The Alphabet That Will Save a People From Disappearing

As kids, two Guinean brothers invented a new script for their native language. Now they’re trying to get it on every smartphone.

Michael Everson / Katie Martin / The Atlantic

Twenty-six years ago, two brothers decided their native language needed a new alphabet. The scripts they’d been using to read and write their native Fulani, an African language spoken by at least 40 million people, weren’t working well.

Fulani’s sounds were rendered imprecisely by the Arabic alphabet, the script most often used to write it; the Latin alphabet presented similar problems. Neither the Arabic nor the Latin alphabets could accurately spell Fulani words that require producing a “b” or a “d” sound while gulping in air, for example, so Fulani speakers had modified both alphabets with new symbols—often in inconsistent ways.

“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.

It’s not every day that a new alphabet is born. The scripts in widest use today—Latin, Chinese, Arabic, Devanagari, Cyrillic—are all at least a thousand years old, and they each evolved from earlier alphabets. Creating a new script and getting it adopted widely is an enormous challenge.

The Cherokee alphabet is a notable success story. In the early 19th century, a man named Sequoyah created a script for writing his native Cherokee, which until then had only ever been spoken aloud. At first, many thought his scribbles were meaningless, and that he was playing tricks on people. But in a blind test not unlike the one Abdoulaye and Ibrahima would complete more than a century and a half later, Sequoyah and his daughter, who had also learned the alphabet, proved that the symbols they’d drawn actually represented words by reading what the other had written.

Sequoyah began to teach others to read and write, and his simple alphabet spread quickly among tribe members. In 1828, just seven years after Sequoyah invented the alphabet, the first-ever Cherokee-language printing press was used to publish the Cherokee Phoenix, a newspaper that was distributed for free to Cherokee speakers who didn’t know English. The press helped standardize and simplify the script, and made it possible to quickly publish newspapers and books.

Today, the computer-age equivalent of casting metal type in the shape of a new alphabet is encoding an alphabet in the worldwide typography standard known as Unicode. The Unicode Consortium, a nonprofit organization, develops standards for the way letters and numbers appear on computer screens. Since 1991, a group of technology companies have worked together to develop a universal character set—a system of coding letters, numbers, and symbols that allows computers to render any supported language in the world.

Cherokee was added to the Unicode standard in 1999, but it wasn’t until 2003 that Apple created a Cherokee keyboard for Mac computers. The first Cherokee keyboard arrived on Windows several years later. Today, Cherokee keyboards exist on iPhones and Android devices, and Facebook, Gmail, and Google Search all support the language.

The Cherokee Nation has put a particular emphasis on language development: The language program has 13 full-time employees, many of whom worked with Google for two years to implement Android support for the alphabet. Even so, its progress toward digital inclusion has been slow.

“A script is not a biological entity,” said Kamal Mansour, a specialist in non-Latin typography at Monotype who represents the company on the Unicode Consortium. “It doesn’t live alone. It has to have acceptance from people.”

“In Africa, there’s been a spate of inventions from very clever people,” he continued. “But many don’t gain acceptance, and they just die off.”

The Barry brothers had a long path ahead of them.

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 word is also an acronym for a phrase that translates to “the alphabet that will save a people from disappearing.”

They started by teaching their friends and family, asking each student to teach three others. They transcribed school books—algebra, geometry, history—using their script, copying them into the very same blank composition books that were home to their first sketches.

In junior high school, they started visiting neighborhoods and markets to teach larger groups of people. Their big break came in 1993, when Abdoulaye—still a high schooler—traveled to Conakry, Guinea’s capital city, during summer vacation. There, he tracked down a famous radio personality and got himself invited onto his show, where he demonstrated the script live on air, as the host described what he saw. “That’s how the whole country learned about Adlam,” Abdoulaye says.

A handwritten document shows the stroke order for writing each Adlam capital letter
A document lays out the stroke order for writing each capital letter in Adlam (Courtesy of Ibrahima and Abdoulaye Barry)

By the mid-1990s, the brothers could walk into a market in a new part of Guinea and see people they’d never met reading and writing their alphabet. Many Fula people are nomadic farmers or traders, and the brothers credit their transitory nature with the writing system’s quick propagation.

