The Third Language.

The Most Common Language In Each US State—Besides English And Spanish: the title is pretty self-explanatory. In Arizona and New Mexico, it’s Navajo, which is unsurprising. In California, Nevada, and Hawaii, it’s Tagalog, which surprised me. In large swaths of the Midwest, it’s German. And in Oregon, it’s Russian! (Thanks, Sven!)

Comments

  1. Here in Minnesota I wonder if Somali will soon pass Hmong; my impression is that the smaller but more recent wave of Somali immigrants contains fewer monolingual English speakers. The city of Minneapolis announces snow emergencies on Facebook in English, Spanish, Hmong, and Somali.

  2. Steve Reilly says:

    Glad I’m not the only one surprised about Tagalog. I would’ve gone for Japanese and Chinese above it, and maybe a few others as well.

    And yeah, so much German? Didn’t realize that.

  3. Are any of those actually the second language rather than the third? That is, more spoken than Spanish in that state?

    As usual DC is missing, but I’m told that the third language for it is Amharic.

  4. Steve Reilly says:
  5. Keith: Quoth the article, the second languages are “Yupik in Alaska, Tagalog in Hawaii, German in North Dakota, and French in Louisiana, New Hampshire, Vermont, and Maine.” In all other cases, the mapped languages are third languages.

  6. Many other maps, plus a warning about how misleading maps can be (there’s a big difference between the biggest band in a U.S. state and the biggest band in a state relative to other states).

  7. OT for this post: Languages of the World returns to life with the Geocurrents posts (sans comments, alas) and a new one or two.

  8. GeorgeW says:

    In Florida, number three is “French Creole.” There are a number of speakers of Haitian Creole. Maybe they lumped all French-based creoles together. I wonder if French Canadian was classified as a “French Creole.”

  9. J.W. Brewer says:

    Obviously the accuracy of this map turns on the accuracy of the underlying data and its subsequent handling, which will often include how lumping/splitting issues are handled either in the framing of the questions or how the respondents answer them. The slate piece claims the source of data is splitterish enough to treat e.g. “Mandarin” and “Cantonese” etc. separately, which might by itself be enough to put Tagalog ahead of undifferentiated “Chinese” in Cal., but then the same map labels the most-common-after-Eng-and-Span language in N.Y. as “Chinese.” So I dunno.

    If you look at the other slate maps, you can see that many of them have pretty random results. The most commonly spoken American Indian language in New Hampshire, for example, is said to be Hopi, which, to put it mildly, is not indigenous to the area. This is presumably the result of very low absolute numbers of all possible contestants combined with imperfect data.

    I’m not sure if the underlying ACS question is open-ended enough to encourage/permit multiple answers when people speak multiple non-English languages “at home” (e.g. Georgian and Russian, which is a combination that sometimes occurs among immigrants in NYC — or perhaps e.g. Tagalog + Cebuano in some Filipino-American households?). If they do accept multiple answers, that creates additional issues with how to aggregate the data since percentages of the different language counts will not add up to 100.

  10. There is a series of “most common language” maps in the original Slate article:

    http://www.slate.com/articles/arts/culturebox/2014/05/language_map_what_s_the_most_popular_language_in_your_state.html

    Some of these results must be due to a very small number of speakers of a particular language but even fewer speakers of other languages in a particular state.

    Come on. Hopi is the most commonly spoken native American language in . . . New Hampshire?

    French is the third most common language in Maryland, Delaware and the Carolinas? Must be kids studying French in high school and college.

    Danish is the most commonly spoken Scandinavian language in West Virginia?

    Also, why are Pashto and Dari excluded from the Indo-Aryan group? In my Northern Virginia neighborhood these language, along with Korean, Amharic and of course Spanish, are widely spoken.

  11. GeorgeW says:

    J.W. Brewer: “Obviously the accuracy of this map turns on the accuracy of the underlying data and its subsequent handling, which will often include how lumping/splitting issues are handled either in the framing of the questions or how the respondents answer them.”

    I would also expect that the map maker would use some semblance of mutual intelligiblity in the lumping and splitting process. Can one say it is the same “language” if the variations lumped together cannot understand each other? I say not.

  12. J.W. Brewer says:

    One issue is just the nature of the data-collection. If you ask the open-ended question and the answer you get is “Chinese,” the protocols may not provide for an automatic follow-up saying “that’s not the right level of generality. Do you mean Mandarin, or . . .” Similarly, two different Haitian-Americans with the same actual linguistic behavior (code-switching between Kreyol and a variety of more standard French, depending on context) may give different answers, with one saying “French” and the other something that will be coded as “French Creole.”

