A number of technologists and economists have predicted that we are on the verge of a series of radical changes in industrial technology which will revolutionize productive processes. The consequences, in terms of human relations and social institutions, of such a revolution are certain to be enormous. But it is not so easy to foresee these consequences. Predictions in this area have to be tentative and subject to constant modification, since they are not based on careful measurement of experimentally controlled observations, but (1) on analysis of social changes following previous technological innovations, and (2) on our general knowledge of the structure and dynamics of contemporary society.
An “industrial revolution” may be said to consist in a relatively sudden and major increase in productivity. No less than three industrial revolutions can be distinguished since the end of the 18th century in England. Until then the technological equipment of Western society had remained substantially unchanged—with a few exceptions, such as the introduction of gunpowder—since Hellenistic times. While social organization had been drastically altered during this time, the only marked increase in material productivity had come as a result of that great expansion of the European economic market in the 16th and 17th centuries which later historians called the “Commercial Revolution.” The first industrial, or technological, revolution came only after the 1770’s, with the introduction of a few simple machines like the spinning jenny and the cotton gin, and the expanded use of water and steam power. The major economic consequence of this first industrial revolution was the growth of the factory system, as opposed to commercial production in the small, family-run shop. About a century later the second industrial revolution began in England with the use of electric power, which made for increasingly complex machines and the relocation of factories. The third industrial revolution emerged in the United States shortly after the beginning of the present century, and has not yet spread to all Western countries. This revolution does not involve technological improvement so much as changes in the use of manpower, its goals being specialization and mass production, and its chief industrial manifestation the assembly line.
Each of these industrial revolutions was followed by radical social changes, largely unanticipated; all were attended by a great deal of dislocation and hardship, but all eventually raised standards of living and made possible a great increase of population. If the knowledge gained from this experience can be applied effectively now to help predict the direction of social changes, it may be possible to take steps to avoid some of their ill effects in the future.
_____________
The mere fact that our society has been experiencing steady technological and other social change for almost two hundred years suggests that a new industrial revolution, such as is now being predicted by technologists, engineers, and professors, will not have the disruptive impact of an industrial revolution that occurs in a static or backward society. There is nothing in sight, as I see it, to justify the fears of a few physical scientists—who happen to be ignorant of social science—that our society will have great difficulty in adjusting to the coming changes, or that there will be widespread disorganization.
Change is a normal condition in any society; rapid change happens to be characteristic of our own; neither the coming engineering changes1 nor the attendant social changes are likely to be extraordinary in extent or speed, even though in some cases they will certainly be different from what we experienced previously.
The technological elements of the new revolution look to be, first, energy derived from fissionable materials that are easily transportable; and second, the introduction of those processes of machine production known as “automation.” A new source of energy involves no basic change in the manner of production, but will profoundly affect cost, location, speed, and other factors.
Automation will have at least three generalized types of effect on the manner of production, according to George B. Baldwin and George P. Schultz, industrial economists of the Massachusetts Institute of Technology, in Monthly Labor Review (February 1955):
1. The linking together of conventionally separate manufacturing operations into lines of continuous production, along which the product will move “untouched by human hands.” This development, which depends primarily on mechanical engineering for its adoption, we shall refer to simply as “integration,” a term already in wide use in the metal-working industries.
2. The use of “feed-back” control devices, or servo-mechanisms, which will allow individual operations to be performed without human control. “Feed-back” refers to a built-in automatic device that will compare the way in which work is actually being done with the way in which it is supposed to be done, and then automatically make any adjustments necessary. “Feed-back” technology is primarily dependent not on mechanical but on electrical engineering and techniques.
3. The development of general- and special-purpose computing machines capable of recording and storing data (usually in the form of numbers) and of performing both simple and complex mathematical operations with such data. This, too, depends primarily on new developments in electrical engineering.
Automation will not affect all industries; and even those in which it can be used are not likely to rely upon it completely. Some industries are more amenable to automation than others—notably those engaged in the production and assembling of parts, those engaged in processing (e.g. of chemicals, flour, oil), and those relying heavily on bookkeeping (e.g. insurance, banking, public utilities). Industries already largely automatic (e.g. canning, oil processing), those in which production requires constant judgment or irregular variation (e.g. slaughtering, construction), and those producing goods the demands for which are continually subject to changes in taste and fashion (e.g. the garment industry) will be least affected by the trends toward geographic decentralization and the concentration of ownership. Agriculture, mining, transportation, service industries, and retailing will be little affected directly by automation.
