"One of the ways we think of our national identities as natural is that we can tell what people's nationality is from their last names," she explains. "Governments have put a lot of effort into deciding what we're named. For example, there's an official list of first names in Switzerland and you have to choose one for your child."

Stevens maintains that in the 11th century, minions of William the Conqueror created surnames in the midst of a census to codify inheritance rules and thereby bolster tax revenue. Later in Europe, surnames were used to control and homogenize various ethnic groups.

"Jews wouldn't take last names because Moses didn't have a last name," she explains. "But when they rebelled, the government assigned them really gross last names like -- Grossman -- which means fat man. If you wanted a pretty name, you had to a pay a bribe."

But how do the nation-building origins of the surname shed light on the personal choices made by modern couples?

"Inheritance laws, political bodies, surnames -- it's all about compensating for men's inability to give birth," Stevens contends. "The surname remains the only way of showing legitimacy. Without it, there's no certainty that the kid has a legal father."

She also hazards a psychological hunch that women still want to demonstrate that they've nabbed a man. "That's especially important if women are keeping their own last names. It's ironic because keeping one's maiden name is supposed to be feminist -- but it may ignite that old anxiety about legitimacy."

But if it's all about legitimacy, why didn't any of the women I spoke to offer that as an explanation? And why did so many of the stories seem so different?

"It is interesting when you get many explanations for the same choice," muses psychoanalyst Nancy Chodorow, author of the ground breaking "Reproduction of Mothering" (UC Press). "One begins to wonder what's going on unconsciously." In her current book, "The Power of Feelings" (Yale University Press), Chodorow addresses this conundrum: How many so-called "personal choices" often have internal and unconscious meanings.

Like Stevens she feels that patrineal surnaming is about a woman giving her child and its father a definite connection. But she casts the choice in a more positive light. "[Giving the man's last name to the child] can be a way of having a sense of two parents," she explains. "It's also a way of trusting in the marriage -- saying, 'This is someone I can count on.' It's about enjoying the good parts of being part of a family, of feeling somehow that this man is making a commitment."

Yet it's interesting that traditionally, the man shows his commitment to the child by giving his name, while the woman shows that same commitment by giving up her own. Why are so many men still so attached to their last names?

"Identification with the father," says Chodorow. "I don't think it's any mystery. It's the classic "in the name of the father" -- in Lacanian psychoanalysis. The mother has the baby in utero but the name is how men get tied to their babies. The tie has to happen somehow that 'This is my baby too.' If she's feeling generous, then this is a way to show it."

Choderow also notes that many young feminists are choosing their battles more carefully. "Women are making choices about where they think it's important [to change] -- maybe they're focused on getting men more interested in child care. They're also learning that every time you do something that's traditional, it doesn't mean that you're not a feminist."

Evolutionary biologist Helen Fisher doesn't dispute Chodorow's notion of patrilineal naming as a linguistic umbilical cord, but she casts the idea in biological terms. "It's tremendously advantageous to think that the father belongs to [the mother and the child] for Darwinian evolutionary reasons. The main reason for marriage is for women to get a man to not only sire her children but to help raise them.

"Even in the age when women can be economically powerful, any way they can build that connection with their husbands [means] they will win VCRs and bicycles and college educations for their DNA."

She notes that studies have shown that mothers and their kin comment more often that a baby resembles its father. "Evolutionary psychologists ended up thinking that this habit is more than just chance; it is a way of building that connection [between father and child]," she explains.

Even with the high rate of divorce, the increasing economic power of working women and the decline in marriage, Fisher doubts that the prevalence of patrilineal naming will change any time soon. Why? Because illegitimacy is not just a paranoid male fantasy.

Order "Mothers Who Think: Tales of Real-Life Parenthood" from the editors of Mothers Who Think.