    Bill W: I believe the underlying ACS question is more or less “do you speak a language other than English at home, if so what is it.” So that’s comparatively unlikely to pick up L1 Anglophones of non-Francophone family background who have learned French at school. (Note as to the “Scandinavian” map the rather embarrassing fact that Danish is the winner in Delaware, which was founded as New Sweden way back when — although I think the descendants of those original settlers had all been linguistically assimilated by some point in the 18th century.)

  13. The Russian speakers in Oregon were not too surprising to me (but then, I grew up in Oregon). There is a substantial community of Old Believers in Woodburn, Oregon.

  14. The nature of the data collection is definitely a big issue (as is its application). I was really surprised that the language for Pennsylvania was Italian, but it looks like that was a mistake; the 2010 American Community Survey for PA, anyway, has German third (with 55,891), with Italian fourth (with 50,658).

    But German is really surprising to me, too, unless it includes Pennsylvania Dutch, with it probably shouldn’t. Looking further down the list, Pennsylvania Dutch is in sixth place with 47,611, and Dutch, while far down the list, has 9,871 speakers.

    I find the German number unlikely, if it completely excludes PA Dutch, but it seems extremely unlikely that there are almost 10,000 speakers of Dutch in Pennsylvania (though I don’t know enough about the situation to say it’s actually impossible). So I’d guess that a big chunk of German & almost all of Dutch is actually Pennsylvania Dutch, in which case I guess that should be the third most common language in PA.

  15. Breffni says:

    J. W.: ‘I believe the underlying ACS question is more or less “do you speak a language other than English at home, if so what is it.”’

    That was the wording of the corresponding 2011 Irish census question: “Do you speak a language other than English or Irish at home?”, with follow-ups “What is this language?” and “How well do you speak English?”.

    It’s hopelessly ambiguous. I imagine the most obvious interpretation, and perhaps the intended one (given “How well do you speak English?”), is: “Is your [by assumption, only] home language something other than English or Irish?”. But it might mean “Do you speak a language instead of or in addition to English or Irish at home?” We speak English and Spanish, and I honestly couldn’t figure out how to respond. The question designers seemed to assume totally monolingual homes, which isn’t likely to be the case in the many homes where there are children and at least one parent is not a native English/Irish speaker.

  16. Are they descended from Alaskans, native or otherwise?

  17. J. W. Brewer says:

    Matt A.: I would tend to assume the Pennsylvania data is largely an artifact of people who speak the same language variety not all giving it the same label when asked an open-ended question, with the data compilers not in this instance being willing to aggregate the data based on what people probably meant (which would likely give an aggregate number well ahead of Italian). Note that lots of Amish have moved west to places like Ohio and Indiana, which may affect the “German” numbers there. How should “Pennsylvania German” (as wikipedia now seems to think it should optimally be called) be treated for lumping/splitting purposes with other varieties of German? I dunno. (Although there is apparently some degree of diglossia among the Amish insofar as they use in religious services hymnals and Bible translations in standard Hochdeutsch alongside sermons preached in their more distinctive dialect.)

    There is another wrinkle which the “at home” questions do not capture. There are not infrequently in the U.S. (and perhaps also in Ireland and elsewhere) mixed marriages where the spouses come from different ethnolinguistic backgrounds, so you may have a situation in which an immigrant spouse never uses his/her L1 “at home” because no one else in the immediate household speaks it, but still regularly speaks Spanish, Korean, or what have you with fellow immigrants in social situations outside the home.

  18. Tagalog in Hawaii shouldn’t have come as a surprise, because the original waves of the Filipino immigration there are more recent (as recent as 1930s) and more connected to the homeland population / less prone to language loss in subsequent generations than Chinese and Japanese immigration. I’m mostly exposed to things Filipino Hawaiian through my main line of work (genetics of the families which turns up, again and again, families with branches on the islands, on the main land, and back in Philippines). One of the clans I worked with had first immigrants to the US in the 1910s and it still has branches in HI, CA, and the Philippines (as well as in the Persian Gulf and in Australia). The 1910s waves of immigration didn’t originated in Northern Luzon like the later waves, but rather in the Visayas (the archipelago in the South-Center of the Phillipines) and many of these migrants must have been native Cebuano etc. speakers, rather than Tagalog; but their relatives in the Philippines now are primarily urban dwellers in the Greater Manila (and occasionally, Cebu City) and they speak Tagalog now, occasionally speckled with Visayan terms.

    But even if you are a recreational visitor to Hawaii, you may notice that the Filipinos are less assimilated than Chinese or Japanese or Korean or Portuguese Hawaiians. Their ethnic foods didn’t end up in the pan-island melange (I’m a sucker for babingka malagkit, BTW ;) ); their farmers markets (one of the best place to buy fresh vegetables / greens on the islands) have much business done in Tagalog; and in the gardening / growing areas of the islands (such as Maui’s North Slope backwoods of Haiku), Filipino communities are a force to be reckoned with, eclipsing the separatists, the landed hippies, and the telecommuter geeks taken together.

  19. Erik Hetzner says:

    Pretty sure the data is from here: https://www.census.gov/hhes/socdemo/language/data/acs/

    You can download the detailed by-state spreadsheets. In California the grouping “Chinese” includes the groupings Chinese, Cantonese, Mandarin, and a few others, and exceeds Tagalog, which has only one entry. As for French creoles, the high percentage of “French Creole” (it’s own heading in the survey) in Florida makes me think that most Haitian Creole speakers fall under that heading.

  20. Tagalog in Hawaii shouldn’t have come as a surprise, because the original waves of the Filipino immigration there are more recent (as recent as 1930s) and more connected to the homeland population / less prone to language loss in subsequent generations than Chinese and Japanese immigration

    I think you mean “Tagalog in Hawaii doesn’t come as a surprise if you’ve studied immigration to Hawaii,” which is doubtless true. If one is omniscient, after all, nothing comes as a surprise. But most of us haven’t studied immigration to Hawaii.

  21. Erik Hetzner says:

    JW Brewer: re. Chinese, the overall grouping is “Chinese” and is divided into Chinese, Mandarin, Cantonese, etc. So yes, it’s weird because Chinese is both a grouping and a language. So that’s why Chinese is top in NY.

    And re. Hopi in NH – there are 36 speakers and the margin of error is 60. Per the footnotes in their spreadsheet, this means there is a 90% probability that the number of Hopi speakers in NH is between 0 and 96. :)

  22. J. W. Brewer says:

    Via those spreadsheets one can double-check the Hopi-in-New-Hampshire oddity noted above. The spreadsheet estimates that New Hampshire has 36 Hopi-speakers, with a margin of error of . . . 60. So the correct number could plausibly be zero. The runner up is Navajo with 25 speakers, plus or minus 33, so . . . They really don’t have any data worth anything on the question. On the other hand the number of Danish-speakers in West Virginia is estimated at 127 plus or minus 106, so I guess they’re statistically confident that there are some rather than none.

  23. So yes, it’s weird because Chinese is both a grouping and a language.

    Usually, the language in question would be called “Chinese (not otherwise specified)”. I mean, if people say or write down that they speak Chinese, there’s usually no mechanism for asking them to be more specific, just because we wanna know.

  24. I think you mean “Tagalog in Hawaii doesn’t come as a surprise if you’ve studied immigration to Hawaii,” which is doubtless true. If one is omniscient, after all, nothing comes as a surprise.

    Swipe not taken :) If you know what’s Tagalog at all, you are already positioning yourself far beyond the nulli-scient crowds. And if you are interested in contemporary ethno-linguistic processes in the US (which is probably true for most readers of this LH entry), then it definitely makes sense for you to know something about how Hawaiian ethnic patchwork came to be.

    Besides Hawaii is a popular vacation spot, and part of its appeal lies in the things ethnic and ethno-historical – a part which I expect to be magnified in significance for any Hatter. No omniscience required.

  25. J. W. Brewer says:

    Re Bill W’s question, the Va. spreadsheet does include a line for Pashto, but it’s grouped under “Other Indo-European languages” rather than “Other Indic languages,” with I suspect a purely linguistic taxonomy having been overridden by political considerations (perhaps that immigrants from Pakistan are presumptively “Asian-American” in the U.S. diversity schema while immigrants from Afghanistan are presumptively “white” even though the national borders do not cleanly track the ethnolinguistic facts on the ground). Dari is probably lumped in with “Persian.” But Urdu is still the statistical winner of its category fair and square, and there are plenty of other Indic languages with more estimated Va. speakers than Pashto.

    Re Keith Ivey’s question, the spreadsheet for D.C. has Spanish in second, and Amharic in fourth, with French falling in between. How much of that French might be e.g. Haitian creole (with speakers self-identifying with the French box rather than the French creole box) I don’t know. There’s also an option for “Patois” which is treated as a subset of “French” rather than a subset of “French creole.” Yeah, I dunno. Any taxonomy like this is going to have some judgment calls and some are going to seem weirder or more arbitrary than others.

