Article by Angus Jenkinson FCybS, June 2021
How is cybernetics relevant to the design of machines and machine-like systems, including software, the internet, robotics, autonomous cars, traffic control systems, and company management dashboards? This page introduces this aspect of cybernetics for the interested. According to the general image, machines are the essence of cybernetics. Are they?
Generally, this is an important cybernetic application. Every such system uses cybernetic techniques (consciously or not). Yet, the current image is mistaken. It is only one important application, and cyberneticians are as concerned about controlling the ethical and social effects as creating machines that control aspects of their own activity through software. Cybernetics ought not to be equated with machines as they are popularly understood. Nevertheless, cybernetics has immense significance for the development smart machines, significance that has already changed the world and will continue to do so.
What is a machine?
The OED (1991) says that it is a structure of any kind, material or immaterial. Versions include a vehicle, ship, technical structures, and a car. It is also a military engine, an apparatus, appliance or instrument, and in the immaterial sense, a device or machination. The definition, an apparatus for applying mechanical power consisting of a number of interrelated parts, each having a definite function is a particularly useful one. This dates back to the 17th century and can be compared with ’system’. A related definition is: A combination of parts, moving mechanically, as contrasted with being, having life, consciousness and will. Hence ‘applied to a person who acts merely from habit or obedience to rule without intelligence, or order, one whose actions have the undeviating precision and uniformity of the machine.’ It is also ‘any instrument employed to transmit force or to modify its application’. By analogy it becomes (first in the US) ‘the controlling organization of a political party, typically applied with disparaging emphasis to organisations of more or less similar character in England.’
The similarity with the common understanding of a system — as an interrelationship of parts that form an entity different from any of the parts by themselves — is perhaps partly why Ashby and others from the Macy generation tended to use the term machine to describe the universal principle of a certain kind of organization: ‘machine’ provides for both material and immaterial versions, like ‘system’. In Ashby’s Introduction to Cybernetics he uses “machine” extensively as a unit of analysis. He makes the point that he is dealing with a universal concept, i.e. one in any form, including the entirely immaterial. In this way, one can talk about the design and operation of Britain’s law as a legal system or a legal machine. The difference might be less technical than rhetorical. No one wants to be part of a machine; but most appreciate the sense of a completely ordered and sensible working arrangement.
A machine is a combination of parts, moving mechanically, as contrasted with being, having life, consciousness and will. (OED)
Moreover, early cyberneticians were particularly fascinated by the ability to demonstrate proofs of their core theories through the behaviour of machines that embodied their designs, like the quasi-living “tortoises” of Grey Walter or the personalities of Braitenberg’s vehicles. This would appear to reinforce the equation of cybernetics with machines. (It is sometimes unfairly labelled first order cybernetics.) Cybernetics has become instrumental in the creation of multiple smart machines, whether physically embodied in designs of mechanical structures from electromagnetic to, quantum operations, or in the virtual realm of software.
Dr Steve Battle FCybS speaks of experimental robot-psychology — applying the Test for the Controlled Variable to a class of robot known as Braitenberg’s Vehicles. Valentino Braitenberg was a cybernetician, a neuroanatomist, and a musician. His book, ‘Vehicles: Experiments in Synthetic Psychology’, inspired many to explore its strange intersection of cybernetics and artistry. The simplest among these robots can be understood as operating by Stimulus-Response; their outputs simply a function of the inputs. But put them in an environment and we see the emergence of behaviour—the control of perception.
The virtual world of modern cybernetic computing design concepts and processes are embodied in physical end-to-end relationships and structures but cannot be reduced to the material. They are universals that transcend any single form but can be translated between material structures, rather as music is translated from one instrument to another. And indeed from one key to another or even to other themes.
The vast domain of software and intelligent machines
When the Teleological Society met1 prior to the formation of the Macy series and the public articulation of cybernetics by its founders, it included several people with strong links to software and computers, including Turing and von Neumann. Turing’s Enigma machine had played a decisive role in the Second World War. A recent novel by Ian McEwan, Machine like Me, assumes that he did not die and later continuing revolution in robotics. The conceit works because Turing is understood as a genius who produced ideas that turned into machines. The ideas were in turn methods of replacing much slower and sometimes less accurate human labour by applying Turing’s analytical genius to a mechanical device that could carry it out at speed. In other words, during exported, to a limited extent, aspects of himself, his way of breaking down problems, importing and comparing data, creating algorithms for computation, leading to an outcome. A translation. In this sense, software as the translation into a form that a machine can execute of a series of steps (an algorithm) that human beings understand by which a particular outcome can be achieved while cancelling out any interference in achieving that outcome so far as possible. That is why an equation can be made between smart machines and cybernetics and human beings — why Norbert Wiener brought the two together in the title of his book and first definition.
This is why the comments were made above. On the one hand, cybernetics can translate into machines and enable them to do immensely useful tasks from satellite navigation in cars enabling agricultural machines to pilot round fields and people around the world. It flies aeroplanes, and will in due course no doubt autonomously control cars. Cybernetic thinking is implicit in most software processes. The range of applications will be known by most readers.
On the one hand, getting to know cybernetics is a way of understanding more deeply how this all works. And in the practical process of designing things so that they do work, we also learn something about both cybernetics and other aspects such as biology.
Walking upright is an immensely difficult problem for life. Only one organism has achieved it properly. Yet almost every child achieves this naturally in the process of growing up. Unless a person becomes injured or very old, walking upright is not difficult. Walking upright seems simple. But it is only simple because we can do it. Again, for many adults, cooking a meal is no problem. But consider how complex it is taken as a whole: not just the myth of Prometheus bringing fire to humans, but an immense global network of supply behind the ingredients. The global economy appears on a dinner plate — and of course much of it is driven by software and machine systems. There is probably no robot on Earth capable of designing and producing a cooked meal from a rich set of available ingredients. The next breakthrough in robotics has been anticipated as the ability to design and produce a robot capable of sorting and putting away children’s toys.
Both represent immensely complicated problems. Compare these with the reverence that we give to Leonardo da Vinci for his pioneer designs of flying machines. It is not uncommon to experience astonishment, even today, at being in a large aeroplane at 10,000 m, that can take off and land, pilot itself and navigate. Pilots are high status individuals. Indeed the entire ecosystem of flying is immensely complicated and structured. But in fact a machine is capable of piloting the entire flight. It can take off, fly, and land under most conditions. It can respond to computerised instructions, like a satellite navigation system. Much of the production processes of making one are automated and codified. Getting a robot to stand up and walk or run appears more difficult than getting a plane to take off, fly and land.
These are interesting questions to explore.
Ethical Questions
The generation of evermore intelligent machines doing evermore in the organisation of our society generates questions about control of those machines, decisions about how they should operate and when, indeed even how such questions should be decided. This is another philosophical aspect of the whole area of the development of machines, which will become no doubt ever more important during the 21st-century.
The aspect of “Machines” in cybernetics therefore includes principles and theory, key persons, research methods, examples and cases, and the significance of these disciplines for cybernetics, e.g.
Robotics
AI
Automation (inc software), autonomous vehicles
Machine Cybernetics
Robotics
AI software
Autonomous vehicles, autopilots…
Smart software systems (feedback and controls)
Assistive tech, eg for surgery
Homeostatic devices and signal systems
The internet
Semantic web and advanced coding
Legal and policy implications
1 Galison, Peter. “The Ontology of the Enemy: Norbert Wiener and the Cybernetic Vision.” Critical Inquiry 21, no. 1 (1994): 228-266. http://www.jstor.org/stable/1343893.