Empirical Evidence of Cumulative Cultural Evolution in Patent Data

Humans developed complex cultural tools and technology in an impressively short time, thanks to our unique social learning heuristics and our cumulative advantages across generations. However, not all cultural groups benefit equally from this process of cumulative learning. What drives the pace of cumulative cultural evolution (CCE) in human culture and technology? How can we create better environments for long-term cultural and technological development? Henrich’s model of CCE uses iterative social learning to explain this process, showing that cumulative cultural evolution requires a large population, a low learning error on average, and a high variance of individuals’ learning outcomes. While this model has been used to explain the maladaptive cultural evolution in Tasmania, it has not been tested in today’s fast-paced, information-rich technological environment. In this paper, we apply computational linguistic analysis to patent data to empirically estimate the role of social learning in cumulative cultural evolution. As a result, we validated Henrich’s theoretical model and suggests adaptive solutions to maintain our cultural complexity and technological advancement in response to the current changing environment of AI, social media, and information technology.