Defines a linear function by the property that its rate of change is constant, which is exactly what makes its graph straight. Computes slope from two points, from a table and from a description, interprets its sign and its units, and introduces the slope-intercept and point-slope forms along with the special cases of horizontal and vertical lines.
Subject: Precalculus · 69 slides · symbolic lesson
Open the interactive version of this deck
Title
Precalculus · Section 2.1
The one shape whose rate of change never changes
Objectives
Chapter 1 asked what a function is. This chapter takes the simplest interesting family of functions and takes it apart completely, because almost every model you meet later starts by pretending something is linear.
OpenStax, Precalculus, §2.1 Linear Functions §2.1, pp. 181-204 — learning objectives 2.1.1 to 2.1.5
Chapter 1 gave you the idea of a function and the idea of an average rate of change. This lesson puts the two together in the simplest possible way: what if the rate of change never changes? Everything in the section falls out of that one condition.
You will meet two ways to write such a function down. Slope-intercept form is the one you already half know, and it is the right form when you are told where the line starts. Point-slope form is the one that saves you when nobody tells you where it starts, which in real problems is most of the time.
Aim for fluency rather than recognition. By the end you should be able to take a slope and a point, or two points, or a graph, or a table, and produce the rule in under a minute, with a check at the end that you actually trust.
Warm-up
Discussion prompt
A car is 10 kilometres from home when the clock starts, 40 kilometres away after one hour, 70 after two hours and 160 after five hours. Work out the average rate of change from the start to each of those three times. What do you notice, and what does it force the graph to look like?
Write your three quotients down before you reveal the answer. The arithmetic is easy, and doing it is what makes the pattern land.
The point of computing three of them, over intervals of one hour, two hours and five hours, is that a linear function is the one kind of function where the length of the interval makes no difference. For anything else, the average rate over five hours would be nothing like the average rate over one.
If you found the third quotient harder than the first two, notice why: you had to remember to divide by five, not just subtract. That slip is the single most common error in this lesson, and it comes back in every table problem.
Section
Part 1
Concept
The Shanghai maglev train reaches a cruising speed and holds it. Suppose it holds 83 metres per second, starting from a point 250 metres past the station.
Figure (svg): The line for the train's distance, starting at 250 metres when the clock starts and climbing steadily; a right triangle on the line shows a run of 10 seconds against a rise of 830 metres.
Two facts make the whole graph. Where it starts, and how much it gains per second. Nothing else is needed, because the gain per second never varies.
\[ D(t) = 83t + 250 \]
OpenStax, Precalculus, §2.1 Linear Functions §2.1, pp. 181-182 — the maglev train
Look at the triangle drawn on the line. It sits between twenty and thirty seconds, but you could slide it anywhere along the line and it would keep the same shape. Ten seconds across always buys eight hundred and thirty metres up.
The pink dot is where the story begins: the train was already two hundred and fifty metres from the station when it reached cruising speed. That number has nothing to do with the speed, and changing it would slide the whole line up or down without tilting it at all.
Notice that the graph is restricted here. Negative times do not mean anything in this story, so the sensible domain is the non-negative numbers. That restriction comes from the situation, not from the algebra, and the algebra would happily accept any real number.
Concept
The book writes this one function four ways, and you will be asked to move between them all year.
Figure (svg): A two-row table of the train function: times 0, 1, 2, 3 seconds above distances 250, 333, 416, 499 metres, with arrows marked plus 1 between the times and plus 83 between the distances.
OpenStax, Precalculus, §2.1 Linear Functions §2.1, pp. 181-183 — the four representations
Four representations, one function. Exams and real problems switch between them without warning, so practise recognising the same object in all four outfits.
The table is the one worth staring at. The top row steps by one and the bottom row steps by eighty-three, and those two step sizes are the slope in disguise: eighty-three divided by one. If the top row stepped by two instead, the bottom row would step by one hundred and sixty-six, and you would still have to divide to recover the rate.
The word form is the one students skip, and it is the one that carries the meaning. Being able to say the rule as a sentence, with units attached, is what turns a formula into a model you can reason about.
Notation
Annotate
On: \( f(x) = mx + b \)
linear function — A function whose rate of change is constant. Equivalently, a polynomial of degree one, and equivalently a function whose graph is a straight line.
OpenStax, Precalculus, §2.1 Linear Functions §2.1, p. 183 — equation 2.2 and the definition box
Step through the four annotations and connect each to the train picture. The letter m is the tilt, the letter b is the height at which the line meets the vertical axis, and x and f of x are the input and output that vary.
The single most useful habit here is to read b as the output at an input of zero rather than as the y-intercept. In a story problem, zero is a real moment in time or a real depth or a real level of production, and asking what the output is at that moment is usually the fastest way to find b.
Keep the units attached to m. The slope of the train function is not eighty-three, it is eighty-three metres per second, and a slope with no units is halfway to being meaningless.
Worked example
Pressure in pounds per square inch on a diver depends on her depth below the surface, in feet.
\[ P(d) = 0.434d + 14.696 \]
Name the initial value
Why: It is the output when the input is zero, and zero depth is the surface of the water.
\[ P(0) = 0.434(0) + 14.696 = 14.696 \]
Name the rate of change
Why: It is the number multiplying the depth, and its units are output units over input units.
\[ m = 0.434 \text{ PSI per foot} \]
Say the rule as a sentence
Why: Restating it in words is the point of the example.
\[ \text{pressure} = 0.434 \cdot \text{depth} + 14.696 \]
Figure (svg): The line for pressure against depth, starting just under 15 pounds per square inch at the surface and rising in a straight line to about 58 at one hundred feet, with the value at 33 feet marked.
Use it at a real depth
Why: Thirty-three feet is roughly one extra atmosphere of water above her.
\[ P(33) = 0.434(33) + 14.696 = 14.322 + 14.696 = 29.018 \]
Check that the slope means what it claims
Why: One more foot down must add exactly 0.434, whatever depth you start from.
\[ P(34) - P(33) = 29.452 - 29.018 = 0.434 \]
OpenStax, Precalculus, §2.1 Linear Functions §2.1, p. 183 — Example 2.1
The book asks only for a restatement in words, which sounds like a soft question. It is not. Saying the rule out loud forces you to name which number is the rate and which is the starting value, and that is the whole skill.
The rate here is small, a little under half a pound per square inch for every foot of depth, but it adds up fast. At thirty-three feet the pressure has roughly doubled from its surface value, which is why that depth is called one atmosphere.
The check at the end is the model for every check in this lesson. It does not re-do the arithmetic; it tests the claim. If the slope really is 0.434 pounds per square inch per foot, then dropping one more foot must add exactly that, and it does.
Picture it
Figure (svg): Two panels of dots. On the left the dots climb by the same amount each step and lie on a straight line. On the right the gaps between the dots grow, one, three, five, seven, and the dots bend upward.
On the left the outputs jump by three every time the input jumps by one, so the dots line up. On the right the jumps grow, so the dots curve away. The test for linearity in a table is exactly this, and it only works when the input steps are equal.
Cover the curves and look only at the numbers between the dots. On the left they are all the same; on the right they grow by two every time. That is the whole test for linearity in a table, and it takes about five seconds.
The trap hidden in this test is that it only works when the input steps are equal. If a table gives you inputs of nought, two, four and then eight, the last output jump covers twice as much ground and will naturally be twice as large. You must divide, not compare.
The right-hand dots come from squaring the input. Their growing differences are the reason you cannot fit a line through them, and in Chapter 3 those same growing differences will be the signature of a quadratic.
Sorting
Sort into buckets
Sort each rule or table by whether it defines a linear function.
OpenStax, Precalculus, §2.1 Linear Functions §2.1, pp. 199-202 — Exercises 7 to 14 and 48 to 49
Sort all eight before revealing. For the rules, the question is whether you can get it into the shape of a number times x plus a number, using nothing but rearranging. For the tables, the question is whether equal input steps give equal output steps.
Two of the rules look non-linear until you move things around, and one looks linear until you notice the square. The rule with y squared is the nastiest of the set, because once you solve it for y you get a square root and a plus-or-minus, which is not even a function.
If you got the two tables the wrong way round, look again at the second one: the outputs rise by twenty-five, then seventy-five, then one hundred and twenty-five. Growing differences over equal input steps mean the rate is not constant.
