Next: Technocracy: Engineers as Rulers

Lecture 21: The Big Machines

Large tasks require large machines. The greatest tasks of antiquity -- the Pyramids, the Great Wall -- required the coordinated efforts of tens or hundreds of thousands of people. Besides those who pulled the ropes or pushed the stones, there are those who supplied them with food, those who quarried and dressed the stones, those who decided what the workers should accomplish each day. This was a management problem comparable with the Manhattan project or the race to the moon -- perhaps made slightly easier by the drastic measures available for handling labour disputes and the flexible finish time. So a great part of the task confronting the engineers who ran these projects was the management of teams of workers.

Since this management is a necessary part of any major engineering task, it must be considered as part of our professional expertise. And we can trace technological progress in management just as we can trace progress in the sciences.

A major step forward occurred two hundred and sixteen years ago, when a Scot named Adam Smith published ``The Wealth of Nations". In the first chapter of that book, Smith described a factory for making pins:

``one man draws out the wire, another straights it, a third cuts it, a fourth points it, a fifth grinds it at the top for receiving the head ... and the important business of making a pin is, in this manner, divided into about eighteen distinct operations, which in some factories are all performed by distinct hands. I have seen a small factory of this kind where ten men only were employed, and where some of them consequently performed two or three distinct operations ... they could, when they exerted themselves, make among them about twelve pounds of pins a day. There are in a pound upwards of four thousand pins of a middling size. Those ten persons, therefore, could make among them upwards of forty-eight thousand pins a day. Each person, therefore, making a tenth part of forty-eight thousand pins, might be considered as making four thousand eight hundred pins in a day. But if they had all worked separately and independently, they certainly could not each of them have made twenty, perhaps not one pin in a day"

This constituted a major advance in the management of human effort. Divide the labour, and give each worker a single repetitive task. And for the pinmakers, it was a life sentence. Once this observation is made, what has the pin maker to do all day but to cut wire, or to grind the head? And the observation cannot be unmade. Pins are cheaper to make this way. Whoever owns the factory, wherever in the world it is, workers doing it this way can undersell workers doing it any other way.

Why should that be? The reason is simple. We say ``Quick as thought", but in fact thought is very slow. After his last fight, Joe Louis, the great boxer, was asked why he decided to retire. He said, ``I noticed I'd started to think about my next punch." While he was world champion, his punches were made at reflex speed. At that level, there's no time to think. Similarly, you can only ride a bicycle when your body learns how; if you try to think about which direction to tilt the wheel, you'll be on the ground.

So if workers are to work more quickly, we must not allow them to think. The worker's brain is an unnecessary extra, a design flaw. Perhaps it was at about this time that we started referring to workers as `hands': mill-hands, factory hands. That was, after all, their only useful part.

This way of talking sounds cruel, but it would be a mistake to lay the blame for the worker's misery on heartless factory-owners. Suppose the workers themselves owned the factory. Once they'd made Adam Smith's observation, what could they do? Unless they work in the most efficient way, they must either pay themselves lower wages or be undersold in the market.

Now let's go forward to the beginning of this century, to 1911, when an American called Taylor published ``The Principles of Scientific Management". Taylor looked at North America and wrote ``We can see our forests vanishing, our water powers going to waste, our soil being carried by floods into the sea; and the end of our coal and our iron is in sight...but our wastes of human effort are greater."

What waste of human effort was he talking about? Well, Mr Taylor, despite the impressive titles we see on the frontpiece of his book, had begun his working life as a machinist, running a lathe in a steel mill. As a result of his hard work, he was soon promoted to be a foreman. And, as a foreman, he had the task of getting more work out of his fellow workers. He soon realised that to do this job properly, he needed to know how much work a man could do, if he worked as efficiently as possible. So he designed an experiment. At that time, he was in charge of workmen loading pig-iron -- large lumps of iron weighing about 100 lbs. He took two of the strongest men, promised them extra wages, and got them to load iron in various different ways and at various different rates. Meanwhile, two college graduates stood by, timed all the men's motions, and made careful notes. The object was not to find out the most a man could do in a day, but how much he could do on a long term basis; that is, how much he could do without becoming exhausted.

