44.  There is a large literature on these issues and on the current consensus on Austronesian history and migration. Large tracts are summarized in Geoffrey K. Chambers and Hisham A. Edinur, “Reconstruction of the Austronesian Diaspora in the Era of Genomics,” Human Biology 92, no. 4 (2020): 247–263. See also Terry Melton, “Genetic Evidence for the Proto-Austronesian Homeland in Asia: mtDNA and Nuclear DNA Variation in Taiwanese Aboriginal Tribes,” American Journal of Human Genetics 63, no. 6 (1998): 1807–1823; Elizabeth Matisoo-Smith, “Tracking Austronesian Expansion into the Pacific via the Paper Mulberry Plant,” Proceedings of the National Academy of Sciences 112, no. 44 (2015): 13432–13433.

45.  Jerry Norman and Tsu-Lin Mei, “The Austroasiatics in Ancient South China: Some Lexical Evidence,” Monumenta Serica: Journal of Oriental Studies 32, no. 1 (1976): 274–301.

46.  AAAS, “Ancient DNA Reveals Genetic History.”

47.  AAAS, “Ancient DNA Reveals Genetic History.”

48.  Chiang et al., “Comprehensive Map,” 2736.

49.  Chiang et al., “Comprehensive Map,” 2736.

50.  Erika E. S. Evasdottir, Obedient Autonomy: Chinese Intellectuals and the Achievement of Orderly Life (Vancouver: University of British Columbia Press, 2004).

51.  Professor Fu is also a coauthor on one of the papers I have discussed above. See Yang et al., “Ancient DNA Indicates Human Population Shifts.”

52.  Chris Buckley, “Brushing Off Criticism, China’s Xi Calls Policies in Xinjiang ‘Totally Correct,’ ” New York Times, September 26, 2020.

53.  “DNA” and “gene” are obviously not the same thing, but the Chinese term for DNA is too complex to be used by politicians in this way. Thus, the acronym DNA and the Chinese word for gene, jiyin, are used interchangeably, as if they referred to the same thing: a buried code that preserves and transmits an essence. The compound literally means “fundamental cause” or “driver.” Like many modern Chinese terms in science, it may have been first coined in Japan.

54.  Jinping Xi, “Jianshe zhongguo tese Zhongguo fengge Zhongguo qipai de kaoguxue. Geng hao renshi yuanyuan liuchang boda jingshen de Zhonghua wenming” [Build an archaeology with Chinese characteristics, Chinese style, and Chinese spirit: Better investigate the ancient, grand, and profound Chinese civilization], Xinhua Net, November 30, 2020, http://www.xinhuanet.com/politics/2020-11/30/c_1126803588.htm. See also the official party newspaper Guangming Ribao’s article in the same month, setting out the official truth on these matters: “Cong kaogu faxian kan baqian nian yilai zaoqi zhongguo de wenhua jiyin” [Looking at the cultural genes of early China in the past 8,000 years from archaeological discoveries], Xinhua Net, November 4, 2020, http://www.xinhuanet.com/politics/2020-11/04/c_1126695556.htm.

55.  Xi, “Jianshe zhongguo.”

56.  See note 41 above. In reality, complicity and self-censorship are widespread, and therefore do not often come to light.

57.  Bai Tiantian, “Official Urges Archeologists to Prove Xinjiang Part of China since Ancient Times,” Global Times, March 23, 2017. On the policies, see Jilil Kashgary and Kurban Niyaz, “Chinese Research on Xinjiang Mummies Seen as Promoting Revisionist History,” Radio Free Asia, June 11, 2022, https://www.rfa.org/english/news/uyghur/melting-pot-06032022104308.html.

58.  As indicated in note 18 above, the so-called white papers are the government’s statements of the official truth on matters of history, the ownership of Tibet, and so on.

59.  State Council Information Office of the People’s Republic of China, Historical Matters Concerning Xinjiang (Beijing: Foreign Languages Press, 2019).

