The inverse functions put to work. Covers approximating the three inverses a calculator does not provide, finding the domain and range of a transformed inverse function by tracking key points, applications where a measured ratio must be converted into an angle, and the main point of the section: solving equations whose values are not special angles, using an inverse function to name the reference angle exactly as a square root names an irrational solution.
Subject: Trigonometry · 68 slides · symbolic lesson
Open the interactive version of this deck
Title
Trigonometry · Section 10.6, subsections 10.6.3 and 10.6.4
Three missing buttons, and how one arcsine names infinitely many angles
Objectives
The last two lessons built the six inverse functions and argued about their ranges. This lesson spends them. Two jobs: getting a number out of a calculator that has only three of the six, and solving an equation whose value is not on the special-angle list.
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, pp. 833-841 — subsections 10.6.3 and 10.6.4
The last two lessons were about building the inverse functions and arguing over what their ranges should be. This one spends them, and it is the part of Section 10.6 that you will actually use again in every later chapter.
There are two separate jobs here, and it is worth keeping them apart in your head. The first is arithmetic: your calculator has three inverse buttons and there are six inverse functions, so three of them have to be reached by a detour. The second is the real point of the section: solving an equation whose answer is not one of the handful of special angles you memorised.
The second job is easier than it sounds, because nothing about the method changes. You will still find a reference angle, still read the sign to pick two quadrants, still attach coterminal families. The only new thing is that the reference angle gets named by a function instead of read off a table.
Warm-up
Discussion prompt
Solve x squared equals 4, and then solve x squared equals 7. Write both answers down. What did you do differently the second time, and is the second answer any less exact than the first?
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, p. 838 — the opening analogy of subsection 10.6.4
Write both answers down before you reveal anything. This takes ten seconds and it sets up the whole lesson.
Almost everybody solves the first one instantly and hesitates on the second. But look at what you actually did: nothing changed except that the second answer needed the square root symbol to name it. You did not approximate, you did not give up, and you would not accept the claim that the second equation has no solution.
Hold on to that reaction. When you meet the sine of theta equals one third in a few slides, the temptation is to say it has no nice answer and reach for a decimal. The honest answer is the same shape as the root of seven: a function applied to the number, written down exactly, and left alone.
Section
Part 1
Concept
Almost every calculator offers arcsine, arccosine and arctangent, usually printed as sine to the minus one and so on. Arccotangent, arcsecant and arccosecant are not there. Each of the three has a reciprocal identity that trades it for one that is.
\[ \operatorname{arccot}(x) = \arctan\!\left(\tfrac1x\right) \quad (x > 0) \]
\[ \operatorname{arcsec}(x) = \arccos\!\left(\tfrac1x\right), \qquad \operatorname{arccsc}(x) = \arcsin\!\left(\tfrac1x\right) \]
| identity | valid when | from |
|---|---|---|
| arccot(x) = arctan(1/x) | x > 0 only | Theorem 10.27 |
| arcsec(x) = arccos(1/x) | |x| ≥ 1 | Theorem 10.28 ranges |
| arcsec(x) = arccos(1/x) | x ≥ 1 only | Theorem 10.29 ranges |
| arccsc(x) = arcsin(1/x) | the same split | both theorems |
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, p. 833 — subsection 10.6.3, opening paragraph
Look at your calculator. You will find keys marked sine to the minus one, cosine to the minus one and tangent to the minus one, and you will not find the other three. This is not an oversight; it is that three buttons are enough, because each missing function is a reciprocal away from one that is there.
The three identities in the table are all doing the same algebraic thing: the cotangent of an angle is one over its tangent, so an angle with cotangent x is an angle with tangent one over x. That part is uncontroversial. What varies from row to row is the condition on the right, and the condition is never about the algebra.
The conditions are about ranges. Two functions can agree on a value and still disagree about which angle to return, and they agree only where their ranges overlap. Notice that the two arcsecant rows carry different conditions even though the formula is identical, which is a direct consequence of the two conventions from the last lesson.
Notation
Annotate
On: \( \operatorname{arccot}(x) = \arctan\!\left(\tfrac{1}{x}\right), \quad x > 0 \)
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, p. 833 — Theorem 10.27, restated
Work through the four annotations in order and make sure the second one lands, because it is the one that gets skipped. The identity is not fragile because of any division by zero; it is fragile because the two functions involved return angles from different intervals.
Here is the clean way to see it. Arccotangent must answer somewhere strictly between zero and pi, which is the upper half of the plane. Arctangent must answer between negative pi over two and pi over two, which is the right-hand half. The only place both can answer is the first quadrant, and a positive input is exactly what sends you there.
Once you have that, the repair is obvious rather than magic. Feed the identity the size of the number, which is positive, get the reference angle, and then place that reference angle in whichever quadrant the range and the sign actually demand. Every example in Part 1 is that one move.
Worked example
Approximate to four decimal places.
\[ \operatorname{arccot}(2) \]
Check the identity's condition
Why: The input 2 is positive, so Theorem 10.27 applies as it stands.
\[ 2 > 0 \;\Longrightarrow\; \operatorname{arccot}(2) = \arctan\!\left(\tfrac12\right) \]
Punch in the arctangent in radian mode
Why: One half is the reciprocal of two.
\[ \arctan(0.5) \approx 0.4636 \]
Verify it lands in the arccotangent's range
Why: The range is everything strictly between zero and pi, and 0.4636 is in it; the cotangent of 0.4636 is 1.9998, which is 2 to the accuracy carried.
\[ 0 < 0.4636 < \pi, \qquad \cot(0.4636) \approx 1.9998 \]
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, p. 833 — Example 10.6.5, part 1a
This is the easy case and it is worth doing slowly, because the habit you build here is the one that saves you later. The first step is not a calculation at all; it is checking that the identity is allowed. Two is positive, so it is.
Make sure your calculator is in radian mode. This section is about real numbers, and a calculator left in degree mode will hand you 26.57 instead of 0.4636 with no warning. A quick sanity check: any arctangent answer in radians has size below about 1.571, so a number like 26 is a mode error, not a maths error.
The last step is the one to keep. Asking whether the answer lies in the function's own range costs one glance, and it will catch every single branch mistake that Part 1 is about. Here 0.4636 is comfortably between zero and pi, so nothing has gone wrong.
Worked example
Approximate to four decimal places.
\[ \operatorname{arcsec}(5) \]
Check the condition under both conventions
Why: Five is at least one, so Theorem 10.28 and Theorem 10.29 both give the same identity here.
\[ 5 \ge 1 \;\Longrightarrow\; \operatorname{arcsec}(5) = \arccos\!\left(\tfrac15\right) \]
Evaluate the arccosine
Why: One fifth is 0.2.
\[ \arccos(0.2) \approx 1.3694 \]
Verify by taking the secant back
Why: The cosine of 1.3694 is 0.2000, so its reciprocal is 5.000; and 1.3694 lies in the first quadrant, which both conventions include.
\[ \sec(1.3694) = \frac{1}{\cos(1.3694)} \approx 4.9991 \]
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, p. 833 — Example 10.6.5, part 1b
The interesting word in the first step is either. Five is at least one, so it is a positive input, and both conventions from the last lesson supply the same identity for it. That means this answer is convention free: whichever book your course follows, the arcsecant of five is the same number.
That is a general fact worth carrying: the two conventions only ever disagree about negative inputs. For positive ones, both ranges include the first quadrant and both send you there, so the whole argument evaporates.
The check runs the calculation backwards. You take the cosine of your answer and see whether its reciprocal is five. It comes out as 4.9991 rather than exactly 5, and that is the rounding in the four-decimal angle showing up, not an error. If it had come out as, say, 1.02, you would know immediately that something had gone wrong.
Intuition
Every one of the six inverse functions contains the whole of the first quadrant in its range. So when the input is positive and the answer is acute, all six agree with each other and with every reciprocal identity, and there is nothing to think about.
A negative input is where the ranges part company. Arccosine goes on into the second quadrant, arcsine turns back into the fourth, arccotangent goes into the second, and arcsecant and arccosecant go wherever the convention in force sends them.
So the rule of thumb is short. Positive input: use the identity. Negative input: find the reference angle from the positive number, then place it yourself.
It is worth pausing on why this is true rather than just remembering it. Every one of the six restricted functions was built to be one to one while still covering all the values, and the cheapest way to start is at the first quadrant, where every trigonometric function is positive and well behaved. So all six ranges contain the first quadrant.
The disagreements start when the restriction has to continue somewhere. Cosine and secant continue into the second quadrant, sine and cosecant turn back into the fourth, and secant and cosecant have an asymptote in the way, which is what forces the two rival conventions.
So the practical rule is short enough to keep in your head under exam pressure. Positive input, use the identity and move on. Negative input, stop and do the detour. If you only remember one thing from Part 1, remember that the sign of the input is the thing that decides how much work there is.
Concept
Suppose you want an inverse function of a negative number and the identity refuses. Call the answer theta. Three facts pin it down, and none of them needs a calculator.
\[ \text{QII: } \theta = \pi - \alpha, \qquad \text{QIII: } \theta = \pi + \alpha, \qquad \text{QIV: } \theta = 2\pi - \alpha \;\text{ or}\; -\alpha \]
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, p. 834 — the reference-angle approach of Example 10.6.5, part 1c
This slide is the engine for the three examples that follow, so read it as a procedure rather than as a fact. The three numbered facts narrow the answer down from an infinite set to a single angle without any calculator work at all.
The order matters. First the range tells you half the plane, because that is what the definition of the function guarantees. Then the sign of the value cuts that half down to a single quadrant, because each trigonometric function is positive in exactly two quadrants and negative in the other two. Only then do you compute anything.
The three placement formulas at the bottom are the same ones you have used since Section 10.2. The reference angle is always measured to the nearest part of the horizontal axis, never to the vertical one, so second quadrant means pi minus it and third means pi plus it. If you find yourself writing pi over two minus something, you have measured to the wrong axis.
Worked example
Approximate to four decimal places. The identity is unavailable, so run the detour.
\[ t = \operatorname{arccot}(-2), \quad \cot(t) = -2, \quad 0 < t < \pi \]
Narrow the quadrant
Why: Cotangent is negative in the second and fourth quadrants, and the range rules out the fourth.
\[ \cot(t) < 0 \;\text{and}\; 0 < t < \pi \;\Longrightarrow\; \tfrac{\pi}{2} < t < \pi \]
Write the reference angle
Why: The Reference Angle Theorem strips the sign, so alpha has cotangent positive two.
\[ \cot(\alpha) = 2, \quad 0 < \alpha < \tfrac{\pi}{2} \]
Now the identity is legal
Why: Its argument is a positive 2.
\[ \alpha = \operatorname{arccot}(2) = \arctan\!\left(\tfrac12\right) \approx 0.4636 \]
Place the reference angle in the second quadrant
Why: A second-quadrant angle is pi minus its reference angle.
\[ t = \pi - \arctan\!\left(\tfrac12\right) \approx 3.1416 - 0.4636 = 2.6779 \]
Figure (svg): A unit circle with a ray into the second quadrant meeting the circle at the point with coordinates about negative 0.894 and 0.447. A yellow arc from the positive x-axis around to the ray is marked theta, and a short green arc between the ray and the negative x-axis is marked alpha, the reference angle.