Most factories that automatize will not do so, as I have said, completely. The engineers think it is often feasible to automatize 80 per cent, say, of a plant’s equipment, but the remaining 20 per cent tends to be too expensive to automatize. Hence automation will not come immediately or completely in the near future. As to the use of nuclear power, the experts differ widely. Some say that it will be possible to produce nuclear power at competitive costs by 1965, while others say that “it will be 1975 or later before as much as one-tenth of America’s electricity will come from atomic plants” (E. V. Murphree, “The Revolutions of Power,” Saturday Review, January 22, 1955). Probably the best estimate is by Sir John Cockcroft, an outstanding authority, in a speech at the UN conference on atomic energy at Geneva in September 1955: “. . . the whole history of engineering development shows how rapidly capital costs fall in the early stages of new developments, and there is good reason to believe that in the second decade the cost of nuclear power will fall below that of power from coal and oil.”
On the assumption that automation and nuclear power will be introduced to a significant extent during the next twenty or thirty years, what effects are they likely to have on the structure of industry, the labor market, the lives of workers, education, government, international relations, and even on what might be called the “mental life” of society?
_____________
Industry should become less dependent on heavy fuels that are expensive to transport, like coal, oil, and water power, and on unskilled manpower in large supply; hence it should become more mobile and easier to decentralize. The present need to be close to markets and sources of raw materials will remain, however, so that decentralization will probably not be radical. Existing investments in real estate and plain inertia will probably serve further to slow down the rate of decentralization. But gradual decentralization there will be—to avoid high land costs, high taxes, congestion, and other undesirable aspects of our present large urban industrial concentrations. Another effect of automation will be to diminish the benefits to industry of moving into areas that offer a large supply of cheap, unskilled labor—the South, for instance. The skilled workers required for automation, and the schools for training skilled workers, are more readily available in the North and Midwest, and the wages for skilled workers show little regional differential. Given the advantage of the proximity of the large Northern markets, it is likely that industries that automatize will tend to locate in the North.
An economist, Wassily Leontief, has made a careful study of the costs of automation and finds it relatively inexpensive, hence capable of being introduced rapidly. He writes (in the Scientific American for September 1952):
The average [cost of automatization] for all industries would be almost six per cent . . . [and] six per cent is far from a formidable figure. Furthermore, the investment in instruments would not necessarily mean a net increase in the total plant investment per unit of output. On the contrary, the smoother and better balanced operation of self-regulating plants has already shown that they can function with less capitalization than a non-automatic plant of identical capacity. And much existing equipment can readily be converted from manual to automatic control. . . . The mechanization of the 19th century required heavy capital investment and proceeded slowly; the new technology, unhampered by such vast capital requirements, can be introduced at a much faster pace.
But larger firms will be readier to spare both the time, however short, and the capital for the change-over; they will also be in a better position to take advantage of the possibilities of decentralization. Thus automation may encourage concentration of ownership unless the government steps in to assist small firms with loans, information, etc., etc.
There will be a need for new kinds of skills, and a sharp drop in the demand for unskilled labor. To keep employed, many unskilled or semi-skilled workers will require schooling, and some categories of skilled workers will have to be retrained. It seems nonsense to expect widespread permanent unemployment as a consequence of machines taking the place of workers. Workers released from one job or industry will be required in another. But they will be needed with different skills and perhaps in different places, so that temporary unemployment and dislocation are likely.
Older workers will suffer greater hardship. Since the expense of their retraining will net a smaller total return to employers, they are less likely to be given the opportunity for reeducation. Untrained older workers will either be downgraded, or—if pensions become available at a lower age—retired earlier. On the other hand, the retirement age for the worker retrained to operate automatic machinery may be raised because of the ensuing shortage of skilled labor—one that will last for many decades—and because work itself will become easier physically than it is now.
Some categories of female labor will find it harder to obtain the kind of unskilled employment now available to them because unskilled and incidental clerical and factory jobs will be greatly reduced in number. (Waitresses and saleswomen, however, will be unaffected by automation.) Women are also less likely to be offered training as machinists and maintenance workers, which are the jobs that will be most plentiful in automatized factories.
