| My First Book and Where It Led N. KATHERINE HAYLES University of California, Los Angeles, USA | Media Theory Vol. 9 | No. 2 | 00-00 © The Author(s) 2025 CC-BY http://mediatheoryjournal.org/ |
Abstract
This essay traces my research trajectory from my first book, The Cosmic Web (1984) to my latest book, Bacteria to AI (2025a). It begins with what David Beer calls the “articulation problem,” that is, the limitations that arise when one speaks from within the field (physical, social, mathematical, theoretical) in which one is located. These include paradoxes of self-reference and the impossibility of constructing a logical system that is both true and complete (Gödel’s Theorem). In How We Became Posthuman (1999), the articulation problem surfaced within cybernetics in the form of second-order cybernetics, which recognized that the observer himself/herself is located within a system that can be observed in turn. With Science and Technology Studies (STS) well-established by the late 1990s, the articulation problem took a wider significance when STS researchers argued that humans themselves always speak from the human umwelt (world-construction), so that the “laws of nature” science identifies are never about nature as such but about our experience of nature. Thus the articulation problem morphed into the species-construction problem, which I explored in my last three books, How We Think (2012), Unthought (2017), and Bacteria to AI (2025a).
Keywords
research trajectory, articulation problem, species construction
David Beer’s insightful and sympathetic article on my first book, The Cosmic Web (1984), rightfully highlights the “problem of articulation” as a major conclusion from that study.
Although the problem has received complex treatments in mathematics, philosophy, cybernetics and systems theory, among others, it is not difficult to understand. If one is located within a field about which one is attempting to speak, inevitably certain paradoxes of self-referentiality will emerge; for example, one is both observer and observed.
Reflecting on Beer’s article, I am reminded of Aldous Huxley’s comment, when in middle age he was offered the chance to revise Brave New World, that he did not want to spend his later years correcting the mistakes of his youth; consequently, he opted to re-publish it unchanged. So with The Cosmic Web. Its shortcomings are even more obvious to me now than at the time, but I think I would not revise it if offered the chance. As Beer suggests, there is indeed a through line from The Cosmic Web to my later books. Beer stops his analysis with my second book, Chaos Bound: Orderly Disorder in Contemporary Literature and Science (1990), but if he had continued his line of thought, more parallels would have come to light. For those interested in my research trajectory, I use this opportunity to discuss how the “problem of articulation” continues to occupy me even as it mutated in form, including up to the present.
The Cosmic Web had the vision—and the ambition—of my later books, but not the method. Published in 1984, it attempted to tackle a huge problem with the (inadequate) resources at my command as a newly credentialed assistant professor. Had I been trained as an intellectual historian—say, like Martin Jay—I could have perhaps been able to trace more fully the idea of an encompassing field that cut across any number of areas in the early twentieth century, including linguistics, philosophy, mathematics, social theory, and the physical sciences. Certainly such an analysis would have included the work of Alfred North Whitehead, of which I was largely ignorant at the time. As it was, I was able to instance only a few parallels that had come into my ken, such as quantum field theory and Saussure’s linguistics. Nevertheless, I intuited that the paradigm shift toward the notion of a field was much broader, if for no other reason than that the ramifications seemed to me evident in both literature and science, two fields that at the time were basically not talking to one another. As the notion of an encompassing field expanded in the early 20th century, so too did the “problem of articulation” with its associated paradoxes of self-reference and resulting limitations on what could be said.
By the time of my second book on chaos and complexity (1990), the recognition of these limitations had already surfaced in such developments as Gödel’s Incompleteness Theorem, Cantor’s set theory, and perhaps most relevant to my work, the emerging field of Science and Technology Studies (STS). If the scientist speaks from within the field that contains him, strict objectivity—what Donna Haraway called the God’s eye view—is not possible. Researchers such as John Law, Bruno Latour, Andrew Pickering, Simon Schaffer, Steven Shapin, and feminist scholars such as Donna Haraway and Sandra Harding, made this point in many different contexts, with devastating effectiveness. Some scientists took note, especially after Gross and Levitt published Higher Superstition: The Academic Left and Its Quarrels with Science (1994), and Alan Sokal succeeded in getting a nonsense article published in the cultural studies journal Social Text (1996). The so-called “Science Wars” were on.
For me personally, this was a painful period, akin perhaps to an adult watching her two parents fight. My lineage included my scientific training in chemistry, which had imbued me with a rock-solid belief that scientific methods were—and are—our best means of reaching reliable and robust knowledges. But my training in literary studies, and the humanities more broadly, had also made me aware that all of us are immersed in our cultural contexts, which have enormous influences over what and how we think—scientists included. I felt that I had to come to grips with how scientific inquiries could at the same time be both culturally conditioned and powerfully reliable.
