1.4 Inverse Functions

One-to-one functions and the horizontal line test, the definition of an inverse through its two cancellation equations, finding an inverse algebraically by swapping and solving, the reflection of graphs in the line y equals x, restricting a domain to force an inverse to exist, and the inverse trigonometric functions with their conventional ranges.

Subject: Calculus I · 66 slides · symbolic lesson

Open the interactive version of this deck

What this lesson covers

The lesson, slide by slide

1. Inverse Functions

Title

Calculus I · Section 1.4

Undoing a function, and paying for it when it cannot be undone

2. What this lesson gives you

Objectives

Calculus keeps asking questions backwards: given the output, which input produced it? An inverse function answers that question, when an answer exists.

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 65 — learning objectives 1.4.1 to 1.4.5

A function takes an input and produces an output. A surprising amount of calculus runs the other way: you know the output, a temperature, a concentration, a sine value, and you want the input that produced it. An inverse function is the tool that answers that question in one step.

The catch is that not every function can be run backwards. If two different inputs land on the same output, the output alone cannot tell you which one you started from. So this lesson begins with a test for when an inverse exists, then shows how to find one, how to draw one, and what to do when the function you have fails the test.

The last part applies all of that to the trigonometric functions, which fail the test as badly as a function can. Their inverses are defined by a convention you will use for the rest of calculus, and the convention has consequences that catch out almost everybody at least once.

3. Before anything new: run a function backwards

Warm-up

\[ f(x) = x^3 + 4 \]

Discussion prompt

Some input x produced the output 12. Find it. Then write, as a formula in y, the input that produces a general output y.

Commit to an answer before revealing it. The first question is ordinary equation solving: subtract four, then take a cube root. The second question asks for the same steps done once and for all, with a letter in place of twelve.

Notice the order you undid things in. The function cubed first and added four second, so you subtracted four first and took the cube root second. Undoing always reverses the order, the way you take off your shoes before your socks come off.

The formula you wrote, the cube root of y minus four, is a new function. It takes an output of the original and returns the input. That is exactly what the rest of the lesson calls the inverse function. The question for the next few slides is whether this always works, and it does not.

4. What it means to undo a function

Section

Part 1

5. Undoing the cube-plus-four machine

Concept

Figure (svg): Two rounded boxes. The left box, the domain of f, holds the inputs minus 1, 0 and 2; the right box, the range of f, holds 3, 4 and 12. Orange arrows labelled f run left to right from each input to its output; a dashed cyan arrow labelled f inverse runs from 12 back to 2.

Every orange arrow can be reversed because no two of them land on the same output. The inverse is nothing more than the arrows run backwards.

\[ y = x^3 + 4 \quad\Longleftrightarrow\quad x = \sqrt[3]{y - 4} \]

Call the reverse machine f inverse. Feed it what f produced and you get back what you started with:

\[ f^{-1}\left(f(x)\right) = f^{-1}\left(x^3 + 4\right) = \sqrt[3]{x^3} = x \]

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 65 — Existence of an Inverse Function

Look at the picture as a set of arrows. The original function sends minus one to three, zero to four and two to twelve. The inverse is those same arrows reversed: it sends twelve back to two, and so on.

Reversing arrows is only possible because no two arrows land on the same output. If two of them pointed at twelve, the reversed arrow from twelve would have to point at two places at once, and a function is not allowed to do that.

The last line is the check that matters. Feed any input x into the original function, feed the result into the inverse, and x comes back untouched. The cube root and the cube cancel, the minus four and the plus four cancel, and nothing is left but x. That cancellation is going to be the definition.

6. The definition: two cancellation equations

Concept

Inverse function — Given a function f with domain D and range R, its inverse (if it exists) is the function with domain R and range D that sends each output back to the input that produced it.

\[ f^{-1}(y) = x \quad\text{if and only if}\quad f(x) = y \]

\[ f^{-1}\left(f(x)\right) = x \text{ for all } x \text{ in } D, \qquad f\left(f^{-1}(y)\right) = y \text{ for all } y \text{ in } R \]

The roles of the two sets swap: the range of f becomes the domain of the inverse, and the domain of f becomes its range.

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 66 — Definition and equation 1.11

The formal definition says the inverse sends y to x exactly when the original sends x to y. Everything else follows from that one sentence.

In particular it gives two cancellation equations. Doing f and then undoing it returns every input of f. Undoing first and then doing f returns every output of f. You need both: one of them alone is not enough, and a later slide shows a pair of functions that passes one and fails the other.

Pay attention to the sets. The inputs of the inverse are the outputs of f, so the domain of the inverse is the range of f. The outputs of the inverse are the inputs of f, so the range of the inverse is the domain of f. You will use this swap every time you are asked for the domain and range of an inverse, and it saves you from working them out from scratch.

7. Reading equation 1.11 piece by piece

Notation

Annotate

On: \( f^{-1}\left(f(x)\right) = x,\; x \in D \qquad f\left(f^{-1}(y)\right) = y,\; y \in R \)

  • Read it 'f inverse'. The raised minus one is a name tag, not an exponent.
  • Do f, then undo it: every input of f comes back unchanged.
  • Undo first, then do f: every output of f comes back unchanged.
  • Each equation is only claimed on its own set. Outside those sets it may fail, and in Part 6 it does.

Click through the notes one at a time. The notation is compact, and each piece is carrying a separate idea.

The raised minus one is the part most worth slowing down on. With numbers, a raised minus one means a reciprocal. Attached to the name of a function it means something else entirely: the function that undoes this one. The two meanings clash, and the next slide is about what happens when you mix them up.

The last note, about the sets, looks like fine print, but it is the most important part of the definition for this lesson. Each cancellation equation is only promised on its own set of inputs. For most functions in this section that never matters. For the inverse trigonometric functions at the end, stepping outside those sets is exactly how the cancellation breaks.

8. Trap: reading the minus one as a reciprocal

Trap

The trap

A natural but wrong reading of the notation:

\[ f^{-1}(x) = \frac{1}{f(x)} \]

Wrong. Test it on the warm-up function at 12.

The fix

The inverse undoes f; the reciprocal divides one by f. For the cube-plus-four function they are completely different numbers.

\[ f^{-1}(12) = 2 \]

\[ \frac{1}{f(12)} = \frac{1}{1732} \]

This slip is common because the notation invites it. In algebra, x to the minus one really is one over x, so it feels natural to read f inverse of x as one over f of x.

The numbers show how different the two things are. The inverse asks which input produced twelve, and the answer is two. The reciprocal first computes the output at twelve, which is seventeen hundred and thirty-two, and then divides one by it. One is a question about going backwards; the other is ordinary arithmetic going forwards.

If you ever want the reciprocal, write one over f of x, with the fraction bar. Keep the raised minus one on a function name for the inverse only. For the trigonometric functions this matters a great deal: one over sine has its own name, cosecant, and it is not the inverse sine.

9. Why squaring cannot be undone

Concept

Figure (svg): The parabola y equals x squared with a dashed horizontal line at height 4 meeting it at two points, x equals minus 2 and x equals 2.

Given only the output 4 you cannot say which input produced it. An inverse would have to return two answers, and a function may return only one.

\[ y = x^2 \;\Longrightarrow\; x = \pm\sqrt{y} \]

For every positive output there are two candidate inputs. A formula that returns two numbers is not a function, so the squaring function has no inverse on the whole real line.

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 66 — the discussion of x squared

Try the same recipe on the squaring function. Solving y equals x squared for x gives plus or minus the square root of y. That plus or minus is the problem.

Look at the dashed line at height four in the picture. It meets the parabola twice, at minus two and at two. If someone tells you the output was four, you genuinely cannot say which input it came from, and no cleverness with algebra will fix that. The information was lost the moment the function squared.

So the squaring function, on the whole real line, has no inverse. This is not a technicality: it is the reason the square root key on your calculator gives only one answer, and the reason you have to add the plus or minus by hand when you solve a quadratic. Part 5 shows how to rescue an inverse by throwing away half of the parabola.

10. One-to-one: different inputs, different outputs

Concept

One-to-one function — A function that never sends two different inputs to the same output.

\[ x_1 \ne x_2 \;\Longrightarrow\; f(x_1) \ne f(x_2) \]

Equivalently, and usually easier to check algebraically: equal outputs force equal inputs.

\[ f(x_1) = f(x_2) \;\Longrightarrow\; x_1 = x_2 \]

A function has an inverse exactly when it is one-to-one.

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 66 — Definition of a one-to-one function

The property the cube-plus-four function had and the squaring function lacked gets a name: one-to-one. Different inputs always give different outputs.

The second way of writing it is the one to use when you prove something with algebra. Assume two outputs are equal, then show the inputs must be equal. If the algebra forces that, the function is one-to-one. If you can find two different inputs with the same output, one example is enough to show it is not.

The last sentence on the slide is the whole point of the definition. A one-to-one function can always be run backwards, because each output came from exactly one input. A function that is not one-to-one never can, because at least one output is ambiguous. So before you try to find an inverse, the first question is always whether the function is one-to-one.

11. The horizontal line test

Section

Part 2

12. The horizontal line test

Concept

Figure (svg): Left: y equals x squared with two horizontal lines, each meeting the parabola at two points. Right: y equals x cubed with two horizontal lines, each meeting the curve exactly once.

Slide a horizontal ruler up and down each graph. On the parabola it catches two points at almost every height; on the cubic it never catches more than one.

Horizontal line test — A function is one-to-one if and only if every horizontal line meets its graph at most once.

A horizontal line is the set of points at one height, one output value. Two hits on that line are two inputs sharing that output.

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, pp. 66-67 — Rule: Horizontal Line Test and Figure 1.38

Here is the quick way to see one-to-one on a graph. A horizontal line is all the points at a single height, so it represents a single output value. Wherever it touches the graph, it marks an input that produces that output.

On the left, horizontal lines through the parabola hit it twice, once on each side of the vertex. Two hits means two inputs sharing an output, so the squaring function is not one-to-one. On the right, the cubic keeps rising, and every horizontal line crosses it exactly once.

The test is exact: a function is one-to-one if and only if no horizontal line meets its graph more than once. It is only as reliable as the graph you are looking at, though. A graph drawn over too small a window can hide a second crossing, so when it matters, back the picture up with the algebra from the previous slide.

