The Cybernetic Brain

by Andrew Pickering

  It is true, I think, as I said earlier in this chapter, that Western educational systems are strongly oriented toward the transmission of positive knowledge, and that this hangs together with a modern ontology of knowability, controllability, enframability. But, oddly enough, there are few people who would defend the proposition that this is a total description of the world. Everyone knows that surprise is a distinguishing mark of science, and that unintended consequences go with engineering initiatives. At the moment, however, such knowledge is, well, marginal. It doesn't amount to much; it continually disappears out of the corner of our eyes. And my suggestion is that it does not have to be this way, even in our pedagogical institutions. We could bring unknowability into focus for our children.

  How? Here I think first about what I have been calling the discovery of complexity. Cellular automata are not at all difficult to draw—by hand, never mind with a computer—and even young children can enjoy the patterns they make. I am sure that there are already schools that incorporate them in their curriculum, but it seems highly likely that they do so as a subfield of mathematics instruction—here is yet another branch of mathematics, in a sequence that would include arithmetic, geometry, groups, or whatever. My suggestion is that one could incorporate cellular automata into the curriculum differently, along the lines of how I have discussed them here, as ontological icons, little models of a world graspable in terms of a nonmodern ontology.17That would be the point of learning about them in the sort of courses I am imagining. Many of the artifacts we have examined along the way would likewise be discussable in this way at school. Gysin's Dream Machines might similarly offer a way into thinking about the explorability (rather than the givenness) of our perceptions and, by extension, our selves. School trips might include going to look at interactive art in local museums, again explicitly discussed as ontological theater, and not: here's a Rembrandt, here's a Picasso, here's a tortoise.

  Beyond that, of course, the history of cybernetics (and all that I assimilated to it a few pages ago) offers a wonderful and teachable set of examples that show that we are perfectly capable of going on in a world of unknowability and becoming, that there is nothing paralyzing about it. The more musical and theatrical students might enjoy playing a Musicolour machine (simulated on a computer) more than the trombone, say. Do-it-yourself kits for making tortoiselike robots are relatively cheap, and students could explore their emergent patterns of interaction. One can easily simulate and play with a multihomeostat setup. Some sort of quasi-organic architectural design methods might be more fun than many of the uses computers are conventionally put to in school. I am in danger of rehearsing the contents of this book yet again, and, well, more advanced students could also try reading it.

  The more I think about it, the more promising this idea becomes.18 In line with my earlier remarks on variety, I am certainly not proposing the unthinkable, to rid the curriculum of its modern elements. Nothing in this book, I repeat, threatens modernity, only its taken-for-granted hegemony and universality. I am suggesting the inclusion of a series of courses in schools and universities that would figure prominently in the curriculum, explicitly conceived as relating to a nonmodern ontology. I teach in a university, not a school, and I would be prepared to argue for one such course as a requirement for all freshmen, whatever their field.19

  I am, in the end, very attracted to this idea of systematically introducing students to a nonmodern ontology, beginning at an early age. If done right, it could easily produce a generation that would automatically say "wait a minute" when presented with the next high-modernist project of enframing, who would immediately see the point of Latour's "politics of nature," and who would, moreover, be in just the right position to come up with new projects that center on revealing rather than enframing. I would enjoy sharing a world with people like that.

  NOTES

  _ _ _ _ _

  Notes to Chapter 1

  1. On the Macy conferences—so called because they were sponsored by the Josiah Macy, Jr., Foundation—see especially Heims (1991) and Dupuy (2000). These meetings did much to sustain the development of American cybernetics under the permanent chairmanship of Warren McCulloch. Their proceedings were published from 1949 onward under the title Cybernetics: Circular Causal, and Feedback Mechanisms in Biological and Social Sciences,edited by Heinz von Foerster, joined from 1950 onward by Margaret Mead and Hans Lukas Teuber. The proceedings have recently been republished with valuable ancillary material by Pias (2003, 2004).

