Complex societies and moralizing gods: Is cultural evolution Darwinian or Lamarckian?

Whitehouse et al. ignited a bit of a controversy with their recent Nature paper which concludes that “Complex societies precede moralizing gods throughout world history”. Their argument is that “belief in morally concerned supernatural agents culturally evolved to facilitate cooperation among strangers in large-scale societies”. They use a very large database to suggest that moralizing gods are unknown in various regions before the rise of complex societies. The idea seems to be that once people started organizing into larger and more complex polities, moralizing gods emerged as a tool to manage behaviour, minimize conflict, and generally keep everyone in line.

Critics were quick to respond, not so much on the paper’s conclusions or theoretical underpinnings, but rather on methodological grounds. I think there are also very good theoretical reasons to question the paper’s conclusion. But let’s discuss methods first.

Methodological concerns

Beheim et al. make the very good point that for nearly all the regions studied by Whitehouse et al, data on the presence of moralizing gods is completely missing before their first appearance. Archaeologically, this is a big problem. It would be one thing if Whitehouse et al. said that they had evidence for the absence of moralizing gods before the rise of social complexity. Their database however, mostly shows that there is no information whatsoever on moralizing gods before the appearance of social complexity. As we should all know by now, in archaeology especially, absence of evidence is not evidence of absence.

By turning their absence of evidence into evidence of absence for purposes of their analysis, Whitehouse et al. commit a serious methodological error. It is entirely possible that moralizing gods are present before large, complex societies in many, or even in all the regions studied. We just don’t know.

Meanwhile, Slingerland and others at the Database of Religious History raise concern about data quality. Of course, it isn’t unusual for area specialists to react badly when generalists try to consider their unique tiles in the context of a wider mosaic, and I was expecting to see some critique on that front. In this case though, the historians make very good points.

They show clearly that for at least some of the regions they include in their study, Whitehouse et al. rely heavily on broad syntheses rather than narrower regional sources for their coding, with the result that there is a fair bit of interpolation in their database, rather than actual observations.

In other cases, they use either viewpoints held by a minority of specialists, or fail to factor-in a real measure of data quality. These are all cardinal sins in the eye of the particularist, but even a woolly-headed generalist like me has to admit that they are serious concerns for this type of study.

Theoretical concerns

My own problems with the paper start with the idea that social complexity precedes moralizing gods, and that moralizing gods evolve as an adaptation to social complexity. I am not surprised to see the serious methodological critiques of the paper, because its conclusion doesn’t make intuitive evolutionary sense to me.

Unless we want to think that cultural evolution is completely Lamarckian, it is much simpler to think that the idea of moralizing gods appeared here and there, and allowed the elaboration of large, complex social structures where they had been impossible or improbable before.

If cultural evolution is Darwinian, moralizing gods didn’t evolve to do anything. The idea popped up at some point, as an innovation. Maybe at many points, at many times. Once it existed, it solved some of the problems of living in large, complex groups. Those groups could then exist.

Moralizing gods may well have existed for a long time. For all we know, they are a Paleolithic innovation. Combined with increased reliance on spatially stable plant resources (i.e. fields) in the Holocene, they could have allowed people to live together in larger numbers, and with more layers of organization and centralization.

Peter Turchin, one of the authors of the original paper, claims that SESHAT (their database) is a “killer of theories”. So far, based on their analysis and its critiques, I see no reason to kill the theory that cultural evolution is Darwinian rather than Lamarckian.