Flexible membranes should crumple. Take a sheet of paper, remove its stiffness, and it collapses into a ball. This is what thermal fluctuations do to elastic surfaces in statistical mechanics — they destroy long-range order. The Mermin-Wagner theorem tells you that continuous symmetries break in low dimensions. A two-dimensional membrane embedded in three-dimensional space shouldn't maintain its flat phase without bending rigidity to enforce it.
Chen, Gandikota, Kim, and Cacciuto show that self-avoidance alone prevents crumpling (arXiv:2602.21714). The surface can't pass through itself, and this single constraint — no self-intersection — is enough to keep the membrane flat in the thermodynamic limit. The size exponent ν=1 regardless of the degree of self-avoidance. Strong or weak, the constraint has the same qualitative effect. Flatness isn't proportional to the constraint's strength; it's a binary consequence of its presence.
This settles a decades-old debate where theory and simulation disagreed. Theoretical arguments predicted crumpling for surfaces without bending rigidity. Simulations consistently found flat phases. The resolution: self-avoidance is a fundamentally different kind of constraint from bending rigidity. Rigidity fights crumpling by making deformations energetically expensive. Self-avoidance fights crumpling by making certain configurations geometrically impossible. The membrane doesn't resist folding because folding costs energy — it resists folding because the folds would require the surface to occupy the same space twice.
The distinction matters because it suggests two categories of order-maintenance: energetic (the system prefers order) and topological (the system can't access disorder). Bending rigidity is energetic — at high enough temperature, it loses. Self-avoidance is topological — no temperature overcomes it. The membrane stays flat not because it wants to, but because the alternative is geometrically forbidden.
There's something clean about constraints that work by elimination rather than preference. A system that maintains order because disordered states don't exist is more robust than one that maintains order because ordered states are cheaper. The robustness comes from the structure of the configuration space itself, not from the energy landscape within it. Shape the space, and the dynamics take care of themselves.