ArticlePDF Available


Where a newly-married couple lives, termed post-marital residence, varies cross-culturally and changes over time. While many factors have been proposed as drivers of this change, among them general features of human societies like warfare, migration and gendered division of subsistence labour, little is known about whether changes in residence patterns exhibit global regularities. Here, we study ethnographic observations of post-marital residence in societies from five large language families (Austronesian, Bantu, Indo-European, Pama-Nyungan and Uto-Aztecan), encompassing 371 ethnolinguistic groups ranging widely in local ecologies and lifeways, and covering over half the world's population and geographical area. We apply Bayesian comparative methods to test the hypothesis that post-marital residence patterns have evolved in similar ways across different geographical regions. By reconstructing past post-marital residence states, we compare transition rates and models of evolution across groups, while integrating the historical descent relationships of human societies. We find that each language family possesses its own best fitting model, demonstrating that the mode and pace of post-marital residence evolution is lineage-specific rather than global.
Accepted Manuscript
Post-Marital Residence Patterns Show Lineage-Specific
Jiří C. Moravec, Quentin Atkinson, Claire Bowern, Simon J.
Greenhill, Fiona M. Jordan, Robert M. Ross, Russell Gray,
Stephen Marsland, Murray P. Cox
PII: S1090-5138(17)30383-5
DOI: doi:10.1016/j.evolhumbehav.2018.06.002
Reference: ENS 6213
To appear in: Evolution and Human Behavior
Received date: 29 November 2017
Revised date: 5 June 2018
Accepted date: 5 June 2018
Please cite this article as: Jiří C. Moravec, Quentin Atkinson, Claire Bowern, Simon J.
Greenhill, Fiona M. Jordan, Robert M. Ross, Russell Gray, Stephen Marsland, Murray
P. Cox , Post-Marital Residence Patterns Show Lineage-Specific Evolution. Ens (2018),
This is a PDF file of an unedited manuscript that has been accepted for publication. As
a service to our customers we are providing this early version of the manuscript. The
manuscript will undergo copyediting, typesetting, and review of the resulting proof before
it is published in its final form. Please note that during the production process errors may
be discovered which could affect the content, and all legal disclaimers that apply to the
journal pertain.
Post-Marital Residence Patterns Show Lineage-Specific
Jiří C. Moraveca, Quentin Atkinsonb, Claire Bowernc, Simon J. Greenhilld,e,
Fiona M. Jordanf, Robert M. Rossf,g,h, Russell Grayb,e, Stephen Marslandi,
Murray P. Coxa,
aStatistics and Bioinformatics Group, Institute of Fundamental Sciences, Massey
University, Palmerston North, New Zealand
bDepartment of Psychology, University of Auckland, Auckland, New Zealand
cDepartment of Linguistics, Yale University, New Haven, CT 06511, Connecticut, USA
dARC Centre of Excellence for the Dynamics of Language, Australian National University,
Canberra, ACT 0200, Australia
eMax Planck Institute for the Science of Human History, D-07745 Jena, Germany
fDepartment of Anthropology and Archaeology, University of Bristol, Bristol, BS8 1TH, UK
gInstitute for Cognitive and Evolutionary Anthropology, School of Anthropology and
Museum Ethnography, University of Oxford, Oxford, OX1 2JD, UK
hARC Centre of Excellence in Cognition and its Disorders, Department of Psychology,
Royal Holloway, University of London, Surrey, TW20 0EX, UK
iSchool of Mathematics and Statistics, Victoria University of Wellington, Wellington, New
Where a newly-married couple lives, termed post marital residence, varies cross-
culturally and changes over time. While many factors have been proposed as
drivers of this change, among them general features of human societies like war-
fare, migration and gendered division of subsistence labour, little is known about
whether changes in residence patterns exhibit global regularities. Here, we study
ethnographic observations of post-marital residence in societies from five large
language families (Austronesian, Bantu, Indo-European, Pama-Nyungan and
Uto-Aztecan), encompassing 371 ethnolinguistic groups ranging widely in local
ecologies and lifeways, and covering over half the world’s population and geo-
graphical area. We apply Bayesian comparative methods to test the hypothesis
that post-marital residence patterns have evolved in similar ways across differ-
ent geographical regions. By reconstructing past post-marital residence states,
Murray P. Cox and Stephen Marsland
Preprint submitted to Evolution and Human Behaviour June 12, 2018
we compare transition rates and models of evolution across groups, while in-
tegrating the historical descent relationships of human societies. We find that
each language family possesses its own best fitting model, demonstrating that
the mode and pace of post-marital residence evolution is lineage-specific rather
than global.
Keywords: Kinship; Post-marital residence; Cross-cultural comparison;
Bayesian phylogenetics
1. Introduction
The decision about who will leave home after marriage and who will stay –
post-marital residence – influences social structures in important ways, including
inheritance of property (Leacock, 1955; Agarwal, 1988), household size (Ember,
1973; Divale, 1977), types of marriage and broader family structure (Divale &
Harris, 1976). From an evolutionary perspective, investment in grand-children
hinges on factors including co-residence (Sear & Mace, 2008), and differential
movements of men and women on marriage even impact genetic variability in
sex-specific DNA (Guillot et al., 2016; Lansing et al., 2017).
Post-marital residence states vary widely, but in ethnographically-attested
societies worldwide, the most common residence pattern is patrilocality (Mur-
dock, 1967), where women move to live with the family of their husband.
Nonetheless, other residence practices are also common, the most frequent of
which are: matrilocality, where women remain with their natal community, while
men move; ambilocality, where a newly-wed couple lives with the family of ei-
ther the husband or wife; and neolocality, where the couple establishes a new
residence separate from their respective families.
Importantly, the social norms of post-marital residence that individuals and
societies follow – their ‘residence rules’ – are not static, but change over time.
Residence is heavily co-articulated with other aspects of descent, marriage and
kinship, but residence itself has commonly been viewed as one of the key driving
forces of broader social structure (Murdock, 1949). Consequently, explanations
for transitions in post-marital residence tend to focus mostly on external factors,
and a number of theories have been proposed to explain when and why residence
patterns change. These factors typically invoke major cultural disruptors; be-
haviours that are sufficiently common globally that they might be expected to
influence residence dynamics in universal ways, such as gender-biased division of
subsistence labour (Lippert & Murdock, 1931; Ember & Ember, 1971), warfare
(Ember & Ember, 1971) and migration (Divale, 1974). Conversely, individual
choices – of people and communities (Ly et al., 2018) – also play a role in creat-
ing these new cultural norms. Here, we set out to explore which of these views
is most supported by the data.
There are multiple reasons why a community might adopt a new post-marital
residence rule; for instance, ecological changes or technological developments
(including transitions to agricultural, pastoral (Aberle, 1961) or wage-labour
(Ember, 1967; Zhang, 2008) lifestyles) often change the gender-productivity
balance (Brown, 1970), and communities may come to favour the more eco-
nomically beneficial sex (Lippert & Murdock, 1931; Murdock, 1949; Ember &
Ember, 1971). Modelling has suggested that these changes in residence can be
evolutionarily stable (Ji et al., 2016).
Warfare can also drive post-marital residence change: war with external par-
ties often disrupts male labour, while feuding within a community can encourage
related men to cluster together for protection (Ember & Ember, 1971; Ember,
1974). Villages at war could have high death rates and thus may switch to
matrilocal residence, replenishing losses by attracting men from allied villages
that are not at war (Divale, 1974, 1984).
It has been suggested that matrilocal societies are more peaceful (Van Velzen
& Van Wetering, 1960), with matrilocal bands perhaps acting as a frontier-
advancing structure (Jones, 2011). Feuding is common in patrilocal societies
(Otterbein & Otterbein, 1965; Ember & Ember, 1971; Divale, 1974, 1984), forc-
ing them to develop explicit peacemaking mechanisms and enacting political
integration to reduce infighting. This in turn leads to the linkage of patrilocal
residence with the increasing political complexity of societies (Murdock, 1949;
Ember & Ember, 1971), thus presupposing a global trend towards patrilocality
with the rise of polities and states. Ambilocality has been considered to be an
adaptive social configuration, especially for forager or hunter-gatherer groups,
who rely on a broad resource base (Marlowe, 2004) or are affected by resource
instability (Kelly, 1995). While most hunter-gatherers seem to be classified as
patrilocal (Ember, 1978) due to their culturally preferred residence, this might
contrast with their actual social flexibility.
