0
$\begingroup$

Let $k$ be a field of characteristic zero and $R$ be the polynomial ring $k[x_1,...,x_n,t_1,...,t_n]$. Let $P_i = (a_{i1}:a_{i2}:a_{i3})$ be $n$ points in the projective plane over $k$, such that not all of them lie on a line. Now define $t = t_1\cdots t_n$, $w_i = x_i t t_i^{-1}$ for $1 \leq i \leq n$, and $z_j = \sum_{i=1}^n a_{ij} w_i$ for $1\leq j \leq 3$. I want to show that $z_1,z_2,z_3,$ and $t$ are algebraically independent over $k$. Note that since the points $P_i$ do not all lie on a line, we have that $z_1,z_2,z_3$ are linearly independent.

Given any relation $F(z_1,z_2,z_3,t) = 0$ with coefficients in $k$ (i.e. a polynomial $F \in k[X_1,X_2,X_3,X_4]$) we can obtain another relation $F'(x_1,...,x_n,t_1,...,t_n) = 0$ with coefficients in $k$, by using the above definitions of $z_1,z_2,z_3,$ and $t$ in terms of the variables in $R$. Since these variables are algebraically independent over $k$, the coefficients of $F'$ vanish. But these coefficients are polynomials in the coefficients of our original relation $F$. Can we use this to show that the coefficients of $F$ vanish? Or, more generally, do you think this is the right way to proceed?

It seems that in order to push this argument further I need to determine the specific relationship between the coefficients of $F'$ and those of $F$, but this would be messy.

$\endgroup$
2
  • $\begingroup$ At the very least, you need some conditions on the $a_{ij}$. As an extreme case, if all of the $a_{ij}=0$, then all $z_j=0$. Or if the $a_{ij}$ are chosen appropriately, you could have the $z_j$ being $k$-linear combinations of one another. So at the least, you'll want to assume that the matrix $(a_{ij})$ has rank 3 in $\text{Mat}_{3\times n}(k)$. Equivalently, assume that $z_1,z_2,z_3$ are $k$-linearly independent. $\endgroup$ Commented Jul 14, 2015 at 0:42
  • 1
    $\begingroup$ @JoeSilverman, I've added the condition that the $P_i$ do not all lie on a line, which guarantees that the matrix $(a_{ij})$ has rank 3, as you suggested. Thanks for the comment; I really enjoyed The Arithmetic of Elliptic Curves. $\endgroup$ Commented Jul 14, 2015 at 17:41

0

You must log in to answer this question.

Browse other questions tagged .