Page 1 →Introduction Locating the Energies of Computational Performance
Dated to the high Middle Ages, the Lokapannati, a text written in Pali (a sacred Buddhist language), tells a tale of steampunk robots, reincarnation, and technological power.1 In the fantastical story, King Ajatasatru solicits the help of an engineer, knowledgeable about the Roman tradition of making self-moving machines. With the engineer’s assistance, he creates a set of robot guards to protect a collection of Buddhist relics stored in a stupa (an underground hollow covered with a clay dome). According to popular interpretation, King Ajatasatru did this because he needed agents that could move as if they were alive, but who would also remain “unattached” to the relics, so they would be available when the rightful inheritor—King Asoka—came to collect. Though they could not speak, Ajatasatru’s robots were treated as “spirit movement machines” infused with their own anima (wind of life), a feat allegedly well-guarded by the Romans as a trade secret.2 The story of King Ajatasatru’s robots is informative in that, despite lacking speech, the robots communicated. Their movements embodied nodal, ambient echoes of the cultural moment in which they were being witnessed. While the ancient Romans might have gazed at the robots as feats of “spring-and-compression-propelled machinery,” those same moving machines, for an ancient Indian Buddhist, might have intuited a discerning removal of the problem of “becoming attached.”3 King Ajatasatru’s robots, while not alive, nonetheless moved in lively ways—they were performing. As such, they enlivened discourse by imbuing it with more-than-human rhetorical energies, which interacted with culturally shaped bodies, manifesting force not just as ideas, but as feelings.
Flash forward to today. We find similarly influential communication in computing machines. But they are not the stuff of legend. Web applications Page 2 →offer us conclusions about science. Twitter bots generate art. Machine-learning systems make fun of politicians. We live in an era where a substantial share of our public communication is machinic.4 Like the robots of King Ajatasatru’s caves, our modern computing machines cannot speak for themselves per se, but they do exert energies, which can impact persons viscerally, preconsciously. Current scholarship, in the spirit of rhetoric of science, technology, and medicine, and that field’s drive to critically engage the monolith of “objective” communication, is turning attention to investigate how software can constrain, or open up, possibilities of rhetorical invention and how we might participate with automated systems to make meaning. And as a consequence, the current literature productively complicates the idea that automated technologies are neutral, passive tools, proposing instead that machines are better understood as ethically loaded, active agents (or actants) helping to comprise the assemblages of rhetorical processes. Not yet pursued at length in the current literature are the rhetorical energies that attend the performances of computing machines as they give advice, entertain, and proffer insight, speaking to human concerns in more-than-human ways. This book is an effort to better capture the opportunities and threats of influence inherent to dynamic computational media—media that can perform for users by carrying out operations “on their own.” Over the course of the book, I will build out a conceptual framework useful for naming, analyzing, and evaluating the physical and emotional energies imbued within computational performances.
While applying the idea of rhetorical energy to the performances of web applications, bots, and voice-based interfaces is the unique contribution of this book, the idea of rhetorical energy itself is not new. George Kennedy first proposed that we think of rhetoric as energy in his 1992 essay, “A Hoot in the Dark,” in which he explores what he calls a “general rhetoric,” or a rhetoric that is not moored in the language, rationality, and intention that we so often equate with human activity, but rather more generally in the energies that impact bodies, including nonhuman ones.5 In his words: “Rhetoric in the most general sense may perhaps be identified with the energy inherent in communication: the emotional energy that impels the speaker to speak, the physical energy expended in the utterance, the energy level coded in the message, and the energy experienced by the recipient in decoding the message.”6 Defined as such, we can recognize that humans are not the only ones that influence others. Flowers communicate desirability with sweet smells, deer communicate position within their social order with displays of strength, and crows sound calls Page 3 →to bring the flock together in congregation. An important point that Kennedy drives home in his essay is that rhetoric is not exclusive to humans, and, perhaps more profoundly, rhetoric is not fixed solely in the verbal or the rational. Rather, at its base, rhetoric is moored in emotional and physical energies as they are expended and experienced, shaped by biology and culture (including that of nonhuman animals).
Though Kennedy’s original discussion of rhetorical energy is interested in locating the energies shared between specific living species (e.g., the suasions of crows), the idea of rhetorical energy is useful for thinking through the communication of computing machines. In the same way that Kennedy pointed our attention beyond human communication to learn more about the rhetoric of nonhuman living things, we can take it a step further by interrogating the rhetorical energy of things that are lively but not alive, such as computational performances. In fact, Kennedy probably did not have computing machines in mind, but he indirectly makes a connection between rhetorical energy and electrical current in his joking proposal of rheme, a unit “analogous to an erg or volt,” with which the “experimentalists” could measure rhetorical energy.7 Kennedy’s passing statement demonstrates a broad-thinking quip, acknowledging the sometimes hard divides between social scientific and interpretive approaches to scholarship, but it also expresses the possibility of rhetoric in configurations not just in animal–animal scenarios, but also in animal–nonanimal scenarios. The key here is in the expenditure and experience of energy. Computing machines, as they run processes, flip switches, create states on a drive, and retrieve memory are expending energy—they are performing as computing machines. And these performances affect us.
As will be discussed in more detail later in this chapter, orienting to the rhetorical energies of computational performances is a means of thickening accounts of procedurality, something that many would identify as the defining characteristic of computational media. In procedural rhetorics the analyst accounts for procedures—the mechanics and decision trees of a given piece of software—in an effort to “read” the argument that it encourages onto users not through words or visuals per se, but through the processes of a given software. Procedural approaches, while often yielding profound and useful insights, also tend to emphasize logic at the expense of understanding other aspects of rhetorical influence in computing, such as feelings. This book offers a necessary thickening of procedurality by turning attention to the visceral “more” that attends the movements of computational performances, which are surely Page 4 →punctuated by procedures, but which also influence at a gestural level, shaped by a deep complex of affective/cultural layerings, accessible through the lens of rhetoric as energy.
