Introduction

In 2018, automation is a concept in disarray.

When Delmar Harder, an executive at Ford Motor Company, introduced the term "automation" in the 1940s, its meaning was clear enough. Where before automobile engines had been manufactured by workers transferring parts between various metal-working machines manually, Harder and his subordinates in the "Automation Department" were evaluating where machines could be introduced into the process to move parts from machine to machine without a worker hefting them¹.

By the mid-1950s, "automation" was no longer a term of art. In the popular press, it had become dramatically more capacious, very nearly a label for technological change as such.

I intend to take a critical look at the various meanings of "automation" as it is used today, particularly in the political discourse around basic income, "bullshit jobs", "precarity", and so forth. But before doing so I thought it would be helpful to make more widely available some primary source materials that have helped me think through what I find so dissatisfying about commentary on automation.

In 1955, a congressional subcomittee held a series of hearings on "Automation and Technological Change". Among those invited to testify were executives of major manufacturing and consulting firms, scientists, engineers, economists, and labor leaders (most notably Walter Reuther). Many offered definitions of "automation" which I've excerpted below².

These definitions provide a useful snapshot of expert views during the first wave of anglophone automation discourse. Surprisingly little has changed since.

Definitions

John Diebold: President, John Diebold & Associates

We have heard automation characterized variously as a potential threat to the national economy; as a key to increased leisure opportunity; as a mystical pseudoscience of robots and giant brains; and as a press-agent's description of automatic operation, from the kitchen toaster to the subway turnstile… As you have said, automation is certainly capturing the popular imagination, and is being very widely - if not always wisely - discussed.

As I understand it, my mission this morning is to make clear some of the characteristics of automation - "the nature of the beast." If I am not mistaken, during the next 2 weeks of hearings you will receive a great many different and conflicting definitions of automation. The most common statement, in this respect, is that automation is merely an extension of mechanization. If this is the case, there certainly seems little justification for a separate new word to describe it.

I think that if the facts are looked to, if what is actually happening at the present time is analyzed, it is possible to come up with a particular set of changes that are occuring in industry and in business that perhaps justify the use of a new word to describe them. It seems to me this more than simply the introduction of new kinds of hardware, or a particular kind of technology. It is a basic change in production philosophy.

Our orginal and traditional attitude toward the organization of industrial processes has been to organize our industry according to a division of labor. When we first organized factories and businesses, we found it necessary to break down our work, to allow for a division of labor according to specific human skills. Later, as we mechanized our activities, and introduced machines into factories, we followed the same breakdown, the same division of labor, and we mechanized around that division of labor. We introduced machines in specific departs of plants and into sections of plants to do activities that were organized orginally according to the devision of labor. Now, through automation - and I think this is perhaps the basic meaning of automation - we are beginning to look at our industrial processes as complete, integrated systems, from the introduction of the raw material until the completion of the final product. It seems to me this is basically a change in production philosophy. It is something analogous to Henry Ford's concept of the assembly line. It is a way of looking at, as much as it is a way of doing, a technology.

One way of defining automation is to say that it is a means of organizing or controlling production processes to achieve optimum use of all production resources - mechanical, material, and human.

There are two basic steps that industry follows in the approach toward automation. The first of these is the organization of each of the several steps of the production process into a fully integrated system. The oil refineries pioneered in this step; the chemical industry, processing industries, and nuclear production have since followed in going through this first step of automation. They have changed what had formerly been batch processes into integrated systems.

The second step of automation is to take the system and to control I think there has been a great deal of confusion upon this point, and you may very well see this reflected in some of the statements during the course of the hearings. People who are not familiar with the process industries will point to oil refineries and say: "This is automation. Other industries are going to develop in the same way." Actually, by the introduction of control systems into oil refineries and other processing industries - paper manufacturing, sugar refining, chemical manufacture, etc. - a second stage of automation is being achieved. Here it is not a case of replacing hand labor by machines, but rather of operating the machines at optimum efficiency all of the time.

I think it is possible to characterize the nature of automation by three simple statements: First, the concept of automation is very simple. It is a welding together, as I have said, of the production steps. It is looking at production processes as closed and integrated systems. Secondly, the technology of automation is incredibly complex. It is easily among the half dozen most advanced technologies of our time. Fundamentally, it deals with the transmission and use of information for the purposes of machine control, and for the purposes of optimizing production. I think you may be interested in the origin of this technology. The individual use of self-regulating mechanisms - devices that can regulate the course of their own activity - is very old. It goes back to the float-controlled valves the Romans used, and devices used by the early Dutch to keep windmills facing into the wind. James Watt devised a regulator, the "flyball governor," to keep his steam engine operating at constant speed, and there have been a number of other uses of this concept of self-regulation, which is known as feedback.

During World War II, the theory and use of feedback was studied in great detail by a number of scientists in this country and in England. The introduction of rapidly moving aircraft very quickly made traditional gun-laying techniques of anti-aircraft warfare obsolte. It was impossible to follow such rapidly moving targets manually. As a result, a large part of the scientific manpower in this country was directed toward the development of self-regulating devicesand systems to control our military equipment. It is out of this work that the technology of automation, as we understand it today, has developed. During the last 10 years, this technology has begun to be applied to industry. Yet it is not so much the technology itself, but rather the way it is applied, that is properly called automation.

