The most thorough way to name a child (twice!)

Warning: This post makes some audacious claims that I can’t actually back up with hard data. Still, I know in my heart they’re true. Audacious Claim™ #1: my wife and I conducted the most exhaustive baby-naming process in human history, and we did it twice.

For our first son, we evaluated 18,353 names. For our second we went even bigger, reviewing 19,494 names.

Sure, there are baby-naming books with more entries, but does anyone actually read them cover to cover?

We didn’t just read each name quietly. I said every single name out loud so we could hear how it sounded before moving on to the next thousand on our list.

You probably have so many questions. Questions like Why would you possibly do that? or Was your wife okay with this?

We’ll get to that, but let’s start at the beginning.

Like many great stories, our journey begins in Iceland.

Inception

For me, the most interesting part of Icelandic culture is that they have a set list of approved baby names mandated by the Icelandic Naming Committee. This committee reviews new name proposals and even hears petitions for exceptions. That might sound controversial, but there are a few reasons behind it.

The first is that Icelandic surnames are patronymic, meaning children’s last names come from their fathers’ first names. If a father’s name is Stefán, his son’s last name would be Stefánsson, and his daughter’s last name would be Stefánsdóttir. Since 2019, Iceland also supports a non-binary option Stefánsbur, which means Stefán’s child.

Because of this system, the name a parent chooses for their child is passed down for multiple generations, potentially having an outsized impact on the overall sound of Icelandic names. The Icelandic government sees it as their responsibility to protect against names far outside the cultural norms. Imagine a non-Icelandic name like Bluey; they could end up with a generation of Blueyssons and Blueydóttirs thirty years from now.

The second reason for the approved name list is that the Icelandic language uses a system called declension for nouns, meaning they change their form based on rules I am absolutely not equipped to explain. The gist seems to be that first names must be structured a certain way in order to fit grammatically into Icelandic. If a child is given a name that literally cannot fit into the language, the concern is that the language could begin to die out. This is especially concerning because a majority of Icelanders speak multiple languages.

Iceland is a remarkably small country: only 360,000 citizens. Wyoming, the least-populated US state, has roughly 50% more people. With such a small population, it wouldn’t take many generations of neglect for the language to disappear altogether.

So that’s our quick foray into Iceland’s naming policies. You can judge it however you want: you might think it’s a noble preservation of culture, or a backward and close-minded stifling of individuality. Ultimately, it doesn’t matter. This isn’t about Iceland. It’s about me.

And what I took away from these policies was curiosity. Specifically, curiosity about what a system like this could look like in the US. I was interested in how something like this could work in for a much larger population with a much less homogeneous cultural makeup.

Now, the Icelandic system has 4,699 approved names. How many names would we need for something like that in America? Luckily, the past 145 years of US naming data is free and available to anyone online. The Social Security Administration keeps records of name frequency back to 1880. You can download these text files for free on their website and dig into any year you want.

Here are the most common boys and girls names for a random year in the 1980s, along with the number of babies given each name that year. This was the mother lode, so I knew I wanted to take a closer look.

I didn’t realize it at the time, but it was also the first step in what would ultimately become my patented five-phase naming process.

Phase 1: Data Cleanup

I was aiming for breadth rather than completeness, so I took a shortcut while exploring these files. I downloaded data for each decade (1880, 1890, 1900, etc.), then wrote a Python script to combine them into a single, workable Excel file.

After deduplicating the results, I was left with sixty-one thousand eight hundred sixty-seven names. The list below shows the most popular names from my dataset across any year I looked at. There may have been more Davids or Michaels over the past 150 years, but the most babies ever named a single name in any year was James.

From there, I went back to the Iceland naming idea. Based on frequency data, how many names would we hypothetically need for a fictional “approved name list”?

So I thought, “What if we only allow names given to over 1,000 babies in any single year?” That gives us fewer than 1,500 names; only a third of the Iceland list. Dropping down to names with over 200 babies brought me closer to the Iceland number, but it still didn’t seem like quite enough.

Ultimately, I landed on names with 20 or more occurrences in any year from my dataset. Given how obscure some of the names were below this threshold, it felt like a reasonable balance between pragmatism and completeness. And because this whole exercise was moderately pointless, it seemed like a good stopping point. I thought that was the end of the story. I put all this away, expecting never to revisit it.

But then my wife and I started talking about baby names, and I remembered that list. Now, 20,000 is a large number. But it’s not an impossibly large number. That gave me an idea.

What if we took my list of names, randomized the order, and used that as the starting point for naming our baby?

Phase 2: Judgement

The goal for the first round of evaluation was to eliminate names we had zero interest in. But before that, we needed to decide what mattered most to us when it comes to names. We didn’t exactly know what we were looking for, but we had a pretty clear idea of what we didn’t want.

First off, we didn’t want a name that was too common. Any of the Davids or Christophers or Michaels in the audience can likely tell you about their experience growing up with a slew of similarly named peers. Those are all great names, but our hope was that our kids could have the option to use either their first or last names as they wish, as opposed to having “Michael M.” or “Chris H.” thrust upon them for life.

Now that I spend a lot of time at local playgrounds, I’ve personally met nearly 500 toddlers named Jack. I feel for them, and for the Liams and Noahs and Olivias and Emmas out there with ultra-popular names.

On the flip side, we didn’t want something too obscure either. We wanted to give our kids names that were recognizable and familiar, names they could find on keychains and coffee mugs on future family vacations.

Finally, I was concerned about the efficiency of whatever name we chose. This is my second unsubstantiated but absolutely true Audacious Claim™: billions of dollars in lost productivity each year can be traced to the inconsistent spelling and pronunciation of names.

