Code Roundtable: Bernard Dionysius Geoghegan

Bernard Dionysius Geoghegan (Department of Digital Humanities, King’s College London) is a media theorist and historian of science researching how digital technologies shape science, culture, and the environment. You can read more about his work here.

It is a great honor to have this roundtable on my book Code: From Information Theory to French Theory appear in the History of Anthropology Review. I extend a warm thanks to the commenters and editors for this opportunity.

Responding to a discussion of Code in HAR feels like a homecoming. Code hinges on the idea that anthropology plays a special role in the intertwined history of informatics, data-driven analytics, and the human sciences. As Rebecca Lemov has noted, anthropologists were early proponents of data-driven human sciences, and their reflections on computing did much to animate mid-century thinking about “new media” in an era of globalization and political strife.^[1]Rebecca Lemov, Database of Dreams: The Lost Quest to Catalog Humanity (New Haven: Yale University Press, 2015). See also Ronan Le Roux, “Lévi-Strauss, Une Réception Paradoxale de La Cybernétique,” L’Homme, no. 189 (2009): 165–90; Poornima Paidipaty, “‘Tortoises All the Way down’: Geertz, Cybernetics and ‘Culture’ at the End of the Cold War,” Anthropological Theory 20, no. 1 (2020): 97–129; and Ginger Nolan, The Neocolonialism of the Global Village (Minneapolis: University of Minnesota Press, 2018). Anthropologists’ key role organizing, facilitating, and popularizing the so-called Macy Conferences on Cybernetics is exemplary of the broader importance anthropologists and ethnographers played in debates around science and technology from around 1930 to 1970. The cybernetics conferences were not a mere product of anthropologists and their fellow-travelers in fields such as linguistics wanting informatic enlightenment from engineers and mathematicians. Of equal importance were mathematicians and engineers who turned to anthropology for help thinking about the human significance of advances in computing and communications engineering. Likewise, officers at philanthropic institutions such as the Josiah Macy Jr. and Rockefeller Foundations saw cooperation across natural and human science, mediated by anthropology, as vital to developing technical (and technocratic) knowledge that could be applied to improving the human condition. Thus, in anthropologists’ struggles around cybernetics, we find a point of entry into a history of larger debates around informatics, technocracy, the role of theory in the human sciences, and scientists’ efforts to counter modern political violence.

Anthropologists also made explicit a concern for planetary politics that frequently remained implicit in the work of many of their colleagues in the natural sciences, such as Warren Weaver and Norbert Wiener. While a thorough examination of these thinkers’ work shows an enduring and deeply felt concern for the Global South and the role of technology in shaping a global political community, only intermittently did Weaver and Wiener make these concerns explicit (we can think here of Weaver’s role in projects like the Green Revolution or Wiener’s work in India as exemplary of their concern for how technical advances would shape the autonomy and well-being of formerly colonized lands). By contrast, anthropologists with ties to cybernetics explicitly thematized the claim that cybernetics had a role to play in re-evaluating the relation among modern and “indigenous” knowledges in an era of decolonization. Indeed, while so-called “German media theory” is often maligned for an overly technical and apolitical outlook, media theorists Ute Holl and Erhard Schüttpelz were among the first writers to reconsruct the centrality of colonial and postcolonial cultures to cybernetics (and I hope my book has expanded on that literature).^[2]Ute Holl, Kino, Trance & Kybernetik (Berlin: Brinkmann & Bose, 2002); Erhard Schüttpelz, Die Moderne Im Spiegel Des Primitiven: Weltliteratur Und Ethnologie (1870-1960) (München: Wilhelm Fink, 2005). This means that anthropology not only figures decisively in the rise of cybernetics, but also casts unique light on geopolitical aspects of that rise.

