Labor and Monopoly Capital The Degradation of Work in the Twentieth Century

Machinery offers to management the opportunity to do by wholly mechanical means that which it had previously attempted to do by organizational and disciplinary means. The fact that many machines may be paced and controlled according to centralized decisions, and that these controls may thus be in the hands of management, removed from the site of production to the office——these technical possibilities are of just as great interest to management as the fact that the machine multiplies the productivity oflabor.\* It is not always necessary, for this purpose, that the machine be a well-developed or sophisticated example of its kind. The moving conveyor, when used for an assembly line, though it is an exceedingly primitive piece of machinery, answers perfectly to the needs of capital in the organization of work which may not be otherwise mechanized. Its pace is in the hands of management, and is determined by a mechanical device the construction of which could hardly be simpler but one which enables management to seize upon the single essential control element of the process.

To explore this subject in somewhat greater detail by means of a very recent instance, let us take as an example the work of the machine shop. This remains the fundamental branch of all industry, not only because of the great role which machine tools play in a great many areas of production, but also because it is in machine shops that the machinery ofindustry is itself fabricated. It is also particularly important because of the recent innovations in machine control which are revolutionizing production methods in machine shops, and which have initiated control systems that are spreading into many quite different industries. It has a further interest for this discussion because in this industry one may see how machinery is utilized to tackle the very machineshop problems with which Taylor wrestled for so many years.

The problem of controlling machine tools presents itself to management largely as a problem of unit or small-batch production. Highly automatic machine systems adapted to mass production or continuous flow processes are of little help in this respect, since they represent huge fixed investments which can return their cost only when applied to a large volume. And it has been estimated that three-fourths of all production in the metal-working industries of the United States takes place in batches of fifty units or less.1º Quantities as small as these must be manufactured on universal or general purpose machine tools, and the tooling, fixtures, and setup costs that may be distributed among these short runs are necessarily limited. Thus this vast area of metal cutting has until recently remained the province of the skilled machinist. Insofar as management has found an answer to the problems of cheapening labor and controlling production, it has taken the form, on the one hand, of breaking down the machinist's craft among machine operators specializing in the lathe, milling machine, and other individual machines, and making machine setup itself a specialty; and, on the other hand, of predetermining operations according to management standards in the Taylor tradition.

The mechanical solution to the problem has taken the form of numerical control,\* which has been called “probably the most significant new development in manufacturing technology since Henry Ford introduced the concept of the moving assembly line."" In application to machine tools alone, this concept is working a revolution in industry, but its applications are spreading beyond machine tools and potentially embrace a great variety of machine and hand operations. It is therefore worth considering in detail, as a prime instance of the managerial use of machinery in the capitalist mode of production, and how this affects the worker and the labor process.

The concept of this form of control over machinery has been traced back to two French inventions: that of Falcon in 1725, a knitting machine controlled by a perforated card, and that of Jacquard in 1804, a knitting and weaving machine controlled in the same way. The principle is like that of the player piano, which is actuated by holes in a paper roll. It was picked up again by an American inventor who in 1916 patented a continuous-path machine for cutting cloth in the garment industry. An application of this method to the control of machine tools was patented in 1930, but the development and application of the concept did not really begin, despite this long history, until after the Second World War. Financing for the research was made available by the United States Air Force and was carried on by the John Parsons Corporation and later by the Massachusetts Institute of Technology, which demonstrated in 1952 a prototype in the form of a vertical milling machine operated under numerical control.12

134 Labor and Monopoly Capital

Machinery

135

\* “One great advantage which we may derive from machinery," wrote Babbage, “is from the check which it affords against the inattention, the idleness, or the dishonesty of human agents."9

\* The term comes from the control (usually by a punched-tape reading device) over the movements of the tool or the work by means of numbers (for instance, 2.375 inches) which represent distances along three axes, and by means of which the tool may be guided to any point in a three-dimensional solid.

The possibility of widespread industrial application of this as well as many other control systems materialized with the electronics revolution of the 1950s and 1960s, which furnished cheap and reliable circuitry for the control instruments. This began with the transistor, which at first simply replaced the vacuum tube on a one-for-one basis. By the early 1960s, however, integrated circuits combined transistors and other components on tiny chips of silicon crystal, so that eventually large-scale integrated circuits offering the functions of hundreds of expensive and bulkier parts were combined on a single chip. As the yield of batch production processes climbed, the cost fell from an average price per circuit function (one transistor) of \$2 in 1965 to under three cents in 1971. The reliability of operation and ease of repair through modular replacement, combined with this cheapening of increasingly complex circuitry, are the basis for the revolution in control technology, and it is here rather than in earlier experiments that the source of this new industrial technique and its broad use must be sought.\*

By 1968, no more than 1 percent of machine tools in industrial use were numerically controlled, but the shape of the future could be seen in the fact that 20 percent of all new machine tools shipped in that year were equipped with this attachment. And in trade showings of machinery, as well as in industry journals, the big majority of machines displayed or advertised are now tools of this type.

