Main FI Index | Main Newspaper Index
Encyclopedia of Trotskyism | Marxists’ Internet Archive
From Fourth International, Vol.15 No.2, Spring 1954, pp.53-58.
Transcription & mark-up by Einde O’Callaghan for ETOL.
THE development of automatic tracking and firing apparatus for anti-aircraft batteries during World War II introduced the germs of a new industrial revolution. The process thus portended is already under way. We have before us the technological possibility of freeing mankind from drudgery, providing mankind with incalculable material abundance and giving humanity the leisure time to develop its full potentialities. We have at hand the beginnings of a self-operating means of production. The name given to this new industrial revolution is automation.
The first industrial revolution came as one of the effects of capitalism’s unquenchable thirst for relative surplus value. Labor-saving machinery, increasing the productivity of labor, cut down the amount of socially necessary labor in a commodity; thereby breaking through a profit barrier. This first industrial revolution brought a persistent lowering of the value of labor power, an increase of constant capital at the expense of variable capital, and an inescapable decline in the rate of profit.
Now a new industrial revolution, automation, has entered upon the scene – a consequence, again, of capitalism’s lust for relative surplus value. Automation raises the contradictions of capitalist industrialization to a new intensity: technological unemployment beyond yesterday’s wildest fear’s, astronomical quantities of constant capital for each worker directly employed, and a plummeting rate of profit. With automation, the capitalist “spider” has taken a “wasp’s egg” under its skin. Automation, a qualitative change in the means of production, hastens the doom of an outdated society. Automation carries with it an intensification of the social and political forces that will drive the working class to take power and reorganize society from top to bottom.
Already the beginnings of this new industrial, revolution are met in such significant areas of the economy as Ford’s engine plant in Cleveland. (See Electrical Manufacturing, August, 1953.) The various journals which circulate among management and engineering staffs are crammed with both ads and articles featuring the “gimmicks,” instruments and methods of automation. Parts and units specifically designed for use in constructing computers are manufactured and offered in quantity by an increasing number of firms. Large banks are advertising their willingness to finance automation in industry. Practically every major manufacturing firm in the country has some kind of automation plans already in development.
Part of this development of automation is a by-product of the munitions industries. Modern jet aircraft fly too fast for the response rates of human reflexes. More and more of the control of these craft is passing from the pilot to various types of electronic “brains.” A supplier of parts receives orders for a certain quantity of such essential computer elements as servo-mechanism units. Soon this supplier is in position to produce more servo-mechanism, units than his Air Force contract requires. He advertises the surplus on the open market. This pattern is reproduced over the entire electronics industry.
The manufacture of radar equipment, television sets, etc., generates productive facilities which are readily convertible to manufacture of control circuits. The general scramble for odds and ends in a shrinking internal market forces the process to a new pitch. We have only to survey the topics of articles and ads in industrial and professional engineering journals during the past few years to see some of this process in operation.
Thus it is the inner, inescapable logic of US capitalism that drives it to wade in the seas of automation – seas in which it cannot swim.
What we shall do here is to show why automation represents the beginning of a new industrial revolution, why it is not merely a continuation of the old industrial revolution. We shall show why capitalism, for the most profound social and economic reasons, cannot complete this revolution. Finally, we shall show how automation relates to the problems of the socialist revolution.
The history of man’s economic, social and political development revolves around his invention and development of tools. It is by the implements of chipped and flaked stone that we identify paleolithic man. The spear, the axe, the bow and arrow raised man’s food-gathering power above that of other animals. The invention of the plow speeded the development of agricultural economy and the social and political forms which evolved from agrarian society. In each stage of man’s social and political development we look for the root in the changes in man’s relationship to tools – changes in the means of production.
