Complex societies precede moralizing gods throughout world history

Complex societies precede moralizing gods throughout world history. Harvey Whitehouse, Pieter François, Patrick E. Savage, Thomas E. Currie, Kevin C. Feeney, Enrico Cioni, Rosalind Purcell, Robert M. Ross, Jennifer Larson, John Baines, Barend ter Haar, Alan Covey & Peter Turchin. Nature, Mar 20 2019. https://doi.org/10.1038/s41586-019-1043-4

Abstract: The origins of religion and of complex societies represent evolutionary puzzles1–8. The ‘moralizing gods’ hypothesis offers a solution to both puzzles by proposing that belief in morally concerned supernatural agents culturally evolved to facilitate cooperation among strangers in large-scale societies9–13. Although previous research has suggested an association between the presence of moralizing gods and social complexity3,6,7,9–18, the relationship between the two is disputed9–13,19–24, and attempts to establish causality have been hampered by limitations in the availability of detailed global longitudinal data. To overcome these limitations, here we systematically coded records from 414 societies that span the past 10,000 years from 30 regions around the world, using 51 measures of social complexity and 4 measures of supernatural enforcement of morality. Our analyses not only confirm the association between moralizing gods and social complexity, but also reveal that moralizing gods follow—rather than precede—large increases in social complexity. Contrary to previous predictions9,12,16,18, powerful moralizing ‘big gods’ and prosocial supernatural punishment tend to appear only after the emergence of ‘megasocieties’ with populations of more than around one million people. Moralizing gods are not a prerequisite for the evolution of social complexity, but they may help to sustain and expand complex multi-ethnic empires after they have become established. By contrast, rituals that facilitate the standardization of religious traditions across large populations25,26 generally precede the appearance of moralizing gods. This suggests that ritual practices were more important than the particular content of religious belief to the initial rise of social complexity.

---
Supernatural agents that punish direct affronts to themselves (for
example, failure to perform sacrifices or observe taboos) are commonly
represented in global history, but rarely are such deities believed to
punish moral violations in interactions between humans
2
. Recent mil-
lennia, however, have seen the rise and spread of several ‘prosocial
religions’, which include either powerful ‘moralizing high gods’ (MHG;
for example, the Abrahamic God) or more general ‘broad supernatu-
ral punishment’ (BSP) of moral transgressions (for example, karma in
Buddhism)
9,12,16–18
. Such moralizing gods may have provided a crucial
mechanism for overcoming the classic free-rider problem in large-scale
societies
11
. The association between moralizing gods and complex
societies has been supported by two forms of evidence: psychological
experiments
3,6,27,28
 and cross-cultural comparative analyses
7,11,14–18,20
.
The contributions of theistic beliefs to cooperation, as well as the his-
torical question of whether moralizing gods precede or follow the estab-
lishment of large-scale cooperation, have been much debated
9,10,12,23,24
.
Three recent studies that explicitly model temporal causality have come
to contrasting conclusions. One study, which applied phylogenetic
comparative methods to infer historical changes in Austronesian reli-
gions, reported that moralizing gods (BSP but not MHG) preceded the
evolution of complex societies
16
. The same conclusion was reached
in an analysis of historical and archaeological data from Viking-age
Scandinavia
18
. By contrast, another study of Eurasian empires has
reported that moralizing gods followed—rather than preceded—the
rise of complex, affluent societies
20
. However, all of these studies are
restricted in geographical scope and use proxies for social complexity
that the authors themselves concede are ‘very crude’
20
 (for example,
thebinary classification of societies as ofeither high or low complexity).
To overcome these limitations, we used ‘Seshat: Global History
Databank’
29
,a repository of standardized data on social structure, reli-
gion and other domains for hundreds of societies throughout world his-
tory. In contrast to other databases that attempt to model history using
contemporary ethnographic data, Seshat directly samples over time as
well as space. Seshat also includes estimates of expert disagreement and
uncertainty, and uses more-detailed variables than many databases.