Exploring the rhetorical energies of computational performance is an activity that joins arms with other scholars who are working to expand assumptions about what rhetoric is and how it occurs by looking beyond (human) intention, rationality, and verbally based discourses.8 Such is a move away from rhetoric as epistemic and toward rhetoric as ontological, a distinction that Scot Barnett and Casey Boyle elaborate in their edited volume, Rhetoric, Through Everyday Things: “Whereas epistemology emphasizes knowledge about things, and thus about their meaning and cultural significance for us, ontology stresses relational being.”9 Integral to Barnett and Boyle’s framing of rhetoric as ontological is an emphasis on embodied experience as it happens between humans and nonhumans, which pushes not only where one might look for influence, but also how influence happens in the first place. For example, in approaching rhetoric as ontological one might push on Aristotelian commitments to locating rhetoric as formal arguments or defining rhetoric as an exclusively human enterprise—to ask, like John Muckelbaurer does, whether “plants turning toward the sun and audiences accepting an argument might well involve the same kind of action/motion,” complicating distinctions between symbolic and natural processes.10 Approaching rhetoric as energy, moreover, is to follow Muckelbaurer in his provocative proposition to consider rhetoric as “heliotropic,” a move that blurs the Burkean distinctions between (symbolic) action as “behavior possible to a typically symbol-using animal” and (natural) motion as “the extrasymbolic or nonsymbolic operations of nature.”11 What conceptualizing rhetoric as energy does in this sense is push toward rhetoric as ontological by locating influence beyond mere symbolism and its attending prioritization of rationality, and instead locate it between action and motion, within the expenditure and experience of energy “at work,” illuminating influence in realms where we might otherwise be reluctant to concede it. Computational performances exist in our homes, the workplace, and in the public, touching nearly every aspect of our lives. This book will show that they influence in ways that are subtly powerful, for their lively (but not alive) movements complicate common assumptions regarding symbols and sensation, artificiality and naturality, and subjects and objects.
By the end of the book, I hope that the reader will come to understand that people can “catch feels” from the energies of computing machines.12 The case studies in particular will demonstrate that computational performances Page 5 →are attended by influential energies that can be leveraged to catalyze sublime artistic experiences, generate false confidence in pseudoscience, and encourage compulsive reactions in politics. In an era where we increasingly communicate with our machines, rather than through them, it is paramount that we understand the energies that attend their performances. In recognizing the rhetorical energies of computing machines, we can begin to appreciate their visceral but deep suasions as well as to account for the ethical implications of their lively performances. Because people can (consciously or unconsciously) deploy the energies of machines to influence, diving deep into computing’s affective / cultural layerings is not simply fun or interesting—it is integral to updating our critical sensibilities of the world.
The Rhetorical Energies of Computing Machines
Voice-based assistants, chatbots, art-making machine-learning systems—our contemporary machines can put words together, respond to their human counterparts, and generate imagery. Even in such examples, where machines are obviously communicating, some may harbor a reservation in calling it real communication and think of it as quasi-communication.13 Such is a common view within the burgeoning field of human–machine communication, an interdisciplinary area of research, which focuses on the implications of machine communicators (such as chatbots or voice-based assistants). For example, as Leopoldina Fortunati and Autumn P. Edwards, editors of Human-Machine Communication, explain in their introduction to the inaugural issue of the journal:
Humans are aware that media agents and social robots are quasi-interlocutor, quasi-communicator, quasi-social, but they play the game and pretend to really communicate and to have social relationships with them. . .. The strength of this illusion depends on the simulating ability of the media agents. Although it is an illusion, and is even consciously recognized as such by the people involved, it can generate all the same feelings of communicative and social satisfaction as interactions with other humans.14
The specific contributions that a rhetorical energies approach offers to human–machine communication will be discussed in more detail in the conclusion of the book. For now, we note that voice-based assistants, or chatbots, or even machine-learning systems (which can be much more autonomous in their behavior) are often conceptualized as miming symbolic interaction, for they are Page 6 →not alive like persons are. This is hard to deny—talking to a computer is talking to a computer and not a person. The computer does not have feelings or stakes in the world—it just drives forward, carrying out its scripts, and reacting to the input of a human (or another machine). And so, computing machines might be understood as agents mindlessly miming symbolic interaction, rather than engaging in real communication.
However, the “quasi” in quasi-communication is loaded with the assumption that real communication is derived from the human. What we overlook in our search for the human is an entire world of energy that matters, buzzing in the background, affecting us, covered over by a narcissistic hubris, wherein homo sapiens (the “wise” ones) look only for the voice, performance, and words of other humans as if everything else is The Nothing. While they might not have stakes in the world, or even experience feelings themselves, machines move us. And, surely, if we are looking for symbolic interactions between persons, we are forced to imagine such communication as quasi.
