CHICAGO, May 1975. For no apparent reason, a stranger began shouting abuse at a young mother who was traveling with her children on an "L" train. After the two uniformed officers present detrained, he threw her to the floor and began dragging her down the aisle, while the conductor and five other passengers did nothing. Drawing a "Saturday night special," she shot him in the shoulder and held him for police.

Chicago again, October 1976. Returning to her apartment, Ms. W found a man who had entered by battering down a wall, and had raped her roommate and thrown her out the 15th-story window. When Ms. W. drew a gun, which she carried because she felt the neighborhood was bad, the attacker fled.

These examples highlight my reasons for disagreeing with my fellow liberals who urge a national handgun ban, a demand that has risen again in the wake of the fatal shootings of Dr. Michael Halberstam and John Lennon. My concern is not with the self-defense value of handguns, but with much more fundamental question inherent in banning them:

Do we really want to punish these women and others like them for keeping handguns to defend themselves and their families? Furthermore, how could the enforcement of such a law possibly be effective or fair?

Perhaps we would be marginally better off if even responsible people didn't have guns for sport or protection -- just as we might be better off if even the responsible didn't drink or use pot socially. But the cure would be worse than the disease, because it would severely punish millions of responsible people -- which is what would be required to make a handgun ban effective.

There are more than 50 million handguns in circulation today, and at a conservative estimate, more than half of the owners would defy a national confiscation law. These figures alone suggest the vast number of problems that would arise if a handgun ban were enacted.

The most obvious inference to be drawn is that the overwhelming majority of gun ban resisters will not be criminals in any ordinary sense. They will be people who believe (whether rightly or not) that they have both a consitutional right and a need to possess handguns for self-defense. As Frank Zimring, the leading academic proponent of a handgun ban, acknowledges, fewer than one-half of 1 percent of all owners misuse their handguns. (of course, even one-half of 1 percent of 50 million allows for a staggering amount of crime, particularly since the same gun may be used to commit 80 or 100 robberies, year after year. But reducing the overall number of handguns will not reduce the number of handguns criminals have because criminals are not about to give them up.) But we aren't talking here about one-half of a percent; as I have already suggested, the number of those who would defy a handgun ban would more like 50 percent of 50 million.

This brings us to the problem of enforcement: Effectively discouraging such non-compliance would require a law with real teeth. But the mere threat of punishment would not be enough to deter handgun owners who believe that they need their weapon for the protection of their family in a violent society.

Compliance could be achieved only by enforcement so vigorous that resisters would be more frightened of the law than of living without the security of a handgun. Simple mathematics demonstrates the staggering cost of such enforcement. Jailing only 1 percent of those likely to resist would be far beyond the resources of the federal prison system. Clearly an across-the-board enforcement of a handgun ban would be too costly -- and in the end, impossible.

We already know the probable result of such an impossible situation: selective enforcement. Our experience with sex and pot laws teaches us that when the cost and injustices of enforcing a ban grossly exceed the benefits, confidence in the law among police and prosecutors, judges and juries -- not to mention the citizenry itself -- rapidly fades. Enforcement becomes progressively more haphazard until at last the laws are used only against those who are unpopular with the police. We hardly need to be reminded of the odious search and seizure tactics police and government agents have often resorted to in order to trap violators of these laws.

But what about all the evidence that handgun bans, if enforceable and rigorously enforced, greatly reduce violence? Alas, I can find no such evidence, nor do I believe any such evidence exists. Gun control is a plausible theory. But there are other equally plausible theories, some of which may even increase violence.

Before we turn a pausible theory into a federal law that jails people for mere ownership of guns, we had better see how that theory has worked out in actual practice. But this is a subject on which gun prohibitionists remain prudently silent.

They prefer to parrot endless statistics on the use of handguns in crime -- leaving the implications that banning them would marterially reduce violence. To determine whether a ban would have the desired effect, the federal government funded at the University of Wisconsin the most massive and sophisticated study ever done of gun laws in practice. Reported in October 1975, its "conclusion is, inevitably, that gun control laws have no individual or collective effect in reducing the rate of violent crime."

Likewise the only in-depth study of Britain's handgun ban, conducted at Cambridge University in 1971, finds that it has had no ascertainable effect upon violence. Indeed the British homicide rate has doubled in the last 15 years while the American rate has risen less than 30 percent in 45 years. (If, despite the fact that theirs are escalating faster, British violence rates remain far below ours, it is only because the British started far lower -- in the era before Britian banned pistols.)

Noting that "the use of firearms in crime was very much less (before 1920) when there were no controls of any sort," the Cambridge report concludes that social and cultural factors, not gun control, determine violence levels. Thus the report recommends "abolishing or substantially reducing controls" because their administration and enforcement divert immense police resources from real crime control.

