Abstract
Images are at the heart of strategies of persuasion. They render certain aspects visible and leave others unrepresented; and they may shape processes of scientific reasoning and imagination. By tracing diagrammatic images in the anthropological sciences throughout the 20th century, the contributions to this special issue highlight some dominant pictorial traditions for rendering human evolution and diversity visible. This article aims to provide an overview of and an introduction to the special issue ‘Visibility Matters’.
Images are no mere ornaments of texts, nor do they simply illustrate what is already textually explained. As many contributions to the iconic turn have emphasized, images are at the heart of strategies of persuasion. They render certain aspects visible and leave others unrepresented; and they may shape processes of scientific reasoning and imagination (Heßler, Adelmann, Frercks et al., 2009; Lynch and Woolgar, 1990, 2014; Pauwels, 2006; Scholz and Griem, 2010). Along these lines, it has been shown for visualizations of human evolution and diversity that they constitute arguments and shape the anthropological sciences. They inform and are informed by popular perceptions of self and other (e.g. Moser and Gamble, 1997; Redknap, 2002). Scientists as well as science studies scholars have discussed how scientific accounts of human evolution inevitably implicate arguments and judgements about contemporary human variation (e.g. Haraway, 1990; Marks, 2006; Proctor, 2003; Wolpoff and Caspari, 1997).
Visualizations of human evolution and diversity have received considerable attention with regard to dioramas and photographs. However, a different kind of imagery has largely escaped systematic analysis: the diagrammatic – the charts that collocate data and the drawings that visualize processes, functions and relationships. This is all the more astonishing in the light of the 19th-century turn towards rigorous measurement and visual presentation of data in tables and graphs (Tufte, 1990, 1997) that coincided with the turn towards evolutionary accounts of human history (Gould, 1996[1981]). Diagrammatic representation typically employs a combination of textual, numerical, symbolic and pictorial elements (Goody, 1977; Tufte, 1997). It relates, clusters, serializes, spatializes and temporalizes data. It allows the synoptic presentation of singular observations and measurements that bring to light otherwise unseen patterns, processes, functions and relationships.
In this special issue, we engage with aspects of the history and present of such diagrammatic tools of knowledge generation and communication in the anthropological sciences including the younger and very powerful technologies of visualization in molecular anthropology. We look at the practices behind the diagrams: the processes and techniques that transform data into tables, graphs, maps and trees. How exactly are diagrams made, from the collecting of bodily parts or fluids, to publication or presentation? What is their epistemic function? How prominent are they in particular publication genres? How do they combine with each other, for example, in a film?
Imag(in)ing the deep past of human diversity
Portraits and Life Scenes
From the second half of the 19th century onwards, popular writers and scientists alike made use of visualizations to purport their views of human prehistory. To our knowledge, the first picture that was loosely based on palaeoanthropological evidence and expressed the idea of a human prehistory was Pierre Boitard’s ‘L’homme fossile’ [‘Fossil Man’] in the Magazin universel of 1836 (Secord, 2001: 312). The visual reconstruction of human ancestors was linked to politics from the start, in Boitard’s case to the aims of radical socialism. Furthermore, Boitard’s fossil man had obviously been inspired by ape and ‘Negro’ imagery as well as by Dumont d’Urville’s descriptions of ‘primitive tribes’. Artists such as Emile Bayard who produced the series of prehistoric scenes for Louis Figuier’s L’homme primitif in 1870 also drew on visual art motifs such as the Graeco-Roman origin myths, the medieval Wild Man, Adam and Eve in Paradise, or ancient Celtic warriors (Moser, 1998). Since the La Chapelle-aux-Saints Neanderthal images based on Marcellin Boule’s and Arthur Keith’s competing reconstructions were widely distributed via L’Illustration and The Illustrated London News in 1909 and 1911 respectively (Moser, 1992; Sommer, 2006), the number of such images has risen exponentially.
Anthropologists and archaeologists themselves, as well as historians and sociologists of science, have analysed the alluring depictions of prehistoric humans and especially of life scenes. They have in particular highlighted the ways in which these are – frequently blatantly – racialized and gendered. Studies have focused on strategies of persuasion and the establishment of genres as well as the manifold cultural preconceptions that are worked into images of human evolution. The figures or the scenes represented may, among other things, be plausible because they represent current social realities and stereotypes. As visual repertoire for the depiction of apes, ‘Negroes’ and Neanderthals, characteristics such as hairiness, dark skin, strong muscles, a protruding jaw, long arms, well-developed brow ridges, large canine teeth, etc., function as markers of the primitive. Conversely, gracility, lightness of skin, prominence of the frontal part of the brain, straight hair, etc., signify the biologically and culturally advanced. In terms of behaviour, the latter characteristics are associated with elaborate material culture, with language, art and complex social organization, pointing in the direction of western civilizations (e.g. Moser, 1993b; Sommer, 2007: 247–62).
