CHAPTER 16

Building Otherwise

Julia Flanders

This chapter is a step in a longer exploration of the ways of reading technical systems as systems of cultural meaning and ideology. It is an inquiry concerning how to understand the relationship between systems and their components, their design, their meaning and the meaning of the things stored in them, their builders, their users, and the products they are used to create. In particular, it is about how otherness and logics of difference animate the tools and systems used in digital humanities; and it is about how we should act, where action requires both a theory of intention and causality and an understanding of where our intervention needs to be directed.

The provocation for this work, for me, has been twofold. First, I feel a strong challenge from work by scholars like Tara McPherson to read technological systems as ideological systems, and to focus not just on their effects as completed systems but on their genesis and development: on the ideologies that shaped their design.[1] This reading simultaneously situates digital humanities practitioners as responsible parties—designers and builders implicated in the design of future systems—and complicates that role of agency by suggesting that the ideological entailments of such systems may not be visible to their builders, and indeed that the design logics that feel most deeply natural and functional to one generation may be revealed as deeply problematic in the next. Situated thus, I want critical digital humanities practitioners to ask whether and how they would have built these systems differently, whether in their current projects they are able and willing to take different approaches, and what we can all learn from these historical examples that might inform an alternative practice. And I also want to ask whether the legacy of these systems’ genesis in a cultural logic of racism also informs their current effects in the world, and whether we need to repudiate and redesign these systems as part of a remedy. The challenge in this analysis is to understand the relationship between individual intentions, individual identities, and this larger economy of power: in what sense does it matter who builds things? It is clear that one’s subject position (as a woman of color, as a white man, etc.) doesn’t necessarily align with ideological commitments, and indeed intersectionality shows how complex even the politically visible subject positioning can be. Putting Clarence Thomas on the U.S. Supreme Court did not necessarily make that body more progressive on issues of race, but it did make race visible as part of the operations of that body. In the same way, this analysis needs to attend both to the perspective of individuals (who can take deliberate action, and can try to inform that action with critical thinking) and to the systemic forces and shifts that constrain and constitute our subject positions as agents.

This sequence of questions leads to a second, more personal provocation to consider my own situation within this matrix. My professional acculturation in digital humanities took place at the Women Writers Project (WWP; at the time, located at Brown University; now at Northeastern University), a major early effort to make gender a category of analysis and visibility in the emerging world of digital scholarly research in the most literal way possible. As a female academic, director of a digital “Women Writers Project,” writing a chapter about gender, I can’t help but be read as significantly female. And these three spaces of gender seem to line up as if they are about the same thing: my personal gender helps literalize “women” as the significant fact about the WWP within this narrative, and also imputes to me a position of authority from which to write about gender.

But the development of the WWP’s research focus mirrored a shift in feminist theory from a second-wave attention to the visibility and rights of women (for the WWP, the discoverability and valuation of women’s writing in the pre-Victorian period) to a third-wave focus on how the structure of discourse enacts and reinforces cultural power dynamics of gender, race, class, coloniality, and other differentials. The WWP’s work involved developing methods of digital text encoding for the representation of early women’s writing: in effect, developing new discursive layers within the digital code through which documentary information could be represented and analyzed, and translating the methods of traditional scholarly editing into the digital medium. The project of gender here thus entailed not only a focus on the gender of physical bodies but also attention to the strong, if less obvious, gender implications carried by technologies of text markup and editorial practice. Work by scholars like Stephanie Jed, Katie King, Martha Nell Smith, Donald Reiman, and many others has shown how deeply editorial methods are implicated in the politics of gender, and it’s also clear that in the digital medium, editorial methods are affected by, and enacted through, technological choices.[2]

In this work, as in McPherson’s, we can see a foundational assumption: there is no such thing as a “merely technical” design decision: technical systems are meaning systems and ideological systems, as far down as we are willing to look. If we are only interested in seeing and understanding the operations of gender in the places where they appear obvious (bodies and personal identity), then we will miss some of its most important operations in spaces where it may be more powerfully at work. And conversely, the places where gender may seem most obviously legible may be misleading or peripheral. My historical impact on the WWP has been to make it less about women and more about digital technologies—more attentive, perhaps, to the occult operations of gender within those technologies, but not necessarily in ways that alter our use of those technologies in response. Should we read that shift of emphasis as progressive (from a gender politics standpoint) in virtue of my own gender—this is the impact of a woman in a digital project!—or as evidence of professional acculturation recasting me as a white male technophile? Does my gender anchor my actions and intentions, or does it simply offer a perspective of difference from which other kinds of positioning (for instance, my race and class) can dislodge and reinscribe me?

