The use of Facebook advertising and fake Twitter accounts by Russian hackers to influence the US election and infiltrate activist groups like Black Lives Matter are the most recent and perhaps most dramatic demonstrations that algorithms, bots, and big data are now at the foundation of our political milieu. These technologies help shape the field of possible political organization and action, for better and worse. The cultural and political theorist, Davide Panagia explores this situation through his concept of #datapolitik. Panagia coined this term because he believes that we lack an adequate critical vocabulary and theoretical tools to articulate the shifting nature of politics today. For Panagia, the horizon of possible forms of political organization and collective action has been transformed by the ubiquity of algorithms and the reduction of political subjects to data points amenable to behavioral analysis and control. In this interview with Contrivers’ Review editor Çağlar Köseoğlu, Panagia discusses both how the “ever-growing governmental infrastructure of non-human agents” constrains our politics and how this same milieu of #datapolitik holds the possibility of “emergent forms of political organization and action, solidarity and emancipation.”

Çağlar Köseoğlu: Recently you have proposed the concept of datapolitik (or #datapolitik) and described it as an “underappreciated form of modern political power” and “the unique constellations of political forms available to our contemporary techno-digital condition.” Datapolitik seems to be in dialogue with other relatively new concepts such as informational politics and algorithmic politics. Could you, to start off, expound on what datapolitik broadly entails and why you have chosen for this particular variation—politik—instead of the more common “politics”? 

Davide Panagia: First off, Çağlar, let me thank you for getting in touch with me and inviting me to do this interview. It’s very generous of you to be interested in my work which is very much in an early stage of development. And quite frankly, your questions are inordinately helpful in giving me the opportunity to orient my thoughts which are, as you suggest, in dialogue with many, many conversations and writings appearing currently in both academic and non-academic spaces. You are absolutely right, in this regard, to note the potential affinities of the term #datapolitik with other kindred terms like informational politics and algorithmic politics. The fact of the matter is that I draw quite substantially from people working with these other terms and ideas—especially thinkers like Colin Koopman, Orit Halpern, Louise Amoore, Jairus Victor Grove, Tiziana Teranova, Richard Grusin, Neal Thomas, Mika La-Vaque Manty, Rita Raley, Brian Massumi, and Antoinette Rouvroy (amongst many others) whose work on “algorithmic governmentality” (another term to add into the mix) I’ve recently discovered.

What these and other thinkers seem to be latching onto as an urgent task of political thinking is the fact that our systems of governance are operated by non-human agents who have the capacity not only to govern everyday life, but more crucially to make autonomous decisions. These non-human, sovereign agents are the algorithmic equations that animate us. This, it seems to me, is a concern that goes well beyond the kind of analysis typically provided of the politics of social networking—analyses that rely on pre-established conceptions of both psychology, action, and sociality. In other words the rise of #datapolitik to me means that our established critical vocabularies, theories, and conceptual innovations are insufficient for the demands posed by this ever growing governmental infrastructure of non-human agents that possess an impressive degree of sovereignty over people.

For instance, it seems limiting to me to exclaim the scandal of neoliberal capital as a symptom of our digital age. Of course neoliberal capital is complicit with the spread of information networks. But that’s hardly a compelling insight because it’s not clear that the “j’accuse” directed at neoliberal capital does anything other than point to yet another instance of exploitation. And once we’ve itemized yet another structure of domination, what have we done? That is to say, are the various sophistications of ideology-critique the best that our political thinking can do in this case? And are the transformations of power, governance, and self under #datapolitik reducible to the image of political thought that ideology-critique perpetuates?

What to me was exciting about many of the critical ventures that developed around media and aesthetics in the twentieth century was the concerted effort on the part of various thinkers to develop their political-theoretical positions in relation to the specificity of technical media. I’m thinking here of Benjamin and Kracauer on film, Adorno on film and music, Foucault on spatial architectures, Derrida on reading and writing, Butler on gender, and so on and so forth. Technical media in the cases of these thinkers weren’t simply physical objects of investigation, they were dynamic perceptual milieus that participated in the disposition of worlds. This is why I very much prefer the term “dispositif” or “dispositive” when speaking of technical objects rather than the term “apparatus.” And I prefer the term dispositif because I share James Chandler’s intuition about the Roman rhetorical sense of the dispositio of technical media, which refers to how matter is disposed, how its component parts are arranged, ordered, and organized.1 That Roman rhetorical sense of dispositio is also at the heart of Foucault’s own notion of governmentality as “the right disposition of things.”2 A dispositio regards the arrangement of peoples, things, and sensibilities and, crucially, the limits of those arrangements and designations. In any case, what happened in twentieth century critical thinking was the emergence and elaboration of philosophical vocabularies that exposed the limitations of dispositifs, but also showed the transformations of worlds that those same techno-mediatic dispositional arrangements enliven. If you go back to the ideas innovated in the twentieth century, it is truly stunning to see the extent to which technical media are implicated in the development of ideas that today we typically take for rote as our terms of political criticism.

