Kaveh Waddell at The Atlantic writes about the development of an indigenous alphabet for the Fulani language by Abdoulaye Barry and his brother Ibrahima. The title, “The Alphabet That Will Save a People From Disappearing,” is idiotic — there are at least 20 million Fula, and they’re not going anywhere — but the story is fascinating:

“Why do Fulani people not have their own writing system?” Abdoulaye Barry remembers asking his father one day in elementary school. The variety of writing styles made it difficult for families and friends who lived in different countries to communicate easily. Abdoulaye’s father, who learned Arabic in Koranic schools, often helped friends and family in Nzérékoré—Guinea’s second-largest city—decipher letters they received, reading aloud the idiosyncratically modified Arabic scripts. As they grew older, Abdoulaye and his brother Ibrahima began to translate letters, too.

“Those letters were very difficult to read even if you were educated in Arabic,” Abdoulaye said. “You could hardly make out what was written.”

So, in 1990, the brothers started coming up with an alternative. Abdoulaye was 10 years old; Ibrahima was 14.

After school, they’d shut themselves in their rooms to draw, filling blank composition books they brought home from the classroom with the shapes that would make up their new alphabet. They’d take turns drawing letters, and together, assigned sounds to the shapes they came up with.

Six months later, they had a working script. Like Arabic, its 28 letters were written right to left. But unlike Arabic, whose short vowels are written as diacritical marks above and below letters, the script assigned its five vowels proper letters. It looked something like a cursive version of Ethiopic. Ibrahima and Abdoulaye’s parents started taking their project seriously, and invited one of their father’s relatives, who had an influential post in the local government, for a demonstration.

The visitor tested them: With Abdoulaye in the other room, Ibrahima would take dictation. When Abdoulaye returned, he read aloud what his brother had written. They switched and repeated the test. Over and over, the brothers consistently read out the right sounds, even those unique to Fulani. Crucially, they spelled the same complicated words in the same ways, independently of one another.

The visitor turned to their father. “Oh, yes, these kids are being serious,” he said. […]

During the decade after that first big test in the brothers’ house, their new alphabet—yet unnamed—spread at an astounding rate. Eventually, it would come to be called Adlam, after its first four letters: the equivalents of a, d, l, and m.

The brothers encountered a lot of obstacles in their efforts to gain official recognition for their alphabet, but it seems to be doing well, and people are spontaneously adopting it and achieving literacy with it (always a good sign, since invented writing systems are many and successes few). You can read more about it at The Randall M. Hasson Blog (first installment of a three part series).


  1. Ian Myles Slater says

    I knew I had seen the name Adlam elsewhere, and fairly recently.

    Checking an obvious source, I found that Adlam is included in the current Unicode Standard (9.0); it is described in Chapter 19 (African Scripts), section 9. (So I assume that is where I encountered it.)

    The account there does not mention the ages of the brothers when they began developing the script, however.

  2. David Marjanović says

    Actually, the first mention is in the second part; the first is just a history of alphabets and fonts.

    Read the comments on the third part, too.

    “The name Adlam is an acronym derived from the first four letters of the alphabet (A, D, L, M), standing for Alkule Dandaydhe Leñol Mulugol (“the alphabet that protects the peoples from vanishing”).”
    The Pffft! of All Knowledge

    Adlam is included in the current Unicode Standard (9.0)


  3. So at least this is one dumbass headline that can’t be blamed on the copy editor.

  4. It seems plausible that a standardized and well-structured writing system could help increase literacy. And aren’t non-literate languages at greater risk of losing speakers?

  5. Yes, except that the Fulɓe already have two perfectly good writing systems, one Latin and one Arabic. Bothe have some minor discrepancies from one country to another, on the lines of Swiss vs. non-Swiss Standard German. Literacy in Fula (which Ethnologue divides into 9 separate languages) is low, but functionally the writing systems are quite usable.

  6. Yeah, that’s one thing I found hard to understand: why were all these people (assuming the linked story is accurate) so grateful to finally be able to achieve literacy in their own language? It’s not as if it were unwritten, and the Latin script (I can’t speak to the Arabic one) is perfectly usable. (And, of course, a lot more helpful to non-native speakers, though that’s irrelevant to native speakers.)

