1
$\begingroup$

Given a field of characteristic not 2 or 3 containing a primitive third root of unity, is it true that every nonsingular cubic curve, i.e. a curve defined by one homogeneous form of degree 3 in 3 variables, equivalent (up to a linear change of variables) to something of the form $X^3+b Y^3+c Z^3=d X Y Z$? If not then what is the necessary and sufficient condition for the cubic curve to be equivalent to something of this form? What is the canonical form for cubic curves?

$\endgroup$
1
  • 2
    $\begingroup$ There is no canonical form for cubic curves because the moduli space of cubic curves is not a scheme. $\endgroup$
    – Will Sawin
    Dec 10, 2014 at 22:14

1 Answer 1

3
$\begingroup$

Over the complex numbers, there are several normal forms. A ternary cubic can be written as $$ \lambda XYZ-(X+Y+Z)^3 $$ with a parameter $\lambda$. This normal form is invariant under the symmetric group $S_3$ of permutations of variables. This leads to the so-called Hesse normal form, given by $$ X^3+Y^3+Z^3-3\mu XYZ. $$ It has the property that the Hessian of any of its members is of the same form.

Hesse showed that the inflection points of a ternary cubic over the rationals are defined over a solvable extension. As a result, any ternary cubic can be brought to the Hesse normal form over a solvable extension of the base field. More precise details can be found in the reference: Ternary Cubic Forms and Etale Algebras, by Raczek and Tignol.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.