I should note that I come to the work on #datapolitik as someone whose research thus far studies the relationship of aesthetics and politics. Through such work, I have come to a deep appreciation of the complicity of technical media with and in our thinking—which is why #datapolitik for me presents an (urgent) opportunity to develop new conceptual artifices and critical vocabularies that attempt to answer, among other things, what are the different modes of political transformation in the digital age. And rather than adapt critical vocabularies from other technical media and transpose them upon our current scenarios, I think it is important that we exploit and investigate the dynamics and limits of the most prevalent technical media of our day. For me, these include the perceptual milieus enabled by the technical medium of the algorithm, the ontology of the feedback loop, and the heterotopia of ubiquity.

So to get at the second part of your question—why the “k” in #datapolitik? In part, this is a heuristic. When I started thinking about these issues, I wanted a term that pressed upon me the political dimensions of the project of inquiry, but that in some sense also distracted me from more conventional uses of the term “politics” for all of the reasons I spell out above. There is to me something unique happening today that definitely has a history, but that is also importantly different from the understandings of power, governance, and humanness that have come before. So, the hashtag matters to me because it signals the various tactics of ubiquity that current technologies enable. And the “k” matters because it rehabilitates an older language of power—of “realpolitik”—prevalent in some corridors of history and political science departments.

But to get back to what I was hinting at earlier, what is new to me about #datapolitik is that it is a form of realpolitik of and by non-human agents. Thus it begs of us the need to rethink our assumptions about the motivations of action in politics. The older term “realpolitik” was designated to describe an account of politics that tried to explain the actions of great statesmen who saw the accumulation of power through territory and war as the motivational purpose of nation-states. So decisions were said to be made on the basis of how to aggrandize power, territory, and domain. #datapolitik is not that, in part because I don’t think the anthropomorphism that assumes states are rational actors fits here. There are other relational dynamics at play. I hesitate to say this, but it feels as if that idea of the state is almost irrelevant to #datapolitik in part because the algorithm is indifferent to the content of any identity—whether state, self, or nation. That is, what kind of a relation is “interest” when non-sovereign agents (to abuse a term from Sharon Krause’s book Freedom Beyond Sovereignty3) are developing domains and modalities of influence that help determine political outcomes. In any case, I am currently exploring the idea that #datapolitik is not a power of sovereignty but it is fundamentally a police power.

CK: One of the people you mentioned, Antoinette Rouvroy, charts—in a lecture given in Amsterdam in 2013—three dimensions that have been significantly and in novel ways transformed by what she terms “algorithmic governmentality”: knowledge production (becomes knowledge without truth), power exercise (becomes power without authority), subjectivation (becomes personalization without subject). You, too, seem to focus on roughly the same categories and map the current shifts in a conceptual language that is reminiscent of Foucault. Could you indicate or perhaps speculate to what extent #datapolitik—in particular its non-human aspect and fairly new techniques like the algorithm—can be grasped in a Foucaldian framework of governmentality?

DP: Your intuition is right to the extent that, like Rouvroy, I am indebted to Foucault’s morphology of aesthetic and political sensibilities. And, like Rouvroy and others, I also consider Gilbert Simondon’s work as important to the study of #datapolitik. But for me it’s less the conceptual language that I find compelling in these and many other thinkers (i.e., Deleuze, Rancière, Barad, Chamayou, Latour, Massumi, Berlant, etc.), than the sites and emphases their work generates, and (especially with Foucault) the relevance of the micropolitical practices that operate in the age of the algorithm. Foucault didn’t begin with concepts but began with practices that were attached to institutions and their forms. And here, an institution is not necessarily an established construct but, again, a series of activities and sensibilities that operate in more or less identifiable and consistent ways, though not always as consistently as we’d like. I like to call such a mode of analysis an affective pragmatics. It is affective because it is interested in dispositions, perceptibilities, and sensibilities; it is pragmatic because it is interested in how all of these things work together and apart.