  7. David Marjanović says

    Yes, except that the Fulɓe already have two perfectly good writing systems, one Latin and one Arabic.

    The blog post makes clear that the Arabic one, at least as used in Guinea, is anything but “perfectly good” or indeed a “system”: people use Arabic letters with idiosyncratic results, apparently much like Middle English with lots of vowels omitted, and find each other’s writings very hard to read.

    The Latin one appears not to be widely known in Guinea at least, and there may also be religious objections to it.

  8. David Eddyshaw says

    The language isn’t in any danger in general, has influenced quite a number of smaller languages, and indeed is a regional lingua franca in Northern Cameroon. Lots of Nigerian Fulani now only speak Hausa, though, and various Fulfulde-speaking groups are often in a far from dominant position in the countries where they live. I don’t know about Guinea-Conakry.

    The Fulbe are a cultural rather than genetic group. Like Tuaregs, they vary a lot in physical appearance. The wikipedia article sadly recycles the evergreen nonsense about them being of North African origin, which basically goes back to pre-Greenberg misidentification of the language as “Hamitic” along with the straightforwardly racist idea that the Fulani must be a bit whiter than other Africans because they were the rulers, and Africans couldn’t possibly be up to creating their own empires. There’s similar stuff out there about the Soninke, for the same reason. (All still there in mainstream history books up until the sixties, alas.)

    The paler-skinned Fulani are in fact the cattle-raising nomads found right across West Africa; the Fulani who ended up in charge are the darker-skinned settled Fulani.

  9. David Eddyshaw says

    Part of the thing about writing and literacy is probably that the Fulbe, especially the empire-building sort, are pretty solidly Muslim. So literacy would mean literacy in Arabic, and if you were to write your own language at all you would naturally use Arabic letters – not a happy solution for any language at all except Arabic itself and maybe its Semitic sisters at a pinch. There may well be some ideological objection to Latin letters, though widespread literacy in pretty much any African language is the happy exception rather than the rule, for all kinds of reasons to do with sheer unavailability of materials and general lack of educational opportunities all round. The idea that an indigenous *script* would help is probably a romantic illusion, given the all too practical reasons why literacy hasn’t taken off except in colonial languages. Mind you, it worked for Cherokee – for a while.

  10. January First-of-May says

    No relation to Adlam, but whichever time period the Book of Kells is actually from, it’s most definitely not 4th century (and certainly not 384 AD, as stated in the linked article).

  11. David Eddyshaw says

    If there *is* any extended and readily obtainable written material in an African language it’s usually the Bible translation. There is no similar pressing incentive to translate the Koran for Muslims, and much less reason for most Muslims to read a translation of the Koran if there is one, than for Christians to read the Bible in their own language. (There is a Hausa version, maybe *because* Hausa literacy is widespread. There is quite a thriving trade in Hausa novels.)

    Why learn to read in your own language if there’s nothing to read once you’ve finished the primer?

  12. : why were all these people (assuming the linked story is accurate) so grateful to finally be able to achieve literacy in their own language?

    I’m pretty sure the missing piece of the picture is N’Ko. The principal languages of Guinea are Manding and Fulani, and the former is more and more often written in N’Ko, by far the most successful and prestigious of the many, usually ephemeral, locally invented alphabets of West Africa, equipped with a self-improvement ideology and an extensive library of religious and didactic texts. In such a context, the feeling that only a unique alphabet will properly express Fulani identity is naturally going to be much stronger than elsewhere.

    There are substantially different Arabic orthographies for Fulani depending on which part of West Africa you look at. Guinea (more specifically Futa Jalon) happens to have a particularly strong and early manuscript tradition of writing Fulani in Arabic script, but, judging from the one example I’ve seen, the orthography used there is less than ideal. Unlike most traditional West African Arabic-based orthographies, it does make it possible to distinguish e and o from i and u, but unfortunately to do this it merely adds a dot that indicates “non-Arabic sound here”, which could also apply to a consonant, so (for instance) bi+dot could be read as be, pi, mbi, pe, or mbe. But it wouldn’t surprise me if they’ve improved on this system since.