I think th third point about automation that may help to give an understanding of its nature is that its application is very widespread. Automation can be applied in many types of of businesses and industries. If automation is regarded as a philosophy, a way of organizing production, it is something that can be applied in areas where only a small amount of actual mechanization is possible. It seems to me that a good example which will clarify this point is the automation of office procedure. We have developed, through the use of new technology, the machines you mentioned in your introduction: computing machines and data-processing equipment. Through the use of this equipment it is possible to automate office operations which hitherto have been conducted entirely by hand.

In the factory, automation means basically two things: It means, first of all, the integration of production machines, which may which may in fact be no more than a new level of mechanization. An often-cited example is the Ford plant in Cleveland, where injection blocks are made quite automatically by the use of a series of special-purpose machines - an automatic mass-production line.

Most of American industry, however, depends upon short runs of product. About 89 to 90 percent of all American production is in lots of less than 25 individual pieces. It is impossible to build special-purpose machines to manufacture these, because the character of the product changes too frequently. This is where the second meaning of factory automation comes in. In such a job-shop operation automation is just beginning to be achieved, in the form of tape-controlled machines tools – machines for which instructions can be provided in a flexible and variable form. This kind of automation is just beginning to have an impact. It gives every impression of taking a very loong time to come about.

All of these facts about the nature of automation – and I am sure you will hear many more during the next few weeks – are different aspects of what is basically a single new philosophy, and a single new technology.

Walter S. Buckingham: Associate Professor of Economics, Georgia Institute of Technology

Since World War II some spectacular discoveries in the fields of electronics and communications have permitted the manufacture of various types of automatic computing machinery. These machines are capable of translating a large body of previously developed, theoretical, economic, and business principles into practical significance. Called electronic computers, they are capable of processing data with almost unbelievable speed. When information is fed into them, usually on tapes, they can perform a series of logical operations can can choose among several previously anticipated courses of action based on built-in criteria. They can even adjust automatically for errors. The operation of these computers to solve scientific or commerical problems is often referred to as automation.

Also, in the last few years of a number of automatic or semiautomatic machines have been constructed to supplement convential assembly-line operations in factories. These machines perform hundreds of individual mechnical functions without direct human intervention. The operation of these machines is likewise commonly called automation.

Finally, sceintists, computer manufacturers, and science fiction writers have shown, hypothetically at least, how the administrative and manufacturing processes of an enterprise could be integrated into a single, silent, automatic monster which could grind out an endless chain of products without a man in sight. This awesome picture has charged the imagination of some and struck terror into the hearts of others. The possibility of such developments is also called automation.

In addition to this definitional confusion many speculations, hypotheses, and fragments of theories concerning the broad economic and social implications of automation are currently being expounded. In this flood of verbiage there is no* shortage of imagination, but there is a notable lack of the Kind of critical thought and careful documenta tion which yields quantitative, scientifically accurate results. There is a great need to collect, sift, classify, and evaluate the empirical evidence which alone can test these generalizations…

The variety of popular uses of the term "automation" necessitates some definition which is both precise and relevant for analysis. Such a definition can best be derived from an examination of the major principles which underlie most if not all of the popular concepts of automation. These are four such major principles - mechanization, feedback, continuous process, and rationalization.

Mechanization means the use of machines to perform work. Some times mechanization substitutes machinery for human or animal muscle. The steam engine did this. Sometimes mechanization sub stitutes machinery for brainwork at the lower, routine levels. The electronic computer does this. Because of the power, compactness or speed of machine operation, mechanization usually permits tasks to be performed which could never be done by human labor alone no matter how much labor was used or how well the enterprise was organized and managed. Mechanization increases wealth and re duces drudgery in the long run but in the short run it may cause hardships to workers whose skills are rendered obsolete, diluted by a further specialization or whose jobs are abolished altogether.

Feedback is the second principle inherent in automation. This is a concept of control whereby the input of machines is regulated by the machine's own output so that the output meets the conditions of a predetermined objective. As in a simple, thermostatically con trolled heating system, the conditions created by the output auto matically control, in turn, the amount of input and hence the per formance of the machine. When controlled by the feedback prin ciple, machines start and stop themselves and regulate quality and quantity of output automatically.

Continuous flow or process is the third principle of automation. This concept is of increasing importance because it is spreading from many individual production processes to the business enterprise itself and on to the entire economy. Mass production, increasing interdependence and now automation all embody this principle which is leading to a concept of the business enterprise as an endless process. Business for the most part has ceased being an operation that can be started and stopped with small loss. The regulation of a constant flow of goods has become a major concern of management…

Rationalization, the fourth principle of automation, means the application of reason to the solution of problems or to the search for knowledge. In a production system it means that the entire process from the raw material to the final product is carefully analyzed so that every oper ation can be designed to contribute in the most efficient way to the achievement of clearly enunciated goals of the enterprise.

