Networks of Critique

Following Marianne de Laet and Annemarie Mol, “critique’s” fluidity—its capacity to be remade and redeployed across a wide variety of spaces while retaining an epistemic wholeness—may be one of the term’s enduring strengths. Critical is a slippery (Law and Lien) term, more so as we see it deployed and redeveloped across a wider array of disciplines. It is now common to find calls, programs, and edited volumes dedicated to critical computing, critical design, and critical methodologies. As Jeffrey and Shaowen Bardzell have noted, across design, the social sciences, and technical disciplines, “critical” can simultaneously stand in for literary criticism, the “capital C” Criticism of the Frankfurt School, and broader movements within social and cultural theory, including intersections of feminism, critical race studies, de/postcolonial literature, queer theory, and disability studies (Bardzell and Bardzell). Across the humanities, too, critique has become fluid terrain, both in terms of its definition and in its perceived usefulness. In The Limits of Critique, Rita Felski largely constructs contemporary critique as a broad set of “suspicious reading” practices that aim to uncover hidden capitalist, misogynist, and heteropatriarchal values woven in between the lines of classic texts. This suspicious reading, Felski argues, rewards the capacity for scholars to develop a critical distancing from their texts, both in terms of focusing on context over text and in terms of forming a negative affective relationship with the text in order to identify its—or its author’s—political failings. Conversely, Sheila Liming has argued that Felski and her “postcritical” stance represent a longing for an imagined precritical institutional past, where the humanities thrive by aligning with, rather than challenging, institutions of power. Borrowing from Laura Wilder, Liming notes that disciplinary and transdisciplinary discourses about critique are often less about defining critique itself and more often about taking public umbrage with other scholars’ definitions or research practices. As such, “critique” and its criticisms become transformed into blunt instruments to be wielded against one’s rivals.

Latour—himself a divisive figure among the humanities and STS—consistently emerges as a source of both diagnosis and intervention with regard to critique across the humanities, social sciences, and design. Latour is often encountered through his 2004 article “Why Has Critique Run Out of Steam?” (henceforth “Steam”), itself a statement of public umbrage against Latour’s sociological rival Pierre Bourdieu (Nelson). Though in “Steam” Latour is writing for a sociological audience, the piece has been widely cited across multiple fields for its tackling of the “double gesture” of the critical stance. The double gesture, Latour argues, is that the critic moves to undermine those who believe an object to be valuable (be it a scientific fact, cultural narrative, or prominent text), first by revealing that value is only projected onto that object by individuals (the fetishizing moment) and second by also revealing the shaping of individual minds by external forces, be it of capitalism, culture, or power. Valuers of an object are thus doubly duped—the objects they value have no intrinsic value, and they only value those objects because they are told by others to value them. “Do you see now why it feels so good to be a critical mind?” Latour pokes. “You can turn all of these attachments into so many fetishes and humiliate all the believers by showing it is nothing but their own projection . . . entirely determined by the action of powerful causalities . . . you alone can see” (“Why Has Critique Run Out of Steam,” 238–39). The critic adopts a distanced stance—they articulate their power as the ability to stand above and apart from the fetishizing loop.

None of us, of course, can truly stand away from or above the social, including the social forces that fetishize “critique” itself as a valuable object. Felski leverages Latour to argue that critique’s power comes not from intellectual robustness but rather through an alliance of powerful theorists, texts, dissertation committees, and advisers who maintain critique’s hegemonic power of orientation in humanities scholarship. Felski thus turns the double gesture back on critics themselves: critique has no value in and of itself, only that which is socially projected, and critical scholars only project that value due to powerful institutional and cultural forces that reward them for doing so.

The way to break critique’s stranglehold in the humanities, Felski argues, may lie in the use of Actor-Network Theory. For Felski, ANT’s focus on tying together webs of relationships—tracing why important texts are powerful and enduring, why they matter—is a counter to the “ethos of negativity” of critique and its invoking of “historical-political contexts” (Felski, “Comparison and Translation,” emphasis in the original).