Finally, human behavioural ecologists have drawn attention to context-specific
inclusive fitness considerations that, in aggregate, may shape community-level
norms of residence (Marlowe, 2004; Wood & Marlowe, 2011; Scelza & Bliege-
Bird, 2008; Kramer & Greaves, 2011). It has been proposed that paternity
uncertainty influences post marital residence (Greene, 1978; Hartung, 1981),
where men in situations of high uncertainty may preferentially choose to invest
in their sister’s children rather than their own. Disentangling inclusive fitness
effects on residence from those on descent and inheritance is difficult (Holden &
Mace, 2003; Mattison, 2011). Furthermore, the costs and benefits of particular
residence norms may vary by the investing sex and over the course of individu-
als’ lives (Wood & Marlowe, 2011). The extent to which such context-specific,
individual-level, adaptive forces might scale up, or be generalisable, across dif-
ferent human groups, and thus influence macroevolutionary patterns, is still a
topic of investigation.
Generic factors can affect any society. For instance, while particular in-
stances of warfare or migration are geographically restricted, their general trends
are often truly global, especially since many geographically-widespread language
families have spread through demographic expansions into previously settled re-
gions. Divale (1974, 1984) suggests that while many drivers of residence change
appear essentially stochastic, they exhibit cycles of change (for instance, from
patrilocal, to matrilocal, to avunculocal and back to patrilocal residence), with
each residence change providing the drivers for its successor.
Regardless of the exact causes of residence change, identifying transitions
in post-marital residence remains challenging, as they are often hard to ob-
serve on a human time scale and leave few direct traces in the archaeological
record. While early studies of residence patterns relied on relatively under-
powered association tests and correlations (Driver, 1956; Aberle, 1961; Tooker,
1968; Blalock, 1971), modern methods aim to explicitly model the evolution
of post-marital residence through time. By using language trees as a proxy for
historical relationships between cultures (Mace & Pagel, 1994), modern phyloge-
netic comparative approaches can infer ancestral post-marital residence states
statistically against a background of phylogenetic divergence within language
families (Currie, 2013). Past residence states, and the rates at which societies
have transitioned between those different states, can therefore be reconstructed
from the present distribution of post-marital residence states using a continuous-
time Markov chain within a Bayesian statistical framework (Pagel et al., 2004).
However, developing methods to analyse patterns across, rather than within,
language trees has proven challenging, and to date the evolution of post-marital
residence has only been studied using phylogenetic comparative methods – sep-
arately – in three language families: Austronesian (Jordan et al., 2009), Bantu
(Opie et al., 2014) and Indo-European (Fortunato & Jordan, 2010; Fortunato,
2011). Now, however, newly available language phylogenies and improved cross-
cultural analyses afford an opportunity to undertake the largest investigation of
cultural evolution in post-marital residence across multiple language families.
Here, we model transitions in post-marital residence across five language
phylogenies, with the aim of testing the hypothesis that a globally common
set of processes has governed changes in post-marital residence states. If the
processes implied by these theories of residence change operate universally, we
would expect to observe similar patterns of residence evolution globally. The
alternative is that individual transitions are instead driven primarily by local
2. Materials and Methods
2.1. Language trees and post-marital residence data
For cross-cultural comparison of post-marital residence evolution, language
families were chosen according to their size and the availability of sufficient lin-
guistic cognate data, resulting in five language families being studied: Austrone-
sian, Bantu, Indo-European, Pama-Nyungan, and Uto-Aztecan. Post-marital
residence has previously been analysed individually for the Austronesian (Jor-
dan et al., 2009), Bantu (Opie et al., 2014) and Indo-European (Fortunato &
Jordan, 2010; Fortunato, 2011) language families, whose phylogenies and post-
marital residence state encodings were obtained from the authors.
For the Uto-Aztecan and Pama-Nyungan language families, a literature
search was performed to determine the primary social norm of post-marital
residence for each language community (see Supplementary Material for de-
tails). The Uto-Aztecan language tree was obtained from Ross and colleagues
(Ross et al., In prep.), while Pama-Nyungan language data were obtained from
the Chirila database (Bowern, 2016) and re-analysed with BayesPhylogenies v
1.1 (Pagel & Meade, 2004) running for 107generations using the m1p model,
in which cognates are lost and gained at the same rate. Trees were pruned
to contain only languages with known residence states. Due to the absence of
calibration points, chronological trees were not obtainable for all language fam-
ilies, and tree branches were scaled by the number of cognate substitutions. A
posterior tree sample (500 <n<1000) was used for all language families, with
variation dictated by the availability of posterior samples for published trees.
A summary of residence states observed for each language family is given in
Supplementary Table 1. Schematics of the distribution of residence states in
the five trees are presented in Figure 1.
2.2. Transition rates
Some authors (Murdock, 1949; Divale, 1974, 1984) suggest that there may
be strong directionality in post-marital residence transitions and thus that some
transitions may not occur at all or only at much lower frequency. Given this
possibility, Reversible Jump Markov Chain Monte Carlo (RJ-MCMC) was ex-
plicitly chosen to fully explore the complex model space. This method aims to
reduce the number of parameters by dynamically setting some to zero, or group-
ing them under a single governing parameter (e.g., setting all transitions to a
single universal rate). Importantly, RJ-MCMC can explicitly test the level of
evidence for different patterns and directions of post-marital residence change,
which is a feature we exploit below.
BayesTraits v 2 (Pagel & Meade, 2006) was used to calculate the transition
rates. Five independent trials of MCMC, each with 108steps, were performed for
each language family with a sampling frequency of 104and an exponential prior
for the frequency of residence transitions Exp(λ).λwas distributed according
to the hyperprior 1
λU(0,200) for all datasets except Pama-Nyungan, for
which the hyperprior was defined as 1
λU(0,400). These values were chosen
from initial maximum likelihood estimates. The convergence of the MCMC
runs was explored using convergence tests implemented in the R package coda
v 0.18-1 (Plummer et al., 2006), and posterior distributions were inspected and
summarized using Rv 3.3.2 (R Core Team, 2017).
To determine whether each language family has its own mode of evolution,
we tested each tree to ascertain whether the transition matrix from any other
tree was as good a fit or better to its data. To do so, we calculated the likeli-
hoods of observed residence states for a particular language family tree given the
rate matrices of each other language family. From these likelihoods, Bayes fac-
tors were calculated by comparing the fit of the original rate matrix with rate
matrices estimated from all of the other datasets in pairwise fashion. These
values indicate whether the likelihoods are significantly different.
2.3. Simulations
To place rates in a more easily interpretable context, we simulated the num-
ber of residence changes on each language tree as defined by its unique transition
matrix. Following Huelsenbeck and colleagues (Huelsenbeck et al., 2003) and
using the R package phytools v 0.5-64 (Revell, 2012), 5000 SIMMAP simulations
of residence evolution were run using the mean rate transition matrices for each
tree. Step-by-step transitions between pairs of states with respect to branch
lengths on the maximum clade credibility tree were inferred using the rate ma-
trix Q, as estimated by BayesTraits. Transitions were generated by first drawing
time from an exponential distribution according to the diagonal elements of the
matrix, followed by choosing the type of transition with probability proportional
to its rate. The probability of transitioning from residence state sito state sj
is defined as Pr(sisj) = qij/Pk6=iqik , where qij is the rate of switching from
state ito j. In other words, probabilities were normalized by the rate of change
from the current state sito any other state. Estimates of the time to each
transition were sampled from an exponential distribution parametrized by the
negative of this normalization factor, and samples were drawn until the branch
length was reached. To save computation time, instead of sampling from the
posterior distribution of the rate matrix calculated by BayesTraits, the posterior
distribution was summarized by the mean rate matrix Q, which accounts for
zero values in the RJ-MCMC. The total number of simulated transitions in each
language family was then normalized by the number of language substitutions
(i.e., the total branch length of each tree).
