The Function A Robot Arm Uses To Decide How Far To Turn
A robot arm figures out how far to rotate by feeding a dot product into inverse cosine. When two directions are known as vectors, their dot product gives the cosine of the angle between them — but the arm needs the angle, not its cosine. Inverse cosine is the step that turns that ratio back into a usable angle in degrees or radians. Get the range wrong and the arm rotates to the wrong pose, which on a factory line means a missed weld or a collision.
The inverse cosine is exactly that reversal. Ordinary cosine takes an angle and returns a ratio; inverse cosine runs the machine backward, taking a ratio and returning an angle. This page frames the function as the reverse of cosine and focuses on its domain, range, and graph. For the same function approached through its notation, identities, derivative, and integral, see the companion arccosine page - same function, different entry point.
What Is The Inverse Cosine Function?
The inverse cosine function, written $\cos^{-1} x$ or $\arccos x$, is the function that undoes cosine. Formally:
$$\arccos x = \theta \quad \text{means} \quad \cos\theta = x, \quad \text{with } \theta \in [0, \pi]$$
In words: $\arccos x$ is the angle whose cosine is $x$. It is one of the inverse trigonometric functions, the family that recovers angles from ratios.
One notation warning first, because it causes real errors. The symbol $\cos^{-1} x$ means the inverse function arccosine - it does not mean $\dfrac{1}{\cos x}$. That reciprocal is secant. So $\cos^{-1} x$ is an angle, while $(\cos x)^{-1}$ is a number; the superscript $-1$ is doing two different jobs depending on where it sits.
Why The Range Must Be Restricted To The Principal Branch
Cosine is a repeating wave, so infinitely many angles share the same cosine: $\cos 60^\circ$, $\cos 300^\circ$, and $\cos 420^\circ$ are all $\dfrac{1}{2}$. If inverse cosine tried to return all of them, it would not be a function at all - a function must give exactly one output per input.
To fix this, mathematicians restrict cosine to the interval $[0, \pi]$, where it is strictly decreasing and hits each value exactly once. That chosen interval is called the principal branch, and the single angle it returns is the principal value. So $\arccos\dfrac{1}{2} = 60^\circ$ (that is, $\dfrac{\pi}{3}$), never $300^\circ$, because only $60^\circ$ lives in $[0, \pi]$.
What Are The Properties Of The Inverse Cosine Function?
The domain and range of arccosine are simply the range and domain of cosine, swapped - which is what happens to every function and its inverse.
Domain: $[-1, 1]$. You can only take the inverse cosine of a number between $-1$ and $1$, because those are the only values cosine ever outputs. $\arccos 2$ does not exist.
Range: $[0, \pi]$, the principal branch. Every output angle lies between $0$ and $\pi$ (that is, $0^\circ$ to $180^\circ$), inclusive.
Decreasing: arccosine falls steadily from left to right, starting at $\arccos(-1) = \pi$ and ending at $\arccos(1) = 0$.
Not odd, not even. Unlike arcsine, arccosine has no origin or axis symmetry. Instead it obeys $\arccos(-x) = \pi - \arccos x$.
Complementary with arcsine: $\arccos x + \arcsin x = \dfrac{\pi}{2}$ for every $x$ in $[-1, 1]$. This links it to arcsin and to the cofunction identities.
Key Inverse Cosine Values
These principal values recur constantly and are worth knowing cold.
$x$ | $-1$ | $-\dfrac{1}{2}$ | $0$ | $\dfrac{1}{2}$ | $1$ |
|---|---|---|---|---|---|
$\arccos x$ | $\pi$ | $\dfrac{2\pi}{3}$ | $\dfrac{\pi}{2}$ | $\dfrac{\pi}{3}$ | $0$ |
Reading the table left to right, the output angle shrinks as $x$ grows - a direct picture of arccosine being a decreasing function.
How Do You Graph The Inverse Cosine Function?
The graph of $y = \arccos x$ is the reflection of the restricted cosine curve across the line $y = x$, which is how every inverse function relates to its original. Here is how to build it.
Start with cosine on $[0, \pi]$. Sketch $y = \cos x$ over just that interval - it falls from $(0, 1)$ down to $(\pi, -1)$. This is the principal branch.
Reflect across $y = x$. Swapping the roles of input and output turns that falling cosine arc into the arccosine curve. The point $(0, 1)$ on cosine becomes $(1, 0)$ on arccosine; the point $(\pi, -1)$ becomes $(-1, \pi)$.
