Varikin Findings: There are no universals in the cultural evolution of kinship terminology

There is huge potential for variation between kinship terminology (systems of words for relatives) cross-culturally. In fact, for a set of 16 relatives, there are 10,480,142,147 theoretically possible ways to categorise them. Given the potential for effectively unbridled variation in kinship terminology, observed diversity seems remarkably constrained, and has previously been categorised into six key types (see Fig. 1). Each type has been hypothesised to align with particular cultural norms of descent, marriage or residence patterns, and it was thought that these social systems acted as drivers of the variation and change of kinship terminology. However, since these conclusions were reached, there have been significant improvements in data and statistics. Will they still stand up when tested using more modern methods?

VariKin members Sam Passmore and Fiona Jordan’s 2020 publication No universals in the cultural evolution of kinship terminology put this to the test. They used phylogenetic comparative methods (PCMs) to explore the cultural evolutionary patterns of kinship terminology diversity, particularly testing so-called drivers of change such as marriage, residence and descent. This study marks the first cross-language-family phylogenetic analysis of the drivers of kinship terminology and investigates the Austronesian, Bantu and Uto-Aztecan language families (selected due to their large size and cultural data availability). In the study, they found no evidence for previously assumed universal drivers of evolution in kinship terminologies, because no presumed universal relationship linking kinship terminology and social structure was supported across all language families. This is not because one or two societies do not follow a trend: for most of the hypotheses tested, they could not conservatively even claim ‘statistical regularity’.

However, rather than assume that the lack of statistical relationship between kinship terminology types and social structure indicates no functional relationship, the authors suggest that the typology might misrepresent kinship terminological diversity by being too crude a characterisation. This raises the important question of how well existing typologies represent known diversity, and the results from this particular study – which was always an integral investigation of the VariKin Project – have helped to clear the ground for more detailed characterisations of kinship organisation in future. For example, as stated in the publication, the creation of the “Kinbank” database as part of the VariKin project will allow us to more realistically characterise types and patterns of change in kinterms without built-in assumptions of the influence from social structure.

Varikin Findings: How Pama-Nyungan Grandparent Naming Systems Change

The VariKin project has examined the cultural evolution of kinship terminology from a range of perspectives and methods. Considerable research by our group and others has focused on the comparison of cousin and sibling terms, but grandparent terms, the topic of this paper, are largely understudied. The social salience of grandparents, due to their frequent role in the care of grandchildren as well as in controlling wealth and cultural knowledge, suggested that a project of this focus was long overdue.

Catherine Sheard and Fiona Jordan from the VariKin team, along with Claire Bowern and Rikker Dockum (Yale University), addressed this gap in an open access paper published earlier this year in Evolutionary Human Sciences. “Pama–Nyungan grandparent systems change with grandchildren, but not cross-cousin terms or social norms” compares grandparent terms across 134 Australian languages from the Pama-Nyungan family. Using Pama-Nyungan as a study system is less common in studies of kinship macroevolution. Australia’s ecological and anthropological history (namely, a continental radiation of primarily hunter-gatherer languages) provides a contrasting example to language families associated with agricultural spreads, such as Austronesian or Bantu.



A standard way of thinking about grandparental kin terms is to consider each of parent of mother and father: MM, MF, FM, and FF. Pitjantjatjara has a two-term system that merges gender like English does, with ‘kami’ for MM and FM (grandmother), and ‘amu’ for FF and MF (grandfather). Other languages have four separate terms, for example, Alyawarra has ‘arrenge’ FF, ‘artartetye’ MF, ‘anyanye’ MM and ‘aperle’ FM. Pama-Nyungan languages also explore other three-term systems that merge parallel grandparents, so that MM=FF as in Djapu, or merge cross grandparents, so that MF=FM, as in Yagara.

The group used Bayesian phylogenetic comparative methods to model the stages of grandparent naming systems (see Fig 3. below). This analysis suggests that the ancestral, proto-Pama–Nyungan system had four separate terms (on the left of the diagram below). In different groups of languages, this shifted to either a two-term system (merging sexes, like English), or a three-term system merging parallel grandparents (MM=FF), with a potential third stage (merging cross, MF=FM) emerging in some cases from there. Note that some of these transitions seem to have occurred without a stable intermediate state.



