मंथिकाManthika
Menu
IN REVIEW

Free while we finish the books. A book in review is free to read and use, and so is its PDF. Once the book is finished, its PDF is ₹299.

Why the inverse of sine needs a rule first

In Class 10 you found angles from ratios. A pole is 6 m tall, and its shadow is 6 m long. Find the angle of the sun. The tan of that angle is 6 over 6, which is 1. You knew the angle whose tan is 1, and wrote 45°. There was one answer, and you never doubted it.

Dignaga
A SCHOLAR INDIA REMEMBERS

Dignaga was born in south India, around the 5th or 6th century CE. He trained in the monastic universities of the north. Part of his work was a theory of how words mean anything at all. Ask his question about “the inverse of sine”. What does it pick out? Until you fix a branch, endless angles. Once you fix one, exactly one.

Your calculator does the same thing. Type 0.5 and press the $\sin^{-1}$ key. In degree mode it shows 30. One ratio goes in, and one angle comes out.

The level line at one half meets the sine curve four times here. Two crossings are labelled: pi over six and five pi over six.
Three branches of sine each meet the line at one half once. Any one could carry the inverse; the middle one was chosen by agreement.

But many angles share a sine. Give it an angle and you get one height. Now ask the question backwards. Which angle has this height? Many do. $\sin 30^\circ = 1/2$, and $\sin 150^\circ = 1/2$ as well. Go once more round the circle and $390^\circ$ gives $1/2$ again. The list never ends.

Six angles, two heights. Sine gives 0 for three angles and 1/2 for three, so it is not one-one.
Cosine at 1/2 and tangent at 1 fail the same way: one level line, four crossings, and more past the frame.

Why sine cannot simply be reversed

In Chapter 1 you saw when a function can be run backwards. It must be one-one and onto. Sine is not one-one, and neither are cosine and tangent. Each one repeats its values for ever. In radians, $\sin 0 = \sin \pi = \sin 2\pi = 0$.

Ask which angle has sine 1/2, and three answer. An inverse must give one, so sine cannot be undone yet.

So which one does the calculator show? It shows 30, never 150 or 390. Everyone has agreed on one branch of angles to answer from.

The principal branch

Here is that agreement, written as a definition. $\sin^{-1} x$ is the angle whose sine is $x$, taken from the agreed branch $-\pi/2$ to $\pi/2$. This is the PRINCIPAL BRANCH. On it, each height belongs to one angle, so the inverse has one answer. Each trigonometric function gets its own branch. On it the function must reach every value it can take. It must reach each value only once.

Keep only the angles from −π/2 to π/2. Every height now has one angle, so sine can be undone.

Here is the question for the whole chapter. The sine of an angle is always one number. The angles behind a sine are many. Which one do you read? By the last page you will know the principal branch of all six functions. You will also know which rules still hold on those branches.

0.5
→
sin⁻¹ key
→
π/6
The calculator’s sin⁻¹ key gives one answer for 0.5: π/6, that is 30 degrees, the angle on the principal branch.
Check yourself
  1. Why is “the inverse of $\sin$” ambiguous before any domain is restricted?
    1. because $\sin$ is undefined for negative angles
    2. because $\sin$ is not one-one
    3. because $\sin$ only ever outputs integers
    4. because inverse functions never exist for periodic functions
    Check your answer
    1. because $\sin$ is undefined for negative angles — $\sin$ takes negative angles fine, $\sin(-\pi/2) = -1$ is a perfectly good value. The ambiguity is that many DIFFERENT angles share the same output, not that some angle is missing an output.
    2. ✓ because $\sin$ is not one-one — (B) $\sin$ repeats its values infinitely often, $\sin(0) = \sin(\pi) = \sin(2\pi) = 0$, so no single output points back to a unique input until a branch is fixed.
    3. because $\sin$ only ever outputs integers — $\sin$ is continuous and takes every value in $[-1, 1]$, not just whole numbers, for instance $\sin(\pi/6) = 1/2$. That is not the reason an inverse is hard to define here.
    4. because inverse functions never exist for periodic functions — A periodic function is not permanently barred from having an inverse, restrict it to one branch where it stops repeating, and the inverse exists there. That restriction is this chapter’s whole move.
  2. What does restricting $\sin$ to a principal branch achieve?
    1. it changes the value $\sin$ takes at each point in the branch
    2. it makes $\sin$ one-one and onto on that branch
    3. it removes every negative value from $\sin$’s range
    4. it makes $\sin$ periodic, which it was not before
    Check your answer
    1. it changes the value $\sin$ takes at each point in the branch — $\sin(\pi/6)$ is $1/2$ whether or not a branch has been chosen, restricting the domain only changes which inputs are under consideration, never the value the function returns at any of them.
    2. ✓ it makes $\sin$ one-one and onto on that branch — (B) Restricted to $[-\pi/2, \pi/2]$, $\sin$ becomes one-one (no two inputs share an output) and onto $[-1, 1]$ (every output is hit), exactly what an inverse function needs.
    3. it removes every negative value from $\sin$’s range — The branch $[-\pi/2, \pi/2]$ keeps negative angles, and $\sin$ is still negative on the left half of it, $\sin(-\pi/6) = -1/2$. Restricting the domain narrows inputs, it doesn’t touch sign.
    4. it makes $\sin$ periodic, which it was not before — $\sin$ was periodic before any restriction was applied, $\sin(0) = \sin(2\pi)$ regardless of which branch we later pick. Restricting a domain narrows where we look; it doesn’t add periodicity.
  3. A student says: “$\sin^{-1}$ should exist at every real input, because $\sin$ itself is defined for every real angle.” What is wrong with this reasoning?
    1. nothing: the student is correct and $\sin^{-1}$ does exist for every real number
    2. being defined everywhere doesn’t make a function one-one
    3. $\sin$ is not actually defined for every real angle, only for angles between $0$ and $2\pi$
    4. $\sin^{-1}$’s domain has nothing to do with where $\sin$ is defined
    Check your answer
    1. nothing: the student is correct and $\sin^{-1}$ does exist for every real number — $\sin^{-1}$’s domain is $[-1, 1]$, it takes SINE VALUES as input, not angles. $\sin^{-1}(2)$ has no meaning, since no angle ever has sine $2$.
    2. ✓ being defined everywhere doesn’t make a function one-one — (B) $\sin$ being defined at every angle only means every input produces SOME output. An inverse additionally needs one-oneness, no two inputs sharing an output, which $\sin$ fails badly without a restricted branch.
    3. $\sin$ is not actually defined for every real angle, only for angles between $0$ and $2\pi$ — $\sin$ really is defined for every real angle, $\sin(100)$, $\sin(-7)$, all give a value. That part of the student’s claim is true; the flaw is elsewhere.
    4. $\sin^{-1}$’s domain has nothing to do with where $\sin$ is defined — The student’s whole argument rests on where $\sin$ is defined, that premise is real, it is just the wrong test. Being defined everywhere never implies being one-one anywhere.

↑ Back to top

Before you start

You met every one of these facts once already, in Class 11. Try each check below.

If any of these felt new, read the page named before going on.

↑ Back to top

The principal branches

●KEY-TERM
A 4 m ladder touching the wall 2 m up.

Picture a ladder leaning on a wall. The ladder is 4 metres long. Its top touches the wall 2 metres up. How steep is it? Height over length gives 2 over 4, which is a half. A half is the sine of 30°. So the ladder makes 30° with the ground.

But 150° also has a sine of one half. Could the ladder lean at 150°? Not against this wall. A ladder leans somewhere between lying flat and standing upright. You never even think about the other angles. You read the one that makes sense.

Cut sine from 0 to π and the line at 1/2 crosses twice, with no negative heights. From −π/2 to π/2 it crosses once.

Writing the habit down

Mathematics writes that habit down. Look at the sine curve from $-\pi/2$ to $\pi/2$. Over that branch it climbs once, from $-1$ at the left end to $1$ at the right end. It never turns back. So every height from $-1$ to $1$ is reached, and reached only once. On that branch sine is one-one and onto $[-1, 1]$.

Sine’s branch runs from minus pi over two to pi over two. Both ends are solid dots, so both ends are allowed answers.
domain
range
sin⁻¹ takes inputs from −1 to 1 and answers from −π/2 to π/2. Every arrow lands on the branch.

This interval, $[-\pi/2, \pi/2]$, is the PRINCIPAL BRANCH of sine. The INVERSE SINE undoes sine on this branch. It takes a height and gives the one angle on the branch with that height. In symbols, $\sin^{-1}: [-1, 1] \to [-\pi/2, \pi/2]$. So $\sin^{-1} y$ is the one $x$ in $[-\pi/2, \pi/2]$ with $\sin x = y$. *Read $\sin^{-1} x$ aloud* as “the angle whose sine is $x$”. It is an angle, never a ratio.

*Whatever sits inside $\sin^{-1}$* must land in $[-1, 1]$. Take $\sin^{-1}(2x - 1)$. 1. The condition: $-1 \leq 2x - 1 \leq 1$, since that is what $\sin^{-1}$ needs. 2. Add $1$ to all three parts: $0 \leq 2x \leq 2$. 3. Halve all three parts: $0 \leq x \leq 1$. So the domain of $\sin^{-1}(2x - 1)$ is $[0, 1]$.

π/2
→
sin
→
sin⁻¹
→
π/2
0
→
sin
→
sin⁻¹
→
0
−π/2
→
sin
→
sin⁻¹
→
−π/2
On the branch, sin⁻¹ undoes sine: π/2, 0 and −π/2 go through sine, then sin⁻¹, and come out unchanged.

Let us check two values. $\sin(\pi/2) = 1$, and $\pi/2$ is the right end of the branch. So $\sin^{-1}(1) = \pi/2$. Next, $\sin 0 = 0$, and $0$ sits in the middle. So $\sin^{-1}(0) = 0$.

range
-1
Both ends of the branch are solid, so −π/2 is a real answer. 5π/6 has the right sine but lies off the branch.

Why both ends are allowed

Both ends belong to the branch, and the square brackets say so. An answer of exactly $\pi/2$ or $-\pi/2$ is allowed. Every other angle with the same sine lies outside the branch. It is simply not a candidate. The ladder’s 30° is $\pi/6$, and it is on the branch. So $\sin^{-1}(1/2) = \pi/6$.

The branch is what makes the inverse exist at all. With it, your calculator’s inverse-sine key gives one answer, and you know which one.

Your turn: what is $\sin^{-1}(1)$? Check that your angle is on the branch before you write it down.

Check yourself
  1. What is the principal branch from which $\sin^{-1}$ answers?
    1. all real numbers
    2. $(-\pi/2, \pi/2)$, open at both ends
    3. $[-\pi/2, \pi/2]$
    4. $[0, \pi]$
    Check your answer
    1. all real numbers — If the output could be any real number, $\sin^{-1}(1/2)$ could be $\pi/6$, $\pi/6 + 2\pi$, or infinitely many others, not a function at all. Restricting to one branch is the whole point.
    2. $(-\pi/2, \pi/2)$, open at both ends — $\sin^{-1}$’s branch is closed: $[-\pi/2, \pi/2]$. Both endpoints are legal outputs, $\sin^{-1}(1) = \pi/2$ and $\sin^{-1}(-1) = -\pi/2$ are both correct answers, not excluded ones.
    3. ✓ $[-\pi/2, \pi/2]$ — (C) $\sin^{-1}: [-1, 1] \to [-\pi/2, \pi/2]$ is the definition, $[-\pi/2, \pi/2]$ is the range every answer must come from.
    4. $[0, \pi]$ — $[0, \pi]$ is $\cos^{-1}$’s branch. $\sin^{-1}$ uses $[-\pi/2, \pi/2]$, mixing the two up is the single most common slip in this chapter.
  2. What is the domain of $\sin^{-1}$?
    1. $[-1, 1]$
    2. $(0, 1)$
    3. $[-\pi/2, \pi/2]$
    4. all real numbers
    Check your answer
    1. ✓ $[-1, 1]$ — (A) $\sin^{-1}$ accepts exactly the values $\sin$ can output, $[-1, 1]$, since anything outside that range was never a sine value to begin with.
    2. $(0, 1)$ — $\sin^{-1}(-1) = -\pi/2$ and $\sin^{-1}(1) = \pi/2$ are both perfectly valid, the domain is the full closed interval $[-1, 1]$, including both endpoints and every negative value in between.
    3. $[-\pi/2, \pi/2]$ — $[-\pi/2, \pi/2]$ is the set of ANGLES $\sin^{-1}$ can output, its range. The domain is what it accepts as input, which is $[-1, 1]$, sine’s own set of possible values.
    4. all real numbers — $\sin^{-1}(5)$ has no meaning, no angle has sine equal to $5$, since $\sin$ never exceeds $1$. The domain of $\sin^{-1}$ is bounded by what $\sin$ can actually output.
  3. Why is $\sin^{-1}(-1) = -\pi/2$ specifically, and not some other angle whose sine is $-1$?
    1. because negative inputs must always produce negative angle outputs
    2. because $-\pi/2$ is the smallest angle for which $\sin$ is ever negative
    3. because $\sin(-1) = -\pi/2$ by direct computation
    4. $-\pi/2$ is the unique branch angle with sine equal to $-1$
    Check your answer
    1. because negative inputs must always produce negative angle outputs — There is no sign-matching rule as such, the definition just finds the unique branch angle with the right sine. It is coincidence, not a rule, that a negative input here gives a negative output.
    2. because $-\pi/2$ is the smallest angle for which $\sin$ is ever negative — There is no “smallest angle” rule, $\sin$ is negative at infinitely many angles ($-\pi/2$, $-5\pi/6$, and more). The definition picks the one angle inside the fixed branch $[-\pi/2, \pi/2]$, not the smallest one overall.
    3. because $\sin(-1) = -\pi/2$ by direct computation — $\sin^{-1}(-1)$ asks “which branch angle has sine $-1$?”, not “what is $\sin$ of $-1$ radian?”, reading the notation as plain $\sin$ swaps the whole question.
    4. ✓ $-\pi/2$ is the unique branch angle with sine equal to $-1$ — (D) $\sin(-\pi/2) = -1$, and $-\pi/2$ lies inside the closed branch $[-\pi/2, \pi/2]$, the definition guarantees exactly one such angle exists in that branch, and this is it.
  4. Which of these could actually be a value that $\sin^{-1}$ returns?
    1. $-\pi$
    2. $\pi$
    3. $\pi/2$
    4. $2\pi/3$
    Check your answer
    1. $-\pi$ — $-\pi$ sits outside $[-\pi/2, \pi/2]$ just as far as $\pi$ does on the positive side, $\sin^{-1}$ can never return it.
    2. $\pi$ — $\pi$ is nowhere near $[-\pi/2, \pi/2]$, no input to $\sin^{-1}$ can ever produce it, since every output is guaranteed to fall inside the branch.
    3. ✓ $\pi/2$ — (C) $\pi/2$ is inside $[-\pi/2, \pi/2]$, in fact $\sin^{-1}(1) = \pi/2$ exactly.
    4. $2\pi/3$ — $2\pi/3$ is a little past $\pi/2$, which is the branch’s own right edge, being close to the boundary does not put it inside the branch.
●KEY-TERM

Cosine needs its own branch

*$\cos^{-1} x$* is the angle whose cosine is $x$. Many angles share a cosine, just as many share a sine. So cosine needs an agreed branch too. Sine’s branch, $-\pi/2$ to $\pi/2$, will not do.

On sine’s branch, cosine repeats itself. $\cos(-x) = \cos x$, so $-x$ and $x$ always give the same height. For example, $\cos(-\pi/3)$ and $\cos(\pi/3)$ are both $1/2$. Cosine also never goes below $0$ there. So no angle on that branch gives $-1/2$.

Cosine on sine’s branch crosses 1/2 twice and never goes below zero. From 0 to π it falls once through every height.

Cosine falls once from 0 to π

Now take the angles from $0$ to $\pi$. $\cos 0 = 1$, $\cos 90^\circ = 0$ and $\cos 180^\circ = -1$. On $[0, \pi]$ cosine falls the whole way, from $1$ down to $-1$. It reaches every value in $[-1, 1]$, and each one only once.

Cosine’s branch runs from 0 to pi and falls the whole way, from height 1 down to minus 1. Both ends are solid.

So cosine gets its own PRINCIPAL BRANCH, $[0, \pi]$. On sine’s branch, cosine is neither one-one nor onto $[-1, 1]$. On $[0, \pi]$ it is both.

domain
range
cos⁻¹ answers from 0 to π. The arrows cross: a bigger input gives a smaller angle, because cosine falls on its branch.

Defining the inverse cosine

The INVERSE COSINE is $\cos^{-1}: [-1, 1] \to [0, \pi]$. It undoes cosine on that branch: $\cos^{-1} y$ is the one angle $x$ in $[0, \pi]$ with $\cos x = y$. The branch starts at $0$, so no answer is ever negative.

Check the two ends. $\cos 0 = 1$, and $0$ is in $[0, \pi]$. So $\cos^{-1}(1) = 0$. $\cos \pi = -1$, and $\pi$ is the far end. So $\cos^{-1}(-1) = \pi$. Both ends are allowed.

sin⁻¹
cos⁻¹
0
0
cos⁻¹(0) must land on cosine’s branch. π/2 does; −π/2 sits on sine’s branch only. Borrowing sine’s branch is the usual slip.

Each inverse gets a branch chosen for its own function. When you meet $\cos^{-1}$, think “0 to $\pi$” before anything else.

