ARIE AND ANTHONY B. of New Orleans were divorced in 1978 after a 10-year marriage. Marie was awarded custody of their 2-year-old son Terry. Within a year of the divorce decree, Anthony sued her to win custody of the boy but did not get it. He sued again in 1980, and again the next year, each time without success.

In 1983 the child, then 8, visited his father for Christmas. Rather than send him back, Anthony enrolled him in a local school, later claiming this was done with Marie's consent. But a court credited Marie's story that it had happened against her will. When the school year ended, Terry returned to his mother. On his next visit to his father, however, Anthony filed again for custody.

The suit ground its way forward and finally came up for trial. The court found that Marie was highly fit as a parent and that the boy had flourished under her care. End of case? Just the beginning. Under Louisiana's 1977 child custody law, the court explained, it mattered not how fine a home Marie had provided; the only question was whether Anthony might perhaps provide one that was even better. Apparently swayed by a wish to avoid another change of schools, and by the boy's expressed inclination to go on living with his father, the court ruled for Anthony. After two appeals by Marie, the court sent the case back to the trial court. Thus, after courtroom fighting stretching over eight years, no end was in sight for this wretched family.

Custody disputes are as old as Solomon, but the fight between Marie and Anthony was in its own way at ground zero of America's litigation explosion. It shows the effects of distinctive changes in what our courts try to do when they hear lawsuits.

Until not long ago, most American courts resolved custody disputes through "mechanical" rules or presumptions. If the parents had agreed on custody at the time they separated, the agreement stuck. Otherwise, the mother beat the father, at least where the children were of "tender years." Both mother and father beat third parties. Once custody was awarded, it tended to stay put.

These rules of thumb were not the last word, and sometimes the calculations got more complex. But judges could dispose of most cases in a fairly rigid way, with an air of looking up the answers as from a cookbook or logarithm table.

The drawbacks were obvious. By disposing of cases wholesale, the law deliberately turned its gaze from individual equities. Surely, critics urged, everychild and parent deserved a searching look at his or her own personal situation, with as much warmth and personal attention as a court could muster.

Through the 1970s and 1980s the mechanical presumptions were weakened, usually with little fanfare. A 1982 case came to symbolize the new approach. In that case, a New York court switched custody to a father on the grounds that while the mother was not at all unfit, the father now appeared more fit -- the "sole criterion" being the "best interests of the child when all applicable factors are considered."

The formula sounded so reasonable that it spread rapidly around the country. The "best interests of the child" standard, it soon became apparent, has three interlocking features. First, it provides what might be called hair-trigger litigability. To get into court, you need not assert that a switch would make the child better off by a mile; a millimeter will do.

The second feature is subjectivity: No two parents, judges or hired experts ever quite agree on the precise content of a child's best interests. Is it better off with a parent who hews to higher moral standards, or one who is more affectionate? With a parent who will spend more time looking after it, or with one who will arrange superior schooling? A lot depends on which judge you draw, and what mood the judge is in this morning.

The third feature is that everything comes to be relevant and nothing, as the lawyers say, dispositive. Does your ex swear? Smoke? Gamble? Watch too many soap operas? Does he roam the beach gathering driftwood? Do the neighbors find her stand-offish? Perhaps none of these peccadillos significantly endangers a child, but all can have some effect and you never know what will tip the balance. So it can't hurt to bring them all up.

As the courts left off trying to be like answer machines, they started to become more like hot-tub discussion groups. ("Tell us all about the relationship.") Civil libertarians were soon expressing unease about custody litigation: parents were feeling pressured toward conformity, and privacy was getting lost in the trample.

The "good-parent" wars turn out to have made surprisingly little difference in custody outcomes. Most children still go to the mother, and the exceptions tend to come in the same old areas. And yet the effects have been profound. To avoid custody battles, for example, many mothers will make major concessions on financial issues. Knowing this, many lawyers for fathers deploy the custody weapon even if their client is not all that keen to take the kids. As for the children, a distinguished battery of psychologists, social workers and sadder-but-wiser judges have by now concluded that the sheer experience of being fought over for years can endanger the emotional health even of a normal child with good parents. The maternal-preference rule may or may not embody any timeless wisdom about the special bond between mother and child. Actually, as a rule, it is of relatively recent vintage: For a long time, it was the father who got presumptive custody. What is important, almost more than which rule prevails, is that there be a rule, and one as clear, knowable and universal -- as mechanical, in short -- as can be. A good rule is comprehensive, disposing of run-of-the-mill cases so courts can concentrate on the unusual. It is objective, so different judges can hope to rule the same way, and the parties themselves can have some hope of agreeing on how the law will treat their case. And it is clearly spelled out in advance, so everyone can know where they must toe the line.

For a very long time the law in this country, and in England before that, showed an overwhelming preference for fixed rules over fuzzy standards. Some rules have survived more or less intact to our own day. Take the idea of an age of majority. When is someone old enough to get married, or sign a binding contract or order a drink in a bar? If the law were fully alive to human realities, it would entertain much litigation on this subject. Instead it promulgates a no-thinking-required rule -- count the birthdays -- lest it cast perennial suspense over the eligibility of a million bachelors, the validity of a million credit cards and the retention of a million liquor licenses.

How should we behave on the highway? One familiar rule holds that if you bump into the driver ahead of you, you pay for the damage, even if he slammed on his brakes for no obvious reason. If courts tried to work out responsibility on a case-by-case basis, they'd find it hard to distinguish the occasional innocent rear-car driver from all the tailgaters. And the simple rule of thumb provides a generally sound maxim: Pay more attention to the car in front than to the car behind. As a group, drivers benefit.

