1.2 Domain and Range

Finds the domain of a rule by hunting for what it forbids — a zero denominator and a negative even radicand cover nearly every case — then writes the survivors in interval and set-builder notation. Reads both domain and range off a graph as its two shadows, and handles piecewise definitions, whose conditions must partition the domain without overlapping or leaving a gap.

Subject: Precalculus · 64 slides · symbolic lesson

Open the interactive version of this deck

What this lesson covers

The lesson, slide by slide

1. Domain and Range

Title

Precalculus · Section 1.2

Which inputs a rule accepts, which outputs come back, and how to write both down

2. What this lesson gives you

Objectives

Section 1.1 said a function pairs each input with exactly one output. It did not say which inputs are allowed. That is this lesson, and it is where most of the algebra of Chapter 1 actually lives.

OpenStax, Precalculus, §1.2 Domain and Range §1.2, pp. 40-61 — learning objectives 1.2.1 and 1.2.2

Section 1.1 gave you the idea of a function: one input, one output. What it did not say is which inputs a rule will actually accept. That question turns out to have a short answer and a lot of practice attached to it, and it is what this whole section is about.

Three skills sit side by side here and it helps to keep them apart in your head. The first is algebraic: look at a formula and work out which inputs would break it. The second is notational: write a set of numbers correctly, with the right brackets in the right order. The third is graphical: look at a picture and read off both the inputs used and the outputs produced.

The last part of the lesson, piecewise functions, is where all three meet. A piecewise rule is several formulas with a slice of the domain each, so you cannot evaluate one, graph one or state its domain without being fluent in the first three skills.

Everything here gets used again immediately. Every graph you sketch for the rest of the course comes with a domain, and in calculus the domain is often the whole difficulty of a problem.

3. Before anything new: which inputs break a rule?

Warm-up

Discussion prompt

Without looking ahead, write down one number you may not put into one divided by x, one number you may not put into the square root of x, and one number you may not put into the cube root of x.

Write your three answers down before you reveal anything. Committing to an answer, even a wrong one, is what makes the correction stick.

The first two are the ones almost everyone gets. Dividing by zero is undefined, so one over x has no value at zero. A square root asks for a number whose square is the input, and no real number squares to something negative, so the square root refuses every negative input.

The third one is the interesting one. If you hunted for a number the cube root refuses, you were hunting for something that is not there. The cube root of negative eight is negative two, because negative two multiplied by itself three times gives negative eight. Odd roots are perfectly happy with negative numbers; it is only even roots that are fussy.

Hold on to that asymmetry. It is the reason the section talks about even roots rather than roots in general, and it is a favourite examination question.

4. What a formula forbids

Section

Part 1

5. The domain is decided by the rule, not chosen by you

Concept

Figure (svg): A diagram: a box of accepted inputs on the left, the rule one over x minus two in a box in the middle, a box of outputs on the right, and the input two drawn below in red bouncing off the machine because it would divide by zero.

The domain is decided by the rule, not chosen by you: it is every input the rule can actually process.

domain — The set of inputs a function accepts. When a rule is given by a formula with no story attached, its domain is taken to be every real number for which the formula returns a real number.

range — The set of outputs the function actually produces, once every input from the domain has been fed through.

So the work is never to find what is allowed. It is to find the small set of inputs that break the rule, and then keep everything else.

OpenStax, Precalculus, §1.2 Domain and Range §1.2, pp. 40-41 — the holding-area picture, Figure 1.18

Look at the picture as a machine with two holding areas. On the left is everything the machine will accept, on the right everything it can produce. The machine in the middle is the rule, and it is the rule that decides both.

That word decides matters. You do not get to choose a domain to make a problem easier, and you do not have to be told one. When a formula arrives with no story attached, its domain is understood to be every real number the formula can handle. Some books call this the implied domain.

Now look at the input two, drawn in red below. It is not refused because anyone dislikes it; it is refused because feeding it in would ask the machine to divide by zero, and there is no number to hand back. Every domain question in this lesson has exactly this shape: find the inputs that would ask for something impossible, and keep the rest.

One more thing worth noticing in the picture. The outputs listed on the right are a strange-looking bunch, and they are much harder to describe than the inputs. That is normal: the range is usually the harder of the two questions, which is why graphs get used to answer it.

6. Only two operations can fail

Concept

At this level a formula built from real numbers can go wrong in exactly two ways, and you can spot both by looking.

Notice that the two jobs pull in opposite directions. A denominator gives you an equation whose solutions you delete. An even root gives you an inequality whose solutions you keep.

OpenStax, Precalculus, §1.2 Domain and Range §1.2, pp. 41-44 — the three forms to remember

This slide is the whole algebraic half of the lesson, so it is worth memorising. At this level, only two things a formula can do will refuse an input: dividing by something that turns out to be zero, and taking an even root of something that turns out to be negative. Logarithms will add a third case in Chapter 4.

Pay attention to the difference in how you handle them, because reversing the two is the commonest mistake in this material. A denominator gives you an equation. You solve it, and the solutions are the enemy: you throw them away. An even root gives you an inequality. You solve it, and the solutions are your friends: you keep exactly those.

So one leaves you with all the real numbers minus a few isolated points, and the other leaves you with a single stretch of the line, or sometimes a couple of stretches. If you ever find yourself with an answer of the wrong shape, for instance a whole half-line removed because of a denominator, you have probably swapped the two.

Everything else is safe. You can add, subtract, multiply and raise to whole-number powers all day without excluding anything, and odd roots never cause trouble either.

7. Example 1.16: the domain of a list of pairs

Worked example

Not every function comes as a formula. This one is just a list, and a list carries its domain on its face.

\[ \{(2, 10), (3, 10), (4, 20), (5, 30), (6, 40)\} \]

Pick out the first coordinate of each pair

Why: In an ordered pair the input is written first.

\[ 2, \quad 3, \quad 4, \quad 5, \quad 6 \]

Collect them into a set

Why: Braces, and no repeats; the order they are written in does not matter.

\[ \text{domain} = \{2, 3, 4, 5, 6\} \]

Check against the outputs

Why: The second coordinates give the range, and 10 appears twice there: two different inputs may share an output, while each input is listed once.

\[ \text{range} = \{10, 20, 30, 40\} \]

OpenStax, Precalculus, §1.2 Domain and Range §1.2, p. 42 — Example 1.16

This is the gentlest possible case, and it is worth doing because it keeps the definition honest. A function does not have to be a formula. Here it is simply a list of five pairings, and the domain is whatever the list says it is.

The only thing to be careful about is which coordinate is which. The input is written first in an ordered pair, so the domain comes from the first coordinates and the range from the second ones. Reading them the wrong way round gives two answers that are both wrong.

Notice what happens with the output ten. It appears twice, from two different inputs. That is entirely allowed: a function may send two inputs to the same output. What it may not do is send one input to two outputs. When you list the range you write ten once, because a set does not count repeats.

There is nothing to exclude here, because nothing is being computed. Keep that in mind when you meet a table of data later: a table carries its domain on its face, exactly like this list.

8. Example 1.17: a polynomial forbids nothing

Worked example

Find the domain of the rule below.

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

Follow what the rule does to a number

Why: Square it, then take one away. Both moves work on every real number.

\[ x \;\longmapsto\; x^2 \;\longmapsto\; x^2 - 1 \]

Look for the two failure modes

Why: There is no denominator and no even root, so nothing is excluded.

\[ \text{domain} = (-\infty, \infty) \]

Do Try It 1.14 the same way

Why: Adding, subtracting and cubing are allowed everywhere, so every polynomial has the same answer.

\[ f(x) = 5 - x + x^3: \quad \text{domain} = (-\infty, \infty) \]

Check an input that looks dangerous

Why: A large negative number is squared into a large positive one; there is no boundary anywhere.

\[ f(-1000) = 1000000 - 1 = 999999 \]

OpenStax, Precalculus, §1.2 Domain and Range §1.2, p. 43 — Example 1.17 and Try It 1.14

Read the rule as a sequence of instructions: take the input, square it, subtract one. Then ask of each instruction whether there is any real number it would refuse. Squaring accepts everything, subtracting accepts everything, so nothing is refused.

The answer is written as one interval running from negative infinity to infinity, with round brackets at both ends because neither end is a number you can reach. You will write that interval a great many times before the course is over.

The Try It underneath is the same reasoning with cubing in place of squaring. Every polynomial, however long and however high its powers, has the same answer: all real numbers. Once you have seen why, you never have to work it out again, you just have to recognise a polynomial when you see one.

The check step is there to build a habit rather than to settle any doubt. Ask yourself what the most awkward input you can think of would do. A large negative number gets squared into a large positive one, which is fine; there is simply no boundary anywhere for the rule to trip over.

9. Example 1.18: a denominator that can vanish

Worked example

Now there is something to find. A denominator is only a problem at the inputs that make it zero, so find those first.

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

Set the denominator equal to zero

Why: This is an equation, and its solutions are the inputs to delete.

\[ 2 - x = 0 \]

Solve for x

Why: Add x to both sides.

\[ x = 2 \]

Delete that one input

Why: Everything to the left of 2 is fine, and so is everything to the right.

\[ x < 2 \quad \text{or} \quad x > 2 \]

Write it with a union

Why: The union symbol glues two separate stretches into one set.

\[ \text{domain} = (-\infty, 2) \cup (2, \infty) \]

Figure (svg): The graph of y equals x plus one over two minus x, in two branches that run away from a dashed vertical line at x equals 2, with a heavy bar along the bottom broken by an open circle at 2.

Two separate branches, one on each side of the forbidden input. The heavy bar below is the domain: the whole line with a single point punched out.

Check on both sides of 2

Why: Just below 2 the outputs are large and positive, just above they are large and negative, and at 2 itself the rule asks for a division by zero.

\[ f(1.9) = 29, \quad f(2.1) = -31, \quad f(2) = \tfrac{3}{0} \text{ is undefined} \]

OpenStax, Precalculus, §1.2 Domain and Range §1.2, pp. 43-44 — Example 1.18

The key move is the very first one, and it is easy to say out loud: set the denominator equal to zero. Notice that this is an equation, even though the answer you are building is a set. The equation is a device for locating the bad inputs, nothing more.

Here it locates exactly one, the input two. The numerator plays no part at all. A numerator of zero is perfectly legal and just produces an output of zero, so do not waste time looking at it when the question is about the domain.

Now look at the picture, because it shows what removing a single input does to a graph. The curve breaks in two, and near the missing input the outputs shoot off in opposite directions on either side. That vertical dashed line is not part of the graph; it marks where the graph is not.

The heavy green bar underneath is the domain drawn on the number line, with one point punched out of it. Get used to reading a domain as a picture like that, because it makes the union notation feel like a description rather than a rule to memorise.

10. Try It 1.15: the same move, a different denominator

Fill the middle

\[ f(x) = \frac{1+4x}{2x-1} \]

Fill in the blanks

2x - 1 = 0 \;\Longrightarrow\; x = 1/2, \qquad \text1/2 = \left(-\infty, 1/2\right) \cup \left(___, \infty\right)

Why: Setting two x minus one equal to zero gives x equal to one half, and that single input is the only one deleted, so the domain is everything below one half together with everything above it. The numerator plays no part at all: a numerator of zero is perfectly legal, and here it would just make the output zero.

