Fix $\ell \geq 3$, $r \geq 2$ and $1 \leq k \leq \ell - 1$ and $z_0, \ldots, z_\ell \in \mathbb{C}$ with $z_i \neq 0$ for all $i$ and $z_i \neq z_j$ for all $i \neq j$. Now consider the (irreducible, non-homogeneous) polynomials

$q_i = z_{\ell}x_i^r - z_ix_{\ell}^r - (z_{\ell} - z_i)$ for $0 \leq i < \ell$ and

$p = z_\ell x_0^{r + k} - x_1\cdots x_k(1 - x_{\ell}^r)$

in $R = \mathbb{C}[x_1, \ldots, x_{\ell}]$. Define $I$ to be the ideal

$I = (q_1, \ldots, q_{\ell - 1}) \subset R$ and let $K$ be the field of fractions of $R/I$. Then $p$ and $q_0$ can be regarded as polynomials in $(R/I)[x_0] \cong \mathbb{C}[x_0, \ldots, x_{\ell}]/(q_1, \ldots, q_{\ell - 1})$, thus we can also regard them as polynomials in $K[x_0]$.

I have an argument that proves the irreducibility of $p$ and $q_0$ over $R/I$ and after having computed examples using Sage/Singular, I think that $p$ and $q_0$ are both irreducible over $K$ as well, but I don't see how to prove this. Proving it directly by writing it as a product in $K[x_0]$ and trying to work out what the factors could be didn't work and as $R/I$ is not a GCD domain, one can't use Gauss' lemma. Does anyone know some argument that could be applied here to deduce the irreducibility of $p$ and $q_0$ over $K$?

Note: As this is a problem that arose when trying to use the Gianni-Trager-Zacharias primality test, we can assume (by induction) that $I$ is a prime ideal in $R$. See here for the original problem.


Your Answer

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

Browse other questions tagged or ask your own question.