When Abdoulaye and Ibrahima moved to Conakry for college, in the late 1990s, their work on Adlam slowed, but Ibrahima dove back in with enthusiasm toward the end of his five-year civil engineering program. Others began getting involved in the development of Adlam through a student group called “Winden Jangen,” Fulani for “reading and writing.” The brothers transcribed more and more books, still by hand, with no help but a supply of carbon-copy paper they received as a gift from their relative in the government.

But the new alphabet wasn’t sitting too well with others in power. During his last year in college, Ibrahima was arrested and put in prison for three months. He says he was never told exactly why he was imprisoned—officers raided a Winden Jangen meeting one day and hauled him away, he recalls—but he suspects that his work had made some people nervous. “Maybe they feared that he was trying to instigate something bigger because they did not understand the script,” Abdoulaye said.

When Ibrahima was released, he was invigorated rather than discouraged. He set about writing with even more gusto, and started a Fulani-language newspaper, writing each issue by hand and distributing photocopies. He traveled to neighboring Senegal, where Fulani is also an official language, to try and spread Adlam—only to find that it had preceded him. Someone told him a palm-oil trader had been distributing his books in Senegal, the Gambia, Sierra Leone, and the Ivory Coast.

The brothers soon realized that Adlam reached a crossroads. “We knew that the only way to spread Adlam faster and publish more books was to have it in computers,” Abdoulaye told me.

To bring the script into the digital world, Ibrahima and Abdoulaye looked overseas.

They didn’t know it yet, but the help the brothers would need was tucked away in a maze of an academic building at the University of California, Berkeley. That’s where Deborah Anderson, a linguistics researcher, manages a program called the Script Encoding Initiative. Since its founding, in 2002, the program has played a role in getting more than 70 scripts added to the Unicode standard, by helping to write detailed proposals for the organization’s technical committee and providing funding for the work.

With money from donors like Google, UNESCO, and the National Endowment for the Humanities, Anderson’s mission is to help encode scripts that might not otherwise end up in front of the Unicode committee.* That includes historic scripts—one of her current projects aims to drastically expand the set of supported Egyptian hieroglyphs, to help Egyptologists who can’t currently use search functions or copy/paste in their research—and scripts used by linguistic or religious minorities.

Preserving those living scripts through type can promote ethnic pride and identification, Anderson said, and encourage literacy and education in native languages. “By not including such scripts, the web will be relegated to the more politically prominent scripts and the languages using them,” she said in an email.

In Unicode’s early years, there was little focus on minority or dead scripts. The original proposal for the system, written in 1988, makes clear that the body would prioritize modern languages over “preserving past antiquities.” But the first Unicode standard, published in 1991, indicated that “less common and archaic scripts” could find their way into future versions. Indeed, Unicode 3.0, which was published in 1999, included Norse runes alongside modern non-Latin scripts like Cherokee and Ethiopic.**

Without Berkeley’s script lab, many of the alphabets currently in Unicode may never have been encoded. Since a lot of funding for technology comes from hubs in the United States, Europe, and China, languages spoken in those areas are best supported on computers and phones, while others haven’t gotten as much attention. Alphabets used to write languages spoken by communities without political clout are especially likely to be glossed over, potentially accelerating their demise as their speakers are forced to depend on technology that operates in English or French or Chinese instead of in their native tongue.

Being added to the Unicode standard is still only the first step in a script’s path toward digital relevance. Next, technology companies have to be convinced to support the alphabet on their desktop and mobile operating systems, and social-networking platforms like Twitter have to translate their apps and websites to work with the script. For users to actually be able to type in the alphabet, developers have to create virtual keyboards for smartphones, and keyboard layouts for computers made to work with other scripts. The entire process can take years.

