How might the work of Karen Barad influence rhetorical algorithm studies?
Rhetorician Daniel Hocutt argues that the relationship between humans and algorithms is fundamentally rhetorical due to how they mutually “influence” each other’s activity through their interactions. Hocutt attributes rhetorical agency in this relationship not to humans, algorithms, or both but rather to the (inter)activity “itself,” or what he calls “emergent assemblage activity; “a product of human and nonhuman entities working alongside each other in complex platforms.” Introna has similarly characterized human and nonhuman algorithmic relations as an “assemblage,” a “temporally unfolding sociomaterial assemblage or process” (Governance 23).
I draw a comparison between Hocutt and Introna not because they are the only two scholars to use this terminology, by any means, but because both identify a similar methodological challenge in their respective academic fields. As a rhetorician, Hocutt argues that while “posthuman and new materialist theories provide a methodological framework for understanding” emergent assemblage activity as rhetorical agency, that rhetorical code studies requires more empirical methodologies to unpack how algorithms persuade us by acting as “gate-keepers” to knowledge, even though patents and other issues keep algorithm code from being entirely “known” to us (Algorithms as Information Brokers). As a social scientist, Introna is also concerned with the “inscrutability” of code, but in the context of how the sheer scope of source code and its influence on every aspect of modern life life makes it impossible for sociologists, cultural anthropologists, and media studies scholars to account for all the effects of an algorithmic assemblage.
The methodological dilemma posed by the “assemblage” model is that if algorithmic action only has meaning in its temporal, embedded, relational unfolding, it is hard to analyze its effects. Introna compares algorithmic iterations to playing a song on a musical instrument:
The ‘playing’ of music is not in the enactment of individual notes but the temporal flow of the notes being played—each note inheriting from the prior, and imparting to, the subsequent note, or figure, motif, cell, phrase, or melody. Certainly, each specific line of code (or note) is important and needs to be correct in its own terms. However, the relational temporal flow is what enacts the ‘‘sort’’ or plays the melody of music” (23)
To extend this metaphor, we can analyze a single note or a melody from a symphony, but this compromises our ability to evaluate harmony. On the other hand, again, the sheer scope of algorithmic systems and the importance that they hold in our world requires “cutting out” part of an algorithmic assemblage for analysis.
“The reasons to cut this temporally unfolding sociomaterial assemblage or process at any point—that is, to make a particular determination of what specifically is being done (such as, comparing, swapping, sorting, administering, managing, etc.)—might be described as technical (e.g., to write the code), it might be described as social (to enact an administrative practice), and so forth. The point is that any such cut in the temporal flow of doing can be made, more or less arbitrarily, to enable us to determine what is ‘‘being done” (Introna “Governance”23).
In the above quote, Introna writes that the object of something “being done” may “described as technical” or “might be described as social” despite previously disavowing a distinction between technical/social, or source/effect on the previous page. The implicit point is that just as Introna separates “technical” and “social,” “writing” and “enacting a practice” in this sentence for the purpose of clarity, so can the scholar make conceptual splits when trying to account for what kinds of calculations are being performed on our world.
For these reasons, the work of Karen Barad on “agential realism” has been influential on scholars in such fields including critical algorithm studies and rhetoric. Barad introduced the concept of agential realism as an alternative to positivist science and social constructionism. Central to both critiques is her rejection of “representationalism,” a term which here refers the belief that there is a distinction between a representation and the thing that is being represented that existed prior to representation. Barad draws on her own background in quantum physics to argue that matter not a “substance,” but intra-activity.” to produce her theory of performative matter.
At this point, let’s turn to Barad.
Karen Barad 101
Drawing from Butler, Barad argues that relata do not pre-exist relationships. Phenomenon crystalize into “things” (thingification) through the establishment of diffractive semantic and ontological “boundaries.” For this crystallization, or “thingification” to occur, Barad writes that “apparatuses enact agential cuts that produce determinate boundaries and properties of ‘entities’ within phenomena” (emphasis mine) (Meeting the Universe Halfway 148).