Actually, rationalistic philosophy is nothing new, having become an important force in the world with the Renaissance. However, the scientific, rationalist philosophy takes on numerous new implications when it can be implemented by modern electronic machinery. The rise of electronic computers has led to a fascination with the possibil ity that superrationalism in the business and scientific spheres might spill over and transform society into an exact mechanism in which all e ements of chance, risk, capriciousness and free will, as well as all spiritual values, would be eliminated. Although this kind of s ecu lation is highly dubious nevertheless it is one logical extension oi this fourth principle of automation.

Following these four principles - mechanization, feedback, continuous process and rationalization - automation can be given a definition precise enough to be useful for logical analysis. It can be said to be any continuous and integrated operation of a rationalized production system which uses electronic or other equipment to regulate and coordinate the quality and quantity of production.

D.J. Davis: Vice President, Manufacturing, Ford Motor Company

[Davis begins with a concise history of changes in production technique at Ford since its founding, including the debut of the assembly line.]

To emphasize that automation, as we consider not new, con sider the well-publicized flour mill that was conceived and built back in 1807, almost 150 years ago. In this flour mill the grain was dumped into hopper that led to scale, where was weighed and dumped again into another hopper that had a screw-type conveyor at the bottom. This conveyor carried the grain to bucket elevator, which raised to the top floor. From the top floor flowed by gravity to another screw conveyor that carried to hoppers feeding the mill stones. As the flour emerged was fed mechanically to screens, then into barrels, and finally carried away by wagon or barge. This was true automation, whether the designer of that mill knew or not.

At Ford we have defined automation as "the automatic handling of parts between progressive production processes." It the result of nothing more than better planning, improved tooling, and the application of more efficient manufacturing methods which take full advantage of the progress made by the machine-tool and equipment industries.

As used at Ford, automation covers a wide variety of material-handling and related devices. In the machining of engine cylinder blocks, for example, automation moves the parts being manufactured into and out of load and unload stations automatically and, at the same time, actuates the machine cycle through electrical interlocking. When it can be applied, automation also is used to index the part, position turn or rotate depending upon the requirements of the succeeding operations.

[Davis then describes the use of electrical interlocks at Ford.]

Robert W. Burgess: Director, Bureau of the Census, Department of Commerce

I think we can say that "automation" is a new word for a now familiar process of expanding the types of work in which machinery is used to do tasks faster, or better, or in greater quantity.

Walter P. Reuther, President, Congress of Industrial Organizations

There has been a great deal of propaganda. Even the Secretary of Commerce has misrepresented the position of American labor with respect to automation…. We welcome any step forward that will make it possible for mankind to create greater economic wealth, with less human effort…

The other day, I am told, one of the vice presidents of the Ford Motor Co. came before your committee. I think he gave you a lot of very valuable information, but I take issue with one conclusion that- he left with your committee when he said that automation is just an extension of the normal technological evolutionary process. I believe that is an understatement of what automation is. Automation is the second phase of the industrial revolution.

James Watt, when he developed the first simple and crude steam- powered machine, which was used in the textile industries of England, made the first step toward the substitution of mechanical power for human power - mechanical muscles replaced human muscles. We took that simple beginning and we developed our mass-production economy. But automation is not just an extension of that technological process. Automation makes a completely new development in the technological process because automation, in addition to substituting mechanical power for human power, begins to substitute mechanical judgment for human judgment - the machine begins to substitute for the thinking process on a mechanical basis for the thinking process which heretofore was done exclusively by the human mind.

The machine has a memory in which you can store all sorts of complex information. With specific impulses you can bring that information out of the memory and with it begin to instruct the machine to do very complicated tasks.

This is why automation is not a normal extension of the technological process which was initiated at the beginning of the industrial revolution, but marks a second phase of the indusrial revolution.

It means that because the machine now cannot only replace human power, but can replace human judgements, its impact will be much greater than the impact of the first phase of the industrial revolution.

Don G. Mitchell: Chairman and President, Sylvania Electric Products, Inc.

Automation is only a more recent term for mechanization which has been going on since the industrial revolution began…

Automation comes in bits and pieces. First the automating of a single process, and then gradually a tying together of several processes to get a group or subassembly complete.

Robert C. Tait: President, Stromberg-Calrson Company; Senior Vice President, General Dynamics Corporation

In my own view, automation is simply a phrase coined, I believe, by Del Harder of Ford Motor Co. in describing their recent supermechanization which represents an extension of technological progress beyond what has formerly been know as mechanization.

Some experts in this field limit the application of automation to control devices that involve what is known as feedback - that system of machines and controls that capable of adjusting its own operation in the direction needed to obtain desired result, rather than simply following preset cycle of operations. One of the earliest applications of the feedback principle James Watt's flyball governor for steam engines. Other similar long-standing applications are windmills, ship-steering engines, thermostats, and so forth. Thus the feedback principle in itself nothing new but the modern application new in that the feedback information handled by electronic means.