Nonetheless, when I first started working on this book, “anthropology,” “human sciences,” and “sciences of man” were not widely seen as crucial to histories of computing or cybernetics. The gifted and insightful historian of science, Steve Heims, frequently presented human scientists as an audience for a cybernetic kernel emanating from natural science and engineering.^[3]Steve J. Heims, Constructing a Social Science for Postwar America: The Cybernetics Group (1946-1953) (Cambridge, MA: MIT Press, 1991). Media theorists and historians were more likely to associate cybernetics with antihumanism, posthumanism, and transhumanism—that is, fields linked to vague ambitions to overcome the human and humanism. As the history of anthropology teaches us, however, the critique of the human and humanism are central concerns of anthropology and the human sciences. Far from desperately clutching onto an antiquated notion of the human, anthropologists interested in cybernetics, such as Bateson, Mead, and later, Lévi-Strauss, embraced its potential to dissolve idealist constructions of a self-possessing, liberal humanist subject.

The turning point in my research on cybernetics and anthropology came from my efforts to understand some of the institutional arrangements and projects that preceded the organization of Macy Conferences. Examining media-technical research and interdisciplinary research projects organized by Mead and Bateson from the 1930s and early 1940s, it appeared to me that much of what has been called “posthuman” about cybernetics was first sketched in anthropological theory, only later winning some interest among engineers and mathematicians. The critique of consciousness, agency, intentionality, individuality, gender, embodiment, and the liberal self, identified with cybernetics by thinkers like Friedrich Kittler, Donna Haraway, and Katherine Hayles, was inherited from progressive anthropologists who viewed indigenous cultures as actively articulating that critique in everyday practice (they would make similar claims about persons dubbed pathological, e.g., schizophrenics and delinquents).^[4]For all that, these authors also identified cybernetics with an ambivalent relation to the liberal humanist subject and the human sciences. See, for example Donna Haraway, “The High Cost of Information in Post-World War II Evolutionary Biology: Ergonomics, Semiotics, and the Sociobiology of Communications Systems,” Philosophical Forum 13, no. 2–3 (1981-1982): 244–78; and N. Katherine Hayles, How We Became Posthuman: Virtual Bodies in Cybernetics, Literature, and Informatics (Chicago: University of Chicago Press, 1999), 85–87. In this revised historical timeline, anthropology was not the conceptual residue of a program superannuated by cybernetics but, on the contrary, an instrument of the radical epistemologies associated with it.           

As already alluded to, a great deal of Code aims at recovering a place for the anthropologists in cybernetics who, much as they shaped the early field, tended to downplay or efface their role in its development. I am thinking chiefly of Mead and Bateson, to some extent Lévi-Strauss, as well as an auxiliary community of linguists, psychoanalysts, semioticians, and sociologists who immersed themselves in anthropological theories—thinkers like Roman Jakobson and Luce Irigaray. I am grateful to the roundtable for recognizing that the book, though critical in its evaluation of how cybernetics and anthropology intersected, is also a circumspect appreciation of anthropology. A major goal of this book was to outline some of the ethical commitments and political dilemmas that led these thinkers to cybernetics; to consider how political problems, rather than technical questions of engineering, directed their interest in new technologies of computation and communication. Recovering those commitments also demands attention to how human problems informed engineers’ understanding of their own work’s social meaning.  The shorthand for that human problem is “technocracy,” or better yet “liberal technocracy,” although the precise formulation of the problem could be differentiated into US and French inflections, as well as Progressive Era, interwar, wartime, and Cold War variations.

Epistemic prestige and funding priorities gave the hard sciences and engineering a special privilege in twentieth century US liberal democracy, creating strong financial and political incentives for anthropologists, linguists, and other human scientists to tout the “formalizing practices” of engineering as their own.^[5]On formalizing practices, see Martha Lampland, “The Illusion of Abstraction,” Comparative Studies in Society and History 65, no. 1 (2023): 4–26. Based on my study of letters and unpublished documents, it seems Mead and Bateson downplayed their contribution to the cybernetics movements, giving their mathematical and engineering colleagues a credit that sometimes outstripped their own conceptual innovations. Perhaps this was due to their self-effacing personalities, or the ethnographer’s propensity for crediting insights to what they found out in the field (and, in many respects, the Macy Conferences were a field experiment in observing behavior in small groups).^[6]On social scientists from Mead’s circle (including Mead herself) using their interlocutors as a model of interaction in small groups, see Jamie Cohen-Cole, The Open Mind: Cold War Politics and the Sciences of Human Nature (Chicago: The University of Chicago Press, 2014), 104–38. But it seems to me that Mead and Bateson were open to a broader tendency in the period to view natural sciences and engineering as the premier fora for scientific invention and formalization. In any case, these sorts of scholars found in mathematician Norbert Wiener (credited with coining the term “cybernetics”) a supremely ambitious synthesizer, eager to claim credit for his unique potential to bridge all manner of scientific, political, ethical, and epistemic questions. But without the likes of Mead and Bateson, or humanistically and socially-minded foundation officers like Frank Fremont-Smith and Lawrence Frank, it seems to me there would have been no “cybernetics moment.”^[7]On the cybernetics moment, see the indispensable book Ronald R. Kline, The Cybernetics Moment: Or Why We Call Our Age the Information Age (Baltimore: Johns Hopkins University Press, 2015).