OpenStax, Precalculus, §1.2 Domain and Range §1.2, p. 44 — Try It 1.15

Do it yourself before you type anything. Set the denominator equal to zero, solve the little linear equation, and you have the one input to remove.

Two x minus one is zero when two x is one, that is when x is one half. So one half is the single input this rule refuses, and everything else survives. The answer splits into the part of the line below one half and the part above it.

The numerator, one plus four x, is again irrelevant. It is worth saying twice, because the temptation to set the numerator equal to zero as well is surprisingly strong. That calculation is useful, but for a different question: it tells you where the graph crosses the horizontal axis.

When you type your answer, notice that both blanks hold the same number. A single forbidden input always shows up twice in the interval notation, once as the right-hand end of the piece below it and once as the left-hand end of the piece above it, and both of those brackets are round.

11. Example 1.19: an even root needs a nonnegative inside

Worked example

A square root asks a different question. Instead of deleting solutions you keep them, and instead of an equation you solve an inequality.

\[ f(x) = \sqrt{7-x} \]

Set the radicand at or above zero

Why: Zero is allowed: the square root of zero is zero, which is real.

\[ 7 - x \ge 0 \]

Subtract 7 from both sides

Why: Both sides drop by the same amount, so the inequality still holds.

\[ -x \ge -7 \]

Divide both sides by negative one

Why: Dividing an inequality by a negative number reverses it.

\[ x \le 7 \]

Write it as an interval

Why: 7 itself is allowed, so its bracket is square; there is no left end, so that bracket is round.

\[ \text{domain} = (-\infty, 7\,] \]

Figure (svg): The graph of y equals the square root of seven minus x, falling from the left and ending at a filled dot on the x-axis at x equals 7, with a heavy arrow along the axis running left from 7.

Left of 7 the inside of the root is positive; at 7 it is zero; right of 7 it is negative and there is no real output at all.

Check the endpoint and one input past it

Why: At 7 the inside is zero and the output is zero, a real number; at 8 the inside is negative and there is no real output at all.

\[ f(7) = \sqrt{0} = 0, \qquad f(8) = \sqrt{-1} \text{ is not real} \]

OpenStax, Precalculus, §1.2 Domain and Range §1.2, pp. 44-45 — Example 1.19

Here is the other failure mode, and the mental script is different. You are not looking for bad inputs to delete; you are describing the good ones directly, by writing down the condition that the inside of the root is not negative.

Zero is allowed inside a square root, which is why the condition uses at or above zero rather than strictly above. That single detail decides whether the final bracket is square or round, so it is worth being deliberate about.

The middle of the algebra hides the classic slip. When you divide both sides of an inequality by negative one, the inequality turns around. If you forget, you will write that x is at least seven, which is precisely the wrong half of the line. The picture is the cheapest possible check: the curve lives to the left of seven, not to the right.

Look at the endpoint on the graph. At seven the inside of the root is zero, the output is zero, and the dot is filled, which is exactly what the square bracket in the answer is saying.

12. Try It 1.16: which way does the inequality point?

Prediction

\[ f(x) = \sqrt{5+2x} \]

Predict first

Solve the condition that the inside of the root is at or above zero. Which inputs survive?

  • every x at least -2.5
  • every x at most -2.5
  • every x at least 2.5
  • every real number

Correct: every x at least -2.5

Why: Five plus two x is at or above zero when two x is at or above negative five, that is when x is at or above negative two point five. Nothing was divided by a negative number here, so the inequality never flipped. Checking one input settles it: at negative three the inside would be negative one, which has no real root, and at negative two the inside is one, whose root is one.

OpenStax, Precalculus, §1.2 Domain and Range §1.2, p. 45 — Try It 1.16

Commit to an option before reading on. The point of this one is to see whether the previous example taught you the rule or just the answer.

The condition is that five plus two x is at or above zero. Subtract five, and two x is at or above negative five. Divide by two, which is positive, so nothing flips, and x is at or above negative two point five.

That is the trap. The previous example flipped because it divided by a negative number; this one does not, because it divides by two. The inequality sign is not decoration and it does not flip out of habit. It flips only when you multiply or divide by a negative number.

Test it in your head if you are unsure. At negative three the inside would be negative one, which has no real square root, so negative three must be out. At negative two the inside is one, whose root is one, so negative two must be in. Only one of the four options agrees with both of those.

13. Exercise 112: a root inside a denominator

Worked example

Both failure modes in one rule. Each contributes a condition, and the domain is what satisfies them at the same time.

\[ f(x) = \frac{5}{\sqrt{x} - 3} \]

The even root needs a nonnegative inside

Why: Here x itself is the radicand.

\[ x \ge 0 \]

The denominator may not be zero

Why: Set it to zero to find the input to delete.

\[ \sqrt{x} - 3 = 0 \]

Solve that equation

Why: Add 3 to both sides, then square both sides.

\[ \sqrt{x} = 3 \;\Longrightarrow\; x = 9 \]

Impose both conditions at once

Why: Start at zero, and punch out the single input 9.

\[ \text{domain} = [0, 9) \cup (9, \infty) \]

Check one good input and both bad ones

Why: At 4 the rule returns a number; at 9 the denominator is zero; at a negative input there is no real root to begin with.

\[ f(4) = \frac{5}{2-3} = -5, \quad f(9) = \frac{5}{0}, \quad f(-1): \; \sqrt{-1} \text{ is not real} \]

OpenStax, Precalculus, §1.2 Domain and Range §1.2, p. 63 — Exercise 112

This is the exercise that separates people who have learned two rules from people who have learned to combine them. The root is in the denominator, so it must satisfy both conditions at once.

From the root you get that x is at or above zero. From the denominator you get that the root itself must not be zero, which happens when x is nine. Notice how that second condition was found: set the denominator equal to zero and solve, exactly as in Example 1.18, even though the denominator now contains a root.

Imposing both conditions gives an answer with a very particular shape: a half-line starting at zero, with one interior point punched out. The zero end is square, because zero is fine to put in, and the two brackets around nine are round.

It is worth noticing what would change if the root were in the numerator instead. Then zero inside the root would be harmless, nine would be an ordinary input, and the answer would just be everything from zero on. The position of the root inside the fraction changes the answer, so read the formula carefully before you start.

14. Which restriction does each rule carry?

Sorting

Sort into buckets

Sort each rule by what it forbids, before solving anything.

nothing is forbidden
x² - 5x + 6; the cube root of (x - 7)
a denominator only
1/(x - 7); (x + 1)/(x² - 4)
an even root only
the square root of (x - 7); the fourth root of (9 - x)
both at once
1 / the square root of (x - 7)
none
A polynomial and an odd root accept every real number. The cube root of a negative number is negative, which is still a real number.
den
Set the denominator equal to zero and delete what you find: x = 7 in the first, and both x = 2 and x = -2 in the second.
root
An even root, so set the inside at or above zero and keep the solutions: x at least 7 in one, x at most 9 in the other.
both
The root must be real AND the denominator must be nonzero. Nonnegative and nonzero together mean strictly positive, so this domain is x greater than 7, with 7 itself excluded.

Do this sort with your eyes, not with algebra. The whole point is to build the reflex of scanning a formula for hazards before you do any solving, because that scan is what tells you how many conditions your answer will have.

Two of these look alike but behave differently: the square root and the cube root of the same expression. The even root restricts, the odd root does not. If you sorted those two into the same bucket, go back and reread the warm-up.

The fourth root is there for the same reason. Fourth, sixth and eighth roots behave exactly like square roots, because they are all even, and the words you say to yourself should be even root rather than square root.

The last bucket, both at once, is the one to be slow with. A root in a denominator has to be real and nonzero at the same time, and those two conditions together are stricter than either alone. That is the whole content of the previous slide, met again in one line.

15. Exercise 118: a cubic under the root

Worked example

Same rule as always: the inside of an even root must be at or above zero. This time the inside is a cubic, so the answer is not one interval.

\[ f(x) = \sqrt{2x^3 - 50x} \]

Set the radicand at or above zero

Why: The condition is unchanged; the solving is harder.

\[ 2x^3 - 50x \ge 0 \]

Factor completely

Why: Take out 2x first, then use the difference of squares.

\[ 2x(x^2 - 25) = 2x(x-5)(x+5) \ge 0 \]

Mark the roots on a line

Why: They cut the real line into four regions, and the sign can only change at a root.

\[ x = -5, \quad x = 0, \quad x = 5 \]

Figure (svg): A sign chart with four rows: the signs of two x, of x plus five and of x minus five in each of the four regions cut by minus five, zero and five, and the product's sign underneath, which is nonnegative between minus five and zero and again from five onward.

Each factor changes sign once. The product changes sign at every root, so the signs alternate, and the shaded stretches are where the radicand is allowed to sit.

Read off the regions where the product is nonnegative

Why: Two negative factors and one positive give a positive product; the roots themselves give zero, which is allowed.

\[ \text{domain} = [-5, 0] \cup [5, \infty) \]

Check one input from each region

Why: Two of the four regions produce a real root and two do not.

\[ f(-6): \sqrt{-132}, \quad f(-3) = \sqrt{96}, \quad f(3): \sqrt{-96}, \quad f(6) = \sqrt{132} \]

OpenStax, Precalculus, §1.2 Domain and Range §1.2, p. 63 — Exercise 118

The condition is the same as always, so the only new thing here is how you solve an inequality whose left side is not linear. The answer is never to divide by anything; it is to factor, find the roots, and let the number line do the work.

Factoring is the step to do carefully. Take out the common factor of two x first, and what is left is a difference of squares, which splits into two more linear factors. Three factors, three roots, four regions.

Now read the sign chart from the top down. Each row tells you the sign of one factor in each region, and the bottom row multiplies them. Signs are easy to multiply: an even number of negative factors gives a positive product, an odd number gives a negative one.

The roots themselves deserve a moment's thought. At negative five, zero and five the radicand is zero, and the square root of zero is zero, which is a perfectly good output. So the roots are included and the brackets around them are square. If the root had been in a denominator, they would all have been excluded instead.

16. Trap: cancelling before the domain is fixed

Trap

The trap

Exercise 117. The factor cancels, so why not cancel it?

\[ f(x) = \frac{x^2-9x}{x^2-81} \]

\[ = \frac{x(x-9)}{(x-9)(x+9)} = \frac{x}{x+9} \]

Wrong conclusion: the domain is every real number except negative nine.

The fix

The domain belongs to the rule you were given, not to the tidier rule you produced. The original formula asks you to divide by zero when x is nine, so nine is out, and it stays out.

\[ \text{domain} = (-\infty, -9) \cup (-9, 9) \cup (9, \infty) \]

Cancelling is still worth doing, because it tells you what the graph looks like near nine: a single missing point rather than a vertical asymptote.

This trap catches good students, because the algebra in it is correct. The factor really does cancel, and the simplified rule really does equal the original one at every input where both are defined. The error is in thinking that the domain travels with the simplification.

