Masculinity Studies and the Digital Humanities: A Review Essay, Some Experiments, and Some Thoughts on Future Directions
By Anna Young, Mellon Graduate Student Fellow for the Digital Humanities 2020-2021
In recent years, digital humanists have been particularly vocal about the gendered politics of their discipline and the academy at large.[1] Digital humanists have also produced a number of projects that look critically at historical constructions of gender. However, a cursory examination of existing gender history projects primarily examines gender as it pertains to women. Since at least the late 1990s, a substantial portion of this work focused on restoring otherwise overlooked sources authored by women or relating to women’s lives and making them accessible online to students and researchers. Example projects include large-scale text encoding initiatives like the Perdita Manuscript Project, begun in 1997, followed by Orlando: Women’s Writing in the British Isles from the Beginnings to the Present and Northeastern University’s Women Writers Project, all of which seek to preserve and center women’s writing.[2]
Beyond simply recovering and amassing repositories of women’s writing, women’s historians have also applied the traditional tools of feminist scholarship, with its attention to presence, absence, and representation, to their digital projects. Jacqueline Wernimont has pointed out how The Women Writers Project, for instance, functions not just as a “recovery” or “repository” project, but also a powerful analytical tool for uncovering latent formalistic trends in premodern women’s writing that do not necessarily track onto male-dominant generic classifications.[3] The Suffrage Postcard Project, similarly, has applied strategies of “feminist data visualization” to its corpus in an effort to track gendered patterns of representation of suffrage proponents, with a particular eye to “the absence of women of color, immigrant women, working-class women, non-cisgender women, and women without children in visual representations of women.”[4]
As excellent and sophisticated as feminist digital scholarship has become, digital humanists have much less often used these tools to examine constructions of maleness or male gender. Digital projects explicitly dedicated to the history of gender as it pertains to men are relatively hard to come by. Of course, men are hardly scarce in the historical record, nor even in digital humanities projects generally. The impetus for most of the women’s history projects I mentioned above in fact developed in response to the over-representation of male-authored sources in early digital humanities textual projects, many of which focused on the canonical “great men” of history, perhaps most notably, Roberto Busa’s Index Thomisticus. However, virtually none of the digital projects in the vein of “great men” literature examined their male subjects with an eye to gender. The implication of this is that men have generally been perceived to be less defined by gendered norms and expectations, whereas women’s writing almost reflexively lends itself to “gender” analysis. That is, while digital collections—and historical scholarship generally—tend to contain lots of sources authored by men, very few are concerned with writing about men and the cultural meanings attached to manhood.
Historians of masculinity and feminist critics have long noted this apparent paradox. The ubiquity of writing by men at once renders men both everywhere and nowhere in the historical record. Historically, maleness has often been propped up as the human “norm” in various ways. Relatively few male writers (in pre-modern European history at least) felt the need to reflect at length on their own gender identities and, furthermore, tended to take the maleness of both their readers and their subjects for granted. Thus, reading for masculine gender is hardly straightforward and often requires a good degree of inference, seeing as historical texts often equate maleness with a generalized, purportedly apolitical and ungendered humanness.[5]
If recent scholarly interventions—not to mention, recent events—have taught us anything, though, masculinity is hardly apolitical, nor should it be overlooked in historical studies of gender. Gender historians thus have an obligation to denaturalize the normative status of historically dominant categories, like maleness and whiteness, as much as we need to recover histories of women and other marginalized groups. As Wernimont has noted, the same operations of “presence and access” that have constrained or marginalized women’s writing also worked to the benefit of those with the “privilege of presence” in the historical record and are thus equally worthy of historical consideration and analysis if we ever hope to gain a more complete picture of gender in the past.[6]
So what place can the digital humanities have in all of this? Many studies of historical masculinities have already made excellent use of traditional “close reading” humanistic methods to uncover past perceptions of men and manhood, which are, again, often somewhat obscured by the long-standing historical tendency to “figure the masculine as the human.”[7] Textual analysis tools may help to get around some of these methodological difficulties that have plagued masculinity scholars. Textual analysis uses computer software to quantifiably analyze digital text. Often, textual analysis tools are used to quantify word frequencies in a corpus of texts, develop word indexes, or otherwise quantify the use of words, phrases, or other components in a collection of texts, either revealing latent patterns that might lend itself to further interpretation or helping to support an existing interpretation.[8] As Hannah Schilperoot has suggested, textual analysis tools can be profitably used to “determine gender assumptions apparent during the cultural and historical time period in which the text was written,” for instance, by examining the use of particular words or phrases by male, as opposed to female authors.[9] Because computational tools are capable of working at a large scale, they may also be useful for extracting what are often tacit assumptions about men or masculinity from a wide range of texts.
