It is not certain who was the first person to apply a feedback mechanism to regulate or control a machine. There is evidence that certain kinds of regulating devices, such as for regulating the level of oil in an oil lamp or the outflow of water reservoirs, were already being used more than 2,000 years ago, and were known in the Middle Ages and the Renaissance. However, knowledge was transferred orally from craftsman to craftsman, and only a few written descriptions exist, none of which treats the theoretical aspects of mechanical engineering. In the 17th century, machines that developed great power, like windmills and steam engines, were first used. It then became essential to provide some mechanism to limit or control their power, so that it would not finally destroy the machine which produced it. One of the first patents granted to a feedback mechanism was for a 'whirling regulator' which controlled the speed of rotation in windmills by means of a centrifugal pendulum (T. Mead: Regulator for Wind and Other Mills, patent no. 1628, London 1787). The engineer James Watt (1736–1819) adapted it and used it as a 'governor' to regulate the velocity of rotation in steam engines, where the output (i.e. the velocity of rotation) regulated the input (i.e. the steam). During most of the 19th century, although regulating devices were widely used in engineering, there were no clear concepts of the dimensions and mechanical properties they should have. They were built and simply had to be tried out — and often failed. There was no theoretical framework which allowed the performance of a given regulator to be calculated. Then James Clerk Maxwell (1831–79), a physicist, reduced the problems to mathematical formulae. However, it was not until the end of the 19th century that there became available the mathematical tools for giving easy and practical solutions to the equations suggested by Maxwell. Only then did the theory of feedback regulation become an established discipline that could be applied to all fields of engineering.
Similarly, in biology, the idea of control cannot be traced to a single person. Several steps, each focusing on special aspects, were important and necessary until it became apparent that the theory of feedback as applied to biological systems is structurally and mathematically the same as that used in engineering. Three physiologists of the 19th century deserve to be mentioned, because each of them drew attention to one important aspect of feedback control. The three aspects are the complex organization of organisms, the relative constancy of certain physiological parameters, and the description of animal behaviour using teleological terms.
In 1828 Charles Bell, the Scottish anatomist and surgeon, published Animal Mechanics, or Proofs of Design in the Animal Frame, in which he discussed the complex organization of organisms. He compared the structure which gives stability to bones with elements of architecture and building engineering. Similarly he compared the mechanism and efficiency of the heart and vascular system with pumps and pipes used in engineering. Although he did not go as far as actually to describe feedback mechanisms, by systematically comparing organisms with machines, and by using the same terminology, he was the first to widely use the concept of models. Models have since been generally employed, and have become an important element in biological cybernetics. One of their purposes is to isolate certain functional aspects of behaviour and to study their logic and limits of operation in order to better understand the mechanisms of biological design and development.
Claude Bernard (1813–78), a French physiologist, introduced the concept of the constancy of the milieu intérieur. By that he understood that within certain limits blood, for example, has a constant composition independent of environmental changes. If the glucose level in the blood falls, animals as well as humans get hungry, eat, and by doing so raise the glucose level again. Bernard saw the results of feedback mechanisms and explored several examples, like the regulation of body temperature as well as glucose levels in the blood. In his book Lessons on Phenomena of Life in Animals and Plants (1878) he extended these experimental results and developed a general theory of how animals are able to maintain the constancy of their milieu intérieur (see homeostasis).
The German physiologist Eduard Pflüger (1829–1910) considered the goal-directed behaviour of feedback mechanisms in biology. In 1877 he published a paper, The Teleological Mechanics of Nature. Teleology is defined as the study of final causes. Feedback mechanisms are characterized by the fact that the input is controlled by the output, and thus stabilizes the output, or makes the performance relatively independent from disturbing influences. One can consider the stability of the output as the 'goal' of the system. To turn the argument round, whenever a behaviour in biology is encountered that can be described as goal directed, i.e. teleological, it is very likely that a feedback mechanism is involved.
