Is nuclear safety a matter of differing perceptions?; The Atom and the Fault, by Richard L. Meehan. Cambridge, Mass.: The MIT Press. 161 pp. $13.95.

Imagine, if you will, a hearing being held in California to determine if a certain site is safe for a nuclear reactor. Power company officials and representatives of antinuclear, community, and other groups wait to testify. Since the crucial question is whether or not an earthquake fault under the proposed site could threaten the reactor, both sides - pro and con - have retained geologists.

But wait a minute - how can geologists be on both sides? Either the fault is active, or it isn't, right? Or as the author puts it: ''Isn't there only one set of 'facts,' one reality? Aren't we scientists and engineers specially trained to perceive that one reality? Or are there several realities out there, each differing, depending on our individual - or is it professional? - background or motives, our personal or collective politics?''

To probe these and other questions, Meehan doesn't need to ''imagine'' a scene like the one described above. He has participated in many such hearings through his work as a consultant on earthquakes and land failures to industry and government. He is also an adjunct professor in the Values, Technology, Science, and Society program at Stanford University. His concern with values is a key element in the book.

Meehan is exploring something far more subtle than the pro- and anti-nuclear-power issues. He is probing the nature of honesty and truth in the evaluation of scientific data.

Besides giving the reader the basic background in earthquake faults and how they relate to reactors, he discusses geological problems that were crucial in the hearings regarding actual and proposed nuclear plant sites, including those at Bodega Head and Diablo Canyon. He also describes the roles of engineers, the law, ordinary citizens, and the government. And he offers in some detail a specific case history: the General Electric Nuclear Center in Vallecitos Valley, Calif.

The nation's first licensed Atomic Energy Commission (AEC) test reactor, it had been operating for 20 years and had a flawless safety record. When its license came up for renewal, however, the Nuclear Regulatory Commission - successor to the AEC - told the company to explore further the geology of the base of the hill where the reactor was located. A new geological map had indicated that a fault might be present.

To do this, two 13-foot-deep trenches two miles apart were dug. The first inspection of the trenches didn't reveal anything unusual, but additional examination opened up a controversy: Did the evidence indicate that a landslide had taken place in the past, or did it reveal an active geological fault? The NRC hearings that followed were an effort to discover the truth. Besides the 525 jobs that hung in the balance, the reactor produced half the free world's medical radioisotopes.

A crucial question that kept coming up was how risk-free should a nuclear reactor be. One hundred percent? Eighty? Fifty? How much loss of life would be tolerable if an earthquake should rupture the reactor containment building? Those on the antinuclear side would probably saythat no risk is acceptable. Many engineerswould prefer to quantify the risk so a formula for evaluating sites could be worked out.

Back in the 1950s, when Congress authorized the civilian development of nuclear power, it didn't include a risk factor in the legislation. After all, the political and military outlook at the time wasn't as conscious of a need for safety standards. Also, the technology was so new that it would have been difficult to set a risk level. Yet this legislative oversight has plagued nuclear reactors ever since.

Meehan observes that in order to operate, the NRC has had to work toward a consensus on the siting of individual plants.

But the psychological and financial cost to those involved in the process has been great. He also points out that we need to understand ''the personal and social benefits of electrical power'' before such standards can be established.

The account of the Vallecitos case is a fascinating picture of democracy in action. But the result of the hearings is bittersweet. After six years of debate and delay, the license was renewed. Unfortunately, by then ''GE had lost its once profitable medical isotope business to the Canadians.''

The final chapter tackles the moral, social, and technical elements of siting nuclear power plants and evaluating geological faults. Among the questions covered are whether faults are important, what role government and ''experts'' should play, what motivates the pro- and the anti-nuclear forces.

Since these questions are being discussed by someone who has ''been there,'' the answers are laced with practical observation while they probe the deeper philosophical and moral issues. If Meehan has an axe to grind, it isn't in evidence. The text seems to be an unusually clear, honest, and open effort to grapple with the questions and to move toward solutions.