Washington — As the debate continues over a freeze on nuclear arms, a related controversy grows: Should Americans prepare to survive an all-out Soviet attack?
The Reagan administration argues that millions of lives could be saved and the nation more quickly recover in the event of the ''unthinkable,'' if civil defense were expanded.
Critics -- including a number of government intelligence experts -- say just the opposite: that the likelihood of nuclear war would be increased by such preparations, and that in any case the world's nuclear arsenal far overwhelms the good that evacuation plans and bomb shelters could do.
The diversity of official opinion has been illustrated in recent days by congressional action. The House Armed Services Committee voted on Friday to give the President everything he wants in 1983 as part of his seven-year, $4.2 billion plan to build up civil defense. The Senate Armed Services Committee, on the other hand, reduced the President's request for 1983 from $252 million to $ 144 million (still $10 million above the 1982 figure). Even the committee's hawkish chairman, John Tower (R) of Texas, questions administration claims about civil defense.
Woven throughout the budget debate is the public uneasiness about the possibility of nuclear war and what some see as frightening administration comments about ''limited nuclear war'' and (as its 1983 budget message states) being prepared to ''successfully fight either conventional or nuclear war.''
Pentagon officials have been softening their rhetoric, emphasizing the deterrent aspects of civil defense and what they call a modest program of increase. In his press conference last week, the President said, ''Everybody would be a loser in a nuclear war.''
Still, the debate here will focus on the specifics of the administration proposal to increase the federal government's long-range civil defense spending plans by more than 60 percent. These include:
* Preparing evacuation plans for 380 high-risk areas (military targets, industrial centers, and cities over 50,000) -- to host areas in small towns or rural regions at least 20 miles away. Such ''crisis relocation,'' government officials estimate, ''might save -- in a large-scale attack preceded by strategic warning -- up to twice as many Americans as the 40 percent expected to survive under present civil defense.''
* Upgrading communications and warning systems and providing emergency instructions for the public.
* Improving bomb-shelter facilities in host areas and increasing the number of radiological defense officers and fallout-detection equipment.
* Providing matching funds to state and local governments so that they can improve their civil defense capabilities.
A major theme in the administration plan is that a sneak attack would be unlikely, that there would be telltale signs (including the evacuation of Russians from cities).
''It is designed to make nuclear war less likely by improving our ability to deter the Soviet Union from an attack on the United States,'' says Louis Giuffrida, head of the Federal Emergency Management Agency, which oversees civil defense. ''In conjunction with our strategic forces, civil defense will help contribute to a strong deterrent posture. It also reduces the ability of the Soviets to coerce the US with nuclear blackmail.''
Pentagon officials say a stronger civil defense could make it more difficult for the Soviet Union to effectively cripple US industrial capacity. A broad relocation program, they argue, would present the USSR with serious retargeting problems.
Another viewpoint comes from former director of the National Security Agency, retired Adm. Noel Gayler. In recent congressional testimony, he warned that civil defense planning in general ''generates a mind-set toward nuclear war.''
''Evacuation of major cities in any reasonable length of time is impracticable,'' says Admiral Gayler, who also served as deputy director of the Joint Strategic Target Planning Staff. ''It would invite attack during evacuation -- actually increasing casualties. And any evacuation area can itself be targeted if an opponent wishes to do so.''
The congressional Office of Technology Assessment (OTA) has reported that a population relocation program would present a ''staggering logistics problem.''
''Although it is true that effective sheltering and/or evacuation could save lives, it is not clear that a civil defense program based on providing shelters or planning evacuation would necessarily be effective,'' the OTA reported in its 1980 study, ''The Effects of Nuclear War.''
In analyzing the USSR's vast civil defense program, the Central Intelligence Agency estimated that with a week's warning, the Soviets would reduce their casualties significantly, but they would still be in ''the low tens of millions'' in the event of nuclear war.
''The most important factor in effective crisis relocation is public cooperation,'' states recent administration literature on civil defense. But like the subject of a nuclear freeze, officials may find that much public opinion is moving in another direction.
Officials in Boulder, Colo. and Sacramento, Calif., have voted against federal evacuation plans. There have been anti-civil defense actions by officials in Cambridge, Mass., Greensboro, N.C, and Houston. The question will be aired soon in Philadelphia, where religious groups and physicians are expected to protest proposed evacuation plans.