In this way, a great deal of data were collected. But it took a suprisingly long time to discover any law at all that would describe the data. Eventually they did discover a law, which seemed to apply to all work involving heavy labor, that is, all work in which the limit of a man's capacity is reached because he is tired out. The law is that, for any given load, there is a fixed percentage of the day for which the man can handle that load; for the remainder of the day his muscles must be entirely free from load. The pig-iron ingots weighed 92 lbs. A man could be under that load for 43% of the day, and, as long as he was free from load for the remainder of the day, he could work at that rate indefinitely without becoming worn-out.

Now this law, as I have said, was hard to discover. And, as Taylor observed, the type of person best suited to handling pig-iron is not usually very clever. So it is very unlikely that any workman would ever discover the most efficient way of working by himself. Therefore, he must be told exactly what to do by a manager who has discovered the scientific laws governing the work. Taylor includes in his book a dialogue between himself and a pig-iron loader -- Schmidt -- who is to be trained in the new method: Schmidt has already been told that, if the method works, he will become a high-priced man; that is, he will earn $1.85 a day instead of $1.15 a day -- a 60% wage increase.

``Schmidt, you have seen this man here before, haven't you?" [pointing to a college graduate]

``No, I never saw him."

``Well, if you are a high-priced man, you will do exactly as this man tells you tomorrow, from morning to night. When he tells you to pick up a pig and walk, you pick it up and you walk, and when he tells you to sit down and rest, you sit down. You do that right straight through the day. And what's more, no back talk. Now, a high-priced man does just what he's told to do, and no back talk. Do you understand that? When this man tells you to walk, you walk; when he tells you to sit down, you sit down, and you don't talk back to him. Now, you come to work here tomorrow morning and I'll know before night whether you are really a high-priced man or not."

So encouraged, Schmidt found he could comfortably load 47.5 tons of pig-iron a day. The average load at that time was 12.5 tons a day.

This is only one example of scientific management in action. Taylor argues that, for any work whatever, there are scientific principles governing how the work should be done, and that these scientific principles will be more efficient than any `know-how' that the worker may have acquired through his experience. Therefore, management must take over the detailed planning of the work, and the worker must be trained -- as Schmidt was trained -- to work according to the plan.

Now, it may seem to us that this training is more like slavery. But Taylor has anticipated this objection. His reply is that the detailed instruction that the worker receives is just like that received by a surgeon or a college-trained engineer, and, in just the same way, it makes the recipient capable of more complex, and hence more satisfying tasks. ``If it were true that the workman would develop into a larger and finer man without all of this teaching, and without the help of the laws which have been formulated for doing his particular job, then it would follow that the young person who now comes to college to have the help of a teacher in mathematics, physics or chemistry would do better to study these things unaided and by himself."

Smith left the worker cutting wire to pin-length. Now Taylor removes any remaining freedom he has; not only is his task specified, but the exact motions he must go through in performing the task are planned for him.

There remain one more step to complete the worker's misery, and that step was made in 1913, when Henry Ford introduced the moving assembly line to produce the Model T. Now the worker is constrained in time as well as in space; his pace of working is set by the line, and the line speed is set by the management. Initially, the automotive workers were very unhappy with this new way of working, until Ford introduced another revolutionary change: he doubled the wages of his workers to $5/day. He was intensely criticised for this by other industrialists, who called him `a traitor to his class', but Ford argued that the combination of cheaper cars and higher wages would create a large new market -- for the first time, the workers would earn enough to buy the cars they were building. For the workers, the higher wages did a lot to make up for their increasingly unpleasant working conditions.

To appreciate the impact that this way of working had when it was first introduced, we need to look at how it appeared to contemporary observers. The anti-human tempo of the assembly line is conveyed by two films made at that period, Fritz Lang's Metropolis and Chaplin's Modern Times.

The plight of Chaplin and the inhabitants of Metropolis can be described as follows:

``Owing to the extensive use of machinery and to the division of labour, work ... has lost all individual character, and, consequently, all charm for the workman. He becomes an appendage of the machine, and it is only the most simple, most monotonous and most easily acquired knack that is required of him."

(From `The Communist Manifesto', Marx and Engels, 1848.)

Chaplin becomes literally a part of the machine. But, the machine works. Productivity per worker increased at an accelerating rate between 1890 and 1990. The worker has more money in his pocket. And after all, there's no alternative. If this is the most efficient way, we must work this way or go out of business.