The general upgrading of the average skill levels of workers may be expected to have some effect on unions. Skilled workers have traditionally favored strong unions and, given the need of labor in general for security amid the dislocations brought on by automation, they will probably become unionized in increasing number. Unions will probably also be strengthened by an increase in their functions. At the same time, some of their traditional bêtes noires will be eliminated: piecework pay, incentive systems, the speedup, and other features associated with individual production. Also, as part of the long-term trend associated with increasing productivity, the service industries will (as Colin Clark foresees) preempt a steadily increasing proportion of the labor force; service workers have been the hardest to unionize, but may become less so as more and more of them are recruited from previously unionized labor.
_____________
The great rise in productivity may be expected to raise real incomes and the general standard of living sharply, while further reducing hours of work.2 Such a change will, of course, have innumerable secondary effects which can only be guessed at. Two are particularly worth speculating about.
The continuing rise in the standard of living will reduce still further the consumption differentials between classes, especially as automatically produced goods are expected to be of higher quality. The lower-income groups will be able to afford more of the things now available only to the wealthy. And there will be a greater flowering of leisure-time activities. Present trends suggest a considerable variation in this development: not only will there be more recreation and social participation of the kinds already familiar to us, but small “sideline” businesses will multiply, along with “do-it-yourself” work around the home, gardening, “creative art,” etc., etc. Since many people are not yet prepared to cope with an abundance of leisure, its sudden increase will probably provoke a great deal of boredom and dissatisfaction at first. New small industries and occupations purely exploitative in purpose may arise.
Education, too, will of necessity be greatly affected by automation. Industries that automatize will in most cases be compelled to train employees themselves, probably under the joint supervision of management and unions, while the training of young people for their future careers will probably remain in the hands of the kinds of schools now in existence. Pressure is already being exerted to change these schools to meet the needs of the new industrialization. At the secondary school level, the need for technical training will be greater than ever before. Since teachers adequately trained in technological disciplines are in increasingly short supply because of the relatively low salaries in their field, not only will teachers’ salaries be raised (by Federal or industry-aid plans), but differential salaries by subjects may be expected to appear.
At the university level, the rise of engineering, natural science, and social science may be expected to continue at the expense of the humanities and “education.” The humanities will be able to retain their importance in general education by subordinating themselves to science as the major subject. But managers and engineers will more than ever need the breadth that only the humanities can provide. Higher education will continue to be increasingly important for social mobility. Leadership in industry, government, and civic life generally will depend more than ever on the ability at least to understand, if not guide or manage, the newer technology and the social readjustments it will entail.
_____________
The effects on government of the changes in technology will be secondary insofar as they will derive from the effects on industry and labor. If industry and population are decentralized, the many material and cultural services now provided by municipalities will have to be assumed by state and Federal government, or by new metropolitan and/or regional authorities not yet in existence. The exodus to the suburbs is already straining the cities’ capacity to collect taxes to provide the services which suburbanites still use. If industrial ownership becomes concentrated among a decreasing number of firms, government controls in industry may be expected to multiply. Since the new “monopolies” will, presumably, not be inefficient or try to restrict production—in fact, they can be expected to do the contrary—the new government controls will move in the direction set by the Federal Communications Commission and the Securities and Exchange Commission, rather than in that of the old-fashioned “trust-busting” of the Sherman and Clayton acts.
A broadened social security program—as safeguard against unemployment, interruptions of income during retraining periods, and in old age—is already being recognized as essential to personal and social stability. Social security will become even more necessary with increasing technological unemployment and the loss of incidental privileges such as pension rights, vacation rights, seniority, etc., etc. Probably not all of the different kinds of security benefits will be provided by the government—some unions have already moved to require industry to provide severance pay and a “guaranteed annual wage.”
Government will play a new role in education, under the pressure of industry and labor alike. Federal aid to education is about to become a reality, and in the coming years it may have to provide a major part of school revenue. It is even likely that the Federal government will grant increasing funds to private colleges; it has already begun to do so in the form of research subsidies and agricultural extension work.
Internal migration policies, and agencies to implement them, will probably become necessary in all industrialized countries. Information about job opportunities, job requirements, and wage levels can parallel the crop-reporting service the Federal government has long provided. Small loans to pay for transportation of families and household goods cannot be provided by the usual lending agencies, limited as they are by their collection facilities and by state law; a Federal agency will probably have to be set up for this purpose. Social services to provide aid in finding housing and in helping migrants adjust to new communities will be greatly in demand. These and other related innovations can be handled by a “Division of Migration Service” in one of the existing departments of the Federal government.