I am gratified by the roundtable’s openness to the proposition that anthropology deserves a more prominent place in the founding of cybernetics and the modern sciences of communications. It brings me satisfaction to read that my analysis prompted Martha Lampland to reconsider the work of anthropologists such as Mead and Bateson in connection with cybernetics (and she is surely right that scientific management deserves a key place in a fuller history of technocracy and the human sciences). I am likewise gratified by historian of science Ekaterina Babintseva’s eagerness to credit the book with giving a place to thinkers such as linguist Roman Jakobson and anthropologist Claude Lévi-Strauss in “twentieth-century information sciences” (though in this regard my work also follows in the steps of Ute Holl, Christopher Johnson, Ronan Le Roux, Céline Lafontaine, and Lydia Liu, among others).^[8]Ute Holl, Kino, Trance & Kybernetik (Berlin: Brinkmann & Bose, 2002); Céline Lafontaine, L’Empire cybernétique: Des machines à penser à la pensée Machine (Paris: Seuil, 2004); Lydia H. Liu, The Freudian Robot: Digital Media and the Future of the Unconscious (Chicago: University of Chicago Press, 2010); Christopher Johnson, “‘French’ Cybernetics,” French Studies 69, no. 1 (2015): 60–78; Ronan Le Roux, Une histoire de la cybernétique en France (1948-1975) (Paris: Classiques Garnier, 2018). To some extent, the scope of my claims rests on seminal critiques of structuralism launched by Henri Lefebvre, Simone de Beauvoir, Michel Foucault, Jean-Paul Sartre, and Félix Guattari, all of whom found in structuralism a poorly masked avatar of US technocratic reason. As Ted Underwood has playfully asserted, “In a just world, every article about GPT-4 would nod toward Barthes and Foucault,” a point also made recently by Bernhard J. Dotzler and Henning Schmidgen in their recent and provocative book on the technical ideals inflecting Foucault’s discourse analysis.^[9]Ted Underwood, “The Empirical Triumph of Theory,” In the Moment (blog), June 29, 2023; Bernhard J. Dotzler and Henning Schmidgen, Foucault, digital (Lüneburg: Meson Press, 2022). Alternatively, as Samuel Gerald Collins puts it in his response, today “We are surrounded by multi-agent systems at every moment in our lives,” which invite explanation in terms suggested by structuralism, post-structuralism, and the distributed agencies adduced by Mead’s and Bateson’s systems-oriented ecologies and kinships. In a sense, besides recovering anthropology’s indispensable role in the history of cybernetics, Code offers clues for understanding how twenty-first-century phenomena like LLMs are belatedly catching up with the ambitions and imagination of mid-twentieth-century anthropology.

