American labor unions are in trouble. The commitment to "scientific management" that gave them an international reputation for progressiveness a few years ago is now making them obsolete.

From World War II through the end of the 1960s, most Americans -- including most businessmen -- accepted unions as a natural and probably useful part of modern society. Our government imposed the "right" to collective bargaining -- along with democracy and constitutionalism -- on the peoples of occupied Germany and Japan after the war. At home and abroad America's industrial unions were seen as a key to our efficiency largely because they relied on, and thus encouraged, a systematic approach to division of labor -- "scientific management." Under this approach, work was broken down into clearly definable tasks, not necessarily related to a particular skill, and raises and promotions were awarded according to rational, predetermined rules.

But in the 1970s, businessmen -- even foreigners who had come to the United States from prosperous, highly unionized countries -- began to complain that in organized American shops they couldn't structure work efficiently. Unions, they said, hindered their ability to assign workers, design jobs, and schedule production. While much of the business press implied that the only real problem was high union wages, managers worried most about the unions' impact on the production process.

The unions' commitment to scientific management was now contributing to conflicts between unions and management and thus to America's decline in productivity -- especially in America's big companies. Old-fashioned scientific management was obsolete; therefore so was a union structure built on it. The only encouraging result was that the same processes now threatening unions and big businesses are probably opening new opportunities for smaller firms.

Unions' problems are profound: Most contemporary American unions have built themselves on the highly specialized division of labor that scientific management implied. Nineteenth-century craft unions lost to management when they fought division of labor. The Sons of Vulcan, for example, the organization of men who controlled the extraction of pure metal from iron ore, had to merge with several other unions to form the Amalgamated Association of Iron, Steel, and Tin Workers Union. The merged union was crippled by bloody clashes with managrement in an unsuccessful strike in 1892, and finally was eliminated.

In the 1930s, however, new industrywide unions arose, such as the United Steelworkers and the United Automobile Workers. Many unions that had originally been craft-oriented, such as the International Association of Machinists and the International Ladies' Garment Workers' Union, began to organize workers without regard for craft distinctions. The new wave accepted the division of labor and structured its demands accordingly.

To prevent petty despotism in the shop, unions insisted, as the proponents of scientific management had long urged, that work tasks be clearly defined, that the grouping of these tasks into jobs be explicitly delineated, and that rules governing wages and the distribution of jobs among workers be written down. Scientific management techniques such as time-and-motion studies -- careful analyses of how many motions and therefore how much time a job should take -- quickly become standard tools for the resolution of disputes. Unions left management free to organize work as it thought best, so long as the organizing principles were codified and adhered to.

Unfortunately, the elaborate division of labor to which so many American managers and unions had become committed is not always efficient. It appears to have worked best in producing long runs of standardized products. In the 1960s Americans joked about how little seemed to distinguish the "new model" Chevrolet every year from the old model of the previous year. But if new models had been truly new, General Motors could hardly have realized the substantial cost-reducing benefits of scientific management. Only when products don't change too often or too substantively does money invested in single-purpose machines or carefully specified individual work tasks pay off.

When the market calls for short runs of specialized products, when demand is unpredictable, or when a product must change rapidly, old-fashioned craftsmanship -- with each worker ready to perform a variety of tasks -- may be more efficient than "modern" mass-production techniques.

Since the beginning of the 1970s, demand for almost everything has been unstable. The mass markets that American industry dominated in the 1940s, '50s, and '60s have changed dramatically. Even if General Motors could produce a 30-foot-long Impala 25% cheaper than in 1968, the car would be unlikely to sell if it couldn't get 20 miles to the gallon.

The reasons for rapid change in the marketplace are diverse. The uncertainty associated with the oil crisis has contributed to the change. So has the growing sophistication of competitors in several foreign countries. Rising consumer incomes may have increased the share of demand for more specialized goods. Possibly the pace of technological change has increased. Certainly the latest technology, especially computer-assisted manufacturing, permits small-scale batch production and rapid, substantive changes of product design more readily than old-style mechanization did.

Whatever the reasons, the rapidity of change is clear. And foreigners have taken advantage of change to launch serious attacks on Americans' traditional markets. The pressure is forcing sectors of American manufacturing that produced standardized products to move backward, if not to craft production, then at least to flexible job descriptions.

But our "progressive" unions find it hard to go along. Because they have built their grievance procedures on a highly specialized division of labor, any move toward flexibility threatens them. Unions don't know how to protect their members without rigid job descriptions and rules that they can enforce simply.

Obviously the unions' bewilderment can disrupt small firms as well as large ones, but smaller firms have always lacked the personnel to put a different person in charge of every possible task. Even in industries such as garments and machinery, in which unions have long tried to make small firms use techniques like time-and-motion studies, small firms had to learn long ago to handle techniques flexibly. Rule-bending that would have caused shop stewards' whistles to blow and work to stop in large plants has long been routine in small unionized printing shops, for instance.

In fact, flexibility is a specialty of small firms. To the extent that the unions' problems indicate the economy of the 1980s will place a premium on flexibility, the outlook for small business is highly favorable. As people experiment more, whether with semiconductor technology or with wood stoves, there's a much greater chance for small businesses of find a role.

For unions themselves the outlook is much less optimistic, but by no means hopeless.

Strong unions need not demand rigid job descriptions. In countries like Germany, Italy, and Japan, unions have developed very different methods. German unions participate in companies' investment decisions; Japanese unions demand that salary increases strictly follow seniority. Neither approach seems to have hurt companies' efficiency.

Today these foreign unions aggravate American problems because their flexibility increases the competitive pressure on American companies. But perhaps in the long run they may provide ideas that contribute to a new approach for unions here.