Getting the Names Right

If you’re a fiction writer, it’s important to choose your names wisely. Anna Karenina works well; Vladlena Ovchinnikova might have been a harder sell. Joe Gargery, Miss Havisham, Estella, Willy Loman. Prospero/a, Rosencranz, Guildenstern, Portnoy, Yossarian, Galadriel, Tywin Lannister, Don Draper. Medea. The right name is magic, the almost-right name is a cap pistol.

No doubt every fiction writer has her or his tricks and recourses, some perhaps as closely guarded as favorite truffle-hunting spots. I’ll share a resource that I ran across recently, tipped off by a friend: the Social Security Administration’s lists of top names for newborns, by year.  The SSA publishes a subset of the data in table form, but fuller data files are also available for download. I enjoy a good data crawl, so I grabbed all the files and sucked them into a data analysis tool I banged together (using FileMaker Pro, if you must know).

The data files run from 1880 to 2014, and they include counts of all babies with a given name in year X, with the caveat that the count for any name must be five (5) or above. A name given to four or fewer children in a year is not recorded, for purposes of anonymity and perhaps also data set size.

The data set is pretty big as it is: about 104,000 unique names, recorded over about 1.8 million “name occurrences” (instances of a name in a given year).

This information is naturally most useful to writers who want to write about America 1880-2014. But that covers a pretty good swath. In my case I began with a simepl question. I had a character born in 1941, whom I wanted to name “Royal”. Was “Royal” a name given to any children in 1941? Why, yes it was. In fact it was given to 108 children in that year. So no anachronism there.

So how popular a name is the name Royal over the entire covered period? Here’s a chart:

Royal

This shows us the number of occurrences of Royal as a male name, by year, normalized to account for changes in population. (The numbers in the chart are normalized to occurrences per billion, if you’re curious.) So “Royal” was very popular in the teens and twenties, went into a long decline, and just very recently is on a huge upswing. The female Royal shows a similar trend. It would be interesting to know what might be driving the upswing. Honestly, something in popular culture is likely to play a role. As an example, let’s look at Hermione:

Hermione

That one pretty well speaks for itself. Popular in 1915, then very spotty until, well, the Harry Potter era.

(What’s with 1915, anyway? We’ll come back to that).

I’m sure no one doubts the celebrity effect, but here are two more examples, just for fun:

Barbra

Britney

Barbra, no middle “a”, was popular in the 50s, but enjoyed a big runup during the early to mid-60s when Barbra Streisand was winning multiple Grammies, peaking in 1965, a year she was nominated again for the Grammy but didn’t win, and began to move into movies. Britney is a relatively new name, late 20th century on, and shows clear trends in popularity as well.

The Rise of Variety, the Fall of the Old Guard

So what are the most popular names overall? The SSA’s table shows us this: names like Mary, Patricia James, John, Jennifer, Robert. But what about trends? These are the overall “winners” (feel free to shudder at the cultural-hegemonic overtones — fundamentals even …), but are they still as popular as they used to be?

Actually, no. Here are James and Mary, the most popular male and female names, respectively, over the life of the whole data set:

James Mary

James peaked around 1947, Mary back in the early 20s, and both have been in long-decline ever since. These, again, are declines as a percentage of population.

The trend is true of all of the names at the “top” of the popularity list. In fact, you have to drill all the way down to #75, Samuel, before you find one of these “all-timers” that is not at the foot of a steep decline.

Samuel

Even Samuel has tailed off a bit very recently, but it’s strikingly different from the 74 above it in that, in the early 21st century, it was in fact more popular than ever.

After Samuel, the rhythm of decline returns. At the same time, we see popularity becoming more concentrated, peaking and then tailing off again more quickly:

KathleenCynthia GregoryThen another Samuel pops out, showing continued strength, only this time it’s Benjamin:

BenjaminIs there some significance to the fact that Samuel and Benjamin are both apparently Biblical? Maybe. But after Benjamin, we’re back to the compact, apparently trend-driven pattern again, with the very spiky Shirley (guess which celebrity drove that one?) and the beautifully symmetrical Amy:

ShirleyAmy

Down, down down.

How Can Everything Be in Decline?

The population keeps growing, right? We have more people every year. So how can the number of all these popular names be declining? Doesn’t something need to be rising?

Well, maybe. But remember that these are not absolute numbers, but ratios of the whole population. (Most name charts look quite similar if you graph the absolute number of names, but it’s important to normalize). So if the population is growing, but the “market share” of many of the top names is falling, what’s going on?

Well, one possibility is market fragmentation: if the absolute number of names in circulation grows, the market share of the average name drops. Instead of letting our eyes glaze over looking at one name after another, let’s ask about the total number of circulating names:

TotalByYear

Well, that’s a bit of a no-brainer. The 1880 data set has 2000 distinct names, while the 21st-century sets have over 30,000. (Interestingly, after seven decades of steady growth, there’s a slight but discernible retreat in uniqueness in the last 6 years or so.)

But is the really the only trend at work? Is the universe of names simply becoming more fragmented? Is there any name or group of names that’s getting “bigger”? And what is really happening between say 1923 and 1940, when the total number of unique names actually shrinks? Doesn’t that mean names are dying out — going extinct, if you like? And if old names are dying, don’t new ones need to replace them?