Your turn: what is $\cos^{-1}(1)$? Before you answer, say which branch your angle must come from.

Check yourself
  1. What is the principal branch from which $\cos^{-1}$ answers?
    1. $(0, \pi)$, open at both ends
    2. $[0, \pi]$
    3. $[-\pi/2, \pi/2]$
    4. $[-\pi, \pi]$
    Check your answer
    1. $(0, \pi)$, open at both ends — $\cos^{-1}$’s branch $[0, \pi]$ is closed at both ends, $\cos^{-1}(1) = 0$ and $\cos^{-1}(-1) = \pi$ are both valid outputs, not excluded ones.
    2. ✓ $[0, \pi]$ — (B) $\cos^{-1}: [-1, 1] \to [0, \pi]$ is the definition, $[0, \pi]$ is the range every answer must come from.
    3. $[-\pi/2, \pi/2]$ — $[-\pi/2, \pi/2]$ is $\sin^{-1}$’s branch. Cosine gets its own: $[0, \pi]$, reusing sine’s branch for cosine is exactly the slip to avoid.
    4. $[-\pi, \pi]$ — On $[-\pi, \pi]$, $\cos(\pi/2) = 0$ and $\cos(-\pi/2) = 0$ too, two different inputs, one output, so $\cos$ is not one-one there. The actual branch $[0, \pi]$ avoids exactly this.
  2. What is the domain of $\cos^{-1}$?
    1. $(-1, 1)$
    2. all real numbers
    3. $[0, \pi]$
    4. $[-1, 1]$
    Check your answer
    1. $(-1, 1)$ — $\cos^{-1}(1) = 0$ and $\cos^{-1}(-1) = \pi$ are both perfectly valid, the domain is the full closed interval $[-1, 1]$, including both endpoints.
    2. all real numbers — $\cos^{-1}(4)$ has no meaning, no angle has cosine equal to $4$, since $\cos$ never exceeds $1$. The domain of $\cos^{-1}$ is bounded by what $\cos$ can actually output.
    3. $[0, \pi]$ — $[0, \pi]$ is the set of ANGLES $\cos^{-1}$ can output, its range. The domain is what it accepts as input: $[-1, 1]$, cosine’s own set of possible values.
    4. ✓ $[-1, 1]$ — (D) $\cos^{-1}$ accepts exactly the values $\cos$ can output, $[-1, 1]$.
  3. Why is $\cos^{-1}(0) = \pi/2$, and never $-\pi/2$, even though $\cos(-\pi/2) = 0$ as well?
    1. because $\pi/2$ is closer to zero than $-\pi/2$ is
    2. $\cos^{-1}$’s branch $[0, \pi]$ never goes negative, ruling $-\pi/2$ out
    3. because $\cos(-\pi/2)$ does not actually equal $0$
    4. because $\cos^{-1}$ always returns the positive of two candidate angles
    Check your answer
    1. because $\pi/2$ is closer to zero than $-\pi/2$ is — There is no distance-from-zero comparison happening, $-\pi/2$ is simply not IN the branch $[0, \pi]$ at all, so it was never a candidate to compare against $\pi/2$ in the first place.
    2. ✓ $\cos^{-1}$’s branch $[0, \pi]$ never goes negative, ruling $-\pi/2$ out — (B) $[0, \pi]$ excludes every negative angle by construction, so $-\pi/2$ is ruled out before any comparison is needed
    3. because $\cos(-\pi/2)$ does not actually equal $0$ — $\cos(-\pi/2) = \cos(\pi/2) = 0$ is correct, cosine is even. That is real; the reason $-\pi/2$ is never the answer is that it falls outside $\cos^{-1}$’s branch $[0, \pi]$, not that the value is wrong.
    4. because $\cos^{-1}$ always returns the positive of two candidate angles — It is not that $\cos^{-1}$ picks the positive one of two options, the branch $[0, \pi]$ simply starts at $0$ and never goes negative, so no negative candidate is ever in play to begin with.
  4. Which of these could actually be a value that $\cos^{-1}$ returns?
    1. $-\pi$
    2. $3\pi/4$
    3. $3\pi/2$
    4. $-\pi/4$
    Check your answer
    1. $-\pi$ — $-\pi$ sits outside $[0, \pi]$ just as $-\pi/4$ does, cosine’s branch never contains a negative angle at all.
    2. ✓ $3\pi/4$ — (B) $3\pi/4$ lies inside $[0, \pi]$, in fact $\cos^{-1}(-1/\sqrt{2}) = 3\pi/4$ exactly.
    3. $3\pi/2$ — $3\pi/2$ is far past $\pi$, the branch’s own right boundary, $\cos^{-1}$ can never return it.
    4. $-\pi/4$ — $\cos^{-1}$’s branch is $[0, \pi]$, which starts at $0$ and never dips below it, a negative output like $-\pi/4$ is impossible for any input, unlike for $\sin^{-1}$.
●KEY-TERM

The slope is a tangent

The slope of a ramp is the tangent of its angle. A slope of 1 means $\tan 45^\circ = 1$. As the angle grows from $-\pi/2$ towards $\pi/2$, the tangent climbs from far below zero to far above it. It reaches every real number, and each one only once.

A ramp rises 1 metre for every 1 metre it runs along the ground. Its slope is 1. A wheelchair ramp is much gentler than that. A slide in a park is steeper. Slope is rise over run, and a slope can be any number at all. Flat ground has slope 0. A road going downhill has a negative slope. A very steep climb might have slope 10 or 100.

Yet the angle of a ramp never reaches upright. A wall’s slope cannot be written as a number, because its run is zero.

On its middle branch tangent climbs from far below to far above, so every level line, 1, 3 or −2, crosses it once.
Tangent’s branch lies between the dashed lines at minus pi over two and pi over two. It climbs without end, and both ends are hollow.

The two ends are left out. At $-\pi/2$ and at $\pi/2$ the tangent has no value, just as a wall’s slope has none. So tangent’s PRINCIPAL BRANCH is open at both ends: $(-\pi/2, \pi/2)$. On it, $\tan$ is one-one and onto all of $\mathbb{R}$.

domain
range
tan⁻¹ takes every real number and answers strictly between −π/2 and π/2. Even 10 lands inside, at about 1.47.

The INVERSE TANGENT is $\tan^{-1}: \mathbb{R} \to (-\pi/2, \pi/2)$. It undoes tangent on that branch. $\tan^{-1} y$ is the one angle $x$ in $(-\pi/2, \pi/2)$ with $\tan x = y$.

0
→
tan⁻¹
→
0
1
→
tan⁻¹
→
π/4
Two quick checks: tan⁻¹(0) is 0 and tan⁻¹(1) is π/4, both inside the open branch.

Two quick checks. $\tan 0 = 0$, so $\tan^{-1}(0) = 0$. $\tan(\pi/4) = 1$, and $\pi/4$ is inside the branch. So $\tan^{-1}(1) = \pi/4$, the ramp with slope 1.

Every real number is a legal input to $\tan^{-1}$, which sine and cosine do not have. You never need a domain check on $\tan^{-1}$ itself. Only what you put inside it must be defined. But an answer of exactly $\pi/2$ is always wrong.

Your turn: a ramp has slope $1000$. Is its angle $\pi/2$? Say why, using the branch.

Check yourself
  1. What is the principal branch from which $\tan^{-1}$ answers?
    1. all real numbers
    2. $[0, \pi]$
    3. $[-\pi/2, \pi/2]$, closed at both ends
    4. $(-\pi/2, \pi/2)$, open at both ends
    Check your answer
    1. all real numbers — $\tan$ repeats every $\pi$ radians, without a restricted range of outputs, many angles would share a tangent value again. $(-\pi/2, \pi/2)$ is what keeps $\tan$ one-one.
    2. $[0, \pi]$ — $[0, \pi]$ is $\cos^{-1}$’s branch. $\tan^{-1}$ uses the open interval $(-\pi/2, \pi/2)$ instead.
    3. $[-\pi/2, \pi/2]$, closed at both ends — $\tan(\pi/2)$ doesn’t exist, tangent has a vertical asymptote there. So $\tan^{-1}$’s branch cannot include $\pi/2$ as an endpoint; it must stay open.
    4. ✓ $(-\pi/2, \pi/2)$, open at both ends — (D) $\tan^{-1}: \mathbb{R} \to (-\pi/2, \pi/2)$ is the definition, the branch is open because $\tan$ itself is undefined at $± \pi/2$.
  2. Why does $\tan^{-1}$ accept every real number as input, unlike $\sin^{-1}$ and $\cos^{-1}$?
    1. because $\tan^{-1}$’s branch is wider than $\sin^{-1}$’s or $\cos^{-1}$’s
    2. $\tan$’s outputs already cover every real number, unbounded both ways
    3. because $\tan$’s formula involves division, which allows any real result
    4. because $\tan$ is defined at every angle, with no gaps at all
    Check your answer
    1. because $\tan^{-1}$’s branch is wider than $\sin^{-1}$’s or $\cos^{-1}$’s — The branch $(-\pi/2, \pi/2)$ is the RANGE of $\tan^{-1}$, its width doesn’t determine what inputs are accepted. The domain is decided by what $\tan$ itself can output.
    2. ✓ $\tan$’s outputs already cover every real number, unbounded both ways — (B) As $x$ ranges over $(-\pi/2, \pi/2)$, $\tan x$ takes every real value, growing without bound near both ends, so $\tan^{-1}$’s domain is unrestricted
    3. because $\tan$’s formula involves division, which allows any real result — It is not the division in $\tan$’s formula that matters, it’s that $\tan$, as $x$ ranges over the open branch, genuinely sweeps through every real value, unbounded in both directions.
    4. because $\tan$ is defined at every angle, with no gaps at all — $\tan(\pi/2)$ is undefined, tangent has infinitely many gaps, at every odd multiple of $\pi/2$. That is not why its inverse accepts all reals.
  3. Which of these can $\tan^{-1}$ actually return as an output?
    1. $\pi/4$
    2. $\pi/2$
    3. $-\pi/2$
    4. $\pi$
    Check your answer
    1. ✓ $\pi/4$ — (A) $\pi/4$ is strictly inside $(-\pi/2, \pi/2)$, in fact $\tan^{-1}(1) = \pi/4$ exactly.
    2. $\pi/2$ — $(-\pi/2, \pi/2)$ is open, $\pi/2$ is excluded, not a valid $\tan^{-1}$ output for any input, however large.
    3. $-\pi/2$ — $(-\pi/2, \pi/2)$ excludes $-\pi/2$ just as it excludes $\pi/2$, neither endpoint is ever reached, only approached.
    4. $\pi$ — $\pi$ is well past the branch’s own boundary $\pi/2$, not a borderline case, simply outside the range $\tan^{-1}$ can ever produce.
●KEY-TERM

Three more functions, built as reciprocals

Cotangent, secant and cosecant are reciprocals. $\cot x = 1/(\tan x)$, $\sec x = 1/(\cos x)$ and $\operatorname{cosec} x = 1/(\sin x)$. Cosine and sine never go beyond $1$ in size. So secant and cosecant never come below $1$ in size. No value strictly between $-1$ and $1$ is ever reached. Cotangent is different: $\cot x = \cos x/\sin x$, a ratio that passes through every real number, including values between $-1$ and $1$. So $\cot^{-1}$ takes every real number as its input, unlike $\sec^{-1}$ and $\operatorname{cosec}^{-1}$.

Four friends share one pizza. Each friend gets a quarter of a pizza. Turn the fact around: four quarters make one pizza, so the reciprocal of a quarter is four.

Take the reciprocal of a small number and you get a big one. Take the reciprocal of a number between minus one and one, other than zero, and you land outside that range. So the reciprocal of a number that started small can never end up small.

Allowed inputs run outward from minus 1 and from 1. The shaded gap between is refused. At 2, secant inverse is pi over three.

That gives the inverses their domains. $\sec^{-1}$ and $\operatorname{cosec}^{-1}$ take only inputs with $|x| \geq 1$. Their domain is $\mathbb{R} - (-1, 1)$. *Check $|x| \geq 1$ first,* every time.

domain
range
cot⁻¹ takes every real number and answers strictly between 0 and π: −1 gives 3π/4, 0 gives π/2, 1 gives π/4.
Six ranges in two groups: three start at 0, three are centred on 0. Hollow rings mark points a range leaves out.

Three branches, three ranges

Each one also gets a PRINCIPAL BRANCH where it is one-one. $\sec^{-1}$ has range $[0, \pi] - \{\pi/2\}$, because $\sec$ has no value at $\pi/2$. $\operatorname{cosec}^{-1}$ has range $[-\pi/2, \pi/2] - \{0\}$, because $\operatorname{cosec}$ has no value at $0$. $\cot^{-1}: \mathbb{R} \to (0, \pi)$, and it never touches $0$ or $\pi$.

FunctionRangePoint left out
$\sec^{-1}$$[0, \pi] - \{\pi/2\}$$\sec$ has no value at $\pi/2$
$\operatorname{cosec}^{-1}$$[-\pi/2, \pi/2] - \{0\}$$\operatorname{cosec}$ has no value at $0$
$\cot^{-1}$$(0, \pi)$never touches $0$ or $\pi$
2
→
sec⁻¹
→
π/3
2
→
1/x
→
cos⁻¹
→
π/3
sec⁻¹(2) by the reciprocal: 2 becomes 1/2, take cos⁻¹, get π/3. Both routes give the same angle.

Let us try one. $\cos(\pi/3) = 1/2$, so $\sec(\pi/3) = 2$. And $\pi/3$ is on secant’s branch. So $\sec^{-1}(2) = \pi/3$.

−√3
→
cot⁻¹
→
5π/6
−√3
→
1/x
→
tan⁻¹
→
−π/6
For a negative input the shortcut fails: cot⁻¹(−√3) is 5π/6, but tan⁻¹ of the reciprocal value gives −π/6.

Board papers ask about these three much less than the first three. You can turn them back into the first three by taking the reciprocal of the input. $\sec^{-1} x = \cos^{-1}(1/x)$ for $|x| \geq 1$. In the same way, $\operatorname{cosec}^{-1} x = \sin^{-1}(1/x)$ for $|x| \geq 1$. For $\cot^{-1}$ the reciprocal needs a sign check. $\cot^{-1} x = \tan^{-1}(1/x)$ only for $x > 0$. For $x < 0$ it is $\pi + \tan^{-1}(1/x)$. For example, $\cot^{-1}(-1) = \pi + (-\pi/4) = 3\pi/4$, which is inside $(0, \pi)$.

So you carry three branches in your head, not six. The other three come from a reciprocal.

Your turn: find $\cot^{-1}(-1/\sqrt{3})$. Which of the two rules for $\cot^{-1}$ do you need?

Check yourself
  1. What is the domain of $\sec^{-1}$?
    1. $[-1, 1]$
    2. $\mathbb{R} - (-1, 1)$
    3. all real numbers
    4. $(-1, 1)$
    Check your answer
    1. $[-1, 1]$ — $[-1, 1]$ is $\sin^{-1}$’s and $\cos^{-1}$’s domain. $\sec^{-1}$’s domain is the opposite kind of set , everything OUTSIDE $(-1, 1)$, not the interval itself.
    2. ✓ $\mathbb{R} - (-1, 1)$ — (B) Outputs strictly between $-1$ and $1$ are never hit by $\sec$, so $\sec^{-1}$’s domain excludes that whole open interval
    3. all real numbers — $\sec x = 1/\cos x$ never lands strictly between $-1$ and $1$, so $\sec^{-1}(1/2)$ has no meaning. The domain excludes that whole open interval.
    4. $(-1, 1)$ — $(-1, 1)$ is exactly what $\sec^{-1}$’s domain LEAVES OUT, not what it consists of, the actual domain is every real number except that open interval.
  2. What is the range of $\cot^{-1}$, and why does it not touch $0$ or $\pi$?
    1. $\mathbb{R} - (-1, 1)$: $\sec^{-1}$’s domain shape
    2. $(-\pi/2, \pi/2)$, matching $\tan^{-1}$’s branch
    3. $[0, \pi]$, closed, matching $\cos^{-1}$’s branch exactly
    4. $(0, \pi)$, open: asymptotes at $0$ and $\pi$
    Check your answer
    1. $\mathbb{R} - (-1, 1)$: $\sec^{-1}$’s domain shape — $\mathbb{R} - (-1, 1)$ is the DOMAIN shape used for $\sec^{-1}$ and $\operatorname{cosec}^{-1}$, not a range of angles. $\cot^{-1}$’s range is the single connected interval $(0, \pi)$.
    2. $(-\pi/2, \pi/2)$, matching $\tan^{-1}$’s branch — $(-\pi/2, \pi/2)$ belongs to $\tan^{-1}$. $\cot^{-1}$’s own branch is $(0, \pi)$, a different open interval, not the same one under another name.
    3. $[0, \pi]$, closed, matching $\cos^{-1}$’s branch exactly — $\cos^{-1}$’s branch is the CLOSED interval $[0, \pi]$. $\cot^{-1}$’s branch shares the same numbers but is OPEN: $(0, \pi)$, $\cot$ has a vertical asymptote at both $0$ and $\pi$, so neither endpoint is ever reached.
    4. ✓ $(0, \pi)$, open: asymptotes at $0$ and $\pi$ — (D) $\cot(0)$ and $\cot(\pi)$ are both undefined, vertical asymptotes, so the branch excludes both endpoints, staying open
  3. A student refuses to learn $\sec^{-1}$, $\operatorname{cosec}^{-1}$, $\cot^{-1}$ separately, saying “I’ll just turn them back into $\sin^{-1}$, $\cos^{-1}$, $\tan^{-1}$ every time.” Is this a workable strategy?
    1. no: NCERT requires all six inverse functions to be memorised independently, with no shortcuts
    2. yes: each of the three reduces to a first-three function via its own reciprocal identity
    3. no: the reciprocal identities only work for $\sec^{-1}$, never for $\operatorname{cosec}^{-1}$ or $\cot^{-1}$
    4. no: reciprocal identities change the VALUE of the answer, so this strategy gives wrong results
    Check your answer
    1. no: NCERT requires all six inverse functions to be memorised independently, with no shortcuts — These three genuinely carry less weight, because most problems reduce back to the first three via a reciprocal identity, that is the intended, sanctioned strategy, not a forbidden shortcut.
    2. ✓ yes: each of the three reduces to a first-three function via its own reciprocal identity — (B) Each of the three follows the same pattern, restricted to a branch where it is one-one, and reduces back to the first three via a reciprocal identity, e.g. $\sec^{-1} x = \cos^{-1}(1/x)$ for $|x| \geq 1$
    3. no: the reciprocal identities only work for $\sec^{-1}$, never for $\operatorname{cosec}^{-1}$ or $\cot^{-1}$ — $\operatorname{cosec}^{-1} x = \sin^{-1}(1/x)$ and a similar identity exist for $\cot^{-1}$ too, the reduction strategy is not special to $\sec^{-1}$, it works across all three.
    4. no: reciprocal identities change the VALUE of the answer, so this strategy gives wrong results — $\sec^{-1} x = \cos^{-1}(1/x)$ is a true equality for $|x| \geq 1$, both sides give the identical value. Using it does not introduce any error.
Three points share the height one half. Only pi over six lies inside the shaded branch, so it is the answer.
Worked example

