The Cybernetic Brain

by Andrew Pickering

  _ _ _ _ _

  So much for the general ontological significance of the homeostat. As in the previous chapter, however, we should confront the point that Ashby, like Walter, aimed at a distinctly modern understanding of the brain: neither of them was content to leave the brain untouched as one of Beer's exceedingly complex systems; both of them wanted to open up the Black Box and grasp the brain's inner workings. Ashby's argument was that the homeostat was a positive contribution to knowledge of how the performative brain adapts. What should we make of that? As before, the answer depends upon the angle from which one looks. From one angle, Ashby's argument was certainly correct: it makes sense to see the homeostat's adaptive structure as a model for how the brain works. From another angle, however, we can see how, even as modern science, the homeostat throws us back into the world of exceedingly complex systems rather than allowing us to escape from it.

  The first point to note is, again, that Ashby's science had a rather different quality from that of the classical modern sciences. It was another instance of explanation by articulation of parts (chap. 2): if you put together some valves and relays and uniselectors this way, then the whole assemblage can adapt performatively. Ashby's science thus again thematized performance, at the level of parts as well as wholes. Second, and again like Walter's, Ashby's science was a science of a heterogeneous universe: on the one hand, the brain, which Ashby sought to understand; on the other, an unknown and cognitively unknowable (to the homeostat) world. Performative interaction with the unknowable was thus a necessary constituent of Ashby's science, and in this sense the homeostat returns us to an ontology of unknowability. And, third, a discovery of complexity also appears within Ashby's cybernetics, though this again requires more discussion.

  In chapter 3 we saw that despite its simplicity the tortoise remained, to a degree, a Black Box, capable of surprising Walter with its behavior. The modern impulse somehow undid itself here, in an instance where an atomic understanding of parts failed to translate into a predictive overview of the performance of the whole. What about the homeostat? In one sense, the homeostat did not display similarly emergent properties. In his published works and his private journals, Ashby always discussed the homeostat as a demonstration device that displayed the adaptive properties he had already imagined in the early 1940s and first discussed in print in his 1945 publication on the bead-and-elastic machine.

  Nevertheless, combinations of homeostats quickly presented analytically insoluble problems. Ashby was interested, for example, in estimating the probability that a set of randomly interconnected homeostats with fixed internal settings would turn out to be stable. In a 1950 essay, he explored this topic from all sorts of interesting and insightful angles before remarking that, even with simplifying assumptions, "the problem is one of great [mathematical] difficulty and, so far as I can discover, has not yet been solved. My own investigations have only convinced me of its difficulty. That being so we must collect evidence as best we can" (Ashby 1950a, 478). Mathematics having failed him, Ashby turned instead to his machines, fixing their parameters and interconnections at random in combinations of two, three, or four units and simply recording whether the needles settled down in the middle of their ranges or were driven to their limits. His conclusion was that the probability of finding a stable combination probably fell off as (1/2)n, where n was the number of units to be interconnected, but, rather than that specific result, what I want to stress is that here we have another discovery of complexity, now in the analytic opacity of multihomeostat setups. Ashby's atomic knowledge of the individual components of his machines and their interconnections again failed to translate into an ability to predict how aggregated assemblages of them would perform. Ashby just had to put the units together and see what they did.

  As in the previous chapter, then, we see here how the modern impulse of early cybernetics bounced back into the cybernetic ontology of unknowability. While illuminating the inner go of the brain, homeostat assemblages of the kind discussed here turned out to remain, in another sense, mini–Black Boxes, themselves resistant to a classically scientific understanding, which we can read again as suggestive icons for a performative ontology. Imagine the world in general as built from elements like these opaque dynamic assemblages, is the suggestion. We can go further with this thought when we come to DAMS, the homeostat's successor.

  Making much same point, the following quotation is from a passage in Design for a Brain in which Ashby is discussing interconnected units which have just two possible states, described mathematically by a "step-function" and corresponding to the shift in a uniselector from one position to the next (1952, 129): "If there are n step-functions [in the brain], each capable of taking two values, the total number of fields available will be 2n. . . . The number of fields is moderate when n is moderate, but rapidly becomes exceedingly large when n increases. . . . If a man used fields at the rate of ten a second day and night during his whole life of seventy years, and if no field were ever repeated, how many two-valued step-functions would be necessary to provide them? Would the reader like to guess? The answer is that thirty-five would be ample!" One is reminded here of Walter's estimate that ten functional elements in the brain could generate a sufficient variety of behaviors to cover the entire experience of the human race over a period of a thousand million years. What the early cyberneticians discovered was just how complex (in aggregate behavior) even rather simple (in atomic structure) systems can be.

  _ _ _ _ _

  The homeostat is highly instructive as ontological theater, but I should also note its shortcomings. First, like all of the early cybernetic machines including the tortoise, the homeostat had a fixed goal: to keep its output current within predetermined limits. This was the unvarying principle of its engagement with the world. But, as I said about the tortoise, I do not think that this is a general feature of our world—in many ways, for example, human goals emerge and are liable to transformation in practice. At the same time, we might note an important difference between the homeostat's goals and, say, the tortoise's. The latter's goals referred to states of the outer world—finding and pursuing lights. The homeostat's goals instead referred inward, to its internal states. One might therefore imagine an indefinite number of worldly projects as bearing on those inner states, all of them obliquely structured by the pursuit of inner equilibrium. This is certainly a step in the right ontological direction beyond the tortoise.