  2. See,forexample,Bowker(1993,2005),ConwayandSiegelman(2005),Cordeschi (2002), Dupuy (2000), Edwards (1996), Galison (1994), Haraway (1981–82, 1985), Hayles (1999), Heims (1980, 1991), Kay (2001), Keller (2002), Mindell (2002), Mirowski (2002), Richardson (1991), and Turner (2006). Much of this work focuses on the history of cybernetics in the United States, and there is relatively little discussion of cybernetics elsewhere.

  3. OnMcCulloch's enduring concern with the brain see McCulloch (1988, 2004), Heims (1991, chap. 3), and Kay (2001). In the 1930s,McCulloch worked at the Rockland State Hospital, a mental institution, and from 1941 to 1952, the key years in the initial development of cybernetics, he was director of research in the Department of Psychiatry of the University of Illinois Medical School in Chicago. In 1952, he moved to the Research Laboratory of Electronics at MIT, and he remained there for the rest of his working life. Wiener had no background in brain science, but the key paper in the genesis of his cybernetics, Rosenblueth, Wiener, and Bigelow (1943), moved immediately from the autonomous weapons systems he had worked on earlier to a general analysis of purposive behavior in animals and machines.

  4. McCulloch, Verbeek, and Sherwood (1966, ix–x) align Craik with James Clerk Maxwell, who numbered, among many scientific achievements, his theoretical analysis of the steam-engine governor.

  5. The Research Laboratory of Electronics was the successsor to the wartime Rad Lab at MIT; on its history see Wildes and Lindgren (1985, chap. 16, which includes much on Wiener, Shannon, and information theory but covers the work of McCulloch and his group in a single paragraph on p. 263, without using the word "cybernetics") and www.rle.mit.edu/about/about_history.html (which mentions the work of Norbert Wiener, Claude Shannon, Jerome Lettvin, and Walter Pitts but says nothing about McCulloch). The Illinois Biological Computer Laboratory and the Brunel Cybernetics Department are discussed further in chapters 4 and 6, respectively. In Britain, two other departments of cybernetics have existed, one at the University of Reading, the other at the University of Bradford (which closed down in summer 2006), but neither of these enters our story.

  6. For a recent encapsulation of the critique, see Harries-Jones (1995, 104): "Social critics across the political spectrum, from the left-leaning Jürgen Habermas to the right-leaning Jacques Ellul, have identified cybernetics and the systems concepts derived from it as the foremost ideology of the military-industrial technocracy that threatens our planet." The point of Harries-Jones's summary is to exempt Bateson from the critique. If one follows Harries-Jones's citations, one finds the young Habermas (1970, 77, 117–18) bivalent: enthusiastic about cybernetics as a medium of interdisciplinary communication between otherwise isolated scientific specialties, but anxious about a "cybernetic dream of the instinct-like self-stabilization of societies." The latter follows a list of possible future technological innovations that begins with "new and possibly pervasive techniques for surveillance, monitoring and control of individuals and associations" and is thus an instance of the "control" critique discussed at some length in the next chapter. We are, of course, monitored much more intensively than we were in the late sixties, but we do not have cybernetics to thank for that. One finds the later Habermas (1987) struggling with the sociology of Talcott Parsons, whose work did incorporate a cybernetic element but is hardly exemplary of the strands of cybernetics at issue in this book. Ellul (1964) mentions cybernetics from time to time, but only within the flow of his overall critique of the "technological society" and the reduction of hu
man life to "technique." Our examples of cybernetics fit Ellul's concept of technique exceedingly poorly.

  7. Though industrial automation is often cited as a paradigmatic instance of the Marxist deskilling thesis, the situation was often seen differently, at least in postwar Britain and France. There, automation was widely imagined to hold out the welcome promise of relieving the human race of many of the demands of repetitive wage labor (while giving rise to the "leisure problem"—what would people do with all their spare time?). More on this in chapter 7. The idea seems quaint and touching now—how could anyone think that?—but it is worth remembering that automation once had this positive valence, too.