It does not. The domain belongs to the rule you were handed. That rule asks you to divide by x minus nine, and at nine that is a division by zero, so nine is not an input and no amount of tidying can make it one.

There is a useful piece of information in the cancellation, though, and it is not the domain. It tells you what the graph looks like near nine. Because the factor cancels, the graph does not fly off to infinity there; it is a perfectly ordinary curve with a single point missing, usually drawn as a small open circle. Near negative nine, where nothing cancels, the graph does fly off.

So the working order is: domain first, from the rule as given, then simplify to understand the shape. Never the other way round.

17. Interval and set-builder notation

Section

Part 2

18. A bracket includes, a parenthesis excludes

Concept

An interval is written as two ends, smallest first, separated by a comma. Only the shape of the brackets says whether the ends themselves belong.

\[ (0, 100\,] \quad \text{means: more than 0 dollars, and 100 dollars is allowed} \]

OpenStax, Precalculus, §1.2 Domain and Range §1.2, pp. 41-46 — the conventions of interval notation, Figure 1.19

Interval notation is short, which is why everyone uses it, and it is unforgiving, which is why it gets marked wrong so often. There are only four conventions and all four are checked in every answer you write.

The one that carries meaning is the shape of the bracket. Square means the end value is in the set; round means it is not. On a graph that is the difference between a filled dot and an open circle, and in an inequality it is the difference between at most and strictly less than. The same distinction, in three notations.

The other three conventions are mechanical but still cost marks. Smaller number first, always. Two numbers, separated by a comma, and no more than two. And an infinity end always takes a round bracket, because infinity is not a value anything could be equal to.

The example on the slide is worth reading as a sentence. Somebody has a hundred dollars to spend: they can spend any amount up to and including the hundred, but spending nothing is not counted, so zero gets a round bracket and one hundred gets a square one.

19. Reading an interval piece by piece

Notation

Annotate

On: \( [-4, \; 0) \)

  • The left end, −4, belongs to the set. On a graph this is a filled dot; in an inequality it is written with ≤.
  • The smaller number is always written first. Writing [0, −4) is not a different set; it is not a set at all.
  • The right end, 0, does NOT belong. On a graph it is an open circle; in an inequality it is a strict <.
  • Every real number strictly between the ends is included — not just the whole numbers. The interval [−4, 0) holds −3.7 and −0.0001 as well.

Click through the four annotations and say each one out loud in words. An interval is a compressed sentence, and the only way to stop it from being a pattern you copy is to be able to expand it back into English.

The two brackets are doing independent jobs. It is entirely normal for one end to be included and the other not, which is exactly what this interval does, and that pattern shows up constantly in this section: the graph of Example 1.21 will produce one of each.

The last annotation is the one people forget under pressure. An interval contains every real number between its ends, not just the whole numbers. There are infinitely many numbers in this interval, and almost all of them are not integers.

One habit that will save you: after you write an interval, test a number from inside it and a number just outside it against the original condition. Two quick tests catch almost every bracket error.

20. Reading set-builder notation piece by piece

Notation

Annotate

On: \( \{x \mid 4 < x \le 12\} \)

  • Read them as 'the set of'. Everything inside describes a collection of numbers, not a single number.
  • Names the typical member. It says what kind of thing is in the set before any condition is imposed.
  • Read it as 'such that'. Some books print a colon instead; they mean the same thing.
  • The test for membership. A number is in the set exactly when it passes: bigger than 4, and no more than 12.

\[ \{x \mid 4 < x \le 12\} = (4, 12\,] \]

OpenStax, Precalculus, §1.2 Domain and Range §1.2, pp. 45-46 — set-builder notation and interval notation

Set-builder notation looks more forbidding than interval notation, but it is really just an English sentence with two abbreviations in it. Read the braces as the set of and the vertical bar as such that, and the whole line becomes a sentence you can say.

Its advantage is that it can describe anything. Interval notation can only describe a stretch of the line, so a condition like all the whole numbers, or all the inputs where some messy formula is positive, has to be written in set-builder form. You will need that in Chapter 3.

Its disadvantage is that it is longer, and that the condition inside it has to be exactly right. Writing the set of x such that x is greater than four when you meant at least four is a much easier mistake to make than choosing the wrong bracket, because there is no picture attached.

The line underneath the annotation is the translation you should be able to do in either direction on sight. Practise going from interval to set-builder as well, since examination questions ask for both.

21. When one interval is not enough

Concept

Figure (svg): A number line from minus six to six with heavy rays running outward from filled dots at minus three and three, and a dashed line joining them across the middle to show the part that is excluded.

One condition, two intervals. Whenever a set falls apart into separate stretches, interval notation needs a union sign to hold them together.

Some conditions describe a set that falls into separate pieces. The union symbol joins them: a number is in the union when it is in either piece.

\[ \{x \mid |x| \ge 3\} = (-\infty, -3\,] \cup [3, \infty) \]

In inequality notation the same job is done by the word or. Keep the pieces in left-to-right order, and never write a union as though it were one interval.

OpenStax, Precalculus, §1.2 Domain and Range §1.2, p. 46 — combining intervals with a union

Sometimes a perfectly ordinary condition describes a set that comes in two separate pieces. That is what the picture shows: being at least three units away from zero means being well to the right or well to the left, and the middle is excluded.

Interval notation cannot describe that with one interval, because an interval is by definition an unbroken stretch. So you write both pieces and join them with the union sign, which simply means that a number qualifies if it is in either piece.

The word for this in inequality notation is or, and it must be or, never and. A number cannot be both at most negative three and at least three, so writing and would describe an empty set. If you find yourself writing and between two intervals, something has gone wrong.

Keep the pieces in left-to-right order, smallest first, exactly as you do inside a single interval. It is not strictly wrong to write them the other way round, but every marker and every textbook writes them left to right, and you want your answers to be easy to check.

22. Example 1.20: one number line, three notations

Worked example

Read the heavy parts of the line from left to right, and describe each one before joining them.

Figure (svg): A number line from minus two to seven: a heavy segment with filled dots at one and three, then an open circle at five with a heavy ray running right.

Filled dot, square bracket; open dot, round bracket. The arrow end is always round, because infinity is not a number you can include.

Describe the first heavy piece

Why: Two filled dots, so both ends belong.

\[ 1 \le x \le 3 \]

Describe the second heavy piece

Why: An open circle at 5 and an arrow going right, so 5 is out and there is no right end.

\[ x > 5 \]

Join them in inequality notation

Why: The word or allows a number to be in either piece.

\[ 1 \le x \le 3 \quad \text{or} \quad x > 5 \]

Rewrite in set-builder notation

Why: Same condition, now written after the bar.

\[ \{x \mid 1 \le x \le 3 \text{ or } x > 5\} \]

Rewrite in interval notation

Why: Square brackets for the filled dots, round for the open circle and for infinity.

\[ [1, 3] \cup (5, \infty) \]

Check four test numbers

Why: Each one lands where the picture says it should.

\[ 3 \in [1,3], \quad 4 \notin, \quad 5 \notin, \quad 5.1 \in (5, \infty) \]

OpenStax, Precalculus, §1.2 Domain and Range §1.2, p. 47 — Example 1.20

Work through this the way the steps do: describe each heavy piece on its own, in words, before you write any notation at all. Most errors here come from trying to write the final line in one go.

The dots do all the work. Two filled dots at the ends of the first piece mean both ends belong, so that piece is an interval with two square brackets. The open circle at five means five is not in, and the arrow means there is no right-hand end, so that piece gets two round brackets, one because the dot is open and one because infinity is never included.

Then the two pieces get joined, with the word or in inequality notation and with the union sign in interval notation. Both say the same thing.

Do not skip the gap. The numbers between three and five, four among them, are in neither piece, and that is the whole reason two pieces were needed. Testing four is the quickest way to prove to yourself that the union is not just a decorative symbol.

23. Try It 1.17: a line that is heavy on the left

Fill the middle

Figure (svg): A number line from minus six to four: a heavy ray coming in from the left and stopping at a filled dot at minus two, then a heavy segment from a filled dot at minus one to an open circle at three.

Two pieces again, but this time the unbounded one is on the left, so it is written first.

Fill in the blanks

\left(-\infty, \; -2\,\right] \cup \left[-1, \; 3\right)

Why: The left arrow has no start, so the first piece opens with negative infinity and a round bracket, and it closes at the filled dot on negative two with a square bracket. The second piece runs from the filled dot at negative one, which is included, to the open circle at three, which is not. In words: every number at most negative two, together with every number from negative one up to but not including three.

OpenStax, Precalculus, §1.2 Domain and Range §1.2, p. 47 — Try It 1.17

This is the same reading exercise with the unbounded piece on the other side, which changes where the infinity goes and nothing else.

Start at the left. The arrow says the heavy part continues forever in that direction, so the first piece opens with negative infinity and a round bracket. It stops at a filled dot on negative two, so that end is square.

The second piece runs between two dots of different kinds: filled at negative one, open at three. So one end is square and the other is round, and the two ends are written smallest first, as always.

Say the whole thing in words as a final check, because the words are what an examiner is really testing: every number at most negative two, together with every number from negative one up to but not including three. If your notation does not say that sentence, fix the notation.

24. Fill in the missing translations

Comparison

Comparison matrix

in wordsinequalityinterval
at least 1 and at most 31 ≤ x ≤ 3[1,3]
more than 5x > 5(5, ∞)
at least -4 and below 0-4 ≤ x < 0[-4,0)
every real number except 2x < 2 or x > 2(-∞, 2) ∪ (2, ∞)
at most 7x ≤ 7(-∞, 7]

Three notations, one meaning, and the exercise is simply to move between them without thinking too hard. Fluency here pays off for the rest of the course, because a question will state a set in whichever notation suits it and expect you to answer in another.

The two bounded rows are the mechanical ones: read the inequality signs, choose square for the ones that allow equality and round for the strict ones, and write the two numbers in order.

The unbounded rows are where the thinking is. There is no number at the far end, so there is nothing to include or exclude there, and the bracket is round by default. Going the other way, from an interval to an inequality, an infinity end simply disappears: you write one inequality instead of two.

The last row, every real number except two, is the shape a denominator always produces. Notice that it needs a union even though the excluded part is a single point, and that both brackets around that point are round.

25. Trap: a square bracket on infinity

Trap

The trap

A domain that runs forever, written like this:

\[ [-4, \infty\,] \]

Wrong. Infinity is not a number, so there is nothing there to include.

The fix

An unbounded end always takes a round bracket. The same goes for the left end of a set that runs down forever.

\[ [-4, \infty) \quad \text{and} \quad (-\infty, 7\,] \]

The other half of this trap is writing the ends in the wrong order. The smaller number is always first, so a set of numbers below seven is written with negative infinity on the left.

This one is pure notation, and it is marked wrong every time, so it is worth fixing now. A square bracket is a promise that the end value belongs to the set. Infinity is not a value, so there is nothing there to belong.

It helps to remember what infinity is doing in the notation. It is not a number that the interval reaches; it is shorthand for the interval never stopping. You cannot include something the set never gets to.

