Portal de Periódicos da CAPES

Theorizing with incorrect data: A new look at the historical inaccuracies of the bioarchaeology of corsets

2023; Wiley; Volume: 125; Issue: 3 Linguagem: Inglês

10.1111/aman.13882

1548-1433

Rebecca Gibson,

Tattoo and Body Piercing Complications

This essay strives to correct the bioarchaeological record of skeletal changes due to corseting, a currently understudied but impactful subject in biological anthropology. Sir Arthur Conan Doyle ([1892] 1994, 7) tells us, via his famous fictional detective Sherlock Holmes, that “It is a capital mistake to theorize before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts.” Similarly, it is as much a mistake to theorize using singular instances of data or data that can be shown to be incorrect after a more than cursory examination. Such has been the case in various anthropological and bioarchaeological events, including Piltdown Man and Doyle's own belief in fairies—both of which I will discuss below—and more recently, conclusions made about corseting via historical radiographic documentation. Using two radiographs from the most commonly discussed source on the medical history of the corset, Le Corset by Ludovic O'Followell (1908),1 I demonstrate that prior to my publication on “Effects of Long Term Corseting on the Female Skeleton: A Preliminary Morphological Study” in NEXUS (Gibson, 2015), the understanding of corset-related change relied on faulty data from O'Followell's book. Given that data from O'Followell were the only radiographic-image-based documentation of the phenomenon of corset-related skeletal changes from O'Followell's time, it constituted the entire factual dataset from which one could draw conclusions about women who corseted prior to it. While other scholars of corseting in the late 1800s and early 1900s did provide images (see the iconic drawing in William Alcott's [1837, 238] The House I Live In as well as similar drawings from the pamphlet The Corset: Questions of Pressure and Displacement by Dickinson [1887]), the physical manifestations of corset alteration are considerably less exaggerated than the drawings of the day; after all, it is quite easy to draw something and call it accurate, particularly if your audience is ignorant, trusting, or credulous, while in fact it is not accurate at all (see discussion below of hoaxes). I would like to note, however, that my subsequent publication of The Corseted Skeleton (Gibson 2020) critiqued O'Followell's text but did not formally recognize his flaws, as discussed here. In 2015, and later in 2020, I adopted a critical view of the historical evidence of corset-related skeletal changes, conducting my own research on skeletal populations from St. Brides Parish, London, and the Musée de l'Homme, Paris—the aforementioned data required in order to successfully develop a theory. The skeletal changes found in both populations—on women whom I could verify would have reasonably practiced long-term corseting—seemed consistent, at first glance, with drawings and radiographs from O'Followell's book. However, I have subsequently undertaken a translation and annotation of Le Corset (forthcoming from Bloomsbury's fashion history imprint under the title The Bad Corset: A Feminist Reimagining), and in the course of said work, discovered that many of the radiographs used by O'Followell were manipulated to fool the viewer into a mistaken interpretation of the image. Below is an examination of two of these radiographs—which detail the evidence of falsification—to correct the interpretation of the image-based bioarchaeological record. Before I discuss them, however, a brief diversion into O'Followell's influence on the subject of corseting. Many prominent authors between 1700 and 1900 wrote on the topic of corseting, from Jean-Jacques Rousseau's (1762) condemnation of the practice in his fictional works to William Berry Lord's ([1868] 1870) masterful examination of stays and corsetry through history, to the aforementioned Dickinson (1887), but they did not have two things that O'Followell did: radiographs and extensive medical knowledge. As shown in my forthcoming translation and annotation of Le Corset, O'Followell's medical knowledge does not stand the test of time, nor was much of it accepted as accurate during his own time, and yet combined with the presence of radiographic imagery, he exuded authority and for decades has been treated as such. Part of this persistence is luck—O'Followell's work was there to be quoted. There may have been other documents laudatory of the corset that may subsequently have been lost to time; we will never know. Mentioned by such early pioneers of corset historical research as David Kunzle (1977, 1982), who began work on the subject in the 1970s, and Valerie Steele (1997, 1998, 2003), whose work on the fashion of eroticism goes back to the late 1990s (who did not necessarily champion his work but who were also not able to assess it from a bioarchaeology standpoint), O'Followell's Le Corset contains many assertions that seem obvious, despite the fact that they reinforce our view of the time as misogynistic. First among these is the assertion that women were oppressed by their corsets—that they corseted due to the patriarchal expectations of the day, and that such corseting made them miserable. This assertion seems reasonable. The literature of the time bears it out, showing misogyny as a staple of the culture of the late 1800s and early 1900s in France. Many famous texts that are often considered foundational, such as Freud's ([1893] 1952) work on hysteria, affirm misogyny as an inherent principle in the medical practice of the day and instantiate it into popular understandings of European and American cultures. And so when we look at O'Followell's body of work, and we use that popular understanding to interpret it, we come to a seemingly inevitable conclusion: the corset must be a violent misogynistic tool of patriarchal oppression. But in employing that interpretation, we replicate the mistake of theorizing without data. Much of what can be found in the literature is composed of anecdotes or case studies, neither of which constitute solid, repeatable, generalizable, numeric data. What, then, does the data tell us? The above figure, the first clinical image in Le Corset, underpins much of O'Followell's arguments (Figure 1). Throughout his book, O'Followell argues that the medical “harms” done to women who corseted were self-inflicted, either by abuse of the corset, or by wearing a bad corset (badly made, ill-fitted, unsupportive, laced “too” tight). O'Followell's major claim revolved around the corset's alteration of the ribs (and thus the compression of all organs and organ systems beneath those ribs). Radiography, using X-rays to create images, had been invented just over a decade earlier, and the use of the technology for medicinal purposes was relatively new, carrying with it both the weight of a physician's authority and the ability to visualize internal structures without surgery, leading to an amazed and enthralled public. In the case of this image, published in Le Corset in 1908, there are several pieces of evidence that, upon detailed and careful examination, indicate that this radiograph was manipulated to show unrealistic representations of the human skeleton. This radiograph shows a “corseted” woman wearing a corset with a metal-lined top edge and metal-wire laces. At first glance, the skeleton is easily visualized, and the metal makes the corset stand out on the radiograph. Yet, for a researcher familiar with how corsets are worn, many things represented in this image can be identified as falsifications, four of them which will be addressed here. I have both professional and personal knowledge of corset lacing, being both a scholar of corseting (Gibson, 2015, 2020) and a wearer of them since 2003, and therefore noticed that the laces in the radiograph (Figure 2) are too slack at the top of the corset. Slack or disordered laces do not provide the proper support and create a messy silhouette. While many people do prefer slackness in their laces, there would be no reason to leave them loose if they had been laced and tightened by someone other than the person wearing the corset. A tight-laced corset with metal-wire lacing and metal grommets that, according to O'Followell, a doctor had laced would not display any slack in the laces. This is commonly seen if the subject of the radiograph put the corset on and laced it themselves, as it is difficult to get enough leverage to pull the posterior top laces tightly, but an individual who had been helped into the corset would show tight and even lacing from top to bottom.2 This does not indicate that the subject of the radiograph was a participant in the lacing—as we shall see below, there was no participant involved—but merely that O'Followell was inconsistent in his descriptions of the processes and methods he followed. O'Followell (1908, 27) states that metal was added to the upper and lower edges of the corset to make the edges more visible (Figure 2). The image shows the outline of the person's body, with the fat/muscle layer clearly visualizable. Not only does the fat/muscle layer show evidence of being edited, so that it creates the silhouette that O'Followell wanted to demonstrate (note the overhanging shadow at the top left of the image, but the metal edge fails to curve to the front, as a true corset would when worn on a body. Additionally, the edge of the corset does not match the characteristics of any corset from this