THE LABOR DEPARTMENT's proposed strategy for protecting workers against cancer-causing substances is bound to stimulate some major legal and scientific fights. That is one of its purposes. Labor Secretary Ray Marshall and Dr. Eula Bingham, administrator of the Occupational Safety and Health Administration want to get the basic policy issues thrashed out and settled once and for all. This is the only way to extricate OSHA from the bogs of its old case-by-case approach, which involves fighting the same policy battles over and over again. That method has been so tedious that after six years, the agency has managed to set final standards for just 17 of perhaps 2,000 suspect substances. At that rate, hundreds of carcinogens might be identified and curbed, eventually, only in the grimmest way - after workers start getting sick, sometimes decades after their exposure to harm.

The new proposal is doubly welcome because its emphasis on precautionary action is so strong. Its goals is nothing less than to limit workers' exposure to all known and suspected carcinogens to the "lowest feasible level." In advancing such a strict and sweeping rule, the Labor Department has rejected arguments that "strong" and "weak" carcinogens should be controlled differently and that modest degress of exposure to some dangerous substances may be "safe." Given the state of scientific knowledge, the department's wariness impresses us as not only defensible but desirable. It's true that, as a spokesman for E.I. du Pont de Nemours and Company said, "all carcinogens are not alike." But scientists cannot yet make detailed distinctions with much assurance. On the other hand, they do know that minimal exposure to some substances can be lethal, and that as people encounter more toxic chemicals, the long-range health dangers do not merely add up; they multiply.

We see more room for argument about some other aspects of the plan. One involves the grounds on which chemicals should be classified as known or suspect carcinogens. As the furor over saccharin shows, there is still great debate among the experts, and even more confusion among laymen, about how various types of tests on animals should be interpreted and even whether they are relevant to humans. Recent disclosure that some tests have been run improperly have added another layer of uncertainty.

Perhaps an even tougher issue is now "feasible" should be defined when specific worker-protection programs are being devised. If the proposed strategy is adopted, OSHA is eventually going to be ordering lost of industries to make extensive changes in their process and work practices. But just how much is going to be required? For those exposed to toxic substances, this could be a life-or-death question; for the affected industries, it is obviously the "bottom line."

Experience suggests that in some cases, automation of hazardous processes or the substitution of safer chemicals will be easier and less expensive than employers may predict. In other cases the balancing of economic and health factors will be very hard. No matter what, the public - especially affected communities - should be fully informed about alternatives before choices are made. If OSHA can find better ways to do this, its future decisions could have much more public acceptance and support.