Thomas Schelling’s 1969 model of residential segregation is one of social science’s most quoted results: give agents on a checkerboard only a mild preference for having some same-type neighbors, let them move when they’re unhappy, and the whole system reliably segregates into large, sharply divided clusters, even though not one individual agent wanted or intended that outcome. It’s since become genuinely, formally connected to statistical physics — not as a metaphor, but as an actual mathematical mapping that physicists have worked out in detail. That connection invites an obvious next comparison: materials scientists have their own famous story about weak, local preferences producing sharp, ordered structure, in block copolymers that self-assemble into precise nanoscale patterns. The connection between Schelling’s model and physics is real. The connection to block copolymers specifically, though, points to the wrong branch of physics.
Scientific Foundation
Schelling’s model places two types of agents on a lattice, each with a threshold for how many same-type neighbors they need to feel comfortable; unhappy agents relocate to any open spot that satisfies their preference. The striking, well-replicated finding is that even very mild individual preferences, agents who’d be perfectly content as a minority in a mixed neighborhood, reliably produce nearly complete segregation at the aggregate level once the system settles. This isn’t just a loose analogy to physics anymore. A landmark PNAS paper by Vinkovic and Kirman built a direct mathematical bridge, replacing Schelling’s economic concept of “utility” with the physics concept of energy and treating the whole system as a clustering process analogous to particles separating in a physical mixture. More recent statistical physics work goes further, formally analyzing Schelling-type dynamics as a Blume-Emery-Griffiths spin-1 system, a specific, well-studied model from magnetism and alloy theory. This is genuine, rigorous common ground, not wordplay.
Cross-Domain Connection
Block copolymers are materials built from two chemically distinct polymer segments, or blocks, covalently bonded together into a single molecule. Left to their own preferences, the two blocks want to separate, driven by a measurable thermodynamic incompatibility captured in the Flory-Huggins interaction parameter. But they can’t actually pull apart into two separate bulk regions, because they’re physically tethered to each other along the same chain. The result is microphase separation: instead of full separation, the material organizes into small, regular, geometrically ordered domains, lamellae, cylinders, spheres, or gyroid networks, depending on the relative size of the two blocks and the strength of their mutual dislike. This bounded, periodic, highly controllable structure is precisely what makes block copolymers industrially valuable, forming the basis of next-generation semiconductor lithography techniques capable of patterning features below 10 nanometers.
What Remains Undemonstrated
Here’s the honest, precise correction. What makes block copolymer microphase separation its own distinct, named phenomenon, worth distinguishing from ordinary phase separation, is specifically that covalent bonding constraint: the two blocks can never fully separate, no matter how strongly incompatible they become, because the chain connecting them physically caps how large any single domain can grow, setting the domain size according to the polymer’s molecular weight rather than allowing unlimited macroscopic separation. Schelling’s model has no equivalent constraint. Its agents move freely across the entire lattice, with nothing tethering a member of one group to a member of the other the way a polymer chain tethers its two blocks together. Left to run, Schelling-model segregation, like real residential segregation, characteristically grows into large clusters limited chiefly by the size of the city or lattice itself, not by any built-in structural cap analogous to a polymer’s chain length. That makes the model’s rigorously established physics kinship, treating it as a spin system or lattice gas describing clustering and phase separation generally, a much better fit for ordinary, unconstrained phase separation, the same broad statistical-mechanics family that describes plain oil-and-water demixing or metal alloys separating into distinct regions, than for block copolymers’ specifically bounded, self-limiting microphase behavior.
Why It Matters
That distinction carries a real, somewhat sobering implication rather than just technical housekeeping. Block copolymers are useful precisely because their self-assembly is self-limiting, constrained by a hard physical fact that guarantees domains stay small and orderly no matter how strong the underlying incompatibility gets. Cities have no equivalent constraint. There’s no covalent bond capping how large a segregated neighborhood can grow once the same basic dynamics that Schelling identified start operating — which means the more accurate physics analogy for residential segregation isn’t the neat, bounded, engineered patterns materials scientists prize in copolymer films. It’s the older, blunter, unconstrained kind of phase separation, the sort that, left alone, simply keeps growing until it runs into the edge of the system itself.
Human Dimension
There’s a genuine, worthwhile lesson in tracing a tempting comparison carefully enough to discover it’s aiming at the wrong branch of a field it was otherwise right to reach for. Schelling’s model earning a real, rigorous place in statistical physics is a remarkable fact on its own, worth taking seriously rather than diluting with an imprecise second comparison. The honest version of the story is less tidy than “cities self-assemble like nanomaterials” — but it’s more useful, because it correctly identifies which physical property, boundedness, is actually doing the protective work in one system and is conspicuously, consequentially absent from the other.
Sources:
1. Wikipedia — “Schelling’s model of segregation” — https://en.wikipedia.org/wiki/Schelling’s_model_of_segregation
2. PNAS — “A physical analogue of the Schelling model” (Vinkovic & Kirman) — https://www.pnas.org/doi/10.1073/pnas.0609371103
3. arXiv — “Statistical physics of the Schelling model of segregation” — https://arxiv.org/pdf/0707.1681
4. arXiv — “Symmetric preferences, asymmetric outcomes: Tipping dynamics in an open-city segregation model” — https://arxiv.org/html/2602.09795
5. arXiv — “Relocation without preference: A destination-agnostic Schelling-type metapopulation model” — https://arxiv.org/pdf/2604.24998
6. PMC (National Institutes of Health) — “Block Copolymers: Synthesis, Self-Assembly, and Applications” — https://pmc.ncbi.nlm.nih.gov/articles/PMC6418972/
7. ScienceDirect — “Directed block copolymer self-assembly for next-generation lithography” — https://www.sciencedirect.com/science/article/abs/pii/S2589597426000183
8. npj Soft Matter (Nature) — “Morphological and phase behaviors of symmetric diblock copolymers: insights from coarse-grained molecular dynamics simulations” — https://www.nature.com/articles/s44431-025-00010-9
9. PMC (National Institutes of Health) — “Exploring Microphase Separation in Semi-Fluorinated Diblock Copolymers: A Combined Experimental and Modeling Investigation” — https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12874174/
Idea originated at artificialideas.org. Article researched and written by Claude Sonnet 5. Published at artificialideas.org.
Artificial Ideas 