These questions bear on the question of how the digital humanities might “build otherwise” because they suggest how complex “building” and “making” are as expressions of intention, identity, and cultural politics. It feels to me like a central paradox of the digital humanities that even while relying on technological systems and narratives of technological improvement, the field maintains a commitment to a “maker culture” and to “building” that means something more than simply a self-reliant desire to be handy. The field approaches making in a critical spirit, as interpreters of process and of ways of engaging with the material. But the status of cultural meaning in this context—both its tensional relation to the practical operations of tools and systems and its genesis in the various parts of those systems—is complicated to tease out. The design of a tool like Mukurtu (an open-source content management system aimed at supporting the cultural heritage needs of indigenous communities) takes seriously the ways in which information systems shape and enable and foreclose cultural meaning. Its designers deliberately foregrounded the rights and needs of those communities in their implementation of the basic functions of “content management.” But its lexicon of features (collection building, metadata tags, batch import and export, and so forth) are in many ways structurally indistinguishable from those of any content management system, and like all twenty-first-century digital tools are indebted to the still deeper features of operating systems, networks, web architectures, and so forth. If, as Tara McPherson suggests, these structural paradigms are themselves racially coded, even down to the level of the operating system, then the challenge of recognizing the cultural ideology of tools, let alone resisting it, is surely immense.

I can frame the specific concerns of this chapter as follows. If digital humanities as a domain of research+praxis is indeed prepared to take seriously the cultural, scholarly, and ideological significance of the full stack of technologies—if we’re prepared to read that entire stack as a cultural text, in addition to attending to the cultural effects of the entire stack as a working system—then three questions seem urgent. First, how do race, gender, and other forms of otherness operate within the scholarly, cultural, and ideological space of technological systems? Second, does it matter where we look for them? Do they operate differently at different places in the system? And third, do we have alternatives? Once we see these operations, what do we do differently?

What concerns and interests me about current attempts to answer these questions is the difficulty they reveal in creating a single coherent account of the operations of diversity in theories about how “building” operates in digital humanities. To illustrate this point concretely I’d like to take as examples four influential interventions that I admire and have found very useful. The first is Miriam Posner’s call, in her keynote at the 2015 Keystone Digital Humanities conference, for a complete rethinking of scholarly technological systems, an intellectual overhaul attentive to the ideological commitments in which these systems implicate us. This piece takes seriously the need for a critical revision: Posner calls for data models that respond both to the complexity of the world and to political and social justice concerns, and she also takes seriously the role that technical knowledge could play in that revision (for instance, in her “How Did They Make That?” blog series where she has been a vocal and creative advocate for greater knowledge of the “under the hood” aspects of digital projects). The examples are provocative: Google maps as an instance of the indebtedness of key information resources to corporate interests; Cartesian mapping systems more generally as an instance of the ways in which the available data models arise from colonialist intellectual and political traditions. Here she argues that two kinds of action are demanded of us. The first is a work of imagination: a move toward alternative ways of seeing, organizing, and analyzing that enables us to deliberately understand and use technical systems as meaning systems rather than as neutral and unalterable parts of the landscape. The very seamlessness of our interface with technology is precisely what insulates us and deadens our awareness of these tools’ significance: “In a similar way, many of the qualities of computer interfaces that we’ve prized, things like transparency, seamlessness, and flow, privilege ease of use ahead of any kind of critical engagement (even, perhaps, struggle) with the material at hand” (Posner, para. 12). The second action she demands is a work of building: creating real systems that use these alternative models and paradigms: “We can scrutinize data, rip it apart, rebuild it, reimagine it, and perhaps build something entirely different and weirder and more ambitious.” This is a brave as well as ambitious agenda, but it leaves open two important questions. For one thing, it is not clear how far down the technological stack this revision is expected to go. The operations of digital humanities “coding” and “building” as Posner describes them (mapping, exhibit building, interface building) are located near the top of the technical stack, leaving untouched the deeper layers: the structural logic of databases, operating systems, stylesheets, algorithms, data representations. She speaks of reimagining data models and the systems that use them, but the examples she offers for emulation demonstrate above all how difficult it is to perform that work of reimagination at any distance below the surface. One layer down, they are still working with conventional databases, content management systems, and metadata standards. A project like Jacqueline Goldsby’s “Mapping the Stacks,” in which a team of graduate students discovers and describes archival materials relating to African American history in Chicago, arguably depends for its impact precisely on the existing mechanisms of systematic visibility that are underwritten by the existing archival infrastructure, even as it rethinks where and how we look for the archival content. Indeed, it’s clear that many of the activities of positive engagement depend on the practical workings of mundanities like contributory interfaces, user authentication, and other systems that are very deeply rooted in the “business applications” side of the operation. My point here is not that Posner’s argument is misguided—far from it—but rather to show the complexities that lie ahead if we explore its implications a bit more fully. And the other, even trickier question this articulation leaves open is the issue of how changes in technical systems (using different data models, different tools) actually effect social change.