Read the endpoints. The finished curve runs from $(-1, \pi)$ on the left, through $(0, \frac{\pi}{2})$ in the middle, to $(1, 0)$ on the right. It is smooth, continuous, and always decreasing.
Because the domain is only $[-1, 1]$, the graph is a single finite arc - it does not extend left or right forever the way many function graphs do. That bounded shape is the visual signature of an inverse trig function. You can also read every one of these values off the unit circle, where the angle in $[0, \pi]$ is the upper half.
Where Is Inverse Cosine Used? - "Recovering an angle from a known ratio"
Inverse cosine exists to answer one question: you know a cosine value, so what was the angle? That is the recurring need whenever an angle has to be reconstructed from measured or computed data.
Angle between two vectors. In robotics, physics, and computer graphics, the dot product gives $\cos\theta = \dfrac{\mathbf{a}\cdot\mathbf{b}}{|\mathbf{a}||\mathbf{b}|}$, and inverse cosine recovers $\theta$. This is how a game engine computes lighting angles and how a robot arm decides its rotation.
Navigation and surveying. When a distance and a projection are known, the bearing angle often comes out as an arccosine.
Solving triangles with the law of cosines. After the law of cosines gives $\cos C$ in terms of the three sides, inverse cosine turns that ratio into the actual angle $C$.
What most explainers gloss over is why the answer is capped at $180^\circ$. It is not an arbitrary ceiling - it is the principal branch $[0, \pi]$ doing its job. Any angle beyond that range shares its cosine with an angle inside it, and arccosine always returns the one inside. For a formal reference on the analytic properties and series, Wolfram MathWorld's inverse cosine entry collects them.
Examples Of The Inverse Cosine Function
Example 1
Evaluate $\arccos\left(\dfrac{1}{2}\right)$.
Ask: which angle in $[0, \pi]$ has cosine $\dfrac{1}{2}$? That angle is $\dfrac{\pi}{3}$ (which is $60^\circ$):
$$\arccos\left(\frac{1}{2}\right) = \frac{\pi}{3}$$
Final answer: $\arccos\left(\dfrac{1}{2}\right) = \dfrac{\pi}{3}$.
Example 2
Evaluate $\arccos\left(\cos\dfrac{4\pi}{3}\right)$. First instinct, then the correct route.
The tempting move is to say inverse cosine cancels cosine, so the answer is just the inside angle: $\dfrac{4\pi}{3}$.
Check it against the range. Arccosine only ever returns a value in $[0, \pi]$, and $\dfrac{4\pi}{3}$ is larger than $\pi$, so it cannot possibly be the answer. The cancellation shortcut fails whenever the inside angle sits outside the principal branch.
Do it properly. First find $\cos\dfrac{4\pi}{3} = -\dfrac{1}{2}$. Then ask which angle in $[0, \pi]$ has cosine $-\dfrac{1}{2}$:
$$\arccos\left(-\frac{1}{2}\right) = \frac{2\pi}{3}$$
Final answer: $\arccos\left(\cos\dfrac{4\pi}{3}\right) = \dfrac{2\pi}{3}$, not $\dfrac{4\pi}{3}$.
Example 3
Find the angle $\theta$ in a right triangle where the adjacent side is $2$ and the hypotenuse is $4$.
Cosine is adjacent over hypotenuse, so $\cos\theta = \dfrac{2}{4} = \dfrac{1}{2}$. Recover the angle with inverse cosine:
$$\theta = \arccos\left(\frac{1}{2}\right) = \frac{\pi}{3} = 60^\circ$$
Final answer: $\theta = 60^\circ$.
Example 4
Evaluate $\arccos(-1)$.
Which angle in $[0, \pi]$ has cosine $-1$? Cosine reaches $-1$ only at $\pi$:
$$\arccos(-1) = \pi$$
Final answer: $\arccos(-1) = \pi$ (that is, $180^\circ$).
Example 5
Use the identity $\arccos x + \arcsin x = \dfrac{\pi}{2}$ to find $\arcsin\left(\dfrac{1}{2}\right)$, given $\arccos\left(\dfrac{1}{2}\right) = \dfrac{\pi}{3}$.
