THE ELEMENTS FOR a serious compromise on nuclear arms control are now on the table in Geneva. But today they are just pieces of broader, separate proposals put forward over the past several months by both the Soviet Union and the United States in the strategic and European missile negotiations.

Only by mixing some parts of each country's proposals can an arguably fair, meaningful, single, new treaty be cobbled together governing the most dangerous nuclear weapons. In truth, there is only one direction to go -- combining the two negotiations now underway: those on strategic weapons, termed START, and those on medium-range weapons, called INF.

In simplified terms, a framework for agreement could work like this:

In the strategic talks, the Soviets have proposed a numerical limit on intercontinental missile launchers of 1,450. (One of arms control's great ironies is that the Soviet START position today is almost exactly what the first Jimmy Carter proposal to Moscow was back in March 1977. That effort at deep cuts was quickly stopped, turned down cold by Soviet Foreign Minister Andrei Gromyko.) Now, the Reagan START plan calls for something like 1,200 launchers, a number very near the Soviet's. So let's agree on something in between.

However, add to it the present Reagan proposal limiting each side to 5,000 warheads on those missiles. Remember, Soviet party boss Yuri Andropov proposed counting warheads in the Euromissile talks, so why not carry that approach over to the strategic arena?

Thus you would get a launcher limit of, say, 1,350, with a total warhead limit of 5,000. Reagan sublimit proposals, designed to force the Soviets to dismantle their large, land- based ICBMs, would be dropped. Each country would retain the freedom to choose what numbers of each type of land and sea-based missiles they wanted to deploy and the warheads mix on land and at sea.

The counting rules of the 1979 Strategic Arms Limitation Treaty (SALT II) would be observed -- i.e. a missile tested with 10 warheads would be counted as deployed with 10 warheads, even though some versions may have as few as one. No new missile could have more than 10 warheads. A small warhead would count the same as a large one, because with present and future development in accuracy, the explosive power of warheads will be equalized.

In those same European missile negotiations, the United States has demanded an equal number of medium-range missiles as the Soviets. Fine, but count them against the 1,350 launcher and 5,000 warhead limit. The Soviets claim their SS20 medium-range missiles compensate in part for the 162 land- and sub- based nuclear missiles in the hands of the British and French. Okay, make all the SS20s the Soviets deploy above 162 (and they are up to 343 now) accountable against Moscow's overall launcher and warhead limits.

There also is a limit of 350 strategic aircraft in the Soviet START position -- which is close to the 400 the U.S. has proposed. Add some sublimit -- say 150 -- to hold down the number of those planes that could carry air-launched cruise missiles. Ship- and sub-launched cruise missiles also should be limited in number and range, say to a maximum of 600 kilometers.

If these reductions are not deep enough for the Reagan group, make them the initial five-year goal. Then build in additional 10 percent reductions in each of the following five- year periods, making it a 15-year agreement overall.

It's time the two superpowers recognize their strategic nuclear forces have been structured in very different ways and that arms agreements should not try to smooth over those differences.

Before the 1972 agreements, the Soviets were recognized as superior in numbers of land-based ICBMs, the Americans in sub-based SLBMs. Then, in an unexpected compromise, Henry Kissinger gave in to the then- Soviet President Leonid Brezhnev's position and gave the Russians a larger number of ballistic missile submarines than the U.S.

Since that time, it's been the unhappy history of arms control negotiations that both nations have tried to use the talks to develop agreements which allow their own weapons production plans to continue uninhibited, while limiting those that might strengthen the other side.

Today, Moscow and Washington are playing that same game. What's needed, however, is compromise which would require each side to give up some of its favorite, irrational, but long-standing strategic myths.

For the Soviets, that means dropping the need to have more sub- launched as well as land-based missiles, and accepting the idea that bombers with cruise missiles are not as dangerous as ballistic missiles.

For Americans, it means forgetting the notion that the Soviet numerical advantage in large, land-based ICBMs makes a meaningful difference. The new Trident II sub-based missile will be the equalizer, carrying eight warheads equivalent in power and accuracy to the land-based MX or Soviet SS18, that country's biggest and best ICBM. U.S. officials also should recognize that Pershing II and ground-launched cruise missiles based in Western Europe have to be considered strategic weapons, as Moscow claims.

It's not difficult to guess that this approach will run into trouble in Moscow. Initially, it would come from a stolid bureaucracy that, as shown by the handling of the Carter, March 1977 proposal, is just not equipped to deal with new initiatives.

In Washington, there would first be opposition from the two existing negotiating delegations. Which one would run the ensuing talks? Then there would be the proponents of positions that were endangered. For the Reagan team, giving up on ending the Soviet "heavy" missile advantage is a no-no, according to key administration officials.

But from a practical and rational point of view, the path toward compromise seems clear. However, who ever said that nuclear arms or their control had either of those qualities?