Steve Ramsay, in a provocative Modern Language Association conference presentation titled “Who’s In and Who’s Out” and its companion piece “On Building,” takes a position complementary to this one, in the sense that he too argues for the importance of “building” as digital humanities practice, but with a much stronger and less ambivalent role marked out for technical expertise: without it, we cannot competently theorize the technological stack. And without that competent theorization, the discipline of digital humanities unravels. “Building” for Ramsay is an “expansive” term that also encompasses “people who theorize about building, people who design so that others might build, and those who supervise building,” but at its heart it signals the “methodologization” of the humanities: the essential link between the work of praxis and that of critical reflection.

The complex gender and racial politics of this emphasis on competence are not lost on Ramsay. In a moving contribution to the long thread of responses to Miriam Posner’s blog post “Some Things to Think About before You Exhort Everyone to Code,” he describes the ways in which the distinctively “hacker” form of expertise translates into a certain kind of social ruthlessness or tone-deafness in which technical competence can seem—in a deceptively egalitarian gesture—to be the only thing at stake.[3] The pedagogical or cultural challenge appears to be simply how to empower women and other underrepresented groups to gain that expertise, in other words, a shift in the culture of those groups to make them comfortable with the techne of that technical space. But as his narrative demonstrates, technical competence isn’t the issue at all; his female undergraduates are perfectly competent as programmers, perfectly confident in their abilities. In their encounter with the programmer culture of the computer science department, the problem is gender, not competence, and it’s a problem of gender difference: of the male programmers not realizing they have a gender until a different one walks in the door. The space of expertise is defined not simply by qualifications but by a proprietary association between those qualifications and a specific tribe that possesses them. As Moya Bailey notes in her contribution to the same thread, the question is “about both making room at the table for everyone and also questioning who is in a position to ‘invite’ folks to the table in the first place.”[4]

This last point is especially important because “making room at the table” assumes a kind of cultural assimilation that doesn’t necessarily change how the table’s affairs are conducted. And while we may recognize that assimilation as a political sidestep around the core issue, it is nonetheless a seductive engagement for both parties. At a personal level, I can attest with some retrospective chagrin to the thrill (as a young scholar/practitioner) of discovering that one can blend in with a group of experts and be accepted as one of them. The culture of programming has remained resilient and recognizable despite the entry of women and other minorities into the field; competence interpellates us powerfully, both technically and socially, framing the status we grant to high-value forms of expertise. So the question Ramsay’s interventions thus raise is how we can treat “building” as a metric of competence that anchors professional identity, while retaining ideological and critical maneuverability.