Let me elaborate. What Barad calls “discursive practices” are not (only) “linguistic expressions” of human language, but extend to other material “specific practices of intra-action…entangled with other intra-actions”(PP 822. Discursive practices “entangled with other intra-actions”(PP 822) “produce material phenomenon” (PP 820) through practices of boundary-making called “agential cuts.”. Barad defines ““intra-action” here as “the mutual constitution of entangled agencies” (Meeting 33) that make up “the material apparatus” in which all discursive practices occur. She defines “agential cuts” as what allows “objects” to (be)come by differentiating phenomenon. Yet such “differentiation” always occurs in relationality. For example, the agential cuts that distinguish phenomenon include separating objects as perceivable “different” “objects” held apart in space and time.
For example, phenomenon crystalize into “things” (thingification) must be “positioned” in spatial, temporal proximity to one another to be apprehendible by the “subject” (also positioned in spatial, temporal relation). Even the dichotomy between “subject” and “object” dichotomy depends on a specific spatial, temporal proximity. As Barad writes, “the boundaries and properties of component parts of” any “phenomenon” that emerges “become determinate only in the enactment of an agential cut delineating the ‘measured object’ from the ‘measuring agent’” (Meeting 337). In short, it is necessary for the concept of separability or difference to be enacted, however performatively, as a condition for apprehension of reality.
The agential cut that enacts this boundary is agential separability — an agentially enacted ontological separability within the phenomenon” (Meeting the Universe Halfway 175) that makes the distinction possible. “Separability” is not “absolute” (339 Barad writes, given how deeply “entangled” intra-acting discursive practices are. At the same time, agential separability is a precondition for ontology precisely because of how deeply “entangled” intra-action is.
Barad’s agential realism posits phenomenon are the ontological inseparability of agentially intra-acting components” (Meeting 33) She writes elsewhere in Meeting the Universe Half Way that “discursive practices are boundary making practices” in general “have no finality in the ongoing dynamics of agential intra-activity” (149). Presumably, agential cuts have no finality within the material apparatus that enacts it. It is the constant iteration of change and action, or self-constitution of material phenomenon in flux, that Barad generations her theory of performative matter.
It is the way agent cuts move through intra-action of/in apparatuses that enact the discursive practices that “divide” “distinguish” and “relate” “objects.” Barad posits that matter is performative, the way Butler describes gender as performative, that these cuts are enacted constantly in such a way forms the perception of a way of being (matter itself, gender subjectivity) in ways often according to discursive boundaries but by way of constant, unintentional reiterations of “the laws” that are nonetheless unbound to completely perceptible “rules.”
(If you thought Butler’s performativity referred to a “performance,” stop trusting Tumblr to give you theory).
To return to Barad’s theory of matter as performative, I repeat: for her, matter is performative because it (re)produces itself through the effects of its own intra-active dynamic material relationships.
“It is only through specific agential intra-action that the boundaries and properties of ‘components’ of phenomena become determinate and that particular articulations become meaningful. In the absence of specific agential intra-actions, these ontic-semantic boundaries are indeterminate. In short, the apparatus specifies an agential cut that enacts a resolution (within the phenomenon) of the semantic, as well as ontic, indeterminacy. Hence apparatuses are boundary-making practices” (Meeting 14)
It should be repeated that in Barad’s theory, that apparatuses are boundary-making practices for phenomenon with no true “interior” or “exterior.” This is why Barad uses the phrase “intra-activity” instead of “inter-activity” to indicate that discursive practices in play are apparatuses enacting agential cuts on/within itself. Barad’s agential realism posits phenomenon are the ontological inseparability of agentially intra-acting components” (Meeting 33)
What Barad calls “discursives practices” are not (only) “linguistic expressions” of human language, but extend to other material “specific practices of intra-action…entangled with other intra-actions”(PP 822. Discursive practices “entangled with other intra-actions”(PP 822) “produce material phenomenon” (PP 820) through practices of boundary-making called “agential cuts.”. Barad defines ““intra-action” here as “the mutual constitution of entangled agencies” (Meeting 33) that make up “the material apparatus” in which all discursive practices occur. She defines “agential cuts” as what allows “objects” to (be)come by differentiating phenomenon. Yet such “differentiation” always occurs in relationality. For example, the agential cuts that distinguish phenomenon include separating objects as perceivable “different” “objects” held apart in space and time.