Yet if we adjust our definition of rhetoric, say, by following Kennedy in his proposal to think of rhetoric not as based in language or argument, or even as solely the enterprise of humans, but rather as “the energy inherent in communication” the persuasiveness of machines as machines begins to emerge.15 Rhetorical energy can include language, but it also includes such things as “physical actions, facial expressions, gestures, and signs generally.”16 What makes Kennedy’s contribution particularly distinctive is his interrogation of the rhetorics of nonhuman animals, which are not located in language per se, but rather in performances. Bucks, demonstrating their rank in the deer social order through public displays of strength—that is rhetorical energy. Flowers making appeals to honeybees with sweet smells—that is rhetorical energy too. The crux of rhetoric as energy is that it locates influence not in language but more broadly in the energies that are “at work” in the interaction between entities. Relocating influence in this way, as Chris Ingraham elaborates, jaunts the analyst into “the biologically hardwired realm of pre-intentional survival mechanisms, those that are communicated through a perceivable code, yes, but not necessarily in symbolic messages governed by a semiotic regime of meaning.”17 Likewise, Catherine Chaput and Crystal Broch Colombini center rhetorical energy as concerned with “the physiological urge to defend, sustain, and thrive . . . our most taken-for-granted modes of engagement.”18 Put differently, while our visceral responses to the world—and the other entities that imbue it with energy—might be “read,” they are also nonconscious, born of entanglements of biology and culture. Although Kennedy is largely concerned with the Page 7 →prospect of locating rhetoric in the realms of nonhuman animals—entities that are alive—it does not take a large leap to consider that machines, as they pulse electricity and carry out processes, are expending energies, beyond words, enlivening discourse in more-than-human ways. This project aims to continue to explore what it means to think of rhetoric as energy by expanding its scope to account for the energies of lively (but not alive) communication, such as that of computational performances.
By bringing the expenditure and experience of energy in computational performances from the background into the foreground, we can begin to appreciate the something “more” that attends computational performances, or instances of persuasion that leverage the movements of machines. Take, for instance, table 0.1, which compares examples of traditional modes of (human) delivery with computationally performed ones. (Note that the contents of this table also track with the case studies of the book.) The Twitter bot and vaccine web application examples stem from my earlier, less-developed work on “machinic rhetorics.”19 All of the examples in the table are mapped over what Leah Ceccarelli has articulated as the “ends” of rhetoric: the aesthetic, the political, and the epistemic. Important to note is that focusing on the rhetorical energy of machines is to take up concern for the ontological as it regards these three ends of rhetoric.20 While machines might not be persons engaging in symbolic interaction with other humans, they might nonetheless add something persuasive as machines, located not in stylistics or lines of argument per se, but rather in the energies that attend the performance of computing “at work” as they carry out processes (like calculating and/or lampooning a president).
The idea that machines can add something to rhetoric can be difficult for some people to accept. Helping us understand this is Carolyn Miller, who, in her 2007 study of the rhetorical abilities and effects of machines, shared a description of a fictional speech grading system—replete with the catchy name, AutoSpeech-Easy—with other teachers of rhetoric.21 Miller’s hypothetical AutoSpeech-Easy is described as being able to track not just the language of students’ speeches, but also their nonverbal movements and intonations—the energies of their performances. Miller, in her proposition of the hypothetical technology, clairvoyantly projects the robust and steadfast developments of such technologies as facial recognition, motion mapping, and sentiment analysis, at a time in which these technologies were only reaching viable marketability. The fact that these technologies are ubiquitous today underscores the power and durability of her contribution. Based on survey responses from teachers who were asked to consider the use of the technology to grade student Page 8 →speeches, Miller notes a qualitative uneasiness in their reactions having to do with the reluctance to attribute agency to computers. In particular, she notes that the act of “speaking strongly resists automation because we understand it intuitively . . . as a performance, meaning that it is dynamic and temporal, that it requires living presence.”22 Furthermore, for the respondents, speaking to AutoSpeech-Easy as opposed to a human audience creates a communicative action vacant of “an Other, someone who may resist, disagree, disapprove, humiliate—or approve, appreciate, empathize, and applaud”—it is devoid of the necessary sharing of connection and relational-being-together between living agents.23
Rhetorical Ends | “Traditionally” Performed | Computationally Performed |
Epistemic Argument | A brochure that details the aluminum contents of various vaccines | A web application, styled in the likeness of an “expert” system, which invites input from the user to detail the aluminum contents of various vaccines |
Aesthetic Experience | A poetry project, which sources US Census data to tell vivid stories of persons, published in a book | A Twitter bot that will run for the next 1,700 years, sharing a biographical statement for a real person every hour on the hour |
Political Critique | Alec Baldwin’s parody of Donald Trump on Saturday Night Live | An autonomous machine-learning system, trained on Donald Trump’s public speaking transcripts to perform his “persona” |
Extended Description
Table that describes examples of person-performed and computationally-performed rhetorics across the epistemic, aesthetic, and political ends of rhetoric.
Miller counters with the suggestion that machines can (and perhaps should) be conceptualized as rhetorical agents at least insofar as they might represent imitations that can invite attributions of (human) agency. Specifically, she notes Page 9 →that rhetorical agency exists “exactly between the agent’s capacity and the effect on an audience” in what she designates as the “kinetic energy of rhetorical performance.”24 If machines can perform like humans then, by consequence, they can also invite the attribution of agency, represented by the kinetic energies of their rhetorical performances. By kinetic energies, she means the potential energies associated with habits of mind and being that we attribute to human behavior and expectation, scripted, animated, put into motion. Because the movements of machines can instantiate performances of living (just like people do), their status as rhetorical agents is illuminated. But as Miller is careful to note, “we understand agency as an attribution made by another agent, that is, by an entity to whom we are willing to attribute agency. It is through this process of mutual attribution that agency does, indeed, produce the agent.”25 Consequently, she concludes: “Do we owe such acknowledgment, such agency-granting attribution, to automated assessment systems? Right now, I suspect that most of us agree that we do not, and moreover that out of respect for our students we should not ask them to make such attributions either.”26 To use AutoSpeech-Easy to grade student speeches, moreover, would be to revoke the attribution of agency—both as capacity and effect—from the students asked to speak to it, by omitting the something more found in the relations between persons: the emergence of one’s own, and others’, agency through mutual attribution. Consequently, Miller demonstrates that although machines might be able to perform the kinetic energies of humans, they nonetheless are not alive, and so we might harbor a reluctance to attribute agency to them—either as effect or capacity.
