Wednesday, November 26, 2014

Keeping African languages out of African schools?

Humiliation of Ugandan students who speak their mother tongue in school, and Malawi's recent decision to move to an English-only instruction policy, reflect the continued low status of African languages in African education. In much of Africa, the first languages of students are formally excluded from African schools by national policies, and/or accorded low or even negative value in school culture. In the extreme, this situation can be seen in terms of denial of human rights (per work by Tove Skutnabb-Kangas), but in any event seems to run contrary to ample research on the benefits of  learning in the mother tongue and of bilingual education.

"Why are schools punishing children for speaking African languages?"

Punishment for speaking Luganda at school.
Image from Facebook via
A recent article by writer and lawyer Bwesigye bwa Mwesigire called attention to ongoing practices in some Ugandan schools of humiliating students who speak African languages. A particular example is requiring students caught speaking a language other than English to wear a sack, until they catch one of their peers transgressing the English-only rule, who will then have to wear the sack.

This is not a new linguistic human rights issue, nor a practice limited to one country. It extends back to colonial rule, though enforcement of "no vernacular" has evidently moved from corporal punishment by teachers to humiliation and enforcement by students. I previously mentioned this in a 2008 blog posting entitled "Burning textbooks, beating schoolchildren."

New "English-only" policy in Malawi

In a decision made last March by then Malawian Minister of Education, Lucious Kanyumba, changed Malawi's longstanding bilingual education policy to one where students will be taught in English only from day one of their schooling. The stated object of the change, which is now implemented, was to improve English language levels of students.

Previous language of instruction policy (per a 2000 paper by Henri G. Chilora) had all students learning in the national language Chichewa for the first four grades, and then shifting to English. However this policy, adopted in 1968, had the effect of eliminating other Malawian languages from schools (see Chilora's paper, and one by Misheck Dickson Issa and Shoko Yamada on perceptions of language of instruction policy).

The decision to move to English-only has been controversial, with arguments against it citing advantages of MTB/MLE, lack of teachers prepared to instruct in English, and questions about equating good English with good education. It elicited an early protest from students of Chancellor College. A good summary of the debate - which begins by acknowledging the students' action - is provided by Steve Sharra (a later version of this article was highlighted by Ndesanjo Macha in his blog).

A study by Helen Abadzi, Radhika Iyengar, Alia Karim and Florie Chagwira of Columbia University's Center on Globalization and Sustainable Development found that even if the goal of education was good English, it would make sense to begin teaching reading in children's first languages (which happen to be written with orthographies where letters have more consistent values than in English).

Language of instruction policy in Malawi has moved away from use of its children's first languages in stages - first using only Chichewa (national language, first language of a majority, and also a second language for some number) in the name of national unity, and now to only English (official language, first language of almost no one, and a second language for some number) in the name of better job opportunities.

Language of Instruction (LoI) in Uganda and Malawi

In the process of researching this blog post, I came across an article (actually a book chapter) that examines and compares language in education policies in the two countries mentioned above: Ismail S. Gyagenda and Wardah M. Rajab-gyagenda, "Examining Ugandan and Malawian Language of Instruction Policies From a Linguistic Human Rights Perspective," in Zehlia Babaci-Wilhite, ed., Giving Space to African Voices: Rights in Local Languages and Local Curriculum (SensePublishers, Rotterdam, 2014). For now I'll just pass on the reference, but one passage from the content available online leaves us with a relevant perspective and question:
"In the human rights approach [to language policy], language is a manifestation of one's identity and cannot be willy-nilly suppressed without deep educational consequences for the students as well as society in general. How do the LoI policies of Uganda and Malawi over the years fare within this human rights perspective?"

Tuesday, November 25, 2014

Writing Bambara right

How to compose text in the Latin-based orthography of the Bambara language of Mali? One question raised by an ebola poster in non-standard Bambara (see previous posting) is whether the modified letters (technically "extended characters") in the Bambara alphabet  discourage use of the standard orthography. There are two potential issues - fonts and keyboards - although noting use of standard Bambara in other materials, these are not the impediment they once were. I'll briefly discuss both below, after a quick intro to written Bambara.

Bambara orthography

The Bambara alphabet today includes the following characters:
a  b  c  d  e  ɛ  f  g  h  i  j  k  l  m  n  ɲ  ŋ  o  ɔ  p  r  s  t  u  w  y  z

Digraph consonants (two letters to represent one sound) have been phased out of use, such that "ny" is now "ɲ" (in Senegalese orthography this would be "ñ"). However, "sh" is still used - although the IPA borrowing for this sound - "ʃ" - is sometimes seen (just today noted it in an email by odd coincidence). Double vowels however are used, for words where the vowel sound is slightly prolonged.

Bambara is a tonal language. The two tones are rarely marked, but when they are, accent marks are used. (A change in the alphabet some years ago from "è" and "ò" to "ɛ" and "ɔ" permitted marking of tones with accents rather than underscores for low tone).


