A non-algebraizable foliation
Paulo Sad · 2025
Read A non-algebraizable foliation by Paulo Sad (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-18472.txt.
Introduction
Paulo Sad gives one sharp case in a 2025 geometry paper. He shows a holomorphic foliation on a surface that is not algebraizable. Yet the foliation is topologically equivalent to an algebraic one. The case splits two traits that might have been thought to go hand in hand. One trait is topological sameness. The other trait is realizability with algebra, the chance to be built by such means.
Catalog record
| CREATOR | Paulo Sad |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 9 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-18472
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A non-algebraizable foliation Paulo Sad Abstract . It is presented an example of a holomorphic foliation of a non-algebraizable surface which is topologically equivalent to an algebraic foliation. In this note we show in a specific example how to change an algebraic foliation to another holomorphic foliation which is topologically but not holomorphically equivalent. Of course several examples of moduli spaces of foliations are known, the simplest one appears with local 2-dimensional singularities in the Poincaré domain; there is also an extensive analysis in [ 1 ] for more general singularities. The difference here is that we change the holomorphic type of the surface where the foliation is defined, but preserving the existence of a similar (from the topological point of view) holomorphic foliation of the new surface. More precisely, we exhibit a pair of holomorphic foliations \mathcal{F} and \mathcal{G} with the following properties: (i) \mathcal{F} is defined in the projective plane {\mathbb{C}}P^{2} ; it has an invariant smooth compact curve C of genus 3 with 16 dicritical singularities. The local holonomy map relatively to C at each singularity is the identity, but the holonomy group of C has infinitely many elements. (ii) \mathcal{G} is defined in some non algebraizable surface containing a copy of C and it is topologically equivalent to the restriction of \mathcal{F} to some neighborhood of C . In the language of [ 1 ] , the moduli space of \mathcal{F} is not trivial. We use the construction presented in [ 2 ] which allows to go from an algebraic surface to a non algebraizable one, but in a different way since we deal with extending holomorphic applications between (non smooth) Stein spaces. We apply the construction to the surface associated to the Neeman’s example which has a foliation with a genus 3 compact curve as a leaf (); that foliation can be transported to the non-algebraizable surface. 1. The Surfaces We start with a smooth plane quartic C\subset{\mathbb{C}}P^{2} and select 16 points: two pairs of points belong to two bitangent lines and the remaining 12 points to the intersection with a smooth cubic. Then we blow-up {\mathbb{C}}P^{2} at these poins to obtain a surface \hat{{\mathbb{C}}P^{2}} ; the remarkable fact here is the existence of a smooth strictly psh function f defined on a neighborhood V=\underset{t>t_{0}}{\bigcup}f^{-1}(t)\cup{\hat{C}}\subset{\hat{{\mathbb{C}}P^{2}}} of the strict transform \hat{C} of C which tends to \infty as we approach \hat{C} . As a consequence, if \bar{t} is such that the real 3-manifold f^{-1}({\bar{t}}) is smooth then by [ 5 ] the set {\hat{{\mathbb{C}}P^{2}}}\setminus\underset{t\geq{\bar{t}}}{\bigcup}f^{-1}(t) is holomorphically convex and contains a maximal exceptional subset that can be contracted to a finite number of points [ 6 ] . We get a (singular) Stein surface M with a finite number of singularities; we denote by M_{\infty} the compact surface M\cup V . Remark \hat{{\mathbb{C}}P^{2}}\setminus{\hat{C}} is an example of a 1-convex complex surface, therefore it is a holomorphically convex set (not necessarily Stein, differently from the construction in Section 6 of [ 2 ] ). Let us introduce the following notation: (i) p_{1},\dots,p_{16} the points of C where {\mathbb{C}}P^{2} is blown-up. (ii) D_{1},\dots,D_{16} the exceptional divisors. (iii) \{{\hat{p}}_{j}\}={\hat{C}}\cap D_{j} (iv) V_{j}\subset\hat{{\mathbb{C}}P^{2}} neighborhood of D_{j} that is blown-down to a small neighborhood of p_{j} The surface \hat{S}:=V\underset{id}{\cup}(V_{1}\cup\dots\cup V_{16})\subset M_{\infty} may be seen as V glued to each V_{j} by the identity map near \hat{p}_{j} . Now we replace one of the identifications, say near \hat{p}_{1} , by another biholomorphism \phi that sends a smooth curve l_{1}\subset V transverse to \hat{C} and D_{1} at \hat{p_{1}} into D_{1} . Let