number h accepts.
  2. Read the range of h from the output row: {4, 1, 8, 6}, or in order {1, 4, 6, 8}.The output row lists every number h actually produces. No number appears twice there, which is why h is a one-to-one function.
  3. Trade them. Domain of h−1={1,4,6,8}. Range of h−1={−2,0,3,5}.h−1 takes each output of h back to the input it came from.
Answer
  • Domain of h−1: {1, 4, 6, 8}
  • Range of h−1: {−2, 0, 3, 5}
Double-check Find the 8 in the output row and look up to its input: h(3) = 8, so h−1(8) = 3. The 8 is in the new domain and the 3 is in the new range. Now try 2: no column has 2 in the output row, so h−1(2) has no answer, and 2 is not in the new domain.

Rung 2 of 5

A straight line on a limited stretch

A taxi charges f(x) = 3x + 4 dollars for a ride of x miles and only takes rides from 0 to 10 miles, so the domain of f is [0, 10]. The inverse f−1 takes a fare and gives back the miles. Find the domain and the range of f−1. In plain words: which fares can you hand to f−1, and which mile counts can it return?

Show the steps and the answer
41934[4, 34]
These are the fares the taxi can charge: the range of f, and so the domain of f−1.
  1. Write the domain of f: [0, 10].The question says the taxi accepts rides of 0 to 10 miles, both ends included.
  2. Find the lowest and highest outputs. f(0)=3·0+4=4 and f(10)=3·10+4=34.The fare climbs 3 dollars every mile, so the smallest fare belongs to the shortest ride and the largest fare to the longest ride.
  3. Write the range of f: [4, 34].A straight line has no gaps, so as x slides from 0 to 10 the fare passes through every number from 4 to 34. It never repeats a fare, so f is a one-to-one function.
  4. Trade them. Domain of f−1=[4,34]. Range of f−1=[0,10].f−1 takes fares in and gives miles out.
Answer
  • Domain of f−1: [4, 34]
  • Range of f−1: [0, 10]
Double-check A 5 mile ride costs f(5) = 3 · 5 + 4 = 19, so f−1(19) = 5. The 19 is in [4, 34] and the 5 is in [0, 10]. Try a fare outside the list: 40 dollars would need 3x + 4 = 40, so 3x = 36 and x = 12 miles, which the taxi does not drive. So 40 is not in the domain of f−1, which agrees with [4, 34].

Rung 3 of 5

A square root that is subtracted

f(x) = 5 − x+4 is a one-to-one function. Find the domain and the range of f−1 without finding its formula. In plain words: which numbers does f produce, since those are the ones f−1 accepts, and which numbers does f accept, since those are the ones f−1 returns?

Show the steps and the answer
(−4, 5)(0, 3)domainrange−44812246
The graph starts at height 5 and only falls, so its shadow on the y-axis is 5 and everything below.
  1. Find the domain of f. The inside of the root must be 0 or more: x+4≥0, so x≥−4. Domain of f: [−4,∞).A negative number has no real square root. Solving x + 4 ≥ 0 finds the smallest input the root accepts. Plug it in to see: f(−4) = 5 − 0 = 5, which works, while f(−5) would need −1.
  2. Find the range of f. The root x+4 starts at 0 and grows without end, so 5−x+4 starts at 5 and falls without end: f(−4)=5,f(0)=3,f(5)=2,f(12)=1. Range of f: (−∞,5].The root is never negative, and here it is taken away from 5. Taking away 0 or more leaves 5 or less. This is the gotcha: the range runs down from 5, not up from it.
  3. Trade them. Domain of f−1=(−∞,5]. Range of f−1=[−4,∞).The outputs of f are the inputs of f−1, and the inputs of f are its outputs. The square bracket at 5 and the square bracket at −4 each stay with their own list.
Answer
  • Domain of f−1: (−∞, 5]
  • Range of f−1: [−4, ∞)
Double-check f(0) = 5 − 4 = 5 − 2 = 3, so f−1(3) = 0. The 3 is in (−∞, 5] and the 0 is in [−4, ∞). Try a number outside: 6 would need 5 − x+4 = 6, so x+4 = −1, and a square root is never negative. So 6 is not in the domain of f−1.

Rung 4 of 5

A fraction with x in the bottom

f(x) = 1x−3 + 2 is a one-to-one function. Find the domain and the range of f−1 without finding its formula. In plain words: find the one number f cannot take in and the one number f cannot give out, then trade the two lists.

Show the steps and the answer
−22468−2246
The graph never sits over x = 3 and never reaches the height y = 2, the two numbers that get thrown out.
  1. Find the domain of f. Set the bottom equal to 0: x−3=0, so x=3. Throw that input out. Domain of f: (−∞,3)∪(3,∞), every number except 3.Setting the bottom equal to 0 finds the one input that would make the fraction divide by zero. Plug it in to see: f(3) = 10 + 2, and 10 has no value.
  2. Find the one output f can never give. The fraction 1x−3 is 1 divided by something, and that is never 0. So f(x)=1x−3+2 is never 0+2=2.A fraction equals 0 only when its top is 0, and the top here is always 1.
  3. See that every other output is reached. To get 3 you need 1x−3=1, which x=4 does: f(4)=1+2=3. To get 1 you need 1x−3=−1, which x=2 does: f(2)=−1+2=1. Any target other than 2 works the same way. Range of f: (−∞,2)∪(2,∞).For a target other than 2, the fraction must equal the target minus 2, which is not 0, and some x always makes 1x−3 equal a number that is not 0.
  4. Trade them. Domain of f−1=(−∞,2)∪(2,∞). Range of f−1=(−∞,3)∪(3,∞).The number f never outputs, 2, is the number f−1 cannot accept. The number f cannot accept, 3, is the number f−1 never gives back.
Answer
  • Domain of f−1: (−∞, 2) ∪ (2, ∞), every number except 2
  • Range of f−1: (−∞, 3) ∪ (3, ∞), every number except 3
Double-check f(5) = 15−3 + 2 = 12 + 2 = 52, so f−1(52) = 5. The input 52 is not 2 and the output 5 is not 3, so both fit. Then test the excluded number: f−1(2) would need 1x−3 + 2 = 2, so 1x−3 = 0, which never happens. So 2 is out of the domain of f−1.

Rung 5 of 5

Only the two lists, with open and closed ends

A one-to-one function g has domain (−3, 7] and range [0, 12). (a) Find the domain and the range of g−1. (b) Does g−1(12) exist? (c) Does g−1(0) exist? (d) Could g−1(5) = 9? In plain words: no table, graph or formula is given, only the list of inputs and the list of outputs of g, and you must say what g−1 can take in and give out.

0512[0, 12)
These are the numbers g−1 accepts: the filled dot keeps 0 in and the open dot leaves 12 out.
Show the steps and the answer
  1. (a) Trade the lists. Domain of g−1=[0,12). Range of g−1=(−3,7].The outputs of g are the inputs of g−1, and the inputs of g are its outputs. Each bracket stays with its own list.
  2. (b) Look for 12 in [0, 12). The round bracket leaves 12 out, so g−1(12) does not exist.g never outputs 12, so there is no input to send 12 back to.
  3. (c) Look for 0 in [0, 12). The square bracket includes 0, so g−1(0) exists. Its value is some number in (−3,7].0 is an output of g, so exactly one input produced it. The two lists alone do not say which input.
  4. (d) g−1(5) exists, because 5 is in [0, 12). Its value cannot be 9, because 9 is not in (−3,7].g−1(5) = 9 would mean g(9) = 5, and g does not accept 9.
Answer
  • (a) Domain of g−1: [0, 12). Range of g−1: (−3, 7]
  • (b) No
  • (c) Yes, and its value lies in (−3, 7]
  • (d) No
Double-check Turn each question into a question about g. Part (b) asks whether g ever outputs 12: no, the range [0, 12) stops short of 12. Part (c) asks whether g ever outputs 0: yes, 0 is in the range. Part (d) asks whether g(9) = 5 is possible: no, 9 is outside the domain (−3, 7]. All three agree with the answers.

Counterexamples

What it is not, and the tempting wrong moves.

✗ Not this: A thermometer function f turns Celsius into Fahrenheit for liquid water, so its domain is [0, 100]. Tempting: f−1 is the same function run backward, so its domain is [0, 100] too.
f−1 takes Fahrenheit readings in, and those run from f(0) = 32 to f(100) = 212. The number 20 is in [0, 100], but f never outputs 20, so f−1(20) does not exist.
✓ Instead: Domain of f−1 = range of f = [32, 212]. Range of f−1 = domain of f = [0, 100].
✗ Not this: For f(x) = x−2 + 1: the outputs are y values, and y can be anything, so the range of f is every real number and f−1 accepts every number.
The range is the list of outputs that actually come out, not every number you can name. A square root is never negative, so f(x) is never below 1. No input gives f(x) = 0, for example.
✓ Instead: The range of f is [1, ∞), so the domain of f−1 is [1, ∞).
✗ Not this: Let f double a number and add 3, but only for the inputs 1, 2 and 3. It is undone by 'subtract 3, then halve'. That rule can be done to any number, so f−1 accepts any number, for example f−1(6) = 1.5.
The undoing rule can be carried out on 6, but f never produced 6: its outputs are only 5, 7 and 9. And 1.5 was never an input of f. Being able to compute a rule does not put a number in the domain.
✓ Instead: The domain of f−1 is {5, 7, 9}, the range of f, and its range is {1, 2, 3}. f−1(6) does not exist.
✗ Not this: g(x) = x2 accepts every number and outputs 0 and up, so g−1 has domain [0, ∞) and range every real number.
The trade applies only when an inverse function exists. Here g(3) = 9 and g(−3) = 9, so two inputs share an output and g is not a one-to-one function. There is no inverse function to have a domain.
✓ Instead: Check that the function is a one-to-one function first. If it is not, there is no inverse function, and so no domain or range to report.

Tips and tricks

  • Memory cue: what f gives out, f−1 takes in. Outs become ins, and ins become outs.
  • You do not need a formula for f−1 to know its domain and range. Everything comes from f.
  • To test one number, turn the question around. 'Is 7 in the domain of f−1?' is the same question as 'Does f ever output 7?'
  • Brackets travel with their list. If the range of f is [1, ∞), the domain of f−1 is [1, ∞), with the same closed end at 1.
  • The trade works in both directions, because f is the inverse of f−1. Knowing the domain and range of either function gives you both for the other.
  • When a later lesson gives you a formula for f−1, still take its domain from the range of f. A formula can look as if it accepts more numbers than the inverse really does.

7 The trap

Writing the domain of f again as the domain of f−1. The two functions trade their lists, and the two lists are often different. For f(x) = x−2 + 1 the tempting answer is 'domain of f−1: [2, ∞)', but [2, ∞) is what f−1 gives back, not what it takes in. To avoid this, write two labeled lines before you answer, 'goes into f−1' and 'comes out of f−1', and fill the first line from the outputs of f.

Flash cards

Say the answer out loud, then flip.

f is a one-to-one function. Where do you get the domain of f−1?
From the range of f. The backward coat check takes in tickets, and only the tickets that were handed out.
You know f(4) = 9, and f is a one-to-one function. Which number goes into f−1, and which comes out?
  • 9 goes in, so 9 is in the domain of f−1.
  • 4 comes out, so 4 is in the range of f−1.
  • Reason: f(4) = 9 says the same thing as f−1(9) = 4.
f(x) = x−2 + 1 has domain [2, ∞) and range [1, ∞). What are the domain and range of f−1?
  • Domain of f−1: [1, ∞)
  • Range of f−1: [2, ∞)
  • The two lists trade places. [2, ∞) is what f−1 gives back, not what it takes in.
Why can you not put a number into f−1 that f never produced?
There is no input for f−1 to send it back to. It is like handing over a ticket the counter never gave out: no coat is waiting for it.