Tara McPherson’s “Why Are the Digital Humanities So White” addresses this point full on, arguing that the ideology of technical expertise—and in fact all of the hallmarks of its execution in the foundational technical systems of daily digital activity—carries a deeply political stamp: it both bears witness to and actively enacts a logic of racism (and I think she would agree that the same logic extends to gender discrimination as well). She offers a provocative exploration of how we might read race with technology, asking why we can’t seem to see them as connected and suggesting that the very design logic of information technology—modular, “clean,” highly formalized—provides a cover story for its presumed imperviousness, or neutrality with respect to ideology. Like Ramsay, she takes seriously the need to critically theorize the full depth of the technical stack, at least down to the operating system, and also the full range of competences and practices that reproduce that stack as it evolves. The terrain that her position leaves unresolved is the question of causality: she explores in detail (as Posner does not) the possibility that causal responsibility for today’s culture of racial and gender inequity might lie in deeply embedded, nearly invisible things like operating systems: “Computers are coders of culture. . . . If . . . Unix hardwired an emerging system of covert racism into our mainframes and our minds, then computation responds to culture as much as it controls it. Code and race are deeply intertwined, even as the structures of code labor to disavow these very connections.” And it also leaves unresolved the question of personal agency: are the builders of Unix involved as intentional agents—who could have done otherwise—in the structural racism of its design logic, and if so, what would be the conditions under which that other agency could have been expressed? In the end, she leaves open the question of how individuals can intervene in their efforts toward social change, and whether changing those systems would have any effect if it were possible to do so.

Moya Bailey, in “All the Digital Humanists Are White, All the Nerds Are Men, but Some of Us Are Brave,” emphasizes the need for social change, and the need to theorize the field and its technologies in terms of race and other forms of diversity. Unlike Ramsay, she is not committed to a specific definition of “digital humanities” which can in turn be used to define the expertise and concerns appropriate to that field; instead, she is committed to a definition by inclusion, saying in effect that the field must be defined so as to include all of the activities, by all of the diverse peoples undertaking them, that bear on questions of cultural digitality. Many of these activities involve appropriating tools for a radical politics that brings race, gender, and other axes of diversity into view precisely for their frictionality within those tools. But this appropriation (through usage) of tools for social justice projects leaves the structure of tool development unquestioned and unaltered: in effect, staging an occupation rather than seeking to rearchitect along different lines. Bailey’s work thus brings us back to the issue of what kinds of expertise would be needed to accomplish that rearchitecting, and what the new purpose of those rearchitected tools should be.

From the spaces opened up between these very different interventions, a set of questions emerges. First, would it help to alter our technological design? If the problem is bad data models (ideologically discriminatory, colonialist, patriarchal, etc.) and tools that reify them, could we build an alternative stack of technological systems to serve some different version of our purposes? Would the effects of doing so in fact reverse or subvert or counter the “culture” to which McPherson tells us “computation responds”? Second, how does the culture of the process affect the culture of the product? McPherson makes a point of noting that the people actually involved in the development of Unix, their intentions with respect to the code, are something quite apart from the structural logic of separation that informed the design of the operating system itself. Ramsay seems to feel that the core problem (with respect to gender) is in large part that women are being turned off and turned away culturally from practicing in a field where they otherwise could make great and useful technical contributions (without changes to the ways software tools are imagined or constructed). If we had a more diverse programming culture, would that result in technical systems that are not, in McPherson’s terms, “white” (or by extension “male”)? And if so, how would this in turn affect the users of such systems and the communities they create? Third, what specific actions can individuals take—as designers of politically implicated systems or as resisters of them? What actions have the potential for structural change rather than merely academic self-reflection (of the sort I’m currently indulging in)? And fourth, who are the beneficiaries in question: whose interests are being served? Is the goal to remedy specific oppressions, or to create a richer critical perspective? Are there achievable changes being proposed? How widely would those changes propagate?

My goal thus far has been to convey a sense of the trickiness of the problem space. I want next to consider how the field of digital humanities has attempted, and how it might attempt, to respond. In the terms set out by my title, how do we “build otherwise”? What are the spaces within the enterprise where gender, race, and other forms of power differentiation are especially operative as ways of making a difference, and how do we respond in light of this understanding? Spoiler alert: I’m not going to be able to answer any of those questions in a satisfying way, but I am going to try to derive something useful from the failure.