As I’ve suggested, with my exploration of #datapolitik I’m less interested in issues of governmentality per se than of governmentality’s service animal, the police. And again, I’m not interested in the police as a specific entity that interpolates, but as a set of operations that regulate circulation through space and time. These operations include modes of thinking (i.e., cause and effect, teleology), technologies (i.e., the cybernetic feedback loop), practices (i.e., predation, tracking, and capture), attitudes (i.e., safety and security), and beliefs (i.e., an immanent threat or potential catastrophe). And so it’s undoubtedly true that whenever we do any kind of conceptual work we do so by invoking an archive of referents: authors, terms, concepts, and intelligibilities. But the question for me isn’t so much the potential repetition that comes with one’s occupation of an archive, but the people, places, events, and activities that one’s own specific displacing and disordering of an archive put on display. The task of intellectual indebtedness for me is one of the transmediation of archive, or what Miriam Hansen refers to as the heaping of broken images that is an “an-archive.”4 This is also why I remain committed to David Hume’s psychology of impressions.5

I want to offer opportunities for giving attention to things, events, forces, and collectivities that have, as of yet, been underprivileged in our thinking about the relation of politics and digital objects. And the way that I wish to proceed is to look at how the development and application of cybernetics today, its history, the motivations behind its emergence, its psychologies, and so forth, are impacting everyday life and, especially given my own political sensibilities, the ways in which we think about political organization and action, solidarity and emancipation. This is why I’ve begun to articulate #datapolitik as a police power, to the extent that it is a coordination of technologies and sensibilities oriented towards the cynegetic predation of actions that take shape as information or data-points, or what we otherwise call “clicks.” This idea came to me by accident.

In the fall of 2015 I was reading a great deal about cybernetics, thanks in part to Orit Halpern’s important book, Beautiful Data.6 It’s that book that sent me to read the works of Norbert Weiner, Claude Shannon, and W. Ross Ashby. Now, one thing you must realize about me is this: I am innumerate. In fact, I came to political theory because I dropped a Calculus class in my freshman year in college and so needed to pick up a credit for the year so that I might graduate on schedule. I decided to take an intro course in Political Science taught by a legal and political theorist, and the rest is pretty much history. This to say that when I read the cyberneticists I focus on the way in which their description of what mathematical calculations are apt to do inform their motivations of what they imagine possible.

At the same time—and here’s the “accident” part of the an-archive story—I began reading an important book by the philosopher Lisa Guenther, Solitary Confinement, that many colleagues had told me I had to read. The book is a phenomenology of solitary confinement punishment in the American prison-industrial complex. It’s an extraordinary work that everyone should read, even though the tenor and site of the work is an incredibly demoralizing topic. But Guenther is the most hopeful and inspirational of writers and she refuses to deny her readers the vitalism necessary to engage the horrors of the practices she describes by being merely disenchanting. I never thought a book on solitary confinement and social death could be so uplifting! It’s one of the beautiful things about the tradition of phenomenology—it really is committed to embodied life.

Moving on. In the book there are a series of chapters devoted to behavior modification, a topic and a literature I hadn’t entertained for some time. And Guenther goes back to Merleu-Ponty’s early work, The Structure of Behavior, and shows its relevance to the ideas of behavioral psychology behind solitary confinement. More importantly for my immediate purposes, however, is that the connection to Merleau-Ponty struck a note as I was also reading of and about Simondon’s critiques of Aristotelian hylomorphism in his engagements with cybernetics, and his theory of individuation. Simondon, of course, was a student of Merleau-Ponty (amongst others) and cybernetics is an offshoot of I. Pavlov and B.F. Skinner’s behaviorism. If you combine all of these messy strands together, you arrive at the possible idea that what we’re dealing with when we consider something called #datapolitik is a collaboration of sensibilities for behavior modification that emerges and derives from a set of techniques, practices, and affections for capturing things in motion that we might otherwise call police sensibilities.

Of course, this should be of no surprise given that cybernetics is one of the many military technologies that find their way into everyday life. And as Weiner and Ashby make clear, cybernetics is a technology that adopts a recursive calculus of repetition for the taming of difference so as to enable the infinite predictability (and thus availability) of a moving target. In fact, it’s impossible for me now to read Gilles Deleuze’s Difference and Repetition—a work that has always been part of my intellectual lexicon—as anything other than a retort to cybernetics, and it’s equally impossible for me to consider the turn to affect that Deleuze’s work procures anything other than a response to behaviorism’s confidence that experience is representable, identifiable, and trackable.