Additionally, the hunters, toolmakers and artists of the scenes – those standing erect and in action – are predominantly male, while their female counterparts, if present, squat at the margins or kneel in front of a fireplace. Such role ascriptions again seem to draw on schemata from art history – conventionalized representations of people and objects – as in the case of man-the-toolmaker (mythical and everyday blacksmith in fine arts), guy-with-a-rock in game killing (mythological hero in Italian Renaissance and later painting), and drudge-on-a-hide (scullery maid in the background of many 17th- and 18th-century realist bourgeois scenes). While women may appear eroticized and ornamental, the crouching woman in particular evokes associations with the animal, the pornographic, the base and unevolved (Gifford-Gonzalez, 1993; also Berman, 1999; Conkey, 1997; Moser, 1993a; Wiber, 1997). Finally, scholars have investigated how aspects such as gender, race and social norms in general are negotiated between different actors in the process of producing a dioramic prehistoric scene, how the three-dimensional space of the museum might have constituted visual regimes, and how the meanings of imag(in)ings of human deeper history have been transformed in circulation – how far racism and sexism rest in the eyes of the beholder (e.g. Sommer, 2010).
Craniometric diagrams
No less ‘constructed’ than their naturalistic counterparts were the diagrammatic images of skulls that were based on techniques from a craniometric tradition that reached back at least to the 18th century, European polygenists like Anders Adolf Retzius in Sweden, Paul Broca in France, James Hunt in England and Rudolf Virchow in Germany were central for the development of the methods and instruments as well as the institutionalization of physical anthropology in the 19th century. The American polygenist Samuel George Morton produced Crania Americana in 1839 and Crania Aegyptiaca in 1844 that gave testimony to a quantitative approach and a great collection of skulls. They inspired a crania genre that turned skulls into collections of immutable mobiles such as figures arranged in tables, curves of skull indices and volumes, as well as drawings and photographs (e.g. G. A. Koeze’s Crania Ethnica Philippinica in 1901). Into this decidedly visual tradition, fossil human remains were integrated, a process exemplified by such hallmarks as Reliquiae aquitanicae in 1875 and Crania ethnica in 1882 that established the fossil races of Neanderthal and Cro-Magnon (Sommer, 2007: 121–3; Sommer, 2015: 45–58).
With the turn of the 20th century the diagrammatic image lost none of its impact. Skulls were transferred onto paper as outlines that showed the points and lines relevant for comparative measurements and they were often arranged in a series. Such images remained an integral part of the establishment of the supposed hierarchy from the white adult male via the female and ‘primitive races’ down to the prehistoric races and apes. Superimpositions of the outlines of skulls to compare prehistoric races to apes, recent ‘primitive races’ and/or recent ‘civilized races’ as previously applied by Thomas Henry Huxley (Man’s Place in Nature in 1863) and Charles Lyell (The Antiquity of Man in 1863) in the study of the Fuhlrott Neanderthal, as well as comparative alignment of jaws and other parts of the skeleton, were widely employed (e.g. Marcellin Boule, Les hommes fossiles in 1921; Arthur Keith, Ancient Types of Man in 1911; Hermann Klaatsch, ‘Entstehung und Entwicklung des Menschengeschlechts’ [‘On the Origin and Development of Humankind’] in 1902). Sometimes the skull of a supposed genius was included to render the hierarchy even more explicit.
The British geologist-turned-anthropologist, William Sollas, too, was a great aficionado of measuring and drawing. In his papers and the influential book Ancient Hunters and Their Modern Representatives in 1911, 1915 and 1924, he among many other images published diagrams with curves that represented the skull-size range of living and extinct hominids, their overlap ‘proving’ his argument that Pithecanthropus (today Homo erectus) was closer to modern humans than apes. Elsewhere, he entered skull sizes as dots in a coordinate system and linked them by a line to ‘show’ the linear increase in brain size in the course of human evolution from a fossil ape, via Pithecanthropus, Neanderthal/Australian Aborigine, to the ‘modern white race’. To give another example: ‘The British Oracle of Anthropology’, Arthur Keith, specialized in arranging the outlines of skulls, jaws and teeth in frames and grids opposite each other to highlight their differences (see his Antiquity of Man in 1915 and 1925). The information these images are intended to convey is usually easily grasped, particularly if the viewer is accustomed to their style. What cannot be seen are the controversies that went on about the right interpretation and reconstruction of the skeletal remains, about the best method of analysis, or indeed about the correct line of alignment of skulls for comparison, and the most reliable points of reference for the measurements (Sommer, 2007: part II).