The partner error is writing the ends in the wrong order. Because the set of numbers below seven is usually read out as x is at most seven, people write seven first. But interval notation is always written in increasing order, so negative infinity goes on the left and seven on the right.

A quick test when you are unsure: point at the two ends on a number line, left one first. Whatever you point at first is what you write first.

26. A graph and its two shadows

Section

Part 3

27. Domain is the shadow below, range is the shadow at the side

Concept

Figure (svg): A downward parabola drawn only from a filled dot at minus two to a filled dot at five, with a heavy bar along the x-axis from minus two to five and a heavy bar up the y-axis from minus four to four, joined to the curve by dashed lines.

Shine a light from above and the curve casts the domain onto the horizontal axis; shine it from the side and the same curve casts the range onto the vertical axis.

A graph shows every input-output pair at once. Flatten it onto the horizontal axis and you see exactly which inputs occur: that is the domain. Flatten it onto the vertical axis and you see which outputs occur: that is the range.

Read the domain left to right and the range bottom to top, because both are written from the smaller value to the larger one.

OpenStax, Precalculus, §1.2 Domain and Range §1.2, pp. 47-48 — finding domain and range from graphs

This is the picture to keep for the rest of the course. A graph is a collection of points; flatten it straight down and you see which inputs occur, flatten it sideways and you see which outputs occur.

Try it with the curve on the slide. Every point of it lies above some input between negative two and five, and no point lies above anything else, so those are the inputs used. Every point lies beside some height between negative four and four, so those are the outputs produced.

Two things people get wrong. The first is reading the range off the ends of the curve instead of off its highest and lowest points; here the top of the range comes from the peak in the middle, not from either end. The second is writing the range backwards; like the domain, it is always written smaller value first, which for a range means bottom to top.

Both ends of this curve are filled dots, so both intervals get square brackets at both ends. The next slide is what happens when they are not dots at all.

28. When the graph runs off the page

Concept

Figure (svg): A curve starting at a filled dot at minus five, height five, wiggling down and then plunging off the bottom of the frame, with a heavy arrow along the top running right without end and a heavy arrow down the right-hand side.

The arrows say the picture is only a window. The domain and the range both continue past the frame, so both are written with an infinity end.

An arrowhead means the graph keeps going. The domain and the range then run past the edge of the picture, and you write an infinity end for whichever side continues.

\[ \text{domain} = [-5, \infty), \qquad \text{range} = (-\infty, 5\,] \]

Notice that the highest output is not at a peak in the middle of the picture; it is at the left endpoint. Always check the ends, not just the turning points.

OpenStax, Precalculus, §1.2 Domain and Range §1.2, p. 48 — Figure 1.24

An arrowhead on a graph is a claim: the curve keeps going the way it was going, past the edge of the picture. That is what turns a domain or a range into an unbounded interval.

Read the horizontal extent first. There is a filled dot at negative five and an arrow heading right, so the inputs start at negative five, including it, and never stop. The domain therefore has a square bracket on the left and an infinity end on the right.

Now the vertical extent. The curve dives off the bottom of the frame, so the outputs go down forever, and the highest output is five. That gives a range with an infinity end at the bottom and a square bracket at five.

Notice where that top value came from. It is not at a peak in the middle of the picture; it is at the left endpoint, and the two peaks inside the window are well below it. Whenever you read a range, check the endpoints as well as the turning points, because either can hold the extreme value.

29. Example 1.21: reading both off one curve

Worked example

One curve, two answers. Take the horizontal extent first, then the vertical extent.

Figure (svg): A curve from an open circle at minus three on the x-axis, dipping to about minus three point seven, rising almost back to the axis and then falling to a filled dot at one, height minus four, with a domain bar above and a range bar to the right.

The height zero is reached only at the open left end, so zero is not in the range. That single open dot is the whole difference between a square bracket and a round one.

Find the leftmost and rightmost points

Why: The left end is an open circle above negative three; the right end is a filled dot above one.

\[ \text{left: } x = -3 \text{ (open)}, \qquad \text{right: } x = 1 \text{ (filled)} \]

Write the domain

Why: Open circle, round bracket; filled dot, square bracket.

\[ \text{domain} = (-3, 1\,] \]

Find the lowest and highest points

Why: The lowest is the filled dot at height negative four; the highest value, zero, happens only at the missing left end.

\[ \text{lowest: } y = -4 \text{ (filled)}, \qquad \text{highest: } y = 0 \text{ (open)} \]

Write the range

Why: Bottom value first, top value second.

\[ \text{range} = [-4, 0) \]

Check one height inside and two outside

Why: Height negative two is reached, height negative five is below everything the curve does, and height zero is only approached at the end that is missing.

\[ f(-1) = -2, \qquad -5 \notin \text{range}, \qquad 0 \notin \text{range} \]

OpenStax, Precalculus, §1.2 Domain and Range §1.2, pp. 48-49 — Example 1.21, Figures 1.25 and 1.26

Take the two questions one at a time and in the same order every time: horizontal extent, then vertical extent. Trying to see both at once is how people end up swapping the answers.

Horizontally, the curve runs from an open circle above negative three to a filled dot above one. Open circle, round bracket; filled dot, square bracket. That gives the domain straight away.

Vertically, the lowest the curve ever gets is the filled dot at negative four, so negative four is in. The highest it ever gets is zero, but look carefully at where that happens: only at the left end, and that end is an open circle. So the height zero is approached but never actually produced by any input that is in the domain, and the range gets a round bracket at the top.

That last point is the whole reason this example is in the book. An output belongs to the range only if some legal input produces it. Being between the lowest and the highest value is not enough on its own, although in this particular case every height strictly between negative four and zero really is reached.

30. An open dot at the top of the range

Prediction

\[ \text{range} = [-4, 0) \quad \text{or} \quad [-4, 0\,] \;? \]

Predict first

On the previous curve the value zero is produced only at the left endpoint, which is an open circle. Is zero in the range?

  • No, because the only input that would produce it is not in the domain
  • Yes, because the curve visibly touches height zero
  • Yes, because zero is between the lowest and highest values
  • It cannot be decided from the graph

Correct: No, because the only input that would produce it is not in the domain

Why: An output belongs to the range only if some input in the domain actually produces it. The curve rises to height zero exactly at the left end, and that end is an open circle, so no input in the domain gives zero. Being between the lowest and the highest value is not enough either, although here every value strictly between negative four and zero really is reached.

Commit to an answer first. This question has one job: to test whether you read the range off the picture or off the ends of the domain.

The value zero sits at the very top of what the curve does, and the curve visibly rises towards it. But an output is only in the range if some input in the domain actually produces it, and the single input that would produce zero is negative three, which was removed by that open circle.

It is the same logic that made the domain open at negative three, applied one axis over. One missing input can remove an output from the range, and here it removes exactly the one that would have been the maximum.

The option about being between the lowest and highest value is tempting and is wrong as a reason, even though it happens to give the right verdict for the interior values. Ranges are built from outputs that are actually produced, not from a region that looks plausible.

31. Example 1.22: thirty-five years of oil production

Worked example

Real data, read the same way. The picture is the function; there is no formula to look at.

Figure (svg): A graph of Alaskan crude oil production in thousands of barrels per day from 1973 to 2008: flat and low until 1977, climbing steeply to a peak just above two thousand in 1988, then falling steadily to under seven hundred by 2008.

Redrawn from the section's figure. Here the domain and the range are read off the data itself, and both are approximate: the extreme values do not land on gridlines.

Name the two quantities

Why: Years along the bottom, thousands of barrels per day up the side.

\[ t = \text{year}, \qquad b = \text{thousand barrels per day} \]

Read the horizontal extent

Why: The data begins in 1973 and stops in 2008.

\[ 1973 \le t \le 2008 \]

Read the vertical extent

Why: The lowest point is about 180 and the peak is about 2010.

\[ 180 \le b \le 2010 \]

Write both in interval notation

Why: Both ends are attained by the data, so both brackets are square.

\[ \text{domain} = [1973, 2008], \qquad \text{range} \approx [180, 2010] \]

Check what the two extremes mean

Why: The peak year is 1988, and production then fell for twenty years; the word approximately matters, because the extreme values do not sit on gridlines.

\[ b(1988) \approx 2010, \qquad b(2008) \approx 680 \]

OpenStax, Precalculus, §1.2 Domain and Range §1.2, p. 50 — Example 1.22

Real data behaves exactly like a drawn curve: the domain is still the horizontal extent and the range is still the vertical extent. The only difference is that you now have to say what the two axes mean before you can write anything down.

Name the variables first, as the book does, because an answer like from 1973 to 2008 is meaningless until you have said that the inputs are years. Once the variables are named, the reading is ordinary.

Then be honest about accuracy. The lowest and highest points do not land on gridlines, so the book writes about one hundred and eighty and about two thousand and ten and uses the word approximately. Pretending to a precision the graph cannot support is a real error, not a stylistic one.

It is worth looking at the shape for a moment too, because this is what a function often is in practice: not a formula but a record of something that happened. Production was tiny until the pipeline opened, climbed for a decade, peaked in 1988 and has fallen ever since.

32. Try It 1.18: a graph with no formula behind it

Estimation

A second data graph, of world population increase, runs from 1950 to about 2002. It rises from roughly 47 million a year, peaks near 89 million around 1989, and falls back to about 77 million.

Predict first

Which pair of intervals describes that graph?

  • domain from 1950 to 2002, range from 47 to 89 million
  • domain from 47 to 89 million, range from 1950 to 2002
  • domain from 1950 to 2002, range from 77 to 89 million
  • domain from 1950 to 1989, range from 47 to 89 million

Correct: domain from 1950 to 2002, range from 47 to 89 million

Why: The inputs are years, so the domain is the stretch of years covered, 1950 to 2002. The outputs are yearly increases in population, and they run from the smallest value shown, about 47 million, to the peak, about 89 million. The third option mistakes the last value for the smallest, and the fourth stops the domain at the peak instead of at the end of the data.

OpenStax, Precalculus, §1.2 Domain and Range §1.2, p. 50 — Try It 1.18, Figure 1.28

Pick an option before reading the explanation, and pick it by asking yourself one question first: which quantity is the input here?

The horizontal axis holds years, so years are the inputs and the domain is a stretch of years. The vertical axis holds the yearly increase in population, so those are the outputs and the range is a stretch of population figures. Two of the options are the right numbers attached to the wrong words.

For the range, read the lowest point and the highest point, not the first and the last. The curve starts near its lowest value, climbs to a peak around 1989 and then comes back down to a value that is higher than where it started, so the smallest output is at the very beginning and the largest is at the peak.

One option stops the domain at the peak year. That is a common slip when a graph has an obvious high point: the peak is interesting, but it is not where the data stops.

33. The toolkit functions

Section

Part 4

34. Three functions that accept everything

Picture it

Figure (svg): Three small graphs side by side: a horizontal line for the constant function, the diagonal line y equals x for the identity function, and the V shape of the absolute value function.

All three accept every input. Only the outputs differ: one height, every height, and every height from zero up.

The constant function returns the same output whatever you feed it, so its range is a single number. The identity returns the input untouched, so its range is everything. Absolute value returns a distance from zero, so its outputs stop at zero and go up.

