The sciences of racialization

The historical phenomenon we refer to as ‘race science’ was not a single institutionalized discipline, but rather the convergence of concepts and methodologies around the shared problematic of human biological variation. It emerged from older ideas in medicine and natural history that eventually fragmented into more specialized disciplines during the nineteenth century: anatomy and physiology, anthropology and ethnology, archaeology, philology and linguistics, psychology and phrenology: a set of fields now sometimes lumped back together under the label ‘human sciences’. Although ‘race science’ did exist as a historical actors’ category, it seems to be more frequently used now by historians than in the past by its practitioners (at least in English – the reverse may be true for German Rassenkunde and French raciologie). For the most part, ‘race science’ was not something one could directly earn a degree in; the majority of race science experts were formally trained in medicine, philosophy or anthropology. However, race science ‘had a history and coherence of its own to its practitioners’ and an ‘internal logic of scientific arguments’ forming the basis of local and transregional intellectual communities.Footnote ⁷

Historians’ efforts to understand these scientific communities under the rubric of ‘race science’ have been deeply influenced by the work of Nancy Stepan, whose 1982 book The Idea of Race in Science popularized ‘race science’ as a historiographical phrase. She described her book as an ‘introduction to the history of scientific racism in Britain’, which has contributed to a lingering tendency among both historians and the general public to treat the terms ‘race science’ and ‘scientific racism’ as interchangeable.Footnote ⁸ Nevertheless, ‘scientific racism’ has a distinct historiographical genealogy rooted in post-war twentieth-century efforts to delegitimize lingering vestiges of race science as pseudoscience and relegate it once and for all to the fringes of professional science.Footnote ⁹ The particular formulation of ‘scientific’ racism (with only the word ‘scientific’ in quotation marks), popularized in the early 1960s, was simultaneously used by the American historian of Africa Philip D. Curtin to critique nineteenth-century British imperial ideology, and by the Spanish anthropologist Juan Comas in his scathing rebuke of the first publication of the journal Mankind Quarterly.Footnote ¹⁰ While ‘scientific racism’ is a meaningful term, for our purposes, it is important to set aside this term's connotations of pseudoscience and speak instead of ‘race science’ to accurately depict a series of inquiries and debates that, in their time, were based upon widely accepted scientific methodologies and theoretical frameworks and expressed in recognized scientific venues.

The origins of race science proper are conventionally dated to the late eighteenth century, when European naturalists constructed taxonomic systems of organisms that included divisions of the human species into ‘varieties’ or ‘races’.Footnote ¹¹ However, the social and intellectual processes of racializing human groups long pre-date formal race science. Ideas of human group differences already existed in all parts of the world during the premodern period; in Europe, West Asia and North Africa these ideas were rooted in humoural medicine, the Graeco-Roman geographic theory of climes, religious genealogies of human origins, cultural practices of recording ancestral lineages, and prejudicial hierarchies of civilizational greatness. Historians still debate the intellectual, social and political processes through which these various forms of pre-racial thinking were gradually transformed into the notion of distinct human races defined by physiological differences detectable in the skin, bones and brain.

Researchers largely agree that early modern medical practice, especially in the context of the African slave trade, played a significant role, leading to Suman Seth's suggestion that race medicine was the immediate predecessor of modern race science.Footnote ¹² Christian theology also fundamentally shaped how physicians and naturalists thought about the possibility of human races having shared or distinct origins.Footnote ¹³ What is indisputable is that race science reached its height in the mid-nineteenth century, as European empires reached the height of their power and scientific methodologies for constructing racial classifications rapidly proliferated. Generalized fields of study like anatomy and philology produced specialized techniques and instruments that lent themselves well to identifying racialized differences: anthropometry, craniometry, phrenology and linguistics all found their way into the emergent disciplines of ethnology and anthropology.Footnote ¹⁴ However, widespread scholarly and public interest around race as a scientific category did not lead to consensus about the definition of race. Every aspect of race science, from the number of human races to the relative value of classifying methodologies, was vigorously contested. Regardless, these intellectual debates were unified by their links to colonial socio-economic structures. Colonial spaces, including settler colonies, became sources of empirical data for the study of racial difference; the resulting theoretical products of race science, in turn, provided the justification for practising racial discrimination in the colonies. In other words, race science as an epistemological domain was co-constituted with the racial discrimination inherent to colonial governance.Footnote ¹⁵ Colonies became ‘laboratories’; that is, sites for experimental practices ranging from medical concentration camps to new educational curricula.Footnote ¹⁶ The different legal status and racial classification of colonized peoples enabled Europeans to conduct research that was unfeasible, unpopular or even banned in the metropole.Footnote ¹⁷

