% pubman genre = article @article{item_3244367, title = {{Population size limits the coefficient of variation in continuous traits affected by proportional copying error (and why this matters for studying cultural transmission)}}, author = {Premo, L. S.}, language = {eng}, issn = {1573-7764}, doi = {10.1007/s10816-020-09464-9}, publisher = {Springer}, address = {Berlin}, year = {2021}, date = {2021}, abstract = {{The accumulated copying error model (ACE) combines findings on the proportional nature of human visual perception error with cultural transmission theory. Previous studies have employed ACE to provide population-level expectations of the coefficient of variation (CV) of a continuous trait, such as the thickness of ceramic vessels. To date, empirical departures from expected CV values have been interpreted as evidence of biased cultural transmission or functional constraints without consideration for how population size might affect population-level cultural variation in the presence of proportional perception error. Here, I employ agent-based simulation experiments to investigate the ways in which population size affects the CV of a continuous cultural trait under different cultural transmission mechanisms. Results show that the CV of a continuous cultural trait is a function of its cultural equivalent N and effective population size (Ne) as well as the relative strength of cultural selection. The results also demonstrate that different combinations of N and cultural transmission yield identical CV values. The study highlights a new set of difficulties with inferring individual-level process{\textemdash}the mechanism(s) by which cultural information is transmitted among individuals{\textemdash}from population-level pattern{\textemdash}the CV of a continuous cultural trait. In light of these results, I identify and discuss one avenue through which we might improve our ability to infer past cultural transmission from archeological data.}}, journal = {{Journal of Archaeological Method and Theory}}, volume = {28}, pages = {512--534}, }