    Note BTW that they have “Jamaican Creole” on their list of “Other Indo-European languages” Dude, thanks for the recognition, but isn’t that sufficiently rooted in English to at least be grouped into “Other West Germanic languages”? Also, they call Latvian “Lettish,” which seems archaic to the point of zaniness.

  26. @John Cowan: Growing up, I had always assumed that the core of the Woodburn Old Believer commubity had come to Oregon via Alaska, but the reality appears to have been quite a bit stranger: http://sites.google.com/a/lclark.edu/rsco/immigrant-communities/old-believers

  27. I had always assumed that the core of the Woodburn Old Believer community had come to Oregon via Alaska, but the reality appears to have been quite a bit stranger

    and in some ways opposite: pretty much all of Alaska Old Believers came from Oregon (via Harbin and Brazil) (having purchased a surplus govt. land tract on Kenai Peninsula in the late 60s). None of the original Alaskan Orthodox belonged to the Old Orders.

    We already discussed the two strains of Alaska Russians in the recent Ninilchik dialect thread (although I mistakenly wrote there that it was Washington State, rather than OR, from where the Old Believers migrated to Nikolaevsk AK, and, following a schism there, on to Kachemak-Selo)

  28. David Marjanović says:

    Also, why are Pashto and Dari excluded from the Indo-Aryan group?

    Because they’re Iranian as opposed to Indic – only the latter gets called “Indo-Aryan”.

    But German is really surprising to me, too, unless it includes Pennsylvania Dutch, with it probably shouldn’t.

    Why shouldn’t it? I have no trouble reading the Pennsylvania Dutch Wikipedia. By the standards of what else counts as German, that’s way more than enough… :-)

  29. Stefan Holm says:

    Most of those maps use what mathematicians call discrete variables for what really should be continous ditto. That is, the winner takes it all. On the Scandinavian map Swedish is strongly dominating. If, say, in a state 100 families answer that they speak Swedish at the kitchen table, 60 answer Danish and 50 Norwegian, such differences would naturally reflect the distribution of speakers in Scandinavia itself. In that case Swedes would be less than 50% of the Scandinavian speakers but look like a huge majority.

    By the way, all current speakers of Swedish in the USA are first generation immigrants. Hopeful attempts to find any reminiscences from earlier periods (in Minnesota and elsewhere) have been in vain.

  30. Why shouldn’t it? I have no trouble reading the Pennsylvania Dutch Wikipedia. By the standards of what else counts as German, that’s way more than enough… :-)

    Ha! I guess that’s right, when you’re talking about German. I was just going on two things. One, I live in Lancaster, PA, & spoken PA Dutch sure seems completely different to me, but I can only read Hochdeutsch, and can’t speak any kind of German at all, so my opinion about that doesn’t really count. And the other is that I once watched a Werner Herzog documentary from the 70s about an auctioneering championship; it looks like he can’t understand the Pennsylvania Dutch speakers he speaks to, and they don’t seem to understand him at all (unless he speaks English). Of course, this could have been played up in the editing.

  31. J. W. Brewer says:

    And re Erik M’s original question, “Cushite” (which presumably includes Somali) is in fourth place in Minn. behind Hmong (which is in 3d behind Spanish) but quite a ways back in percentage terms. Of course if one group is still getting new immigrants at a substantial clip and the other isn’t (an empirical question I don’t know the answer to) or there are differences in the rate in which US-born kids do or do not pick up the heritage languages, those numbers could still shift fairly rapidly relative to each other in a fairly short period of time.

    Note, btw that this fine-grained breakdown at the state level seems to be from data collected in 2006-08, so it may already be out of date in some particulars. New data was collected in 2011 and there have been higher-level Census Bureau releases of that starting last year – so perhaps an equivalently detailed breakdown of 2011 data will be out soonish.

  32. J. W. Brewer says:

    Russian in 6th place in Alaska (after English, Yupik, Spanish, Tagalog, and Inupik), but I’m assuming that’s pretty much all the via-Oregon Old Believers rather than a survival of pre-1867 usage. I suspect that some non-Old-Believer Orthodox parishes there (whether the L1 of the parishioners is English or Yupik or Aleut or Tlingit or whatever) may still use some Church Slavonic in services out of respect for tradition but that doesn’t mean anyone including the priest could necessarily carry a conversation in any Slavic tongue.

Speak Your Mind

*