Repository logo

A multi-proxy inference of Jōmon population dynamics using bayesian phase models, residential data, and summed probability distribution of 14C dates

Accepted version



Change log


Crema, ER 
Kobayashi, K 


We introduce a new workflow for analysing archaeological frequency data associated with relative rather than absolute chronological time-stamps. Our approach takes into account multiple sources of uncertainty by combining Bayesian chronological models and Monte-Carlo simulation to sample possible calendar dates for each archaeological entity. We argue that when applied to settlement data, this combination of methods can bring new life to demographic proxies that are currently under-used due to their lack of chronological accuracy and precision, and provide grounds for further exploring the limits and the potential of the so-called “dates as data” approach based on the temporal frequency of radiocarbon dates. Here we employ this new workflow by re-examining a legacy dataset that has been used to describe a major population rise-and-fall that occurred in central Japan during the Jōmon period (16,000 - 2,800 cal BP), focusing on the temporal window between 8,000 and 3,000 cal BP. To achieve this goal we: 1) construct the first Bayesian model of forty-two Jōmon ceramic typology based cultural phases using a sample of 2,120 radiocarbon dates; 2) apply the proposed workflow on a dataset of 9,612 Jōmon pit-dwellings; and 3) compare the output to a Summed Probability Distribution (SPD) of 1,550 radiocarbon dates from the same region. Our results provide new estimates on the timing of major demographic fluctuations during the Jōmon period and reveal a generally good correlation between the two proxies, although with some notable discrepancies potentially related to changes in settlement pattern.



Bayesian chronological modelling, Monte-carlo simulation, Jomon chronology, Prehistoric demography

Journal Title

Journal of Archaeological Science

Conference Name

Journal ISSN


Volume Title



Elsevier BV
European Research Council (801953)
ERC Starter Grant (Project N. 801953)