period, as the edges overlap the metal boning of the corset, and in a normal corset of the period, the boning would not go all the way to the top. The corset itself, from the lacing to the edges, is manipulated to show a false version of the corseted human body. Turning to the image itself (Figure 2), while it is the most cleanly focused of all the radiographs included, it is still very muddy and shadowed due to the limitations of technology at the time. However, in personal correspondence with radiography expert Rory Langton at Indiana University South Bend (email exchange with author, August 11–16, 2022), Langton and I were able to confirm that while there could be internal organs, none can be clearly visualized. Furthermore, in a true radiograph from this angle, the organs would not obscure the posterior view of the anterior part of the ribs, as the overall shading does in the thoracic cavity in this image. Such complete opacity is more consistent with the rib cage having been stuffed with something to maintain its shape. Next, one must examine the vertebrae (Figure 2). The twelfth thoracic vertebra and all five lumbar vertebrae are visible on the image, but with careful examination, it becomes clear that they are not in the proper orientation for a posterior view. The vertebrae are all rotated 90 degrees from the vertical and shown from their superior views, not their posterior views. Additionally, each now superiorly viewed vertebra is rotated approximately 45 degrees clockwise, meaning that the left transverse process is just slightly off of horizontal in the resulting image, and the spinous process is angled down and to the left. The vertebral foramen are visible where the intervertebral discs and the superior and inferior articular facets should be, on the far-right edge of the image, as blank holes. Such rotations in vertebrae from their normal positions in a living human would result in instant paralysis by spinal cord compression and most likely death from the resulting pain and shock. Furthermore, Langton states, “the vertebra definitely seem to be misaligned/misplaced/artificial. It also appears that this is either a radiograph of a deceased patient or phantom (radiography term used when we image an artificial body part)” (email exchange with author, August 11–16, 2022). This radiograph has been deliberately manipulated to show incorrect anatomical features, as has the second example seen below. The use of a metal hook-and-eye busk (Figures 3 and 4) demonstrates the backward placement of the corset, as it was always, without exception, the anterior closure of the garment. Its use as the anterior, and only anterior, closure is seen throughout advertisements (O'Followell, 1908, frontmatter) and museum collections (Gibson, 2020; Figures 5 and 6). It was never used as the posterior closure, for one primary reason: it takes strength, dexterity, and finesse to close a busk of this type (none of which are achievable by reaching one's hands around the back of one's body). The image is, according to O'Followell's (1908, 53) caption, “Fig. 31—Radiographie du corset Ligne (devant)” or “Figure 31, radiography of the straight corset (from the rear).” This is simply not accurate. The posterior of a corset would be laced through a series of eyelets, as in the first example above (Figures 1 and 2), or the corset would have no back closure, being just one contiguous piece and closing in the front. Furthermore, full corsets’ posteriors from this period had top edges that formed a straight line across the back of the wearer (see Figure 5), not a V-shape, as is evident in the above image. As for the body itself, the ribs are clearly seen from a posterior view. Langton did not rule out the presence of internal organs but stated that he could not clearly visualize the heart and liver, indicating an ambiguity as to the nature of the subject, alive or dead. From this starting point, the viewer can assume that the image is not to be taken as canonically accurate. Next, the viewer moves down the body to the pelvic girdle (Figures 3 and 4), which is at a 180-degree antero-posterior rotation from the torso. As with the previous image, the posterior of the ribs can be visualized quite clearly. However, one can also clearly see the anterior view of the pelvic girdle, due to the depth of field in the image, with the anterior portion of the iliac crests, the obturator foramina, subpubic angle, and the anterior edge of the acetabula, from the back to the front of the image, respectively. Were this an image taken from the posterior, the viewer would be able to see the sacrum and coccyx, both of which are blocked from the front by the corset's hook-and-eye busk; in fact, it is clear that the lower point of the hook-and-eye busk slightly overlaps