  Second, I described the homeostat as exploring its environment openendedly, but this is not strictly true. My understanding of open-endedness includes an indefinitely large range of possibilities, whereas the homeostat had precisely twenty-five options—the number of positions of its uniselector. A four-homeostat setup could take on 254 = 390,625 different states in all.17 This is a large number, but still distinctly finite. As ontological theater, therefore, we should think of the homeostat as pointing in the direction of open-ended adaptation, without quite getting there.

  Third, and most important, as the word "uniselector" suggests, adaptation in the homeostat amounted to the selection of an appropriate state by a process of trial and error within a combinatoric space of possibilities. This notion of selection appears over and over again in Ashby's writings, and, at least from an ontological point of view, there is something wrong with it. It leaves no room for creativity, the appearance of genuine novelty in the world; it thus erases what I take to be a key feature of open-endedness. It is easiest to see what is at stake here when we think about genuinely cognitive phenomena, so I will come back to this point later. For the moment, let me just register my conviction that as models of the brain and as ontological theater more generally, Ashby's homeostats were deficient in just this respect.

  _ _ _ _ _

  One final line of thought can round off this section. It is interesting to examine how Ashby's cybernetics informed his understanding of himself. As mentioned above, a multihomeostat assemblage foregrounded the role of time—adaptation as necessarily happen
ing in time. And here is an extract from Ashby's autobiographical notebook, "Passing through Nature" (Ashby 1951–57), from September 1952 (pp. 36–39):

  For forty years [until the mid-1940s—the first blossoming of his cybernetics] I hated change of all sorts, wanting only to stay where I was. I didn't want to grow up, didn't want to leave my mother, didn't want to go from school to Cambridge, didn't want to go to hospital, and so on. I was unwilling at every step.

  Now I seem to be changed to the opposite: my only aim is to press on. The march of time is, in my scientific theorising, the only thing that matters. Every thing, I hold, must go on: if human destiny is to go on and destroy itself with an atomic explosion, well then, let us get on with it, and make the biggest explosion ever!

  I am now, in other words a Time-worshipper, seized with the extra fervour of the convert. I mean this more or less seriously. "Time" seems to me to be big enough, impersonal enough, to be a possible object of veneration—the old man of the Bible with his whims & bargains, & his impotence over evil, and his son killing, has always seemed to me to be entirely inadequate as the Spirit of All Existent, if not downright contemptible. But Time has possibilities. As a variable it is utterly different from all others, for they exist in it as a fish lives in the ocean: so immersed that its absence is inconceivable. My aim at the moment is to reduce all adaptation to its operation, to show that if only Time will operate, whether over the geological periods on an earth or over a childhood in an individual, then adaptation will inevitably emerge. This gives to time a position of the greatest importance, equalled only by that "factor" that called space & matter into existence.

  This passage is interesting in a couple of respects. On the one hand, Ashby records a change in his perspective on time and change (in himself and the world) that is nicely correlated with the flourishing of his cybernetics. On the other, this passage returns us to the relation between cybernetics and spirituality that surfaced in the last chapter and runs through those that follow. Walter made the connection via his discussion of the strange performances associated with Eastern spirituality, which he assimilated to his understanding of the performative brain and technologies of the self. There are also definite echoes of the East in this passage from Ashby—one thinks of Shiva indifferently dancing the cosmos into and out of existence—though now the bridge from cybernetics to spirituality goes via time and adaptation, the key themes of Ashby's cybernetics as exemplified in the homeostat, rather than technologies of the self.18

  The self does, however, reappear in a different guise in this passage. "The old man of the Bible with his whims & bargains" is the very paradigm of the modern, self-determined, centered, human subject writ as large as possible. And it is interesting to note that Ashby's rejection of this image of the Christian God went with a nonmodern conception of himself. Just as a multihomeostat setup dramatized a decentered self, not fully in control and constitutively plunged into its environment, so "Passing through Nature" begins (Ashby 1951–57, pp. 1–3) with the story of a meeting in January 1951 at which Warren McCulloch was present. Realizing how important McCulloch was to his career as a cybernetician, Ashby took the initiative and shook hands with him, but then immediately found himself going back to a conversation with someone of "negligible . . . professional importance." "What I want to make clear is that I had no power in the matter. The series of events ran with perfect smoothness and quite irresistibly, taking not the slightest notice of whatever conscious views I may have had. Others may talk of freewill and the individual's power to direct his life's story. My personal experience has convinced me over and over again that my power of control is great—where it doesn't matter: but at the important times, in the words of Freud, I do not live but 'am lived.' "

  By the early 1950s, then, Ashby's understanding of himself and God and his cybernetics all hung together, with questions of time and change as their pivot. I take this as another instance of the fact that ontology makes a difference—here in the realm of spirituality and self-understanding, as well as brain science and much else: time worship and "I am lived" as an ontology of performative becoming in action.19

