First steps toward nuclear arms control; Ex-CIA official explodes myths of 'superiority' and 'limited war'; Russian Roulette: The Superpower Game, by Arthur Macy Cox. New York: Times Books. 248 pp. $14.95.

In the jaunty jargon of nuclear holocaust, ''launch 'em or lose 'em'' is part of the lingo. It is used to describe the awesome expedient of counterattack while the enemy's missiles are still in the air. ''Launch on warning'' is the more technical term for this ruinous possibility. As the time factor involved becomes less and less, such a launching becomes more and more likely.

This is just one of the many warning notes sounded by Arthur Macy Cox in his important book. Cox is a former CIA official who in recent years has made a serious study of weaponry and arms control. Early in the book he sounds his siren blast:

''We and the Soviets are on the brink of developing strategic counterforce capabilities, which means that we are both moving toward launch on warning. When we have arrived at that point, our nuclear weapons systems will be on hair-trigger alert. Both sides will know that the other side has the capability to launch a first strike which could destroy not only missiles but the command and control and communication system of the other side as well.''

To drive this home, Cox tells us some impressive horror stories of computer error and mechanical failure that have almost led to accidental nuclear war. These bring home how near we already are to launch on warning. During a recent eight-month period there were three serious false alerts. Two of them, within days of each other, were caused by the failure of an electronic chip about the size of a dime and costing 46 cents. Because of these failures, the US early-warning system reported what seemed to be a large-scale Soviet attack by both land-based missile and strategic submarine, with the first wave due to reach target areas in 10 minutes!

Fortunately in each case the computer error was discovered before a counterattack by missile, submarine, and bomber was airborne. But the author reminds us that the elapsed time for delivery either way in the key European area can now be as little as six minutes.

In seven hard-hitting chapters Cox builds a case, in which he seeks to demolish two prevalent theories: (1) the need for American nuclear superiority so that Washington can negotiate from strength, and (2) the fallacy that there can ever be a limited nuclear war.

On the question of superiority, he reminds us repeatedly and convincingly that superiority is not the way to make any progress in reducing the clear and increasing dangers of all-out war. The Soviet leaders believe that today they have achieved parity, and it is an achievement they will never relinquish. ''The Soviets are no more likely to negotiate from a position of weakness than we are, '' he writes. ''In order to cut nuclear arsenals, it is essential that both sides believe that there is parity or relative equality.''

On the fallacious theory of the possible limited battlefield use of atomic weapons (an eventuality envisaged by President Reagan and others), Cox hits hard: ''The Soviets have made it abundantly clear that if they ever are attacked with nuclear weapons they will respond with all the nuclear power at their disposal. This is why the recently perfected US theories of limited nuclear war and counterforce weapons make no sense at all. They are based on a false assessment of Soviet policy and intentions.''

In another chapter Cox is highly critical of the Soviet blunder into third-world forays, both direct and indirect.

To bolster and clarify his own strong views, toward the end of his book Cox asks a leading Soviet expert on US affairs to comment on his major tenets. The expert, Georgy Arbatov, is an adviser to President Brezhnev. After some circling and snarling, like two mastiffs, Arbatov and Cox find basic agreements on the ultimate issue of survival. They agree that arms-control talks are the only hope and that progress in such talks can be made only if there is mutual recognition that the arms buildup is a standoff and that there must be some safety valve to third-world adventure.

The last word belongs to Cox, and his conclusion cannot be said too widely or too often:

''The main barrier to strategic arms control has been the American hawks, who oppose the concept of nuclear equality. Until recently the US has maintained nuclear superiority. The American hawks want to restore that superiority, but that is an unrealistic and very dangerous course. The sane approach would be to negotiate a treaty which would reduce and control nuclear weapons systems to a point where neither power threatened the survival of the other, where neither side could gain any conceivable advantage by striking first in a surprise attack.''