Androgynous names in the USA

Billet publié le 24/08/2014

Very often, boys have boys’ names and girls, girls’ names. But sometimes, the same name (Leslie, Dana, Sammie, Alva, Lou…) is given to boys and to girls. Those “androgynous” or epicene names are interesting : most of the time, they are unstable, they begin as male names and end as female names. [See Lieberson, Stanley, Susan Dumais, and Shyon Baumann, ‘The Instability of Androgynous Names: The Symbolic Maintenance of Gender Boundaries’, The American Journal of Sociology, 105 (2000), 1249–87 jstor]

Let’s take Leslie :
At the end of the 19th century, it is given to baby boys more than 9 times out of 10. Around 1950, it is given at the same frequency among boys and girls. But now, male Leslies are much less frequent than female Leslies.
It is difficult to find the opposite evolution, where a female name is masculinized.

Let’s consider that a name is epicene if the babies born year N and receiving this name are girls more than 10% of the times and less than 90%. This definition is restrictive, I consider that some names — such as Leslie now — aren’t really epicene anymore even if they were epicene before (because in 2013, there were fewer than 1 boys for 10 Leslie). This definition focuses on the current use of epicene names.
10/90 are arbitrary boundaries, one could use 1/99 or 30/70 (and it is easy to do, see the R code below).


In 1880, 2% of the babies had an epicene name (and there were very few such names). During most of the 20th century, around 3.5% of babies received an epicene. Since 1960 (or 1980) this proportion is increasing : 8% of the babies born in 2010 received an epicene. And today (dotted line) more than 1500 names are epicene. The consequence of these number : epicene names are “small” names, given to a small number of babies each year.


The real proportion of epicene babies is higher : names given to less than 5 male or female babies are not included in the database, and we lack information about 10% of the babies. And very rare names are more likely to be epicene than common names.

Let’s focus now on the population of babies receiving an epicene. From 1900 until 1950 (black line), more than 50% of epicene babies are male (which means that parents are more often than not giving “male” names to their daughters when they give them an epicene). From 1950 until 1990, the epicene babies are mostly female.
As you can see (dotted red line) there are always more “male” names than “female” names in epicene names [a “male” name is a name given to a higher proportion of male babies than female babies].


There seems to be an interesting evolution of the Gini coefficient. The Gini coefficient is a measure of inequality (most often used to describe inequality of income in a country). Here, it is used to describe the distribution of name frequency.

Notes : I relied on Social Security Administration’s applicants numbers and first name. They are closely related to birth for the current period, but not before the 1930s : I very crudely corrected the skewed sex ratio. I used the ‘babynames’ package for R.
You can download the R code (it is not pretty) : epicene-usa-web.R

I'm a non-native speaker, and English grammar is often confusing. Please help me improve (you can comment, I will correct)