8
$\begingroup$

Suppose $R$ is a regular local ring of dimension at least 3, and suppose $P_1, P_2$ are dimension 1 primes. Does there necessarily exist a dimension 2 prime $Q$ contained in both?

In other words, is every pair of "little curves" contained in a "little surface"?

If this is false, will it hold if the ring is complete?

I am particularly interested in the mixed characteristic case.

$\endgroup$
2
  • $\begingroup$ Can you first show that $P_1\cap P_2\neq 0$? $\endgroup$
    – Mohan
    Commented Mar 21, 2020 at 2:35
  • 2
    $\begingroup$ Yes, since their product $P_1 P_2$ will be nonzero. $\endgroup$
    – Danny
    Commented Mar 21, 2020 at 13:01

1 Answer 1

3
$\begingroup$

I prove below that the answer is positive in the following special case: $R$ is the localization of a finitely generated polynomial ring over a field $k$. By Cohen's structure theorem this will also cover the case that $R$ is finite dimensional, complete and equicharacteristic.

For the proof it suffices to consider the case that $k$ is algebraically closed and $R$ is the localization of $k[x_1, \ldots, x_n]$ at the maximal ideal generated by $x_1, \ldots, x_n$. let $C_1, C_2$ be two irreducible curves on $k^n$ passing through the origin. It suffices to show that there is an irreducible surface $S$ containing both $C_1$ and $C_2$.

Claim: there is a curve $C$ and non-constant (and therefore dominant) maps $\phi_j: C \to C_j$, $j = 1, 2$.

Proof of the claim: Take any polynomial $f$ which is non-constant on both $C_j$. The fields $k(C_j)$ of rational functions on $C_j$ are finite extensions of $k(f)$, so that there is a common finite extension $L$ of $k(C_1)$ and $k(C_2)$ (see this for a neat trick). Now define $C$ to be the (unique) nonsingular curve such that $k(C) = L$.

It is easy to construct $S$ with $C, \phi_1, \phi_2$ from the claim: take the closure of the image of $\phi: C \times k \to k^n$ given by $\phi(x, t) := t\phi_1(x) + (1-t)\phi_2(x)$.

$\endgroup$
1
  • $\begingroup$ Thanks for that -- I am actually mostly interested in the mixed characteristic case though! I will clarify this in the question. $\endgroup$
    – Danny
    Commented Mar 22, 2020 at 20:10

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .