The Cybernetic Brain
Page 39
Second, I want to emphasize that with Eno's interest in cellular automata and complex systems we are back in the territory already covered at the end of chapter 4, on Ashby's cybernetics, with systems that stage open-ended becomings rather than adaptation per se. Indeed, when Eno remarks of It's Gonna Rain that "you are getting a huge amount of material and experience from a very, very simple starting point" (Eno 1996b, 3) he is singing the anthem of Stephen Wolfram's "New Kind of Science." In this sense, it would seem more appropriate to associate Eno with a line of cybernetic filiation going back to Ashby than with Beer—though historically he found inspiration in Brain of the Firmrather than Design for a Brain. We could also recall in this connection that no algorithmic system, in mathematics or in generative music, ever becomes in a fully open-ended fashion: each step in the evolution of such systems is rigidly chained to the one before. Nevertheless, as both Eno and Wolfram have stressed in their own ways, the evolution of these systems can be unpredictable even to one who knows the rules: one just has to set the system in motion and see what it does. For all practical purposes, then, such systems can thematize for us and stage an ontology of becoming, which is what Eno's notion of riding the system's dynamics implies.77
Third, we should note that Eno's ambient music sounds very different from the music we are used to in the West—rock, classical, whatever. In terms simply of content or substance it is clear, for instance, that three notes repeating with different delays are never going to generate the richness, cadences, and wild climaxes of Roxy Music. Whatever aesthetic appeal ambient music might have—"this accommodates many levels of listening attention and is as ignorable as it is interesting" (Whittaker 2003, 47)—has to be referred to its own specific properties, not to its place in any conventional canon.78 And further, such music has a quality of constant novelty and unrepeatability lacking in more traditional music. In Cvaries each time it is performed, according to the musicians who perform it and their changing preferences; It's Gonna Rain depends in its specifics on the parameters of the tape players used, which themselves vary in time; the computerized system Eno described above is probabilistic, so any given performance soon differs from all others even if the generative parameters remain unchanged.79 Perhaps the easiest way to put the point I am after is simply to note that Eno's work, like Alvin Lucier's biofeedback performances (chap. 3), raises the question, Is it music? This, I take it, again, is evidence that ontology makes a difference, now in the field of music. I should add that, evidently, Eno has not been alone in the musical exploitation of partially autonomous dynamic systems, and it is not the case that all of his colleagues were as decisively affected by reading the Brain of the Firmas he was. My argument is that all of the works in this tradition, cybernetically inspired and otherwise, can be understood as ontological theater and help us to see where a cybernetic ontology might lead us when staged as music.80
Fourth, these remarks lead us, as they did with Beer himself, into questions of power and control. Usually, the composer of a piece of music exercises absolute power over the score, deciding what notes are to be played in what sequence, and thus exercises a great deal of power over musical performers, who have some leeway in interpreting the piece, and who, in turn, have absolute power over the audience as passive consumers. In contrast, "with this generative music . . . am I the composer? Are you if you buy the system the composer? Is Jim Coles and his brother who wrote the software the composer? Who actually composes music like this? Can you describe it as composition exactly when you don't know what it's going to be?" (Eno 1996b, 8). These rhetorical questions point to a leveling of the field of musical production and consumption. No doubt Eno retains a certain primacy in his work; I could not generate music half as appealing as his. On the other hand, the responsibility for such compositions is shared to a considerable extent with elements beyond the artist's control—the material technology of performance (idiosyncratic human performers or tape players, complex probabilistic computer programs)—and with the audience, as in the case of computer-generated music in which the user picks the rules. As in the case of Beer's social geometries, a corollary of the ontology of becoming in music is again, then, a democratization, a lessening of centralized control, a sharing of responsibility, among producers, consumers, and machines.81
My fifth and final point is this. It is ironic that Eno came to cybernetics via Beer; he should have read Pask. The musical insights Eno squeezed out of Beer's writings on management are explicit in Pask's writings on aesthetics. As we can see in the next chapter, if Pask had handed him the torch of cybernetics, Eno would not need to have equivocated. Pask, however, was interested in more visual arts, the theater and architecture, so let me end this chapter by emphasizing that we have now added a distinctive approach to music to our list of instances of the cybernetic ontology in action.
7
_ _ _ _ _
GORDON PASK
from chemical computers to adaptive architecture
NOW, WE ARE SELF-ORGANIZING SYSTEMS AND WE WANDER AROUND IN A WORLD WHICH IS FULL OF WONDERFUL BLACK BOXES, DR. ASHBY'S BLACK BOXES. SOME OF THEM ARE TURTLES; SOME ARE TURTLEDOVES; SOME ARE MOCKING BIRDS; SOME OF THEM GO "POOP!" AND SOME GO "POP!"; SOME ARE COMPUTERS; THIS SORT OF THING.
GORDON PASK,"A PROPOSED EVOLUTIONARY MODEL" (1962, 229)
THE DESIGN GOAL IS NEARLY ALWAYS UNDERSPECIFIED AND THE "CONTROLLER" IS NO LONGER THE AUTHORITARIAN APPARATUS WHICH THIS PURELY TECHNICAL NAME COMMONLY BRINGS TO MIND. IN CONTRAST THE CONTROLLER IS AN ODD MIXTURE OF CATALYST, CRUTCH, MEMORY AND ARBITER. THESE, I BELIEVE, ARE THE DISPOSITIONS A DESIGNER SHOULD BRING TO BEAR UPON HIS WORK (WHEN HE PROFESSIONALLY PLAYS THE PART OF A CONTROLLER) AND THESE ARE THE QUALITIES HE SHOULD EMBED IN THE SYSTEMS (CONTROL SYSTEMS) WHICH HE DESIGNS.
GORDON PASK,"THE ARCHITECTURAL RELEVANCE OF CYBERNETICS" (1969A, 496)
Now for the last of our cyberneticians. Andrew Gordon Speedie Pask (fig. 7.1) was born in Derby on 28 June 1928, the son of Percy Pask, a wealthy fruit importer and exporter, and his wife Mary, and died in London on 28 March 1996, at the age of sixty-seven.1 Gordon, as he was known, was much the youngest of three brothers. His brother Alfred, who trained as an engineer but became a Methodist minister, was twenty years older. The other brother, Edgar, was sixteen years older and was Gordon's "hero and role model" (E. Pask n.d., n.p.), and it is illuminating to note that Gar, as he was known, distinguished himself by bravery in research verging on utter recklessness in World War II. He left his position as an anesthetist at Oxford University to join the Royal Air Force in 1941 and then carried out a series of life-threatening experiments on himself aimed at increasing the survival rate of pilots: being thrown unconscious repeatedly into swimming pools to test the characteristics of life jackets; again being thrown repeatedly, but this time conscious, into the icy waters off the Shetlands to test survival suits; hanging from a parachute breathing less and less oxygen until he became unconscious, to determine at what altitude pilots stood a chance if they bailed out; being anesthetized again and again to the point at which his breathing stopped, to explore the efficacy of different modes of resuscitation. He "won the distinction of being the only man to have carried out all [well, almost all] of his research while asleep," and the Pask Award of the British Association of Anaesthetists for gallantry and distinguished service was instituted in his honor in 1975 (Pain 2002). Gar was a hard act for young Gordon to follow, but he did so, in his own unusual way.2
Figure 7.1.Gordon Pask in the early 1960s. (Reproduced by permission of Amanda Heitler.)
Gordon was educated at Rydal, a private school in Wales, where he also took a course in geology at Bangor University. He was called up for military service in 1945, but "Gordon's career in the RAF was extremely brief. During his first week at camp, he passed out while doing the mandatory session of push-ups, and was returned home in an ambulance" (E. Pask n.d., n.p.). Pask then studied mining engineering at Liverpool Polytechnic, before taking up a place at Downing College, Cambridge, in 1949, wh
ere he studied medicine and gained a BA in physiology in the natural science tripos in 1953 (Pask 1959, 878). In 1956, he married Elizabeth Poole (E. Pask [1993] describes their unconventional courtship), and they had two daughters: Amanda (1961) and Hermione (adopted in 1967). In 1964, Pask was awarded a PhD in psychology from University College London and in 1974 a DSc in cybernetics by the Open University. In 1995, the year before his death, Cambridge awarded him an ScD (Scott and Glanville 2001; Glanville and Scott 2001b).
His first book, An Approach to Cybernetics,was published in 1961 and was translated into Dutch and Portuguese, and several other books followed (Pask 1975a, 1975b, and 1976a were the major ones; also Pask and Curran 1982; and Calculator Saturnalia[Pask, Glanville, and Robinson 1980]—a compendium of games to play on electronic calculators). A list of his publications (journal articles, chapters in books and proceedings, technical reports) runs to more than 250 items (in Glanville 1993, 219–33). At different times he was president of the Cybernetics Society and the International Society for General Systems; he was the first recipient of the Ehrenmitgleidof the Austrian Society for Cybernetic Studies and was awarded the Wiener Gold Medal by the American Society for Cybernetics.
From the 1950s onward, Pask enjoyed many university affiliations, including professorial chairs at Brunel University (in the Cybernetics Department, part-time, beginning in 1968) and the University of Amsterdam (in the Centre for Innovation and Co-operative Technology, beginning in 1987; Thomas and Harri-Augstein 1993, 183; de Zeeuw 1993, 202).3 He also from time to time held visiting positions at several institutions: the University of Illinois, Old Dominion University, Concordia University, the Open University, MIT, the University of Mexico, and the Architecture Association in London. But the principal base for Pask's working life was not an academic one; it was a research organization called System Research that he founded in 1953 together with his wife and Robin McKinnon-Wood.4 There Pask pursued his many projects and engaged in contract research and consulting work.5
So much for the bare bones of Pask's life; now I want to put some flesh on them. Before we get into technical details, I want to say something about Pask the man. The first point to note is that he was the object of an enormous amount of love and affection. Many people cared for him intensely. There are two enormous special issues of cybernetics journals devoted entirely to him, one from 1993 (Systems Research[Glanville 1993]), the other from 2001 (Kybernetes[Glanville and Scott 2001a]), and both are quite singular in the depth and openness of the feelings expressed. And this was, no doubt, in part because he was not like other men—he was a classic "character" in the traditional British sense (as were Grey Walter and Stafford Beer in their own ways). There are many stories about Pask. His wife recalled that "Gordon always denied that he was born, maintaining that he descended from the sky, fully formed and dressed in a dinner jacket, in a champagne bottle, and that the Mayor and aldermen of Derby were there to welcome him with a brass band and the freedom of the city." It is certainly true that he liked to dress as an Edwardian dandy (bow tie, double-breasted jacket and cape). At school, he built a bomb which caused considerable damage to the chemistry lab (which his father paid a lot of money to put right), and he claimed that "the best thing about his school was that it taught him to be a gangster." At Cambridge, he would cycle between staying awake for forty-eight hours and sleeping for sixteen (E. Pask n.d.). Later in life he became more or less nocturnal. His daughter Amanda told me that she would bring friends home from school to see her father emerge as night fell. Pask's ambition in studying medicine at Cambridge was to follow in his brother Edgar's footsteps, but as one of his contemporaries, the eminent psychologist Richard Gregory, put it, this story "is perhaps best forgotten." Pask apparently tried to learn anatomy by studying only the footnotes of the canonical text, Gray's Anatomy,and (Gregory 2001, 685–86) "this saved him for two terms—until disaster struck. He was asked to dissect, I think an arm, which was on a glass dissecting table. Gordon was always very impetuous, moving in sudden jerks. Looking around and seeing that no-one was looking at him, he seized a fire axe, swung it around his head, to sever the arm. He missed, There was an almighty crash, and the arm fell to the floor in a shower of broken glass. Perhaps it is as well that Gordon did not continue in medicine." Pask's partner in that ill-fated anatomy lab was Harry Moore, who later worked with Pask at System Research on many of the projects discussed below (Moore 2001). Among Pask's many behavioral quirks and conceits, one friend "marvelled at his perfect cones of cigarette and pipe ash that he appeared to preserve in every available ash-tray" (Price 2001, 819).6 Richard Gregory (2001, 685), again, recalls that Pask "was forever taking pills (his brother was an anaesthetist so he had an infinite supply) for real or imagined ailments. These he carried in a vast briefcase wherever he went, and they rattled." Pask apparently felt that he understood medicine better than qualified doctors, which might have had something to do with the decline of his health in the 1990s. Other stories suggest that some of these pills were amphetamines, which might have had something to do with Pask's strange sleeping habits and legendary energy.
Musicolour
Pask's engagement with cybernetics began when he was an undergraduate at Cambridge in the early 1950s. Many people declared themselves cyberneticians after reading Wiener's 1948 Cyberneticsbook, but Pask took his inspiration from the man himself (E. Pask n.d.):
The epiphany of his Cambridge life came when he was invited by Professor John Braithwaite, Professor of Moral Philosophy, to look after Professor Norbert Wiener, who was visiting Cambridge [and lecturing there on cybernetics]. Gordon who had been struggling for some years to define what he wanted to do, found that Wiener was describing the very science he had longed to work on, but had not known what to call. He had known for some time that what he wanted to do was to simulate how learning took place, using electronics to represent the nervous system . . . [and] in order to study how an adaptive machine could learn. Gordon decided to use his expertise in theatrical lighting to demonstrate the process.
