A review of Harnessing Non-Modernity: A Case Study of Artificial Life, by Christine Aicardi.
Twenty-five years in, and Artificial Life soldiers on, a research field bent toward that receding horizon practitioners like to term “life.” In her ambitious doctoral thesis, Christine Aicardi offers readers a grand tour of a big, noisy, and heterogeneous community of researchers who build and operate robotic and virtual computational systems in order to probe the characteristics of life in digital media. Neither an ethnographic account nor a bird’s-eye view of a dispersed discipline, Aicardi maps the terrain of ALife research. She then zooms in, anchoring her macro-study in an ethnographically-oriented portrayal of the Centre for Computational Neurosciences and Robotics (CCNR) at the University of Sussex. In her update of prior ethnographies of Artificial Life research (the work of Stefan Helmreich and Lars Risan are both touchstones), Aicardi argues that “the Artificial Life research community shares a culture of simulation, and more precisely, a culture of ‘agency-rich’ simulation. Their defining cultural trait, their cultural ‘glue’, is the conviction that synthetic systems, which design involves computer simulations, is relevant to understanding life and its distinctive characteristics… The resulting Artificial Life systems are expected to display agency in a strong sense; that human authors willfully engage their empowered nonhuman creations into collaborative play” (p.15).
After an introductory chapter, Aicardi introduces readers to the broad contours of the field by analyzing journal articles and reporting on international conferences. Aicardi here recognizes the importance of play to ALifers. As avowed SF fans and skilled RPG gamers (as, notably, Aicardi also is), simulation is second nature to these scientists, who play with their simulations in order to elicit agential interactive responses from their machines: “In a culture of simulation,” she writes, “it seems that objects-to-think-with are also objects-to-play-with” (p.68). Chapters 2 and 3 focus on the at-once adhesive and hybridizing effects of such simulations. Andrew Pickering’s open-ended and “non-modern” “dance of agency” and Peter Galison’s “trading zones” inform her analysis. Such simulations, she concludes, function as “messiness engines” (pp.17, 160) that anchor a diverse field while maintaining its instability.
Turning next to her field-site at the University of Sussex, Aicardi examines the most intellectually prominent ALife community, asking what is distinctive about it. Based on attendance at conferences, interviews with scientists, online listservs, and observation, she concludes that Sussex ALifers are marked and sustained by an elitist enrollment that privileges reflexivity, critical thinking, autonomy, and individualism, as well as a mentor-protégé pedagogical model that fosters loyalty among its members. Further, Aicardi diagnoses three manifestations of an interdisciplinary approach to ALife: (1) a phenomenological pragmatics among researchers who recognize first-person lived experience as a form of scientific data; (2) a historical sensibility that claims mid-century British cyberneticians as predecessors; and (3) an overlapping of scientific and artistic rationales and approaches to the question of Artificial Life. I here focus on the first two criteria.
Aicardi worries that the first interdisciplinary mode, the influence of continental philosophy, might be a theoretical weakness in her work, astutely noting that the same continental philosophers (Heidegger, Husserl, Merleau-Ponty, Bergson) inform her work and that of her interlocutors. What is to be made of the importance of first-person experience as a methodological approach for both the Sussex brand of ALife and for laboratory studies? How to make sense of the fact that her informants recruit Shapin and Schaffer to justify their own methods? Far from being a weakness in her work, this methodological looping seems key to her story, and I hope she will ruminate further in her future publications on how this shared methodological approach might reveal something about both ALife and ethnography as research practices — what might fieldwork be a simulation of? How can ALife shed light on our own approaches as STS scholars? Is Aicardi, like her informants, at play?
In the origin story Sussex ALifers tell, they profess kinship with prominent British computer scientists and cyberneticians, among them Alan Turing, Grey Walter, Ross Ashby, Gordon Pask, and Stafford Beer. What correspondences do they draw between their simulations and the earlier cybernetic menagerie of tortoises, homeostats, and ears? Aicardi interprets her informants’ amateur historical inclinations, in part, as a way of rejecting a more recent provenance of ALife that begins in Los Alamos in 1987. She also suggests that such stories allow them to recognize themselves as inheritors of a much longer history she shorthands as “mind-as-machine.” I am curious, however, as to what assumptions about mind, machine, life, and simulation are assumed by the stories ALifers tell themselves about themselves. To borrow a phrase from Christopher Kelty’s ethnography of computer geeks, what do such “usable pasts” — that is, folklore that embeds values, attitudes, beliefs and imaginations — tell us about how AL researchers understand both the A and the L in their name (Two Bits: The Cultural Significance of Free Software and the Internet. Durham: Duke University Press, 2008)?
To begin addressing that question, let me first return to another question, one posed by Lars Risan in the title of an unpublished paper delivered at a 1997 ALife conference, and later restated by Aicardi in her thesis: “Why are there so few biologists here?” Indeed, even when theoretical biologists and synthetic biologists make cameo appearances, biologists — and life — are strangely absent from Aicardi’s tale. Her interlocutors may be seeking to simulate the characteristics of life, but what those characteristics might be remains unclear. As she reports, many scientists she spoke with espouse the catchphrase “life = cognition,” where cognition encompasses any flexible response to environmental stimuli. What of the many other characteristics life exhibits: metabolism, adaptation, homeostasis, reproduction, organization, mutation, growth? It seems that, once again, life has collapsed onto information. If ALife once sought to simulate life because intelligence had proven too intractable a problem for AI, contemporary ALife seems to have fallen back upon — embraced! — artificial intelligence. Perhaps ALife is now post-life, its researchers captured by the agency and playfulness of their own simulations, forgetting what it was they were trying to simulate in the first place.
Sophia Roosth
Department of the History of Science
Harvard University
roosth@fas.harvard.edu
Sources
Bibliographic analysis of primary sources (peer-reviewed journal articles, websites, listservs, conference proceedings)
Participant-observation and semi-structured interviews with prominent Artificial Life scientists and affiliates of the University of Sussex CCNR
Dissertation Information
University College London. 2010. 362pp. Primary advisor: Joe Cain.
1 comment