The next step in combining biology and technology was taken by Felix Lincke (1840–1917), professor of mechanical engineering at the Institute of Technology in Darmstadt, Germany. He was probably the first who saw the outlines of a unifying theory of feedback control that is applicable to machines as well as to organisms. In 1879 he published a lecture, The Mechanical Relay, in which he classified the different feedback mechanisms used in mechanical engineering and listed the necessary elements of any feedback loop. These are:
(i) the 'indicator', which continuously measures the output,
(ii) the 'executive' organ, which modifies the input of the feedback loop,
(iii) , the 'transmitter' which connects the 'indicator' and the 'executive' organ, and
(iv) the 'motor', which supplies the energy.
Applied to an arm movement, the 'indicator' is the sensory nerve endings that sense the position of the arm, the 'executive organ' is the motor nerves with the muscles that perform the movement, the 'transmitter' is the brain which establishes a connection between the sensory and motor nerves, and the 'motor' is the alimentary tract supplying the energy for the whole system. The action of the indicator, executive organ, transmitter, and motor in a feedback loop can be described in the same way whether they are identified in biological systems or in machines. It is always the difference between the intended goal and the measurement given by the indicator that modifies the input to the feedback loop and thus brings the output of the system nearer to its goal. Although Lincke's paper was published in a widely recognized journal of engineering (Zeitschrift des VDI), and also as a book, nobody picked up his ideas and he was virtually unknown until well after Wiener's success in 1948.
In 1940, 60 years after Lincke's lecture, another engineer, Hermann Schmidt (1894–1968), a professor at the Institute of Technology in Berlin, published a series of papers in which he independently developed ideas similar to those of Wiener. Wiener focused his ideas more on the mathematical problems involved, while Schmidt based his theory more on the historical development of engineering. First, with primitive tools like an axe, man determines the exact action. In a second phase, energy is provided, but man still has to control the action by constantly monitoring the state of the machine, as in an automobile. In a third stage, energy and control over that energy are provided, as in aeroplanes with automatic landing control. (The pilot approaching the airfield makes only the decision to land; the exact monitoring of height, velocity, etc. is taken over by a computer.) In this third stage of development, energy and the immediate control over it are transferred to a machine, even though man still has to determine the goal of action for the machine. Only in science fiction, like Samuel Butler's novel Erewhon (1872), is a fourth stage envisioned in which machines would develop to such a stage that they would determine their own goals — be it to the benefit of humans or not.
For a discussion of the application of cybernetics to human social behaviour, see also .
Fig. 1. Original figure from T. Mead, patent no. 1628,1787: 'A regulator on a new principle for wind and other mills, for the better and more regular furling and unfurling of the sails on windmills without the constant attendance of a man, and for grinding corn and other grain, and dressing of flour and meal, superior in quality to the present practice, and for regulating all kind of machinery where the first power is unequal.' Right up one sees a speed regulator. During rotation centrifugal forces will lift the two spherical weights. They pull on strings which then furl the sails so that the effective area exposed to the wind is reduced.
Fig. 2. Original figure from F. Lincke: 'Das mechanische Relais' (the mechanical relay), VDI Zeitschrift, 23, 509–24, 577–616 (1879). The 'indicators' are sensory nerves (S, Ns), the 'executive' organ (motor nerves Ne and muscle B), the 'transmitter' (brain with ganglion cells G), and the 'motor' (stomach M, heart H, lung L). 'The activity we use to direct our human "machine" to its goal results from the difference between the will and the observed or imagined reality, i.e. the difference between the intention and the result of the execution.'
(Published 1987)
— Volker Henn
- Bibliography
- Mayr, O. (1970). The Origins of Feedback Control.
- Muses, C. (2002). 'Recollections of Norbert Weiner, Warren McCulloch and Stafford Beer'. Kybernetes, 31.
- Wiener, N. (1948). Cybernetics.
- — — (1956). I am a Mathematician.
