Metaglossia: The Translation World

"Featherstone's show highlights the diversity and artistry of Island youth.

By Abby Remer -June 5, 20240
See the Garden Gate at Featherstone's 25th anniversary art show now through June 16. —Dena Porter
Austin Awad, 6, and his parents Mark and Anastacia were full of pride as Austin showed his colorful art work entitled "The Volcano." —Dena Porter
Gianni Hejduk-Petrucci created this mixed-media piece entitled "Beauty in the Eye of the Reptile." —Dena Porter
Megan and Joe Harper stood proudly with their 6-year-old daughter Savannah who created the acrylic painting "Magical Tree." —Dena Porter
Featherstone president emeritus Marilyn Wortman, contributors Preston and Connie Williams, and board director Bob Avakian enjoyed this milestone event. —Dena Porter
Consultant Takiema Bunch, left, with Garden Gate co-founders Dawn Warner and Leigh Ann Yuen. —Dena Porter
“The 100 Languages of Children” teems with the creativity of Garden Gate Child Development

Center students and alumni, as well as youth from Featherstone’s Children’s Art Program, the Martha’s Vineyard Regional High School, M.V. Community Services, the Wampanoag Tribe, and the homeschool community.

The show at Featherstone Center is a glorious celebration of Garden Gate’s 25th anniversary.

The school, situated on Featherstone’s campus, is founded on a creative approach to learning. Its classrooms are designed as studio spaces where students use traditional and unexpected art materials to explore, create, and learn. Garden Gate embraces an arts-based approach to early education that nurtures creativity, problem-solving, and social competence.

Drawings, photographs, sculptures, ceramics, jewelry, and textiles adorn the light-filled Francine Kelly Gallery, honoring the creativity of Island youth. The diversity of the 135 pieces reflect the many-faceted “languages” Vineyard youth use to express themselves.

Two collaborative works grace the entryway. To the right is an immensely endearing, expressively colored baby cow, which the students saw daily in the field adjacent to their classroom. The tenderness between mother and calf melted everyone’s hearts and became a welcome grace during the pandemic. The children painted it with artist and Garden Gate alum Jack Yuen, and the work has come to symbolize the school, serving as a testament to the power of arts and nature to soothe and inspire.

To the left is “Neighborhood,” an impressive watercolor inspired by Romare Bearden’s “The

Block.” Dawn Warner, co-director at the preschool, explains that one of the teachers had seen Bearden’s collage in New York, which, in true Garden Gate fashion, spawned a project where students created a life-size cardboard bakery, bathroom, museum, and library, as well as this piece, which shifts in scale and perspective. “We work in our interests and hope to inspire the kids with the larger pieces of art in the world,” Warner explains.

A museum trip prompted a grouping of renderings of ships in all shapes and sizes. Students visited the M.V. Museum to learn about what life was like going out to sea during the whaling era. Learning about shipwrecks in a more modern era led to studying the Titanic. Some of the imaginative works that spun out from the topic are truly intricate white and brown pen drawings on black paper of giant ships, including John Frank Fiorito’s “The Floating Hotel,” which could be a Steamship Authority ferry on steroids.

A grouping of mixed-media works done on the easels reflects the young 3- and 4-year-old artists’ keen vision. “Each started as a painting on the easel, which are always open,” says Garden Gate co-director Leigh Ann Yuen. “They then added a second layer with tempera paint sticks; the third was oil pastels. It was the first time we worked with these kids to title their work, and they’re my favorite part.” Arlo McCarty’s “Rainbow Princess” and “We Eat Candy” by Emmett Piper-Roche are just two of the intriguing set. “That was a really interesting process for them to think of a title,” Yuen says.

There are also engaging titles among the delightful panels of individual interpretations of flowers, including Anastasia Ledden’s “Magical Beauty” and Dorian Markovic’s aptly named “Famous Dorian’s Flower.”

Featherstone’s art students shine as well. Reagan Ready’s “Mystical Maple” reaches toward the heavens with the spirit of angels simmering between a recognizable image and abstraction. Orion Thibodeau’s acrylic dabs of paint in “The Meadow of Color” swim across the canvas like lightning bugs. Ellis Flynn’s deftly controlled spray of pigments reflects a keen eye and finesse of technique.