When does a binding contract come into existence if you are negotiating by mail? In principle, perhaps, when there's a "meeting of the minds." But minds are annoyingly inscrutable. Hence the law's "mailbox rule," which provides a fairly objective trigger: You're both locked in when you drop your letter of acceptance into a mailbox.

Of course, no preannounced rule of thumb ever quite fits the perceived merits of all the cases that come in. When the immensely influential group of thinkers known as the Legal Realists came along early in this century, they launched a devastating attack on the effort to maintain clear, knowable-in-advance legal rules. In fact, they said, judges were strongly tugged by their sense of the equities of a case, the policy of society as a whole, and other factors outside the letter of the law; and they had a dozen ways to manipulate cases to make them come out according to these lights.

Language itself, for starters, was slippery: The town elders might think they were being plain enough when they banned vehicles in the park, but did that mean a court had to apply the law to bicycles? Wheelchairs? A statue of the general in his Jeep? There was a malleable quality as well in the way a case's facts were "characterized" among legal categories. Most devastating, there were so many different precedents, maxims and canons of interpretation that by picking and choosing, a judge could steer cases to any desired destination point. Why not be candid? Why not drop the answer-machine pose, and announce openly that the reasons for decisions were not easily pinned down?

Even the traditionalists had resigned themselves to a considerable tincture of indeterminacy in a few areas of the law. The whole concept of "negligence" in injury cases was sadly amorphous, with many outcomes unguessable until a jury came back with a verdict. A dispute over whether someone named in a will exerted "undue influence" over its maker requires a study, by circumstantial evidence, of what commentators have called the psychological world of the dead person. The antitrust laws from early on were vague and open-ended in their commands.

This was not an encouraging sign, since negligence, will-contest and antitrust disputes were all notorious for their expense, vexation and scope for litigious zeal. But the realists carried the day. From the courts' undeniable failure to provide perfect guidance or apply wholly neutral tools of analysis, it was deduced not that they should try harder but that they should stop trying so hard. The basis of the custody-challenge explosion thus turns out, with appropriate substitution of subject matter, to be the basis of most of the rest of the litigation explosion. Vagueness creeps into the law on the padded feet of words and phrases like fairness, equitableness, good cause, good faith -- pillowy expressions that tend to soften the blow of what is in fact a grant of wide judicial discretion over some area. Consider the splitting up of the property in a divorce. Most states have embraced the "equitable distribution" of property, a nice-sounding phrase that calls for a long fight on everything of conceivable relevance.

Then there's another current favorite, the "balancing test" that instructs future judges to weigh 10, 20 or 30 factors No method is provided for figuring out which factors should trump which others, or what to do when six factors cut one way and eight the other.

Lawsuits against manufacturers over injuries in the use of products, one of the biggest growth areas in litigation, is handled through one of the most amorphous balancing tests yet invented. Someone hurts himself using a caustic drain cleaner or trying to pull a carving knife out of its original holder. Should the product maker have to pay? The most widely used set of modern guidelines invites the jury to consider at least a dozen factors including the likelihood that users will hurt themselves with the product; the probable seriousness of those injuries; the ability of users to avoid dangers by being careful; the user's likely awareness of the dangers; and the general public's knowledge of the same thing.

Got that figured out? There is much more. The bewildered jury must then consider the product's usefulness; whether it could have been made safer without making it less useful or "too expensive to maintain its utility"; whether other products might serve the same need and not be as unsafe; and on and on. Not surprisingly, many companies have decided to flee the uncertainty by declining to market useful products often found on the scene in injuries, from football helmets to life-saving drugs.

Today's legislators are making things far worse. Federal appeals judge Alex Kozinski writes that it "seems as if legislators now pass statutes because of, not despite, their lack of clarity. By using vague language, legislators can avoid making the difficult political choices that they have to confront when drafting a statute precisely."

Many books could be written on the impossibility of getting clear answers to single questions like, "Is this a legally adequate environmental impact statement?" or "Is this a lawful way to draw the bounds of city council districts?" or "Does this firefighting exam unlawfully discriminate against the handicapped?"

Theory often follows practice, and now some members of the leftish Critical Legal Studies movement have come along to proclaim that indeterminacy is not just a tolerable evil in the discipline of the citizenry but a downright good. Leading C.L.S. light Duncan Kennedy of Harvard, in a well-known article, has suggested that vague standards are preferable to definite rules precisely because they give citizens the sense that the totality of their behavior is under scrutiny by the tribunes of society. But while lofty ideals of jurisprudence can sometimes seem remote from the needs of everyday life, legal determinacy and predictability have the homely as well as the transcendent virtues. They allow us to plan our dealings in times of legal peace. They take on new importance at the first signs of a quarrel, when we want to know whether we have fallen short of our duties, so we can make amends, or stand on firm ground, so we can frame demands. If the quarrel goes to litigation, reliable law encourages a prompt settlement or, if worst comes to worst, cabins the bounds of what is at stake in a trial. Afterward it provides reassurance that the verdict hung on objective factors rather than bias or luck and quells the nagging suggestion that fighting just a bit harder might have turned the outcome.

The surest way to destroy predictability is to tell citizens and judges alike to decide what to do when interests clash by looking into their consciences. Until that happy day on Canaan's shore when all consciences converge, the only result can be a Babel of disagreement. Today's voluminous case reports are filled with lawsuits between respectable citizens whose sense of fairness inevitably differs but who would not for that reason, in an era of clearer law, have had to come to legal blows. Where people of honest intention are suing each other in large numbers, it is because they have been baffled in their efforts to learn in any other way what the law expects of them. "No profound social theory is needed to explain why people are more litigious today than ever before," as Richard Epstein of the University of Chicago puts it. Legal uncertainty "breeds litigation . . . . It's that simple."

Walter Olson, a senior fellow at the Manhattan Institute, is the author of "The Litigation Explosion" published this month by Dutton, from which this article is adapted.