Demographic trends and ngrams

Do words follow families?

I previously used Google ngrams to identify the arrival of terms such as “parenting” and “sibling rivalry.” And I took a shot at tracking family relation words in relation to family structure in histoty. But what about specific demographic trends that have captured the public’s attention and sparked debate? Here are two: cohabitation and divorce.


Cohabitation existed as a concept in law and culture for a long time before it appeared as a common household structure – a man and a woman living in what the Census Bureau used to call a “close personal relationship” (by which they didn’t mean any old close personal relationship). Looking at some of the old Google Books uses, it’s clear they are referring to men and women living together unmarried.

Here are the ngram for cohabitation (above) and the percent of U.S. households that include cohabitors. I’ve scaled them so the horizontal gridlines line up with the decades from 1960 to 2010.

You can’t see the scales, but they are similar up to 1980. That is, next to nothing till 1970, then a doubling to 1980.  But after that the demographic trend continued upward while the language trend plateaued.


Divorce appears to be something like a social panic, with the hype not often matching the facts, except in the most general sense that there is more divorce now than in the old days. In fact, divorce rates have had their ups and downs, as you can see below. Again, I’ve lined up the gridlines.

Here it appears that the word database doesn’t pick up the post-WWII spike in divorces. But the run-up in the 1970s is well represented. Then it took about 15 years for the gradual decline in divorce rates to be reflected in the word database. That’s reasonable, since this crude divorce rate is not quite reflected in the popular visibility of divorce (for example, the aging of the population will tend to reduce the crude divorce rate, as still-married people live on and on, adding to the denominator of the rate).

Anyway, I’m satisfied to conclude tentatively that ngrams trends may follow (or even drive) demographic trends, and I’m interested in possibility that the disparities in the timing of fluctuations might be useful.

4 thoughts on “Demographic trends and ngrams

Comments welcome (may be moderated)

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s