Cybernetic innovations involve the application of technical objects (i.e., algorithms) for the tracking and capture of difference, pure and simple. Their field of application is undetermined, because ubiquitous, and they work on exactly the same basis and principal that Skinner’s approach to the study of human action does; namely, the idea that an action is independent of any content, or thick description of the human essence (like soul, or an identity, or a will), because what counts is automatic stimulus (i.e., the clicking of the “like” button on Facebook) and not the biography of the agent. And this attitude towards action’s automaticity also has a history—and that history is an enlightenment inheritance of post-humanism that finds progenitors in people like Descartes (of course), but more dramatically in the French philosophe La Mettrie who attempted to defend atheism (and thus deny the concept of the soul) by imagining the possibility that the human animal is simply an animated machine. Skinner confirmed the autonomic nature of animation when he learned how a salamander’s tail would move when pricked, even after it had been cut off from the salamander’s body. And he (as well as everyone else) had had practice with the idea of human automata given that automata were amongst the most prevalent form of entertainment throughout the modern period,7 especially if you consider the fact that cinema is a ubiquitous technology that animates humanoids on a screen automatically.8

Like I said, it’s a big mess, it’s an an-archive; but what I’m sensing is a way of describing what (today) we experience on an everyday basis—that is, the development and implementation and adoption of behavioral systems of control that track and capture movement on a micro-temporal basis. And this to me sounds a lot like a police sensibility of cynegetics and predation, and I want to learn more about it. And I don’t know (yet) if my learning will produce a different kind of critique altogether, as you suggest, or not. In part because, as I mentioned earlier, it’s inordinately difficult for me to conceive of critique in light of the fact of ubiquity and in light of our unwillingness (or, perhaps, the impossibility) of giving up on our collective technical existence. As I see it, critique is always based on technology; it is responsive to systems that produce things: the printing press, industrialization, photography, etc. And in part, what is implicit in the notion of critique is the possibility of having some distance between subject and object, between critic and work. I may be critical of “x” because it is collusive to freedom and thus will commend its demolition or, at the very least, one’s turning away from it. But I don’t believe that is possible with contemporary digital objects given the immersive nature and ubiquity of this abstract things like data and algorithms and, indeed, given our entangled complicity with their perpetual workings.

CK: In your previous research you have been expressly committed to an aesthetic understanding of politics. On your view, how could one make sense of the fundamentally cynegetic techniques and processes of the “police” in aesthetic terms? In other words, what is the exact aesthetic dimension, status or workings of some of the key digital objects of our present? And, on the basis of this, do you think that #datapolitik points toward a particular form, mode or site of political action?

DP: This is a very interesting question. Here’s how I would begin answering it: We live in the age of ubiquity that has superseded—or at the very least, is surpassing—the age of the particular.

I think that one way to begin to tackle the spirit of your questions, then, is to think about our relationship to the dispositif of #datapolitik that I am isolating, namely the algorithm and its ubiquity in everyday life. The #datapolitik project is a continuation and extension of my research on the relationship between aesthetics and politics in contemporary life given its commitment to pursuing a study of the dispositifs of our digital age. So first, a few words on this notion “dispositif” that is in some circulation today. Allow me to add to the comments about the dispositif I’ve already made (see the answer to Question 1). The term has an interesting aesthetic and political history that isn’t quite available to non-French readers. In English, dispositif is typically translated as “apparatus” and the substitution is rarely, if ever, noted. But as Alain Brossat has shown, the language of the dispositif has a historical specificity that marks the emergence of a novel conceptual and theoretical imaginary that, “to put it bluntly, represents a movement from ‘science’ to ‘politics.’”9 Now Brossat is referring specifically to Michel Foucault’s lexicon. But I think that his point can be generalized, and we can talk about a general shift in sensibilities from the idea of a technical apparatus to that of the dispositif. And that shift regards a shift in our conceptions of power, its operations, and the perceptibilities that emerge from an ensemble of technical and human forces. The apparatus (exemplified, for instance, in Althusser’s mechanical engineering metaphor of ideological state apparatuses) are devices that control from the top down and impose a certain order of operations in society. Dispositifs are, to use the language of the sentiments, dispositional arrangements that dispose a milieu of perceptibilities and sensibilities.

In his recent Aisthesis Jacques Rancière offers a helpful formulation: “A medium is neither a basis, nor an instrument, nor a specific material. It is the perceptible milieu of their coexistence.”10 Now I may be making the mistake of using the term “medium” and “dispositif” interchangeably, but for the time being I think we can allow ourselves that slippage only so that we might emphasize this domain of coexistence of a variety of technical and perceptual forces that are disposed in a particular way and produce certain effects. Thus, when we translate dispositif as apparatus we end up missing this aesthetic and political specificity, this shift between the idea of technical media as tools of domination to talking about them as participants in the arrangements of sensibilities and perceptibilities of a techno-human milieu.