Concept
Start from the rate and derive the formula, rather than being handed it.
\[ \frac{f(x) - f(0)}{x - 0} = m \quad \text{for every } x \ne 0 \]
Multiply both sides by x
Why: The denominator is just x, since the second input is zero.
\[ f(x) - f(0) = mx \]
Add the starting output to both sides
Why: Write f(0) as b, the initial value.
\[ f(x) = mx + b \]
Every input obeys the same equation with the same two constants, and that equation is the equation of a line. So the constant rate, the degree-one formula and the straight graph are three descriptions of one property.
OpenStax, Precalculus, §2.1 Linear Functions §2.1, pp. 185-186 — the slope formula in function notation
This is short, and being able to reproduce it is worth more than memorising the formula it produces. Start from the definition of a linear function, which is a constant rate of change, and pick the two inputs to be zero and a general x.
The algebra is two moves. Multiply both sides by x to clear the fraction, then add the output at zero to both sides. What comes out is the slope-intercept form, with b identified as the output at an input of zero, exactly as the annotation slide claimed.
Read the conclusion carefully. It says every input obeys one equation with the same two constants. That is what makes the graph a line rather than something that happens to look straight in the region you plotted.
Intuition
A general function needs infinitely much information to pin down: an output for every input. A linear function needs two numbers.
That is why linear models are everywhere. Two measurements are usually cheap to take, and two measurements are exactly enough to build the whole rule and then predict outputs you never measured.
It is also why the assumption has to be checked. If the real rate drifts, the two numbers you fitted will be right near your measurements and wrong far away from them.
It is easy to skate past how strong this is. Most functions need infinitely much information to specify, because you have to say what the output is at every input. A linear function needs two numbers, and then every other value is forced.
That economy is why linear models are the default first attempt at almost anything, from a company's costs to a drug's clearance rate. Two measurements are cheap, and two measurements are exactly the budget a line requires.
The flip side is that the assumption does real work and must be earned. If the true rate drifts, your fitted line will agree with reality near your two measurements and drift away from it elsewhere, which is the usual way a confident-looking prediction goes wrong.
Section
Part 2
Concept
Given two points on the line, the slope is the change in output divided by the change in input.
\[ m = \frac{\text{change in output (rise)}}{\text{change in input (run)}} = \frac{\Delta y}{\Delta x} = \frac{y_2 - y_1}{x_2 - x_1} \]
Figure (svg): A straight line with two marked points on it; a horizontal leg from the first point to below the second is labelled run, and a vertical leg up to the second point is labelled rise.
In function notation the same thing is written with the outputs named by the function, which is the form you will meet whenever a problem gives you two function values instead of two points.
\[ m = \frac{f(x_2) - f(x_1)}{x_2 - x_1} \]
OpenStax, Precalculus, §2.1 Linear Functions §2.1, pp. 185-186 — equation 2.3 and Figure 2.7
Two subtractions and one division. The only difficulty is bookkeeping, and the picture is there to remind you what each subtraction means: the run is how far right you walked, the rise is how far up you ended.
The second version, written with the function evaluated at the two inputs, is the same formula. You will meet it whenever a problem hands you function values rather than coordinate pairs, and the first move in that situation is always to rewrite them as points.
Notice the picture is drawn with the two axes at the same scale. That matters more than it looks: on squashed axes a gentle line can look ferocious, and you cannot judge a slope by eye unless you know what the scales are.
Notation
Annotate
On: \( m = \frac{y_2 - y_1}{x_2 - x_1} \)
The subscripts confuse people more than anything else in this formula, and the fix is to read them as labels rather than as operations. Point one is simply whichever point you decided to write first. There is no rule about which point deserves the honour.
The second annotation is the one that costs marks. Reversing both subtractions flips two signs and leaves the quotient alone, so both orders are legal. Reversing only one flips the answer, and the answer is then wrong by a minus sign, which is exactly the trap on the next few slides.
The third annotation is the deep one. Any two points on the line give the same answer, and that is not true of any other kind of graph. It is what licenses you to choose the two points with the friendliest numbers.
Picture it
Figure (svg): One straight line with three right triangles of different sizes resting on it: legs two and one, then four and two, then six and three. Every triangle has the same shape.
Three different pairs of points, three different rises and runs, one identical quotient. The triangles are different sizes and the same shape, which is what equal ratios look like in a picture.
Three pairs of points, three different rises, three different runs, one quotient. If you have met similar triangles in geometry, this is the same fact wearing algebra's clothes: the triangles have the same shape, so corresponding sides are in the same ratio.
This is why a line can be said to have a slope at all. On a curve the answer would depend on which two points you picked, which is why average rate of change on a curve always has to say over which interval.
Practically, it means you should hunt for the two points whose coordinates are whole numbers. Picking a hard pair when an easy pair is available is self-inflicted arithmetic.
Worked example
A linear function has these two points on its graph. Find the slope, and say whether the function is increasing or decreasing.
\[ (3, -2) \qquad \text{and} \qquad (8, 1) \]
Subtract the outputs, second minus first
Why: This is the rise. Watch the double sign: subtracting a negative adds.
\[ \Delta y = 1 - (-2) = 3 \]
Subtract the inputs in the same order
Why: Second minus first again. This is the run.
\[ \Delta x = 8 - 3 = 5 \]
Divide the rise by the run
Why: One number, no units here because none were given.
\[ m = \frac{3}{5} = 0.6 \]
Read the sign
Why: A positive slope means the output grows as the input grows.
\[ \tfrac35 > 0 \;\Longrightarrow\; f \text{ is increasing} \]
Figure (svg): The line through the points three comma minus two and eight comma one, with a run of five and a rise of three drawn as a right triangle beneath the line.
Check by reversing the order of the points
Why: Both subtractions flip sign, so the quotient must be unchanged.
\[ \frac{-2 - 1}{3 - 8} = \frac{-3}{-5} = \frac{3}{5} \]
OpenStax, Precalculus, §2.1 Linear Functions §2.1, p. 187 — Example 2.3
The only place to go wrong is the first line. The output at the first point is negative, so the subtraction on top is one minus a negative, which is one plus two, which is three. Write the brackets in; do not do it in your head.
Once you have three fifths, two questions are answered at once. The value tells you the line climbs three units for every five to the right, and the sign tells you the function is increasing. You never had to plot anything.
The final line is the check you should copy into every slope problem you ever do. Swap the two points and recompute. Both differences flip sign, the quotient comes out the same, and if it does not you have made exactly the mistake the next slide is about.
Prediction
\[ (2, 3) \qquad \text{and} \qquad (0, 4) \]
Predict first
A linear function passes through these two points. What is its slope, and is the function increasing or decreasing?
Correct: Minus one half, and decreasing
Why: Taking the second point first, the rise is four minus three, which is one, and the run is zero minus two, which is minus two. One divided by minus two is minus one half. A negative slope means the output falls as the input rises, so the function is decreasing. Choosing minus two instead means the fraction was written upside down.
OpenStax, Precalculus, §2.1 Linear Functions §2.1, p. 187 — Try It 2.1
Commit to one option before revealing. This checkpoint is deliberately awkward, because the point with the larger input is written first and the point with the input of zero is written second.
If you took the points in the order given, the rise is three minus four, which is minus one, and the run is two minus zero, which is two, giving minus one half. If you took them the other way round, you get one over minus two, which is the same number. Either is fine; mixing them is not.
The two distractors with two in them are what happens if the fraction goes upside down. A useful guard is to ask whether the answer should be big or small before computing: the outputs here barely move while the inputs move by two, so the slope must be a fraction smaller than one.
Trap
Finding the slope through the two points below:
\[ (3, -2), \quad (8, 1) \]
\[ m = \frac{1 - (-2)}{3 - 8} \]
\[ m = \frac{3}{-5} = -\frac{3}{5} \]
Wrong. The sign came out backwards.
The outputs were subtracted second minus first, but the inputs were subtracted first minus second. Choose an order and use it on both lines.
\[ m = \frac{1 - (-2)}{8 - 3} = \frac{3}{5} \]
A quick sanity check catches it every time: the second point is higher and further right, so the line climbs and the slope must be positive.
This is the most common error in the whole section, and it does not announce itself: the arithmetic looks fine, the fraction reduces nicely, and only the sign is wrong.
