What are the inverse trigonometric functions, and what are their domains, ranges and graphs?
Define and sketch the inverse trigonometric functions arcsin, arccos and arctan, including their domains and ranges
A focused answer to the HSC Maths Extension 1 dot point on inverse trigonometric functions. Restricted domains for sin, cos and tan to define arcsin, arccos and arctan, the stage-by-stage construction of the arcsin graph by reflection, their domains, ranges, exact values and identities, with worked examples.
✦ Generated by Claude Opus 4.8·16 min answer·
Reviewed by: AI editorial process; not yet individually human-reviewed
NESA wants you to know how the inverse trigonometric functions are defined as inverses of suitably restricted trig functions, to state and sketch their principal-value graphs (with domains and ranges), and to evaluate or simplify expressions involving them. The crucial idea, and the one most students rush past, is that sin, cos and tan are not one-to-one, so an inverse only exists once we cut each function down to a single rising or falling branch. Everything else, the ranges, the graphs, the "why arcsin(sinx) is not always x" trap, follows from that single decision.
The answer
Why we have to restrict the domain first
A function has an inverse only if it is one-to-one: each output must come from exactly one input. But sinx, cosx and tanx are periodic, so a single output value comes from infinitely many inputs. Geometrically, a horizontal line cuts the sine curve over and over.
To get an inverse we choose a principal branch: a largest interval on which the function is strictly increasing or strictly decreasing (hence one-to-one) and still hits every output value once. For sine, the natural choice is [−2π,2π].
The inverse is then the reflection of that branch in the line y=x (reflecting any function in y=x swaps the roles of x and y, which is exactly what taking an inverse does). Reflecting the restricted sine branch produces the arcsin curve.
Reading the reflected curve on its own gives the graph you must be able to sketch from memory: the domain is what was the range of the restricted sine ([−1,1] on the x-axis) and the range is what was the domain ([−2π,2π] on the y-axis).
The vertical tangents at the endpoints x=±1 are the mirror image of sine's horizontal tangents at its peak and trough: reflection turns a flat tangent into an upright one. That is why arcsin rises ever more steeply as x→±1.
arcsin (also written sin−1)
Restrict sinx to x∈[−2π,2π], where it increases from −1 to 1. The inverse arcsin then maps [−1,1] onto [−2π,2π]. The graph passes through (−1,−2π), (0,0) and (1,2π); it is an odd function (it has 180∘ rotational symmetry about the origin), strictly increasing, with vertical tangents at the endpoints.
arccos (also written cos−1)
Restrict cosx to x∈[0,π], where it decreases from 1 to −1. The inverse arccos maps [−1,1] onto [0,π].
The graph passes through (−1,π), (0,2π) and (1,0), is strictly decreasing, and has vertical tangents at the endpoints. Unlike arcsin, it is neither even nor odd: it has point symmetry about (0,2π), which is the geometric content of the identity arccos(−x)=π−arccosx.
arctan (also written tan−1)
Restrict tanx to x∈(−2π,2π), where it increases from −∞ to +∞. The inverse arctan maps all of R onto (−2π,2π).
The graph passes through the origin, is odd, strictly increasing, and approaches the horizontal asymptotesy=±2π without ever reaching them. The asymptotes are the reflection of tangent's vertical asymptotes at x=±2π: reflecting in y=x turns a vertical asymptote into a horizontal one. Note the domain is all real numbers, so unlike arcsin and arccos there is no "out of domain" restriction to check on the argument of arctan.
The identities you are expected to know
The complementary identity links arcsin and arccos:
arcsinx+arccosx=2π,x∈[−1,1].
It is true because the two angles are the acute angles of the same right triangle (one has sine x, the other has cosine x), and they add to a right angle. For symmetric (negated) arguments:
The reverse order only returns x when x is already inside the principal range:
arcsin(sinx)=xonly when x∈[−2π,2π].
For an x outside that interval you must first reduce sinx using periodicity or a reflection identity, then take arcsin of the resulting in-range value. This is the single most-tested subtlety on the topic.
Composing across different functions: the right-triangle move
To simplify something like tan(arccosx) or cos(arcsinx), set θ equal to the inverse-trig part, draw a right triangle (or use a Pythagorean identity) for the known ratio, and read off the ratio you want, keeping the sign correct for the principal range. The worked examples below do this twice.
How exam questions ask about inverse trig functions
"Find the exact value of arcsin(…) / arccos(…) / arctan(…)": identify the unique angle in the principal range with that ratio. State the range you are working in to justify the sign.
"State the domain (and range) of f(x)=arcsin(expression)": set the inside between −1 and 1 and solve; for arctan the domain is all reals. The range of a transformed inverse function shifts and scales the base range.
"Sketch y=arccosx (or a transformation of it)": draw the base curve through its three key points, then apply shifts and stretches. Mark endpoints and asymptotes.
"Evaluate arcsin(sinθ) for a given θ": a principal-range trap. Reduce the inner value first if θ is outside the range.
"Simplify cos(arcsinx) / tan(arccosx)": the right-triangle move, watching the sign.
"Show that arctana+arctanb=…": the arctan addition formula, with the ±π correction when ab>1.
Edge cases worth knowing
Endpoints are included for arcsin and arccos but the range of arctan is open.arcsin(±1)=±2π and arccos(∓1)=0 or π are defined, but arctan never actually outputs ±2π (those are asymptotes).
arccos is not odd. The reflex arccos(−x)=π−arccosx (not −arccosx) trips up students who assume all inverse trig functions behave like arcsin.
Transformed graphs shift the range, not just the curve. For y=2arctanx the range becomes (−π,π) with asymptotes at y=±π; for y=arcsinx+2π the range becomes [0,π].
Exam-style practice questions
Practice questions written in the style of NESA exam questions on this dot point, with worked answer explainers. The year tag is the paper they imitate, not the source.
2021 HSC Q42 marksFind the exact value of arccos(−21)+arcsin(23).
Show worked answer →
arccos(−21)=32π because cosine of 32π equals −21 and the principal range of arccos is [0,π].
arcsin(23)=3π because sine of 3π equals 23 and the principal range of arcsin is [−2π,2π].
Sum: 32π+3π=π.
Markers reward stating the principal-value ranges, evaluating each inverse separately, and combining to a clean exact answer.
2023 HSC Q21 marksWhat is the domain of the function f(x)=arcsin(2x−1)?
Show worked answer →
For arcsinu to be defined, −1≤u≤1.
So −1≤2x−1≤1.
Add 1 throughout: 0≤2x≤2.
Divide by 2: 0≤x≤1.
Domain: [0,1].
Markers reward the constraint on the inner argument, the algebra to isolate x, and the correct interval.