13. Two line tests, two different questions

Intuition

testquestion it answerswhat a failure means
vertical lineis this graph a function at all?one input with two outputs
horizontal lineis this function one-to-one?one output with two inputs

The two tests are mirror images of each other, which is no accident: the graph of the inverse is the mirror image of the graph of f, so the horizontal test on f is the vertical test on its inverse.

You have met the vertical line test before. It decides whether a graph is the graph of a function at all: a vertical line is a single input, and a function may give only one output for it.

The horizontal line test asks a different question about something that is already a function. It is easy to confuse the two under pressure, so tie each to its question. Vertical: can one input have two outputs? Horizontal: can one output have two inputs?

There is a neat reason the two tests look like reflections of each other. You will soon see that the graph of an inverse is the graph of f reflected in the diagonal line. Reflecting turns horizontal lines into vertical ones. So a function passes the horizontal line test exactly when its reflection passes the vertical line test, that is, exactly when the inverse is a genuine function.

14. Example 1.28(a): a staircase

Worked example

Use the horizontal line test on the staircase graph, the floor function that rounds each x down to a whole number.

Figure (svg): A staircase graph: for each whole number n from 0 to 7, a horizontal segment at height n from x equals n to n plus 1, closed dot on the left end and open dot on the right. The horizontal line y equals 2 lies along a whole step.

The line y = 2 does not just touch the graph twice; it runs along a whole step, so infinitely many inputs share the output 2.

Pick a horizontal line at a whole-number height

Why: Take height 2.

\[ y = 2 \]

Find where it meets the graph

Why: It lies along the whole step from 2 up to, but not including, 3.

\[ f(x) = 2 \quad\text{for every } x \in [2, 3) \]

Conclude

Why: Any whole-number height meets the graph infinitely often.

\[ \text{not one-to-one} \]

Check with two explicit inputs

Why: Two different inputs, one output, which is exactly what one-to-one forbids.

\[ f(2) = 2 = f(2.5), \quad 2 \ne 2.5 \]

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, pp. 67-68 — Example 1.28a

The staircase is the floor function: it takes any number and rounds it down to the whole number at or below it. Every input from two up to, but not including, three rounds down to two.

That is why the horizontal line at height two does something more dramatic than cross twice. It lies along an entire step. Infinitely many inputs share the output two, and the same happens at every whole-number height.

The check step is the kind of evidence a proof needs. A graph is persuasive, but naming two specific different inputs with the same output settles the matter completely: two and two and a half both give two. One such pair is all it takes to show a function is not one-to-one. You never need more than one counterexample.

15. Example 1.28(b): one over x

Worked example

\[ f(x) = \frac{1}{x}, \quad x \ne 0 \]

Figure (svg): The two branches of y equals one over x. Three dashed horizontal lines at heights 2, one half and minus 1.5 each meet the graph at exactly one point; the x-axis itself is never reached.

Each nonzero height is reached exactly once, at x equal to its reciprocal. Height zero is never reached, which is why zero will be missing from the inverse's domain.

Read the graph

Why: Every horizontal line meets the graph at most once.

\[ \text{one-to-one} \]

Back it up algebraically

Why: Suppose two inputs give equal outputs.

\[ \frac{1}{x_1} = \frac{1}{x_2} \]

Cross-multiply

Why: Both inputs are nonzero, so this is allowed.

\[ x_2 = x_1 \]

Check by solving for the input

Why: A height c other than zero is hit only at x equal to one over c, and height zero is never hit.

\[ \frac1x = c \;\Longrightarrow\; x = \frac1c \quad (c \ne 0) \]

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 68 — Example 1.28b

The book settles this one from the graph: each horizontal line crosses at most once. The picture shows three such lines, each with a single crossing, and the x-axis, which is never crossed at all.

The algebra confirms it in two lines. Assume two inputs give the same reciprocal, cross-multiply, and the inputs come out equal. That is the definition of one-to-one, verified for every pair of inputs at once rather than for the handful a picture can show.

The final step does even more: it solves for the input. A height c is reached only at one over c, which is already a formula for the inverse. One over x is its own inverse. And height zero is never reached, which tells you in advance that zero will be missing from the domain of the inverse, because it is missing from the range of f.

16. Checkpoint 1.23: x cubed minus x

Prediction

Figure (svg): The cubic y equals x cubed minus x for x from minus 2 to 2: it rises, turns down through the origin, and rises again, with a small hump above the x-axis on the left and a dip below it on the right.

Before answering, imagine a ruler lying along the x-axis itself.

Predict first

Is f(x) = x³ − x one-to-one?

  • Yes: it is a cubic, and cubics pass
  • Yes: it rises overall
  • No: some horizontal line meets it three times
  • Cannot tell from a graph

Correct: No: some horizontal line meets it three times

Why: The x-axis is itself a horizontal line, and it meets the graph at minus 1, 0 and 1, since the cubic factors as x times x minus 1 times x plus 1. Three inputs share the output zero. Not every cubic is one-to-one; the plain cube is, because it never turns around.

\[ f(-1) = f(0) = f(1) = 0 \]

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 68 — Checkpoint 1.23

Make your prediction before revealing it. The tempting answer is yes, because the plain cube passed the test on an earlier slide and this is also a cubic.

Look for a horizontal line that crosses three times. The easiest one to see is the x-axis itself. The cubic crosses it where x cubed minus x is zero, and factoring gives x times x minus one times x plus one, so the crossings are at minus one, zero and one. Three different inputs, one output.

The lesson is that the family a function belongs to does not decide whether it is one-to-one. What matters is whether the graph ever turns around. The plain cube never turns; this one rises, turns down, and turns up again, and any turn-around creates a horizontal line with more than one crossing. The next slide lets you watch that turn appear and disappear.

17. Slide c and apply the horizontal line test

Tweak it

Parameter explorer

The graph is y = x³ + cx. Slide c from negative to positive. For which values of c does the graph pass the horizontal line test?

\[ f(x) = x^3 + {c}x \]

  • c — from -3 to 3: coefficient c

\[ c < 0: \quad f\left(-\sqrt{-c}\right) = f(0) = f\left(\sqrt{-c}\right) = 0 \]

Start at the default, which is Checkpoint 1.23, and you will see the hump and the dip. Now slide c up toward zero and watch them shrink toward the origin.

At c equal to zero the curve is the plain cube, which flattens for an instant at the origin but never turns back. For every positive c it only rises. Those graphs all pass the horizontal line test.

For every negative c, however small, there is a turn. The line under the slider makes it precise: the curve crosses zero at zero and at plus and minus the square root of minus c, three different inputs. So the dividing line is sharp: one-to-one exactly when c is zero or positive. In Chapter 4 you will be able to find the turning points with a derivative instead of a slider.

18. One-to-one or not?

Sorting

Sort into buckets

Sort each function, on its natural domain, by whether it has an inverse.

One-to-one: an inverse exists
f(x) = 3x − 4; f(x) = x³ + 1; f(x) = √(x − 1); f(x) = 1/(x + 2)
Not one-to-one
f(x) = x²; f(x) = |x|; f(x) = sin x; f(x) = x⁴ − 1
yes
3x − 4 and x³ + 1 always rise; √(x − 1) rises on its domain; 1/(x + 2) is 1/x shifted left, which passes.
no
x², |x| and x⁴ − 1 give the same output at x and −x; sin x repeats every 2π, so sin 0 = sin π.

Sort each function before checking, and for every one you put in the second bucket, name the two inputs that share an output. That habit turns a guess into an argument.

The pattern among the functions that fail is symmetry or repetition. The square, the absolute value and the fourth power are all even: they give the same output at x and at minus x. Sine repeats itself every full turn. Any of those features guarantees a horizontal line with two crossings.

The functions that pass either always rise or are built from one-to-one pieces. The square root of x minus one looks as if it might share the square's problem, but its domain starts at one and it only ever rises there. One over x plus two is one over x shifted, and shifting does not change whether a horizontal line crosses more than once.

19. Is one-to-one the same as always increasing?

Counterexample

Discussion prompt

A function that always increases (or always decreases) is one-to-one. Is the converse true? Find a one-to-one function that neither always increases nor always decreases on its domain.

Think about this before revealing the answer. A function that always rises can never return to an earlier height, so it is automatically one-to-one, and the same goes for one that always falls. The question is whether the reverse is true.

One over x is the classic counterexample. On each branch it falls. But compare an input on the left branch with one on the right: from minus one to one the output jumps from minus one up to one. So it is not decreasing overall, and it is not increasing either, yet the earlier example proved it is one-to-one.

The gap at zero is what makes this possible. A function that is continuous on a whole interval and one-to-one there really must be always increasing or always decreasing, but a break in the domain lets it jump. Keep that in mind: monotone is a sufficient reason for one-to-one, not the only one.

20. Finding an inverse

Section

Part 3

21. The strategy: solve, then swap

Concept

  1. Solve the equation y = f(x) for x. Because f is one-to-one, there is exactly one solution.
  2. Interchange x and y and write y = f⁻¹(x), so the inverse uses x as its input like every other function.

\[ y = f(x) \;\xrightarrow{\text{solve}}\; x = g(y) \;\xrightarrow{\text{swap}}\; f^{-1}(x) = g(x) \]

The swap changes nothing about the function, only the name of its variable. It matters when you draw f and its inverse on the same axes.

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 69 — Problem-Solving Strategy: Finding an Inverse Function

The book's strategy has two steps, and the first is where all the work is. You solve y equals f of x for x. Because f is one-to-one, the equation has exactly one solution for each y in the range, so the solution is a function of y.

The second step is bookkeeping. By convention a function's input is called x, so you rename: swap the letters and write y equals f inverse of x. Nothing mathematical changes; the same rule has been given new variable names.

The reason for bothering with the swap is graphing. When f and its inverse are drawn on the same axes, both must use the horizontal axis for their input. That is what makes the reflection picture in Part 4 work. In applied problems, as you will see with temperature, it is often better not to swap, because the letters carry meaning.