Two defences. The first is mechanical: write the subtraction on top and immediately write the matching subtraction underneath, in the same order, before evaluating either. The second is a sanity check on the picture: if the second point is both higher and further right, the line climbs, so a negative answer is impossible.
Get in the habit of doing the sanity check even when you are confident. It costs two seconds and it catches this error every single time.
Worked example
A city grew from 23,400 people in 2008 to 27,800 in 2012. Assuming the change was steady, find the growth per year.
Subtract the populations
Why: This is the change in output over the whole interval.
\[ 27{,}800 - 23{,}400 = 4{,}400 \text{ people} \]
Subtract the years in the same order
Why: This is the change in input.
\[ 2012 - 2008 = 4 \text{ years} \]
Divide, and carry the units through
Why: People on top, years underneath, so the slope reads people per year.
\[ m = \frac{4{,}400 \text{ people}}{4 \text{ years}} = 1{,}100 \text{ people per year} \]
Figure (svg): Two dots four years apart, at twenty-three thousand four hundred and twenty-seven thousand eight hundred, joined by a straight line whose four equal yearly steps of 1,100 are marked.
Check by rebuilding the later figure
Why: Four years of growth added to the 2008 count must return the 2012 count.
\[ 23{,}400 + 4(1{,}100) = 23{,}400 + 4{,}400 = 27{,}800 \]
OpenStax, Precalculus, §2.1 Linear Functions §2.1, p. 187 — Example 2.4
This is the first problem where the units do real work. Neither the inputs nor the outputs are abstract numbers: one is years and the other is people, so the slope is people per year and nothing else.
The word steady in the problem is doing the heavy lifting. Real populations do not grow in a straight line, and the question is explicitly asking you to assume they do so that a single rate exists. Whenever a problem says assume the change was constant, it is telling you that a linear model has been imposed.
The check rebuilds the 2012 figure from the 2008 figure and four years of growth. That is a real test, because if you had divided by five years instead of four, or subtracted the wrong way, the rebuilt number would miss.
Fill the middle
A small town grew from 1,442 people in 2009 to 1,868 in 2012. Fill in the two missing numbers, assuming steady growth.
Fill in the blanks
m = \frac426142 = \frac___}___ = ___
Why: The population rose by 426 over three years, so the rate is 426 divided by 3, which is 142 people per year. Always write the units after the number: the answer is 142 people per year, not just 142.
OpenStax, Precalculus, §2.1 Linear Functions §2.1, p. 187 — Try It 2.2
Fill both blanks before checking. The numerator is the change in population and the denominator is the change in years, and the years are the ones easiest to get wrong: from 2009 to 2012 is three years, not four.
Counting year gaps by subtracting is safer than counting on your fingers. Two thousand and twelve minus two thousand and nine is three, and that is the number you divide by, whatever your instinct about how many years are involved.
Finish by saying the answer out loud with its units: one hundred and forty-two people per year. A bare one hundred and forty-two would not tell anyone what happened to the town.
Worked example
Find the slope of the line through the two points below.
\[ (6, 11) \qquad \text{and} \qquad (-4, 3) \]
Subtract the outputs, second minus first
Why: Taking the leftmost point second is allowed; only consistency matters.
\[ \Delta y = 3 - 11 = -8 \]
Subtract the inputs in the same order
Why: Second minus first again, and this one is negative too.
\[ \Delta x = -4 - 6 = -10 \]
Divide
Why: A negative divided by a negative is positive, and the fraction reduces by two.
\[ m = \frac{-8}{-10} = \frac{4}{5} \]
Check with the other order
Why: Swapping the points makes both differences positive and must give the same answer.
\[ \frac{11 - 3}{6 - (-4)} = \frac{8}{10} = \frac{4}{5} \]
OpenStax, Precalculus, §2.1 Linear Functions §2.1, p. 199 — Exercise 27
This is the case that makes people abandon their chosen order halfway through. Both differences come out negative, which feels wrong, and the temptation is to quietly flip one of them to make the arithmetic prettier.
Resist it. Two negatives divide to a positive, and the positive answer is correct: the point with the larger input also has the larger output, so the line climbs. The picture agrees with the algebra.
The check swaps the order and gets two positive differences instead. Same quotient, less alarming-looking arithmetic. If you find negatives distracting, choosing the order so that the larger input comes first is a perfectly legal habit.
Concept
Strip the units off a slope and you have thrown away most of what it tells you. The units are forced: output units on top, input units underneath.
| the situation | input | output | the slope reads |
|---|---|---|---|
| the maglev train | seconds | metres | 83 metres per second |
| the diver | feet of depth | pounds per square inch | 0.434 PSI per foot |
| the city | years | people | 1,100 people per year |
| a phone bill | minutes talked | dollars | 0.10 dollars per minute |
The last row is Exercise 72, where the monthly charge is 24 dollars plus a tenth of a dollar a minute. Reading the two constants out of a formula in context is half of what this section is for.
OpenStax, Precalculus, §2.1 Linear Functions §2.1, pp. 186, 203 — the units box and Exercise 72
Read down the last column of the table and notice that every entry has the shape something per something. That is not decoration; it is what a rate is.
Getting the units right also tells you when you have divided the wrong way round. If a problem is about dollars and minutes and your answer comes out in minutes per dollar, you inverted the fraction, and you will know before you ever check the number.
The phone-bill row is worth remembering, because that style of question appears constantly: a fixed monthly charge plus a per-unit charge. The fixed charge is the initial value and the per-unit charge is the slope, every time.
Real world
Discussion prompt
Jessica is walking home. Two minutes after leaving she is 1.4 miles from home; twelve minutes after leaving she is 0.9 miles from home. What is her rate, in miles per hour? Do the whole conversion yourself before revealing.
OpenStax, Precalculus, §2.1 Linear Functions §2.1, p. 199 — Exercise 3
Do the whole thing before revealing, including the conversion. The question asks for miles per hour and the data are in minutes, and half the marks in a question like this live in the unit conversion.
The slope comes out negative, and that surprises people who expect a walking speed to be positive. Look at what is being measured: it is her distance from home, and that distance is falling. A rate of change is negative exactly when the quantity is shrinking.
To convert, multiply by sixty because there are sixty minutes in an hour, and think about whether the number should get bigger or smaller. She covers more miles in an hour than in a minute, so the magnitude must grow, and 0.05 becoming 3 is the right direction.
Intuition
A slope does two separate jobs, and it is worth keeping them apart in your head.
The sign says which way the line tilts: up to the right, down to the right, or not at all. The size, meaning the distance from zero, says how sharply it tilts. A slope of minus seven is a far steeper line than a slope of two, even though minus seven is the smaller number.
When a question asks which line is steeper, it is asking about size. When it asks which function grows faster, it is asking about size and sign together, because a falling line does not grow at all.
OpenStax, Precalculus, §2.1 Linear Functions §2.1, p. 185 — the greater the absolute value, the steeper the line
Slopes carry two independent pieces of information and questions usually ask about only one, so read the question carefully before answering.
Which line is steeper is a question about size only: ignore the signs and compare distances from zero. Which function is increasing is a question about sign only. Which function grows faster needs both, because a falling line is not growing at all.
A quick numerical anchor: a slope of minus seven is much steeper than a slope of two, even though minus seven sits far to the left of two on the number line. Ordering slopes by steepness is not the same as ordering them as numbers, and mixing those two orderings up is the next trap in this deck.
Section
Part 3
Concept
Figure (svg): Three small graphs side by side: the first line slants upward, the second slants downward, and the third is flat.
\[ f(x) = mx + b: \quad m > 0 \text{ increasing}, \quad m < 0 \text{ decreasing}, \quad m = 0 \text{ constant} \]
There is no fourth case for a function, because a straight graph cannot rise on one stretch and fall on another. That is what having a single rate of change rules out.
OpenStax, Precalculus, §2.1 Linear Functions §2.1, p. 184 — the increasing and decreasing functions box
Three pictures, one number. Once you can read the sign of the coefficient of x, you know which of the three you are looking at without plotting a single point.
The third case deserves a moment. A constant function is a linear function, with slope zero. It feels like a degenerate case, and it is, but it obeys every rule in the section and it turns up constantly in modelling as a fixed fee or a flat charge.
There is no fourth case for a function, and that is worth saying explicitly, because curves have plenty: they rise here, fall there, level off somewhere else. Having a single rate of change is exactly what rules that out.