2.4. Scaling dynamics
To test how post-marital residence evolves relative to language branch lengths,
a scaling parameter κ(Pagel, 1999) was added to the length of tree branches,
such that tnew =tκ
old. If κ1, then the branch length reflects the evolution
of post-marital residence, while κ > 1or <1indicate that longer branches
are scaled more than shorter branches. At the extreme, κ= 0 would suggest
that there is no relationship with branch length, and thus post-marital residence
would evolve independently of the branches on which changes are observed to
occur (i.e., cultural change would be independent of linguistic change).
3. Results
Our analysis focuses on five language families where data are sufficient to
explore the evolution of post-marital residence: the previously reported Aus-
tronesian (Island Southeast Asia and the Pacific), Bantu (Sub-Saharan Africa)
and Indo-European (Eurasia), together with new data for Uto-Aztecan (West-
ern USA and Mesoamerica) and Pama-Nyungan (Australia) (for a overview of
residence states in these datasets, see Supplementary Table 1). Cumulatively,
these languages cover over half the world’s population and geographical area
(see Supplementary Figure 1). Several of these language families have been as-
sociated with Neolithic farming expansions, and they include communities that
currently are, or were until very recently, farmers, foragers or pastoralists, with
a geographic range from the tropics to temperate regions, and from islands to
We assigned ethnographically observed states of residence pattern norms to
contemporary ethnolinguistic groups (Figure 1). To begin, we tested whether
language trees with branches scaled by cognate changes are appropriate for
analysing post-marital residence. Branch lengths reflect observed language
change and are a proxy for evolutionary time. We rescaled branches using
Pagel’s κ(Pagel, 1999) to measure the extent to which the observed branch
lengths can be rescaled without changing the variability in residence patterns.
This simple metric scales all branch lengths by raising them to the same expo-
nent, κ. A value of κclose to zero would suggest that a model with all branches
the same length would fit the residence data better; a value close to one pro-
vides justification for the current model; while higher values of κmake the tree
more star-like, which would mean that the branches effectively have independent
random lengths. While inferred κvalues (Supplementary Table 9) have large
credibility intervals, they strongly centre around 1, supporting the hypothesis
that language trees with branches delimited in shared cognates provide a robust
basis for inferring post-marital residence change.
From the trees (Figure 1), it is clear that residence patterns vary widely,
even among groups that speak closely related languages. Even a cursory ex-
amination suggests great variation in the underlying processes; for instance,
not all residence states are found in every language family. Estimated rates of
transitions between residence states also indicate differences between language
families (Table 1, Supplementary Table 2), with comparatively little change in
Bantu in contrast to frequent change in Pama-Nyungan.
Figure 2 further suggests that patterns of residence change differ between
language families. To explicitly test this, we fitted the estimated mean rate
matrix for a given tree to every other tree and calculated the likelihood of the
fit to the observed residence data. In each case, the tree’s own rate matrix fitted
significantly better than the rate matrix from any other language family (see
Supplementary Table 3 and Supplementary Table 4).
The best statistical support for residence transitions in the language trees
occurs from patrilocality to matrilocality, and back. The Uto-Aztecan tree is
interesting because there is strong evidence against most directions of residence
change (Supplementary Table 2). A benefit of RJ-MCMC, as mentioned previ-
ously, is that all directions of change are tested explicitly in our models.
SIMMAP (Huelsenbeck et al., 2003) simulations of residence transitions us-
ing the observed rate matrices provide additional insight into patterns of change
(Supplementary Table 5). In all datasets except Bantu and Uto-Aztecan, sev-
eral transitions are typically seen to occur along each individual branch. Bantu
and Uto-Aztecan are exceptions because estimated rates of residence evolution
are low and the number of languages in the tree is small, respectively.
To place these values within a more intuitive conceptual framework, we can
make ‘back of the envelope’ estimates of how these changes map on to the ap-
proximate time depth of each language family (Supplementary Table 6). If we
assume that the families studied here (or the parts of them represented in the
trees) are somewhere around 4,000 to 7,000 years old, post-marital residence
transitions seem to have occurred once along any given lineage every 425
years in the Austronesian and Indo-European trees and every 1280 years
in the Bantu tree (see Supplementary Material). The similar estimates for
the Austronesian and Indo-European language families are striking, given that
they differ in many key aspects, such as age, magnitude of residence rates and
amount of language change. However, the less well-studied Pama-Nyungan and
Uto-Aztecan language families give a wider range of values (Supplementary Ta-
ble 6). More rapid changes in residence in the Pama-Nyungan family might be
explained by the fast demographic spread of the language family through Aus-
tralia, quickly colonizing a wide range of ecological regions (Bouckaert et al.,
2018), as well as the social flexibility of indigenous Australian groups, as evi-
denced by the rapid spread of ‘section’ kinship systems (Dousset, 2005).
Across all the trees, there is a tendency for patrilocality to be the most com-
mon and persistent state, both from the perspective of simulated transition rates
and the time spent in each residence state. 64% of communities are patrilocal,
and unlike matrilocality, ambilocality or neolocality, patrilocality appears in all
five language trees. The importance of this residence state can be measured by
comparing estimated transitions to and from each residence state (Supplemen-
tary Table 7), with patrilocality acting as a culturally favoured state (Ji et al.,
Patrilocal residence may stabilize a set of social-structural axes by central-
izing both authority and the inheritance of property; for instance, in many
matrilocal and/or matrilineal societies, women’s brothers still act as heads of
household over many decisions (Richards, 1950; Schneider, 1961; Schlegel, 1972;
Divale & Harris, 1976). This apparent conflict between descent and decision
power was termed the matrilineal puzzle by Audrey Richards (1950) (reviewed
by Mattison, 2011). However, this does not mean that matrilocality is neces-
sarily unstable or non-favoured (see review by Mattison, 2016), as it is still the
second most common state in the Austronesian and Bantu trees. Transitions
from matrilocality to patrilocality, and back, and the generally low frequency
of ambilocality, suggest that the primary role of ambilocality is not simply as
an intermediate state. While ambilocality can occur when the frequency of
patrilocal and matrilocal marriages is similar (see Murdock (1949) and Goode-
nough (1956) for field examples), our analyses predominantly support the role
of ambilocality as a separate functional state with its own dynamics.
As with transition rates, exploring post-marital residence change through
time using SIMMAP simulations (here measured in terms of language change)
suggests that patrilocality is cumulatively the most common state, found almost
90% of the time in Pama-Nyungan to around half the time in Austronesian and
Uto-Aztecan (Supplementary Table 8). Matrilocality is the next most common
residence state, but does not occur at all in the Indo-European family. Neolo-
cality also occurs reasonably often, but the length of time spent in this state is
usually short. The exception is Indo-European, where societies are estimated
to have spent 23% of their time practising neolocality, which is comparable to
the time spent in ambilocal or matrilocal residence in other language families.
An unusually high rate of switching is observed from patrilocality to neolocality
in Indo-European (Table 1), in line with findings that suggest a special role
for neolocality as an alternative residence strategy in Indo-European prehistory
(Fortunato, 2011).
Other cultural dynamics unique to particular language families are observed.
For instance, transition rates are inferred robustly for Bantu, but are relatively
infrequent, as is clear by visual inspection of the tree (Figure 1). This suggests
that there were surprisingly few switches between residence states compared to
the other language families in our dataset, which is especially interesting as the
Bantu tree is relatively large (here, 120 languages), and yet using SIMMAP
simulations parameterized on the transition rate matrix, only 20–36 transitions
between residence states were inferred, compared to 255–351 transitions in the
Austronesian tree (134 languages). The Austronesian tree also shows evidence
for all twelve possible transitions between the four residence states, a property it
shares only with the much smaller Uto-Aztecan tree (25 languages). At the other
extreme, the Pama-Nyungan tree only exhibits two residence states, patrilocality
and matrilocality. However, in contrast to the Bantu tree, a very fast rate of
residence change was estimated for Pama-Nyungan, even though relatively few
transitions appear on visual inspection of the tree (Figure 1).