22. Example 1.29: the inverse of 3x minus 4

Worked example

Find the inverse, state its domain and range, and verify it.

\[ f(x) = 3x - 4 \]

Write the output as y

Why: Start from the equation that defines f.

\[ y = 3x - 4 \]

Add 4 to both sides

Why: Undo the last thing f did first.

\[ 3x = y + 4 \]

Divide by 3

Why: Undo the multiplication.

\[ x = \frac13 y + \frac43 \]

Interchange x and y

Why: Rename the variables so the inverse takes x.

\[ f^{-1}(x) = \frac13 x + \frac43 \]

Swap the domain and range

Why: f is defined and takes every value on the whole real line.

\[ \text{domain of } f^{-1} = (-\infty, \infty), \quad \text{range of } f^{-1} = (-\infty, \infty) \]

Verify the first cancellation equation

Why: Substitute f into the inverse.

\[ f^{-1}\left(f(x)\right) = \frac13(3x - 4) + \frac43 = x - \frac43 + \frac43 = x \]

Check the second cancellation equation too

Why: An inverse must cancel in both orders.

\[ f\left(f^{-1}(x)\right) = 3\left(\frac13 x + \frac43\right) - 4 = x + 4 - 4 = x \]

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 69 — Example 1.29

Cover the words and read only the maths: each line undoes one operation of f. The function multiplied by three and then subtracted four, so you add four first and divide by three second, exactly reversing the order.

The domain and range need no work here. A line with nonzero slope is defined everywhere and reaches every height, so both f and its inverse have the whole real line as domain and range. Swapping them changes nothing.

The last two steps verify the two cancellation equations separately, and it is worth doing both. The book's note after this example makes the point: to be an inverse, both compositions must return x. For a line, if one works the other always does, but later in the lesson you will meet a pair where only one works, and checking both is how you catch it.

23. Checkpoint 1.24: a rational function

Worked example

Find the inverse and state its domain and range.

\[ f(x) = \frac{3x}{x - 2} \]

Multiply both sides by x minus 2

Why: Clear the fraction.

\[ y(x - 2) = 3x \]

Expand

Why: Distribute y.

\[ xy - 2y = 3x \]

Collect the x terms on one side

Why: Everything with x to the left, everything else to the right.

\[ xy - 3x = 2y \]

Factor out x

Why: Now x appears only once.

\[ x(y - 3) = 2y \]

Divide by y minus 3

Why: Allowed exactly when y is not 3.

\[ x = \frac{2y}{y - 3} \]

Interchange x and y

Why: The inverse, in standard form.

\[ f^{-1}(x) = \frac{2x}{x - 3} \]

Swap the domain and range

Why: f never reaches 3, since 3x over x minus 2 equals 3 plus 6 over x minus 2; and f is not defined at 2.

\[ \text{domain of } f^{-1}: x \ne 3, \qquad \text{range of } f^{-1}: y \ne 2 \]

Check with a number

Why: f sends 4 to 6; the inverse must send 6 back to 4.

\[ f(4) = \frac{12}{2} = 6, \qquad f^{-1}(6) = \frac{12}{3} = 4 \]

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 69 — Checkpoint 1.24

When x appears in both the numerator and the denominator, you cannot undo operations one at a time. The method instead is to clear the fraction, gather every term containing x on one side, and factor x out, so that it appears exactly once.

Watch the step where you divide by y minus three. It is only legal when y is not three, and that restriction is information: it says three is not in the range of f. You can see why by rewriting f as three plus six over x minus two. The second part is never zero, so f never equals three.

So the domain of the inverse excludes three, and its range excludes two, the number that was missing from the domain of f. The check uses one round trip with friendly numbers: f takes four to six, the inverse takes six back to four. A single round trip does not prove the formula, but a wrong formula almost always fails it.

24. Put the inverse-finding steps in order

Ranking

Put in order

Order the steps for finding the inverse of f(x) = 1/(x + 2), Exercise 194.

  1. Confirm f is one-to-one (it is 1/x shifted left 2)
  2. Write y = 1/(x + 2)
  3. Take reciprocals: x + 2 = 1/y
  4. Subtract 2: x = 1/y − 2
  5. Swap: f⁻¹(x) = 1/x − 2, with domain x ≠ 0 and range y ≠ −2

Why: Existence comes before algebra, because solving an equation that has two solutions would give a formula that is not a function. The domain and range come last, read off by swapping those of f.

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 77 — Exercise 194

Arrange the steps before checking. Most people can do the algebra in the middle; the steps that get forgotten are the first and the last.

The first step is to confirm the function is one-to-one. If you skip it and the function is not, the algebra will usually produce a plus or minus, or two solutions, and you might pick one without realising that no true inverse exists.

The last step, the domain and range, comes almost free by swapping. One over x plus two is never zero, so zero is missing from the domain of the inverse; and the original could not take the input minus two, so minus two is missing from the range of the inverse. You can confirm both straight from the formula one over x minus two.

25. Exercise 190: undo a cube root

Fill the middle

\[ f(x) = \sqrt[3]{x - 4} \]

Fill in the blanks

y = \sqrt[3]3 \;\Longrightarrow\; y^3 = x - 4 \;\Longrightarrow\; f^4(x) = x^___} } + ___

Why: Cube both sides to remove the cube root, add 4, then swap the variables. This is the warm-up function run in the other direction: the cube-plus-four function and the cube-root-of-x-minus-four function are inverses of each other, and both have the whole real line as domain and range.

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 77 — Exercise 190

Fill in the blanks before checking. The function takes a cube root last, so undoing starts by cubing.

Once both sides are cubed, adding four isolates x, and the swap gives the inverse. You have seen the answer already: it is the warm-up function. The cube-plus-four function and the cube-root-of-x-minus-four function undo each other, so each is the inverse of the other.

That symmetry is always true. If g is the inverse of f, then f is the inverse of g, because the two cancellation equations are the same pair of statements read in the other order. Also notice there are no domain restrictions here: cube roots, unlike square roots, accept every real number, and cubes take every real value.

26. Exercise 205: verifying a pair by composition

Worked example

Decide by composition whether these two functions are inverses.

\[ f(x) = x^2 + 2x + 1,\; x \ge -1, \qquad g(x) = -1 + \sqrt{x},\; x \ge 0 \]

Factor f

Why: It is a perfect square.

\[ f(x) = (x + 1)^2 \]

Compose f after g

Why: Put g into f.

\[ f\left(g(x)\right) = \left(-1 + \sqrt x + 1\right)^2 = \left(\sqrt x\right)^2 = x \]

Compose g after f

Why: Now put f into g.

\[ g\left(f(x)\right) = -1 + \sqrt{(x+1)^2} = -1 + |x + 1| \]

Use the restriction on f

Why: Because x is at least minus 1, the quantity inside the absolute value is not negative.

\[ -1 + |x + 1| = -1 + (x + 1) = x \]

Check with a number

Why: Both cancellations hold, so the two are inverses. At x equal to 3: f gives 16, and g sends 16 back to 3.

\[ f(3) = 16, \qquad g(16) = -1 + 4 = 3 \]

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 78 — Exercise 205

Sometimes you are handed a candidate inverse and asked only to confirm it. The tool is composition in both orders, and this pair shows why both orders matter.

The first composition is clean: factoring f as x plus one squared, the minus one and plus one cancel and the square undoes the root. The second composition is where the care is needed. The square root of x plus one, all squared, is not x plus one in general; it is the absolute value of x plus one.

The restriction on f rescues it. Because x is at least minus one, x plus one is zero or positive, and its absolute value is itself. Without that restriction, the second composition would fail for every x below minus one. The numerical check at three confirms the whole thing: sixteen out, three back.

27. Find the error: only one cancellation checked

Error analysis

Annotate

On: \( f(x) = x^2,\; g(x) = \sqrt{x}: \quad f\left(g(x)\right) = x \;\Longrightarrow\; g = f^{-1} \)

  • True for every x at least 0: the square of a square root gives back the number.
  • Never checked. g(f(−3)) is the root of 9, which is 3, not −3. The second cancellation fails.
  • The squaring function on all real numbers is not one-to-one, so it has no inverse at all. The square root is the inverse only of x² restricted to x ≥ 0.

Click through the notes. The first step of the argument is correct, and that is what makes the error easy to miss.

Squaring a square root really does give back the number, for every number the square root accepts. So one cancellation equation holds perfectly. But the other order starts by squaring, and squaring throws away the sign. Minus three squared is nine, and the root of nine is three, not minus three.

The deeper problem is that the claimed inverse was never possible: the squaring function on all real numbers fails the horizontal line test, so nothing can be its inverse. The square root is the inverse of the squaring function restricted to x at least zero. That is exactly why the definition demands both equations, each on its own set.

28. Inverse pair or not?

Discrimination

Sort into buckets

Exercises 199 to 203: compose each pair and sort it.

Inverses: both compositions give x
f = 8x, g = x/8; f = 8x + 3, g = (x − 3)/8; f = 1/(x − 1), g = 1/x + 1
Not inverses
f = 5x − 7, g = (x + 5)/7; f = (2/3)x + 2, g = (3/2)x + 3
yes
8(x/8) = x and (8x)/8 = x; 8·(x − 3)/8 + 3 = x and ((8x + 3) − 3)/8 = x; for the last pair f(g(x)) = 1/((1/x + 1) − 1) = x and g(f(x)) = (x − 1) + 1 = x.
no
5·(x + 5)/7 − 7 = (5x − 24)/7, not x: the inverse of 5x − 7 is (x + 7)/5. And (2/3)((3/2)x + 3) + 2 = x + 4, not x: the inverse is (3/2)x − 3.

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 78 — Exercises 199 to 203

These five pairs come straight from the exercise set. Compose each pair at least one way before sorting, and for any pair that fails, work out what the correct inverse should have been.

The two failures are classic slips. For five x minus seven, the partner divides by seven and adds five, mixing up which number multiplies and which is added. The correct inverse adds seven and then divides by five. For two thirds of x plus two, the partner flips the fraction but keeps the plus sign on the constant, and the composition leaves x plus four behind.

The last pair is the least obvious and it passes. One over x minus one, and one over x plus one, do not look like mirror images, but composing them in either order cancels cleanly. When your intuition is unsure, composition is the judge.