Trees and maps
The most frequent diagram is likely the phylogenetic tree, because it has found such a wide distribution outside expert circles. It has its roots in medieval genealogical representations of ‘noble descent’ and sacred genealogy (the Tree of Jesse), in the visualization of pedigrees in animal breeding, and in comparative historical philology. In anthropology, it has appeared in many avatars, from Ernst Haeckel’s oak – that conveys evolutionary progress and possibly the idea that ontogeny recapitulates phylogeny 1 – to naked line diagrams (Alter, 1999; Bouquet, 1994; Brace, 1981; Müller-Wille and Rheinberger, 2007). Trees have changed with the dominant view of human evolution in the course of history, and the message varies with the visual metaphor: a bush is less hierarchical than a tree, a chart less organic than a rhizome (Sommer, 2007: 187–96).
Possibly because of its rich cultural history, the visual language of the phylogenetic tree seems straightforward. However, in reality it condenses elaborate theories and scenarios without revealing the decisions on which the particular pattern of branching is based (Tattersall and Eldredge, 1977). Furthermore, already in the early decades of the 20th century, the ways in which phylogenetic diagrams were understood by non-specialists might well have differed from at least the messages consciously intended by their expert producers (Clark, 2001). Even if biologists tried to convey a more complex idea of the evolutionary process, the public was so accustomed to the linear hierarchies from ape to man that evolutionary diagrams tended to be read as indicating a goal-oriented and hierarchical development (Gould, 1995: 42–60).
Tree diagrams have gained renewed power in the second half of the 20th century through the molecularization of research into human evolution (Sommer, 2008). In Sommer’s contribution to this issue, she reconstructs their history from the 1960s when mathematical and computational tools were developed to arrive at human-population trees from various kinds of serological and molecular data. Sommer argues that the visual language of the tree is a diaspora, and shows how the population tree was actually mapped on the surface of the earth. She sees this visual language of the tree doubled in the discourse of unity in diversity that often structures especially popular textual renderings of modern human evolution. This anti-racist liberal discourse has at times been criticized as running counter to the socio-political effects of human population genetics – most notably in the Human Genome Diversity Project. Often based on so-called isolated peoples, and continuing to use some of the earlier population labels, population-genetic tree diagrams are meant to represent a state before the great historical population movements. Focusing on the work of the population geneticist Luigi Luca Cavalli-Sforza, Sommer follows the practices of tree-building and -mapping from the early blood-group studies to the current genome-wide admixture research.
The map is indeed a close ally of the tree when it comes to visualizing how human groups have evolved and diversified on their migratory routes all over the globe. The racial cartography of human origins may be traced to the 6th century (Livingstone, 2010), and the history of thematic maps that are based on numerical data reaches back into the late 19th century (e.g. Friendly, 2012). However, the seroanthropological analysis of blood helped population geneticists to set a new trend in mapping so-called racial differences (Gannett and Griesemer, 2004). In this issue, Susanne Bauer engages with the techniques of mapping human genetic diversity during late Soviet and early post-Soviet times. Since the 1970s, anthropologists and population geneticists at the Institute of General Genetics in Moscow have worked on a project called ‘Genofond and Genogeography of the Population in the Soviet Union’ that aimed at an inventory of all data ever collected on different populations. They took up a Soviet tradition of genogeography from the 1920s to develop statistical mapping techniques in collaboration with computer scientists. The visualizations of the resulting genogeographic atlas published in 2003 render the past present and space dynamic. Bauer shows how molecular markers were integrated with early biological data and inquires after the politics of difference in the process.
Rendering contemporary human diversity visible
A long-standing tradition in visualizing human diversity draws on a biblical topos: Noah’s three sons have often been depicted as representing the oldest ancestors of the three ‘human races’. And indeed, assembling three, four, or five different faces in one picture, and thereby demonstrating that humans are diverse, is one of the oldest and most pervasive strategies of persuasion in this field. One finds this kind of representation even today – one face representing one ‘race’ (or ethnicity) – together with other well-established visualizations, classifications and descriptions of contemporary human diversity in school books and children’s books, in atlases and handbooks, in encyclopedias and all sorts of educational material with a history dating back to the 19th century. 2 To be sure, physical anthropologists often used other, more abstract ways of visualizing diversity, at least in their journal contributions (Hanke, 2007), and one needs to view their attempts to visualize their understanding of divisions of humankind against this backdrop of a widespread, well-established visual culture that drew heavily on faces. One could call this a visual culture that was meant to make the – presumably – already visible even more visible.