I am flattered by Collins’s generous assertion that “Geoghegan’s book demonstrates […] that we are all heirs to ‘code.’” He identifies, here, a “we” that I very much had in mind when writing the book, but which I was cautious about naming. I see Code as our history, a singular yet broadly relevant history of data-driven human sciences, the far-reaching inheritances they bequeathed to critique and cultural theory, and aspirations they lent to technology and technocracy.^[10]I also see Code as reconstructing some of the proving grounds that came to privilege “third-party funding” and “digital humanities” as means for nudging the humanities and social sciences in our own moment. Code is a history of how the beneficiaries and participants in these developments shaped, inverted, complicated, and reoriented the resources of technocracy, sketching elements of social theory that would guide later movements like antipsychiatry and critical semiotics. What Haraway termed “the informatics of domination” was only secondarily and derivatively about technology.^[11]Donna Haraway, “Manifesto for Cyborgs: Science, Technology, and Socialist Feminism in the 1980s,” Socialist Review 80 (1985): 161. In the first instance, it was about strategies of knowledge, including the manner in which science, technocracy, technology, and capitalism remake social problems as questions of input and output, as information and communications. Those tasks cannot be realized by engineering or mathematics, understood as strictly technical enterprises. They demand social, political, and cultural agendas, developed through institutions and personnel. Universities, foundations, NGOs, museums, government programs, professors, social workers, grant officers, and experimental filmmakers are neither sufficient nor necessary to the informatics of domination. In the history considered by Code, these actors also play parts in actualizing that domination. To observe this fact and offer an outline of its dimensions is not to impugn particular individuals or institutions, and even less to assert an ethical distance between these historical subjects and us. Rather, I elicit a proximity between these actors’ situatedness and our own, to find in my subjects’ conceptual innovations paradoxes and complicities that we may also recognize in our own work.

The problem of how politically motivated funding shaped the logic of the human sciences, central to this book, is more urgent today than it was in the heyday of cybernetics. These thinkers’ struggles and inquiries prefigure our own. In many respects, their efforts to tackle the political force of informatics, to incorporate it into their work, reflects a faith in the relevance of the human sciences that exceeds what is commonly held by these fields’ practitioners today. Amrina Roysada generously credits the book with locating the “human sciences within a complex network of non-academic intellectuals and technocrats.” This phrase counterbalances—without contradicting—Lampland’s remark that the “cybernetic vision of the social world is flattened into a two-dimensional landscape of 1’s and 0’s” (much as I agree with her analysis). That is to say, if cybernetics seems sometimes to lead to a sort of flattened and depoliticized ontology, it also gave tools for organizing heterogeneous networks that cut across many professional and institutional lines. We can think here of the more upbeat evaluations of cybernetic possibilities offered in books by Eden Medina and Andrew Pickering.^[12]Eden Medina, Cybernetic Revolutionaries: Technology and Politics in Allende’s Chile (Cambridge, MA: MIT Press, 2011); Andrew Pickering, The Cybernetic Brain: Sketches of Another Future (Chicago: University of Chicago Press, 2010).

In the case of what I call “the cybernetic apparatus,” cybernetics also circumvented, at least somewhat, the strictures of US anticommunism. Cybernetics offered progressive scientists and policymakers in the US an informatic language that had an air of dispassionate rigor, even as they embarked on the analysis of politically contentious issues such as race, income inequality, and gender. These strange juxtapositions are part of the weird, often generative ambiguity of cybernetic ontology, of its ability to do multiple and contradictory things at once. It’s notable that some of the most successful appropriations of cybernetic discourse were by Marxist thinkers known for their lively and passionate takes on human culture. I’m thinking here of Roland Barthes’s essay on the press photograph or Stuart Hall’s essay on encoding and decoding.^[13]Roland Barthes, “The Photographic Message,” in Image, Music, Text, trans. Stephen Heath (New York: Hill and Wang, 1977), 15–31; Stuart Hall, “Encoding/Decoding,” in Culture, Media, Language (London: Hutchinson, 1980). Early communication studies theorists such as Wilbur Schramm and David Berlo had read Claude E. Shannon’s communication schema, with its tidy model of sender and receiver, encoder and decoder, as a model for staving off popular misunderstandings of mass communications.^[14]Wilbur Schramm, “Information Theory and Mass Communication,” Journalism Quarterly 32 (Spring 1955): 131–46; David K Berlo, The Process of Communication: An Introduction to Theory and Practice (New York: Holt, Rinehart and Wilson, 1960). Barthes and Hall identified this way of thinking as complicit in valorizing industrial and elite thought and a defense of its imposition on the public, and offered up an alternate reading in which so-called noise, error, and misunderstanding expressed the agency and critical impetus of readers and watchers. Against the temptation to secure a unity or coherence for cybernetics, Code understands these competing interpretations as elements in a common history—wherein something like “tirading,” understood as contest over meaning and imposition of difference, is integral to the establishment of what Galison termed a “trading zone.”^[15]Peter Galison, Image and Logic: A Material Culture of Microphysics (Chicago: University of Chicago Press, 1997).

