Smooth Markov partition conjecture for expanding toral endomorphisms
Let be an expanding integer matrix, and let be the induced toral endomorphism. A smooth Markov partition conjecture asserts that admits a smooth, in fact linear, Markov partition if and only if there is some such that is diagonalizable with integer eigenvalues.
Smooth Markov partition conjecture. The induced toral endomorphism admits a smooth, in fact linear, Markov partition if and only if for some , is diagonalizable with integer eigenvalues.
This conjectures a generalization to arbitrary dimensions of the preceding two-dimensional characterization. The paper leaves this generalization for future work; no resolution is given here.
References
- Chayce Hughes and Huub de Jong, “Smoothness of Markov Partitions for Expanding Toral Endomorphisms”, arXiv:2604.00257 (2026).
Progress summary
A 2026 paper proves one direction in every dimension and the full statement in two dimensions, but the general conjecture remains open.
The conjecture asserts that an expanding toral map has a smooth, linear Markov partition exactly when some power is diagonalizable with integer eigenvalues. It extends the established two-dimensional characterization to arbitrary dimension.
Known results
- In every dimension, the stated algebraic condition implies existence of a linear Markov partition.
- In dimension , the converse also holds, giving the full characterization.
- In dimension , checking suffices for the criterion, since the relevant orders are .
- Other two-dimensional cases exhibit nowhere-smooth or hybrid partition behavior.
April 2026 partial results
The paper Smoothness of Markov Partitions for Expanding Toral Endomorphisms presents the arbitrary-dimensional converse as Conjecture 1.5 and explicitly leaves it for future work. No proof, counterexample, or independent verification of the full conjecture was found.
Current status (as of October 2026): The sufficient direction in arbitrary dimension and the complete two-dimensional case are established, while the converse in dimensions greater than remains open.
Solutions 0
No solutions have been posted yet.