Thomas–Yau–Joyce conjectures

In the appropriate Fukaya-category setting for a compact Calabi–Yau manifold XX, if a Lagrangian submanifold L⊂XL\subset X defines an object ELE_L that is stable with respect to the relevant Bridgeland stability condition, then there should exist a special Lagrangian submanifold L′⊂XL'\subset X whose Fukaya-category object is isomorphic to ELE_L, i.e. EL′≅ELE_{L'}\cong E_L. The conjectures also predict that mean-curvature flow starting from an unstable Lagrangian evolves through controlled singularities and converges, after the corresponding surgeries, to a configuration of special Lagrangians.

References

Primary source

arXiv

Additional references

Progress summary

Refreshed
Claimed progress

A new preprint reports conditional progress toward the conjectures in K3 surfaces, but the general problem remains open.

The Thomas–Yau–Joyce conjectures seek stability criteria for when a Lagrangian has a special-Lagrangian representative and predict its behavior under mean-curvature flow. The general conjectures are not solved.

Known results

  • Stoppa (2025) proved a restricted equivalence for Lagrangian sections in certain Calabi–Yau fibrations mirror to toric weak Fano manifolds, including selected weak del Pezzo cases.
  • In specified hyperkähler 44-manifolds, mean-curvature flow was shown to undergo finitely many neck-pinches and converge to chains of special-Lagrangian spheres; the full Thomas conjecture remains open.

September 2026 K3 development

Jacopo Stoppa’s preprint claims a stability-to-special-Lagrangian implication for a class of K3 objects, with unconditional results for selected quartic and sextic surfaces. It presents a controlled compact-manifold realization of the conjectural picture, but its principal theorem assumes strong expected properties of homological mirror symmetry; the claim is unverified.

Current status (as of October 2026): restricted and conditional results are available, while the general Thomas–Yau–Joyce conjectures remain open and the K3 preprint is an unverified progress claim.

Sources

Solutions 0

No solutions have been posted yet.