Cultural Diffusion Simulator — Axelrod's Model of Cultural Dissemination

Trait Diversity

Traits per feature (q) 8

Features per agent (F) 5

Interaction

Interactions / frame 40

Live Readouts

Interaction steps

Cultural regions

—

Largest region

—

Mean neighbor overlap

—

How it works

This is Robert Axelrod's 1997 model of the dissemination of culture ("The Dissemination of Culture: A Model with Local Convergence and Global Polarization"). Each agent on the grid holds a culture vector of F features, and each feature can take one of q possible traits — no meaning, just a discrete label:

culture(i) = (t1, t2, …, tF),  tk ∈ {0, 1, …, q-1}

Every interaction step: pick a random agent i and a random neighbor j. Their cultural overlap is the fraction of features they already share:

overlap(i,j) = (# features where t_i = t_j) / F

With probability equal to that overlap, i and j interact: a random feature on which they still differ is chosen, and i copies j's trait for it. Higher similarity makes further convergence more likely — social influence grows with similarity, exactly the mechanism behind real cultural assimilation and homogenization.

q (traits per feature) — the diversity of the trait pool. Low q converges fast into one dominant culture (homogenization); high q crosses a critical threshold q_c where local convergence and finite trait diversity instead freeze the grid into many stable, mutually foreign cultural regions — polarization from purely local imitation, no central authority involved.
F (features) — more shared dimensions of culture make partial overlap, and therefore convergence, more likely.
4-neighbor / 8-neighbor — von Neumann vs. Moore neighborhoods; more neighbors speeds mixing.
Two adjacent agents belong to the same rendered region once their culture vectors are identical in every feature — a stable cultural boundary, exactly as in the original paper's "zone" definition.

Real-world relevance: this same local-imitation mechanism is used to model language change, opinion polarization, and why nation-states end up with a handful of stable dialect/culture regions rather than one smooth gradient or total fragmentation.

Each block = one agent's culture vector Same color = identical culture (one region)

⚠ Couldn't load the 3D engineThree.js failed to load from the CDN. Check your connection and reload.