But even if we are reluctant to attribute agency to machines, they continue to impact our interconnected, increasingly automated discourse ecologies.27 While machines may not be alive, they nonetheless participate in communication, whether or not we grant them agency. Krista Kennedy makes this point through archival examination of Wikipedia articles while shedding light on the machines that human actors collaborate with to compose content on the Wiki platform.28 Namely, she poignantly describes the compositional process of writing articles about small towns, which in traditional encyclopedic writing were often omitted due to labor constraints. However, with the development of bots, such as Rambot (named after its creator, Derek Ramsey), a software-based agent that sourced raw US Census data to write thousands of articles a day about small towns, we find an impact on the very compositional trajectory of Wikipedia. From there, humans (and other bots) could interact with those articles, adding their contributions and revisions, exposing that even if we Page 10 →refuse to attribute agency to machines, they nonetheless plug away, performing the work of writing. And this is exactly Kennedy’s point: By studying the compositional processes of Wikipedia articles, she teaches us that writing, an activity that we often wish to attribute to humans, is something that emerges between the human and the machine. Kennedy explains: “These bots react to their environment, initiate action with it, and affect change both within the texts and sometimes within the broader scope of the project, as when Wikipedia rather suddenly expanded exponentially to cover thousands of towns.”29
The power of Kennedy’s contribution is that even if we refuse to attribute agency to machines, they nonetheless continue to generate articles and police the writing of other authors—they are members of the discourse ecology from which the genre of Wikipedia articles emerges. Surely, creation is an enterprise that we wish to attribute to human ingenuity, but Wikipedia is a creation that would not exist as it does today without the contributions of its bots. The reality of this is that, as much as we might be reluctant to acknowledge it, humans are not the only ones “choosing the right words.” Furthermore, and as this book hopes to show, the performances of computing machines also instantiate suasions that go beyond words, for they are attended by rhetorical energies—movements and processes that speak beyond words and from beyond the human.
Scaffolding our understanding of rhetorical energies as occurring between humans and nonhumans is Debra Hawhee, whose Rhetoric in Tooth and Claw offers a virtuoso reading of classical rhetorical texts, informed by George Kennedy’s rhetorical energy. Specifically, she zeros in on (non)human animals as they appear in discussions of language to draw out the relationship between language and sensation. “The snapping of the curs, the crying children, the rushing lion, the leaping puppies,” Hawhee says, “are all doing something with intensity of feeling; they embody and convey pathos, and that embodiment constitutes their likeness to the human animals being described.”30 Hawhee tracks the distinction that Aristotle makes in De Anima between dunamis (potential energy) and energeia (kinetic energy) into his Rhetoric. For Aristotle, she maintains, the invocation of movement—kinetic energy—is what elevates metaphors from the level of a thought to an experience, a “bringing before the eyes.”31 Animals, and especially nonhuman animals, bring “more” than just ideas to language: they bring energy, which activates the body, adding sensation, exposing the abilities of animals to “enliven the imagination.”32 Rather than using metaphors to designate the movements of humans, these metaphors draw on the embodied experience of animal movement, further supporting Page 11 →the claim that the rhetorical energies of nonhumans can be leveraged to communicate in ways that go beyond the human. Where the moving snarls of tigers or the quiet hops of rabbits are instances in which the energies of non-humans elevate lingual metaphors to the status of bringing “more,” computing machines similarly add sensations. And surely this can happen in lingual constructions that draw on the movements of machines. For instance, “processing,” or even a joking, “error 404, file not found,” when a colleague asks when the next meeting is scheduled, are metaphors that imbue the energies of computing machines. However, computational media—web and mobile applications, bots, video games, voice-based assistants—are unique in that they are distinctly characterized by energies of dynamic movement across the senses. Their performances are not moored solely in the verbal, or the visual, but also in sound and touch—they are multisensorial.
One’s phone, buzzing in their pocket, alongside a cleverly designed ding is a case in point: the energies of computing machines move across sensory media, finding expression in other senses beyond sight, such as touch and sound. Casey Boyle, James J. Brown, and Steph Ceraso offer the metaphor of transduction to conceptualize an approach to rhetoric that is willing to explore not sight, or any other single sense, but rather the multiplicity of senses that span a wide sector of the digital sensorium that characterizes daily life. In their words: “Transduction refers to how a signal moves across disparate registers of relations: neural firings move to fingers to perform keystrokes that then transform into electrical charges that then become digital bits and are delivered to a screen by software or saved to a hard drive that becomes transcoded again whenever someone opens a file.”33 Accounting for the multiplicity of energies of a given computational performance, manifested as converging and diverging light, sound, and/or physical movement, is a means for accounting for the digital sensorium; not individual visuals added to sounds and touches, but rather as a more holistic energy that attends the convergence of those stimuli transduced (converted) into the body. As multisensorial complexes of movement, the performances of computing machines, while not alive, can be approached as lively.
To regard computing machines as lively, rather than alive, is to recognize that computing machines do not enact agency in the sense of an invested “mind” influencing another. However, they are attended by energies in the form of movements entangled with the concerns of humans, animating communication in distinctly nonhuman, nonverbal, and visceral ways. For example, as I have written about elsewhere, the “steam engine,” as it bellowed steam and moved rods, “echoed the spirit of science, come to save us from our Page 12 →human shortcomings,” and the flipping of switches and the flashing of neon bulbs of early computers moved with the reassuring (alarming?) character of “mechanical Frankenstein’s monsters, birthed from the work of groups of mathematicians, scientists, and engineers recurrently visiting a problem in order to systematically remove the human from the idea in order to inscribe it into the program of a given circuit.”34 Approaches to rhetorical agency that orient toward humans, and words, and arguments will surely find these things in computing machines. However, such approaches also problematically overlook those instances of impact that move beyond human control, verbal articulation, and argumentative rationality, wherein the energies transduced are not merely artificial or natural—they are an amalgamation of both. People make computing machines, but computing machines move on their own.
