The adoption of moralising religious beliefs may have originated earlier in history and by much smaller populations than a previous study has shown, reveals a fresh analysis of the original data, published in Nature.
In 2019, a study published in Nature systematically coded records from 414 societies and 30 regions around the world, spanning the past 10,000 years into the Seshat dataset. Through their analysis the authors claimed that beliefs in moralising gods coincided with the growth of large, complex societies of over 1 million people – a “megasociety threshold”.
Crucially, 61% of all moralising god observations in this data were labelled as “unknown”. The authors treated these “missing” values as the absence of belief.
New research, led by Dr Bret Beheim (Max Planck Institute for Evolutionary Anthropology, Germany) and co-authored by Dr Michael Muthukrishna and Dr Rachel Spicer (LSE), highlights that in this treatment of the data, populations such as Hawaii with a well-documented archaeological history, would only have an appearance of religious beliefs with the arrival of the Europeans who were there to record it. As such, their assumption produced a powerful forward bias, pushing the estimation of the adoption of religious beliefs to a much later date in history.
Beheim et al claim this treatment of the “Unknown” data is misleading and set about to re-analyse the Seshat data across the 12 key regions by moving the first appearance of a moralising god back by one century, which is the smallest time unit in the data. This minimal correction reversed their main result: moralising gods precede dramatic rises in social complexity.
Dr Bret Beheim, Max Planck Institute for Evolutionary Anthropology, said:
“The rise of quantitative approaches to study historical dynamics is a young and exciting field, but source provenance is as important as it has always been. We’ve shown how you deal with the data you don’t have can sometimes be more important than the data you do have.”
Dr Michael Muthukrishna, LSE, said:
“The relationship between moralising religious beliefs and the size and complexity of societies is hotly debated in cultural evolution. This re-analysis re-opens the debate: the belief in big, moralising supernatural punishers—from karma to Yahweh and Allah—may have co-evolved with large civilizations.”
This re-analysis has resulted in the authors of the original 2019 study to retract their paper.
Dr Rachel Spicer, LSE, said:
“How data is handled before it's analysed can completely change the results. In this case the mistake in data handling affected the results enough that the original study was retracted. We hope this work highlights the importance of data handling decisions for future research.”
The new research is available here: https://www.nature.com/articles/s41586-021-03655-4