Time was, the lack of fonts (and before Unicode became the dominant standard, character encoding behind the fonts and the lack of compatibility among different 8-bit fonts) presented the main problem for creating and sharing text in Bambara with the extended characters ɛ, ɲ, ŋ, and ɔ.

Font support for extended Latin characters is still uneven, though current operating systems can substitute a missing character from another font (all being encoded in Unicode). As I compose this posting, I note Blogger's default font lacking 3 of the 4 extended characters from the obvious substitution (per figure above from screenshot; background color added). On the other hand, the font for the published posting does include these characters. So no substitution is necessary.

Basically this means that most of the time, one can display the needed characters, but for aesthetic reasons, fonts that include all of those characters would be preferable. In finding fonts, it is helpful to know that the needed extended characters may be spread among several Unicode "blocks." For Bambara this means a font will have to have, in addition to the basic Latin blocks common to any font with Latin characters, the following extended blocks:
Latin Extended-A is fairly common in fonts, but not the other three above. (If needed, the "ʃ" and its capital form would be covered by the IPA and Extended-B ranges.)

Alan Wood's extensive list of "Unicode character ranges and the Unicode fonts that support them" is an excellent resource for finding fonts for specific Unicode ranges. (Sill looking for a resource that would allow one to choose several Unicode blocks and get a list of fonts that cover them.)


Since display of extended characters is no longer the impediment it used to be, the big issue now seems to be how to efficiently compose text with extended characters that are not supported by computer keyboards (i.e., not via inserting symbol in a wordprocessor or cutting and pasting characters from another source). This means use of alternative keyboard drivers or onscreen keyboards or character pickers.

In the latter category, there are a couple of websites worth noting. In both, one types from one's keyboard and then clicks on extended characters, producing text onscreen that can be copied and pasted elsewhere:
  • Lexilogos has a page for Bambara, featuring a window where one can type basic Lain characters and then click on the extended ones onscreen (and diacritics for accents).
  • Richard Ishida has a more complex IPA Character Picker enabling input of many more extended Latin characters.
Keyboard drivers enable one to use one's existing keyboard in any application. These generally use either Tavultesoft Keyman or Microsoft Keyboard Layout Creator (MSKLC). A short list of links to current keyboard drivers useful for composing in Bambara follows (with thanks to Valentin Vydrin, who shared this on the Translating Hope list). I'll add to this but encourage comments with additional information:
  • Via Mali Pense site (see under "ÉCRIRE LE BAMBARA - ka bamanankan sɛbɛn"). Note also a spell checker ("vérificateur orthographique"; see under "POUR ÉCRIRE SANS FAUTES - Fililatilennan sɛbɛnni na")
  • Via LLACAN site (see under "Saisir des caractères spéciaux sous windows.")

Monday, November 24, 2014

Does spelling matter in Bambara ebola materials?

International SOS ebola poster in Bambara
A Bambara version of an ebola information poster - one of a number of translations of an original done by International SOS - shows why review of technical health materials translated into or composed in African languages is necessary (per "2Ds & 4Rs"). While giving due credit for the translation, it is also important for the common ultimate goal of communicating about ebola in the most widely spoken language in Mali to offer constructive criticism towards improving this product and guiding similar efforts.

In the case of this poster, it is immediately apparent that the standard Bambara orthography is not used. This is important since the standard orthography is used in adult literacy, some primary education, and various publications including on ebola and other health related topics. In a previous posting on this blog - "More on standard orthographies of African languages"- I discussed in more detail the reasons to respect and make use of the alphabets and rules of spelling adopted in many African countries for writing their languages. In the case of critically important health information on a deadly virus, adherence to standard ways of writing, as well as to as much standard terminology as there may be, would seem obligatory.

The resolution of the poster available at this time (from an image accompanying a tweet embedded below, which also includes a French version) is not sufficient to review the content in detail, but is is possible to provide some tentative revisions to the headings (see below the embedded tweet).

Proposed revisions

Here are the original headings from the poster followed by tentative revisions. In this I have relied on my rusty knowledge of Bambara with reference to the online dictionary at Any corrections or better renditions are invited:
(NB- Doubled vowels are pronounced differently than single ones. "Baana" by itself may mean "rich person," although the context here would tell one that what is meant is "bana"=disease.)
(NB- Clear influence of French orthography - further comment below)
(NB- Bambara has a 7 vowel system; "ɛ" and "ɔ" are different from "e" and "o")
(NB- Some issues with nasalized terminal vowels, which are indicated with an "n" following:  "bɛ" is generally not written with a terminal "n", but it may be that the writer hears the open-e ("ɛ") as nasalized; on the other hand, "mun" [what] is written with a terminal "n", though in common speech, the nasalization may not be heard. Also, "ni i" [if you] is often contracted as "n'i".)

