Saturday, April 15, 2017

Second most spoken languages in Africa, part 3

A couple of years ago, I wrote a couple of posts about a map of second languages in Africa, within a set of similar maps for all continents, that came out of Olivet Nazarene University (ONU). Well there's another version by Max Holloway on MoveHub that actually dates to 2014, which is making the rounds again now thanks to Digg.

Here's the Africa map, snipped from the set (click to enlarge), and with the legend modified slightly to make the image narrow enough to fit here in a larger view:

What are we talking about?

I actually think that it's great that people try to produce such different ways of looking at facts we sometimes take for granted. Not meaning "alternative facts" here, but maybe alternative ways of looking at facts - different angles which help understand a complex whole.  And I especially appreciate the effort that has gone into doing this with regard to languages. Also, it is worth noting that the map of Africa above covers more countries than the ONU map - no small effort in either case, but kudos to Max Holloway and MoveHub for taking it further.

All that said, the first issue with this effort is the same as with the other one: Is it "second most common first language" (L1) or "most common second language" (L2)? I think it's intended to be the first, but that's muddied by mention "second language." Or is it really something like the second most commonly spoken language (L1+L2)? It would help to begin such efforts by mentioning these alternatives, and making it clear what one is and and which ones aren't being referred to.

What counts as first most spoken?

Second, there are questions about assumptions made and data used. Is the assumption here - like in the ONU map - that the "official language" (a legal or sometimes constitutional category) is the most spoken (first) language? That cannot be assumed to be the case, especially in African countries where official languages generally are those inherited from the colonial period. So for example, as I discussed previously (in "part 1"), Bambara would not be the second most spoken language in Mali, but rather the first (L1+L2, and probably L1 only), with the official French probably being second (counting L1+L2 speakers, but definitely not first or second counting L1 only).

Similar issues arise in many countries. I won't go into all of them (having done so previously, in "part 2"), but will note the interesting case of Ethiopia. The two most spoken languages there are Amharic and Oromo, and figures vary on their respective numbers of speakers. The ONU map showed Oromo as the second language, but in the accompanying article cited figures that Oromo was spoken by more of the population. The map above from MoveHub reflects the latter (Amharic as second). The figures in Ethnologue are very close, with about 100k more Amharic speakers (L1+L2) than speakers of all varieties of Oromo (which are generally taken together, thought that couold be another discussion), However it looks like a larger percentage of Oromo speakers are L1 speakers, there being a significant number of L2 speakers of Amharic. I go into all this as an indication of the kinds of complexities one gets into when trying to make a simple declaration of which language is the second in the country - as well as the need mentioned above to be very clear what criteria one is using.

Data and interpretation

But what about the data on which the map is based - where did the information used come from? Perhaps from a list something like this one from InfoPlease? Many of the labels on this map look like the languages listed in second position for various countries, including "Sudanic" for Burkina Faso and "Bantu" for Angola, which are language families and not languages (and Sudanic is not currently used as a linguistic classification). So a major issue is the quality of data relied on, and its interpretation on the map.

On the topic of language groupings, Fon in Benin is a Gbe language, like Ewe in Togo and southeast Ghana. In all three of these countries (among many, as noted above) the rankings of languages should be reviewed - although it is interesting to imagine the unexplored implications of three states in West Africa having major numbers of Gbe language speakers.

Again, I won't review this in detail, as much is similar to the ONU map already reviewed, though with some differences that are interesting (such as Mende probably correctly the second most spoken language in Sierra Leone after Krio) or puzzling (such as Kiunguja, a dialect of Swahili in Zanzibar, for Tanzania).

Three maps

Where to go with all this, and why spend three blog posts on it? On the latter question, I think that this map concept is a useful way to look at languages in Africa - and the world (remembering that both the ONU and MoveHub efforts covered all continents). However, nice graphics have a way of circulating and if the information in them is not accurate, or otherwise presents a confusing picture, they don't serve the purpose they were created for.

Yet in the case of Africa, at least, this is a complicated subject based on often imperfect data that can be interpreted variously. So anyone's map of a clearly defined "second (most spoken) (first) (first & second) languages" by country could be critiqued on details.

What I would propose for Africa is a set of three maps, based on a bonafide source, showing for each country:

  • the most spoken language (L1+L2),
  • the second most spoken language, and
  • the third most spoken (inspired on the latter by an interesting map of third most-spoken languages by state in the US). 

Put these three side-by-side with the standard map showing official language(s) by country and you have the basis for some interesting discussions.

* "Second most spoken languages in Africa" (1 May 2015) discusses the problems with the ONU map, while "Second most spoken languages in Africa, part 2" (8 May 2015) comments country by country. See also, "How many people speak what in Africa?" (7 May 2015).


Unknown said...

IsiZulu is the most commonly spoken language in South Africa. You have isiXhosa. You ether have isiNguni or isiZulu but not idiXhosa

Anonymous said...

This information is false. I see no where you mentioned kinyarwanda and it's the second most spoken language after Swahili in east and central Africa. We do have our local languages we speak indo-european languages as SECOND LANGUAGES stop misguiding the world

Don said...

Thanks for these comments. Three quick clarifications before responding:

1) This map, is not mine, but one apparently put together by Max Holloway of MoveHub based on a reading of linguistic profiles of countries on the CIA website (page gone from their site but preserved on

2) This is intended to be a map of second most spoken languages, not first most spoken language (hence isiXhosa is listed for South Africa, not isiZulu, for example.

3) This map is country by country, so it wouldn't treat the ranking of languages on a regional or multi-country level (altho that might be interesting.

In my opinion, the data on the original CIA page is not presented in as systematic a way as would be needed to accurately interpret what is the first and second most spoken language in all countries. In the case of Mali, for example, it says French is official (with no estimate of % of people speaking it) and then follows with Bambara as having 46.3% (which is much lower than other estimates I'm familiar with). In fact, it is generally agreed that Bambara is spoken by more people in Mali than French, but the map indicates Bambara as the second language of the country. French is widely spoken as a second language, and it is reasonable to assume it is the second language (Fulfulde probably being the third most spoken).

There are a lot of such issues with this map (and the data on which it was based).

For Rwanda, however, the original CIA page has French as second after Kinyarwanda by percentage of the population, and similarly for South Africa, isiXhosa is second after isiZulu (and before Afrikaans).