29. Complete the inverse table

Comparison

Comparison matrix

f(x)one-to-one? (yes/no)f⁻¹(x)domain of f⁻¹
∛(x − 4)yesx³ + 4all reals
x² − 4, x ≥ 0yes√(x + 4)[−4, ∞)
x² − 4, all xnononenone
x³ + 1yes∛(x − 1)all reals
√(x − 1)yesx² + 1[0, ∞)
sin x, all xnonone until restrictednone

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 77 — Exercises 189 to 193

Fill in the blanks by typing yes, no or all reals. The table is a summary of the moves you now know, applied to five of the book's exercises and one warning case.

Look at the two rows for x squared minus four. The formula is the same; only the domain differs. Restricted to x at least zero it is one-to-one and has the square root of x plus four as its inverse. Allowed all real x, it has no inverse at all. The domain is part of the function, not a footnote.

The cube and cube-root rows have the whole real line everywhere, while the square-root rows carry restrictions, because square roots only accept numbers that are not negative and only return numbers that are not negative. And sine has no inverse on its whole domain; Part 6 fixes that by restricting it, exactly as you restrict the parabola.

30. The graph of an inverse

Section

Part 4

31. Swap the coordinates, reflect the graph

Concept

Figure (svg): Equal-scale axes. The curve y equals 2 to the x in cyan and its inverse in orange, mirror images across the dashed line y equals x. The point (1, 2) on the first curve is joined by a short dashed segment to the point (2, 1) on the second.

Every point (a, b) on f has a partner (b, a) on the inverse. Swapping coordinates is exactly a reflection in the diagonal.

\[ (a, b) \text{ on } f \iff b = f(a) \iff f^{-1}(b) = a \iff (b, a) \text{ on } f^{-1} \]

Every point of the inverse's graph is a point of f with its coordinates swapped, and swapping coordinates reflects the plane in the diagonal line.

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, pp. 69-70 — Graphing Inverse Functions and Figure 1.39

The chain of equivalences on the slide is the whole argument. A point a, b is on the graph of f exactly when f of a is b, which by definition happens exactly when the inverse sends b to a, which says the point b, a is on the graph of the inverse.

In the picture, the point one, two sits on the exponential curve because two to the first is two. Its partner two, one sits on the inverse. The dashed segment between them crosses the diagonal line at a right angle, and the two curves are mirror images of each other across that diagonal.

This gives you a way to draw the inverse of any one-to-one function without a formula. Sketch the diagonal, reflect a few key points by swapping their coordinates, and join them. Here the inverse of two to the x has no formula you know yet; Section 1.5 names it the logarithm base two.

32. Why the mirror is the diagonal

Picture it

Figure (svg): Equal-scale axes from 0 to 5 with the dashed diagonal y equals x. The points (1, 4) and (4, 1) are joined by a segment that crosses the diagonal at a right angle at its midpoint (2.5, 2.5), marked with a small square.

The segment from (a, b) to (b, a) has slope minus 1, perpendicular to the diagonal, and its midpoint has equal coordinates. That is the definition of a mirror image.

\[ \text{slope} = \frac{a - b}{b - a} = -1, \qquad \text{midpoint} = \left(\tfrac{a+b}{2}, \tfrac{a+b}{2}\right) \]

The segment joining a point to its swapped partner is perpendicular to the diagonal, and the diagonal cuts it in half. Those two facts are what a reflection means.

It is worth seeing why swapping coordinates is a reflection, rather than taking it on trust. Take any point a, b and its partner b, a.

The slope of the segment joining them is a minus b over b minus a, which is minus one. The diagonal has slope one, and slopes whose product is minus one belong to perpendicular lines. So the segment meets the diagonal at a right angle, as the small square in the figure shows.

The midpoint has both coordinates equal to a plus b over two, so it lies on the diagonal. A segment that is perpendicular to a line and bisected by it joins a point to its mirror image. That is precisely what reflection in a line means, and the argument works for every point at once.

33. Example 1.30: reflecting a square root

Worked example

Sketch the inverse by symmetry, give its domain and range, then confirm the formula.

\[ f(x) = \sqrt{x + 2}, \quad x \ge -2 \]

Record the domain and range of f

Why: The root needs x plus 2 not negative, and a root is never negative.

\[ \text{domain } [-2, \infty), \qquad \text{range } [0, \infty) \]

Swap them for the inverse

Why: Domain and range trade places.

\[ \text{domain of } f^{-1} = [0, \infty), \qquad \text{range of } f^{-1} = [-2, \infty) \]

Solve for x

Why: Square both sides; y is not negative, so nothing is lost.

\[ y^2 = x + 2 \;\Longrightarrow\; x = y^2 - 2 \]

Interchange x and y

Why: Keep the restriction that came from the range of f.

\[ f^{-1}(x) = x^2 - 2, \quad x \ge 0 \]

Figure (svg): Equal-scale axes. The curve y equals the square root of x plus 2 starts at (minus 2, 0) and rises slowly; its inverse y equals x squared minus 2, for x at least 0, starts at (0, minus 2) and rises steeply. They are mirror images across the dashed line y equals x and meet at (2, 2).

The starting point (−2, 0) of f reflects to the starting point (0, −2) of the inverse, which is why the inverse lives only on x ≥ 0.

Check with a point and its mirror

Why: f sends 7 to 3, so the inverse must send 3 to 7; and (2, 2) lies on the diagonal and on both graphs.

\[ f(7) = \sqrt 9 = 3, \quad f^{-1}(3) = 9 - 2 = 7 \]

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 70 — Example 1.30

The book's instruction is to sketch the inverse by symmetry. Do that first: draw the diagonal, reflect the starting point minus two, zero to zero, minus two, and reflect a couple of other points. The picture already tells you the inverse is half of a parabola.

The domain and range come from swapping, before any algebra. The inverse accepts only numbers from zero up, because those are the only outputs a square root ever produces. That restriction is essential: the formula x squared minus two on its own would describe a whole parabola, which is not one-to-one.

When you solve by squaring, you might worry about losing information, since squaring is exactly the operation that was not one-to-one. Here nothing is lost, because y is known to be zero or positive. The check confirms a round trip at seven and three, and the point two, two lies on the diagonal, so it is on both graphs.

34. Checkpoint 1.25: a line and its mirror

Worked example

\[ f(x) = 2x + 3 \]

Solve for x

Why: Subtract 3 and divide by 2.

\[ y = 2x + 3 \;\Longrightarrow\; x = \frac{y - 3}{2} \]

Interchange x and y

Why: The inverse is again a line.

\[ f^{-1}(x) = \frac{x - 3}{2} \]

Reflect the intercepts

Why: The y-intercept of f becomes the x-intercept of the inverse.

\[ (0, 3) \text{ on } f \;\longleftrightarrow\; (3, 0) \text{ on } f^{-1} \]

Figure (svg): Equal-scale axes. The line y equals 2x plus 3 and its inverse y equals x minus 3 over 2, mirror images across the dashed diagonal. The intercepts (0, 3) and (3, 0) are marked as partners, and the two lines cross at (minus 3, minus 3) on the diagonal.

Steepness 2 becomes steepness one half: reflecting in the diagonal swaps rise and run. The crossing point lies on the mirror because it is its own partner.

Find where they meet

Why: A point on the diagonal is its own mirror image.

\[ 2x + 3 = x \;\Longrightarrow\; x = -3 \]

Check both lines pass through it

Why: Substitute minus 3 into each.

\[ f(-3) = -6 + 3 = -3, \qquad f^{-1}(-3) = \frac{-6}{2} = -3 \]

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 71 — Checkpoint 1.25

For a line, the algebra is quick: subtract three, divide by two, swap. The interesting part is what the picture shows.

Reflection swaps the roles of the axes, so the y-intercept three of the original becomes the x-intercept three of the inverse. It also swaps rise and run, so a slope of two becomes a slope of one half. That relationship between slopes of a function and its inverse will return in Chapter 3, where the derivative of an inverse is one over the derivative of the function.

The two lines cross at minus three, minus three. A point on the diagonal has equal coordinates, so swapping them leaves it where it is, and it must lie on both graphs. Solving the original equal to x finds it, and substituting into both formulas checks it.

35. Must a function and its inverse meet on the diagonal?

Prediction

\[ f(x) = -x^3, \qquad f^{-1}(x) = -\sqrt[3]{x} \]

Predict first

Every crossing in the examples so far lay on y = x. Do these two graphs cross anywhere OFF the diagonal?

  • No, never: crossings are always on y = x
  • Yes: at (1, −1) and (−1, 1)
  • They never cross at all
  • Only at the origin

Correct: Yes: at (1, −1) and (−1, 1)

Why: Minus the cube of 1 is minus 1, and minus the cube root of 1 is also minus 1, so the point (1, minus 1) lies on both graphs, and by symmetry so does its mirror (minus 1, 1). For a decreasing function the graph and its reflection can cross off the diagonal. For an increasing one, crossings do sit on the diagonal.

Predict before revealing. Every crossing so far was on the diagonal, and it is easy to conclude that this is always so.

Test the point one, minus one. Minus the cube of one is minus one, so it is on the graph of f. Minus the cube root of one is also minus one, so it is on the graph of the inverse too. The two graphs cross there, and it is not on the diagonal. By the mirror symmetry, minus one, one is another crossing.

The difference is that this function is decreasing. For an increasing function, the graph and its reflection can only meet on the diagonal, which is why the earlier examples all did. A decreasing function's graph crosses the diagonal from above to below, and its mirror image can cross it at partner points on either side. So when you solve for crossings, do not simply set f of x equal to x unless f is increasing.

36. Restricting a domain

Section

Part 5

37. Half a parabola is invertible

Concept

Figure (svg): Two panels on equal-scale axes. Left: the right half of the parabola, x at least 0, drawn solid, with its inverse, the square root of x, reflected across the dashed diagonal. Right: the left half of the parabola, x at most 0, with its inverse, minus the square root of x, below the x-axis. The unused half of the parabola is dotted in each panel.

The same parabola, cut in half two different ways, gives two different inverses. The choice of half is a decision, and the inverse's formula records it through the sign of the root.

\[ g(x) = x^2,\; x \ge 0 \;\Longrightarrow\; g^{-1}(x) = \sqrt{x} \]

\[ h(x) = x^2,\; x \le 0 \;\Longrightarrow\; h^{-1}(x) = -\sqrt{x} \]

Restricted domain — A subset of the domain on which the function is one-to-one. The restricted function is a new function, and it has an inverse.