4. Discussion
The analyses presented here represent a new design for tests of evolutionary
and cross-cultural hypotheses using cultural phylogenetic methods. Examining
the dynamics of post-marital residence in five language families has been made
possible by nearly two decades of innovation in the study of language variation
via phylogenetic modelling (Gray & Jordan, 2000; Gray et al., 2009; Grollemund
et al., 2015; Kolipakam et al., 2018). This approach is further enabled by re-
cent moves to make these language trees, as well as cultural and environmental
datasets that map to the relevant ethnolinguistic groups, openly available via
resources such as D-PLACE (Kirby et al., 2016) and others. When hypotheses
speak to the evolution of human behaviour as a whole, we urge other researchers
to test their ideas across multiple language families. Phylogenetic methods cir-
cumvent old qualms about Galton’s Problem (e.g., Ross & Homer, 1976; Mace &
Pagel, 1994; Korotayev & Munck, 2003), and when these modern computational
approaches are combined with spatial and environmental data, this approach
re-enables the use of global cross-cultural data to inform our understanding of
the processes that drive cultural evolution.
In the specific context of post-marital residence, transitions between res-
idence states have been associated with many different factors, such as in-
tense warfare (Ember & Ember, 1971; Divale, 1974, 1984), prolonged male ab-
sence (Murdock, 1949; Korotayev, 2003; Ember, 2011), sudden depopulation
(Murdock, 1949; Ember, 1967, 2011), changing economic conditions (Murdock,
1949; Ember, 1967), new technological developments (Murdock, 1949; Ember,
1967), inclusive fitness considerations such as paternity certainty and kin altru-
ism (Shenk & Mattison, 2011), post-colonial contact (Ember, 1967; Korotayev,
2003), and even the spread of new dominant cultural practices, like religions
(Goody, 1983; Fortunato & Archetti, 2010). However, the most influential
theories for macro-evolutionary patterns have emphasized warfare (Ember &
Ember, 1971), migration (Divale, 1974) and changes in the gender-based divi-
sion of subsistence labour (Lippert & Murdock, 1931; Murdock, 1949; Ember
& Ember, 1971), all of which are commonly observed globally. As with previ-
ous studies that have used phylogenetic comparative methods (Jordan et al.,
2009; Fortunato & Jordan, 2010; Opie et al., 2014), we do not attempt to model
these putative causal factors directly, but instead employ a probabilistic model
that treats transitions in post-marital residence states as a stochastic process
with many possible causes. We recognize, however, that not all transitions were
necessarily independent; for example, contact with Papuan groups was likely
an ongoing driver of the switch to patrilocality among Austronesian-speaking
groups (Jordan et al., 2009), and Christianity changed the nature and form of
family structures in Europe (Goody, 1983), crossing deep relationships in the
Indo-European language tree. Both speak to contact-induced versus internally-
driven change. The patterns of post-marital residence that we observe likely
represent the cumulative outcome of many interlinked processes, and detailed
coevolutionary testing has the potential to tease many of these factors apart in
the future.
Overall, our results provide strong evidence that each language family has
its own unique dynamics of post-marital residence change, providing little sup-
port for the view that common factors have driven similar processes of change
in residence states globally. Instead, the evolution of societies seems to be
dominated more by local causes, potentially including common factors acting
within locally specific contexts. This is especially apparent from estimates of
transition rates, presence/absence of residence states and different patterns of
robustly inferred rates, all of which vary widely among the language families.
Even groups with similar historical trajectories, such as the rapid agriculturally-
driven expansions of Bantu and Austronesian speakers, show very different past
and modern patterns of post-marital residence. These findings echo the lineage-
specific patterns observed for linguistic structural features, such as word order
(Dunn et al., 2011). Far from arguing for global commonality in the processes
underlying post-marital residence change, these results lend support to the idea
that a suite of causal factors, many perhaps local in origin, have driven past
shifts in post-marital residence.
We thank Laura Fortunato (Institute of Cognitive and Evolutionary Anthro-
pology, University of Oxford, UK) and Kit Opie (Department of Anthropology,
University College London, UK) for access to analyses from their published
work; Lyle Campbell (Department of Linguistics, University of Hawai’i Manoa,
USA) and Jane Hill (School of Anthropology, University of Arizona, USA) for
cognancy coding in the Uto-Aztecan dataset; and Eilis Donnelly (Victoria Uni-
versity of Wellington) for assistance with summarizing the ethnographic litera-
ture of Uto-Aztecan and Pama-Nyungan speakers.
This research was supported by the Royal Society of New Zealand through a
Rutherford Fellowship (RDF-10-MAU-001) to M.P.C., by an Allan Wilson Cen-
tre for Molecular Ecology and Evolution grant to M.P.C. and R.G., and by an
Alexander von Humboldt Stiftung fellowship to M.P.C. The Max Planck Insti-
tute for the Science of Human History funded J.C.M. to attend its 2016 Spring
School on quantitative methods for studying linguistic and cultural evolution.
S.M. and M.P.C. are partially supported by the New Zealand Centre of Re-
search Excellence Te P¯unaha Matatini. R.M.R. was supported by a grant from
the European Union Horizon 2020 Research and Innovation Programme (grant
agreement No 644055 [ALIGNED,]). F.M.J. re-
ceived funding from the European Research Council (ERC) under the European
Union’s Horizon 2020 research and innovation programme (grant agreement No
639291, Starting Grant VARIKIN), and was supported by a Leverhulme Re-
search Fellowship (47690).
Aberle, D. (1961). Matrilineal descent in cross-cultural perspective. In
D. Schneider, & K. Gough (Eds.), Matrilineal Kinship. University of Cal-
ifornia Press: Berkeley.
Agarwal, B. (1988). Who sows? Who reaps? Women and land rights in India.
The Journal of Peasant Studies,15 , 531–581.
Blalock, H. (1971). Causal Models in the Social Sciences. Methodological per-
spectives. Aldine·Atherton: Chicago.
Bouckaert, R. R., Atkinson, Q. D., & Bowern, C. (2018). The origin and ex-
pansion of Pama-Nyungan languages across Australia. Nature Ecology &
Evolution,In Press.
Bowern, C. (2016). Chirila: Contemporary and historical resources for the
indigenous languages of Australia. Language Documentation & Conservation,
10 , 1–44.
Brown, J. K. (1970). A note on the division of labor by sex. American Anthro-
pologist ,72 , 1073–1078.
Currie, T. E. (2013). Cultural evolution branches out. Cross-Cultural Research,
47 , 102–130.
Divale, W. (1984). Matrilocal Residence in Pre-Literate Society. Studies in
Cultural Anthropology. UMI Research Press: Stockton.
Divale, W. T. (1974). Migration, external warfare, and matrilocal residence.
Cross-Cultural Research,9, 75–133.
Divale, W. T. (1977). Living floor area and marital residence: A replication.
Behavior Science Research ,12 , 109–115.
Divale, W. T., & Harris, M. (1976). Population, warfare, and the male
supremacist complex. American Anthropologist,78 , 521–538.
Dousset, L. (2005). Assimilating identities: social networks and the diffusion of
sections. Oceania Monographs.
Driver, H. (1956). An Integration of Functional, Evolutionary, and Historical
Theory by Means of Correlations. Number 12 in Indiana University publica-
tions in Anthropology and Linguistics. Waverly Press: Baltimore.
Dunn, M., Greenhill, S. J., Levinson, S. C., & Gray, R. D. (2011). Evolved
structure of language shows lineage-specific trends in word-order universals.
Nature,473 , 79–82.
Ember, C. R. (1973). Feminine task assignment and the social behavior of boys.
Ethos,1, 424–439.
Ember, C. R. (1974). An evaluation of alternative theories of matrilocal versus
patrilocal residence. Cross-Cultural Research,9, 135–149.
Ember, C. R. (1978). Myths about hunter-gatherers. Ethnology,17 , 439–448.
Ember, C. R. (2011). What we know and what we don’t know about variation
in social organization: Melvin Ember’s approach to the study of kinship.
Cross-Cultural Research,45 , 16–36.
Ember, M. (1967). The emergence of neolocal residence. Transactions of the
New York Academy of Sciences,30 , 291–302.
Ember, M., & Ember, C. R. (1971). The conditions favoring matrilocal versus
patrilocal residence. American Anthropologist,73 , 571–594.
Fortunato, L. (2011). Reconstructing the history of residence strategies in Indo-
European-speaking societies: neo–, uxori–, and virilocality. Human Biology,
83 , 107–128.
Fortunato, L., & Archetti, M. (2010). Evolution of monogamous marriage by
maximization of inclusive fitness. Journal of Evolutionary Biology,23 , 149–
Fortunato, L., & Jordan, F. (2010). Your place or mine? A phylogenetic com-
parative analysis of marital residence in Indo-European and Austronesian
societies. Philosophical Transactions of the Royal Society of London B: Bio-
logical Sciences,365 , 3913–3922.
Goodenough, W. H. (1956). Residence rules. Southwestern Journal of Anthro-
pology,12 , 22–37.
Goody, J. (1983). The Development of the Family and Marriage in Europe .
Cambridge University Press: Cambridge.
Gray, R., Drummond, A., & Greenhill, S. (2009). Language phylogenies reveal
expansion pulses and pauses in Pacific settlement. Science,323 , 479–483.
Gray, R. D., & Jordan, F. M. (2000). Language trees support the express-train
sequence of Austronesian expansion. Nature,405 , 1052–1055.
Greene, P. J. (1978). Promiscuity, paternity, and culture. American Ethnologist,
5, 151–159.
Grollemund, R., Branford, S., Bostoen, K., Meade, A., Venditti, C., & Pagel,
M. (2015). Bantu expansion shows that habitat alters the route and pace
of human dispersals. Proceedings of the National Academy of Sciences USA,
112 , 13296–13301.
Guillot, E. G., Hazelton, M. L., Karafet, T. M., Lansing, J. S., Sudoyo, H., &
Cox, M. P. (2016). Relaxed observance of traditional marriage rules allows
social connectivity without loss of genetic diversity. Molecular Biology and
Evolution,32 , 2254–2262.
Hartung, J. (1981). Paternity and inheritance of wealth. Nature,291 , 652–654.
Holden, C. J., & Mace, R. (2003). Spread of cattle led to the loss of matrilineal
descent in Africa: a coevolutionary analysis. Proceedings of the Royal Society
of London B: Biological Sciences ,270 , 2425–2433.
Huelsenbeck, J. P., Nielsen, R., & Bollback, J. P. (2003). Stochastic mapping
of morphological characters. Systematic Biology,52 , 131.
Ji, T., Zheng, X.-D., He, Q.-Q., Wu, J.-J., Mace, R., & Tao, Y. (2016). Kinship
as a frequency dependent strategy. Royal Society Open Science,3, 150632.
Jones, D. (2011). The matrilocal tribe. Human Nature,22 , 177–200.
Jordan, F. M., Gray, R. D., Greenhill, S. J., & Mace, R. (2009). Matrilocal res-
idence is ancestral in Austronesian societies. Proceedings of the Royal Society
of London B: Biological Sciences ,276 , 1957–1964.
Kelly, R. L. (1995). The Foraging Spectrum: Diversity in Hunter-Gatherer
Lifeways. Washington, DC: Smithsonian Institution Press.
Kirby, K. R., Gray, R. D., Greenhill, S. J., Jordan, F. M., Gomes-Ng, S., Bibiko,
H.-J., Blasi, D. E., Botero, C. A., Bowern, C., Ember, C. R., Leehr, D., Low,
B. S., McCarter, J., Divale, W., & Gavin, M. C. (2016). D-PLACE: A global
database of cultural, linguistic and environmental diversity. PLOS ONE ,11 ,
Kolipakam, V., Jordan, F., Dunn, M., Greenhill, S., Bouckaert, R., Gray, R., &
Verkerk, A. (2018). A Bayesian phylogenetic study of the Dravidian language
family. Royal Society Open Science,5, 171504.
Korotayev, A. (2003). Form of marriage, sexual division of labor, and post-
marital residence in cross-cultural perspective: A reconsideration. Journal of
Anthropological Research,59 , 69–89.
Korotayev, A., & Munck, V. (2003). Galton’s Asset” and “ Flower’s Problem”:
Cultural networks and cultural units in cross-cultural research: (or, male
genital mutilations and polygyny in cross-cultural perspective). American
Anthropologist,105 , 353–358.
Kramer, K. L., & Greaves, R. D. (2011). Postmarital residence and bilateral
kin associations among hunter-gatherers: Pumé foragers living in the best of
both worlds. Human Nature,22 , 41–63.
Lansing, J. S., Abundo, C., Jacobs, G. S., Guillot, E. G., Thurner, S., Downey,
S. S., Chew, L. Y., Bhattacharya, T., Chung, N. N., Sudoyo, H., & Cox,
M. P. (2017). Kinship structures create persistent channels for language
transmission. Proceedings of the National Academy of Sciences USA,114 ,
Leacock, E. (1955). Matrilocality in a simple hunting economy (Montagnais-
Naskapi). Southwestern Journal of Anthropology,11 , 31–47.
Lippert, J., & Murdock, G. (1931). The Evolution of Culture. George Allen &
Unwin: Crows Nest, Australia.
Ly, G., Alard, B., Laurent, R., Lafosse, S., Toupance, B., Monidarin, C., Dif-
floth, G., Bourdier, F., Evrard, O., Pavard, S., & Chaix, R. (2018). Residence
rule flexibility and descent groups dynamics shape uniparental genetic diver-
sities in South East Asia. American Journal of Physical Anthropology,165 ,
Mace, R., & Pagel, M. (1994). The comparative method in Anthropology.
Current Anthropology,35 , 549–564.
Marlowe, F. W. (2004). Marital residence among foragers. Current Anthropol-
ogy,45 , 277–284.
Mattison, S. M. (2011). Evolutionary contributions to solving the “Matrilineal
Puzzle”. Human Nature,22 , 64.
Mattison, S. M. (2016). Matrilineal and matrilocal systems. In The Wiley
Blackwell Encyclopedia of Gender and Sexuality Studies. John Wiley & Sons,
Murdock, G. P. (1949). Social Structure . Macmillan: London.
Murdock, G. P. (1967). Ethnographic Atlas. University of Pittsburgh Press:
Opie, C., Shultz, S., Atkinson, Q. D., Currie, T., & Mace, R. (2014). Phy-
logenetic reconstruction of Bantu kinship challenges Main Sequence Theory
of human social evolution. Proceedings of the National Academy of Sciences
USA,111 , 17414–17419.
Otterbein, K. F., & Otterbein, C. S. (1965). An eye for an eye, a tooth for
a tooth: A cross-cultural study of feuding. American Anthropologist,67 ,
Pagel, M. (1999). Inferring the historical patterns of biological evolution. Na-
ture,401 , 877–884.
Pagel, M., & Meade, A. (2004). A phylogenetic mixture model for detecting
pattern-heterogeneity in gene sequence or character-state data. Systematic
Biology,53 , 571.
Pagel, M., & Meade, A. (2006). Bayesian analysis of correlated evolution of
discrete characters by Reversible-Jump Markov Chain Monte Carlo. The
American Naturalist,167 , 808–825.
Pagel, M., Meade, A., & Barker, D. (2004). Bayesian estimation of ancestral
character states on phylogenies. Systematic Biology,53 , 673–684.
Plummer, M., Best, N., Cowles, K., & Vines, K. (2006). CODA: Convergence
diagnosis and output analysis for MCMC. R News ,6, 7–11.
R Core Team (2017). R: A Language and Environment for Statistical Com-
puting. R Foundation for Statistical Computing Vienna, Austria. URL:
Revell, L. J. (2012). phytools: An R package for phylogenetic comparative
biology (and other things). Methods in Ecology and Evolution,3, 217–223.
Richards, A. I. (1950). Some types of family structure amongst the Central
Bantu. In A. R. Radcliffe-Brown, & D. Forde (Eds.), African Systems of
Kinship and Marriage (pp. 207–251). Oxford University Press London.
Ross, M. H., & Homer, E. (1976). Galton’s problem in cross-national research.
World Politics ,29 , 1–28.
Ross, R., Greenhill, S., List, J., Anderson, C., Botero, C., Vilela de Moraes e
Silva, B., & Gray, R. (In prep.). The origins and dispersal of the Uto-Aztecan
Scelza, B., & Bliege-Bird, R. (2008). Group structure and female cooperative
networks in Australia’s Western Desert. Human Nature,19 , 231–248.
Schlegel, A. (1972). Male Dominance and Female Autonomy: Domestic Au-
thority in Matrilineal Societies. Comparative Studies. Human Relations Area
Files Press: New Haven.
Schneider, D. (1961). The distinctive features of matrilineal descent groups. In
D. Schneider, & K. Gough (Eds.), Matrilineal Kinship. University of Califor-
nia Press: Berkeley.
Sear, R., & Mace, R. (2008). Who keeps children alive? A review of the effects
of kin on child survival. Evolution and Human Behavior ,29 , 1–18.
Shenk, M. K., & Mattison, S. M. (2011). The rebirth of kinship. Human Nature,
22 , 1–15.
Tooker, E. (1968). Masking and matrilineality in North America. American
Anthropologist,70 , 1170–1176.
Van Velzen, H., & Van Wetering, W. (1960). Residence, power groups and intra-
societal aggression. International Archives of Ethnography,49 , 169–200.
Wood, B. M., & Marlowe, F. W. (2011). Dynamics of postmarital residence
among the Hadza. Human Nature,22 , 128–138.
Zhang, W. (2008). State, gender and uxorilocal marriage in contemporary rural
North China. The China Journal ,60 , 111–132.
Austronesian Bantu
Indo-European Pama-Nyungan Uto-Aztecan
Figure 1: Ethnographic observations of post-marital residence states mapped on to five lan-
guage trees: Austronesian, Bantu, Indo-European, Pama-Nyungan, and Uto-Aztecan. Ter-
minal branches are coloured according to the main post-marital residence state recorded for
each society. Branch lengths of each maximum clade credibility tree are drawn proportional
to the number of observed lexical substitutions. To show the residence states clearly, trees are
not drawn to the same scale.
27 114
83 55
Austronesian Bantu
Indo-European Uto-Aztecan
13 27
Figure 2: Graphs showing transition rates between post-marital residence states for each
language family. M, matrilocality; P, patrilocality; A, ambilocality; N, neolocality. Arrow
weights indicate mean transition rates inferred from the analysis (with values shown adjacent),
while shading indicates how frequently the rate is inferred to be zero (lighter shades indicate
less certainty). Node colours indicate post-marital residence states, as in Figure 1.
Table 1: Rates of transitions between post-marital residence states. Means and 95% credible intervals are reported (rounded to the nearest integer);
dashes indicate transition states that are not observed in a given language tree. Note that zeros were removed from each distribution and are reported
separately (see Supplementary Table 2).
Austronesian Bantu Indo-European Pama-Nyungan Uto-Aztecan
mean 95% CI mean 95% CI mean 95% CI mean 95% CI mean 95% CI
ambi matri 123 (10; 292) 28 (2; 72)
ambi neo 106 (10; 267) 47 (0; 185) 26 (2; 70)
ambi patri 101 (8; 220) 55 (0; 235) 25 (2; 57)
matriambi 122 (18; 283) 32 (2; 79)
matrineo 66 (6; 218) 3 (1; 6) 30 (2; 78)
matripatri 78 (8; 201) 4 (1; 12) 286 (1; 801) 27 (2; 71)
neo ambi 117 (9; 292) 48 (0; 188) 31 (2; 78)
neo matri 110 (9; 295) 10 (1; 51) 30 (2; 77)
neo patri 114 (9; 291) 13 (1; 76) 55 (0; 258) 27 (2; 69)
patri ambi 47 (7; 129) 38 (0; 185) 16 (1; 40)
patri matri 63 (6; 172) 3 (1; 5) 45 (1; 131) 15 (1; 42)
patri neo 27 (5; 71) 2 (0; 5) 83 (0; 334) 13 (1; 34)
mean 89 6 54 165 25
Figure 1
Figure 2
... Prehistoric kinship systems, flexible and adaptive through time [1][2][3][4] and integral to the organization and evolution of early complex societies, 4,5-9 exhibit both ancestral and adaptive dynamics in correspondence with prehistoric human dispersals. 7,10-13 As a result, certain general tendencies seem to have facilitated, though not determined, directional changes in post-marital residence system. ...
... [27][28][29][30][31][32][33][34][35][36] Additionally, post-marital residence systems in prehistoric Austronesian societies appear to have been more fluid, potentially transitioning back and forth between states on a time scale of centuries. 4,24 Cultural phylogenetic analyses 4 indicate that patrilocality and matrilocality, respectively, are the first and second most common (likely) states in the Austronesian tree. ...
... [27][28][29][30][31][32][33][34][35][36] Additionally, post-marital residence systems in prehistoric Austronesian societies appear to have been more fluid, potentially transitioning back and forth between states on a time scale of centuries. 4,24 Cultural phylogenetic analyses 4 indicate that patrilocality and matrilocality, respectively, are the first and second most common (likely) states in the Austronesian tree. ...
Full-text available
As adaptive systems, kinship and its accompanying rules have co-evolved with elements of complex societies, including wealth inheritance, subsistence, and power relations. Here we consider an aspect of kinship evolution in the Austronesian dispersal that began from about 5500 BP in Taiwan, reaching Melanesia about 3200 BP, and dispersing into Micronesia by 1500 BP. Previous, foundational work has used phylogenetic comparative methods and ethnolinguistic information to infer matrilocal residence in proto-Austronesian societies. Here we apply Bayesian phylogenetic analyses to a data set on Austronesian societies that combines existing data on marital residence systems with a new set of ethnographic data, introduced here, on initiation rites. Transition likelihoods between cultural-trait combinations were modeled on an ensemble of 1000 possible Austronesian language trees, using Reversible Jump Markov Chain Monte Carlo (RJ-MCMC) simulations. Compared against a baseline phylogenetic model of independent evolution, a phylogenetic model of correlated evolution between female and male initiation rites is substantially more likely (log Bayes factor: 17.9). This indicates, over the generations of Austronesian dispersal, initiation rites were culturally stable when both female and male rites were in the same state (both present or both absent), yet relatively unstable for female-only rites. The results indicate correlated phylogeographic evolution of cultural initiation rites in the prehistoric dispersal of Austronesian societies across the Pacific. Once acquired, male initiation rites were more resilient than female-only rites among Austronesian societies.
... Recent interest in past kinship has come to reflect a range of research ranging from gender, sexuality, and social history to evolutionary theory (Apicella et al. 2012, Cronk & Gerkey 2007, Ember 2011, Ensor 2011, Hrdy 2009, Jordan et al. 2009, Levine 2008, Moravec et al. 2018, Shenk et al. 2019. The wide range of evidence-new and old ethnographies, archaeology, genetics, linguistics-requires a hermeneutic approach toward the inevitable layers of bias in each record. ...
... After comparing thousands of simulation results to the observed societies at the branch tips, Bayesian phylogenetic analysis yields the set of transition likelihoods among the four states. Cultural phylogenetic techniques can reveal the likelihood of cultural states at the convergence, or root, of the tree, including the origins of major agricultural dispersals (e.g., Fortunato & Jordan 2010, Moravec et al. 2018). ...
... The diversity of kinship in Pacific, Melanesian, and Island Southeast Asian societies illustrates the scope of possibility (Brück 2021, Miller et al. 2021, Souvatzi 2017). Much of this Pacific diversity in kinship likely developed in the past 5,000 years, with each major language group having its own unique history (Moravec et al. 2018, Rácz et al. 2020. In Island Southeast Asia, linguistic (Hage & Marck 2003) and cultural phylogenetic evidence (see the sidebar titled Prehistoric Kinship Evolution and Cultural Phylogenetics) suggests that the initial Austronesian dispersal across the western Pacific Ocean originated with matrilineal groups ( Jordan et al. 2009). ...
