2
$\begingroup$

hallo,

I have the following question: Let $M$ be a $n-$dimensional complex manifold and $X \subset M$ be a compact $n-$dimensional totally real analytic Riemannian submanifold. Let furthermore $\alpha$ be a 2-form on $X$. Is it possible to extend $\alpha$ to a holomorphic 2-form in a neighbourhood of $X$ in $M$? I hope for answers. Thanks in advance.

bruno

Improve this question
$\endgroup$
9
  • $\begingroup$ are there any references ? $\endgroup$
    – bruno
    Commented Sep 5, 2012 at 6:42
  • $\begingroup$ does anyone have an idea ? $\endgroup$
    – bruno
    Commented Sep 5, 2012 at 10:07
  • 2
    $\begingroup$ Wouldn't the existence of such an extension require $\alpha$ to be real-analytic on $X$? $\endgroup$
    – Andreas Blass
    Commented Sep 5, 2012 at 13:34
  • $\begingroup$ @bruno: what's a totally real submanifold? $\endgroup$
    – Sándor Kovács
    Commented Sep 5, 2012 at 18:18
  • 1
    $\begingroup$ When $\alpha$ is a function the answer is affirmative (the analytic expansion of $\alpha$ converges in a neighborhood of any point of X). For $\alpha$ a form, locally extend it using local coordinates, and then observe that by uniqueness the extension is global. $\endgroup$
    – AndreA
    Commented Sep 6, 2012 at 21:07

1 Answer 1

Reset to default
3
$\begingroup$

Such extension exists. This follows from the followoing three facts.

1) For every point $x\in X$ there is a neighbourhood $U(x)\subset M$ and a holomorphic map $\phi: U(x)\to \mathbb C^n$ such that $\phi (U(x)\cap X)\subset \mathbb R^n$.

2) Let $\alpha$ be a real analytic two-form defined on $\mathbb R^n$, then it can be extended to a holomorphic two-form on a neighbourhood of $\mathbb R^n$ in $\mathbb C^n$ that retracts on $\mathbb R^n$

3) If you have two extensions of $\alpha$ to a same neighbourhoods of $\mathbb R^n$ then they coincide.

I don't know a reference for 1) but it should be something standard, to get 2) and 3) you can just use Taylor decomposition of $\alpha$.

Once you know 1,2,3 the statement follows easily.

Improve this answer
$\endgroup$

You must log in to answer this question.

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