Evaluate $\sin^{-1}(1/2)$

  1. $\sin^{-1}(1/2)$
    You want the angle $x$ in $[-\pi/2, \pi/2]$ with $\sin x = 1/2$.
  2. $\sin(\pi/6) = 1/2$
    A standard value. And $\pi/6$ is inside the branch $[-\pi/2, \pi/2]$.
  3. *$\sin^{-1}(1/2) = \pi/6$*
    It is the one angle on the branch with sine $1/2$. So it is the answer.
  4. $\sin(5\pi/6) = 1/2$ too, but $5\pi/6$ lies outside $[-\pi/2, \pi/2]$
    Another angle has the same sine. It lies outside the branch, so it is not a candidate.
  5. $\pi/6 \approx 0.52$, and $\sin(\pi/6) = 1/2$
    Check both facts. The angle lies between $-1.57$ and $1.57$, and its sine is $1/2$.
Worked example

The same step, on a ray at the critical angle

  1. Light inside glass ($\mu = 3/2$) meets the glass-air boundary, and $\sin i_c = 1/\mu$ (Physics 12, Chapter 9, Ray Optics and Optical Instruments)
    Past the critical angle $i_c$, every ray already inside the glass reflects back in instead of refracting out.
  2. $\sin i_c = 1/(3/2) = 2/3$
    Substitute the glass’s own refractive index into that relation.
  3. $i_c = \sin^{-1}(2/3) \approx 41.8^\circ$
    The critical angle is exactly the inverse sine of that ratio, the same step as $\sin^{-1}(1/2) = \pi/6$ above, applied to a different input.
Check yourself
  1. Evaluate $\sin^{-1}(\sqrt{3}/2)$.
    1. $\pi/6$
    2. $-\pi/3$
    3. $\pi/3$
    4. $2\pi/3$
    Check your answer
    1. $\pi/6$ — $\pi/6$ solves $\sin^{-1}(1/2)$, not $\sin^{-1}(\sqrt{3}/2)$, this answer was carried over from a different input by mistake.
    2. $-\pi/3$ — $\sin(-\pi/3) = -\sqrt{3}/2$, not $+\sqrt{3}/2$, flipping the sign changes the value being evaluated, giving the wrong angle entirely.
    3. ✓ $\pi/3$ — (C) $\sin(\pi/3) = \sqrt{3}/2$, and $\pi/3$ lies inside the principal branch $[-\pi/2, \pi/2]$
    4. $2\pi/3$ — $\sin(2\pi/3) = \sqrt{3}/2$ is correct, but $2\pi/3$ is outside $[-\pi/2, \pi/2]$, the branch angle with that sine is $\pi/3$, not $2\pi/3$.
  2. $\sin(5\pi/6) = 1/2$ as well as $\sin(\pi/6) = 1/2$. Why is $\sin^{-1}(1/2)$ still just $\pi/6$, with no second correct answer?
    1. because $5\pi/6$ and $\pi/6$ are actually the same angle in disguise
    2. because $\sin(5\pi/6)$ does not really equal $1/2$
    3. $5\pi/6$ lies outside the branch $[-\pi/2, \pi/2]$, so only $\pi/6$ qualifies as the branch answer
    4. because $\sin^{-1}$ is allowed to return either answer, and $\pi/6$ is just the conventional choice
    Check your answer
    1. because $5\pi/6$ and $\pi/6$ are actually the same angle in disguise — $\pi/6$ and $5\pi/6$ are different angles that happen to share the same sine, that is exactly why $\sin$ is not one-one over its full domain, the whole reason a branch is needed.
    2. because $\sin(5\pi/6)$ does not really equal $1/2$ — $\sin(5\pi/6)$ really does equal $1/2$, that part is true. The reason it isn’t the answer is that $5\pi/6$ falls outside the branch, not that the sine value is wrong.
    3. ✓ $5\pi/6$ lies outside the branch $[-\pi/2, \pi/2]$, so only $\pi/6$ qualifies as the branch answer — (C) The branch $[-\pi/2, \pi/2]$ contains exactly one of the two angles with sine $1/2$, and it is $\pi/6$
    4. because $\sin^{-1}$ is allowed to return either answer, and $\pi/6$ is just the conventional choice — A function returns exactly one output per input, by definition. $\sin^{-1}$ is not free to return either $\pi/6$ or $5\pi/6$, only the branch angle, $\pi/6$, is the answer.
  3. Evaluate $\sin^{-1}(-1/2)$.
    1. $5\pi/6$
    2. $-\pi/6$
    3. $\pi/6$
    4. $7\pi/6$
    Check your answer
    1. $5\pi/6$ — $5\pi/6$ is both outside the branch AND has the wrong sign of sine ($\sin(5\pi/6) = +1/2$), two separate reasons this cannot be the answer.
    2. ✓ $-\pi/6$ — (B) $\sin(-\pi/6) = -1/2$, and $-\pi/6$ lies inside the branch $[-\pi/2, \pi/2]$
    3. $\pi/6$ — $\sin(\pi/6) = +1/2$, not $-1/2$, dropping the negative sign from the input silently changes which value is being evaluated.
    4. $7\pi/6$ — $7\pi/6$ sits nowhere near $[-\pi/2, \pi/2]$, $\sin^{-1}$’s output is guaranteed to stay inside the branch, so this can be ruled out immediately.
Cosine meets −1/√2 three times in the frame. Only 3π/4 lies inside 0 to π.
−1/√2
→
drop the sign
→
cos⁻¹
→
π minus
→
3π/4
Drop the sign, read π/4, then take it from π: cos⁻¹(−1/√2) is 3π/4.
Worked example

Evaluate $\cos^{-1}(-1/\sqrt{2})$

  1. $\cos^{-1}(-1/\sqrt{2})$
    You want the angle $x$ in $[0, \pi]$ with $\cos x = -1/\sqrt{2}$.
  2. $\cos(\pi/4) = 1/\sqrt{2}$
    A standard value. But it is positive, and you need a negative.
  3. Cosine is negative in the second quadrant, so $\cos(\pi - \pi/4) = -\cos(\pi/4) = -1/\sqrt{2}$
    Going from $\pi/4$ to $\pi - \pi/4$ changes the sign of the cosine.
  4. $\pi - \pi/4 = 3\pi/4$, and $3\pi/4 \in [0, \pi]$
    This angle is inside cosine’s branch $[0, \pi]$.
  5. *$\cos^{-1}(-1/\sqrt{2}) = 3\pi/4$*
    It is the one angle in $[0, \pi]$ with cosine $-1/\sqrt{2}$.
  6. $3\pi/4 \approx 2.36$, and $\cos(2.36) \approx -0.71 \approx -1/\sqrt{2}$
    Check with decimals. The angle lies between $0$ and $3.14$, and its cosine has the right sign and size.
Check yourself
  1. Evaluate $\cos^{-1}(-1/\sqrt{2})$.
    1. $5\pi/4$
    2. $3\pi/4$
    3. $\pi/4$
    4. $-\pi/4$
    Check your answer
    1. $5\pi/4$ — $5\pi/4$ overshoots past $\pi$, the branch’s right boundary, $\cos^{-1}$’s output is guaranteed to stay inside $[0, \pi]$.
    2. ✓ $3\pi/4$ — (B) $\cos(3\pi/4) = -1/\sqrt{2}$, using the second-quadrant reduction $\cos(\pi - \pi/4)$, and $3\pi/4$ lies inside $[0, \pi]$
    3. $\pi/4$ — $\cos(\pi/4) = +1/\sqrt{2}$, not $-1/\sqrt{2}$, dropping the sign answers the wrong question. The branch angle with the NEGATIVE value is $3\pi/4$, further along the branch.
    4. $-\pi/4$ — $\cos^{-1}$’s branch is $[0, \pi]$, which never dips below $0$, $-\pi/4$ is not even a legal output, regardless of the input.
  2. Why does evaluating $\cos^{-1}$ of a negative value use $\pi - \theta$ rather than $-\theta$, unlike changing the sign on sine’s side?
    1. because $\cos(\pi - \theta)$ and $-\cos(\theta)$ happen to be unrelated by coincidence
    2. because subtraction is only mathematically valid for cosine, never for sine
    3. cosine’s branch $[0, \pi]$ has no room for a negative angle, unlike sine’s
    4. because $\pi - \theta$ and $-\theta$ always give the exact same angle
    Check your answer
    1. because $\cos(\pi - \theta)$ and $-\cos(\theta)$ happen to be unrelated by coincidence — $\cos(\pi - \theta) = -\cos(\theta)$ is a genuine, standard identity, it is exactly why $\cos(\pi - \pi/4)$ lands on $-1/\sqrt{2}$. It’s not a coincidence; it’s the mechanism the whole reduction relies on.
    2. because subtraction is only mathematically valid for cosine, never for sine — Subtraction is not a cosine-only tool, $\sin(\pi - \theta) = \sin \theta$ uses it too, in the worked example on reducing $\sin^{-1}(\sin(3\pi/4))$. The real reason for $\cos^{-1}$’s $\pi - \theta$ form is about which branch cosine has, not which operation is allowed where.
    3. ✓ cosine’s branch $[0, \pi]$ has no room for a negative angle, unlike sine’s — (C) Cosine’s branch $[0, \pi]$ never contains a negative angle, so $-\theta$ could never be a legal branch answer to reduce into, $\pi - \theta$ lands back inside the branch instead
    4. because $\pi - \theta$ and $-\theta$ always give the exact same angle — $\pi - \pi/4 = 3\pi/4$, not $-\pi/4$, these are two different angles in general, not interchangeable names for the same reduction step.
  3. Evaluate $\cos^{-1}(-1/2)$.
    1. $\pi/6$
    2. $-\pi/3$
    3. $2\pi/3$
    4. $\pi/3$
    Check your answer
    1. $\pi/6$ — $\pi/6$ solves $\cos^{-1}(\sqrt{3}/2)$, a different input, this answer was carried over from a different evaluation by mistake.
    2. $-\pi/3$ — $\cos^{-1}$’s branch is $[0, \pi]$, which never dips below $0$, $-\pi/3$ is ruled out on sight, regardless of the input.
    3. ✓ $2\pi/3$ — (C) $\cos(2\pi/3) = -1/2$, using $\cos(\pi - \pi/3)$, and $2\pi/3$ lies inside $[0, \pi]$
    4. $\pi/3$ — $\cos(\pi/3) = +1/2$, not $-1/2$, the branch angle with the negative value is further along the branch, at $2\pi/3$.
Tangent meets −1 once on every branch. The principal branch gives −π/4, the answer.
Inverse tangent gives pi over four at 1 and minus pi over four at minus 1. The joining line passes through the origin.
Worked example

Evaluate $\tan^{-1}(-1)$

  1. $\tan^{-1}(-1)$
    You want the angle $x$ in $(-\pi/2, \pi/2)$ with $\tan x = -1$.
  2. $\tan(\pi/4) = 1$
    A standard value. It is positive, so the sign still needs fixing.
  3. $\tan$ is an odd function, so $\tan(-\pi/4) = -\tan(\pi/4) = -1$
    Tangent is odd. Change the sign of the angle and the tangent changes sign too.
  4. $-\pi/4 \in (-\pi/2, \pi/2)$
    The angle is inside tangent’s open branch.
  5. *$\tan^{-1}(-1) = -\pi/4$*
    It is the one angle on the branch with tangent $-1$.
  6. $-\pi/4 \approx -0.785$, and $\tan(-0.785) \approx -1$
    Check with decimals. The angle lies between $-1.57$ and $1.57$, and its tangent is $-1$.
Check yourself
  1. Evaluate $\tan^{-1}(-1)$.
    1. $-3\pi/4$
    2. $-\pi/4$
    3. $3\pi/4$
    4. $\pi/4$
    Check your answer
    1. $-3\pi/4$ — $-3\pi/4$ sits beyond $-\pi/2$, the branch’s left boundary, outside the interval $(-\pi/2, \pi/2)$ entirely.
    2. ✓ $-\pi/4$ — (B) $\tan(-\pi/4) = -1$, using $\tan$’s odd symmetry, and $-\pi/4$ lies inside the open branch $(-\pi/2, \pi/2)$
    3. $3\pi/4$ — $3\pi/4$ is well past $\pi/2$, the branch’s own boundary, $\tan^{-1}$’s output is guaranteed to stay strictly inside $(-\pi/2, \pi/2)$.
    4. $\pi/4$ — $\tan(\pi/4) = +1$, not $-1$, the branch angle with the negative value is $-\pi/4$, using $\tan$’s odd symmetry.
  2. Why does $\tan(-\pi/4) = -1$ follow directly from $\tan(\pi/4) = 1$, with no separate calculation needed?
    1. because $\tan$ happens to repeat the value $1$ at every angle related to $\pi/4$
    2. because dividing by a negative number always produces $-1$ regardless of the original value
    3. $\tan$ is an odd function, so $\tan(-\theta) = -\tan(\theta)$ holds for every angle in its domain
    4. because $-\pi/4$ and $\pi/4$ are the same angle measured in opposite directions, and so give equal tangents
    Check your answer
    1. because $\tan$ happens to repeat the value $1$ at every angle related to $\pi/4$ — Saying $\tan$ “repeats $1$” near $\pi/4$ isn’t a mechanism, and it’s not even accurate, $\tan(-\pi/4) = -1$, the negative of $1$. The real reason is $\tan$’s odd symmetry.
    2. because dividing by a negative number always produces $-1$ regardless of the original value — There is no division by a negative number in this reduction, the value $-1$ comes from $\tan$’s own odd symmetry, $\tan(-\theta) = -\tan(\theta)$, not from any division step.
    3. ✓ $\tan$ is an odd function, so $\tan(-\theta) = -\tan(\theta)$ holds for every angle in its domain — (C) $\tan$ is an odd function, so $\tan(-\theta) = -\tan(\theta)$ holds for every angle
    4. because $-\pi/4$ and $\pi/4$ are the same angle measured in opposite directions, and so give equal tangents — $-\pi/4$ and $\pi/4$ are different angles with different tangent values, $\tan(-\pi/4) = -1$ while $\tan(\pi/4) = 1$. Odd symmetry relates the two values; it doesn’t make the angles the same.
  3. Evaluate $\tan^{-1}(-\sqrt{3})$.
    1. $-2\pi/3$
    2. $2\pi/3$
    3. $\pi/3$
    4. $-\pi/3$
    Check your answer
    1. $-2\pi/3$ — $-2\pi/3$ sits beyond $-\pi/2$, the branch’s left boundary, outside $(-\pi/2, \pi/2)$.
    2. $2\pi/3$ — $2\pi/3$ is past $\pi/2$, the branch’s boundary, outside $(-\pi/2, \pi/2)$ regardless of its tangent value.
    3. $\pi/3$ — $\tan(\pi/3) = +\sqrt{3}$, not $-\sqrt{3}$, the branch angle with the negative value is $-\pi/3$, by $\tan$’s odd symmetry.
    4. ✓ $-\pi/3$ — (D) $\tan(-\pi/3) = -\sqrt{3}$, using $\tan$’s odd symmetry, and $-\pi/3$ lies inside the open branch $(-\pi/2, \pi/2)$

↑ Back to top

Graphs of the inverse functions

▲CONCEPT

Make a table with two columns, angle and sine. One row reads 30° and one half. Another reads 90° and 1. Now swap the two columns. The new table starts from a height and gives an angle. That swapped table is the inverse.

Swapping is a mirror

Plot both tables on graph paper. Every point moves to the place with its two numbers swapped. The point (3, 5) would move to (5, 3). Swapping like this is a mirror. The mirror is the diagonal line through the origin.