Worked example
A teenager sends about 60 texts a day. For each scenario, write the rule and say whether it is increasing, decreasing or constant.
Part (a): total texts sent after x days
Why: Sixty join the total each day, and the total starts at zero.
\[ f(x) = 60x, \quad m = 60 > 0 \;\Longrightarrow\; \text{increasing} \]
Part (b): texts left in a 500-text plan
Why: Start at 500 and take away sixty a day, so the rate is negative.
\[ f(x) = 500 - 60x, \quad m = -60 < 0 \;\Longrightarrow\; \text{decreasing} \]
Part (c): monthly cost of an unlimited plan
Why: The number of days does not appear in the rule at all.
\[ f(x) = 50, \quad m = 0 \;\Longrightarrow\; \text{constant} \]
Figure (svg): Three lines over eight days: one climbing from zero by sixty a day, one falling from five hundred by sixty a day and reaching zero before day nine, and one flat at fifty.
Check part (b) against the story
Why: The allowance should run out sometime during the ninth day, and the graph must cross zero there.
\[ 500 - 60x = 0 \iff x = \frac{500}{60} \approx 8.33 \text{ days} \]
OpenStax, Precalculus, §2.1 Linear Functions §2.1, pp. 184-185 — Example 2.2
Three scenarios, and the work is in translating each sentence into a rule rather than in any algebra. Ask yourself two questions each time: what is the output when the input is zero, and how much does the output move per day?
Part (b) is where the negative appears, and it appears because texts are being used up rather than accumulated. Notice that the rule is written starting with five hundred, so the coefficient of x is minus sixty and not five hundred. Reading coefficients off a rule written back to front is a classic slip.
Part (c) is the one people argue with. The cost does not depend on the number of days, so the input does not appear in the rule at all, and the slope is zero. The graph is a flat line, and it is still a linear function.
Discrimination
Sort into buckets
Sort each rule by whether it is increasing, decreasing or constant.
OpenStax, Precalculus, §2.1 Linear Functions §2.1, p. 199 — Exercises 15 to 24
Sort all seven before checking. You only need one thing from each rule: the sign of the number multiplying x. Everything else on the slide is a distraction.
Two of the rules are written with the constant first, which is where mistakes come from. In the one that reads five minus two x, the coefficient of x is minus two, not five and not two. Rewrite it with the x term first if that helps.
The gentle slopes, one half and one quarter, are increasing just as surely as the steep one. This bucket is about direction, not speed, and a line that climbs slowly still climbs.
Tweak it
Parameter explorer
Set the starting value to one and drag the slope down from three to minus three. Watch the line pivot about the point where it meets the vertical axis. What is the one slope at which the line stops tilting altogether, and what happens to the function there?
\[ y = {m}x + ({b}) \]
Set the starting value to one and leave it there for a moment. Now drag the slope slowly down from three. Watch how the line pivots about the point where it meets the vertical axis: that point never moves, because it is the output at an input of zero and the slope has nothing to say about it.
At the moment the slope reaches zero the line stops tilting and becomes flat. Push past it and the line starts falling. There is no gap and no special behaviour at the crossing; the flat line is just the one case in the middle.
Now leave the slope alone and drag the starting value instead. The line slides up and down without changing its tilt at all. That independence is the reason you can find the two constants separately, which is exactly what the modelling problems in the next part do.
Trap
Which of these two lines is steeper?
\[ y = -5x + 1 \]
\[ y = 2x + 1 \]
Answer given: the second, because two is bigger than minus five.
Wrong. Steepness ignores the tilt direction.
Steepness is the distance of the slope from zero, not its position on the number line.
\[ |-5| = 5 > 2 = |2| \]
The first line is the steeper of the two. The sign only reports which way it leans: the first falls as you read left to right, the second rises. Both facts matter, and they are different facts.
This one catches people who are being careful in the wrong way. They remember that the sign matters, so they compare the signed numbers, and the comparison they perform is a legitimate comparison of the wrong thing.
Say it to yourself as two separate sentences. The first line falls, and it falls steeply. The second line rises, and it rises gently. Direction and steepness are not competing answers to one question; they are answers to two questions.
The bars in the corrected line mean distance from zero, which you met as absolute value in Chapter 1. If a question asks which line is steeper, it is asking you to strip the signs off before comparing.
Ranking
Figure (svg): Five lines through the origin at different tilts: two rising, one gently and one very steeply, and three falling at different angles.
Put in order
Put these five lines in order, from the one closest to flat up to the steepest.
Why: Ignore the signs and the constants and read only the distance of each slope from zero: one quarter, one, two, five, seven. The starting values move the lines around the plane but never change how steep they are, and two of the steepest four are falling lines.
Drag them into order before checking, and use the picture to sanity-check your ordering once you have it.
The recipe is to strip off the constant, strip off the sign, and read what is left: one quarter, one, two, five, seven. The constants move the lines around the plane but cannot tilt them, so they never enter this decision.
Notice that the two steepest lines in the list are both falling. If you ordered by the signed values you would have put them at the bottom, which is the error from the previous slide showing up in a new costume.
Section
Part 4
Concept
\[ f(x) = mx + b \]
Figure (svg): Three parallel lines of the same tilt crossing the vertical axis at three heights, marked at plus three, zero and minus four.
The slope turns the line; the starting value slides it. Because the two act independently, you can find them independently, and most modelling problems give you one of them directly and make you work for the other.
OpenStax, Precalculus, §2.1 Linear Functions §2.1, pp. 183, 195 — the definition box and the how-to for evaluating a model
The picture is the whole idea: three lines, one slope, three starting values. Turning and sliding are independent operations, and that is why the two constants can be hunted separately.
Work through the three bullets and match each to a problem you have already seen. The train gave you both numbers directly. A table with an input of zero hands you the starting value. Two data pairs hand you neither and make you compute both.
In every modelling problem from here on, your first question should be which of those three situations you are in. That decides the whole method, and getting it right takes about five seconds of reading.
Worked example
Ben pays 1,250 dollars a month in overhead and 37.50 dollars to produce each item. Write the cost of producing x items in a month.
Name the fixed part
Why: It is charged whether or not anything is produced, so it is the output at an input of zero.
\[ b = 1250 \]
Name the part that varies
Why: Each item adds the same amount, so that amount is the rate; the book calls it the marginal cost.
\[ m = 37.5 \text{ dollars per item} \]
Assemble the rule
Why: Rate times input plus initial value.
\[ C(x) = 37.5x + 1250 \]
Evaluate at one hundred items
Why: Substitute and multiply before adding.
\[ C(100) = 37.5(100) + 1250 = 3750 + 1250 = 5000 \]
Check that the rate behaves like a rate
Why: One extra item must raise the cost by exactly the marginal cost, at any production level.
\[ C(101) - C(100) = 5037.5 - 5000 = 37.5 \]
OpenStax, Precalculus, §2.1 Linear Functions §2.1, p. 194 — Example 2.8
Cost problems of this shape are everywhere, and they are all the same problem. A fixed cost that appears whether or not anything happens is the initial value; a cost that repeats per item is the rate.
The book calls the per-item cost the marginal cost, and that word is worth knowing, because economics uses it constantly and it means exactly the slope of the cost function.
The check is the interesting part. Producing one more item must raise the cost by exactly the marginal cost, at any level of production. That is precisely the assumption a linear cost model makes, and it is also its main weakness: real costs per item usually fall as volume rises.
Worked example
Marcus has 200 songs and adds 15 every month. Write the number of songs as a function of the number of months, then find how many he has after a year.
Read the initial value
Why: He already owns 200 at time zero, so that is the output at an input of zero.
\[ N(0) = 200 \;\Longrightarrow\; b = 200 \]
Read the rate
Why: Fifteen songs join the collection every month, and the units follow.
\[ m = 15 \text{ songs per month} \]
Write the rule
Why: Rate times months plus the starting collection.
\[ N(t) = 15t + 200 \]
Evaluate at twelve months
Why: A year is twelve months, so substitute twelve.
\[ N(12) = 15(12) + 200 = 180 + 200 = 380 \]
Check with half the time
Why: Half as many months should add half as many songs, because the rate is constant.
\[ N(6) = 15(6) + 200 = 290, \quad 290 - 200 = 90 = \tfrac{180}{2} \]
OpenStax, Precalculus, §2.1 Linear Functions §2.1, pp. 195-196 — Example 2.10
This is the same shape as the cost problem with friendlier numbers, so use it to rehearse the routine rather than to learn anything new. Initial value, rate, assemble, evaluate, check.