The first industrial revolution under capitalism, through mechanization, took the motive power and the tools from the workman and transferred them to the machine, but kept the workman as an appendage of the machine. (Marx, Capital, vol.I, chap.15.) Automation changes this relationship between man and his tools. From machines to make machines, our technology has now advanced to machines that control machines – and thereby machines that control themselves. The appendage of the machine – the worker at the machine – is junked. A small portion of this displaced labor force will find jobs in the control room and on maintenance teams. Further steps in automation will eliminate even these.
Automation eliminates the human appendage
- by devices which automatically transfer work from one machine in a sequence to another;
- by built-in guides and feeds;
- by “monitoring” devices which constantly or regularly inspect the operation’s output and automatically make the adjustments to the machine that meters, gauges, etc., indicate to be necessary;
- by coordinating devices, such as computers, which integrate each machine into a larger whole, and which make the whole production line work as one self-operating machine. In general, the devices and techniques for accomplishing this already exist.
Some further features are not yet developed:
- constructing machines which are effectively “self-repairing”;
- constructing machines which can add new elements to themselves automatically.
Neither of these two problems is more than an engineering problem; automation hasn’t yet reached that stage of practical development where these two features have received much attention. Nonetheless, a fully automatic factory must incorporate these features.
Thus it is technologically possible. through automation, to eliminate most of the labor force in industry today. This is not science-fiction; it is fact, as more and more workers will realize shortly.
The one key principle that underlies automation is the principle of feedback. We shall attempt to make this principle clear through illustration.
Manufacturer Jones walks into the office of his engineering staff and announces a problem he wishes solved. In his factory there are a number of electroplating tanks which are cooled by water running through coils placed in the tanks. In this case it is important to keep the tanks at a temperature of 70 degrees Fahrenheit, with a tolerance of plus or minus two or three degrees. This means that the amount of water flowing through the cooling coils must be carefully regulated. If too much water runs through the coils, the tank temperature will drop too low; if not enough water flows through the coils, the temperature of the tank will rise too high.
It is not possible to run a fixed amount of water through the coils, for two main reasons: first, the tendency of the tanks to heat up varies with the amount of water and the temperature of the shop air; second, the temperature of the water flowing through the coils varies. So the manufacturer has to hire labor to regularly adjust the valves on the cooling coils.
| |
Since Manufacturer Jones, like any successful business man, is money-hungry, there is a strong probability that he will perforate an ulcer unless his engineering staff finds a way to drop that “extra” labor from the payroll.
Fortunately for Jones’ ulcers, his engineering staff pops up with a quick and easy solution. On each valve they install a small, reversible motor, geared way down. In each tank they place a thermo-couple, which puts out varying currents according to the tank temperature. Now they take the wires from the thermo-couples and run them to control boxes which regulate the valve motors. When the temperature starts to rise, the electronic control starts the motor to open the valve. When the temperature starts to fall, the electronic control starts the motor in the opposite direction to close down the valve ... and the “extra” workers get their notice.
A simpler but similar device operates the automatic oil-fired central-heating system in a modern home or office-building.
What has been done? Electrical “information” – a quantity of voltage, amperage or phase difference – from the output (in these cases, temperature) is transmitted to the device regulating the input (in these cases, a source of cooling or heating).
The same problem might be solved hydraulically. The engineers at Jones’ might have put a ball of fluid in the tank. Assuming that this fluid expanded and contracted in volume fairly rapidly with changes in temperature, they could connect the ball to a spring-loaded valve on the water line, so that when the tank temperature went up the fluid would expand, opening the valve; and when the temperature went down in the tank, the fluid would contract, allowing the spfing to force the valve partly closed.
In this kind of solution to Jones’ problem, hydraulic “information” about the temperature would have been used to control the device regulating the input.
In either case, this transmission of information about outputs to control devices regulating inputs is called feedback. Whether the information is electrical, hydraulic or mechanical, the principle is the same. Many modern computers use all three kinds of information, according to whichever kind is the most efficient and least costly.
Obviously, the same principle can be used to “tie” a reading from a meter, gauge, micrometer, etc., to a small motor or valve on a machine – to make the machine “self-adjusting.”
