## Cybernetics and Systems Theory Defined

Right off, let's dispense with the childish belief that words "have" meanings. Charles Dodson / Lewis Carroll was close to the mark with this dialogue between Alice and Humpty Dumpty:- Humpty Dumpty: [Having just proved it is 364 times better to celebrate Alice: I don't know what you mean by 'glory.'

Humpty Dumpty: Of course you don't -- 'till I tell you. I meant Alice: But 'glory' doesn't mean 'a nice knock-down argument.'

Humpty Dumpty: When I use a word, it means just what I choose it to Alice: The question is whether you can make words mean so many Humpty Dumpty: The question is which is to be the master -- that's all.

- Information is "a difference that makes a difference," to use Gregory Bateson's definition. It is a measure of the reduction of uncertainty (entropy) that results from receiving a message. The term was rigorously defined by Claude Shannon of Bell Labs in 1950. Its unit is the bit, which is the amount by which your uncertainty is decreased when you find out that something happened which had a probablity one half. The bit measures in units of the minus the log to the base two of probablity: -log2(p) By this definiton news that of event which had a probablity of one fourth yields two bits of information, news that of event which had a probablity of one eighth yields three bits of information, and so on.
- System is a word that we use to describe any "experience-cluster" that we can map as a set of interacting elements over time. Typically a system is mapped by identifying the pathways of information flow -- as well as possibly the flow of energy, matter and other variables. But the "flow" of information is special, because only information can go from A to B while also staying at A. (Consider that photocopy machines would be useless if you didn't get to keep your original.)
- Cybernetics is the study of systems which can be mapped using loops (or more complicated looping structures) in the network defining the flow of information. Systems of automatic control will of necessity use at least one loop of information flow providing feedback. (Appropriately the word cybernetics is derived from the Greek for "pilot," as in "auto pilot" and "pilot light.") The word was first coined in 1945 by Nobert Wiener in the book Cybernetics (listed below), in which he defined it as "the study of control and communication in the animal and the machine." Gregory Bateson clarifies the methodology in "Cybernetic Explanation" (one of the essays in the book
__Steps to an Ecology of Mind__, listed below.) In a nutshell, he says: - Systems theory is the study of systems which can be mapped using any kind of network to define the flow of information. This includes the study of systems whose emergent properties we cannot yet predict due to a lack of plausable mechanisms, rigorous mapping techniques and/or robust mathematical treatment. By these definitions systems theory includes and is more abstract than cybernetics. In General System Theory (listed below), Ludwig Bertalanffy draws this distinction:
- Cybernetics is a theory of control systems based on communication (transfer of information) between systems and environment and within the system, and control (feedback) of the system's function in regard to environment... The model is of wide application but should not be identified with "systems theory" in general.

... Cybernetic systems are a special case, however important, of systems showing self-regulation.

## Why We Need Cybernetics and Systems Theory Now

"When we try to pick up anything by itself

we find it is attached to everything in the universe."

-- John Muir

I keep having this image of a survey course in human physiology, in which the syllabus covers each body system in turn, and in the final week it is all put together into a whole person. But, as the semester drags on the instructor gets behind in the material, until suddenly it is time for finals and the integrative material has not been covered. "Well," says the teacher, "I will leave it to the more ambitious students as an extra credit project." ## Where Cybernetics and Systems Theory Came From

"What is a man, that he may know a number,

or a number, that a man may know it?"

-- Warren McCulloch ` `

In 1868 James Clerk Maxwell (author Maxwell's equations of the electrodynamics, and inventor of the mental construct Maxwell's Demon) was invited by steam engineers to help them figure out why the governors on their engines didn't always work right: sometimes the steam engines exploded. Maxwell analyzed the "steam-engine-with- governor under a changing load" as a system of non-linear differential equations, and concluded the system would do one of five things based on the coefficients of the equations. (1) It corrected the speed back to the desired level fairly smoothly (the most desired response):Or (2) it corrected the speed back to the desired level after some overshoot (also a desired response):

Or (3) it oscillated continuously -- an annoying and inefficient repsonse later called "hunting" in the 1930's by electronics researches (actually this behavior was not explicitly described by Maxwell, probably because it is unstable, but it is implicit in his analysis):

Or (4) it oscillated with increasing amplitude until it blew up (also sometimes called "hunting":

Or (5) it just blew up:

This was the first explicitly cybernetic analysis of a system I can find. Thanks to it the engineers were able to design their governors so that the steam engines didn't explode so often. The governor became a metaphor for some 19th century visionaries: Samuel Butler in Erewhon predicted thinking machines evolving out of governors, and Alfred Russell Wallace came up with the idea of evolution by natural selection -- independently from Darwin -- when he realized that adaptive restraints operated like a governor on a steam engine.