As well as making discoveries regarding systems of grandparental terminology, Catherine and colleagues note that we still do not know why it is that these systems transitioned in that particular order. In their paper, the authors tested potential social factors that might drive these transitions (community marriage organisation, and post-marital residence) but found no significant correlations. There was minimal co-evolution with the rest of the terms in the kinship system, and it remains an open question what linguistic or social forces shape grandparent terms.

New paper on Murrinhpatha children’s understanding of kinship lexicon and grammar (post by Dr Alice Mitchell)

Among the vast number of things children need to learn about language is how to appropriately refer to other people. One way to do this is to use kinship terms—words like ‘great-grandmother’, ‘brother-in-law’, or ‘sister’. The particular set of kinterms a child needs to know will obviously depend on the languages they’re learning to speak.


English-speaking children learn different words for mother and mother’s sister (‘aunt’), while speakers of many other languages learn a single term for these relatives.For example, in Murrinhpatha, a language spoken in Wadeye in Australia’s Northern Territory (see map), the word kale can refer to both mother and mother’s sister, among other relations. How do children figure out who can be grouped together under the label kale? From a broader perspective, how do children learn the kinship-related language used in their community?


A new open-access paper just published in Language, ‘Acquiring the lexicon and grammar of universal kinship’, explores this question in Murrinhpatha. The paper was authored by Joe Blythe (Macquarie University), Jeremiah Tunmuck (Yek Yederr), Alice Mitchell (University of Cologne), and Péter Rácz (Central European University). Alice and Peter joined the project while they were post-docs working on the Varikin project, having met Joe at a kinship workshop held at the MPI for Social Anthropology in Halle, Germany. Joe then visited the EXCD lab in January 2018 as a Benjamin Meaker Visiting Professor. After that their collaboration continued online across distant time zones.


One of the special things about Murrinhpatha kinterms is their ‘universal’ character, by which we mean that almost anyone in the community can be referred to with a kinterm. Working out how to refer to a newly introduced individual involves a kind of mental calculation based on telescoping chains of relatedness: if I know someone is my mother’s mother’s sister’s daughter, I can reduce that down to ‘mother’ and also refer to her as kale.[1] For Murrinhpatha people, information about kinship relations is so important that it’s expressed not just in vocabulary but also in the grammar of their language. In a sentence about two people doing something, part of the verb indicates whether those two people are siblings or not. For example, parraneriwakthadharra means, roughly, ‘They were following’, where it is grammatically specified that ‘they’ refers to two siblings.


Interested in how children learn this kind of kinship-related vocabulary and grammar, Joe designed two tasks: the ‘kinterms’ task, testing understanding of kinterms, and the ‘kintax’ task, testing understanding of the grammatical categories relating to kinship. In the kinterms task, Joe and Jeremiah showed children photos of their own relatives and asked them questions about kinship relations. There were three different types of questions: first, the researchers asked children straightforward questions, in Murrinhpatha, like ‘is this your cousin?’, where the children were expected to answer ‘yes’ or ‘no’. The researchers then went through a new series of photos and asked children what they called the person in each photo. In the third part, children were asked what the person shown in the photo would call them. Other research has shown that young children often struggle to identify who they are from someone else’s perspective, so we expected this part of the task to be the hardest. This vocabulary-oriented task was carried out with 24 children aged between five and sixteen.


Unsurprisingly, older children performed better on this task. We didn’t see any great age-related breakthroughs but rather a gradual improvement in children’s understanding of kinterms. We also found evidence that closer kin are easier for children to classify. This result is fairly intuitive, too, but what is notable here is that children’s learning partly follows culture-specific ideas about the relative ‘closeness’ of kin: children made fewer mistakes labelling a parent’s same-sex sibling than a parent’s opposite-sex sibling. While these two kinship categories can be considered similarly ‘close’ from a genealogical perspective, the language differentiates them, categorising a parent’s same-sex sibling as a parent, while providing a different term for a parent’s opposite-sex sibling. Nonetheless, biological parents were the easiest of all to identify, emphasising the importance of ‘closeness’ for learning kinship concepts, whether defined genealogically, culturally, or experientially. Our results also supported earlier research in other languages that taking someone else’s perspective on kinship relations is cognitively challenging.