What, then, does a worker do to his machine? He reads a meter, gauge, micrometer, counter or blueprint, takes an order from some central authority – “information” – and adjusts, starts, sets up, stops the machine. The worker, then, represents to the machine a small amount of power used to adjust controls, and a “nervous system” to handle the nerve impulses (information) to control those iron muscles. The worker does not use all his intelligence, but only a very small part of it, to do his job on the machine. (Cf. Marx, op. cit., pp.461-462.) The factory uses the principle of feedback, built into the human being, to control the variable features of the machine. (Cf. Wiener, Cybernetics, 1948, chap.5.)
Now, by utilizing this principle of feedback in control and computing devices, it is possible to eliminate the production worker from his last function. Naturally, where the job is the most dull and repetitive, the possibility of relatively cheap automation is the greatest. Where a greater amount of skill is involved, it may be more expensive to automate.
It is clear that most workers are threatened with replacement by a machine. The practical question that faces the average factory worker is: How far can capitalism go with automation, and how long will it take them to get to me?
Unquestionably we are going to see a lot of automation in the next few years. Ford’s Cleveland engine plant is a clear and unmistakable warning of things to come. A project called “Operation Tinker Toy” threatens to eliminate a large percentage of the workers in the electronics parts manufacturing industries, leaving out the effect of the rate of growth of the electronics production industries. The major tool industries are showing a galloping interest in producing automation equipment, with a growing army of engineer-hucksters peddling the automation-products from door to door in industry. Can Wall Street go all the way to essentially man-less production lines?
We may answer that categorically: No.
The Bureau of Standards is now analyzing “Operation Tinker Toy,” according to a note in a recent issue of Electronics. “Operation Tinker Toy” presents a new method of assembling the components of radio, TV, radar, etc., in a manner suited for automatic production. Here are some of the figures given for the cost of producing 400 units of a certain assembly unit per hour:
|
|
Direct Manufacturing |
|
||
|---|---|---|---|---|
|
Method |
Materials |
Labor |
Overhead |
Total |
|
Conventional |
$ 35.85 |
$ 5.60 |
$ 5.44 |
$ 46.80 |
|
MDE (hand) |
20.56 |
2.27* |
2.27* |
28.23 |
|
MPE (auto) |
20.56 |
2.83 |
2.86 |
26.25 |
|
* (Source gives $2.27 for each – undoubtedly an error. – L.M. |
||||
What is the capital investment in each method? Electronics’ sources do not give figures for conventional methods, but a comparison of MDE (hand) with MPE (auto) gives us the kind of information we seek. MDE requires a capital investment of $82,000 for a productive capacity of 400 units per hour. MPE requires a capital investment of $665,000 (sic) for a productive capacity of 405 units per hour. For less than 10 percent decrease in the cost of production, your electronics capitalist must increase his capital investment by over 700 percent!
“Operation Tinker Toy” consists of light operations: dip-soldering, tube-insertion, etc. Now turn to the behemoths of steel and auto production – what fantastic ratios of capital investment are required to automate these giants? Only 8 to 1? We may well doubt it.
In other words, automation means, in the first place, a tremendous increase in constant capital relative to variable capital. Staying on the conservative side, let us assume that it only halves the labor force at only eight times the investment: the effect on the rate of profit becomes starkly clear. (Cf. Marx, op. cit., pp.444-445.)
How much capital would be required to automate a major part of the US economy? There is no precise estimate available, for obvious reasons; but from the preceding discussion, it is clear that the amount would be “astronomical.”
Thus automation presents those sections of the capitalist class lacking sufficient backing to convert to the new process with a potential squeeze many times more severe than similar past developments which sank powerful robber barons in the competitive struggle. Finally, automation signifies a great increase in technological unemployment, and therefore an ultimate narrowing of the market. Since automation represents the next stage of development in the means of production, we may scratch our heads in vain for a way for capitalism to escape these consequences.