Check the value and the range
Why: The cotangent of 2.6779 is negative 1.9998, which is negative two to four figures, and 2.6779 is between pi over two and pi as required.
\[ \cot(2.6779) \approx -1.9998, \qquad 1.5708 < 2.6779 < 3.1416 \]
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, p. 834 — Example 10.6.5, part 1c
This is the full detour, and it is longer than the calculation deserves precisely because every step is written out. In practice you will compress it to two lines, but you should be able to expand it again when a problem surprises you.
The first step does the real work. Cotangent is negative in the second and fourth quadrants, but the arccotangent's range is only from zero to pi, which rules the fourth quadrant out. One quadrant survives, so the answer is determined before any button is pressed.
Look at the figure and find the two arcs. The long yellow arc from the positive horizontal axis round to the ray is theta, the number you want. The short green arc from the ray to the negative horizontal axis is the reference angle, and its cotangent is positive two because the triangle it sits in has both legs positive in length. The whole detour is just reading the long arc as a half turn minus the short one.
Worked example
The book gives a second route to the same number, and it is the one worth remembering because it is shorter.
\[ \beta = \arctan\!\left(-\tfrac12\right), \quad \tan(\beta) = -\tfrac12, \quad -\tfrac{\pi}{2} < \beta < 0 \]
Notice what the wanted angle's tangent is
Why: Cotangent negative two means tangent negative one half, since the two are reciprocals.
\[ \cot(t) = -2 \;\Longleftrightarrow\; \tan(t) = -\tfrac12 \]
Read where the calculator's answer sits
Why: A negative tangent sends the arctangent into the fourth quadrant.
\[ \beta \approx -0.4636 \]
Move a half turn to the wanted quadrant
Why: Angles a half turn apart have the same tangent, and t is the second-quadrant one.
\[ t = \pi + \beta = \pi + \arctan\!\left(-\tfrac12\right) \]
Evaluate
Why: Add pi to the calculator's negative answer.
\[ t \approx 3.1416 - 0.4636 = 2.6779 \]
Figure (svg): The same unit circle with a full line through the origin. One end of the line points into the fourth quadrant at the angle beta, which is arctangent of negative one half; the other end points into the second quadrant at beta plus pi, which is the angle wanted. A yellow arc of half a turn joins the two directions.
Verify the two routes agree
Why: The reference-angle route gave pi minus 0.4636 and this one gives pi plus negative 0.4636. These are the same arithmetic, because the arctangent is odd, so both give 2.6779.
\[ \pi - \arctan\!\left(\tfrac12\right) = \pi + \arctan\!\left(-\tfrac12\right) \approx 2.6779 \]
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, pp. 834-835 — Example 10.6.5, part 1c, second solution
The book gives this second solution and it is the one you should actually use, because it is two lines rather than five. The idea: instead of stripping the sign off the value, keep the negative and let the calculator answer in the wrong quadrant, then correct by a half turn.
Why is a half turn the right correction? Because tangent and cotangent both have period pi. Look at the figure: the dashed line through the origin has two ends, and both ends have the same tangent, since tangent is the slope of the line and both ends lie on one line. So adding pi to any answer gives another angle with the same tangent, and exactly one of the two lies in the range you need.
Convince yourself the two routes are the same arithmetic. One says pi minus the arctangent of one half; the other says pi plus the arctangent of negative one half. Arctangent is an odd function, so those two expressions are identical. If you ever get different answers from the two routes, one of them has a sign error.
Trap
Asked for the arccotangent of negative two, a student types the reciprocal and presses the arctangent key:
\[ \arctan\!\left(-\tfrac12\right) \approx -0.4636 \]
\[ \operatorname{arccot}(-2) \overset{?}{=} -0.4636 \]
Wrong. The number is a real angle, but it is not this function's value.
Arccotangent returns a number strictly between zero and pi. Negative 0.4636 is not in that interval, so it cannot be the answer whatever the algebra suggested. Add a half turn to land in the range.
\[ 0 < \operatorname{arccot}(x) < \pi \]
\[ -0.4636 + \pi \approx 2.6779 \]
This is the single most common mistake in Part 1, and it is not carelessness. The algebra genuinely does say that an angle with cotangent negative two has tangent negative one half, and negative 0.4636 genuinely is such an angle. Nothing in the calculation is false.
What is false is the claim that this angle is the value of the arccotangent function. A function returns one specific output, chosen in advance by its range, and negative 0.4636 is not in the interval from zero to pi. Getting a true statement about angles is not the same as answering the question asked.
Build the range check into your reflexes. After every inverse evaluation, look at your answer and ask whether it lies where that function is allowed to land. It takes a second and it catches this every time.
Worked example
This one has two different right answers, because the two conventions disagree about negative inputs.
\[ \operatorname{arccsc}\!\left(-\tfrac32\right) \]
Under the Theorem 10.28 ranges, use the identity
Why: There the identity holds for the whole domain, including negatives.
\[ \operatorname{arccsc}\!\left(-\tfrac32\right) = \arcsin\!\left(-\tfrac23\right) \approx -0.7297 \]
Under the Theorem 10.29 ranges, the identity fails
Why: There it is stated for inputs at least one only, so start from the definition.
\[ t = \operatorname{arccsc}\!\left(-\tfrac32\right), \quad \csc(t) = -\tfrac32, \quad \pi < t \le \tfrac{3\pi}{2} \]
Write the reference angle
Why: Stripping the sign gives cosecant three halves, so sine two thirds.
\[ \alpha = \operatorname{arccsc}\!\left(\tfrac32\right) = \arcsin\!\left(\tfrac23\right) \approx 0.7297 \]
Place it in the third quadrant
Why: A third-quadrant angle is pi plus its reference angle.
\[ t = \pi + \arcsin\!\left(\tfrac23\right) \approx 3.1416 + 0.7297 = 3.8713 \]
Figure (svg): A unit circle with a ray into the third quadrant meeting the circle where the height is negative two thirds. A yellow arc runs from the positive x-axis all the way round past pi to the ray, marked theta, and a short green arc between the ray and the negative x-axis is marked alpha.
Check both answers against their own ranges
Why: The cosecant of negative 0.7297 and the cosecant of 3.8713 are both exactly negative 1.5, and the two angles differ by pi. Each sits inside its own convention's range and outside the other's, so each is right in its own book and wrong in the other.
\[ \csc(-0.7297) = \csc(3.8713) = -1.5, \qquad 3.8713 - (-0.7297) \approx \pi + \pi \]
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, p. 835 — Example 10.6.5, part 1d
This is the example that shows why the convention argument in the last lesson was not academic. The same three symbols on the page, arccosecant of negative three halves, mean two different numbers depending on which book you are holding, and both numbers are correct.
Follow the two routes separately. Under the Theorem 10.28 ranges the identity is stated for the whole domain, negatives included, so you simply take the arcsine of negative two thirds and you are done. Under the Theorem 10.29 ranges the identity is only for inputs at least one, so you fall back to the definition, find the reference angle from the positive input, and place it in the third quadrant.
The check at the end is the important beat. Both answers have cosecant exactly negative 1.5, and they differ by a half turn. Neither is more correct than the other; each is right in its own book. If a question ever asks you for an arcsecant or arccosecant of a negative number without saying which convention, the honest answer is to give both and say so.
Picture it
Figure (svg): A horizontal axis of angle values from negative pi over two to three pi over two. An upper orange band covers the range of Theorem 10.28, from negative pi over two up to pi over two with zero removed, and carries a dot at negative 0.7297. A lower purple band covers the range of Theorem 10.29, from zero to pi over two and from pi to three pi over two, and carries a dot at 3.8713.
Nothing mathematical is in dispute here. Both dots name an angle whose cosecant is negative three halves, and both are correct answers to the question what is the arccosecant of negative three halves. What differs is which band of angles the author decided to call the range.
The practical consequence: whenever a negative input goes into an arcsecant or an arccosecant, say which convention you are using, or avoid the function altogether. The last part of this lesson takes the second option.
Read this picture as two menus rather than two answers. Each band is one author's decision about which angles the arccosecant is allowed to return, and the dot on each band is the member of that menu whose cosecant is negative three halves.
Notice that both bands contain the whole of the first quadrant, from zero to pi over two. That is the overlap that makes positive inputs safe, and you can see at a glance that there is no disagreement anywhere in that stretch.
The practical lesson is the one in the second paragraph on the slide, and it is what Part 5 is built on. You are almost never forced to use these two functions. Every secant equation is a cosine equation in disguise, so you can reciprocate at the start and stay in territory that everybody agrees about.
Fill the middle
The detour, once more, with a different number. Fill in the two blanks.
Fill in the blanks
\operatorname3\!\left(-\tfrac13\right) = \pi - \arctan(1.2490) \approx 3.1416 - ___ \approx 1.8925
Why: The reference angle has cotangent positive one third, so its tangent is 3 and it equals the arctangent of 3, about 1.2490. A negative cotangent puts the answer in the second quadrant, so subtract that reference angle from pi and get about 1.8925, which is safely between zero and pi.
Do not reveal until you have both blanks written down. The first blank is the trap: the reference angle has cotangent one third, and it is the tangent you need for the calculator, so you must flip it to three rather than typing one third.
The second blank is the arctangent of three, about 1.2490, and you should be able to sanity-check it without a calculator. A tangent of three is steep, so the angle should be well past halfway to a right angle; a right angle is about 1.5708 radians, and 1.2490 is comfortably below it. If you had typed one third by mistake you would get about 0.3217, which is far too small.
Finish by checking the range, as always. The answer 1.8925 lies between zero and pi, so it is a legal arccotangent value, and it is past pi over two, which is where a negative cotangent belongs.
Sorting
Sort into buckets
Sort each request by whether the reciprocal identity can be used straight away, or whether a reference angle has to be found first.
Sort all six before revealing. The fastest way is not to recall the theorems but to ask one question of each: is the input positive? If it is, there is nothing to think about. Four of the six go straight into the first bucket on that test alone.
