Family 042 — Oka K3 surfaces: extending holomorphic maps with freedom

← All digests

Family 042 · Algebraic and complex geometry
Importance 4/5 · Provisional editorial rating
Sources consulted 2026-10-12

1. The question

Suppose a holomorphic map is defined only near a small closed region in complex Euclidean space. Can it be approximated there, as accurately as desired, by a holomorphic map defined on the whole Euclidean space?

For complex-valued functions on a convex set, polynomial approximation offers a familiar model. If the values must lie on a curved complex surface, however, one cannot simply approximate each coordinate independently: the resulting coordinates might fail to describe a point of the surface. Global restrictions on the target can obstruct the approximation altogether.

An Oka manifold is a target with a particularly strong form of holomorphic flexibility. Family 042 claims that every complex K3 surface has this property, including nonprojective K3 surfaces that contain no algebraic curves. It then derives consequences for entire curves and several classes of compact complex surfaces.

The claim concerns approximation of every local map in every source dimension. Merely producing a single interesting holomorphic curve on the surface is much weaker.

2. Background and essential definitions

K3 surfaces

A complex K3 surface is a compact connected complex surface that is simply connected and has trivial canonical bundle. The latter condition means it possesses a holomorphic two-form that never vanishes. Every complex K3 surface is Kähler, but not every one can be embedded in projective space.

A smooth quartic surface in $\mathbb P^3$ is a standard projective example. Nonprojective examples can have no curves at all, so a proposed argument based solely on moving algebraic curves cannot immediately handle the whole class.

“Surface” here means complex dimension two, hence real dimension four. A holomorphic map from $\mathbb C$ traces a complex curve in it; a map from $\mathbb C^2$ can have full-rank derivative and explore a whole local neighborhood.

Convex approximation

A compact set $K\subset\mathbb C^m$ is convex if it contains every real line segment joining two of its points. Let $f$ be holomorphic on a neighborhood of $K$, with values in a complex manifold $X$.

The convex approximation property says that for every $\epsilon>0$, there is an entire holomorphic map $F:\mathbb C^m\to X$ with

$$\sup_{z\in K}d_X(F(z),f(z))<\epsilon.$$

Here $d_X$ is a distance from a smooth Hermitian metric. For compact targets, the qualitative property does not depend on the chosen metric. “Entire” means defined on all of $\mathbb C^m$; it does not mean polynomial, injective, or surjective.

Requiring this in every positive integer dimension $m$ characterizes the Oka property. Its equivalent formulations include deforming continuous maps from suitable complex source spaces into holomorphic ones, with approximation and prescribed holomorphic data. The standard sources are Stein spaces, the complex-analytic counterparts of affine geometric spaces, with enough global holomorphic functions to distinguish points and control compact sets.

Several different kinds of flexibility

A surface is dominable by $\mathbb C^2$ if one entire map has full-rank derivative somewhere. It is strongly dominable if this can be arranged through each target point. These are existence statements about particular maps.

An entire curve is a map $\mathbb C\to X$. It is an immersion when its derivative is never zero. Prescribing a jet means prescribing its value and finitely many derivatives in local coordinates. The Oka property controls such interpolation and approximation in a much wider setting than dominability alone.

Likewise, the Oka-1 property concerns maps from open Riemann surfaces, which have complex dimension one. It does not automatically imply approximation from all $\mathbb C^m$.

3. Development history and previous partial results

The Oka principle developed from the observation that some holomorphic existence problems are governed principally by topology. Gromov’s work on holomorphic sprays gave a powerful geometric mechanism for this flexibility. A spray is a holomorphic family of maps that can move a point in all tangent directions.

Forstnerič’s convex approximation theorem turned a local-looking approximation test into a characterization of the global Oka principle. It is this characterization that makes the main theorem of M1 a full Oka statement.

K3 surfaces posed a difficult test because their geometry combines strong cohomological structure with potentially few algebraic subvarieties. Buzzard–Lu proved dominability for elliptic and Kummer K3 surfaces. Forstnerič–Lárusson strengthened the Kummer conclusion to strong dominability and studied flexibility throughout the compact-surface classification.

Other work removed obstructions without proving the full property. Kamenova–Lu–Verbitsky showed that the Kobayashi pseudodistance vanishes on every complex K3 surface, using period dynamics with a later correction. This pseudodistance measures resistance to holomorphic discs. Its vanishing is compatible with great flexibility, but does not construct the approximating maps demanded by Oka theory. Alarcón–Forstnerič obtained Oka-1 results for elliptic and Kummer K3 surfaces.

A close predecessor is Xie–Zhao’s 2026 work. It proves Oka properties for explicit K3 examples, including smooth hypersurfaces of multidegree $(2,2,2)$ in $(\mathbb P^1)^3$, and establishes dense sets of Oka K3 parameters. These are substantial full-Oka results, beyond the earlier dominability and one-dimensional statements.

Density still leaves a logical gap: a dense set can omit particular points, or even many points. M1 seeks to cover every K3 period, including exceptional configurations not reached by a generic argument. It also uses Chen–Gounelas’s families of genus-one curves for the projective case, rather than treating algebraic curves as a new ingredient invented here.

4. What the new manuscripts claim

Every complex K3 surface is Oka

Theorem 1.1 of M1 states the convex approximation property for every complex K3 surface, every $m\ge1$, every nonempty compact convex $K\subset\mathbb C^m$, every local holomorphic map near $K$, and every positive error tolerance.

There is no restriction on projectivity, the number of divisor classes, or the existence of an elliptic fibration. The target is the given K3 surface, not merely a nearby deformation of it.

