Generative inference for cultural evolution.

Area of Science:

• Cultural Evolution
• Human Behavior
• Social Learning

Background:

• Understanding social learning in human populations is a key challenge in cultural evolution.
• Theoretical work often lacks empirical data, focusing on idealized learning processes.
• Population genetics has advanced using generative frameworks for demographic and selection inference.

Purpose of the Study:

• To demonstrate the utility of generative inference for analyzing cultural evolution data.
• To apply data-driven methods to infer social learning strategies in human populations.
• To address challenges like equifinality in cultural data analysis.

Main Methods:

• Utilizing generative frameworks, including coalescent-based simulation and approximate Bayesian computation (ABC).
• Applying these methods to observed population-level frequency data.
• Inferring the presence or absence of hypothesized social learning strategies.

Main Results:

• Generative inference successfully applies to cultural evolution, using data to test learning strategies.
• The approach provides robust inferences without assuming equilibrium.
• Excluding inconsistent processes is a valuable outcome with sparse cultural data.

Conclusions:

• Generative inference offers a powerful, data-driven approach to studying cultural evolution.
• This framework enhances understanding of social learning mechanisms across human populations.
• The methodology aids in reconstructing the human cultural past.