However, by the early twentieth century, contestations regarding race science methodologies and categories reached a critical point. The reliability of key craniometric measures like the cephalic index came under attack by American anthropologist Franz Boas at the same time as emerging methods like serology, undergirded by Mendelian theories of heredity, seemed to contradict earlier systems of racial classification.Footnote ¹⁸ Elazar Barkan links together these scientific developments with the activism of American and European scientists and physicians against Nazi racial hygiene as evidence of the ‘retreat of scientific racism’ by the 1930s.Footnote ¹⁹ This chronology, with its implication that race science was already and everywhere in decline by the outbreak of the Second World War, has been challenged on multiple fronts. For example, new histories of race science centring Asia and Africa have revealed that race science continued to flourish in these regions well after the 1930s, practised not only by European colonial administrators but also by indigenous elite intellectuals for their own anti-colonial and elite nationalist purposes. A narrative of decline rooted in the North Atlantic thus overlooks the ongoing role of race science in decolonizing movements, from Egypt to India to Indonesia, during the early to mid-twentieth century.

On the one hand, Asian and African nationalist practitioners of sciences such as anthropology, archaeology and statistics rejected European claims about the racial inferiority of non-Europeans, attacking the methodological and analytical inconsistencies of race science to undermine the racial hierarchies proposed by European scientists. On the other hand, elite nationalists used methods and discourses similar to those of race science to demonstrate the existence of a coherent national unit capable of self-government. Measurements from racial anthropology and serology were leveraged in international decisions about the territorial boundaries of new nation states, and were used to arbitrate communal claims of national belonging. In the process, anti-colonial race scientists tended to reinforce and naturalize the social and religious hierarchies of dominant communities, contributing to forms of systematic exclusion and even ethnic cleansing. Thus race science functioned as a double-edged sword in the hands of nationalist movements, and continues to sustain ongoing racialized discrimination within postcolonial nation states.Footnote ²⁰

Equally significant to this history are non-European regimes of imperial expansion and colonial occupation, such as those of China and Japan. Han Chinese anthropologists conducted anthropological research on non-Han ethnic groups to argue that they comprised a racial unity, supporting Han settler colonialism in other regions under Chinese rule.Footnote ²¹ Japanese scientists were pioneers in race science methods like serology and dermatoglyphics, deployed on the inhabitants of their colonial territories stretching from Taiwan to Korea and Manchuria.Footnote ²² Japanese racial taxonomies of aboriginal peoples in Taiwan have had an enduring afterlife, shaping recent nationalist genetic research aiming to distinguish Chinese people in Taiwan from those living in the mainland People's Republic of China.Footnote ²³ Furthermore, post-war Japanese scientists involved in the International Biological Programme pursued personal collaborations with Carleton Coon and Reginald Ruggles Gates in their research on ‘hybrid’ Japanese American children, despite American anthropologists’ ostracization of Coon and Gates as figureheads of outdated scientific racism.Footnote ²⁴

Interdisciplinary scholarship demonstrates that North Atlantic race science endures in contemporary forms of biological and medical theories, methods and practices. The malleability of what constitutes the marker of race – that is, the physiological site in which ancestry can be seen and measured – has enabled the persistence of racial research through population genetics, biomedicine, pharmacogenomics and forensics. Notwithstanding scientists’ terminological shifts from racial type to ethnic group, population, and biogeographical ancestry, the notion that racial categories capture something biologically meaningful continues to persist. While individuals are no longer considered to belong to pure racial types, specific groups have been targeted as especially desirable for the study of human genomic diversity: namely geographically and socially ‘isolated’ groups whose histories of marginalization make them vulnerable to extinction as distinct communities.Footnote ²⁵ Meanwhile, individuals considered to belong to ‘admixed’ populations may have their DNA analysed into a precisely calculable mixture of African, Asian or European ancestry, thus entrenching these categories as distinct typological populations rather than gradients of genetic variation.Footnote ²⁶ Social scientists like Dorothy Roberts and Nadia Abu El Haj, among others, show how the biological reductionism of nineteenth-century race science has successfully adapted itself to the profit logics of neoliberal capitalism and its allied imagination of biological citizenship. Thus, even while neoliberal states recognize the claims and possibilities of multiracial identities, the biological reality of race is reinscribed through the recognition of medical interventions into racially specific diseases and drugs.Footnote ²⁷