Latour and his contributions to ANT are often posited as ways of flattening the analysis of humans and nonhumans in order to develop a more “realist” account of social and material phenomena. This realism can be used as a way of dismissing critical and contextual scholarship and can become leveraged in support of anti-critical political positions (Lossin), often through the argument that contemporary humanities have placed too much value on “the social” or “the contextual.” Latour himself writes that when analyzing phenomena, “the social” is what is to be explained, rather than the explanatory factor (Reassembling the Social). But herein lies the issue, especially when relying on Latour’s “Steam” and his specific spin on ANT to decenter political and normative analysis. Latour desires to do critique better, not less.

ANT provides a methodological framework that allows STS scholars to speak of the agency of concrete material conditions on the social practices of technical and knowledge workers, while also articulating the contingency and co-determinacy of both those material conditions and social practices. In Latour’s words:

To try to follow an actor-network is a bit like defining a wave corpuscle . . . any entity can be seized either as an actor (a corpuscle) or as a network (a wave). It is in this complete reversibility—an actor is nothing but a network, except that a network is nothing but actors—that resides the main originality of this theory. (“Networks, Societies, Spheres,” 5)

As Venturini and colleagues note, the “hyphen” in actor-network is not a relational connector, not a way of understanding actors connected within networks. Rather, it is an equals sign: both an actor and a network, and therefore neither fully an actor nor a network (Venturini, Munk, and Jacomy). This point is, unfortunately, often mischaracterized by both STS and humanities scholars, who imagine actors as “individual entities that assert force while interacting with other entities” (Van Gorp and Bron, para. 8), a characterization that results from the historical and methodological coupling of ANT with social network analysis (Venturini, Munk, and Jacomy). While such a stance can be descriptively useful, it also assumes prebuilt entities that interact, rather than actors that come to be through their relations.

ANT is not anti-political or anti-contextual; rather, it asserts that what counts as the “context” of an object or a text comes to be through relations of actor, network, and analyst. Text and context are not fixed. Rather, they are more like an autostereogram—a “magic eye” puzzle—where background and foreground have as much to do with where we fix our eye as with the printed image itself. As we move our eyes, our heads, our hands, as we reorient our relationship to the page, new shapes, contours, and figures emerge—they reach out to us. The image only becomes fixed when we hold our gazes and bodies stable.

It matters what we hold stable in our critical analyses. Felski, for example, holds stable a certain orientation of the humanities in her writings; it is, as Liming notes, a classical institutional definition of the humanities. We again see Felski’s particular stabilization when discussing the challenges to her own desire to bring ANT into literary studies pedagogy:

And yet, while an occasional course on actor-network theory may sneak its way onto an English syllabus, the chances of most classes on the Victorian novel or contemporary women’s fiction being refurbished as classes in the sociology of mediation are close to nil. That is not, after all, what most teachers and students come to literature for. (The Limits of Critique, 184)

The pragmatic points here are well made: It is unlikely that Victorian literature classes will become deeply entangled with sociological theory. However, the argument papers over some of the more conceptual nuances of the very framework Felski is calling for. Again, for ANT, neither space nor context preexist the actor-networks that “inhabit” them. Rather, they are productions of those actor-networks, part and production of particular historical, material, and social arrangements (Dourish). “Space,” then, is itself an actor-network, in that it entangles and becomes entangled by other pieces of the network from which it is constituted; it coproduces its own stability.

Similarly, students do not come to literature classes. Rather, those students, classes, theories, and texts produce one another, and narratives of what those classes and students are for emerge from the practices and material arrangements of classroom-student-institutional networks. The epistemic stability of literature classes is as much a result of social and material arrangements of power as any other stable network. Similarly, texts or works of art on their own do not produce webs of entanglement. Rather, they come to be—as texts, as sources of political and cultural power, as objects of criticism—through the actor-networks they coproduce.