To approach the agency of computing machines this way is to follow Karen Barad in taking up an agential realist approach, which recognizes that mattering (in the sense of making meaningful) and matter (in the sense of the stuff of physics) are entangled with respect to agential outcomes.35 The feelings and ideas of humans, moreover, are not the results of direct human action, but rather the results of intra-actions between humans and nonhumans, and even nonhumans and nonhumans, including matter, catalyzing outcomes.36 Within this framing, rather than look only for the interactions between a human and another human, mediated by a chatbot, for instance, one can also look for the intra-actions in which the chatbot emerges as an active participant, attended by catalyzing energies. A chatbot may be made by a person. But a chatbot moves on its own, representing not just the (human) preprogrammed joke that “Pineapple does not belong on pizza,” but also the (machinic) movement of code—the energy of mathematics incarnate—a human concern, expressed in a more-than-human way. The mattering of matter in the chatbot, put differently, is not merely the outcome of thoughtful, eloquent humans; the chatbot is also moving.
Accounting for the energies that attend computing machines is to enact what Barad calls a posthuman approach to performativity—not meant to reject human impact, but rather to add nuance to how mattering is negotiated between bodies (human and nonhuman).37 Particularly productive is Barad’s advice to focus on performance, rather than words, as a means for capturing the impacts of the nonhuman. As Krista Kennedy succinctly articulates: “Conceptualized as performance rather than as inherent capacity, agency is not a property that a subject can possess . . . but rather it arises through response to a situation composed of parameters beyond the control of any single actor.”38 Page 13 →That is, rather than focus solely on the salty and sweet words of the chatbot, a rhetoric as energy approach to computing machines would pay particular attention to the sorts of intra-actions of energy taking place in the performance of the chatbot as it draws on traditions of chatbot design and electricity to move according to its scripts, illuminates pixels, and chimes with “Bleebloop” sounds. While machines may not be invested minds, making appeals to others, they do instantiate performances, intermingling in the wider discourse ecology, moving on their own, reacting to situations, imbuing them with energies.
To explore the rhetorical energy of machines, then, is to drive at the multisensory, visceral “more” that attends computational performances. Rhetoric as energy does not necessarily displace verbal eloquence or logical rationality. Those aspects of rhetoric remain. However, in conceptualizing rhetoric as energy, one’s focus of analysis is made broader, zooming out to account for the background feelings and visceral responses of bodies to sensory stimuli, which might be hidden in ambience, but nonetheless are moving (both in the sense of physical movement and in the sense of impacting bodies). We need to be willing to go deeper than asking how machines may, or may not, mime humans, and take seriously the idea that, despite being nonhuman and nonanimal, they also bring lively movements as machines.
Beyond the Front and the Back Ends of Computing and Toward the Deep End
In computing, there are two general realms, divided by a metaphoric expression describing that which is visible and that which is hidden from view.
The front end: The realm of computing that deals with the user interface (i.e., recieving input and giving output).
The back end: The realm of computing that deals with the databases, functions, and networking from which a given program operates (i.e., information processing and storage).
However, there is another end of computing—buried in ambience—existing as an intertextual flurry of cultural beliefs, affects, and practices, where the entanglements of feeling and culture are catalyzed by the energies of computational performances.
The deep end: The realm of computing that deals with the performative expenditure and experience of machinic rhetorical energies (i.e., the catalyzing of visceral feelings).
Page 14 →Orienting to the deep end is to point attention to the energies that animate computational performances manifest not in the front end or the back end alone, but between them, as they are entangled with wider ecologies of discourse. The deep end of computing is characterized by enculturated habits of response regarding nature and artifice, Terminator, patriarchy, neural networks, tropes of prophets and seers, software ecologies, and the material compositions of networks; happenings, tropes, and material realities that ambiently shape the animating energies of machines. The overlooking of the deep end is, at least partially, supported by what David Gunkel would point out is an apparent effort to cling to the notion that computing machines are merely media of human communication, rather than communicators themselves.39 It is for precisely this reason that pursuing the deep end is a critical move for understanding the suasions of machines, for they increasingly govern expression,40 perpetuate stereotypical representations of marginalized identities,41 and even take on the characters of “racist sociopath[s].”42 As such, we must understand the rhetorical energies that attend their performances by diving deep into the ecologies that animate them.
To dive into the deep end requires tacking back and forth between the front and back ends of a given computational performance, discovering connections between its sensory output and background processes to “thicken” an understanding of its energies.43 Throughout the case studies of this book, I explore the gap between the front and the back ends not only by analyzing the user interfaces of machinic performances, but also the procedures that drive and characterize them. In some cases, where it is accessible, I also incorporate “readings” of code from the back end with my analyses of front-end output. To tack back and forth between the back end and the front end is a move toward what Annette Vee has identified as “full stack rhetoric” or scholarship that explores rhetorics between the front and the back end, for both the front and back ends are part and parcel of the enterprise of computing.44 The energies of computational performances are not simply expended when they render “outputs” on the front end. They are also expended in the processes of the machine as it evaluates variables and loops through conditional states on the back end. Consequently, a full stack rhetoric—a rhetoric concerned with both the front and the back ends—is required for thickening understanding of the deep ends that animate the energies of machines, fore-fronting the ambient environs from which they emerge.