  13. I think Lameen has found the key factor: the success of N’ko. In turn I think the success of N’ko needs to be explained: I can’t help but note that Colin Masica wrote that in South Asia it is common for minority language activists to create new scripts because the multiplicity of languages and scripts in South Asia created the perception that, in order to be deemed “respectable”, a language needed its own writing system.

    I suspect that in West Africa the fact that the two major prestige languages (Arabic on the one hand, French/English (according to country) on the other) each have their own script had, in like fashion, played a major role in the success of N’ko: since Arabic has its own script, as does French or English, it was widely accepted (although probably never stated outright) that to be taken seriously a written African language needs to have its own script too. Whether this will allow Adlam to become a widespread script remains to be seen, of course.

    And I do not believe the existing imperfections found in the Latin- or Arabic-based orthographies of Fulani suffice to explain the rise of a wholly new writing system: David Marjanović ‘s point about the Fulani in Arabic script in Guinea being “Middle English-like” in its phonological (in)adequacy does suggest this, since from the Middle English period to the present no non-Latin script was ever a serious competitor to the Latin script.

  14. The bafflement over why a new writing system was received so gratefully when there were already two adequate ones seems to me a kind of linguistic nerdview, to be honest. Writing systems are about so much more than the ability to record or reproduce utterances.

    Imagine a country where there were three groups of people, one of which got to use iPhones, the other jPhones, and a third which had to buy secondhand i- or jPhones and use laborious hacks and workarounds to get any use out of them. The invention of a convenient and user-friendly kPhone for the third group would be greeted with jubilation by that third group, even if it didn’t technically allow them to do anything they couldn’t already have done with slightly more effort using their janky jailbroken jPhones.

    (In this analogy, China has a tradition of beautifully hand-scrimshawed cases for gigantic “brick” satellite phones that has become part of the national identity, Japan has the same thing except everyone also carries a walkie-talkie for sidechannel clarifications, and Korea uses a form of telepathy invented by a king a couple of centuries ago.)

  15. a third which had to buy secondhand i- or jPhones and use laborious hacks and workarounds to get any use out of them

    Is this Japan?

  16. The bafflement over why a new writing system was received so gratefully when there were already two adequate ones seems to me a kind of linguistic nerdview, to be honest.

    A good point that had never occurred to me (being a language nerd).

  17. This is why the Latin alphabet never displaced runes, and why the Finns collectively embraced silence in preference to the glyphs of their warrior-cleric conquerors, probably.

  18. Literacy in Finnish wasn’t a grass-roots movement: it resulted from the adoption of Finnish as a national symbol by socially dominant swedophones who expected literacy (and the Latin script) to go with a national language. “Swedes we are no longer, Russians we don’t want to be, so let us be Finns!”

  19. David Marjanović says

    from the Middle English period to the present no non-Latin script was ever a serious competitor to the Latin script

    The difference there is that (with the runes long forgotten) approximately nobody knew that any other script even existed. In Guinea everyone knows at least two exist (and that’s without even counting N’Ko).

  20. Is this Japan?

    Clearly not, because when a speaker of the language invented a sensible and attractive orthography that suited the language perfectly, everyone was delighted and cheerfully adopted it! (Or so the article suggests.)

  21. @David, you should know that His Exquisiteness Count-Baron-Professor-Doctor von Bladet has room for an ox tongue in each cheek, and is not afraid to use them.

  22. Clearly not, because when a speaker of the language invented a sensible and attractive orthography that suited the language perfectly, everyone was delighted and cheerfully adopted it!

    Ah, yes…. mine was a silly question.

  23. Adlam covers not just native Fulani phonemes, but also Arabic pharyngealized consonants and velar and dental fricatives, as well as characters for labial-velar stops. The latter are presumably needed to represent loan words from other languages spoken in Guinea. The extensions for representing sounds in Arabic loans are mostly written by adding diacritics to letters for ordinary Fulani phonemes that are pronounced similarly.

  24. I wonder if the people suggesting here that the fact that a people like the Fula are Muslim means they must use the Arabic script for literacy are aware that a majority of the world’s Muslims (and a great majority of non-Arab Muslims in Africa) already write their native languages in non-Arabic scripts.

