Complete model spaces do not force Lie-group regularity

About 44 years old · traced to

Is every infinite-dimensional Lie group modelled on a complete locally convex space regular?

References

Progress summary

Refreshed
Claimed solved

A July 2026 preprint claims a counterexample showing that completeness of the modelling space does not guarantee regularity, but the result is not yet independently confirmed.

The problem asks whether every infinite-dimensional Lie group modelled on a complete locally convex space is regular. The reported construction gives a negative answer to this question and to the related Glöckner–Neeb question about automatic multiplication growth.

July 2026 counterexample

The preprint constructs a contractible complex analytic BCH–Lie group HH, modelled on the complete Silva space φC\varphi_{\mathbb{C}}, with homeomorphic exponential map exp⁡H:L(H)→H\exp_H:L(H)\to H, yet HH is not even C0C^0-semiregular. It exhibits smooth curves γλ\gamma_\lambda converging to zero in C∞([0,1],W)C^\infty([0,1],W) for a finite-dimensional WW, while U′(t)=γλ(t)U(t)U'(t)=\gamma_\lambda(t)U(t), U(0)=1U(0)=1, has no HH-valued C1C^1 solution on [0,1][0,1]. The associated groups are claimed not to be CkC^k-semiregular for any k∈N0∪{∞}k\in\mathbb{N}_0\cup\{\infty\}.

Current status (as of July 2026): The universal regularity assertion is claimed false by an explicit complete-space counterexample, but independent verification and publication status remain open.

Sources

Solutions 0

No solutions have been posted yet.