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 71 — Restricting Domains and Figure 1.40

The squaring function fails the horizontal line test, but each half of it passes. Keep only the right half, x from zero up, and you have a new function with a perfectly good inverse, the square root. Keep only the left half, and the inverse is minus the square root.

Notice that these are genuinely different functions, even though they share a formula. The book gives them different names, g and h, to make that point. A function is a formula together with a domain, and changing the domain changes the function.

In each panel of the figure, the half you kept is solid, the half you threw away is dotted, and the inverse is the reflection of the solid half only. The sign in front of the root is how the formula remembers which half was kept. Get the sign wrong and your inverse returns inputs from the wrong half.

38. Example 1.31: a shifted parabola

Worked example

\[ f(x) = (x + 1)^2 \]

Show f is not one-to-one

Why: The graph is y equals x squared shifted left 1; two inputs equally far from minus 1 share an output.

\[ f(-3) = 4 = f(1) \]

Restrict to the right half

Why: On x at least minus 1 the parabola only rises, so f is one-to-one there.

\[ f(x) = (x+1)^2, \quad x \ge -1 \]

Swap the domain and range

Why: On that half the outputs are all numbers from 0 up.

\[ \text{domain of } f^{-1} = [0, \infty), \qquad \text{range of } f^{-1} = [-1, \infty) \]

Take square roots

Why: Both signs are possible algebraically.

\[ y = (x+1)^2 \;\Longrightarrow\; x + 1 = \pm\sqrt{y} \]

Choose the sign from the restriction

Why: x at least minus 1 means x plus 1 is not negative, so the plus sign.

\[ x + 1 = \sqrt y \;\Longrightarrow\; x = -1 + \sqrt{y} \]

Interchange x and y

Why: The inverse on the restricted domain.

\[ f^{-1}(x) = -1 + \sqrt{x} \]

Figure (svg): Equal-scale axes. The parabola y equals x plus 1 squared, with its left half dotted and its right half, x at least minus 1, solid. Its inverse, minus 1 plus the square root of x, starts at (0, minus 1) and rises slowly. The vertex (minus 1, 0) and its partner (0, minus 1) are marked.

Keeping the half with x ≥ −1 means the inverse's outputs are at least −1. The plus sign in front of the root is what guarantees that.

Check with a round trip

Why: Start inside the restricted domain.

\[ f(2) = 9, \qquad f^{-1}(9) = -1 + 3 = 2 \]

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, pp. 71-72 — Example 1.31

The first step proves that f is not one-to-one with a specific pair: minus three and one are both two units from the vertex at minus one, so they give the same output four. That is the algebraic version of a horizontal line crossing twice.

Restricting to x at least minus one keeps the right half of the parabola, where it only rises. On that half, the outputs are every number from zero up, and that becomes the domain of the inverse.

The step that deserves the most attention is choosing the sign. Taking square roots honestly gives plus or minus. The restriction says x plus one is not negative, and only the positive root can equal something that is not negative. So the plus sign stays and the minus sign goes. The figure shows the result: the inverse starts at zero, minus one, the reflection of the vertex, and its outputs stay at minus one or above, exactly the restricted domain of f.

39. Before solving: which sign?

Step zero

\[ f(x) = (x - 1)^2, \quad x \le 1 \]

Discussion prompt

Exercise 192. Solving y = (x − 1)² gives x − 1 = plus or minus the root of y. Before you write anything else, which sign is correct, and how do you know?

Answer this before doing any algebra. Solving always produces two candidate signs, and the whole difficulty of the exercise is choosing between them.

The rule is to look at the restriction, not at the formula. The restriction says x is at most one, so x minus one is zero or negative. The square root of y is never negative, so the only way to write x minus one as a root is with a minus sign in front.

This is the reverse of Example 1.31, where the right half was kept and the plus sign won. It is worth saying out loud: right half, plus sign; left half, minus sign. Once you see that the sign is fixed by the restriction, you will never have to guess.

40. Exercise 192: the left half of a parabola

Worked example

\[ f(x) = (x - 1)^2, \quad x \le 1 \]

Find the range of f

Why: The left half falls from infinity down to the vertex value 0.

\[ \text{range of } f = [0, \infty) \]

Solve with the correct sign

Why: From the previous slide: x minus 1 is not positive.

\[ x - 1 = -\sqrt{y} \;\Longrightarrow\; x = 1 - \sqrt{y} \]

Interchange x and y

Why: Write the inverse in x.

\[ f^{-1}(x) = 1 - \sqrt{x} \]

Swap the domain and range

Why: The domain of f was x at most 1.

\[ \text{domain of } f^{-1} = [0, \infty), \qquad \text{range of } f^{-1} = (-\infty, 1] \]

Check with a round trip

Why: Minus 1 is in the restricted domain; the plus sign would have returned 3 instead.

\[ f(-1) = (-2)^2 = 4, \qquad f^{-1}(4) = 1 - 2 = -1 \]

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 77 — Exercise 192

With the sign settled on the previous slide, the rest is routine, but read the range step carefully. The left half of this parabola comes down from very large values to zero at the vertex, so its outputs are every number from zero up.

The inverse, one minus the square root of x, starts at one when x is zero and falls from there. Its outputs are one and everything below, which is exactly the restricted domain of f, as the swap predicts.

The round trip check is where the sign pays off. Starting at minus one, f gives four. The inverse with the correct sign sends four to one minus two, which is minus one: back where you started. If you had used the plus sign, you would get one plus two, which is three, a number that was never in the restricted domain. A single round trip catches a wrong sign every time.

41. Checkpoint 1.26: one over x squared, for negative x

Worked example

\[ f(x) = \frac{1}{x^2}, \quad x < 0 \]

Show f is one-to-one on this domain

Why: Suppose two negative inputs give the same output.

\[ \frac{1}{x_1^2} = \frac{1}{x_2^2} \;\Longrightarrow\; x_1^2 = x_2^2 \;\Longrightarrow\; x_1 = \pm x_2 \]

Rule out the minus sign

Why: Both inputs are negative, so they cannot be opposites.

\[ x_1 = x_2 \]

Find the range of f

Why: For negative x, x squared takes every positive value, and so does its reciprocal.

\[ \text{range of } f = (0, \infty) \]

Solve for x, choosing the negative root

Why: The domain says x is negative.

\[ x^2 = \frac1y \;\Longrightarrow\; x = -\frac{1}{\sqrt{y}} \]

Interchange and state the domain and range

Why: They are the range and domain of f.

\[ f^{-1}(x) = -\frac{1}{\sqrt{x}}, \quad \text{domain } (0, \infty), \quad \text{range } (-\infty, 0) \]

Figure (svg): Equal-scale axes. The left branch of y equals one over x squared, for negative x, drawn solid in the second quadrant; the right branch is dotted. Its inverse, minus one over the square root of x, lies in the fourth quadrant, a mirror image across the dashed diagonal. The partner points (minus 2, one quarter) and (one quarter, minus 2) are marked.

The inputs of f were negative, so the outputs of the inverse must be negative: that is where the minus sign in front of the root comes from.

Check with a round trip

Why: Start at minus 2.

\[ f(-2) = \frac14, \qquad f^{-1}\left(\tfrac14\right) = -\frac{1}{1/2} = -2 \]

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 72 — Checkpoint 1.26

One over x squared is even, so on its whole domain it fails the horizontal line test: minus two and two give the same output. Restricted to negative x, it passes, and the first two steps prove it: equal outputs force equal squares, so the inputs are equal or opposite, and two negative numbers cannot be opposites.

For the range, think about what happens as x runs over the negatives. Near zero, one over x squared is huge; far out, it is tiny; and it takes every positive value in between. That becomes the domain of the inverse.

The negative root appears for the same reason as in the last exercise: the answer must be a negative number, because the inputs of f were negative. In the figure, the solid branch in the upper left reflects to the branch in the lower right. The round trip at minus two confirms it: one quarter out, minus two back.

42. Trap: the root of a square is not the number

Trap

The trap

Undoing a square with a square root, without looking at signs:

\[ \sqrt{x^2} = x \]

Wrong for every negative x.

The fix

The root is never negative, so it returns the size of x, not x itself. At x equal to minus 3 the left side is 3.

\[ \sqrt{x^2} = |x| \]

The square root undoes squaring only on the half x at least 0, the restriction it was built for.

This slip is behind many of the sign errors in this part of the lesson. It looks like the square root simply cancels the square, and for zero and positive numbers it does.

For a negative number it cannot. Square minus three and you get nine; the square root of nine is three, because the square root symbol always means the root that is not negative. The sign of the original number has been lost, and the root cannot recover it. What survives is the size of the number, its absolute value.

So the square root undoes squaring only on the right half of the number line, the half it was designed to invert. Whenever a solution involves the square root of something squared, write the absolute value first, then use the restriction in the problem to remove the bars with the correct sign, as Exercise 205 did.

43. The inverse trigonometric functions

Section

Part 6

44. Restrict sine, then reflect it

Concept

Figure (svg): Equal-scale axes. The sine curve drawn dotted over a wide interval, with the piece from minus pi over 2 to pi over 2 drawn solid in cyan. Its reflection across the dashed diagonal, the inverse sine, runs from (minus 1, minus pi over 2) to (1, pi over 2) in orange.

Only the solid piece of sine gets inverted. It rises once from −1 to 1, so every value between them is reached exactly once.

Sine repeats, so it fails the horizontal line test badly. On the interval from minus a quarter turn to plus a quarter turn it rises once through every value from minus 1 to 1. That is the standard restriction.

\[ \sin^{-1}(x) = y \quad\text{if and only if}\quad \sin y = x \text{ and } -\frac{\pi}{2} \le y \le \frac{\pi}{2} \]

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, pp. 72-73 — Inverse Trigonometric Functions and equation 1.12

Sine is about as far from one-to-one as a function can be: every value between minus one and one is taken infinitely often. So there is no inverse on the whole real line, and the fix is the one you just practised: restrict the domain.

There are infinitely many intervals where sine is one-to-one. The conventional choice is from minus a quarter turn to plus a quarter turn. On that interval sine rises steadily from minus one to one, it passes through zero at zero, and it includes the first-quadrant angles you know best.