Other schools are represented, too. “Matisse Fish Bowl,” by Kennedy Garrison from Edgartown, captures the French modern artist’s bold use of black outline and intense color to create distinct imagery. While nodding to Matisse’s famous painting of the same subject, Garrison has made it wholly unique.

“Pink Flower” by Abigail Buckley from the Oak Bluffs School has the impact of 1960s Pop Art, as the single, spiraling purple blossom pushes against the boundaries of the frame that can barely contain its organic energy.

Austin Awad, from the Martha’s Vineyard Public Charter School, animates the crowing avian in “Rooster” through skillfully applied fields of color and robust outlining.

Among the many impressive Martha’s Vineyard Regional High School artworks is Brody Royal’s “Fish in Net.” The slim ceramic fish are artfully caught in a bronze wire net hanging between two wooden dowels. The creatures cannot move up, down, or sideways, suspending them in time and place.

Caleb Dubin’s “Moon Glow,” “50/50,” and “Orbit” show an astounding versatility of creative talent in the choice of materials, glazing techniques, and forms. Tessa Schultz makes us smile with the crisp, large-format photograph “Vacation-Over.” A single arm, hand outstretched, grasping for help or at least attention, sticks straight up from the sand on the otherwise completely abandoned beach.

“Summer Waves,” by Hollis Oliver from Vineyard Montessori School, is a stunning close-up in which each swooping brushstroke conveys the power of the ocean in all its glory.

The exhibition’s title, “The 100 Languages of Children,” comes from a poem by Loris Malaguzzi, describing how children communicate and make sense of the world. The first stanza speaks to them having a hundred languages, thoughts, ways of thinking, playing, and speaking. The second beautifully reflects the wealth visible in the gallery:

 

Always a hundred

ways of listening

of marveling, of loving

a hundred joys

for singing and understanding

a hundred worlds

to discover

a hundred worlds

to invent

a hundred worlds

to dream.

 

“When the children walk in and see their work hung in the gallery, they just swell,” Warner shares. “They see themselves as artists — people who are part of their community and have value. They feel on very even footing with folks on this campus. We don’t want that lifelong love of art to fade. We want them to have this start that keeps them going in the arts.”

 

“The 100 Languages of Children: Celebrating Creativity and Community” is on view through June 16 at Featherstone Center for the Arts. Open daily, 12 to 4 pm."

#metaglossia_mundus

"05/06/2024 Rachael Davies LEGO Braille Bricks expands to six more languages with top-up packages, available to buy now on the Pick a Brick online service.
We spotted LEGO Braille Bricks on Pick a Brick earlier this week and now, thanks to a press release from the LEGO Group, we know why just a few bricks were added. It’s part of a way to expand the LEGO bricks for the blind and visually impaired to more languages through top-up packages. By adding top-up packs to existing English and Spanish sets, this tactile building experience is now accessible for Danish, Swedish, Finnish, Norwegian, Dutch and Portuguese-speaking families.
40655 Play with Braille – French and 40656 Play with Braille – English first launched in September 2023, soon to be followed by 40722 Play with Braille – German Alphabet, 40723 Play with Braille – Italian Alphabet, and 40724 Play with Braille – Spanish Alphabet, available for £79.99 / $89.99 / €89.99.
Now, you can buy pre-selected LEGO Braille Bricks featuring special characters from the additional alphabets on Pick a Brick to allow for more language combinations. Danish, Swedish, Norwegian, Finnish, and Dutch top-up packs are compatible with 40656 Play with Braille – English while the Portuguese top-up pack is compatible with 40724 Play with Braille – Spanish Alphabet. An element guide helps users combine the top-up packs and LEGO Braille Bricks sets for optimal play value, which includes removing some of the existing bricks and exchanging them with the new elements.
lego
“Play has the power to change lives,” said Rasmus Løgstrup, LEGO Group Lead Designer on LEGO Braille Bricks. “When children play, they learn vital skills, so we were thrilled by the reception that LEGO Braille Bricks received in educational settings and the thousands of requests to make them broadly available meant we just had to make it happen!”
Support the work that Brick Fanatics does by buying your LEGO sets using our affiliate links. Thank you!
Rachael Davies
I write about all the very best fandoms – and that means LEGO, of course. Spending so much time looking at and talking about LEGO sets is dangerous for my bank balance, but the LEGO shelves are thriving. You win some, you lose some."