Racial economies of science

The imperial turn in historiography has shown the extent to which science is part and parcel of global histories of slave labour, colonialism and other dimensions of economic exploitation. In this regard, modern science is irrevocably embedded in the rise of racial capitalism. Drawing on the work of Destin Jenkins, Justin Leroy and other scholars of the Black Marxist tradition, we consider all capitalism, ‘from the beginnings of the Atlantic slave trade and the colonization of the Americas onward’, to constitute racial capitalism ‘in material profitability and ideological coherence’.Footnote ²⁸ We therefore use this term not to imply the existence or possibility of a non-racial capitalism, but rather to highlight that ‘domination based on difference [is an] inciting incident’ for the history of capitalism.Footnote ²⁹ The expansion of European empires’ racialized domination of the global economy simultaneously generated both the funds and the motivation to carry out extensive new investigations of the natural world. With this in mind, we embrace James Poskett's suggestion that ‘rather than thinking of some sciences as racial, and others as not, we should therefore think of modern science as racial from the start’.Footnote ³⁰

Some Anglocentric examples can illustrate this phenomenon. Several eminent members of the Royal Society of London, including but not limited to Robert Boyle, Hans Sloane and Isaac Newton, had significant ties to slave-trading companies and slave-worked plantations until the abolition of slavery in the British Empire in 1833. They financed their scientific activities using stock dividends and other forms of profit from slavery. Furthermore, the role of slavery in the work of these British natural philosophers was not merely financial: the logistics of slave-based economies were inextricable from their intellectual interests and output.Footnote ³¹ For example, Boyle, Newton, Robert Hooke and their contemporaries were preoccupied with optics and the relation of light to colour. As such, they were interested in determining the underlying basis of different human skin colours, and the Royal Society avidly solicited observations on skin colour from travellers to Africa and the Americas throughout the seventeenth century.Footnote ³² The scientific stakes of this debate over skin colour were not solely about racialized anatomical differences, but also about physics, for which experiments with skin could test theories about the nature of colour perception more broadly.

European colonies in the Americas, sustained by slave-holding plantation economies, became sites for experimental research into this question. Eighteenth-century Creole scholars like physician–naturalist John Mitchell and physician–chemist Edward Bancroft took advantage of their access to enslaved Africans and indigenous peoples in Virginia and Dutch Guiana, respectively, to debate the physical and chemical underpinnings of colour. Both Mitchell and Bancroft engaged deeply with Newton's work on optics, derived from the latter's experiments refracting light through prisms. Mitchell's gruesome experiments on the skin of enslaved Africans led him to declare that there was no pigment or other substance within human skin producing colour variations. Rather, hailing Newton's theories of colour, Mitchell argued that the skin of Europeans, Africans and Native Americans had distinct variations in physical structure that produced different degrees of light reflectance and thus created the optical effect of skin colour variation.Footnote ³³

Bancroft did not work directly on the question of skin colour, but nonetheless his contributions to chemistry repeatedly invoked racial differences. Bancroft, who made his fortunes directly from colour in the form of dyes, rejected Newton's proposition that colour was solely a visual effect of light and not an inherent property of matter. According to Bancroft, different kinds of matter possessed colour according to ‘chemical affinities’ through which they reacted to light rays and to oxygen. Bancroft's perspective was partly influenced by his observations of the effects of certain plant dyes upon the skin of indigenous peoples and European colonists in Guiana, which had dramatic but always temporary darkening effects. Furthermore, Bancroft's account of the history of dyeing reinforced a racial hierarchy of chemical knowledge: while he acknowledged that even ‘savages’ were capable of manipulating colour and Indians of producing calico prints, only Europeans, through their understanding of philosophical chemistry, had been able to rationalize and perfect the technical art of dyeing.Footnote ³⁴

Much of Newton's own work did not concern humans at all, yet he nevertheless benefited from an early modern ‘information order’ that was strongly shaped by the colonial networks of racialized plantation economies.Footnote ³⁵ Newton never travelled outside Britain, so the insights of his vaunted Principia Mathematica depended on astronomical data and experimental observations of gravitational force collected by others travelling across the world. One significant example is Newton's use of the observations of Jean Richer and Jean Deshayes, French astronomers who experimented with pendulums in Gorée, Cayenne and Guadeloupe, major end points of the French trade in enslaved Africans in Senegal, Guiana and the Caribbean respectively. Their lengthy voyages to collect data were directly enabled by the slave-trading companies, who carried them aboard ships with human cargo. Newton's groundbreaking conclusions about geographical variations in gravitational force, based on the data of Richer and Deshayes, are thus inextricably linked to European enslavement of Africans.Footnote ³⁶