The worker is efficient, but the worker is also miserable. And a miserable worker causes trouble. He may drink or use drugs to relieve the monotony of his job. Just making the same repetitive motion for eight hours will leave him pretty dopey, whether he uses drugs or not. So let's get rid of him. We've simplified his task till it's almost mechanical. And the new sciences of robotics and artificial intelligence give us machines more intelligent than the average worker. So why not replace the workers by robots?

The company I worked for between 1983 and 1988, General Motors, decided to do just that. Morale among the workforce was a chronic problem; there were periodic strikes, but even between strikes, the quality of workmanship on the cars was poor. Occasionally the disaffected workers would produce `Friday afternoon specials' -- cars with crushed beer cans pre-installed in the cylinders. The management knew the sources of the quality problems -- the problem was the quality of the workers. So they would create automated factories, in which every operation would be under electronic control. Chairman Roger Smith wrote up a detailed shopping list -- EDS, Hughes Electronics, forty billion dollars worth of automation. No other company in the world could come up with this kind of investment. This would not only put us ahead of the domestic competition, it would put us ahead of Japan and Germany.

Here is the result. The new technology was a disaster. Inside the new plant in Hamtramck, Detroit, spray-painting robots were painting each other. A robot designed to install windshields was systematically installing them two inches through the windowframe. The assembly lines spent more time down than they spent up. In 1988, MIT researchers studied the plant, and found it was half as efficient as comparable Japanese facilities. And strangely, the Japanese facilities didn't place much reliance on robots or computer control. Moreover, their plants were not only more efficient, but the products they produced were of higher quality. And this was true even when the Japanese plants were located in the US, often hiring the same workers as had been laid off from closed GM plants. For example, NUMMI (New United Motor Manufacturing, Inc.), a joint GM venture with Toyota, was established in 1984 in Fremont, California. Toyota took over a GM plant that had closed due to labour troubles and low productivity. Despite the fact that they were working with largely the same labour force as GM, Toyota greatly increased the plant's efficiency and the quality of its products, outperforming any other unionized factory in North America, and achieving productivity comparable to Toyota's Japanese plants.

Paul S. Adler, a sociologist who interviewed employees at NUMMI, wrote in "Time and Motion Regained" (Harvard Business Review January-February 1993), that workers are made to feel that their views are taken seriously. By contrast, at a GM plant in Linden, New Jersey, foremen yelled at workers when they tried to stop the production line. Although there are some complaints at NUMMI about peer pressure from other team members to maintain a fast pace, there is a broad commitment to the philosophy of continuous improvement.

This difference was not peculiar to the car industry. A 1981 study compared TV sets produced by Motorola with those produced by Matsushita in Japan. Japanese sets contained 5 defects per thousand. Motorola's sets contained 1500 defects per thousand.

So what went wrong? Why didn't GM's automation work, and what is the Japanese secret?

To answer that, let's go back to Taylor and `the P.O.S.M.' I forgot to present some of Taylor's conclusions. But then, almost all of North America forgot those conclusions too.

Taylor noted that workers were initially very reluctant to cooperate with his studies. They didn't want management to discover that they could do more work than they had done in the past. And this was because they believed, probably with good cause, that the employer would always pay them the same amount, regardless of how much they worked. Therefore it was not in their interests to increase productivity. But when the increase in productivity was shared with the workers, it was not at all difficult to get their cooperation. Taylor summarised his experience in four principles: It was on the second and third points that GM fell down. There was always an adversarial spirit between management and the workers, and one of the causes of this was certainly that productivity increases were not shared. For example, in 1984 management persuaded the union to forego a pay increase, on the grounds that no money was available. The following day, the top managers met and voted themselves a 17% pay increase.

This part of Taylor's theory, as I've said, was largely ignored in North America. One person who did pick up on it was W. Edwards Deming. Deming began his work in the 1930's. He was concerned with quality, and identified principles that would allow any company to improve the quality of its products, and at the same time increase the efficiency of its operations. However, very few people in management paid any attention to him. Then, in 1950, Deming was invited to Japan to advise on rebuilding their economy after the war. Management in Japan was ready to listen; they felt, perhaps, that since the West had won the war, they must have valuable knowledge. Gradually, the knowledge of his techniques spread through the country, and the results showed that they worked. Many Japanese business leaders credit Deming with starting their climb towards economic success. In 1960, the Emperor awarded him the Second Order Medal of the Sacred Treasure. Every year the Union of Japanese Science and Engineering gives a very prestigious prize to the company that has been most successful in improving the quality of its products, and that prize is called the Deming prize.