Full-text available
As observed in recent centuries, the contemporary variety of kinship systems reflects millennia of human migration, cultural inheritance, adaptation, and diversification. This review describes key developments in prehistoric kinship, from matricentric hominin evolution to the Neolithic transition to agriculture and the heterogeneous resilience of matriliny. Starting with our hominin ancestors, kinship evolved among a cooperative breeding species to multilevel group structure among human hunter-gatherers, to substantial kinship changes brought on by the origins of intensified farming, to permanent settlements and unequal resource access. This review takes the approach that new forms of subsistence facilitated new equations of reproductive success, which changed cultural norms of kinship systems and heritable wealth. Subsequently, the formation of complex societies diminished kinship as the primary organizing principle of society. The article describes new methodologies and theoretical developments, along with critiques of bioarchaeological interpretations of prehistoric kinship. Expected final online publication date for the Annual Review of Anthropology Volume 51 is October 2022. Please see for revised estimates.
... These studies suggested Proto-Indo-European was virilocal or neolocal (Fortunato, 2011;Fortunato and Jordan, 2010), proto-Malayo-Polynesian was uxorilocal, and matrilocality was prevailing in proto-Austronesian populations (Fortunato and Jordan, 2010;Jordan et al., 2009). A phylogenetic study using linguistic data from five large language families (Austronesian, Bantu, Indo-European, Pama-Nyungan, and Uto-Aztecan) suggested that evolutionary trajectories of post-marital residence systems are likely to vary within each language family (Moravec et al., 2018). Some light has also been shed on the coevolution of sex-specific dispersal and various ecological or social factors. ...
Sex-biased dispersal has long been of interest to anthropologists and biologists, as it can structure populations and determine patterns of kinship, relatedness and cooperation. In most contemporary human societies, females usually disperse at marriage. In a minority of human societies, male dispersal, bisexual philopatry, or both sexes dispersing is practiced. Previous studies suggest that emergence of either agriculture, cattle pastoralism, or patriliny is associated with female-biased dispersal in certain language families. The ancestral patterns of sex-specific dispersal and its ecological correlates in Sino-Tibetans remain uncertain. Here we use comparative phylogenetic methods to infer the evolutionary history of sex-specific dispersal in Sino-Tibetan groups, and tested for coevolution between subsistence (agriculture and cattle-keeping), descent and sex-specific dispersal. We use a variety of ethnographic and historical sources to identify dispersal strategies across Sino-Tibetan phylogenetic trees (n = 97). We found that 1) earliest Sino-Tibetan groups were likely patrilocal; 2) agriculture likely co-evolved with only female dispersal patterns, but the result is sensitive to alternative coding strategy; 3) there is no evidence that domestic cattle co-evolved with dispersal patterns of either gender; and 4) kinship descent likely co-evolved with female dispersal, but not with male dispersal. Moreover, change from state of "patrilineal" to "non-patrilineal" triggered change in female dispersal patterns, from "female non-stay" to "female stay". Our results suggest that change in descent drove change in female-specific dispersal pattern in Sino-Tibetans. Our findings illustrated how subsistence or descent can play different roles in shaping sex-biased dispersal patterns.
... In contrast with the evidence for patriliny and patrilocality in the European Neolithic and Bronze Age (e. g., Bentley et al., 2012;Mittnik et al., 2019;Feinman and Neitzel, 2019), complex and sedentary hunter-gatherer communities existed in the MSEA region for millennia prior to the arrival of farming , and kinship through time appears to have been more flexible and heterogeneous (Fox, 1983;Jordan et al., 2009;Ledgerwood, 1995;Malinowski, 1932;Mills, 1997;Rácz et al., 2019;Reid, 1988). In Island Southeast Asia (ISEA), genetic, cultural phylogenetic and linguistic evidence suggests the initial Austronesian dispersal across the Western Pacific Ocean originated with matrilineal groups (Hage and Marck, 2003;Jordan et al., 2009;Moravec et al., 2018;Oota et al., 2001). ...
An exciting topic in the prehistory of Mainland Southeast Asia (MSEA) is where, when and why matriliny/matrilocality existed in certain regions during the rise of complex societies. Linguistic and genetic evidence have been used to infer kinship questions on a broad, continental scale. To contribute to this discussion on the site scale, here we summarize and compare, for the first time, almost two decades of isotopic work on prehistoric skeletons from across the region. For comparability of mobility patterns between sexes and between sites, we introduce simple normalized conversions of the (strontium, oxygen and carbon) isotopic data. Taking the evidence together, we find support for our hypothesis that matriliny/matrilocality was most likely at times and places with maritime or riverine trade, as exhibited at several coastal/deltaic sites of Thailand and Vietnam. In any case, there is an exciting future ahead for multi-method investigations into prehistoric kinship and social differentiation in Southeast Asia.
... They can be applied to estimate the ancestral states of a cultural trait, and test whether the transmission of cultural traits was functionally linked to particular ecological circumstances or geographical proximity 64 . Recent phylogenetic comparative studies have provided key insights into cultural evolution in Austronesian, Bantu, Indo-European, Pama-Nyungan and Uto-Aztecan populations 65 . Our reconstructed phylogeny has been applied to a quantitative cross-cultural database to study the cultural evolution of kinship and subsistence among Sino-Tibetan cultures (Ji et al., forthcoming). ...
Full-text available
An accurate reconstruction of Sino-Tibetan language evolution would greatly advance our understanding of East Asian population history. Two recent phylogenetic studies attempted to do so but several of their conclusions are different from each other. Here we reconstruct the phylogeny of the Sino-Tibetan language family, using Bayesian computational methods applied to a larger and linguistically more diverse sample. Our results confirm previous work in finding that the ancestral Sino-Tibetans first split into Sinitic and Tibeto-Burman clades, and support the existence of key internal relationships. But we find that the initial divergence of this group occurred earlier than previously suggested, at approximately 8000 years before the present, coinciding with the onset of millet-based agriculture and significant environmental changes in the Yellow River region. Our findings illustrate that key aspects of phylogenetic history can be replicated in this complex language family, and calls for a more nuanced understanding of the first Sino-Tibetan speakers in relation to the “early farming dispersal” theory of language evolution.
The relationship between residence, gender and mobility is central to the study of early social complexity. And yet, until recently, it was deemed as archaeologically intractable. The recent combination of strontium data and genomics with other methods has opened up entirely new possibilities for the archaeological study of human mobility, but these advances are not without problems. Theoretical framing, empirical accuracy and data interpretation remain controversial. In this paper we address the relationship between residence patterns, gender and mobility among early complex societies, combining both ethnographic and archaeological evidence. Our approach focuses on Chalcolithic Iberia, a period in which the stage for emerging social complexity was set. The possible existence of male-centered residential patterns and their possible connection with conflict, social complexity and gender inequalities is examined. The available data on strontium isotopes suggest women were more frequently buried in places different from those where they grew up, which can be linked to bilocality biased to patrilocality, especially in the so called ‘mega-sites’. While preliminary, this body of evidence opens up fresh lines of enquiry for the study of early complex societies, highlights the benefits of combining different kinds of evidence, and underlines the centrality of gender in the social analysis.
Full-text available
Makpa marriage practice among Manang Himalayan people of Nepal.
Full-text available
Makpa marriage strategies among the Manang Himalayan community of Nepal.
Phylogenetic analyses increasingly take centre-stage in our understanding of the processes shaping patterns of cultural diversity and cultural evolution over time. Just as biologists explain the origins and maintenance of trait differences among organisms using phylogenetic methods, so anthropologists studying cultural macroevolutionary processes use phylogenetic methods to uncover the history of human populations and the dynamics of culturally transmitted traits. In this paper, we revisit concerns with the validity of these methods. Specifically, we use simulations to reveal how properties of the sample (size, missing data), properties of the tree (shape) and properties of the traits (rate of change, number of variants, transmission mode) might influence the inferences that can be drawn about trait distributions across a given phylogeny and the power to discern alternative histories. Our approach shows that in two example datasets specific combinations of properties of the sample, of the tree and of the trait can lead to potentially high rates of Type I and Type II errors. We offer this simulation tool to help assess the potential impact of this list of persistent perils in future cultural macroevolutionary work. This article is part of the theme issue ‘Foundations of cultural evolution’.