The deep end of computing is an ambient rhetoric, which Thomas Rickert articulates as being located in the chora—the place—from which rhetorical action and thinking emerges. Locating rhetoric this way dethrones the rational Page 15 →human subject as manipulator of nature and instead redistributes agency across an ecology of actants, interconnected by ambient ideas and processes.45 Put in terms of an allegory that Rickert uses at the outset of his Ambient Rhetoric, the impactful character of a wine does not emerge solely from the hand of the vintner; it is also a matter of the soil, the grapes, and the stories of the land from which they grow, fermenting impact on the embodied mind. The character of the wine is as much a product of the environment as it is the human who processed the grapes. Such a reorientation to rhetoric expands the scope of rhetoric beyond words, beyond human actors, and beyond isolated “situations,” urging the analyst to dig deeper into the otherwise invisible substrate of rhetorical processes, covered over with human hubris, forgetfulness, and shortsighted yearnings to uphold the superiority of rationality.
We gain valuable insight when we adjust the aperture of rhetorical scholarship in ways that flatten and bring into focus the background with which embodied minds and things are entangled. In the specific case of machine communication, importing the chora of computing—its ambiance of ideas, wires, codes, networks, interfaces, and bodies—allows for tacking back and forth between the front end and back end, pushing us toward the deep end that animates the rhetorical energies of computational performance. Orienting oneself to the deep end of computing is to orient to the conditions that shape the energies of machines as they run conditional loops and evaluate variables on the back end, expressing output on the front end.
Particularly useful for approaching the feelings that attend the rhetorical energies of machines, as they are steeped in ambience, is Rickert’s operationalization of Martin Heidegger’s concept of Stimmwng—“attunement” (sometimes translated as “mood”).46 While it can be frustrating to pin down consistent definitions of just what scholars mean when they are using terms to talk about feelings and sensations, here I borrow from Rickert’s explanation to define some important terms to help in clarifying what attunement is. Broadly speaking, one can conceptualize a distinction between emotions and moods. Emotions are specific, articulable, states (e.g., happiness, anger, sadness, surprise, fear). Moods, on the other hand, are more general (often longer lasting), vaguer states (e.g., contentedness or angst). While moods do not necessitate specific emotional states, moods do impact the range of emotional possibility. Anyone who has awoken in an angsty mood, for instance, might note that, while it is not impossible to be happy, it is nonetheless harder to be. And, conversely, waking in a contented mood one might find it harder to be angry, but easier to be happy. One’s moods impact how they emerge into the world. This tracks Page 16 →with what Heidegger means by attunement, only his account is more robust in its recognition that one’s feelings in the world are not constituted merely of emotional content, but also lifeworld content. Ideas, assumptions, feelings, culture, previous experience—they shape, and are shaped by, our feelings. Thus, attunement is a means of catching glimpses at those affects that precede cognition, constituting our being-in-the-world, shaped by that which is beyond, but also within, the body. In this sense, when we are speaking of the deep end of computing, we are not attempting to name the specific emotional states invoked by the energies of computational performances. We are, however, attempting to dive deep into their ambient conditions to explain the possibilities of attunement that attend the energies of those performances.
Where some modernist visions of rhetoric might locate emotion as secondary to logic, approaching feelings as something added to arguments, Heidegger posits that emotion is more constitutive. According to Daniel Gross, Heidegger’s positioning of feelings as prior to logic places emphasis on the necessary energies implied in rhetorical action, for “without affect our disembodied minds would have no heart, and no legs to stand on. We would have no grounds for concern, no time and place for judging, no motivation to discourse at all.”47 As such, attunements are mishmashes of feelings and logical content (e.g., “I feel like something is incorrect here.”). Feeling is primary to, but nonetheless refracted by, logic.
Heidegger’s attunement, moreover, helps us recognize that messy estuary between sensation and cognition, wherein a piece of music does not simply afford a sad state but also is attended by an anxious attunement (which also feels like being in space, hurtling through a galaxy), informed by our cultural experiences with analog synthesizer sounds, cinema, and fast, erratically paced syncopation. Heidegger’s attunement is an answer to the problem that Hawhee raises concerning rhetorical scholarship of sensation—those processes by which bodies take up stimuli through the senses: “How to write about sensation without positing an individual as opposed to a collective, or of thinking in terms of communal sensation, without presuming sameness?”48 Attunement affords terminology for the analyst to orient to the feelings / logics that attend a given energy without falling into the trap of attempting to declare the specific emotional states or sensations experienced by persons or that all persons within a given public are the same. Instead of assessing feelings as happy or sad, attunement involves an orientation to the general states of affect that attend a given set of energies, not as individually experienced emotion, but rather as communal mood, while at the same time, bearing recognition that it is analytically Page 17 →productive to approach feeling and logic as entangled with one another. Where feeling can absolutely be conceived as something distinct from (and perhaps less than) logic, or that it is merely an individually experienced phenomenon, Heidegger’s attunement captures the fact that we are social animals that emerge into the world through culturally shaped bodies that both exert and experience energies, physical and emotional, including those of nonhumans.
Byron Hawk helps us navigate the complex of nonhuman and human energies. In his essay on the keyword of resonance in the context of sonic rhetorics, he forwards a materialist approach to rhetoric with a penchant for the trans-duction of energy. In his words: “Bodies are entangled through circulatory waves of energy and force, where resonance layers and amplifies multiple vibrations and sustains them up through knowledge production and cultural circulation.”49 Particularly helpful here is breaking apart energy and force as they are construed in the process of resonance. Energy is the thing that is “at work”—light or motion, for instance—whereas force is the vibratory resonance of that energy on some other thing(s), such as a person. Catherine Chaput’s “affective energies” approach makes a similar distinction between “affect” (which in her frame is equated with energy) and “sensation” (the apprehension of energy by the body), which positions “emotion” as the conscious rationalization of a given sensation.50 Chaput is largely concerned with human rhetorical energies, which makes sense, given that she is developing affective energies in the context of human systems of labor and capital.