Ibrahima and Abdoulaye moved from West Coast to West Coast: Abdoulaye went first, traveling from Guinea to Portland, Oregon, with his American wife in 2003, and got a night job as a revenue auditor for Hilton Hotels and Resorts while he studied finance. Ibrahima joined him in Portland in 2007.

In between work and school, the brothers saved up enough to pay for the development of the first Adlam keyboard and font. They found a software company in Seattle that makes specialty fonts, and, in 2008, they were able to type in Adlam for the first time. It was a huge step, but it came with disappointment: Since Adlam was not yet supported by Unicode, if they sent a document written in Adlam to someone without the font installed, the recipient would see only a nonsensical jumble of random, disconnected Arabic letters. That’s when the importance of being included in Unicode sank in.

For the next few years, development largely stalled. The brothers worked and studied, and continued supporting Adlam schools at home and in the United States. Ibrahima signed up for a calligraphy class in Portland, hoping to streamline the shapes of the letters he and his brother had first sketched on notebook paper decades ago.

At a calligraphy conference in Portland, he met an artist named Randall Hasson, who invited him to present Adlam at another event in Colorado Springs in 2013. (Hasson wrote about his first meeting with Ibrahima on his blog.)

In Colorado, an audience member asked Ibrahima what he wanted for his alphabet. “I said I want only one thing,” Ibrahima told me. “I just want the script to be encoded. To be in Unicode. That is my main goal.”

A type designer in attendance knew just who could make that happen. He introduced the brothers to Anderson, the linguist at Berkeley, who in turn put them in touch with Michael Everson, one of the co-authors of the Unicode Standard and a prolific writer of script proposals. As a part of the Script Encoding Initiative, Everson had already written upward of 60 proposals that had been accepted by Unicode—and he said he was willing to take this one on, too.

I asked Everson if any of the alphabet’s quirks presented challenges—the fact that it’s written right to left, for example, or the fact that the letters are connected to one another—but he said it was nothing new. What was new was that the script wasn’t yet set in stone.

Students learn to read and write Adlam in a classroom in Sierra Leone
Students learn to read and write Adlam in a classroom in Sierra Leone (Courtesy of Ibrahima and Abdoulaye Barry)

“Mr. Barry had learnt calligraphy and was working really hard on optimizing stroke order and [duration],” Everson said. “I was saying, ‘You have to stop tinkering with this.’ If he wanted this to be stable he needed to stop altering it.”

With partial funding from the Berkeley encoding lab—Everson says he volunteered his time and resources to cover the rest of the cost—he worked with the brothers and the Fulani community in the U.S. to create a proposal. “It’s very well developed and well designed,” Everson said of the script. “These guys should be really proud of themselves.”

Abdoulaye and Ibrahima traveled to the 2014 meeting of the Unicode Technical Committee in Sunnyvale, California, to iron out the last kinks in their proposal. That same month, the committee voted to approve Adlam.

“I can’t even describe the feeling,” Abdoulaye said. He tried anyway: “When we got the email that we were officially encoded, we were very, very happy.”

Open up the settings on any smartphone and scroll through the keyboard options. If you’re holding an iPhone, you’ll find Macedonian and Telugu and Catalan—but not Adlam. The alphabet was included in Unicode 9.0, which was published this summer, but technology companies haven’t leapt at the chance to use it in apps and keyboards.

Like Cherokee during the four-year lull between its induction in Unicode 3.0 and the first official support for its alphabet on Macs, Adlam is in purgatory. Neither Windows, nor Apple, nor Google supports it on their operating systems yet, and although Facebook recently added an option to use the network in Fulani, it used the Latin script rather than Adlam.

Last month, some hope came in the form of an enormous font family called “Noto,” the culmination of a five-year collaboration between Google and Monotype that aims to create a typeface that supports more than 800 languages, all in a unified style. It already covers more than 30 alphabets—every single one released in Unicode 6.1, including Cherokee—and in its next phase of development, Google says it will cover 100 percent of Unicode 9.0, which includes Adlam. The Google font is an important step, but it doesn’t yet mean Adlam will be supported on an operating system anytime soon.