Topic modeling is one such technique that might lend itself to this kind of analysis. Topic modeling works by identifying underlying “topics” in a large corpus, by calculating the probability that certain clusters of words will appear together in a document.[10] The technique may also be profitably applied to the subject of gender by examining the alignment of particular topics with masculine or feminine-coded words. Joshua Catalano and Briana Pocratsky, for instance, have recently used topic modeling tools to analyze conceptions of masculinity and whiteness as presented in the History Channel’s programming, suggesting the utility of topic modeling for gender historical research.[11]
I’ve recently been playing with MALLET myself, because I am curious whether or not topic modeling tools can draw out latently “gendered” topics in pre-modern texts—for instance, by identifying certain words that frequently cluster around gendered identifiers like “man,” “woman,” “he,” or “she.” To test this, I ran a model on a set of 176 texts, published between 1490 and 1694, tagged as “medical” on Early English Books Online to see if medical texts from this period discussed men and women’s medical topics in distinctly gendered ways. I chose this corpus for the sake of convenience because the texts are not only in English, but they have already been transcribed and require minimal cleaning for textual analysis. I set the model to identify 40 “topics” or strings of words likely to co-occur together in this corpus. The following is a sample of the results:
5 0.1867 vrine disease great selfe learned diseases physitian vrines onely time meanes physitians colour good patient physicke reason ignorant parts opinion
22 0.05637 moone mars man venus saturne houre day iupiter luna sol dye mercury sygne sicke signe person dayes sycke wilt houres
While the lists of words above may seem a bit nonsensical at first glance, it is possible to infer from most the unifying “topic” that unites each list of words. Topic 5 refers to uroscopy, or the practice of examining a patient’s urine for signs of disease, while later words in the string suggest that this was a topic subject to some debate about the professional authority of physicians. Topic 22 clearly encompasses astrological topics and astrological healing, a common topic in early modern medical texts. These results suggest that the model successfully identified things that a human reader would also recognize as distinct topics in this corpus. Now for topics that include gendered identifiers:
3 0.1246 spirit god blood hee life man spirits things saith nature world body earth vertue cure weapon master selfe onely power
7 0.30839 nature heate body things naturall men vse time reason good man great bodie cold medicine bloud hot substance bodies matter
11 0.16694 man things men god body nature great mind reason life force grow good child children parts people time bodies humours
12 0.09906 man thinges greate declare bodye partes declareth parte long necke good nose saye sayeth handes meane founde rounde great lyfe
19 0.08042 named man shew english latin remedy mans head impediment vse word shewe looke matter body called bones member good infirmitie
22 0.05637 moone mars man venus saturne houre day iupiter luna sol dye mercury sygne sicke signe person dayes sycke wilt houres
25 0.06233 man good body meate colde men whan meates humours vryne vse stomacke hym thynges slepe stomake make drynke wyne longe
33 0.11952 oyle great make fire vse wounds cured cure helpeth water order oile wound good wine helpe time man medicines times
38 0.10547 god tree man lord himselfe state body hir hee mynde brought men people wicked christ order prophet vse minde trees
39 0.11134 good water hym geue hys wyne litle make tyme colde man called drye tymes payne hote body warme drinke thynges
The identified topics overwhelmingly favored identifiers like “man,” “men,” “hys,” and “hee,” all in bold above—“hir” only shows up once, in Topic 38. No topic mentions “women,” which is curious because this corpus included midwifery and gynecological texts that explicitly named the “diseases of women” as their subject, though these results suggest that this genre was relatively marginal within the corpus as a whole. Running the same corpus through Voyant for word frequencies also suggests that mentions of “man” were much more common than women. “Man” was the second-most commonly-used word (common stopwords excluded) in the corpus, while “woman” came in one hundred and thirty-eighth place.