The two arcsecant of negative eight rows are the interesting ones, and they are deliberately identical except for the convention. Under the Theorem 10.28 ranges the arcsecant continues into the second quadrant, exactly where the arccosine goes, so the identity covers negatives too. Under the Theorem 10.29 ranges it continues past pi toward three pi over two, where no arccosine can reach, so the identity is restricted to positive inputs and a negative one needs the detour.
If you got that pair the wrong way round, go back to the previous slide's picture and look at which band reaches past pi. Reaching past pi is exactly what the arccosine cannot do.
Prediction
You ask a calculator, by way of the arccosine, for the arcsecant of 1.001.
Predict first
Roughly what should come back?
Correct: Close to zero
Why: The secant of zero is exactly 1, so the arcsecant of exactly 1 is zero. An input a thousandth above 1 gives an angle a little above zero, about 0.0447. Both conventions include zero in the range, so there is no disagreement here and certainly no error, since 1.001 is inside the domain.
\[ \operatorname{arcsec}(1.001) = \arccos\!\left(\tfrac{1}{1.001}\right) \approx 0.0447 \]
Commit to an option before revealing. The reasoning takes one sentence, and it does not need any inverse function at all.
Start from a value you know: the secant of zero is one, because the cosine of zero is one. So the arcsecant of exactly one is zero. Nudging the input up to 1.001 nudges the output up from zero by a small amount, and the answer is about 0.0447.
If you were tempted by the error option, check the domain. The arcsecant accepts every number of size at least one, and 1.001 is comfortably inside. It is inputs between negative one and one, like 0.75, that produce an error, because no angle has a secant that small in size.
Section
Part 2
Concept
A function such as three times the arctangent of four x is an inverse function with things done to it. Domain and range can both be read off, and each comes from a different part of the formula.
\[ f(x) = A\,F(Bx + C) + D \]
\[ B, C \;\longrightarrow\; \text{the domain}, \qquad A, D \;\longrightarrow\; \text{the range} \]
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, p. 836 — Example 10.6.5, part 2, and Theorem 1.7
This division of labour is not special to trigonometry. It is the same rule you used for shifting and stretching parabolas, and it is stated in the book as Theorem 1.7. What makes it feel new here is only that the parent functions have unusual domains and ranges to begin with.
Say the rule to yourself in the form that makes it obvious: whatever is inside the function decides which inputs are legal, and whatever is done to the answer afterwards decides which outputs appear. Nothing inside can change what the parent is capable of producing, and nothing outside can make an illegal input legal.
The single most useful habit for this whole part is to write the argument on its own line before you do anything else. Half the errors in domain questions come from imposing a condition on the whole expression instead of on the argument, and writing the argument down separately makes that mistake nearly impossible.
Concept
What you track through the outer operations depends on which parent function you started from, because the two pairs are shaped differently.
| parent | what to track | the three or two features |
|---|---|---|
| arccos(x) | three key points | (−1, π), (0, π/2), (1, 0) |
| arcsin(x) | three key points | (−1, −π/2), (0, 0), (1, π/2) |
| arctan(x) | two horizontal asymptotes | y = −π/2 and y = π/2 |
| arccot(x) | two horizontal asymptotes | y = 0 and y = π |
Points give closed ends, so an arcsine or arccosine range uses square brackets. Asymptotes are approached and never reached, so an arctangent or arccotangent range uses round ones.
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, p. 836 — the procedure of Theorem 1.7 applied to the inverses
The two pairs behave differently and it is worth knowing why, not just which is which. Arccosine and arcsine come from restricting cosine and sine to closed intervals, so their graphs genuinely start and stop: there are endpoints to track.
Arctangent and arccotangent come from restricting functions that run off to infinity, so their graphs never arrive anywhere. What you track instead is the two horizontal lines they creep toward. Three points versus two asymptotes is the whole difference.
That difference shows up in your final answer as brackets. An endpoint is reached, so it gets a square bracket and the range is closed. An asymptote is never reached, so it gets a round bracket and the range is open. Marking a transformed arctangent's range as closed is a small slip that is easy to avoid once you know where the brackets come from.
Worked example
Find the domain and the range, and check the answer against the graph.
\[ f(x) = \frac{\pi}{2} - \arccos\!\left(\frac{x}{5}\right) \]
Put the argument in the arccosine's domain
Why: The arccosine accepts inputs from negative one to one inclusive.
\[ -1 \le \frac{x}{5} \le 1 \]
Solve for x
Why: Multiply all three parts by five.
\[ -5 \le x \le 5 \;\Longrightarrow\; \text{domain } [-5, 5] \]
List the parent's three key points
Why: These are the ends and the middle of the arccosine graph.
\[ (-1, \pi), \quad \left(0, \tfrac{\pi}{2}\right), \quad (1, 0) \]
Send the inputs through the division by five
Why: Only the first coordinate changes.
\[ (-5, \pi), \quad \left(0, \tfrac{\pi}{2}\right), \quad (5, 0) \]
Send the outputs through the negate-and-add
Why: Each height y becomes pi over two minus y.
\[ \left(-5, -\tfrac{\pi}{2}\right), \quad (0, 0), \quad \left(5, \tfrac{\pi}{2}\right) \]
Read the range off the tracked heights
Why: The three heights run from negative pi over two up to pi over two, and the curve is continuous, so it covers everything between.
\[ \text{range } \left[-\tfrac{\pi}{2}, \tfrac{\pi}{2}\right] \]
Figure (svg): Two graphs side by side. On the left, y equals arccosine of x, running from the point negative one comma pi down to one comma zero, with its three key points marked. On the right, y equals pi over two minus arccosine of x over five, running from negative five comma negative pi over two up to five comma pi over two, with the three tracked points marked.
Check against a known identity
Why: Pi over two minus the arccosine of u is exactly the arcsine of u, so this function is the arcsine of x over five. Its range should therefore be the arcsine's range, and it is.
\[ \frac{\pi}{2} - \arccos(u) = \arcsin(u) \;\Longrightarrow\; f(x) = \arcsin\!\left(\tfrac{x}{5}\right) \]
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, p. 836 — Example 10.6.5, part 2a
Notice how the two halves of the question never touch each other. The domain came from one inequality about x over five and nothing else. The range came from three heights pushed through two operations and nothing else. If you find yourself using the outer minus sign while computing the domain, you have crossed the wires.
Track the three points slowly. First the inputs get multiplied by five, because setting x over five equal to negative one means x is negative five. The heights are untouched by that. Then the heights get negated and shifted, and the inputs are untouched by that. Two independent passes.
The check is the best part of this example. Pi over two minus the arccosine of something is the arcsine of that something, by the cofunction relationship, so this whole function was the arcsine of x over five wearing a disguise. That is why the range came out so tidily, and it is a good reminder that recognising a result is a legitimate way of checking it.
Notation
Annotate
On: \( f(x) = \tfrac{\pi}{2} - \arccos\!\left(\tfrac{x}{5}\right) \)
Step through the four annotations and connect each one to the picture on the previous slide. The first two describe the inside of the function and they are the only ones that had any effect on the domain.
The minus sign is the one people undersell. It does two things at once: it flips the graph upside down, which is why the curve now rises from left to right instead of falling, and it reverses the order of the range, so the old top becomes the new bottom. If you track the endpoints rather than trying to keep the interval in order, this looks after itself.
The shift by pi over two is the last operation applied, so it is the last one you apply to the heights. Order matters here in the same way it matters for any composition: negate first, then shift, because that is how the formula is written.
Worked example
Same question, different parent. Here there are no endpoints to track, only asymptotes.
\[ f(x) = 3\arctan(4x) \]
Put the argument in the arctangent's domain
Why: The arctangent accepts every real number, and four x is defined for every real x.
\[ \text{domain } (-\infty, \infty) \]
Write down the parent's range
Why: The arctangent is squeezed between two asymptotes and reaches neither.
\[ -\frac{\pi}{2} < \arctan(u) < \frac{\pi}{2} \]
Multiply through by three
Why: The inequality keeps its direction because three is positive.
\[ -\frac{3\pi}{2} < 3\arctan(4x) < \frac{3\pi}{2} \]
State the range
Why: Both ends stay open, because an asymptote stays an asymptote when you stretch.
\[ \text{range } \left(-\frac{3\pi}{2}, \frac{3\pi}{2}\right) \]
Figure (svg): Two curves on one set of axes. A faint dashed curve is y equals arctangent of x, flattening toward horizontal lines at plus and minus pi over two. A bold curve is y equals three times arctangent of four x, much steeper near the origin and flattening toward horizontal lines at plus and minus three pi over two.
Check what the four actually did
Why: Put x equal to one quarter: the inside becomes 1 and the output is three times arctangent 1, which is three pi over four, about 2.3562. The same height on the plain arctangent needs x about 2.4142, so the four compressed the picture sideways and left the two ceilings alone.
\[ f\!\left(\tfrac14\right) = 3\arctan(1) = \tfrac{3\pi}{4} \approx 2.3562 \]
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, p. 836 — Example 10.6.5, part 2b
Compare this with the previous example line by line and notice what changed. The domain question got easier, because the arctangent accepts every real number, so there is no inequality to solve at all. The range question changed shape, because there are no endpoints to track.
The three-line inequality in the middle is the whole argument. Start from the parent's range written as an inequality, multiply through by the outside factor, and read off the result. Multiplying by a positive number keeps the inequality pointing the same way; if the outside factor had been negative, you would have to flip it.
Look at the figure and find the four dashed lines. The inner pair are the parent's ceilings and the outer pair are the new ones, three times as far out. The bold curve rushes to its ceilings much faster than the faint one, and that speed is entirely the work of the four inside. Speed is not range.
Tweak it
Parameter explorer
The curve is A times the arctangent of B times x. Slide A, then slide B. One knob changes how high and how low the curve can ever get; the other only changes how fast it gets there. Which is which, and so which one controls the range?
\[ y = {A}\arctan({B}\cdot x) \]
Play with this before you read the answer. Take A up and down first, with B left alone, and watch the two flat ends of the curve move apart and together. Then put A back and take B from small to large, and watch the curve stand up steeply near the middle while the flat ends stay exactly where they were.
That is the division of labour made visible. The outside factor A is multiplying the output, so it stretches the set of outputs, which is the range. The inside factor B is multiplying the input, so it changes which x gives which output, and the collection of available outputs is untouched.
The curve here is drawn without axes on purpose, so do not try to read numbers off it. Watch the shape only: how far apart the flat ends are, and how quickly the curve gets to them. Those two features are exactly what A and B control.
Intuition
Think of the transformed function as a pipeline. A number enters, the inside operations reshape it, the parent inverse function converts it into an angle, and then the outside operations reshape that angle.