At the present level of productivity, even the mighty US industrial apparatus cannot produce enough wealth to undertake automation at a leap. This holds true for either a capitalist US or a US under a Workers and Farmers Government.
What is possible is the planned achievement of automation through definite stages. Let us arbitrarily call these stages A, B, C, etc. – with stage A representing the amount of automation possible now. By accomplishing stage A we will raise the national productivity to a higher level, which will permit the advance to stage B. Then B will permit C, and so on.
Now, each stage of automation has the characteristic of linking together ever larger areas of the industrial capacity. Stage A might represent the automation of assembly lines within the factory. Stage B might represent the tying together of the whole factory as one automatic machine. Stage C might be the linking of steel mills to iron and coal mines, etc., as one automatic, man-less assembly, and so on.
Obviously, in order to put these assemblies together as one machine, they must be compatible. This requires national standardization of equipment, so that parts later to be fitted together will actually fit together. Otherwise, the rate of automation will be slowed down to a snail’s pace by the necessity for junking large quantities of useful national wealth and replacing it over and over again, simply because there was no planning.
Automation is going to produce great social changes, whether under capitalism or under a Workers and Farmers Government. Millions of jobs rationalized out of existence mean that workers have to have new jobs, jobs which should be at least as good or better than the jobs they have lost. If national planning is in play, it is feasible to coordinate the wiping out of old, outdated jobs with the creation of the new jobs which have to be done. Also, as the amount of socially necessary labor decreases, national planning will enable the general reduction of hours with planned increases in “pay.”
Again, automation not only wipes out jobs, it wipes out the need for old-style, repetitive factory labor. In place of production workers, we will need an equal or greater number of engineers and scientists. Our whole educational system will be hopelessly outdated by these changes in the means of production. Educational changes must be made so that we may have the skills we need. That is another problem of national planning.
At the level of productivity which automation brings about, the problem of natural resources – already an acute world problem – becomes a major issue. The archaic and wasteful use of coal for fuel, wood for houses and metal for products destined for junk, must be brought to a halt. The thoroughgoing national and international planning of the conservation and replacement of world resources is absolutely essential if we are to survive and raise our standard of living.
We have shown that full automation under capitalism is impossible. Some bourgeois sources agree with this conclusion – but for the wrong reasons.
From the Philadelphia engineering offices of Minneapolis-Honeywell, a firm which presumes to know something about control systems, we hear that “push-button” factories are not foreseeable. Minneapolis-Honeywell engineers point out that a computer with an “intelligence” comparable to that of an ant would require a structure the size of the Pentagon building in Washington. They inform us that it would require the amount of power necessary to service a city the size of Philadelphia. Also, they assert, it would require a cooling system equivalent to the flow of the Mississippi River.
How does this analogy apply to the problem of automation? What lies behind their thinking here? It is not only Minneapolis-Honeywell engineers who are laboring under a delusion here; many other major controls manufacturing firms make the same error. Their difficulty arises not from a slip of the slide-rule, but from their abysmal ignorance of capitalism and the vulgar social prejudices which they drag into the engineering laboratory.
Marx remarked of the bourgeois political economist:
“... when considering the capitalist mode of production, he ... treats the work of control made necessary by the cooperative character of the labor process as identical with the different work of control, necessitated by the character of that process and the antagonism of interests between capitalist and laborer.” (Marx, op. cit., pp.364-365.)
Since engineers learned their little mishmash of economics from third-rate poll-parrots of the same variety, it is not to be wondered that they miss this all-important point. They mistake the class role of the boss for the necessary direction of the productive apparatus:
“It is not because he is a leader of industry that a man is a capitalist; on the contrary, he is a leader of industry because he is a capitalist. The leadership of industry is an attribute of capital, just as in feudal times the functions of general and judge were attributes of landed property.” (Ibid.)
In other words, the Minneapolis-Honeywell engineers implicitly assume that the function of the boss is essential to the means of production. From that they assume that automation requires the imposition of intelligence on the production line from the top down, that automation requires a hierarchy of “capitalist” thinking-machines to replace the existing hierarchy of bosses.