She writes elsewhere in Meeting the Universe Half Way that “discursive practices are boundary making practices” in general “have no finality in the ongoing dynamics of agential intra-activity” (149). Presumably, agential cuts have no finality within the material apparatus that enacts it. It is the constant nature of the iteration of change and action, or self-constitution of material phenomenon in flux, that Barad generations her theory of performative matter.
Barad’s Ethical Considerations:
In conclusion, she notes that agental cuts have ethical consequences. Because “different agential cuts produce different phenomena” (Meeting 175), these agential cuts make possible different ways of becoming/being, different possibilities for the future, and establish “what matters” in multiple senses of the phrase. To return to Barad’s claim that “being” is “not a static relationality but a doing” (“Posthumanist Performativity” 803)” “that always entails constituting exclusions” (“Posthumanist Performativity” 803),” engagement with “world-making” discursive practices must not only account for where and how agental cuts are made but also for their ethical stakes. We are all connected to each other: what really divides and separates us?
Barad is firm that her rejection of individual human agency neither diminishes does not absolves us of responsibility to intervene in the discursive “world-making” practices.
“We are responsible for the cuts that we help enact not because we do the choosing (neither do we escape responsibility because “we” are “chosen” by them), but because we are an agential part of the material becoming of the universe. Cuts are agentially enacted not by willful individuals but by larger material arrangement of which which ‘we’ are a ‘part.’ The cuts that we participate in enacting matter” (Meeting 178).
Karen Barad in Rhetorical Algorithm Studies:
Tarleton Gillespie writes, we often “find ourselves more ready to proclaim the impact of algorithms than to say what they are (“Algorithm”18). For one thing, the term “algorithm does not have a generally accepted formal definition (Wikipedia “Algorithmic Characterization”) even in mathematics, where “the notion is expanding” to the point where it “cannot be rigorously defined in full generality (Gurevich “What are Algorithms?”).” In computer science, attempts to formalize the term through methods of “algorithm characterizations” (attempts to formalize the word algorithm by their various kinds, characteristics, and functions) often struggle because the various formulas, procedures, or forms of computation we call “algorithms” share no fundamental formal qualities exclusive to the set, “no shared logic of algorithms (Neyland “Organizing Algorithms” 134).” Like viruses, it can be easier to identify a software algorithm by “process of elimination” (what is “not” an algorithm) or by perceived intent of the programmer (in the case of the algorithm, “to solve a problem”) than any stable ontological category that necessarily defines and limits the term.
For another thing, it is impossible to abstract a “single” algorithmic operation outside the technological relational environments “it” interacts with(in) because algorithms are both participation in this setting and mediate it. Even the most formulaic definition of “algorithm” cannot be abstracted from cultural logics. For example, Bernhard Rieder writes that in Knuth’s class definition of algorithm as “‘a finite set of rules which gives a sequence of operations for solving a specific type of problem,” Knuth does not specify what constitutes a “problem” (or who decides) and or clarifies what action that ‘some’ modifies in ‘some action to be performed or some decision to be made, or by who or what (Rieder “PageRank”); non-trivial questions in distinguishing accountability from responsibility. Similarly, for cultural anthropologist Nick Seaver, even the current standard textbook definition of algorithms as “a sequence of computational steps that transform the input into the output (Cormen qtd. Seaver 1)” depends on external “inputs” to operate and “translates,” or really, transforms information through computation. As processes of transformation and transmediation “only have an influence when they are running (McKelvey “Algorithmic Media Needs Democratic Methods” 599),” algorithms are temporally “entrenched” (Roberge and Seyfert “What are Algorithmic Cultures?” 2), unfolding and embodying (within) time (Miyazaki “Algorithmic Ecosystems” 129). Thus, according to Jonathan Roberge and Robert Seyfert in their introduction to the essay collection Algorithmic Cultures: Essays on Meaning, Performance, and New Technologies, algorithms are arguably best conceived as action, not in the sense of action as “something purposive and straightforward,” but action as ‘outward expansion’ “producing numerous outputs from multiple inputs” in ways open to human and nonhuman interruption regardless of a “single” algorithm’s programming purpose or the programmer’s intentions (Roberge and Seyfert 3).