Lesson 5 of 7

Reading an inverse from a table and from a graph

Subject: MathematicsLeans on: Reading a pointLeans on: Negative coordinatesLeans on: What the notation f^{−1} meansLeans on: One-to-one function and the horizontal line testLeans on: Domain and range trade placesLeans on: Checking inverses by composingOpen this lesson
Leans on 1Comes back in 0

1 In plain words

Picture the sign-in sheet at a gym. It has two columns: a locker number on the left and the person using that locker on the right. Read left to right, the sheet answers 'who has locker 12?' Read right to left, the same sheet answers the opposite question: 'which locker does Dana have?' You only traded the column you start from. A table of a function works the same way. Trade its two rows and you are holding the table of its inverse function. A graph is a picture of those same pairs, so trading the two numbers in every point draws the graph of the inverse function. That trade looks like a flip across the line y = x, as if the line were a mirror.

2 Picture it

(1, 3)(3, 1)(2, 5)(5, 2)(−1, −1)−2246−2246(1, 3)(3, 1)(2, 5)(5, 2)(−1, −1)−3−2−1123456−2246
The solid line is f and the dashed line is f−1.

Reminder

Reminder: Reading a point, Negative coordinates, What the notation f−1 means, One-to-one function and the horizontal line test, Domain and range trade places, Checking inverses by composing
  • Reading a point. (a, b) means go a across, then b up. (3, 1) is 3 right and 1 up. (1, 3) is 1 right and 3 up. The order matters, which is why swapping makes a new point unless the two numbers are equal.
  • Negative coordinates. A negative first number means left of zero and a negative second number means below zero. (−1, 2) is 1 left and 2 up. (2, −1) is 2 right and 1 down.
  • What the notation f−1 means. f(a) = b gives f−1(b) = a. If f(2) = 9, then f−1(9) = 2. The −1 is part of the name of the inverse function. It is not a power.
  • One-to-one function and the horizontal line test. A one-to-one function never gives two inputs the same output. On a graph, that means no horizontal line meets the graph twice. Outputs 5, 8, 5 fail. Outputs 5, 8, 6 pass.
  • Domain and range trade places. The inputs of f−1 are the outputs of f. If the outputs of f are 4, 7 and 9, then 4, 7 and 9 are the only numbers f−1 accepts.
  • Checking inverses by composing. Feeding the answer back must return the starting number: if f−1(7) = 3, then f(3) must be 7, so f(f−1(7)) = 7.

3 Why

Earlier in this unit you saw that f(a) = b means f−1(b) = a. A table column with a on top and b underneath says f(a) = b, so the same column turned upside down says f−1(b) = a. A point (a, b) on a graph says the same thing a column does, so the inverse function owns the point (b, a). Why a mirror? Fold the page along the line y = x. The fold lays the x-axis onto the y-axis, so 'a across and b up' becomes 'b across and a up'. That is the point (b, a).

4 The rule

If f is a one-to-one function, then: (1) the table of f−1 is the table of f with its two rows traded; (2) the point (a, b) is on the graph of f exactly when the point (b, a) is on the graph of f−1; (3) the graph of f−1 is the graph of f reflected (flipped, as in a mirror) across the line y=x.

The same idea, five ways

Say it

'the point a comma b on f becomes the point b comma a on f inverse.' Or say: 'the graph of f inverse is the graph of f flipped over the line y equals x.'

Write it

Every input and output pair of f, written in the opposite order, is an input and output pair of the inverse function, so the two graphs are mirror images of each other in the line y = x.

In math
  • f(a) = b means f−1(b) = a
  • (a, b) on the graph of f gives (b, a) on the graph of f−1
  • table of f−1 = table of f with the two rows traded
  • graph of f−1 = graph of f reflected across the line y = x
Like

Draw a curve in wet ink, fold the paper along the diagonal crease, and press. The print that appears on the other side of the crease is the graph of the inverse function.

See it
(1, 3)(3, 1)(2, 5)(5, 2)(−1, −1)−2246−2246
The line drawn is the line y = x, the mirror.

The same idea, other ways

As a picture

Fold the page along the line y = x and the graph of f lands on the graph of f−1. This works for curves as well as for straight lines. In the picture, the solid curve is a function and the dashed curve is its inverse function. The point (2, 8) on the solid curve has a partner (8, 2) on the dashed curve. The point (−2, −8) has the partner (−8, −2). Every point has a partner with the same two numbers in the opposite order.

(2, 8)(8, 2)(−2, −8)(−8, −2)−8−448−8−448
A curved graph (solid) and its inverse function (dashed) are the same shape, flipped across the diagonal.
As a story

A phrase card for travelers has English on the left and Spanish on the right. Start from the left column and it turns English into Spanish. Start from the right column and the same card turns Spanish back into English. You did not write a new card. You changed which column you enter by. One warning comes with the story: if two English words on the card shared one Spanish word, the backward reading would not know which English word to give you. That is why the function must be a one-to-one function before you trade anything.

With small numbers

Look at the table. The column under 2 holds 4, so f(2) = 4. Read that same column from the bottom up: 4 goes back to 2, so f−1(4) = 2. Do the same with the other two columns: 6 goes back to 1, and 9 goes back to 3. Those three backward readings are the whole inverse function. As points, (1, 6), (2, 4) and (3, 9) on the graph of f become (6, 1), (4, 2) and (9, 3) on the graph of f−1.

input xoutput f(x)162439↑ solve: output given, read every input above it
The highlighted 4 in the output row points up to the input 2, which is how you read f−1(4) = 2 from the table…
Why it must be true

A mirror sits exactly halfway between a thing and its reflection. Test that on the pair (1, 4) and (4, 1). Halfway between them is (2.5, 2.5), and a point with two equal numbers is on the line y = x. Try (−1, 2) and (2, −1): halfway is (0.5, 0.5), on the line again. This always happens, because the halfway number a+b2 is the same whichever order you add in. Also, the trip from (1, 4) to (4, 1) goes 3 right and 3 down, which crosses the line y = x squarely, not at a slant. Halfway and square on: that is a mirror.

(1, 4)(4, 1)halfway (2.5, 2.5)(−1, 2)(2, −1)halfway (0.5, 0.5)−224−224
The halfway point of each swapped pair lands on the line y = x, which is what makes that line the mirror.
Point on the graph of fWhat it saysSwapped point, on the graph of f−1What it says
(0, 1)f(0) = 1(1, 0)f−1(1)=0
(1, 3)f(1) = 3(3, 1)f−1(3)=1
(2, 5)f(2) = 5(5, 2)f−1(5)=2
(−1,−1)f(−1)=−1(−1,−1)f−1(−1)=−1

5 How

  1. Check that f is a one-to-one function. In a table, no number may appear twice in the output row. On a graph, use the horizontal line test: no horizontal line may meet the graph twice.
  2. To find one value, such as f−1(7), say what it asks: 'f gave the output 7; which input was it?' Find 7 in the output row of f, or at height 7 on the y-axis, and read back to the input.
  3. To get the whole table of f−1, trade the two rows. The old output row becomes the new input row.
  4. If you want the new inputs in order from smallest to largest, slide whole columns. Never separate an input from its own output.
  5. To get the graph of f−1, pick several clear points on the graph of f, write each one as (a, b), and swap it to (b, a).
  6. Plot the swapped points and join them with the same kind of shape the original had. Draw the line y = x lightly: each point and its swap must sit straight across that line from each other, the same distance away.
  7. Check one value both ways: if you wrote f−1(b) = a, the table or graph of f must show f(a) = b.

Reading an inverse from a table or from a graph

1
Is f a one-to-one function (no output repeated in the table, no horizontal line meeting the graph twice)?
YesThe inverse function exists. Go on.
NoStop. Trading rows or swapping points would give one input two outputs, so there is no inverse function to read.
↓
2
Do you need only one value, such as f−1(7)?
YesFind 7 in the output row of f, or at height 7 on the y-axis, go to the graph or the column, and read the input.
NoGo on.
↓
3
Is f given as a table?
YesTrade the two rows. Slide whole columns if you want the new inputs in order.
NoGo on.
↓
4
Is f given as a graph?
YesPick clear points, turn each (a, b) into (b, a), plot them, and join them as the mirror image in the line y = x.
NoIf f is a list of pairs, swap the two numbers in every pair. It is the same move.
  1. Say what is being asked: f−1 takes an output of f and returns the input it came from.
  2. Make sure f is a one-to-one function: no repeated number in the output row, or a graph that passes the horizontal line test.
  3. For one value f−1(k), find k among the outputs of f (the output row, or height k on the y-axis) and read back to its input.
  4. For a whole table, trade the two rows, then slide whole columns into order if you like.
  5. For a whole graph, swap each clear point (a, b) to (b, a), plot the new points, and join them as the mirror image in the line y = x.
  6. Check one answer forward: f−1(b) = a is right only if f(a) = b.

6 Worked example

Worked example

A table read backward

The first table shows a function g. In plain words, g−1 reads that table backward: it starts from an output of g and returns the input it came from.
(a) Does g have an inverse function?
(b) Write the table of g−1.
(c) Find g−1(7) and g−1(0).
(d) The point (3, 7) is on the graph of g. Which point does that put on the graph of g−1?

input xoutput g(x)−24011−33760
x:−2,0,1,3,6;g(x):4,1,−3,7,0.
Show the steps and the answer
input xoutput g(x)−24011−33760↑ solve: output given, read every input above it
The table of g, with the output 7 highlighted and an arrow up to its input 3.
input xoutput g⁻¹(x)−3106104−273↓ evaluate: input given, read the output below it
The table of g−1.
  1. Look along the g(x) row of the first table and see whether any number appears twice.Only a one-to-one function can be undone, and one-to-one means no two inputs share an output. The five outputs are all different, so g is a one-to-one function and g−1 exists.
  2. Trade the two rows of the first table. The g(x) row becomes the new input row and the x row becomes the new output row. For example, the column with −2 over 4 becomes a column with 4 over −2.Each column of g says g(a) = b. The inverse function sends b back to a, so b goes on top as the input and a goes underneath as the output.
  3. Slide whole columns so the new inputs run from smallest to largest. The second table is the result. Its first column, −3 over 1, came from the old column 1 over −3.The order of the columns does not change what a table says, as long as each input keeps its own output. Inputs in order are quicker to read.
  4. Find g−1(7). Look for 7 in the g(x) row of the first table. It sits under x=3, so g−1(7)=3.g−1(7) asks which input of g produced the output 7. The first table says g(3) = 7.
  5. Find g−1(0). Look for 0 in the g(x) row of the first table, not in the x row. It sits under x=6, so g−1(0)=6.The 0 in the x row belongs to the fact g(0) = 1, which answers a different question. For the inverse function, 0 has to be an output of g, and the first table says g(6) = 0.
  6. Swap the two numbers in (3, 7) to get (7, 3).The point (3, 7) on the graph of g says g(3) = 7. That is the same fact as g−1(7) = 3, which is the point (7, 3) on the graph of g−1.
Answer
  • (a) Yes. No output of g repeats, so g is a one-to-one function.
  • (b) The second table: g−1(−3) = 1, g−1(0) = 6, g−1(1) = 0, g−1(4) = −2, g−1(7) = 3.
  • (c) g−1(7) = 3 and g−1(0) = 6.
  • (d) (7, 3)
Double-check Feed each answer back into g. Composing a function with its inverse function must return the number you started with: g(g−1(7)) = g(3) = 7, and g(g−1(0)) = g(6) = 0. Both come back correctly. A second look: the inputs in the table of g−1 are −3, 0, 1, 4 and 7, which are exactly the outputs of g, as they must be because domain and range trade places.

Ladder: from easy to exam-hard

Try each rung before you open its steps.

Rung 1 of 5

One fact, read both ways

A one-to-one function h has h(4) = 9. In plain words: the input 4 gives the output 9.
(a) What is h−1(9)?
(b) Name the point this fact puts on the graph of h, and the point it puts on the graph of h−1.

Show the steps and the answer
  1. Write the fact h(4)=9 as the point (4, 9).In a point on a graph, the first number is the input and the second number is the output.
  2. Read the fact backward: h−1(9)=4.The inverse function takes each output of h back to the input it came from, and the output 9 came from the input 4.
  3. Swap the two numbers in (4, 9) to get (9, 4).h−1(9) = 4 has input 9 and output 4, so on the graph of h−1 it is the point (9, 4).
Answer
  • (a)h−1(9)=4
  • (b) (4, 9) is on the graph of h, and (9, 4) is on the graph of h−1.
Double-check Compose: h(h−1(9)) = h(4) = 9. You are back at the 9 you started with.