I resolved the dilemma through the concept of what I called “constrained constructivism” (1991). The core insight was that whatever theories the sciences construct are not reality as such but models of reality. Any scientist would of course acknowledge this immediately. The inability to have objective knowledge of the world as it “really” is means that it is not possible to say a model is “correct,” since there is no viewpoint from which to make such a judgment. The most that could be said—and this turns out to a highly nontrivial condition—is that a model is consistent, or not, with our experiences of reality. If a model can be shown to be not consistent, then it has been falsified and will end up in the dustbin of history, as many models have—the geocentric solar system, the ether theory, the belief that bumps on the head reveal intrinsic personality characteristics, and so forth.
Like almost all good ideas, I was of course not the only one to arrive at it; Karl Popper had spent his career making this point (1959; 1962). In my terms, constrained constructivism meant that STS research was correct. Scientific models are not only culturally influenced but culturally constructed, partaking of contemporary ideologies and fully immersed in what I called the cultural matrix. The crucial point is that they are not merely culturally constructed; they are also constrained by scientific methods that test whether or not they are consistent with reality as we experience it. As Sandra Harding observed when arguing for what she called “strong objectivity,” what would we do with an objective viewpoint even if it were possible? We are human and see with human eyes, think with human brains. Notwithstanding all the technologies we have developed that extend our perceptions beyond our natural senses, in the end we always and only process the world as humans.
Once I arrived at the concept of constrained constructivism, the Science Wars began to seem to me nothing more than a tragic misunderstanding. Scientists were understandably irate that cultural critics would cavalierly disregard the safeguards they had instituted to ensure that empirical claims would be fairly tested and adjudicated. What they failed to understand, however, is that the STS position that there are no “laws of nature” did not mean that our experiences can never be recognized as following regular patterns that can be reliably expressed in scientific and mathematical models. Rather, the point was that such observations are always indexed to our human ways of experiencing and understanding the world. The so-called “laws of nature” are not inherent in reality but rather in our experiences of reality.
My third and best-known book, How We Became Posthuman (1999), focused on telling three narratives: how information lost its body in the post-World War II period; how the cyborg was constructed technologically, culturally and socially in the same period; and how traditional ideas about the human were shifting in the face of information technologies, especially virtual and computational media. I argued strongly against fantasies of disembodiment, and posited against Shannon’s disembodied notion of information a counter-proposal by British physicist Donald Mackay for a fully embodied theory of information. In retrospect, my insistence on human embodiment would become crucial to my later work.
A central chapter analyzed the Macy Conferences on Cybernetics from 1946-1953. With cybernetics, the “articulation problem” resurfaced with the advent of so-called second-order cybernetics. Presented in Heinz von Forester’s seminal work with the punning title Observing Systems, the argument pointed out that from one perspective, an observer views a system, but from another perspective, the observer himself is inside a system that can be observed in turn. At last, here was a version of the “articulation problem” located within a scientific discipline itself, and I was delighted to expand on its implications.
Another influential version of the “articulation problem” circulating at the time appeared in Humberto Maturana and Francisco Varela’s Autopoiesis and Cognition: The Realization of the Living (1980), which argued that any system processes the world solely in terms of its own organization and structure. Autopoiesis and Cognition took the extreme position that no information from the outside ever enters a cognitive system, but only triggers a response; causation, in this view, is only in the eye of the observer, not objectively present in the system as such. Although I was sympathetic with the autopoietic approach, seeing in it another version of the “articulation problem,” I thought then—and still think—that eliminating causation results in unnecessarily convolutedrhetoric and twisted-pretzel thinking. Nevertheless, the plot was thickening; now what was at stake was not only cultural constructions but species constructions.
Niklas Luhmann used Maturana and Varela’s work as a basis for his social systems theory (1984), but he wisely changed the idea of informational closure to operational closure. In his theory, a system is always exponentially less complex than its surrounding environment. To keep itself from being overwhelmed, the system processes information only in terms of its own structure, and in this sense, remains operationally closed. It achieves a kind of wrenched referentiality, however, by proliferating differences within its own categories, thus mimicking the complexities of its environment, but in its own terms. An important point of departure for Luhmann, the necessary gesture that makes any system possible, is drawing a distinction between a system and its environment. We can think of this gesture in its simplest form as a line drawn on paper: on one side is the system, on the other its environment. Luhmann understood that by drawing the line, one not only creates a system but at the same time and with the same gesture, inevitably introduces a blind spot—that is, all the alternative ways in which that original line might have been drawn differently, resulting in different environments and different systems.
