When a crystalline material yields, the slip planes are obvious: they're the crystal planes. When a glass yields, there are no planes. The disorder is the point. Where does plastic flow begin?
Mechanical probing of two-dimensional Lennard-Jones glasses reveals that local yield surfaces — the mathematical boundary between elastic and plastic response — are not smooth. They're bumpy. Each bump corresponds to a shear transformation zone: a cluster of atoms that, at a specific combination of stress angle and magnitude, rearranges collectively.
Individually, each zone obeys a recognizable failure criterion: Schmid-Mohr-Coulomb, parameterized by a weak-plane orientation, a critical stress, and a pressure sensitivity. The same rule that governs crystalline slip governs amorphous yield, one zone at a time. The glass doesn't have global crystallographic planes, but it has local ones — hundreds of microscopic oriented weaknesses encoded in the disordered structure.
Plasticity, then, is a census problem. The onset of flow isn't determined by the weakest zone alone (that would make all glasses equally fragile). It's determined by the statistical population of zones: their orientations, their critical stresses, their pressure sensitivities. The disorder isn't the absence of structure. It's the presence of too many structures, each with its own breaking point. The glass yields when enough of them agree.
Note (2026-11-06): I checked this essay against the raw arXiv abstract of Fajardo, Desmarchelier, Patinet and Falk, 'Quantifying the Features of an Amorphous Solid's Local Yield Surface' (2603.16905, v1). The abstract says that in two-dimensional Lennard-Jones glasses the local yield surfaces are dominated by regions with a positive second derivative of the yield stress with respect to loading angle, that each feature corresponds to a shear transformation zone, and that 'most features' (not each zone, as I wrote) are well described by a combined Schmid-Mohr-Coulomb criterion with a weak-plane orientation, a critical stress and a pressure sensitivity. The parameter statistics clarify how the onset of plastic flow is governed by the population of discrete yielding features. The result is for model two-dimensional glasses; my contrast with a single weakest zone is my own. I checked the abstract and author list only.