The racialized coercive regimes of manual and intellectual labour hidden within Newton's physics are rather easier to detect in the history of botanical sciences, where the importance of non-European agricultural and pharmaceutical knowledge was more readily acknowledged for its potential economic and medicinal value. Indeed, Europeans highly valued the botanical knowledge of Asian peoples, to the point of staking major economic ventures upon their racialized understanding of plant–cultivator relationships. In 1621 the Dutch East India Company, seeking to create a monopoly over nutmeg in the Maluku archipelago's Banda islands, massacred or exiled almost the entire indigenous population of 15,000 people and divided the islands into Dutch-controlled plantations to be worked by people enslaved and imported from other parts of South and South East Asia. The small number of Bandanese allowed to remain were also enslaved and ‘deliberately distributed about the islands to make use of their expertise in cultivation and spice production, with several hundred individuals initially exiled to Batavia being returned to the Bandas for this very purpose’.Footnote ³⁷

Attempts by other European empires to break the Dutch monopoly and cultivate smuggled nutmeg elsewhere, such as the French colony of Mauritius, were often foiled by Europeans’ lack of knowledge about the plant and its peculiar botanical features and ecological needs. Assuming that this knowledge was racially embodied, French officers brought two Maluku men to Mauritius in the 1770s, expecting that the nutmeg plants would thrive under their care; however, they were disappointed when the men (who had not previously worked in nutmeg cultivation) were unsuccessful. The French had to accumulate botanical knowledge of nutmeg through experimental acclimatization, translation of Dutch texts and spying missions to the Banda islands.Footnote ³⁸ A similar phenomenon occurred in the early nineteenth century, when the British East India Company (EIC) began to set up tea plantations in Assam. The EIC initially aspired to recruit Chinese tea cultivators, whom they believed to have leaf-processing skills, botanical expertise and a racial disposition toward industry that neither Europeans nor the region's ‘primitive’ natives possessed. Like the French in Mauritius, the British were disappointed that the Chinese workers they recruited from South East Asian trade ports did not live up to these racialized expectations, and soon turned to migrant labour from central India to harvest Assamese tea.Footnote ³⁹

Historians of botany in colonial Latin America and the Caribbean have highlighted European ‘bioprospecting’, the search for new plants with medicinal and agricultural value. Such bioprospecting similarly hinged on the racialization of knowledge, with European physicians and naturalists seeking to extract everything that indigenous or African diasporic communities knew about the local plant life.Footnote ⁴⁰ These communities were not always willing to share their knowledge of herbal remedies, which were often administered within the rituals of healing systems that Europeans openly ridiculed and rejected. European physicians therefore coerced indigenous and African-descent healers to give up their botanical secrets through means ranging from cash bribes and spying on their activities to physical force and threats to their lives.Footnote ⁴¹ In early colonial Mexico, Iberian physicians valued direct observation of indigenous informants in the preparation of medicinal and food plants.Footnote ⁴² This fetishization of ‘untainted’ indigenous knowledge only increased in the following centuries as indigenous peoples themselves were ‘lost’ to the ravages of disease or amalgamated into Mexico's culture of mestizaje. By the late eighteenth century, colonial botanists of the Royal Botanical Expedition to New Spain lamented their difficulty in finding indigenous people ‘uncorrupted enough to still have the old wisdom’.Footnote ⁴³ Lance Thurner shows that the expedition's objectives and epistemology centred on ‘careful attention to race’, their task being ‘to discern truths and medical discoveries’ achievable only through identifying racialized indigeneity, through which they would access ‘secrets of nature’.Footnote ⁴⁴

The process of racializing Mexico's knowledge resources was not limited to humans, as Helen Anne Curry shows through her analysis of twentieth-century botanical taxonomic research into Mexican ‘races of corn’. The term ‘race’, of course, can be applied to plants and animals just as well as to humans, and historically this was a common usage (recall the subtitle of Darwin's On the Origin of Species by Means of Natural Selection, or the Preservation of Favoured Races in the Struggle for Life, which refers more often to races of cabbages, dogs and pigeons than to humans). Yet even as human race science was allegedly declining in the mid-twentieth century, botanists documented and classified Mexico's diverse maize varieties through explicit analogies to physical anthropology and the racialization of indigenous peoples in Mexico.Footnote ⁴⁵