An early response, whose limitations are now clearly visible to us, makes gender and race and “difference” visible as cultural content: as significant and overlooked categories in the construction of a cultural past that become visible where they are aggregated and intensified. Many early projects formed, as the Women Writers Project did, around various categories of invisibility and disenfranchisement, with gender and race very significantly among them. If the value of these early reclamation efforts lies in the ways they create distinctive intellectual spaces for the study of “other” cultural production, a value whose importance shouldn’t be underestimated even now, then their limitation lies precisely in that distinctiveness: they don’t permit the study of these “others” as anything but a separate category. They also typify the approach Moya Bailey has characterized as “add and stir” (where the addition might be any demoted category, that is, women, aboriginal peoples, etc.): the idea that if we add the missing special element back into the default culture, somehow we will end up with something whole and neutral.[5]

What’s striking about this “reclamatory” way of framing the problem is that the only category we can study is the marked, demoted one; as with the nomenclature of “women’s studies” and similar programs, the reclamatory approach focuses on the marked category and concedes neutrality and centrality to the unmarked category. Furthermore, these marked categories are not visible as part of the advanced search interfaces for major comprehensive digital resources such as EEBO, ESTC, NINES, or Google Books. And to the extent that categories of personal identity such as gender and race are visible (for instance, in WorldCat Identities), it is only as a marked category: female authors, authors of color, and authors with disabilities all bear the informational traces of their difference, but these markings are not part of a systematic regime of information; there is in every case a “null” unmarked value, a default setting (whiteness, maleness, ableness, straightness). More subtly, as Hope Olson has argued in “The Power to Name,” in the subject cataloging systems prevalent in the United States (i.e., Library of Congress subject headings), these categories of identity are located several steps down in a taxonomic system whose primary divisions are things like “literature” and “history”; identity is treated only as a qualifier on other more salient informational categories (“American fiction—women authors”), rather than as a primary category of discovery.[6]

The attempt to foreground categories of identity also assumes that we can represent these categories as part of a clear-cut and unproblematic descriptive vocabulary for identity, and that we have (and wish to apply) clear criteria for discovering whom it applies to. The Text Encoding Initiative (TEI) considered this issue a few years ago when Melissa Terras agitated successfully for an expansion of the options for representing “sex” as a characteristic of persons. The TEI had previously used the ISO standard “codes for the representation of human sexes” whose permitted values are 0, 1, 2, and 9 (not known, male, female, and not applicable). From an information-retrieval standpoint, those values have a certain kind of brutal utility (setting aside the humor value of “1” and “2”): to realize the kind of basic discoverability of “women writers” in WorldCat, some simplistic representational standard is needed, though perhaps not quite that simplistic. But as descriptors these values are obviously impoverished; the TEI now permits projects to define their own descriptive vocabulary or to use an externally defined standard (such as ISO).

But descriptive adequacy is not the only goal here. Amber Billey and colleagues present a detailed critique of the cataloging rules expressed in the Resource Description and Access (RDA) standard concerning the representation of the gender of authors, pointing out that the cataloger’s imperative to classify and make visible the gender of an author (based on evidence such as the cultural gendering of names, etc.) may run counter to an author’s sense that their gender is not a relevant or easily categorized fact:

RDA rule 9.7 poses problems on two grounds. First, the rule directs the cataloger to describe the gender of the author as part of the project of constructing access points and relationships between bibliographic entities. In this sense, the gender marker is like format or the number of pages: an objective description of reality. The author really has a single gender that could really be captured by the cataloger. Queer theory, as well as the lived experience of authors of non-normative genders, tells us this is not so. The second problem concerns retrieval. By marking the gender of the author using a fixed category, the LC interpretation of RDA reifies contemporary understandings of gender as a binary system with only two acceptable gender markers (male or female). Even if catalogers indicate gender using alternate labels, RDA’s insistence on the relevance of gender as a descriptive attribute reifies regressive social binaries and is passively hostile to transgender individuals. (Billey et al., emphasis added)[7]

Indeed, even what might look from a third-wave feminist perspective like a very progressive development—making gender visible as a category in metadata, and recording the fact that gender assignment may change, rather than treating it as a permanent and self-evident category—in fact looks to Billey and fourth-wave feminism quite retrogressive. For one thing, it reifies the oversimplification of “male | female,” and in fact any version of a controlled vocabulary for this purpose, however extended, is definitionally going to be an oversimplification—all the more so with categories like race that have for much longer been understood as fluid and local. And it also enforces the requirement that gender be treated as a category of identity at all, which these authors assert is not necessarily something everyone wants or benefits from.