us explain briefly how this glueing is done. We take a bidisk \{(x,y);|x|\leq r,|y|\leq r\} such that \hat{p_{1}}=(0,0) , \{(x,y);|x|\leq r,y=0\} is contained in \hat{C} and \{(x,y);|x|\leq r,|y|=r\} is outside \bar{V} . We assume that l_{1} is parametrized as x=x(y) with |x(y)|<r/2 and x(0)=0 . In V_{1} we take the bidisc \{(x^{\prime},y^{\prime});|x^{\prime}|\leq r,|y^{\prime}|\leq r\} such that \hat{p_{1}}=(0,0) , \{(x^{\prime},y^{\prime});|x^{\prime}|=r,|y^{\prime}\leq r\} is contained in \partial V_{1} , \{(x^{\prime},0);|x^{\prime}|\leq r\}\subset{\hat{C}} and \{(0,y^{\prime});|y^{\prime}|\leq r\}\subset D_{1} . The mapping (x^{\prime},y^{\prime})=\phi(x,y) satisfies: \phi sends |x|=r to |x^{\prime}|=r , \phi(x,0)=(x,0) , it preserves the horizontals y^{\prime}=c and \phi(x(c),c)=(0,c) for each |c|\leq r ( \phi can be taken as a homography for each horizontal). It is easy to see that we can define a continuous isotopy from \phi|_{\{(x,y);|x|=r\}} to the identity restricted to \{(x,y);|x|=r_{1}\} ( r_{1} close to r ) along the horizontals \{(x,y);r_{1}\leq|x|\leq r,|y|=c\} , c sufficiently small. We put \hat{S}^{\prime}=(V\underset{\phi}{\cup}V_{1})\underset{id}{\cup}(V_{2}\cup\dots\cup V_{16}) ; it is a smooth complex surface. Let us denote by V^{\prime} and D_{1}^{\prime} the copies of V and D_{1} contained in {\hat{S}}^{\prime} and by i the natural biholomorphism which takes V to V^{\prime} . Lemma Suppose that l_{1} is contained in some irreducible algebraic curve L_{1} of \hat{{\mathbb{C}}P^{2}} which intersects \hat{C} in another point \hat{p} different from \hat{p_{1}} . Then there is no embedding of {\hat{S}}^{\prime} into a compact holomorphic surface. Proof. Without any loss of generality we will prove that {\hat{S}}^{\prime} is not an open subset of a compact holomorphic surface T . The function f^{\prime}=f\circ{i}^{-1} is also a strictly smooth psh function of V^{\prime} . We have V^{\prime}=\underset{t>t_{0}}{\cup}{f^{\prime}}^{-1}(\bar{t}) and {f^{\prime}}^{-1}(\bar{t}) is a smooth real 3-manifold. Consequently T\setminus\underset{t\geq{\bar{t}}}{\cup}{f^{\prime}}^{-1}(t) is a holomorphically convex set with boundary {f^{\prime}}^{-1}(\bar{t}) . We contract the exceptional subset of T\setminus\underset{t\geq{\bar{t}}}{\cup}{f^{\prime}}^{-1}(t) to get a (singular) Stein surface M^{\prime} . We write M_{\infty}^{\prime}=M^{\prime}\cup V^{\prime} . Consider now i restricted to \underset{t_{0}<t<{\bar{t}}}{\bigcup}f^{-1}(t)\subset M and taking values in \underset{t_{0}<t<{\bar{t}}}{\bigcup}{f^{\prime}}^{-1}(t)\subset M^{\prime} . By a generalized Hartogs Theorem for holomorphic applications between Stein surfaces ( [ 3 ] ) we can extend i to a holomorphic application I defined in M and taking values in M^{\prime} , and a fortiori from M_{\infty} to M^{\prime}_{\infty} . Now we look to I^{-1}(D_{1}^{\prime}) (we keep using the same notation for D_{1}^{\prime} after the contractions done in T\setminus\underset{t\geq{\bar{t}}}{\cup}{f^{\prime}}^{-1}(t)) . It contains L_{1} , so that \{\hat{p}_{1},\hat{p}\}\subset I^{-1}(D_{1}^{\prime}) , contradiction. ∎ As a consequence, \hat{S} is not biholomorphically equivalent to {\hat{S}}^{\prime} when we use special curves for l_{1} . For the sake of clarity, let us indicate these surfaces as {\hat{S}_{1}}^{\prime} Let S_{1}^{\prime} be the surface obtained from {\hat{S}_{1}}^{\prime} blowing-down all the (-1) divisors. The Lemma implies Theorem The germ of S_{1}^{\prime} along C is not algebraizable. 2. The Foliations It is known ( [ 4 ] ) that we can select cubics in the Neeman’s construction such that when the blow up’s occur at the 12 points of intersection with the quartic C (besides the points lying in the bitangents) the resulting surface \hat{{\mathbb{C}}P^{2}} has a global foliation \hat{\mathcal{F}} which contains \hat{C} as a leaf (withouth singularities!). The surface \hat{S} we considered in the last Section has therefore a regular foliation \hat{\mathcal{F}} . Let us go back to the construction of the surface \hat{S^{\prime}} of the last Section. In the description of the glueing map \phi we used bidiscs in V and V_{1} ; their horizontals can be supposed to be contained in the leaves of \hat{\mathcal{F}} . Since \phi preserves horizontals, we have a new foliation \hat{{\mathcal{F}}^{\prime}} defined in S^{\prime} . We write {\mathcal{F}}_{1}^{\prime} for such a foliation in S_{1}^{\prime} . Proposition \hat{\mathcal{F}} and \hat{\mathcal{F}}^{\prime} are topologically equivalent. Proof. We consider \hat{S} as the glueing of V to V_{1} by the identity. We define a topological equivalence \Psi as