Howard Coughlin: President, Office Employees' International Union, AFL

Automation in the office consists of the development of general and special purpose computing machines capable of recording and storing information, and of perfomring both simple and complex mathematical operations on such information. It is apparent that in the long run, automation will be of benefit to all of us. However, through the development of automation, dislocations of personnel will occur which will present many problems to be resolved.

James B. Carey: Secretary-Treasurer, CIO; President, International Union of Electrical Workers

When I speak of automation, I am referring to the use of mechanical and electronic devices, rather than human workers, to regulate and control the operation of machines. In that sense, automation represents something radically different from the mere extension of mechanization. Of course, it is not a sudden, full-blown appearance. It rests upon previous developments, especially upon Government- sponsored research and development in electronics and radar. But it is a new departure from the older methods of machine production by machine operators, since it represents the automatic operation of machines and of entire industrial and clerical processes.

The first industrial revolution, usually identified with Watt’s steam engine, replaced animal and human muscle power with steam power and electric power; it replaced the handicraft worker with the machine tender or machine operator. Automation uses control devices that result in the automatic production and processing of goods and data; it tends to replace the human regulation and control of machines and thereby changes the machine operator into a supervisor of an automatically controlled operating system.

Automation is a new technology, arising from electronics and elec¬ trical engineering. It is not a new machine, or even a new industry. It is rather, a new and revolutionary technology that is applicable to almost all, if not all, types of industrial and clerical operations. It makes possible the automatic office, as well, as the automatic factory. There is a likelihood that entire departments, offices, and plants, in the major parts of the economy will be using automation equipment within the coming 10 years.

William W. Barton: President, W. F. & John Barnes Company

Personally, I prefer to think of automation in the larger sense - as an innovation created by man to increase his production; technocracy, if you will.

In a broad sense, it can be stated that any feat that hands and body can perform can be duplicated automatically, given enough time and money.

James P. Mitchell: Secretary of Labor

Taking the first aspect of the word, we find that in some ways automation means precisely what each individual man on the street thinks it means, for to a great extent it is a word which produces various sorts of fears in various sorts of individuals—fear of change, fear of technology itself, fear of displacement, fear of unemployment, fear of machines, fear of science in general. In this sense, automation is nothing new, since these same fears have been with us in one form or another ever since the first caveman resented fire.

I am reminded, in fact, of the hearings held by the House Labor Committee less than 20 years ago. The discussion centered around a resolution that would have required the Secretary of Labor to draw up a list of laborsaving devices and a parallel estimate of the number of people probably unemployed as a result of their use.

So the present discussion is not within an unfamiliar context.

Taking the second aspect of the word, the technical one, we find that in a general way the word represents technological change, which surely is nothing new. It represents a movement certainly as old as the industrial revolution and probably older. Now, technological change varies, as we all know, in rate and degree. Its latest manifestation, coming as it has in a favorable setting of growth and prosperity, has appeared with relative swiftness and in some ways spectacularly.

It has come with such devices as complex automatic systems, electronic controls and regulators, feedback systems, transfer machines, conveyors, and the like. It has been attached to many self-regulating processes which have reduced the number of workers needed to perform a given job. When you get down to specifics there is some argument about the precise definition of this second aspect of the word. However, we can all agree that it is the latest development in the progress of industrial technology, the latest step in the long search for ways to replace human energy with mechanical energy.

Joseph A. Beirne: President, Communications Workers of America

We in the telephone industry have lived with mechanization and its successor automation for many years. They have been accompanied by fear and job insecurity. They have been our constant companions and loom larger today than ever before.

[Throughout the remainder of his testimony, Beirne favors the term "mechanization".]

Cledo Brunetti: Director, Engineering Research and Development, General Mills, Inc.

First, I want to point out that automation is not a revolutionary technique, but a continuation of our progress in mechanization…

Automation, a newly coined word, to describe an old, old process, has created in some minds the impression that this is not a process of natural growth. This "newness" interpretation of the word is confusing, because the word seems to set end limits on a continuous growth process. Automation cannot be said to have begun on any certain date, nor can it be said that it will end at any definite time. Automation is in truth but a phase df our continuing technological advance. It is just more of the kind of stuff which has been taking work out of work, and creating more and better jobs all the time.

Gentlemen, we had automation in this country in 1784. This book, which I hold here, by Oliver Evans, was published in 1795, and it describes an automatic flour-milling plant which was built near Philadelphia at that time…

But automation, in effect, had its start in the days of the Neanderthal man, when that early ancestor learned that the animals of the forest could be subdued more effectively with a club than with his hands. As a matter of fact, he even used it for communication in convincing his wife what to do.

The point is that the desire to lessen work, to do things more quickly, effectively, and easily, came into physical being as the wheel in that time. Man invented the wheel way back in those days and the wheel was the basis for the machine.

Ultimately then man learned to make combinations of wheels and shafts and other parts into machines capable of even more complex operation. To the machine was added cheaper, more efficient, more abundant power.

Then came programing [sic], which was telling the machine what to do. Such as using a switch to turn it on and off, but more complicated programing. Finally, control, to check the quality and quantity of machine output. Put them all together — machine, power, programing and control — and you have a definition of automation in four words, which gives you also the history of automation.