Exactly the contrary is true. The development of the means of production has outdated the capitalist economically and socially. That means we can dispense with the boss and his equivalents altogether.
Let us recall our illustration of Manufacturer Jones’ plant. A man was removed from production. Did we replace him with a machine with the equivalent of human intelligence? By no means. A motor, a thermo-couple, and a few wires and tubes did the job quite nicely – even better than the human operator. In principle, we shall have no greater scientific problem in “laying off” the bosses: automation makes them “extra labor.” In fact, getting rid of the boss is equivalent to an essential technological improvement in the means of production.
Let us take our Minneapolis-Honeywell engineer sluggards to the ant and teach them a few lessons.
The individual ant is not particularly intelligent as insects go; an ant is a muscle-head through and through. However, the ant colony displays a marvelous degree of over-all intelligence. From whence this intelligence? From a super-ant “boss” hidden among his bonds and coupons down in the hill? Not at all. The intelligence of the ant colony is greater than the intelligence of all its members; it is the product of all the ants functioning in a social organism; this intelligence is a social product.
The intelligence of the ant colony is derived from the organic evolution of ant colonies, as the intelligence of a human being is a function of matter organized through organic evolution, an effect of natural selection. There is nothing mystical about it; as any dialectician knows, when you put a lot of similar objects together you obtain a whole which is something quite different from its parts. The intelligence of the ant colony does not reflect an average intelligence in each ant, but arises from the particular organization of ants as a whole.
A similar “law” holds when individual workers are put together in a factory. Turning again to Marx:
“When numerous laborers work together side by side, whether in one a,nd the same process, or in different but connected processes, they are said to cooperate, or to work in cooperation.
“Just as the offensive power of a squadron of cavalry, or the defensive power of a regiment of infantry, is essentially different from the sum of the offensive or defensive powers of the individual cavalry or infantry soldiers tnken separately, so the sum total of the mechanical forces exerted by isolated workmen differs from the social force that is developed, when many hands take part simultaneously in one and the same undivided operation, such as raising a heavy weight, turning a winch or removing an obstacle ... Not only have we here an increase in the productive power of the individual, by means of cooperation, but the creation of a new power, namely, the collective power of masses.” (Marx, op. cit., pp.357-358.)
This may be instanced in the prob-lem-solving power of certain groups of individuals, in relation to the problem-solving power of the same persons working individually. The whole is greater than the sum of its parts; the intelligence of the group is greater than the intelligence of the sum of its members taken separately. We may say that this organization produces a new intelligence, a social intelligence, a social product, which, for lack of a proper term, we may call a “social brain.”
Needless to say, not all kinds of organizations of human beings produce an efficient increase in social intelligence, just because they seem to be cooperating groups. Fifteen real Bolsheviks, organized as a “problem-solving group” in a Bolshevik organization, generally present more collective social intelligence than a thousand Mensheviks, organized in one of their bleating conventions. This is historical fact, as any candid and intelligent student of history will admit. The source of the difference lies largely in the fact that Bolshevik organization represents a group integrated about the performance of a function, whereas Menshevik organization is linked to the performance of no continuous practical function. The cooperation of workers in the means of production forms them, into an efficient problem-solving group, evolving on the basis of its efficacy in improving productive output. It is for this reason that we place so much justified confidence in Bolshevik organization and in the social intelligence potential of factory workers in cooperative productive groups.
Now, applying these principles to automation, when we “slave” a machine to the output of a standard of production, we have given that machine enough “intelligence” to do its job, and not much more. We don’t concern ourselves with the pretty problem of making the machine as “intelligent” as the man it replaces; we merely construct the machine to do the job required.