With numerical control, the machine process is subjected to the control of a separate unit, which receives instructions from two sources: in numerical form from an external source, and in the form of signals from monitoring devices which check the ongoing process at the point of contact between tool and work. Using this information, the control unit originates signals which activate power drives controlling the work, tool, coolant, etc.

From a technical point of view, the system offers several advantages. Complex metal cutting—for example, the machining of surfaces to compound curves—slow and demanding when the calculations are made in the course of cutting, may be coded with relative ease and cut with assurance; this was one of the features which, because of its applicability to the shaping of dies and other parts used in aircraft production, interested the air force in the method. The coding of any job is quickly completed when separated from machine execution, and once coded a job need never be analyzed again: the tape may be kept on file and used whenever a remake is called for. The processes of metal cutting are virtually automatic, relieving the worker of the need for close control of the machine while cutting is in progress. The separation of conceptualization and calculation from the machine means that the tool itself is in more constant use for metal cutting; at the same time, it goes through its continuous cutting path without interruption, which also makes for more efficient use of these expensive pieces of equipment.

The unity of this process in the hands of the skilled machinist is perfectly feasible, and indeed has much to recommend it, since the knowledge of metal-cutting practices which is required for programming is already mastered by the machinist. Thus there is no question that from a practical standpoint there is nothing to prevent the machining process under numerical control from remaining the province of the total craftsman. That this almost never happens is due, of course, to the opportunities the process offers for the destruction of craft and the cheapening of the resulting pieces of labor into which it is broken Thus, as the process takes shape in the minds of engineers, the labor configuration to operate it takes shape simultaneously in the minds of its designers, and in part shapes the design itself. The equipment is made to be operated; operating costs involve, apart from the cost of the machine itself, the hourly cost of labor, and this is part of the calculation involved in machine design. The design which will enable the operation to be broken down among cheaper operators is the design which is sought by management and engineers who have so internalized this value that it appears to them to have the force of natural law or scientific necessity.\*

Labor and Monopoly Capital

Machinery

137

\* That engineers think in this fashion, or are guided in this direction by all the circumstances of their work, will not appear strange to anyone with the slightest familiarity with engineering as it has developed from its nineteenth-century beginnings. “The monogram of our national initials, which is the symbol for our monetary unit, the dollar, is almost as frequently conjoined to the figures of an engineer's calculations as are the symbols indicating feet, minutes, pounds, or gallons," said Henry R. Towne, the industrialist and pioneer of shop management, in a paper read to the American Society of Mechanical Engineers in 1886."The dollar," he said on a later occasion, "is the final term in almost every equation which arises in the practice of engineering. . . ."14 Or, in the words of a chemist in more recent years: “I'm no longer really interested in problems that don't involve economic considerations. I’ve come to see economics as another variable to be dealt with in studying a reactionthere's pressure, there's temperature, and there's the dollar."15

\* On this, see Business Week's recent special issue on productivity, which characterizes the goals of increased productivity through solid-state electronic controls as "fewer parts, less-skilled, low-cost labor, and fewer manufacturing steps."13

Numerical control is thus used to divide the process among separate operatives, each representing far less in terms of training, abilities, and hourly labor costs than does the competent machinist. Here we see once more the Babbage principle, but now in a setting of technical revolution. The process has become more complex, but this is lost to the workers, who do not rise with the process but sink beneath it. Each of these workers is required to know and understand not more than did the single worker of before, but much less. The skilled machinist is, by this innovation, deliberately rendered as obsolete as the glassblower or Morse code telegrapher, and as a rule is replaced by three sorts of operatives.

First, there is the parts programmer. The process of taking the specifications of an engineering drawing and recording them on a planning sheet is essentially the same work as that previously done by the machinist when, drawing in hand, he approached a given job. But the parts programmer is not required to know everything else the machinist knew: that is to say, the actual craft of cutting metals in its execution on the machine. He learns instead merely the shadow of the process in tabulated and standardized form, and with that his learning ceases. He is taught to work this information up in a manner suitable for coding.

“The planner," in the words of one description, “simulates the machining done in the shop. . . . He goes through every step in very much detail, leaving no decision to be made later at the machine. The planner determines the feeds and speeds, the cutters required, and even the miscellaneous functions such as when coolant is to be on or off. Importantly, he determines the feed rates and the depths of cut to be taken."l6 To convert the specifications on the drawing into a planning sheet requires, in the case of much of the equipment used for numerical control, only a knowledge of blueprint reading such as the craftsman acquires in his first months of apprenticeship, plus the basic arithmetic of adding and subtracting, plus the use of standard data on machine capabilities. This may be made more or less difficult by the type of equipment used and the complexity of the job in hand. In a recent development, the machining specifications are themselves stored on computer tape and the programmer need only make a description of the part (both rough form and finished form) by simply converting the engineering drawing into a listing of dimensions, using nothing but simple machine-shop terms. The computer delivers a machine control tape, a printout of what is on the tape, a tool list, and the computed job cycle time. This system, it is claimed, reduces program time for a part that once required four or five hours to only twenty or thirty minutes. One company advertises another programming system that uses a“10-word program vocabulary" and says: "“Within a week a shop man can program effectively."17 An article in Monthly Labor Review says: "Most of the functions of the skilled machinist have been shifted to the parts programmer. Consequently, skilled machinists often staff these positions."18 This may have been so at the outset, but the job of parts programmer now increasingly falls to the technical college graduate (often from the two-year junior college) who fits the “labor profile" of this desk job more closely than the machinist, especially in being cheaper.

The next job is that of converting the planning sheet into machine-readable form—usually a paper tape punched on a simple coding machine. Here the candidate is immediately selected by unanimous choice: the “girl" machine operator who learns her job in a few days, attains optimum efficiency in a few weeks or months, and is drawn from a large pool at hardly more than half the pay of the machinist.

So far as the machine operator is concerned, it is now possible to remove from his area of competence whatever skills still remain after three-quarters of a century of “rationalization." He is now definitively relieved of all the decisions, judgment, and knowledge which Taylor attempted to abstract from him by organizational means. The true “instruction card"—Lillian Gilbreth's “self-producer of a predetermined product"—is at last fully revealed in the program tape. “Numerically controlled machines," we are told by one authority, “are fundamentally easier to operate. The skills required of an operator are less than with conventional machines, where he must often be a trained machinist. With numerical control equipment, the operator must, of course, know his equipment. He must have the training and intelligence required to perform several rather straightforward prescribed routines, but he does not possess the technical skills of the experienced machinist. The intelligence corresponding to these latter skills is on the tape in numerical control."

The difference between the trained machinist—even one so limited in craftsmanship that he can operate only a single machine—and the operator of a numerically controlled machine is often understated, both by managers, who are prepared (in public at least) to conceal the downgrading in the interest of a smoother transition and for public relations reasons, and also unions, for whom the exaggeration of the “demanding nature” and “increased responsibility" of the new process is a routine part of contract bargaining. But some idea of the manner in which managers view this difference can be seen in the following response to a University of Michigan survey of companies using numerical control: "Cost of developing and training an operator to produce identical parts by conventional methods and machines compared with NC machining system is approximately 12 to 1.20 This would mean that if it takes four years to give a machinist his basic training, an operator of the sort required by numerically controlled machine tools may be trained in four months. Experience bears this out.\*

138 Labor and Monopoly Capital

Machinery

139

\* This is not to say that, in unionized situations, the pay of machinists is immediately reduced to operator levels the moment numerical control is introduced. In some exceptional instances, where very few numerically controlled machine tools have

140 Labor and Monopoly Capital

Control in this form, and in other forms that are developing out of it, is by no means limited to metal-cutting machinery; the principle has general applicability in many crafts. In this connection, the evolution of boiler shop and other heavy plate construction work is of interest. In these trades the best-paid craft has long been that of layout. The layout man takes off the blueprint the specifications of each part and inscribes them on the plate stock, along with directions for flame-cutting or shearing, punching or drilling, bending or rolling, etc. It was at one point noted that the layout man spent a certain amount of his time simply marking his layout with close-spaced center-punched marks. In those shops that had enough work to warrant a subdivision of tasks in the layout department, this task was taken from the layout mechanic and vested in “markers," at a far lower pay rate. Then, in the 1950s, a method was devised for drawing each piece, in the drafting room, to accurate scale on a transparency which could be projected on the steel from a slide projector mounted high above the layout trestles. Now the layout man became nothing more than a marker himself; after he had adjusted the focus so that a single dimension was correct, he had no more to do than mark. But with numerical control, the steel can go straight to the flame-cutting tables, where the cutting torch is guided by the control tape, so that not only is the work of layout eliminated—or rather transferred to the office—but so is the so-called semi-skill of the flame cutter.

In the same way, in sheetmetal shops numerically controllea macmues are now used for cutting required shapes (with “nibbling" tools, multiple-tool turrets, etc.) without the necessity of shop layout or skilled sheetmetal workers. One furniture manufacturer makes kitchen cabinets and bedroom furniture out of vinyl-clad particle board by numerical control, using a so-called mitre-fold process. The boards are cut in such a way that they fold to make complete units and are held together by the “wood-grained vinyl covering"; folded and glued, each “unit" is complete. It takes just one hour of labor to assemble each piece of this furniture, one-third the time taken by woodworkers (our source leaves to the imagination of the reader the quality and appearance of these surpassingly modern products).21 And it remains only to be noted that the process of making an engineering drawing has itself been found to be readily susceptible to the same attack, so that there now exist drafting machines which draw plans from tape under numerical control. In each of these cases, the public unveiling of the new devices is accompanied by much self-congratulation and by philanthropic phrases about the lightening of the toil of the worker, the ease with which laborious tasks are accomplished, and so forth. Few write so plainly about the way in which functions are distributed, and the effect this is having upon the world of work, as did Thilliez, the engineer who introduced numerical control into the Renault plants, in his 1967 technical book, La Commande numérique des machines:

But in addition the technique of numerical control implies an effect which might be called extraordinary, on the level of the philosophy of the organization of the enterprise. It separates the intellectual work from the work of execution, just as has for a long time been the case with the fabrication of long runs on special purpose machines, and this separation allows the execution of both functions under the technical conditions best adapted to a superior organization, thus in the final accounting most profitable.22

Such a separation of “intellectual work from the work of execution" is indeed a “technical condition" best adapted to a hierarchical organization, best adapted to control of both the hand and the brain worker, best adapted to profitability, best adapted to everything but the needs of the people. These needs, however, are, in the word of the economists, “externalities," a notion that is absolutely incomprehensible from the human point of view, but from the capitalist point of view is perfectly clear and precise, since it simply means external to the balance sheet.

While the forms of utilization of machinery—the manner in which labor is organized and deployed around itare dictated by the tendencies of the capitalist mode of production, the drive to mechanize is itself dictated by the effort to increase the productivity of labor. But the increasing productivity of labor is neither sought nor utilized by capitalism from the point of view of the satisfaction of human needs. Rather, powered by the needs of the capital accumulation process, it becomes a frenzied drive which approaches the level of a generalized social insanity. Never is any level of productivity regarded as sufficient. In the automobile industry, a constantly diminishing number of workers produces, decade by decade, a growing number of increasingly degraded products which, as they are placed upon the streets and highways, poison and disrupt the entire social atmosphere—while at the same time the cities where motor vehicles are produced become centers of degraded labor on the one hand and permanent unemployment on the other. It is a measure of the manner in which capitalist standards have diverged from human standards that this situation is seen as representing a high degree of “economic efficiency."

Machinery

141

been brought into a shop, the union has been able successfully to insist that the entire job, including programming and coding, be handled by the machinist. In many other cases, the pay scale of the machinist has been maintained or even increased by the union after the introduction of nurerical control, even though he has become no more than an operator. But such pay maintenance is bound to have a temporary character, and is really an agreement, whether formal or not, to “red circle" these jobs, as this is known in negotiating language; that is, to safeguard the pay of the incumbents. Management is thus sometimes forced to be content to wait until the historical process of devaluation of the worker's skill takes effect over the long run, and the relative pay scale falls to its expected level, since the only alternative to such patience is, in many cases, a bitter battle with the union.

The most advanced methods of science and rational calculation in the hands of a social system that is at odds with human needs produce nothing but irrationality; the more advanced the science and the more rational the calculations, the more swiftly and calamitously is this irrationality engendered. Like Captain Ahab, the capitalist can say, “All my means are sane, my motives and object mad."

The drive for increased productivity inheres in each capitalist firm by virtue of its purpose as an organization for the expansion of capital; it is moreover enforced upon laggards by the threats of national and international competition. In this setting, the development of technology takes the form of a headlong rush in which social effects are largely disregarded, priorities are set only by the criteria of profitability, and the equitable spread, reasonable assimilation, and selective appropriation of the fruits of science, considered from the social point of view, remain the visions of helpless idealists.\* Each advance in productivity shrinks the number of truly productive workers, enlarges the number of workers who are available to be utilized in the struggles between corporations over the distribution of the surplus, expands the use of labor in wasteful employment or no employment at all, and gives to all society the form of an inverted pyramid resting upon an ever narrower base of useful labor. Yet no matter how rapidly productivity may grow, no matter how miraculous the contributions of science to this development, no satisfactory level can ever be attained. Thus, a century after the beginning of the scientifictechnical revolution and almost two centuries after the Industrial Revolution, the problem for capitalism which towers over all others, and which takes the form of a crisis threatening survival itself, remains: more productivity. In Business Week we read: "Five years of inflation, recession, and uncertain recovery have forced the men who manage U.S. business and the men who make U.S. economic policy to a painful conclusion: Somehow the nation must make a quantum jump in efficiency. It must get more output from its men and machines."24 The very “efficiency” which produced the crises is here seen as the only answer to it. The machine which, working at top speed, threatens to fly apart is to be preserved from that threat by running it even faster. Each capitalist nation will further degrade its own working population and social life in an attempt to save a social system which, like the very planets in their orbits, will fall to its destruction if it slows in its velocity. Here we have the reductio ad absurdum of capitalist efficiency, and the expression in concrete terms of the insoluble contradiction that exists between the development of the means of production and the social relations of production that characterize capitalism.

