An Exchange on Automation

To the Editors:

I am writing this on vacation so I apologize for not having the references in front of me, in particular David Noble’s book, Forces of Production, but I remember well a previous article of his on the machine tool industry. (What I remember most was the depth of detail he provided on the NC [Numerical Control] industry, my own field of specialization.)

The problem with his argument then and, as I gather from James Fallows’ review [NYR, September 27], repeated now in the book, is that it just does not work for the chosen industry: the machine tool industry (or more correctly the Tool & Die industry which supplies the machine tool industry). The argument states that Numerical Control technology has advanced at the expense of other, comparable technologies because it offers management (or “capitalism” in the large as I recall from the previous article) effective, white collar control over the shop floor.

By sheer coincidence, in the same week as the NYRB piece appeared, the journal Manufacturing Engineering, a publication of the Society of Manufacturing Engineers, published a special feature on the Tool & Die industry entitled “How Fares the Journeyman Now?” by R.P. Bergstrom. The point brought home in this article, and well known to anyone in manufacturing, is that the overwhelming majority of companies in this business are “job-shops” and, far from capitalism in the large, these enterprises represent capitalism in the very small—90 percent have fewer than fifty employees, over 50 percent have fewer than ten! (See Dept. of Commerce, Bureau of the Census, Annual Survey of Metal Working Industry, Standard Industrial Code 3544.) These are “Ma and Pa” operations, with very few exceptions, but they employ some 85,000 production workers.

Now the point is this: these are the users of NC technology. The military and aerospace industry may have pioneered it, for their own deep purposes, but the job shop has embraced it. Why? Because NC offers a practical solution to the production of 50 to 2,000 flat or axi-symmetrical parts (say a holding plate with dozens of holes drilled in precise locations). This is the typical size and shape of production for this industry. Larger batches would justify a special machine (“hard” automation) dedicated to the production of that part alone. Smaller batches make skilled, hand work still profitable. But for medium-sized batches, programmable technology seems to be the option of choice today.

About 90 percent of the NC machines in production can be found in job shops (National Machine Tool Builders Association estimate). We might like to get behind the scenes in this little world and reexamine David Noble’s argument, to determine if its introduction is having the effect he describes. Is it used as a ploy to deskill the machinist? An interesting quote appears in the Manufacturing Engineering article cited above, which may be interpreted as supporting this argument: “Take an average guy, and give him the [NC] equipment,” says one job-shop operator. “Then the other guys in the shop will see him using the new technology and they’ll think that they, too, can do it. What you’re doing is creating a little competition out there—which is good.”

Now this can no doubt be read as a diabolical plot to divide the workforce and enlist the workers. But a careful reading will show that the power balance has not really shifted. The man on the shop floor still controls the process—it’s the same “guy”—only now he has more sophisticated tools. I have personally never seen the transition to the new technology work successfully without having it under the control of an experienced machinist. What he controls is not just the metal-removal process, but the conceptualization process that precedes it…the abstract process defining how the metal will be removed to produce the desired shape. It’s the opposite of deskilling (from bee to architect).

Clearly this is not the case for the aerospace industry studied. There the NC programming and metal removal functions have been divorced. But my point again is that these bureaucracies are not typical NC installations: they represent fewer than 10 percent of the installed NC base. On the whole NC is the jobshop’s tool, and that’s where its impact is being felt. And it’s not all that bleak.

B. Kuttner

President, Computer Tool and Die Systems

Ann Arbor, Michigan

James Fallows replies:

As I tried to indicate in my review, Forces of Production is a mixture of the masterful and the overstated. What is impressive is David Noble’s careful history of the introduction of “numerically controlled” machine tools. What I found far less persuasive was his implication that numerical control, and automation in general, almost inevitably oppress and “deskill” the work force.

Mr. Kuttner’s letter is also a mixture. He eloquently explains how numerical control, far from imposing rigidity, can often add flexibility to machine-shop operations. This is especially valuable to job shops, which produce a constantly changing variety of products and therefore need equipment that can adapt to different tasks. But his suggestion that small job shops are the principal sites of numerically controlled machines does not square with any other figures I have seen. Job shops can also exist within large plants—a large General Motors facility, for example, may contain several small job shops—but Mr. Kuttner says he is referring to factories with fewer than fifty employees. The chart below, taken from American Machinist’s 13th Annual Inventory of Metalworking Equipment, in 1983, indicates that plants with more than one hundred employees, which contain 60 percent of all machine tools, also contained 60 percent of numerically controlled equipment (referred to as NC on the chart).

NUMERICALLY CONTROLLED EQUIPMENT, 1983

Given the history of numerical control, anything except this pattern would be a surprise. As David Noble emphasized, the machines were initially so expensive that companies would use them only when the Air Force subsidized the cost. Recently, the price of numerical control, like other computerbased equipment, has been coming down, and smaller shops have been able to enter the market. (The American Machinist survey showed that the smallest establishments are increasing their share of NC machines. As the setting for NC shifts to the smaller shops, Mr. Kuttner’s assessment of their effects may become more representative.