the anterior view of the pubic symphysis, which it could only do were it in front of it in the X-ray machine's field of view. This radiograph contains the anterior view of the pelvis and femora and the posterior view of the rib cage—a physical impossibility on a living person. It is a composite image, created from at least two radiographic images. This new information is vital to the bioarchaeological discussion of changes to the skeleton from corseting, due to the reliance by Stone (2012, 2020) on O'Followell and his radiographs to describe the effects of corseting.3 This needs to be noted especially regarding the radiograph shown in Figures 3 and 4, which was until recently the only medical evidence utilized in bioanthropological scholarship of this type of skeletal abnormality and which had been assumed by researchers to be factual without the consultation of skeletal remains from the same era.4 Ironically, as my work shows, corseting does alter the ribs and vertebrae in specific predictable and verifiable ways; O'Followell had no justifiable reason to falsify his radiographs (Figure 7, 8). And in fact, there is radiological evidence of modern skeletal changes due to corseting, that of Cathie Jung, current holder of the title of world's smallest waist (Steele, 2003, 66), yet the first bioarchaeological publication to discuss corseting's effect on the rib cage (Stone, 2012) relies almost entirely on O'Followell's radiograph. It is this alone that is vital to Stone's argument that corseting was patriarchal violence and that it inherently and intrinsically harmed people, an argument which she continues in 2020 (Stone, 2020). It was not until 2015, with my publication in NEXUS, that any current bioarchaeologist examined skeletal material to verify or contest O'Followell's evidence, and not until my 2020 monograph The Corseted Skeleton that such an examination was conducted on a skeletal sample with enough individuals to distinguish rib and spinal variations between male and female skeletons and between skeletal remains from Europe or Europeanized countries and those with minimal histories of colonization. While I did verify alteration due to corseting on European and Europeanized female skeletons, my examination showed no correlation with other diseases or death (Gibson, 2020). In 3,815 death records, culled from the church register of St. Bride's Parish, London, there was not a single record of corsets, corsetry, or corset-related changes to the body as the cause of death.5 This analysis, combined with the demonstrable manipulation of the above radiographs, must lead to changes in the current scholarly discourse surrounding skeletal changes due to corseting. Additionally, this research significantly intersects with other crucial bioarchaeological and bioanthropological areas, most notably those involving metabolic disorders, as the altered shape of the ribcage due to long-term corseting is most often mistaken for, and dismissed as, rickets/osteomalacia (for information on the presentation of rickets, see Brickley and Ives, 2008; Ives, personal correspondence). The presentation of rickets, however, does not show changes in the vertebral spinous processes, which were first shown in my work in NEXUS and in The Corseted Skeleton. Conversely, corseting does not cause pigeon-breastedness, or a medial indentation of the ribs, both of which are characteristic of rickets. This conflation of the two—of rickets being mistaken for corset-related alterations—was first seen in the only pre-2012 anthropological exploration of the topic of corseting, a conference paper given by J. Lawrence Angel and his coauthors Jennifer O Kelley and Gretchen Schneider, all of the Smithsonian Institution. In the only extant evidence of this presentation, an abstract, Angel et al. (1985) write, “stays with straight supports brace the ribs with the pelvis; skeletons (Colonial cemeteries) show side-flattening of ribs 6–8… . Modern back-braces … [show] ‘healed fracture’ waviness of ribs 6–12).”6 While he also describes flattening of various vertebral spines, and the women whose skeletons he examined most certainly corseted, my dataset shows that no waviness or side-flattening occurred if the person did not also have rickets. That it remains so easy to confound the manifestations of the two skeletal alterations is proof that evaluation of corseted bodies requires precise data-gathering methods from skeletal material and suggests the adoption of a skeptical eye toward the work of O'Followell and other nineteenth- and twentieth-century physicians. This same skeptical viewpoint has been used through anthropological history to root out other hoaxes, as well. While it is impossible for any research to be without bias, one goal of anthropology is to describe phenomena, not