It is frequently theorized that murders would not occur if handguns weren't lying around for ordinarily law-abiding citizens to grab when they lose their tempers. To test this, the Wisconsin study compared homicide rates of handgun ownership, state by state. It found no correlation. In other words, rates of homicide do not decrease as rates of handgun ownership decrease, nor do they increase as rates of handgun ownership increase. Note that this finding undercuts the whole theory of handgun prohibition. If reducing handgun ownership does not reduce homicide or other violence, a handgun ban is just one more diversion of police resources from real crime to victimless crime.

The Wisconsin study's national evidence is reinforced by the Harvard study of a 1974 Massachusetts law providing a mandatory year in prison for having a handgun without a permit. The study found that during 1975 this widely publicized law sharply reduced the carrying of firearms and the number of assaults with firearms. Having swallowed the prohibitionist theory that this would sharply reduce homicide, the authors were surprised to find no corresponding reduction in any kind of violence. As previous criminological studies have suggested, deprived of a handgun, a momentarily enraged citizen will resort to the far more deadly long gun. Deprived of all firearms, he will prove almost as deadly with knives, hammers, etc.

On examination, the "evidence" for handgun prohibition boils down to the fact that most of the progressive intellectual or political figures of our day detest guns and seek their elimination. But opinions and emotions, however unanimous, are not evidence. And they certainly don't justify jailing otherwise law-abiding people, who act upon different opinions.

Turning to the enforceability issue, the gun prohibitionist argument so closely parallels Prohibition theory that it might have been taken directly from the speeches of William Jennings Bryant or Jane Addams. Antigun liberals seem to have learned nothing from our Prohibition experience or even, amazingly enough, from their own cogent arguments against marijuana bans.

They swallow whole the myth that no matter how deeply valued a commodity is, it can be eliminated by drastic penalties massively enforced. Like the prohibitionists of old, their faith is never shaken by the persistent failure of state legislation based upon this myth. To their minds this only argues for a national gun ban which would, they are confident, be enforceable.

The most favorable possible climate for enforcement of gun bans is surely England, an island, relatively small, isolated and well policed, which has not had a major war in more than 30 years. Yet the Cambridge study found that "50 years of very strict controls on pistols has left a vast pool of illegal weapons" in Britain.

Consider our own federal law against the submachine gun, a weapon which never had many civilian sales when it was legal, since it has no legitimate sporting or self-defense use and was inordinately expensive. Nevertheless, submachine guns are easily acquired today by lunatic-fringe groups and at comparatively lower prices than before 40 years ago. How then could we ever hope to coerce the elimination of handguns, which have been available since 1835 and of which there are more than 50 million in America today?

The problems of enforcing a handgun ban far exceed those we experienced in enforcing the Prohibition Amendment. Unlike liquor, a handgun is a single-purchase item; properly cared for, it can last hundreds of years. Admittedly, the effective life of ammunition often will exceed 30 years. But it would be impossible for police to stop handgun owners from purchasing a handful of ammunition each generation, when Prohibition agents could not prevent the distribution, sale and resale of millions of gallons of liquor each month.

Since compliance cannot be coerced by enforcement, the question becomes one of voluntary compliance. The Cambridge report found that compliance occurred among law-abiding Britons only because England was so peaceful when handgun prohibition was adopted that few considered guns necessary for self defense. As I have already suggested, 50 percent noncompliance is a very conservative estimate of the likely response in our violent society to a national gun law.

How would Congress react to the massive defense of a handgun ban? The example of foreign countries -- and of states like Michigan and Missouri that have British-type permit laws -- is not encouraging. The Cambridge study notes that every new British firearms act has brought the police unheard-of new powers of search and seizure. In Holland, Taiwan, Northern Ireland, Rhodesia, South Africa, Jamaica and the Phillippines (to name a representative few), gun laws are enforced by indiscriminate search.

Liberal prohibitionists will protest that they do not propose extreme penalties or unconstitutional enforcement. But the escalation of penalties and of repugnant police practices is virtually inevitable when laws enacted with great expectations (but in defiance of empirical evidence) fail to stop crime.

For just as the zealots now cling to the plausible theory of gun bans despite the steadily increasing factual evidence of unworkability in practice, so will they continue to believe that "just one more" escalation will bring us to their uptopia, in which no one has guns and no one murders.

We will approach that utopia only when violence rates have dropped to the point at which people no longer feel they need guns for protection. That, as the Cambridge study tells us, was the happy situation that enabled Britain in 1920 to go -- with minimum resistance -- from virtually no gun control to complete prohibition. Unfortunately neither right nor left-wing ideologues are willing to deal with the fact that utopia cannot be brought about by easy, mechanistic solutions -- whether these be gun bans, abrogating defendants' constitutional rights or quantum leaps in the numbers of snoopers and informers.

Violence can be decreased only through painful, basic, long-term change in the mores and institutions that produce such an unusually high percentage of violence-prone individuals in our society. If we are unwilling to make such sweeping change, we must reconcile ourselves to living in a violent society -- a society that has no right to prevent victims from arming themselves against the violence it could eliminate, but has chosen not to.