Nonexpansive directions in the Jeandel-Rao Wang shift (Q6165953)

In this work, the authors study the Jeandel-Rao tilings (see [\textit{E. Jeandel} and \textit{M. Rao}, Adv. Comb. 2021, Paper No. 1, 37 p. (2021; Zbl 1478.05020)]) and the associated two-dimensional subshift, investigating the dynamics of the subshift via the set of \textit{slopes of nonexpansive directions}. These are the slopes for which the translation action on the subshift in this direction is nonexpansive, in the sense that for all \(t > 0\) there exist distinct elements of the subshift \(x, y\) such that in a neighbourhood of radius \(t\) around the subspace of \(\mathbb{R}^2\) of the given slope, \(x\) and \(y\) agree. Nonexpansive directions are a way of formalising the famous Conway worms, first identified in Penrose tilings and then many others. The set of slopes of nonexpansive directions is a conjugacy invariant for \(\mathbb{Z}^2\) subshifts. The authors calculate the set of slopes for the Jeandel-Rao tilings, finding them to be given by \(\{0, \phi+3, -2\phi+2,-\phi+\frac{5}{2}\}\), where \(\phi = \frac{1+\sqrt{5}}{2}\) is the golden ratio. Of particular interest is that three of these slopes are irrational. This results relies on previous work of the first author [Ann. Henri Lebesgue 4, 283--324 (2021; Zbl 1492.37022)] in which a Markov partition is found for the unique minimal subshift, which encodes much of the dynamics of the shift itself. From this result, the authors are then able to identify all Conway worms for the Jeandel-Rao tilings and describe their resolutions in terms of elements in the Fibonacci subshift. This gives a new context for and extends the key observation made by E. Jeandel and M. Rao in their work [loc. cit.] that the tilings can be partitioned into rows of height \(4\) and \(5\) arranged consecutively in the order prescribed by some element of the Fibonacci subshift, hence proving aperiodicity of the tilings. The proofs in the article are technical and calculation-heavy, but the authors are careful to make them readable and provide many diagrams to illustrate their definitions and methods. The paper ends with a short section on \textit{octopods}, an analogue for the Jeandel-Rao tilings of the decapods for Penrose tilings. The authors pose two questions: one about a specific configuration with a hole and whether it constitutes a configuration that can be extended to the entire plane except for the hole (an essential hole); the other asks how many octopods exist for the Jeandel-Rao tilings and how many of those are also essential holes.