More recently, historians have exposed how the development of geological sciences is likewise indebted to the racialized economies of European empires. The forms of environmental and chemical knowledge that underlie the modern science of geology, such as mining and metallurgy, underwent significant shifts following Iberian imperial expansion in the sixteenth century. As Allison Bigelow writes, ‘the collision of African, Indigenous, and Spanish knowledge in the Americas’ enabled the conceptual emergence of a technical expert with ‘skilled knowledge of mineral extraction’ distinct from general manual labour or artisanal techniques.Footnote ⁴⁶ This skilled knowledge incorporated notions of human difference into its technical language and practices. Not only did Spanish colonial mining and metallurgy operate through highly racialized divisions of labour, just like agricultural plantations; Spanish colonial scholars also used racialized language, including colour-based terminology and categorical terms like casta and mulato, to classify different grades of metals and ores.

Robyn d'Avignon traces similar patterns through the history of West African gold mining and French colonial ‘exploration geology’. West African peoples possessed a dynamic tradition of geological knowledge and mining practices that often outperformed European colonial excavation technologies. In the early decades of the twentieth century, d'Avignon writes, ‘French prospectors benefited from the labor and mineral expertise of generations of African miners, [who, in turn,] acquired new techniques from expatriate geologists’, creating an ‘economy of subterranean knowledge that blurred any facile distinction between African and European mining practice’.Footnote ⁴⁷ Yet for decades, French racism denied the importance of this ‘fluid’ knowledge exchange, casting African gold prospecting and mining techniques as primitive and static, and legally restricting African miners to work only in designated areas.

In between these historical examples, geology's systematization as a distinct scientific discipline in the nineteenth century was firmly intertwined with racial ideas. Studying the physical properties of the Earth meant constant confrontations with evidence of past worlds and lives in the form of sedimentary deposits and the fossils they contained. By the 1830s, British scientist Charles Lyell offered an expansive definition of geology as the study of changes in both ‘organic and inorganic kingdoms of nature’, requiring knowledge not only of mineralogy and chemistry but also of zoology, botany and anatomy. Lyell drew direct analogies between the proper methodologies of geologists to understand the past changes leading to the Earth's present state, and of historians to understand the ‘present condition of nations’ on the basis of ‘various peculiarities of national character’ and ‘different degrees of moral and intellectual refinement’.Footnote ⁴⁸

Indeed, the linkage of geology to human history went beyond mere analogy. Questions about the age of the Earth and the speed and timing of geological events were addressed through the accumulation of human and primate fossils and the archaeological remains of past human activity. For example, archaeologists working in Egypt began applying the geological method of stratigraphy to their excavation practices, while geologists used their new relative dating of human habitation in Egypt to dispute biblical chronologies of the Earth.Footnote ⁴⁹ The framework of ‘deep time’ that coalesced during the nineteenth century, based on intensive analysis of geological formations, fossils and prehistoric human remains was not simply contemporaneous with European colonization of other parts of the world; rather, this development was a direct result of colonization and its associated processes of racialization.Footnote ⁵⁰

As Pratik Chakrabarti details for colonial India, aboriginal and indigenous populations in the present were seen as relics of the prehistoric past. The geological history of the Earth was approached and analysed in terms of human racial evolution, as in the naming of the prehistoric supercontinent of Gondwanaland after the ‘primitive’ Gond tribe, designated by British ethnologists as a branch of the aboriginal ‘Dravidian’ peoples who inhabited the Indian subcontinent before its hypothesized invasion by Aryans.Footnote ⁵¹ Finally, the very nature of geological fieldwork enabled not only European colonial scientists, but also anti-imperialists outside Europe, to make claims about present-day racial and civilizational superiority. Sigrid Schmalzer and Grace Yen Shen show that geological expeditions in twentieth-century China had racial implications both in terms of the potential discovery of human fossils or other evidence of prehistoric human activity, and in terms of the physical capabilities of Chinese geologists themselves to document and analyse the national landscape.Footnote ⁵²

