Cardan's Cubic Formula
/Back when I taught university math, students would remember the quadratic formula and then naturally ask if there was a cubic formula, and how that works. Well here goes:
We start with the general cubic equation ax3+bx2+cx+d=0 and clean it up a bit:
x^3+b x^2+c x+d=0,\text
where b/a,c/a, and d/a have been replaced with b, c, and d, respectively. Keep in mind the fact that x3+bx2+cx+d is a monic polynomial
and that reducing to the monic case has no effect on the roots. Now, multiplying by z3we conclude that y3+py+q=0 is equivalent to the
equation
z6+qz3−=0.
This equation is called the cubic resolvent of the reduced cubic y3+py+q=0, and this can be written as
(z3)2+qz3−=0.
By the quadratic formula we obtain
z^3=\frac\left(-q\pm \sqrt\right)
so that
\left.z=\sqrt[3]\left(-q\pm \sqrt\right) \\
Substituing this gives a root of the reduced cubic y3+py+q, and then x=y−b/3 is a root of the cubic x3+bx2+cx+d.
Question: is this derivation correct, and if not why? By setting y3+py+q=0, we essentially assumed that a solution exists. What justifies this assumption? A cubic equation has three roots, yet the cubic resolvent has degree 6. Why? We assumed z≠0, what happens when z=0?