Whatever the inside does, it hands the parent a number, and the parent's output set is fixed once and for all by its own range. Squeezing or shifting the inputs changes which x produces which output, but not the collection of outputs available.
The outside sits downstream of the parent, so it is the only thing that can change that collection. That is why the division of labour is exact, not approximate, and why it is worth writing the argument down on its own line before imposing any condition.
If the rule still feels like something to memorise, this pipeline picture is worth twenty seconds of thought, because after it the rule stops being arbitrary.
The parent inverse function is a gate. Whatever reaches it, it emits an angle from its own fixed range and nothing else. Reshaping the numbers before the gate changes which of them arrives when, but the gate's repertoire is unchanged, so the set of possible outputs is unchanged.
Everything after the gate is a different matter, because it operates on the outputs themselves. Stretch them and the set spreads out; shift them and the set slides. That is the only way the range can move, so the only place to look when a range question is asked is outside the function.
Worked example
The last part of the example, and the one that forces a piecewise formula on a calculator.
\[ g(x) = \operatorname{arccot}\!\left(\frac{x}{2}\right) + \pi \]
Find the domain
Why: The arccotangent accepts every real number, and so does half of x.
\[ \text{domain } (-\infty, \infty) \]
Take the parent's asymptotes through the shift
Why: Adding pi lifts both of them by pi.
\[ 0 < \operatorname{arccot}(u) < \pi \;\Longrightarrow\; \pi < g(x) < 2\pi \]
Rewrite for a calculator when x is positive
Why: There the reciprocal identity is legal, since half of x is positive.
\[ x > 0: \quad g(x) = \arctan\!\left(\frac{2}{x}\right) + \pi \]
Rewrite for a calculator when x is negative
Why: The arccotangent of a negative needs the half-turn correction from Part 1, which adds another pi.
\[ x < 0: \quad g(x) = \arctan\!\left(\frac{2}{x}\right) + 2\pi \]
Handle x equal to zero separately
Why: Two over zero is undefined, so neither arctangent formula reaches this point; the arccotangent of zero is pi over two.
\[ g(0) = \operatorname{arccot}(0) + \pi = \frac{\pi}{2} + \pi = \frac{3\pi}{2} \]
Figure (svg): The graph of y equals arccotangent of x over two, plus pi. It falls smoothly from just below the horizontal line at two pi on the far left to just above the horizontal line at pi on the far right, passing through the point zero comma three pi over two, which is circled.
Verify the three pieces join up
Why: As x comes down to zero from the right, two over x runs off to infinity and the arctangent creeps up to pi over two, so g creeps up to three pi over two. From the left it creeps down to the same value. The single point matches both one-sided limits, so the graph has no break. Note that the book prints this middle value as pi; three pi over two is correct, and the printed graph and range both agree with three pi over two.
\[ \lim_{x\to 0^{+}} g(x) = \frac{\pi}{2} + \pi = \frac{3\pi}{2} = \lim_{x\to 0^{-}} g(x) \]
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, pp. 836-837 — Example 10.6.5, part 2c
The domain and range here take two lines, and then the example turns into something else: a lesson in what a calculator cannot do. Do not skim the last three steps, because they are the reason the book included this part.
The problem is that a graphing calculator has no arccotangent key, so you have to feed it arctangents. For positive x the identity works. For negative x the identity lands you a half turn too low, so you add pi to correct it, exactly as in Part 1. And at x equal to zero, two over x does not exist, so neither formula says anything and you have to supply the value yourself.
One warning about the book here. It prints the value at zero as pi, but the arccotangent of zero is pi over two, so the correct value is three pi over two. The printed graph and the printed range both agree with three pi over two, so this is a slip rather than a different convention. It is a useful slip to notice: whenever you write a piecewise formula, check that the middle piece agrees with what the two outer pieces are heading toward.
Socratic
Discussion prompt
The arccotangent of half x, plus pi, is a single smooth curve. Yet to graph it a calculator needs three separate arctangent formulas. Explain what goes wrong with using just the positive-x formula everywhere.
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, p. 837 — the piecewise calculator formula
Answer in writing before revealing. The question is really about a jump, and the quickest way to find it is to ask what the arctangent of two over x does as x approaches zero from each side.
From the right, two over x is enormous and positive, so the arctangent creeps up to pi over two. From the left, two over x is enormous and negative, so the arctangent creeps down to negative pi over two. The formula jumps by a whole pi across zero, even though the function it is supposed to represent is perfectly smooth there.
So the extra pi in the negative piece is not a fudge. It is precisely the amount needed to cancel the arctangent's own jump, and once it is added the two pieces line up at three pi over two and the graph is continuous. The isolated point at x equal to zero is there only because the formula divides by x.
Error analysis
A student is asked for the domain of twice the arcsine of x, plus one, and writes this.
Annotate
On: \( -1 \le 2\arcsin(x) + 1 \le 1 \;\Longrightarrow\; \text{domain} = [-1, 0] \)
Write the argument on a line of its own before imposing anything. That one habit removes this error permanently.
Read the wrong line before the annotations and see whether you can say what is wrong in one sentence. Most people spot that something is off but take a while to name it.
The instinct was right and that is what makes the error durable. Sandwiching something between negative one and one is exactly what an arcsine problem needs. The mistake is purely about which something: the condition belongs to the argument, and here the argument is just x.
Notice what the written inequality actually describes. Solving it tells you where the output of the whole function sits between negative one and one, which is a real question but a completely different one. Getting a sensible-looking interval out of an incorrect setup is why this error survives self-checking, and why writing the argument on its own line first is worth the ten seconds.
Comparison
Comparison matrix
| function | domain | range |
|---|---|---|
| arcsin(x/3) | [−3, 3] | [−π/2, π/2] |
| 3 arctan(4x) | all real numbers | (−3π/2, 3π/2) |
| π/2 − arccos(x/5) | [−5, 5] | [−π/2, π/2] |
| arccot(x/2) + π | all real numbers | (π, 2π) |
| arcsec(12x) | |x| ≥ 1/12 | the arcsecant's own range |
The last row is Exercise 230. The arcsecant's domain is everything of size at least one, so twelve x must be at least one in size, which makes x at least one twelfth in size. Nothing outside the function appears at all, so the range is untouched.
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, p. 847 — Exercises 222 to 233
Fill every blank before revealing, and say out loud which part of each formula you are using. For a domain blank you should be looking only at what is inside; for a range blank only at what is outside.
The first three rows are the examples you have just done, in a different order, so they should be quick. The fourth is the one from Example 10.6.5 part 2c, where the shift moved the pair of asymptotes and the round brackets stayed round.
The last row is different in kind and it is worth dwelling on. The arcsecant's domain is not an interval; it is everything of size at least one, which is two separate pieces. Putting twelve x into that condition gives x of size at least one twelfth, again in two pieces. Nothing is done to the output, so the range is whatever your course's arcsecant convention says, untouched.
Fill the middle
Find the domain of the arccosine of the quantity three x minus one, over two.
Fill in the blanks
-1 \le \frac-11 \le 1 \;\Longrightarrow\; -2 \le 3x-1 \le 2 \;\Longrightarrow\; ___ \le 3x \le 3 \;\Longrightarrow\; -\tfrac13 \le x \le ___
Why: Multiply all three parts by two, then add one to all three parts, then divide all three by three. The domain is the closed interval from negative one third to one. Notice that only the argument was ever touched.
This is a pure domain question, so nothing outside the arccosine matters, and in fact there is nothing outside it. Work on the argument alone.
The mechanics are a three-part inequality, which you handle by doing the same thing to all three parts: multiply by two, add one, divide by three. The only place to slip is forgetting to apply an operation to the middle part as well as the two ends.
Check the answer by testing an endpoint. Put x equal to one: the argument becomes one, which the arccosine accepts, and the answer is zero. Put x equal to negative one third: the argument becomes negative one, also accepted, and the answer is pi. Anything outside that interval pushes the argument past one in size and the function stops existing.
Section
Part 3
Concept
Almost every practical use of an inverse trigonometric function has the same shape. Two lengths have been measured, so a ratio is known, and an angle is wanted. Which inverse function you reach for is not a matter of taste: it is decided by which two sides were measured.
| what you measured | the ratio is a | so the angle is |
|---|---|---|
| opposite and adjacent | tangent | arctan(opp / adj) |
| opposite and hypotenuse | sine | arcsin(opp / hyp) |
| adjacent and hypotenuse | cosine | arccos(adj / hyp) |
For an acute angle none of the convention arguments from Part 1 can bite, because every inverse function's range already contains the whole of the first quadrant. That is why applied problems are the easy case.
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, pp. 837-838 — the paragraph introducing Example 10.6.6
This is the most common way you will meet an inverse trigonometric function outside a textbook, and the shape of the problem is always identical. Somebody measured two lengths, somebody wants an angle, and the only decision is which of the three ratios you are holding.
The decision is not yours to make on grounds of preference. If a surveyor gives you a slope distance and a height, you have the hypotenuse and the opposite side, so you use the arcsine, even if you would rather use the tangent. Using the wrong pair means inventing a length that nobody measured.
The good news is in the last paragraph. Applied angles in a right triangle are acute, and the first quadrant sits inside every inverse function's range, so none of the convention trouble from Part 1 can reach you here. Applied problems are genuinely the easy case.
Worked example
A roof has a six over twelve pitch: seen from the side, it rises six feet over a run of twelve feet. Find its angle of inclination in degrees, to the nearest hundredth.
\[ \text{rise } = 6\,\text{ft}, \qquad \text{run } = 12\,\text{ft} \]
Identify the two sides
Why: The rise is opposite the angle of inclination and the run is adjacent to it.
\[ \tan(\theta) = \frac{\text{opposite}}{\text{adjacent}} = \frac{6}{12} \]
Simplify the ratio
Why: Six twelfths is one half.
\[ \tan(\theta) = \frac12 \]
Apply the arctangent
Why: The angle is acute, so the arctangent returns it directly with no quadrant work.
\[ \theta = \arctan\!\left(\frac12\right) \approx 0.4636\,\text{rad} \]
Convert to degrees
Why: Multiply by 180 and divide by pi.
\[ \theta \approx 0.4636 \times \frac{180}{\pi} = 26.5651^\circ \approx 26.57^\circ \]
Figure (svg): A right triangle drawn to scale with a horizontal run of 12 feet along the bottom, a vertical rise of 6 feet on the right, and the sloping roof line joining them. The angle of inclination at the bottom left corner is marked theta and the right angle at the bottom right is marked with a small square.
