Skip to content

Coarsening

Version
v1 · 2026-08-24 · History
Prime #
705
Origin domain
Chemistry & Materials Science
Subdomain
phase dynamics → Chemistry & Materials Science

Core Idea

Coarsening is the drift of a population of discrete units toward fewer and larger because the cost is concentrated at the boundary (scaling with surface) while content scales with volume — so small units pay a higher cost per content, and material flows from small to large.

How would you explain it like I'm…

Big Bubbles Eat Small

When soap bubbles sit together in the sink, the little ones slowly disappear and the big ones get bigger. That happens because the skin of a bubble costs something, and little bubbles have too much skin for how little they hold. So little ones lose out and big ones win, until just a few big bubbles are left.

Fewer, Bigger Over Time

Coarsening is when a bunch of small things slowly turn into fewer, bigger things over time. The reason is that the cost is at the edges — the skin or border of each unit — not the inside. Small things have a lot of edge compared to how much they hold, so they're 'expensive' that way, while big things are more efficient. Because of that, material drifts from the small units into the big ones, either by leaking across or by merging, until you're left with fewer, larger units holding the same total amount with much less edge overall.

Boundary-Cost Merging

Coarsening is the pattern where a population of discrete units evolves toward fewer and larger units over time, because the cost is concentrated at the boundary between units, not in the bulk. Smaller units pay a higher boundary-to-bulk ratio per unit of content, and that asymmetry drives material, energy, or organizational mass to flow from small units into large ones — either by transfer through the medium or by direct merger. The boundary cost (surface energy, interface friction, coordination overhead) scales with surface area, while content scales with volume, so the surface-to-content ratio falls as a unit grows: bigger is cheaper per unit content. Given a transport mechanism and enough time, the system trends toward one large unit, though real constraints usually stop it short. It's distinct from preferential attachment ('rich get richer'): coarsening is driven purely by small units paying relatively more at the boundary, and needs no attractiveness rule at all.

 

Coarsening is the structural pattern by which a population of discrete units evolves toward fewer and larger units over time because the cost is concentrated at the boundary between units, not in the bulk. Smaller units pay a higher boundary-to-bulk ratio per unit content; this asymmetry drives material, energy, or organisational mass to flow from small units into large ones — by transfer through the medium or by direct merger — until the population is reduced to fewer, larger units that together carry the same total content with less aggregate boundary. The commitments are five. The system is populated by discrete units with a definite count and size distribution. There is a boundary between units and the surrounding medium carrying a definite cost — surface energy, interface friction, coordination overhead, regulatory overhead. That boundary cost scales with surface while content scales with volume, so the surface-to-content ratio falls as a unit grows and small units carry a higher cost per unit content. There is a transport mechanism by which content moves between units — diffusion in materials, capital and labour flows in economics, member migration in social systems, refactoring in software. And the system has time to evolve toward lower total boundary, the asymptotic state being one large unit, though kinetic constraints typically stop the process short. The pattern is structurally distinct from preferential attachment: coarsening is driven by the boundary-cost asymmetry between small and large — already-small units pay relatively more — not by a rich-get-richer rule, and it can occur with no attractiveness asymmetry at all, purely from interface-cost minimisation.

Broad Use

  • Materials science (origin): Ostwald ripening and grain coarsening, with LSW theory giving a power-law growth of mean size.
  • Biology: lipid-droplet coarsening and ecological-patch coarsening as edge-to-area ratio shapes patch persistence.
  • Urban form: small land parcels aggregate into larger ones as per-parcel fixed costs fall on a per-area basis.
  • Economics: small firms consolidate when fixed costs of operation are high relative to variable costs.
  • Software architecture: small services consolidate when per-service overhead is high relative to the work — the same logic as grain coarsening.
  • Information: sparsely-used categories are absorbed into larger neighbours when per-category maintenance outweighs bulk usage.

Clarity

Forces the cost-locus question: is consolidation driven by boundary cost (coarsening), an attractiveness rule (preferential attachment), or bulk cost (economies of scale)? The same observable admits opposite remedies.

Manages Complexity

Compresses a wide class of consolidation dynamics into five parameters — size distribution, boundary cost, bulk cost, transport, and kinetic regime — that transfer across substrates.

Abstract Reasoning

The boundary-cost-dominated equilibrium is a single unit, so observed plurality is kinetic, not thermodynamic; balanced against a bulk diseconomy, the two opposed scalings set an optimal unit size.

Knowledge Transfer

  • Economics: LSW scaling ports to industry consolidation, with antitrust acting as the boundary-pinning that arrests it.
  • Software: grain-coarsening intuition ports to service consolidation, with enforced bounded contexts as the pinning.
  • Ecology to urban form: the small-patches-lose-first diagnostic maps to small-parcels-bear-more-overhead.

Example

A polycrystalline metal anneals as atoms diffuse from small grains to large; the mean grain size grows as a power of time, and second-phase particles can pin the boundaries and freeze the distribution short of a single grain.

Relationships to Other Abstractions

Local relationship map for CoarseningParents appear above the current abstraction, mutual partners to the right, and children below. Node labels state whether each abstraction is prime or domain-specific; colors identify relation types.CoarseningPRIMEPrime abstraction: Scaling and Scale Dependence — is a kind ofScaling andScale DependencePRIMEDomain-specific abstraction: Ostwald Ripening — is a kind ofOstwald RipeningDOMAINDomain-specific abstraction: Reeb graph — is a kind ofReeb graphDOMAIN

Current abstraction Coarsening Prime

Parents (1) — more general patterns this builds on

  • Coarsening is a kind of Scaling and Scale Dependence Prime

    Coarsening is 'one specific dynamical member' of the scale-dependence family — the boundary-cost-driven consolidation of a population toward fewer-and-larger over time, with characteristic kinetics.

Children (2) — more specific cases that build on this

  • Ostwald Ripening Domain-specific is a kind of Coarsening

    Ostwald Ripening is the curvature-solubility and diffusion-mediated species of Coarsening in which large dispersed units grow as small units dissolve through a matrix.

  • Reeb graph Domain-specific is a kind of Coarsening

    The proposed strict upward parent is prime:coarsening.

Hierarchy path (1) — routes to 1 parentless root

Not to Be Confused With

  • Coarsening is not Diseconomies of Scale because coarsening is driven by boundary cost falling with size whereas diseconomy is bulk cost rising with size; they are opposed forces that jointly set the optimal unit size.
  • Coarsening is not Economies of Scale because economies lower the bulk (per-content) cost whereas coarsening turns on the boundary (per-unit) cost — a different cost-locus with a different remedy.
  • Coarsening is not generic Scaling and Scale Dependence because scale dependence is the broad fact that behaviour changes with size whereas coarsening is the specific consolidation dynamic boundary-cost asymmetry produces over time.