Dense entire immersions with a prescribed tangent vector

For any $x\in X$ and any nonzero $v\in T_xX$, Corollary 1.2 claims an entire immersion satisfying

$$f:\mathbb C\longrightarrow X, \qquad f(0)=x, \qquad f'(0)=v, \qquad \overline{f(\mathbb C)}=X.$$

The closure is in the ordinary complex topology. In the projective case this implies Zariski density as well. Density is not surjectivity, and immersion is not injectivity. The prescribed derivative is the actual vector $v$, not only its direction.

The manuscript also states approximation and discrete jet interpolation from open Riemann surfaces in a prescribed continuous homotopy class, and strong dominability by $\mathbb C^2$.

Consequences for compact surfaces

M1 derives the Oka property for every Enriques surface, whose unramified double cover is a K3 surface. Together with established cases for tori, bielliptic surfaces, and Kodaira surfaces, this gives the assertion that every connected minimal compact complex surface of Kodaira dimension zero is Oka.

Here minimal means there is no exceptional curve that can be contracted to a smooth point. Kodaira dimension measures the growth of sections of powers of the canonical bundle. These qualifications delimit a part of the surface classification, not all compact complex surfaces.

For minimal class VII surfaces, characterized here by first Betti number one and Kodaira dimension $-\infty$, M1 claims that precisely the Hopf and Enoki surfaces are Oka. This classification additionally uses the global spherical shell theorem claimed in family 060. That companion is not an input to the K3 theorem itself.

Finally, projective K3 surfaces and Enriques surfaces have global dominating holomorphic sprays, by combining the claimed Oka result with an existing theorem for projective Oka manifolds. The manuscript does not extend this spray conclusion to nonprojective K3 surfaces.

5. Why the advance matters

Importance: 4/5. If correct, the main theorem would settle a prominent flexibility question for a fundamental class of complex surfaces and complete important portions of the Oka surface picture.

Its force comes from the quantifiers. Every local map can be approximated, not just a carefully chosen map; every complex Euclidean source dimension is allowed; every K3 surface is covered. These features would supply a reusable tool for constructing holomorphic maps with specified behavior.

The nonprojective case is especially striking. A surface may contain no compact curves while admitting entire curves dense in its ordinary topology. There is no contradiction: the image of a noncompact source need not be a closed algebraic curve.

The result also connects areas that often use different languages: approximation in several complex variables, algebraic families of curves, symplectic gluing, and arithmetic dynamics on period spaces.

6. The main idea

Extend in one direction while retaining all parameters

M1 first isolates a local operation that enlarges the domain of a holomorphic map in one complex variable. The remaining variables are retained as holomorphic parameters. Two transverse strip maps supply complete directions of motion, one along each coordinate.

The extensions initially disagree on an overlap. Splitting and contraction estimates correct that disagreement. The argument then enlarges polydiscs and removes the real faces of a containing polytope one at a time, passing from the planar operation to approximation on arbitrary convex sets.

Keeping the passive parameters is essential. Without it, the construction would establish only a one-dimensional approximation property.

Find the required directions geometrically

For projective K3 surfaces, two families of genus-one curves provide complete complex directions. Their different degrees force independence at general points, and a propagation argument handles the exceptional set.

For nonprojective surfaces, M1 builds long chains of annuli and joins them by symplectic coordinate changes. Naively correcting the joins could consume all the available transverse width. The proposed sewing theorem keeps the estimates uniform in the length of the chain. Exactness removes a period obstruction, and a quadratic iteration retains a positive width through infinitely many joins.

This constructs suitable open regions in the space of K3 periods. A period plane is the real two-plane spanned by the real and imaginary parts of the holomorphic two-form, after identifying second cohomology with a fixed lattice.

Reach every period, not merely a dense collection

The final transfer separates period planes according to whether they contain zero, one, or two independent rational directions. Arithmetic orbit results and Torelli theory carry the local approximation property through the appropriate orbit in each case. Rational planes fall into the projective case; the other cases require the different open regions built earlier.

Thus the argument does not simply say that approximation survives passage to an arbitrary limit of Oka surfaces. It supplies a route for each type of period plane.

Once Oka interpolation is available, the dense entire immersion is obtained by prescribing nonzero first derivatives and values at a discrete sequence of source points, choosing the target values to form a dense sequence.

7. Scope, limitations, and open questions

The principal claim concerns complex K3 surfaces. It does not prove every higher-dimensional Calabi–Yau or hyperkähler manifold is Oka. Numerical similarities between these classes do not transfer the approximation theorem automatically.

The surface-classification consequences are restricted to the minimal cases stated above. The manuscript makes no classification claim for arbitrary blowups or all properly elliptic surfaces. Its class VII conclusion has the separate family 060 dependency.

Entire maps here are holomorphic, not necessarily algebraic or given by an explicit finite formula. The approximation statement gives no effective complexity bound for finding one. Dense immersions need not be embeddings, and strong dominability is not a global covering map.

The projective restriction on global sprays is also retained. Understanding stronger global structures on nonprojective K3 surfaces, and how these methods extend to higher dimensions, remains beyond the stated conclusions.

8. Sources and further reading

M1’s Theorem 1.1 gives convex approximation, Corollary 1.2 gives dense entire immersions, and Section 10 separates the interpolation, surface-classification, and spray consequences. Its introduction distinguishes the earlier dominability, Oka-1, and full-Oka results.

The papers of Forstnerič, Forstnerič–Lárusson, and Xie–Zhao linked above provide direct accounts of the relevant earlier advances. Background on K3 surfaces helps locate the examples, while the approximation inequality in Section 2 is the central definition needed to read this digest.

Source manuscripts

• M1. OpenAI, Every complex K3 surface is Oka (September 23, 2026).

Sources consulted on 12 October 2026 at repository revision fd4aeeb. Download the manuscript citations.