OpenStax, Precalculus, §1.2 Domain and Range §1.2, pp. 51-52 — Figures 1.29, 1.30 and 1.31

These nine toolkit graphs are the ones you are expected to know without deriving them, and it is much less work than it sounds, because their domains and ranges follow from shapes you can picture.

The constant function is the strange one. It accepts every input and produces a single output, so its range is a set containing one number, written with braces. Some books write it as an interval from that number to itself, which is the same thing.

The identity function hands back whatever you give it, so its domain and its range are both everything. It is the only toolkit function for which the two answers are the same for an interesting reason rather than by accident.

Absolute value accepts everything too, but it reports a distance from zero, and a distance is never negative. So the inputs are the whole line and the outputs are the half-line from zero up, including zero itself, which comes from the input zero.

35. Powers and roots: which half do you lose?

Picture it

Figure (svg): Three small graphs: the upward parabola of the squaring function, the S-shaped curve of the cubing function, and the half-curve of the square root function starting at the origin.

An even power folds the negatives upward, so it costs you half the outputs. An even root refuses the negatives outright, so it costs you half the inputs.

Squaring sends negative inputs to positive outputs, so half the outputs are lost and the range stops at zero. The square root has the opposite problem: half the inputs are lost, because a negative number has no real square root. Cubing loses nothing at either end.

OpenStax, Precalculus, §1.2 Domain and Range §1.2, pp. 52-53 — Figures 1.32, 1.33 and 1.36

Look at these three side by side and ask, for each, whether it loses inputs or outputs. That single question organises the whole toolkit.

Squaring loses outputs. Every input is fine, but negative inputs come back positive, so no negative number is ever produced and the range stops at zero. The graph shows it: nothing below the horizontal axis.

The square root loses inputs instead. Negative numbers have no real square root, so the graph simply does not exist to the left of zero, and the domain starts at zero. Its outputs are also zero and up, because the square root symbol means the nonnegative root.

Cubing loses nothing at either end, and neither does the cube root. The reason is the same in both cases: an odd power keeps the sign of its input, so negative inputs give negative outputs and nothing gets folded or refused. Even means fussy, odd means relaxed, and that one sentence covers all four of these graphs.

36. The two that break at zero

Picture it

Figure (svg): Three small graphs: the two branches of one over x, the two upward branches of one over x squared, and the gentle S of the cube root function.

The two reciprocals lose the input zero. The cube root does not: an odd root of a negative number is a perfectly good negative number.

One over x refuses the input zero, and it also never produces the output zero, since a fraction with numerator one is never zero. One over x squared refuses zero as an input and produces only positive outputs. The cube root, by contrast, is perfectly happy with negative numbers.

toolkit functiondomainrange
constant, f(x) = c(-∞, ∞){c}
identity, f(x) = x(-∞, ∞)(-∞, ∞)
absolute value(-∞, ∞)[0, ∞)
squaring, x²(-∞, ∞)[0, ∞)
cubing, x³(-∞, ∞)(-∞, ∞)
reciprocal, 1/x(-∞, 0) ∪ (0, ∞)(-∞, 0) ∪ (0, ∞)
reciprocal squared, 1/x²(-∞, 0) ∪ (0, ∞)(0, ∞)
square root[0, ∞)[0, ∞)
cube root(-∞, ∞)(-∞, ∞)

OpenStax, Precalculus, §1.2 Domain and Range §1.2, pp. 53-55 — Figures 1.34, 1.35 and 1.37

The two reciprocals are the only toolkit functions with a hole in the domain, and it is the same hole in both: the input zero, because that is the one input that makes the denominator zero.

Their ranges differ, though, and that is the detail to notice. One over x produces every number except zero, positive ones from positive inputs and negative ones from negative inputs. One over x squared has a positive denominator whatever the input, so it only ever produces positive numbers, and its range loses the whole negative half as well as zero.

Why can neither produce zero? Because a fraction is zero only when its numerator is zero, and the numerator here is one. That argument recurs in Example 1.24 and is worth having ready.

The cube root is on this slide as a contrast. It looks vaguely like the square root graph reflected and extended, and it restricts nothing at all. Use the table underneath to test yourself: cover the last two columns and see how many of the nine you can state from the picture alone.

37. Match each toolkit function to its range

Matching

Match the pairs

  • a. f(x) = |x|
  • b. f(x) = x³
  • c. f(x) = 1/x
  • d. f(x) = 1/x²
  • e. f(x) = 7
  • w. [0, ∞)
  • x. (-∞, ∞)
  • y. (-∞, 0) ∪ (0, ∞)
  • z. (0, ∞)
  • v. {7}

Why: Absolute value produces zero and everything above it. Cubing produces every real number. One over x produces every number except zero, positive and negative both. Squaring the denominator forces the output positive, so zero drops out from below as well. A constant function produces one single number, and its range is the set containing just that number.

Do this from the shapes, not from memory of a table. For each function, picture its graph and ask how high and how low it goes, and whether it ever misses a value in between.

The two that need care are the reciprocals. Both lose zero from the range, but only one of them keeps the negative numbers. Squaring the denominator forces every output positive, which is why one of the answers is an interval with a round bracket at zero and the other is a union of two intervals.

The constant function is the odd one out on purpose. Its range is a set with a single element, so it is written with braces rather than brackets, and it is a good reminder that a range is a set of numbers and not always an interval.

If you got absolute value and cubing mixed up, go back a slide and look at the two pictures again. One folds the negative side up and one does not.

38. Example 1.23: a cubic with nothing to forbid

Worked example

With the toolkit shapes known, many domains and ranges can be written down without solving anything.

\[ f(x) = 2x^3 - x \]

Look for the two failure modes

Why: No denominator and no even root, so every input is allowed.

\[ \text{domain} = (-\infty, \infty) \]

Ask which outputs can occur

Why: A cubic with a positive leading coefficient falls away to the bottom left and climbs away to the top right, with no break in between.

\[ x \to -\infty \Rightarrow f(x) \to -\infty, \qquad x \to \infty \Rightarrow f(x) \to \infty \]

Write the range

Why: A continuous curve that reaches arbitrarily low and arbitrarily high hits every height on the way.

\[ \text{range} = (-\infty, \infty) \]

Check three heights

Why: A large positive output, a large negative one, and zero are all produced.

\[ f(2) = 14, \qquad f(-2) = -14, \qquad f(0) = 0 \]

OpenStax, Precalculus, §1.2 Domain and Range §1.2, p. 55 — Example 1.23

By now the domain should take you about two seconds. Scan for a denominator, find none; scan for an even root, find none; every real number is in.

The range takes a little more thought, and the argument is worth learning because it applies to every odd-degree polynomial. Far to the right the cubed term dominates and the outputs run off upward; far to the left they run off downward. The graph is unbroken, so on its way from very low to very high it must pass through every height in between.

That last sentence is doing real work, and you will meet it again as the intermediate value theorem in calculus. For now it is enough to see that a continuous curve cannot get from below a height to above it without crossing it.

Be careful not to over-generalise. This works because the degree is odd. An even-degree polynomial, like the squaring function, turns around and has a smallest or largest output, so its range is only half the line.

39. Example 1.24: one input missing, one output missing

Worked example

This one is the reciprocal toolkit function, shifted. Both the domain and the range lose exactly one number, and they are not the same number.

\[ f(x) = \frac{2}{x+1} \]

Find where the denominator is zero

Why: Set it to zero and solve.

\[ x + 1 = 0 \;\Longrightarrow\; x = -1 \]

Write the domain

Why: Delete that single input.

\[ \text{domain} = (-\infty, -1) \cup (-1, \infty) \]

Ask whether the output can be zero

Why: A fraction is zero only when its numerator is zero, and the numerator here is the constant 2.

\[ \frac{2}{x+1} = 0 \;\Longrightarrow\; 2 = 0, \text{ impossible} \]

Write the range

Why: Every height except zero is produced.

\[ \text{range} = (-\infty, 0) \cup (0, \infty) \]

Figure (svg): The graph of y equals two over x plus one, with two branches either side of a dashed vertical line at x equals minus one, flattening toward the x-axis without ever touching it.

One input is missing from the domain and one output is missing from the range, and they are different numbers: the missing input is minus one, the missing output is zero.

Check by solving for the input

Why: Given any target height other than zero, this formula names an input that produces it.

\[ y = \frac{2}{x+1} \;\Longrightarrow\; x = \frac{2}{y} - 1, \qquad y = 0.5 \Rightarrow x = 3 \]

OpenStax, Precalculus, §1.2 Domain and Range §1.2, p. 56 — Example 1.24

This example exists to stop you from assuming that the domain and the range answer the same question. Both of them here lose exactly one number, and the two numbers are different.

The domain is routine: set the denominator equal to zero, find negative one, delete it. The range needs the little argument about fractions. This fraction has numerator two, and a fraction is zero only when its numerator is zero, so the output zero is impossible. Every other height is produced.

The picture makes both facts visible at once. The dashed vertical line at negative one is the missing input, and the curve flattening towards the horizontal axis without ever touching it is the missing output. Two different kinds of gap, drawn in two different directions.

The check step is the honest way to confirm a range. Rather than staring at the graph, solve the equation for x in terms of the target height. If the formula you get makes sense for every height except zero, then every height except zero is in the range.

40. Example 1.25: a stretched square root

Worked example

The square root shape again, moved four to the left and stretched to twice the height.

\[ f(x) = 2\sqrt{x+4} \]

Set the radicand at or above zero

Why: The root is the only thing that can fail.

\[ x + 4 \ge 0 \;\Longrightarrow\; x \ge -4 \]

Write the domain

Why: The endpoint is allowed, so the bracket is square.

\[ \text{domain} = [-4, \infty) \]

Find the smallest output

Why: A square root is smallest when its inside is zero.

\[ f(-4) = 2\sqrt{0} = 0 \]

Ask how large the output can get

Why: The inside grows without bound, and so does its root, and doubling does not change that.

\[ x \to \infty \;\Longrightarrow\; 2\sqrt{x+4} \to \infty \]

Write the range

Why: From zero upward, with zero included.

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

Figure (svg): The graph of y equals two times the square root of x plus four, starting at a filled dot at minus four on the x-axis and rising to the right, with a heavy arrow along the x-axis from minus four and a heavy arrow up the y-axis from zero.

The graph never turns back and never comes down, so the smallest output is the one at the left end and there is no largest.

Check two inputs against the picture

Why: Both land on the curve where the graph says they should.

\[ f(0) = 2\sqrt{4} = 4, \qquad f(5) = 2\sqrt{9} = 6 \]

OpenStax, Precalculus, §1.2 Domain and Range §1.2, p. 56 — Example 1.25

Recognise the shape first. This is the square root toolkit graph, shifted four units to the left and stretched to twice the height, which is enough to predict both answers before doing any algebra.

The domain comes from the only hazard, the root. The inside must be at or above zero, so the graph begins at negative four and continues right forever, with a square bracket at the start because zero inside a root is fine.

For the range, notice that the curve never turns around: it starts at its lowest point and climbs from there. So the smallest output is the one at the left endpoint, which is zero, and there is no largest, because the inside of the root grows without bound and so does the root.