In pursuit of this “solution," industry, trade, and offices rationalize, mechanize, innovate, and revolutionize the labor process to a truly astonishing degree. The methods used are as various as the resources of science itself. And since these resources are so vast, where they cannot accomplish a large saving of labor by a revolution in production they achieve the same effect by a degradation of the product.

The construction industry, for example, divides its efforts between the destruction of sound buildings and their replacement with shoddy structures whose total life span will not equal the useful life remaining to the demolished buildings. This industry, which because of the nature of its processes is still largely in the era of hand craftsmanship supplemented by powered hand tools, the lowest level of mechanization, makes continual and determined efforts to climb out of this disadvantageous position. It favors new materials, especially plastics, painting and plastering with spray guns (a single spray plasterer keeps a number of workers busy smoothing), and the pre-assembly of as many elements as possible on a factory basis (a carpenter can install six to ten prefabricated door assemblies, pre-hung in the frames with hardware already in place, in the time it takes to hang a single door by conventional methods; and in the process becomes a doorhanger and ceases to be a carpenter). The trend of dwelling construction is best exemplified by the rapidly growing “mobile home" segment of the industry. The “mobile home" is a mass-produced factory product; of the three parties involved—the workers, the manufacturers, and the residents—only the middle one has any advantage to show from the transaction. Yet mobile homes are spreading over the landscape triumphantly, and one may easily predict for them a still greater future because of the high degree of “efficiency" with which they allocate labor and capital.

A quarter-century ago, Siegfried Giedion described the transformation of the crusty, wholesome loaf of bread into a “product" with the “resiliency of a rubber sponge."25 But the production process for the manufacture of this bread is a triumph of the factory arts. Continuous mixing, reduction of brew fermentation time, dough which is metered, extruded, divided, and panned to the accuracy of a gram in the pound, conveyorized baking and automatic depanning, cooling, slicing, wrapping, and labelling have effectively rid the bakery of the troublesome and unprofitable arts of the baker, and have replaced the baker himself with engineers on the one hand and factory operatives on the other. The speed with which the operation is conducted is a marvel of efficiency, and, apart from its effects on the worker, if only it were not necessary for the people to consume the “product" the whole thing could be considered a resounding success.

142 Labor and Monopoly Capital

Machinery

143

\* The occasional dream that flickers in the minds of those who have assimilated the capitalist way of looking at all problems is quickly extinguished by “practical" considerations. Thus Alfred Marshall, early in the present epoch: “In fact, if all the world were a single people, with one purpose and that the highest, it might be well to put some check on this rapid supersession of human skill; even at the expense of delaying the increase of material comforts and luxuries. But Britain can exist only by obtaining her necessary supplies of food and raw products in return for the exportation of manufactures: and her hold on external markets can be maintained only by her use of the most effective processes known."23 The impulse to view society from a human standpoint, “a single people, with one purpose and that the highest," does Marshall credit, but he can entertain it no longer than is necessary to dismiss it from the standpoint of British capitalism, whose single impulse, and that the lowest, requires first of all the sacrifice of the British people.

Furniture production is being remade in the image of the automobile industry. It has increasingly become a mass-production process in which the skills and effects of woodworking and cabinetmaking are disappearing. Shaping is done on automatic contour-profilers operated by unskilled labor. Routers for grooving and cutouts are done by template-controlled machines with a cam and sensing system of programming. “Unskilled operators need only to feed material to the machine; in some applications, even feeding is automatic."26 “The use of pneumatic power clamps and assembly machines is speeding the assembly of frames, case ends, drawers, and chairs, while requiring fewer and less-skilled workers. One machine takes parts directly from a tenoner, feeds metal parts from hoppers, inserts these parts, drives pins or nails, and ejects a completed shelving onto a conveyor at a rate of 7 to 10 per minute, and only one operative is required to load the hoppers and pinners. Another machine takes panels, aligns and joins them perfectly square, drives staples to hold the assembly true while the glue dries, and permits the assembly of a kitchen cabinet by one man every 60 seconds."27 Painting has been mechanized, using automatic spray and flow-coating techniques. Upholstery work is now done with precut, preformed, and stretch materials, thus putting an end to the traditional skills of that trade.

In meatpacking, the industry which was the first in the United States to introduce conveyor lines, the on-the-rail dressing system has displaced the older conveyor. “In rail systems, stunned cattle are hoisted to a high conveyor rail, on which they are slaughtered and then moved through all dressing operations to the chill cooler. Workers, stationed on mechanized platforms which move vertically and horizontally according to the requirements of each task, use power knives and saws. Mechanical hide strippers, which grasp and peel the hide from the carcass, substantially reduce the skilled hand-cutting operations once necessary to remove a high-quality hide without damage. Labor savings per unit may be between 25 and 60 percent on the kill line. These savings are in reduction of waiting time between performance of individual tasks, which are now machine-paced and synchronized, and elimination of constant repositioning of the carcass necessary in the older bed’ system."28 There is a machine which permits an image of each carcass to be projected on a screen so that workers can be guided in making major cuts by a pointer which indicates them; far fewer workers need to know how and where to cut a carcass. Machines emulsify, form, smoke, cook, chill, peel, wash, and package frankfurters. Electronic machines that weigh and package poultry along with mechanical pluckers process up to 9,000 chickens an hour, although, we are told, the prevalence of "low wage rates in poultry packing plants" is among the factors which “tend to retard technological change" in that branch of the industry.29