Check against a landmark pitch
Why: A twelve over twelve pitch has rise equal to run, so its angle is exactly 45 degrees. Halving the ratio gave 26.57 degrees, which is more than half of 45, exactly as it must be, since the tangent curves upward. Note that the book prints 26.56 degrees; the exact value is 26.56505 degrees, which rounds to 26.57 at the nearest hundredth.
\[ \arctan(1) = 45^\circ, \qquad \arctan\!\left(\tfrac12\right) \approx 26.57^\circ > \tfrac{45^\circ}{2} \]
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, pp. 837-838 — Example 10.6.6
A pitch is a ratio that builders quote as rise over run, so a six over twelve pitch rises six feet for every twelve feet across. The first step is only translation: rise is the side opposite the angle, run is the side next to it, and opposite over adjacent is the tangent.
Make sure you convert at the right moment. Work in radians until the answer is found, then multiply by 180 and divide by pi at the very end. Converting early and then doing more trigonometry in mixed units is a reliable way to produce nonsense.
Two things to take from the check. First, the landmark: a twelve over twelve pitch is a right isosceles triangle, so exactly 45 degrees, and halving the ratio does not halve the angle. Second, the book prints 26.56 degrees, but the exact value is 26.56505, which rounds to 26.57 at the nearest hundredth as the question asks. That is a printing slip, not a different method, and it is the kind of thing worth checking yourself whenever a rounded answer looks like it was truncated.
Estimation
A walking path gains one metre of height for every twenty metres travelled horizontally.
Predict first
Estimate the angle of inclination before computing it.
Correct: About 3 degrees
Why: The tangent is one twentieth, or 0.05. For a small angle the tangent is close to the angle measured in radians, so the answer is close to 0.05 radians, and 0.05 radians is about 2.86 degrees. The exact value is 2.8624 degrees. The trap is reading 0.05 as a number of degrees when it is a number of radians.
\[ \arctan(0.05) \approx 0.0500\,\text{rad} \approx 2.8624^\circ \]
Estimate before computing. This is worth practising because the small-angle relationship is genuinely useful and because the wrong answer here is so tempting.
For a small angle, the tangent of the angle is very nearly the angle itself measured in radians. So a gradient of one in twenty, which is a tangent of 0.05, corresponds to an angle of about 0.05 radians. The trap is to see 0.05 and think degrees, which would give the first option and would be wrong by a factor of about 57.
Convert the estimate and you get about 2.86 degrees, and the exact answer is 2.8624 degrees. That is the version of the same idea you should carry into any physics or engineering course: for small angles, the tangent, the sine and the angle in radians are all within a fraction of a percent of one another.
Worked example
A guy wire 1000 feet long runs from the top of a tower to the ground, touching level ground 360 feet from the base. Find the angle it makes with the ground, to one decimal place.
\[ \text{hypotenuse } = 1000\,\text{ft}, \qquad \text{adjacent } = 360\,\text{ft} \]
Identify which two sides are known
Why: The wire is the slanted side, so it is the hypotenuse; the 360 feet lies along the ground next to the angle, so it is the adjacent side.
\[ \cos(\theta) = \frac{\text{adjacent}}{\text{hypotenuse}} = \frac{360}{1000} \]
Simplify and apply the arccosine
Why: Three hundred and sixty thousandths is 0.36.
\[ \theta = \arccos(0.36) \]
Evaluate in radians, then convert
Why: The arccosine of 0.36 is 1.2025 radians.
\[ \theta \approx 1.2025 \times \frac{180}{\pi} \approx 68.9^\circ \]
Figure (svg): A right triangle drawn to scale representing a tower and a guy wire. The vertical side is the tower, about 933 feet tall; the horizontal side along the ground is 360 feet; the slanted side is the 1000 foot wire. The angle between the wire and the ground is marked theta.
Verify with the third side
Why: The tower's height must be the square root of one million minus 129600, which is 932.95 feet. The tangent of the angle should then be 932.95 over 360, and the arctangent of that is 68.9 degrees, the same answer by a different route.
\[ \sqrt{1000^2 - 360^2} \approx 932.95, \qquad \arctan\!\left(\frac{932.95}{360}\right) \approx 68.9^\circ \]
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, p. 846 — Exercise 211
Read the wording carefully before reaching for a function. A wire 1000 feet long that touches the ground 360 feet from the base gives you the slanted side and the side along the ground. That is hypotenuse and adjacent, so cosine, so arccosine.
The single most common error on problems like this is to use the arctangent out of habit, treating the 360 feet and the 1000 feet as the two legs. Look at the picture: the 1000 foot wire is visibly longer than either leg, which is what a hypotenuse always is, and putting it in as a leg would give a triangle that cannot exist.
The verification at the end is worth copying as a technique. Compute the third side with the Pythagorean theorem, then redo the angle with a different ratio. Getting the same 68.9 degrees from a completely different pair of sides is strong evidence that you read the problem correctly, not just that you pressed the right buttons.
Matching
Match the pairs
Why: Rise and run are the two legs, so the tangent. A kite string is the slanted side and the height is opposite the ground angle, so the sine. Ground distance next to the angle with the slanted line of sight is the cosine. A canyon depth and a horizontal distance are again two legs, which is Exercise 212 and gives about 7.7 degrees for a depth of 7117 feet and a distance of ten miles.
Pair all four before revealing. For each one, name the three sides out loud first: which is the hypotenuse, which is opposite the angle, which is next to it. The function follows automatically once those names are attached.
Two of the four use the arctangent, and that is not a trick. It happens whenever both known lengths are legs, which is the commonest situation in surveying, because horizontal distance and height are both easy to measure while a slant distance usually is not.
The canyon item is Exercise 212, and it is worth noticing how large the numbers are and how small the angle is. A canyon 7117 feet deep sounds dramatic, but from ten miles away the line of sight drops only about 7.7 degrees below the horizontal. Ratios of this kind are why angles of depression in real terrain are almost always single digits.
Intuition
Builders talk in pitches, surveyors in gradients, and both are ratios. It is tempting to treat a ratio as if it were an angle in disguise, and for small values you can almost get away with it, because the tangent of a small angle is very close to the angle in radians.
It stops working quickly. Doubling a pitch from six over twelve to twelve over twelve takes the angle from 26.57 degrees only to 45 degrees, not to 53.14. Doubling again, to twenty-four over twelve, gives 63.43 degrees rather than 90.
So ratios cannot be added or averaged as though they were angles. If two slopes have to be combined, convert each to an angle first. That is the point of the second homework transfer prompt, and it is a real source of error in construction and in navigation.
This slide exists because treating a ratio as an angle is a real and expensive mistake, not a pedantic one. For gentle slopes the two behave so similarly that the habit forms before anything goes wrong.
Run the numbers in the second paragraph yourself and watch the pattern break. Six over twelve gives 26.57 degrees, twelve over twelve gives 45, twenty-four over twelve gives 63.43. The ratios double each time; the angles gain 18.43, then 18.43 again, purely by coincidence at these values, and then far less. There is no constant of proportionality to find, because the tangent is not a straight line.
The practical consequence is simple: convert first, combine afterwards. If you need the average of two slopes, or the angle between two surfaces given their pitches, turn each into an angle before you do any arithmetic on them. This is exactly the second transfer prompt in the homework, and it is worth doing in numbers rather than words.
Section
Part 4
Concept
In Sections 10.2 and 10.3 you solved equations whose answers were on the list of common angles. If the value is not on that list, the reference angle simply gets a name instead of a value, and every later step is word for word the same.
| step | sine of theta equals one half | sine of theta equals one third |
|---|---|---|
| reference angle | π/6 | arcsin(1/3) |
| quadrants | I and II | I and II |
| first family | π/6 + 2πk | arcsin(1/3) + 2πk |
| second family | π − π/6 + 2πk | π − arcsin(1/3) + 2πk |
The left column is exact and the right column is exact. The only difference is that one reference angle can be read off a table and the other has to be named by a function, exactly as the root of seven has to be named by a square root.
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, p. 838 — the opening of subsection 10.6.4
Read the table across, row by row, and notice that only the first row differs. Everything below it is the same structure with a different name plugged in. If you can solve the left column, you can already solve the right one.
The claim in the last paragraph is the one students resist. It feels as though pi over six is an answer and the arcsine of one third is an unfinished calculation. It is not. Both are exact real numbers, and one of them has a shorter name only because somebody long ago decided to memorise a table of the angles with tidy sines.
Think about where pi over six came from. You know it because you learned a special triangle, not because there is anything easier about it than the arcsine of one third. The notation is the only difference, and once you accept that, this whole part is a review of Section 10.2 with new labels.
Concept
Why two families and not one? Because a horizontal line at a height strictly between negative one and one cuts the unit circle at exactly two points, and those two points are the two angles in a single revolution with that sine.
After that the circle repeats. Adding a whole revolution to either angle gives another angle with the same sine, and subtracting one does too, forever in both directions.
\[ \sin(\theta + 2\pi k) = \sin(\theta) \quad \text{for every integer } k \]
So a complete answer is two starting angles, each carrying the phrase plus two pi k for integers k. Leaving off the families answers a different, smaller question.
This slide answers the question that the next worked example will otherwise raise: why two families rather than one answer? The reason is geometric and you should be able to see it on the circle without any algebra.
Draw a horizontal line at any height strictly between negative one and one and you will cut the unit circle at exactly two points. Those two points are two directions, so two angles, and there is no way to get from one to the other by whole revolutions. They are genuinely separate and each one needs its own family.
The two boundary heights behave differently, which is worth knowing. At height exactly one the line touches the circle at a single point, so there is one family instead of two, and above one it misses entirely, so there are no solutions. Anything strictly between gives two.
Worked example
Find every angle theta whose sine is one third.
\[ \sin(\theta) = \frac13 \]
Put the value on the unit circle
Why: The terminal side meets the circle where the height is one third, and a horizontal line at that height cuts the circle twice, in the first and second quadrants.
\[ y = \frac13: \quad \text{one crossing in QI, one in QII} \]
Name the acute solution
Why: The arcsine is defined to return the angle between negative pi over two and pi over two with this sine, and for a positive value that is the first-quadrant one.
\[ \alpha = \arcsin\!\left(\frac13\right), \quad 0 < \alpha < \frac{\pi}{2} \]
Write the first-quadrant family
Why: Every angle coterminal with alpha has the same terminal side, hence the same sine.
\[ \theta = \arcsin\!\left(\frac13\right) + 2\pi k \]
Write the second-quadrant family
Why: Alpha is the reference angle for the second-quadrant crossing, which is pi minus alpha.
\[ \theta = \pi - \arcsin\!\left(\frac13\right) + 2\pi k \]
Figure (svg): A unit circle cut by a horizontal line at height one third. The line meets the circle at two points, one in the first quadrant and one in the second. A ray runs from the origin to each point; the first-quadrant angle is marked alpha and the second-quadrant angle is marked pi minus alpha.