Bambara in French orthography?

It is possible to write Bambara (or any language for that matter) in French orthography. The New York Times, for instance, used the French spelling of the Bambara and Manding term "jatigiya" (basically meaning hospitality)  - "diatiguiya" - in two recent stories on ebola in Mali (on Nov. 10 and Nov. 12). And it is true that Bambara speakers schooled only in French (which for long has been the general rule in Mali) may resort to orthographic rules of the latter when transcribing their first language - although it should be pointed out that this often results in different spellings based on the user's hearing and rendering of sounds. This juxtaposition of a language (Bambara) with a historically recent standardized written form, and a large number of the literate population taught only in a second language (French), raises legitimate questions about how best to render Bambara in print for ebola messaging.

However the fact that Mali has chosen a particular orthography (much like that used for similar Manding languages in neighboring countries in the region), and that people taught to read the language have been taught in this orthography, would support using this system rather than one based on French phonetics. Keep in mind that people taught to read Bambara in adult literacy classes or in mother tongue based bilingual schools would likely be in villages where they could read the material - aloud i necessary for any community members who cannot read it.

As an example of use of the Bambara orthography, note the translation of an ebola factsheet done by the Dokotoro Project and the Mali Health Organizing Project, which was highlighted in an earlier posting on this blog.

International SOS ebola poster in Yoruba
I also should note here that Bambara can also be written in the N'Ko alphabet, and that there are people literate in that instead or as well. Altogether this is a complicated situation similar to that in the rest of the multilingual West African region, but not one so complex as to make use of first languages and local lingua francas like Bambara in their written forms problematic for implementation.

Better localization also means pants

Translation also needs to be accompanied by attention to cultural appropriateness of the material and also, where images are involved, the visual literacy of the audience. Together, these considerations are part of full "localization" of material. It is worth noting in this context that when Translators Without Borders produced localizations of the same International SOS ebola posters in Nigerian languages, they added pants to the figures (which although abstract, appear in the original to be dressed only in t-shirts). A revised Bambara version should probably follow suit.

(For a discussion of technical aspects of producing text documents in Bambara, see the next posting: "Writing Bambara right." For two other translations of the same poster, which use standard Bambara, see:  "Two more ebola posters in Bambara." For a similar issue with a poster translated into the related Malinke language, see: "Does spelling matter in Malinke ebola materials?")

Thursday, November 06, 2014

Uganda ebola posters, in which languages?

Here are three ebola posters from Uganda - produced by the Health Promotion and Education Division of the Ministry of Health - for which I'm requesting help to identify the languages. Have had no luck with other channels so will post here and hope for input from a wider audience.

The three were part of a 3-page PDF document, which I converted into separate image files (these were also posted on Twitter):

1. Yega ebindi ebikwataine na EBOLA

2. Omanya ebikwete aha EBOLA


 3. Minya binene'bihambengene okwa EBOLA

There was a suggestion on Twitter by @IndigenousTweet that the language of the third one is Olukonzo.

Please feel free to add a comment if you can identify the languages of any of the three. 

Addendum (13 Nov. 2014)

A tweet from @IndigenousTweet offers tentative identification of the languages:

  • Rutooro is one of four closely related languages of southwestern Uganda, for which a common standardized version - Runyakitara - has been developed. It's worth checking whether the poster is actually in Runyakitara.
  • Lusoga is spoken in southern Uganda, to the east of the area where Luganda (to which it is closely related) dominates.
  • Olukonzo is also spoken in southwest Uganda and apparently also across the border in DRC.
A follow on question is whether these three posters are actually a subset of a larger number of translations with other Ugandan languages. (Am seeking more info, which I'll post when available.)

Tuesday, November 04, 2014

Time for a Krio Wikipedia?

According to a recent tweet, one of the speakers at the recently concluded Localization World Conference in Vancouver (29-31 Oct. 2014), pointed out a specific need in the area of African language Wikipedia editions:

The speaker was probably referring to Krio, a creole language that is the most widely spoken lingua franca in Sierra Leone. In my limited understanding, the pidgin forms of Liberian English spoken in Liberia, however, are not the same, even as they share some characteristics. None of the above are spoken significantly in Guinea, the third country in West Africa most impacted by the ebola epidemic.

Nevertheless, this raises an important issue, since Krio is a language with a written form that is said to be spoken by 90% of Sierra Leoneans. It is evidently already used in at least some kinds of ebola communication, and would be a logical immediate focus for developing and disseminating further information on ebola and other health topics.

A response to the above tweet informs us that there is already the foundation of a Wikipedia edition in Krio:

So the next question is how to encourage development of this nascent Krio Wikipedia? Perhaps starting with ebola and other health-related info via the WikiProject Medicine's Translation Task Force and the Ebola translation task force?