Full-text available
The Dravidian language family consists of about 80 varieties (Hammarström H. 2016 Glottolog 2.7) spoken by 220 million people across southern and central India and surrounding countries (Steever SB. 1998 In The Dravidian languages (ed. SB Steever), pp. 1textendash39: 1). Neither the geographical origin of the Dravidian language homeland nor its exact dispersal through time are known. The history of these languages is crucial for understanding prehistory in Eurasia, because despite their current restricted range, these languages played a significant role in influencing other language groups including Indo-Aryan (Indo-European) and Munda (Austroasiatic) speakers. Here, we report the results of a Bayesian phylogenetic analysis of cognate-coded lexical data, elicited first hand from native speakers, to investigate the subgrouping of the Dravidian language family, and provide dates for the major points of diversification. Our results indicate that the Dravidian language family is approximately 4500 years old, a finding that corresponds well with earlier linguistic and archaeological studies. The main branches of the Dravidian language family (North, Central, South I, South II) are recovered, although the placement of languages within these main branches diverges from previous classifications. We find considerable uncertainty with regard to the relationships between the main branches.
Full-text available
It remains a mystery how Pama-Nyungan, the world's largest hunter-gatherer language family, came to dominate the Australian continent. Some argue that social or technological advantages allowed rapid language replacement from the Gulf Plains region during the mid-Holocene. Others have proposed expansions from refugia linked to climatic changes after the last ice age or, more controversially, during the initial colonization of Australia. Here, we combine basic vocabulary data from 306 Pama-Nyungan languages with Bayesian phylogeographic methods to explicitly model the expansion of the family across Australia and test between these origin scenarios. We find strong and robust support for a Pama-Nyungan origin in the Gulf Plains region during the mid-Holocene, implying rapid replacement of non-Pama-Nyungan languages. Concomitant changes in the archaeological record, together with a lack of strong genetic evidence for Holocene population expansion, suggests that Pama-Nyungan languages were carried as part of an expanding package of cultural innovations that probably facilitated the absorption and assimilation of existing hunter-gatherer groups.
Full-text available
Objectives: Social organization plays a major role in shaping human population genetic diversity. In particular, matrilocal populations tend to exhibit less mitochondrial diversity than patrilocal populations, and the other way around for Y chromosome diversity. However, several studies have not replicated such findings. The objective of this study is to understand the reasons for such inconsistencies and further evaluate the influence of social organization on genetic diversity. Materials and methods: We explored uniparental diversity patterns using mitochondrial HV1 sequences and 17 Y-linked short tandem repeats (STRs) in 12 populations (n = 619) from mainland South-East Asia exhibiting a wide range of social organizations, along with quantitative ethno-demographic information sampled at the individual level. Results: MtDNA diversity was lower in matrilocal than in multilocal and patrilocal populations while Y chromosome diversity was similar among these social organizations. The reasons for such asymmetry at the genetic level were understood by quantifying sex-specific migration rates from our ethno-demographic data: while female migration rates varied between social organizations, male migration rates did not. This unexpected lack of difference in male migrations resulted from a higher flexibility in residence rule in patrilocal than in matrilocal populations. In addition, our data suggested an impact of clan fission process on uniparental genetic patterns. Conclusions: The observed lack of signature of patrilocality on Y chromosome patterns might be attributed to the higher residence flexibility in the studied patrilocal populations, thus providing a potential explanation for the apparent discrepancies between social and genetic structures. Altogether, this study highlights the need to quantify the actual residence and descent patterns to fit social to genetic structures.
Full-text available
Languages are transmitted through channels created by kinship systems. Given sufficient time, these kinship channels can change the genetic and linguistic structure of populations. In traditional societies of eastern Indonesia, finely resolved cophylogenies of languages and genes reveal persistent movements between stable speech communities facilitated by kinship rules. When multiple languages are present in a region and postmarital residence rules encourage sustained directional movement between speech communities, then languages should be channeled along uniparental lines. We find strong evidence for this pattern in 982 individuals from 25 villages on two adjacent islands, where different kinship rules have been followed. Core groups of close relatives have stayed together for generations, while remaining in contact with, and marrying into, surrounding groups. Over time, these kinship systems shaped their gene and language phylogenies: Consistently following a postmarital residence rule turned social communities into speech communities.
Full-text available
From the foods we eat and the houses we construct, to our religious practices and political organization, to who we can marry and the types of games we teach our children, the diversity of cultural practices in the world is astounding. Yet, our ability to visualize and understand this diversity is limited by the ways it has been documented and shared: on a culture-by-culture basis, in locally-told stories or difficult-to-access repositories. In this paper we introduce D-PLACE, the Database of Places, Language, Culture, and Environment. This expandable and open-access database (accessible at ) brings together a dispersed corpus of information on the geography, language, culture, and environment of over 1400 human societies. We aim to enable researchers to investigate the extent to which patterns in cultural diversity are shaped by different forces, including shared history, demographics, migration/diffusion, cultural innovations, and environmental and ecological conditions. We detail how D-PLACE helps to overcome four common barriers to understanding these forces: i) location of relevant cultural data, (ii) linking data from distinct sources using diverse ethnonyms, (iii) variable time and place foci for data, and (iv) spatial and historical dependencies among cultural groups that present challenges for analysis. D-PLACE facilitates the visualisation of relationships among cultural groups and between people and their environments, with results downloadable as tables, on a map, or on a linguistic tree. We also describe how D-PLACE can be used for exploratory, predictive, and evolutionary analyses of cultural diversity by a range of users, from members of the worldwide public interested in contrasting their own cultural practices with those of other societies, to researchers using large-scale computational phylogenetic analyses to study cultural evolution. In summary, we hope that D-PLACE will enable new lines of investigation into the major drivers of cultural change and global patterns of cultural diversity.
Full-text available
Humans divide themselves up into separate cultures, which is a unique and ubiquitous characteristic of our species. Kinship norms are one of the defining features of such societies. Here we show how norms of marital residence can evolve as a frequency-dependent strategy, using real-world cases from southwestern China and an evolutionary game model. The process of kinship change has occurred in the past and is also occurring now in southwestern China. Our data and models show how transitions between residence types can occur both as response to changing costs and benefits of co-residence with kin, and also due to the initial frequency of the strategies adopted by others in the population: patrilocal societies can become matrilocal, and neolocal societies can become duolocal. This illustrates how frequency-dependent selection plays a role both in the maintenance of group-level cultural diversity and in cultural extinction.
Here I present the background to, and a description of, a newly developed database of historical and contemporary lexical data for Australian languages (Chirila), concentrating on the Pama-Nyungan family (the largest family in the country). While the database was initially developed in order to facilitate research on cognate words and reconstructions, it has had many uses beyond its original purpose, in synchronic theoretical linguistics, language documentation, and language reclamation. Creating a multi-audience database of this type has been challenging, however. Some of the challenges stemmed from success: as the size of the database grew, the original data structure became unwieldy. Other challenges grew from the difficulties in anticipating future needs, in keeping track of materials, and in coping with diverse input formats for so many highly endangered languages. In this paper I document the structure of the database, provide an overview of its uses (both in diachronic and synchronic research), and discuss some of the issues that have arisen during the project and choices that needed to be made as the database was created, compiled, curated, and shared. I address here the major problems that arise with linguistic data, particularly databases created for diverse audiences, from diverse data, with little infrastructure support.
Matriliny is a relatively rare system of kinship in which kin on the maternal side are recognized more strongly than kin on the paternal side in defining social organization. Matrilineal kinship was once considered problematic by scholars focused on the tensions felt by men whose allegiance was split between natal and affinal kin groups. More recent scholarship focuses on the benefits afforded by matrilineal kinship, including its association with harmonious relationships among natal kin and the benefits to parents of transmitting property along the female line.