3
$\begingroup$

Suppose we have polynomials $f_1,\dots,f_r\in k[X_1,\dots,X_N]$ defining a complete intersection in $\mathbb{A}^N$. I suspect that it is then true that $f_1(X)+f_1(Y),\dots,f_r(X)+f_r(Y)\in k[X_1,\dots,X_N,Y_1,\dots,Y_N]$ defines a complete intersections in $\mathbb{A}^{2N}$. Is this true? If so, is there a generalization of this to commutative $k$-algebras and regular sequences?

$\endgroup$
4
$\begingroup$

This is not true. We can have $N=r+1$ and $f_i = X_{i+1} X_1 - 1$ for all $i$. Then the intersection is the locus where $X_i = X_1^{-1}$ for all $i$ from $2$ to $r+1$, thus has dimension $1$ and is a complete intersection. However $f_i(X_1,\dots,X_n) + f_i(Y_1,\dots,Y_n)$ will vanish as soon as $X_1=Y_1=0$ and has dimension $2N-2$ so i snot a complete intersection as long as $r>2$.

However, it is true for homogeneous polynomials. The reason is that by a degeneration argument, if the vanishing set of $f_1,\dots,f_r$ is a complete intersection, then the vanishing set of $f_1-c_1,\dots, f_r-c_r$ is a complete intersection for any $c_1,\dots,c_r$. Because of this, we can see that the vanishing locus of $f_i(X) + f_i(Y)$ has dimension $r+ (N-r) + (N-r) = 2N-r$ and thus is a complete intersection.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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