While Code critiques agendas that shaped the rise of cybernetics, I hope it also conveys a degree of admiration, fidelity even, for my subjects’ intellectual efforts. I see my account as furthering and elaborating concerns central to my actors, like Mead, Bateson, and Lévi-Strauss. None of these scholars wished to be put on a pedestal outside history and critique. In essential comments, Roysada notes the “invisible labor” of Mead and Bateson’s assistant I Madé Kalér, work that made possible much of their proto-cybernetic, media-technical research in Bali with film, photography, and tape recorders. The invisibility of that labor was part of a larger erasure of the political and administrative violence of Dutch colonialism, up to and including the perverse way “Balinization” tended to stage Balinese tribes as idealized cultural forms for multimedia capture by Mead’s microphones and Bateson’s cameras. In their effort to re-situate phenomena like childcare and mental illness in an extended network of social and material relations, Mead and Bateson arguably erased some of the most decisive social and material relations from their analysis. To call attention to these oversights is to expand and better realize an impetus at work in their initial inquiries. Nonetheless, as Roysada notes, this does not entirely redress the invisibility.^[16]See also Amrina Roysada, “Unsung Native Collaborators in Anthropology,” Sapiens, October 13, 2022.

The entire colonial administrative apparatus, including researchers like Kalér, that buttressed proto-cybernetic research, shapes the archive but only wins fleeting representation in it. This archival problem is interesting, in large part, for its contemporary relevance. In our era of big data and digital archiving, Mead and Bateson are not only predecessors for their landmark efforts to document cultural patterns and submit them to a sort of quasi-algorithmic analysis. They are also predecessors in the temptation to erase or overlook the politics of archiving, the manner in which capturing and analyzing a subordinated culture is never innocent or unproblematic. To see Mead and Bateson, or Lévi-Strauss and Jakobson for that matter, as predecessors to the digital humanities (DH) is also to invite reflection on the complex and subtle forms of violence that may shape our work today. There’s nothing inherently pernicious about DH projects that promise to overcome the biases of Western canons by establishing expanded data-driven methods that allow for a certain degree of global and multicultural studies. Nor should DH projects proposing to use digital arts and media to represent and promote indigenous values be viewed entirely in terms of the administrative and archival exercises of earlier scholars. Nonetheless, these kinds of ambitions for using computing and its offshoots to better include and recognize subordinate cultures are broadly familiar from the kinds of projects Mead, Bateson, and Lévi-Strauss imagined in their day.

Perig Pitrou calls attention to what was among my most enduring interests in writing this book: the ambivalent but integral role of instruments in the cybernetic apparatus. The book aims at documenting what Pitrou calls “the intricate and ambivalent role of technical objects and devices” in knowledge production, including in the human sciences. Pitrou also recalls one of the technical devices I discuss, the Human Relations Area Files (HRAF), installed at the laboratory founded by Lévi-Strauss. Pitrou writes “I worked ‘next to’ [the files], but not directly ‘with’ them.” He adds “it would be instructive to evaluate the actual knowledge produced by this apparatus.” Indeed, many of the human scientists I discuss worked alongside scientific media and instruments, referenced them, cited them, but the extent to which they actually influenced the core of their research findings is an open question. At least one scientist who worked alongside my research subjects told me, confidentially, they never really understood the technical aspects of things like information theory. Bateson’s double-bind theory of mental illness or Lévi-Strauss’s proposed analogy between electronic sciences and indigenous thought in The Savage Mind stand out as the kinds of accounts that relied more on informatic inspiration than necessarily working with computers or deductive mathematics. Nonetheless, it seems to me that their accounts were deeply marked by information sciences that impart analytical frames emphasizing the formal analysis of cultures as communicative systems and so on.