The second task targeted participant’s understanding of ‘kintax’. Children saw a brief animation of people doing something, e.g., waving. At the same time they heard a sentence describing the activity, where the verb in the sentence indicated whether the people were siblings or not. Children were then shown two photos of people in the community, one showing siblings and the other non-siblings, and children were asked to choose the appropriate photo (see right for an example slide).


They were also shown slides and heard similar sentences that tested the understanding of number and gender, so these contrasts could be compared to their understanding of siblinghood. Joe and Jeremiah conducted this task with 39 Murrinhpatha speakers ranging in age from five to 40 years old. The most important finding here was that children’s understanding of kinship grammar progresses at similar rates to other grammatical categories like participant gender and number. Based on results from both the tasks, a complex picture emerges in which linguistic categories, general cognitive abilities, cultural practice, and individual experience all play a role in learning kinship vocabulary and grammar.


This new study moves our field forward on several fronts. It presents the first quantitative investigation of the acquisition of an ever-expanding kinship system; it’s the first study to investigate children’s acquisition of kinship-related grammar; and it’s also innovative in the way it tailored the experimental stimuli to each participant’s own family. This design feature meant that responses were not always directly comparable, which in turn restricted the statistical power of our analysis. Nonetheless, our results showed that children build up a gradual understanding of kinship that focuses on their closest relatives and expands to others as they get older. This supports a ‘focal’ theory of kinship terms, where at least some kinship categories are built from central exemplars and then extend to include more peripheral members. While our approach measured what children know, the next step is to explore how children learn kinship terms—a question that will involve more qualitative methods.


As part of my fieldwork with Datooga-speaking children in Tanzania, I’ve been conducting similar studies of children’s understandings of kin terms. Watch this space for updates on more papers addressing this topic.


[1] For the curious, this works via what anthropologists call ‘merging’ principles: a mother’s mother’s sister is classified as a mother’s mother, and a mother’s mother’s daughter is classified as a mother.

A second-hand treasure

Finding classic ethnography in second-hand bookstores or charity shops is one of my great pleasures. This weekend I rumbled a copy of Te Rangi Hiroa Sir Peter Buck’s “The Coming of the Maori” for the bargain sum of £2.50.

It’s a bit mildew-spotted on the outside but in great condition inside.



I wonder who Barbara was to get such a scholarly and interesting present from her Dad in 1958? (That’s my bookstamp there, yes, isn’t it great? I like ferns. A gift from my ever-thoughtful sweetheart.)



One of the plates at that back: a collection of woven flax kits (kete). I remember trying to do even the most basic weaving as a kid and being astonished at the fine detail that skilled weavers can produce. The middle-left is particularly lovely.

Te Rangi Hiroa was an amazing man – an anthropologist, politician, doctor, health campaigner, and served in the armed forces – and all this at a time when discrimination against Māori people in public life was routine in New Zealand. The article linked here mentions the epilogue from “Vikings of the Sunrise” where Te Rangi Hiroa considers the passing of a traditional Polynesian way of life:

“The old net is full of holes, its meshes have rotted, and it has been laid aside.

What new net goes afishing?”

I had not realised before that this was where Witi Ihimaera took the title of his book of short stories “The New Net Goes Fishing”. Those stories were classic high-school reading when I was growing up, and I can see how the stories resonated with Te Rangi Hiroa’s theme.

[scrapbook] On generalizing from case studies

Köbben (1970 and in other papers) was concerned with the folly of making general statements based on case studies.

In his study of the Siriono of Bolivia, Holmberg (1950) concluded that hunting and gathering tribes tended to be underfed and obsessed by food. From his celebrated study of Western European nations during the 1880s and 1890s, Durkheim (1951 [1897]) concluded that in general social isolation tends to drive a person to suicide. Hauser’s study (1959) of the Thai led him to believe that in general the more atomistic a society, the more it would resist modernization. Raulin (1959), studying the people of Gagnia and Daloa, concluded that uprooted peoples would be more interested in modernization than those still at home in the land of their ancestors.