These developments and their implications provided the impetus to go beyond the conclusions of How We Became Posthuman to my later books. The “articulation problem” had morphed into what I might call the species cognition problem. Why stop with the influence of culture when the biological capacities of a species have an even greater determinate influence on how a species constructs (its version of) the world? Thus my focus shifted to the implications of human cognition in relation to the cognitive capacities of nonhuman living creatures.
The first point I wanted to make from this modified perspective was that, for humans, our biological, cognitive, and psychological capacities have, since the beginning of Homo sapiens, coevolved through strong feedback loops with our technologies. Following the lead of Bernard Stiegler in Technics and Time, 1 (1998), I called this evolutionary spiral technogenesis (thus combining the idea of technics with the genesis of the human). Analyzing the work of French anthropologist Leroi-Gourhan, Stiegler had made the case for prehistoric humans coevolving with their technologies. The domestication of fire, for example, changed the diets of early humans, resulting in changes in the jaw and mouth that later facilitated the emergence of language. The fabrication and transport of tools conveyed such strong evolutionary advantage that it facilitated bipedalism, which in turn extended and strengthened tool use, and so on. My focus was on contemporary technogenesis, for it seemed to me that human co-evolution with tools had become exponentially faster and stronger with the development of cognitive technologies, specifically computational media. These arguments comprised the core of How We Think: Digital Media and Contemporary Technogenesis (2012).
The next point concerned the nature of human cognition itself. In Unthought: The Power of the Cognitive Nonconscious (2017), I drew on research in neuroscience, cognitive science, psychology and other fields to argue that most of human cognition is not conscious. Thus it is necessary, despite centuries of commentary during which consciousness was considered virtually synonymous with cognition, to make a strong distinction between them. Moreover, empirical research from the late 1990s on had demonstrated that nonconscious cognition played central roles in human perceptions and actions. Operating much faster than consciousness and with the capacity to process information too dense and noisy for consciousness to apprehend, nonconscious cognition (or implicit cognition, as it came to be called) plays crucial roles in filtering and pre-processing internal and external information, which otherwise would overwhelm consciousness, resulting in psychic breakdown. Granting that consciousness has unparalleled abilities to analyze and focus on specific problems, most of human everyday activities are guided not by it but by nonconscious cognition.
The crucial implication I wanted to draw from this work was that a creature does not need to possess consciousness to have cognitive abilities. Indeed, in my most recent book Bacteria to AI: Human Futures with our Nonhuman Symbionts (2025a), I make the claim that all living creatures have cognitive abilities. As research on plant cognition and cognitions of microorganisms such as bacteria proceeds apace, this claim is sounding more and more like common sense. More recently, I have an article, ‘Modes of Cognition’ (2025b), identifying five criteria for a system to count as cognitive, showing first how the criteria apply to minimally cognitive organisms such as plants and microorganisms, and then using them with artificial intelligences such as Large Language Models to argue that they also count as (synthetic) cognitive systems.
For me the importance of this perspective lies in what it implies about the relation of humans both to the nonhuman creatures with whom we share the earth and the artificial cognizers we are now inventing. For millennia, humans have claimed superiority because of our cognitive abilities. We are not the most physically impressive species, but our cognition, so we claim, overcomes everything else, enabling us to invent machines faster than cheetahs, stronger than elephants, more agile than monkeys.
With the invention of synthetic cognizers, human evolutionary trajectories have entered a new and unprecedented era. No longer are we the supreme cognizers on the planet; computational media, algorithms, and AI can now match or surpass human achievements in many respects. It is time—indeed, past time—for us to reassess our relations with other cognitive entities, both nonhuman and artificial. We must understand more deeply and broadly all the ways in which human existence is possible only because of our nonhuman symbionts, from the bacteria in our gut without which we could not digest food, to the viruses in our reproductive organs without which we could not reproduce our species. This wholesale reassessment must also include our relations to the artificial intelligences we create. The results, I very much hope, will be to strengthen our realizations that we are not and never were self-sufficient, that symbiosis is pervasive throughout the living world, and that new forms of symbiosis with intelligent machines will be crucial to our human futures.
This is how the “articulation problem” grew from my first book and mutated over time, still guiding my research efforts and if anything becoming more urgent as its implications have broadened and deepened. I hope this account will show the accuracy of Beer’s insightful article on The Cosmic Web, as well as make additional connections with my research trajectory as a whole. Of course, this account is told from a retrospective viewpoint, and it therefore tends to impute more clarity and coherence than may have been present at the time. There were plenty of false starts, intuitions imperfectly grasped, and inadequate intellectual resources along the way. Nevertheless, in many ways I have not deviated from the vision I had as a young professional, hoping to use whatever talents I possess and my unusual multi-disciplinary training to make positive differences in the world.