Instead of waiting for official support for Adlam to arrive, some Fulani speakers have gone ahead and created their own apps. Two Android apps already allow smartphone users to text each other in Adlam (as long as both have the “Adlam SMS” app installed) and learn the Adlam alphabet on their device.

For their part, the brothers—now 36 and 40 years old—have focused recently on expanding the alphabet’s role in education. Adlam is now being taught in more than 10 countries in West and Central Africa, and it’s making it easier for speakers of different dialects of Fulani, from Senegal to Sudan to Cameroon, to communicate with one another.

Ibrahima recently finished writing the first-ever Adlam grammar book, and he’d like to create a Fulani dictionary in Adlam. Both brothers want to open more schools in Guinea and in nearby countries, with the hope of augmenting the French-speaking education system with one that teaches kids in their native language. The formal education system is exclusionary, they say, because it considers people illiterate if they can’t read and write French—the colonial language—rather than asking whether or not they can read or write their own native tongue.

“That’s one of the things that’s really hindering the progress in Africa and Guinea in particular,” Ibrahima says. “We spend so many resources learning new languages, even though we already have languages that we can use to develop our countries.”

Facebook is the next big tech target for the brothers. “Today, the world is connected through the different platforms,” Abdoulaye said. “Until these platforms support our script, it won’t be acceptable, especially to the young generation.”

Google, Apple, Microsoft, and Facebook all declined to comment on whether they planned to support the script. Given how long it took to get Cherokee onto phones, tablets, and computers—and how many other languages still lack official support—Adlam may not fully inhabit the digital world for quite a while still.

I asked Roy Boney, the manager of the Cherokee Language Program, what advice he’d give the Barry brothers. He said the biggest hurdle his project faced was his own community’s impatience. “It takes a while, especially for minority languages,” Boney said, adding that “some people thought this wasn’t necessary for the language.”

Howard Gutowitz, the CEO of Eatoni, a company that makes apps in hundreds of different languages, was less hopeful. Five years ago, Gutowitz was involved in an effort to get N’Ko, an alphabet for writing another of Guinea’s national languages, supported by cellphone manufacturers like Nokia. I called him up to see how it turned out.

“After banging my head against that wall for many years, I decided not to bother anymore,” Gutowitz told me. He thought he might have better luck with apps, and released a calculator app in 200 languages to see which would be most popular. The N’Ko version did pretty well, he said, as did a couple other surprising languages, like Tibetan. The app was a sort of market research, to see which communities are hungriest for localized software.

His advice for the Barrys isn’t too upbeat. “Well, the first thing that comes to mind is, ‘Give up,’” he sighed. “The less despairing remark is, ‘Yeah, build an app; that’s what we did.’”

If an app does well, he says, it can signal to tech companies that it’s worth sinking resources into supporting a particular language. “The appeal to equity or rightness or preservation of languages, I think that’s going to fall on deaf ears when it comes to manufacturers that care about market or money.”

But take the long view and things might not seem quite so bleak. Mansour, the Monotype designer, has been involved in type development for decades. He says it’s easy to forget how far things have come since the ’90s, when the internet was dominated almost entirely by the basic Latin alphabet.

“You couldn’t even type Baltic languages like Estonian or Lithuanian. You couldn’t! You had to buy an Eastern European supplement [software package],” Mansour said. “Now, nobody thinks of that. We collectively forget the stages of development that have happened. Now, you could even type Icelandic on an Arabic system, if you wanted.”

For speakers of Fulani, there’s still no easy way to type the only alphabet designed for their language. The script’s inventors—still its chief advocates—hauled their writing system all the way from hand-drawn shapes on school notebooks to the doorstep of tech companies, neatly packaged into a Unicode block. All that’s left is to hammer on the door until it opens.

* This article originally stated that the Script Encoding Initiative is funded in part by the National Endowment for the Arts; in fact, it's the National Endowment for the Humanities that contributes funding. We regret the error.

** This article originally stated that Syriac Aramaic is a dead language; in fact, modern Syriac dialects are spoken by small groups in several countries. We regret the error.