But, Needham, (1954), studying the Punan of Borneo, concluded that hunting and gathering tribes were usually well fed and unobsessed by food. Asuni (1962), studying the people of western Nigeria, concluded that social isolation had nothing to do with suicide. Adair and Vogt (1949), studying the Zuni, concluded that the less atomistic a society, the more it would resist modernization. De Waal Malefijt (1963), studying the Javanese, concluded that up-rooted peoples would be less interested in modernization than would stay-at-homes.

Köbben, A. 1973 [ 1970] Comparativists and Non-Comparativists in Anthropology. In A Handbook of Method in Cultural Anthropology, Eds. Naroll, R. & Cohen, R. pp. 581- 596.

There’s an interesting discussion of Köbben’s approach to comparison, along with other 20th-century Dutch anthropologists in this chapter:

de Wolf, J. J. 2002. Conditions of comparison : a consideration of two anthropological traditions in the Netherlands. In  Anthropology, by comparison. Eds. Gingrich, A. and Fox, R.G. London; New York : Routledge.

[scrapbook] the fundamental and the freakish

From Adam Kuper’s Huxley lecture:

Kuper, A. 2008. “Changing the Subject About Cousin Marriage, Among Other Things.” The Journal of the Royal Anthropological Institute, 14, 717-735.

Malinowski never delivered a Huxley lecture, but I will give him the last word in mine. In a note for a never completed textbook he remarked that when he came into anthropology the emphasis had been on the differences between peoples. ‘I recognised their study as important, but underlying sameness I thought of greater importance & rather neglected. I still believe that the fundamental is more important than the freakish’ (cited in Young 2004: 76).

“The fundamental is more important than the freakish”.  A good essay topic?


[scrapbook] On the value of comparative anthropology

From: Burton, M, C Moore, J Whiting, A Romney. 1996. “Regions Based on Social Structure.” Current Anthropology 37 (1) (February 1): 87–123.

I often go back to this paper. The analyses themselves are interesting (bottom-up culture regions based on the data rather than geography) but problematic (correspondence analysis on non-phylogenetically-controlled data). But the remarks in the Reply are excellent arguments for comparative analysis.

Scholarly work is by necessity done slowly and carefully. We cannot have an instantaneous image of all of the world’s contemporary societies ready for comparative analysis. There will always be a long lag time from the collection of data to the time when systematic analysis is possible. If our profession allows the findings of long-term projects to be ruled out of court as “old-fashioned,” it will discourage the collection of systematic data in large long-term projects. In our view the trendiness of anthropology is one of the major problems of our field.

And then to end:

The distrust that many anthropologists show toward comparative research is based on misinformation, logical errors, or perceptions of methodological problems that either have been corrected or are in the process of being corrected. While there are always changes in scientific standards over time, the value of cumulative empirical research, which necessarily has a long gestation period, outweighs any possible costs to the use of data that may not have been collected according to a currently fashionable theoretical program. There is no need for false dichotomies between text and numbers, between old data and new data, between description and comparison, or between microscopic and macroscopic approaches.

[scrapbook] Morgan’s Question.

From LH Morgan’s introduction to “Systems of Consanguinity and Affinity of the Human Family“:

The tables, however, are the main results of this investigation. In their importance and value they reach far beyond any present use of their contents which the writer may be able to indicate. If they can be perfected, and the systems of the unrepresented nations be supplied, their value would be greatly increased. The classification of nations is here founded upon a comparison of their several forms of consanguinity. With some exceptions, it harmonizes with that previously established upon the basis of linguistic affinities. One rests upon blood, the preponderance of which is represented by the system of relationship; the other is founded upon language, the affinities of which are represented by grammatical structure. One follows ideas indicated in a system of relationship and transmitted with the blood ; the other follows ideas indicated in forms of speech and transmitted in the same manner. It may be a question which class of ideas has been perpetuated through the longest periods of time.

It strikes me that I am working on both Morgan’s Question and Galton’s Problem.

on culture and language [scrapbook]

If it can be shown that culture has an innate form, a series of contours, quite apart from subject-matter of any description whatsoever, we have a something in culture that may serve as a term of comparison with and possibly a means of relating it to language. But until such purely formal patterns of culture are discovered and laid bare, we shall do well to hold the drifts of language and of culture to be non-comparable and unrelated processes.

Sapir (1921) Language: An Introduction to the Study of Speech.