The figure shows the restricted piece solid, the rest of sine dotted, and the reflection of the solid piece, the inverse sine. Its domain is from minus one to one, and its outputs lie between minus and plus a quarter turn. That output interval is part of the definition, and every evaluation of an inverse sine must respect it.

45. The six conventional ranges

Concept

functionalso writtendomainrange: the angle it returns
sin⁻¹ xarcsin x[−1, 1][−π/2, π/2]
cos⁻¹ xarccos x[−1, 1][0, π]
tan⁻¹ xarctan xall real x(−π/2, π/2)
cot⁻¹ xarccot xall real x(0, π)
csc⁻¹ xarccsc x|x| ≥ 1[−π/2, π/2], not 0
sec⁻¹ xarcsec x|x| ≥ 1[0, π], not π/2

Each range is an interval of angles, containing the first quadrant, on which the original function is one-to-one and takes every value it ever takes.

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 73 — equations 1.12 to 1.14

Each inverse trigonometric function has a range fixed by convention, and the table collects them. You should know the first three without looking, because they are the ones calculus uses constantly.

Notice the patterns. Sine, tangent, cosecant use an interval centred on zero, from minus a quarter turn to plus a quarter turn. Cosine, cotangent, secant use the upper half turn, from zero to a half turn, because cosine is not one-to-one on an interval centred at zero: it is even.

The endpoints matter. Tangent is undefined at plus and minus a quarter turn, so the inverse tangent's range is open and never reaches them. The cosecant and secant ranges each exclude one angle where the original function is undefined. Other books sometimes choose different ranges for the inverse secant and cosecant, so when those appear, check which convention is being used.

46. Reading the inverse cosine definition

Notation

Annotate

On: \( \cos^{-1}(x) = y \iff \cos y = x \text{ and } 0 \le y \le \pi \)

  • An angle. Read it as 'the angle whose cosine is x'. Not one over cos x, which is sec x.
  • The equation every candidate angle must satisfy. It has infinitely many solutions.
  • The tie-breaker: the upper half of the unit circle, where cosine runs once from 1 down to −1.
  • x must be between −1 and 1, because no angle has a cosine outside that interval.

\[ \cos^{-1}\left(\tfrac12\right) = \frac{\pi}{3} \quad\text{because}\quad \cos\frac{\pi}{3} = \frac12 \text{ and } 0 \le \frac{\pi}{3} \le \pi \]

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 74 — evaluating the inverse cosine of one half

Click through the notes. The key habit is in the first one: read the inverse cosine of x as the angle whose cosine is x. The output is an angle, not a number to be multiplied.

The definition has two conditions. The equation cosine of y equals x has infinitely many solutions, one family in each full turn. The second condition, y between zero and a half turn, picks exactly one of them.

The worked line at the bottom is the book's own first evaluation. Many angles have cosine one half, including minus a sixth of a half turn and five thirds of a half turn. Only a third of a half turn, pi over three, lies in the upper half circle, so that is the answer. Every inverse trigonometric evaluation in this lesson follows the same two-part check: right value, right range.

47. Three more inverse graphs

Picture it

Figure (svg): Three panels. Left: the inverse cosine, falling from (minus 1, pi) to (1, 0). Middle: the inverse tangent, rising across all x and trapped between dashed horizontal asymptotes at minus pi over 2 and pi over 2. Right: the inverse secant, two pieces for x at most minus 1 and at least 1, with a dashed asymptote at pi over 2.

Each output range is a single interval of angles on which the original trig function was one-to-one. The horizontal asymptotes of the inverse tangent are the vertical asymptotes of tangent, reflected.

Each graph is the restricted piece of a trig function, reflected in the diagonal. Notice the asymptotes of the inverse tangent: its outputs approach, but never reach, a quarter turn in either direction.

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 73 — Figure 1.41

These three panels complete the picture from the book's Figure 1.41, alongside the inverse sine you have already seen. Each is the solid restricted piece of a trigonometric function, reflected in the diagonal.

The inverse cosine falls from a half turn at minus one to zero at one, because cosine on the upper half turn falls from one to minus one. The inverse tangent is defined for every real number and levels off at plus and minus a quarter turn. Those horizontal asymptotes are the reflections of tangent's vertical asymptotes.

The inverse secant has two separate pieces, because secant is never between minus one and one. Its outputs avoid a quarter turn, where secant is undefined, and approach it from both sides. You will use the inverse tangent most in calculus: it appears as the antiderivative of one over one plus x squared in Chapter 5.

48. Example 1.32(a): inverse sine of minus root 3 over 2

Worked example

\[ \sin^{-1}\left(-\frac{\sqrt3}{2}\right) \]

Translate into an angle question

Why: Find the angle with this sine that lies in the allowed range.

\[ \sin\theta = -\frac{\sqrt3}{2}, \qquad -\frac{\pi}{2} \le \theta \le \frac{\pi}{2} \]

Find the reference angle

Why: The sine of a third of a half turn is root 3 over 2.

\[ \sin\frac{\pi}{3} = \frac{\sqrt3}{2} \]

List the angles with the negative sine

Why: Sine is negative below the x-axis: the third and fourth quadrants.

\[ \theta = -\frac{\pi}{3} \quad\text{or}\quad \theta = \frac{4\pi}{3} \]

Figure (svg): A unit circle with its right half, angles from minus pi over 2 to pi over 2, drawn thick in green. A dashed horizontal line at height minus root 3 over 2 meets the circle twice: at angle minus pi over 3 on the green arc, marked with a solid dot, and at angle 4 pi over 3 on the left, marked with a hollow red dot.

Two angles on the circle have sine −√3/2. Only one lies on the green arc, the range the inverse sine is allowed to answer from.

Keep the one in the range

Why: Only minus pi over 3 is on the allowed arc.

\[ \sin^{-1}\left(-\frac{\sqrt3}{2}\right) = -\frac{\pi}{3} \]

Check both conditions

Why: Sine is odd, and the angle is inside the interval.

\[ \sin\left(-\frac{\pi}{3}\right) = -\sin\frac{\pi}{3} = -\frac{\sqrt3}{2}, \quad -\frac{\pi}{2} \le -\frac{\pi}{3} \le \frac{\pi}{2} \]

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 75 — Example 1.32a

Every evaluation has the same shape. First rewrite it as an angle question with both conditions: the sine must be minus root three over two, and the angle must lie between minus and plus a quarter turn.

Then find the angle in two stages. The reference angle, ignoring the sign, is a third of a half turn, from the triangle you know. Sine is negative in the third and fourth quadrants, which gives two natural candidates on the circle.

The unit-circle picture makes the choice visible. The green arc is the allowed range. The candidate at minus pi over three lies on it; the one at four pi over three does not. The final check confirms both conditions, and it is worth writing both, because an angle with the right sine but in the wrong place is the most common wrong answer.

49. Example 1.32(b): tangent of an inverse tangent

Worked example

\[ \tan\left(\tan^{-1}\left(-\frac{1}{\sqrt3}\right)\right) \]

Evaluate the inner inverse tangent

Why: The tangent of minus pi over 6 is minus one over root 3, and that angle is in the open interval of a quarter turn either side of zero.

\[ \tan^{-1}\left(-\frac{1}{\sqrt3}\right) = -\frac{\pi}{6} \]

Apply tangent

Why: Tangent of that angle is the number you started with.

\[ \tan\left(-\frac{\pi}{6}\right) = -\frac{1}{\sqrt3} \]

Check against the cancellation rule

Why: The inverse tangent accepts every real number, so tangent of inverse tangent returns its input for every y, with no exceptions.

\[ \tan\left(\tan^{-1} y\right) = y \text{ for all real } y \;\Longrightarrow\; -\frac{1}{\sqrt3} \checkmark \]

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 75 — Example 1.32b

Work from the inside out. The inverse tangent of minus one over root three asks for the angle, strictly between minus and plus a quarter turn, whose tangent is that number. Tangent of a sixth of a half turn is one over root three, and tangent is odd, so the angle is minus pi over six.

Applying tangent to that angle gives back minus one over root three, the number you started with. That was always going to happen.

The check explains why. The inverse tangent accepts every real number, and for any input it returns an angle whose tangent is that input. So tangent of inverse tangent of y equals y for every y. When the outer function is the original and the inner one is its inverse, the composition cancels on the whole domain of the inverse. The other order is a different story, and it is the next slide.

50. When the cancellation fails

Concept

\[ \sin\left(\sin^{-1}\frac{\sqrt2}{2}\right) = \sin\frac{\pi}{4} = \frac{\sqrt2}{2} \]

\[ \sin^{-1}\left(\sin \pi\right) = \sin^{-1}(0) = 0 \ne \pi \]

Figure (svg): The graph of y equals inverse sine of sine x for x from minus 2 pi to 2 pi: a zigzag of straight segments between heights minus pi over 2 and pi over 2. It coincides with the dashed line y equals x only between minus pi over 2 and pi over 2. The point (pi, 0) is marked.

Inside the shaded strip the inverse sine really does undo sine. Outside it, the output is folded back into the strip, so at x = π you get 0.

The inverse sine inverts only the restricted sine, so it can only return angles in the restricted interval. Starting outside the interval, the round trip lands somewhere else.

\[ \sin^{-1}(\sin x) = x \text{ only if } -\tfrac{\pi}{2} \le x \le \tfrac{\pi}{2}, \qquad \sin(\sin^{-1} y) = y \text{ for } -1 \le y \le 1 \]

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 74 — compositions of trigonometric functions and their inverses

The first line behaves: the inverse sine of root two over two is pi over four, and the sine of pi over four is root two over two. The second line is the surprise. The sine of pi is zero, and the inverse sine of zero is zero, not pi.

Nothing has gone wrong with the arithmetic. The inverse sine is the inverse of the restricted sine only, so it can only ever return angles between minus and plus a quarter turn. Pi is outside that interval, so no matter what, the inverse sine cannot give it back.

The zigzag graph shows the whole story. Inside the shaded strip, the inverse sine of the sine of x is x, and the graph lies on the diagonal. Outside, every angle is replaced by the one angle inside the strip with the same sine, so the graph folds back and forth. The summary line gives the two rules: one order cancels on the whole domain of the inverse, the other only on the restricted interval.

