Last month I proposed a new solution to the problem of choosing family names: When you’re starting a family, you and your partners construct and adopt an entirely new middle name, a ‘union name‘ symbolizing your shared life and shared values. If you have children, this union name then becomes their surname.

Many people voiced enthusiasm about the idea, but many also raised interesting concerns and criticisms. I’ve collected them here, with my responses.

Objection 1: It’s better for women to be subordinate, and patrilineal family names help reinforce that. Patriarchal families and societies are happier, stabler, and more successful.

Response: Most women seem to want more autonomy, not less (Pew 2010). That’s very surprising, if autonomy makes them worse off. In fact, that bit of evidence on its own mostly settles the question, until we get strong evidence that women are systematically wrong in this highly specific way about their own interests. We find ourselves in a position similar to an abolitionist trying mightily to refute the claim that Africans love being slaves. Five minutes of talking to people, in a setting where they can talk freely, does the job, and we can move on to more interesting matters.

If there’s compelling evidence to the contrary, I’ll need to see it before I can say much more. On the political claim, too — I need some reason to doubt the surface-level appearance ‘gender equity makes societies more prosperous (Dollar & Gatti 1991; Brummett 2008), healthy (Kawachi et al. 1999), and just (Melander 2005)’.

Objection 2: Union names are too convenient. We should retain an annoying, difficult system, because then it will be more diagnostic of future relationship woes. If people have to fight over whose name gets passed on to the kids, that will ruin relationships that wouldn’t, or shouldn’t, have lasted.

Response: In general, it’s bad policy to make people’s lives worse as a test or trial, unless you are in desperate need for the data that such a test is likely to provide. (And have no other way to acquire such data.) There may be two false assumptions going into the above objection:

(i) Small inconveniences don’t matter. Lots and lots of small inconveniences distributed over a population will add up to have a big impact. And if by chance a lot of them happen in your life at once, they can certainly feel big! People are often under an unusual amount of pressure when they’re deciding whether to have kids or begin a serious long-term relationship. If there’s anything we can do to make their challenges at that stage in life more fun, inspiring, and pleasant, we should jump on the chance.

(ii) If you break up for dumb reasons, you would have broken up eventually anyway. That’s not how relationships work. First, relationships don’t remain at the exact same strength at all times; they can grow in strength. (Or shrink, or oscillate.) Second, failing to overcome a low-level challenge isn’t proof that you would also have failed at all high-level challenges. Bad break-ups can occur just because the wrong thing happened at the wrong time. Life is chaotic, and love’s dynamics are not constrained by what should have happened.

The take-away from this is that we should have compassion and try to make people’s lives better, in small ways and large ones. People don’t deserve solitude or angst just because we find the reason behind their relationship troubles silly.

Also: Union names are challenging. They do help test the strength of people’s commitment. But they do so in a way that tests a relevant skill for romantic and familial relationships. The ability to collaborate, make mutual compromises, come up with imaginative solutions, and find common ground — that’s what union names are training and testing. The ability to be dominant or subordinate, to demand unequal sacrifices, to adhere to out-of-date social norms — that’s what more traditional naming systems are training and testing. I think the former skills are more important for more people.

Objection 3: Naming your children from scratch is hard. Our naming conventions should streamline the process, not add more complexity.

Response: I’d expect social conventions to arise that give people obvious standard choices for surnames — name X after loved one Y, give X common popular name Y, … — so that most people don’t end up inventing names from scratch. That’s how given names currently work, so it’s probably how union names will work too.

As for why we should add even a small amount of work to the process: Human names actually matter. They can have a much bigger and more direct impact on our self-image and social relations than inanimate object names can. If union names encourage people to think and talk more carefully and cooperatively about what identity they want for themselves and their children, great!

Objection 4: Parents can’t be trusted to make up entirely new names for their children. Look how terrible they are just at coming up with decent first names!

Response: It’s certainly a shortcoming of union names that they allow parents to screw up their kids’ lives in more drastic ways. However, if we have mechanisms in place for keeping parents from choosing seriously socially harmful first names for kids, then those mechanisms should generalize to socially harmful surnames.