  Notes to Chapter 2

  1. Continuing a line of thought from chapter 1, I should note that not everything that can be counted as cybernetic speaks to the ontological concerns at issue here. The interest, for me, of the British cyberneticians is precisely that their work draws one in an ontological direction. Again, there are aspects of the work of Walter, Ashby, and others that do not obviously engage with my ontological concerns, and I will not go into them in any detail. Ontological interest is thus another principle of historical selection that informs this book.

  2. "Modern" is a word with more resonances and associations than I need, but I cannot come up with a better one. This paragraph and those that follow try to define my own usage. In the history of Western philosophy, Descartes's articulation of the duality of mind and matter—the idea that they are made of different stuff—was a key moment in the history of modernity thus conceived. Likewise the Scientific Revolution: Newton's laws of motion as specifying regular properties of matter, independent of any human knower; also the Enlightenment's emphasis on reason as a special and defining property of the human. Our scientific and commonsense stances toward the world remain, I think, largely within the space thus defined, and hence they are "modern" on this definition.

  3. The dominant strand in twentieth-century philosophy of science portrayed the material world as a passive substrate which was the origin of observation statements, themselves either the basis for scientific induction or a deductive constraint on theory building. On such a view, there is no space to think about the performative aspects of the world and our constitutive entanglement with them in knowledge production.

  4. This characterization of modern science is crude, though not, I think, misleading. A properly nuanced and exemplified discussion would require another book. I should stress therefore that my object here is not to do full justice to modern science, but to set up an illuminating contrast that will pick out what I take to be key features of cybernetics—and these are the topic of this book. Thematizing this contrast serves to intensify my own tendency to dichotomize, while, from the cybernetic side, too, there are nuances that might be considered to blur the distinction—some of which are discussed below. Canguilhem (2005 [1983]) notes the different ontologies espoused by scientists and historians of science—one fixed, the other fluid—but does not develop this observation, treating it as simply a fact about the two camps; I thank Hans-Jörg Rheinberger for bringing this essay to my attention.

  5. This echoes an argument due to van Gelder (1995).

  6. From yet another angle, the suggestion in the following chapters is not that ontological claims came first and somehow gave rise to specific cybernetic projects, and neither is it the reverse. I am interested in how the ontology and these projects hung together and informed one another.

  7. Thus, the argument of The Mangle of Practice was that the material world as explored by modern science is itself an exceedingly complex system in Beer's terms. The scientists, however, adopt a peculiar stance in that world, construing their experience in terms of fixed entities (whose ascribed properties turn out themselves to evolve in practice). From my ontological perspective, the modern sciences read the world against the grain—and The Mangle undid this by reading the sciences against their grain.

  8. In the second postscript to The Mangle I floated the idea that my analysis might be a Theory of Everything, as the physicists like to say of their theories. Later I realized that the idea of a theory of everything could be understood in two rather different ways. The physicist's sense is that of a complete mathematical theory from which all worldly phenomena can be deduced, in principle at least. This is the sense of a theory of everything as the end of science—all that remains is to fill in the details. The nonmodern ontology of The Mangle instead suggests an always infinite horizon of constructive engagement with the world—in modern science as a never-ending finding out; in cybernetics as an endless staging of the ontology in this situation or that, in this way or that.

  9. In Walter's work especially, the interest in strange performances and altered states can be seen as part of an attempt to map input-output relations of the brain as a Black Box, and I can make another connection back to The Mangle here. In the first postscript to the book I commented on the apparently never ending range of material powers and performances that have manifested themselves in the history of science, and I contrasted this with the apparently unvarying historical parameters of human agency. In an attempt to symmetrize the picture, I mentioned a couple of examples of "non-standard human agency" but was unable to offer any very convincing documentation. I did not know then that cybernetics itself had plunged into explorations in this area, and I had forgotten that the sixties were a golden age for the same sort of experimentation in everyday life.

  10. Kauffman (1971) explicitly identifies this style of explanation as "cybernetic"; I return to his work below.

  11. The preceding paragraphs are not meant to suggest that the cyberneticians were the first to discover the existence of unpredictably complex systems. The three-body problem was around long before cybernetics. The argument is that the discovery of complexity in cybernetics emerged in a specific way within the frame of key projects. Furthermore, we will see that the cyberneticians addressed this problematic in very different ways from the authors usually associated with "complexity," including Kauffman and Wolfram. For popular accounts of the recent history of work on chaos and complexity, see Gleick (1987) and Waldrop (1992), neither of whom mentions cybernetics.