#metaglossia_mundus

"3 June 2024 Darién migration goes global, language becomes a major challenge

‘Sometimes officials demand things, but we have no idea what they want.’

 

The Darién Gap: The reality behind the numbers: This ongoing series explores the risks, complexities, and dynamics surrounding one of the world’s busiest but most neglected migration routes, on the border of Colombia and Panama. For more context and background, read this story. 

NECOCLÍ, Colombia

Two dozen people with Chinese passports in hand and heavy luggage in tow line up at Necoclí dock ready to depart for the Darién Gap. Boats wait to take them across the bay to the town of Acandí – the gateway to the treacherous 100-kilometre jungle trek from Colombia into Panama that is the only overland route towards the United States.

The Chinese migrants pay street vendors in US dollars to wrap their belongings in plastic and avoid damage during the 40-minute ride. None of them speak Spanish. Instead, they negotiate via hand signals: Vendors charge them three to four times more than the Spanish-speaking migrants paying in Colombian pesos.

The group is managed by a Colombian handler, who bundles their passports and speaks with the workers at the docks, organising passage, sorting their tickets and checking off the names he has on a clipboard.

Until last year, the vast majority of migrants crossing the Darién Gap came from Latin America and the Caribbean, but while Venezuelans and Ecuadorians still make up the two largest groups, the jungle stretch has now turned into a global migration hotspot. 

In 2023, migrants from over 100 countries made the passage, those from China being the fourth biggest group and the largest outside the Americas, according to official Panamanian data. Afghanistan came in at number seven, and India in tenth spot. The list increasingly includes countries from across Asia, Africa, and the Middle East.

The annual numbers attempting the journey have soared from tens of thousands to hundreds of thousands in just a few years: Last year, more than 520,000 people took the path, and more than 110,000 traversed the route in the first three months of 2024.

 

The internationalisation of the influx through the Darién has been sudden and is putting pressure on aid groups and governments to adapt to the specific needs these new migrants have. Their lack of Spanish language skills and different cultural mindset often put them at even greater risk than Latin American migrants. 

“Numbers in the Darién Gap, in a sense, act as an international barometer, and what it is telling us is clear,” said Bram Ebus, consultant in Colombia for the International Crisis Group (ICG). “This is an international crisis.”

The key drivers

The United States has become an ever more attractive destination. The number of migrants encountered at the Mexico-US border – both those apprehended and those expelled – soared from approximately 1.7 million in 2021 to nearly 2.5 million in 2023. 

A number of factors have been contributing to this trend.

Post-pandemic inflation rises and economic slowdowns have been important push factors in many countries, including China.

Lack of integration, restrictive migration policies, and mounting xenophobia across South America have also been pressuring many of the more than 7.7 million Venezuelans who fled their country’s economic collapse since 2015 to migrate for a second time. Other countries have their own particular circumstances. In Afghanistan, for example, the return of the Taliban and a series of droughts, floods, and earthquakes are adding to an already complex humanitarian crisis, driving many to migrate. 

Conflict is a key factor for many. According to global conflict monitoring group ACLED, 2023 saw 12% more conflicts compared to 2022, and a 40% increase from 2020. This amounts to one in six people living in active conflict zones around the world.

Migrants from Somalia, Haiti, and Syria, all described conflict – or fear of conflict – as one of the main reasons for their decision to apply for asylum in the United States.

Climate change has also become a key driver: extreme weather patterns causing natural disasters and turning more regions into “climate hotspots” where biodiversity is disappearing and agriculture has become almost unviable. Nepal, Bangladesh, and India – high on the list of Darién Gap nationalities – are among the countries more vulnerable to climate risks.

Once in the Americas, the increasing trend to impose visa restrictions and to militarise borders in several Latin American countries makes the Darién Gap the only overland pathway for those wanting to cross into Central America and head northwards. 