In the same non-diagrammatical vein, spectacles like anthropological displays or exhibitions of ‘African villages’ in European zoos (Dreesbach, 2012) and dioramas were also meant to make the already visible even more visible, or to point out visible differences between humans, by putting them in spatial proximity and, thereby, in immediate contrast. Such performances have been scrutinized, and their production, visual codes and mythologies situated in the context of colonial science and public discourse (e.g. Arnoldi, 1999; Bal, 1992; Hale, 2008; Haraway, 1984–5; Qureshi, 2011). Recent studies have also highlighted the specific role of photography in anthropological and ethnographical research (Edwards, 1992; Edwards and Morton, 2009; Hight and Sampson, 2004) and tackled the issue of visibility and invisibility in relation to racial thinking (Poole, 1997; Stoler, 1997). Yet before colour photography allowed for genres such as comparative collections of portraits, visual comparison of human faces or bodies heavily rested on drawings, sometimes coloured, and photographs in black and white; and accordingly, on features such as shape of the eye or eyelid, hair texture, face shape, etc. (e.g. Nesturch, 1959: 24, table I). In combination with categories or textual descriptions of complexion, such representations aimed at evoking visual types of humans (cf. Groebner, 2003).
However, for scientific inquiries, these typological strategies hardly sufficed when new quantitative methodologies and new visual ways of documenting were introduced from the late 19th century onwards (Hoßfeld, 2005; Laukötter, 2007). The materiality of the very objects of study contributed to these new forms: whereas bones and skulls made up specific objects that could be revisited, transported, stored and traded, studying living humans brought new challenges, but also opportunities. Large numbers of individuals could be recruited under certain circumstances, such as schools or military barracks (Hartmann, 2011) but they could hardly be revisited. This made the challenge of mass data processing and interpretation ever more difficult. Data visualization was one promising way to deal with this challenge.
Moreover, it was particularly the issue of complexion that fascinated life scientists in the 19th century, as a visual clue to human variation and ancestry. While body height and other measurements were acknowledged to be difficult to assess and compare, due to differences in age, gender, or living conditions, complexion seemed to promise to be a relatively gender and age neutral trait, and one that particularly promised to yield visual clues. Objectifying and standardizing the visual impression of colour became one of the preoccupations of those who worried about anthropometric methodologies. 3
Beyond these concerns about visible traits, scientists became increasingly interested in the quantitative analysis of differences that were not so easy to grasp visually. The study of diversity of living humans considered an ever-growing number of potentially relevant, comparable and measurable traits, and produced an endless stream of numerical data from physiology, biochemistry, serology, pathology, genetics and other disciplines. 4 Making the differences in distribution between human populations visible was a challenge that scientists tried to meet with novel visualizing techniques. Yet the development of statistical and cartographic tools for the analysis and visualization of human diversity not only went hand in hand with new ways of analysing genealogical relations (Parnes, 2007); it was also accompanied by intense debates about the adequacy of these tools (Porter, 1995). There was by no means consensus on which of those traits were in fact measurable, quantifiable, comparable and meaningful enough to yield important insights into human variation (Massin, 1996). Neither was it univocally agreed upon as to how the data could be collated, handled and interpreted. And finally, as the field of theory of graphical representations grew in the early 20th century, there were plenty of possibilities as to how the findings could be best presented (Auerbach, 1913; Tufte, 1997, 2001).
From these possibilities, scientists and science communicators chose different strategies for visualizing their claims about human difference. Some of the most relevant strategies are discussed in the case studies assembled here, along with their impact and issues regarding their wider distribution: as discussed above, maps and phylogenetic trees, the focus of Marianne Sommer’s and Susanne Bauer’s contributions, were among those images that had for a long time circulated with great ease, and continued to be the most preferred visualization in the second half of the 20th century; however, now with a heavy leaning towards diagrammatic forms. Abstract visualizations such as graphs, as in the case study of Amir Teicher, did not circulate very easily. This finding is confirmed in Veronika Lipphardt’s article that contextualizes abstract diagrammatic visualizations in the visual repertoire of the field in the first half of the 20th century. Jenny Bangham’s article examines some of the strategic and aesthetic decisions made in the graphical communication of genetics in anti-racist campaigns after the Second World War.
