Sounding the Alarm
Most Notorious Victory.
by Ben B. Seligman.
Free Press. 441 pp. $7.95.
Automation, that vague and portentous word, has come to signify two major processes. One is the profusion of information created by more and more powerful generations of computers; the other is the movement, rapid in industries like oil refining and chemicals and slower in most others, toward the creation of nearly automatic factories equipped with masses of machinery, animated by sensitive control devices, and overseen by a few—a very few—human custodians. When the computer is linked to the factory, as in the almost uncanny maneuvers involved in numerical control, little occasion seems to remain for the exercise of human judgment and human skill.
Almost intuitively, it is hard for anyone to escape the sense that such technical events as these are fated to transform the society that permits them to occur. This is indeed the opinion of a good many students of American society. Automation was prominently featured in the manifesto of the Ad Hoc Committee on the Triple Revolution, and the guaranteed annual income was identified as the major way to cope with automation’s consequences. The Santa Barbara group led by W. H. Ferry at the Center for the Study of Democratic Institutions has taken an equally serious view of the impact of automation and plumped for extensive national planning as the sovereign remedy. Ben Seligman’s excellent book supports this position.
Since at least the English Industrial Revolution of the 18th century, technical change has been a defining characteristic of Western societies. Hence to argue that automation substantially differs from the incessant intrusion of technical novelties upon routine industrial processes is to allege that the speed of contemporary innovation has increased to the point where the disruptive consequences of technical innovation strain the absorptive capacities of the social system. The impact of automation, in other words, is distinguishable from previous technological effects because in so many industries, offices, banks, universities, schools, and government agences all at once work has been redefined, old skills and specialties consigned to the scrap heap (along with their proprietors), and new capacities demanded of human beings too elderly or too set in their vocational habits to stand a chance of acquiring these new talents. Accordingly, the adjustment problem is far too grave now and potentially almost catastrophic to permit ourselves the complacencies of such received solutions as tax cuts, wars on poverty (on the cheap), and manpower retraining schemes. In this view, the effective remedies for the present, and still more for the future, must be as substantial as the social effects of automation themselves. Possibly we require new concepts of work and leisure. Certainly we need to invent mechanisms for the distribution of the material abundance which automation creates; and these mechanisms should be quite independent of any need to labor at “jobs” as they are now defined.
This “alarmist” case is accepted by only a minority of professional economists. The official professional view is quite different and quite comforting. For example, summarizing and endorsing the analysis of last year’s report on the National Commission on Technology, Automation, and Economic Progress, three well-regarded economists in a Brookings study write:
The Commission found: (1) that appropriate use of fiscal-monetary policy could ensure a continuing compatibility between rapid technological progress and high levels of employment; (2) that there was no evidence that the increasing demand for highly educated and skilled workers was generating structural imbalance in the labor market; and (3) that the evolution of private and public policies had done much to reduce the cost of adjustment for individual workers, but there was considerable room for future progress.
If the Commission (which included the distinguished MIT economist, Robert Solow) is correct, then the automation problem is, despite all appearances to the contrary, not really very different from our historical experience of technical change. If, as the Commission’s own researches imply, we are in the 1960’s moving a little more rapidly than in the past and the typical lag between the invention of a new process and its general diffusion is shortening, this measure of acceleration is still far from great enough to swamp the effective responses of conventional economic weapons. The mixture as before—a combination of prudent tax reduction, responsible expansion of public spending, and alert loosening of credit—will maintain or restore full employment. When employment is high, we can depend upon the adaptability of employers and employees alike to smoothe individual transitions from old to new jobs. Where the mechanism falters, an improved program of publicly supported job retraining and labor mobility will see us through.
This “conventional” view is more or less accepted by economists whose political preferences range from Left of Great Society to Right of Milton Friedman. And certainly the steady downward drift (painfully slow and incomplete though it has been) of unemployment during the last six years seems a testimonial both to the failure of automation to cause massive job displacement and the capacity of fiscal policy to maintain a near-full employment level of Gross National Product.
Until recently I have myself found the official position reasonably convincing. What has somewhat shaken me now is Ben Seligman’s persuasive book. Let me say at the outset that this is by far the most impressive study of the impact of automation that I have seen. Seligman provides a useful history of the new computer and factory technology and a lucid sketch, case by case, of how the new devices operate in a considerable variety of blue- and white-collar situations. The peril in generalizing from example is notorious: as the proverb goes, “for instance” is no argument. However, when the instances are themselves important, their number is large, and they occur in variegated activities—in this case as variegated as baking, printing, machine tools, shipbuilding, medical diagnosis, teaching, and legal research—the aphorism loses its power to convince. On the evidence, computer technology can be applied to almost any productive activity. How soon it will be applied is in part a question of cost and in part a question of identifying the areas where human beings still possess a comparative advantage over their electronic rivals.
Seligman’s exposition makes it plain that even if the conventional confidence in our national capacity to maintain full employment is generally valid, there are still enormous difficulties in converting assembly-line workers into computer technicians and hand bookkeepers into programmers, let alone the unskilled into any marketable category. Non-emotive, economists’ labels like flexibility, adaptability, and mobility conceal an uncomfortably large incidence of shattered pride in suddenly worthless skills, assured statuses rudely revised, and community relationships broken. At the very best, this is what the medication of job retraining in a high-employment economy promises. At worst, the medication fails and displaced workers never again find a satisfactory place in the workaday world.
I have nothing but praise to contribute to the pages in this study which discuss the scope and prospective incidence of automation’s impact upon work, skills, and attitudes. I am less happy about the polemical skirmishes into which Seligman enters with such upholders of the “conventional” view as Herbert Simon, Daniel Bell, Charles Silberman, and Peter Drucker. The points at issue are frequently both technical and moot. Certainly this is true of the argument over the adequacy of our productivity measures. My impression, for what it may be worth, is that Seligman scores some shrewd points, particularly against Silberman, but that too much space and polemical energy are wasted upon issues which are beginning to assume only historical significance.
The conventional wisdom, I am inclined to think, does have the better of the argument about the effects to date of automation. More significant is the question of what is likely to happen next. And here I find myself increasingly convinced by the kind of evidence marshaled by Seligman that automation is on the threshhold of creating dramatic social change. In the end I do not differ with Seligman on the issue of the need for social invention. My remaining point of separation from Seligman and other “alarmists” is only this: they think that the storm has already broken. I see the heavy clouds which portend the oncoming storm. Soon, it appears we will all be on the same side.