For some nationalities, Nicaragua provides one alternative. The government eliminated visa requirements for Cubans in 2021, leading to a significant drop in the numbers of Cubans crossing the Colombia-Panama border. In a move that spurred tensions with the US government, Nicaragua began offering entry to Haitians and an increasing number of African and Asian countries via charter flights. But these flights and visa costs are steep and out of reach for many. 

US visa requirements and domestic migration policies are also making it harder for migrants from outside the continent to enter the United States legally, forcing them to go through Latin America.

This all means that a growing number of people from Asia, Africa, and the Middle East are feeling compelled to travel north through the Darién Gap to the US border to ask for asylum, joining hundreds of thousands of Venezuelans, Haitians, and Ecuadorians.

The language barrier

Ahmed is one of them. A recent graduate in his mid-twenties, he has travelled from Kismayo, a port in southern Somalia, to Las Tecas – a migrant camp at the entrance of the jungle trek on the Colombian side of the border, about eight kilometres from Acandí.

When The New Humanitarian encountered him in April, he was waiting at a small aid station run by Médecins du Monde (Doctors of the World). The woman he was travelling with, also Somali, wanted to take advantage of the health services before they embarked on the gruelling hike into Panama, which takes three to five days. 

Ahmed speaks English, Somali, and a bit of French. He translated for her as she interacted with the doctors, from English to Somali. It was the first opportunity she had had to speak with medical professionals since their trip began. 

Médecins du Monde has workers who speak English and Spanish on site. For other languages, they have access to translators who work by phone, but they are not always available, which sometimes forces them to communicate via Google Translate.

Aid workers told The New Humanitarian that the majority of humanitarian groups operating in the region are yet to adapt to the influx of non-Spanish-speaking migrants.

Médecins du Monde is among those trying. It is in the process of expanding its translation capacities by posting French-speaking staff at its station in Las Tecas and employing more translators in Necoclí, where many aid organisations operate.

A worker from GIFFM – the Colombian organisation that coordinates aid responses between the government, UN agencies, and private efforts – told The New Humanitarian that the language barrier is particularly problematic during medical emergencies, when using a translator over the phone can be far from ideal.

Speaking on condition of anonymity, they described the case of a woman from Kenya who sought treatment for a mental health emergency in Necoclí. Her distress prevented her from communicating clearly what language she spoke, delaying treatment by hours until a translator could arrive in person, which resulted in additional anguish for her and complicated her care.

Compounding the issue is the fact that communicating with those in authority is also difficult. According to Ahmed, migration officials in Necoclí spoke exclusively in Spanish, as did most police during his journey. “Sometimes they demand things, but we have no idea what they want,” he said.

A way around visa requirements

In early 2023, Ahmed had never heard of the Darién Gap, but he was already trying to leave his native country out of fear of being forced to join a gang or extremist group.

“They don’t give you a choice,” he said.

He tried to move to Kenya, but was denied asylum. Then he saw “a TikTok video posted by a Somali friend in America”, recorded in the Darién Gap. After researching online and in migration group chats, he decided to follow his friend’s example.

 

Ahmed organised the trip himself, but many migrants in the Darién region plan their intercontinental odysseys with “travel agencies” – some of which are run by transnational criminal organisations, according to Ebus.

A Chinese migrant in his early twenties who spoke to The New Humanitarian and plans to go by the name of “Bobby” when he arrives in the United States used one such “agency” to plan his route. 

Bobby, who asked that his real name be withheld, is among a growing number of middle-class Chinese nationals who have chosen to leave their country due to strict lockdowns, dire economic conditions, and restrictions on political freedom. 

Between January and April 2024, more than 24,000 Chinese people were registered at the southern border of the United States – more than during all of 2023, according to the US Department of Homeland Security.

Of the dozens of migrants from outside of Latin America who spoke to The New Humanitarian in the span of three days, most said they began their land journeys in Peru or – more commonly – Ecuador, because it doesn’t require visas for most nationalities, and migrants often only need to provide proof of onward travel before being granted a tourist visa.

Easy targets 

Migrants from Africa, Asia, and the Middle East don’t only face linguistic challenges. They are also more vulnerable to corrupt security forces and criminal groups, especially as the police and the cartels know they tend to travel with more money.