The aesthetic and political point, for me, is to acknowledge this experiential milieu and to consider its terms of operations, its workings, but also its distensions, its capacities, and its transformations. And, finally, how we might engage it critically: What are the dispositifs of the age of ubiquity, we might therefore ask? And we may begin to answer this question by noting that one of the principal effects of #datapolitik is its transformation of the field of operations of everyday life from whatever it may have been to one of tracking and capturing changes and alterations; in short, the age of ubiquity generates a cultural politics of cynegetics and predation. To answer your question explicitly—“how could one make sense of the fundamentally cynegetic techniques and processes of the “police” in aesthetic terms?”—one thus needs to consider these diverse techniques in terms of the dispositional milieu they arrange; that is, the dynamics of sensation, perception, movement, and transformation they effect, the capacities they make available, their structures of support, and their limits.

Given this, the important thing to note is that with #datapolitik we’re not talking about those familiar, Benthamite, surveillance and disciplining procedures and tactics. The specificities of the dispositif have changed substantially from Foucault’s studies on utilitarianism. Because frankly, there is no looking and no surveying going on today; this despite our heightened anxiety about NSA spying techniques. The reality is that nobody actually cares about what you look like, how you look, and what you are doing. Nobody is listening, transcribing, or recording. We are eons away from The Wire (2002-2006) or The Conversation (1974). What matters today is your movements and how they may be tracked by a highly sophisticated set of algorithmic calculations that monitor the differential positions of your signal transmitters—whether your cell phone, or your computer, or your smartwatch, or your body. What we’re talking about is what Rita Raley refers to as “dataveillance,”11 a concept initially proposed by Roger Clarke12 in 1988 that deserves a full research agenda of its own.

Now, it’s crucial to realize that dataveillance is nothing new, to the extent that it has been at the heart of so many of our modern practices in marketing and economics, political science, policy analysis, and everyday life. Once the human was recognized as being the bearer of information by, for instance, Francis Galton (1822-1911) or Alphonse Bertillon (1853-1914), then it became possible to track and capture humans as never before. Hence the development of what Grégoire Chamayou identifies as our papered subjectivity in the passport that is, for him, central to the emergence of police dataveillance.13

As we’ve noted, the basic operation of dataveillance is to measure differentials in human movements, gestures, ticks, and clicks—and there are a variety of technical instruments, some digital, some not, that enable this. This is different from classic Benthamite surveillance for a variety of reasons: At one very basic level, dataveillance is not motivated by any correctional norm or ambition. It doesn’t want you to change your behavior, at all. It is a classic example of Rancière’s police qua traffic cop: it doesn’t interpellate but says “Move along! There is nothing to see here!”14 Dataveillance just wants to measure a differential, and collect data on that, not to correct anything that may be wrong about your behavior—whether that behavior is that of the criminal, or that of the student, or that of the delinquent. And this is evident by the fact that while the Benthamite panoptic scenario required the subject’s knowing (or at the very least intuiting) that they were being observed, in the case of dataveillance no such self-knowledge or subjectivity is necessary in order for its operation to be effective. This doesn’t mean that we don’t suspect that we’re being watched, especially after Snowden’s release of NSA documents; but our knowledge of surveillance doesn’t change dataveillance’s operation or, for that matter, ours. No one I know has ever stopped using computers, email, cell phones, or any digital channel of transmission as a result of Snowden’s revelations.

In short, one major difference between panoptic surveillance and dataveillance is that the latter has absolutely no interest in the interior lives of humans, or in changing those lives. And this is a consequence of its behavioral origins, as noted earlier. Behavioral psychology offers a way of thinking about human movement as automatic and independent of will, or intention, or soul. Human movement is automatic and it provides data about how movement works. It does not offer insight into the inner workings of the soul—and, frankly, why does that even matter? Hence the great and groundbreaking formula of cybernetics: information behaves. For #datapolitik, then, there is absolutely no difference between the tracking and capturing of information about a NASDAQ transaction, the tracking and capturing of a terrorist’s movements, or the tracking and capturing of consumer trends. Our subjectivity is indistinguishable from our objectivity in that our existence matters because we are bearers and transmitters of data differentials. The human in #datapolitik is a human derivative, fully automated, and indistinct from any other automated object.

Given this, let’s look at an object of dataveillance that opens up a potential milieu for the experience of #datapolitik. If you consider the example of any random customer survey, like the end of semester class survey North American undergraduates are asked to take, you see dataveillance at work. A classroom survey charts an external expression of experience—not the motivations of your experience, but its naked externality—the fact that there is a differential between a before and an after, between where the student was and where she is now. I can’t know how satisfied you are with my class lectures, for instance, or how I can make it better for you, and so my teaching institution generates a set of questionnaires that can be easily answered by coloring in, or clicking, a small bubble that registers a condition. That is, I have no access to your interiority and frankly, who cares? All I have to do is prompt a sequence of gestures that register a position in space and time. Those movements delimit a differential that can be tracked—the location “before” the movement and the location “after” the movement. And the click marks the physical difference of that differential: you were somewhere between 1-5 on the ‘what you know’ spectrum before you took the class, now (hopefully) you are somewhere between 5-10.

