A Washington Post column by Chris Cillizza shows and discusses a remarkable chart that “details how the 17 most common non-English languages in 1980 have fared over the past 30 years” in the US; Cillizza summarizes it thus:
In 1980, the five most common non-English languages spoken in the United States were (in order): Spanish, Italian, German, French and Polish. Thirty years later, the top five are (in order): Spanish, Chinese, French, Tagalog and Vietnamese.
He points out that “Yiddish, which was the 11th most common non-English language in the U.S. in 1980, has fallen to last over the last two decades” and “Russian … started at 14th in 1980 but has soared to the eighth most common language in 2010,” and has other interesting observations. Thanks, Kobi!
Looking at the underlying data, I think “French” scores so high only because of lumping together the categories “French” and “French Creole.” The distinction is in any case tricky because the data is probably coming almost entirely from self-identification/self-reporting and you could ask two Haitian-American immigrants whose language use is identical to each other what non-English language they speak at home and one might label it French and the other Kreyol. And of course “Chinese” scores so high only by lumping Mandarin, Cantonese, Taiwanese, etc etc. into one undifferentiated mass.
It’s worth clarifying that this refers to languages other than English spoken “at home”.
“People who knew languages other than English but did not use them at home, who only used them elsewhere, or whose usage was limited to a few expressions or slang were excluded.”
So it doesn’t refer to languages learned in school, for example. The reason I wondered about this is because of the prominence of French.
Although note (if you click around until you find the Census Bureau report that the chart is derived from) that the percentage of “French” speakers who also spoke English “extremely well” is very high, so it’s likely not predominantly recent immigrants. There’s a bit of a self-reporting issue about exactly how much of the non-English language you would need to speak at home to answer yes. For some groups (e.g. for French Cajuns, descendents of Quebecois immigrants to New England) one might hypothesize that they would be embarrassed if they said “nah, we never ever speak French at home anymore – that was Grandma’s thing” and thus are inclined to say yes even if the non-English language makes up a fairly low percentage of total intra-household utterances.
It’s pretty simple to interpretation. Falling curves are old people dying, rising curves are immigration. Only Armenian, Portuguese and French are more or less stable.
http://www.census.gov/prod/2013pubs/acs-22.pdf appears to be the source of the underlying data (in its table 2). Note that in pure percentage terms, Yiddish has had only the second-greatest drop from 1980 to 2010 – slightly less steep than the drop in Italian-speakers (a significantly larger group in absolute terms in both years). Because of the massive growth in aggregate speaking-something-other-than-English headcount, those “stable” languages like Armenian etc had to grow quite a lot to retain their relative position.
The only two languages that don’t have a completely one-way path (of either increase or decline in absolute headcount) are the French/Creole combination (up through 2000 than mildly down, presumably as new immigration fails to fully replace old people dying off) and The Former Yugoslav Language of Serbo-Croatian (down significantly from ’80 to ’90 as the oldsters from a previous wave of immigration some decades previous pass on, but then up sharply from ’90 to ’00 and again the following decade as a brand-new wave of immigration starts).
It would have been nice had they given absolute numbers along with relative ranking. It would throw some light on how steady state the state of the given language is.
And only thirty years of data! Interesting, of course, but it does make one hunger for more. Consider the status of Texas Germanover the course of the twentieth century
BWA: there are absolute numbers in the Census document I linked to in an earlier comment. The Census document I think also has an explanation of how the form of questions asked has varied historically from census to census, which may mean that the data they do have from earlier censuses is not directly comparable in an apples-to-apples way.
There was a serious undercount of Yiddish-speaking communities in 2010. It seems possible that if that had not occurred Yiddish would actually be shown to have increased since 2000. I can attest, however, that there was a radical drop-off in the number of Yiddish speakers born in prewar Europe in that decade, a drop-off that will, of course, continue until there are none left (this is also true of English speakers born in the prewar United States).
@Ben:
“There was a serious undercount of Yiddish-speaking communities in 2010.”
Why do you say that?
Though the number of Yiddish speakers is still declining, the Hasidic communities are still raising children in large numbers who speak Yiddish fluently, so at some point (barring unforeseen changes) the number will bottom out and start increasing.
FWIW, these questions were not asked on the Census proper in 2010, but the 2010 and 2011 numbers are from the ACS, which does not attempt to be universal but does attempt to be a statistically valid cross-sample of the larger population. How well it does, and whether it undersamples Yiddish speakers, are technical/methodological questions I don’t have much insight into.
@dw: I saw a map a year or two ago that I unfortunately can no longer find that showed census undercounts (judged by unreturned forms) by census district in NYC. South Williamsburg and Borough Park stood out starkly.