One small habit to build: convert the units before you substitute. The question asks about a year and the rule counts months, so twelve goes in, not one. Substituting the wrong unit is a much more common error than getting the arithmetic wrong.
The check exploits constancy of the rate. Half the months must add half the songs, and ninety is half of one hundred and eighty. Any proportional check like that will catch a slip in the multiplication.
Worked example
Ilya earns a base salary plus commission. Last week he sold 3 policies and earned 760 dollars; the week before he sold 5 and earned 920. Find the rule and interpret both constants.
\[ (3, 760) \qquad \text{and} \qquad (5, 920) \]
Find the rate from the two pairs
Why: Change in income over change in policies.
\[ m = \frac{920 - 760}{5 - 3} = \frac{160}{2} = 80 \]
Attach the units
Why: Dollars on top, policies underneath, so the rate is his commission.
\[ m = 80 \text{ dollars per policy} \]
Substitute one pair into the form
Why: Either pair will do; take the first.
\[ 760 = 80(3) + b \]
Solve for the initial value
Why: Subtract the commission earned from the total paid.
\[ b = 760 - 240 = 520 \]
Write the rule
Why: This is his pay for a week with no sales at all, plus eighty for each sale.
\[ I(n) = 80n + 520 \]
Figure (svg): A line of income against the number of policies sold, starting at five hundred and twenty when nothing is sold and passing through the two recorded weeks at three policies and five policies.
Check against the pair that was not used
Why: The second week must also land on the line, or the model is wrong.
\[ I(5) = 80(5) + 520 = 400 + 520 = 920 \]
OpenStax, Precalculus, §2.1 Linear Functions §2.1, pp. 196-197 — Example 2.11
This is the first problem where neither constant is given, and it is the template for almost every two-data-point question you will meet. Find the rate first, then reach back for the starting value.
Follow the units through the slope calculation. One hundred and sixty dollars over two policies is eighty dollars per policy, and that per-policy figure is exactly what the word commission means. Naming the slope in the language of the story is half the interpretation the question asks for.
The starting value comes out at five hundred and twenty, and it is a real thing: what Ilya is paid in a week when he sells nothing. In the picture it is the height at which the line meets the vertical axis. Always finish by checking against the data pair you did not use, as the last line does here.
Worked example
The table gives the number of rats after w weeks. Write the linear rule.
\[ \begin{array}{c|cccc} w & 0 & 2 & 4 & 6 \\ \hline P(w) & 1000 & 1080 & 1160 & 1240 \end{array} \]
Read the initial value straight off
Why: The table happens to contain an input of zero, which is the easy case.
\[ P(0) = 1000 \;\Longrightarrow\; b = 1000 \]
Take one step of the table
Why: Two weeks pass and eighty rats appear, so divide to get the weekly rate.
\[ \frac{1080 - 1000}{2 - 0} = \frac{80}{2} = 40 \]
Confirm with a step between two non-neighbouring rows
Why: If the rate really is constant, any pair of rows must give the same number.
\[ \frac{1240 - 1080}{6 - 2} = \frac{160}{4} = 40 \]
Write the rule
Why: Forty rats per week on top of the thousand already there.
\[ P(w) = 40w + 1000 \]
Check against a row the rule was not built from
Why: The third column is the honest test, since neither step above used it.
\[ P(4) = 40(4) + 1000 = 1160 \]
OpenStax, Precalculus, §2.1 Linear Functions §2.1, p. 197 — Example 2.12
This table is generous: it contains an input of zero, so the starting value can be read straight off with no work. Always scan for that before doing anything else.
For the rate, the temptation is to read eighty off the table and call it the slope. It is not: eighty rats appear over two weeks, so the weekly rate is forty. Dividing by the input step is the step people skip, and unequal input steps punish them for it.
Two things make this a solid solution. It confirms the rate with a second, non-adjacent pair of rows, which is how you verify that the table really is linear. And it checks the finished rule against a row that neither computation used.
Socratic
\[ \begin{array}{c|ccccc} x & 0 & 2 & 4 & 8 & 12 \\ \hline H(x) & 12.5 & 13.5 & 14.5 & 16.5 & 18.5 \end{array} \]
Discussion prompt
The table gives a young tree's height in feet, x months after a new plant food was introduced. Write the rule H(x), and then say what each of your two constants means for the tree. Notice that the input steps here are not all equal.
OpenStax, Precalculus, §2.1 Linear Functions §2.1, p. 198 — Try It 2.6
Do this one on paper before revealing, and read the top row of the table carefully first. The inputs go nought, two, four, eight, twelve, so the steps are two, two, four and four. They are not all equal.
That is exactly why you must divide rather than compare output differences. From four months to eight the height rises by two feet, which looks like four times the earlier rise of half a foot per row, but it covers four months. Divide and the rate is half a foot a month, the same as everywhere else.
Finish by saying what each constant means for the tree: twelve and a half feet tall when the experiment started, growing half a foot a month afterwards. If a question gives you a real object, it wants the interpretation as well as the formula.
Error analysis
Annotate
On: \( 3x + 5y = 15 \;\Longrightarrow\; m = 3, \; b = 15 \)
OpenStax, Precalculus, §2.1 Linear Functions §2.1, p. 199 — Exercise 10
Look at the claim before opening the annotations and see whether you can name the problem in one sentence. The two numbers were lifted out of an equation that was not in the right form.
Slope-intercept form has the output alone on one side. Until that is true, the coefficient of x is not the slope and the constant is not the starting value. Here you must subtract the x term and then divide everything by five, and both constants come out different from the ones claimed.
The last annotation is the habit to steal. Testing one point takes ten seconds and settles the argument: substitute an input into the original equation and into your rearranged rule, and if the outputs differ, the rearranging went wrong.
Comparison
Comparison matrix
| the rule | slope m | starting value b |
|---|---|---|
| y = 5x − 2 | 5 | -2 |
| y = 3 − 7x | -7 | 3 |
| 2y = 8x + 6 | 4 | 3 |
| y = 9 | 0 | 9 |
| 3x + 5y = 15 | −3/5 | 3 |
Two of these rows need work before anything can be read: the third must be divided through by two, and the last must be rearranged. Only then do the coefficient and the constant mean what you want them to mean.
Fill in every blank before checking. Three of these rows can be read straight off, and two cannot, and telling which is which is the skill being practised.
The row with two on the left of the output needs dividing through by two, which halves both the coefficient and the constant. Doing only one of them is the usual slip.
The row with y equal to nine is worth a second look. There is no x at all, so the coefficient of x is zero, which means the slope is zero. It is the constant function again, and it does fit the general form with the rate set to nothing.
Matching
Match the pairs
Why: Look for the two numbers separately. The amount present before anything happens is the constant: 60 litres, 3 dollars, nothing, 40 dollars. The amount added or removed per unit of input is the coefficient, and it is negative only when the quantity is shrinking, as in the draining tank.
Match all four before revealing, and pick out the two numbers separately in each story rather than trying to see the whole rule at once.
The amount present before anything happens is the constant: sixty litres in the tank, three dollars on the meter, nothing in the pool, forty dollars for the gym. The amount added or removed per unit of input is the coefficient.
Only one story has a negative coefficient, and it is the one where the quantity is shrinking. The gym is the odd one out in a different way: the input never appears, because going more often costs nothing extra, so the slope is zero and the rule is a bare number.
Section
Part 5
Concept
Suppose you know one point on the line and the slope, but not the intercept. Write the slope formula using a general point on the line and the point you know.
\[ m = \frac{y - y_1}{x - x_1} \qquad (x \ne x_1) \]
Multiply both sides by the denominator
Why: The right-hand side cancels completely.
\[ m(x - x_1) = y - y_1 \]
Swap the sides
Why: Nothing has changed except which side is written first.
\[ y - y_1 = m(x - x_1) \]
point-slope form — The equation of the line through the point with coordinates x-one and y-one whose slope is m. Substituting the known point and slope gives the line immediately, with no algebra needed first.
OpenStax, Precalculus, §2.1 Linear Functions §2.1, p. 188 — equation 2.4 and its derivation
This derivation is three lines and it is worth being able to reproduce, because then you never have to remember which coordinate goes where: you can rebuild the form on the spot.