Let me unpack that for you. If your name is Megyn with a Y, every time you give your name or email address to a doctor’s office, a dentist, or a plumber, you have to stop and spell it out. If your name is spelled like Anna but pronounced Ahh-na, every time someone gets it wrong is another moment when you have to correct them. All these little exchanges and the myriad clerical errors they cause add up to tens of thousands of human hours spent dealing with the idiosyncrasies of unconventional names each year.

I don’t begrudge people who have these names. Your name is deeply personal, and it makes sense to want people to get it right. But when it comes to my kids, I didn’t want to set them up to deal with those little frustrations, especially if it meant correcting adults on their names all the time. So one of our guiding principles was that you can read a name and know how to say it, and hear it and know how to spell it.

Once we had alignment on what was important to us, we got into the fun part: evaluating every name! We sat on the couch and read through the list, saying each name out loud one by one. This might sound slow or painful, but it’s actually pretty easy to skip past a lot of names. Once we’d seen a few hundred names, we identified a few obvious themes that we could quickly pass over.

Here are just a few examples of names that were easy to eliminate for our kids:

Cultural Fit

There are lots of beautiful names from around the world, but my wife and I have no familial or cultural connection to most of them.

Nontraditional Spellings

Next up were spellings that didn’t pass the efficiency test. To be honest, I’m not sure which of these were cultural variations, which were deliberate changes made by parents, and which could, in fact, be typos. Every one of these variations was given to at least 20 people in a given year. Some were much higher than that.

Deep Cuts

We came across quite a few names we didn’t really recognize at all. Some of these are certainly from cultures we’re less familiar with, while others might be more recent inventions.

Virtues

There are way more of these than you might think, and my wife hated all of them.

Pop Culture

These are nontraditional names that spike shortly after a book, TV show, or movie becomes popular. We didn’t want to tie our kids to a pop culture reference that might not age well by the time they reach high school.

Zeitgeist

These are names that aren’t necessarily tied to pop culture, but appear and quickly gain notoriety quickly based on their novelty. Sometimes, news reports about how strange they are end up as cultural superspreader events, allowing them to reach even wider audiences.

Mix & Match

Finally, there’s a popular genre of names that didn’t do much for us. The general idea is that you can mix and match any of dozens of first syllables with dozens of second syllables to create a trendy new name.

The Data

Of course, I couldn’t resist collecting a little meta-data on our own naming process.

I kept detailed records of when we evaluated each name so we could review our progress over time.

Here’s how many names we evaluated each day. It’s interesting to notice how flipped the patterns are. The first time around, we were excited about a fun new process, whereas the second time we struggled to build momentum and powered through at the end.

My wife was quite sick during both pregnancies, and you can identify those gaps here in red where I was not going to make her sit and listen to a bunch of names.

Finally, we tracked things like boys’ vs. girls’ names and the number of syllables in each name on our Maybe list.

Phase 3: Evaluation

At this point, we’d spent more than half a year each time evaluating the list of names, but we ended up with a shortlist that we could whittle down. It’s a subjective process, but I still wanted to add as much structure as we could.

I randomized the shortlist of names three times, and set up a 5-point rating scale. My wife and I each evaluated the names without looking at each other’s scores or our previous ones. I hoped that reviewing the names multiple times would help reduce any wild swings in opinion we might have from day to day.

After the third round, I converted the scores to numbers and totaled them. From there, we were able to see which names we both liked, which names we both hated, and where our opinions were misaligned. The second time around, there were a lot of 0s because my score and my wife’s score canceled each other out. We picked an arbitrary threshold of 4 points or higher, since that meant my wife and I both liked the name at least once. Those names moved on while the rest were discarded.

Phase 4: Review

The first time around, we ended up with 21 names at this point. The second time, just 11.

From there, we turned to Wolfram Alpha to check the current rank, most common age, and popularity over time for each of our finalist names. This was especially helpful the first time around, when we eliminated a third of the options that were more popular than anticipated. Wolfram Alpha has some free APIs you can pair with Google App Script to easily populate charts like this in Google Docs, which is how I built this phase of the project.

Phase 5: Fight About Names (There is No Magic Solution)

After all those months of tedious process, the journey ended the same way most parents pick a name. We just talked about it until we got to something we both liked. We weighed things like initials, nicknames, and middle names until we landed on front-runner names for both a boy and a girl.

Frequently Asked Questions

Q. Why didn’t you tell us the names?!

A. Every time I’ve told people the actual names, they end up disappointed at how normal and relatively unremarkable they are.

Q. Why didn’t you just use the Maybe names from the first time around for your second kid?Every time I’ve told people the actual names we picked, they end up disappointed at how normal and relatively unremarkable they are.

A. The truth is, tastes change. Of the 82 names we said Maybe to the second time, only 37 were on the first list. Names we liked the first time around became much more popular, or we met a bratty kid at the playground with that name and it ruined it for us.

Q. How did you pick middle names?

A. The answer isn’t very satisfying. We just tried to find things that fit nicely with the first names we liked.

Q. Would you do this process for a third child?

A. Would I? Yes. Will I? No. We’re all set with two kids, and I doubt she would agree to this process if we ever get a dog.

Q. Would you recommend this process to a friend or colleague?

A. Of course not. This was a ridiculous way to spend a few dozen hours, but at least it makes for a good story.

Q. Is {{name}} on your list?

A. Almost any name you’ve ever heard of (or imagined) is likely on the list. I recently revisited this data set for some recreational SQL exploration, so I’ll be sharing a follow-up post about that project soon.