Rung 2 of 5

One value from a table, with a tempting wrong column

The table shows a one-to-one function p. Find p−1(5) and p−1(1). In plain words: p gave the output 5, which input was it? Then the same question for the output 1.

input xoutput p(x)18233104551
x: 1, 2, 3, 4, 5; p(x): 8, 3, 10, 5, 1.
Show the steps and the answer
input xoutput p(x)18233104551↑ solve: output given, read every input above it
The 5 that matters is the one in the output row, and the arrow from it points up to the input 4.
  1. For p−1(5), find 5 in the p(x) row. It sits under x=4. So p−1(5)=4.The input of p−1 is an output of p, so it must be found in the output row. The column says p(4) = 5.
  2. Do not use the column that has 5 on top.That column says p(5) = 1. It answers 'what does p do to 5?', which is the forward question, not the backward one.
  3. For p−1(1), find 1 in the p(x) row. It sits under x=5. So p−1(1)=5.The column says p(5) = 1, so the output 1 came from the input 5. The column with 1 on top says p(1) = 8, which is the forward reading again.
Answer
  • p−1(5)=4
  • p−1(1)=5
Double-check Read the table forward at each answer: the column under 4 holds 5, so p(4) = 5, and the column under 5 holds 1, so p(5) = 1. Both match.

Rung 3 of 5

Flipping the points of a graph

The solid line in the first picture is the graph of a function f. It passes through (0, 3), (2, 2), (4, 1) and (6, 0). In plain words: each point is an input and output pair of f, and the inverse function owns the same pairs in the opposite order.
(a) Give four points on the graph of f−1.
(b) Where does the graph of f−1 cross each axis?
(c) Which point is on both graphs?

(0, 3)(2, 2)(4, 1)(6, 0)246246
The graph of f with four clear points to swap.
(0, 3)(3, 0)(4, 1)(1, 4)(6, 0)(0, 6)(2, 2)246246
The dashed line through the swapped points is f−1, and the two lines meet at (2, 2), a point on the line y =…
Show the steps and the answer
  1. Check that f can be undone. The graph is a slanted straight line, so every horizontal line meets it exactly once.Passing the horizontal line test means f is a one-to-one function, so the inverse function exists.
  2. Swap the two numbers in each point: (0, 3) becomes (3, 0), (2, 2) becomes (2, 2), (4, 1) becomes (1, 4), and (6, 0) becomes (0, 6).A point (a, b) on the graph of f says f(a) = b, which is the same fact as f−1(b) = a, the point (b, a).
  3. Find the axis crossings. (3, 0) has second number 0, so it is on the x-axis. (0, 6) has first number 0, so it is on the y-axis.The swap trades the two axes. f crossed the y-axis at (0, 3), so f−1 crosses the x-axis at (3, 0). f crossed the x-axis at (6, 0), so f−1 crosses the y-axis at (0, 6).
  4. Notice that (2, 2) did not change.Its two numbers are equal, so it lies on the line y = x, the mirror. A point on the mirror is its own reflection, so both graphs pass through it.
  5. Plot the four swapped points and join them. The second picture shows the result as a dashed line.The mirror image of a straight line is a straight line, so the swapped points line up.
Answer
  • (a) (3, 0), (2, 2), (1, 4) and (0, 6)
  • (b) It crosses the x-axis at (3, 0) and the y-axis at (0, 6).
  • (c) (2, 2), which lies on the line y = x.
Double-check Read one fact both ways: the solid line gives f(4) = 1 and the swapped point (1, 4) gives f−1(1) = 4. Compose them: f−1(f(4)) = f−1(1) = 4, back where you started. Also list the swapped points left to right: (0, 6), (1, 4), (2, 2), (3, 0). Each step of 1 to the right goes down 2, a steady pattern, so they do sit on one straight line.

Rung 4 of 5

Reading f−1 straight from the graph of f

A function f is given only by its graph. The curve passes through (0, 1), (1, 2), (4, 3) and (9, 4). Find f−1(3) and f−1(1) without drawing the graph of f−1. In plain words, f−1(3) asks: the output of f is 3, what was the input?

(0, 1)(1, 2)(4, 3)(9, 4)24681024
Start at height 3 on the y-axis.
Show the steps and the answer
  1. Check that f can be undone. The curve keeps rising as you move to the right, so no horizontal line can meet it twice.A graph that passes the horizontal line test belongs to a one-to-one function, so the inverse function exists.
  2. For f−1(3), start at 3 on the y-axis, not on the x-axis.The input of f−1 is an output of f, and the outputs of f are heights, measured on the y-axis.
  3. Move straight across at height 3 until you hit the curve. You arrive at the point (4, 3). Drop straight down to the x-axis and read 4. So f−1(3)=4.The point (4, 3) says f(4) = 3, so the output 3 came from the input 4.
  4. For f−1(1), start at height 1 on the y-axis. The curve is right there, at the point (0, 1). The input is 0, so f−1(1)=0.The point (0, 1) says f(0) = 1. Do not use the point (1, 2): it says f(1) = 2, which is the forward reading of the number 1.
Answer
  • f−1(3)=4
  • f−1(1)=0
Double-check Read the graph the usual way, up from the x-axis. Above x = 4 the curve is at height 3, so f(4) = 3. Above x = 0 the curve is at height 1, so f(0) = 1. Both agree with the answers. The matching points on the graph of f−1 would be (3, 4) and (1, 0).

Rung 5 of 5

You are handed the graph of f−1 and asked about f

This time the solid line in the first picture is the graph of f−1, not of f. It passes through (−4, −1), (−2, 0), (0, 1), (2, 2) and (4, 3).
(a) Find f(3).
(b) Find f(0).
(c) Where does the graph of f cross the x-axis?
In plain words: you hold the backward machine and must answer questions about the forward one.

(−4, −1)(−2, 0)(0, 1)(2, 2)(4, 3)−4−224−4−224
This line is the graph of f−1, so in each labeled point the second number is an input of f and the first…
Show the steps and the answer
(4, 3)(3, 4)(−2, 0)(0, −2)(0, 1)(1, 0)(2, 2)−4−224−4−224
Swapping the points of f−1 gives the dashed line, which is f, and the two lines meet at (2, 2).
  1. Turn the question around. f(3) asks for the output of f when the input is 3. On the graph of f−1, the inputs of f appear as the second numbers of the points.A point (b, a) on the graph of f−1 says f−1(b) = a, which is the same fact as f(a) = b. Undoing works in both directions: f undoes f−1 as much as f−1 undoes f.
  2. Find the point whose second number is 3. It is (4, 3). So f(3)=4.(4, 3) on the graph of f−1 swaps to (3, 4) on the graph of f, and (3, 4) says f(3) = 4.
  3. Find the point whose second number is 0. It is (−2,0). So f(0)=−2.(−2, 0) swaps to (0, −2), which says f(0) = −2. Do not use (0, 1): its first number is 0, and it says f−1(0) = 1, a different fact.
  4. The graph of f crosses the x-axis at a point of the form (a, 0). Its swap, (0, a), is on the graph of f−1 and on the y-axis. That point is (0, 1). Swap it back to get (1, 0).Swapping trades the axes, so the place where f−1 crosses the y-axis tells you where f crosses the x-axis.
Answer
  • (a) f(3) = 4
  • (b)f(0)=−2
  • (c) The graph of f crosses the x-axis at (1, 0).
Double-check Swap all five points to get points on the graph of f: (−1, −4), (0, −2), (1, 0), (2, 2), (3, 4). Read the answers from this list: f(3) = 4, f(0) = −2, and (1, 0) is on the x-axis. Then compose: f−1(f(3)) = f−1(4) = 3, read from the point (4, 3). You are back at 3, as a function and its inverse function require.

Counterexamples

What it is not, and the tempting wrong moves.

✗ Not this: A table of f has one column with 2 over 9 and another column with 9 over 4. Tempting answer: f−1(9) = 4, because 9 sits over 4.
The column 9 over 4 says f(9) = 4. That is f read forward. It tells you nothing about which input produced the output 9.
✓ Instead: Find 9 in the output row. It sits under 2, so f(2) = 9 and f−1(9) = 2.
✗ Not this: To draw f−1, flip the graph of f over the x-axis, so (1, 3) becomes (1, −3).
Flipping over the x-axis changes the sign of every output. It does not trade inputs with outputs. The point (1, −3) says 'the input 1 gives −3', which has nothing to do with undoing f(1) = 3.
✓ Instead: Trade the two numbers: (1, 3) becomes (3, 1). The mirror is the slanted line y = x, not the x-axis and not the y-axis.
✗ Not this: A function h has h(1) = 5, h(2) = 8 and h(3) = 5. Trade the rows of its table to get the table of h−1.
After the trade, the input 5 would have two outputs, 1 and 3. A function gives each input exactly one output, so the traded table is not a function. The rule applies only to a one-to-one function, and h repeats the output 5.
✓ Instead: h has no inverse function. Always check for a repeated output, or use the horizontal line test, before you trade anything.
✗ Not this: Swapping (−2, 5) gives (2, −5), because swapping moves the minus sign.
The swap moves whole numbers, and each number keeps its own sign. Nothing changes sign.
✓ Instead: (−2, 5) becomes (5, −2). In words: f(−2) = 5, so f−1(5) = −2.

Tips and tricks

  • Memory cue: 'Inverse: trade places.' The rows of the table trade places, the two numbers in each point trade places, and the two axes trade places.
  • Say each point as a sentence. (1, 3) on f is 'f sends 1 to 3'. Its swap (3, 1) is 'f inverse sends 3 back to 1'. If the two sentences do not undo each other, the swap was done wrong.
  • A point with two equal numbers, such as (2, 2), does not move when you swap it. So wherever the graph of f touches the line y = x, the graph of f−1 passes through the same spot.
  • Axis crossings trade axes. If f crosses the y-axis at (0, 3), then f−1 crosses the x-axis at (3, 0).
  • You can read f−1(k) from the graph of f without drawing anything new: start at height k on the y-axis, go across to the graph, then straight down to the x-axis.
  • To check a drawn inverse, turn the paper until the line y = x points straight up. The two graphs should look like the left and right wings of a butterfly. Use the same spacing on both axes, or the mirror will look bent.
  • After trading the rows of a table, the new input row is the old output row. If your new inputs are not exactly the old outputs, something was copied wrong.

7 The trap

Reading in the forward direction. For f−1(5), the 5 is an output of f, so look for it in the output row of the table, or at height 5 on the y-axis of the graph. If you look for 5 in the input row, or at 5 on the x-axis, you find f(5), the answer to the forward question. Before you read anything, say the question aloud: 'f gave 5; what went in?'

Flash cards

Say the answer out loud, then flip.

You have the table of a one-to-one function f. How do you get the table of f−1?
Trade the two rows. The old output row becomes the new input row, like reading a gym sign-in sheet from the person back to the locker. If you reorder, slide whole columns so each input stays with its own output.
What does f−1(5) ask, and where do you look for the 5?
It asks: f gave the output 5, so which input went in? Look for 5 in the output row of the table of f, or at height 5 on the y-axis of its graph, then read back to the input.
The point (2, 9) is on the graph of a one-to-one function f. Which point must be on the graph of f−1, and why?
The point (9, 2). The point (2, 9) says f(2) = 9, which means f−1(9) = 2. Swap the two numbers in the point.
The graph of f−1 is the graph of f flipped across which line, and how do you check a swapped point against it?
The line y = x, which acts as a mirror. A point (a, b) and its swap (b, a) must sit straight across that line from each other, the same distance away.

From earlier

Domain and range trade places

f is a one-to-one function. Where do you get the domain of f−1?
Show answer
From the range of f. The backward coat check takes in tickets, and only the tickets that were handed out.

Domain and range trade places

You know f(4) = 9, and f is a one-to-one function. Which number goes into f−1, and which comes out?
Show answer
9 goes in, so 9 is in the domain of f−1.
4 comes out, so 4 is in the range of f−1.
Reason: f(4) = 9 says the same thing as f−1(9) = 4.