_____________
A major factor in world politics today is the difference in productive capacity between the industrialized and the so-called “underdeveloped” countries. The new technology will only increase the difference. The first effect of the newer technology will be to make it even more difficult for the underdeveloped countries to “catch up” with the West. The abundance of capital and of workers with specialized skills required for automation are lacking in backward countries. But if capital and training are provided for them by more advanced countries like our own, they may be able to modernize, and even ultra-modernize, their economies with relative swiftness.
The underdeveloped countries have, however, one advantage: they are not saddled with large capital investments in technological plants inherited from the first, second, and third industrial revolutions. Such investments may retard the advanced countries unless they are willing to scrap much of their present “wealth”; England is the notable example of a country that was held back because her capital commitments to the first and second industrial revolutions prevented her from taking advantage of the third. Some engineers (according to A. H. Raskin in the New York Times on April 8, 1955) estimate that 70 per cent of all machine tools now in use in the United States have become outmoded. That is quite a large investment to scrap.
Automation will also be helped in underdeveloped countries by the fact that most of their towns and cities are small and scattered. Automatic production does not need the large urban agglomerations developed in the West under industrialism over the past one hundred and fifty years. Introduced into hitherto un-industrialized countries, automation can avoid the dislocations caused by extreme and rapid urbanization.
Atomic energy will also be a special advantage for backward countries once it becomes available at competitive prices. Fissionable materials (and materials that are likely to become fissionable under the attack of science in the coming decades) are much more widely scattered over the earth’s surface than coal and oil, and in any case much cheaper to transport. Thus the coming industrial revolution should be more easily exportable to Asia, Africa, and South America than were the earlier industrial revolutions (though the training of labor will still remain a major problem in these areas). Technologically advanced countries like Italy and Japan that lack coal and oil will likewise be helped by the shift to atomic energy, and the prospects are for a further lessening of the great differential in productivity between the West and the rest of the world that has been characteristic since 1775.
_____________
We have heard much of the ruinous effects of over-specialization and the meaninglessness of work on the assembly line. The fourth industrial revolution may be expected to check and even reverse this tendency. The typical new worker will be skilled and technically educated, not the unskilled or semiskilled hand that the earlier industrial revolutions called for. In an automatized factory he will be able to “see” a given productive process from beginning to end. His main job will be to control, tend, and repair machines, and this active role will not permit the machines to set the rhythm of his work.
There is a dangerous tendency current to treat automation and the technological changes it will entail as things mysterious and superhuman. Scientists, engineers, and some businessmen are among those who most frequently conjure up such spectres. Labor leaders have tended on the whole to avoid Sunday-supplement language in dealing with the subject. Economists and other social scientists seem the most restrained.
The mathematician Norbert Wiener (of “cybernetics” fame) has perhaps been the one most guilty of using scare language in speaking of automatic machinery and its effects. A typical and widely quoted statement from his book The Human Use of Human Beings (Houghton Mifflin, 1950) is: “It is perfectly clear that this [automation] will produce an unemployment situation in comparison with which . . . the depression of the 30’s will seem a pleasant joke.” Another mathematician, John von Neumann, has “proved mathematically” that a machine can be constructed which will “reproduce itself” and that its offspring will likewise be capable of “reproduction,” but by this he only means that a machine could be constructed and programmed so that, when the proper parts are inserted in it, the machine would assemble them into another machine just like itself.
A tendency to personify automatic machines by lending them the attributes of human beings likewise serves to make them seem more frightening. Edmund Callis Berkeley has entitled his book on the new computing machines Giant Brains. The machines themselves—which, after all, do nothing but add and subtract at a tremendous speed—have been given such names as “Mark” and “Edna,” which are not frightening in themselves but do imply that such machines have human powers. At least two electronic processes have been described by their inventors as “learning.” Out of the Bell Telephone Laboratories has come a mechanical “mouse” that can “learn” to find a piece of cheese in a maze with increasing speed. (The first time this little mobile electronic gadget bumps into the walls of the maze almost at random, but because it records these “errors” it is able the second time around to thread the maze without a false move.) At the Cambridge Mathematical Laboratory in England, Dr. Maurice V. Wilks has invented an electronic device that can be “conditioned” gradually to respond to only one out of all the digits; this is also described as “learning.”3 Without disparaging such devices, an eminent neurologist, Professor Warren McCulloch of the University of Illinois, has estimated (New York Times, August 8, 1954) that to build a gadget with as much “intelligence” as an earthworm, it would take all the electronic power that could be generated by Niagara Falls, and would require all that cataract’s water to cool the electronic tubes.
