Wednesday, December 31, 2014

Does spelling matter in Malinke ebola materials?

An earlier post looked at problems with non-standard Bambara in an ebola poster. Here I'll look at similar issues with a poster in Malinke (Maninka), a closely related Manding language spoken mainly in upper Guinea, and also southwestern Mali.

The Malinke poster at right is a translation of a CDC poster on ebola symptoms, one of various different materials translated into a number of languages by Translators Without Borders. It appears to be written in the old, pre-1985 Guinean orthography with some French spellings.

After independence, Guinea adopted a common alphabet to transcribe all its languages, using the keys available on the typewriters of the day. Many people learned to read and write with these spelling rules (I personally encountered people and Pular materials using this system, and used it myself, in the mid-1980s), and to some degree their use persists.

When the government transitioned after the death of longtime president Sékou Touré in the mid-1980s, Guinea adopted an orthography harmonized with that of the other countries of the region. For Malinké that meant for example the following changes: dyj ; èɛ ; nhŋ ; nyɲ ; öɔ ; and tyc (the latter sound is represented by "ch" in English and variously by "tch" or "thi" in French).

The French writing influences in the poster are dj instead of dy or j, and ou instead of u. The accent aigu on é in the poster is probably gratuitous (the letter e in the old and new Malinke orthographies indicates the same sound) - unless a tone was intended by the writer (Manding languages are tonal, but usually tones are not marked in writing them in Latin-based script).

To see what the current orthography would look like, a simple set of substitutions would be straightforward, however I'm not sure on some word divisions (what knowledge I have of Bambara is an imperfect guide for that task).

The product as is, is presumably readable - and read-aloud-able - by many in Guinea since it is based on the older transcription. I don't have information on how widespread use of the new orthography is, but that is a question that should be considered.

And to the extent it is used consistently, conversion to the new system would be easy - which would facilitate use with Malinké populations outside of Guinea, or comparison with materials in other Manding languages (for terminology development etc.). Malike is also written in the N'Ko script (which has been the subject of some previous posts) - transcription between that and the Latin-based systems for communication or for comparison  is not as simple.

So in answer to whether spelling matters - yes, especially if the material is to be re-used or reviewed in comparison to other materials in Malinke or in other Manding languages. The fact that there has been a change in orthographies adds a complication, but one easily dealt with if each system is used consistently. And as indicated above, the matter of relative knowledge and use of the old and new orthographies for Malinke - and other languages of Guinea - is a question.

Addendum (2 January 2015)


I am told that the ebola poster above has numerous mistakes. This again points to the need to review materials in African languages (per "2Ds & 4Rs"), preferably before publication.

See also follow up posting: "More on written Malinke."

Friday, December 26, 2014

When do Liberian languages bear mentioning in ebola reporting?

A recent Oxfam piece about door-to-door ebola education volunteers in the West Point township of Liberia's capital, Monrovia, is an interesting story about local level efforts to combat the epidemic...
However it tells us nothing about the important linguistic aspects of this communication and research effort in multilingual Liberia. What languages does a volunteer like the profiled Agnes Nyantie use in speaking with various people in West Point? What kind of training did she and colleagues receive in those languages? When she "meticulously" fills out visit forms, is she doing so verbatim, or translating from one spoken language to another written one (presumably English)?

It may be that all is done seamlessly in English (the official language), but given that Liberia is a country where 31 languages are spoken (per Ethnologue), that no first language is spoken by more than 20% of the population (per Aménagement linguistique dans le monde), and standard English is spoken by only about 20% of the population (per Fact Monster), that seems unlikely.

Liberian (Pidgin) English is spoken as a second language by around half the population, playing a role analogous to Krio in Sierra Leone, though not as widespread, and apparently not as standardized. So it may be that a lot of the conversations that would be part of work like that being done in West Point take place in the pidgin. However, would a non-standardized pidgin be used to fill out visit forms?

What about the first languages of Liberia in ebola work? They are used on radio stations, for instance. A story last October in Jeune Afrique highlighted local radio in first languages of Liberia, such as Bassa. A WHO/Africa story discusses traditional chiefs of the National Traditional Council of Liberia recording "several radio messages in different languages telling their people to take steps to avoid Ebola" - these recordings were also used on radio.