By contrast, the current inquiry is interested in the energies contributed by computational performances—lively, but nonhuman, machinic energies. Furthermore, distinguishing between energy and force (breaking matters “at work” apart from the resonant moods that attend them) clears space for the energies of nonhumans to be acknowledged as they contribute to the cultivation of attunements. This is what Rickert means when he says that “rhetoric is not . . . energy, per se, but rather a need and a capacity in relation to the world and other entities that has to be discovered and deployed,” a view that resonates with Annette Vee and James J. Brown, in their brief connection to rhetorical energy in the introduction to the “Rhetoric and Computation” special issue of Computational Culture; they note that rhetorical energy “is not primarily about meaning but is rather that which triggers discourse.”51 Rhetorical force does not exist in the energies of machinic performances or discourse alone, it exists in the animating resonances between them.
Computational performances emerge as nodal bursts of energy, which we might deny or confuse either as “just the way it is,” or a simple trick, “making Page 18 →us believe it is an animal,” overlooking that they nonetheless affect us, encouraging attunements as machines. To analyze the deep end of computing is to mobilize Rickert’s ambient rhetoric, by accounting for the lively energies of computational performances as well as the antecedents of discourse that shape them. As the case studies in this book illustrate, the machinic energies of computational performances can cultivate attunements of contentedness or angst. But before moving into the case studies, it is important to first articulate what is added by rhetorical energy as a term of art in rhetorical analysis of computing by discussing it in relation to what many would identify as the defining rhetorical characteristic of computational media: procedurality.
Thickening Procedurality with the Rhetorical Energies of Computational Performance
“Procedural rhetoric,” as Ian Bogost first articulates the idea in his Persuasive Games, “is a technique for making arguments with computational systems and for unpacking computational arguments others have created.”52 Video games, for instance, can invite users to take on a particular worldview by way of presenting decisions and giving feedback on those decisions (e.g., being “rewarded” with forward progress by exploiting natural resources to “craft” a more lovely island). Such is an argument moored squarely in an interactive experience, which encourages the player to accept a conclusion not stated through verbal claims and evidence but rather through decision trees.
Although Bogost tends to favor video games as his examples of procedural rhetoric, James J. Brown is quick to note that procedural rhetoric is applicable to the broader realm of software while making the extended point that rhetorical outcomes are mutually shaped between people and machines amid networked technologies, resulting in discourses that demonstrate “the predicament of hospitality” in that proceduralization cannot open to a habit of being without closing off from, or constraining, other habits of being.53
Procedural arguments involve explicit statements (“If X, then do Y”), but this does not mean that the arguments themselves are explicit. Like an enthymeme that omits one of its premises, a procedural argument has embedded assumptions, and this invites the audience to interact and interpret. Engaging with a procedural argument involves more than reading content—it involves reading the rules that generate that content and understanding how those rules express certain worldviews. Further, procedural arguments simultaneously insist on the execution of Page 19 →sets of instructions and invite interaction with those instructions. Once again, we are presented with the predicament of hospitality—procedural rhetoric both invites interaction and attempts to hold it at a distance.54
Kevin Brock and Dawn Shepherd extend this line of thinking in what they coin as “procedural enthymemes,” or logical structures in software, which encourage users to infer conclusions from the premises implied in the operation of a program as it “analyzes potential connections across multiple points of data to ‘understand’ particular subjects.”55 Carbon footprint software, which encourages the user to conclude that they need to change their daily recycling habits, based on the premise that the software’s analysis has discovered a negative impact on the environment from the user’s inputs, would be an example of a procedural enthymeme. At the outset of their article, Brock and Shepherd draw a passing connection between rhetorical energy and computing machines, wherein they cite Kennedy’s “Hoot in the Dark” to clear space for nonhumans—such as algorithms—to be considered as rhetorical. While Brock and Shepherd are more interested in articulating the premises of argument that exist in a given software than they are in the energies of lively machinic performances, their connection to Kennedy is generative in the sense of moving toward the computational machine not merely as a medium of arguments (as in the decision trees of video games), but as a performer of arguments (as in the performance of “data analysis”). Put differently, Brock and Shepherd’s work puts us on a path to recognizing that the movement—the energy—of the machine is just as important to its influence as the line of argument that it represents.
For example, in the case of a carbon footprint calculator, the front end might be adorned by clever graphic designs and user-interface elements that allow one to indicate details such as how many people live in their household and whether they drive a car. Those elements may convey the generally “busy” aesthetic of a scientific organization, wherein, alongside the user-interface elements, text blobs explain the methodologies used to design the application. Procedurally, the user is asked to share their daily behaviors by entering numbers of miles driven in a car or flown in an airplane, offering a logic that directly ties one’s personal behaviors to environmental impact. The user can even play with the numbers, demonstrating variance in their carbon output, which reifies the notion that even small changes in one’s daily living can have concrete impacts on climate change, based on scripts operating on the back end of the application.
Page 20 →But if we articulate the rhetorical energies of the carbon footprint application, we discover “more.” In particular, as the case study of the first chapter will elaborate in more detail, within the deep end of computing exists the trope of the prophet as well as the rituals of interaction endemic to expert systems, which can afford an attunement, wherein the user is encouraged to feel that the conclusions of the application are true, realized not just in the procedures the user is asked to interact with, but also in the performance of “running calculations,” enlivening an argument about environmental relations amid the Anthropocene, by leveraging the energies of a computing machine to speak to human concerns in a more-than-human way. The application is attended by a feeling just as much as it instantiates an argument, explainable not simply by way of the front or back end, but rather by diving into the deep end of computing, where longstanding discourses (like those of oracles and rituals of expert systems) ambiently shape, and are shaped by, the movements evident in the front and the back end processes—the energies—of the performance.