Lesson 6 of 7

Restricting a domain so an inverse exists

Subject: MathematicsLeans on: The horizontal line testLeans on: Squaring a negative numberLeans on: Where a squared expression is smallestLeans on: Domain and range trade placesLeans on: Interval notationLeans on: Absolute valueLeans on: Order of operationsLeans on: Solving a one-step equationOpen this lesson
Leans on 2Comes back in 0

1 In plain words

Picture a town on a long straight road. A sign tells you how far you are from town, but not which side you are on. "Five miles" could mean five miles east or five miles west, so you cannot work back from the sign to one spot. Now agree to use only the road east of town, with the town itself included. On that half, "five miles" names exactly one spot, and you can always work back. Restricting a domain is the same move. When a function gives the same output to two different inputs, you keep only a piece of its inputs, a piece where every output comes from one input. On that piece the function can be undone.

2 Picture it

(0, 0)(2, 4)domainrange22468(0, 0)(2, 4)domainrange12313579
The right half of y = x2, starting at (0, 0): the horizontal line at height 4 meets it once, at (2, 4).

Reminder

Reminder: The horizontal line test, Squaring a negative number, Where a squared expression is smallest, Domain and range trade places, Interval notation, Absolute value, Order of operations, Solving a one-step equation
  • The horizontal line test. A function is a one-to-one function when no horizontal line meets its graph more than once. y = x2 fails: the line at height 4 meets it at x = −2 and at x = 2.
  • Squaring a negative number. A negative times a negative is positive: (−3)2 = (−3) × (−3) = 9. That is why −3 and 3 share the output 9.
  • Where a squared expression is smallest. A square is never negative, so it is smallest when the part being squared is 0. (x − 2)2 is smallest when x − 2 = 0, which is x = 2, and there it equals 0.
  • Domain and range trade places. The domain of f−1 is the range of f, and the range of f−1 is the domain of f. If f has domain [2, ∞) and range [1, ∞), then f−1 has domain [1, ∞) and range [2, ∞).
  • Interval notation. [0, ∞) means every number from 0 up, with 0 included. A square bracket includes the end, a round one leaves it out, and ∞ always gets a round one.
  • Absolute value. |x| is the distance from x to 0 on the number line, so it is never negative: |−6| = 6 and |6| = 6.
  • Order of operations. Work inside the parentheses first, then the power, then add or subtract: (5 − 2)2 + 1 = 32 + 1 = 9 + 1 = 10.
  • Solving a one-step equation. Undo what was done to x on both sides: x + 4 = 0 becomes x = −4 after subtracting 4 from both sides.

3 Why

An inverse must send each output back to one input. The function y = x2 sends both −3 and 3 to 9, so undoing 9 would need two answers, and a function may give only one. Nothing repairs that while both inputs stay. So you remove one input from every such pair by keeping only x ≥ 0. On that half a bigger input always gives a bigger square, so no output repeats. Every output is still reached, since each one from a negative input also comes from its positive partner. Keeping the right half is a convention: the left half works as well.

4 The rule

A restricted domain is a smaller set of inputs than the full domain of a function.

If f fails the horizontal line test, choose a restricted domain on which f is one-to-one. On that piece f has an inverse function. The usual choice also keeps every output of f, so the inverse accepts all of them. For f(x)=x2 the usual restricted domain is x≥0, written [0,∞). The domain of f−1 is then the range of f, and the range of f−1 is the restricted domain.

The same idea, five ways

Say it

f of x equals x squared, for x greater than or equal to zero. Or: x squared, restricted to x at least zero.

Write it

The function f squares its input, and the only inputs allowed are the numbers that are zero or more.

In math
  • f(x)=x2,x≥0
  • f(x) = x2 on [0, ∞)
  • domain: {x | x ≥ 0}
  • graph words: the right half of the parabola, starting at (0, 0) and climbing
Like

A road with the west half closed: every distance sign now points to exactly one spot.

See it
0[0, ∞)
The restricted domain x ≥ 0: a filled dot at 0, because 0 is kept, and shading to the right without end.

The same idea, other ways

As a picture

Look at the whole parabola y = x2. The horizontal line at height 4 meets it twice, at (−2, 4) and at (2, 4). Now lay your hand over everything left of the y-axis. The same line meets what is still showing only once, at (2, 4). Slide the line to any other height of 0 or more: one meeting point each time. Covering the left half is the restriction x ≥ 0.

(−2, 4)(2, 4)−222468
Cover the left half and the line at height 4 meets the graph once in place of twice.
As a story

A building has floors above the lobby and basement levels below it. A note says: meet me 3 levels from the lobby. That could be 3 up or 3 down, so the note cannot be traced back to one place. If the basement is closed, the same note has one meaning: floor 3. Closing the basement did not change how levels are counted. It only removed the second place a note could point to.

With small numbers

In the table, the output row reads 9, 4, 1, 0, 1, 4, 9. Every output except 0 appears twice, once under a negative input and once under its positive partner. Cross out the columns under −3, −2 and −1. What is left is 0, 1, 2, 3 on top and 0, 1, 4, 9 underneath. No output repeats, and no output is missing.

input xoutput x²−39−24−1100112439↑ solve: output given, read every input above it
The output 9 points back to two inputs, −3 and 3, until the negative columns are crossed out.
Why it must be true

Flip every point (a, b) of the whole parabola to (b, a), the way an inverse graph is made. The point (2, 4) becomes (4, 2), and (−2, 4) becomes (4, −2). Now the input 4 has two outputs, 2 and −2, so the flipped graph fails the vertical line test and is not a function. Flip only the kept half and (4, −2) never appears. Each input has one output, so the flipped graph is a function.

(1, 1)(2, 4)(4, 2)241234
The kept half (solid) and its flipped copy (dashed) mirror each other in the line y = x, and each passes the…
As a rule of thumb

For a graph with a single turn, such as a parabola: cut at the turn, keep one side. Find the point where the graph stops falling and starts climbing, or stops climbing and starts falling. Keep that point and everything on one side of it. Either side gives an inverse; this unit keeps the right side when nobody says which.

Piece of y = x2 keptPasses the horizontal line test?Still reaches every output y ≥ 0?Verdict
all xNo: −2 and 2 both give 4YesNo inverse
x≥0YesYesWorks. The usual choice
x≤0YesYesWorks. Use it when the problem names the left side
x > 0YesNo: the output 0 is lostHas an inverse, but loses the output 0. The usual choice keeps the turning point
x≥1YesNo: the outputs below 1 are lostHas an inverse, but loses the outputs below 1
x≥−1No: −1 and 1 both give 1YesToo big. Still no inverse

5 How

  1. Run the horizontal line test on the whole graph. If no horizontal line meets it twice, stop: the function already has an inverse and needs no restriction.
  2. If the test fails and the graph turns around once, as a parabola or an absolute value graph does, find that turn. For a squared expression, the turn is where the part being squared equals 0. For x2 that is x = 0. For (x − 2)2 it is x = 2.
  3. For a graph with that one turn, keep the turning point and everything on one side of it. Unless the problem names a side, keep the right side: x ≥ the turning value.
  4. Run the horizontal line test again on the kept piece only. Every horizontal line must now meet it at most once.
  5. Check whether any output was lost. The usual choice keeps the range of the whole function; a piece with a smaller range still has an inverse, but that inverse accepts fewer numbers.
  6. Write the restricted domain next to the formula, for example f(x) = x2, x ≥ 0. It is now part of the function.
  7. Trade places: the domain of f−1 is the range of f, and the range of f−1 is the restricted domain. Every answer the inverse gives must lie in the restricted domain.

Restricting a domain so an inverse exists

1
Does any horizontal line meet the graph more than once?
YesThe function is not a one-to-one function yet. Go on to restrict it.
NoStop. It already has an inverse, and no restriction is needed.
↓
2
Does the graph turn around at one lowest or highest point?
YesCut there. Keep that point and everything on one side of it.
NoKeep a stretch of inputs on which the graph only climbs or only falls, so no height repeats. Then work out the range of that stretch.
↓
3
Does the problem say which side to keep?
YesKeep the side it names.
NoKeep the right side: x ≥ the turning value. This unit uses that convention.
↓
4
Does the kept piece pass the horizontal line test?
YesIt has an inverse. If it also reaches every output of the whole function, it is the usual choice. Write the restricted domain beside the formula.
NoAn output still repeats, so the piece is too big. Make it smaller and test again.
  1. Run the horizontal line test on the whole graph.
  2. If it fails and the graph turns around once, find the turning point: for a squared expression, set the part being squared equal to 0 and solve.
  3. For a graph with one turn, keep the turning point and one side of it. Use the side the problem names, or the right side if none is named.
  4. Run the horizontal line test again on the kept piece.
  5. Check whether the kept piece still has the range of the whole function, as the usual choice does.
  6. Write the restricted domain beside the formula, then trade domain and range to describe the inverse.

6 Worked example

Worked example

Restricting y = x2 and undoing an output

The function f(x) = x2 fails the horizontal line test. Choose a restricted domain so that f has an inverse, give the domain and range of f−1, and find f−1(25).

Show the steps and the answer
(−5, 25)(5, 25)−6−4−224639152127
On the whole parabola the line at height 25 meets the graph twice, once on each side of 0.
input xoutput f(x)00112439416525↑ solve: output given, read every input above it
On the kept piece each output appears once, and 25 points back to the single input 5.
  1. Say what the question asks in plain words: pick a piece of the inputs on which no two inputs share an output. Then, for the output 25, find the one kept input that gives it.An inverse sends each output back to one input, so shared outputs have to go before anything can be undone.
  2. Show the problem with numbers: f(−5)=(−5)2=25 and f(5)=52=25. In the first picture, the horizontal line at height 25 meets the graph at (−5,25) and at (5, 25).Two inputs share the output 25, so f is not a one-to-one function on all x. Undoing 25 would have two answers.
  3. Find where the graph turns around. x2 is smallest when x=0, where it equals 0. Left of 0 the graph falls. Right of 0 it climbs.A square is never negative, so 0 is the lowest output. The repeats happen because the climb on the right retraces the heights of the fall on the left.
  4. Keep the turning point and the right side: x≥0, written [0,∞).No side is named in the problem, so use the convention. This removes one input from every pair that shared an output.
  5. Test the kept piece. For inputs 0 or more, a bigger input always gives a bigger square, so the graph only climbs and every horizontal line meets it at most once.A graph that only climbs can never come back to a height it has already passed.
  6. Check the range. The kept piece starts at height 0 and climbs without end, so its range is y≥0, written [0,∞). The whole parabola had the same range.Each output that came from a negative input also comes from its positive partner, so nothing was lost.
  7. Trade places. Domain of f−1: [0,∞), the range of f. Range of f−1: [0,∞), the restricted domain.The inputs of the inverse are the outputs of f, and the outputs of the inverse are the kept inputs of f.
  8. Find f−1(25). The output is 25. The kept input that gives it is 5, since 52=25 and 5≥0. So f−1(25)=5.The input −5 also squares to 25, but −5 is not in the restricted domain, so it cannot be an answer.
Answer
  • Restricted domain: x ≥ 0, or [0, ∞).
  • Domain of f−1: [0, ∞).
  • Range of f−1: [0, ∞).
  • f−1(25)=5.
Double-check A second way: look at the table of the kept piece. In the output row no number appears twice, and 25 sits in one column only, the column under the input 5. So the point (5, 25) is on the kept graph, and flipping it gives (25, 5) on the graph of the inverse, which says f−1(25) = 5.

Ladder: from easy to exam-hard

Try each rung before you open its steps.

