Cybernetics ...
  "the science and art of understanding"... - Humberto Maturana
  "interfaces hard competence with the hard problems of the soft sciences" - Heinz von Foerster


Pre-History of Cybernetics

on the shoulders of giants
BACK: History
THEME: Control and Regulation in Machines

Learning about Regulation through Practical Applications   Control and regulation were indeed focal topics in the research being done by the people who first defined the field. However, interest and work on these topics dates back as far as historical records permit us to see.

The earliest tangible work on control was motivated by practical concerns. There were devices whose operation and maintenance could be simplified by imposing capacities for automatic regulation. We know that self-regulating devices were constructed far back in ancient times. A Greek named Ktesibios in Alexandria invented a float regulator for a water clock circa 270 BC, and another Greek named Philon of Byzantium used a float regulator to keep the level of oil in a lamp constant circa 250 BC. By the time of the first century AD, float regulators and similar devices had been employed for a variety of applications such as: self-closing cisterns, automatic wine dispensers, syphons to maintain constant water level differences between two vessels, and (semi-)automated operation of temple doors.

For the next thousand years, such devices would continue to be designed to exploit physical phenomena such as water flow, buoyancy, and magnetism. Around 1100 AD a south-pointing compass was linked to the wheels of a chariot so as to keep the vehicle steered southward. During the European Dark Ages, more sophisiticated float regulators continued to be developed in the Arab Empire. By the 13th century Arab craftsmen had added rudimentary 'on/off' switching and refined float regulators to produce remarkably accurate clocks.

The 14th century marked an interruption in this line of progress. The Mongols' sack of Baghdad destroyed the foremost center of float regulation technology. Meanwhile, back in Europe the mechanical clock arrived, and the best engineering of the day was devoted to it. Because these clocks' operations were tied to a constant time flow, their variable or adaptive self-regulation was not an issue.

Self-controlling mechanisms became interesting again in the late 16th century, in the context of volume agricultural operations (e.g., milling). Such farming and food operations were the organizational precedents for the later Industrial Revolution in manufacturing. For example, mechanisms were appearing to regulate the rotational speed of a millstone. In the Low Countries the mill-hopper (a device for controlling grain flows based on the millstone's speed) were widespread by 1600. From 1600 onward, such devices percolated to smaller-scale and more mundane applications. The earliest furnace temperature regulator appeared circa 1624, and similar such regulators were built mainly for use with incubators.

The earliest prototype steam engines appeared in the first decade of the 1700's. However, it was not until 1769 that James Watt produced the first really promising steam engine. HIs key innovation in making it useful wasn't the engine per se, but the means for effectively regulating its operation. By 1788 Watt had perfected his flyball governor mechanism (the whirling gadget that became the paradigmatic icon for 'control'). This innovation provided the operational stability and safety required to place steam engines into everyday usage. It was therefore regulation that finally validated engine technologies for deployment, and their deployment put the Industrial Revolution into high gear.

Two sophisticated controller elaborations appeared as the 19th century got underway. The first was the invention of mechanical devices that employed what would later be termed a 'feedback loop'. In 1793 the Swiss watchmaker Breguet employed a closed-loop feedback system to synchronize pocket watches in his pendule sympathique. This involved mechanically linking watch mechanisms so that they mutually influenced one another. Around 1840, the British Astronomer Royal - G.B. Airy - came up with a means for keeping his telescope directed at a celestial object over time. His solution involved a simple feedback device for continuously manuevering the telescope to compensate for the earth's rotation.

The second significant elaboration on control was the appearance of devices whose controlling behaviors could be easily changed - i.e., devices that could be 'programmed'. Joseph Jacquard invented his eponymous loom by adding a mechanism through which the loom's operations would be guided by a stack of punched 'cards'. This represented the first production machine to exploit reconfigurable or swappable 'coding'. This theme of programmability was reflected in the work of Charles Babbage, who reified his logical schemata into the specifications for a general purpose Analytical Engine around 1833. By the early years of the 20th century Hollerith punch cards (the direct descendants of Jacquard's cards) were controlling calculators in large enterprises.

For the next 100+ years, control and regulation devices would continue to proliferate and evolve. New forms of exploitable phenomena (e.g., hydraulics, gyroscopes, electrical circuits) would make for more widespread and more effective regulation and control. Entire classes of specialized controller devices (e.g., servomechanisms) appeared. By the time the 1940's and WWII arrived, Norbert Wiener, Arturo Rosenblueth, and Julian Bigelow were tasked to address the cutting edge of control technology - controlling complex weapons systems in accordance with electronic sensors. At the same time, Warren McCulloch and Warren Pitts were grappling with the problem of self-regulation in artificial neurons. Regulation was central in these scholars' seminal 1943 papers - "Behavior, purpose and teleology" and "A logical calculus of the ideas immanent in nervous activity", respectively. It's therefore no surprise that the first meeting among those who would become the 'cybernetics group' was on the subject of 'cerebral inhibition' (i.e., a regulatory mechanism).

BACK to History

Noted Contributors
to Cybernetics

  For feedback, suggestions or contributions, please email the ASC Webmaster.
  ASC HOME TOP of Page
The Subject of Cybernetics

on the shoulders of giants
This essay contributed by Randall Whitaker, March 2003