More on US Census Bureau & African languages

Last July I posted on the U.S. Census Bureau's coverage of African languages spoken in the United States. That focused on names and categories used (which I understand will be reviewed for possible revision), and included a map from Slate based on the Bureau's data showing the most spoken African languages or categories by State.

In October, the Census Bureau released its detailed data on over 300 languages (and language categories) spoken in the U.S. A summary table of this data was featured in an article written by Nikhil Sonnad last month on Quartz (and on CityLab under a different title). There is a small error in the text of that article, where it mentions "Sudanese" (spelled like the nationality) as a language - that should actually be "Sundanese" (spelled correctly in the table), which is spoken in Indonesia.

Below is a table with information excerpted from the Census Bureau's data, showing the numbers for the African languages discussed in my previous post, sorted by number of speakers. I have added Krio and Pidgin, which were omitted from that post. "African" stands for "African (not further specified)" in the Bureau's list of languages. The total number of speakers of African languages as defined by the Bureau - all of the below except Afrikaans, Arabic (which of course is spoken in Southwest Asia as well as North Africa), Krio, Malagasy, and Pidgin - is 894,499.

	Number of speakers1	Margin of Error2	Speak English less than "Very Well"1	Margin of Error2
Arabic	924,374	13,743	341,425	5,888
Kru, Ibo, Yoruba	322,255	7,681	64,690	2,487
Amharic	195,260	6,368	81,385	3,479
Cushite	122,445	4,437	59,495	2,817
Swahili	88,685	3,414	22,055	1,913
Bantu	56,685	2,641	16,635	1,574
Fulani	30,475	2,022	11,745	1,193
Mande	29,835	2,461	10,370	1,171
Afrikaans	23,010	1,525	1,885	318
African	12,320	1,508	5,000	997
Krio	10,560	1,240	2,820	718
Chadic	8,565	991	2,275	426
Sudanic	8,510	1,317	3,935	710
Nilotic	6,890	1,184	2,235	490
Efik	5,620	775	905	305
Pidgin	4,445	636	1,100	352
Berber	2,940	756	1,630	472
Gur	1,310	529	405	272
Nilo-Hamitic	1,275	644	575	327
Malagasy	720	231	225	101
Mbum (and related)	715	353	370	269
Nubian	305	234	185	175
Nilo-Saharan	270	183	155	127
Saharan	80	95	(D)	(D)
Khoisan	55	89	20	32

Notes:
1. Detailed-language estimates are rounded to the nearest multiple of five. Aggregate estimates (in this selection from the original, only Arabic) are unrounded and appear in table B16001 (http://factfinder.census.gov/bkmk/table/1.0/en/ACS/13_5YR/B16001/0100000US). Detailed-language estimates may not sum to aggregate estimates because of rounding.
2. Data are based on a sample and are subject to sampling variability. The degree of uncertainty for an estimate arising from sampling variability is represented through the use of a margin of error. The value shown here is the 90 percent margin of error. The margin of error can be interpreted roughly as providing a 90 percent probability that the interval defined by the estimate minus the margin of error and the estimate plus the margin of error (the lower and upper confidence bounds) contains the true value. In addition to sampling variability, the ACS estimates are subject to nonsampling error (for a discussion of nonsampling variability, see Accuracy of the Data at http://www2.census.gov/programs-surveys/acs/tech_docs/accuracy/MultiyearACSAccuracyofData2013.pdf). The effect of nonsampling error is not represented in these tables.

Source: 
U.S. Census Bureau, 2009-2013 American Community Survey, Table 1. Detailed Languages Spoken at Home and Ability to Speak English for the Population 5 Years and Over for United States:  2009-2013. Release Date: October 2015. http://www2.census.gov/library/data/tables/2008/demo/language-use/2009-2013-acs-lang-tables-nation.xls

US Census Bureau and African languages

The number of people speaking an African language at home in the United States* increased by 111% in the decade before the last census. Which African languages are most common among them? Looking at a map of most spoken African languages by state in the US from an article by Ben Blatt in Slate last year kind of answers that, but then raises more questions. To begin with, the names listed in the map include two languages (Amharic and Swahili), two branches of language families (Bantu and Cushitic), one language group (Nilotic), and a seemingly eclectic grouping of "Kru, Ibo, and Yoruba."

According to the Census Bureau site, it tabulates non-English speaking households by 328 languages or language categories (see the full list of codes). A list of African language categories excerpted from the Census Bureau's list raises more questions (the below includes headings and subheadings from the list; those categories marked with "◄" are listed in the Slate article as the "choices"; those highlighted in yellow are listed in one or more states on the map):

Source: "Tagalog in California, Cherokee in Arkansas: What language
does your state speak?" by Ben Blatt, Slate, 13 May 2014

INDO-EUROPEAN (601-678)
...
Germanic languages (607-618)
...
611    Afrikaans
...

ASIAN AND PACIFIC ISLAND LANGUAGES (684-695, 698-771)
...
Pacific Islands languages (730-776)
...
   West Indonesian (732-741)
...
738    Malagasy
...

OTHER LANGUAGES (679-683, 696-697, 777-999)
Semitic languages (777-780)
777    Arabic
...
780    Amharic ◄

Afro-Asiatic languages (781-783)
781    Berber ◄
782    Chadic ◄
783    Cushite ◄

Nilo-Saharan languages (784-789)
784    Sudanic ◄
785    Nilotic ◄
786    Nilo-hamitic ◄
787    Nubian ◄
788    Saharan ◄
789    Nilo-saharan
790    Khoisan ◄

Niger-Congo Languages (791-798)
791    Swahili ◄
792    Bantu ◄
793    Mande ◄
794    Fulani ◄
795    Gur ◄
796    Kru, Ibo, Yoruba ◄
797    Efik ◄
798    Mbum (and Related) ◄
799    African (not further specified)

Time for the Census Bureau to review its African language categories?

It is hard to know what to say about this list. It is obvious that some thought has gone into it (despite the puzzling combination of two major and distinct Volta-Niger languages - Yoruba and Igbo - with the linguistically and geographically distant Kru languages). And indeed the Bureau indicates it has based the categories "generally on Classification and Index of the World's Languages (Voegelin, C.F. and F.M., 1977)" and that it updates them "constantly using linguistic books and online resources." But without further  background it is hard to get away from the feeling that it is a bit random.

Another unknown is what the process is for determining who is counted under what category. The Census questions include an open-ended one for people who speak a language other than English at home: "What is this language?" Only the few languages listed directly on the list would be simple to tabulate - such as Amharic and Swahili. But if someone answers, say "Hausa," they would at some point (presumably not that of the enumerator) have their answer tabulated under "Chadic," since Hausa is in the Chadic branch of the Afro-Asiatic family of languages. If "Zulu," then "Bantu," and so on. The reason for tabulating under broader categories, as I understand it, has to do with confidentiality concerns (when there are few people speaking a language in an area).

However, if "Wolof" or "Akan" is the answer, there doesn't seem to be a category under Niger-Congo that applies, other than the catch-all 799. How many cases are there like this?

Another issue is that a couple of categories may not be current - Sudanic and Nilo-Hamitic.

Altogether, it would seem that an expert review and revision of the categories would be timely, as we are still five years from the next census. That would make the results more useful.
_________
* According to U.S. Census Bureau, Statistical Abstract of the United States: 2012, "Table 52. Population by Selected Ancestry Group and Region: 2009," a total of 777,553 people speak African languages (not specified) at home. This total does not include Arabic, which was counted separately at 845,396.