These forms of attention to difference get us a certain distance in understanding how gender, race, and other formations inhabit digital systems, but they ignore something deeper, namely, the power dynamics inscribed in those information systems themselves. Directives on how to read those power dynamics are available to us from many quarters, and I have scope to mention only a few here in hopes that their further implications will be clear by extrapolation. As I noted at the outset, early humanities computing projects explored the ways that a gender politics (and the politics of other power vectors) might be embedded in editorial theory, quite apart from considerations of the literal gender of authors or editors. We see this enacted in practice in digital archives and editions that seek to revise an Anglo-American tradition of critical editing focused on producing authoritative editions that informationalize their sources and produce a kind of textual master knowledge. Examples include the Dickinson Electronic Archives at the University of Maryland, the Women Writers Project, and arguably also editions that pursue a similar editorial agenda but with an editorial politics focused on power dynamics other than gender: for instance, “fluid-text” editions like the Melville Electronic Library.

This work of revision draws on an existing strand of research among traditional (nondigital) textual editors and theorists interested in Western theorizations of the body and what we might call “the gender politics of abstraction”: for example, scholars including Naomi Schor, Stephanie Jed, Joan Scott, and others who examine what Joan Scott calls “those long traditions of (Western) philosophy that have systematically and repeatedly construed the world hierarchically in terms of masculine universals and feminine specificities.”[8] As scholars like Felicity Nussbaum and Terry Eagleton have argued, the same logic of physicalized otherness also extends to race and to class.[9] These same roots in neoclassical aesthetics are those from which the digital humanities also draws very heavily in its conceptualization of things like schemas, data models, and ontologies, and this work suggests that we need to be equally attentive to the politics of difference that inhabit these instruments.

It should thus come as no surprise that the same neoclassical aesthetic that brings us the feminized, particularized body and complementary narratives about race, class, and forms of labor should deeply inform the ways practitioners and theorists in digital humanities think about building and making. The politics of praxis in the digital humanities are illuminated by metaphors like “getting one’s hands dirty” with tools and coding systems, just as much as by those in which computation is the “handmaiden” of scholarship. In another place it would be well worth a digression into the details of these metaphors, but a few key points are worth unpacking. First, if we’re interested in seeking out the architecture of difference within the logic of “building” that animates the “maker culture” in DH, we need to consider what the idea of “building” commits us to as a cultural meme. It puts the self-reliant maker at the center: as a heroic small producer, fascinated and satisfied by the process of creation, legible as an artisan but also as an entrepreneur, perhaps also as a crafter, and also as a figure whose intentions and desires and self-determination matter.[10] Within this individualistic space, identity and the right to self-expression are an adequate basis for ethics. As a result, the “maker” figure writes individualism into the maker space in ways that make it harder to think simultaneously about structural factors, such as how the maker might be implicated (in the complex ways McPherson lays out) in a design logic animated by structural racism. This version of the maker asks us to see the individual as an ignorant victim of system, or as complicit with system, or as an embodiment of free will that invalidates critique at the system level altogether, but it doesn’t give us a way of seeing the intentions of the maker as politically irrelevant and structurally ineffectual. This characterization also offers a form of cultural critique that tries to access the appealing preindustrial space of the workshop (mirroring current fascination with artisanal food, furniture, and so forth), while turning its gaze away from the ways in which the “raw materials” of the DH maker space, like the prepackaged “curated box of do-it-yourself electronics” of the AdaBox, are themselves industrial products (and require an industrial-grade system of global transport and information dissemination to bring them to market).[11]