Marshall G. Munce: Vice President, York Corporation

Automation is a new word, and to many people it has become a scare word. Yet it is not essentially different from the process of improving methods of production which has been going on throughout human history—ever since men first took up jagged pieces of flint to perform operations better than they could be performed with bare hands. Since mankind learned to think, people have sought to amplify their efforts and produce more for their needs by using tools instead of muscles…

What is considered by some to be new in mechanization today, and the reason this hearing has been called, is the development of ingenious control mechanisms, such as the electric eye, mechanical brains, and other intricate electronic and radiation devices, which can direct and control the operation of machines.

The production engineer has at his disposal a variety of machines and devices for controlling them which, when put together in proper sequence, can turn out a continuous flow of mass-produced products or materials without human hands touching them during the manufacturing process.

This is "automation," and to some people it calls up a specter of robot, workerless factories, and great numbers of people deprived of their jobs and their means of livelihood.

Yet these control devices are not new. Automation is not a subject which can be discussed only in the future tense. Its essential features have been applied in a number of iields for many years. It is the word that is new, not the principles involved.

Modern petroleum refineries fit very closely the definition of the "automatic factory," which is regarded as the ultimate in automation. The substances to be treated in the refinery flow continuously through the several processes, controlled automatically by regulating devices. Human labor is used only for maintenance and for control at a few critical points. Oil refining has been on this continuous-flow basis for over 30 years.

Automatic manufacture has been highly developed for some time in a number of other processing industries. The continuous-flow technique has become the standard method of producing certain chemicals, some kinds of food, paper, and for the refining of ores. The production of cigarettes is an almost wholly automatic process.

The dial system of routing telephone calls is an example of automation which has been with us for some time. The New York Telephone Co. began installation of dial telephones in 1922.

Even in retailing, automation has a long history. Vending machines for selling various forms of merchandise date back as far as most of us can remember. So far such devices have never captured more than a small fringe of total retail trade. Whether automatic selling will become more prevalent in the future I would not venture to predict, but if it does, it will not be a brand new development.

Ralph J. Cordiner: President, General Electric Company

For practical purposes in planning manufacturing facilities, General Electric defines automation as "continuous automatic production," largely in the sense of linking together already highly mechanized individual operations. Automation is a way of work based on the concept of production as a continuous flow, rather than processing by intermittent batches of work.

There are those who try to make automation a catchall term to apply to every improvement, whether new or thoroughly familiar, that occurs in a factory or office. It is important to recognize that automation is only one phase in the process of technological progress, a natural evolutionary step in man’s continuing effort to use the discoveries of science in getting the world’s work done…

Thus, in the factory and in the office, you could say that progress toward greater automation is nothing new; only the expression "automation" is new.

Technological change in industry is a gradual process. Most products are first made by hand, or with hand tools. Then industry mechanizes : It introduces machines for some parts of the process, although many hand operations usually remain. As the economics of the situation warrant, the machines are made more and more automatic. Finally, where we can do it technically and where economic considerations warrant the investment, we link together parts of the process to achieve more continuous and automatic operation. As industry moves its operations up this scale toward automation, there is a greater demand for more highly trained people to handle the larger responsibilities.

Most of industry is still in the lower mechanization stage. There are literally millions of hand operations in manufacturing today. Highly skilled machinists, for example, usually spend much of their time and effort placing material in machines, and removing the material. Only part of their time is used as skilled machinists. There are many other wasteful and burdensome hand operations that add to the ultimate cost of goods going to the consumer…

Thomas J. Walsh: Chemical Group Director, Automation Project; Professor, Chemical Engineering, Case Institute of Technology

The tendency of material such as fluids to flow means that they can be handled in pipes in a continuous process. This continuous operation is the chemical plant equivalent of what I believe automation amounts to in the mechanical-process industry.

Ralph E. Cross: Executive Vice President, The Cross Company

To begin, I think I would like to start out as most witnesses do, I presume, by giving a definition of automation. I like to think of automation as being the application of cost-reducing machines and techniques. I would like to put some emphasis on the words "cost-reducing" and distinguish it from labor-saving, which I will get into later…

Automation in the metalworking industry has been in existence for over 100 years. It actually goes back to the time when the first basic machine tools were invented…

The difficulty with these machines is that they simply do not reduce production costs. The secret of bringing a machine like this into existence, or any other, for that matter, is getting the proper balance between mechanization and labor. Every time a new machine is purchased a decision must be made between more mechanization and less labor, on the one hand, and less mechanization and more labor, on the other…

So the objective is not to reduce labor, but to get the proper balance between mechanization and labor, so that we will get the lowest possible part cost.

In all of my experience in creating automated machines, I don’t believe I have ever sold a machine that could not have been made more automatic, or that might not have reduced labor further than it did. I think I can sum it up by saying that I have never found a situation to exist where all mechanization and no labor would provide the lowest operating cost. I think that sort of a situation is as extreme as a situation with all labor and no mechanization.