Check both families numerically
Why: The arcsine of one third is 0.3398, whose sine is 0.333333. The second family starts at 3.1416 minus 0.3398, which is 2.8018, and the sine of 2.8018 is also 0.333333. Both genuinely solve the equation, and adding 6.2832 to either changes nothing.
\[ \sin(0.3398) = \sin(2.8018) \approx 0.3333 = \tfrac13 \]
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, pp. 838-839 — Example 10.6.7, part 1
Do not let the length of this fool you: there are only two ideas here, the acute solution and the second-quadrant reflection, and everything else is bookkeeping. Compare it with how you solved the sine of theta equals one half and you should find the two identical except for step two.
Step two deserves a moment. The arcsine is defined to return the angle between negative pi over two and pi over two with the given sine. For a positive value like one third, that is the first-quadrant angle, so the arcsine hands you the acute solution directly and you do not have to strip any signs.
Look at the figure and locate both dots. They are at the same height, one third, which is exactly what the equation asks for. The right-hand one is at angle alpha and the left-hand one is its mirror image across the vertical axis, which is what pi minus alpha means. Once you see them as a mirror pair you will never write pi plus alpha by accident.
Picture it
Figure (svg): The graph of y equals sine x drawn over three full periods, from negative two pi to four pi, with the horizontal line y equals one third laid across it. Six crossings are marked with dots. An arrow points to the crossing nearest zero and is labelled as the only one a calculator reports.
Here is why one number cannot be the answer. The horizontal line at one third crosses the sine curve over and over, and the picture continues in both directions without end. Six crossings are drawn; there are infinitely many.
Look at the spacing. Within each period there is a near pair, the first-quadrant crossing and its second-quadrant partner, and then a gap before the pattern repeats one whole revolution later. That is exactly the shape of the two families: two starting angles, each stepping by two pi.
Spend a moment on this picture, because it is the answer to the question your calculator cannot answer. Every place the horizontal line meets the wave is a solution, and the picture continues off both edges forever.
Count the crossings inside one period, say from zero to about 6.28. There are two: an early one near 0.34 and a later one near 2.80. Then nothing until the pattern begins again one full revolution later. Two starting angles, each repeating every two pi, is exactly what the two families say.
Follow the arrow to the crossing nearest zero. That is the one, and the only one, that pressing the arcsine key gives you. It is a correct solution; it is roughly one infinity-th of the answer. The habit to build is to draw this picture, or the circle version, before you write anything down.
Trap
Asked to solve the sine of theta equals one third, a student presses the arcsine key and stops.
\[ \theta = 0.3398 \]
Wrong, twice over. It is one solution out of infinitely many, and it is a decimal where an exact answer was asked for.
A calculator is built to return one number, so it must pick one angle out of each family. Picking is its job; completing the answer is yours.
Draw the unit circle, or the graph, and mark every solution. A sine or cosine equation gives two crossings per revolution, so two families with step two pi. A tangent equation gives two crossings that happen to be half a revolution apart, so one family with step pi.
\[ \theta = \arcsin\!\left(\tfrac13\right) + 2\pi k \]
\[ \theta = \pi - \arcsin\!\left(\tfrac13\right) + 2\pi k \]
Then give a decimal only if one was asked for, and say which interval it belongs to.
This is the recurring trap of the whole lesson, and it is worth naming it clearly: a calculator is a function evaluator, and a function returns one value. It is doing its job correctly and it is still not answering the question you asked.
Two separate things went wrong on the left, and it is worth keeping them apart. The completeness problem is that infinitely many solutions are missing. The exactness problem is that the one solution given has been rounded when an exact name was available. A marker will take marks for each, and they have different fixes.
The fix for completeness is always the same and always cheap: draw the unit circle, mark every place the equation is satisfied, and read off one solution per quadrant plus a family for each. Sine and cosine give two families with step two pi; tangent and cotangent give one with step pi. If you draw the picture, you will not forget.
Notation
Annotate
On: \( \theta = \pi - \arcsin\!\left(\tfrac13\right) + 2\pi k, \quad k \in \mathbb{Z} \)
Every complete answer you write this lesson has these four parts, so it is worth being able to point at each of them. Work through the annotations and then try covering the slide and reconstructing the expression from its description.
The part that quietly matters most is the last one. Writing that k is an integer means every integer, including the negative ones, and the negative ones are not optional decoration: they are the solutions to the left of zero on the graph you just looked at. Restricting to non-negative k halves the answer.
The pi minus in front is the quadrant instruction, and it changes with the function. For a cosine equation it would be a plain minus sign, giving the negative of the reference angle. For a tangent equation there is no second expression at all, because the two quadrants collapse into one family. Read the front of the expression as an instruction about where, and the back as an instruction about how often.
Worked example
Find every real number t whose tangent is negative two. Visualise them as angles.
\[ \tan(t) = -2 \]
Find the two quadrants
Why: Tangent is negative where sine and cosine have opposite signs, which is the second and fourth quadrants.
\[ \tan(t) < 0 \;\Longrightarrow\; \text{QII or QIV} \]
Let the arctangent name the fourth-quadrant one
Why: The arctangent returns an angle between negative pi over two and zero for a negative input, which is a fourth-quadrant direction.
\[ \beta = \arctan(-2), \quad -\frac{\pi}{2} < \beta < 0 \]
Locate the second-quadrant solution
Why: The second-quadrant direction is the opposite end of the same line through the origin, a half turn away.
\[ \text{QII solution} = \beta + \pi \]
Figure (svg): A unit circle with the straight line y equals negative two x drawn right through the origin. It meets the circle at one point in the second quadrant and one in the fourth. The fourth-quadrant angle is marked beta, measured clockwise from the positive x-axis, and the second-quadrant point is marked beta plus pi.
Collapse both quadrants into one family
Why: Starting at beta and stepping by pi visits the fourth quadrant, then the second, then the fourth again, and so on.
\[ t = \arctan(-2) + \pi k, \quad k \in \mathbb{Z} \]
Figure (svg): The graph of y equals tangent of t drawn in four branches between its vertical asymptotes, from about negative two pi to two pi, with the horizontal line y equals negative two across it. Each branch is cut by the line exactly once, and the four crossings are marked with dots at equal spacing.
Check two members of the family
Why: The arctangent of negative two is negative 1.1071, whose tangent is negative 2 exactly. Adding 3.1416 gives 2.0344, and the tangent there is also negative 2. One family, both quadrants.
\[ \tan(-1.1071) = \tan(2.0344) = -2 \]
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, pp. 839-840 — Example 10.6.7, part 2
Work through this one and keep asking how it differs from the sine example, because the difference is the whole point. The quadrant reasoning is the same, the reference-angle idea is the same, and then the two families collapse into one.
Look at the circle figure first. The dashed line is the set of all points whose height is negative two times their first coordinate, and it passes through the origin, so both of its ends are legitimate directions with tangent negative two. Those two directions are exactly opposite, a half turn apart, which no sine or cosine equation ever gives you.
Now look at the graph figure. Every branch of the tangent is crossed exactly once by the horizontal line, and the branches are spaced by pi. The dots are equally spaced, and that equal spacing is the family. It is a good picture to keep, because it makes the single family look inevitable rather than like a special rule to memorise.
Intuition
A sine value fixes a height, and a horizontal line at that height cuts the circle at two points that are generally not opposite each other. Two unrelated directions, so two families.
A tangent value fixes a slope, and a line of that slope through the origin cuts the circle at two points that are always exactly opposite. Opposite directions differ by a half turn, so the second family is the first shifted by pi, and stepping by pi from the start sweeps up both.
Said another way: sine and cosine have period two pi, so their solution families step by two pi. Tangent and cotangent have period pi, so theirs step by pi. The period of the function is the step of the family, always.
This is the explanation behind the rule, and it is short enough to reconstruct in an exam if you forget which function gets which step.
A sine or cosine value picks out a horizontal or vertical line, and such a line cuts the circle at two points that are mirror images, not opposites. Two unrelated directions, so two families. A tangent or cotangent value picks out a line through the origin, and a line through the origin always hits the circle at two opposite points. Opposite means half a turn, so the second solution is the first plus pi, and stepping by pi covers both.
The last sentence is the version to memorise, because it is general: the period of the function is the step of the family. Sine and cosine have period two pi and their families step by two pi. Tangent and cotangent have period pi and their families step by pi. Secant and cosecant have period two pi, which is another reason reciprocating them to cosine and sine loses nothing.
Error analysis
A student solves the sine of theta equals one third and hands in this.
Annotate
On: \( \theta = 0.3398 + 2\pi k \;\text{ or }\; \theta = 2.8018 + 2\pi k \)
Write the exact form first, every time. A decimal can always be produced from it afterwards; exactness cannot be recovered from a decimal.
This answer is much better than the one in the earlier trap, which is what makes it worth examining. The structure is completely right: two families, correct quadrants, correct step. Someone who wrote this understands the method.
What they lost is exactness, and for no gain. The arcsine of one third is a perfectly good name for a number, and writing 0.3398 instead replaces it with something that is merely close. The second family is worse, because 2.8018 is a rounded pi minus a rounded arcsine, so two roundings are stacked.
The last annotation says when a decimal is right, and it matters because your homework will ask for them. Exercises 188 to 207 say solve, then approximate the solutions in one revolution. That is two deliverables. Give the exact families first, then the decimals, clearly labelled as approximations, and you have answered both.
Prediction
You are asked to solve the tangent of t equals seven, for every real t.
Predict first
How many families of solutions will the complete answer have?
Correct: One
Why: A tangent equation always yields one family. The two quadrants that hold solutions are opposite one another, so they are half a revolution apart, and a single family stepping by pi reaches both. The answer is the arctangent of seven plus pi times k, for integers k.
\[ t = \arctan(7) + \pi k, \quad k \in \mathbb{Z} \]
Commit before revealing. The question is not about the number seven at all, which is a hint in itself.
Any tangent equation, whatever the value on the right, has solutions in two opposite quadrants, and opposite means half a turn apart. So a single family stepping by pi always suffices, and the answer here is the arctangent of seven plus pi times k.
If you chose two, you have imported the sine rule. Check yourself with the graph from a few slides back: a horizontal line crosses each branch of the tangent once, and the branches are a distance pi apart, so the crossings are evenly spaced. Evenly spaced crossings are one family by definition.
