March 23, 2020

Blog All Dog-eared Pages: Cybernetic Revolutionaries by Eden Medina

Cybernetic Revolutionaries by Eden Medina tells the history of Chile and Stafforrd-Beer's attempt at the start of the 1970s to build a socialist democracy with help from cybernetics and networked computers. An alternative, more decentralized approach to the same sort of issues that today's "big data" and "smart city" initiatives aim to address.

Here are my dog-eared pages notes from reading it. (If you want another contemperaneous history of socialist technology approaches in the UK, check out The Plan a documentary)

Page 3
In Chile, I know that I am making the maximum effort towards the devolution of power. The government made their revolution about it; I find it good cybernetics.
—Stafford Beer, February 1973

Page 8
Through sociotechnical engineering practices, Chilean and British technologists tried to make Project Cybersyn implement and uphold principles of Chilean democratic socialism. For example, the system included mechanisms to preserve individual liberty within a context of greater state control. Some Cybersyn technologists also tried to use Project Cybersyn as a vehicle for increasing worker participation in economic management and proposed having workers collaborate with Chilean operations research scientists. I argue that, for the system to support values such as worker participation or decentralized control, Cybersyn needed to implement and maintain the social, organizational, and technical relationships specified by its designers. Yet the reverse was also true: changing these social, organizational, and technical relationships could cause the system to produce configurations of political power, including totalitarianism, that were very different from Chilean democratic socialism.

Page 22
From its earliest days cybernetics valued the cross-disciplinary pollination that occurred when experts from a variety of fields convened to discuss a common problem. The conferences organized by the Josiah Macy Foundation from 1946 to 1953, which laid the groundwork for the field of cybernetics, are the most notable example of such collaboration. For example, the attendance list at the first Macy conference included the anthropologist Gregory Bateson, neurophysiologist Warren McCulloch, mathematician John von Neumann, anthropologist Margaret Mead, logician Walter Pitts, Rosenblueth, Bigelow, and Wiener, among others. Attendees at the Macy conferences drew inspiration from cybernetics' encouragement of the use of common metaphors to describe biological and mechanical systems and took this innovation back to their home disciplines.

Page 25
To put it another way, Beer was more interested in studying how systems behaved in the real world than in creating exact representations of how they functioned. Furthermore, he was centrally concerned with developing mechanisms to help these systems self-regulate and survive. He stressed that cybernetics and operations research should drive action, not create mathematical models of increasing complexity and exactitude.

Page 28
Beer challenged the common definition of control as domination, which he viewed as authoritarian and oppressive and therefore undesirable. It was also "naive, primitive and ridden with an almost retributive idea of causality." What people viewed as control, Beer continued, was nothing more than "a crude process of coercion," an observation that emphasized the individual agency of the entity being controlled. Instead of using science to dominate the outside world, scientists should focus on identifying the equilibrium conditions among the subsystems and developing regulators to help the overall system reach its natural state of stability. Beer emphasized creating lateral communication channels among the different subsystems so that the changes in one subsystem could be absorbed by changes in the others. This approach, he argued, took advantage of the flexibility of each subsystem. Instead of creating a regulator to fix the behaviour of each subsystem, he found ways to couple subsystems together so that they could respond to each other and adapt. Such adaptive couplings helped maintain the stability of the overall system.

Page 35
The Viable System Model devised ways to promote vertical and lateral communication. It offered a balance between centralized and decentralized control that prevented both the tyranny of authoritarianism and the chaos of total freedom. Beer considered viable systems to be largely self-organizing. Therefore, the model sought to maximize the autonomy of its component parts so that they could organize themselves as they saw fit. At the same time, it retained channels for vertical control to maintain the stability of the whole system.

Page 39
In contrast to the centralized planning found in the Soviet Union, Allende's articulation of socialism stressed a commitment to decentralized governance with worker participation in management, reinforcing his professed belief in individual freedoms. Yet he also acknowledged that in the face of political plurality the government would favour the "interest of those who made their living by their own work" and that revolution should be brought about from above with a "firm guiding hand."

Page 54
Beer also felt that CORFO needed to change its practices and worried that the "quickest solution" to managing the enterprises "will be imposed before any new thinking has time. Unless cyb[ernetics] can move faster."

Page 61
The exponential increase in the number of punch cards processed annually by the Chilean treasury paralleled the rise in the number of pages it generated annually, as well as the greater length of the annual report that accompanied the presidential address to Congress (105 pages in 1965; 496 pages in 1967; and 1,075 pages in 1970, illustrated each year by ever more graphs and tables).

Page 63
[Beer] regarded the centralized approach [of the Soviet Union] as bureaucratic and overly complex, as well as vulnerable to manipulation. Factory managers and government bureaucrats could easily change the value of the data they submitted to the computer centers in order to put their management skills in a more favorable light.

Page 73
Second, Cyberstride, like the Viable System Model, tried to find a balance between autonomy and cohesion. For example, when the system detected a production anomaly, the National Computer Corporation would alert both CORFO and the factory interventor. The government would then give the interventor a limited window of time to resolve the problem on his own. The enterprises therefore maintained their autonomy to a reasonable degree. If the interventor could not resolve the problem within this limited period, CORFO would intervene. Such intervention would limit the autonomy of the factory, but Beer reasoned it was nonetheless essential for preserving the viability of the entire economic system.

Third, the system's design reflected Allende's commitment to raising employment levels, a key part of the government program. Unlike concurrent uses of computer technology in industry, Cyberstride would use computers in a way that did not lead to unemployment. In industrial settings, computers are often linked to factory automation, which can raise productivity levels but also allow companies to downsize their workforce. Rather than automating labor or replacing management, Cyberstride would offer factory managers and CORFO a tool to help them increase factory productivity using the human and material resources available.

Page 79
The program also implemented a new and untried method of Bayesian statistical forecasting known as the Harrison-Stevens Approach, which first appeared in the December 1971 edition of Operational Research Quarterly. Dunsmuir stumbled onto the method while performing a literature review for the project. He convinced Beer that the new method would recognize significant variations in the production data and predict whether these initial data points signified the beginnings of a linear trend, and exponential trend, a step function, or an anomaly that would return to normal.

Page 87
Even in its final version, however, the model would not function as a predictive black box that gave definitive answers about future economic behaviour. Rather, it offered a medium with which economists, policy makers, and model makers could experiment and, through this act of play, expand their intuition about economic behaviour and the interplay of price controls, foreign exchange reserves, import and export rates, and other factors. Thus, the simulator was not meant to replace human expertise but to enhance it.

Page 89
Beer used the word algedonic to describe a signal of pleasure or pain. An algedonic meter would allow the public to express its pleasure or pain, or its satisfaction or dissatisfaction with government actions.

Unlike polls or surveys, these algedonic meters would not limit or prompt answers by asking set questions. The user simply moved a pointer on a dial somewhere between total dissatisfaction and absolute happiness. This design "uses the [human] brain as a computer," Beer wrote, "structured and programmed by individuality." Reminiscent of Beer's attention to autonomy and broad participation, the meter permitted users to construct their own scale of happiness and did not impose a standardized definition. Unlike many survey techniques, the meter did not require users to rationalize their level of happiness or normalize it to fit on a uniform scale. Instead, the meter recorded the user's gut feeling at a particular moment; the position of the knob on the meter would determine the voltage output on the device. Beer wrote that the meters could be installed in any location with a television set, such as in a Chilean home or in select community centers.

Page 93
According to Gerrity, the CIA planned to bring Chile to the point of economic collapse by urging companies to "drag their feet in sending money, in making deliveries, in shipping spare parts, etc." and to "withdraw all technical help and not provide any technical assistance in the future." We may never know the full extent of U.S. intervention in Chile during the Allende government or the entire role played by companies such as ITT, but we do know that the U.S.-led economic blockade created shortages of spare parts, caused significant problems for Chilean industries, stopped industrial machinery, and affected the production and repair of consumer goods.

Page 109
[...] the deputy director of the State Technology Institute explained, "it is important for Chile to be selective about the technologies it adopts, because in the long run they may determine social values and the shape of society—as the automobile has in the United States, for example.

Page 125
Beer imagined that the individuals sitting in the operations room would be either members of the government elite or factory workers, individuals who did not know how to type—a skill typically possessed by trained female secretaries. With little instruction, occupants could use the large "big-hand" buttons on each armrest. Participants could also "thump" these buttons if they wished to emphasize a point. Beer claimed that an interface of large geometrical buttons made the room more accessible for workers and prevented it from being a "sanctum sanctorum for a government elite." Through this design decision, the system allowed for worker participation.

Page 132
Looked at from a different angle, the Cybersyn system could even be read as disempowering Chilean workers. The timing charts printed in the study of the Easton Furniture factory are reminiscent of the time studies that characterized the Taylor system of management, which had been introduced in a number of Chilean factories before Allende came to power.

Page 145
In many ways its relative obscurity had benefited Cybersyn. The team had enjoyed a high degree of autonomy that was rarely challenged. Using his informal web of contacts, Flores had found it fairly easy to secure most of the financial, human, and material resources that the project needed and had done so without bureaucratic delays.

In October 1972, truck owners and then many other business owners across the country went on strike, in an attempt by the owning classes to bring down the government..
Page 149
After the strike, Silva said, "two concepts stayed in our mind: that information helps you make decisions and, above all, that it [the telex machine] helps you keep a record of this information, which is different from making a telephone call. [Having this record] lets you correct your mistakes and see why things happened."

Page 150
To put it another way, the network offered a communications infrastructure to link the revolution from above, led by Allende, to the revolution from below, led by Chilean workers and members of grassroots organizations, and helped coordinate the activities of both in a time of crisis. During the strike, workers in the state-run factories found ways to maintain production while simultaneously defending their workplaces from attack. They also transformed factory machine shops into spaces for repairing the trucks the government owned or requisitioned. Meanwhile, the telex network helped the government direct raw materials, fuel, and transportation resources to the places that most needed them. It also helped the government keep track of its trucks and provided information about which roads were blocked and which roads were open.

Beer estimated that the telex network transmitted two thousand messages daily during the strike. "The noise was indescribable," Beer said [...]

Page 158
[Beer] advised the government to ask the Chilean people what goods they considered essential, the quantity they desired, and the quantity that they already had, and use that information to determine which goods were of "primary necessity."


He urged the government to use the pamphlet he had developed and Angel Parra's folk song to teach the Chilean people about cybernetic thinking in government and to promote the government's use of science and technology in the service of the Chilean people. He pushed the government to create a "proper proletarian channel" for Chilean television and radio that could shape public opinion of the government and offset the influence of the media outlets run by the opposition and funded by the CIA.

Page 159
Beer insisted that workers should control the use of Cybersyn and argued that this was within their capabilities. Since he reasoned that Cybersyn was like any other "automated machine tool," he concluded that Chilean workers could reasonably use the system "without understanding the electronics," such as the mainframe computer. Beer also felt that the operations room should function as "the shop-floor of Total Industry" and reiterated that it should be "a place for the workers." In addition, Beer wanted to change the power relations between Chilean scientists and technologists to that of advising the workers when needed and performing supporting roles, such as system maintenance.

Page 161
In contrast, Beer's report envisioned a more substantial form of participation. He wanted to change how management decisions were made, whose knowledge was used to make these decisions, and how workers, technologists, and managers interacted. And he believed that Project Cybersyn could change all this for the better.

Page 163
In his writings Beer often cited nature as a complex system that remains viable through its self-organization. He argued that such systems do not need to be designed because they already exist. To modify the behaviour of such a system, one need not control its every aspect but rather change one subsystem so that the overall system naturally drifts toward the desired goal. Perhaps the injection of worker action could drive Chile toward a new point of homeostatic equilibrium, one that was congruent with the overall goal of socialist transformation. Worker improvisation on the ground could, moreover, supplement Allende's directives from above. beer viewed this redundancy as another prerequisite for self-organization and system viability. He wanted to encourage self-organization both by having Chilean workers participate in the actual design of Cybersyn and by using cybernetics to enhance the new forms of participation that were developing in Chilean communities. Such participatory activities would not only increase worker freedom but also create a more participatory working life and a more democratic society.

Page 166
Because the team grew in this decentralized, "opportunistic" way, most new recruits were not introduced to Beer and had little or no awareness, let alone understanding, of cybernetic principles such as the Viable System Model or the Law of Requisite Variety. Nor did they understand how Beer's approach to decentralized control was congruent with the principles of Allende's democratic socialism. Increasingly, Cybersyn was becoming a technological project divorced from its cybernetic and political origins.

Page 168
Examining how beer's and Flores's views changed in the aftermath of the strike shows that individual experiences, historical moments, and geographies all contribute to the ways that technologies have politics and politics are shaped by technology.

Page 178
To these criticisms, Beer responded that the system used simple technologies such as telex machines, drew from excellent programming talent in London and Santiago, and relied on many "human interfaces," meaning it was not automated. He also said that he was tired of hearing the assertion that such a system could be built only in the United States, and stressed that building the futuristic control room required only "the managerial acceptance of the idea, plus the will to see it realized." But, he added, "I finally found both the acceptance and the will—on the other side of the world." This final comment was a not-so-subtle jab at his British compatriots, who over the years had questioned the legitimacy and feasibility of his cybernetic ideas.