Sine’s branch and its inverse mirror each other in the dashed line y = x. Each point’s two coordinates swap places.

*That line is $y = x$. A function and its inverse are always mirror images in it. You saw this in Chapter 1. Here it works* on three curves you already know.

Let us test one point. $(\pi/6, 1/2)$ is on sine’s branch. Swap it and you get $(1/2, \pi/6)$, which is on the graph of $\sin^{-1} x$.

Turn cosine’s falling branch over the line y = x and you get cos⁻¹, falling from π at −1 to 0 at 1.

Three shapes, mirrored

The graph of $\sin^{-1} x$ rises from $(-1, -\pi/2)$ through the origin to $(1, \pi/2)$. It climbs the whole way, like sine’s own branch on $[-\pi/2, \pi/2]$. The graph of $\cos^{-1} x$ falls from $(-1, \pi)$ to $(1, 0)$. It falls the whole way, like cosine’s branch on $[0, \pi]$. A mirror keeps rising curves rising and falling curves falling.

Tangent’s two walls at ±π/2 become two ceilings when the graph is turned over. tan⁻¹ climbs toward them and never arrives.

*$\tan^{-1} x$ is different.* It climbs across all of $\mathbb{R}$, but it never reaches $\pi/2$ or $-\pi/2$. The lines $y = \pi/2$ and $y = -\pi/2$ are horizontal asymptotes. The curve flattens towards them as $x \to \infty$ and $x \to -\infty$, and never touches either.

Inverse sine climbs and inverse cosine falls, both over minus 1 to 1. Inverse tangent runs the full width, never reaching the dashed lines.
sin⁻¹ keeps the order of its inputs, so its arrows run side by side. cos⁻¹ reverses it, so its arrows cross.
Inverse cotangent falls from π toward 0 across every input. Inverse secant and cosecant each have two pieces, none between −1 and 1.

The other three follow their own branches. $\cot^{-1} x$ falls from $\pi$ towards $0$ across all of $\mathbb{R}$. $\sec^{-1} x$ and $\operatorname{cosec}^{-1} x$ each break into two pieces. Nothing is drawn between $-1$ and $1$.

So if a branch slips your memory, sketch the function’s branch and reflect it in $y = x$.

Check yourself
  1. What general fact explains why the graph of $\sin^{-1} x$ looks the way it does?
    1. the graph is found by solving $\sin^{-1} x = 0$ for every possible $x$
    2. a function and its inverse mirror each other in $y = x$
    3. $\sin^{-1}$ has a completely unrelated shape to $\sin$’s own graph
    4. every inverse trig graph is a straight line, since angles are linear quantities
    Check your answer
    1. the graph is found by solving $\sin^{-1} x = 0$ for every possible $x$ — A graph plots the function’s OUTPUT at every input, not a single equation solved once, “$\sin^{-1} x = 0$” only describes one point on the curve ($x=0$), not the whole shape.
    2. ✓ a function and its inverse mirror each other in $y = x$ — (B) This is the general mirror-image fact for $f$ and $f^{-1}$, applied here to three familiar curves
    3. $\sin^{-1}$ has a completely unrelated shape to $\sin$’s own graph — $\sin^{-1}$’s graph is not some independent shape, it is $\sin$’s restricted-branch graph, reflected across the line $y = x$. The two are tightly related, not unrelated.
    4. every inverse trig graph is a straight line, since angles are linear quantities — None of these graphs are straight, $\sin^{-1} x$ rises but curves, and $\tan^{-1} x$ flattens toward horizontal asymptotes as $x$ grows. “Angles are linear” does not make their graphs linear.
  2. Which of the three inverse trig graphs is strictly DECREASING throughout its domain?
    1. $\cos^{-1} x$, from $\pi$ to $0$
    2. $\tan^{-1} x$
    3. $\sin^{-1} x$
    4. none of the three, since each curve eventually flattens out
    Check your answer
    1. ✓ $\cos^{-1} x$, from $\pi$ to $0$ — (A) $\cos^{-1} x$ falls from $(-1, \pi)$ to $(1, 0)$, strictly decreasing throughout, the one exception among the three
    2. $\tan^{-1} x$ — $\tan^{-1} x$ increases across its whole domain, approaching $-\pi/2$ and $\pi/2$ as horizontal asymptotes , it never decreases.
    3. $\sin^{-1} x$ — $\sin^{-1} x$ climbs the whole way, from $(-1, -\pi/2)$ through the origin to $(1, \pi/2)$, it is strictly increasing, not decreasing.
    4. none of the three, since each curve eventually flattens out — $\cos^{-1} x$ falls the whole way, from $(-1, \pi)$ down to $(1, 0)$, it is the one of the three that decreases strictly throughout its domain.
  3. As $x \to \infty$, what happens to the graph of $\tan^{-1} x$?
    1. it starts decreasing back toward $0$
    2. it becomes undefined, since $\tan^{-1}$’s domain is bounded
    3. it flattens toward $y = \pi/2$, never reaching it
    4. it crosses $y = \pi/2$ and keeps climbing past it
    Check your answer
    1. it starts decreasing back toward $0$ — $\tan^{-1} x$ increases the whole way as $x$ grows, it flattens near the asymptote but never reverses direction back toward $0$.
    2. it becomes undefined, since $\tan^{-1}$’s domain is bounded — $\tan^{-1}$ accepts every real $x$ as input, its domain is unbounded. It is the output (range) that is confined to $(-\pi/2, \pi/2)$, never the input.
    3. ✓ it flattens toward $y = \pi/2$, never reaching it — (C) The graph flattens toward the horizontal asymptote $y = \pi/2$ as $x \to \infty$, matching the open upper end of $\tan^{-1}$’s range
    4. it crosses $y = \pi/2$ and keeps climbing past it — The branch $(-\pi/2, \pi/2)$ is open, so $\tan^{-1} x$’s output can get arbitrarily close to $\pi/2$ but is guaranteed never to reach or cross it, that is exactly what an asymptote means.

↑ Back to top

When sine inverse of sine is not x

✕MISCONCEPTION
First try

I saw $\sin^{-1}(\sin(3\pi/4))$ and cancelled it straight to $3\pi/4$, since the inverse undoes the function.

Second look

$3\pi/4$ is not inside the principal branch $[-\pi/2, \pi/2]$, so that cancellation never applied here.

$\sin^{-1}(\sin x) = x$ only when $x$ already sits inside $[-\pi/2, \pi/2]$. Reduce first, then cancel.

In Chapter 1, an inverse undoes its function: $f^{-1}(f(x)) = x$. With sine, the undo works only on the branch. Start from $3\pi/4$, which is off it. Its sine is $1/\sqrt{2}$, and $\sin^{-1}(1/\sqrt{2}) = \pi/4$. You come back with $\pi/4$, not $3\pi/4$.

Inside the shaded branch, the zigzag matches the line y = x. At three pi over four, the zigzag gives pi over four.

Check one that works. Take $x = \pi/6$. Its sine is $1/2$, and $\sin^{-1}(1/2) = \pi/6$. *You get $x$ again,* because $\pi/6$ was on the branch.

π/6
→
sin
→
sin⁻¹
→
π/6
3π/4
→
sin
→
sin⁻¹
→
π/4
π/6 goes through sine, then sin⁻¹, and comes out as π/6. 3π/4 comes out as π/4. Which step lost it?
−π/3
→
cos
→
cos⁻¹
→
π/3
3π/4
→
tan
→
tan⁻¹
→
−π/4
Cosine and tangent do the same: −π/3 comes out as π/3, and 3π/4 as −π/4. Each lands on its own branch.
sin⁻¹
cos⁻¹
tan⁻¹
0
−π/3 lies on sine’s and tangent’s branches, so those inverses undo it. It misses cosine’s branch, so cos⁻¹ does not.

The same check applies to the other two. $\cos^{-1}(\cos x) = x$ only for $x$ in $[0, \pi]$. And $\tan^{-1}(\tan x) = x$ only for $x$ in $(-\pi/2, \pi/2)$. Each one carries its own branch.

*When $x$ is off the branch,* bring it back first. 1. Sine: $\sin^{-1}(\sin x) = \pi - x$ for $x$ in $[\pi/2, 3\pi/2]$, and $x - 2\pi$ for $x$ in $[3\pi/2, 5\pi/2]$. For example, $4\pi/5$ is in $[\pi/2, 3\pi/2]$, so $\sin^{-1}(\sin(4\pi/5)) = \pi - 4\pi/5 = \pi/5$. 2. Cosine: $\cos^{-1}(\cos x) = 2\pi - x$ for $x$ in $[\pi, 2\pi]$. For example, $3\pi/2$ is in $[\pi, 2\pi]$, so $\cos^{-1}(\cos(3\pi/2)) = 2\pi - 3\pi/2 = \pi/2$. 3. Tangent: $\tan^{-1}(\tan x) = x - k \pi$, with the integer $k$ chosen so the answer lands in $(-\pi/2, \pi/2)$. For example, $2\pi/3$ needs $k = 1$, so $\tan^{-1}(\tan(2\pi/3)) = 2\pi/3 - \pi = -\pi/3$.

Check yourself
  1. A student writes: “$\sin^{-1}(\sin x) = x$ for every real $x$, since the inverse undoes the function.” What is the flaw?
    1. $\sin^{-1}(\sin x)$ is undefined whenever $x$ is negative
    2. only holds when $x$ already lies inside the branch $[-\pi/2, \pi/2]$
    3. nothing: the cancellation genuinely holds for every real $x$
    4. $\sin$ and $\sin^{-1}$ do not actually undo each other under any circumstances
    Check your answer
    1. $\sin^{-1}(\sin x)$ is undefined whenever $x$ is negative — $\sin x$ always produces a value in $[-1, 1]$ for any real $x$, so $\sin^{-1}(\sin x)$ is always defined , negative $x$ causes no problem here.
    2. ✓ only holds when $x$ already lies inside the branch $[-\pi/2, \pi/2]$ — (B) The cancellation only holds when $x$ already lies inside the principal branch $[-\pi/2, \pi/2]$, outside it, $\sin^{-1}(\sin x)$ returns the branch angle with the same sine, not $x$ itself
    3. nothing: the cancellation genuinely holds for every real $x$ — $\sin^{-1}(\sin(\pi)) = \sin^{-1}(0) = 0$, not $\pi$, the cancellation genuinely fails once $x$ leaves the branch $[-\pi/2, \pi/2]$.
    4. $\sin$ and $\sin^{-1}$ do not actually undo each other under any circumstances — Inside the branch $[-\pi/2, \pi/2]$, $\sin^{-1}(\sin x) = x$ is completely correct, the functions do cancel there. The flaw is only in claiming it holds EVERYWHERE.
  2. Is $\cos^{-1}(\cos(-\pi/3)) = -\pi/3$?
    1. the expression is undefined: $-\pi/3$ is negative
    2. yes: cosine and its inverse always cancel
    3. no: $\cos^{-1}(\cos(-\pi/3)) = \pi/3$
    4. no: because $\cos(-\pi/3)$ does not actually equal $\cos(\pi/3)$
    Check your answer
    1. the expression is undefined: $-\pi/3$ is negative — $\cos$ is defined at every real angle, including negative ones, $\cos(-\pi/3) = 1/2$. Nothing about this expression is undefined; it just doesn’t cancel to $-\pi/3$.
    2. yes: cosine and its inverse always cancel — $\cos^{-1}(\cos x) = x$ requires $x \in [0, \pi]$. $-\pi/3$ is outside that branch, so the cancellation does not hold as claimed, the answer is $\pi/3$, not $-\pi/3$.
    3. ✓ no: $\cos^{-1}(\cos(-\pi/3)) = \pi/3$ — (C) $\cos(-\pi/3) = 1/2$, and the branch angle with cosine $1/2$ is $\pi/3$, since $-\pi/3$ itself is outside $[0, \pi]$
    4. no: because $\cos(-\pi/3)$ does not actually equal $\cos(\pi/3)$ — $\cos(-\pi/3) = \cos(\pi/3)$ is correct, cosine is even. That equality is fine; the reason the answer isn’t $-\pi/3$ is the branch condition, not this identity.
  3. Is $\tan^{-1}(\tan(3\pi/4)) = 3\pi/4$?
    1. no: equals $-\pi/4$, since $3\pi/4$ lies outside the open branch
    2. the expression is undefined, since $\tan(3\pi/4)$ does not exist
    3. no: because $\tan(3\pi/4)$ is actually positive, not negative
    4. yes: $\tan$ and $\tan^{-1}$ always cancel to give back the original angle
    Check your answer
    1. ✓ no: equals $-\pi/4$, since $3\pi/4$ lies outside the open branch — (A) $\tan(3\pi/4) = -1$, and the branch angle with tangent $-1$ is $-\pi/4$, since $3\pi/4$ itself is outside $(-\pi/2, \pi/2)$
    2. the expression is undefined, since $\tan(3\pi/4)$ does not exist — $\tan(3\pi/4) = -1$ is a perfectly ordinary value, $\tan$ is only undefined at odd multiples of $\pi/2$, and $3\pi/4$ is not one of them.
    3. no: because $\tan(3\pi/4)$ is actually positive, not negative — $\tan(3\pi/4) = -1$ is negative, as expected in the second quadrant. This sign is correct; the issue is only that $3\pi/4$ itself lies outside the branch.
    4. yes: $\tan$ and $\tan^{-1}$ always cancel to give back the original angle — $\tan^{-1}(\tan x) = x$ requires $x \in (-\pi/2, \pi/2)$. $3\pi/4$ is outside that branch, so the answer is not $3\pi/4$, it is $-\pi/4$, the branch angle with the same tangent.
  4. For which $x$ is $\sin^{-1}(\sin x) = x$ GUARANTEED to hold?
    1. for every real $x$, with no restriction at all
    2. only for $x = 0$, since that is the one value both functions agree on trivially
    3. only for $x \in [-\pi/2, \pi/2]$
    4. only for positive $x$, since $\sin^{-1}$’s branch starts near $0$
    Check your answer
    1. for every real $x$, with no restriction at all — The cancellation is NOT unconditional, once $x$ leaves $[-\pi/2, \pi/2]$, $\sin^{-1}(\sin x)$ returns the branch angle with the same sine, which is a different number from $x$.
    2. only for $x = 0$, since that is the one value both functions agree on trivially — The cancellation holds throughout $[-\pi/2, \pi/2]$, not only at $x = 0$, for instance it also holds at $x = \pi/4$, $x = -\pi/3$, and every other point in the branch.
    3. ✓ only for $x \in [-\pi/2, \pi/2]$ — (C) The branch $[-\pi/2, \pi/2]$ is exactly where $\sin$ is one-one, so $\sin^{-1}$ can recover $x$ exactly there and nowhere else
    4. only for positive $x$, since $\sin^{-1}$’s branch starts near $0$ — $[-\pi/2, \pi/2]$ includes negative angles, and the cancellation holds there too, for instance $\sin^{-1}(\sin(-\pi/4)) = -\pi/4$ genuinely holds.
Pi over four and three pi over four have the same height, one over root two. Only pi over four is inside the shaded branch.
3π/4
→
π minus x
→
sin⁻¹ of sin
→
π/4
Swap 3π/4 for π/4, the angle on the branch with the same sine. Then sin⁻¹ undoes sin.
Worked example

