In a news article published by this week’s Science magazine, Wealth may have driven the rise of today’s religions, Lizzie Wade writes:
Today’s most popular religions all have one thing in common: a focus on morality. But the gods didn’t always care whether you are a bad person. Researchers have long puzzled over when and why religions moved away from a singular focus on ritual and began to encourage traits such as self-discipline, restraint, and asceticism. Now, a new study proposes that the key to the rise of so-called moralizing religions was, of all things, more wealth.
She refers to the article, Increased Affluence Explains the Emergence of Ascetic Wisdoms and Moralizing Religions, by Nicolas Baumard, Alexandre Hyafil, Ian Morris, and Pascal Boyer just published in Current Biology.
I know two of the authors personally, and with another one we have communicated by e-mail. All of them are good people and excellent scientists, but I have both good and bad things to say about this article.
Let’s start with the good news. What is novel about the study is that the authors advance several rival hypotheses and then bring in historical data to evaluate the relative merits of each. In most previous analyses of this kind, researchers have been limited to ‘static’ data. Such data provide us with a ‘snapshot’ of where different societies are (or were), but do not tell us where they came from, or what happened to them after the snapshot was taken. The highly productive anthropological databases, George Murdock’s Ethnographic Atlas and the Standard Cross-Cultural Sample, put together by Douglas White and George Murdoch, which provided data for literally hundreds (perhaps even thousands) of articles, are of this kind. They have been very useful, but not as useful as dynamical databases that trace the development of societies over time.
Baumard and co-authors, instead, looked at the dynamics. They selected eight most advanced areas of antiquity, ranging from Mesopotamia to Mesoamerica, and looked at the time trajectories of several variables in each area. As I have explained in previous blogs (for example, here), moving from static to dynamic databases will allow us to test theories much more robustly. Because causes typically precede the effects, being able to resolve how things change with time is key to testing mechanisms of social evolution.
So far so good. But now comes the criticism. Although I like where Baumard et al. are going, I am much more skeptical of their first steps. I’ll try to make my critique constructive, because I share the goals of the authors. However, getting where we want will require much more work.
First, let’s think of what Baumard et al. want to explain. It’s the rise of World Religions during the Axial Age (which most scholars date 800–200 BC, although these authors narrow it down to 500–300 BC). Baumard et al. select three Axial developments: Greece (e.g., Stoicism, Skepticism), North India (e.g. Buddhism, Jainism), and North China (Confucianism, Daoism). They treat them as all-or-nothing developments, which simply doesn’t make sense.
For example, they exclude Zoroastrianism from their list of Axial Religions. Yet, Zoroastrianism is arguably one of the first Monotheistic religions. Contrary to what these authors say, most scholars treat it as an Axial religion (and I agree with that). Baumard et al. also exclude Egypt from their list of Axial breakthroughs. That’s wrong, wrong. This is a big subject, so I’ll blog about it later.
The problem with this binary, all-or-nothing approach is that it puts the actual history into a Procrustean bed. It seems to me that it would be much better to use a more nuanced approach – gathering data on different aspects of the Axial breakthroughs separately. This is the approach that we have adopted in our Seshat database. We score separately such things as morally concerned supernatural beings, monotheism, concern for equity and against oppression, etc. The Axial transition was really a multidimensional development, and it wasn’t concentrated within 200 years in just three locations. By abandoning binary, black-and-white characterizations we achieve a richer dataset that will allow us to test various theories much better.
I also must take issue with the explanatory variables that Baumard et al. used. Their algorithm for calculating the main one, energy capture per capita, is never clearly explained (and I delved into the Supporting Materials to check it). I much admire Ian Morris’s vision and pioneering spirit, but as he himself admits in numerous places in The Measure of Civilization, his estimates are very tentative. And while the overall trajectory that he traces from the Mesolithic to the present day is largely correct, the details during the Iron Age (which is the period relevant to the Axial Age developments) are much shakier than the article in Current Biology portrays. In fact, I think that the picture will change quite a lot as we refine these estimates.
So I wouldn’t yet jump to the conclusion that it was really affluence that resulted in the rise of World Religions. In Part II I will continue with my critical comments, as well as suggestions as to how we can (and will) improve on the Baumard et al. study.