In his classic study on cybernetics, W. Ross Ashby asserts that:

The most fundamental concept in cybernetics is that of ‘difference’, either that two things are recognizably different or that one thing has changed with time… All the changes that may occur with time are naturally included, for when plants grow and planets age and machines move some change from one state to another is implicit.15

So the class survey is a classic instance of the customer satisfaction report that provides differential data that can be examined and analyzed. It marks a change in one’s dispositions to the scene of learning, and the learning subject (i.e., the student) is like a machine to the extent that its satisfaction is tracked by that gesture which marks a differential—a change—in its sense of how its world is arranged. The classroom survey is a classic dispositif of dataveillance that enables the tracking of success or failure of the learning experience by the eighteen year-old undergraduate student. This data is then collected and used to ascertain the positive or negative value of the instructor who isn’t in the classroom to occasion learning but is in the classroom as a prompt for the student’s charting a differential that is measured exclusively as a “+” or “-”.

Now, this is a charged example and a controversial one too. But I use it to offer the sense of the extensiveness of the police dispositif I am trying to assemble. And with it we see aesthetics and politics at work at all levels and in different modes, modalities, and media. Your last question asks me about a particular form of #datapolitik’s expression—but I don’t think that’s possible; because what I’m trying to chart is a generalized series of dispositifs that distend and extend throughout our contemporary condition, that are manifest in such a rich and complex variety of forms, that any one form is simply an instance of a general operation of our cynegetic milieu. The example of the classroom survey is a manifestation of what I’m looking at, but I’m sure we can come up with a million plus one other examples of that milieu of forces, perceptibilities, and effectivities. And that perhaps is the point, that the age of the network has made the specific or the particular more and more difficult to value because for #datapolitik, any particular is collectable as a part of a ubiquitous whole that doesn’t require any specific rule or identity for organizing its collectivity. More than the age of the network, then, we should speak of the age of ubiquity and how the operational logic of ubiquity raises problems for political and aesthetic collectivization.

CK: Your description of the dispositif calls to mind Rancière’s distribution of the sensible, which too refers to the simultaneous constitution and delimitation of our perceptibilities and sensibilities. For him, as you know, a given distribution of the sensible can be disrupted and reconfigured (i.e. redistributed) by the appearance of a political subject, which he defines by the (speech) act of partaking. How much possibility do you see for political agency under #datapolitik and what kind of modalities and planes would this involve? What would a political practice “against” #datapolitik look like as we are so dependent on #datapolitik’s countless technical milieus, #datapolitik is so indifferent to the content of our actions and #datapolitik has permeated many of the operations that make up our daily lives? The final sentence of your last answer indicates that the ubiquity of the dispositif of #datapolitik complicates a practice of collectivization. Could you expound on this a bit more, in particular on its implications for what you have just called “political organization and action, solidarity and emancipation”?

DP: Honestly, I feel that anything I could say to reply to your question will be thoroughly dissatisfying—in part because I think the task is precisely to consider not only the implications of #datapolitik for collective politics, but also the transformations wrought to our ambitions and sensibilities of collectivization by an entirely new (at least for the enterprise of political thinking) arrangement of techno-human forces. And the first challenge (at least for me) is that it’s not clear that a position of negation—the idea of developing political practices “against” #datapolitik—is an option on the table. I’m sufficiently persuaded by the work of Jonathan Sterne, Orit Halpern, Richard Grusin, and others that write about the operational milieus that shape contemporary perceptual technics (to use Sterne’s wonderful expression) that I’m challenged to conceptualize a position of negation vis-à-vis #datapolitik. The simple fact that I would never consider replying to any of your questions on pen and paper and then mailing them to you (rather than typing my replies on a keyboard and sending the file electronically) suggests to me that any “against” strategy needs to come terms with a certain inevitable hypocrisy.

