EDIT: Let $f\colon X\to Y$ be a morphism of complex analytic spaces (not necessarily smooth or reduced). Assume that

(a) $f$ is injective on points;

(b) $f$ is local imbedding near each point $x\in X$;

(c) for any holomorphic germ $\gamma\colon (\mathbb{C},0)\to Y$ with image contained in $f(X)$ there exists a holomorphic germ $\tilde\gamma\colon (\mathbb{C},0)\to X$ such that $f\circ \bar \gamma=\gamma$.

(d) $f(X)$ is a closed subset of $Y$.

Question. Is it true that $f$ is proper?

Remark. (1) In the algebraic situation the answer is positive by the standard valuative criterion of properness.

(2) A priori it is not clear to me that locally on $Y$ the image $f(X)$ is an analytic subset.

If there is any analytic version of the valuative criterion of properness, I would be curious to know anyway. A reference would be helpful.

EDIT: Conditions (a),(c), (d) imply the following property which looks closer to the standard valuative criterion of properness. For any holomorphic germs $\gamma\colon (\mathbb{C},0)\to Y$ and $\beta\colon (\mathbb{C},0)\backslash\{0\}\to X$ such that $f\circ \beta=\gamma|_{(\mathbb{C},0)\backslash\{0\}}$, there exists a holomorphic germ $\bar\gamma\colon (\mathbb{C},0)\to X$ such that $$\bar\gamma|_{(\mathbb{C},0)\backslash\{0\}}=\beta,\, f\circ \bar\gamma=\gamma.$$

  • $\begingroup$ This map is not injective, but I will add some extra assumptions on $f$ which are satisfied in my situation. $\endgroup$
    – makt
    Jan 20, 2015 at 15:38
  • 1
    $\begingroup$ Are you sure condition (c) is what you mean? The inclusion of an open subset of $Y$ satisfies (a), (b) and (c). $\endgroup$ Jan 20, 2015 at 15:57
  • $\begingroup$ @LaurentMoret-Bailly: Thanks, you are right. I added also the condition (d) that $f(X)$ is closed (but not necessarily analytic). This is still satisfied in my situation. $\endgroup$
    – makt
    Jan 20, 2015 at 16:14

1 Answer 1


Here is a counterexample: $Y=\mathbb C^2$, $X=(\mathbb C\setminus\{0\})\cup\mathbb C$, and $f$ is given by $z\mapsto (z,e^{1/z})$ on $\mathbb C\setminus\{0\}$ and is given by $w\mapsto(0,w)$ on the second copy of $\mathbb C$. This map is certainly not proper. It clearly satisfies conditions (a),(b),(d). Let's check condition (c).

Let $\gamma:(\mathbb C,0)\to Y=\mathbb C^2$ be a germ whose image is contained in $f(X)$. If the germ sends $0$ to a point whose first coordinate is nonzero, then it clearly lifts, so it suffices to check the case when $\gamma(0)=(0,w)$ for some $w\in\mathbb C$. Consider the first coordinate of $\gamma$: if this first coordinate is not identically zero, then it covers an open neighborhood of zero by the open mapping theorem. On the other hand, the function $e^{1/z}$ takes arbitrarily large values near zero since it has an essential singularity there, and this contradicts the fact that the image of $\gamma$ is contained in the image of $f$. Thus the first coordinate of $\gamma$ is identically zero, and thus it trivially lifts to $X$.


Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.