The problem: trigonometric functions are not invertible
From Chapter 1, only a one-one and onto function has an inverse. The six trigonometric functions fail spectacularly: sin:R→[−1,1] repeats its values endlessly (sin0=sinπ=sin2π=0), so it is many-one — no inverse exists on the full domain.
The fix is the same trick that rescued x2: restrict the domain until the function becomes bijective. Restricted to [−2π,2π], the sine function climbs steadily from −1 to 1, hitting each value exactly once — one-one and onto [−1,1], hence invertible.
Branches and the principal value branch
Sine restricted to [−23π,−2π], or [2π,23π], or any such interval is also bijective onto [−1,1] — each choice produces a different branch of the inverse. To make sin−1 a single well-defined function, one branch is fixed as the principal value branch:
sin−1:[−1,1]→[−2π,2π]
The same story, told six times, gives the whole table:
Function
Domain
Range (principal value branch)
sin−1x
[−1,1]
[−2π,2π]
cos−1x
[−1,1]
[0,π]
tan−1x
R
(−2π,2π)
cot−1x
R
(0,π)
sec−1x
R−(−1,1)
[0,π]−{2π}
cosec−1x
R−(−1,1)
[−2π,2π]−{0}
How to remember the ranges:sin−1,tan−1,cosec−1 (the "odd" trio) live in the symmetric window around 0; cos−1,cot−1,sec−1 live in the window [0,π]. Open vs closed endpoints follow the original function: tan blows up at ±2π (open), sec is undefined at 2π (deleted point).
Graphs: reflection in y=x
As for any inverse, the graph of y=sin−1x is the mirror image of the restricted y=sinx in the line y=x — interchange the axes and the point (a,b) becomes (b,a). The figure shows the principal branch (dark) and its reflection.
Notation warning (a guaranteed 1-mark trap):sin−1x is the inverse function, NOT sinx1. The reciprocal is written (sinx)−1. The alternative name arcsinx avoids the ambiguity entirely.
Computing Principal Values
The template
The principal value of an inverse trigonometric expression is the one value lying in the principal range. To find sin−1(−21):
Step 1 — set the equation: let y=sin−1(−21), so siny=−21.
Step 2 — recall the reference angle:sin6π=21, so sin(−6π)=−21.
Step 3 — check the branch:−6π∈[−2π,2π] ✓.
Answer:sin−1(−21)=−6π. Among the infinitely many angles whose sine is −21, the principal branch selects exactly this one.
Negative inputs: where each function puts them
The branch geometry decides how minus signs come out:
sin−1,tan−1,cosec−1 of a negative number are negative(in [−2π,0))cos−1,cot−1,sec−1 of a negative number lie in (2π,π](second quadrant)
So tan−1(−1)=−4π but cos−1(−21)=π−3π=32π — the odd trio flips the sign, the even trio subtracts from π.
For cosec−1 and sec−1, convert to the reciprocal first if it helps: cosec−1(2) asks for the angle whose cosecant is 2, i.e. whose sine is 21.
Combining principal values
Sums of principal values are evaluated term by term:
tan−1(1)+cos−1(−21)+sin−1(−21)=4π+32π−6π=123π+8π−2π=43πcos−121+2sin−121=3π+2⋅6π=32π
Common mistakes to avoid
Mistake 1 — giving a non-principal answer.sin−121=65π? No: 65π has the right sine but lies outside [−2π,2π]. Only 6π qualifies.
Mistake 2 — flipping the sign for cos−1.cos−1(−21) is NOT −3π — the range [0,π] contains no negative angles. It is π−3π=32π.
Mistake 3 — reading sin−1x as sinx1. The notation trap; see the warning above.
Mistake 4 — domain slips.sin−1(2) does not exist (2∈/[−1,1]), and sec−1(21) does not exist either (21 lies in the excluded gap (−1,1)).
Solved Examples
Example 1 — The model principal value
Find the principal value of sin−1(21).
Step 1 — set up: let y=sin−121, so siny=21.
Step 2 — reference angle:sin4π=21, and 4π∈[−2π,2π] ✓.
Answer:4π.
Example 2 — A second-quadrant cotangent
Find the principal value of cot−1(−31).
Step 1 — set up:coty=−31 with y required in (0,π).
Step 2 — negative cotangent puts y in the second quadrant:cot3π=31, so
y=π−3π=32π
Step 3 — check:32π∈(0,π) ✓.
Answer:32π — for the [0,π]-family, negative inputs mean "subtract the reference angle from π", never "attach a minus sign".
Example 3 — The odd trio with negatives
Find the principal values of (i) tan−1(−3), (ii) cosec−1(−2).
Step 1 — (i):tan3π=3, and tan−1 of a negative is negative: −3π∈(−2π,2π) ✓.
Step 2 — (ii):cosec4π=2, so the answer is −4π, legal in the cosecant branch.
Answer: (i) −3π, (ii) −4π.
Example 4 — Reciprocal conversion for sec and cosec
Find the principal value of sec−1(32).
Step 1 — convert:secy=32 means cosy=23.
Step 2 — reference angle:cos6π=23, and 6π∈[0,π]−{2π} ✓.
Answer:6π.
Example 5 — A three-term combination
Evaluate tan−1(1)+cos−1(−21)+sin−1(−21).
Step 1 — each principal value separately:tan−1(1)=4π,cos−1(−21)=π−3π=32π,sin−1(−21)=−6π
Step 2 — add over denominator 12:123π+128π−122π=129π=43π
Answer:43π — note how the same input −21 produced a second-quadrant answer under cos−1 and a negative answer under sin−1.
Example 6 — Another combination
Evaluate cos−1(21)+2sin−1(21).
Step 1 — principal values:cos−121=3π and sin−121=6π.
Step 2 — combine:3π+2⋅6π=3π+3π=32π.
Answer:32π.
Example 7 — Reading domains
Which of the following exist? (i) sin−1(1.5), (ii) sec−1(0.8), (iii) tan−1(1000), (iv) cot−1(−5).
Step 1 — check each against the domain table: (i) 1.5∈/[−1,1] — does not exist. (ii) 0.8∈(−1,1), the excluded gap for sec−1 — does not exist. (iii) tan−1 accepts all reals — exists (a value just under 2π). (iv) cot−1 accepts all reals — exists (a second-quadrant value).
Answer: only (iii) and (iv) exist. The domains [−1,1] (for sin−1,cos−1) and R−(−1,1) (for sec−1,cosec−1) are complementary — a number in the open interval (−1,1) feeds the first pair only, a number beyond it feeds the second pair only, and ±1 feed all four.
Ready to test your knowledge?
Take a quick interactive quiz on this topic —
free, works without login.