If we map Latour’s “Steam” through Ahmed, we may find that critique as a mode of inquiry is not necessarily defunct but that it allows us to trace the shifting orientations of critique, field, and the digital humanities themselves. Tracing what counts as actor and what counts as network in any given moment—and how those accounts shift from moment to moment and from analyst to analyst—is what provides ANT its analytic and political power. An ANT approach would argue that there is no bringing of critique or the political “into” a space. Rather, critique is the disruption of networked space—the recognition that what counts as spatially stable and what counts as fluid or contested is always in negotiation. Critique may be disruptive, but disruption need not be destructive or negative; disruption in an ANT sense means reorientation. Critique becomes redefined—not as a suspicious reading of a text, nor as a form of public airing of grievances, but as a vector for new spatial and disciplinary production. Critique in an ANT sense is not separate or distant from the objects it analyzes, the authors who wield it, or the students who read it—they are all a part of one another.

To return to spaces of critical DH: The field has long since moved past narratives of existing only in English departments (Kirschenbaum) or as a “big tent” cordoned off from more “traditional” humanities research (Svensson). Gold and Klein (“Introduction: Digital Humanities”) illustrate how since its inception, DH has permeated through the fields of book history, Black studies, art history, and archaeology, among others. As noted above, however, even these distinct fields have overlapping disciplinary and institutional networks and share many scholarly commitments. There are possibilities for DH to assemble critical spaces even further afield. As Liming argues, critique and critical scholarship’s future “continues to bravely take shape outside of traditional, institutional containers.” What can DH be if we assemble it elsewhere and allow it to be reassembled by those elsewheres?

A Critical DH Elsewhere

Here I turn to a critical DH project, “Critical Computer Science,” an experimental pedagogical-scholarly reorientation of both digital humanities and computer science (CS). Though institutionally this project was housed between a department of computer science and a department of science and technology studies, it has since its inception been articulated as a DH effort and was supported through the National Endowment for the Humanities (NEH) Office of Digital Humanities and Humanities Connections programs. I want to highlight this project here to show what new networks of critical DH “outside” of the humanities can provide both institutionally and epistemically. Institutionally, Critical Computer Science reoriented the roles of both humanities and science departments, including producing new imaginations of the relations the sciences can have in developing cultures of critique and the role humanities faculty can play in the sciences. Epistemically, the project served to produce ways of analyzing relations of politics and code, the actor-networks that produce the ontological boundaries of programming, and how neoliberalism shapes academic structures and scholarly outcomes. Through DH, computer science spaces became reoriented into vectors for humanistic critique.

I undertook the Critical Computer Science project in collaboration with several graduate and undergraduate CS students who felt marginalized by their computer science classes and curricula. Feelings of marginalization occurred for a variety of reasons among the research collaborators, but many had to do with a perceived lack of space to productively address oppressive algorithmic structures from within their computer science curricula and training. It is no accident that the first stages of this project came into being in the wake of the Cambridge Analytica scandal, the (sometimes enthusiastic) participation of Big Tech in the development of a migrant tracking and detention network in the United States during the Trump administration, and the growing public concern over the use of facial recognition and artificial intelligence (AI) technologies in policing and governance. It is also not an accident that the student researchers who voiced these concerns represented a broader array of genders, races, and (sexual) orientations than typically imagined in computer science classrooms.

These students did have some spaces to talk through these concerns in their humanities and social science electives. It was through these electives, in fact, that I met several of the students who would become the first cohort of Critical Computer Science undergraduate researchers. But, by and large, while these students were thankful for the conversational space in their STS and cultural studies classes to discuss algorithmic injustice, they were disappointed that they had to leave their homes in CS in order to do it. These students viewed computer science as a part of their disciplinary and institutional identity. Contra Felski, the critical impulse at work here was not a desire among these students to distance themselves from their object of critique—computer science as a political institution—but rather a desire to build networks that allowed for a deeper entanglement with it. These students did not want to tear apart their CS education, but they did want to reorient it.