60.  State Council Information Office, Historical Matters.

61.  State Council Information Office, Historical Matters.

62.  Wenjun Wang et al., “Ancient Xinjiang Mitogenomes Reveal Intense Admixture with High Genetic Diversity,” Science Advances 7, no. 14 (2021): 1. This research comes from the new aDNA laboratory at the Institute for Vertebrate Paleontology and Paleoanthropology in Beijing. While the institute has long languished among the paucity of human fossils found in China, it has been reinvigorated with its new mission in genetics and, especially, aDNA. Like all research institutions in China, it is closely controlled by the ruling Communist Party.

63.  Wang et al., “Ancient Xinjiang Mitogenomes,” 9.

64.  Qidi Feng et al., “Genetic History of Xinjiang’s Uyghurs Suggests Bronze Age Multiple-Way Contacts in Eurasia,” Molecular Biology and Evolution 34, no. 10 (2017): 2572.

65.  Chunxiang Li et al., “Evidence that a West-East Admixed Population Lived in the Tarim Basin as Early as the Early Bronze Age,” BMC Biology 8, no. 15 (2010): 1–12.

66.  Shi-Zhu Gao et al., “Ancient DNA Reveals a Migration of the Ancient Di-Qiang Populations into Xinjiang as Early as the Early Bronze Age,” American Journal of Physical Anthropology 157, no. 1 (2015): 71–80.

67.  See, for instance, James Patrick Mallory and Victor H. Mair, The Tarim Mummies: Ancient China and the Mystery of the Earliest Peoples from the West (London: Thames & Hudson, 2000). This book was written before the emergence of archaeogenomics. There have been many criticisms of these arguments, taking issue with the authors’ naming of the mummy populations as “Westerners” or “Europeans.”

68.  Fan Zhang et al., “The Genomic Origins of the Bronze Age Tarim Basin Mummies,” Nature 599, no. 7884 (2021): 256–261.

69.  Bruce Bower, “DNA from Mysterious Asian Mummies Reveals Their Surprising Ancestry,” Science News, October 27, 2021, https://www.sciencenews.org/article/ancient-dna-asian-mummies-bronze-age-ancestry-mystery; Holly Chik, “Xinjiang Mummies’ Origins Closer to Home than We Thought, Study Finds,” South China Morning Post, October 27, 2021.

70.  Chik, “Xinjiang Mummies’ Origins.”

8   Concluding Commentary

K. Ann Horsburgh

A great deal of ink has been spilled over the problems exhibited by researchers attempting to incorporate ancient DNA (aDNA) data into anthropological models of prehistory. The problems are both ethical and interpretive. Several of the chapters in this volume have compellingly and comprehensively discussed the ethical violations abundantly seen in aDNA research. They show that since there continue to be significant professional rewards for rapid and sensational publication—a phenomenon explored in detail in Andreas Nyblom’s chapter on the fuss surrounding the “the female Viking warrior”—and few professional consequences for riding roughshod over ethical standards and the concerns of descendant and stakeholder communities, researchers will continue to make their own culture-bound value judgments. The research community itself must decide whether the ethical requirements of engaging in aDNA research should have teeth, by whom those teeth should be wielded, and what kinds of wounds those teeth should inflict.

I want to thank the editors of this volume for the opportunity to provide this commentary. I will use the space available to me to discuss two phenomena related to anthropological aDNA research: one relevant to the community of scholars interested in molecular approaches to the past, and one relevant to both public understandings of the kinds of research discussed in this book and to the ways in which scholars untrained in genetics understand the data generated. First, however, I would like to be explicit about my positionality in this space, because the peculiarities of my academic background significantly shape my views on the current state of affairs. I am a molecular anthropologist trained in the four-field Americanist tradition, in which the subdisciplines of biological anthropology, archaeology, linguistics, and sociocultural anthropology sit together under the umbrella of anthropology. I was trained in this way first in my home country of Aotearoa New Zealand, which—by virtue of a historical quirk—follows an Americanist anthropology rather than the British anthropology more typical of Commonwealth countries, and then later in the United States. While I hold undergraduate degrees in both anthropology and the biological sciences, all my advanced training is in anthropology, not biology. I came to molecular anthropology having read while in high school popular accounts of Allan Wilson’s and Vince Sarich’s immunological research devoted to dating the evolutionary split between humans and the rest of the great apes, and of Rebecca Cann’s PhD research, which resulted in the naming of Mitochondrial Eve.

