The doctors are at it again. It's not exactly clear whom they're addressing, but they're out to prove that nuclear war would be bad for public health. Yet another symposium on "the medical implications of nuclear war" has just been held, this time at the National Academy of Sciences.

The result was a ghoulish competition in number production. Fourteen million to 15 million dead, says the government. No, 36 million to 56 million dead, says Theodore Postol of Stanford. Thirty-eight million, says the "von Hippel group" of Princeton -- and for an attack on just U.S. military-industrial targets. If only strategic targets are hit: 19 million.

It has always struck me as odd that whenever there is a disaster the press becomes fixated on ascertaining exactly how many died. "It had been thought that 36 people perished in the train wreck. The toll has now been put at 38." These numbers are ostensibly about people. But since none of them is known to us, they are really only about numbers. What can such numbers possibly mean?

And what is the point of such precision? True, the order of magnitude (whether tens or hundreds or millions) gives an idea of the scale of the suffering. It is a way of establishing hierarchies of grief. (We do the same in words by using such graded terms as accident, disaster, catastrophe and holocaust.) But that does not explain the need to say 38 instead of 36. The precision here, I suspect, has a psychological basis. It is a desperate attempt to keep busy in the face of helplessness. And it fills a moral need too. It helps fend off the shame and guilt that attends those whose job it is merely to record suffering.

Thirty-six million or 14 million. Does anything change if you change the number? Nuclear war remains the ultimate calamity in either case. The idea is to prevent it. Obsessive, preemptive body-counting is busy work and a confession that one does not know what to do.

Except for the newly minted numbers, there is nothing very new here. Why then did the symposium make page 27 of The New York Times? The headline will give you a clue: "Study Sees Atom War as Causing AIDS Epidemic." One research team delivered up the news that nuclear war would cause "a marked increase in the incidence of AIDS and AIDS-related diseases."

And just when you thought it was safe to send your kids back to school. Is the fear of nuclear war in such alarming decline that it needs a transfusion of AIDS hysteria?

What next? The effects of nuclear war on the Checker cab population of New York? I can see it now: "Estimates for a one-megaton attack range from between 70 to 80 percent of the cabs being knocked out of service. The Von Braun group has found that of the remaining 20 to 30 percent, most would lose air conditioning. Health effects are incalculable, but with the help of the new Gulag supercomputer we estimate 10,000 casualties from the combined effects of heat stroke and violence among people competing for the same cab. (We assume no pre-nuclear gun control.)"

Am I being frivolous about a serious subject? No more than the doctors. This esoteric competition for new and more horrible discoveries about the post-nuclear world is a transparently political exercise dressed in scientific garb. The garb permits the doctors to deploy white-coated authority, but does little for the rest of us. Compared with the question of how best to prevent nuclear war (arms control, "Star Wars," early warning, capitulation?) or how best to combat AIDS (today, not after the apocalypse), body- counting and refined end-of-the-world diagnostics is indeed frivolous.

In the paper on nuclear AIDS, immune suppression is traced to various causes. Among these are radiation, malnutrition and psychological factors, such as "stress, depression and bereavement." And what do we know about the psychological health of the post-apocalyptic world? "Chazov and Vartanian have estimated" -- I'm not kidding now, I'm quoting -- "that at least one-third of the population surviving a nuclear war would suffer from severe mental and behavioral disturbances."

Well, it happens that of these two authorities, the original claim to fame of one (Chazov) is that he was Brezhnev's cardiologist. The other (Vartanian) is the operational head of the Soviet psychiatric institute that devised the criteria for classifying political dissidents as mentally ill. Come to think of it, perhaps they are experts on depression.

Okay. Let's leave personalities out of it. Authority is authority. One-third, is it? Not one-fifth or one-half? Do the other two-thirds carry on as usual? Such precision. A famous epidemiological study of mental disturbance once reported that 81 percent of the population of midtown Manhattan suffered from some mental impairment. And that was 1962. But cabs were hard to find even then.