Worked example
Solve, then approximate the solutions that lie in the interval from zero up to but not including two pi, to four decimal places.
\[ \sin(x) = \frac{7}{11} \]
Name the reference angle
Why: Seven elevenths is positive and less than one, so the arcsine returns a first-quadrant angle.
\[ \alpha = \arcsin\!\left(\frac{7}{11}\right) \]
Write the exact answer
Why: Positive sine means first and second quadrants.
\[ x = \arcsin\!\left(\frac{7}{11}\right) + 2\pi k \;\text{ or }\; x = \pi - \arcsin\!\left(\frac{7}{11}\right) + 2\pi k \]
Approximate the reference angle
Why: Seven elevenths is 0.636364.
\[ \arcsin(0.636364) \approx 0.6898 \]
Pick out the members inside one revolution
Why: Take k equal to zero in each family; both results are already between zero and two pi.
\[ x \approx 0.6898 \quad \text{or} \quad x \approx 3.1416 - 0.6898 = 2.4518 \]
Check both decimals
Why: The sine of 0.6898 is 0.636362 and the sine of 2.4518 is 0.636366, both agreeing with seven elevenths, 0.636364, to five decimal places; the small discrepancy is the rounding in the angles, not an error in the method.
\[ \sin(0.6898) \approx 0.63636 \approx \sin(2.4518) \]
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, p. 845 — Exercise 188
This is the exact form your homework will take, so treat it as a template. The question has two parts: solve, which wants exact families, and approximate, which wants decimals inside one revolution. Answer both, in that order, and label them.
The step that people rush is picking out the members inside the interval. Take k equal to zero in each family and check whether the result is already between zero and two pi. Here both are, so there is nothing more to do. In other problems one of them will be negative and you will need to add a revolution, which is what happens in Exercise 195.
Look at the check. The two sines come out as 0.636362 and 0.636366 rather than exactly 0.636364, and that small spread is the rounding in the four-decimal angles working its way back through. Knowing that the discrepancy should be in the sixth decimal place, and not the third, is how you tell rounding from a genuine mistake.
Section
Part 5
Concept
An equation in secant or cosecant is a trap waiting to happen. Solving it with an arcsecant forces you to declare a convention, and a reader who uses the other one will read your answer as a different angle. The book's advice is to get rid of the secant in the first line.
\[ \sec(x) = -\frac53 \;\Longleftrightarrow\; \cos(x) = -\frac35 \]
Arccosine and arcsine have universally agreed ranges, so an answer written with them means the same thing to every reader. The conversion costs one line and removes the problem entirely.
A cotangent equation has no convention problem, since the arccotangent's range is agreed as well, but converting to a tangent equation is still worth doing because it lets you use the single-family shortcut.
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, p. 840 — Example 10.6.7, part 3, and the advice from Section 10.3
This is the book's own advice and it is one line of work for a permanent saving. An equation in secant or cosecant is the only place in this section where the convention argument can reach you, and reciprocating removes it before it starts.
Be clear about what reciprocating does and does not cost. It is an equivalence, not an approximation: the secant of x equals negative five thirds exactly when the cosine of x equals negative three fifths, because neither side is ever zero where the other is defined. Nothing is lost and no extra solutions appear.
The cotangent case is a slightly different argument, so do not merge them. There is no convention problem with arccotangent, since everybody agrees its range is from zero to pi. You convert a cotangent equation to a tangent one for a different reason: to use the one-family shortcut and save yourself writing two.
Intuition
Cosine is one to one on a single unbroken interval from zero to pi, and sine on a single unbroken interval from negative pi over two to pi over two. Each is the obvious choice, each covers the whole output range once, and everybody picked the same one.
Secant and cosecant cannot manage that, because each has a vertical asymptote sitting in the middle of any interval long enough to cover all its values. Whatever you restrict them to has a hole in it, and there are two sensible ways to patch around the hole. That is the entire source of the disagreement.
Since every secant equation is a cosine equation in disguise, and every cosecant equation is a sine equation in disguise, you never have to enter that argument.
It is natural to wonder why mathematics tolerates two conventions for one pair of functions and none for the others, and the answer is not politics.
Cosine takes every value from negative one to one exactly once on the interval from zero to pi, and it is continuous there, so that interval is the obvious restriction and there is no competing candidate. Sine does the same on the interval from negative pi over two to pi over two. Everybody made the same choice because there was really only one to make.
Secant and cosecant cannot be restricted that cleanly. Any stretch long enough to produce all their values has one of their vertical asymptotes in it, so the restricted domain comes in two pieces and you must decide where to put the second piece. Two reasonable answers, two conventions, one avoidable argument.
Worked example
Solve for every real x, and give an answer that does not depend on any convention.
\[ \sec(x) = -\frac53 \]
Reciprocate both sides
Why: The secant is one over the cosine, so flipping both sides turns this into a cosine equation.
\[ \cos(x) = -\frac35 \]
Find the two quadrants
Why: A negative cosine means a negative first coordinate, which is the second and third quadrants.
\[ \cos(x) < 0 \;\Longrightarrow\; \text{QII or QIII} \]
Let the arccosine name the second-quadrant solution
Why: The arccosine of a negative value already returns an angle between pi over two and pi.
\[ \beta = \arccos\!\left(-\frac35\right), \quad \frac{\pi}{2} < \beta < \pi \]
Get the third-quadrant solution by reflecting
Why: The vertical line at negative three fifths cuts the circle at two heights that are exact opposites, so the two angles are negatives of each other.
\[ \text{QIII solution} = -\beta \]
Figure (svg): A unit circle cut by a vertical line at x equals negative three fifths. The line meets the circle at one point in the second quadrant and one in the third, at heights four fifths and negative four fifths. Rays run to both; the upper angle is marked beta and the lower one negative beta.
Write both families
Why: Each starting angle carries its own coterminal family.
\[ x = \arccos\!\left(-\tfrac35\right) + 2\pi k \;\text{ or }\; x = -\arccos\!\left(-\tfrac35\right) + 2\pi k \]
Check one member of each family
Why: The arccosine of negative three fifths is 2.2143, whose cosine is negative 0.6 and whose secant is therefore negative 1.666667, which is negative five thirds. Cosine is even, so negative 2.2143 gives the same value; adding two pi puts that member at 4.0689, inside one revolution.
\[ \sec(2.2143) = \sec(-2.2143) = -\tfrac53, \qquad -2.2143 + 2\pi \approx 4.0689 \]
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, p. 840 — Example 10.6.7, part 3
Compare the first step with what you would have had to do without it. An arcsecant answer would need a stated convention, a reader who uses the other one would misread it, and the second family would be easy to lose. One reciprocation removes all three problems.
The quadrant reasoning after that is ordinary. A negative cosine means the first coordinate of the point on the circle is negative, which is the left half, which is the second and third quadrants. The arccosine of a negative value already lands in the second quadrant, so it names one solution outright and you do not need a reference angle at all.
The third-quadrant solution is where a cosine equation differs from a sine equation, so look at the figure. The vertical line cuts the circle at two points with equal and opposite heights, so the two angles are negatives of each other. That is why the family is written with a plus or minus, and why a cosine answer looks tidier than a sine one.
Worked example
Same move, positive value, and this time finish with the two decimals inside one revolution.
\[ \sec(x) = \frac32 \]
Reciprocate
Why: Flip both sides to get a cosine equation.
\[ \cos(x) = \frac23 \]
Find the two quadrants
Why: A positive cosine means a positive first coordinate, so the first and fourth quadrants.
\[ \text{QI or QIV} \]
Write the exact answer
Why: The arccosine names the first-quadrant angle, and its negative is the fourth-quadrant one.
\[ x = \arccos\!\left(\tfrac23\right) + 2\pi k \;\text{ or }\; x = -\arccos\!\left(\tfrac23\right) + 2\pi k \]
Approximate inside one revolution
Why: The arccosine of two thirds is 0.8411; the negative member needs a revolution added to enter the interval.
\[ x \approx 0.8411 \quad \text{or} \quad x \approx 6.2832 - 0.8411 = 5.4421 \]
Verify both by taking the secant
Why: The cosine of 0.8411 is 0.666667 and the cosine of 5.4421 is 0.666667, so both secants are 1.5, which is three halves. No convention was ever declared, and no arcsecant appears in the answer.
\[ \sec(0.8411) = \sec(5.4421) \approx 1.5 \]
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, p. 845 — Exercise 196
Same method, positive value, and this time the question also wants decimals. Notice that the reciprocation is just as worthwhile even though the value is positive: the point is not the sign, it is that no arcsecant appears anywhere in your final answer.
The quadrants change because the sign changed. A positive cosine means a positive first coordinate, so the right half of the circle: the first and fourth quadrants. The structure of the answer, a plus or minus in front of the arccosine, is unchanged.
The interesting step is the last approximation. The second family's k equal to zero member is negative 0.8411, which is outside the interval from zero to two pi, so you add one revolution to get 5.4421. That habit, taking the family and sliding a member into the window that was asked for, is what Exercises 188 to 207 are drilling.
Fill the middle
Reciprocate to a tangent equation, then use the single-family shortcut. Fill in the two blanks.
Fill in the blanks
\cot(x) = -12 \;\Longrightarrow\; \tan(x) = -1/12 \;\Longrightarrow\; x = \arctan\!\left(-\tfrac3.0585___\right) + \pi k \approx ___ \text___ 6.2000
Why: The reciprocal of negative twelve is negative one twelfth. The arctangent of that is negative 0.0831, which is outside the interval from zero to two pi, so add pi to get 3.0585 and add two pi to get 6.2000. Both are in one revolution, and the cotangent of each is negative twelve.
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, p. 845 — Exercise 195
Fill both blanks before revealing. The first is a reciprocal, and the only way to get it wrong is to forget to flip, which would give negative twelve and an angle nowhere near the right size.
The second blank is the interesting one. The arctangent of negative one twelfth is about negative 0.0831, a tiny negative angle just below the positive horizontal axis, and it is outside the interval you were asked for. Add pi to get 3.0585, and add two pi to get 6.2000. Both are inside one revolution and there are exactly two, as there always are for a tangent or cotangent equation over a full revolution.
Check the plausibility of the answer rather than just the arithmetic. A cotangent of negative twelve is a very large value in size, which means the angle is very close to a multiple of pi, and 3.0585 is indeed just short of 3.1416. If your answer had come out near 1.5 you would know something was wrong.
Sorting
Sort into buckets
Sort each equation by whether reciprocating it before solving is worth the line it costs.
Sort all six before revealing. The test is quick: does the equation use a function whose inverse everybody agrees about, and does it use one whose solutions collapse into a single family?
The secant and cosecant items are the clear cases; reciprocate them and the convention question never arises. The cotangent item is the one people disagree about, and either answer can be defended, but converting is still the better habit because it buys you the one-family shortcut.