(In fact, giving parents more leeway might force bureaucrats to take this problem more seriously and put more laws on the books. So the end result could well be fewer irresponsible name choices.)

Objection 5: Giving people so much control makes it likely they’ll later be less happy with it. If you give them less freedom, they’ll grow attached to their choice and rationalize it more readily.

Response: Entirely true! In general, giving people more freedom lets them select more personalized options, but also makes them more indecisive, anxious, and likely to regret their decision. See Dan Gilbert’s excellent talk on synthetic happiness:

I accept this as a cost, but I think it’s worth it for all the advantages union names confer.

Ultimately, we’ll just have to try them and see how they work. If binding families together in a more free, egalitarian, imaginative, and collaborative way doesn’t end up having as many (foreseen or unforeseen) benefits as one might suspect, then a much simpler, more automated system may turn out to be superior.

If people really just don’t care that much about surnames, then you could, for example, flip a coin to decide whose name gets taken on by everyone else. But my suspicion is that trivializing family bonds in that way isn’t the best solution available. (For instance, the parent who randomly has to change eir name may not be the one in the better position to bear the associated social costs.)

Objection 6: So why not just use a coin flip to decide which surname the children get, but let the parents have completely different names? Or leave the parents’ names intact, but use some arbitrary system to assign surnames to the children? For example, you could give the first child the alphabetically earliest surname of its parents, the second child the second-earliest surname, then keep cycling through.

Response: Coin flips and arbitrary conventions are admirably fair. But they still bear the cost of making the whole process seem meaningless and impersonal. Why not humanize and personalize our naming conventions, if we’ve found a relatively easy and simple way to do so?

I’m also wary of systems that give different surnames to the children, even randomly. First, I don’t want to encourage parents, even a little bit, to choose how many children they have based on an implicit desire to pass on their name, or on an implicit desire to equalize the distribution of names, or what-have-you. People’s decision-making is capricious and destructive enough without society going out of its way to distract them with shiny gold Name coins.

Second, I don’t want to factionalize families. These proposals all have the disadvantage of frequently leaving one family member excluded from an important symbolic tie that binds the rest of the family together. Compared to other systems, unity names are just what they sound like. They encourage familial unity more than any alternative does. They create a symbol that ties everyone in the group together, with no one left out in the cold, favored over the rest, or cut off into separate tribes; and they do so without any reliance on pointless infighting or dominance hierarchies.

My own parents went with: ‘The kids take on the father’s surname, but the mother’s name stays unchanged.’ In some ways that’s progress, but it’s still sexist and awkward. It means my mother’s forever cut off from the rest of the family. It means we can’t all rally together under one banner, lest we incur dissonance. It’s a small thing, but some small things matter.

Objection 7: Your system requires partners to come to an agreement on challenging, highly personal issues with many degrees of freedom. That’s a recipe for disaster.

Response: It’s true that union names demand some maturity and willingness to compromise in order to work. I don’t think that’s a bad thing. The alternative is to make our naming conventions unequal (so one person gets final say) or arbitrary (so nobody gets final say).

That said, if two partners are completely unable to agree on a single name, they can still fall back on creating a union name that’s a hyphenated version of their two top choices. This may not be ideal, but it’s one of a variety of compromises the system allows. And since it gets replaced by the next generation’s union name (rather than merged with it), it doesn’t run into the problem of accumulating more and more names over time, and doesn’t become unmanageably large.

Objection 8: What about single parents?

Response: For simplicity’s sake, let’s assume a parent who has never been in any unions. (Though if ey has, that doesn’t solve this problem; you probably don’t want to falsely suggest that your child is in the same family as an unrelated ex of yours.) So the parent’s name is A B.

The simplest answer would be to just name the child C B, like most English speakers do today. But that will introduce confusion, because — assuming siblings are more common than single parents in this union-name-using community — people will initially think that A B and C B are siblings, rather than parent and child.

So I recommend sticking with the union system, and having the parent make up a new name D, change eir name to A D B, and name the child C D.

This has the advantage of allowing you to later ‘adopt’ a spouse Y Z into the same union — say, if you marry someone when the kid is still very young who ends up acting as the child’s caregiver. That new spouse will then take on the middle name D, becoming Y D Z.