Racialization and its afterlives in data sciences and technology

How did race as an epistemic object of study, as well as a scientific category of analysis, enable the production and practice of emerging scientific disciplines? In this section, inspired by the recent work of Iris Clever, we examine the methodological and technological underpinnings of the new fields called data science and ‘artificial intelligence’ or machine learning.Footnote ⁵³ These fields rely on a foundation of statistics, as well as technologies to generate and process data, such as the spirometer, the camera and the electronic computer. Both statistics and these associated technologies have enjoyed a public perception of providing ‘objective’ and trustworthy information about the natural world, overcoming the inevitable human bias of individual researchers.Footnote ⁵⁴ But in fact race played a significant role in the disciplinary emergence of statistics and the evolving notion of quantitative objectivity as a key scientific value. For example, Suman Seth highlights the medical statistics recorded for European military forces deployed in overseas colonies as the site of a numerical reification of racial difference.Footnote ⁵⁵ Alain Desrosières, Theodore Porter and Ian Hacking have explored the linked rise of statistical thinking and race science within nineteenth-century Europe.Footnote ⁵⁶ The entangled emergence of race science and statistics was not incidental. The availability of ever-increasing volumes of racial anthropometric data and the demand of race scientists for more nuanced and precise techniques to sort racial categories facilitated the development of new data management practices and statistical formulae to determine similarity and difference between groups in quantitative terms.Footnote ⁵⁷

These statistical developments were often intertwined not only with the interests of colonial officials, but also with those of anti-colonial nationalist scientists.Footnote ⁵⁸ In 1922, Nelson Annandale, the Scottish director of the Zoological Survey of India (ZSI), collected anthropometric data (such as nasal index, head length and stature) from two hundred Anglo-Indians. In 1922, he asked Indian statistician Prasanta Chandra Mahalanobis to statistically analyse the data since they represented ‘a true biologically mixed population’.Footnote ⁵⁹ Mahalanobis used Karl Pearson's method of calculating the racial proximity of two groups in analysing the Anglo-Indian data set. In doing so, however, Mahalanobis encountered the limitations of Pearson's coefficient of racial likeness (CRL). While the CRL could determine whether or not two groups were racially connected, it failed to determine the magnitude of racial proximity. The CRL was therefore inadequate for Mahalanobis and Indian anthropologists in this period, who were focused on finding out the extent to which each caste, tribe and religious group in the subcontinent had racially mixed with others.Footnote ⁶⁰ Following his work with Annandale, Mahalanobis tasked himself with developing a statistical formula that could measure racial intermixture. He devised the coefficient of generalized distance, later popularized as the ‘Mahalanobis distance’ by English statistician Ronald A. Fisher.Footnote ⁶¹

The Mahalanobis distance is a calculation of the distance between a data point within one population group and the mean value of another population group's data set. As one of India's major original contributions to international statistics, the Mahalanobis distance formula has since been applied to many other statistical contexts, such as population growth, agricultural output and financial calculations.Footnote ⁶² Recently it has been incorporated into machine learning and facial-recognition technologies, discussed below.Footnote ⁶³ Yet institutional and biographical histories of Indian statistics tend to overlook or deliberately ignore the relationship between statistical innovation and race science. The statistical tools that Mahalanobis developed were shaped by the specific problem he aimed to solve, namely the extent to which Anglo-Indians had racially mixed with other Indian communities. Racial questions and data sets often served as conditions for the fact that many of these methods are now applied for other purposes does not make their origins in race science irrelevant. Rather, it reveals the extent to which our contemporary forms of allegedly objective quantitative data analysis were made possible through investigations of racial difference.

Projects of statistical race making involved not only abstract formulae, but also concrete technologies of measurement. For instance, Lundy Braun shows that the spirometer, a mechanical device used to measure lung capacity, served as both a conduit and a consequence of nineteenth-century racial ideas. Lung capacity was considered an important measure of health and strength, and the spirometer enabled the large-scale accumulation of data for the calculation of statistical averages for different social groups, which in turn were used to determine group members’ life insurance assessments and the suitability for particular occupations. Following its application to comparing the lung capacity of white and black farmers, spirometry data came to be seen as objective evidence for the hereditary nature of racial differences, and therefore a legitimate basis for racialized divisions of labour. By reinforcing expectations of different physiological norms of lung capacity for different races, the spirometer led to the introduction of statistical mechanisms of ‘race correction’ in the collection of medical data. Modified spirometry devices are still in use today, as is their associated statistical legacy of race correction, which advocates now try to corroborate with genetic markers.Footnote ⁶⁴