  25. a majority of the world’s Muslims … already write their native languages in non-Arabic scripts.
    Are you sure? I know that there are major Muslim populations that use non-Arabic scripts (Latin in Indonesia, Latin and Cyrillic for many Turkic people), but do they really sum up to a majority?

  26. Interesting question!

  27. Unlikely, I’d say. 200MM in Indonesia + 145MM in Bangladesh + 75MM in Nigeria + 75MM in Turkey + 30MM in Sudan + 28MM in Ethiopia + 27MM in Uzbekistan + 23MM in China + 17MM in Malaysia + 16MM in Russia + 16MM in Malaysia + 16MM in Niger + 10MM in Somalia = 668MM, or about 45% of the (conservatively) 1500MM Muslims worldwide.

  28. The interesting question may be how to allot Indonesian Muslims, numbering over 210 million (the most of any country). They speak many languages, and while probably the vast majority of Indonesian Muslims will be familiar with seeing their L1s in modified Arabic script, it is by no means the primary written form for all of them.

  29. Hat, John Cowan: it’s a very interesting question, but solving it is a little more complicated than simply comparing population figures! Speaking of which, John Cowan: I don’t see how one could consider the Sudan to be non-Arabic script-using. Conversely, one would need to add the 17 million Latin script-using Muslims of Tanzania (where did you get your list from?) : also, I suspect that the situation for Muslims in India must be messy, inasmuch as my understanding (can anyone confirm or deny this?) is that for those who have an L1 other than Urdu or Kashmiri (i.e. about two-thirds of the Muslims of India) the language of literacy will be another Indian language (i.e. one using a non-Arabic script). Which would mean adding over a hundred million people to John Cowan’s balance sheet, making non-Arabic script users a weak majority of Muslims.

    Things get messier still if we look at the number of actual literate individuals: let’s take two Muslim countries, one of which uses the Arabic script (Pakistan) and the other does not (Indonesia), both with comparable populations (circa 200 million). The literacy rate in Indonesia being about 90% and that of Pakistan 60%, I’d say we need to compare 120 million users of Arabic script in Pakistan to 180 million users of the Latin script in Indonesia, rather than 200 million people in Pakistan and Indonesia.

    Unfortunately this factor doesn’t seem (prima facie) to clearly “tip the balance” in either direction: Afghanistan and Yemen use Arabic script and have above-average illiteracy, for example, but above-average illiteracy is also true of most non-Arabic script script-using countries in sub-Saharan Africa. Muslims living in first world countries would also need to be counted as non-Arabic script users. Hmm. Considering the global increase in literacy worldwide and the fact that sub-Saharan African countries are the fastest-growing countries today, I suspect that a majority of the world’s Muslims will become literate in a non-Arabic script in the near future (assuming it isn’t true already, see above).

  30. The list comes from Wikipedia, s.v. “Muslim World”.

  31. David Marjanović says

    The list contains Malaysia twice, once as 17 and once as 16 million. Is the first a glitch for Tanzania?

  32. David: Good catch! I suspect you’re right.

    An additional comment to my earlier one: for China’s Muslims things are as messy, linguistically, as they are for the Muslims of India: the circa 10 million Hui are L1 speakers of Chinese and thus should be counted as non-Arabic script users, but the circa 8 million Uyghur speakers *are* Arabic script users. I have no idea as to what the linguistic situation (regarding literacy) is for the smaller Muslim ethnic groups of China: can anyone help?

    Oh, and a clarification: in addition to Urdu and Kashmiri there is another constitutional language of India which uses the Arabic script, Sindhi, but the bulk of its speakers (in India!) are not Muslim.

  33. To calculate this, the first-approximation best strategy is to ignore L1s (that’s what the schools do, after all) and simply look at official languages, since that’s what the published literacy rates will reflect in most cases. But that strategy ignores the fact that in most Muslim countries kids learn to write at least a few chapters of the Qur’an in Arabic, so a lot of people who are officially only literate in Latin script actually know Arabic script as well. This fact is especially important for West Africa, where a lot of people counted as illiterate for census purposes actually write their own language regularly in Arabic characters.

  34. Osman Barry says

    I really love this Adlam


  1. […] Hat reports on a new script for the Fulani […]

Speak Your Mind