prescribe them. Practitioners observe, we participate when possible, but we should interpret with a light hand, leaving the culture to define its own parameters. We must approach both historic and contemporary cultures as full contributing members in the process of our scientific endeavors. However, we must also keep in mind the missteps of the past and the repeated crafting of evidence to support a researcher's viewpoint. This is particularly important when approaching conceptualizations of gender, gender roles, and the interplay between agency and cultural coercion that determines the nuanced participation of people in those roles. Much work has been done recently on redefining how anthropologists view the historic presentations of gender and gender roles in the archaeological record, from Gifford-Gonzalez's (1995) incredible examination of our prejudices in her piece “The Drudge-on-the Hide” to Knudson and Stojanowski's (2008) observations on “New Directions in Bioarchaeology: Recent Contributions to the Study of Human Social Identities,” DiGangi and Moore's (2013) thorough textbook Research Methods in Human Skeletal Biology (2013), Redfern's (2017) extensive work on violence patterns in skeletal remains, and Agarwal and Wesp's (2017) comprehensive Exploring Sex and Gender in Bioarchaeology. We cannot root out misogyny from the canonically accepted data if we are reifying it ourselves. One way this manifests is an interpretation of grave good typology that reinforces strict gender-binary differences, despite written historical records to the contrary, as has occurred in various so-called Viking warrior burials, which have been historically classified as uniformly male, despite inconclusive skeletal evidence. A recent reevaluation of one such burial created a slightly less sensational article titled “A Female Viking Warrior Confirmed by Genomics,” (Hedenstierna-Jonson et al., 2017). To this end, one might also look at the doctoral work of Slotten (2020) on combining analysis types into a truly biocultural examination of gender-role expectations, skeletal trauma, and burial good types to reassess our understanding of how to “see” gender in the bioarchaeological record. We must not let our assumptions about gender get in the way of our interpretations. In much of the analysis of corseting as a cultural practice, the assertion that women were controlled by the patriarchy into wearing a harmful, dangerous, deadly garment is taken as fact. However, this assertion does not account for nuance, variation, or the give and take between a person's agency and their feelings of obligation to their culture of origin. Part of this obligation was to the concept of “civilization,” with corsets shaping the body away from its natural state and toward a more cultured appearance (Balch, 1904; Fields, 1999; Gruber, 2016; Ribeiro, [1986] 2003). However, in France in the early 1900s, women had considerable choice in how to wear their corset and how tight to lace it. Whether they liked the practice or abhorred it, beginning mid-century, they were speaking up in droves, writing to magazines rejecting the notion that their natural shape was uncivilized and rejecting the idea that corsets were unhealthy (Anonymous, 1866; Anonymous, 1882; Anonymous, 1883) and doing university-level work on the topic (see Dr. Inès Gaches-Sarraute's [1900] Êtude Physiologique et Pratique and the awe-inspiring oeuvre of Dr. Justyna Budzińska-Tylicka). The corset became, if not villain, then antihero for which some could root, against what most of society and most of the medical community believed and advised. While women in the La Belle Époque France existed in an overarchingly patriarchal society, so, it can be argued, do we in the year 2023—and yet, individual buy-in to participation in that society was almost as varied then as it is now, even if tilting at hegemonic windmills was, and remains, exhausting. However, this is far from the only mistaken interpretation or use of dubious evidence in anthropology's history. In many cases, inaccuracies were passed to researchers, or from researchers to the reader, for various reasons—the culture may have had their own agenda (Lee, 1969), the research may have been predisposed toward a certain viewpoint due to the agenda of the anthropologist (Tierney, 2000), or, as with the above radiographs, the data itself may have been flawed or deliberately fabricated by an author who was part of a culture that is no longer around to consult (O'Followell, 1908; in Stone, 2012, 2020). In all of these cases, the preconceived notions about the research topics failed to be set aside so that the evidence might speak for itself. We see this repeatedly throughout the history of our discipline. To return to Conan Doyle, while he is most