In the manufacture of wearing apparel, every aspect of the production process is being energetically attacked. Since this is an industry which is characterized by the existence of many shops, most of them relatively small, a great many are still in the stage of traditional “rationalization," breaking down operations into a large number of smaller and simpler steps. At the same time these steps are being speeded up by the introduction of a variety of devices, chiefly attachments to sewing machines such as needle positioners, automatic thread cutters, pleaters, and hemmers. The use of two- or three-layer bonded materials, which eliminate separate linings, and synthetic fabrics, which may be processed by novel methods such as the electronic fusing of seams in place of sewing, opens up new vistas for cheapening and transforming mass-produced clothing. Advanced production methods are copied from sheetmetal and boiler-shop techniques: die-cutting to replace hand cutting, patterngrading equipment which produces different size copies of a master pattern, etc. There is a photoline tracer which guides a sewing head along the path of a pattern placed in a control unit. Improving on this, a photoelectric control is used to guide a sewing head along the edge of the fabric. In these latter innovations we see the manner in which science and technology apply similar principles to dissimilar processes, since the same control principles may be applied to complex contours, whether on steel or cloth.

In typesetting, the printing industry first took the path of eliminating the linotyper's skill through the use of tape-controlled linecasting, the essence of which was the separation of the keyboard from the casting of lead slugs. The operator prepares a control tape on a machine much more rapid and simpler to operate than the linotype. But the use of photocomposition in tandem with the electronic computer has enabled the industry to begin the elimination of metal type altogether, and along with it the need for the operator to justify lines and hyphenate words, since these functions can be performed by a computer which utilizes a record of the syllabification of almost every word in the language. That this rids the typesetter of one more load of useless knowledge has already been amply pointed out, but no one has yet pointed to the knowledge acquired in its place.

Despite the variety of means used in all the innovations we have been describing, their unifying feature is the same as that which we noted at the outset of this discussion: the progressive elimination of the control functions of the worker, insofar as possible, and their transfer to a device which is controlled, again insofar as possible, by management from outside the direct process. It is this which dominates the new place of the worker in production processes, and it is this above all which is slighted or entirely neglected in conventional assessments. The knowledge of labor and production processes which the outsider may gain from a study of nontechnical sources is limited by the vague and imprecise impressions which pass for information, and from which sociologists and journalists, already too eager to achieve optimistic conclusions, glean their notions about the trend of labor in modern society. Morris A. Horowitz and Irwin L. Herrenstadt did an exhaustive survey of such materials in the course of their attempt to study changes in skill requirements of various occupations, and they report the following:

Labor and Monopoly Capital

Machinery

145

Limiting our focus to the past 15 years, we examined various bibliographies and indexes in an effort to pull together the literature relevant to our field of interest. Over 500 bibliographic titles were selected for careful scrutiny and analysis. The overwhelming majority of these speculated about the effects of automation based upon general impressions, discussions with a few industrialists or union leaders, or a few very limited case studies conducted by others. If an article or book discussed automation and manpower, it typically referred to employment opportunities resulting from technological change, the effects on the occupational structure of the plant or industry, or the effects on the skill composition of the labor force. Only a small number of studies made any effort to investigate and analyze the effects of automation on job content and the worker characteristics required of the changing jobs.\*

This very paucity of systematic information and analysis makes all the more important James R. Bright's unique study, referred to earlier in this chapter. In 1954, research was begun at the Harvard Business School on the "managerial implications" of “automation." In 1958, Bright published a volume entitled Automation and Management, which begins with a survey of the evolution of mechanization in manufacturing (with special attention to the electric lamp and shoe industries) and then analyzes in great detail thirteen of the most advanced production systems in operation at the time of the study. These included the Ford Cleveland Engine Plant; a highly automated bread bakery; a small integrated oil refinery with the reputation of being an outstanding example of automatic control; a new automatic production line for the manufacture of oil seals; the foam rubber mattress department of a rubber company, in which the entire process, from raw material stores to finished product warehouse had been integrated into a single radically new system; a chemical plant making commercial fertilizer; a feed and grain plant with a high degree of automaticity; a small coal mine which attempted extensive use of automatic equipment; a plating plant in which a complex production sequence was subjected to a highly automatic handling system; an instrument manufacturer with a unique work-feeding system; an electrical parts manufacturer who applied automatic methods to assembly operations; another plating plant with a differing system of work organization; and a V-8 engine plant. Bright later wrote several articles (see, for example, the July-August 1958 issue ofHarvard Business Review) and, most important, a summary of his conclusions with regard to skill written in 1966 for the National Commission on Technology, Automation, and Economic Progress.