Rung 1 of 6

A table with repeated outputs

The table shows a function p. Two of its outputs appear twice. Which inputs should you keep so that p has an inverse, without losing any output? Then find p−1(7).

input xoutput p(x)−27−14031427
x:−2,−1,0,1,2;p(x):7,4,3,4,7.
input xoutput p⁻¹(x)304172
x: 3, 4, 7; p⁻¹(x): 0, 1, 2.
Show the steps and the answer
input xoutput p(x)−27−14031427↑ solve: output given, read every input above it
The output 7 appears twice, under −2 and under 2, so 7 cannot be sent back to one input yet.
input xoutput p⁻¹(x)304172↓ evaluate: input given, read the output below it
The kept columns with the rows swapped: each input has one output, and 7 goes to 2.
  1. Say what the question asks in plain words: choose a set of inputs on which no output repeats. Then, for the output 7, find the one kept input that gives it.An inverse sends each output back to one input, so the repeats must go first.
  2. Look along the output row of the first table. The two 7s sit under the inputs −2 and 2. The two 4s sit under −1 and 1. The 3 sits under 0 alone.Each repeated output marks a pair of inputs that share it, so p is not a one-to-one function yet.
  3. Find where the outputs turn around. Reading left to right they fall 7, 4, 3 and then climb 4, 7. The turn is at the input 0.The repeats come from the climb retracing the fall, so the cut belongs at the turn.
  4. Keep the inputs 0, 1 and 2. Their outputs are 3, 4 and 7.This keeps the turn and the right side, the usual choice. The three outputs are all different, and 3, 4 and 7 are every output the full table had. Keeping −2, −1 and 0 would also work.
  5. Find p−1(7). In the kept columns, the output 7 sits under the input 2. So p−1(7)=2.The input −2 also gave 7, but it was thrown out, so only 2 remains.
Answer
  • Keep the inputs 0, 1 and 2.
  • p−1(7)=2.
Double-check Swap the two rows of the kept table, as you do to read an inverse from a table. The second table shows the result: the new inputs 3, 4 and 7 each have one output, 0, 1 and 2. That is a function, and it sends 7 to 2.

Rung 2 of 6

The problem names the left side

Restrict g(x) = |x| to x ≤ 0. Does g have an inverse on that piece? Give the domain and range of g−1, then find g−1(6).

Show the steps and the answer
(−6, 6)(0, 0)−6−4−2246
The left arm of y = |x|: the line at height 6 meets it once, at (−6, 6).
  1. Say what the question asks in plain words: the side to keep is given, the left half x≤0. Test that half, trade domain and range, and then, for the output 6, find the kept input that gives it.When a problem names the side, you use that side, not the usual right side.
  2. See why a restriction is needed. On the whole V-shaped graph, g(−6)=|−6|=6 and g(6)=|6|=6.Absolute value is distance from 0, and distance ignores direction, so −6 and 6 share the output 6.
  3. Keep x≤0: the left arm of the V, with its corner at (0, 0). On that arm, moving left always raises the graph.A graph that only rises as you move one way never reaches the same height twice, so g is a one-to-one function on this piece.
  4. Find the range of the kept piece: y≥0, written [0,∞).Distances are 0 or more, and the left arm reaches every one of them. The height 6, for example, is reached at x = −6.
  5. Trade places. Domain of g−1: [0,∞). Range of g−1: (−∞,0].The inverse takes the outputs of g as its inputs and hands back the kept inputs of g, which are 0 or less.
  6. Find g−1(6). Of the two numbers at distance 6 from 0, only −6 satisfies x≤0. So g−1(6)=−6.The answer of the inverse must lie in the restricted domain.
Answer
  • Yes. On x ≤ 0, g is a one-to-one function.
  • Domain of g−1: [0, ∞).
  • Range of g−1: (−∞, 0].
  • g−1(6)=−6.
Double-check Feed the answer back in: g(−6) = |−6| = 6, the output we started from. Also −6 lies in (−∞, 0], the range of g−1, so it is an allowed answer. The number 6 does not.

Rung 3 of 6

A piece that is too big

Someone restricts f(x) = x2 to −1 ≤ x ≤ 3 and says it now has an inverse. Is that right? If not, repair the restricted domain and give the domain and range of the inverse.

Show the steps and the answer
(−1, 1)(1, 1)22468
On −1 ≤ x ≤ 3 the line at height 1 still meets the graph twice, so this piece is too big.
  1. Say what the question asks in plain words: do any two inputs between −1 and 3 share an output? If they do, cut the piece down until none do.A restricted domain works only if the function is a one-to-one function on all of it.
  2. Try a pair on opposite sides of 0: f(−1)=(−1)2=1 and f(1)=12=1. Both −1 and 1 lie between −1 and 3.The piece reaches past the turning point at 0, so part of the left side is still there to share outputs. The horizontal line at height 1 meets this piece twice, so it fails.
  3. Cut at the turn: keep 0≤x≤3, written [0, 3].x2 is smallest at x = 0. From 0 to 3 the graph only climbs, so no height is reached twice.
  4. Find the range of the repaired piece. The lowest output is f(0)=0 and the highest is f(3)=9, so the range is 0≤y≤9, written [0, 9].The piece climbs from (0, 0) to (3, 9) and passes every height in between.
  5. Compare with the range before the repair. On −1≤x≤3 the lowest output was also 0 and the highest was also 9.The inputs that were removed, from −1 up to 0, gave outputs between 0 and 1. The inputs from 0 to 1 still give those outputs, so nothing was lost.
  6. Trade places. Domain of f−1: [0, 9]. Range of f−1: [0, 3].The domain of the inverse is the range of f, and the range of the inverse is the restricted domain.
Answer
  • No. The inputs −1 and 1 both give 1.
  • Repaired restricted domain: 0 ≤ x ≤ 3, or [0, 3].
  • Domain of f−1: [0, 9].
  • Range of f−1: [0, 3].
Double-check Test one of the removed inputs. f(−12) = (−12) × (−12) = 14, since the tops multiply to 1 and the bottoms multiply to 4. The kept input 12 also gives 14. So the output 14 is still reached, and now by one input only.

Rung 4 of 6

The turn is not at 0

Let h(x) = (x − 2)2 + 1. Find the usual restricted domain, give the domain and range of h−1, and find h−1(10).

Show the steps and the answer
(−1, 10)(2, 1)(5, 10)−224624681012
The whole graph turns at (2, 1), and the line at height 10 meets it at x = −1 and at x = 5.
input xoutput h(x)213245510617↑ solve: output given, read every input above it
On the kept piece the outputs only climb, and 10 points back to the single input 5.
  1. Say what the question asks in plain words: find where this graph turns, keep the right side, trade domain and range, and then, for the output 10, find the one kept input that gives it.The method is the same as for x2. Only the place of the turn has moved.
  2. Show that a restriction is needed: h(1)=(1−2)2+1=(−1)2+1=2, and h(3)=(3−2)2+1=12+1=2.The inputs 1 and 3 share the output 2, so h is not a one-to-one function on all x.
  3. Find the turn. Set x−2=0 and add 2 to both sides: x=2. This finds the one input where the squared part is as small as it can be. Plug it in: h(2)=(2−2)2+1=0+1=1.A square is never negative, so its smallest value is 0. That makes (2, 1) the lowest point of the graph, the place where it stops falling and starts climbing.
  4. Keep x≥2, written [2,∞).No side is named, so keep the turn and the right side. From x = 2 on, the graph only climbs.
  5. Find the range of the kept piece: y≥1, written [1,∞).The lowest output is 1, at the turn, and the graph climbs without end from there. The whole graph had the same range.
  6. Trade places. Domain of h−1: [1,∞). Range of h−1: [2,∞).The inverse takes outputs of h, which are 1 or more, and hands back kept inputs, which are 2 or more.
  7. Find h−1(10). Look at the table of the kept piece. The output 10 sits under the input 5, so h−1(10)=5.h(5) = (5 − 2)2 + 1 = 32 + 1 = 9 + 1 = 10, and 5 ≥ 2, so 5 is a kept input.
Answer
  • Restricted domain: x ≥ 2, or [2, ∞).
  • Domain of h−1: [1, ∞).
  • Range of h−1: [2, ∞).
  • h−1(10)=5.
Double-check On the whole graph, the other input that gives 10 is −1, since h(−1) = (−1 − 2)2 + 1 = (−3)2 + 1 = 9 + 1 = 10. But −1 is less than 2, so it was thrown out. That leaves 5 as the only kept input with output 10, which agrees with the table.

Rung 5 of 6

A shifted graph, left side kept

Restrict k(x) = (x − 1)2 to x ≤ 1, the left side this time. Give the domain and range of k−1 and find k−1(16).

Show the steps and the answer
input xoutput k(x)−316−29−140110↑ solve: output given, read every input above it
The kept piece x ≤ 1: no output repeats, and 16 points back to the single input −3.
  1. Say what the question asks in plain words: the problem names the piece, x≤1. Confirm that it works, trade domain and range, and then, for the output 16, find the kept input that gives it.The side is given, so do not switch to the right side out of habit.
  2. Find the turn. Set x−1=0 and add 1 to both sides: x=1. This finds the input where the squared part is 0, its smallest value. Plug it in: k(1)=(1−1)2=0.The named piece ends exactly at the turn, (1, 0). Moving left from there, the graph only climbs, so no output repeats on x ≤ 1.
  3. Find the range of the kept piece: y≥0, written [0,∞).The lowest output is 0, at the turn, and the piece climbs without end as you move left.
  4. Trade places. Domain of k−1: [0,∞). Range of k−1: (−∞,1].The inverse hands back kept inputs of k, and every kept input is 1 or less.
  5. Find k−1(16). In the table of the kept piece, the output 16 sits under the input −3. So k−1(16)=−3.k(−3) = (−3 − 1)2 = (−4)2 = 16, and −3 ≤ 1, so −3 is a kept input.
Answer
  • Domain of k−1: [0, ∞).
  • Range of k−1: (−∞, 1].
  • k−1(16)=−3.
Double-check The whole graph also reaches 16 at x = 5, since k(5) = (5 − 1)2 = 42 = 16. But 5 > 1, so 5 is outside the restricted domain. The answer has to lie in (−∞, 1], the range of k−1, and −3 does.

Rung 6 of 6

Decide whether a restriction is needed at all

Two functions: a(x) = x3 and b(x) = (x + 4)2 − 5. Decide for each whether it needs a restricted domain. For any that does, give the usual one and the domain and range of the inverse.

Show the steps and the answer
(−5, −4)(−4, −5)(−3, −4)−8−6−4−2−6−4−2246
The whole graph of b turns at (−4, −5), and the line at height −4 meets it on both sides of the turn.
input xoutput b(x)−4−5−3−4−2−1−14011
On the kept piece x ≥ −4 the outputs only climb, so none repeats.
  1. Say what the question asks in plain words: for each function, find out whether two inputs can share an output. Restrict only the one where they can.A restriction repairs a failed horizontal line test. A function that passes needs no repair.
  2. Test a. Try a pair of opposites: a(−2)=(−2)×(−2)×(−2)=−8 and a(2)=2×2×2=8. The outputs differ. The graph of x3 climbs the whole way from left to right.A negative number cubed stays negative and a positive one stays positive, and a bigger input always gives a bigger cube. So a passes the horizontal line test and needs no restriction.
  3. Test b. b(−5)=(−5+4)2−5=(−1)2−5=−4, and b(−3)=(−3+4)2−5=12−5=−4.The inputs −5 and −3 share the output −4, so b fails the horizontal line test and needs a restricted domain.
  4. Find the turn of b. Set x+4=0 and subtract 4 from both sides: x=−4. This finds the input where the squared part is 0, its smallest value. Plug it in: b(−4)=(−4+4)2−5=0−5=−5.The lowest point of the graph is (−4, −5). The shared outputs come from the two sides of that point.
  5. Keep x≥−4, written [−4,∞).No side is named, so keep the turn and the right side, where the graph only climbs.
  6. Find the range of the kept piece: y≥−5, written [−5,∞).The lowest output is −5, at the turn, and the graph climbs without end from there.
  7. Trade places. Domain of b−1: [−5,∞). Range of b−1: [−4,∞).The domain of the inverse is the range of b, and the range of the inverse is the restricted domain.
Answer
  • a(x) = x3 needs no restriction. It is already a one-to-one function.
  • b needs one. Usual restricted domain: x ≥ −4, or [−4, ∞).
  • Domain of b−1: [−5, ∞).
  • Range of b−1: [−4, ∞).
Double-check Look at the table of the kept piece of b. The output row reads −5, −4, −1, 4, 11. It only climbs, so no output appears twice. The output −4 now comes from −3 alone, because −5 lies left of −4 and was thrown out.