By the late 1940's, thanks mostly to the growth of electronics, a lot of people were running around with the idea that "feedback" was somehow important. One of them was Warren McCulloch, a pioneer brain researcher who first proposed the mathematical modeling of neurons. He was approached by the Macy Foundation to chair a conference on the nervous system. The Macy Foundation, funded by the family that ran Macy's department store (and its famous Thanksgiving Day parade), funded conferences on medicine; they had done the heart, lungs, skin, etc. but never the brain or nerves. But McCulloch was determined to make these meetings more than a typical medical conference. He invited physiologists, electronics specialists, mathematicians, physicists, even social scientists -- including husband and wife anthropologists Gregory Bateson and Margaret Mead. The participants met for a few days every six months over a period of several years. At first McCulloch only let the "neuro" people talk; he wanted everyone to understand the great questions facing them before they started looking for answers. But eventually a cautious collaboration developed, as the participants probed their intuition of what was missing from their knowledge of minds.

In 1948 one of the attendees, mathemetician Norbert Wiener, published a book in which he purported to name the new field of inquiry they were investigating: Cybernetics was the name of the field and the book. This move received mixed reviews from the other participants in the conference. However, many of the attendees did return to their disciplines and begin using the new set of tools provided by the conference, and by Wiener (including two who I had the good fortune to meet and study under: Gregory Bateson and Heinz von Foerster).

Meanwhile, in Germany, Ludwig von Bertalanffy began publishing papers on the theory of general systems, in which he (prophetically in many cases) laid down some of the criteria of such a theory. He pointed out that the fundamental tool of general systems theory was the system of differential equations, but any such set of equations robust enough to describe non-trivial systems was unsolvable. Therefore, intuition and computer simulation should play important roles in a theory of complex systems. But his work had little impact initially.

In 1950 Shannon and Weaver at Bell Labs published their first paper on what has been called "information theory" and "communication theory," but I would prefer to call "transmission theory." It is the study of how to get bits reliably over an unreliable channel -- a topic of great interest to Bell Telephone at that time. In academia this work eclipsed cybernetics, probably because it was less intellectually threatening; cybernetics advocated connecting outputs to inputs, which had been forbidden since the ancient Greeks, while information theory dealt with the familar model of:

+-------+ +---------+ +--------+

| INPUT |---->| PROCESS |---->| OUTPUT |

+-------+ +---------+ +--------+

Also, it didn't help that the popular press picked up and began to abuse the word "cybernetics," as if it meant "the study of computers, robots, and electronic gizmos," or that the book Psycho-Cybernetics was published by plastic surgeon Maxwell Maltz in 1960; it was a useful pop-psychology self-help book about auto-suggestion, but had little to do with cybernetics. With the advance of digital computers in the 50's and 60's, the field of "information science" was heralded, which included the study of computer languages and their compilers, as well as Shannon's work. But cybernetics mostly suffered benign neglect by information science departments. By way of an example, in 1966 Scientific American published an entire issue devoted to the new technology of information, and later re- issued it as one of their theme paperbacks, called Information. Every diagram in this book has the same structure as the one above: INPUT, PROCESS, OUTPUT.

Yet, while the so-called information scientists ignored cybernetics (which they could because they designed systems and were free to design them without feedback), those scientists investigating the biochemistry of cell metabolism, the principles of nervous systems, and the population biology of ecologies (all pre-existing systems) were drawn to cybernetics because it offered more accurate models of the systems they were studying.

Some progress was made by topologists in the late 60's in classifying systems in terms of all possible behaviors they could exhibit. Initially this work, called the theory of dynamical systems, simply refined the distinctions drawn by Maxwell's governor paper.

The biggest methodological barrier to the advance of cybernetics in the 50's and 60's was the expense of computer time. But in the 70's pocket calculators became affordable, and it was on such a calculator that some of the earliest examples of chaos were discovered. This term is not used here in the every-day sense, but to describe a fifth category of system behavior besides the four illustrated above: non- periodic deterministic behavior. The discoveries of chaos, along with increasingly available computer power, sparked a renaissance of interest in cybernetics and systems theory in the late 80's.

Won't it be exciting to see what the 90's will bring?

## How I Got Into Cybernetics and Systems Theory

"Watch out -- you might get what you're after."

-- David Byrne, 1983

"Burning Down the House" ` `

In the fall of 1969, while a Junior in high school in southern California, I got my hands on an early Whole Earth Catalog. I was attracted to it because the title sounded integrative, and sure enough the first section was entitled Understanding Whole Systems. Here I was exposed to the ideas of Bucky Fuller, Norbert Wiener, Marshall McLuhan, and Paul Ehrlich. I went to a college where I could design my own major program (University of California at Santa Cruz, Kresge Collge), and searched for three years for a faculty member who would sponsor my major in "Understanding Whole Systems." Dr. Gregory Bateson arrived at Kresge in the fall of 1973, and I had even studied his English accent for a school play, but it wasn't until he was featured on the first page of the Understanding Whole Systems section of the new Whole Earth Epilog in summer of 1974 that I realized he was who I had been looking for. Starting in my senior year I took all of his classes, and he sponsored a student-directed seminar which I taught in early 1975 on Understanding Whole Systems. I'd planned to extend this activity into my major program, but I too quickly reached the end of my senior year, I hadn't met the university breadth requirements to graduate, and I was out of scholarship money. As I explained above, some teachers recently received, through a mutual friend, a copy of my class handouts from that course in "Understanding Whole Systems" which Bateson sponsored me in teaching sixteen years ago at the University of California at Santa Cruz (UCSC). They asked how I would revise those notes based on what I have learned in the intervening years. I am happy to answer that question.