These results attest to the tendency of early modern medical texts to, again, refer to “man” in the sense of mankind, to refer to medical conditions presumably universal to all humans. For instance, references to the preparation of medicines like in Topic 33 were probably not intended to be gender-specific to men. This is further reflected above in topics that co-locate “man” or “men” with the gender-neutral “people” or “person” or that seem to be referring to “man” in the abstract, such as in Topic 3 and 38, which seem to be discussing the relationship between God, nature, and humankind. Topic 38 especially sounds a lot like the Genesis creation story, in which “God made man in his own image”—a phrase not far off from the string “god tree man lord himself state body.” The implied reference to the Fall and the “tree” (of knowledge?) might also explain why this is the only topic in which a woman’s presence, “hir,” is implied.
So if men are overwhelmingly discussed in the abstract, to refer to all humankind, is it possible to draw out anything gender-specific about men from these results? I’d contend, in the first place, that the treatment of maleness as the normative state of humankind is a gendered construct that speaks volumes about constructions of masculinity in early modernity. Some topics also frequently cluster “man” together with terms that I know from my own research to be strongly associated with male, rather than female, bodies. References to “heate” and “hot” that occur above most likely reflect the common association between men and humoral heat, a quality strongly associated with activity and vitality, as opposed to cold, which was associated with sickness, weakness, and, of course, femininity. Other terms grouped with “men” or “man” above are also suggestively masculine for the time period, such as “weapon,” “master,” “power,” “reason,” “mind,” “force,” “meate,” and “great.”
Without a comparison with what topics tend to cluster around “women,” though, it is hard to say from just these results how distinctively masculine these words may have been coded. The topics retrieved, after all, don’t strongly suggest a link between coldness and women, which I can only infer from my own reading, because they only mention “men.” Further runs with more topics beyond the 40 I trained for here may successfully catch the more marginal references to women and women’s health issues, which might enable this kind of comparison. Further training with more topics may also capture more information about male sexuality and male sexual health, or relationships between men and women, which are strikingly absent from the retrieved topics, though commonly discussed in the sources. In any case, these initial results I think point to some possible applications for textual analysis in the field of masculinity studies and, with further training, may help to identify large-scale gendered patterns in historical sources.
Another digital arena in which digital humanities tools might prove useful to scholars interested in the history of men or masculinity is that of digital exhibits or repositories. A few existing projects have, like the women’s writing projects mentioned earlier, created digital exhibits or repositories centered on sources that have often been used by scholars of masculinity and identity. The Philip C. Van Buskirk Archive houses the digitized diary of a nineteenth-century Navy sailor, whose writing has often been analyzed by scholars of masculinity, much like the seventeenth-century diary of Samuel Pepys. The subscription-based Defining Gender also gathers together primary sources related to both men’s and women’s historical experiences of gender. James N. Green’s book Beyond Carnival (2001) is also accompanied by a companion website featuring sources related to the experiences of homosexual men in twentieth-century Brazil, along with a number of other collections dedicated to the sources of LGBTQ+ and trans histories.[12] Though these latter few groupings include sources that discuss constructions of masculinity in relationship to queer or non-binary identities, digital collections specifically organized around the subject of masculinity in history are still, however, relatively scarce.
Among existing projects, the Suffrage Postcard Project is perhaps most focused on consciously analyzing and deconstructing historical masculinities, as much as collecting and exhibiting sources related to the history of masculinity. Along with examining the suffrage postcards for depictions of women, the project has used social-identity tagging to examine illustrations of masculinity and fatherhood that circulated in early twentieth-century pro- and anti-suffrage postcards (Fig. 1).
Fig. 1. Examples of social-identity tags used by the Suffrage Postcard Project.