  The Social Basis of Ashby's Cybernetics

  Turning from ontology to sociology, it is evident already that there are again clear parallels between Ashby and Walter. Ashby was telling no more than the truth when he described his early work—up to 1940, say—as having no social basis, as "a hobby I could retreat to": something pursued outside his professional life, for his own enjoyment. Even after 1940, when he began to publish, his work for a long time retained this extraprofessional, hobbyist quality, very largely carried on in the privacy of his journals. In an obituary, his student Roger Conant (1974, 4) speaks of Ashby building the homeostat "of old RAF parts on Mrs Ashby's kitchen table" and of writing his two books "in Dr. Ashby's private padded cell" at Barnwood House.20

  When he did begin to publish his protocybernetic theorizing, Ashby submitted his work initially to the journals in which his earlier distinctively psychiatric papers had appeared. His very first paper in this series (Ashby 1940) appeared in the leading British journal for research on mental pathologies, the Journal of Mental Science. It appears that there was no great response to Ashby's work within this field, outside the narrow but important circle de- fined by himself, Grey Walter, Frederick Golla, and G. W. T. H. Fleming, the editor of the journal in question. And one can understand why this might have been: clinical psychiatrists and psychologists were concerned with the practical problems of mental illness, and, besides its oddity as engineering, Ashby's work sticks out like a sore thumb in the pages of the psychiatric journals— his theoretical work offered little constructive input to psychiatric practice (though more on this below).

  Conversely, in seeking to create a community of interest for his work, Ashby. like Walter, systematically looked beyond his profession. A journal entry from early June 1944 (p. 1666) records that "several of my papers have been returned recently & it seems that there is going to be considerable difficulty in floating this ship."21 At this point he began writing to other scholars with whom he appears to have had no prior contact about his and their work, and it is notable that none of the people he addressed shared his profession. Thus, the small existing collection of Ashby's correspondence from this period includes letters to or from the experimental psychologists Kenneth Craik and E. Thorndike in 1944, and in 1946 the anthropologist-turned-cybernetician Gregory Bateson, the eminent neurophysiologist E. D. Adrian, the doyen of American cybernetics, Warren McCulloch, the British mathematician Alan Turing, and Norbert Wiener himself. In most cases it is clear that Ashby was writing out of the blue, and that he identified this extraprofessional and protocybernetic community from his reading of the literature. Through these contacts, and also by virtue of something of an explosion in his publication record—around twenty cybernetic essays appeared in various journals between 1945 and 1952—Ashby quickly assumed a leading position in the nascent cybernetic community, though, as we saw in the previous chapter, this was itself located outside the usual social structures of knowledge production. In Britain, its heart was the Ratio Club, the dining club of which Ashby was a founder member; Ashby was an invited speaker at the 1952 Macy cybernetics conference in the United States, and he regularly gave papers at the Namur cybernetics conferences in Europe. As far as knowledge dissemination was concerned, Ashby's route into the wider social consciousness was, like Walter's and Wiener's, via the popular success of his books.

  Ashby's cybernetics largely existed, then, in a different world from his professional life, though that situation began to change in the late 1950s. Through what appears to be a certain amount of chicanery on the part of G. W. T. H. Fleming, who was chairman of the trustees of the Burden Neurological Institute as well as director of Barnwood House, where Ashby then worked, Ashby was appointed in 1959 to succeed Golla as the director of the Burden. His ineptitude in that position—including trying to purge the library of outdated books, setting exams for all the
staff, and setting private detectives on Grey Walter—remains legendary in British psychiatric circles, and Ashby was saved from a disastrous situation by the opportunity to flee to the United States (Cooper and Bird 1989, 15–18). Stafford Beer's diary for 1960 records the circumstances of an offer from Heinz von Foerster to join the faculty of the University of Illinois, made while Beer, Pask, and Ashby were all on campus for a conference on self-organization—an offer which Ashby understandably accepted without hesitation (Beer 1994 [1960], 299–301).

  At Illinois, Ashby's formal position was that of professor in the Department of Electrical Engineering with an associated position on the biophysics committee. His primary affiliation was to von Foerster's Biological Computer Laboratory, the BCL. The BCL was an independently funded operation housed within the Electrical Engineering Department and was, during the period of its existence, 1958–75, the primary institutional basis for cybernetics in the capitalist world.22 At the BCL Ashby became the only one of our cyberneticians to enjoy full-time institutional support for his work, both in research and teaching. Ashby retired from the BCL in 1970 at the age of sixty-seven and returned to England, and Conant (1974, 4) records that "the decade spent in the United States resulted in a host of publications and was in his own estimation the most fruitful period of his career." It seems clear that this time of singular alignment between paid work and hobby was also one of the happiest periods of Ashby's life, in which he could collaborate with many graduate students on topics close to his heart, and for which he is remembered fondly in the United States (unlike the Burden) as "an honest and meticulous scholar . . . a warm-hearted, thoughtful, and generous person, eager to pass to his students the credit for ideas he had germinated himself" (Conant 1974, 5).