The first output of our project occurred, like many critical DH initiatives do, in the classroom. With permission from the CS department head and support from sympathetic faculty members, the students and I designed Critical Computer Science 1 (CCS1), an alternative section of standard CS1. CCS1 centered critical race theory and feminist technoscience studies in the computer science classroom and developed alternative homework assignments and lectures, incorporated classroom reading discussions of STS and DH articles, and formed outside-of-class marginalized student support groups. In effect, Critical CS1 taught the basics of computer science in a contextualized way: Students still learned “core” programming skills—such as lists, loops, conditionals, dictionaries, and classes—via Python, but they did so through engagements with literature and datasets that outlined the histories and cultural values embedded within computational logics and systems. The course also included interrogations of “pre-digital” computationalism (Golumbia), such as the use of pen-and-paper algorithmic decision-making tools in the nineteenth century to identify and ghettoize Irish migrants (Shrout), in order to give students a “long view” of the digitized present. Many of these conversations were made possible by leveraging prior examples of critically oriented DH work, like the collaboratively produced Torn Apart/Separados site and Anelise Hanson Shrout’s “(Re)Humanizing Data” project, as tutorials or starting points of projects and assignments. Perhaps most institutionally important was the fact that students who satisfactorily completed CCS1 received the same administrative credit as students who completed the standard CS1 course. This meant that students did not risk missing their CS requirements or being locked out of upper-level CS classes if they chose to enroll in CCS1. It also meant that CCS1 was required to meet the learning objectives and standards of “rigor” of standard CS1, while also maintaining the critical orientations we demanded of it.

Though I am not a computer science faculty member, I was given permission, through an unpaid teaching overload, to teach the class, under the condition that students had to specifically request being placed in the “critical” section and could choose to leave it at any time. The result was a required core curricular science course, structured through STS and DH readings, supported by the NEH, taught by a humanities faculty member, codesigned and co-taught by CS undergraduate students critical of their own education, and delivered to CS students who self-selected based on their desire for a politically oriented computer science. The actor-network analysis is already useful here: This unconventional network coproduced a DH research project, a critical community space, and an accredited computer science course. It allowed critical humanities work to leverage the institutional capital and epistemic infrastructures (Malazita, “Epistemic Infrastructure”) of multiple scholarly spaces, including those more integrated into the resource and funding wells of the university. It also, as I will show, allowed for students who may not necessarily have come to a traditional humanities course to imagine productions of critique beyond networks of textual interpretation, which in turn led to political action.

Changing Capacities

It would be easy enough to characterize the Critical Computer Science project as an effort to bring “ethics” via STS and DH into a CS classroom, rather than itself being a DH project. The former was, for the most part, how CCS1 was interpreted by sympathetic faculty in the computer science department, and it was a narrative that was useful in articulating some of the political and institutional implications of the project to administrative audiences. But CCS1’s successes—including students in the first cohort transitioning into project researchers, as well as publications in humanities and social sciences journals coauthored with CS undergrads—stemmed from the project’s reorientation of concepts and institutions of computing in ways deeply aligned with humanistic critique. In doing so, CCS1 produced students with critical humanistic capacities who would not otherwise have had the networks to develop them. Moreover, the kinds of critique these students were capable of was distinct from those we may see in a more “traditional” humanities student. It was not just the ability to “suspiciously read” code and software—though there is plenty to be suspicious about in contemporary algorithmic culture. Rather, critique came in the students’ capacity to continuously reorient the disciplinary conventions of computer science in ways that enabled more politically engaged work to be done.

I mentioned above that since CCS1 awarded formal CS credit, we were made subject to faculty oversight so that we met standards of CS education. This was quite reasonable, and even welcome, because we did not want graduates of CCS1 to be at a disadvantage in their later CS1 classes compared to students who had passed the standard section. However, what became quickly evident was that what was “counted” by overseers as within the actor-network of CS1 consistently shifted. For example, when Critical CS1 began, the research team was told that Computer Science 1 was not about teaching programming; rather, that course is designed as a foundational introduction to computer science principles, such as abstraction, basic logic gates, basic database parsing, and problem-solving skills. That students learned the Python programming language when exploring these epistemic dimensions was considered almost tangential. This was initially a boon for the project, as defining computer science education epistemically rather than instrumentally—as developing knowledge practices rather than as developing a set of technical skills—dovetailed nicely with the critical race and feminist science studies approaches that we sought to introduce to the course. But after we presented our assignments, which taught these computing fundamentals alongside critical perspectives about their raced and gendered dimensions, the oversight group reconstructed the course as “just about learning Python.” Within this revised framework, core computational concepts like abstraction were labeled as too advanced, thus coupled STS or DH readings that critiqued foundational assumptions of computer science were positioned as out of scope for the class.

