I’ll get to Taylor and Kim, but first more general data.
How gender binary is the practice of naming babies in the U.S.? Very. In 2018, 76% of babies were given names that were more than 99% male or female, according to data from the Social Security Administration (which releases name counts for only two sex categories).
That looks extreme (kurtosis = 1.06!), but 76% is actually the lowest that number has ever been. Here is the trend in babies with >99%-typed names back to 1880 (note the y-axis starts at 70%):
How important are the trends in name binaryness?
In her New York Times article on the rise nonbinary gender identities among young Americans, and a follow-up, Amy Harmon interviewed nonbinary people named Flynn, Keyden, and Charley. (In 2018, 85% of the babies given the name Charley were identified as girls at birth, compared with 0.2% of those named Charles and 52% of those named Charlie — the most androgynous spelling of the three).*
One notable development in the striking rise of nonbinary identities has been the supportiveness of some parents. But are such parents reacting positively to their children’s development, or — not waiting to be prompted — giving their babies more androgynous names at birth? Extreme sex-dominance of names has become less common, but still dominants. And truly androgynous names, say, between 40% and 60% associated with one sex, are very rare.
Over the long run, the U.S. is becoming a less sex-binary society, but that evolution is far from direct. From 1950 to 1975 (the period featured in Jo Paoletti’s book on the unisex movement in fashion), the percentage of babies given names that were less than 95% associated with a dominant sex almost doubled, to 7.4%. And since then it has increased to 13%. However, the percentage given names that are between 40% and 60% sex-dominant remains barely over 1%. Here are those trends, back to 1940, using data from the Social Security Administration.
Are the parents giving androgynous names even doing it on purpose? I’m not sure how we can tell. Despite phonetic cues, which are guides but not rules, the gender of a name is ultimately determined by the gender of the people who have it. When names are very rare, it’s likely parents just don’t know the sex of the other babies getting the name. Maybe parents giving the names Charlie, Finley, and Dakota — the most popular androgynous names — chose them because they like their androgynousness. But others, like Justice or Ocean, probably just don’t have stable genders attached to them. And the conventional wisdom (from Stanley Lieberson and colleagues) is that androgynous names are not stable — they either swing toward one gender or fade away.
Here are the most common names between 40% and 60% sex dominant in 2018. Maybe blog readers can say something about the motives of the parents using these.
In that 2000 paper by Lieberson et al., which used data on Whites only from Illinois, through 1989 (how did people ever do sociology with such paltry data available to them?), they reported that the parents of girls are more likely to assign them androgynous names than the parents of boys are. That is consistent with the idea that the penalty for gender non-conformity is greater for boys than for girls, that femaleness is the contaminant more than non-conformity — which is why the move toward gender equality meant women wearing pants more than men wearing dresses. But now that may have reversed. Boys are now more likely to be given names that are less than 95% sex-dominant.
I think this is a good avenue for exploring changes in gender attitudes, including regarding nonbinary identities and gender conformity. This will require looking beyond name count trends, obviously.
Kim and Taylor
Another avenue for research involves name contamination (another Lieberson idea, which Tristan Bridges and I have written about; see also earlier posts). From a wide angle, it’s easy to see that androgynous names usually don’t stay that way, or they disappear. But the specific mechanism may be that parents of boys are spooked by the rising femininity of a name and thus turn away from it.
In that Lieberson et al article they cite the case of Kim, which (among Whites in Illinois) was increasing among both boys and girls before Kim Novak burst on the scene in 1954, as a sexy female movie star. And they also observe the rise of Taylor, just beginning by the end of their dataset, in 1989. Now we can update that, and expand it to the whole country, to see the amazing similarity of the cases. Amazing similarity, that is, if you remember who Taylor Dayne is.
Taylor Dayne was a big deal very briefly, at the end of the 1980s, with three gold singles, “Tell It to My Heart”, “I’ll Always Love You,” and “Love Will Lead You Back.” She was nominated for a Best R&B Vocal Performance Grammy for “I’ll Always Love You,” in 1988 (losing to Aretha Franklin). Did Taylor Dayne kill Taylor — right after giving us Taylor Swift (born 1989)? I’m open to other suggestions, but I think it fits. She was a big star briefly, and the music she made (no offense) didn’t turn out to be the most memorable of the period, which was awkwardly sandwiched between decades. There is a difference in scale between the cases, as Taylor peaked at the #6 most popular girls’ name and the 51st most popular boys’ name in the mid-1990s. Also, Taylor still ranks, and is still 18% male, while Kim virtually disappeared. So maybe the dynamic is a little different now.
Anyway, I love the idea that Taylor Dayne killed Taylor, because she isn’t even a real Taylor — she was born Leslie Wunderman (were any other Jews nominated for R&B vocalist Grammys?), and only chose the name Taylor in 1987, as it was already spiking upward. It also raises an issue relevant to the question of nonbinary-supporting parents: name changes. If gender identities are increasingly fluid, maybe names will be, too. In addition to being less sex-typed, names may also become less permanent. Just a thought.
* In the original version of this post I mistakenly wrote that 20% of Charles’s were girls, it’s actually 0.2% (I read .19 as a proportion instead of a percent).
Data and code for this analysis are on the Open Science Framework here: https://osf.io/m48qc/.