Is there an example of a complex analytic space $X$ that doesn't have any (not necessarily open or closed) positive dimensional subspace $Y$ which is analytically isomorphic to (the complex analytic space associated to) a scheme?

Edit: after D.Arapura's comment, we require $X$ to have dimension $>1$.

If I remember correctly, there are non "Abelian" complex tori $X=\mathbb{C}^n/\mathrm{Lattice}\;\;$ that do not have any positive dimensional analytic subvariety. Can a counterexample be derived from this?

Also, any complex algebraic space has an open subspace which is a scheme. So the counterexample (if it exists) must be searched outside algebraic spaces.

What if the question is modified by requiring that $X$ has no $Y$ that is locally closed in the analytic Zariski topology (where opens are complements of analytic subspaces)?

  • 2
    $\begingroup$ A disk will work, but this is probably not what you want. Did you want $X$ to be compact? $\endgroup$ May 8, 2011 at 13:00
  • $\begingroup$ +1 to DonuArapura's comment! Perhaps I should've required $X$ to be of dimension > 1 ? $\endgroup$
    – Qfwfq
    May 8, 2011 at 13:07
  • $\begingroup$ Does a polidisk work? If so, I'll add the requirement "$X$ compact" as suggested by D.Arapura. $\endgroup$
    – Qfwfq
    May 8, 2011 at 13:26
  • $\begingroup$ Yes, I suspect that a polydisk does work, although this would be much harder. Off the top of my head, I might argue that polydisks have many bounded holomorphic functions, while the algebraic examples probably don't (reduce to the quasiprojective case, and apply suitable Riemann/Hartogs extension theorems to extend to the projective closure). $\endgroup$ May 8, 2011 at 13:40
  • $\begingroup$ I deleted my answer, because the first half of it was wrong: Torus -point is not complex analytic equivalent to a scheme if the tours is non-algebraic. For the second example, I can not prove that non-algebraic torus can not contains something 2-dimensional complex analytic subset X isomorphic to a scheme, so that torus-X is a one dimensional Riemann surface of infinite topological type. $\endgroup$ May 8, 2011 at 18:17

1 Answer 1


Take a complex 2-torus $X$ without curves, and hence, without non-constant meromorphic functions (see e.g. Shafarevich, Basic algebraic geometry, chapter VIII, \S 1, example 2). The only locally closed subsets of $X$ will be $X$ itself, $\varnothing$, finite subsets and the complements of finite subsets. $X$ minus a finite subset can't be algebraic: if it were, it would have a non-constant meromorphic function, and then so would $X$.

[upd: locally closed here means locally closed in the analytic Zariski topology, so this answers the second question and not the first.]

  • $\begingroup$ Why can not torus have a locally analytic curve of infinite genus, say a leaf of a foliation? $\endgroup$ May 21, 2011 at 1:24
  • $\begingroup$ Dmitri -- I'm not sure I understand what a locally analytic curve of infinite genus is. But an analytic Zariski closed subset of a compact complex manifold is homeomorphic to a finite CW-complex $\endgroup$
    – algori
    May 21, 2011 at 1:44
  • $\begingroup$ @dmitri. The holomorphic foliations on $X$ are defined by vector fields. To see it, take a vector field and look at the tangency locus with a given foliation. Since $X$ has no curves in it, the foliation and the vector field are everywhere tangent or transverse. But of course we can start with a vector field tangent to the foliation at a given point. $\endgroup$ May 21, 2011 at 2:00
  • $\begingroup$ algori sorry for the confusion. I agree with your remark 100%, but it does not remove my concerns. I had in mind the following. We know that on a hyperbolic genus $g$ surface we can have a collection of geodesics whose complement is a collection of open ideal hyperbolic triangles. This collection is called lamination. Each geodesic is not compact in the surface. So why can not you have a lamination in the torus whose complement is analytically isomorphic to something algebraic? $\endgroup$ May 21, 2011 at 8:53
  • $\begingroup$ jvp, you are right about foliations, when I was writing "foliation" I was thinking also about laminations... Is it clear that they don't exist? (by lamination I mean something like an infinite collection of vertical lines on C^2 given by equations $x_n=\frac{1}{n}$) $\endgroup$ May 21, 2011 at 8:59

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.