Do not let the factor of two worry you. Doubling every output stretches the graph vertically but changes neither which inputs are legal nor the fact that the outputs start at zero and go up. Multiplying by a negative number would be a different matter, and the Try It on the next slide does exactly that.

41. Try It 1.19: the same root, reflected

Fill the middle

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

Fill in the blanks

\text2 = \left(-\infty, \; 0\,\right], \qquad \text___ = \left(-\infty, \; ___\,\right]

Why: Two minus x is at or above zero exactly when x is at most two, so the domain stops at two and two itself is allowed. For the range, the square root produces zero and everything above, and the minus sign in front flips all of that below the axis, so the outputs are zero and everything below. The largest output, zero, happens at the input two.

OpenStax, Precalculus, §1.2 Domain and Range §1.2, p. 56 — Try It 1.19

Two blanks, and they come from two different pieces of reasoning, so do them separately.

For the domain, the inside of the root is two minus x, and it has to be at or above zero. That happens when x is at most two. Notice that the variable is being subtracted here, so the allowed inputs run to the left rather than to the right, exactly as in Example 1.19.

For the range, start from what the square root alone produces, which is zero and everything above. The minus sign in front then flips all of that to the other side of zero, so the outputs are zero and everything below. The largest output is zero, and it comes from the input two.

It is a good habit to sketch this in your head: the ordinary square root graph, reflected left to right by the minus inside, and reflected top to bottom by the minus outside. Two reflections, two answers, and neither of them is the answer you would get by guessing.

42. Slide the inside of the root

Tweak it

Parameter explorer

This is the square root of the quantity x minus h. Drag h and watch where the curve begins. What is the domain in terms of h, and does the range change at all?

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

  • h — from -4 to 6: shift h

Drag the slider slowly and watch only one thing: where the curve begins. The left end of the curve moves exactly with the parameter, because the curve begins where the inside of the root is zero.

So the domain is every input from the parameter onward, with a square bracket at that point. When the parameter is positive the graph shifts right, and when it is negative the graph shifts left, which is the direction most people guess wrong. Subtracting inside the function moves the graph in the opposite direction, and this is the cleanest place in the course to see why: the starting point is wherever the inside becomes zero.

Now watch the vertical extent, which does not move at all. Whatever you do to the inside of the root, the outputs are still zero and up, because the square root symbol cannot return a negative number.

That is the takeaway to carry into Section 1.5 on transformations: changes inside a function affect the domain, and changes outside it affect the range.

43. Trap: assuming the range is whatever the domain is

Trap

The trap

A very common pair of lines about the reciprocal squared function:

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

\[ \text{domain} = (-\infty,0) \cup (0,\infty) \;\Rightarrow\; \text{range} = (-\infty,0) \cup (0,\infty) \]

Wrong. The domain and the range are answers to two different questions.

The fix

Squaring the denominator makes every output positive, so the whole negative half of the range is gone as well as zero.

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

Get into the habit of finding the domain from the formula and the range from the shape of the graph. They agree for some functions, such as the identity and the cube root, and disagree for most.

This trap is a habit of reasoning rather than an algebraic slip. Having done the work to find the domain, it is tempting to reuse the answer, and for a couple of toolkit functions that even works, which is what makes the habit stick.

It fails here for a reason you can see in one line. Squaring the denominator makes it positive whatever the input, so the whole fraction is positive whatever the input, and every negative output is impossible as well as zero.

The general lesson is that the two questions have different sources. The domain comes from the formula, by hunting for hazards. The range comes from the behaviour of the outputs, which usually means the shape of the graph. Two questions, two methods.

When you do find that the domain and the range agree, make sure you can say why. For the identity and the cube root it is because the graph covers the whole plane diagonally; for one over x it is because the graph is symmetric in a way that swaps the two axes.

44. Piecewise-defined functions

Section

Part 5

45. Absolute value is two rules wearing one coat

Concept

Sometimes one formula cannot describe a function over its whole domain. The absolute value is the first example you have already met: it leaves nonnegative inputs alone and flips negative ones.

\[ |x| = \begin{cases} x & \text{if } x \ge 0 \\ -x & \text{if } x < 0 \end{cases} \]

piecewise function — A function defined by more than one formula, each formula carrying its own piece of the domain. The domain of the whole function is the union of those pieces.

Rules like this appear whenever a situation changes at a boundary: a bulk discount above a certain order size, a higher tax rate above a certain income, a different postage rate above a certain weight.

OpenStax, Precalculus, §1.2 Domain and Range §1.2, pp. 56-57 — graphing piecewise-defined functions

You have used absolute value for years without thinking of it as two rules, but that is exactly what it is. Nonnegative inputs come back unchanged; negative inputs come back with the sign flipped. There is no single ordinary formula that does that.

Read the notation carefully, because the layout is doing work. The left column holds formulas and the right column holds conditions, and the brace says that these are alternatives, not a system of equations to solve together.

The reason to bother is that lots of real situations change their rule at a boundary. Postage jumps at a certain weight, a discount starts at a certain quantity, a tax rate changes at a certain income. None of those can be written with one formula, and all of them can be written with a brace.

Everything you already know still applies. A piecewise function is still a function, so it must give exactly one output per input, and that requirement is what forces the conditions to be written carefully.

46. Reading a piecewise definition piece by piece

Notation

Annotate

On: \( f(x) = \begin{cases} 5n & \text{if } 0 < n < 10 \\ 50 & \text{if } n \ge 10 \end{cases} \)

  • One function, not two. The brace collects the alternatives, and exactly one of them applies to any given input.
  • The left column is what to compute. Each line may be any rule at all — a line, a parabola, a constant.
  • The right column says when to use that line. Read the condition FIRST, decide which line you are on, and only then compute.
  • The conditions must not both catch the same input (that would break the one-output rule) and any input they both miss is simply not in the domain.

Step through the annotations and notice the reading order, because it is the opposite of the order people naturally use. You read the conditions on the right first, to decide which line applies, and only then read the formula on the left.

Doing it the other way round is the single commonest error with piecewise functions, and there is a slide on it later in this part. The formula on the top line is not the default; it is just the one that happens to be written first.

The last annotation carries two requirements that are easy to state and easy to forget. No input may satisfy two conditions, or the rule would hand out two outputs and stop being a function. And any input satisfying none of the conditions is simply not in the domain, which is a legitimate thing for a rule to do, as long as you notice it.

When you write a piecewise function of your own, check both requirements explicitly. Look at each boundary value and ask which line, if any, claims it.

47. Example 1.26: the museum's two price rules

Worked example

A museum charges 5 dollars a head for a guided tour of 1 to 9 people, or a flat 50 dollars for a group of 10 or more. Write the cost as a function of the group size.

Split the inputs at the boundary

Why: Small groups obey one rule, large groups another; the boundary is 10 people.

\[ 0 < n < 10 \qquad \text{and} \qquad n \ge 10 \]

Write the rule for a small group

Why: Five dollars for each of the n people.

\[ C = 5n \]

Write the rule for a large group

Why: A flat fee, whatever the size.

\[ C = 50 \]

Put both in one definition

Why: One brace, one line per rule, each line carrying its own condition.

\[ C(n) = \begin{cases} 5n & \text{if } 0 < n < 10 \\ 50 & \text{if } n \ge 10 \end{cases} \]

Figure (svg): A cost graph against group size: a straight line from the origin rising five dollars per person up to ten people, then a flat line at fifty dollars, with a filled dot where they meet.

Two rules, one graph. The join happens to be smooth here because five times ten is also fifty, but a piecewise function is under no obligation to join up.

Check on either side of the boundary

Why: Nine people pay 45 dollars under the first rule and ten pay 50 under the second, so nobody is charged twice and nobody is missed; the two rules happen to agree at ten.

\[ C(9) = 45, \qquad C(10) = 50, \qquad 5(10) = 50 \]

OpenStax, Precalculus, §1.2 Domain and Range §1.2, pp. 57-58 — Example 1.26

This is the modelling direction: a situation in words, and your job is to produce the brace. Start by finding the boundary, because that is what makes the rule piecewise in the first place. Here it is a group of ten.

Then write each rule separately, in the simplest form you can. Small groups pay five dollars a head, which is five times the number of people. Large groups pay a flat fifty, which does not involve the number of people at all, and a formula is allowed to ignore its input.

Now write the conditions to match the words exactly. A group of one to nine gets the first rule, and ten or more gets the second, so the second condition allows equality and the first does not. Getting that boundary wrong by one is the usual slip.

The check at the end is worth doing every time. Nine pays forty-five, ten pays fifty, and no group size gets asked to pay twice or is left with no price at all. Here the two rules happen to agree at ten, which makes the graph join up smoothly, but a piecewise function is under no obligation to do that.

48. Example 1.27: the cost of four gigabytes

Worked example

A phone company charges 25 dollars for anything under 2 gigabytes, and 10 dollars for each gigabyte after the second. Find the cost of 1.5 gigabytes and of 4 gigabytes.

\[ C(g) = \begin{cases} 25 & \text{if } 0 < g < 2 \\ 25 + 10(g-2) & \text{if } g \ge 2 \end{cases} \]

Locate 1.5 among the conditions

Why: 1.5 lies between 0 and 2, so the first line applies.

\[ 0 < 1.5 < 2 \;\Longrightarrow\; C(1.5) = 25 \]

Locate 4 among the conditions

Why: 4 is at least 2, so the second line applies.

\[ 4 \ge 2 \;\Longrightarrow\; C(4) = 25 + 10(4-2) \]

Do the arithmetic

Why: Two gigabytes past the boundary, at ten dollars each.

\[ C(4) = 25 + 20 = 45 \]

Figure (svg): A cost graph against gigabytes: a flat line at twenty-five dollars up to two gigabytes, then a straight line rising ten dollars per gigabyte, with marked points at one and a half gigabytes and at four gigabytes.

The rule changes at two gigabytes, which is why the graph changes direction there. Reading a value means first deciding which side of two your input is on.

Check the boundary itself

Why: At exactly two gigabytes the second line applies and charges nothing extra, so the graph bends without jumping.

\[ C(2) = 25 + 10(0) = 25 \]

OpenStax, Precalculus, §1.2 Domain and Range §1.2, pp. 58-59 — Example 1.27

This is the evaluation direction, and there is a script for it that you should follow every single time: find which condition your input satisfies, then use that line, then compute.

For one and a half gigabytes the first condition holds, and the first line is a constant, so the answer is twenty-five dollars with no arithmetic at all. That feels too easy and is nonetheless correct: under two gigabytes, everyone pays the same.

For four gigabytes the second condition holds. The formula charges the base twenty-five plus ten dollars for each gigabyte beyond the second, and four is two gigabytes beyond, so the extra is twenty dollars. Notice that the formula subtracts two rather than using four directly, because the extra charge starts from the boundary, not from zero.

The check at the boundary is the interesting one. At exactly two gigabytes the second line applies and charges nothing extra, giving the same twenty-five dollars as the first line. That is why the graph bends at two instead of jumping, and it is what a sensible pricing plan looks like.