Counterexamples

What it is not, and the tempting wrong moves.

✗ Not this: After restricting f(x) = x2 to x ≥ 0, writing f−1(9) = 3 or −3.
−3 is not in the restricted domain, so the inverse can never hand it back. An inverse function gives one answer, and that answer must be a kept input.
✓ Instead: f−1(9) = 3, because 32 = 9 and 3 ≥ 0.
✗ Not this: Giving x > 0 as the usual restriction of f(x) = x2, leaving out 0.
The piece passes the horizontal line test and has an inverse, but the output 0 is lost, because only the input 0 gives it. The range shrinks from y ≥ 0 to y > 0.
✓ Instead: The usual restriction keeps the turning point: x ≥ 0, and the range stays y ≥ 0.
✗ Not this: Restricting every function before finding an inverse, for example cutting f(x) = 2x + 1 down to x ≥ 0.
A line that is not flat already passes the horizontal line test. Cutting it throws away outputs for no reason: with x ≥ 0 the outputs below 1 are gone.
✓ Instead: Test first. Restrict only when some horizontal line meets the graph more than once.
✗ Not this: Trying to repair y = x2 by cutting the outputs, for example "keep only y ≥ 0".
The range of x2 is already y ≥ 0, so this changes nothing. The trouble is two inputs sharing one output, and that lives on the input side.
✓ Instead: Cut the inputs: keep x ≥ 0.

Tips and tricks

  • Memory cue: cut at the turn, keep one side.
  • The cover check: lay your hand over the part of the graph you threw out, then run the horizontal line test on what is still showing.
  • The restricted domain is part of the function. f(x) = x2 with x ≥ 0 and f(x) = x2 with x ≤ 0 share a formula, but they are different functions with different inverses. Write the restriction beside the formula every time.
  • Keeping the right side is a convention, not a law. Nothing in the mathematics breaks if you keep the left side, but your inverse would hand back the inputs on the other side of the turn, and your answers would not match anyone who followed the convention. If a problem names a side, use the side it names.
  • A fast check on any answer from the inverse: it must lie in the restricted domain. If it does not, you picked the partner that was thrown out.
  • The kept piece and its inverse still mirror each other in the line y = x, as every function and its inverse do.

7 The trap

The most common mistake is to restrict the domain on paper and then forget it. Once you keep x ≥ 0 for f(x) = x2, the inputs −1, −2, −3 and the rest are gone. So f−1(9) is 3 only. It is not "3 or −3", and it is not −3. To avoid the mistake, write the restricted domain beside the formula every time you write the function, and check that each answer from the inverse lies inside the restricted domain.

Flash cards

Say the answer out loud, then flip.

Why can f(x) = x2 not be undone while it keeps all of its inputs?
Two different inputs share one output: −3 and 3 both give 9. An inverse may send 9 back to only one input. It is like a road sign that says "five miles from town" without saying which side.
For f(x) = (x − 2)2, where does the graph turn around, and which restricted domain do you usually keep?
  • The turn is where the part being squared equals 0, so x = 2.
  • Keep the turning point and the right side: x ≥ 2.
With f(x) = x2, x ≥ 0, what is f−1(9)?
3 only. The input −3 was removed by the restricted domain, so the answer is not "3 or −3".
After you restrict the domain of f, what are the domain and the range of f−1?
  • They trade places.
  • The domain of f−1 is the range of f.
  • The range of f−1 is the restricted domain.

From earlier

Domain and range trade places

f is a one-to-one function. Where do you get the domain of f−1?
Show answer
From the range of f. The backward coat check takes in tickets, and only the tickets that were handed out.

Domain and range trade places

You know f(4) = 9, and f is a one-to-one function. Which number goes into f−1, and which comes out?
Show answer
9 goes in, so 9 is in the domain of f−1.
4 comes out, so 4 is in the range of f−1.
Reason: f(4) = 9 says the same thing as f−1(9) = 4.

Lesson 7 of 7

Finding the inverse of a formula by swapping and solving

Subject: MathematicsLeans on: Undoing steps in reverse orderLeans on: Clearing a fractionLeans on: DistributingLeans on: Factoring out a common letterLeans on: Two square rootsLeans on: A negative on the bottom of a fractionLeans on: Domain and range trade placesLeans on: Checking that two functions are inversesOpen this lesson
Leans on 2Comes back in 0

1 In plain words

Think of wrapping a gift. First the box, then the paper, then the ribbon. To unwrap it you make the opposite moves in the opposite order: ribbon off, paper off, box open. A formula is a wrapping recipe. f(x) = 2x + 3 says: take the input, double it, then add 3. The inverse function is the unwrapping recipe: subtract 3, then halve. This lesson gives one method that writes the unwrapping recipe for you, even when the formula is too tangled to unwrap in your head. Write y for the output, trade the letters x and y, then get y alone. What stands on the other side is the inverse function.

2 Picture it

1. Write y = f(x)

2. Swap x and y

3. Solve for y

4. If ± appears, the domain of f picks the sign

5. Name the result f−1(x)

6. Check: a number through f, then through f−1, returns to the start

Reminder

Reminder: Undoing steps in reverse order, Clearing a fraction, Distributing, Factoring out a common letter, Two square roots, A negative on the bottom of a fraction, Domain and range trade places, Checking that two functions are inverses
  • Undoing steps in reverse order. To solve 2x + 3 = 11, undo the last step first: subtract 3 to get 2x = 8, then divide by 2 to get x = 4.
  • Clearing a fraction. Multiply both sides by the bottom and the bottom cancels: if 6 = 12n, then 6n = 12, so n = 2.
  • Distributing. A number or letter outside a bracket multiplies every term inside: x(y − 3) = xy − 3x. With numbers: 2(5 − 3) = 2 · 5 − 2 · 3 = 4.
  • Factoring out a common letter. When every term has the same letter, pull it to the front: xy − 2y = y(x − 2). Check by distributing back: y · x − y · 2.
  • Two square roots. Both 32 = 9 and (−3)2 = 9, so y2 = 9 has two answers, y = 3 and y = −3, written y = ±3 and read plus or minus 3. The symbol 9 alone means only 3, the one that is not negative.
  • A negative on the bottom of a fraction. Multiply the top and the bottom by −1 to move the sign: x−9−2 = 9−x2. Every sign on top changes.
  • Domain and range trade places. The inputs of f−1 are the outputs of f, and the outputs of f−1 are the inputs of f. If f takes only x ≥ 0, then f−1 gives only y ≥ 0.
  • Checking that two functions are inverses. Compose both ways and get x each time. For f(x) = x + 6 and g(x) = x − 6: f(g(x)) = (x − 6) + 6 = x and g(f(x)) = (x + 6) − 6 = x.

3 Why

The equation y = f(x) is a list of pairs: each input x with its output y. The inverse function holds the same pairs turned around, so each old output becomes an input. Trading the letters x and y in the equation turns every pair around at once. That swapped equation already is the inverse, but it is written backwards, with the new output y buried inside the formula. Solving for y digs it out, so the equation reads output = rule again. Calling the input x and the output y is a convention, and the swap keeps that convention for the inverse.

4 The rule

For a one-to-one function f: write y=f(x), swap x and y to get x=f(y), solve for y, and name the result f−1(x).

If solving produces ±, the domain of f picks the sign, because the outputs of f−1 are the inputs of f. The domain of f−1 is the range of f.

The same idea, five ways

Say it

when f is one-to-one, find f inverse by swapping x and y, then solving for y.

Write it

The inverse function of a one-to-one f is found by writing y = f(x), exchanging the letters x and y, and solving the new equation for y.

In math
  • y = f(x) becomes x = f(y)
  • x = f(y), solved for y, reads y = f−1(x)
  • f(a) = b means f−1(b) = a
  • f(x) = x2 with x ≥ 0 gives f−1(x) = x
Like

Like turning driving directions around for the trip home: read the list from the bottom up, and turn every left into a right.

See it
=x2y + 3do the same thing to both sides
The swapped equation is a level scale, and every move you make to get y alone must be made on both sides.

The same idea, other ways

As a story

In the morning you put on socks, then shoes. At night you take off shoes, then socks: the opposite moves in the opposite order. f(x) = 2x + 3 doubles, then adds 3. Its inverse function subtracts 3, then halves, so f−1(x) = x−32. Swapping and solving does this reversing for you. When you solve x = 2y + 3 for y, your first move is to subtract 3 and your second is to divide by 2.

With small numbers

f(x) = 2x + 3 sends 4 to 11, because 2 · 4 + 3 = 11. The inverse function must send 11 back to 4. Ask: which y gives 11 = 2y + 3? Subtract 3 to get 8, then halve to get 4. Now ask the same question for any output x in place of 11: x = 2y + 3, so y = x−32. Swapping and solving is that one question, answered for every output at once.

42x + 3(x − 3) ÷ 2firstsecond114
The first machine turns 4 into 11 and the second machine, the inverse, turns 11 back into 4.
As a picture

An earlier lesson showed that the graph of an inverse function is the mirror image of the original graph in the line y = x, because each point (a, b) becomes (b, a). Swapping the letters x and y in the equation is that same flip, done in algebra. The point (1, 5) on y = 2x + 3 becomes the point (5, 1) on the inverse, y = x−32.

(1, 5)(5, 1)−448−448
The solid line is y = 2x + 3 and the dashed line is its inverse.
Why it must be true

The first lesson of this unit said that f−1(b) = a exactly when f(a) = b. Put x in place of b and y in place of a: y = f−1(x) exactly when x = f(y). So the swapped equation is the definition of the inverse function, written out for this formula. Solving for y changes how the equation looks, not which pairs satisfy it.

As a rule of thumb

Write, swap, solve, check. One y after the swap: undo in reverse order. Two y's: gather them and factor. A squared y: take the root, then let the domain of f pick the sign.

Kind of formulaAfter the swapKey moveInverse
f(x) = 2x + 3x = 2y + 3Undo in reverse order: subtract 3, then divide by 2f−1(x)=x−32
f(x)=2x+1x−3x=2y+1y−3Clear the fraction, gather the y terms, factor y outf−1(x)=3x+1x−2
f(x)=x2,x≥0x=y2,y≥0Square root of both sides, keep the plus signf−1(x)=x
f(x)=x2,x≤0x=y2,y≤0Square root of both sides, keep the minus signf−1(x)=−x

Part 1: A linear function: f(x) = mx + b

A linear function does two things to its input: it multiplies by a number, then adds a number. After the swap, y sits in one place only. So you undo the two steps in reverse order, the same way you solve any linear equation.

  • Shape: f(x) = mx + b, with m ≠ 0
  • After the swap: x = my + b
  • Inverse: f−1(x) = x−bm. Rebuild it each time, it takes ten seconds.
  • Domain and range of f and of f−1: all real numbers
  • If m = 0 the function is constant. Every input shares one output, so there is no inverse function.
(6, 8)(8, 6)48124812
The solid line is f(x) = x/2 + 5 and the dashed line is its inverse.

Worked example

A linear function with a fraction in it

Find f−1(x) for f(x) = x2 + 5. In plain words: f halves its input, then adds 5. Find the rule that takes each output back to its input.

Show the steps and the answer
  1. Confirm an inverse function exists.The graph is a slanted straight line. A horizontal line meets a slanted line once, so f is a one-to-one function.
  2. Write y=x2+5.y is a short name for the output f(x).
  3. Swap the letters: x=y2+5.The inverse function uses the same pairs turned around. Now x stands for an old output and y for the input it came from.
  4. Subtract 5 from both sides: x−5=y2.Adding 5 was the last thing done to y, so it is the first thing to undo.
  5. Multiply both sides by 2: 2(x−5)=y, which is y=2x−10.Multiplying by 2 undoes the halving. Distributing gives 2 · x − 2 · 5 = 2x − 10.
  6. Rename y: f−1(x)=2x−10.y alone on one side is the output of the inverse function.
Answer
f−1(x)=2x−10
Double-check With a number: f(6) = 62 + 5 = 3 + 5 = 8, and f−1(8) = 2 · 8 − 10 = 6, the number we started with. By composing: f(f−1(x)) = 2x−102 + 5 = (x − 5) + 5 = x.