Rearrange the identity:
$$\arcsin\left(\frac{1}{2}\right) = \frac{\pi}{2} - \arccos\left(\frac{1}{2}\right) = \frac{\pi}{2} - \frac{\pi}{3} = \frac{\pi}{6}$$
Final answer: $\arcsin\left(\dfrac{1}{2}\right) = \dfrac{\pi}{6}$.
Example 6
Find the angle between the vectors $\mathbf{a} = (1, 0)$ and $\mathbf{b} = (1, 1)$.
The dot product is $\mathbf{a}\cdot\mathbf{b} = (1)(1) + (0)(1) = 1$. The magnitudes are $|\mathbf{a}| = 1$ and $|\mathbf{b}| = \sqrt{2}$. So:
$$\cos\theta = \frac{\mathbf{a}\cdot\mathbf{b}}{|\mathbf{a}||\mathbf{b}|} = \frac{1}{\sqrt{2}}$$
Recover the angle:
$$\theta = \arccos\left(\frac{1}{\sqrt{2}}\right) = \frac{\pi}{4} = 45^\circ$$
Final answer: the vectors meet at $45^\circ$.
The first-instinct error students reach for across these is assuming inverse cosine cancels cosine no matter what - it only returns the angle inside $[0, \pi]$, so any inside angle beyond that range gets folded back in.
Common Mistakes With The Inverse Cosine Function
Mistake 1: Treating $\cos^{-1} x$ As One Over Cosine
Where it slips in: Reading the $-1$ superscript as a reciprocal, the way it works in algebra.
Don't do this: Writing $\cos^{-1} x = \dfrac{1}{\cos x}$.
The correct way: $\cos^{-1} x$ is the inverse function arccosine, an angle. The reciprocal $\dfrac{1}{\cos x}$ is secant. The memorizer who learned "$-1$ means flip it" from exponents carries a rule that does not apply to inverse function notation.
Mistake 2: Assuming Inverse Cosine Cancels Cosine Every Time
Where it slips in: Simplifying $\arccos(\cos\theta)$ when $\theta$ is outside $[0, \pi]$.
Don't do this: Writing $\arccos(\cos 240^\circ) = 240^\circ$.
The correct way: Arccosine returns only the principal value in $[0, \pi]$. Since $240^\circ$ is outside that range but shares its cosine with $120^\circ$, the answer is $120^\circ$. The rusher who cancels without checking the range gets a value the function can never output.
Mistake 3: Forgetting The Domain Is Only $[-1, 1]$
Where it slips in: Feeding a number bigger than $1$ into inverse cosine.
Don't do this: Trying to evaluate $\arccos(1.5)$.
The correct way: Cosine never exceeds $1$, so $\arccos(1.5)$ has no answer - it is undefined. The second-guesser who trusts a calculator's error message here is right; the input is genuinely out of domain.
Key Takeaways
The inverse cosine $\arccos x$ (or $\cos^{-1} x$) returns the angle in $[0, \pi]$ whose cosine is $x$.
Its domain is $[-1, 1]$ and its range is the principal branch $[0, \pi]$.
The graph is a single decreasing arc from $(-1, \pi)$ to $(1, 0)$, the reflection of cosine on $[0, \pi]$ across $y = x$.
$\cos^{-1} x$ is the inverse function, not $\dfrac{1}{\cos x}$ (which is secant).
It obeys $\arccos x + \arcsin x = \dfrac{\pi}{2}$ and is used to recover angles from ratios.
To take inverse cosine further with a teacher, explore Bhanzu's trigonometry tutor sessions, a high school math tutor for graphing practice, or live math classes online with peers from 20+ countries.
A Practical Next Step
Practice these to solidify your understanding: evaluate $\arccos(0)$ and $\arccos\left(-\dfrac{1}{2}\right)$ from the unit circle, then simplify $\arccos(\cos 200^\circ)$ by folding it into $[0^\circ, 180^\circ]$. If a value lands outside the range, come back to the principal-branch explanation above. Want a live Bhanzu trainer to work these through with you? Book a free demo class.
Read More
Arctan — the inverse tangent, with its own principal branch and range.
Trigonometric Functions — the primary functions inverse cosine reverses.
Sin Cos Tan — the three ratios and how inverses recover their angles.
Cos2x — a cosine identity that pairs with inverse-cosine problems.
Trigonometric Table — the special-angle values behind every arccosine result.
Was this article helpful?
Your feedback helps us write better content