Evaluate $\sin^{-1}(\sin(3\pi/4))$

  1. $\sin^{-1}(\sin(3\pi/4))$
    The working above cancelled here. $3\pi/4$ is outside the branch $[-\pi/2, \pi/2]$, so the answer cannot be $3\pi/4$.
  2. $\sin(3\pi/4) = \sin(\pi - \pi/4) = \sin(\pi/4) = 1/\sqrt{2}$
    Reduce first. $3\pi/4$ and $\pi/4$ have the same sine, because $\sin(\pi - \theta) = \sin \theta$.
  3. $\sin^{-1}(1/\sqrt{2}) = \pi/4$
    Now use the definition on that value. $\pi/4$ is inside the branch.
  4. *$\sin^{-1}(\sin(3\pi/4)) = \pi/4$*
    So the answer is $\pi/4$, the angle on the branch with the same sine as $3\pi/4$.
  5. $\sin(\pi/4) = \sin(3\pi/4) = 1/\sqrt{2}$, and $\pi/4 \approx 0.79$ lies in $[-1.57, 1.57]$
    Check the two things the answer needs. It has the same sine, and it is on the branch.
Check yourself
  1. Evaluate $\sin^{-1}(\sin(3\pi/4))$.
    1. $\pi/4$
    2. $3\pi/4$
    3. $-3\pi/4$
    4. $-\pi/4$
    Check your answer
    1. ✓ $\pi/4$ — (A) $\sin(3\pi/4) = \sin(\pi - \pi/4) = \sin(\pi/4) = 1/\sqrt{2}$, and $\sin^{-1}(1/\sqrt{2}) = \pi/4$, which IS in the branch
    2. $3\pi/4$ — $3\pi/4$ does not lie in $[-\pi/2, \pi/2]$, the cancellation to “the original angle” only applies inside the branch, and $3\pi/4$ is outside it.
    3. $-3\pi/4$ — $-3\pi/4$ sits well beyond $-\pi/2$, the branch’s own boundary, outside $[-\pi/2, \pi/2]$ regardless of its sine.
    4. $-\pi/4$ — $\sin(3\pi/4) = 1/\sqrt{2}$, positive, $\sin(-\pi/4) = -1/\sqrt{2}$ is a different, negative value, so it can’t be the branch angle here.
  2. Evaluate $\sin^{-1}(\sin(5\pi/6))$.
    1. $\pi/6$
    2. $5\pi/6$
    3. $7\pi/6$
    4. $-\pi/6$
    Check your answer
    1. ✓ $\pi/6$ — (A) $\sin(5\pi/6) = \sin(\pi - \pi/6) = \sin(\pi/6) = 1/2$, and $\sin^{-1}(1/2) = \pi/6$, which is inside the branch
    2. $5\pi/6$ — $5\pi/6$ is outside $[-\pi/2, \pi/2]$, so it can’t be the branch answer, the reduction process is needed first, the same as for $3\pi/4$.
    3. $7\pi/6$ — $7\pi/6$ is well outside $[-\pi/2, \pi/2]$, $\sin^{-1}$’s output is guaranteed to stay inside the branch.
    4. $-\pi/6$ — $\sin(5\pi/6) = 1/2$ is positive, so the branch angle it reduces to must also have a positive sine , $-\pi/6$ has the wrong sign.
  3. Why does evaluating $\sin^{-1}(\sin(3\pi/4))$ take two separate steps, rather than cancelling directly?
    1. $3\pi/4$ is outside the branch, so direct cancellation fails; the angle must be reduced first
    2. because the two-step process is just a stylistic convention, not something the branch restriction actually forces
    3. because $\sin$ must first be computed as a decimal before $\sin^{-1}$ can be applied to it
    4. because $\sin^{-1}$ can only be applied to positive numbers, and $\sin(3\pi/4)$ must first be checked for its sign
    Check your answer
    1. ✓ $3\pi/4$ is outside the branch, so direct cancellation fails; the angle must be reduced first — (A) Because $3\pi/4$ lies outside the branch, so the direct-cancellation shortcut does not apply, the inner sine value has to be found and re-evaluated through $\sin^{-1}$ instead
    2. because the two-step process is just a stylistic convention, not something the branch restriction actually forces — The two-step process is not a matter of taste, since $3\pi/4$ is outside $[-\pi/2, \pi/2]$, direct cancellation is simply invalid there, and reduction is the only correct path.
    3. because $\sin$ must first be computed as a decimal before $\sin^{-1}$ can be applied to it — The whole reduction stays in exact form, $\sin(3\pi/4) = 1/\sqrt{2}$ is used exactly, never converted to a decimal. That is not why two steps are needed.
    4. because $\sin^{-1}$ can only be applied to positive numbers, and $\sin(3\pi/4)$ must first be checked for its sign — $\sin^{-1}$ is defined on all of $[-1, 1]$, negative values included, e.g. $\sin^{-1}(-1/2) = -\pi/6$ works fine. There is no sign restriction forcing an extra step.

↑ Back to top

Two different meanings for the same exponent

✕MISCONCEPTION

A car has a gear marked R. It does not make the car go slower. It does not turn the speed into one over the speed. It runs the car backwards, along the road it came by. The R names a reverse.

Dignaga
A SCHOLAR INDIA REMEMBERS

Before you ask whether $\sin^{-1} x$ equals $1/(\sin x)$, ask what kind of thing each one is. $\sin^{-1} x$ is an angle, the one in $[-\pi/2, \pi/2]$ whose sine is $x$. $1/(\sin x)$ is a ratio, and it already has a name, $\operatorname{cosec} x$.

The undo mark, in symbols

Sine is a function too. An angle goes in and a height comes out. Run it backwards and a height goes in and an angle comes out. Mathematics marks that undo with a small raised minus one.

2
→
f: 3x−2
→
f⁻¹
→
2
π/6
→
sin
→
sin⁻¹
→
π/6
The −1 is the undo mark, as in Chapter 1: f⁻¹ undoes f, and sin⁻¹ undoes sine on its branch.

So $\sin^{-1}$ is the INVERSE SINE, sine run backwards. It is the same $-1$ that names $f^{-1}$ in Chapter 1.

Inverse sine of one half is about 0.52, on the branch of angles. One over sine of one half is about 2.09, past its end.

The trouble is that a raised $-1$ on a number does mean “one over”. $2^{-1} = 1/2$. So $\sin^{-1} x$ looks as if it means $1/(\sin x)$. It does not. The number $1/(\sin x)$ already has its own name. It is $\operatorname{cosec} x$.

1/2
→
sin⁻¹
→
≈0.52
1/2
→
sin
→
1/x
→
≈2.09
sin⁻¹(1/2) is an angle, about 0.52. One over sin(1/2) is a number, about 2.09. Different functions, different answers.

Putting numbers on both

Put numbers on both. $\sin^{-1}(1/2) = \pi/6 \approx 0.52$. That is an angle in radians. Now work out $1/(\sin(1/2))$. Here $1/2$ is an angle of half a radian. Its sine is about $0.48$, and one over that is about $2.09$. That is a plain ratio. Different kinds of answer, and far apart.

Try it yourself. Is $\cos^{-1}(1/2)$ an angle or a number? And $1/(\cos(1/2))$?

This gives you a quick test. An inverse trigonometric answer is always an angle. A reciprocal is always a ratio. Ask which one the question wants.

Your turn: is $\sin^{-1}(1)$ equal to $1/(\sin 1)$? Work out both, in radians, and compare.

Check yourself
  1. A student claims $\sin^{-1} x$ means $1/(\sin x)$, since a “$-1$ exponent usually means reciprocal.” What is wrong with this claim?
    1. $\sin^{-1} x$ is undefined whenever $\sin x$ is zero, so the comparison never even makes sense
    2. $\sin^{-1} x$ is an angle; $1/(\sin x)$ is a number, already named $\operatorname{cosec} x$
    3. nothing: $\sin^{-1} x$ and $1/(\sin x)$ are genuinely the same thing
    4. $1/(\sin x)$ has no name of its own, so the notation clash never actually causes confusion
    Check your answer
    1. $\sin^{-1} x$ is undefined whenever $\sin x$ is zero, so the comparison never even makes sense — $\sin^{-1} x$’s domain restriction is on $x$ itself ($[-1, 1]$), not on whether some OTHER quantity $\sin x$ is zero, that condition doesn’t apply to $\sin^{-1}$ at all.
    2. ✓ $\sin^{-1} x$ is an angle; $1/(\sin x)$ is a number, already named $\operatorname{cosec} x$ — (B) $\sin^{-1} x$ is the INVERSE SINE FUNCTION, an angle, while $1/(\sin x)$ is a NUMBER, already named $\operatorname{cosec} x$
    3. nothing: $\sin^{-1} x$ and $1/(\sin x)$ are genuinely the same thing — $\sin^{-1}(1/2) = \pi/6 \approx 0.52$, while $1/(\sin(1/2)) \approx 2.09$, two very different numbers, one an angle and one a ratio.
    4. $1/(\sin x)$ has no name of its own, so the notation clash never actually causes confusion — $1/(\sin x)$ is already named $\operatorname{cosec} x$, cosecant. Its having a name is exactly the point: reusing $\sin^{-1} x$ for the same idea would be redundant and wrong, not harmless.
  2. Which expression actually equals $\operatorname{cosec} x$?
    1. $\sin^{-1} x$
    2. $1/(\sin x)$
    3. $\sin(-x)$
    4. $-\sin x$
    Check your answer
    1. $\sin^{-1} x$ — $\sin^{-1} x$ names an angle, not a number computed by dividing $1$ by $\sin x$ , it is not $\operatorname{cosec} x$, however similar the notation looks.
    2. ✓ $1/(\sin x)$ — (B) $\operatorname{cosec} x$ is defined as exactly $1/(\sin x)$, the reciprocal of the sine value
    3. $\sin(-x)$ — $\sin(-x) = -\sin x$ negates the sine value via oddness, it never involves dividing $1$ by anything, so it cannot be the reciprocal $\operatorname{cosec} x$.
    4. $-\sin x$ — $-\sin x$ flips the sign of $\sin x$, still not a reciprocal. $\operatorname{cosec} x$ specifically means $1$ divided by $\sin x$, a different operation from negation.
  3. Does $\cos^{-1} x$ mean $1/(\cos x)$?
    1. yes, but only when $x$ is a positive number
    2. yes: the same “$-1$ means reciprocal” rule that fails for sine works correctly for cosine
    3. no: $\cos^{-1} x$ is an angle; $1/(\cos x)$ is $\sec x$
    4. the question doesn’t make sense, since $\cos^{-1} x$ has no defined meaning at all
    Check your answer
    1. yes, but only when $x$ is a positive number — Whether $x$ is positive or negative has no bearing on this, $\cos^{-1} x$ names the inverse cosine function either way, never the reciprocal of cosine.
    2. yes: the same “$-1$ means reciprocal” rule that fails for sine works correctly for cosine — The same distinction applies across every trig function, $\cos^{-1} x$ is an inverse function (an angle), never a reciprocal, exactly like $\sin^{-1} x$. There is no exception where the shortcut becomes correct.
    3. ✓ no: $\cos^{-1} x$ is an angle; $1/(\cos x)$ is $\sec x$ — (C) No, $\cos^{-1} x$ is the inverse cosine function (an angle); $1/(\cos x)$ is already named $\sec x$
    4. the question doesn’t make sense, since $\cos^{-1} x$ has no defined meaning at all — $\cos^{-1} x$ is a completely standard, well-defined notation, the inverse cosine function, domain $[-1, 1]$, range $[0, \pi]$. It is not meaningless; it is simply not a reciprocal.

↑ Back to top

Complementary angle identities

▲CONCEPT

Go back to the ladder against the wall. The ladder, the wall and the ground make a right-angled triangle. Its other two angles add up to 90°. One sits at the foot of the ladder. The other sits at the top, against the wall.

Reading one number two ways

The ladder was 4 metres long, with its top 2 metres up. At the foot, height over length is 2 over 4. That is the sine of the foot angle, 30°. At the top, the same 2 over 4 is the cosine of the top angle, 60°. One number, read as a sine at one corner and as a cosine at the other. The two angles add to 90°.

Inverse sine climbs, inverse cosine falls, and their sum stays at pi over two. At one half, pi over six plus pi over three.

That is a COMPLEMENTARY pair: two angles that fill a right angle. In radians, $\sin^{-1} x + \cos^{-1} x = \pi/2$ for every $x$ in $[-1, 1]$.

Let us see why it holds for every such $x$, negatives too. 1. Call $\theta = \sin^{-1} x$. Then $\theta$ is in $[-\pi/2, \pi/2]$. 2. So $\pi/2 - \theta$ is in $[0, \pi]$, which is cosine’s branch. 3. And $\cos(\pi/2 - \theta) = \sin \theta = x$, the Class 11 co-function rule. 4. So $\cos^{-1} x = \pi/2 - \theta$, the definition of $\cos^{-1}$ on its branch. 5. Add $\theta$ to both sides: $\sin^{-1} x + \cos^{-1} x = \pi/2$.

1/2
→
sin⁻¹
→
+ cos⁻¹(1/2)
→
π/2
0
→
sin⁻¹ + cos⁻¹
→
π/2
−1
→
sin⁻¹ + cos⁻¹
→
π/2
Three inputs, one total: at 1/2, 0 and −1, sin⁻¹ and cos⁻¹ add to π/2 every time.

A quick check at $x = 0$. $\sin^{-1}(0) = 0$ and $\cos^{-1}(0) = \pi/2$. They add to $\pi/2$.

sin⁻¹, cos⁻¹
sin⁻¹cos⁻¹
tan⁻¹, cot⁻¹
tan⁻¹cot⁻¹
sec⁻¹, cosec⁻¹
sec⁻¹cosec⁻¹
Three pairs add to π/2: sine with cosine, tangent with cotangent, secant with cosecant.
sin⁻¹ + cos⁻¹
tan⁻¹ + cot⁻¹
sec⁻¹ + cosec⁻¹
Each pair holds only where both are defined: −1 to 1, every real number, or outside −1 to 1.

Three pairs, same rule

Two more pairs work the same way, each on its own domain.

Each is proved by the same step as the first pair. 1. Tangent and cotangent: call $\theta = \tan^{-1} x$, in $(-\pi/2, \pi/2)$. Then $\pi/2 - \theta$ is in $(0, \pi)$, cotangent’s branch, and $\cot(\pi/2 - \theta) = \tan \theta = x$. So $\cot^{-1} x = \pi/2 - \theta$. 2. Secant and cosecant: for $|x| \geq 1$, call $\theta = \sec^{-1} x$. Then $\pi/2 - \theta$ is in cosecant’s branch, and $\operatorname{cosec}(\pi/2 - \theta) = \sec \theta = x$. So $\operatorname{cosec}^{-1} x = \pi/2 - \theta$.

The secant and cosecant pair needs $|x| \geq 1$, because neither inverse exists between $-1$ and $1$. The branches do not break any of these rules. They only fix which angle each side names.

Complementary pairDomain
$\sin^{-1} x + \cos^{-1} x = \pi/2$$x$ in $[-1, 1]$
$\tan^{-1} x + \cot^{-1} x = \pi/2$every real $x$
$\sec^{-1} x + \operatorname{cosec}^{-1} x = \pi/2$$|x| \geq 1$

This saves you work. Given one member of a pair, the other is $\pi/2$ minus it. You never compute the harder one.

Your turn: check $\tan^{-1} x + \cot^{-1} x = \pi/2$ at $x = -1$. Does it still hold for a negative input?

Check yourself
  1. For $x \in [-1, 1]$, what does $\sin^{-1} x + \cos^{-1} x$ equal?
    1. $\pi/2$
    2. $0$
    3. it depends on $x$, with no fixed sum
    4. $\pi$
    Check your answer
    1. ✓ $\pi/2$ — (A) $\sin^{-1} x + \cos^{-1} x = \pi/2$ for $x \in [-1, 1]$, following from the co-function pair $\cos((\pi/2) - \theta) = \sin \theta$
    2. $0$ — $\sin^{-1} x$ and $\cos^{-1} x$ do not cancel, e.g. at $x = 1/2$, $\sin^{-1}(1/2) = \pi/6$ and $\cos^{-1}(1/2) = \pi/3$, summing to $\pi/2$, not $0$.
    3. it depends on $x$, with no fixed sum — The sum $\sin^{-1} x + \cos^{-1} x$ is exactly $\pi/2$ for every $x$ in $[-1, 1]$, constant, not varying , which is what makes the identity useful for swapping one hard value for an easier one.
    4. $\pi$ — $\sin^{-1} x + \cos^{-1} x = \pi/2$, not $\pi$, $\pi$ would be the SUPPLEMENTARY total, which this pair does not follow.
  2. Which complementary pair requires $|x| \geq 1$, rather than holding for every real $x$ or every $x \in [-1, 1]$?
    1. $\tan^{-1} x + \cot^{-1} x = \pi/2$
    2. $\sin^{-1} x + \cos^{-1} x = \pi/2$
    3. none of the three pairs carries any domain condition, since all three simply always hold
    4. $\sec^{-1} x + \operatorname{cosec}^{-1} x = \pi/2$, for $|x| \geq 1$
    Check your answer
    1. $\tan^{-1} x + \cot^{-1} x = \pi/2$ — $\tan^{-1} x + \cot^{-1} x = \pi/2$ holds for every real $x$, no restriction whatsoever, unlike the $\sec$ and $\operatorname{cosec}$ pair which does need $|x| \geq 1$.
    2. $\sin^{-1} x + \cos^{-1} x = \pi/2$ — $\sin^{-1} x + \cos^{-1} x = \pi/2$ holds for $x \in [-1, 1]$, the closed interval, not the $|x| \geq 1$ condition that belongs to the $\sec$ and $\operatorname{cosec}$ pair.
    3. none of the three pairs carries any domain condition, since all three simply always hold — Each pair DOES carry its own domain condition, $x \in [-1, 1]$ for the sin-cos pair, all reals for the tan-cot pair, and $|x| \geq 1$ for the sec and cosec pair. None is unconditional.
    4. ✓ $\sec^{-1} x + \operatorname{cosec}^{-1} x = \pi/2$, for $|x| \geq 1$ — (D) $\sec^{-1}$ and $\operatorname{cosec}^{-1}$ are only defined for $|x| \geq 1$, so their sum identity inherits that same condition
  3. Where does the complementary identity $\sin^{-1} x + \cos^{-1} x = \pi/2$ actually come from?
    1. it follows from squaring both $\sin^{-1} x$ and $\cos^{-1} x$ and adding the results
    2. it follows from the Class 11 co-function pair $\cos((\pi/2) - \theta) = \sin \theta$
    3. it is simply defined to be true, as a convention chosen for convenience
    4. it is an entirely new fact about inverse trig functions, unconnected to anything in Class 11
    Check your answer
    1. it follows from squaring both $\sin^{-1} x$ and $\cos^{-1} x$ and adding the results — No squaring step is involved, the derivation runs through the co-function pair $\cos((\pi/2) - \theta) = \sin \theta$, applied at the angle each inverse function names.
    2. ✓ it follows from the Class 11 co-function pair $\cos((\pi/2) - \theta) = \sin \theta$ — (B) Each of the three complementary pairs follows from a co-function pair already known from Class 11, such as $\cos((\pi/2) - \theta) = \sin \theta$
    3. it is simply defined to be true, as a convention chosen for convenience — The identity can be derived from $\cos((\pi/2) - \theta) = \sin \theta$, it is proved, not simply declared true by convention.
    4. it is an entirely new fact about inverse trig functions, unconnected to anything in Class 11 — This identity follows from $\cos((\pi/2) - \theta) = \sin \theta$, a co-function pair already known from Class 11, it is a consequence of prior trigonometry, not a new invention.
Worked example