That said, you’re correct in sensing a proximity between my use of the term dispositif and Rancière’s partage du sensible. As I’ve suggested throughout our exchange, with the term “dispositif” I am also invoking the rhetorical tradition of dispositio; and as we’ve also noted the term dispositif has an interesting genealogy in the twentieth century, especially in the French context. Of course, the term is mostly associated with Foucault’s studies on governmentality and biopolitics. But Foucault is, in adopting and adapting this term, himself indebted to Georges Canguilhem who, as far as I know, was the first to develop the idea of technical dispositifs in a brilliant essay entitled Machine and Organism. And then, of course, we have Louis Althusser who abandons his own commitment to the term “apparatus” and picks up on the language of dispositif in his later writings on Macchiavelli, and in his explorations of aleatory materialism. In short, dispositif is a political concept—and for my interest in #datapolitik, it is a central political concept. This because a dispositif registers the entanglement of human and machine forms of sensorial arrangement. A dispositif both generates and constrains techno-human ensembles. More than merely extending human capacities (McLuhan) or being an oppressive influence machine (Tausk/Mulvey/Metz/and the early Althusser), technical dispositifs generate what Gilbert Simondon calls associational milieus that arrange what is and what isn’t available to one’s political sensibilities.16

All of this preamble to say: Yes, without a doubt there is political agency with #datapolitik. Of this I am certain. The question remains as to what political agency looks like in and with #datapolitik. There is the agency of non-human bots, of auto-genetic algorithms, and all sorts of automated systems, there is the distensive agency of fake news (about which we have learned a lot recently); there is also a more classical form of agency in #datapolitik of communication control, which is the cybernetic ambition of much of our data world; there is the agency of hacking, of cyber-espionage, and cyber-terrorism; and there are many other forms of political agency too numerous to list here that include the whole world of virtuality and military training, of drone warfare, and immigration control.

In short, I don’t want to give the impression that #datapolitik is oppression and domination all the way down. The task of political theorizing is (for me) to think beyond the normative and deontological and to explore the potential power of collectivization in #datapolitik. Despite the celebrated marriage of technical media and collective political movements that followed the Tahrir Square uprising and Occupy Wall Street, there has been surprisingly little inquiry into how these collective events that were enabled by #datapolitik are different from more traditional forms of collectivization that we associate with the pre-digital world. I’m thinking here (for instance) of Jason Frank’s important work on the incipience of the people in revolutionary America, and his insightful readings of the literary excess of words that both enabled and produced what he calls “constituent moments,”17 and I want to ask how we can think of collectivization and incipience—of radical democratic constituency—with our contemporary technical milieus?

I don’t think one needs to ontologize technology in order to ask whether contemporary media operate on different levels of transmission and interruption than earlier forms of papered or visual media. And if they do, then one must assume that the forms of collectivization that can emerge out of #datapolitik constellate differently than those that might arise from, say, the nailing of 95 theses on a church door, or the penning of a revolutionary pamphlet. At a very basic level, it might be the case that there is a “radical mediation” in our conception of publics as spaces of spectatorship and/or readership. This is a point that Richard Grusin has recently raised in revising his term “remediation” and introducing the idea of “radical mediation.” And I take great inspiration from his insight that “mediation should be understood not as standing between preformed subjects, objects, actants, or entities but as the process, action, or event that generates or provides the conditions for the emergence of subjects and objects, for the individuation of entities within the world. “Mediation,” he says, “is not opposed to immediacy but rather is itself immediate. It names the immediacy of middleness in which we are already living and moving.”18 Following from this, the question to me isn’t whether collectivization is possible because #datapolitik is so pervasive, ubiquitous, and dominating. But, rather, how do ubiquity and #datapolitik mediate (in Grusin’s sense of the term) forms of solidarity and emancipation, what are the limits of solidarity and emancipation in #datapolitik, and what are the permeabilities of those specific limits?

The issue I currently consider most politically constraining of #datapolitik isn’t the fact of ubiquity so much as the ubiquity of prediction as the logistical metaphysic of everyday life. This is a particular behaviorist malady that despite decades of criticism nonetheless persists without end. At its basis, this is what an algorithm is designed to do: namely, to predict future outcomes and to coordinate action (and therefore movement, and therefore logistical operations, and therefore futures) so as to attain a goal (that typically translates into profit or some analogous form of earthly salvation); more specifically, an algorithm has in its DNA the ambition to strike a target, and to adapt in every possible way so as to achieve target success. Richard Grusin has called this premediation, Brian Massumi calls it the operational logic of preemption, Louise Amoore has called this the politics of possibility, and Orit Halpern has identified it as the new (post-cybernetic) rationalism.

So the issue for me isn’t how we stand against #datapolitik, but how we can collectively enable modalities of incipience that glitch the urge and ambition to predict. What temporal forms are possible in and through #datapolitik other than the strict Aristotelian teleology implicit in operational logistics and predictive analytics. Interestingly, this same issue was taken up (within a different media context) by Nouvelle Vague filmmakers of the 1950s who developed technical editing practices (most famously, Jean-Luc Godard’s jump cut) to challenge the dominance of Aristotelian mimesis that governed much French filmmaking at the time. I’ve written about this elsewhere, but the basic point was that one of the sites of critical attack that emerges from the Cahiers du Cinema directors of the period was the reigning commitment to Aristotelian narrative emplotment defined in terms of the right action at the right time. And these figures, rather than simply denouncing cinema, took on the challenge of altering what the medium could do and thus what the associative powers of the technical milieu could be.