Ahmed’s group was robbed by Peruvian police, who pulled them off a public bus at a security checkpoint and demanded money. They didn’t have enough on them to satisfy the officers, who then took the group to an automatic teller machine under threats to withdraw more. When one of Ahmed’s companions refused, he was beaten and had to go to the emergency room, Ahmed said.

Venezuelans who enter Colombia formally have some limited access to services, but migrants who don’t are generally excluded from healthcare, which operates under a public-private hybrid system that relies on health insurance coverage for payment. The government provides this insurance for free to less wealthy Colombians, and also offers a system of subsidies, but most migrants don't qualify for these.

“If a transient migrant shows up at an emergency room with life-threatening injuries, medical officials are legally bound to provide treatment,” explained Luisa Fernando Gómez, communications director for the health secretary in Apartadó – a city that serves as a regional hub in northern Colombia for migrants travelling on to the Darién. 

“But for other conditions, including pregnancies, or complications from long-term health conditions, they will be denied treatment and referred to private doctors,” she added.

Prenatal care and paediatric services are becoming vital for migrants in the Darién region as families increasingly undertake the journey. More than 30,000 children crossed the Darién Gap in the first four months of 2024, according to a May report from UNICEF – a 40% rise compared to 2023.

Fernando Gómez said the Apartadó mayor’s office – in conjunction with the Colombian Red Cross and UN agencies – is trying to provide basic services via roving “mobile healthcare units”, especially for prenatal care and malnutrition.

“But these are for everyone,” she explained. “While they provide treatment to some migrants, they are not specifically targeted at migrant populations,” she added, urging NGOs to step in more and fill the gap. 

These obstacles, however, are not stopping migrants from persevering. 

At the time of publication, Ahmed was still trapped in Tapachula on the Guatemalan-Mexican border. He had been robbed twice more in his trip – once by Guatemalan police, and again by other migrants when he arrived in Mexico. 

“I am going to have to beg my uncle for money,” he told The New Humanitarian via text messages. “I don’t have the resources to continue north.”

Edited by Daniela Mohor."

#metaglossia_mundus

"TEZU, 3 Jun: The Forum of Library Activists (FLA), Medo along with Bamboosa Library, Tezu and the Lohit unit of Arunachal Pradesh Literary Society (APLS) organized a two-day translation workshop for production of ‘Pratham’ books in Kaman Mishmi language here in Lohit district from June 1. They also celebrated the birthday of Lummer Dai.

Inaugurating the workshop, indigenous affairs director Sokhep Kri spoke on the brilliant literary achievements of Dai and expressed hope that the younger generation will follow his path to develop Arunachalee languages.

 

He also exhorted the teachers to start teaching the Mishmi language textbooks from this academic year.

Informing about the ‘translation workshop’ that was conducted in December 2023 at Bamboosa Library, FLA secretary Keselo Tayang appealed to all the participants to contribute their best.

“Pratham books, under the copy-left arrangement, freely permits translations of their books into any language by anybody, without any formal permission from Pratham. Under four age groups, books can be freely downloaded from the Pratham  website www.storyweaver.org and translated. This is ideal for minor script-less languages, like Arunachal,” the organizers stated in a release.

During the workshop, the participants prepared manuscripts of six books and also prepared audio and video versions of the books. They were guided by Kri, who is also the convenor of Lohit Youth Libraries and APLS Lohit unit president Alenso Chai."

#metaglossia_mundus

"Irene Solaiman spoke with Rest of World about tackling racial bias in AI, low-resource languages, and why there are so many AI girlfriends.

Irene Solaiman on tackling racial bias, language equity, consent, and why there are so many AI girlfriends.

Photography by Greg Kahn for Rest of World

By RUSSELL BRANDOM

3 JUNE 2024

TRANSLATE

 

Q & A

What the AI boom is getting wrong (and right), according to Hugging Face’s head of global policy

As the competition between massive artificial intelligence companies heats up, the repository Hugging Face has emerged as a rare point of neutral ground. Built as a GitHub-style clearinghouse for open-source data sets and models, the site has become a vital resource for anyone working in AI. Without the regulatory baggage of a giant like Meta or Google, Hugging Face has also become a voice of reason in the policy world, advising regulators around the world on the unique promise and risk of AI, while leading its own technical work on bias assessment and watermarking.

