Yankel said:
I see...
but why did they do it in the first place ?
Am I wrong that the cubic root of 1 is 1 ?
What about the quarter root of 1 ?
You're not wrong, but you're not entirely right, either.
First, what do we MEAN by: "a cube root of 1"?
We mean some "number" $x$ such that: $x^3 = 1$. Clearly, 1 works, since: $1^3 = 1$.
Another way to phrase this is:
$x^3 - 1 = 0$
Now, $x^3 - 1 = (x - 1)(x^2 + x + 1)$.
Taking $x = 1$, we see the left factor is 0, so the whole product is 0.
But what if $x \neq 1$? Could it be possible that $x^2 + x + 1 = 0$?
Well, if we use the quadratic formula, with $a = b = c = 1$, we obtain:
$x = \dfrac{-1 \pm \sqrt{1^2 - 4(1)(1)}}{2} = \dfrac{-1 \pm \sqrt{-3}}{2}$
which can be written in the form:
$x = -\dfrac{1}{2} \pm i\dfrac{\sqrt{3}}{2}$.
Note that this is:
$x = \cos\left(\dfrac{2\pi}{3}\right) \pm i\sin\left(\dfrac{2\pi}{3}\right)$
in other words, in the complex plane the "other two cube roots of 1" lie at the angles:
1/3 around the circle, and 2/3 (-1/3) around the unit circle.
You can verify, by direct computation, that if:
$\omega = -\dfrac{1}{2} + i\dfrac{\sqrt{3}}{2}$
that $\omega^2 = \overline{\omega}$, and also that:
$x^2 + x + 1 = (x - \omega)(x - \omega^2)$
(remember, complex solutions to a REAL quadratic come in conjugate-pairs).
The situation is quite analogous for the polynomial:
$x^n - 1$
the roots are:
$\cos\left(\dfrac{2k\pi}{n}\right) + i\sin\left(\dfrac{2k\pi}{n}\right)$
for $k = 0,1,2,\dots,n$.
For $n = 4$ (the fourth roots of 1), we get:
for $k = 0,\ \cos(0) + i\sin(0) = 1 + i0 = 1$.
for $k = 1,\ \cos\left(\dfrac{\pi}{2}\right) + i\sin\left(\dfrac{\pi}{2}\right) = 0 + i1 = i$
for $k = 2,\ \cos(\pi) + i\sin(\pi) = -1 + i0 = -1$
for $k = 3.\ \cos\left(\dfrac{3\pi}{2}\right) + i\sin\left(\dfrac{3\pi}{2}\right) = 0 +i(-1) = -i$.
Indeed, we have: $x^4 + 1 = (x^2 + 1)(x^2 - 1) = (x + i)(x - i)(x + 1)(x - 1)$.
It turns out there is a DEEP connection between $n$-th roots of a number, and $\frac{1}{n}$-th of a circle. The geometrical reason for this is that complex multiplication is "part stretching" and "part rotating".
The circle is a profound mathematical object. I cannot stress this enough. In mathematics we have two "big ideas": the line, and the circle. The extrapolation of these two simple things, leads to a vast array of interesting structures.