51. Example 1.32(c): inverse cosine of cosine of 5 pi over 4

Worked example

\[ \cos^{-1}\left(\cos\frac{5\pi}{4}\right) \]

Evaluate the inner cosine

Why: Five quarter-turns-of-a-half lands in the third quadrant, where cosine is negative.

\[ \cos\frac{5\pi}{4} = -\frac{\sqrt2}{2} \]

Ask the inverse cosine question

Why: The answer must lie in the upper half of the circle.

\[ \cos\theta = -\frac{\sqrt2}{2}, \qquad 0 \le \theta \le \pi \]

Figure (svg): A unit circle with its upper half, angles from 0 to pi, drawn thick in green. A dashed vertical line at x equals minus root 2 over 2 meets the circle twice: at angle 3 pi over 4 on the green arc, marked with a solid dot, and at 5 pi over 4 below, marked with a hollow red dot.

The cosine of 5π/4 and of 3π/4 is the same number. The inverse cosine must answer from the upper half, so it returns 3π/4, not the angle you started with.

Choose the angle in range

Why: The second-quadrant angle with the same cosine.

\[ \theta = \frac{3\pi}{4} \]

Check the answer and why it differs

Why: Its cosine is right and it lies in the range; the input 5 pi over 4 lies outside the range, so no cancellation was due.

\[ \cos\frac{3\pi}{4} = -\frac{\sqrt2}{2}, \quad 0 \le \frac{3\pi}{4} \le \pi, \quad \frac{5\pi}{4} > \pi \]

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 75 — Example 1.32c

If the cancellation always worked, the answer would be five pi over four. That is the answer the question is designed to catch.

Compute the inner cosine honestly. Five pi over four is in the third quadrant, where cosine is negative, so its cosine is minus root two over two. Now ask the inverse cosine question: which angle in the upper half circle has that cosine?

The figure shows the vertical line at minus root two over two meeting the circle twice. The lower point is where you started; the upper point, three pi over four, is on the allowed arc. The check confirms the cosine and the range, and states the reason for the mismatch: five pi over four is bigger than pi, outside the range, so the inverse cosine was never able to return it.

52. Example 1.32(d): inverse sine of a cosine

Worked example

\[ \sin^{-1}\left(\cos\frac{2\pi}{3}\right) \]

Evaluate the cosine

Why: Two thirds of a half turn is in the second quadrant.

\[ \cos\frac{2\pi}{3} = -\frac12 \]

Ask the inverse sine question

Why: An angle with sine minus one half, between minus and plus a quarter turn.

\[ \sin\theta = -\frac12, \qquad -\frac{\pi}{2} \le \theta \le \frac{\pi}{2} \]

Answer it

Why: Sine of pi over 6 is one half, and sine is odd.

\[ \theta = -\frac{\pi}{6} \]

Check both conditions

Why: The sine is right and the angle is in range.

\[ \sin\left(-\frac{\pi}{6}\right) = -\frac12, \quad -\frac{\pi}{2} \le -\frac{\pi}{6} \le \frac{\pi}{2} \]

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 75 — Example 1.32d

This one mixes two different functions, so no cancellation is even tempting. Work from the inside out, as always.

The cosine of two pi over three is minus one half: the angle is in the second quadrant, where cosine is negative, with reference angle pi over three. Then the question becomes an inverse sine question: which angle between minus and plus a quarter turn has sine minus one half?

Sine of pi over six is one half, and sine is odd, so minus pi over six works, and it sits inside the allowed interval. The check confirms both conditions. Notice that the answer is a small negative angle, even though you started with a large positive one. Outputs of inverse sines always live in that narrow band around zero.

53. Trap: an answer outside the range

Trap

The trap

Finding an angle with the right cosine, and stopping there:

\[ \cos^{-1}\left(-\frac12\right) = -\frac{\pi}{3} \]

Wrong. Cosine of minus pi over 3 is plus one half, and the angle is outside the range anyway.

The fix

The inverse cosine answers from the upper half circle only. The angle there with cosine minus one half is in the second quadrant.

\[ \cos^{-1}\left(-\frac12\right) = \frac{2\pi}{3} \]

\[ \cos\frac{2\pi}{3} = -\frac12 \checkmark \]

This is the most common wrong answer in the section, and it comes from treating the inverse cosine as if it behaved like the inverse sine. For the inverse sine, a negative input gives a negative angle. It is natural to expect the same here.

It fails twice over. First, cosine is even, so the cosine of minus pi over three is plus one half, not minus one half: the angle does not even satisfy the equation. Second, the inverse cosine only answers from zero to a half turn, and a negative angle is never in that range.

The correct method is the unit circle. Cosine is the horizontal coordinate, so you need a point in the upper half circle with horizontal coordinate minus one half. That is in the second quadrant, at two pi over three. The inverse cosine of a negative number is always an obtuse angle, between a quarter turn and a half turn.

54. Two round trips, two fates

Fill the middle

Fill in the blanks

\sin\left(\sin^0.3 0.3\right) = 0 \qquad \sin^5(\sin \pi) = ___ \qquad \cos^___\left(\cos\tfrac______\right) = \frac______} }

Why: Sine after inverse sine always returns its input when the input is between minus 1 and 1, so the first is 0.3. The second starts outside the restricted interval and comes back as 0. The third starts inside the range of the inverse cosine, from 0 to pi, so it cancels cleanly and returns pi over 5.

Fill in all three blanks before checking. Each is a composition of a trigonometric function with its inverse, but the order and the starting value decide whether it cancels.

The first starts with a number, 0.3, which is between minus one and one. The inverse sine turns it into an angle whose sine is 0.3, and sine turns it straight back. That order always cancels on the inverse's domain.

The second and third start with an angle. Pi is outside the inverse sine's range, so it comes back as zero. Pi over five is inside the inverse cosine's range, which runs from zero to pi, so it comes back unchanged. Whenever you see an inverse trig function applied to a trig function, ask first whether the starting angle is in the range.

55. Cosine of an inverse sine, from a triangle

Worked example

Simplify the composition for x between minus 1 and 1. It will return when you differentiate the inverse sine in Chapter 3.

\[ \cos\left(\sin^{-1} x\right) \]

Name the angle

Why: Let theta be the inverse sine of x.

\[ \theta = \sin^{-1} x \;\Longrightarrow\; \sin\theta = x, \quad -\frac{\pi}{2} \le \theta \le \frac{\pi}{2} \]

Figure (svg): A right triangle with hypotenuse 1, the side opposite the angle theta equal to x, and the adjacent side equal to the square root of 1 minus x squared. The right angle is at the bottom right.

Name the angle θ = sin⁻¹ x. Its sine is x over 1, so label the opposite side x and the hypotenuse 1; Pythagoras gives the third side, and cosine is that side over 1.

Use the Pythagorean identity

Why: Cosine squared is one minus sine squared.

\[ \cos^2\theta = 1 - \sin^2\theta = 1 - x^2 \]

Take the root with the right sign

Why: On the allowed interval cosine is never negative, so take the positive root.

\[ \cos\left(\sin^{-1} x\right) = \sqrt{1 - x^2} \]

Check with two values

Why: At one half the angle is pi over 6; at minus 0.6 the root is 0.8.

\[ \cos\frac{\pi}{6} = \frac{\sqrt3}{2} = \sqrt{1 - \tfrac14}, \qquad \cos\left(\sin^{-1}(-0.6)\right) = 0.8 = \sqrt{1 - 0.36} \]

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 78 — compare Exercises 212 and 213

This simplification is not in the worked examples of the section, but Exercises 212 and 213 ask for its relatives, and you will need it in Chapter 3 to differentiate the inverse sine.

Name the inner angle theta. Then you know its sine is x, and you want its cosine. The triangle in the figure is a way to remember the result: sine is opposite over hypotenuse, so label the opposite side x and the hypotenuse one; Pythagoras gives the adjacent side as the root of one minus x squared.

The one subtle point is the sign. The Pythagorean identity gives cosine squared, and taking the root allows either sign. The range of the inverse sine settles it: between minus and plus a quarter turn, cosine is never negative, so the positive root is right. The check tests a positive and a negative x, and both agree.

56. Match each expression to its exact value

Matching

Match the pairs

  • a. tan⁻¹(√3/3)
  • b. cos⁻¹(−√2/2)
  • c. cot⁻¹(1)
  • d. sin⁻¹(−1)
  • e. cos(tan⁻¹(√3))
  • p. π/6
  • q. 3π/4
  • r. π/4
  • s. −π/2
  • t. 1/2

Why: Tangent of pi over 6 is root 3 over 3. The upper-half angle with cosine minus root 2 over 2 is 3 pi over 4. Cotangent of pi over 4 is 1. The lowest point of the sine range, minus pi over 2, has sine minus 1. And the inverse tangent of root 3 is pi over 3, whose cosine is one half, a number, not an angle.

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 78 — Exercises 207, 208, 209, 210 and 212

Match them all before checking, and for each inverse, say both conditions to yourself: the value and the range. Four of these are exercises straight from the book.

The one to be careful with is the last. It is cosine of an inverse tangent, so the inverse tangent produces an angle, pi over three, and then cosine turns that angle back into a number, one half. Every other expression on the left returns an angle.

The inverse cotangent may be new. Its range is from zero to a half turn, open at both ends, and cotangent of pi over four is one, so the answer is pi over four. The inverse sine of minus one sits at the very bottom of the inverse sine's range, minus pi over two, which is allowed because that range includes its endpoints.

57. Exercise 221: the Mach angle

Estimation

Figure (svg): The Mach angle mu equals 2 times the inverse sine of 1 over M, in degrees, plotted against the Mach number M from 1 to 5. It falls from 180 degrees at M equals 1, steeply at first and then slowly. Dots mark M equals 1.4, 2.8 and 4.3 at about 91, 42 and 27 degrees.

The faster the plane, the narrower the cone of the shock wave. The inverse sine converts the speed ratio back into an angle.

Predict first

With μ = 2 sin⁻¹(1/M), estimate the Mach angle for a plane at M = 2.8, to the nearest few degrees.