Clifton W. Phalen: President, Michigan Bell Telephone Company

As the subcommittee pointed out last April in its release announcing its study, "automation" is a relatively new word. It is variously defined. To me it means general technological progress of the kind that has been taking place in our industry and in others for many years. It is in this broad sense that I would prefer to use the word.

Automation, however, apparently suggests to some people that factories and even some industries of the future will run automatically, with only a few persons needed to control operations.

S.R. Hursh: Chief Engineer, Pennsylvania Railroad Company

Your witness today is at a loss to describe automation in the true sense of the word. The only real automation to my mind is the human heartbeat which is not subject to conscious control of man. Automation, therefore, is nothing more than controlled mechanization whether that control be electronic or mechanical.

The word "automation" is a comparatively new word. It has in a short period of time become a new science that is having a great effect in many industries especially those that lend themselves to automatic control like the chemical and oil industries and the utilities, such as light and power companies.

It is, however, not new to the railroad industry, especially in the field of railway signaling. For years we in the railroad industry and the manufacturers of signaling devices have thought of it as system control engineering and it dates back to the inception or invention of the track circuit as early as 1872. Since that time, steady growth and development of signaling equipment by close collaboration between the railway industry and the makers of signaling equipment, such as Union Switch and Signal, Division of Westingnouse Air Brake Co., and General Railway Signal, there has been developed the following: Centralized traffic control, automatic train control, continuous cab signaling, route interlocking, automatic freight car classification; the latter, of which, I will touch on in detail, by an explanation of what the Pennsylvania Railroad is now constructing at Conway, Pennsylvania…

Dr. A. V. Astin: Director, National Bureau of Standards

Automation is a relatively new word. It has been defined in many ways by various people. It probably means, the process of rendering automatic. From this point of view, the newest thing about automation is the word itself. The development of devices to perform functions automatically is a very old activity. For example, the ancient Romans invented a hydraulic float valve to control automatically the level of water in storage tanks. I would prefer to consider the subject of mechanization which is a broader area of technology, with automation as one of its important subdivisions. The general goal of mechanization is increased productivity; to use machines to aid man in producing more goods and services. Increased mechanization and increased productivity have expanded together. This has been especially noteworthy over the past 150 years, and particularly in the past 50 years.

[The "phases of automation" are described as follows.]

The first and probably the most basic is the replacement of physical energy provided by humans or animals by energy provided by machines powered from mechanical, electrical, or chemical sources. A primitive example is the use of hydraulic energy to operate a flour mill. More recent examples are the use of gasoline to propel automobiles and tractors and electrical energy to turn, the wheels of factories. The importance of this phase of mechanization is attested by the tremendous expansion of electrical and petroleum energy sources in recent decades. Even more phenomenal expansion can be expected as atomic-energy sources become available.

A second phase of mechanization is the use of physical measurement. In the older days of hand craftsmanship, items were fabricated by fitting mating parts together or by adjusting an item, such as garment, to the individual size of the user. With the development of the science of measurement, together with instruments for making measurements, it became possible to fabricate items according to a specification. When items are fabricated according to physicaf characteristics, as defined on a blueprint or specification, they can be used interchangeably. Deyelopments along these Iines led to mass production techniques, the essence of which is interchangeability of parts. One of the first persons to demonstrate this technique was Eli Whitney, who showed approximately 150 years ago that rifles could be assembled from interchangeable parts, each one of which was fabricated in accordance with carefully prescribed procedures of measurement. More recent advances in the science or physical measurement have brought about instruments that make certain measurements automatically, as well as instruments which make measurements more precisely.

A third phase in mechanization is the use of mechanical handling techniques whereby materials are carried by machine from one processing station to another. The conveyor belt is probably the best known example of this phase of mechanization.

A fourth phase of mechanization involves the concept of feedback. Feedback merely means the use of the result of a physical measurement at one stage of a mechanical process to alter or control an operation at in earlier stage of the process. A commonly used example of feedback is the thermostat on an automatic furnace. The thermostat measures the temperature at a particular location and feeds the result of this measurement to the controls of the furnace, shuttingit off or turning it on according to the temperature observation. Feedback can be accomplished entirely mechanically or with the intervention of human operators in varying degrees.

In the furnace example, a man can be both the temperature-measuring device and the feedback element; he shovels coal or closes the furnace drafts, depending on whether he is cold or hot. In steel production, it is customary to run analyses on the composition of the melts at various steps along the processing line. A human operator may make the measurements with mechanical instruments which can result in an orally conveyed order to modify a phase of the production process as a result of the measurement. In modern petroleum refining, the measuring process, the order giving, and the resulting control modification can all be accomplished mechanically.

