A forward trig function takes an angle and returns a ratio: sin30°=0.5. The inverse runs backward: arcsin0.5=30°. But sine is not one-to-one. On the unit circle, the height 0.5 is reached at 30° and again at 150°, and every full turn repeats both. A function must give a single output, so arcsin reports only angles in [−90°,90°], the right half of the circle, where each height from −1 to 1 appears exactly once.
Each inverse gets a range on which its trig function takes every value once. Sine covers −1 to 1 on [−90°,90°]; cosine does the same on [0°,180°], the top half; tangent covers every real number on (−90°,90°). So arccos(−0.5)=120°, in the top half, while arcsin(−0.5)=−30°, in the right half. One consequence: arcsin(sin150°)=arcsin0.5=30°, not 150°, because 150° lies outside arcsine's range.
The notation sin−1x means the inverse function, not sinx1. The reciprocal of sine is cosecant, cscx=sinx1, which still takes an angle in: csc30°=2. The inverse takes a ratio in: sin−1(0.5)=30°. Look at the input: if it is a ratio and the answer should be an angle, you want the inverse.
Think of it like this. If sine is a machine that takes the tilt of a ramp and reports how high it rises per metre of ramp, arcsine is the reverse machine: give it the rise per metre and it names the tilt. Ramp tilts lie between −90° and 90°, exactly the range arcsine reports; angles outside it have to be found by symmetry.
Core idea. arcsin, arccos, arctan undo the trig functions on a restricted range so the answer is a single angle.