One of the sanest public statements about the new electronic devices has come from Robert Bendiner, a political writer (and contributor to Commentary), in the New York Times for January 23, 1955:
To appreciate the accomplishments of talented gadgets . . . is one thing; to endow them with human attributes . . . is quite another matter. The human brain is not out moded. In any discussion of these so-called Giant Brains . . . it is strangely necessary to make this point at the outset—because if there is one thing more remarkable than the machines themselves, it is the readiness of people to imagine that these assemblies of metal, vacuum tubes, and wire literally think, remember, learn from experience, use judgment and occupy some new and shadowy ground between the inanimate and the human. In sober truth they do none of these things, but even a cursory review of their behavior and achievements shows how tempting it is to engage in just this fantasy.
Mr. Bendiner goes on to point out that the machines do only what they are “programmed” to do, and are incapable of meeting an unforeseen situation; that they do not truly recognize but merely compare; and that they are incapable of generating an idea or anything else new. . . .
_____________
There will probably be fewer heavy, dirty, and boring jobs in automatized factories than in the equivalent contemporary plants, and factory work generally will be cleaner and safer. Many workers will be able to wear white collars instead of overalls. John Diebold goes so far as to say:
In an odd and entirely unexpected way, automation may bring us back to the human and psychological values of the self-respecting craftsman. Electrical and mechanical repair work, instrument adjustment, and general mechanical tinkering can provide challenges, pleasures, and satisfactions very much like those enjoyed by the swordsmith or cabinetmaker of old.
The engineers, Gordon S. Brown and Donald P. Campbell, writing in Scientific American (September 1955), make this more specific:
Even in the most robotized of the automatic functions there will be many men, and they will have interesting and responsible jobs. They will be freed from the timing, nerve-racking or even boring jobs of today’s mass manufacturing. To win this freedom, however, they will have to upgrade themselves in skill and sophistication. The new controllers and instruments will call for a higher level of precision of repair and maintenance. A $50,000 controller cannot be hit with a hammer if the shaft doesn’t fit into the hole on the first try. Men who have heretofore thought of electronic equipment as simply a metal chassis with tubes will become conversant with switching, flip-flop, peaking, and other circuits. They will have to judge when to repair and when to throw away rather than stop production. We have here a paradox; today we cannot afford not to have lots of control, because a half-day shutdown of a plant may mean a $100,000 loss in potential sales.
These robots are not hurting the workman—they merely coax him none too gently into taking more responsible jobs, making bigger decisions, studying and using his mind as well as his hands.
On the other hand, such downright unpleasant jobs as those in the mining and slaughtering industries will probably not be susceptible to automatization in the next few decades. But hours of work will go down in all industries, and this will make even the most unpleasant of occupations more bearable. The mental life of the worker of the future will be much more determined by what he does during his leisure time than what he does during his working hours. As his job ceases to dominate his life, political, family, religious, and “cultural” activities may take on greater significance. It is difficult to see how a democratic state can channel leisure-time activities, except by making the facilities for them, whatever their nature, more freely available. But it is likely that the things that fill the leisure time of the average citizen will shape the future to a greater extent than any other set of factors.
All in all, the net prognosis for the coming industrial revolution under automation seems to be for a happier, more vital people, not a Huxleyan “brave new world” of mechanical monstrousness. Planning, foresight, and a modicum of intelligent action on the part of the public and private organizations should see to that.
_____________
1 See John Diebold, Automation: The Advent of the Automatic Factory (D. Van Nostrand & Co., 1948).
2 It is difficult to predict the extent of the rise of wages and reduction in working hours, as these will be a function not only of the increase in productivity, but also of (1) the extent to which the proceeds of increased productivity are spread throughout the economy; (2) the probable increase in economic demand; (3) the change in the age structure of the population (it is estimated that the population will increase by 20 per cent by 1965, but that the number of those of working age will rise by only 10 per cent). Still, a responsible economist has predicted a four-day work week with a 20 per cent increase in average income by the 1970’s (George L. Ridgeway, Director of Economic Research for International Business Machines Corporation, in the New York Times, March 27, 1955).
3 It seems that the psychologists are as much to blame for this distortion as the physicists and mathematicians, since they have for years been referring to “conditioning” and “trial and error learning” as equivalent to all of human learning, rather than merely as one type.
_____________