A fifth and most recent phase of mechanization is the utilization of advanced and automatic computational techniques. Frequently the completion of the feedback process requires computations of varying degrees of complexity on the results of a physical measurement before a proper control order can be determined. Automatic machines which make possible rapid and reliable computations provide extremely important supplements to the feedback process. It is in this area that the striking recent advances have taken place, providing a basis for extensive future developments. Advances in the tools of computation have another important implication. Hitherto the trend in mechanization has been confined primarily to the processing of physical materials and devices. Automatic, computing machines are, on the other hand, concerned with processing numbers. The processing of numbers is an important operation in many activities other than production lines. Business offices such as banks, insurance companies, and retailing organizations, Government agencies, and scientific laooratories, are all concerned with processing numbers on increasingly larger scales. Hence, this phase of the process of mechanization appears to have applications in many areas other than the production processes where so much mechanization has already taken place.

I mentioned earlier that the goal of mechanization is increased productivity, and increased productivity is, I believe, an objective of primary concern to our Nation.

M. A. Hollengreen: President, National Machine Tool Builders' Association

Actually, machine-tool engineers do not talk about automation. They have been building automatic devices into machine tools for 50 years. They naturally take pride in their most recent accomplishments, but they expected them. They do not regard automation as something entirely new and unforeseen.

I think I can make this point clearer by telling you about metalcutting lathes. Back in the days after the Civil War when lathes were first developed, the operator had to put the workpiece in and take it out by hand in order to make a single outside cut. Then a turret lathe was developed on which two or more cuts, both internal and external, could be made without removing the workpiece from the machine.

Next, the industry developed lathes with more than one spindle. It devised means for moving the workpiece automatically from position to position in the machine, and from spindle to spindle. This was what is called the automatic chucking machine. By the 1920's it was possible to perform 10 or 15 operations on such a machine without moving the work by hand.

Little by little, electric and automatic controls were added to chucking machines so that by World War II there were hundreds of these machines turning out thousands of identical parts a day. Then the word "automation" was coined, sometime just before Korea.

Since we first heard of automation there have been further improvements in chucking machines. They consist mainly in devices for automatic gaging which enable the machine to correct its own errors. But, largely, they represent continuations of previous developments.

Chucking machines involve a great deal of automation, but it is not what has been referred to as Detroit-type automation. To the machine-tool builder this means merely a transfer machine. In such a machine a raw piece is loaded at one station and held in place for the first operation. It then moves automatically to a second station where a second operation is performed. This process can be repeated through many stations. Transfer machines have been built as long as football fields.

Before these [transfer machine] installations [for production of automobile engine valves] were developed, the pieces had to be loaded and moved from operation to operation by hand. This took time, several minutes per piece, and over long production runs these minutes add up to hours and days and months. Before this kind of installation was available, a maximum of 275 valves could be ground in an hour. We can now manufacture 450 in an hour. In addition, human errors have been eliminated. We achieve much greater accuracy. Scrap has been reduced to a minimum.

Dr. Vannevar Bush: President, Carnegie Institution of Washington; Founder of Raytheon; Wartime head of Office of Scientific Research and Development

Automation may be defined as relegation to a machine of the function of performing operations previously performed, manually. Any operations which are repetitive in nature, and which may be specified by a formula or a schedule, may be thus turned over to be performed by a group of interconnected devices - mechanical, electrical, pneumatic, and the like. It is an entirely different question whether it will pay, or whether it is socially desirable, for work thus to be transferred.

Even mental work may be thus relegated to a machine, if it comes within a similar definition, as has been well illustrated by Dr. Burgess. It is being done to great advantage today. This is not usually called automation, but it is of equal importance to the automation which substitutes for manual work. In fact, in the long run it may be of even greater importance to our progress, for it enables us to know more and to know it more accurately.

But automation may be thought of in a much broader sense. Man now has the dream of making machines which are like himself, and which can hence become his slaves. And he has progressed a long distance toward this objective, and will progress further. Yet there are inherent limits, and severe conditions to be met, as the evolution of machines proceeds under man's direction.

The functions of the human mechanism for which he would find substitutes are primarily physical and mental; he would replace manual operations and the mental processes that go with them. He is not much concerned with one property of living organisms, the ability to take in food and convert it into energy. In fact, he assigns this function to a separate category of machines, his engines; for food he uses coal and oil, and thus he produces power to drive his automobile or electricity to run his factory. He will no doubt use atomic and solar energy for the same purpose. In fact, he is doing so well at this that he has more convenient and powerful means than nature has introduced. And when he builds a machine to replace a man he has little worry about its metabolism, for he can readily supply the energy needs.

For doing mechanical things he has a host of devices: Motors, relays, hydraulic cylinders, pneumatic devices. With these he can accomplish movements at far greater speed than can a man’s arms or legs; he can exert far more powerful forces; and he can perform the movements far more precisely and reliably. His machines never tire of their work. Thus even although he has never produced the marvelous physical structure and chemistry of the human muscle, he has many, and in some ways, better substitutes.

When the operation to be performed is strictly repetitive, and no real thinking is involved, the substitution of a machine for a man is straightforward. One merely has to reproduce the mechanical movements involved. But it is possible to go beyond this. Feedback may be introduced, and the machine thus caused to exercise a sort of judgment. Mr. Reuther has given a good example of this in the form of a lathe which notes the variations of its product and readjusts its tools accordingly. And for this purpose we nave many remarkable instruments ; photocells which can sense the exact position of elements, radioactive devices which can instantly sense the thickness of a sheet of metal, gages to which the thickness of a hair is a gross error, and a host of others. Thus he can make his machines self-regulating and self-correcting.