Surely the procedural argument of the carbon footprint calculator is integral to its influence, but much in the same way that the prosody of one’s voice is integral to the force of their articulations of reason so too is rhetorical energy to the procedures of computational performances (see table 0.2). And those energies impact the culturally shaped body to animate attunements, comportments of feeling. Casey Boyle, drawing on Jeffrey Walker’s classical rhetorical work, notes that despite common assumptions about classical rhetoric, which might imagine that enthymemes deal simply in logic and meaning, they also deal in emotion. And if we take such a realization a step further to focus on rhetorical practice from a posthumanist perspective, we will see that enthymemes are better understood not as “missing” a conclusion to be filled in by an audience, but rather as nodal flashpoints, where pre-existing ideas and feelings are transduced from a complex of practices, tied together by an ever-evolving ecology of objects, including humans.56 The carbon footprint calculator may very well instantiate an argument, manifest as a procedural enthymeme. But the carbon footprint calculator also instantiates a performance, manifest as “running the numbers.” Consequently, its effect is not simply meaning or logic changing one’s mind: it is also an attunement, catalyzed from the energies of machinic movement, resonating with existing discourses of the prophet and expert systems to animate an embodied consideration of the relationship between daily practices and CO2 output.
We experience the rhetorical energies of computational performance in ways that we are unaccustomed to naming, for they exist beyond the paradigm of rhetoric as the enterprise of influence had between humans as symbol-using animals.57 Humans obviously use symbols; but we also feel in the world, and, at least some of the time, we are affected by nonhumans, who may not influence us as invested minds making appeals to us, but who absolutely influence us at the level of conveying lively rhetorical energies that interact with our culturally shaped bodies. Brown, in his reflective response to the essays of Rhetorical Machines: Writing, Code, and Computational Ethics, makes a nod to the shortcomings of logocentric rhetorical scholarship of computing, noting that “like any symbol system, code will always elude our grasp, and digital rhetoricians’ move along a continuum from rhetoric (back?) to logic may put us at risk of forgetting this.”58
| Page 21 → Procedure | Argument(s) | Energy | Deep End | Attunement |
If the carbon footprint is greater than X, suggest Y behavior | One’s behavior matters to the environment Changing one’s individual behavior can have a desirable impact on CO2 numbers | The performance of information analysis and retrieval | The rituals of interaction, endemic to the tradition of expert systems, manifest in the front and back ends of the application, further resonating with the trope of the prophet (wherein an agent interprets truth directly from the ether) | Contentedness |
Extended Description
Table that describes the procedure, arguments, energy, deep end of computing, and attunement in a carbon footprint calculator web application. The table is rotated ninety degrees counterclockwise.
Page 22 →Similarly, outside of rhetorical studies, in the realm of human–machine communication, where scholarship is largely quantitative (rather than interpretive), one of the earliest leaders of the field, Steve Jones, has identified “context” as well as “symbolic and affective dimensions” as integral to accounting for the communication between humans and machines.59 This book is an attempt to move toward a necessary thickening of the idea of procedural rhetoric, which locates not just the arguments and logics of software, but also its sensations and feelings, offering an interpretive framework useful for accounting for the deeper contexts from which the affective influence of computing machines emerge.
In subsequent chapters, I argue that the movements of a web application can imbibe the prophetic energy of an all-seeing oracle, a Twitterbot can radiate the sublime energy of nature, and machine-learning systems can signal the “purity” of mathematics. I will also demonstrate that taking seriously the influential energies of computational performance requires revisiting long-standing assumptions about communication ethics, which might problematically define responsibility in terms of the living, and not the lively, resulting in designs of machine communication that do not adequately “hedge bets” against moments in which the machine might catalyze negative energies, beyond the intentions of designers.
It is important to note that the goal of this book is to equip readers with concepts useful for analyzing and thinking through computational performance, even if the technologies of computational performance might be evolving rapidly. That is, while the web applications, bots, and voice assistants discussed in this book might reach a state of technological obsolescence, the concepts built from studying them establish a working foundation from which to integrate, and adapt to, shifts in the technological landscape. This book is about pushing Page 23 →rhetoric to put us on a path toward grappling with computational performance, even if, by its very nature, computational media will likely evolve beyond the scope of the individual case studies discussed herein.
The chapters of the book are organized around what Leah Ceccarelli has articulated as the three ends of rhetoric: the epistemic, the aesthetic, and the political.60 The first case study focuses on the manufactured processing of a simple web application designed by vaccine denialists. The second case study interrogates the processual magnitude of a piece of data art. And the third case study examines the processual signaling of a neural-net-based political performance.
Beyond organization around science, art, and politics, the reader will also notice that the objects of each case study increase in complexity and range of autonomous movement to further illustrate that, while entangled with procedures, computational performances do not necessarily require interaction in the sense of a user or player making choices through a decision tree. The web application of the first case is bound to user interaction. But the bots in the second and third case studies do not require user interaction to carry out their processes—they are, to use the parlance of rhetoric, more monologic than dialogic. Further, this sequencing also tracks with increasing autonomy on the part of the machines. The web application of the first case moves only in response to user interaction. The bot of the second case autonomously generates messaging within the constraints of pre-programmed parameters independently of user interaction. And the neural network of the third case autonomously generates messaging with much looser constraints imposed by the programmer without requiring user interaction.
Following the case studies is a meditation on ethical critique and design of computational performances with concern for the good or ill that computational performances can imbue to the social ecology with their lively (but not alive) movements. The book concludes with a discussion of the value of a rhetoric as energy approach to the growing field of human–machine communication by examining the lively movements of machine communicators amid the COVID-19 pandemic. For now, to further parse the rhetorical energies of computing machines from the procedures that punctuate them, I examine the prophetic energies of Vaccine Calculator.