While the Bright studies dealt in general with the “management" aspects of automation, the principal focus was the “skill requirements" of increasingly mechanized industries. It must be pointed out that Bright nowhere indicates a concern with this aspect of his subject from the point of view of the worker, but views the problem entirely from the management standpoint. His approach is detached and rigidly factual, and his concern is expressed in his final conclusion: “I suggest that excessive educational and skill specification is a serious mistake and potential hazard to our economic and social system. We will hurt individuals, raise labor costs improperly, create disillusion and resentment, and destroy valid job standards by setting standards that are not truly needed for a given task. . . ."3i

In the preface to his book, Bright notes: “A controversial area of this study will lie, no doubt, in my conclusions regarding the skill required of the work force in the automated plant. The relationship of skill requirements to the degree of automaticity as a declining rather than increasing ratio is not commonly accepted, or even considered." Nevertheless, after exploring his tentative conclusions with three to four hundred industrialists, and in presenting his findings to “at least a dozen industrial audiences totaling perhaps three thousand persons," he notes that “in general, these conclusions have not been strongly challenged" except with regard to plant maintenance skills, and even these challenges he attributes to “intense personal experiences" peculiar to special situations.32

146 Labor and Monopoly Capital

Machinery

147

\* Horowitz and Herrenstadt are to be thanked for this survey, but unfortunately the remainder of their article is just as useless as the literature they criticize. It is an attempt to assess “worker characteristics required of the changing jobs" entirely on the basis of the descriptions of “job content" in the second and third editions, 1949 and 1965 respectively, of the Dictionary ofOccupational Titles published by the Department of Labor. A more arid and unrewarding exercise can hardly be imagined, and the result is that after scores of pages of meticulous tabulation and statistics, the authors conclude that “the overall or net change in the skill requirements" during these fifteen years was "remarkably small"; that the "small net change" was "the product of numerous offsetting changes"; and that the result "on balance" is “either inconsequential or inconclusive with respect to overall skill levels."30

Bright's work is not only informative but especially useful for the analytical framework it provides, since he sets up a “mechanization profile" of seventeen levels (see Bright's chart of mechanization levels, p. 149). Apart from the first twowork with the hand and with a hand tooleach level deals with a specific machine function and its operating characteristics. With this "profile," Bright was able to chart the entire series of operations in every production system he studied, thus affording a far more realistic glimpse of so-called automatic production systems than that furnished by the glowing self-tributes of managers or the breathless prose of journalists.

On mechanization levels 1 to 4 Bright concludes that since control is entirely up to the worker, skill is increasing (see Bright's chart, "Changing Contribution Required of Operators," pp. 151-52). On levels 5 to 8, where control is mechanical but still dependent upon the worker, some skills are increasing but a number have turned downward, resulting, in Bright's opinion, in an overall decrease in total skill required. In levels 9 to 11, where the machine has been put under external control at least to the extent of signalling its own needs, most skills turn downward. And finally, in the top six levels, which are characterized by self-modifications of machine action and therefore correspond to advanced methods of automatic production, every indicator of skill used by Bright, from knowledge and experience through decision-making, plunges downward sharply, and the indicators of "Worker contribution" all read either "Decreasing-nil," or flatly, “Nil" (with vague exceptions only for “Responsibility” and “Education").33 The result is summarized by a curve which Bright calls the "Hump in Skill Requirements."(See Bright's curve of “Skill versus Automation," p. 154.)³⁴ It describes a “suggested average experience as mechanization increases," and shows an increase only through the first four levels, a decrease thereafter, and a plunge into the nether regions with the installation of those elements of mechanization which are associated with the popular term “automation." He outlines the idea as follows:

Consider a metalworker. Using hand tools, such as a file, he requires considerable dexterity. As power is added but the tool guidance is left in the operator's hands, he needs new levels of dexterity and decision-making to control the machine action, and these grow in importance, A high degree of attention is required. Knowledge requirement, hence training and/or experience requirement, grows with the introduction of the power tool, for he must know how to adjust and direct the more complex machine of Level 4. He must become a "machinist."

When the mechanically controlled machines of levels 5 and 6 are encountered, job knowledge may not be reduced but attention, decision-making, and machine control requirements are partially or largely reduced. In many instances, the technical knowledge requirement of machine functioning and adjustment is reduced tremendously. This is why “machine operators," rather

Levels of Mechanization and Their Relationship to Power and Control Sources (as charted by James R. Bright)

148

Labor and Monopoly Capital

Machinery

149

than machinists are so frequently quite adequate. The job becomes more nearly one of simple machine actuation, workfeeding, patrolling, and inspecting. In moving to higher levels of mechanization where the machines supply control signals there is a further reduction in the attention-judgment-decisionaction activity demanded of the worker. Of course, this may be offset by increasing technical complexity of the equipment and its adjustment, which requires additional knowledge by the worker. Yet the reverse seems more common.