The company’s head of global policy, Irene Solaiman, is at the center of that work. A former public policy manager at OpenAI, Solaiman was the first person to test ChatGPT for social-impact bias. Now, her team at Hugging Face is advising regulators from the U.S. to the European Union on how best to approach the nascent AI industry, and how to navigate thorny issues of bias, consent, and existential risk along the way.

Rest of World spoke with Solaiman about the promise and risk of the new generation of  large language models — and how to build them with the rest of the world in mind.

This interview has been edited for length and clarity.

I want to start off with the major AI scandal of the month, which is the lawsuit between Scarlett Johansson and OpenAI. It seems like an alarming case for AI policy professionals. Is there anything that surprised you about the case?

On the legal side, the biggest implication I see is establishing precedents for future action. There’s this question of how much Scarlett Johansson has a right to likeness as a voice actor. But no case explicitly like this has existed, so we’re going to find out as things shake out what we can expect in the future.

At the same time, the whole case brings to light the emotional connections people are having with AI. There’s a lot of AI girlfriends out there. Particularly girlfriends and not boyfriends. I’m not a psychologist but I do think that has serious implications for how we engage with AI in a loneliness epidemic.

You’ve said before that you actually had your voice cloned without your consent, as part of a product demo that backfired. Do you think there’s a bigger problem with how the tech world thinks about consent?

Part of it is just the tools: We’re way behind where we need to be in being able to sort through these massive amounts of training data. And that comes out when you think about consent, but also accuracy. 

So when I had my voice cloned, the platform does say that you need to get consent of the data subject. But the enforcement is just a little checkbox that says, “I received consent.” And sometimes it’s ambiguous who owns the video or is in a position to consent. In my case, the training data was from a public video of a talk I’d given — actually, a talk about the importance of getting consent from data subjects.

Bias has emerged as one of the trickiest problems in AI development. On one side, we’ve reported on the stereotyping effect in image-generating models. On the other, Google ended up publicly apologizing for Gemini’s image generator after it introduced a range of genders and ethnicities into queries where that diversity didn’t make sense. Is there a way to handle this that makes sense? Or are companies just going to keep stumbling into political problems here?

So, one important part is moving beyond the models to the larger system. Image generators like Gemini are, by default, incredibly visual. And while I can’t say specifically how Gemini was built, I think it’s unlikely that people are directly prompting the model. There’s layers of systems that come with how people interface with a consumer-facing product. It’s really hard to find where biases are introduced. And probably the answer is, at every point. But this is part of why people are investing in evaluations, interpretability, and red-teaming.

When you look at the whole system, it also includes the data set. I’ve said for a while that we need to glamorize data-set research a whole lot more. It’s not my area, but Dr. Abeba Birhane has done some of the best work around looking at multimodal data sets. 

 

Because you’re drawing from data that human beings created, a lot of it comes down to amplifying existing social norms. So, what proportion is representative? Are we overindexing on one specific population and its history and its infrastructure? In the end, you can never be fully unbiased because perspectives will differ, especially across the world, but even within one country, one city, or one family. What I view as unbiased might look very different from somebody with different political beliefs, from a different upbringing. 

This is where AI is facing some similar parallels to social media. I remember discussing these issues with social media platforms 10 years ago. How do we treat existing norms? We don’t want to do social engineering, but if you’re only amplifying the existing norms, isn’t that still a kind of social engineering? Most of these problems are not specific to AI, but the way that we build, measure, and mitigate them look very different in an AI context.

 

The Gemini case was focused on image, but I imagine it’s even more complicated in text, which is where even more AI development is happening.

Right, the way that we react to consume, measure, and mitigate biases and stereotypes is quite different in image than what we would do in text. I did the first social-impact bias stereotype testing on OpenAI systems way back in the day. I did the first non-Latin character testing on OpenAI and GPT systems. I did it in Bangla [Bengali] because it was the only other language that I knew that didn’t use Latin characters. Which is to say, representation matters.

“There’s so much infrastructure that leads to more data being available on specific languages, even down to the keyboards that the internet was built for.”