The move that makes it work is using a general point, written with plain x and y, together with the one specific point you were given. The slope formula between them equals m, and clearing the denominator gives the form.
Notice the condition attached to the first line. You may only multiply by that denominator when it is not zero, which quietly excludes the case where the two points share an input. That exclusion is a vertical line, and you will meet it again shortly.
Notation
Annotate
On: \( y - y_1 = m(x - x_1) \)
The thing to fix in your head is which letters stay and which get replaced. Plain x and y stay as letters, because they stand for any point on the line. The subscripted pair get replaced by the numbers you were given.
The second annotation is where the marks are lost. Each given coordinate sits behind a minus sign, so a negative coordinate becomes a plus once you simplify. Substitute with the brackets in place and simplify afterwards, never both at once.
The last annotation explains why this form exists at all. There is no letter for the intercept anywhere in it, so you can write the line down knowing nothing about where it crosses the vertical axis. That is exactly the situation two data points leave you in.
Picture it
Figure (svg): Four lines all passing through one marked point, each at a different tilt; the one with slope three is drawn brightest.
Fixing a point leaves a whole fan of lines, one for each slope. Point-slope form is the algebra of that picture: name the point, name the slope, and exactly one line is left standing.
The bright line and the two dashed ones all pass through the same point, and there are infinitely many others you could draw through it. A point alone does not determine a line.
Add a slope and exactly one of the fan survives. Point-slope form is the algebra of that sentence: the point kills every line not through it, and the slope kills every line with the wrong tilt.
Keep this picture in mind when you are given two points instead. The first point picks out the fan, and the second one chooses the slope, so two points determine a line for the same reason.
Worked example
Write the point-slope form of the line with slope 3 through the point below, then convert it to slope-intercept form.
\[ (6, -1) \]
Write the general form first
Why: Getting it on paper before substituting is what protects the minus signs.
\[ y - y_1 = m(x - x_1) \]
Substitute the slope and the two coordinates
Why: Keep the coordinate inside brackets so its own sign stays visible.
\[ y - (-1) = 3(x - 6) \]
Tidy the left side
Why: Subtracting a negative is adding. This is the point-slope answer.
\[ y + 1 = 3(x - 6) \]
Distribute the 3
Why: Multiply it across both terms in the bracket.
\[ y + 1 = 3x - 18 \]
Subtract 1 from both sides
Why: Now y stands alone and the form is slope-intercept.
\[ y = 3x - 19 \]
Check that the given point satisfies the rule
Why: Substituting the input must return the given output.
\[ 3(6) - 19 = 18 - 19 = -1 \]
OpenStax, Precalculus, §2.1 Linear Functions §2.1, p. 189 — Example 2.5
Write the general form down before substituting anything. It takes three seconds and it is the single best protection against dropping a sign, because you can see the minus signs waiting for their numbers.
The output coordinate is negative, so the left side becomes y minus a negative, which is y plus one. That is the point-slope answer, and the question asks for it, so do not race past it on the way to the other form.
Then two ordinary algebra moves, distribute and subtract, and you are in slope-intercept form. The check substitutes the given input and must return the given output, which it does. If yours does not, the error is almost always in the first substitution.
Fill the middle
Write the line with slope minus two through the point below in point-slope form, then in slope-intercept form.
\[ (-2, 2) \]
Fill in the blanks
y - 2 = -2\bigl(x + 2\bigr) \quad\Longrightarrow\quad y - 2 = -2x - 4 \quad\Longrightarrow\quad y = -2x + -2
Why: The form subtracts the input coordinate, and the coordinate is minus two, so the bracket contains x minus minus two, which is x plus two. After distributing you have y minus 2 on the left and minus 2x minus 4 on the right; adding 2 to both sides leaves minus 2x minus 2.
OpenStax, Precalculus, §2.1 Linear Functions §2.1, p. 189 — Try It 2.3
Fill both blanks before checking, and think about the first one carefully. The form subtracts the input coordinate, and the coordinate is minus two, so the bracket contains x minus minus two, which is x plus two.
That is the whole difficulty of this checkpoint, and it is the same difficulty as in Example 2.5, now on the other coordinate. Getting a plus inside the bracket from a negative coordinate feels wrong the first ten times and then becomes automatic.
For the second blank, distribute the minus two to get minus two x minus four, then add two to both sides. The constant lands on minus two, which happens to equal the input coordinate here by coincidence rather than for any reason.
Worked example
Write the line through the two points below in point-slope form, then rewrite it in slope-intercept form.
\[ (5, 1) \qquad \text{and} \qquad (8, 7) \]
Find the slope first
Why: Point-slope form needs a slope, and two points is exactly what the slope formula wants.
\[ m = \frac{7 - 1}{8 - 5} = \frac{6}{3} = 2 \]
Substitute the slope and either point
Why: Take the first point; the printed solution does the same.
\[ y - 1 = 2(x - 5) \]
Distribute the 2
Why: Two times x, and two times minus five.
\[ y - 1 = 2x - 10 \]
Add 1 to both sides
Why: Isolating y finishes the conversion.
\[ y = 2x - 9 \]
Figure (svg): The line through the points five comma one and eight comma seven, with a run of three and a rise of six drawn between them, extended down to cross the vertical axis at minus nine.
Check the point that was not substituted
Why: It has to satisfy the rule too, or the line misses it.
\[ 2(8) - 9 = 16 - 9 = 7 \]
OpenStax, Precalculus, §2.1 Linear Functions §2.1, pp. 190-191 — Example 2.6
This is the workhorse problem of the section: two points in, a rule out. The order is fixed, because point-slope form needs a slope and two points is exactly what the slope formula consumes.
The printed solution substitutes the first point, and so does this one. Substituting the second point instead gives a different-looking point-slope equation that simplifies to exactly the same slope-intercept rule, which is the subject of a slide shortly.
One note on the book. The printed text of this example shows the point-slope answer with stray subscripts on the variables, reading as though there were an x-two and a y-two in the equation. That is a typesetting slip; the correct equation is the one on this slide, with plain x and y.
Step zero
\[ (-1, 3) \qquad \text{and} \qquad (0, 0) \]
Discussion prompt
Before computing anything, look at the two points. One of them makes the rest of the problem noticeably shorter. Which one, and why? Then finish the problem.
OpenStax, Precalculus, §2.1 Linear Functions §2.1, p. 191 — Try It 2.4
Answer before revealing, and answer the question actually asked, which is about strategy rather than arithmetic. Look at the two points and decide which will cost you fewer lines.
The point with an input of zero is the vertical intercept, so substituting it makes both brackets collapse and the point-slope form becomes the slope-intercept form immediately. No distributing, no adding at the end, no chance to lose a sign.
Using the other point is not wrong and gives the same line, as the last part of the answer shows. It just costs two extra lines, each of which is an opportunity for a slip. Choosing the friendly point is a real skill, not laziness.
Worked example
No slope and no formula are given; only the picture. Read what you need off it.
Read two points where the line crosses a grid corner exactly
Why: Guessing fractions of a square is where graph-reading goes wrong.
\[ (0, 2) \qquad \text{and} \qquad (-2, -4) \]
Compute the slope
Why: Second minus first on both lines; both differences come out negative.
\[ m = \frac{-4 - 2}{-2 - 0} = \frac{-6}{-2} = 3 \]
Substitute the second point into point-slope form
Why: Both of its coordinates are negative, so both brackets gain a plus.
\[ y - (-4) = 3\bigl(x - (-2)\bigr) \]
Clean up the two double signs
Why: Minus a negative twice.
\[ y + 4 = 3(x + 2) \]
Distribute the 3
Why: Across both terms of the bracket.
\[ y + 4 = 3x + 6 \]
Subtract 4 from both sides
Why: And the slope-intercept form appears.
\[ y = 3x + 2 \]
Figure (svg): A line on a grid passing through the lattice points minus two comma minus four and zero comma two, rising three units for every one unit to the right.
Check against the intercept you can see on the graph
Why: The line crosses the vertical axis at 2, which must equal the constant in the rule.
\[ f(0) = 3(0) + 2 = 2 \]
OpenStax, Precalculus, §2.1 Linear Functions §2.1, pp. 192-193 — Example 2.7
Nothing is given but the picture, so the first job is to extract two points from it, and the first rule of extracting points is to use lattice points only. A point read as roughly one and a bit will poison everything that follows.