Now, this example is one that I like to bring up often because, being a cinephile, I’m always enthused by it. But I think it’s a relevant example not because it is especially instructive of what to do, but because it points to the possibility of forms of incipient collectivization that both acknowledge and engage the limits of technical practices and technical media and push these limits beyond the extant modes of media handling. It is at this juncture of techno-human incipience that I want to explore emergent forms of political organization and action, solidarity and emancipation in #datapolitik. But this is far from a solitary task. The age of individualized, monastic, theorizing and critique is over. New collective forms must emerge from collectivized thinking and doing, and this can’t simply be an abstract endeavor. The pen was mightier than the sword; but the mouse (or trackpad, or touchpad) is mightier than the pen or the sword ever was. Let’s see what queer adjacencies and practices are now possible with the mouse that neither pen nor sword could enable.


  1. James Chandler, An Archaeology of Sympathy: The Sentimental Mode in Literature and Cinema (University of Chicago Press, 2013). 

  2. Graham Burchell, Colin Gordon, and Peter Miller, eds., The Foucault Effect: Studies in Governmentality (University of Chicago Press, 1991). 

  3. Sharon Krauss, Freedom Beyond Sovereignty: Reconstructing Liberal Individualism (University of Chicago Press, 2015). 

  4. Miriam Bratu Hansen, Cinema and Experience: Siegfried Kracauer, Walter Benjamin, and Theodor W. Adorno (Berkeley: University of California Press, 2011), 36. 

  5. Davide Panagia, Impressions of Hume: Cinematic Thinking and the Politics of Discontinuity (Rowman & Littlefield Publishers, 2013); Davide Panagia, “A Theory of Aspects: Media Participation and Political Theory,” New Literary History 45, no. 4 (2014): 527–48. 

  6. Orit Halpern, Beautiful Data: A History of Vision and Reason since 1945 (Duke University Press, 2015). 

  7. Minsoo Kang, Sublime Dreams of Living Machines (Harvard University Press, 2011); Adelheid Voskuhl, Androids in the Enlightenment: Mechanics, Artisans, and Cultures of the Self (University of Chicago Press, 2013). 

  8. Stanley Cavell, The World Viewed: Reflections on the Ontology of Film (Harvard University Press, 1979). 

  9. Alain Brossat, “La notion de dispositif chez Michel Foucault,” in Miroir, appareils et autres dispositifs, ed. Soko Phay-Vakalis (Editions L’Harmattan, 2009), 201. 

  10. Jacques Rancière, Aisthesis: Scenes from the Aesthetic Regime of Art (Verso, 2013), 193. 

  11. Rita Raley, ed., “Dataveillance and Countervailance,” in “Raw Data” Is an Oxymoron (MIT Press, 2013), 121–45. 

  12. Roger Clarke, “Information Technology and Dataveillance,” Commun. ACM 31, no. 5 (May 1988): 498–512. 

  13. Grégoire Chamayou and Kieran Aarons, “Fichte’s Passport - A Philosophy of the Police,” Theory & Event 16, no. 2 (2013). 

  14. Jacques Rancière, Rachel Bowlby, and Davide Panagia, “Ten Theses on Politics,” Theory & Event 5, no. 3 (2001). Available online as a PDF

  15. W. Ross Ashby, An Introduction to Cybernetics (Martino Fine Books, 2015), 9. 

  16. Simondon asserts this: “On peut donc affirmer que l’individualisation des êtres techniques est la condition du progress technique. Cette individualization est possible per la recurrence de causalité dans un milieu que l’être technique creé autour de lui meme et qui le conditionne comme il est conditionné par lui. Ce milieu à la foie technique et naturel peut être nomé milieu associé.” Gilbert Simondon, Du mode d’existence des objets techniques (Editions Aubier, 2012), 70. [“We can therefore affirm that the individualization of technical beings is the condition for technical progress. This individualization is possible through the recurrence of a form of causality within a milieu created by the technical being, with which it surrounds itself, and that it conditions as well as being conditioned by it. This milieu at once technical and natural may be called an associational milieu.” DP, trans.] 

  17. Jason Frank, Constituent Moments: Enacting the People in Postrevolutionary America (Duke University Press, 2010). 

  18. Richard Grusin, “Radical Mediation,” Critical Inquiry 42, no. 1 (September 1, 2015): 129. Available online as a PDF