6
$\begingroup$

Brieskorn (1970) showed that for semiuniversal deformation of rational double points surface singularities $X\to S$, there is a finite base change $S'\to S$, such that the new family $f:X\times_{S}S'\to S'$ admit simultaneous resolution, i.e., there is a a smooth variety $W$ and commutative diagram $\require{AMScd}$ \begin{CD} W @>{\sigma}>> X\times_SS'@>>>X\\ @VV{\sigma}V @VVfV@VVV\\ S' @=S'@>>>S \end{CD} ​ where $\sigma$ is a birational morphism which desingularizes each fiber of $f$ and $\pi$ is a smooth morphism. Moreover, Brieskorn showed that the base change is identified with Weyl group action on Cartan subalgebra. (A good reference would be P. Slodowy's Four Lectures on simple groups and singularities.)​ ​

For example, semiuniversal deformation of $A_2$ singularity is $$\{x^2+y^2+z^3+tz+s=0\}\to \mathbb C^2, \ (x,y,z,t,s)\mapsto (t,s),\tag{1}\label{1}$$ with the discriminant locus the cuspidal curve $27s^2+4t^3=0$. Now, permutation action on the hyperplane $\{u+v+w=0\}\subset \mathbb C^3$ gives a $6:1$ map $$\{u+v+w=0\}\to \mathbb C^2, \ (u,v,w)\mapsto (uv+uw+vw,uvw).$$ Brieskon's theorem says base change of $(\ref{1})$ along this $6:1$ cover, the total family admits a simultaneous resolution.

I'd like to know if it is possible to generalize this construction to a family with proper base.

More precisely, assume $T$ is a smooth projective variety and $Y\to T$ is a flat family of algebraic surfaces with discrimant locus (let's say a divisor) $D\subset T$, and the singularity on each fiber is at worst rational double point. Moreover, suppose for each point $x\in D$, there is a neighborhood $U$ of $x$ in $T$ such that the restricted family $Y_U\to U$ is semiuniversal deformation of rational double point singularities (cross some trivial directions).​ ​

Question: Is there a finite base change $T'\to T$ such that $Y\times_TT'\to T'$ admits simultaneous resolution?​ ​

$\endgroup$
3
  • 1
    $\begingroup$ Look in Kollár and Shepherd-Barron, Threefolds and deformations of surface singularties, Inventiones 1988, Section 2, or Kollár and Mori, Birational geometry of algebraic varieties, CUP 1998, Section 4.3. I am sure there are other sources; these are the ones I have at hand. $\endgroup$
    – Balazs
    Commented Mar 16, 2020 at 10:35
  • $\begingroup$ Dear @Balazs: Thank you very much for the suggested reference. I was not aware of this development after Brieskorn. If I understand correctly, "simultaneous resolution" as I defined above is the same as "very weak simultaneous resolution" in Kollár and Shepherd-Barron, and according to Theorem 2.4 (due to Laufer) in the paper, the answer to my question is yes because $K^2_{rel}$ is locally constant. Am I right? $\endgroup$
    – AG learner
    Commented Mar 18, 2020 at 5:55
  • $\begingroup$ Dear @Balazs, here is an update of the research on this problem. In the two references you mentioned, there are relevant theorems, e.g., Kollár and Mori, Thm 4.43, Kollár and Shepherd-Barron, Thm 2.4. However, these theorems are about local deformations, i.e., the base is not global. We also consulted Kollar, and he thinks the answer is negative: "the simultaneous resolution exists locally analytically, so when we glue then we get something that is not projective, no matter what base change you do..." But he mentioned there will be a stack represent simultaneous resolution, but not separated. $\endgroup$
    – AG learner
    Commented Jun 23, 2020 at 14:47

0

You must log in to answer this question.