Do not over-apply the rule. The three items already in sine, cosine or tangent form need nothing done to them, and reciprocating them would only add a step and another place to drop a sign. Convert when there is something to gain, which is when a secant, a cosecant or a cotangent is on the page.
Trap
A complete-looking answer to the secant of x equals three:
\[ x = \operatorname{arcsec}(3) + 2\pi k \]
Wrong twice. One family where there should be two, and a function whose value the reader cannot pin down.
A positive secant has solutions in the first and fourth quadrants, so two families are needed. And arcsecant of three means different angles in different books, so even the one family given is ambiguous unless the convention is stated.
\[ \cos(x) = \tfrac13 \]
\[ x = \pm\arccos\!\left(\tfrac13\right) + 2\pi k \]
The converted answer is complete and needs no footnote.
There are two independent faults here and it is worth separating them, because a student who fixes only one still loses marks.
The first is incompleteness. A positive secant, like a positive cosine, has solutions in two quadrants, the first and the fourth, so two families are needed. Writing one family and stopping is the same error as the calculator trap earlier, dressed differently.
The second is ambiguity. Even the one family that was given does not pin down an angle, because the arcsecant of three means one thing under the Theorem 10.28 ranges and something else under the Theorem 10.29 ranges. The converted answer on the right has neither problem, and it took one extra line to write.
Section
Part 6
Pattern
Figure (svg): A flow diagram. A secant, cosecant or cotangent equation is first reciprocated into a sine, cosine or tangent one. The sign of the value then picks two quadrants. An inverse function names the reference angle, one solution is written per quadrant, and a coterminal family with step two pi, or step pi for tangent, is attached to each.
OpenStax Algebra and Trigonometry 2e, §9.5 Solving Trigonometric Equations §9.5 — the same procedure in OpenStax
This is the procedure for every equation in this lesson and for the first stage of every equation in Section 10.7. Read it as a chain where each link decides what the next one does.
Follow the flow diagram and notice how early the decisions happen. The reciprocation decision comes first because it determines which inverse function the rest of the work will use. The sign of the value comes next because it determines the quadrants, and only then does anything get evaluated.
The last step is a discipline, not a calculation. Approximating early feels efficient and then quietly corrupts everything downstream, and it also answers a smaller question than the one asked. Finish the exact answer, then produce decimals if they were wanted, and say which interval they came from.
Ranking
Put in order
Order the steps of a complete solution to a secant equation.
Why: Reciprocating comes first because it decides which inverse function the rest of the work will use. The quadrants come before the reference angle, since they are what the reference angle will be placed into. Approximating comes last, after the exact answer is complete, so that no rounding is carried through the algebra.
Order all six before revealing, and then defend each adjacency to yourself: why must this step come before that one?
Two of the orderings really matter. Reciprocating has to be first, because until it is done you do not know whether you are solving a cosine equation or a secant one, and that changes which inverse function everything else uses. Approximating has to be last, because rounding partway through carries error into every later line.
The middle four are more flexible in practice than the list suggests, and experienced solvers do the quadrants and the reference angle almost simultaneously. But writing the quadrants down first is a good discipline while you are learning, because the quadrants are what tell you whether to use pi minus, pi plus, or a negative sign.
Check
Check your understanding
A calculator gives the arctangent of one quarter as about 0.2450. What is the arccotangent of negative four?
Answer: B
Why: The cotangent is negative, so the angle is in the second quadrant. Its reference angle has cotangent four, hence tangent one quarter, so the reference angle is about 0.2450. A second-quadrant angle is pi minus its reference angle, giving about 2.8966, which lies between zero and pi as the arccotangent's range demands.
Answer before revealing. The calculator value has been handed to you deliberately, so the only work is deciding what to do with it.
Two things have to happen. The four has to be flipped to one quarter, because the calculator speaks tangent and the question speaks cotangent. Then the resulting acute angle has to be placed, and a negative cotangent with a range from zero to pi means the second quadrant, so pi minus it.
Each wrong option is a real mistake with a name. Option A forgot the placement, option C forgot to flip, and option D forgot the sign of the input entirely. If you picked one of them, find which of the two operations you skipped, because that is the one to watch next time.
Check
Check your understanding
A guy wire 30 metres long is anchored 12 metres from the base of a mast. What angle does the wire make with the ground?
Answer: B
Why: The 12 metres lies along the ground next to the angle, so it is the adjacent side, and the 30 metre wire is the slanted side, so it is the hypotenuse. Adjacent over hypotenuse is the cosine, so the angle is the arccosine of 0.4, about 66.42 degrees.
Decide before revealing, and name the three sides first: the wire is the hypotenuse, the 12 metres along the ground is adjacent to the angle, and the mast is the opposite side, which nobody measured.
Adjacent over hypotenuse is the cosine, so the arccosine of 0.4, about 66.42 degrees. A quick plausibility check: the anchor is much closer to the mast than the wire is long, so the wire must be steep, and a steep wire makes a large angle with the ground. Anything under 45 degrees should have looked wrong immediately.
Option D is the dangerous one, and it is worth understanding why. It is wrong in two ways that partly cancel, so it lands only two degrees from the truth. An answer that is nearly right for the wrong reason is worse than one that is obviously wrong, because nothing prompts you to check it.
Check
Check your understanding
Which describes every real solution of the tangent of t equals negative four?
Answer: B
Why: A negative tangent has solutions in the second and fourth quadrants, and those two directions are opposite ends of one line, so they are exactly pi apart. Starting at the arctangent and stepping by pi therefore visits every solution, which is why a tangent equation needs only one family.
Choose before revealing. This question is about the shape of the answer, not the arithmetic, and the four options are the four shapes students actually write.
A tangent equation gives one family with step pi, because its two solution quadrants are opposite ends of one line. Option A has the right start and the wrong step, so it finds half the solutions. Option C imports the sine equation's shape, and pi minus an angle preserves the sine, not the tangent. Option D is the calculator trap in its purest form.
If you can look at an equation and predict the shape of its answer before doing any work, you will catch most of your own errors for free. Sine or cosine: two families, step two pi. Tangent or cotangent: one family, step pi. Secant or cosecant: reciprocate first, then whichever of those applies.
Explain it to yourself
Discussion prompt
The book compares the arcsine of one third with the square root of seven. Explain in three or four sentences how close the analogy is, and name one respect in which the two are genuinely different.
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, p. 838 — the square-root analogy
Write an answer before revealing. The analogy is genuinely good, which is why the book leads with it, and finding its limit is a better test of understanding than repeating it.
Where it holds: both the square root and the arcsine are exact names for numbers with no tidy decimal form, both leave the solving method untouched, and turning either into a decimal is a loss rather than a simplification. If you accept the root of seven, you have to accept the arcsine of one third, and most people who resist the second have simply not noticed they already accepted the first.
Where it breaks: a quadratic has two roots and once you write plus or minus the root of seven you are finished. A trigonometric equation has infinitely many solutions, because the circle repeats, so naming the reference angle is only the first of three jobs. The other two are choosing the quadrants and attaching the families, and no amount of good notation does those for you.
Real world
Discussion prompt
A navigation program is given how far east and how far north a destination lies, and computes a bearing with a single-argument arctangent of north over east. Explain which destinations it will report wrongly, by how much, and what the program should do instead.
Write something before revealing, even if it is only a guess about which destinations go wrong. This is the same half-turn problem as Example 10.6.5 part 1c, in a setting where it costs money.
The heart of it: two destinations in exactly opposite directions have north and east components that are both negated, so their ratio is identical. A function that sees only the ratio cannot possibly tell them apart, and the single-argument arctangent sees only the ratio. Every westward destination comes back as its exact opposite, wrong by 180 degrees.
The fix is the one you have been doing by hand all lesson: look at the sign of the east component and add pi when it is negative. Most programming languages provide a two-argument arctangent that does exactly this, taking the two components separately so it can inspect their signs. When you meet that function later, you will know precisely what problem it exists to solve.
Exit ticket
\[ \csc(x) = 5 \]
Discussion prompt
Solve this for every real x without ever writing an arccosecant, then give the two solutions inside one revolution to four decimal places. Say in one line why the answer needs no note about conventions.
Do this one on paper with the slide covered. It uses every idea in the lesson and it should take about three minutes.
The first move is the reciprocation, and it is the move the ticket is really testing. A cosecant equation solved with an arccosecant would need a convention declared; solved as a sine equation it needs nothing. After that it is the standard sine routine: positive value, first and second quadrants, two families with step two pi.
Finish the way the whole lesson has asked you to. Give the exact families first, then the decimals, and check at least one of them by taking the sine back. If your two decimals do not both give 0.2000, find the arithmetic slip before you close the deck.
Recap
| if the question asks | your first move |
|---|---|
| approximate an arccotangent of a negative | reference angle from the positive input, then π minus it |
| approximate an arcsecant or arccosecant of a negative | say which convention, then identity or reference angle |
| the domain of a transformed inverse | put the argument inside the parent's own domain |
| the range of a transformed inverse | track key points, or asymptotes, through the outside only |
| an angle from two measured lengths | name the two sides, then pick the matching inverse |
| solve an equation with an unfriendly value | reciprocate if needed, then quadrants, reference angle, families |
\[ \sin\theta = c: \quad \arcsin c + 2\pi k \;\text{ or }\; \pi - \arcsin c + 2\pi k \]
\[ \tan t = c: \quad \arctan c + \pi k \]
That completes Section 10.6. Every equation here was a single trigonometric function set equal to a constant. Section 10.7 takes on the general case, where the function appears more than once or is buried inside something else, and the first move there is always to reduce the equation to one of the shapes you have just solved.
Stitz & Zeager, College Trigonometry, Ch. 10 Foundations of Trigonometry — §10.6 The Inverse Trigonometric Functions §10.6, pp. 833-841 — subsections 10.6.3 and 10.6.4 in full
OpenStax Algebra and Trigonometry 2e, §8.3 Inverse Trigonometric Functions §8.3 — the same inverse functions in OpenStax
Read the table as a lookup rather than as a summary. Each row is a question you might be handed and the one move that gets you started, and the first move is where most marks are won or lost.
If you keep only one line from this lesson, keep the second formula: a sine or cosine equation gives two families with step two pi, a tangent equation gives one with step pi. Everything else in Part 4 and Part 5 is decoration on that.
Section 10.7 is next, and it starts where this one stops. There the trigonometric function will appear twice, or squared, or with something else inside it, and the first job will always be to manipulate the equation until it has the shape you have just spent this lesson solving. Nothing here is being replaced; it is being reused.
Want this taught 1-on-1? Alexander tutors Trigonometry — $55/session, free consultation.