It would seem likely that these languages are used commonly in local level person-to-person interactions about ebola - whether of the sort described in West Point or in treatment facilities, etc.

However, given the lack of standardized writing for some languages (different situation from some other countries of the region) and use of indigenous scripts which are not yet fully supported on computers/mobile devices for writing some of them, there may not be much ebola messaging in text format - though that is a question to answer.

So back to the Oxfam article... Recognizing the importance of that organization's contributions to the fight against ebola, and the courageous work of local volunteers working with it to inform and research in ebola striken areas, it would be informative to have more details on the linguistic dimensions of their communication and survey efforts.

Thursday, December 18, 2014

Ebola materials in Temne (Temen, Themine, Themne, Timene, Timini, Timne, Timmannee)

Temne translation by TWB.
Source: International SOS.

A posting on this blog in September featured an ebola message in several languages of Sierra Leone, but not Temne, which is widely used mainly in the north of the country (over a million speakers). Here I'll highlight some ebola materials in Temne (with notes about the name of the language at the end).

First is a translation of the the widely translated and circulated ebola poster by International SOS (on right). This was dated October 7, and done by Translators Without Borders (TWB, which also was involved in producing Mende and Krio versions of the poster).

TWB also helped Humanitarian Response with production of Temne versions of "Social Mobilisation Key Messages" and "Messages for children and caregivers on Ebola - Child Protection and Education."

Other material I'm aware of is audio, beginning with Temne versions of the CDC's radio spots. These recordings - which now are in 17 African languages or varieties of those languages (aside from English and French versions) - have been mentioned several times previously on this blog. It is worth recalling that these audio spots are not accompanied by text version in these languages, and there is no information about the approach(es) taken to producing the translations.

Scientific Animations Without Borders (SAWBO) has produced an animation on ebola, and one of the voice-overs is in Temne. SAWBO's products are available in formats for viewing on cell phones and computer (though there do not appear to be text versions / scripts available).

I am not aware of any material on ebola developed directly in Temne (written or recorded), although on the levels of national or community radio there may have been some significant use of the language in messaging.

What's in a spelling?


A quick comment on nomenclature. The name for the language is rendered in quite a few variant spellings (perhaps including some misspellings), several of which I've listed in the title. These have been observed on the web and/or taken from Ethnologue. Such variation in spelling can't help in managing information in the language. This is not problem unique to Temne (I discussed it in another posting), but seems to be a bit extreme in this case.

The most widely used spellings in English seem to be "Temne" and "Themne," although I remember hearing the language called "Timini" by a native speaker some time ago (the name in the language itself is apparently "Themnɛ"). Would it be possible to settle on one or two spellings that all organizations working in/with this language would use to refer to it?

Tuesday, December 16, 2014

Two more ebola posters in Bambara

Here are two other Bambara translations of the widely translated and circulated ebola poster by International SOS.

In a previous posting, I called attention to another version of the same poster in non-standard Bambara. Both of the ones discussed here use the current standard orthography, but also appear to be a better and more complete translations. However there are differences, which provide an opportunity to review them in comparison and ask questions about methods and outcomes.

The one on the left below was posted on Facebook on October 2 (original provenance not clear); the one to its right was translated by NGO Miriyawa and came from the International SOS site, dated December 1.

To facilitate parallel review of differences between the two translations, the headings (only) are presented in the table below - along the lines of what I did for the previous Bambara translation (which can be opened in a separate screen for comparison):

EBOLA VIRISI (BANAKISԐ) EBOLA
EBOLA YE MUN YE, A BԐ YԐLԐMA COGODI? EBOLA YE MUYE A BԐ MϽGϽ SϽRϽ COGODI?
MUN YE EBOLA TAAMASHԐNW YE?
Tamashԑnw bԑ daminԑ tile 2 magalen kᴐ Ebolatᴐ walima a fure la
EBOLA TAMASԐNW YE MUN YE?
Tamasԑnw bese ka daminԑn tile 2-21 i magalen kᴐfԑ banabagatᴐ la walima a su la
Tamashԑn Kunfᴐlᴐw - Tamashԑn Labanw > Tamasԑn Fᴐlᴐ - Tamasԑn Wԑrԑw >
EBOLA KUNBԐNI ANI WALI KԐTA N'I BANARA? EBOLA KUNBԐ COGO ANI ANW KAN KA MUNKԐ