The performance of the machine need not be fixed once for all, it can be to a certain extent flexible, and the instructions given to it may be altered, and may be very complex. Thus a milling machine may be controlled by a magnetic tape and caused to perform any sequence of operations that such a machine is capable of, time after time, without error and with precision. And by merely altering the tape a* wholly new sequence or operations may lie introduced. We hardly approach the versatility of the human mechanism in this way, for the same man can ride a bicycle, drive a car, swim a river, or add a column of figures. We could make versatile machines, no doubt, that would do many things. But here, in particular, one meets the difference between what it is possible to do, and what it pays to do.

Machines may be made to cooperate so that the performance of one is correlated with the work of the others. In particular it is possible to arrange a line of machines to perform a sequence of operations, passing work on from one to the next. We here encounter what is usually called systems engineering in one of its phases.

It is in the mental area that the most fascinating progress is now occurring. When a man sits down to solve a mathematical problem a very large part of what he does, not all, is repetitive and carried out in accordance with fixed formulas. When he manipulates a mass of statistics, fills out forms, draws curves of relationships, prepares statements of inventory, earnings, and the like, a large traction of his work is simply the routine application of fixed rules. All of this can be done for him by machine, much faster and more accurately than he can do it himself. It is being done, largely by digital machines, with great memories, high precision, and enormous speed. The results of such calculations can sometimes be used directly to control his machines. But more often they simply present partial results for his judgment. A machine can even calculate the progress of an election, and present interesting trends. But it cannot predict the result, unless told what rules to use in so doing, and, as we know from experience, the human instructions in that regard may be faulty. The human brain is a marvelous mechanism. Machines can excel it in precision of memory and operations, and in speed. But they by no means match its extraordinary complexity and flexibility. And, in my opinion, they never will, at least within our lifetimes. From Vice President Davis, of the Ford Motor Co. on the first day to Dr. Bush, a scientist and engineer of extraordinary experience, academic, industrial, and governmental, this morning you have had traced for you the story of automation as mechanization. From this aspect it consists in a large and growing array of mechanical and electronic devices that give producers new machines for doing new things or old things better and equipping old machines with new gadgets. The essential feature of this new step in mechanization is the application of electronics to the control of mechanical and chemical processes. The mechanical brain or electronic computer is the central feature of this development, and the practical result is the substitution of mechanical for manual controls at many points in the physical process of production. As has repeatedly been pointed out, these mechanical controls also make it possible to do things we could not do at all by manual methods, notably the production of atomic power.

In a very able and illuminating paper in the opening session of these hearings, professor Buckingham, of the Georgia Institute of Technology, differentiated four phases of what is labeled "automation"; namely, mechanization, feedback, continuous process, and rationalization. I am somewhat bothered by having "continuous process" put third in this sequence instead of second, because continuous process as been a part of the story of mechanization from early flour-milling through meatpacking to the automobile assembly line, where gravity, or steam, or electric power gave continuity but where manual controls were still heavily relied on.

The real change came when we passed from this kind of continuous-process mechanization to that in which electronic devices make it possible to dispense to considerable extent with the mental element in manual control and to use the feedback principle extensively. Under this principle, electronic mechanisms make it possible to conduct more elaborate, more economical, and more precise continuous-production operations because the outcome of the process controls the process itself, starting, altering, or stopping it so as to make it produce a desired result. This should dispose of the cliche that automation is nothing new, just more mechanization. It has its roots in mechanization, to be sure, but something new was added when electronic devices made possible the widespread application of the feedback principle.

The three earlier phases of industrialism, mechanization, continuous process, and rationalization, all continue but have been given a new dimension. When we introduce this word “rationalization” we pass over from the primarily technological to the primarily economic meanings of automation, though, of course the two are interrelated. Though Professor Buckingham included this area of discussion in his outline statement he did not develop it very far. I am therefore taking up where he left off. He said:

Rationalization in a production system means that the entire process from the raw material to the finished product is carefully analyzed so that every operation can be designed to contribute in the most efficient way to the achievement of clearly enunciated goals of the enterprise * * * The scientific, rationalist philosophy takes on numerous new implications when it can be implemented by modern electronic machinery.

In this aspect also automation is not something altogether new. It is essentially a continuation of the scientific management movement which took shape about 50 years ago. Indeed that movement toward cost-reduction or efficiency increase was even then often called rationalization, particularly in European comment. It gave rise to much agitation about technological unemployment, and to the cult of technocracy in the period between World War I and World War II. It led also to the movement toward economic and social planning. Whereas scientific management was directed toward rationalization of the operations of the individual plant or company, national economic planning had as its goal the rationalization of the operation of the whole economy. The point I want to make here is that this new development of electronic computers and controls, and this enlarged concept of continuous process makes contributions to the objectives of national economic efficiency set out in Employment Act.

Footnotes