49. Example 1.28: graphing three pieces

Worked example

Graph each piece as though it were the whole function, then keep only the part of it that its condition allows.

\[ f(x) = \begin{cases} x^2 & \text{if } x \le 1 \\ 3 & \text{if } 1 < x \le 2 \\ x & \text{if } x > 2 \end{cases} \]

Mark the boundaries on the x-axis

Why: The rule changes at 1 and again at 2.

\[ x = 1 \qquad \text{and} \qquad x = 2 \]

Draw the first piece up to its boundary

Why: A parabola, stopped at x equal to 1, where the condition allows equality, so the dot is filled.

\[ f(1) = 1^2 = 1 \quad \text{(filled dot)} \]

Draw the middle piece between the boundaries

Why: A flat height of 3, open at the left end because 1 is not allowed there, filled at the right end because 2 is.

\[ f(1) \ne 3 \quad \text{(open)}, \qquad f(2) = 3 \quad \text{(filled)} \]

Draw the last piece from the second boundary on

Why: The line y equals x, starting just after 2, so its left dot is open.

\[ f(2) \ne 2 \quad \text{(open)} \]

Figure (svg): A piecewise graph: a parabola up to a filled dot at one comma one, a flat segment at height three from an open circle at x equals one to a filled dot at x equals two, then the line y equals x starting from an open circle at two comma two.

Above the input 1 there are two dots, but only one is filled, so the vertical line test still passes. The same is true above the input 2.

Check the vertical line test at both boundaries

Why: Above the input 1 there is a filled dot at height 1 and an open circle at height 3, so only one point is really on the graph, and the same happens above the input 2.

\[ f(1) = 1, \qquad f(2) = 3 \]

OpenStax, Precalculus, §1.2 Domain and Range §1.2, pp. 59-60 — Example 1.28, Figures 1.41 and 1.42

The method is exactly what it sounds like: graph each formula as though it were the whole function, then rub out everything its condition does not allow. All three formulas here are toolkit shapes you already know, so no plotting of points is needed.

The care goes into the endpoints. At each boundary, look at which piece is allowed to include that input. The first condition allows equality at one, so the parabola gets a filled dot there; the middle condition does not, so the flat piece gets an open circle at the same input but at a different height.

That pair of dots above the input one is worth staring at. One filled, one open, and the graph passes the vertical line test because only the filled one is really a point of the graph. The same thing happens above the input two, with the roles of the pieces swapped.

When you sketch one of these yourself, draw the boundary lines lightly first. It makes it much harder to let a piece run past its condition, which is the other common error here.

50. Try It 1.20: conditions that leave two gaps

Worked example

Graph this one, then write its domain. The three conditions do not quite cover the real line.

\[ f(x) = \begin{cases} x^3 & \text{if } x < -1 \\ -2 & \text{if } -1 < x < 4 \\ x & \text{if } x > 4 \end{cases} \]

Graph each piece on its own condition

Why: A cubic, then a flat height, then a diagonal line.

\[ x^3 \text{ on } x<-1, \quad -2 \text{ on } -1<x<4, \quad x \text{ on } x>4 \]

Make every endpoint an open circle

Why: Every condition here is a strict inequality, so no boundary input belongs to any piece.

\[ f(-1) \text{ and } f(4) \text{ are undefined} \]

Figure (svg): A piecewise graph with open circles at both boundaries: a cubic coming up from the bottom left to an open circle at minus one, a flat segment at height minus two between open circles at minus one and four, then the line y equals x from an open circle at four.

Every endpoint here is an open circle, so the two boundary inputs belong to no piece at all and are simply not in the domain.

Write the domain as the union of the conditions

Why: Three intervals, with two single inputs missing from the real line.

\[ \text{domain} = (-\infty, -1) \cup (-1, 4) \cup (4, \infty) \]

Check that no input receives two rules

Why: The conditions never overlap, so every input in the domain gets exactly one output, and this is still a function despite the holes.

\[ -1 \notin \text{domain}, \qquad 4 \notin \text{domain} \]

OpenStax, Precalculus, §1.2 Domain and Range §1.2, p. 60 — Try It 1.20

Graph it the same way as the last example, and then look hard at the boundaries, because this one is different in a way that is easy to miss.

Every condition here is a strict inequality. The first uses below negative one, the second uses strictly between, the third uses above four. So neither negative one nor four is claimed by any piece, and neither of them is in the domain at all.

That means every endpoint on this graph is an open circle: four of them, in pairs above the two boundary inputs. The picture is the quickest way to see it, and the red dashed lines mark the two inputs where the function simply has no value.

So the domain is three intervals joined by unions, with two single points missing from the real line. This is still a perfectly good function. Nothing requires a domain to be an unbroken stretch, and nothing requires a piecewise rule to cover everything, although in a modelling problem a gap like this usually means somebody has written the conditions carelessly.

51. Exercise 141: finish the four evaluations

Faded example

\[ f(x) = \begin{cases} 5x & \text{if } x < 0 \\ 3 & \text{if } 0 \le x \le 3 \\ x^2 & \text{if } x > 3 \end{cases} \]

Fill in the blanks

f(-1) = -5, \qquad f(0) = 3, \qquad f(2) = 3, \qquad f(4) = 16

Why: Choose the line before you compute. Negative one is below zero, so five times negative one gives negative five. Zero and two both satisfy the middle condition, so both outputs are the constant three, even though the inputs differ. Four is above three, so square it to get sixteen. The most common error is squaring the input two, which uses a rule that does not apply there.

OpenStax, Precalculus, §1.2 Domain and Range §1.2, p. 64 — Exercise 141

Four inputs, three conditions, and the only skill being tested is choosing the right line before computing. Say the condition out loud for each input.

Negative one is below zero, so the first line applies and the answer is five times negative one. Zero and two both satisfy the middle condition, which allows equality at both ends, so both give the constant three. Four is above three, so the last line applies and the answer is four squared.

The most common wrong answer here is squaring the input two, because the squaring formula is sitting there in the definition and looks inviting. It belongs to inputs above three, and two is not above three.

Notice the pair of inputs that give the same output. Zero and two both return three, which is fine: a function may repeat an output. What would not be fine is one input returning two different values, which is what overlapping conditions would cause.

52. Find the error: the wrong line at the boundary

Error analysis

Annotate

On: \( f(x) = \begin{cases} x^2 - 2 & x < 2 \\ 4 + |x-5| & x \ge 2 \end{cases} \quad\Longrightarrow\quad f(2) = 2^2 - 2 = 2 \)

  • The input is 2. The first condition asks for x < 2, which is false at x = 2. The second asks for x ≥ 2, which is true.
  • The top formula was chosen because it was written first, not because its condition was met. Read the right-hand column before the left-hand one.
  • Use the second line: 4 + |2 − 5| = 4 + 3 = 7, not 2.
  • At a boundary the two formulas usually disagree. Here one gives 2 and the other gives 7, so picking the wrong line is not a small error.

OpenStax, Precalculus, §1.2 Domain and Range §1.2, p. 64 — Exercise 140

Click through the annotations in order and diagnose the error before reading the last one. Everything about this calculation is arithmetically correct; the mistake happened before any arithmetic started.

The input is two. The first condition demands strictly less than two, which is false. The second allows equality, which is true. So the second line is the one that applies, and the first line has no claim on this input at all.

The reason the error happens is worth naming: the top formula was chosen because it was written first. A piecewise definition is not a priority list; it is a set of alternatives, each guarded by its own condition, and the guard is what you read first.

Look at how much the error costs. The two formulas give two and seven at this input, so this is not a small slip. Boundary inputs are exactly where the formulas disagree most visibly, which is exactly why examination questions ask for them.

53. Trap: conditions that overlap

Trap

The trap

A piecewise rule written in a hurry:

\[ f(x) = \begin{cases} x + 1 & x \le 3 \\ 2x & x \ge 3 \end{cases} \]

Wrong. At the input 3 both conditions hold, and the two lines disagree.

The fix

The first line gives 4 and the second gives 6, so the input 3 would have two outputs, and the rule would not be a function at all. Make one of the two conditions strict.

\[ f(x) = \begin{cases} x + 1 & x < 3 \\ 2x & x \ge 3 \end{cases} \]

Check both directions whenever you write a piecewise rule: no input caught twice, and no input in the intended domain missed altogether.

This is the other half of the boundary problem, and it happens when you write a piecewise rule rather than read one. Both conditions here allow equality at three, so the input three is caught twice.

That is fatal, not untidy. The first line would give four and the second would give six, so a single input would have two outputs, and the rule fails the definition of a function before you even start graphing it.

The fix is always the same and always small: make one of the two conditions strict. Which one you choose is up to you, and the two choices give genuinely different functions, differing at exactly one input.

Whenever you write a brace, run two checks along the boundaries. Is any input claimed twice? Is any input in the intended domain claimed by nobody? The first is an error; the second may be deliberate, as in the last Try It, but it should be deliberate rather than accidental.

54. Two tax brackets, one function

Real world

Figure (svg): A graph of tax against income: a shallow line at ten percent up to ten thousand dollars, then a steeper line at twenty percent, with a marked bend at ten thousand and a marked point at forty-five thousand dollars.

The graph bends but does not jump: crossing into the higher bracket changes the slope, not the total, so a raise never leaves you worse off.

Discussion prompt

Income up to 10 thousand dollars is taxed at 10 percent, and anything above that at 20 percent. Write the tax as a piecewise function of the income S, then use it to find the tax on 45 thousand dollars.

OpenStax, Precalculus, §1.2 Domain and Range §1.2, p. 57 — the tax-bracket discussion

Try this before revealing it, because it is the piecewise idea in the setting where most people first meet it in real life, and because there is a widespread misunderstanding attached to it.

The structure is the same as the phone plan. Below the boundary there is a simple rule, ten percent of everything. Above the boundary, you pay the full tax on the first ten thousand, which is one thousand, plus the higher rate on the part above the boundary only. That is why the second line subtracts ten thousand before multiplying.

On forty-five thousand of income the tax comes to eight thousand, which is about seventeen point eight percent of the income. Notice that this sits between the two bracket rates and is below the higher one. That is always true, and it is the answer to the common worry that earning slightly more could leave you worse off.

Look at the graph. It bends at the boundary but does not jump, because the higher rate applies only to the income above the boundary. A rule that jumped would be one where crossing a line by a single dollar costs you a lump sum, and tax systems are deliberately not written that way.

55. When the story restricts the domain further

Edge cases

\[ h(t) = -16t^2 + 96t \]

Discussion prompt

Exercise 152. This gives the height in feet of a projectile t seconds after launch. The formula is a polynomial, so it accepts every real number. What is the domain of the actual situation, and why?

OpenStax, Precalculus, §1.2 Domain and Range §1.2, p. 65 — Exercise 152

Everything so far has treated a formula as if it arrived on its own. This slide is about what changes when the formula is modelling something.

Algebraically this is a polynomial, so it accepts every real number, including negative ones and very large ones. But the variable is time since launch and the output is a height, so negative times are meaningless and times after landing are describing a projectile that is no longer in the air.