Both Roberge and Seyfert’s introduction and the body of social science interdisciplinary work represented in the collection draw upon Lucas Introna (2016)’s notion of “algorithms as performative.” To see algorithms as performative goes further to see algorithmic “actions” as iterative “citations” within a knowledge-production material discourse that characterizes algorithms as having an “ontology of becoming.” For Introna, the algorithm’s ontology of becoming occurs with/through certain ‘‘mutually constitutive nature of problems, domains of knowledge, and subjectivities enacted through governing practices’’ (“Algorithms, Governance, and Governmentality”) that include “algorithm” as a domain of knowledge, as enacting nature of the “problems” (that they are programmed to solve, that they supposed identify or pose, etc.), and even production of subjectivities. In other words, algorithms “emerge” as “objects of interest” (Gillespie and Seaver Social Media Collective x]) in multiple discourses alongside and through the emergence of other cultural objects, such as “the digital,” “the material,” and even “the human,” as well as alongside notions of “agency” “inscrutability” and “normativity” that culturally ascribed to algorithms “themselves.”
In conclusion, algorithms are not abstracted machines, but, by raison d’etre, activity embedded in the social prior to any computational process. They are “knowledge, ideas, skills, tools, methodology, habits, and values that permeate practices embedded in layers of social organization, cultural configurations, economic rationales, and political struggles (Rieder “PageRank)” and are “part of a broader array of performativities that includes, for example, rituals, narratives, and symbolic experiences (Rauer “Drones:The Mobilization of Algorithms” 142). Their ontology of becoming participates in the development of the epistemological frameworks that make them “tactible,” just as processes of sorting, organizing, circulating, predictive modeling and content production, content production through prediction (Napoli “On Automation in Media Industries”), their sorting of different publics, and their embedded “values” (what “problems” they solve and why) position them them as both interpretative frameworks and interpreters. Their practices and “embeddedness” within other practices come to “moralize” us (Pfister “Animalizing Technologies in a Posthumanist Spirit” 223) and manipulate, or more gently, persuade us through their ambient presence in our lives.
Barad’s “agental cut” allows rhetoricians to account for how actions and ways of being become “legible” within an algorithmically-mediated rhetorical ecology. Given the difficulty in identifying a “single algorithm” as a thing rather than a protocol, a process, an enacting procedure, understanding algorithmic processes in terms of activity and intra-activity within a digitally mediated apparatus helps us account for quote on quote “what we colloquially call algorithms” filter, produce, and make legible facets of our world in ways that are not always straight-forward. The ethical imperative, then, of adapting Barad’s work to critical algorithm studies (as many interdisciplinary scholars have), is that we ask questions like who is made visible? What processes are being enacted at what consequences? Who is effected and how (intentionally or none intentionally?). We ask these questions knowing the performative nature of algorithms, and, indeed, the apparatuses that structure our lives mean we must constantly revaluate these ethics without expecting a “final” solution.
Barad, Karen. Meeting the Universe Halfway: Quantum Physics and the Entanglement of Matter and Meaning (Durham: Duke University Press, 2007.