Both coordinates of the second point are negative, which means two double signs appear when you substitute. Clean them up one at a time, in a separate line, rather than trying to simplify while you substitute.
The check here is the nicest in the lesson, because it uses information you did not use to build the rule. The graph visibly crosses the vertical axis at two, and the finished rule has a constant of two. If those disagreed, you would know at once.
Worked example
A linear function is described by two of its values. Find its rule in slope-intercept form.
\[ f(3) = -2, \qquad f(8) = 1 \]
Turn the function values into points
Why: An input and its output are exactly a coordinate pair.
\[ f(3) = -2 \to (3, -2), \qquad f(8) = 1 \to (8, 1) \]
Compute the slope
Why: This is Example 2.3 again, with the same two points.
\[ m = \frac{1 - (-2)}{8 - 3} = \frac{3}{5} \]
Substitute the first point into point-slope form
Why: Keep the negative output inside brackets.
\[ y - (-2) = \tfrac{3}{5}(x - 3) \]
Tidy the left side
Why: Subtracting a negative is adding.
\[ y + 2 = \tfrac{3}{5}(x - 3) \]
Distribute the three fifths
Why: Three fifths of x, and three fifths of minus three.
\[ y + 2 = \tfrac{3}{5}x - \tfrac{9}{5} \]
Subtract 2, written as ten fifths
Why: A common denominator makes the last subtraction safe.
\[ y = \tfrac{3}{5}x - \tfrac{9}{5} - \tfrac{10}{5} = \tfrac{3}{5}x - \tfrac{19}{5} \]
Check the second value
Why: The value that was not substituted is the real test.
\[ f(8) = \tfrac{3}{5}(8) - \tfrac{19}{5} = \tfrac{24}{5} - \tfrac{19}{5} = \tfrac{5}{5} = 1 \]
OpenStax, Precalculus, §2.1 Linear Functions §2.1, pp. 194-195 — Example 2.9
The first step is the only new idea: function values are points in disguise. An input and its output are precisely a coordinate pair, so write them as pairs before doing anything else and the problem becomes one you have already solved.
The rest is fraction arithmetic, and fractions are where this example gets its difficulty. Distributing three fifths across the bracket gives three fifths of x minus nine fifths, and then you have to subtract two from a fifth-denominated expression.
Rewriting two as ten fifths before subtracting is the safe route. Subtracting a whole number from a fraction in your head is where minus nineteen fifths turns into minus eleven fifths, and the final check on the second value is what catches it.
Trap
Point-slope form through this point, with slope 4:
\[ (-2, 5) \]
\[ y - 5 = 4(x - 2) \]
Wrong. The input coordinate is negative, and its sign was dropped.
The form subtracts the coordinate. Substitute with the brackets still in place, then simplify.
\[ y - 5 = 4\bigl(x - (-2)\bigr) \]
\[ y - 5 = 4(x + 2) \]
Checking the point catches this instantly. The wrong line gives an output of 5 at an input of 2, not at the input the problem gave you.
This is the point-slope version of the sign trap, and it is the most common error in the second half of the section. Notice how plausible the wrong line looks: it has the right slope, the right shape, and one right number.
The cure is procedural. Substitute with the brackets still written out, so that x minus minus two appears on the page before you simplify it to x plus two. Doing both steps at once is what loses the sign.
The cure is also diagnostic. Substitute the given input into your finished rule. The wrong line here passes through the point with input two, not the point you were given, and the check catches that in one line.
Intuition
Example 2.6 substituted the first of its two points. Nothing would have changed if it had used the second: the point-slope equations look different, and they simplify to the same slope-intercept rule.
\[ y - 1 = 2(x - 5) \quad \text{and} \quad y - 7 = 2(x - 8) \quad \text{both give} \quad y = 2x - 9 \]
Think of it geometrically. Both equations say the same thing, that the line has slope two and contains a particular point, and the two points are on the same line. So the sets of points they describe are identical, even though the algebra in the middle looks different.
The practical consequence is that you may always choose the point with the friendlier numbers, and a point with a zero coordinate is friendliest of all.
Students often ask which point they are supposed to use, as though one of them were the official choice. There is no official choice, and this slide is the proof.
Both point-slope equations say the same thing in different words: this line has slope two and contains a particular point. The two points are on the same line, so the two equations describe the same set of points, and the algebra confirms it by simplifying to one rule.
The practical consequence is that you may always pick the point with the friendlier numbers. A point with a zero coordinate is friendliest, a point with whole numbers is next, and a point with negatives and fractions is last.
Section
Part 6
Concept
Figure (svg): A flat line at height two and an upright line at input three, crossing at the point three comma two; two hollow points are marked on the upright line above and below the crossing.
| the line | equation | slope | a function? |
|---|---|---|---|
| horizontal | y = k | 0 | yes: one output, the same one, for every input |
| vertical | x = h | undefined | no: one input with infinitely many outputs |
\[ \text{horizontal: } m = \frac{k - k}{x_2 - x_1} = 0, \qquad \text{vertical: } m = \frac{y_2 - y_1}{h - h} = \frac{\text{something}}{0} \]
The two cases look symmetric and they are not. Zero on top is a perfectly good answer; zero underneath is not an answer at all, because division by zero is undefined.
OpenStax, Precalculus, §2.1 Linear Functions §2.1, p. 184 — the constant function case
Look at the two lines in the picture and then at the two rows of the table. They feel like a symmetric pair and they are not, and noticing why is the point of this slide.
For the flat line, the outputs are all equal, so the rise is zero and the slope is zero divided by something. Zero is a perfectly ordinary answer, and it makes the constant function. For the upright line, the inputs are all equal, so the run is zero, and something divided by zero is not a number at all.
The hollow points show why the upright line is not a function. At the single input three it has an output at every height, and the definition of a function allows exactly one. A vertical line is a line, and it is not the graph of any function.
Worked example
Find the equation of the line through the two points below, where a and b are any numbers.
\[ (a, b) \qquad \text{and} \qquad (a, b+1) \]
Try the slope formula
Why: Subtract the outputs on top and the inputs underneath, as always.
\[ m = \frac{(b+1) - b}{a - a} = \frac{1}{0} \]
Name what went wrong
Why: The run is zero, and nothing may be divided by zero, so this line has no slope.
\[ \frac{1}{0} \text{ is undefined} \]
Describe the points instead of their slope
Why: Both points have the same input, and so does every point directly above or below them.
\[ x = a \]
Test a concrete case
Why: Take a equal to 4 and b equal to 1, and plot the two points.
\[ (4, 1) \text{ and } (4, 2) \;\Longrightarrow\; x = 4 \]
Check that no slope-intercept form could work
Why: Every rule of that shape returns exactly one output per input, and this line has two outputs at the input a, so no such rule exists.
\[ x = a \text{ cannot be written as } y = mx + b \]
OpenStax, Precalculus, §2.1 Linear Functions §2.1, p. 203 — Exercise 65
The letters make this look harder than it is. Treat a and b as ordinary numbers you happen not to know, and run the usual procedure.
The numerator comes out as one, which is fine, and the denominator comes out as zero, which is not. Write the word undefined rather than infinity: the slope does not exist, and infinity is not a value it takes.
When a line has no slope, describe it by what its points have in common instead. Every point on this line has input a, so the equation is x equals a. The concrete case with a equal to four is worth drawing, because it turns an abstract answer into something you can see.
Edge cases
Discussion prompt
Take the line through (2, 1) and (2 + h, 5) and compute its slope as a formula in h. Then let h shrink towards zero through the values 1, 0.1 and 0.01. What happens to the slope, and what does that say about the line you get at h equal to zero?
Compute the formula in h before reading on, and then evaluate it at the three values. Watching the number grow is more convincing than being told.
The slopes come out four, forty, four hundred, and they keep going. They do not settle on a large value; they escape. That is why a vertical line is not a line of very large slope, and why calling its slope zero, which is the other common guess, is exactly backwards.
This also explains the little condition attached to the point-slope derivation. That derivation multiplied both sides by the run, which is only legal when the run is not zero. The one case it excludes is precisely the one you have just watched break.
Counterexample
Discussion prompt
Someone says: every straight line in the plane is the graph of a linear function. Give a counterexample, and say exactly which requirement of the definition of a function it fails.
