Laura “Namerology” Wattenberg writes:
Rankings offer no sense of scale. There is always a #1 name, even if that name is only a fraction as popular as #1s of past eras. That’s a problem when scale itself—the changing level of consensus—is part of the story.
In the past, popularity charts served as a solid snapshot of everyday name style, especially for boys. America’s top 10 names accounted for about a third of all boys born. But the popularity curve has flattened. Today’s top 10 accounts just one-fourteenth of American boys.
Here’s how we put it in Regression and Other Stories:
Wattenberg summarizes:
America’s top-10 list of popular baby names always looks comforting. Or, depending on your perspective, boring. The names change slowly, and their style remains mostly traditional. The overall impression is of gradual evolution and cultural continuity. And it’s a lie.
Change in 21st-century naming has been anything but gradual. It’s been revolutionary, a splintering of style that breaks with the past in dramatic ways. But the top of the popularity charts cannot tell this story. A top-10 list is a perfect instrument to create the illusion of stability in the midst of a cyclone. . . .
By consistently reporting name data in scaleless rankings, news reports obscure the fact that the top 10 is losing relevance as a portrait of how we name children.
Last year, an incredible 31,538 different names registered in U.S. baby name stats. Yet 8% of babies received a name that was too rare to even register in the count—more than received a top-10 name. That’s the real story of American names today. But as long as we continue to focus on the top of the popularity charts, we’ll continue to see only the most traditional sliver of our increasingly freewheeling name culture.
Good point, as sociology, statistics, and statistical communication.
My understanding of US naming rules is that there aren’t any, so effectively the underlying probability space is infinite (but of course countable), as pretty much any random string will do.
So, in a sense, this is to be expected as new names are practically costless to generate.
These statistics would be more interesting for countries where you have to pick a name from a (potentially long, but usually below 5-10k items) list, and introducing new ones is costly (you have to convince some representative of the state, or eventually a judge).
Isn’t that the point of this post? It appears that the tendency to invent new permutations for names has been increasing. Despite the large range of potential choices/inventions, past decades seem to be more conformist. I think this mirrors many changes in lifestyle choices – media choices come to mind, where television and reading choices seem to have become more diverse over time. Part of this is due to technological change (perhaps much of it, though it isn’t always easy to distinguish between cause and effect). But there does seem to be a willingness to part with tradition and stand out more than in the past (though it makes me wonder why our politics continues to be dominated by two parties).
Hard to say what the cause is without a model and more data. You could get a decline in the top 10% share without any new names being invented/added.
Even with less than 500-1000 names to go around, one could find it equally puzzling why the top 10 names would capture anything in the double digits, and the large overlap between popular names decades apart. Why does the data have so much hysteresis, yet at the same time so much variation?
Tamás,
The U.S. didn’t have naming rules 50 or 100 or 150 years ago either so your “this is to be expected” statement does not explain the trend. Also, sure, nobody’s stunned by this trend, but . . . look at the different lines for the boys and girls! Something’s going on here.
And the names are not “pretty much any random string.” Consider, for example, the dramatic increase in boys’ names ending in “n,” which Wattenberg pointed out several years ago.
When enough children are the children of Elon Musk (X Æ A-12) you may have to change your mind about that random string thing.
And thus Arrow’s Theorem was born
But the US population has doubled since 1970 or so, and a lot of that growth is immigrants, who bring different popular names with them. Us boring white folks with our boring names ain’t a minority yet, but we will be…
And that would also explain the boys vs. girls lines getting closer to one another.
This, as well as other posts from the website, suggest another driver of diversity in naming: inventing never-before-used names. For example there has been an explosion of girls’ names ending in “-ee”. Or increasing appearance of weird words (i.e. those previously not thought as suitable) chosen for boys, like “Dominance”, “Furious”, “Savage”, “Mayhem”, or concoctions with “xx” (like “Brixx” and “Maxximus”)…