The Death and Life of Names

With the SSA data, it’s easy to see a “lifespan” for each name: where it first appears in the data set, and where it last appears. For any year, we can see how many names were “born” in that year, and how many “died”.

Here’s the data on extinctions. It has to be taken with a large grain of salt. For example, since 2014 is the last year in the data set, any name still “alive” at that time will appear to go “extinct”. Data for the 5-10 years before 2014 is also a bit suspect. Some names are spotty, and don’t appear every year. A name that appears in 2013 but not in 2014 may well be back in 2015.

ExtinctByYear

Even with these caveats, it does seem like there are certain real trends. Starting around 1915, names start “dying” in much larger proportions than before. The rate of extinction tails off into the late 40s, but then begins a slow and steady trend upward. This comports with some of the more compact charts we started to see above where names have shorter lifespans. Probably the high extinction numbers from 2005 on will come down over the next few years as the sporadic names sputter in and out, extending their lifespans. We’ll see.

How about births? Here’s the chart of names that are new in a given year.

NewInYear

That is quite something. The very earliest data are suspect for just the same reason that the later extinction data are. The name Hertha appears for the first time in our data set in 1886, which is only to say it didn’t appear with at least 5 usages in 1885. Maybe there were only 4 Herthas in ’85, or maybe there were none at all, but 11 in 1884.

That said, the sudden rise in new names in the early 20th century, culminating in a giant spike in 1915, seems pretty unmistakable. This trend is just slightly ahead of the rise in extinctions in the same period. And both trends taper off to a period of maximum stability about 1935. And both then increase from there, suggesting that name turnover has continued to increase — at least until quite recently. In the last 5-6 years we can see a significant reduction in the rate of new names, perhaps paralleling the aforementioned dropoff in total distinct names in the pool.

Names that Died

Other and voiceless are there shadows there
Still in the little light as winter crickets
Torpid with old death …

Some names are not just dead, but long dead. Here’s a list of some of the names that went extinct in 1950. These are pretty reliable extinctions: these names haven’t been seen since, at least not in numbers over 5 per year.

Bula (F), Alverda (F), Eino (M), Fronie (F), Selmer (M), Orma (F), Leocadia (F), Wauneta (F), Fernande (F), Lockie (F), Lulla (F), Foy (F), Vergia (F), Lattie (F), Elner (F), Vinie (F), Earnie (F), Dorothye (F), Marinus (M)

Overall, the “biggest” name ever to go extinct was Flossie (F), d. 1987, followed by Gay (F) d. 1982 and Bertie (F) d. 1986. The biggest male name ever to die out was Gail (M), d. 1989, followed by Carol (M), also d. 1989 and Verne (M), d. 1990. Gayle (M) also finished up in 1989. Seems like the late 80s were hard on female-sounding male names. Most of these died after a long decline. Pearl was a popular male name in the late 19th century. Dorothy was also a popular male name, peaking around 1935 before dying altogether in 1989. The female name Gay was very popular in the mid-20th century but declined rapidly from the 60s on.

GailM GayF

It’s certainly just speculative, but perhaps American became less and less tolerant of gender-ambiguous names in the later 20th century.

What Happened in 1915?

1915, as we saw above, showed a gigantic proportion of new names relative to other years. Most of these names didn’t burst onto the stage with a gigantic showing right away in 1915, but built up slowly over time. In fact, some of the ones that made the biggest splash in 1915 are not so familiar anymore:

Wanza (F), Orel (F), Audra (F),  Laberta (F), Augusto (M), Hassell (M), Montague (M), Czeslawa (F), Hanako (F), Arlee (M), Faris (M), Shermon (M)

On the other hand, quite a number of the 1915 names had huge staying power. Many sound so modern it’s surprising to learn they debuted before the first World War:

Michelle (F),  Randy (M), Trevor (M), Stacey (F), Jenna (F), Kristina (F), Paige (F), Bethany (F), Karla (F), Greg (M), Cesar (M), Joey (M), Conner (M), Shelia (F), Dakota (F), Angelique (F), Zander (M)

Right now I don’t even have a theory about the sudden explosion of novelty around 1915.

It’s Not All Fragmentation

Earlier I suggested that the reason so many names were in free fall was because of overall fragmentation. But that’s not the whole story. Some names are indeed growing. Some of these are new, some of them older, some of them very old indeed.

I mined the data for names where the usage has increased in each of the last two decades. Then I selected the top fifty male and female names in this group, ranked by their recent usage.

fifty-females2

For the females at least, there’s a clear trend. Names that were popular in the Twenties, then fell into a long trough, are now being revived. How do the males look? Similar, actually.

fifty-males2

So, many of these “trendy” names aren’t really new. As we saw in the 1915 explosion, some names that feel new to us have been around for a while.

An interesting subgroup of this set are the Biblical names. Common Biblically-derived names like Michael and David are in the same decline as a lot of the old guard. But less common ones are discernibly on the rise. Here’s an unscientific sampling.

biblical-names-rising

We could continue this analysis at almost any length. I have some more trends I’d like to explore, but I’ll leave them for future chapters.

See Also

There’s no shortage of web sites that explore this same data set, some with spiffy interactive charts, data mining tools and more. The spiffiest I’ve seen so far is nametrends.net. A similar site is babynamestats.com.

Some interesting discussion in this Huffington Post article.

Of course there is also the SSA source website itself.