This connection to the theater and the arts is one of the themes that we can pursue in several sections of this chapter. Pask had fallen in love with this world in his schooldays, largely through a friend who ran a traveling cinema in North Wales. At Cambridge, Pask "joined the Footlights club and became a prolific lyric writer for the smoker's concerts where numbers and sketches were tried out. [He also contributed] strange, surreal set design and inventive lighting for shows in Cambridge and in London. Gordon had made friends with Valerie and Feathers Hovenden, who ran a small club theatre in the crypt of a church on Oxford Street." In the same period Pask and McKinnon-Wood, also a Cambridge undergraduate, formed a company called Sirenelle dedicated to staging musical comedies. Both were fascinated with the technology of such performances: "Gordon used to come back [to Cambridge] with bits of Calliope organ, I would come back . . . with bits of bomb sight computer" (McKinnon-Wood 1993, 129). From such pieces, the two men constructed a succession of odd and interesting devices, running from a musical typewriter, through a self-adapting metronome, and up to the so-called Musicolour machine. As we shall see, Pask continued his association with the theater, the arts, and entertainment for the rest of his life.7
What, then, of Pask's first sally into cybernetics, the theatrical lighting machine just mentioned? This was the contrivance called Musicolour, for which, as his wife put it, "there were no precedents" (E. Pask n.d.): "Gordon had to design all the circuits used in the machine without any outside assistance.
Figure 7.2: Musicolour logic diagram.The original legend reads, "Outline of a typical Musicolour system. P = Performer, I = Instrument and microphone, A = inputs, yi, to visual display that specify the symbol to be selected, B = inputs, xi, to the visual display that determine the moment of selection, PF = property filter, AV = averager, AT = adaptive threshold device. Memories hold values of (yi). Control instructions for adjusting the sequence of operation are not shown. Internal feedback loops in the adaptive thresh
old devices are not shown." Source: G. Pask, "A Comment, a Case History and a Plan," in J. Reichardt (ed.), Cybernetics, Art, and Ideas (Greenwich, CT: New York Graphics Society, 1971), 79, fig. 26.
Figure 7.3.A, the Musicolour machine; B, its power supply; and C, a still photograph of a light show. Source: G. Pask, "A Comment, a Case History and a Plan," in J. Reichardt (ed.), Cybernetics, Art, and Ideas (Greenwich, CT: New York Graphics Society, 1971), 82, fig. 28.
Figure 7.4.Musicolour display at Churchill's Club, London. Source: G. Pask, "A Comment, a Case History and a Plan," in J. Reichardt (ed.), Cybernetics, Art, and Ideas (Greenwich, CT: New York Graphics Society, 1971), 86, fig. 32.
He built the original machine in his bedroom, on a large, old dining table, which took up most of the room. The process took him several years, during which he took his Tripos examinations and graduated from Cambridge. . . . Gordon had sincerely wanted to be a doctor, like Gar, but once he had begun to work on his Musicolour machine, medicine took second place." Musicolour was a device that used the sound of a musical performance to control alight show, with the aim of achieving a synesthetic combination of sounds and light.8 Materially, the music was converted into an electrical signal by a microphone, and within Musicolour the signal passed through a set of filters, sensitive to different frequencies, the beat of the music, and so on, and the output of the filters controlled different lights. You could imagine that the highest-frequency filter energized a bank of red lights, the next-highest the blues, and so on. Very simple, except for the fact that the internal parameters of Musicolour's circuitry were not constant. In analogy to biological neurons, banks of lights would only be activated if the output from the relevant filter exceeded a certain threshold value, and these thresholds varied in time as charges built up on capacitors according to the development of the performance and the prior behavior of the machine. In particular, Musicolour was designed to get bored (Pask 1971, 80). If the same musical trope was repeated too often, the thresholds for the corresponding lighting pattern would eventually shift upward and the machine would cease to respond, encouraging the performer to try something new. Eventually some sort of dynamic equilibrium might be reached in which the shifting patterns of the musical performance and the changing parameters of the machine combined to achieve synesthetic effects.9 Musicolour was central to the subsequent development of much of Pask's cybernetics, so I want to examine it at some length here, historically, ontologically, aesthetically, and sociologically, in that order.