The vast majority of the research in which I have been involved has used modern and ancient DNA of both human ancestors and nonhuman animals in Africa and the islands of the Pacific Ocean. In the last few years, however, my research interests have substantially shifted. Laboratory skills are significantly transferable, and I am now using molecular approaches to interrogate the complex interactions between human health and cultural and social phenomena, like those discussed in Venla Oikkonen’s chapter about how pathogens in the thawing permafrost alter our conceptions about the future. I am therefore an insider. I have spent many years generating aDNA data, and I love the lab. But I am an insider with one foot out the door. It is from this vantage point, then, that I first want to comment on a category error that is a routine part of discussions of the use and misuse of aDNA data in reconstructions of prehistory, and second, to discuss the relationship between the public’s love affair with anthropologically relevant aDNA and the meteoric rise of direct-to-consumer (DTC) genomics.

A Category Error

The category error is routinely made in the discussion of interdisciplinarity and transdisciplinarity in the generation and interpretation of aDNA data in anthropological contexts. The category error is that which frames this kind of research as necessarily a collaboration between anthropologists/archaeologists and geneticists. As shown in Charlotte Mulcare and Mélanie Pruvost’s chapter about the problematic nature of interdisciplinarity in aDNA studies, the discussion is typically framed around the divergent norms of anthropologists trained in the social sciences and geneticists trained in the life sciences. These differences are said to result in difficulties in communication, nonalignment of priorities, and failures to understand the nuances and complexities of data generated in a different discipline. All these things are indeed difficult, and it is certainly true that the largest, best-funded aDNA laboratories are run by geneticists. However, discussions framed in this way render invisible the many anthropologists who run aDNA laboratories.

The reasons why archaeologists prefer to collaborate with geneticists rather than molecular anthropologists (or anthropological geneticists) are both real and understandable.¹ Geneticists specializing in aDNA tend to work in departments of biological sciences (which come with a variety of names) or medical schools. They have considerably lower teaching loads than people based in anthropology departments, and therefore have more time available for research activities. They are also almost universally better resourced than people in social science departments, both in terms of the support services provided and in terms of research money. Collaborating with geneticists, then, is simply easier than collaborating with molecular anthropologists.

Meanwhile, archaeologists lamenting the lack of real engagement by geneticists are commonplace. Frequent complaints are heard about geneticists ignoring the input of social scientists, and even about manuscripts being published without the archaeologists seeing (never mind approving) the text. I have a great deal of sympathy for archaeologists who are frustrated with the ways in which genetic data are interpreted, but they are not powerless here. If proper, thoughtful, fully contextualized interpretation of anthropologically relevant genetic data matters to them, then there are two sensible courses of action. They can continue to collaborate with geneticists but make their full participation in interpretation and make the writing of the resulting papers a condition of facilitating access to archaeological remains. Archaeologists and the curators of collections have considerable power to dictate terms and demand meaningful collaborations. The field will be better off if they begin wielding that power more widely.

The other option open to archaeologists is to collaborate with anthropologists rather than biologists. This is the more difficult road. Vanishingly few molecular anthropologists have standing budgets, and even fewer have the armies of postdoctoral fellows and laboratory technicians enjoyed by the largest and best-funded genetics labs. This approach requires winning grants before work can begin. It also requires an appreciation that with a smaller workforce come necessarily slower results. I am not an archaeologist, but I do suggest that archaeologists might find collaborations with other anthropologists rewarding, despite the inherent challenges.