Check $\sin^{-1} x + \cos^{-1} x = \pi/2$ at $x = 1/2$

  1. $\sin^{-1}(1/2) = \pi/6$
    You found this value in the first worked example.
  2. $\cos^{-1}(1/2) = \pi/3$
    $\cos(\pi/3) = 1/2$, and $\pi/3$ is inside cosine’s branch $[0, \pi]$.
  3. $\pi/6 + \pi/3 = \pi/6 + 2\pi/6 = 3\pi/6 = \pi/2$
    Add the two. Put both over $6$ first.
  4. *$\sin^{-1}(1/2) + \cos^{-1}(1/2) = \pi/2$*
    The identity holds at $x = 1/2$.
  5. $\sin^{-1}(1) + \cos^{-1}(1) = \pi/2 + 0 = \pi/2$
    Check at a second value, $x = 1$. One match does not prove an identity. It only shows it is not plainly wrong.
Check yourself
  1. Given $\sin^{-1}(1/2) = \pi/6$, use the complementary identity to find $\cos^{-1}(1/2)$.
    1. $2\pi/3$
    2. $\pi/3$
    3. $5\pi/6$
    4. $\pi/6$
    Check your answer
    1. $2\pi/3$ — $\pi/2 - \pi/6 = \pi/3$, not $\pi - \pi/6 = 5\pi/6$’s neighbor $2\pi/3$, the identity subtracts from $\pi/2$, the complementary total, not from $\pi$.
    2. ✓ $\pi/3$ — (B) $\cos^{-1}(1/2) = \pi/2 - \sin^{-1}(1/2) = \pi/2 - \pi/6 = \pi/3$
    3. $5\pi/6$ — $\pi - \pi/6 = 5\pi/6$ uses the wrong total, the complementary identity’s total is $\pi/2$, not $\pi$.
    4. $\pi/6$ — $\pi/6$ is the GIVEN value, $\sin^{-1}(1/2)$, copying it doesn’t answer the question, which asks for $\cos^{-1}(1/2)$ via the identity $\pi/2 - \pi/6$.
  2. Why can the complementary identity let you swap a hard $\cos^{-1}$ evaluation for an easier $\sin^{-1}$ one?
    1. because the swap only works for $x = 1/2$, the one value the worked example happened to use
    2. because $\cos^{-1}$ and $\sin^{-1}$ always give exactly the same numeric value for any input
    3. the two sum to the fixed $\pi/2$, so one gives the other
    4. because $\cos^{-1}$ is simply harder to compute in general, with no actual mathematical relationship to $\sin^{-1}$ involved
    Check your answer
    1. because the swap only works for $x = 1/2$, the one value the worked example happened to use — The identity holds for every $x$ in $[-1, 1]$, $x = 1/2$ was just the worked example’s chosen value, not a special case where the identity happens to be true.
    2. because $\cos^{-1}$ and $\sin^{-1}$ always give exactly the same numeric value for any input — $\sin^{-1}(1/2) = \pi/6$ and $\cos^{-1}(1/2) = \pi/3$, different values. They are related by a fixed SUM, $\pi/2$, not by being equal to each other.
    3. ✓ the two sum to the fixed $\pi/2$, so one gives the other — (C) The two always sum to the fixed value $\pi/2$ on their shared domain, so knowing one instantly gives the other by subtraction
    4. because $\cos^{-1}$ is simply harder to compute in general, with no actual mathematical relationship to $\sin^{-1}$ involved — There IS a real, provable relationship: $\sin^{-1} x + \cos^{-1} x = \pi/2$. That fixed sum, not a subjective sense of difficulty, is what makes the swap valid.
  3. Given $\tan^{-1}(1) = \pi/4$, find $\cot^{-1}(1)$ using the complementary identity.
    1. $3\pi/4$
    2. $\pi/4$
    3. $\pi/2$
    4. $0$
    Check your answer
    1. $3\pi/4$ — The identity requires SUBTRACTING $\tan^{-1}(1)$ from $\pi/2$, not adding, $\pi/2 - \pi/4 = \pi/4$, not $\pi/2 + \pi/4$.
    2. ✓ $\pi/4$ — (B) $\cot^{-1}(1) = \pi/2 - \tan^{-1}(1) = \pi/2 - \pi/4 = \pi/4$
    3. $\pi/2$ — The identity says $\tan^{-1}(1) + \cot^{-1}(1) = \pi/2$, $\pi/2$ itself is the SUM, not the answer for $\cot^{-1}(1)$ alone.
    4. $0$ — If $\cot^{-1}(1) = 0$, the sum $\tan^{-1}(1) + \cot^{-1}(1)$ would be $\pi/4$, not $\pi/2$, contradicting the identity, which fixes the sum at $\pi/2$ for every input.

↑ Back to top

Negating the input, two different ways

▲CONCEPT

A thermometer scale runs from minus 10 to 10, with 0 in the middle. Change a reading’s sign to put it on the other side of zero. 3 becomes minus 3, and it is still on the scale.

Two kinds of scale

Now take a test marked out of 100. Change the sign of 30 to get minus 30 and you fall off the scale. On this scale, taking it from 100 is what stays on: 30 becomes 100 minus 30, which is 70.

sin⁻¹
0
0
cos⁻¹
0
0
On sine’s branch, 1/2 and −1/2 give answers mirrored about 0. On cosine’s, they are mirrored about π/2.

A scale centred on zero changes sign. A scale that starts at zero takes the reading away from the top. The six inverse functions split the same way, three and three.

−1/2
→
sin⁻¹
→
−π/6
1/2
→
sin⁻¹
→
minus
→
−π/6
sin⁻¹(−1/2) is −π/6: the same as taking sin⁻¹(1/2) and putting a minus in front.

Three branches centred on zero

Three branches are centred on $0$. Negating the input just negates the answer.

−1/2
→
cos⁻¹
→
2π/3
1/2
→
cos⁻¹
→
π minus
→
2π/3
cos⁻¹(−1/2) is 2π/3: the same as taking cos⁻¹(1/2) away from π, not putting a minus in front.

Three branches that start at zero

Three branches start at $0$. Negating the input takes the answer away from $\pi$.

1/2
→
cos⁻¹
→
minus
→
−π/3
−1/2
→
cos⁻¹
→
2π/3
Put a minus in front for cosine and you get −π/3, off cosine’s branch. The right answer is 2π/3.

Why can cosine, secant and cotangent not use a minus sign? A minus sign would give a negative angle, and their branches hold no negative angles at all. Here is the step that works instead. 1. Cosine: call $\theta = \cos^{-1} x$, in $[0, \pi]$. Then $\cos(\pi - \theta) = -\cos \theta = -x$, and $\pi - \theta$ is in $[0, \pi]$ too, so it is the branch angle with cosine $-x$: $\cos^{-1}(-x) = \pi - \theta = \pi - \cos^{-1} x$. 2. Secant: the same step gives $\sec^{-1}(-x) = \pi - \sec^{-1} x$, since $\sec(\pi - \theta) = -\sec \theta$ and $\pi - \theta$ stays off $\pi/2$ whenever $\theta$ does. 3. Cotangent: the same step gives $\cot^{-1}(-x) = \pi - \cot^{-1} x$, since $\cot(\pi - \theta) = -\cot \theta$ and $\pi - \theta$ is in $(0, \pi)$ whenever $\theta$ is.

Let us check one. $\cos^{-1}(1/2) = \pi/3$. So $\cos^{-1}(-1/2) = \pi - \pi/3 = 2\pi/3$. That is inside $[0, \pi]$, as it must be.

Inputs 0.6 and minus 0.6 on both curves. Inverse tangent’s two values balance about the origin. Inverse cosine’s two values balance about pi over two.
centred on 0: minus sign
sin⁻¹tan⁻¹cosec⁻¹
starts at 0: π minus
cos⁻¹cot⁻¹sec⁻¹
Sort the six by branch: centred on 0 takes a minus sign, starting at 0 takes π minus.

Sort the six by shape once, and you have two rules to hold instead of six: a minus sign for the three centred on 0, and $\pi$ minus for the three that start at 0.

Your turn: find $\cos^{-1}(-1)$ with the rule. What would a plain minus sign have given?

Check yourself
  1. Which pattern does $\sin^{-1}(-x)$ follow?
    1. $\pi - \sin^{-1} x$
    2. $\pi/2 - \sin^{-1} x$
    3. $-\sin^{-1} x$
    4. $\sin^{-1} x$, unchanged
    Check your answer
    1. $\pi - \sin^{-1} x$ — $\pi - \sin^{-1} x$ is the supplement-style pattern, the one used by functions whose branch runs from $0$ rather than being centred on it. $\sin^{-1}$’s branch IS centred on $0$, so it follows the odd-style pattern instead.
    2. $\pi/2 - \sin^{-1} x$ — $\pi/2$ is the total from the COMPLEMENTARY identity ($\sin^{-1} x + \cos^{-1} x = \pi/2$), a different rule from how $\sin^{-1}$ handles a negated input.
    3. ✓ $-\sin^{-1} x$ — (C) $\sin^{-1}$’s branch $[-\pi/2, \pi/2]$ is symmetric about $0$, so it follows the ODD-style pattern: $\sin^{-1}(-x) = -\sin^{-1} x$
    4. $\sin^{-1} x$, unchanged — $\sin^{-1}(-1/2) = -\pi/6$, which is NOT the same as $\sin^{-1}(1/2) = \pi/6$, negating the input genuinely negates the output here.
  2. Why does $\cos^{-1}(-x) = \pi - \cos^{-1} x$, rather than simply $-\cos^{-1} x$ like sine’s pattern?
    1. because simply changing the sign is mathematically impossible to apply to any inverse trig function
    2. because $\pi - \cos^{-1} x$ and $-\cos^{-1} x$ always happen to be numerically equal anyway
    3. $\cos^{-1}$’s branch $[0, \pi]$ isn’t symmetric about $0$
    4. because cosine and sine are entirely unrelated functions with no shared structure at all
    Check your answer
    1. because simply changing the sign is mathematically impossible to apply to any inverse trig function — Changing the sign is exactly the correct pattern for $\sin^{-1}$, $\tan^{-1}$, $\operatorname{cosec}^{-1}$, their branches ARE centred on $0$. It only fails for $\cos^{-1}$, $\cot^{-1}$, $\sec^{-1}$, whose branches are not.
    2. because $\pi - \cos^{-1} x$ and $-\cos^{-1} x$ always happen to be numerically equal anyway — $\pi - \cos^{-1} x$ and $-\cos^{-1} x$ only coincide at one special value, in general they differ, so the choice of pattern genuinely matters, and only the supplement-style form stays inside cosine’s branch.
    3. ✓ $\cos^{-1}$’s branch $[0, \pi]$ isn’t symmetric about $0$ — (C) $\cos^{-1}$’s branch $[0, \pi]$ is not symmetric about $0$, so simply changing the sign would land outside the branch
    4. because cosine and sine are entirely unrelated functions with no shared structure at all — Sine and cosine are tightly related, the complementary identity itself connects them. The difference in negative-argument pattern comes specifically from their different branch shapes, not from having nothing to do with each other.
  3. Which pattern does $\cot^{-1}(-x)$ follow?
    1. $\pi/2 - \cot^{-1} x$
    2. $-\cot^{-1} x$
    3. $2\pi - \cot^{-1} x$
    4. $\pi - \cot^{-1} x$
    Check your answer
    1. $\pi/2 - \cot^{-1} x$ — $\pi/2 - \cot^{-1} x$ mixes in the complementary identity’s total, a different rule from how $\cot^{-1}$ handles a negated input, which uses $\pi$, not $\pi/2$.
    2. $-\cot^{-1} x$ — $\cot^{-1}$’s branch $(0, \pi)$ is not centred on $0$, it runs from $0$ to $\pi$. That shape calls for the supplement-style pattern, $\pi - \cot^{-1} x$, not a plain sign flip.
    3. $2\pi - \cot^{-1} x$ — $\cot^{-1}$’s branch $(0, \pi)$ has width $\pi$, the supplement-style pattern uses that same total, $\pi - \cot^{-1} x$, not $2\pi$.
    4. ✓ $\pi - \cot^{-1} x$ — (D) $\cot^{-1}$’s branch $(0, \pi)$ runs from $0$ rather than being centred on it, so it follows the SUPPLEMENT-style pattern: $\cot^{-1}(-x) = \pi - \cot^{-1} x$

↑ Back to top

The addition formula for tangent inverse

▲CONCEPT

Turn a camera on its stand through 30°. Then turn it a further 15°. In all it has turned 45°. Angles add. Slopes do not. A 30° slope is about 0.58, and a 15° slope is about 0.27. They add to about 0.85. But a 45° slope is exactly 1.

Dignaga
A SCHOLAR INDIA REMEMBERS

The formula looks like plain algebra. So what did you take on trust? The condition $x y < 1$. Try $x = 1/4$ and $y = 3/5$. Then $x y = 3/20$, under $1$, and the formula gives $\tan^{-1}(1) = \pi/4$. When $x y > 1$, the true sum lies outside $(-\pi/2, \pi/2)$, where $\tan^{-1}$ cannot reach. Then a $\pi$ term is needed. Check $x y$ first.

So if you know two slopes and want the slope of the two angles added together, you need a rule. You met it in Class 11 as the tangent addition formula.

['1/2', '1/3']
→
add the two tan⁻¹
→
π/4
['1/2', '1/3']
→
(x + y)/(1 − xy)
→
tan⁻¹
→
π/4
For 1/2 and 1/3, adding the two angles and using the formula both give π/4.

The addition formula, for inverses

Written for inverses, it is the ADDITION FORMULA for tangent inverse. For $x y < 1$, $\tan^{-1} x + \tan^{-1} y = \tan^{-1}((x + y)/(1 - x y))$.

Shaded: every pair with x times y below 1. The worked pair, one half and one third, is inside. The pair 2, 2 is not.

The condition $x y < 1$ is part of the formula. Call $\alpha = \tan^{-1} x$ and $\beta = \tan^{-1} y$. The reason is in the sign of $\tan(\alpha + \beta) = (x + y)/(1 - x y)$, which splits into three cases. 1. $x, y$ both positive: $x y < 1$ keeps $1 - x y$ positive, so $\tan(\alpha + \beta)$ is positive; since $\alpha + \beta$ is between $0$ and $\pi$, that puts it under $\pi/2$, inside tangent’s branch, where $\tan^{-1}$ undoes $\tan$. 2. $x, y$ opposite signs: one angle is above $0$ and the other below it, so the sum stays inside the branch on its own. 3. $x, y$ both negative: the same argument as case 1 runs in mirror image, below $0$. So the formula holds for every pair with $x y < 1$.

A pair that breaks the formula

Let us try a pair that breaks it. Take $x = y = 2$. Then $x y = 4$. Each $\tan^{-1}(2)$ is about $1.107$, so the true sum is about $2.214$. That is more than $\pi/2$. The fraction gives $(2 + 2)/(1 - 4) = -4/3$, and $\tan^{-1}(-4/3)$ is about $-0.927$. The formula’s answer is off by $\pi$.

['2', '2']
→
add the two tan⁻¹
→
≈2.21
['2', '2']
→
(x + y)/(1 − xy)
→
tan⁻¹
→
≈−0.93
At x = y = 2, xy is 4. The two routes give about 2.21 and about −0.93, exactly π apart.

*So when $x y > 1$* the plain formula lands in the wrong place. A $+ \pi$ brings it back when $x$ and $y$ are both positive, and a $- \pi$ when both are negative. At $x y = 1$ the bottom of the fraction is $0$, and the formula gives nothing at all.

['1/2', '1/3']
→
xy
→
1/6
Multiply first: 1/2 times 1/3 is 1/6, under 1, so the plain formula applies.

*So work out $x y$* on its own line first. Then decide whether the plain formula applies.

Your turn: you want $\tan^{-1}(3) + \tan^{-1}(1/2)$. Work out $x y$ first. Can you use the formula as it stands?