Page 183
The ease with which Hanlon painted Cybersyn as upholding political values that were the opposite of those of Chilean democratic socialism is a good example of the difficulty engineers and other technologists face in designing political values in a technological system. In fact, Hanlon was not alone in recognizing Cybersyn's potential for centralized control.

Page 187
It could also be argued that Project Cybersyn used computers in ways that could not be replicated in more industrialized nations. In 1973 the Chilean economy was more than twenty times smaller than that of Germany and almost one hundred times smaller than that of the United States. The Chilean economy was also composed of only a few core industries. Building a computer system to manage the Chilean economy was thus a much simpler task than trying to build a comparable system in Germany or the United States [...]

Page 198
Members of the military still loyal to the constitutional government quickly put down the coup attempt. However, the attempt made visible the divisions within the military and signaled that the president was losing the support of the armed forces. In response to these events, members of the Cybersyn team were understandably concerned about Beer's safety and their ability to protect him. Despite their efforts, Beer remained conspicuous in Las Cruces: he was a six-foot tall Englishman living in a small Chilean town—and he accidentally set fire to the mayor's house shortly after he arrived.

Page 207
Why Project Cybersyn, a technological project that outwardly tried to decentralize Chilean power structures and support the revolution from below, was frequently read as a tool for centralized government control is a more complicated question. In some cases this interpretation was the result of misinformation, as was the case with the secrecy charges. In other instances it was the result of a willful attempt to cast the Allende government in a negative light. The Soviet embrace of cybernetics in the late 1950s and 1960s might also have influenced some to see Cybersyn as a centralizing technology.

Page 212
Computer and communications technologies have often been linked to processes of political, economic, and social transformation. But claims that these technologies can bring about structural change in society—like the frequent assertion that computers will bring democracy or greater social equality—are often made in the absence of historical analysis.

Page 213
More often, Cybersyn team members attempted to embed political values in Cybersyn through sociotechnical engineering, meaning that they tried to build values not only into the function of the technology itself but also into the social and organizational relations of its construction and use.

Page 215
For example, Allende charged the Project Cybersyn team with building a system that supported worker participation. Yet the scientific techniques Chilean engineers used to model the state-controlled factories resembled Taylorism, a rationalized approach to factory production that disempowered workers and gave management greater control over labor. Time analysis, for example, emerged in the context of capitalist production, prioritizing efficiency and productivity over other values, such as the quality of shop floor life. By using time-analysis techniques, Cybersyn engineers could have inadvertently created production relationships that were counter to the Popular Unity platform and then solidified them in the form of a computer model.

Page 220
This story of technological innovation also challenges the assumption that innovation results from private-sector competition in an open marketplace.

Page 231
"The ideal of an objectively knowledgeable expert must be replaced with a recognition of the importance of background," Winograd and Flores write. "This can lead to the design of tools that facilitate a dialog of evolving understanding among a knowledgeable community." Building on this observation, the authors propose that computers should not make decisions for us but rather should assist human actions, especially human "communicative acts that create requests and commitments that serve to link us to others."

Page 233
This last quote specifically for Jackie Pease...
[Beer's] last book, Beyond Dispute (1994), proposed a new method for problem solving based on the geometric configurations of the icosahedron, a polygon with twenty equilateral triangle faces. He called this new method "syntegrity" and argued that it could serve as a new approach to conflict resolution in areas of the world such as the Middle East.

Posted by Adrian at 03:37 PM | Comments (0) | TrackBack

March 02, 2020

Interesting Things on the Internet: March 2nd 2020

  • How to help someone use a computer. Words of wisdom. Worth applying to other related teaching avenues too, it reminds me of the good examples of fixing I see at the DoES Liverpool Repair Cafes: where the person with the broken item fixes it themselves, with help from the expert, rather than stood to one side watching the expert fix it for them.
  • In The Eternal Inferno, Fiends Torment Ronald Coase With The Fate Of His Ideas. "Carillion consisted, essentially, of a sales and contract management organisation that hunted public-sector service contracts and then hired subcontractors to carry them out." Metcalfe's law is great when you apply it to the number of people with phones in your phone network, or how many computers are connected to the Internet; it's terrible when you're counting the number of entities contracting between each other to deliver something. This is why we need to rid so many areas of our life of marketisiation and financialisation - not (just) because it's not a humane way of arranging affairs, but because it doesn't work!
Posted by Adrian at 10:55 AM | Comments (0) | TrackBack