Collaborations across disciplines, or even subdisciplines, is necessarily time-consuming if it is to be done well. Much of the archaeological dissatisfaction in the interpretations of ancient DNA data stems from the difficulties in communication across disciplines, particularly around the uncertainties and messiness inherent in archaeological data. Genetic data from archaeological sources, which make up most of the data discussed in this volume, are also necessarily messy.² Many aDNA researchers describe what they do in the laboratory as experiments. While aDNA lab work does have the veneer of the experimental, it is by necessity a historical science, not an experimental one.³ While some biologists are trained in the historical sciences (e.g., paleontology), most are not, and geneticists are especially unlikely to be. As the shared social science training of archaeologists and molecular anthropologists includes a common training in the (pre)historical, they can begin their collaborations with a shared appreciation of the methods, limitations, and complexities of historical data.

Molecular Chauvinism

As demonstrated by several chapters in this volume, and perhaps most clearly in Magnus Fiskesjö’s chapter on the political abuse of archaeogenetic research in China, aDNA data are published by people incentivized to go beyond their data and make grand—but unsupported—claims. These claims are then understood by both producers and consumers of genetic data with a bias I have elsewhere named “molecular chauvinism.”⁴ Molecular chauvinism is the tendency to treat genetic data as though they are untarnished by the messiness attendant on other data in the social sciences and, consequently, to privilege them over other kinds of data. The implicit assumption is that the distance between data and interpretation is shorter and less complicated, and so the interpretation is more likely to be accurate. You need only contemplate the claim of geneticist David Reich that “human genome variation has surpassed the traditional toolkit of archaeology—the study of the artifacts left behind by past societies—in what it can reveal of changes in human populations in the deep past” to be convinced that I am not overstating the case here.⁵

Ancient DNA data are consumed by people largely untrained in genetics, embedded in a culture excited about the possibilities of personalized medicine and captivated by recreational DTC genomics. The Human Genome Project was completed in 2003 and it was supposed to usher in a healthcare revolution in which we would all benefit from personalized medicine tailored to the specifics of our individual genomes.⁶ While the tangible health benefits flowing from genome sequencing have been few and far between, the public enthusiasm for genetic data has shown no sign of abating.

One of the largest DTC recreational genomics companies, 23andMe, was founded only three years after the completion of the Human Genome Project. It was followed by Ancestry DNA in 2012. By 2018, the global market value for DTC genomics reached US $830 million, and it is anticipated that by 2025, the global DTC genomics market will value more than US $2.5 billion.⁷ People spit in tubes and hand over money because of a perception that the results they obtain from these companies are meaningful. As Daniel Strand and Anna Källén observe in the introduction to this volume, the popular discourse on DNA has long reproduced the assumption that our genes offer a source of truth about who we “really” are. So pervasive is the notion that our essential selves are to be found in our DNA that is has become a metaphor for realness—characteristic X is “baked into our DNA.” In a way, this idea also comes to the fore in Amade M’charek’s chapter about the changing representations of the Neanderthal: as soon as it is proved that modern humans share DNA with Neanderthals, this hominin group ceases to be portrayed as a monkey or a troglodyte and is instead depicted as “one of us.”

The meteoric rise of DTC companies and their broad uptake by the general public for both ancestry estimation and detection of health-related variants exacerbate a tendency toward genetic determinist thinking that is planted in peoples’ thinking in high school biology classes about Gregor Mendel. You could be forgiven for thinking that we are quite good at examining DNA sequences and predicting important phenotypic traits. We use human characteristics to teach Mendelian inheritance in both high school and at the university. We teach that, for example, the ability to roll your tongue is dominant over the inability, and so if you have TT or Tt combinations of alleles you will be able to roll your tongue and will be unable to roll your tongue only if you have tt alleles. We further teach that if two tongue rolling Tt people have children, about a quarter of their kids will be tt and thus unable to roll their tongues. Unfortunately for this favorite of the classroom, tongue rolling ability does not work like that at all. Instead, it is likely influenced by multiple genes, and there is probably a significant environmental influence as well.⁸

