Earlier this month, an attempt to undo gay marriage in Massachusetts went down in flames. State legislators rejected a proposed constitutional amendment that would have voided a 2003 ruling by the state's Supreme Judicial Court extending marriage rights to gay and lesbian couples. The measure lost by a stunning 157-to-39 margin. In 2004, the amendment (which must be repeatedly approved to be adopted) had narrowly passed the legislature, 105 to 92, before the ruling went into effect.

Even in notoriously liberal Massachusetts, opinion among the general public is more closely divided than this month's vote might suggest. A survey earlier this week found 52 percent of respondents opposed to a ban on gay unions; depending on which of the countless dueling polls you prefer to believe, that may represent an increase over the past year. A 2003 Pew poll suggests that margin will grow as the idea of gay marriage becomes normalized not only in the Bay State but across the country. As people become accustomed to seeing and reading about married gay couples, with few signs of seas boiling or the moon going blood-red, the conviction that marriage just means "the union of a man and a woman" begins to fade. Milton Friedman coined a term for the tendency of government programs, even those highly controversial in their inception, to become sacrosanct: the"tyranny of the status quo." Call this the friendly correlate of that pheonomenon: the benign dictatorship of the status quo. It is the curse of all arguments from How Things Have Always Been: They come with built-in expiration dates.

For conservatives accustomed to insisting that "marriage is what it is," this may seem perplexing. Hasn't marriage been a relationship between one man and one woman since the dawn of time some 6,000 years ago?

That perplexity might fade if they considered the incredible variety chronicled in Stephanie Coontz's excellent survey Marriage, a History: From Obedience to Intimacy, or How Love Conquered Marriage.

Statistically speaking, Coontz observes, the most "traditional" form of marriage is not "one man, one woman" but "one man, as many women as he can afford." Many Native American groups cared about diversity of gender in marriage rather than of biological sex: A couple had to comprise one person doing "man's work" and one person doing "woman's work," regardless of what their genitals looked like. The unique Na people in southwestern China live not in couples but in sibling clusters, with groups of brothers and sisters collaboratively raising the children conceived by the women during evening rendezvous with visitors.

What's most striking, however, is the panoply of purposes marriage has served at different times and in different places, even when Coontz focuses on the more familiar one-on-one heterosexual union.

Among early hunter-gatherer bands, the trading of members to other bands as spouses was, above all, a means of establishing networks of trade and economic cooperation. Once each group had members with loyalties and ties to both, barter became a safer bet. As Claude Lévi-Strauss put it, the "reciprocity which is the basis of marriage is not established between men and women, but between men by means of women, who are merely the occasion of this relationship."

That's not to say that the husbands were in full control either: In ancient Rome, married sons and daughters both lived under control of the patriarch until his death, and in ancient civilizations more generally, marriage decisions were seen as far too important to be left up to the whims of the marrying couple. Here marriage was, above all, about acquiring advantageous networks of in-laws.

In the medieval period, too, marriage might be a handy means of cementing an alliance—or sealing a truce—among rulers. In other times and places, marriage was seen primarily as a means of regulating inheritance or succession. Often—especially where simple market sales of land were tightly restricted—it was the primary means of transferring landed property, and that was seen as the decisive factor in marriage decisions. And not just among the nobility: peasant farmers often held land in separate strips. A marriage might be arranged on the grounds that it allowed adjoining parsels to be united.

Among the working classes in later pre-industrial Europe, marriage was seen as more centrally about the married couple—though a village was apt to intervene if a wedding brought a poor worker into the fold—but as business partners, not star-crossed lovers. Marriage was a way of establishing economic division of labor, and a new widow or widower was seen as a job opening. While it was certainly a bonus if love later developed between a wedded couple, men and women who seemed too infatuated with their partners were widely considered, well, a little odd.

The love marriage, in which people more or less freely chose partners on the basis of mutual affection, was really an 18th century invention, Coontz argues. As late as the mid-19th century, French wags were still bemused at the new fashion of "marriage by fascination." Among some opponents of gay marriage, of course, this is seen as the central problem: the idea that marriage is centrally about uniting a loving couple, from which the notion that it ought to be equally available to gay couples follows. Conservatives, though, sometimes talk as though this is an innovation of last Tuesday's vintage, rather than a transformation that's been ongoing for centuries. As Coontz notes, during the 1950s, conservatives' golden age of marriage, it was precisely this prospect of finding personal fulfillment through marriage to one's soul mate that led to the valorization of married life and its central place in the social imagination.

What emerges above all from Coontz's account is the realization that marriage has no "essence." There is no one function or purpose it serves in every time and place—and indeed, for each function marriage does serve, there is a time and place where some other institution primarily did that work. This shouldn't come as any surprise to readers of F.A. Hayek, who in The Mirage of Social Justice, the second volume of his Law, Legislation and Liberty trilogy, wrote:

Like all general purpose tools, rules serve because they have become adapted to the solution of recurring problem situations and thereby help to make the members of the society in which they prevail more effective in the pursuit of their aims. Like a knife or a hammer they have been shaped not with a particular purpose or view but because in this form rather than some other form they have proved serviceable in a great variety of situations. They have not been constructed to meet foreseen particular needs but have been selected in a process of evolution. The knowledge which has given them their shape is not knowledge of particular future effects but knowledge of the recurrence of certain problem situations or tasks, or intermediate results regularly to be achieved in the service of a great variety of ultimate aims...

Even a knife, though, is not just good at one very abstract sort of thing: You can use it to slice, to puncture, to pry, to crush with the flat of the blade, or, if you're an action movie hero, to jam into a wall and use as a handhold. Institutional evolution, like its biological counterpart, is opportunistic: A structure that serves one function at one stage may be co-opted for a very different function at another stage.

It seems safe to say, then, that marriage in some form or another will be around for a good long while—the fears of conservatives and hopes of a few radical theorists notwithstanding. But the way we understand it is almost as certain to continue changing: Tradition is apt to be a thin bulwark when tradition itself is a chronicle of flux. The meaning of "marriage," as of life, will be the one we give it.