The differences show how different people or teams can come up with alternative ways of rendering the same text - presumably comparable in meaning, but not necessarily equal in accuracy or comprehensibility. Or, if the two were actually translated from different versions (for example, one from French and the other from English), this shows how that can result in different outcomes in the target language (Bambara in this case). It is useful in any event, I think, to first compare the two versions in terms of the language in which they are written: Leaving aside copy-editing issues (spelling; typography), how accurate and understandable are the alternative headings - and the rest of the text?

Going back a step, one question is whether there was any back-translation in either case to verify the accuracy of the translations. Also, what kind of proofing (copy-editing) was done on each version? There are some errors and inconsistencies, although probably not enough to affect meaning.

Another question is whether either of these versions were "field tested" to get an idea of how an average reader - or a listener to the text being read aloud - would understand and react to the text. And visually, how people react to the illustrations. If that was not done, it would be informative to field test the two versions together (perhaps adding the previously discussed one as well).

When doing translations for public education about a health emergency, there is obviously a concern with getting materials out there promptly. However, translation shouldn't be a "once and done" thing. I've previously outlined a case for review of materials (in the context of 2Ds & 4Rs), but review - including quality control - needn't delay initial roll-out of material. Some level of review can be built into the initial process, and then follow-up review can follow. In the case of Bambara there are several materials on ebola available now (remember also the MHOP/Dokotoro factsheet and the Bambara Wikipedia article), which could provide the basis for a more in-depth analysis of use of Bambara for communicating about this virus and other diseases.

Finally, a quick word about "banakisɛ," which appears in one of the above posters. Composed of "bana" (sickness) and "kisɛ" (seed), the literal sense seems very evocative, and in any case more tangible than the abstract loan "virisi." Definitely would wash my hands thoroughly to get rid of those banakisɛw.

Thursday, December 11, 2014

Languages & communication in Senegal's ebola success

Ebola prevention poster in Wolof.
Y'en a marre and An@fa
Source: TWB via Alexandria blog

Senegal was declared ebola-free in mid-October, just before a similar declaration for Nigeria. Having previously looked at languages and communication in Nigeria's success, this post will do the same (belatedly) for Senegal. Language is a key consideration in ebola messaging in the multilingual states of West Africa, even if that fact is sometimes overlooked (such as by a National Geographic article on four lessons from Nigeria and Senegal, or a US Embassy Dakar sponsored conference for journalists on responsible ebola reporting).

By way of background, French is the official language of Senegal, and Wolof the most widely spoken. Of the country's 38 languages (by Ethnologue's count). Wolof and five other languages - Jola (Diola in French spelling), Malinke, Pulaar (as Fula is known there), Serer, and Soninke - are the main national languages. I'll return below to more detailed discussion of categorizating indigenous languages in Senegal, as it could be relevant to planning messaging strategies in similar multilingual countries like Mali, which is the latest country in the region to have an ebola crisis.

Enter, ebola


Prior to the sole ebola case in the Senegal - an infected man who came from Guinea to Dakar before the closing of the Guinea-Senegal border on August 21 (he ultimately recovered) - the country had been "well-prepared with an Ebola response plan as early as March." From what information I've been able to gather, information and public education aspects of this plan used a number of languages, but mainly Wolof and French. (An earlier posting included some links to videos and other web-based info in Wolof and Fula/Pulaar.)

A Senegalese source I consulted (Dr. Ibra Sene) indicated that Wolof and French were used in various media as part of the "massive campaign" to educate people about ebola. However, a report in International Business Times (25 September), characterized radio and television broadcasts of ebola prevention messages as "all day long, in all languages" (without specifying which). If the latter is accurate, it is likely that local community radio stations extended the campaign in additional languages (more on community radio below).