  12. For the pre-Katrina version of this story, see Pickering (2002); for a post-Katrina version, Pickering (2008b).

  Notes to Chapter 3

  1. I am indebted to Rhodri Hayward for much enlightenment on Walter and the relevant literature, including his own writings prior to publication. I have drawn upon two sets of archival holdings below, both in London: John Bates's papers at the Wellcome Library, and the papers of the Burden Neurological Institute at the Science Museum.

  2. This biographical sketch is based on Hayward (2001a) and Holland (2003).

  3. Walter spent the first six months of his Rockefeller fellowship in Germany, including a visit to Hans Berger in Jena (see below), and he was then based at the Maudsley and at the Hospital for Epilepsy and Paralysis in Maida Vale (Wellcome Library, GC/179/B.35, p W.2a; Walter 1938, 6). On Golla and the Maudsley, see Hayward (2004, forthcoming).

  4. Walter (1966) acknowledges the importance of his collaboration with the electrical engineer Geoffrey Parr. At the time of Walter's earliest EEG work, Parr was "head of the special products division of Edison Swan and could let me have special valves. . . . He also had his own workshop at home and every time he came to see me he brought a little present—a specially mounted and calibrated meter, or a set of matched resistors or an engraved graticule." Parr and Walter collaborated on the development of Britain's first commercial electroconvulsive therapy machine (below). As the editor of Electrical Engineering it was Parr who encouraged Ashby to play up the futuristic aspects of his first publication on the homeostat (next chapter).

  5. On Walter's place in the institutional history of EEG research, see Cobb (1981).

  6. Besides Golla and Walter, the research staff comprised Dr. Reiss, a physiologist, and Mr. Tingey, a chemist (salaries £350 and £250, respectively), plus a laboratory me
chanic, a laboratory boy, and a woman dispenser and clerk. A clinical director, Dr. E. L. Hutton, was appointed in November 1939 at £800 a year (on a par with Walter). In 1949, the Burden had eighteen beds for inpatients at eight guineas per week, and Golla himself established a clinical practice that "straddle[d] the boundary between neurology and psychiatry" (Cooper and Bird 1989, 15–16, 21, 61).

  7. Sometimes referred to as The Swerve of the Cornflake. The snowflake of the title is actually Walter's word for what would now be known as a fractal geometry, which allows characters in the book to travel into the future.

  8. Another story branches off here, that of Walter's determined attempts to correlate EEG spectra with human personality traits. See Hayward (2001a).

  9. Of course, any materialist conception of the brain is liable to destabilize this split, including the prewar experimental psychology mentioned below.

  10. Walter employed this tactic of electromechanical modelling frequently. Besides the tortoise, CORA, and the pattern-tracing device mentioned in note 11, The Living Brainincludes an appendix on an artificial nerve that Walter constructed (1953, 280–86). In 1954,Walter (1971 [1954], 40–44) displayed a gadget that he call IRMA, for innate releasing mechanism analogue, that he had built following a 1953 visit to the Burden by Konrad Lorenz. IRMA, like CORA. was intended as an attachment to the tortoise, and one begins to glimpse here a more extensive cyborgian project of building a realistic synthetic animal.

  11. Yet another of Walter's electrical contraptions further illuminated scanning. "The manner in which such a system might work in the brain is illustrated in a device which we had developed for quite another purpose, for reconverting line records, such as EEG's, back into electrical changes" (Walter 1953, 109–10). This device consisted of a cathode-ray oscilloscope coupled via an amplifier to a photoelectric cell which viewed the oscilloscope screen. The spot was arranged to oscillate along the x axis of the screen—scanning—and the circuitry was cleverly arranged so that along the y axis the spot would trace out the edge of any opaque object placed in the screen. Thus, a geometric image was converted into a time-varying electrical signal.