When the variable control levels (Levels 11-17) are reached, we find that the worker contributes little or no physical or mental effort to the production activity. More of the functions are mechanized. The inspecting devices feed corrective information into the machine and thus relieve the operator of mental effort, decision-making, judgment, and even the need to adjust the machine. By its very definition the truly automatic machine needs no human assistance for its normal functioning. "Patrolling" becomes the main human contribution. The “operator," if he is still there, becomes a sort of watchman, a monitor, a helper. We might think of him as a liaison man between machine and operating 35 management.

That this conclusion is not simply the result of an abstract schematization of the problem, but corresponds to real conditions, is made clear by Bright through numerous examples: "During the several years that I spent in field research on managerial problems in so-called automated plants and in exploring automation with industrialists, government personnel, social scientists, and other researchers, I was startled to find that the upgrading effect had not occurred to anywhere near the extent that is often assumed. On the contrary, there was more evidence that automation had reduced the skill requirements of the operating work force, and occasionally of the entire factory force, including the maintenance organization. 536

Bright's reference to the maintenance organization reflects a considerable study he made of maintenance and repair work in the plants he analyzed. He found that the effect of increased mechanization, particularly in its more developed stages, upon the need for skilled maintenance mechanics is not so simple as is usually supposed. On the one side, it is certainly true that the mechanization of a larger span of the production process, the novelty of the equipment used, the electronic circuitry and electro-hydraulic-pneumatic actuating mechanisms, and other such factors tend to increase the need for maintenance and for new skills. But on the other side, he points to a great many factors which tend in the other direction. For example: “One of the effects of automation is to compress the production line and literally to reduce the total physical amount of machinery for a given output, eveņ though that machinery may be more complex. Hence, in several instances the maintenance force was reduced simply because the total volume of machinery was reduced. This reduction more than compensated for the increase in the complexity of the equipment."37 This tendency is reinforced by the mechanization of trouble prevention itself with monitoring devices that anticipate difficulties, by the simplification and standardization of control mechanisms, and so on.

At the same time, the novelty of the control equipment affects only small parts of the maintenance staff. “I found no evidence," says Bright, "that tinsmiths, pipefitters, welders, and carpenters required increased skill, some evidence that hydraulic and pneumatic repairmen need better training because of the increased complexity of the control circuitry, and much evidence that a significant proportion of electricians need extensive additional training." But even in the last instance, Bright points out, the need for electronic maintenance skills requires retraining a very limited number of mechanics. In one plant, in a maintenance crew of seven hundred, eighty were electricians, and the plant engineer found that he needed only three or four competent electronic repairmen per shift. “In other words, only about 10 percent of his electricians needed specialized skilland these amounted to only 1 percent of his total maintenance force."38Overall, while Bright found a number of plants that increased their maintenance staffs substantially, he also found contrasting instances, such as the following:

The most automatic small refinery in the U.S. in 1954 had a maintenance force amounting to 21 percent of the total work force. Conventional refineries show a 50 to 60 percent ratio.

Two major parts-manufacturing plants, each employing over 10,000, have devoted their attention to automatic production since 1946. Both are well known in engineering circles for outstanding automation accomplishments and use literally hundreds of highly automatic machines. Both maintenance forces are characterized by one peculiarity—lack of change. The maintenance force has remained a steady 3.5 to 5 percent of the work force in one firm, and 6 to 8 percent in the other, over the last dozen years of aggressive mechanization with automatic machinery. 39\*

The entire evolution is marked by the very same design characteristic that the consumer sees in home appliances or automobiles: the modular construction of equipment for easy replacement of entire assemblies. While the consumer finds it expensive to buy an entire new assembly in order to replace a part worth a few cents, and also finds the consequent deterioration of repair skills among servicemen exasperating, in industry, where the length of time the production system is shut down for repairs is the most important and expensive factor, replacing entire assemblies is by far the cheapest way. But this tendency further reduces the number of mechanics who are able to do anything but replace the entire module after the source of the malfunction has been locatedand this is something advanced electronic machinery increasingly does for itself. Moreover, even the work of the repair mechanic is now being studied and standardized in much the same fashion as that of the production worker. One such system is called Universal Maintenance Standards:

151

150

Labor and Monopoly Capital

Machinery

\* John I. Snyder, Jr., president and chairman of U.S. Industries, Inc., manufacturers of automatic machinery and controls, writes: “Another myth is that automation will create jobs, not only in the running of machines, but in the building and maintenance of them. Of course this is true to a degree, but not nearly to the degree that some would have us believe. Experience has shown that after the initial ‘debugging’ of automated machines, they require relatively very little maintenance." 40

152 Labor and Monopoly Capital

Machinery

153