References
Beer, D. (2025) ‘N. Katherine Hayles, The Cosmic Web and the Problem of Articulation’, Media Theory (this issue).
Gross, P.R. and N. Levitt (1994) Higher Superstition: The Academic Left and Its Quarrels With Science. Baltimore: Johns Hopkins University Press.
Harding, S. (1995) “‘Strong Objectivity’: A Response to the New Objectivity Question’, Synthese, 10(3): 331-349.
Hayles, N.K. (2025a) Bacteria to AI: Human Futures with our Nonhuman Symbionts. Chicago: University of Chicago Press.
———- (2025b) ‘Modes of Cognition: Implications for Large Language Models,’ Antikythera (May 2025), https://modesofcognition.antikythera.org/
———- (2017) Unthought: The Power of the Cognitive Nonconscious. Chicago: University of Chicago Press.
———- (2012) How We Think: Digital Media and Contemporary Technogenesis. Chicago: University of Chicago Press.
———- (1999) How We Became Posthuman: Virtual Bodies in Cybernetics, Literature and Informatics. Chicago: University of Chicago Press.
———- (1991) ‘Constrained Constructivism: Locating Scientific Inquiry in the Theater of Representation’, New Orleans Review, 18: 76-85. Reprinted in G. Levine, ed. (1993) Realism and Representation: Essays on the Problem of Realism in Relation to Science, Literature, and Culture. Madison: University of Wisconsin Press, pp.27-43.
———- (1990) Chaos Bound: Orderly Disorder in Contemporary Literature and Science. Ithaca: Cornell University Press.
———- (1984) The Cosmic Web: Scientific Field Models and Literary Strategies in the Twentieth Century. Ithaca: Cornell University Press.
Popper, K. (1962) Conjectures and Refutations: The Growth of Scientific Knowledge. London: Routledge.
———- (1959) The Logic of Scientific Discovery. London: Routledge.
Luhmann, N. (1984) Social Systems, translated by John Bodnarz, with Dirk Baecker. Stanford: Stanford University Press.
Maturana, H.R. and F.J. Varela (1980) Autopoiesis and Cognition: The Realization of the Living. Dordrecht: D. Reidel.
Sokal, A. (1996) ‘Transgressing the Boundaries: Towards Transformative Hermeneutics of Quantum Gravity,’ Social Text 46/47 (Summer): 217-252.
Stiegler, B. (1998) Technics and Time, 1: The Fault of Epimetheus. Stanford: Stanford University Press.
Von Foerster, H. (1981) Observing Systems. The Systems Inquiry Series. Cambridge MA: Intersystems Press.
N. Katherine Hayles is the Distinguished Research Professor at the University of California, Los Angeles, and the James B. Duke Professor Emerita from Duke University. Her research focuses on the relations of literature, science and technology in the 20th and 21st centuries. Her twelve print books include Postprint: Books and Becoming Computational (Columbia, 2021), Unthought: The Power of the Cognitive Nonconscious (Univ. of Chicago Press, 2017) and How We Think: Digital Media and Contemporary Technogenesis (Univ. of Chicago Press 2015), in addition to over 100 peer-reviewed articles. Her books have won several prizes, including The Rene Wellek Award for the Best Book in Literary Theory for How We Became Posthuman: Virtual Bodies in Literature, Cybernetics and Informatics, and the Suzanne Langer Award for Writing Machines. She has been recognized by many fellowships and awards, including two NEH Fellowships, a Guggenheim, a Rockefeller Residential Fellowship at Bellagio, and two University of California Presidential Research Fellowships. She is a member of the American Academy of Arts and Sciences. Her most recent book is Bacteria to AI: Human Futures with our Nonhuman Symbionts (2025).
Email: katherine.hayles@duke.edu
Website: nkhayles.com
Supplementary material
This is N. Katherine Hayles’ response to a commentary by David Beer published in the same issue of the Media Theory journal. Readers may also be interested in an interview between the two authors, which is available on the Media Theory blog. See:
Beer, D. (2025) ‘N. Katherine Hayles, The Cosmic Web and the Problem of Articulation’, Media Theory (this issue).
Beer, D. and N. Katherine Hayles (2025) ‘N. Katherine Hayles, In Conversation With David Beer’ (video and written transcript): https://mediatheoryjournal.org/2025/06/14/n-katherine-hayles-in-conversation-with-david-beer-video/
Leave a Reply