But this is not to say that attention to “building” in DH forecloses all access to the more complex politics of that identity. It usefully foregrounds the relation of the individual to the meaning systems of code, and the ways that praxis embeds individual bodies in work processes. It also fruitfully transgresses or transcends the traditional professional identities of the scholar, the developer, the librarian, and thereby draws our attention to different forms of expertise and knowledge that could constitute “scholarship.” Finally, it’s worth noting that the use of “code” as a proxy for digital humanities activities of “building” has some interesting slippages. The word is aligned in common usage with the maker space, with “thingness” and praxis (just as we say that code is a “building block” of a program, part of its “architecture”), but it is also aligned with “encoding” and thereby with discourse, notation systems, the realm of the symbolic. In the politics of DH, “code” is deeply polysemous. It can stand for a kind of machismo, the domain of the hacker whose credibility rests precisely on an uncritical but unarguably expert facility with tools, and for a place of self-empowerment and individual agency, offered equitably, as a kind of leveling oppositionality. This oppositionality is realized through the proliferation of workshops and self-guided teaching resources by which DH practitioners are encouraged to become builders and thereby authenticate their critical relation to conventional formations of the academy, including its traditional power structures of gender, class, race. And “code” also stands for the place where language is most deeply and mysteriously operative in our systems (and hence where our critical and interpretive attention might find its most fruitful object).

The reason I bring us around by this convoluted route to this set of points is to show, first, how fully cultural politics proliferate within all technical structures and practices, and second, how polymorphous those cultural politics are: how difficult they are to map onto a single problem like gender or race or class, and how thoroughly the human, the social, and the technological are mutually implicated.

As an experiment in reading the individual in relation to systems, I’d like to offer a brief case study that looks at two significant female figures in the history of technical systems: Grace Hopper and Jean Sammet. Grace Hopper has been rediscovered by digital humanists as one of the early female computer scientists; she held a PhD in math and had a lifetime of service in the U.S. Navy. She was also the developer of the first compiler, and was a contributor to the COBOL language and the developer of one of its progenitors (a language called FLOW-MATIC). Jean Sammet was another notable early female computer scientist who was closely involved in the development of COBOL. In the 1970s, she was the first female president of the Association for Computing Machinery, and she designed and taught some of the earliest graduate-level courses in computer programming. Extrinsically, it surely matters that these two figures are women, and we could unpack (as many articles have done) the ways in which their gender affected their educational opportunities, their working conditions, the expectations their colleagues had of them, their relationships to systems and institutions of power, and so forth. But what explanatory value does gender hold for their work as designers of programming languages? Or, to come at this from another direction, what are the salient qualities of FLOW-MATIC and COBOL, and how might we begin to read them in cultural and political terms?

Both languages were distinctive at the time for being written for comprehensibility rather than pure arithmetic brevity: FLOW-MATIC was the first programming language to use natural-language-like words rather than symbols, and COBOL extended this approach. In both cases, the goal was for programs to be legible not only to programmers but also to managers. These systems are thus also framing the problem of code notation as a problem of pedagogy and documentation: in other words, situating program code in a work ecology that includes nonexperts, broadening its intelligibility, making it less of a guild knowledge. Jean Sammet in her history of programming languages describes COBOL as being designed both for “the relatively inexperienced programmer for whom the naturalness of COBOL would be an asset” and “essentially anybody who had not written the program initially.”[12] As she goes on to observe, “the readability of COBOL programs would provide documentation to all who might wish to examine the programs, including supervisory or management personnel” (335).

These characterizations suggest an emerging sociality of code, which embeds it in a broadening population of users and readers; we could read these historical signposts as pointing toward the more collaborative working environments of the digital humanities. But we might also say that Hopper’s and Sammet’s work situates program code in a work ecology that is corporate and military rather than scientific: in other words, environments where technical work needs to be consumed and evaluated within systems of hierarchical power where technical expertise exists only in a limited stratum, rather than within an intellectual peer group.[13] And it is surely also interesting that COBOL was also commissioned as a standard language (to eliminate the unmanageable diversity of machine-specific languages), reflecting the fact that computers were becoming numerous enough that portability of code could be useful. That portability signals a set of design concerns that a few years later motivate the very modularity that Tara McPherson marks as part of the deep logic of racial separation she sees at the heart of Unix.

