Fundamental groups of special complex varieties and root orbifolds

The explanation

Take a closed curve of genus g, with B points removed. Count 2g − 2 + B. For a smooth curve, specialness means this count is at most zeroSource 1. A doughnut-shaped elliptic curve has g = 1 and B = 0, so the count is 0. It is special. The sphere with two points removed gives −2 + 2 = 0, also special. The sphere with three points removed gives −2 + 3 = 1. It is not specialSource 1.

Now look at loops. The fundamental group of a space records the loops in it, up to deformation, with joining loops as the operation. For the doughnut it is pairs of whole numbers, and the order of joining does not matter. Such a group is abelian. For a curve with two or more holes, the group is far from abelian. So, among closed curves, special shapes have abelian loop groups.

Campana’s abelianity conjecture predicts this pattern in every dimension. A special compact Kähler manifold should have a fundamental group that becomes abelian after passing to a finite unramified coverSource 2. Such a group is called virtually abelian. Here special means, roughly, that no holomorphic forms detect a map onto a base of general typeSource 3. A compact Kähler manifold is a compact complex space with a compatible way to measure lengths.

The first paper claims to prove the conjecture in every dimensionSource 4. The claim covers the whole fundamental group. It needs no assumption of linearity, residual finiteness, projectivity or a particular Kodaira dimensionSource 5. The finite index may depend on the manifoldSource 5. One consequence covers Kodaira dimension zero, a standard class that includes Calabi–Yau manifolds. Campana had shown such manifolds are special. So their fundamental groups are virtually abelianSource 6.

What was known before? Campana and Claudon proved the conjecture for compact Kähler threefoldsSource 7. For manifolds with zero first Chern class, Beauville’s decomposition already gave the answerSource 8. Results on linear quotients were also known. But an infinite group need not have an infinite linear imageSource 9. A linear image means a copy of the group acting by matrices. Controlling those leftover quotients is the key obstacle for a claim about the whole groupSource 9. The paper’s new step gets enough positivity from holomorphic tensors with values in infinite-dimensional Hilbert spacesSource 10.

The second paper moves to open, non-compact varieties. There, nonabelian groups of a mild kind can appear. Take the Heisenberg group of 3 × 3 upper-triangular whole-number matrices with ones on the diagonal. Let a have a 1 in the top middle, and b a 1 at the right of the middle row. Their commutator, a b a⁻¹ b⁻¹, has a single 1 in the top-right corner. That commutator commutes with both a and b. A group where every commutator commutes with everything has nilpotency class at most twoSource 11.

The second paper claims an independent proof about such groups. Take a smooth special quasi-projective variety. Every complex matrix image of its fundamental group is virtually of class at most twoSource 12. Cao, Deng, Hacon and Păun announced this conclusion first, and the paper gives them prioritySource 13. The bound is optimal. An earlier example has a Heisenberg-type group that is not virtually abelianSource 14.

The third paper treats a root orbifold. This is a fourfold X with one smooth divisor D counted with weight one half. The paper claims virtual abelianity of the full orbifold fundamental group when this orbifold is specialSource 15. It is conditional: it assumes the first paper’s theoremSource 16. It builds a special eightfold whose fundamental group maps onto the needed groupSource 17.

How sure is this? No Lean formalization is listed for these papers in the openai/math release. So no computer proof checker has checked them. The third paper depends on the first. No specialist review has been published.

What it doesn't show. This does not bound the index of the abelian subgroup or cover orbifolds beyond the one-divisor fourfold case.