To test the moralizing gods hypothesis, we coded data on 55var-
iables from 414polities (independent political units) that occupied
30geographical regions from the beginning of the Neolithic period
to the beginning of Industrial and/or colonial periods (Fig.1 and
Supplementary Data). We used a recently developed and validated
measure of social complexity thatcondenses 51social complexity
variables (Extended Data Table5) into a single principal component
thatcaptures three quarters of the observed variation, which we call
‘social complexity’
8
. The remaining four variables were selected to test
the MHG and BSP subtypes of the moralizing gods hypothesis. The
MHG variable was coded following the MHG variable used as stand-
ardin the literature on this topic
11,14–17,30
, whichrequires that a high
god who created and/or governs the cosmos actively enforces human
morality. Because the concept of morality is complex, multidimensional
and in some respects culturally relative—and because not all moralizing
gods are ‘high gods’—we also coded three different variables related to
BSP that are specifically relevant to prosocial cooperation: reciprocity,
fairness and in-group loyalty. For analysis, these three variables were
combined into a single BSP variable. TheMethods, Supplementary
Information and http://seshatdatabank.info/methods/codebook pro-
vide further methodological details, definitions and justifications,
including adiscussion of the relationship between MHG, BSP and big
gods.
Figure1 and Extended Data Table1 show the temporal and geo-
graphical distribution of the appearance of moralizing gods in our
sample. Although societies in all 30regions possessed beliefs about
appeasing supernatural agents through the performance of rituals, in 10
out of the 30regions, there was no evidence for moralizing gods before
their introduction under colonial powers. The remaining 20regions displayed a diverse range of 15different systems of belief in moraliz-
ing gods: in some, the first evidence of moralizing gods came in the
form of MHG and in others it came in the form of BSP (Extended
Data Table1). The first appearance of moralizing gods in our sample 
was in Egypt, where the concept of supernatural enforcement of
Maat (order) is attested by the Second Dynasty, around 2800. This
was followed by sporadic appearances in local religions throughout
Eurasia (Mesopotamia (around 2200), Anatolia (around 1500)
and China (around 1000)) before the wider spread of transnational
religions began during the first millennium  with Zoroastrianism
and Buddhism, followed later by Christianity and Islam. Although
Christianity and Islam would eventually become the most widespread
religions, local forms of moralizing gods were present well before they
arrived in most regions (for example, Roman gods were believed to
punish oath-breaking from as early as 500, almost a millennium
before Christianity was adopted as the official Roman religion). The
diverse range of religious systems represented in our global sample
makes it possible to draw more general conclusions about religion than
have previously been possible.
Although our sampling scheme reduces non-independence, our
polities still cannot be considered statistically independent because
of the historical relationships among them. We controlled for these
using a logistic regression model to account for temporal, geographi-
cal and cultural dependencies in the global distribution of moralizing
gods (seeMethods). This analysis revealed that social complexity was
a stronger predictor of moralizing gods than temporal, geographical or
linguistic relationships, and remained highly significant even after con-
trolling for these relationships (z=6.8, degrees of freedom (d.f.)=800,
P<1×10
−11
; Extended Data Table2), conceptually replicating pre-
vious studies
7,11,14,15
.
The moralizing gods hypothesis posits a ‘statistical causal relation-
ship’
10
 in which moralizing gods facilitate the evolution of complex
societies
9,12,16–18
. This indicates that, on average, social complexity
should increase more rapidly following the appearance of moralizing
gods. To test this prediction, we conducted time-series analyses of the
12regions for which social complexity data were available both before
and after the appearance of moralizing gods (Fig.2, Extended Data
Table1 and Extended Data Fig.1). Notably, average rates of increase
of social complexity were over five times greater before—not after—
the appearance of moralizing gods (paired t-test, t=−6.6, d.f.=199,
P<1×10
−9
; Fig.2). This trend was significant both globally and
individually for 10 out of the 12regional time-series analyses (Extended
Data Table1 and Extended Data Fig.1). None of these 12regions dis-
played a significantly greater rate of increase in social complexity after
the appearance of moralizing gods than before. Robustness analyses
showed that our primary finding of higher rates of increasing social
complexity before the appearance of moralizing gods was present
regardless of the type of moralizing gods (MHG or BSP), the choice of
variables used to estimate social complexity, uncertainty in the timing
of appearance of moralizing gods, or the time windows used to estimate
rates of change in social complexity (Extended Data Table4).