As a final case, we examine photographic technologies, which historians of science and medicine have analysed as key to the concept of mechanical objectivity in scientific imagery. Mechanical cameras rely upon physical and chemical principles to capture light and create an endlessly reproducible image. For nineteenth-century scientists, these features seemed to make the camera able to faithfully document true appearances and therefore exceptionally robust to human errors of perception. Scientific researchers, from physicists to microbiologists, quickly incorporated cameras into their laboratory equipment.Footnote ⁶⁵ Yet the allegedly objective camera likewise soon became ‘a major medium in the scientific study of race’ through the creation of anthropometric photography, which aimed to provide unbiased and even standardized evidence of physical racial differences.Footnote ⁶⁶ Meanwhile, art historians have analysed how the physical and chemical processes of photography engaged British and American cultural anxieties about racial difference, particularly skin colour. In the nineteenth century, the chemical properties of silver nitrate, which darkens upon light exposure, led British writers to coin a literary trope that cameras temporarily transformed their users from racially white to ‘Negroes’ through the form of the photographic negative (and sometimes through staining the skin of amateur enthusiasts who were careless with the substance).Footnote ⁶⁷

Following the mid-twentieth-century rise of colour film photography, camera film manufacturers like Kodak were confronted with the fact that their product was not capable of realistically depicting white and dark-skinned people in the same frame. The film's chemical emulsion was designed to capture the high reflectivity of white complexions, and rendered nearly invisible any others in the frame with darker skin tones. Kodak attempted to solve this problem by issuing ‘multiracial’ colour-calibration reference cards (called ‘Shirley cards’).Footnote ⁶⁸ In another instance, during the 1970s, the Polaroid Corporation was accused of serving the interests of South Africa's apartheid regime through sale of its I.D.-2 camera system, which produced instant photographs for identification cards, and included an enhanced flash allegedly designed to better capture dark skin tones. This technology was used to enforce the notorious passbook system to control and surveil the movements of black South Africans. In response, a number of Polaroid employees formed the Polaroid Workers Revolutionary Movement, which campaigned for boycotting the company.Footnote ⁶⁹

Such instances unfortunately cannot be relegated to a racist past in contrast to a progressive present. The technological transition from film to digital cameras, relying on electronic sensors and computer visualization beginning in the 1990s, did not eliminate racial bias in the camera itself nor in its social uses. Digital photography paved the way for artificial-intelligence (machine-learning) programs for facial recognition; the higher failure rates of facial recognition technologies for dark-skinned people, including Africans, are now well documented in the literature of science and technology studies.Footnote ⁷⁰ In 2009, a Taiwanese American blogger noted how every time she attempted to take a photograph of herself or her family with her newly purchased Nikon S630 digital camera, the camera's user interface would ask if someone in the frame had blinked.Footnote ⁷¹ This feature was programmed into the camera's facial-recognition system to ensure that subjects had their eyes fully open before taking a photograph. Such features, as James Bridle notes, clearly ‘failed to account for the different physiognomy of non-Caucasians’.Footnote ⁷² Thus racial biases are embedded into technology through dominant notions of what constitutes the normal, which inevitably reflect ‘the systemic inequalities still present within today's technological workforce, where those developing and testing the systems are still predominantly white’.Footnote ⁷³ In other words, depending on where the technology is coded and who does the coding, what is normal becomes racially inflected.

Contemporary governments’ increasing reliance on facial recognition as a biometric technology to control and surveil people and borders does not reduce the long-standing impact of racial (and other) biases in policing. For example, in India's north-east, the 1958 Armed Forces Special Powers Act (AFSPA) enables the state to detain and arrest people who ‘look like’ a militant. Discussing racial profiling in this context, scholars note that machine learning in facial recognition ‘has the potential to extend and embolden these systems of occupation. Routine arrest and detention add more facial profiles to the database, making matching easier and justifiable in ongoing surveillance’, and creating a self-reinforcing loop of suspicion and matching individuals.Footnote ⁷⁴ These ongoing situations demonstrate that modern technology does not eliminate racial biases, despite the promise of biometric technologies to capture personalized individual identities rather than group categories. Even if the output of machine learning is progressively more independent of human intervention, it remains inherently dependent on historical data imbued with racialized biases about how ‘normal’ humans look, speak and act. Technologies are produced within contexts of power imbalance and social hierarchies, and consequently tend to reinforce these biases rather than neutralize them.