Write your counterexample before revealing. There is essentially only one family of them, so if you thought of any vertical line at all, you have it.
The second half of the question is the more important half. Saying it fails the vertical line test is true but shallow; saying that at a single input it has infinitely many outputs, and at every other input none, is the actual reason, stated in the language of the definition.
Keep the asymmetry straight. Every linear function has a line for a graph. Not every line is the graph of a linear function. The vertical lines are the only exceptions, and they are the reason the word function keeps appearing in the definitions in this section.
Two truths and a lie
Sort into buckets
Sort each statement about horizontal and vertical lines into true or false.
Sort all seven before checking. Several of them differ from each other by one word, so read each statement to the end before deciding.
The pair to be careful with is the first two. A horizontal line does have slope zero. A vertical line does not: it has no slope at all, and undefined is a different answer from zero, not a fancier way of saying it.
The last statement is the one that ties the section together. A rule of the shape m times x plus b hands back exactly one output for every input, so no such rule can ever produce a vertical line. That is a structural fact, not a limitation of your algebra.
Section
Part 7
Pattern
Figure (svg): A branching diagram: from the question what are you given, four branches lead to reading the slope and intercept, substituting into point-slope, computing the slope first, or reading two lattice points off a graph, and all four arrive at the slope-intercept form.
OpenStax, Precalculus, §2.1 Linear Functions §2.1, pp. 192, 195 — the two how-to boxes
This is the whole section on one slide, arranged by what you were handed rather than by what the book called it. Read the first step of the diagram as a question you ask yourself before writing anything.
Three of the five routes pass through point-slope form, and the fourth and fifth reduce to one of the others. So if you only ever remember one form, remember that one: it costs one extra line when you already know the intercept, and it saves you when you do not.
The last bullet is not decoration. Checking a point the rule was not built from catches every sign error in this lesson, and it takes one line. Two-point problems give you a free check, because the point you did not substitute is exactly such a point.
Real world
Figure (svg): Two lines: the Fahrenheit-from-Celsius line rising steeply and the line where the two readings are equal; they cross at minus forty, and the freezing and boiling points are marked.
Discussion prompt
Zero degrees Celsius is 32 degrees Fahrenheit, and 100 degrees Celsius is 212. Write Fahrenheit as a linear function of Celsius, find and interpret the value at 28 degrees Celsius, and then find the one temperature at which the two scales show the same number.
OpenStax, Precalculus, §2.1 Linear Functions §2.1, p. 204 — Exercise 76
Work all three parts before revealing. The first two are a routine two-point problem dressed in thermometers, and the third is the interesting one.
The slope of 1.8 says that one Celsius degree is 1.8 Fahrenheit degrees, which is why Fahrenheit numbers change faster and why the two scales drift apart as the temperature rises. Reading the slope as a conversion rate rather than as a number is the whole point of the interpretation.
For the third part, set the two readings equal and solve. Two lines with different slopes cross exactly once, so there is exactly one temperature where the scales agree, and it is minus forty. That is why weather reports at that temperature do not bother to say which scale they mean.
Check
Check your understanding
A line passes through the points (−2, 5) and (4, −7). What is its slope?
Answer: A
Why: Second minus first on both lines: the rise is minus seven minus five, which is minus twelve, and the run is four minus minus two, which is six. Minus twelve divided by six is minus two, and the negative sign agrees with the fact that the second point is lower than the first.
This is the computation from the first half of the lesson with both traps present: a negative input in the second coordinate and outputs that fall as inputs rise.
Do the sanity check before you commit. The second point is further right and lower, so the line falls and the slope must be negative. That alone rules out two of the four options.
Of the remaining two, one is the run divided into the rise and the other is the rise divided into the run. Rise over run, always, and the rise is the change in output.
Check
Check your understanding
Average annual income in dollars from 1990 to 1999 is modelled by I(x) = 1054x + 23286, where x is the number of years after 1990. Which sentence correctly interprets the slope?
Answer: C
Why: The slope is the change in output for a one-unit change in input, and the input is measured in years. So 1,054 is the yearly increase, not a total and not a starting level. The units give it away: dollars per year.
Two of these four options are true statements about the model. Only one of them is a statement about the slope, and the question asks about the slope.
Let the units settle it. The input is years and the output is dollars, so the slope is in dollars per year, and the only option that reads as an amount per year is the right one.
The option about a total over the whole decade is the tempting distractor, and it is worth computing what that total would actually be: ten times the yearly rate, so about ten and a half thousand dollars rather than one thousand and fifty-four.
Check
Check your understanding
Which statement about the line x = −6 is correct?
Answer: B
Why: Every point on this line has input minus six, so any two of them give a run of zero and the slope formula divides by zero. Division by zero is undefined, and the line assigns many outputs to the single input minus six, so it is not a function.
Read the equation carefully first. It fixes the input, not the output, and that single observation decides the whole question.
Because the input is fixed, any two points on the line have a run of zero, and the slope formula divides by zero. That is undefined, and the line also fails the definition of a function because the single input has many outputs.
The last distractor is the one to watch for under time pressure. The horizontal line at height minus six and the vertical line at input minus six share a minus six and nothing else; they meet at one point and are otherwise opposites.
Explain it to yourself
Discussion prompt
In two or three sentences, explain why knowing any two points of a linear function determines the entire function, and why that would fail for a function whose rate of change was not constant.
Write your sentences before revealing. Being able to say why a method works, and not just run it, is what turns this section into something you can rebuild when you have forgotten the details.
The two halves of a good answer are these. Two points give the slope, because on a line every pair of points gives the same slope, and the slope plus a point gives the whole rule through point-slope form. So there is exactly one linear function through two points with different inputs.
For the second half, imagine a wiggly graph through the same two points. Your quotient would be the average rate over that one interval and would say nothing about what happens between or beyond, so infinitely many functions would still fit. Constancy of the rate is what removes that freedom.
Exit ticket
\[ f \text{ is linear}, \quad f(2) = -11, \quad f(4) = -25 \]
Discussion prompt
Find the rule for f in slope-intercept form, state whether it is increasing or decreasing, and check your rule against the value you did not substitute.
OpenStax, Precalculus, §2.1 Linear Functions §2.1, p. 195 — Try It 2.5
This combines the whole lesson in one problem: function values to points, points to slope, slope and point to rule, rule back to a check.
The slope comes out at minus seven, and the negatives need care: minus twenty-five minus a minus eleven is minus fourteen, not minus thirty-six. Write the brackets in.
Finish with the check on the value you did not substitute. Substituting the same pair you built the rule from proves nothing; substituting the other one is a genuine test, and it is free.
Recap
| what you are given | which form | what to write |
|---|---|---|
| slope and the starting value | slope-intercept | y = mx + b |
| slope and any one point | point-slope | y − y₁ = m(x − x₁) |
| two points | slope first, then point-slope | m = (y₂ − y₁)/(x₂ − x₁) |
| a graph | read two lattice points exactly | then as for two points |
| a table | check equal steps, then read or solve for b | P(w) = mw + b |
| a horizontal line | constant function | y = k, slope 0 |
| a vertical line | not a function at all | x = h, slope undefined |
\[ m = \frac{y_2 - y_1}{x_2 - x_1}: \quad m > 0 \text{ increasing}, \quad m < 0 \text{ decreasing}, \quad m = 0 \text{ constant} \]
Section 2.2 takes these lines and starts moving them: shifting, stretching and reflecting graphs, and finding where two lines cross.
OpenStax Algebra and Trigonometry 2e, §4.1 Linear Functions §4.1 — the same material in Algebra and Trigonometry
The table is organised the way problems actually arrive: by what you were handed. Learn it in that direction and you will never again stare at a question wondering which form to reach for.
Two facts underneath everything. A linear function is one whose rate of change is constant, and that one condition forces the straight graph, the degree-one formula and the slope-intercept form all at once. And the slope does two jobs, sign for direction and size for steepness, which questions ask about separately.
The last row is the exception that proves the point. A vertical line is a line, it has no slope, and it is not a function, which is why it needs an equation of a different shape. Section 2.2 picks these lines up and starts moving them around: shifting, stretching and reflecting graphs, and finding where two of them cross.
Want this taught 1-on-1? Alexander tutors Precalculus — $55/session, free consultation.