I too have recently started experimenting with visual collections in Omeka as a way of exploring how men or normative masculinity were depicted in sixteenth- and seventeenth-century European sources, as a companion to my dissertation research. My site-in-progress, Masculinity and Male Embodiment in Early Modern Europe: Visual Sources, centers on visual depictions of men and male bodies in early modernity. Right now, the site only features one collection, “Man as Microcosm,” which is intended to directly interrogate how maleness has been equated to humanness in the past and the implications of this relationship for constructions of normative masculinity (Fig. 2). It features sixteenth- and seventeenth-century medical and scientific diagrams that depicted humanity as a whole in the form of a male figure. It includes astronomical representations of the cosmos that depicted the relationship between “man” and the celestial order, as well as other representations of male-centered figurations of human bodies, like the medical images of the archetypal “Zodiac Man,” “Blood-letting Man,” and “Wound Man,” Vitruvian models of human proportions, and anatomical depictions of human bodies. In the future, I hope to add more collections that highlight how specific (as opposed to generalized) male bodies, or non-normative or ambiguously male bodies, were depicted in medical and scientific sources for further comparison.
Fig. 2. Example images from the Omeka collection “Man as Microcosm”
In sum, it seems that there is room for more digital projects focused specifically on historical constructions of masculinity or maleness, following in the vein established by women’s history projects and collections. This is particularly true for historians of the pre-modern period, as most existing projects focus on nineteenth- and twentieth-century sources. So far, my tentative experiments with topic modeling suggest that textual analysis tools might be profitably applied to some of the major methodological problems within masculinity studies and may help to enhance traditional “close reading” methods of historical scholarship. Digital exhibit curation might also allow for side-by-side reading of textual or visual sources that might produce further useful comparisons or directions for further analysis.
[1] See, for example, Moya Z. Bailey, “All the Digital Humanists Are White, All the Nerds Are Men, but Some of Us Are Brave,” in Intersectionality in Digital Humanities, 9-12 (Leeds: ARC Humanities Press).
[2] More examples of digital projects focused on collecting women’s writing may be found in the University of Michigan’s LibGuide dedicated to the subject of “Women’s and Gender Studies”: https://guides.lib.umich.edu/c.php?g=282777&p=1884212
[3] Jacqueline Wernimont and Julia Flanders, “Feminism in the Age of Digital Archives: The Women Writers Project,” Tulsa Studies in Women’s Literature 29, no. 2 (2010): 427.
[4] “About,” the Suffrage Postcard Project, https://thesuffragepostcardproject.omeka.net/about
[5] See Todd W. Reeser, Masculinities in Theory: An Introduction (Chichester, England: Wiley-Blackwell, 2010).
[6] Jacqueline Wernimont, “Whence Feminism? Assessing Feminist Interventions in Digital Literary Archives,” Digital Humanities Quarterly 7, no. 1 (2013): http://www.digitalhumanities.org/dhq/vol/7/1/000156/000156.html
[7] Kenneth Gouwens, Brendan Kane, and Laurie Nussdorfer, “Reading for Gender,” European Review of History: Revue européenne d’histoire 22, no. 4 (2015): 527.
[8] See also Ted Underwood, “7 ways humanists are using computers to understand text,” https://tedunderwood.com/2015/06/04/seven-ways-humanists-are-using-computers-to-understand-text/
[9] Hannah Schilperoot, “Feminist Markup and Meaningful Text Analysis in Digital Literary Archives,” Library Philosophy and Practice (January 12, 2015): 7.
[10] For a fuller explanation, see Ted Underwood, “Topic Modeling Made Just Simple Enough,” https://tedunderwood.com/2012/04/07/topic-modeling-made-just-simple-enough/
[11] Joshua Catalano and Briana Pocratsky, “What’s on History?: Tuning In to Conspiracies, Capitalism, and Masculinity,” Current Research in Digital History 3 (2020): https://doi.org/10.31835/crdh.2020.02
[12] See, for example, Gale Online’s Archives of Sexuality and Gender and the Digital Transgender Archive