For more hands-on applications of this sort to which Pitrou alludes, we could look to applications of information processing by Bateson’s interlocutors at Rockland Hospital (as mentioned in my book and explored in more detail by Seth Watter), to Lemov’s studies of the social relations program at Harvard, and to recent applications of data-driven computing to Lévi-Straussian analysis of myths.^[17]Seth Barry Watter, “Interaction Chronograph: The Administration of Equilibrium,” Grey Room 79 (Spring 2020): 40–77; Lemov, Database of Dreams; and Albert Doja, Laurent Capocchi, and Jean-François Santucci, “Computational Challenges to Test and Revitalize Claude Lévi-Strauss Transformational Methodology,” Big Data & Society 8, no. 2 (July 1, 2021): 20539517211037862. In each case, we find a practical and scholarly enactment of some of the same principles of predecessors in the cybernetic human sciences—for example, a presumption of computational patterns and informatic reason undergirding indigenous thought or mental states. Even as harsh a critic of cybernetic human sciences as Félix Guattari shows a deep and abiding—albeit materialist, and at times ironic—alliance to the kinds of cybernetic analogies populating Lévi-Strauss’s work and an interest in the kinds of media-technical investigation of mental illness spearheaded by the Palo Alto Group.^[18]For some of these cybernetic analogies, given an ironic and materialist twist, see Félix Guattari, “Machine and Structure,” in Molecular Revolution: Psychiatry and Politics, Peregrine Books (New York: Penguin, 1984), 111–18. At La Borde, Guattari took up a number of methods with hallmarks of the Palo Alto Group’s work, including the use of recording media in an analysis aimed at putting the therapist and patient on a more or less even playing field. A brief mention of these methods can be found in Gary Genosko, “Introduction to Félix Guattari’s ‘Project for a Film by Kafka,’” Deleuze Studies 3, no. 2 (2009): 146. And Jakobson, of course, worked at MIT with sound processing expert Gunnar Fant in his attempt to uncover informatic principles in phonemic distribution.^[19]Roman Jakobson, Gunnar Fant, and Morris Halle, Preliminaries to Speech Analysis: The Distinctive Features and Their Correlates (Cambridge, MA: MIT Press, 1963).

In a similar vein of ambivalence, Babintseva notes that it is unclear how much human scientists in my book engaged with the notions and inflections peculiar to cybernetic and information theorists, such as “prediction.” This is, indeed, a thorny topic worthy of further study. To the present day, notions of statistical predictability sketched in work of Wiener and Shannon figure prominently in a wide arrange of computational and DH projects. Just recently Ted Underwood related to me his efforts to correlate computational predictability in novels with experiences of surprise by readers. There is a long history of these sorts of applications, as when Jakobson sought to determine the likelihood or statistical probability of transitions between phonemes. These kinds of efforts are indebted to the work of Wiener and Shannon—but to what extent do they belong to the same conceptual program? How would sameness or analogy be established—by similar techniques, by shared mathematics, by common conceptual assumptions? It is a complex question that recurs throughout the history of ideas circulating among the cyberneticians.

Indeed, within a single Macy meeting it was not uncommon to see multiple and irreconcilable definitions attach to key terms such as “information,” without discussants reaching final consensus on what their shared vocabularies denote. Code makes the analytical choice to not impose a final or stable meaning in such cases. Collins writes, correctly I think, that “this is not a book about who influenced whom,” it is more focused on “multiple […] feedback loops” and the diffuse effects of political economy and imperialism that bound twentieth-century natural, technical, and human sciences together in cybernetic figures and informatic discourse. What does seem clear to me—and what I try to suggest in the book—is that this ambiguous rapport between technical and human sciences, including invocations of technical instrumentation and knowledge as idealized epistemic things, reflects organized efforts to accommodate the human sciences to technocratically-minded social and political ends.

Even if the effort to reconcile these diverse sciences was ambivalent, superficial, or confused, a history of the effort at such a communion offers lessons for our political and technical present. Too often, critical theory and cultural theory have been imagined as standing in opposition to technocracy, working to undermine its more facile assumptions. An argument of Code is that these theories are implicated in complex and variable ways in technocracy, and this involvement is part of what makes these theories relevant and responsive to the contemporary world. I thank the reviewers and HAR for granting time, space, and words to exploring that implication.