I proposed Hopper and Sammet as a case study in how we read the meaning of individual interventions, and the sketchiness of the detail here clearly suggests how much more research would be necessary to complete that reading. But I think we can see nonetheless the kinds of questions such a study prompts. Did these women accomplish anything that is recognizable to us as “building otherwise”? If not, why not? If so, how? To what extent is their agency as individual designers or builders visible to us in the systems to which they contributed, to what extent does that agency bear the stamp of their identity as women, and to what extent can we trace effects that are somehow constructive from the perspective of gender politics? Or some other politics? What did they construe as the opposition? For Grace Hopper, “the establishment” comes up as a repeating figure; what perspective (or failure of perspective) does her status as a woman (in the armed services, in business, in the field of mathematics, and so forth) give her on constructing oppositionality in useful ways?[14] Does it matter, for these purposes, that the types of feminism that might have been visible to her are not necessarily forms we now feel empowered by? What would it mean for current debates in digital humanities to take a more intersectional approach to the examination of the political logic of technical structures, acknowledging how deeply gender and race are implicated in class and economic formations? What does it mean if functional determinants of quality—the application of expertise, the consensus of users, increased efficiency, and so forth—lead us in directions we can clearly see are culturally fraught, in the ways that Miriam Posner and Tara McPherson highlight so clearly? Does reversing that developmental trajectory make a difference? If we build systems that are frictional and self-dismantling, that refuse the design logic arising from male-dominated culture, will they help us build a better society? If women (or people of color, or people from the Global South, etc.) are involved in setting technical directions and establishing those functional determinants (as Grace Hopper and Jean Sammet clearly were), how does that affect our assessment of the political valence of those systems, or do we first need to consider the acculturation and initiation processes by which individuals are “invited to the table”? Does an improved process guarantee a good outcome? And if so, is that because “objectively” (whatever we mean by that) the outcome is better, or because having a better process validates whatever outcome we arrive at?

These reflections are at best a frame for a further research agenda. But in an eleventh-hour addition to this piece it may be relevant to note some preliminary outcomes from an initiative that is taking up that agenda. At an October 2017 forum titled “Design for Diversity,” participants shared a set of case studies investigating the ways in which information systems—digital interfaces, metadata standards, online exhibits, and other tools and components—are animated by forms of cultural hegemony.[15] Facing the question of where such tools and systems express or enforce such cultural norms, and how they might be designed otherwise, the group repeatedly pointed to the shaping force of social processes and relationships in determining the ethical shape of technical outcomes. Social processes such as decision making, information sharing, and strategic planning, to the degree that they included all those implicated in a project’s outcomes, tended to result in systems that were more resilient and more accommodating of diverse cultural positions; sadly, many of the case studies illustrated the corollary position with examples of brittle or failed designs arising from poorly planned processes. (It is worth noting that the typical language of “stakeholders” to describe inclusivity points to the heart of the problem, since it treats the means to ante up, rather than ethical entitlement, as the criterion of inclusion at the table.) Similarly, the group pointed to the importance of building trust relationships within and outside the project that reflect the genuine ethical entailments of the project’s impact in the world (rather than the limited set of entailments representing the project’s own self-interest). In case after case, in a remarkable variety of ways, the specific work of technical design—and the expertise and intentions through which it is effected—was shown to be strongly shaped by these broader contextual factors.

What this discussion suggests is that the project of remaking tools may depend for its success on the social processes employed, and further that the social significance of technical systems lies not only in their overt functioning (what they enable us to do) but also in the social effluent, so to speak, of their construction processes. When a system like Unix, or a language like COBOL, or a resource like the Women Writers Project is created, what does it “give off” in terms of expertise, power relations, installed systems and dependencies, professional advancement or subordination, knowledge and empowerment—and for whom? The successful processes portrayed at Design for Diversity were inclusive in very significant ways, but in particular their principle of diversity had to do with a genuine diversification of the allocation of power: the power to say what is most important about the design of a tool or system, the power to update a record, the power to define vocabularies, the power to say what should be visible or invisible, the power resulting from increased knowledge or expertise. Not only was the tribe of “coders” being diversified, but also that tribe’s understanding of mission—what is being built, for whom, why, under what design imperatives, with what specific stipulations—was being shaped by diversified constituencies, operating under radically different assumptions about whose needs matter. Building otherwise, in digital humanities, may thus require that we understand building as a deeply embedded expression of social justice: that a tool or artifact that is “for” a purpose or an audience needs to involve those it affects in the full ecology of its design, and that we are never building only for, or as, ourselves.

Notes

Bibliography