Graphs
Most abstract visualizations aimed at displaying correlations between data. Novel kinds of diagrams, tables and graphs were continuously proposed and discussed by scientists as useful tools of visualization. Statistical tables, for example, became a preferred visual tool for comparing, for example, diverse demographic data derived from the study of ethnic groups in order to highlight the dynamic relationships between these groups, or between the members and sub-groups of one particular group. The visual ordering of data in tables deserves particular attention: it conveys statements about what is the norm, what is the golden mean, or the most progressive variant. And yet, the informational complexity a table can convey is limited: tables are often kept quite simple in order to allow a quick take-up of information by the reader (Hanke, 2007; Lipphardt, 2009: 160, 176–80).
In contrast, statistical graphs were meant to allow for the display of the most complex data relations. Amir Teicher shows how identifying clusters and group-differences among a collection of numerical data is one of the veteran problems of statistical analysis. In 1907, the physical anthropologist Theodor Mollison came up with a novel ‘deviation-curve’, a compelling visual tool that allowed curves to be generated from average measurements on populations which were plotted on a single coordinate system and compared to finally point to phylogenetic similarities between human races. In the early 20th century Mollison’s method proliferated in German-speaking anthropology and was used in journal publications to bring order into otherwise opaque numerical results. The article shows that the method of visualization inherent to Mollison’s method was highly problematic, and that indeed many of the scholars who used this tool, although partly aware of its drawbacks, found themselves falling into the pitfalls of its misleading visual dimension. To account for the rise – and fall – of the ‘deviation curve’ in the anthropological scientific community, the ‘visual culture’ of racial anthropologists is examined. What made anthropologists cling to this problematic tool, so it is suggested, was that it enabled the graphic reaffirmation of the alleged existence of human races at a time when their existence was put under severe scholarly doubt.
Circulating images and visual communication
Certain forms of diagrammatic visualizations enjoyed more popularity than others in the visualization of human variation, and travelled more readily than others from domain to domain. Veronika Lipphardt’s article gives an overview of the visual culture shared by a number of scientists focusing on human variation – or rather ‘race’ in actors’ terms – in the first half of the 20th century. Clearly, drawings – and later photographs – of people from all over the world constitute a crucial part of this visual culture. Lipphardt’s article focuses on those parts of the visual repertoire that were meant to visualize aspects of human variation that were not so obvious to the eye of the observer. Starting from 1900, the article demonstrates how scientists built up a rich visual repertoire to help in understanding human diversity, one that integrated maps, tables, photos, drawings, diagrams, abstract and non-abstract elements, but increasingly rested on diagrammatic visualizations. Notably, although these diagrammatic visualizations gained in sophistication and methodological rigour, they were rarely used in textbooks. More often, mixed strategies of visualizations – diagrammatic elements integrated into non-diagrammatic images – were chosen as representations of human diversity.
As Sommer’s and Bauer’s articles suggest, due to their ubiquity and origins in wider visual culture, the tree and the map lend themselves to analyses of the circulation of knowledge. But it is in Jenny Bangham’s contribution that issues of communication take centre stage. She engages with the images published by UNESCO in 1952, in the picture book What Is Race? Evidence from Scientists (1952), which was intended as a popular exposition of UNESCO’s Second Statement on Race, made that year. Ostensibly written for an audience of schoolchildren, What Is Race? sought to undermine racial prejudices through the visual depiction of ‘scientific facts’ about genetics. Bangham examines the design and visual language of the picture book in relation to the political agendas of UNESCO’s model of science communication. Affirming UNESCO’s belief in the universal legibility of the images, they were redeployed in other media, including a film strip for use in schools, in Life magazine, and on the BBC television programme Race and Colour. Bangham also looks more closely at the visual and textual arguments about genetics, especially the genetics of blood groups, which, as racially meaningful but non-visible human traits, were strategically deployed to deflect attention away from physical differences such as skin and hair colour.
If one aims to compare or view together these various developments, perhaps one striking aspect is that rather than depicting humans, or body parts, a major trend in this history of visualization was the turn to rendering relations visible. More abstract visualizations however, like graphs and diagrams, may seem stunningly sophisticated, yet their strategy of persuasion seems to have failed. Finally, what about critique, doubts and resistances to visualizing human difference in the 20th century? As the articles of Lipphardt, Sommer and Teicher suggest, a number of scientists paid critical attention to visualizations and their potential to mislead understandings of human diversity. And there have been controversies about the right way of visualizing different understandings of it. Again, diagrams turn out to be technologies of persuasion not only of ‘lay publics’ but also of members of the scientific community who hold contrary views of human evolution.