Check yourself
  1. For $x, y$ with $x y < 1$, what does $\tan^{-1} x + \tan^{-1} y$ equal?
    1. $\tan^{-1}(x + y)$
    2. $\tan^{-1}((x - y)/(1 + x y))$
    3. $\tan^{-1}((x + y)/(1 - x y))$
    4. $\tan^{-1}(x y)$
    Check your answer
    1. $\tan^{-1}(x + y)$ — $\tan^{-1}$ is not linear, $\tan^{-1} x + \tan^{-1} y \neq \tan^{-1}(x + y)$ in general. The combined formula involves the fraction $(x+y)/(1-x y)$, from the tangent addition identity.
    2. $\tan^{-1}((x - y)/(1 + x y))$ — $(x - y)/(1 + x y)$ is the combined form for $\tan^{-1} x - \tan^{-1} y$, a sign pattern flipped from the addition case, which uses $(x + y)/(1 - x y)$.
    3. ✓ $\tan^{-1}((x + y)/(1 - x y))$ — (C) $\tan^{-1}((x + y)/(1 - x y))$, the tangent addition formula re-expressed for inverses
    4. $\tan^{-1}(x y)$ — The numerator of the combined fraction is $x + y$, the SUM, not $x y$, the product. Only the denominator involves the product $x y$.
  2. Why does the plain addition formula for $\tan^{-1} x + \tan^{-1} y$ only work when $x y < 1$?
    1. because $x y < 1$ is simply an arbitrary restriction with no mathematical reason behind it
    2. because $x$ and $y$ must both individually stay below $1$ for the formula to make sense
    3. because the formula’s fraction is undefined by division by zero throughout the whole region $x y < 1$
    4. beyond $x y > 1$ the true angle sum leaves $\tan^{-1}$’s branch. It then needs a $± \pi$ fix
    Check your answer
    1. because $x y < 1$ is simply an arbitrary restriction with no mathematical reason behind it — $x y < 1$ is not an arbitrary rule, it is exactly the boundary where the true sum of the two angles starts landing outside $\tan^{-1}$’s branch $(-\pi/2, \pi/2)$, requiring a correction term beyond it.
    2. because $x$ and $y$ must both individually stay below $1$ for the formula to make sense — The condition is $x y < 1$, a statement about the PRODUCT, not a cap on $x$ or $y$ individually. $x = 5$ and $y = 0.1$ satisfy $x y = 0.5 < 1$ even though $x$ alone exceeds $1$.
    3. because the formula’s fraction is undefined by division by zero throughout the whole region $x y < 1$ — Division by zero happens only exactly AT $x y = 1$, the boundary, not throughout the whole region $x y < 1$, where the formula works fine. The real issue beyond the boundary is a branch problem, not an undefined fraction.
    4. ✓ beyond $x y > 1$ the true angle sum leaves $\tan^{-1}$’s branch. It then needs a $± \pi$ fix — (D) Beyond $x y > 1$, the true sum lands outside $\tan^{-1}$’s branch, needing an extra $± \pi$ term to correct it
  3. Before applying the plain addition formula to $\tan^{-1}(2) + \tan^{-1}(3)$, what should be checked first?
    1. whether either $2$ or $3$ individually exceeds $\pi/2$
    2. whether the product $2 \times 3$ is less than $1$
    3. nothing: plain formula always applies regardless of input
    4. whether $2 + 3$ is less than $1$
    Check your answer
    1. whether either $2$ or $3$ individually exceeds $\pi/2$ — $x = 2$ and $y = 3$ are the numeric INPUTS to $\tan^{-1}$, not angles themselves, comparing them against $\pi/2$ confuses the input values with the angle outputs the formula produces.
    2. ✓ whether the product $2 \times 3$ is less than $1$ — (B) $2 \times 3 = 6$, and since $6 > 1$, the plain formula does not apply directly here
    3. nothing: plain formula always applies regardless of input — The condition $x y < 1$ genuinely gates this formula, here $2 \times 3 = 6$, well past $1$, so the plain formula cannot be applied directly; a correction term would be needed.
    4. whether $2 + 3$ is less than $1$ — The condition $x y < 1$ is about the product, not the sum, checking $2 + 3 < 1$ tests the wrong quantity and would never be the right gate for this formula.
Worked example

Find $\tan^{-1}(1/2) + \tan^{-1}(1/3)$

  1. $x = 1/2$, $y = 1/3$
    Name both inputs before you test the condition.
  2. $x y = 1/6 < 1$
    The condition holds, so the plain formula applies.
  3. $(x + y)/(1 - x y) = (1/2 + 1/3)/(1 - 1/6) = (5/6)/(5/6) = 1$
    Put the values into the formula and simplify.
  4. *$\tan^{-1}(1/2) + \tan^{-1}(1/3) = \tan^{-1}(1) = \pi/4$*
    $\tan^{-1}(1) = \pi/4$ is a standard value.
  5. $\tan^{-1}(1/2) + \tan^{-1}(1/3) \approx 0.4636 + 0.3218 = 0.7854 \approx \pi/4$
    Check with a calculator in radians. The two angles add to $\pi/4$ to four places.
Check yourself
  1. Compute $\tan^{-1}(1/2) + \tan^{-1}(1/3)$.
    1. $\pi/4$
    2. $5/6$
    3. $\pi/6$
    4. $\pi/2$
    Check your answer
    1. ✓ $\pi/4$ — (A) $x y = 1/6 < 1$, so $(1/2 + 1/3)/(1 - 1/6) = (5/6)/(5/6) = 1$, and $\tan^{-1}(1) = \pi/4$
    2. $5/6$ — $5/6$ is only the top of the fraction, $x + y$. The whole fraction is $(5/6)/(1 - 1/6) = 1$, and the answer still needs one more step, $\tan^{-1}(1) = \pi/4$.
    3. $\pi/6$ — $\pi/6$ is $\sin^{-1}(1/2)$’s value from an earlier atom, not the answer to this tangent addition , the two computations are unrelated.
    4. $\pi/2$ — Both $\tan^{-1}(1/2)$ and $\tan^{-1}(1/3)$ are individually smaller than $\pi/4$, their sum reaching $\pi/2$ would be far too large; the actual sum is $\pi/4$.
  2. In the worked computation of $\tan^{-1}(1/2) + \tan^{-1}(1/3)$, why is checking $x y = 1/6 < 1$ done before applying the formula, rather than after?
    1. because the check is only a formality required by convention, with no real mathematical consequence either way
    2. checking first confirms the plain formula is even valid to use, before any computation based on it is trusted
    3. because computing $x y$ afterward would give a different, incorrect numeric value
    4. the order genuinely does not matter: checking before or after makes no difference to the final answer
    Check your answer
    1. because the check is only a formality required by convention, with no real mathematical consequence either way — The check is not decorative, for $x y \geq 1$, the plain formula’s raw output is genuinely wrong without a $\pm \pi$ correction. Skipping the check risks reporting that wrong value.
    2. ✓ checking first confirms the plain formula is even valid to use, before any computation based on it is trusted — (B) Checking first confirms the plain formula is even valid to use, before any computation based on it is trusted
    3. because computing $x y$ afterward would give a different, incorrect numeric value — $1/2 \times 1/3 = 1/6$ no matter when it’s calculated, the VALUE of $x y$ is fixed. What changes is whether you’ve confirmed the formula was valid before trusting its output.
    4. the order genuinely does not matter: checking before or after makes no difference to the final answer — If the check fails ($x y \geq 1$), the plain formula’s raw output is wrong without a correction term , checking afterward risks reporting that wrong value before catching the problem.
  3. Compute $\tan^{-1}(1) + \tan^{-1}(0)$ using the addition formula.
    1. $\pi/2$
    2. $\pi/8$
    3. $\pi/4$
    4. $0$
    Check your answer
    1. $\pi/2$ — $\tan^{-1}(0) = 0$ adds nothing, the sum is just $\tan^{-1}(1) = \pi/4$, not $\pi/2$. $\pi/2$ would require the OTHER term to also be $\pi/4$-sized, which $0$ is not.
    2. $\pi/8$ — The formula computes a real sum, not an average, $(1 + 0)/(1 - 0) = 1$, and $\tan^{-1}(1) = \pi/4$, not half of that value.
    3. ✓ $\pi/4$ — (C) $x = 1$, $y = 0$, $x y = 0 < 1$, and $(1 + 0)/(1 - 0) = 1$, so $\tan^{-1}(1) = \pi/4$
    4. $0$ — $\tan^{-1}(1) = \pi/4$ is a real, nonzero contribution to the sum, it can’t be dropped just because the OTHER term happens to be zero.

↑ Back to top

Inverse trig functions in integrals

▲CONCEPT

You learn a new word in an English lesson. A year later it turns up in a science paper. You do not have to learn it again. You only have to recognise it. The inverse functions in these pages will turn up again in the same way. Their new role seems to have nothing to do with angles.

sin⁻¹ x
→
derivative, Chapter 5
→
integral, Chapter 7
→
sin⁻¹ x + C
Chapter 5 turns sin⁻¹ x into 1 over √(1 − x²); Chapter 7 turns it back. Here you only learn the name.

They come back as answers to integrals. In Chapter 5 you will find that the derivative of $\sin^{-1} x$ is $1/\sqrt{1 - x^2}$. In Chapter 7 you will run that backwards. You will integrate $1/\sqrt{1 - x^2}$ and get $\sin^{-1} x + C$.

You need no integration here.

√(1 − x²)
sin⁻¹cos⁻¹
1 + x²
tan⁻¹cot⁻¹
x√(x² − 1)
sec⁻¹cosec⁻¹
Shapes to know later: √(1 − x²) means sin⁻¹ or cos⁻¹, 1 + x² means tan⁻¹ or cot⁻¹, x√(x² − 1) means sec⁻¹ or cosec⁻¹.

A square root of $1 - x^2$ in the bottom of a fraction is a strong hint. An inverse trigonometric function is hiding behind it.

Check yourself
  1. Where do inverse trig functions reappear later in this course?
    1. only as decorative examples in later chapters, with no functional role
    2. as coefficients in later matrix problems, unrelated to their angle meaning here
    3. they don’t reappear: this chapter is the only place they are used
    4. as answers to integrals in Chapter 7
    Check your answer
    1. only as decorative examples in later chapters, with no functional role — The later appearance is a genuine functional one, $\sin^{-1} x + C$ is the actual computed answer to integrating $1/\sqrt{1 - x^2}$, not a decorative mention.
    2. as coefficients in later matrix problems, unrelated to their angle meaning here — There is no such role for inverse trig functions in matrix chapters, their genuine later role is specifically as antiderivatives in integration, Chapter 7.
    3. they don’t reappear: this chapter is the only place they are used — Chapter 7 (Integrals) shows $\sin^{-1} x$ reappearing directly as an antiderivative, this chapter is not the only place these functions are used.
    4. ✓ as answers to integrals in Chapter 7 — (D) Inverse trig functions reappear as answers to integrals, not just as functions in their own right
  2. What derivative-based fact links $\sin^{-1} x$ to a later integration result?
    1. derivative of $\sin^{-1} x$ is $1/\sqrt{1 - x^2}$
    2. the derivative of $\sin^{-1} x$ is $\sin x$ itself
    3. $\sin^{-1} x$ has no derivative, since it is only defined on a restricted branch
    4. the derivative of $\sin^{-1} x$ is $\sqrt{1 - x^2}$, without any division involved
    Check your answer
    1. ✓ derivative of $\sin^{-1} x$ is $1/\sqrt{1 - x^2}$ — (A) The derivative of $\sin^{-1} x$ is $1/\sqrt{1 - x^2}$, so integrating $1/\sqrt{1 - x^2}$ gives back $\sin^{-1} x + C$
    2. the derivative of $\sin^{-1} x$ is $\sin x$ itself — The derivative of $\sin^{-1} x$ is $1/\sqrt{1 - x^2}$, not $\sin x$, the two expressions look unrelated because they are.
    3. $\sin^{-1} x$ has no derivative, since it is only defined on a restricted branch — A restricted domain is not an obstacle to differentiability, $\sin^{-1} x$ has a perfectly well-defined derivative, $1/\sqrt{1 - x^2}$, on the open interval $(-1, 1)$.
    4. the derivative of $\sin^{-1} x$ is $\sqrt{1 - x^2}$, without any division involved — The derivative is $1/\sqrt{1 - x^2}$, a reciprocal, not $\sqrt{1 - x^2}$ alone, which drops the division and gives a completely different expression.
  3. Does this chapter teach the technique for integrating $1/\sqrt{1 - x^2}$?
    1. yes, but only for the special case $x = 0$
    2. yes: the full integration technique is worked out in this chapter’s later sections
    3. no: only the integral fact is named here
    4. the question doesn’t apply, since $1/\sqrt{1 - x^2}$ is never mentioned in this chapter at all
    Check your answer
    1. yes, but only for the special case $x = 0$ — The forward-pointer fact is stated generally, for the function $1/\sqrt{1 - x^2}$ across its whole domain, not restricted to $x = 0$ or any other single value.
    2. yes: the full integration technique is worked out in this chapter’s later sections — This chapter names the antiderivative fact but works no integration technique, that is left for Chapter 7, where the method is actually taught.
    3. ✓ no: only the integral fact is named here — (C) No integration technique is needed here, this chapter only names $\sin^{-1} x$ as a fact to recognize when it resurfaces in Chapter 7
    4. the question doesn’t apply, since $1/\sqrt{1 - x^2}$ is never mentioned in this chapter at all — $1/\sqrt{1 - x^2}$ is directly named in this chapter’s forward-pointer note, as the expression whose antiderivative is $\sin^{-1} x$, not as something unmentioned.

↑ Back to top

Restriction, and checking the branch first

■RECAP

The question at the start was this. The sine of an angle is one number, but many angles share it. Which one do you read? You read the one on the principal branch. $\sin^{-1} x$ is the angle whose sine is $x$, taken from that branch. Each function has its own branch, and on it the inverse has one answer. Each rule works only where its condition holds.

Inverse sine and inverse tangent cover the same branch, but only inverse sine keeps its ends. Inverse cosine runs from 0 to pi.
1 m, 12 m
→
rise ÷ run
→
tan⁻¹
→
≈4.8°
A ramp rising 1 m in 12 m: the ratio 1/12 goes through tan⁻¹ and comes out as one angle, about 4.8 degrees.

Both halves of the answer are the Chapter 1 idea, used on real curves. A function can be run backwards only where it is one-one and onto. The branch is where that is true. And every identity here carries a condition. Skip the check and you get a wrong answer that looks right. So fix the branch first, then check the condition, every time.