So, what are these techniques?

The workers must understand what the work they're doing is meant to accomplish, and be able to measure its results. One way of doing this is to plot a control chart. This does not show a goal to be met. It is only a way of seeing what is going on. With such a chart, the worker can distinguish between unusual causes of failure and the normal operation of the system. Defects -- losses of quality in a system -- can have two causes: normal causes and special causes. If the system is operating at its normal level, it makes no sense to reward or blame workers for chance variations in level. What we see here is what the system is capable of. If we want a different level of performance, we have to change the system. Perhaps the machine the worker is using requires maintenance, or the raw material is not of the right quality. The worker and management must consult together to change the system, since neither can change it alone.

The other kind of failure is due to special causes -- for example, here, to the spark plugs wearing out. By keeping a chart of this kind, the worker can detect special causes and stop the process when they are detected. The worker must have the power to halt the assembly line, and keep it stopped until the special cause has been eliminated.

For this to work, the worker must know some statistical techniques, and be able to plot and interpret a control chart. Something else is also necessary: the worker must be able to report the problem without fear. Why should there be fear? In many companies, fear is the normal state of affairs. If there is something wrong, someone must be to blame. So there will be a battle between the worker who says the machine needs maintenance and the foreman, whose job is to keep the cost of maintenance below a threshhold. And both of them are the enemies of the inspector, who reports defective items. This system of fear pervades American companies. One of the results of fear is an institutional blindness; managagement sends down a message ``This is the result we want''. Maybe the result they want is impossible, but the person who tells them this will lose his or her job. So each level hides what's really happening from the level above. [slide of Challenger explosion] One of the clearest pictures of this process inside a large organisation is given by Nobel-prizewinner Richard Feynman in his description of NASA managament in the period leading up to the launch of Shuttle 51C. Feynman asked the managers and the lower level engineers to estimate the probability of a failure in a shuttle flight. The engineers said ``About one in three hundred''. Management said ``One in a hundred thousand" That's quite a difference. Where did the figure of one in a hundred thousand come from? As far as Feynman could tell, the managers made that figure up because they wanted it to be true. Top management at NASA wanted there to be no problems with the booster rockets, although there was accumulated evidence that the booster seals developed problems whenever the weather got cold. The engineers who knew the problems with the seals were afraid to report them further up the hierarchy, because that would disturb management's insistence that everything was all right. So they never got fixed.

So we get back to Taylor's third principle: a spirit of cooperation and trust must be developed between the workers and management. This can be done by giving job security, as is done in Japan, and by reducing the number of levels of management and the salary differentials between them. The ratio between the lowest and highest paid employees in Japan is about four to one; in GM it was a hundred to one.

You'll remember that Taylor talked about the waste of human effort involved in turn-of-the century manufacture. But he didn't seem to notice that his approach wasted the uniquely human aspect of the worker: his or her mind. The human mind has a capability that no automatic system has: to step outside the work it's doing, look at its context, and find ways of doing it better. There is a major difference between Taylor's world and ours. That difference is the availability of universal education. That allows the worker to draw up control charts, to understand the process he's involved in, and to take a part in planning it. The worker is no longer trapped inside something he can't control.

And now, at last, some good news about General Motors. The newest division of GM, the Saturn corporation, has used Deming's principles from its inception. General Motors and the United Auto Workers union have shared in all the decision-making. Here's a report on Saturn from `Technology Review':

``Virtually every worker at Saturn feels he or she is responsible for the success of the division; workers constantly make suggestions for improving styling, engineering and manufacturing techniques; and each pays diligent attention to the quality of every part that passes his or her workstation."

And furthermore, I suggest, these workers are not miserable.