Part 2: A fraction with x on the top and the bottom: f(x) = (ax + b) ÷ (cx + d)

Here the input appears twice, once on the top and once on the bottom. After the swap, y appears twice, so you cannot peel the steps off one at a time. The plan has four moves: clear the fraction, bring every term that has y to one side, pull y out as a common factor, then divide by what is left in the bracket.

  • Shape: f(x) = ax+bcx+d, with c ≠ 0 and ad − bc ≠ 0. When ad − bc = 0, f is constant wherever it is defined and has no inverse.
  • After the swap: x = ay+bcy+d
  • Clear the fraction: x(cy + d) = ay + b
  • Gather and factor: y(cx − a) = b − dx
  • Inverse: f−1(x) = b−dxcx−a. Do not memorize this. Rebuild it with the four moves each time.
  • The bottom of f−1 is zero at x = ac. That is the one output f never reaches, so it is the one input f−1 cannot take.
(6, 8)(8, 6)−6612−6612
The solid curve is f(x) = (x + 2) ÷ (x − 5) and the dashed curve is its inverse.

Worked example

X on the top and on the bottom

Find f−1(x) for f(x) = x+2x−5. In plain words: find the formula that takes each output of f back to the input it came from.

Show the steps and the answer
  1. Confirm an inverse function exists.The graph of f is a shifted, stretched copy of y = 1x, and no horizontal line meets that shape twice. So f is a one-to-one function.
  2. Write y=x+2x−5, then swap: x=y+2y−5.The swap turns every pair around. y is now the old input we want.
  3. Multiply both sides by (y − 5): x(y−5)=y+2.Multiplying by the bottom cancels it on the right, so no fraction is left.
  4. Distribute on the left: xy−5x=y+2.x multiplies each term inside the bracket.
  5. Subtract y from both sides and add 5x to both sides: xy−y=5x+2.This puts every term that has y on the left and every term without y on the right.
  6. Factor y out on the left: y(x−1)=5x+2.Both left terms contain y. Distribute back to confirm: y · x − y · 1 = xy − y.
  7. Divide both sides by (x − 1): y=5x+2x−1.The bracket is multiplying y, so dividing by it leaves y alone.
  8. Rename and state the domain: f−1(x)=5x+2x−1, with x≠1.x = 1 would make the bottom zero. This matches f, which never gives the output 1.
Answer
  • f−1(x)=5x+2x−1
  • Domain of f−1: x ≠ 1
Double-check f(6) = 6+26−5 = 81 = 8, and f−1(8) = 40+28−1 = 427 = 6. A second number: f(4) = 6−1 = −6, and f−1(−6) = −30+2−6−1 = −28−7 = 4. Both return to the start.

Part 3: x2 on the restricted domain x ≥ 0

The full squaring function cannot be undone, because 3 and −3 both square to 9. Keep only the inputs that are zero or more, and each output comes from one input. After the swap you solve x = y2, which has two candidate answers, x and −x. The restricted domain tells you which sign to keep.

  • Function: f(x) = x2 with restricted domain x ≥ 0
  • Range of f: y ≥ 0
  • After the swap: x = y2 with y ≥ 0
  • Undo the square: y = x or y = −x
  • The condition y ≥ 0 keeps the plus sign: f−1(x) = x
  • Domain of f−1: x ≥ 0. Range of f−1: y ≥ 0.
  • The same formula on x ≤ 0 keeps the other sign: f−1(x) = −x
(2, 4)(4, 2)2412345
The right half of y = x2 is solid and its inverse y = x is dashed.

Worked example

Squaring, right half only

Find f−1(x) for f(x) = x2 with x ≥ 0. In plain words: f squares numbers that are zero or more. Find the rule that takes each square back to the number it came from.

Show the steps and the answer
  1. Confirm an inverse function exists.On x ≥ 0 the graph is the right half of the bowl, which rises all the way. No horizontal line meets it twice, so on this restricted domain f is a one-to-one function.
  2. Write y=x2, with x≥0 beside it.The domain condition is part of the function and must travel with the equation.
  3. Swap the letters everywhere: x=y2, with y≥0.After the swap, y stands for an old input, and old inputs were zero or more.
  4. Take the square root of both sides: y=±x.Both x and −x square to x. Write both until you have a reason to drop one.
  5. Keep the plus sign: y=x.The condition y ≥ 0 rules out the minus sign, since −x is below zero whenever x > 0.
  6. Rename and state the domain: f−1(x)=x, with x≥0.The domain of f−1 is the range of f, and squares are never below zero.
Answer
  • f−1(x)=x
  • Domain of f−1: x ≥ 0
Double-check f(3) = 9 and f−1(9) = 9 = 3. By composing: f(f−1(x)) = (x)2 = x, and f−1(f(x)) = x2 = x because x ≥ 0. With a negative number that last step would fail: (−3)2 = 9 = 3, not −3. That is why the domain was restricted.

5 How

  1. Make sure f is a one-to-one function on its domain. Use the horizontal line test, or use the restricted domain you are given. If two inputs share an output, stop: there is no inverse function yet.
  2. Write y in place of f(x). If the function comes with a domain condition such as x ≥ 0, write it beside the equation.
  3. Swap the letters. Every x becomes y and every y becomes x, in the equation and in the domain condition. So x ≥ 0 becomes y ≥ 0.
  4. Solve the swapped equation for y. One y: undo the steps in reverse order. Two y's in a fraction: clear the fraction, gather the y terms on one side, factor y out, divide. A squared y: take the square root of both sides and write ±.
  5. If ± appeared, keep the one sign that fits the condition on y.
  6. Write f−1(x) = your result, and state its domain, which is the range of f.
  7. Check. Send one number through f, then send the result through f−1. You should land on the number you started with.

Finding an inverse by swapping and solving

1
Does f pass the horizontal line test on its domain?
YesGo on. An inverse function exists.
NoStop. Restrict the domain to one piece first, then go on with that piece.
↓
2
After the swap, does y appear in only one place, with no square on it?
YesUndo the steps around y in reverse order, last step first.
NoGo on.
↓
3
Does y appear on the top and the bottom of a fraction?
YesMultiply both sides by the bottom, distribute, move every y term to one side and everything else to the other, factor y out, divide by the bracket.
NoGo on.
↓
4
Is y squared?
YesGet the squared piece alone, take the square root of both sides, write ±, then keep the sign that fits the condition on y.
NoGo back to the swap and look again at where y sits.
  1. Confirm that f is a one-to-one function on its domain.
  2. Write y = f(x), with the domain condition beside it if there is one.
  3. Swap x and y everywhere, in the domain condition too.
  4. Solve for y, using the branch of the map below that fits the formula.
  5. If ± appeared, keep the sign that fits the condition on y.
  6. Write f−1(x) = the result. Its domain is the range of f.
  7. Check with a number: through f, then through f−1, back to the start. For full proof, compose both ways and get x each time.

6 Worked example

Worked example

First full run: a linear function

Find the inverse function of f(x) = 4x − 7. In plain words: f multiplies its input by 4, then subtracts 7. Find the formula that takes each output back to the input it came from.

Show the steps and the answer
=x4y − 7do the same thing to both sides
After the swap the scale reads x on one side and 4y − 7 on the other.
  1. Confirm an inverse function exists.The graph of f is a slanted straight line, and a horizontal line meets a slanted line once. So f is a one-to-one function.
  2. Write y=4x−7.y is a short name for the output f(x). It makes the swap possible to write.
  3. Swap the letters: x=4y−7.The inverse function holds the same pairs turned around. From here on, x stands for an old output and y for the old input we are hunting.
  4. Add 7 to both sides: x+7=4y.Subtracting 7 was the last thing done to y, so it is undone first. This leaves the 4y term alone on its side.
  5. Divide both sides by 4: x+74=y.The 4 multiplies y, so dividing by 4 leaves y alone. The whole of x + 7 is divided, not only the 7.
  6. Rename y: f−1(x)=x+74.y alone on one side is the output of the inverse function, written as a rule in x.
Answer
f−1(x)=x+74
Double-check With a number: f(5) = 4 · 5 − 7 = 13, and f−1(13) = 13+74 = 204 = 5, the number we started with. By composing both ways: f(f−1(x)) = 4 · x+74 − 7 = (x + 7) − 7 = x, and f−1(f(x)) = (4x−7)+74 = 4x4 = x.

Ladder: from easy to exam-hard

Try each rung before you open its steps.

Rung 1 of 8

One step, adding

Find f−1(x) for f(x) = x + 6. In plain words: f adds 6. Find the rule that takes each output back.

Show the steps and the answer
10add 6inputoutput16
f turns 10 into 16, so the inverse must turn 16 back into 10.
  1. Write y=x+6.y names the output. The graph is a slanted line, so f is a one-to-one function and an inverse function exists.
  2. Swap the letters: x=y+6.The inverse function uses the pairs turned around.
  3. Subtract 6 from both sides: x−6=y.Subtracting 6 undoes adding 6 and leaves y alone.
  4. Rename y: f−1(x)=x−6.y alone is the output of the inverse function.
Answer
f−1(x)=x−6
Double-check f(10) = 16 and f−1(16) = 16 − 6 = 10. By composing: f(f−1(x)) = (x − 6) + 6 = x.

Rung 2 of 8

Two steps, with a negative number in front of x

Find f−1(x) for f(x) = 9 − 2x. In plain words: f multiplies its input by −2, then adds 9. Find the rule that runs this backwards.

Show the steps and the answer
  1. Write y = 9 − 2x, then swap: x=9−2y.The graph is a slanted line, so f is a one-to-one function. The swap turns every pair around.
  2. Subtract 9 from both sides: x−9=−2y.This removes the 9 from the side that holds y. The minus sign stays with the 2.
  3. Divide both sides by −2: y=x−9−2.The number multiplying y is −2, not 2, so that is what you divide by.
  4. Tidy the signs: y=9−x2.Multiplying the top and the bottom by −1 keeps the value and changes every sign: −(x − 9) = 9 − x.
  5. Rename y: f−1(x)=9−x2.y alone is the output of the inverse function.
Answer
f−1(x)=9−x2
Double-check f(1) = 9 − 2 = 7 and f−1(7) = 9−72 = 22 = 1. A second number: f(4) = 9 − 8 = 1 and f−1(1) = 9−12 = 82 = 4. By composing: f(f−1(x)) = 9 − 2 · 9−x2 = 9 − (9 − x) = x.

Rung 3 of 8

A linear formula written as one fraction

Find f−1(x) for f(x) = 2x+53. In plain words: f doubles its input, adds 5, then divides by 3. Find the rule that undoes all three steps.

Show the steps and the answer
  1. Write y=2x+53, then swap: x=2y+53.The bottom is only a number, so this is still a slanted line and a one-to-one function.
  2. Multiply both sides by 3: 3x=2y+5.Dividing by 3 was the last step done, so it is undone first.
  3. Subtract 5 from both sides: 3x−5=2y.Adding 5 was the step before that.
  4. Divide both sides by 2: y=3x−52.Doubling was the first step done, so it is undone last.
  5. Rename y: f−1(x)=3x−52.y alone is the output of the inverse function.
Answer
f−1(x)=3x−52
Double-check f(2) = 4+53 = 93 = 3 and f−1(3) = 9−52 = 42 = 2. By composing: f(f−1(x)) = (3x−5)+53 = 3x3 = x, since 2 · 3x−52 = 3x − 5.

Rung 4 of 8

X on the bottom only

Find f−1(x) for f(x) = 3x−2. In plain words: f subtracts 2 from its input, then divides 3 by the result. Find the rule that takes each output back.