The nongenetic component of tongue rolling ability points to the fact that the actions of genes are influenced by the environments in which they find themselves; but further, the actions of genes are impacted by the genomes in which they find themselves. To illustrate this point, consider another favorite of the classroom, which is indeed inherited in a Mendelian fashion: the ABO blood system. At the end of the long arm of chromosome 9 is a gene which encodes an enzyme that comes in three alleles.⁹ If the A allele is present, a sugar called N-acetylgalactosamine is added as a fifth sugar on the surface of the red blood cell. If the B allele is present, then the sugar galactose is added to the existing string of four sugars. The O allele is nonfunctional, and so does nothing. Thus, if you have inherited an O allele from each of your parents, you have no fifth sugar on any of your red blood cells and have type O blood. When considered in this way, it is then easy to see both why A and B alleles are dominant to O, and why A and B alleles are co-dominant to each other.

Genes do not, however, act is isolation. They act in functional biological organisms, as part of nested physiological pathways. There is another gene, on the long arm of chromosome 19, which is relevant to a person’s ABO status. This gene (FUT1) similarly encodes an enzyme which moves sugars around.¹⁰ In most people, the enzyme encoded by FUT1 adds the sugar fucose to a string of three sugars on red blood cells. In people who inherited the recessive, nonfunctional version of this gene (genotype hh), that sugar is never added. People carrying this combination of alleles have what is known as the Bombay phenotype. It does not matter what alleles are carried at the ABO gene, because those enzymes have no substrate upon which to act. Regardless of the alleles carried at ABO, a person with the Bombay phenotype will have functionally O-type blood.¹¹ There are no reported health effects associated with the Bombay phenotype, until someone needs a blood transfusion. Transfused with typical O-type blood with red blood cells with four sugars on their surfaces, the alleles of the ABO system suddenly have a task they can complete, and add fifth sugars to the newly transfused red blood cells. The patient’s system now recognizing the transfused blood as nonself launches an aggressive immune response.¹²

The ABO system, then, is a very straightforward Mendelian system. Except when it is not. Genes are obviously important in the building of bodies, but the relationships between genes and traits are not simple or unidirectional and almost never involve single genes and single characteristics. We know much more about the genetics of disorder than we do about the genetics of the average, so genetic disease is a useful place to think about the relationships between genotype and phenotype. Of the diseases we know to show patterns of inheritance, only about 2 percent show patterns consistent with simple Mendelian inheritance.¹³ The rest are influenced by many genes interacting in complex reticulated networks with each other and with other contributors to the developmental system, including the environment and nongenetic inheritance.¹⁴ Their influence is probabilistic, not deterministic.

Final Thoughts

Despite grandiose claims to the contrary, aDNA data have not, and will not, revolutionize reconstructions of the past. Certainly, there has been a dramatic increase in the volume of genetic data relevant to reconstructions of prehistory, but our powers of interpretation have lagged behind the truly impressive improvements in laboratory-based techniques for DNA recovery and analysis.

Further, genetic data are meaningful only to the extent to which they articulate with other kinds of data, which means that engagement across specialties has to be careful, equitable, and undertaken with a humility about the boundaries of one’s own knowledge and a respect for the epistemological rigor in diverse disciplines. As suggested by Stewart B. Koyiyumptewa and Chip Colwell’s chapter about the conflicts between paleogeneticists and Indigenous groups in the United States, a productive future for aDNA research hinges on an engagement with descendant communities and their inclusion as core members of research teams, drivers of research agendas, and lead investigators.

Finally, the field of aDNA studies would benefit enormously from a recalibration of its core mission and its core values. If the research community is interested in accurate, nuanced, and rich reconstructions of the past, we need to stop reaching for the flashy press release, slow down, and truly grapple with the intricacies of synthesizing multiple, complex datasets. As made clear in Marianne Sommer and Ruth Amstutz’s chapter about the graphic representations employed in aDNA research, complex historical processes must not be reduced to simple figures or models.