remembered for his fictional detective Sherlock Holmes, he was actually instrumental in the development of certain early aspects of forensic anthropology and managed to assist Scotland Yard, and several regional police forces, in the solution of multiple crimes, including clearing the name of an innocent man (Fox, 2018). And yet, he was not immune to falling astray of photo manipulation, false evidence, or believing what he wanted to believe. He, and many others in the United Kingdom, were taken in by “The Cottingley Fairies” (Huntington, 1997; Owen, 2003), photos of two girls with dazed, vacant expressions, gazing in wonderment at fairies. The fairies were cardboard cutouts that were superimposed onto the primary image, an extremely common and easy form of photographic manipulation, which had been in use for many years by their publication in 1902 and is likely at least partially how O'Followell manipulated the images seen above in Figures 1-4. Neither the Cottingley fairies nor O'Followell's radiographic lies were as egregious as anthropology's most notorious hoax, Piltdown Man (for an exploration of how paleoanthropology embraced the hoax, see, among others, Stringer, 2012). But similar to Chagnon's depiction of the Yanomami (for the controversy itself, see Tierney, 2000; for reactions to the controversy, see Coello de la Rosa, 2018; Cronk, 2013; Hill and Hurtado, 2004; for reaction to Chagnon's death, see the following endnote),7 the intent to deceive was present in the presentation of “data” that had been cherry-picked, manipulated, and presented as factual without any support. Motives may differ, yet the reliance on small sample sizes, case studies, or unconfirmed anecdotal “evidence” leads to inaccuracies in our interpretations. Again, in all of these cases, careful examination of the available evidence, or the eventual production of counterfactual claims, produced more-accurate results. In the case of corseting damage, it is necessary to work directly with the skeletal material (Gravlee, Bernard, and Leonard, 2003). In examining the skeletal remains of hundreds of women from multiple locations and socioeconomic backgrounds, it is evident that while corseting changed the skeleton, it did not do so in the way presented by O'Followell. It did not shorten their lives significantly or impair their day-to-day functionality, and having carefully considered all genres of historic writing on the topic for their accuracy and ability to contextualize the skeletal material, we can see that it did not emotionally traumatize them or render them agentless vassals of the patriarchy. And through deeply examining the main source of today's corseting “knowledge,” namely O'Followell, we can develop a clear understanding of the biases he brought to his work. To thoroughly interrogate O'Followell, it is necessary to look beyond the radiographs and other pictures in Le Corset to the text itself. Most previous publications make use of small portions of his text, a few paragraphs at most, and this can be misleading. Much in the same way O'Followell himself excerpted and paraphrased the French surgeon Ambroise Parè (who wrote in the 1500s, in the early modern period), O'Followell himself has been reduced to a few soundbites, a few images, which often stand in for the larger discussion on the history of the corset, or the misogyny within that history. He certainly hates the corset and thinks it dangerous, but a casual reading, or a reading of only a paragraph or two, will miss the clear and repeated agenda: his disdain for female autonomy, his desire to control female fertility, and his belief that women who corseted “wrong” —at the wrong age, while fertile, to be pretty, etc.—should be punished.8 This agenda, with which he ends all 15 chapters of Le Corset (see The Bad Corset, forthcoming) is likely pivotal in the radiographs’ falsification, laying the brickwork for twentieth- and twenty-first-century anthropologists’ interpretations with misleading “evidence” of the skeletal damage of corseting and the impact on the health and comfort of the women who wore them. Couching the discussion in the idea of caring about women and protecting them from doing harm to their own bodies is yet another form of misogyny; as with tattoos, piercings, plastic surgery, and other types of body modification, it should be a person's own choice to endure pain or discomfort for their appearance. With a surface-level interpretation of these radiographs, at first glance they look accurate, but deeper analysis, as shown above, brings their deliberate trickery to light. Let us now stop relying on O'Followell's lies, misrepresentations, faulty medical science, and misogyny and lay these radiographs to rest as yet another in a long line of debunked anthropological hoaxes.