If you tried to ‘adopt’ someone into you and your child’s family name without constructing a new union name, then you’d end up having to either: (a) look silly by doubling your own name and becoming A B B to match your spouse Y B Z; (b) look like your spouse’s child by remaining A B and having your spouse become Y B Z; or (c) have your spouse completely change surnames to Y B or Y Z B, which abandons the union name system and all its special advantages.

Just sticking to the union system in all cases seems easier, once it’s well-established. A family with one parent is just as real a family as any other, and deserves just as much to be commemorated with whatever rituals a society uses to honor familial ties.

Objection 9: Your system doesn’t allow traditionalists to pass on the torch of their name with any staying power. All trace of our names will be erased within two generations. That means that legacy names like ‘John Jones VII’ aren’t just discouraged; they’re impossible.

Response: This is true, but I’m not sure it matters very much. Names should be first and foremost about the individuals named. If those names refer to some historical event or lineage, that should be because the lineage is of unusual personal significance to the individual, not because the individual has been pressured into conforming to an arbitrary tradition. It’s a good thing if union names encourage people to construct their own identities as they build their deepest personal bonds and carry out the project of their lives, rather than encouraging people to base their identities primarily on the echoes and expectations of distant ancestors.

That said, union names don’t forbid ancestral naming traditions. If you really want to preserve your name across two or more generations, you can use an alternating system: Sam Boutros Ghali can beget Uma Ghali Boutros, who begets Shashi Boutros Ghali…. You’d just need to start families with people willing to take on one of your traditional names.

As for the impossibility of giving your child your exact name under this system… that’s definitely a feature, not a bug. Union names are a relatively poor choice if domineering creepiness or ambiguity are the things you want from your naming system.

Objection 10: But doesn’t that just reintroduce the problem of one partner getting to impose eir will on the other?

Response: Yes. This will be possible on any nonrandom system. Selfishness and inequality happen in relationships. Union names don’t make it impossible for partners to pressure each other into things they don’t want to do. Union names just make inegalitarian solutions unnecessary, and make the products of name negotiations more interesting and meaningful.

Objection 11: How do we tell ordinary middle names apart from union names?

Response: Well, we could stop giving children middle names so much. If we unambiguously use them only for unions, then we have a very convenient way of knowing people’s relationship history at a glance. Perhaps most people will be satisfied expressing their naming ideas through union names themselves, shrinking the desire for other bonus names.

Then again, maybe some ambiguity is good. Middle names add noise that creates a bit more privacy for people.

Another solution is to have an optional convention for marking the transition from personal names to union / sur-names in one’s full name. For instance, although this wouldn’t be required, you could inject the word ‘of’ before the first union name, if you really want to be clear about your name’s meaning. If personal middle names ever die out, though, this convention should die with it.

Objection 12: Union names make people think of their identities as tied to their partners’ and children’s identities. That’s unhealthy and/or unrealistic.

Response: I disagree. Our identities are tied to our loved ones. They shape our experiences, and draw out of us a specific persona. Both of those factors affect our personality on a deep level. It’s healthy to have some space from one’s family, but it’s also healthy to recognize how indebted we are to our friends, family, and community for who we are.

Hiding from your environment is not rediscovering what’s Authentically You; it’s refusing to acknowledge the part of the Authentic You that’s ineradicably bound up in the outside world.

Objection 13: Union names give parents total control over their children’s names, and very little control over their own names. The reverse makes far more sense. Children should pick their names as a rite of passage, reinforcing their autonomy and self-determination and discouraging parents from thinking of their children as possessions or works of art.

Response: This is a good objection! I do worry about all naming systems that simply impose the parents’ will on the next generation. Children should have a say in their identity — by default, not just if they go out of their way to buck social pressure. But they also need to be called something before they’re old enough to self-name. Some sort of compromise is needed.

My personal suggestion is to encourage children to legally change their first name when they reach a certain age. If this coming-of-age ritual generally leaves the surname intact, then it will remain consistent with the union name system.

I’ll keep expanding the above list as people keep having new ideas!