Let us take the example of several automated machines in a sequence. When machine A does its job on a part, it ejects its finished product, which then goes into B. But what if A begins to outproduce B? B simply doesn’t pick up any more parts from A than it can handle. This prevents A from ejecting its parts ahead of schedule – so A has to slow itself down to the proper pace. No boss is needed. By simply putting the machines together in the proper fashion, we have machines which “automatically” solve their problems.
The viewpoint of Minneapolis-Honeywell engineers would not permit this. They would say: “Let’s build a computer which will handle this kind of problem, and tell A, in that case, to slow down.” What would they do? They would build an analog (the electrical or other equivalent of a working-model of A and B). Now, if A “wished” to step up its production, it would first have to send that information to the computer-boss. The computer would “try out” A’s higher output on the scale-models. Then, discovering that B couldn’t handle the work, it would tell A to slow down. What has been accomplished? Minneapolis-Honeywell engineers have built a very expensive and by no means fool-proof machine to tell A what B could tell A all along without a computer-boss: Slow down.
You don’t have to have computer-bosses to make automation work. A machine is the best working-model of itself that can be built. Once you have built sufficient, information and control networks into the machine for it to do its own job automatically, it is able to transmit, at the same time, any essential information about its performance to other machines – the computer can do no better. Merely by hooking these machines up in the proper way we have made them capable of doing group problem-solving. The automated factory thus completed has enough “intelligence” to handle all the jobs for which it is constructed. If we discover that this is far less than the intelligence of a human being, that only reveals how much we have been wasting human intelligence under capitalism.
Finally, it is important to emphasize the separation of truth from claptrap concerning the “intelligence” of the machine. The capitalist notion of robot “intelligence” springs from a society where capitalist bookkeeping reduces the worker to a cost item along with oil, electricity, tools and raw materials. From the capitalist point of view, the worker is nothing more than he is on the production line, an appendage of the machine. Thus, when the capitalist replaces a worker with a robot, the capitalist assumes that the robot is as intelligent as the worker replaced.
In fact, the intelligence of the worker, as history proves, is of another sort. The worker’s intelligence fits him not only for slavery to his master or his master’s machine. The worker’s intelligence is sufficient to accomplish what no machine has yet threatened to do: topple a whole rotting society and replace it with a new society. In automation we do not reproduce the “intelligence” of a worker: the “intelligence” of our automated machines is less than what the biologist meets under the microscope in that simplest of animals, the amoeba.
What kind of organization of the labor process occurs under automation? It parallels the social organization of a workers’ state. This is not an accident. We have reached the point of historical development where the means of production have outgrown capitalist society. At this point the working class is impelled by the unbearable contradiction between social production and private appropriation to take the leadership of society and reconstruct it along patterns compatible with the new development of the means of production – automation.
Automation, under capitalism, will be distorted, gross and brutal. The falling rate of profit, impelled by the tremendous investments that automation requires, will drive the capitalists to take fresh advantage of competition among the workers over vanishing jobs. For the worker, it means the constant threat of unemployment, destruction of whole sectors of skilled work, and a savage drive to lower the standard of living.
The establishment of a Workers and Farmers Government in the United States, on the other hand, means that automation will be used to lower the average working day and raise the standard of living. Socially, automation under a Workers and Farmers Government will free mankind from degrading slavery to the factory production line, and give humanity the leisure to raise the cultural level to a point which would now seem impossible.
On the international scale, an automated US industrial power will be able to end hunger and poverty on a world scale, to really raise the level of life of colonial peoples to our own level. We shall be able to break down all national boundaries and make a universal reality of that dignity of man which capitalism preaches on Sunday occasions, that dignity of man which capitalism works so mightily to suppress and corrupt 365 days a year.
Since freedom and culture cannot exist in the face of want and exploitation, automation in the hands of the socialist revolution represents the only real solution to the ills of humanity. With that knowledge it cannot be much longer before the workers and farmers of America exert themselves politically to undertake both tasks.
Main FI Index | Main Newspaper Index
Encyclopedia of Trotskyism | Marxists’ Internet Archive
Last updated on: 31 March 2009