Skip to main content
3 of 4
Fixed mistake at the end of the proof (T fibers over the circle but it doesn't have to induce an automorphism on each fiber)

I'm assuming a Li-Yorke pair for $T$ is a pair $x,y$ such that $\limsup_{n\to \infty} d(T^n(x), T^n(y)) > 0 $ and $\liminf_{n\to \infty} d(T^n(x), T^n(y)) = 0$.

For toral automorphisms, I think the answer to your question is no, based on this:

Claim: A linear toral automorphism has a Li-Yorke pair if and only if it is ergodic.

Since ergodic linear toral automorphisms have positive entropy (see for instance here), the claim answers your question negatively.

Proof of the claim: The if part is easy, since a point with a dense orbit and a fixed point give a Li-Yorke pair. Now assume $T$ is not ergodic but has a Li-Yorke pair. Being non-ergodic implies that there is an eigenvalue which is a root of the unity. We may replace $T$ by some power of $T$ and assume that $1$ is an eigenvalue. Moreover you can find an integer eigenvector $v$ with eigenvalue $1$ (because the coefficients of the matrix corresponding to $T$ are integer), so passing to the quotient by $\mathbb{R}v$ you get a torus of dimension $n-1$. Since $T$ had a Li-Yorke pair, so does the automorphism induced by $T$ on the quotient, so you have reduced the dimension. The claim is trivial in dimension $1$, so the proof is completed by induction.