This is a 2D-native reading of Robert Axelrod's model of cultural dissemination — the same interaction rule as the 3D grid-of-boxes version, but read here through four native 2D plots instead of one oblique 3D camera view. Each agent holds a culture vector of F features, each taking one of q traits:
culture(i) = (t1, …, tF), tk ∈ {0, …, q-1}
overlap(i,j) = (# features where t_i = t_j) / F
with prob. overlap(i,j): i copies j's trait on one feature where they differ
Higher similarity makes further convergence more likely — social influence grows with similarity.
- Spatial grid (top-left) — the physical 26×26 lattice, unavoidable since Axelrod regions are defined by spatial adjacency; colored by a hash of each agent's full culture vector.
- Culture-space scatter (top-right) — every agent re-plotted by its own (trait 1, trait 2) coordinates instead of its grid position. This is the model's actual state space: points pile onto the same spot once agents share those two traits, revealing cluster formation independent of where agents sit on the lattice — something the 3D version never shows.
- Order-parameter time series (bottom-left) — region count and largest-region share plotted against interaction step count, live. Region count is not monotonic: a copy that aligns i with j can simultaneously misalign i from a third neighbor k, splitting a region — this plot is where that shows up as a small uptick.
- Neighbor-overlap histogram (bottom-right) — the distribution of overlap(i,j) across every adjacent pair right now. A model converged to consensus collapses this to a single spike at overlap=1; a polarized, frozen grid instead shows a spread of intermediate values with almost nothing left at the extremes.
- q (traits per feature) — low q converges toward one dominant culture; high q crosses a critical threshold into many permanently distinct regions — polarization from purely local imitation.
Real-world relevance: the same local-imitation mechanism models language change, opinion polarization, and why populations settle into a handful of stable cultural regions rather than one smooth gradient or total noise.