The principal branches Graphs of the inverse functions π/2 Complementary angle identities 0 0 π Negating the input, two different ways xy < 1 The addition formula for tangent inverse Inverse trig functions in integrals
Five ideas fan out from the principal branches: the graphs, the complementary pairs, negation, the addition formula and the integrals ahead.
Check yourself
  1. What are the two big ideas that run through this whole chapter?
    1. memorising six branches by rote, and computing decimal approximations quickly
    2. proving each identity from first principles, and avoiding all shortcuts
    3. fixing the branch first, then checking each identity’s condition
    4. graphing every function accurately, and using a calculator correctly
    Check your answer
    1. memorising six branches by rote, and computing decimal approximations quickly — Precision under restriction is the through-line here, fixing the branch and checking domain conditions, not rote memorisation or decimal speed.
    2. proving each identity from first principles, and avoiding all shortcuts — This reduction is the openly endorsed shortcut: reducing $\sec^{-1}$, $\operatorname{cosec}^{-1}$, $\cot^{-1}$ back to the first three via reciprocal identities, rather than demanding everything be proved from scratch each time.
    3. ✓ fixing the branch first, then checking each identity’s condition — (C) Fixing the principal branch first, and checking every identity’s domain condition before using it
    4. graphing every function accurately, and using a calculator correctly — Graphs appear here only as an illustration of a general fact; the two ideas that matter are branch-fixing and domain-condition checking, not graphing technique or calculator use.
  2. How does this chapter’s branch-restriction idea connect back to Chapter 1?
    1. Chapter 1 also restricted domains, but for an unrelated reason: to make counting problems easier
    2. it doesn’t connect to Chapter 1 at all: this chapter’s core idea is entirely new
    3. Chapter 1’s bijective-iff-invertible rule, applied to $\sin$, $\cos$, $\tan$
    4. Chapter 1 covered an entirely different topic, relations, with no bearing on invertibility at all
    Check your answer
    1. Chapter 1 also restricted domains, but for an unrelated reason: to make counting problems easier — Chapter 1’s restriction was specifically about achieving bijectivity so a function becomes invertible , not about counting problems, which is a different topic altogether.
    2. it doesn’t connect to Chapter 1 at all: this chapter’s core idea is entirely new — This connection is named directly up front, restricting the domain is how $\sin$, $\cos$, $\tan$ are MADE bijective, which is exactly Chapter 1’s own invertibility criterion applied here.
    3. ✓ Chapter 1’s bijective-iff-invertible rule, applied to $\sin$, $\cos$, $\tan$ — (C) Chapter 1’s invertible-functions rule, a function is invertible exactly when it is bijective, is applied here concretely to $\sin$, $\cos$, $\tan$
    4. Chapter 1 covered an entirely different topic, relations, with no bearing on invertibility at all — Chapter 1 covers both relations AND invertible functions, the bijective-iff-invertible criterion from its second half is exactly what this chapter’s branch-restriction idea reuses.
  3. What kind of real device or situation relies on picking one principal-value angle, the way this chapter’s inverse trig functions do?
    1. nothing outside pure mathematics: this chapter’s ideas have no real-world analogue at all
    2. any device converting a measured ratio back into one angle
    3. any device that measures an angle directly with a protractor, with no computation involved
    4. anything that only ever needs to add two numbers together
    Check your answer
    1. nothing outside pure mathematics: this chapter’s ideas have no real-world analogue at all — Concrete real-world cases exist directly, a ramp’s incline, a roof’s pitch, a GPS bearing from north, each relying on a principal-value branch to pick one angle, not infinitely many.
    2. ✓ any device converting a measured ratio back into one angle — (B) Any device converting a measured ratio back into a single angle, a ramp’s incline, a roof’s pitch, a GPS bearing from north
    3. any device that measures an angle directly with a protractor, with no computation involved — The relevant devices measure a RATIO (like rise over run for a ramp), not an angle directly, converting that ratio back into a single angle is exactly the $\sin^{-1}$/$\tan^{-1}$ computation this chapter teaches.
    4. anything that only ever needs to add two numbers together — Adding two numbers involves no ambiguity to resolve, and no branch to pick, it shares nothing with the principal-value problem this chapter’s real-world analogy is about.
Check yourself: the whole chapter
  1. $\sin(0) = \sin(\pi) = \sin(2\pi) = 0$. Many different angles share the same sine. What must happen before $\sin^{-1}$ can be defined as a function?
    1. Nothing can be done: $\sin^{-1}$ is undefined, because no function can undo one that repeats its values.
    2. $\sin^{-1} x$ is defined as the smallest non-negative angle with that sine.
    3. Sine is restricted to a branch where it is one-one, and the inverse answers only from that branch.
    4. $\sin^{-1} x$ must return every angle with that sine, as a set rather than a single value.
    Check your answer
    1. Nothing can be done: $\sin^{-1}$ is undefined, because no function can undo one that repeats its values. — Restriction is the fix, not a dead end. Once sine is confined to one branch where it is one-one, a real inverse exists.
    2. $\sin^{-1} x$ is defined as the smallest non-negative angle with that sine. — There is no smallest-non-negative rule. The fixed branch is $[-\pi/2, \pi/2]$, and it includes negative angles such as $-\pi/2$ itself.
    3. ✓ Sine is restricted to a branch where it is one-one, and the inverse answers only from that branch. — (C) Restricted to a branch where it is one-one, for sine, $[-\pi/2, \pi/2]$ , the function becomes invertible, and $\sin^{-1}$ answers only from that restricted branch.
    4. $\sin^{-1} x$ must return every angle with that sine, as a set rather than a single value. — A function must give exactly one output. Choosing a branch is exactly what makes that single answer well defined.
  2. What is $\sin^{-1}(-1)$?
    1. $-\pi/2$: the closed branch includes both endpoints.
    2. $3\pi/2$: another angle whose sine is also $-1$.
    3. Undefined: the endpoints of the branch are excluded.
    4. $\pi/2$: the angle whose sine has magnitude $1$.
    Check your answer
    1. ✓ $-\pi/2$: the closed branch includes both endpoints. — (A) $\sin$ is one-one and onto $[-1, 1]$ on the closed branch $[-\pi/2, \pi/2]$, and $\sin(-\pi/2) = -1$. Both endpoints of this branch are legal outputs.
    2. $3\pi/2$: another angle whose sine is also $-1$. — $3\pi/2$ does have sine $-1$, but it sits outside $[-\pi/2, \pi/2]$. $\sin^{-1}$ always reports the one angle inside that branch.
    3. Undefined: the endpoints of the branch are excluded. — Sine’s branch is closed, not open. Both $-\pi/2$ and $\pi/2$ are legitimate outputs of $\sin^{-1}$.
    4. $\pi/2$: the angle whose sine has magnitude $1$. — The sign carries through. $\sin^{-1}(-1) = -\pi/2$, not $\pi/2$, check the sign before writing the answer.
  3. What is $\cos^{-1}(-1)$?
    1. $-\pi$: cosine inverse can return negative angles, the same way sine inverse does.
    2. $0$: the other endpoint of the branch, paired with the input $-1$.
    3. Undefined: the domain of cosine inverse excludes its endpoints $-1$ and $1$.
    4. $\pi$: inside $[0, \pi]$, the branch used for cosine inverse.
    Check your answer
    1. $-\pi$: cosine inverse can return negative angles, the same way sine inverse does. — Cosine’s branch is $[0, \pi]$, which never dips below $0$. Only sine inverse and tangent inverse can return negative angles.
    2. $0$: the other endpoint of the branch, paired with the input $-1$. — Check which endpoint pairs with which input. $\cos(\pi) = -1$, so $-1$ pairs with $\pi$, not $0$.
    3. Undefined: the domain of cosine inverse excludes its endpoints $-1$ and $1$. — The branch $[0, \pi]$ is closed. Both $0$ and $\pi$ are legitimate outputs, matching the domain endpoints $1$ and $-1$.
    4. ✓ $\pi$: inside $[0, \pi]$, the branch used for cosine inverse. — (D) Cosine is restricted to $[0, \pi]$, and $\cos(\pi) = -1$. So $\cos^{-1}(-1) = \pi$, the right endpoint of this closed branch.
  4. Which of these can be an input to $\tan^{-1}$?
    1. Only numbers in $[-1, 1]$: the same restriction as sine inverse and cosine inverse.
    2. Any real number at all: $\tan^{-1}$ has domain $\mathbb{R}$.
    3. Only numbers in $(-\pi/2, \pi/2)$: the branch tangent inverse itself outputs into.
    4. Any real number except $0$, since $\tan(0) = 0$ would make the inverse ambiguous.
    Check your answer
    1. Only numbers in $[-1, 1]$: the same restriction as sine inverse and cosine inverse. — That restriction belongs to sine and cosine inverse, whose domain is $[-1, 1]$. Tangent ranges over every real number, so its inverse accepts every real number too.
    2. ✓ Any real number at all: $\tan^{-1}$ has domain $\mathbb{R}$. — (B) Tangent is onto all of $\mathbb{R}$ on its branch $(-\pi/2, \pi/2)$, so its inverse accepts every real number as input, unlike sine and cosine inverse, which need $[-1, 1]$.
    3. Only numbers in $(-\pi/2, \pi/2)$: the branch tangent inverse itself outputs into. — $(-\pi/2, \pi/2)$ is where $\tan^{-1}$ sends its answers, not what it accepts as input. The input side is all of $\mathbb{R}$.
    4. Any real number except $0$, since $\tan(0) = 0$ would make the inverse ambiguous. — $\tan^{-1}(0) = 0$ is a perfectly good value. There is no excluded input.
  5. Which of these numbers is NOT a valid input to $\sec^{-1}$?
    1. $-2$
    2. $2$
    3. $0$
    4. $-1$
    Check your answer
    1. $-2$ — $-2$ is accepted: $\sec^{-1}(-2)$ is a valid value, since $|-2| \geq 1$. Secant inverse does not exclude negative numbers.
    2. $2$ — $2$ is accepted, not rejected. Secant inverse rejects the interval strictly between $-1$ and $1$, everything outside it, including $2$, is valid.
    3. ✓ $0$ — (C) $0$ lies strictly between $-1$ and $1$, exactly the interval $\sec^{-1}$ rejects, since $|\sec \theta| \geq 1$ always.
    4. $-1$ — $-1$ is accepted: the domain is closed, $|x| \geq 1$, so both $-1$ and $1$ count as valid endpoints.

↑ Back to top

Where you will meet this

You have already guessed at angles from ratios without knowing the name for the step.

Your turn. A kite string is $50$ m long and the kite sits $30$ m above the ground. Which inverse trig function gives the string’s angle with the ground, and from what ratio?

Answer: The inverse sine of $30/50$, since sine is the opposite side over the hypotenuse, the string itself.

LATER

Where the critical angle goes next

First-year engineering optics and fibre-design courses compute critical and Brewster angles routinely, for every material a design uses, not just the one glass sample evaluated here. Critical angles are an inverse-sine step and Brewster angles (the angle at which reflected light becomes fully polarised) an inverse-tangent one, and the branch choice behind each still matters there.

↑ Back to top

Practice set: Exercise 2.1

Exercise 2.1 is about principal values. You evaluate all six inverse functions at standard values like $1/2$, $1/\sqrt{2}$ and $\sqrt{3}/2$, and at their negatives. Every answer must come from the principal branch.

Exercise 2.1: Principal values
  1. practice Evaluate $\sin^{-1}(1/2)$.
  2. practice Evaluate $\cos^{-1}(1)$.
  3. practice Evaluate $\tan^{-1}(-1)$.
  4. practice Evaluate $\sin^{-1}(-1)$.
  5. practice Evaluate $\cos^{-1}(-1/2)$.
  6. practice Evaluate $\tan^{-1}(\sqrt{3})$.
  7. practice Evaluate $\cot^{-1}(1)$.
  8. practice Evaluate $\sec^{-1}(-\sqrt{2})$.
  9. practice Evaluate $\operatorname{cosec}^{-1}(-2)$.
  10. practice Evaluate $\cot^{-1}(-\sqrt{3})$.
  11. practice Evaluate $\sin^{-1}(\sqrt{3}/2) + \cos^{-1}(1/2)$.
  12. practice Which of these equals $\tan^{-1}(1/\sqrt{3})$?
    1. $\pi/6$
    2. $\pi/4$
    3. $\pi/3$
    4. $\pi/2$
  13. practice Evaluate $\cos^{-1}(1/2) - \sin^{-1}(-1/2)$.
  14. practice Evaluate $\sec^{-1}(2) + \operatorname{cosec}^{-1}(2)$.
  15. practice Which of these equals $\sin^{-1}(-1) + \cos^{-1}(-1)$?
    1. $\pi$
    2. $3 \pi/2$
    3. $\pi/2$
    4. $-\pi/2$
  16. practice A ray of light inside glass (refractive index $\mu = 3/2$) strikes the glass-air boundary. Using $\sin i_c = 1/\mu$ (Physics 12, Chapter 9, Ray Optics and Optical Instruments), write the critical angle $i_c$ as an inverse sine expression, and evaluate it to one decimal place in degrees.
  17. practice A second glass sample has refractive index $\mu = 1.52$. Find its critical angle $i_c$ as an inverse sine expression, and evaluate it to two decimal places in degrees.
  18. practice Find the domain of $\sin^{-1}(2x - 1)$.
Answers
  1. $\pi/6$
  2. $0$
  3. $-\pi/4$
  4. $-\pi/2$
  5. $2 \pi/3$
  6. $\pi/3$
  7. $\pi/4$
  8. $3 \pi/4$
  9. $-\pi/6$
  10. $5 \pi/6$
  11. $2 \pi/3$
  12. (A) $\pi/6$.
  13. $\pi/2$
  14. $\pi/2$
  15. (C) $\pi/2$.
  16. $i_c = \sin^{-1}(2/3) \approx 41.8^\circ$.
  17. $i_c = \sin^{-1}(1/1.52) \approx 41.14^\circ$.
  18. $[0, 1]$

↑ Back to top

Practice set: Exercise 2.2

Exercise 2.2 practises the properties. You simplify $\sin^{-1}(\sin x)$ when $x$ is outside the branch. You also use the complementary pairs, and you add two or three inverse tangents with the formula.

Exercise 2.2: Properties
  1. practice Simplify $\sin^{-1}(\sin(2 \pi/3))$.
  2. practice Simplify $\cos^{-1}(\cos(4 \pi/3))$.
  3. practice Simplify $\tan^{-1}(\tan(3 \pi/4))$.
  4. practice Given $\sin^{-1}(x) = \pi/6$, use the complementary identity to find $\cos^{-1}(x)$.
  5. practice Given $\tan^{-1}(x) = \pi/6$, use the complementary identity to find $\cot^{-1}(x)$.
  6. practice Using $\cos^{-1}(-x) = \pi - \cos^{-1}(x)$, evaluate $\cos^{-1}(-1/\sqrt{2})$.
  7. practice Using $\sin^{-1}(-x) = -\sin^{-1}(x)$, evaluate $\sin^{-1}(-\sqrt{3}/2)$.
  8. practice Using $\cot^{-1}(-x) = \pi - \cot^{-1}(x)$, evaluate $\cot^{-1}(-1)$.
  9. practice Simplify $\cos^{-1}(\cos(7 \pi/6))$.
  10. practice Simplify $\sin^{-1}(\sin(7 \pi/6))$.
  11. practice Apply the addition formula to find $\tan^{-1}(1/5) + \tan^{-1}(2/3)$.
  12. practice Apply the addition formula to find $\tan^{-1}(2) + \tan^{-1}(-1/3)$.
  13. practice Which of these equals $\sin^{-1}(\sin(5 \pi/6))$?
    1. $5 \pi/6$
    2. $\pi/6$
    3. $-\pi/6$
    4. $\pi/3$
  14. practice Simplify $\sin^{-1}(\sin(5 \pi/4))$.
  15. practice Evaluate $\tan^{-1}(1/6) + \tan^{-1}(5/7) + \cot^{-1}(1)$.
  16. practice Prove that $\tan^{-1}(1/2) + \tan^{-1}(1/3) = \pi/4$, stating the condition under which the addition formula applies.
  17. practice Simplify $\tan^{-1}(1/2) + \tan^{-1}(1/5) + \tan^{-1}(1/8)$ by applying the addition formula twice.
Answers
  1. $\pi/3$
  2. $2 \pi/3$
  3. $-\pi/4$
  4. $\pi/3$
  5. $\pi/3$
  6. $3 \pi/4$
  7. $-\pi/3$
  8. $3 \pi/4$
  9. $5 \pi/6$
  10. $-\pi/6$
  11. $\pi/4$
  12. $\pi/4$
  13. (B) $\pi/6$.
  14. $-\pi/4$
  15. $\pi/2$
  16. Proved: $\tan^{-1}(1/2)+\tan^{-1}(1/3) = \pi/4$, since the condition $x y < 1$ holds ($1/6 < 1$).
  17. $\pi/4$

↑ Back to top

Practice set: miscellaneous

The Miscellaneous Exercise mixes the whole chapter. A typical problem reduces an angle into its branch, then uses an identity, then simplifies. Nobody tells you which identity to reach for first.

Miscellaneous: the whole chapter
  1. practice Evaluate $\sin^{-1}(\sin(5 \pi/3))$.
  2. practice Evaluate $\cos^{-1}(\cos(5 \pi/3))$.
  3. practice Evaluate $\tan^{-1}(\tan(2 \pi/3))$.
  4. practice If $\sin^{-1}(x) = \pi/5$, find $\cos^{-1}(x)$.
  5. practice If $\tan^{-1}(x) = 2 \pi/9$, find $\cot^{-1}(x)$.
  6. practice Evaluate $\sin^{-1}(-1/2) + \cos^{-1}(-1/2)$, and check it against the complementary identity.
  7. practice Apply the addition formula to find $\tan^{-1}(1/9) + \tan^{-1}(4/5)$.
  8. practice Apply the addition formula to find $\tan^{-1}(3) + \tan^{-1}(-1/2)$.
  9. practice Evaluate $\sec^{-1}(-2) + \operatorname{cosec}^{-1}(\sqrt{2})$.
  10. practice Which of these equals $\cos^{-1}(-1) - \sin^{-1}(1)$?
    1. $\pi$
    2. $\pi/2$
    3. $0$
    4. $3 \pi/2$
  11. practice Simplify $\tan^{-1}(\tan(5 \pi/6))$.
  12. practice Evaluate $\sin^{-1}(\sin(5 \pi/6)) + \cos^{-1}(\cos(5 \pi/6))$.
  13. practice Evaluate $\sin^{-1}(\sin(11 \pi/6)) + \cos^{-1}(\cos(11 \pi/6))$.
  14. practice Evaluate $\tan^{-1}(1/9) + \tan^{-1}(4/5) + \cot^{-1}(1)$.
  15. practice Evaluate $\tan^{-1}(3) + \tan^{-1}(-1/2) + \sec^{-1}(-1)$.
Answers
  1. $-\pi/3$
  2. $\pi/3$
  3. $-\pi/3$
  4. $3 \pi/10$
  5. $5 \pi/18$
  6. $\pi/2$
  7. $\pi/4$
  8. $\pi/4$
  9. $11 \pi/12$
  10. (B) $\pi/2$.
  11. $-\pi/6$
  12. $\pi$
  13. $0$
  14. $\pi/2$
  15. $5 \pi/4$
Board paper practice
  1. practice Which of these is the principal value of $\cos^{-1}(-1/2)$? (1 mark)
    1. $\pi/3$
    2. $2\pi/3$
    3. $-\pi/3$
    4. $4\pi/3$
  2. practice Assertion (A). $\sin^{-1}(\sin(3\pi/4)) = \pi/4$. Reason (R). $\sin^{-1}(\sin x) = x$ only when $x$ lies in $[-\pi/2, \pi/2]$; outside it, $\sin^{-1}(\sin x)$ returns the branch angle sharing that sine. (1 mark)
    1. Both A and R are true and R is the correct explanation of A.
    2. Both A and R are true but R is not the correct explanation of A.
    3. A is true but R is false.
    4. A is false but R is true.
  3. practice Evaluate $\sin^{-1}(1) + \cos^{-1}(1)$. (1 mark)
  4. practice Find the domain of $\cos^{-1}(x^2 - 4)$. (2 marks)
  5. practice Find the principal value of $\sin^{-1}(\sin(3\pi/5))$. (1 mark)
  6. practice A skateboard ramp rises $1.5$ m over a horizontal run of $6$ m. (i) Write the ramp’s angle as an inverse tangent expression. (1 mark) (ii) Name the branch this angle lies in, and say why. (1 mark) (iii) A second ramp keeps the same rise but has a run of only $3$ m. Write its angle too, and decide whether it is double the first ramp’s angle. (2 marks)
  7. practice Find the value of $\cos^{-1}(-1/2) + 2 \sin^{-1}(1/2)$. (2 marks)
Answers
  1. (B) $2\pi/3$.
  2. (A) Both A and R are true and R is the correct explanation of A.
  3. $\pi/2$.
  4. $[-\sqrt{5}, -\sqrt{3}] \cup [\sqrt{3}, \sqrt{5}]$.
  5. $2\pi/5$.
  6. (i) $\theta_1 = \tan^{-1}(1/4) \approx 14.0^\circ$. (ii) $(-\pi/2, \pi/2)$, since the ramp’s own angle always sits inside it. (iii) $\theta_2 = \tan^{-1}(1/2) \approx 26.6^\circ$, not double $\theta_1$.
  7. $\pi$.

↑ Back to top