In The Pink Panther Strikes Again, when Peter Sellers's Inspector Clouseau blunderingly demolishes a grand piano, a horrified onlooker exclaims, "That's a priceless Steinway!" Replies Clouseau: "Not anymore."

More than a few Americans now find themselves wondering whether marriage is that piano. On Monday, barring the unexpected, the state of Massachusetts will begin issuing marriage licenses to same-sex couples, under orders from the state's Supreme Court. For the first time, gay marriage will enjoy clear statewide legality. Voters will get the last word in a statewide constitutional referendum, but the earliest that can happen is in 2006.

In the United States and the rest of Western civilization, marriage has always been between a man and a woman. As Clouseau said: Not anymore.

More than two dozen other states are rushing to write gay-marriage bans into their constitutions. Some of the bans are inspired by panic, or by dislike of homosexuality. But even many people of goodwill toward their gay and lesbian fellow citizens blanch at redefining society's most basic institution. Gay marriage, to them, seems risky.

They have a point. Gay marriage is risky. But not trying gay marriage is riskier.

To many of its supporters, gay marriage is a civil-rights issue: Marriage is a right, and every couple should have it. To many of its opponents, gay marriage is a moral issue: Homosexuality is wrong, and society should not condone it. Well, gay marriage is a civil-rights issue and a moral issue, but it is also, perhaps most importantly, a family policy issue. Right now, Americans are deciding the shape of marriage—the basic legal and social framework of family—for years to come. Risk, therefore, is just as relevant as rights or as right and wrong. What, then, is the balance of risks?

Begin with what we know for a fact: Something like 3 to 5 percent of the population—all gay and lesbian Americans—are locked out of marriage, which is life's most stabilizing and enriching institution. Even after accounting for differences between the married and unmarried populations, married people are healthier, happier, more prosperous, more secure; they even live longer. To shut millions of Americans off from those benefits is to inflict a very real harm. Moreover, many same-sex couples are raising children: several hundred thousand, at least, and possibly more (there are no firm figures). Presumably, those children would be better off with married parents.

So same-sex marriage would benefit gay people and the children they are raising. That much meets with little dispute. But what about the rest of society? Here the debate turns to what economists call "externalities": harms or benefits to society at large that flow from private decisions.

Opponents of same-sex marriage insist it will bring grave, perhaps catastrophic, negative externalities that will hurt millions of American families. They have yet to explain, however, precisely how allowing same-sex couples to marry would damage anyone else's marriage or family. More plausible is a second common view, which is that same-sex marriage will have little or no impact on straight families. No-fault divorce changed the terms of marriage for heterosexual couples, which was plainly a big deal. The only thing that same-sex marriage does, by contrast, is to expand by a few percentage points the number of people who are eligible to marry their partner.

Less often noticed is a third possibility: positive externalities. Today, a third of all American children are born out of wedlock, cohabitation is soaring, and nearly half of marriages end in divorce. Marriage's problem is not that gay couples want to get married but that straight couples don't want to get married or don't manage to stay married. At long last, gay marriage provides an opportunity to climb back up the slippery slope by reaffirming marriage's status as a norm—not just as a right but as a rite, the gold standard for committed relationships. Gay marriage dramatically affirms that love, sex, and marriage go together—that if you really care, you marry. No exclusions, no excuses.

So gay marriage entails potential social benefits as well as potential risks, even apart from the unquestioned benefits for gay couples. And there is a further element, as important as it is overlooked. Banning gay marriage entails its own risks to marriage. And those are not small risks.

Because society has an interest in seeing same-sex couples settle down and look after one another, and because gay couples' friends and family care about their well-being, committed gay couples are winning increasing social support. One way or another, legal support will follow. Banning gay marriage guarantees that the country will busy itself creating gay-inclusive alternatives to marriage (which will be tempting to heterosexuals) and bestowing legal rights and social recognition on cohabitation (which is open to heterosexuals by definition). The result will be to diminish marriage's special status among a plethora of "lifestyle alternatives"—the last thing marriage needs.

Moreover, the gay exclusion risks marginalizing marriage by tainting it as discriminatory. A March Los Angeles Times poll finds that more than 80 percent of young people (ages 18 to 29) favor anti-discrimination protections for gay people. More than 70 percent believe gays should receive the same kinds of civil-rights protections that are afforded to racial minorities and women. More than half favor gay adoption, three-fourths believe that "a gay person can be a good role model for a child," and more than 70 percent can "accept two men or two women living together like a married couple." Seventy percent describe themselves as sympathetic to the gay community (versus 43 percent of people 65 and older). And three-fourths support gay marriage or civil unions—with the plurality favoring marriage.

In other words, America's young are much more hostile to discrimination than to gays or gay marriage. They will increasingly view straights-only marriage the way their parents have come to view men-only clubs: as marginal, anachronistic, even ridiculous. This is not conjecture; it is already beginning. San Francisco regarded its decision to marry gay couples as a protest against discrimination, and Benton County, Ore., recently stopped issuing marriage licenses altogether, on the grounds that it wanted no part of a discriminatory institution.

"We are genuinely running the risk of making marriage uncool," Frank Furstenburg, a University of Pennsylvania sociologist, said last month, in an Associated Press article about straight couples who are boycotting marriage to protest discrimination. Today, such couples are rare. But in 10 years? Twenty? So there are risks, large risks, on both sides of the equation. Banning same-sex marriage is no safe harbor. Given that fact, it is irresponsible not to try gay marriage, at least if protecting marriage is the goal.

Banning same-sex marriage nationally, as President Bush and many conservatives would do, is hardly a conservative approach; it risks putting marriage on the road to cultural irrelevance. On the other hand, national enactment would be an irreversible leap into the unknown. There ought to be a way to try same-sex marriage without betting the whole country one way or the other. And there is. Try gay marriage in a state or two. Say, Massachusetts.

Massachusetts is one of only a handful of states where gay marriage can legally happen (most states have enacted pre-emptive bans). Its law prohibits marrying out-of-state gay couples, so the experiment will be local. Massachusetts is gay-friendly, allowing same-sex marriage a fair trial. And it gives the final say to the voters, not judges or politicians or bureaucrats. In short, Massachusetts is the perfect laboratory for an experiment that needs to happen.

Starting Monday, and probably for years to come, America will no longer have a uniform national definition of marriage. That is nobody's first choice. Conservatives wish the issue had never arisen and hope, unrealistically, that a constitutional amendment will put the cork back in the bottle. Many gay-marriage proponents wish, just as unrealistically, that the courts could settle the issue quickly by fiat.

But neither a constitutional amendment nor a Supreme Court order could resolve what is, at bottom, a fundamental schism in the social consensus: Older people see same-sex marriage as a contradiction, and younger people see opposite-sex-only marriage as discrimination. Reconciling marriage with homosexuality, equality, and society's needs will be messy, but, as Robert Frost said, the only way out is through. Massachusetts is as good a starting place as the country could have hoped for.