1
$\begingroup$

A holomorphic disk is the image of an injective holomorphic map $f:\mathbb D \to \mathbb C^n$ from the unit disk $\mathbb D \subset \mathbb C$ to $\mathbb C^n$.

Let $\Omega$ be a pseudoconvex domain in $\mathbb C^n$. Let $D=f(\mathbb D)\subset \overline\Omega$ be a holomorphic disk in the closure of $\Omega$ that is the uniform limit of holomorphic disks in $\Omega$. The Continuity Theorem implies, that if the boundary $\partial D$ is contained in $\Omega$, then so is $D$.

What if an inner point $p\in D$ of the disk lies in the boundary? Can another inner point of the disk lie in $\Omega$?

Preliminary results:

By the continuity theorem, an inner point of $D$ in the boundary of $\Omega$ immediately implies that there is also at least one point of $\partial D$ in $\partial \Omega$. Points of $\partial D$ in $\partial \Omega$ are not enough to force $D$ into the boundary, as $\mathbb D \times{0} \subset \mathbb D \times \mathbb D$ shows.

If both $D$ and $\partial \Omega$ are smooth (at $p$), then at $p$ we have a complex tangent vector to both $D$ and $\partial \Omega$, so by Levi-pseudoconvexity a real tangent direction in which $D$ enters inside $\Omega$ necessitates a corresponding real tangent direction in which $D$ exits $\overline \Omega$. So in this case the answer is no.

Passing to smaller subdisks around the point in $\partial \Omega \cap D$ shows that such a point is never isolated, by the continuity theorem.

Since a singularity of $D$ is isolated, we can always find a regular point of $D$ in $\partial \Omega$ in our setting.

Bonus question:

What if $D$ is not a limit of inner disks?

$\endgroup$
0

3 Answers 3

2
$\begingroup$

Here is an example with a "negative" result: take a plurisubharmonic function $u:\Bbb C^n\to[0,\infty)$, not identically zero, with $u(\lambda z)=|\lambda|u(z)$, $\lambda\in\Bbb C, z\in\Bbb C^n$, and its zeroes are dense in $\Bbb C^n$. Such a function exists according the work of J. Siciak. Put $h(z):=u(z)+\|z\|$ and $D:=\{z\in\Bbb C^n: h(z)<1\}$. Then $D$ is pseudoconvex sitting in the unit ball. Take points $a, a_j$ in the boundary of the unit ball with $u(a_j)=0$ and $u(a)> 0$ such that $a_j\to a$. Then the discs $\lambda\to\lambda a_j$ are sitting in $D$ and converge uniformly to the disk $\phi(\lambda):=\lambda a$. Note that the limit disc is sitting partially in $D$, partially on the boundary of $D$.

$\endgroup$
1
  • $\begingroup$ Great example, thank you. It seems the discontinuous boundary is key. Now one could wonder if there is such an example with continuous boundary, but in light of your other answer, this would mean finding an example of a $C^0$ pseudoconvex domain that is not hyperconvex. To my knowledge, the existence of such a domain is an open question. $\endgroup$
    – hife
    Commented Oct 4, 2023 at 8:48
2
$\begingroup$

Take a bounded pseudoconvex domain $D\subset\Bbb C^n$ which allows a continuous function $u:\overline D\to(-\infty,0]$, plurisubharmonic on $D$, not identically zero, and zero on the boundary. Let $\phi_j:\Bbb D\to D$ be holomorphic and let $\phi_j\to\phi:\Bbb D\to\overline D$ uniformly. Assume there is a point $\lambda_0\in\Bbb D$ with $\phi(\lambda_0)\in\partial D$. Since $u$ is uniformly continuous on $\overline D$, we get that $u\circ\phi_j$ converges locally uniformly to $u\circ\phi$. Since the functions $u\circ\phi_j$ are subharmonic, the limit function is also. Then by the maximum principle $u\circ\phi$ vanishes identically, implying that the limit disk lies totally in $\partial D$. Such a function $u$ exists if $D$ is hyperconvex according a result of Blocki. Moreover, if $D$ has a H"older-continuous boundary, then it is hyperconvex according to a recent result by Bo-Yong Chen.

$\endgroup$
0
1
$\begingroup$

Take the so-called Hartog's triangle $\{(z,w)\in C^2:|w|<|z|<1\}$ and the disk $\lambda\to (\lambda,0)$.

$\endgroup$
1
  • $\begingroup$ Thanks, this is a nice and simple example for my bonus question, but it is not a limit of inner disks, which my main question was about. This was not clear from the title. I have now edited the title to emphasise the limit property. $\endgroup$
    – hife
    Commented Sep 28, 2023 at 15:50

You must log in to answer this question.

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