Unfortunately, GM does not seem to have been able to transfer the lessons of NUMMI or Saturn to the rest of the corporation. GM has sent managers to NUMMI, hoping they could transfer its secrets back to their own plants, but with little success. This may be because the needed improvements require a transformation of approach, rather than adding particular practices to an existing system. Deming's philosophy permeates NUMMI's operations. While engineers at other GM plants in North America saved money by replacing humans by robots, NUMMI makes use of automation to facilitate people's jobs. The idea is that technology should be a tool and that workers should add a higher level of intelligence to the machines. Toyota emphasizes the importance of creating trust and facilitating learning - key points from Deming's approach, but ones that U.S. manufacturers have neglected. At NUMMI, cooperation is fostered by rewarding those workers who help and teach others with team leadership roles. If a team finds that they can become more efficient by eliminating a job, they can recommend its elimination without fear that one of them will then be laid off: there have been no layoffs at either NUMMI or Saturn. Rather, the company treats the worker whose job has been eliminated as a valuable resource to be re-trained and used elsewhere.

To summarise: large engineering enterprises require the management of large teams of workers. The skills required for this management are thus part of the subject matter of engineering. And, particularly in this century, engineers have developed effective techniques for management. In addition to Deming's methods, these include time planning, using the Gantt chart; and the algorithms of Operations Research, initially developed for the optimal planning of military operations during the Second World War, subsequently adapted to planning production.

We use the best available techniques for planning the operations of large companies. This involves planning on a very large scale; the biggest companies, like General Motors, have the population and the budget of small countries. And yet, if we go one step larger in scale, we find there is very little in the way of planning for the country as a whole. Why not?

It was Adam Smith's view that the free market eliminated the need for explicit planning. The price system allows demand to be communicated quickly and efficiently from the consumer to the producer; by competitive bidding, consumers set the price of a good to a level that accurately represents its value to them, and this set price then invites entrepreneurs to find the most efficient way of producing the good at that price. In this way society always produces the most needed goods in the most efficient way, without the need for any outside intervention.

By the first decades of the twentieth century, though, it began to seem that this simple formula had its limitations. In particular, the Wall Street crash of 1929, and the great depression that followed, suggested that the `invisible hand' of Adam Smith's free market could not be relied on to create employment or to provide the essentials of life to the whole population.

According to Adam Smith's model, unemployment should not exist. Wherever men could not find a job at the current wage, the wage rate should just fluctuate downwards until it reached a level at which it became profitable for a company to hire more employees. But automation changed this. Now the cost of a machine sets an upper limit on what a man can be paid to do its job; and if that limit is below the cost of a week's food, the man is unemployed. This leads to a positive feedback: the more people unemployed, the smaller the market for the goods and services that the economy can produce, so the economy will shrink further. Throughout the 1930's, factories in full working order sat idle, because there was no-one to buy their output.

The irrationality of this situtation struck many observers, engineers in particular. On the one hand was the need of millions of unemployed for food, clothing and shelter; on the other were the machines and factories that could easily satisfy that need. Only an outmoded economic system stood in the way of the obvious solution.

The most radical criticisms of the current system held that money itself was responsible for the irrationality of classical economics. Considered simply as an engineering parameter, money has several drawbacks: its value is not constant from one day to the next, it can be created by anyone with a printing press, and economists can't even agree on how to measure the amount in circulation. Contrast this with a quantity such as `energy'; this can be measured accurately and repeatably, everyone agrees on what a Joule is, and it stays the same at all times.

These ideas were gradually crystallised as an explicit alternative to the existing order. Frederick Taylor, and his assistant Henry Gantt, inventor of the Gantt chart, had criticised the inefficiencies of the traditional system even before the First World War. Gantt organized a group of fifty engineers into an organization called `The New Machine', which advocated moving to an economy planned and rationally organized by engineers.

Gantt's group didn't last long; it dissolved shortly after the First World War began. A more articulate spokesman appeared at about the same time: Thorstein Veblen, a radical economist. [See the reading list for some of his books.] Veblen contrasted `industry' with `business'. Industry he saw as the natural activity of the engineer: the intelligent organization of resources to get a job done. Business he saw as the search for a profit by selfish individuals with little understanding of or interest in the activity that would generate this profit. Thus his conclusion: the control of the nation's factories must pass from businessmen to engineers.

This is the program of the movement that has come to be called technocracy, and which we will discuss in detail on Friday.