This is not to complain about the tensions that can arise during interdisciplinary work. Rather, these events reinforced for CCS researchers (including myself) the shifting co-constitutive actor-networks of code and programmer identity. The fluidity of these networks has been well documented in STS literature. As Stéphane Couture illustrates in his ethnographies of open-source developers, what counts as legitimate coding practice, as well as who counts as an expert, is heavily dependent on how one’s subject position and technical practice are produced within an actor-network. As Couture shows, however, there is no single definition or quality of code; rather, what becomes constructed as code is both paratextual (Genette and Maclean) and also produced by the identity of the coder. Related documents, such as technical documentation, pseudocode, and graphical “maps” of the program’s structure are often included by developers as a part of the actor-network of a piece of code. The developers in Couture’s study would often refer to source code as both “text” and “textual.” Also contributing to the ever-shifting boundaries of source code were gendered dynamics of expertise—women were less likely to identify their technical practice as source code development than men.

In this project, the nature of code became a contested network, both in terms of “what counts” as code and its role in computer science classrooms, as well as what counts as “core” skills CS students need to develop and internalize. The assignments that CCS students codeveloped would pass in and out of “real CS assignment” status, depending on the reader, the challenge of the assignment, the amount of reading (the more an assignment required reading, the less it was a “real” CS assignment), and the identity of the assignment presenter. It was, for example, easier for a CCS assignment to get approved when a CS student presented it to the oversight group than when I (as a humanities faculty member) did. CCS students, in turn, learned to develop a flexible epistemic orientation to their CS identities throughout the course of the project and their remaining student careers that would allow them to smoothly interact with the institutions of computer science while also challenging them. They had to recognize when it was strategic to wrap themselves in their computer science identities, particularly when participating in student-led CS department committees and organizations, and recognize the cachet it provides when their technical capabilities are called into question by other students (as the story at many engineering schools goes, politically oriented students and technically oriented students are mutually exclusive).

However, CCS researchers also internalized fundamentally humanistic questions: What are the shifting boundaries that determine whether or not one is a programmer or a “legitimate” computer scientist? What were the entanglements of the “text” of the course—Python programming and introductory procedural literacy—and the politics and institutional structures of the course, and how did they come to shape one another? Why did the definitions and discourse of what an introductory computer science course was “supposed to be” change as different people—many of whom represented marginalized identities—attempted to intervene in those courses? Students were able to articulate these questions to other student audiences, leading to the creation of student activist groups, reading groups, and protests of software and engineering firms that contract with the border security military-industrial complex—something almost completely unheard of in the history of our engineering institute.

This, I believe, is exactly the kind of institutional capacity that critical DH can exert if it begins making itself a part of the actor-networks outside of traditional humanistic disciplinary spaces. Such forms of critique are not negative, in that they are not distant or dismissive of texts or technical practice. Nor are the analysts and students who wield such critique distant, as their efforts are oriented toward further entanglement with a broader array of institutional and disciplinary activity. This critique is co-constitutive: It destabilizes our own familiar institutional ground so that we create new understandings of the power of humanities inquiry. This reorientation, in turn, can be transformed into reflexive practice, redefining the boundaries of what it means to be a humanities scholar and creating opportunities for the composition of new and effective critical scholarship and political impact.

Note

Activities described in this chapter were sponsored by the National Endowment for the Humanities (grant numbers AK-255350–17 and HD-248450–16). The described Critical Computer Science assignments were developed by Audrey Beard, Korryn Resetar, Brookelyn Parslow, Naya Murdock, Jesse Ellin, Chris Reed, Xavier Marshall, Ohad Nir, Eryn Buhat, Damiel Faxon, and James Malazita. I want to give a special acknowledgment to Lee Nelson for our long conversations about Bruno Latour.

Bibliography