Factoring finds the landing time: the height is zero at launch and again at six seconds. So the domain of the model is from zero to six, with both ends included, because launch and landing are both real moments. Feed the formula eight seconds and it politely reports a height of negative two hundred and fifty-six feet, which is not a fact about the world.

The general rule is worth stating: a story can only narrow the implied domain, never widen it. If the algebra forbids an input, no amount of context makes it legal; but context can rule out inputs that the algebra was perfectly happy with.

56. Putting it together

Section

Part 6

57. Pattern: finding a domain, then a range

Pattern

Figure (svg): A flow diagram: from the rule, three questions branch out - is there a denominator, is there an even root, is there neither - each feeding into a box that says combine every surviving condition, which leads to a final box about writing the answer in intervals and then reading the range.

One pass down the formula answers the domain. The range is a separate question, and a picture usually answers it faster than algebra does.
  1. Scan the formula for a denominator. Set it equal to zero, solve, and delete those inputs.
  2. Scan it for an even root. Set the radicand at or above zero, solve, and keep those inputs.
  3. Impose every condition at once, not one after the other. A root in a denominator must be positive, not merely nonnegative.
  4. Write the survivors in interval notation, smallest first, unions where the set breaks apart, round brackets at any infinity.
  5. Then go after the range from the shape: a toolkit graph, the endpoints, and whether the curve turns around.
  6. If there is a story, check whether it narrows the domain further, as negative time or a negative group size would.

This is the whole lesson as a procedure, and the diagram is the order to work in. Scan, condition, combine, write; and only then start thinking about the range.

The step people skip is the third one. Conditions have to hold at the same time, not one after the other. That matters most when a root sits in a denominator, where nonnegative and nonzero combine into strictly positive, and the endpoint that would have been included is quietly lost.

The range gets its own arrow on the diagram because it is a different kind of question with a different method. There is no scan you can do to a formula that produces a range. What you have instead is the shape: recognise the toolkit graph, find the endpoints, notice whether the curve turns around, and check whether any output is impossible for a reason like a nonzero numerator.

The last line of the list is the one that separates a good answer from a full one. If the function is modelling something, ask what the situation forbids as well as what the algebra forbids.

58. Put a domain calculation in order

Ranking

Put in order

Order the steps for finding the domain of a rule with a square root in its denominator.

  1. Read the formula and spot both the root and the denominator
  2. Require the radicand to be at or above zero
  3. Require the denominator not to be zero
  4. Combine the two conditions into one
  5. Write the surviving set in interval notation

Why: Spotting the two hazards comes first, because it decides how many conditions there will be. Each hazard is then turned into its own condition, and only after that are they combined, since combining is where nonnegative and nonzero become strictly positive. Writing the interval is the last step, not the first.

Order these before reading the explanation. The order is not arbitrary; each step needs the one before it to have been done properly.

Reading the formula comes first because it tells you how many conditions you are going to have. Someone who starts solving immediately usually finds one condition, writes it up, and loses the marks for the second.

The two conditions then get written down separately, one from the root and one from the denominator. Keeping them apart at this stage is deliberate: each is easy on its own, and mixing them up while solving is where errors creep in.

Combining comes fourth, and it is the step with the most thinking in it, because that is where nonnegative and nonzero become strictly positive. Writing the interval is last, a translation of a decision already made, not the decision itself.

59. Check: a root over a difference

Check

Check your understanding

What is the domain of the square root of the quantity x plus four, all divided by x minus four?

  • A. [-4, ∞)
  • B. [-4, 4) ∪ (4, ∞) (correct)
  • C. (-∞, 4) ∪ (4, ∞)
  • D. (-4, 4) ∪ (4, ∞)

Answer: B

Why: Two conditions. The root needs x plus four at or above zero, so x is at least negative four. The denominator needs x minus four not to be zero, so four is deleted. Imposing both gives everything from negative four up, with the single input four punched out.

Why A tempts people
This uses only the root condition and forgets that the denominator vanishes at four.
Why C tempts people
This uses only the denominator condition; inputs below negative four would make the root imaginary.
Why D tempts people
Negative four is allowed: the radicand is zero there, which is fine, and the denominator is negative eight, which is also fine.

Two hazards, so two conditions, and a correct answer has to respect both. Work them out separately and then combine them, exactly as the pattern slide says.

The root needs its inside at or above zero, which starts the domain at negative four and includes it. The denominator must not be zero, which removes the single input four. Combining gives everything from negative four up, with a hole at four.

Two of the wrong options are each half of the right answer, which is the point of the question. If you apply only one condition you get an answer that looks perfectly respectable and is incomplete.

The remaining wrong option is about the endpoint. At negative four the radicand is zero, which is allowed, and the denominator is negative eight, which is also fine, so negative four really is in the domain and its bracket really is square.

60. Check: reading both answers off a graph

Check

Check your understanding

A graph starts at a filled dot at (-2, 1), rises smoothly to a highest point at height 5, then falls to an open circle at (6, 0). What are the domain and the range?

  • A. domain [-2, 6), range (0, 5] (correct)
  • B. domain [-2, 6], range [0, 5]
  • C. domain (0, 5], range [-2, 6)
  • D. domain [-2, 6), range [1, 5]

Answer: A

Why: The horizontal extent runs from a filled dot at negative two to an open circle at six, so the domain is closed on the left and open on the right. The outputs climb from one to five and then fall all the way down toward zero, so every height above zero and up to five occurs, with zero itself missed because its dot is open.

Why B tempts people
The right-hand end is an open circle, so six is not an input and zero is not an output.
Why C tempts people
This swaps the two answers: the domain is read along the horizontal axis, the range along the vertical one.
Why D tempts people
The curve comes back down past one on its way to zero, so the range does not stop at the starting height.

There is no picture here on purpose: the description contains everything you need, and reading a description carefully is a skill of its own.

Domain first. A filled dot at the left end and an open circle at the right end, so the interval is square on the left and round on the right. That eliminates two options immediately.

Range next, and this is where the question bites. The outputs start at one, climb to five, and then fall all the way down towards zero, so the outputs cover everything from just above zero up to five. The starting height of one is not the smallest output, because the curve comes back down past it on the way to the open circle.

The fourth option makes exactly that mistake, treating the left endpoint as the bottom of the range. Endpoints are worth checking, but so is everything the curve does in between.

61. Check: which piece applies

Check

Check your understanding

For f(x) equal to x squared minus two when x is below two, and 4 plus the absolute value of x minus five when x is at least two, what is f(2)?

  • A. 2
  • B. 7 (correct)
  • C. 4
  • D. undefined, because 2 is a boundary

Answer: B

Why: The input two fails the first condition, which demands strictly less than two, and satisfies the second, which allows equality. So use the second formula: four plus the absolute value of two minus five is four plus three, which is seven.

Why A tempts people
This uses the first formula, whose condition is false at the input two.
Why C tempts people
This drops the absolute-value term instead of evaluating it; the absolute value of negative three is three, not zero.
Why D tempts people
A boundary input is undefined only when no condition catches it. Here the second condition does, because it allows equality.

Read the conditions before you read the formulas. This is the same test as the error-analysis slide, asked in a form where you have to do the whole job yourself.

The input two fails the first condition, which demands strictly less than two, and satisfies the second, which allows equality. So the second formula is the one to use, and the answer is four plus the absolute value of negative three, which is four plus three.

The most tempting wrong answer comes from the first formula, which happens to give a clean number. A clean answer from the wrong line is still wrong, and there is no partial credit for arithmetic done in the wrong place.

The last option is worth a moment. A boundary input is undefined only when no condition catches it, as in the Try It with two gaps. Here a condition does catch it, so the function is perfectly well defined at two.

62. Why an equation for one and an inequality for the other

Explain it to yourself

Discussion prompt

In two or three sentences, explain why a denominator leads you to solve an equation and then delete what you find, while an even root leads you to solve an inequality and then keep what you find.

Write two or three sentences before revealing anything. Being able to say why a method works is what turns it from a ritual into a tool, and this is the single most useful why in the section.

The short version is about how many inputs each hazard ruins. A denominator is zero at only a handful of inputs, the solutions of one equation, so the efficient move is to find those few and delete them from a domain that was otherwise everything.

An even root, by contrast, is unusable on a whole half-line of inputs. Listing the bad ones is hopeless, so you turn the question around and describe the good ones with an inequality. The solutions of that inequality are the domain, not the enemies of it.

If you can say that out loud, you will never again solve a radicand equal to zero and delete the answer, which is the error this distinction exists to prevent.

63. Exit ticket

Exit ticket

\[ f(x) = \frac{2x+1}{\sqrt{5-x}} \]

Discussion prompt

Give the domain of this rule in interval notation and in set-builder notation, and say in one sentence why the endpoint five is treated the way it is.

OpenStax, Precalculus, §1.2 Domain and Range §1.2, p. 63 — Exercise 113

Everything from this lesson is in this one rule, so take it slowly and write out the reasoning rather than just the answer.

There is a root, so its inside must be at or above zero, which gives inputs at most five. The root is also in the denominator, so it must not be zero, which removes five itself. Nonnegative and nonzero together mean strictly positive, so the domain is everything strictly below five.

That is why the bracket at five is round rather than square, and it is the difference between this rule and the one in Example 1.19, where the root stood on its own and five would have been allowed.

Write the answer in both notations while you are here. Interval notation is quicker to write and set-builder notation is easier to justify, and examiners ask for either without warning.

64. Recap

Recap

questionwhere the answer comes fromwhat to watch
domain of a formulathe two failure modes: a zero denominator, a negative even radicandimpose every condition at once
domain from a graphthe horizontal extent, left to rightfilled dot in, open circle out
range from a graphthe vertical extent, bottom to topcheck the endpoints, not only the turning points
a toolkit functionmemory: the nine shapes and their two answersthe range is rarely the same as the domain
a piecewise rulethe condition column, read before the formula columnno input caught twice, and note any input missed
a modelled situationthe story, on top of the algebrathe story can only narrow the domain

\[ \text{domain} = \{x \mid f(x) \text{ is a real number}\}, \qquad \text{range} = \{f(x) \mid x \in \text{domain}\} \]

Next, Section 1.3 takes rates of change and behaviour of graphs, where reading an interval off a picture becomes a daily job.

OpenStax Algebra and Trigonometry 2e, §3.2 Domain and Range §3.2 — the same material in OpenStax Algebra and Trigonometry

Six questions, six sources of answer. If you can look at the left column and say where the answer comes from without reading the middle column, this lesson has done its job.

The two habits most worth keeping are these. Find a domain by hunting for hazards in the formula, not by testing numbers. Find a range from the behaviour of the outputs, usually from a graph, and never by copying the domain.

The notation is not a decoration on top of the mathematics; it is where most of the marks are lost. Smaller value first, square brackets only where the end value is genuinely attained, round brackets at every infinity, and a union whenever the set falls into pieces.

Next section, rates of change and the behaviour of graphs, asks where a function is increasing and where it is decreasing, and every one of those answers is an interval. You will be writing these brackets for the rest of the course.

Sources

  1. OpenStax, Precalculus, §1.2 Domain and Range
  2. OpenStax Algebra and Trigonometry 2e, §3.2 Domain and Range

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

Book on Wyzant · Text (657) 465-8108