Show the steps and the answer
  1. Note the domain of f: x≠2. Then write y=3x−2 and swap: x=3y−2.x = 2 would make the bottom zero. The graph is y = 1x stretched and shifted, which no horizontal line meets twice, so f is a one-to-one function.
  2. Multiply both sides by (y − 2): x(y−2)=3.Multiplying by the bottom cancels it, so y is no longer trapped under a fraction bar.
  3. Divide both sides by x: y−2=3x.x multiplies the bracket, so dividing by x frees it. This is allowed because x is never 0 here: x stands for an output of f, and 3 divided by a number is never 0.
  4. Add 2 to both sides: y=3x+2.This undoes the subtraction and leaves y alone.
  5. Rename and state the domain: f−1(x)=3x+2, with x≠0.The domain of f−1 is the range of f, and f never gives the output 0.
Answer
  • f−1(x)=3x+2
  • Domain of f−1: x ≠ 0
Double-check f(5) = 35−2 = 33 = 1 and f−1(1) = 31 + 2 = 5. A second number: f(3) = 31 = 3 and f−1(3) = 33 + 2 = 1 + 2 = 3.

Rung 5 of 8

X on the top and the bottom

Find f−1(x) for f(x) = 2x+1x−3. In plain words: find the formula that takes each output of f back to its input. After the swap, y will show up twice.

Show the steps and the answer
(4, 9)(9, 4)−8−44812−8−44812
The solid curve is f and the dashed curve is its inverse.
  1. Note the domain of f: x≠3. Write y=2x+1x−3 and swap: x=2y+1y−3.The graph is a shifted, stretched copy of y = 1x, so f is a one-to-one function and an inverse function exists.
  2. Multiply both sides by (y − 3): x(y−3)=2y+1.This clears the fraction.
  3. Distribute on the left: xy−3x=2y+1.x multiplies each term inside the bracket. Now you can see both y terms.
  4. Subtract 2y from both sides and add 3x to both sides: xy−2y=3x+1.Every term with y is now on the left and every term without y is on the right.
  5. Factor y out on the left: y(x−2)=3x+1.Both left terms contain y. Two y's have become one, which is the point of this move.
  6. Divide both sides by (x − 2): y=3x+1x−2.The bracket multiplies y, so dividing by it leaves y alone.
  7. Rename and state the domain: f−1(x)=3x+1x−2, with x≠2.x = 2 would make the bottom zero. It is also the one output f never reaches.
Answer
  • f−1(x)=3x+1x−2
  • Domain of f−1: x ≠ 2
Double-check f(4) = 8+14−3 = 91 = 9 and f−1(9) = 27+19−2 = 287 = 4. A second number: f(2) = 5−1 = −5 and f−1(−5) = −15+1−5−2 = −14−7 = 2.

Rung 6 of 8

Top and bottom, with numbers in front of every x

Find f−1(x) for f(x) = 3x−42x+5. In plain words: the same job as rung 5, with more signs to watch.

Show the steps and the answer
  1. Write y=3x−42x+5 and swap: x=3y−42y+5.The graph is again a shifted, stretched copy of y = 1x, so f is a one-to-one function.
  2. Multiply both sides by (2y + 5): x(2y+5)=3y−4.This clears the fraction.
  3. Distribute on the left: 2xy+5x=3y−4.x multiplies both terms in the bracket.
  4. Subtract 3y from both sides and subtract 5x from both sides: 2xy−3y=−5x−4.Terms with y go left, terms without y go right. Each term changes sign when it crosses over.
  5. Factor y out: y(2x−3)=−5x−4.Check by distributing back: y · 2x − y · 3 = 2xy − 3y.
  6. Divide both sides by (2x − 3): y=−5x−42x−3.This leaves y alone.
  7. Tidy the signs: y=5x+43−2x.Multiplying the top and the bottom by −1 changes every sign and keeps the value. Either form is correct.
  8. Rename and state the domain: f−1(x)=5x+43−2x, with x≠32.The bottom 3 − 2x is zero when x = 32, so that input is thrown out.
Answer
  • f−1(x)=5x+43−2x
  • Domain of f−1: x ≠ 32
Double-check f(−2) = −6−4−4+5 = −101 = −10 and f−1(−10) = −50+43+20 = −4623 = −2. A second number: f(−3) = −9−4−6+5 = −13−1 = 13 and f−1(13) = 65+43−26 = 69−23 = −3.

Rung 7 of 8

X2 on the other half, x ≤ 0

Find f−1(x) for f(x) = x2 with x ≤ 0. In plain words: f squares numbers that are zero or negative. Find the rule that takes each square back to the number it came from.

Show the steps and the answer
0(−∞, 0]
The inputs of f are zero and everything to its left, so the outputs of the inverse must land there too.
  1. Confirm an inverse function exists.On x ≤ 0 the graph is the left half of the bowl, which falls all the way. No horizontal line meets it twice, so on this restricted domain f is a one-to-one function.
  2. Write y=x2 with x≤0, then swap everywhere: x=y2 with y≤0.After the swap, y stands for an old input, and old inputs were zero or negative.
  3. Take the square root of both sides: y=±x.Both x and −x square to x. Write both, then choose.
  4. Keep the minus sign: y=−x.The condition y ≤ 0 rules out the plus sign, since x is above zero whenever x > 0.
  5. Rename and state the domain: f−1(x)=−x, with x≥0.The domain of f−1 is the range of f, and squares are never below zero.
Answer
  • f−1(x)=−x
  • Domain of f−1: x ≥ 0
Double-check f(−4) = (−4)2 = 16 and f−1(16) = −16 = −4, the number we started with. The plus sign would have returned 4, which is not in the domain of f.

Rung 8 of 8

A shifted square on a restricted domain

Find f−1(x) for f(x) = (x − 1)2 + 2 with x ≥ 1. In plain words: f subtracts 1, squares, then adds 2, and only inputs of 1 or more are allowed. Find the rule that runs this backwards, and its domain.

Show the steps and the answer
(1, 2)(4, 11)2424681012
Only the right half of the bowl is kept.
  1. Confirm an inverse function exists.For x ≥ 1 the bracket x − 1 is zero or more, and two different numbers that are zero or more never have the same square. So no two inputs share an output.
  2. Find the range of f: outputs are 2 or more.(x − 1)2 is never below zero, and it equals zero at x = 1. Adding 2 makes the smallest output 2.
  3. Write y=(x−1)2+2 with x≥1, then swap everywhere: x=(y−1)2+2 with y≥1.The domain condition travels with the swap.
  4. Subtract 2 from both sides: x−2=(y−1)2.Adding 2 was the last step done, so it is undone first. The squared piece is now alone.
  5. Take the square root of both sides: y−1=±x−2.Both x−2 and −x−2 square to x − 2. The whole of x − 2 sits inside the root.
  6. Keep the plus sign: y−1=x−2.y ≥ 1 means y − 1 is zero or more, so it cannot equal a negative number.
  7. Add 1 to both sides: y=1+x−2.This undoes the subtraction inside the bracket and leaves y alone.
  8. Rename and state the domain: f−1(x)=1+x−2, with x≥2.The domain of f−1 is the range of f, which is 2 or more. It also keeps the inside of the root from going below zero.
Answer
  • f−1(x)=1+x−2
  • Domain of f−1: x ≥ 2
Double-check f(4) = (4 − 1)2 + 2 = 9 + 2 = 11 and f−1(11) = 1 + 11−2 = 1 + 9 = 1 + 3 = 4. At the edge: f(1) = 0 + 2 = 2 and f−1(2) = 1 + 0 = 1.

Counterexamples

What it is not, and the tempting wrong moves.

✗ Not this: f(x) = 4x − 7, so f−1(x) = 14x−7.
The −1 in the notation f−1 is a name, not a power, as the first lesson of this unit said. Test it: f(5) = 13, but 14·13−7 = 145, which is not 5.
✓ Instead: Swap and solve: x = 4y − 7 gives f−1(x) = x+74, and 13+74 = 5.
✗ Not this: f(x) = 4x − 7 multiplies by 4 and subtracts 7, so the inverse divides by 4 and adds 7: f−1(x) = x4 + 7.
The opposite steps are right but the order is not reversed. Test it: f(5) = 13, but 134 + 7 = 10.25, not 5.
✓ Instead: Undo the last step first: add 7, then divide by 4. f−1(x) = x+74.
✗ Not this: For f(x) = x2 with x ≤ 0, the inverse is f−1(x) = x, because a square root undoes a square.
f(−3) = 9, but 9 = 3, which is not −3 and is not even in the domain of f. The outputs of f−1 must be the inputs of f, which are zero or negative.
✓ Instead: The condition y ≤ 0 keeps the minus sign: f−1(x) = −x, and −9 = −3.
✗ Not this: For f(x) = 2x+1x−3: from x(y − 3) = 2y + 1, divide by x and add 3 to get y = 2y+1x + 3, and stop there.
y is alone on the left but still appears on the right. You cannot compute y from x with this, so it is not solved.
✓ Instead: Distribute, gather the y terms, factor: xy − 2y = 3x + 1, so y(x − 2) = 3x + 1 and f−1(x) = 3x+1x−2.
✗ Not this: f(x) = x2 on all real numbers: swap to x = y2, solve to y = ±x, so f−1(x) = ±x.
This is a case where the method does not apply. ±x gives two outputs for one input, so it is not a function. The cause is that f fails the horizontal line test: 3 and −3 share the output 9.
✓ Instead: Restrict the domain first. On x ≥ 0 the inverse function is x. On x ≤ 0 it is −x.

Tips and tricks

  • Four words to carry: write, swap, solve, check.
  • The ten-second check: send one number through f, then send the result through f−1. If you do not land on your starting number, there is a slip. Pick a starting number that keeps the arithmetic clean, and avoid 0 and 1, which can hide mistakes.
  • A number check catches errors but does not prove the answer. Composing both ways and getting x each time is the proof.
  • For a linear function you can skip the swap: list the steps f does, reverse the order, and replace each step with its opposite. Use the swap method to confirm.
  • With y on the top and the bottom, say the plan out loud: clear, gather, factor, divide.
  • After any square root, look back at the domain of f. Inputs of f are outputs of f−1, so keep the sign whose outputs lie in that domain. For x2 itself, x ≥ 0 keeps +√ and x ≤ 0 keeps −√.
  • The domain of f−1 is the range of f, not whatever the finished formula seems to allow. Example: f(x) = x has outputs that are zero or more, so its inverse is x2 with x ≥ 0 only, even though the formula x2 accepts any number.
  • Do not memorize a formula for the inverse of ax+bcx+d. The four moves rebuild it in under a minute, and a memorized formula with one wrong sign is worse than none.

7 The trap

The most common mistake comes with the square root. After x = y2, people write y = x out of habit, or leave y = ±x as the answer. Neither is safe. Write ± first, then look at the condition on y that came from the domain of f, and keep the one sign that fits it. For x ≥ 0 that is x. For x ≤ 0 it is −x. Then test with a number from the domain of f: it must come back to itself.

Flash cards

Say the answer out loud, then flip.

What are the three main moves for finding the inverse function of a one-to-one formula?
  • Write y in place of f(x).
  • Swap the letters x and y.
  • Solve for y, then name the result f−1(x).
Why does trading the letters x and y produce the inverse function?
The equation y = f(x) is a list of pairs, each input with its output. Trading the letters turns every pair around at once. Solving for y then makes the equation read output = rule again.
f(x) = x2 with x ≤ 0. After the swap you have x = y2 with y ≤ 0. What is f−1(x)?
  • f−1(x)=−x.
  • Write y = ±x first, then keep the sign that fits y ≤ 0.
  • Check: f(−3) = 9, and −9 = −3, the number you started with.
f(x) = 2x + 3 wraps the input: double it, then add 3. What is the unwrapping recipe, and how do you check it?
  • Opposite moves in the opposite order, like taking the ribbon off before the paper: subtract 3, then halve.
  • f−1(x)=x−32.
  • Check: f(1) = 5, and f−1(5) = 5−32 = 1.

From earlier

Domain and range trade places

f is a one-to-one function. Where do you get the domain of f−1?
Show answer
From the range of f. The backward coat check takes in tickets, and only the tickets that were handed out.

Checking that two functions are inverses

What two things must both be true for f and g to be inverses of each other?
Show answer
f(g(x)) = x for every input that g accepts.
g(f(x)) = x for every input that f accepts.
Like two bus routes: both round trips must end where they began.