Still, an SIL-led initiative for translation in southern Senegal was a response to the perception of insufficient attention to communicating about ebola in less-widely spoken languages. It focused on developing ebola materials in the Bandial Gusilay, Jola (Jola-Fonyi), and Manjaku languages of Casamance. Two expat literacy workers in Senegal who facilitated the workshop, Clare Orr and Elisabeth Gerger,  framed it this way
"These days, a lot of awareness-raising about Ebola is going on across Senegal on the radio and TV. However, many people in the villages don't speak enough French, Senegal's official language, or Wolof, the most widely-used national language, to understand the message well. This is why we are trying to reach them through documents and information in their own languages. Those who are able to read in their language can always read the information aloud for those who can't."

 Community radio


Community radio is a key component of development communication in much of Africa (see here and here, for example), including Senegal. A grant by OSIWA in September to the Senegalese Union des Radios Associatives et Communautaires (URAC) for an ebola education program on 73 radio stations reflects the importance of this medium.

A couple of articles have highlighted community radio broadcasts about ebola in first languages of other countries - Guinea, and Liberia and West Africa in general - so one may assume that something similar has been happening in on community radio stations in Senegal. Questions relating to African language broadcasts about ebola on community radio include: which languages?; how was the material developed/translated?; and is any of it available in recordings or transcriptions?

An interesting dimension of communication via local radio in West Africa is that of associated radio listeners' clubs. The UN FAO noted use of these (and a training for club leaders) to address ebola in communities in Senegal.

Social media


On social media, an interesting initiative launched at the end of August as a Facebook group called SenStopEbola, with an associated Twitter account @senstopebola, translated notices and advice on ebola into seven Senegalese languages (see also here; the languages were not specified, but presumably included French and the 6 main national languages). .

Senegal made use of SMS to send text messages about ebola, but reports I've seen have not indicated what languages were used or the content of messages. Also, it's not clear whether the multilingual GooglePlay app - "About Ebola" (which was created last April, and mentioned in a previous posting on this blog) - is counted as part of this effort. Another GooglePlay app, SenStopEbola, which is evidently related to the initiative on Facebook, was mentioned as having Wolof content, but again, no information on how it has figured in wider efforts.

The national languages of Senegal


Of the 38 languages of Senegal, the six most widely spoken - Wolof, Pulaar, Seser, Soninke, Malinke, and Jola - were denoted "national languages" first by presidential decree in 1971, and are specifically named as such in the Senegalese constitution of 2001. The constitution allows for other national languages once "codified" (which includes establishing an "official" standardized orthography).

From 2001, the policy has been to expand the number of national languages, and to that end, several others were codified by 2002: Hassaniya Arabic; Balanta; Bassari (a.k.a. Oniyan); Manjaku; Mankanya; Noon; and Saafi (a.k.a. Safen). Codification of others has apparently proceeded since then.

The result is three "groups" of Senegalese languages that reflect history and number of speakers (without implying a hierarchy of value):
  • Those recognized early as national languages, which tend to have the largest numbers of speakers, and for which there is some notable amount of written material.
  • Those codified more recently, but which still need further research 
  • Those still needing further study towards codification.

Such a grouping of languages in a multilingual state might be a useful parameter to consider when planning out messaging on health topics such as ebola, or other development communication. The first group is readily used, with standard orthography, perhaps also having resources available to consult on the subject area involved - and it reaches the largest number of people. It may be possible even to work on materials outside of the country. The second group may require more effort, including consultation with experts in country. The third may require work mainly on the local level, and where there is not time to develop a standard orthography, use of phonetic transcriptions.

The experience in Senegal may have followed the above to a degree, but further information is needed. In any event, the first group is divided between Wolof, which was widely used, and the other 5 established national languages, which presumably had varying use (ebola info in Pulaar was on the web, for instance, and the others were  probably used on community radio).


Spelling again


As a small contribution to review of ebola materials, the ebola poster in Wolof above has one apparent misspelling - "Fagaroul" - with the French "ou" taking the place of the Wolof "u." It is interesting to note, however, that the same title and spelling is used on the French version of the poster (i.e., a Wolof title, with French influenced spelling, on a French language ebola poster).