In summary, although our analyses are consistent with previous stud-
ies that show an association between moralizing gods and complex
societies
7,11,14–18,30
, we find that moralizing gods usually follow—rather
than precede—the rise of social complexity. Notably, most societies that
exceeded a certain social complexity threshold developed a conception
of moralizing gods. Specifically, in 10 out of the 12regions analysed,
the transition to moralizing gods came within 100years of exceeding a
social complexity value of 0.6 (which we call a megasociety, as it corre-
sponds roughly to a population in the order of one million; Extended
Data Fig.1). This megasociety threshold does not seem to correspond
to the point at which societies develop writing, which might have sug-
gested that moralizing gods were present earlier but were not preserved
archaeologically. Although we cannot rule out this possibility, the fact
that written records preceded the development of moralizing gods in
9 out of the 12regions analysed (by an average period of 400years;
Supplementary Table2)—combined with the fact that evidence for
moralizing gods is lacking in the majority of non-literate societies
2
—
suggests that such beliefs were not widespread before the invention
of writing. The few small-scale societies that did display precolonial
evidence of moralizing gods came from regions that had previously
been used to support the claim that moralizing gods contributed to the
rise of social complexity (Austronesia
16
 and Iceland
18
), whichsuggests
that such regions are the exception rather than the rule.

Conversely, of the societies in the tenregions that did not develop
precolonial moralizing gods, only one exceeded the megasociety
threshold (the short-lived Inca Empire, social complexity=0.61).
This suggests that, even if moralizing gods do not cause the evolution
of complex societies, they may represent a cultural adaptation that is
necessary to maintain cooperation in such societies once they have
exceeded a certain size, perhaps owing to the need to subject diverse
populations in multi-ethnic empires to a common higher-level power
9
.
This may explain why moralizing gods spread when large empires con-
quer smaller—but still complex—societies (for example, the Spanish
conquest of the Incas). In some cases, moralizing doctrines may have
helped to stabilize empires, while also limiting further expansion;for
example, when emperor Ashoka adopted Buddhism and renounced
war following his final conquest of the Kalinga Kingdom that estab-
lished the maximum extent of the Mauryan empire.
Although our results do not support the view that moralizing gods
were necessary for the rise of complex societies, they also do not
support a leading alternative hypothesis that moralizing gods only
emerged as a byproduct of a sudden increase in affluence during a
first millennium  ‘Axial Age’
19–22
. Instead, in three of our regions
(Egypt, Mesopotamia and Anatolia), moralizing gods appeared before
1500. We propose thatthe standardization of beliefs and practices
via high-frequency repetition and enforcement by religious authorities
enabled the unification of large populations for the first time, establish-
ing common identities across states and empires
25,26
. Our data show
that doctrinal rituals standardized by routinization (that is, those per-
formed weekly or daily) or institutionalized policing (religions with
multiple hierarchical levels) significantly predate moralizing gods, by an
average of 1,100years (t=2.8, d.f.=11, P=0.018; Fig.2a). Doctrinal
rituals precede moralizing gods in 9 out of the 12regions analysed,
and even precede written records in 6 of these cases (by as much as
4,000years in the case of Çatalhöyük in Anatolia; see Supplementary
Table2). Although analyses of rates of change of social complex-
ity before and after the appearance of doctrinal rituals do not offer
conclusive support for the hypothesis that doctrinal rituals facilitate
increasing social complexity (Extended Data Table3), these data do at
least suggest that doctrinal rituals led to the establishment of large-scale
religious identities. In the future, higher-quality and higher-resolution
archaeological data may allow for a more nuanced understanding of
the timing and possible coevolution of the rise of doctrinal rituals and
moralizing gods. Such data appear unlikely to affect our primary claim
that complex societies preceded moralizing gods, but this is an empiri-
cal question open to future testing.
We demonstrate how quantifying cultural characteristics of past
societies can contribute to longstanding debates about the evolution
of social complexity. Our results suggest that belief in moralizing gods
was not the only or even the main factor that enabled the expansion
of human societies, but may have occurred along with other features
of ritual practices and religion to facilitate cooperation in increasingly
complex social systems. In particular, an increase in ritual frequency
and doctrinal control may have facilitated the establishment of large-
scale collective identities before the spread of beliefs in moralizing
gods. Thus, when it comes to the initial rise of social complexity, how
you worship may ultimately have been more important than who you
worship.