Guided Variation

Guided variation is a mechanism of cultural evolution identified by Boyd and Richerson in which individuals modify cultural traits through individual learning or experience before transmitting them to others. Unlike faithful imitation — where cultural traits are copied with high fidelity — guided variation introduces directional change into the cultural stream by filtering traits through each learner's individual problem-solving.

The mechanism is analogous to Lamarckian inheritance but with a critical difference: the modifications produced by guided variation are not automatically adaptive. An individual's learning may improve a trait, degrade it, or leave it unchanged. The direction of change depends on the learner's cognitive abilities, the structure of the problem, and the quality of environmental feedback. Where environmental feedback is reliable and immediate — as in toolmaking or agriculture — guided variation tends to produce cumulative improvement. Where feedback is noisy or delayed — as in medical beliefs or political strategies — guided variation may produce systematic error.

Guided variation interacts with other transmission biases in complex ways. When combined with conformity bias, it can accelerate local adaptation while preventing cross-population learning. When combined with prestige bias, it can concentrate innovation in high-status individuals, producing elite-driven cultural change. The net effect depends on the relative strengths of the biases and the structure of the social network through which transmission occurs.

The synthesizer's claim: guided variation is the forgotten engine of cultural progress. The field has fixated on social learning biases — conformity, prestige, content — because they are analytically tractable. But guided variation is where culture actually improves. Without it, cultural evolution is just a random walk through trait space. With it, cultural evolution becomes a hill-climbing algorithm. The question is not whether guided variation matters. The question is why we have so little empirical research on when, where, and how it operates.