  • About 21°
  • About 42°
  • About 70°
  • About 91°

Correct: About 42°

Why: One over 2.8 is about 0.357, whose inverse sine is about 20.9 degrees; doubling gives about 41.8 degrees. The answer 21 forgets the factor of 2, and 91 is the angle at Mach 1.4.

\[ M = 1.4: \;\mu \approx 91^\circ, \qquad M = 2.8: \;\mu \approx 42^\circ, \qquad M = 4.3: \;\mu \approx 27^\circ \]

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 79 — Exercise 221

When a plane flies faster than sound, the pressure waves it makes pile up into a cone, and the Mach angle measures how wide the cone is. The model in the exercise is two times the inverse sine of one over the Mach number.

Make an estimate before revealing the answer. One over 2.8 is a bit more than a third. The inverse sine of one half is thirty degrees, and a third is smaller, so expect around twenty degrees for the inverse sine, then double it.

The graph shows the trend across all speeds. Just above the speed of sound the cone is wide open, and it narrows quickly as the speed grows. The inverse sine is doing the essential job here: it converts a ratio of speeds back into an angle. Notice that the formula only makes sense for M at least one, because the inverse sine needs its input between minus one and one.

58. Exercise 216: converting temperatures back

Real world

\[ C = T(F) = \frac59(F - 32) \]

Discussion prompt

This function turns degrees Fahrenheit into degrees Celsius. Find the inverse function, say what it is used for, and test it on the boiling point of water.

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 78 — Exercise 216

Write your own answer before revealing. This is the strategy from Part 3 with meaningful letters: solve for F by multiplying by nine fifths and adding thirty-two.

The boiling point check is a round trip with real meaning. Two hundred and twelve degrees Fahrenheit becomes a hundred Celsius, and the inverse takes a hundred back to two hundred and twelve.

Notice that the book does not swap the letters here. If you swapped, you would write F equals nine fifths of C plus thirty-two with the letters renamed, and the new function would take a temperature called F that is really in Celsius. In applications, keep the names attached to their units, and swap only when you want to graph the function and its inverse on the same axes.

59. Putting it together

Section

Part 7

60. Pattern: finding and trusting an inverse

Pattern

Figure (svg): A flow diagram. First ask whether f is one-to-one. If not, restrict the domain and return. If so, solve y equals f of x for x, swap x and y, swap the domain and range, and verify both compositions.

Existence first, formula second, bookkeeping third, and a check that runs in both directions.
  1. Is f one-to-one? Use the horizontal line test, or show that equal outputs force equal inputs.
  2. If not, restrict to an interval where it is, and remember which half you kept.
  3. Solve y = f(x) for x, choosing any sign by the restriction, then swap x and y.
  4. Swap the domain and range of f to get those of the inverse.
  5. Check both cancellation equations, or at least a round trip with a number.

The diagram and the list are the same procedure. Existence comes first: if the function is not one-to-one, no algebra will produce an inverse, and you must restrict first.

The restriction then does a second job later on. When the solving step produces a plus or minus, the restriction is what tells you which sign to keep. It also becomes the range of the inverse when you swap the domain and range.

The final step is the check, and it runs in both directions. For a formula, compose both ways. For a quick test, pick a number in the domain, send it through f and then the inverse, and see that it comes back. For the inverse trigonometric functions, the check is always the same pair: the value is right and the angle is in the range.

61. Check: which function has an inverse

Check

Check your understanding

Which of these functions is one-to-one on the given domain?

  • A. f(x) = x², all real x
  • B. f(x) = cos x, 0 ≤ x ≤ π (correct)
  • C. f(x) = |x − 2|, all real x
  • D. f(x) = sin x, 0 ≤ x ≤ π

Answer: B

Why: On the interval from 0 to pi, cosine falls steadily from 1 to minus 1, so no output repeats. That is exactly why this interval is the conventional restriction for the inverse cosine.

Why A tempts people
Two opposite inputs give the same square, for example 2 and minus 2.
Why C tempts people
The absolute value gives the same output 1 at x equal to 1 and at x equal to 3.
Why D tempts people
Sine rises and then falls on this interval: the sine of pi over 6 equals the sine of 5 pi over 6.

Decide before looking at the options, and name a repeated output for each function you reject.

Cosine from zero to pi is the one that passes, and it is not an accident: it is the conventional restriction you just used to define the inverse cosine. On that interval cosine only falls.

The most tempting wrong answer is sine on the same interval. It looks similar, but sine rises for the first quarter turn and falls for the second, so each height between zero and one is reached twice. That is why sine uses a different interval, centred on zero, for its inverse.

62. Check: finding an inverse

Check

Check your understanding

What is the inverse of f(x) = (x + 5)/2?

  • A. f⁻¹(x) = 2x − 5 (correct)
  • B. f⁻¹(x) = 2/(x + 5)
  • C. f⁻¹(x) = (x − 5)/2
  • D. f⁻¹(x) = 2x + 5

Answer: A

Why: Solve y equals x plus 5 over 2: multiply by 2 to get 2y equals x plus 5, subtract 5, then swap. Check: f of 1 is 3, and 2 times 3 minus 5 is 1.

Why B tempts people
That is the reciprocal of f, not its inverse; the minus one in the notation is not an exponent.
Why C tempts people
This undoes the steps in the wrong order: f adds 5 first and divides by 2 last, so the inverse must multiply by 2 first.
Why D tempts people
The sign of the 5 must flip when undoing an addition; with plus 5, f of 1 equals 3 would go to 11.

Work it out before choosing. The function adds five and then divides by two, so its inverse must multiply by two first and subtract five second.

The quickest way to decide between the options is a round trip. The original sends one to three. The correct inverse must send three back to one, and two times three minus five is one.

The distractors are each a real mistake. One reads the minus one as a reciprocal. One undoes the operations in the original order instead of reversing it. One forgets to change the sign of the five. A round trip with any single number catches all three.

63. Check: an inverse trigonometric value

Check

Check your understanding

What is sin⁻¹(sin(5π/6))?

  • A. 5π/6
  • B. π/6 (correct)
  • C. −π/6
  • D. 1/2

Answer: B

Why: The sine of 5 pi over 6 is one half. The angle between minus and plus a quarter turn with sine one half is pi over 6. The original angle lies outside the restricted interval, so the round trip does not return it.

Why A tempts people
This assumes the cancellation always works, but 5 pi over 6 is outside the range of the inverse sine.
Why C tempts people
Minus pi over 6 has sine minus one half, the wrong sign.
Why D tempts people
One half is the inner sine value; the inverse sine then turns it back into an angle.

Work from the inside out before choosing. The sine of five pi over six is one half, since it is in the second quadrant with reference angle pi over six.

Then the inverse sine asks for the angle between minus and plus a quarter turn with sine one half, which is pi over six. The starting angle was outside that interval, so it could not come back.

If you chose five pi over six, you assumed the cancellation always works. Revisit the zigzag graph: the round trip returns x only inside the shaded strip. If you chose one half, you stopped halfway, after the inner sine.

64. Explain the price of a restriction

Explain it to yourself

Discussion prompt

The inverse sine of the sine of x is not always x, but the sine of the inverse sine of y is always y (for y between −1 and 1). Explain in two or three sentences why one order always cancels and the other does not.

Write your explanation before revealing the model answer. Try to use the word range in it, because that is where the explanation lives.

The asymmetry comes from what each function is allowed to return. Sine accepts any angle and returns a number. The inverse sine accepts a number and must return one particular angle, the one in its range.

So starting from a number and going through the inverse sine first, you land on a genuine angle with that sine, and applying sine returns the number. Starting from an angle, you lose information when you apply sine, because many angles share that sine, and the inverse sine can only pick the representative in its range. That lost information is the price of the restriction.

65. Exit ticket

Exit ticket

\[ f(x) = x^2 - 4, \quad x \ge 0 \]

Discussion prompt

Exercise 189. Find the inverse, give its domain and range, and verify it with one round trip.

OpenStax Calculus Volume 1, §1.4 Inverse Functions §1.4, p. 77 — Exercise 189

Do this one without looking back. It uses every step of the pattern: check the restriction, solve with the right sign, swap, state the domain and range, and verify.

The restriction x at least zero means the positive root is the right one. The domain of the inverse is the range of f: the right half of the parabola starts at minus four and rises, so the inverse accepts every number from minus four up.

The round trip at three confirms the formula. If yours did not come back to three, look first at the sign of the root, and then at whether you added or subtracted the four.

66. Recap

Recap

ideawhat to remember
one-to-oneequal outputs force equal inputs; exactly the functions with inverses
horizontal line testevery horizontal line meets the graph at most once
definitionf⁻¹(f(x)) = x on the domain of f and f(f⁻¹(y)) = y on its range
finding f⁻¹solve y = f(x) for x, swap the letters, swap domain and range
graphthe reflection of the graph of f in the line y = x
restrictingkeep a one-to-one piece; the sign of the root records which piece
inverse trigoutputs are angles in fixed ranges; sin⁻¹(sin x) = x only on [−π/2, π/2]

Next, Section 1.5 meets the most important inverse pair in calculus: the exponential function and the natural logarithm.

Stewart, Calculus: Early Transcendentals 8e, §1.5 Inverse Functions and Logarithms §1.5, pp. 55-67 — the same material in Stewart

Everything in this lesson follows from one idea: an inverse sends each output back to the input that produced it. That is only possible when no two inputs share an output, which is what one-to-one means and what the horizontal line test detects.

Finding an inverse is solving and swapping, and its graph is the reflection in the diagonal line. When a function is not one-to-one, you restrict it, and the restriction decides the sign of any root and the range of the inverse. The inverse trigonometric functions are exactly this, with conventional restrictions that you must respect when you evaluate them.

Section 1.5 introduces the exponential functions and their inverses, the logarithms. Everything here carries over: logarithms exist because exponentials are one-to-one, the graph of a logarithm is the reflection of an exponential, and the cancellation equations become the rules you use to solve exponential equations.

Sources

  1. OpenStax Calculus Volume 1, §1.4 Inverse Functions — Strang & Herman, OpenStax / Rice University, CC BY-NC-SA 4.0, pp. 65-79
  2. Stewart, Calculus: Early Transcendentals 8e, §1.5 Inverse Functions and Logarithms — James Stewart, Cengage Learning, 2016, pp. 55-67

Want this taught 1-on-1? Alexander tutors Calculus I — $55/session, free consultation.

Book on Wyzant · Text (657) 465-8108