There’s also the question of low-resource languages. We’ve seen models really struggle with basic operations in languages like Bengali and Tamil, simply because there isn’t enough online text to train on. How do you think about that kind of bias?

This is particularly close to my heart. I learned my heritage language of Bangla as an adult, and it is Sanskrit-derived. And I learned how difficult it is to start to understand the script. There’s 56 characters and they shift and it doesn’t translate well to a keyboard. The internet was originally created in the Western world and a lot of it was created for Latin characters, specifically English. There’s so much infrastructure that leads to more data being available on specific languages, even down to the keyboards that the internet was built for.

Something that I learned this year is that it’s actually more expensive by token to train on and generate non-English languages. You’re paying a higher price to process each unit of data, especially non-Latin character languages. But that’s getting a lot better as costs come down. OpenAI with their latest GPT-4o launch has shown huge reductions in the cost of tokenization for many Indic languages. Cohere Command R announced halving the cost of tokenization for many languages as well. Oftentimes people don’t report or even measure the financial cost. But, you know, money is a big deal.

We’ve also seen certain communities actively withhold training data. We reported specifically on a group of writers in Singapore who refused to make their work available, out of a concern that the resulting model would be used against them. Do you think actions like these are effective, or do they just widen the gap between high-resource and low-resource languages?

I think this is one of the hardest questions. I adore Singapore. I met my fiancé there. I think Singlish is a beautifully rich language, and it makes sense that the authors should have a right to opt out, because that is their work. But it gets a lot fuzzier when we get to the question of who actually represents the language.

We’ve seen this play out in some Indigenous communities. For example, a couple of years ago, a representative of the Maori community spoke out against training models on the language and selling it back to the Maori people. But shortly after, another group and the Maori community trained their own language model and preferred to have full ownership. So I think part of the question here is, who owns the language and who is able to benefit from it?

India is another interesting case. The Indian government is actually working on funding more Indic-language data sets for AI through their Bhashini program. And Hugging Face is working with them on an open-source Hinglish-language model. We’re doing more evaluations, so we launched an evaluation leaderboard in Arabic and Korean. And I think the signaling can help push people towards not just training, but measuring performance across languages."

#metaglossia_mundus

"NEW YORK (AP) — A newly released ad promoting Apple's new iPad Pro has struck quite a nerve online.

The ad, which was released by the tech giant Tuesday, shows a hydraulic press crushing just about every creative instrument artists and consumers have used over the years — from a piano and record player, to piles of paint, books, cameras and relics of arcade games. Resulting from the destruction? A pristine new iPad Pro.

“The most powerful iPad ever is also the thinnest,” a narrator says at the end of the commercial.

Apple's intention seems straightforward: Look at all the things this new product can do. But critics have called it tone-deaf — with several marketing experts noting the campaign's execution didn't land.

“I had a really disturbing reaction to the ad,” said Americus Reed II, professor of marketing at The Wharton School of the University of Pennsylvania. “I understood conceptually what they were trying to do, but ... I think the way it came across is, here is technology crushing the life of that nostalgic sort of joy (from former times).”

The ad also arrives during a time many feel uncertain or fearful about seeing their work or everyday routines “replaced” by technological advances — particularly amid the rapid commercialization of generative artificial intelligence. And watching beloved items get smashed into oblivion doesn't help curb those fears, Reed and others note.

Several celebrities were also among the voices critical of Apple’s “Crush!” commercial on social media this week.

“The destruction of the human experience. Courtesy of Silicon Valley,” actor Hugh Grant wrote on the social media platform X, in a repost of Apple CEO Tim Cook's sharing of the ad.

Some found the ad to be a telling metaphor of the industry today — particularly concerns about big tech negatively impacting creatives. Filmmaker Justine Bateman wrote on X that the commercial “crushes the arts."

Experts added that the commercial marked a notable difference to marketing seen from Apple in the past — which has often taken more positive or uplifting approaches.

“My initial thought was that Apple has become exactly what it never wanted to be,” Vann Graves, executive director of the Virginia Commonwealth University's Brandcenter, said.

Graves pointed to Apple’s famous 1984 ad introducing the Macintosh computer, which he said focused more on uplifting creativity and thinking outside of the box as a unique individual. In contrast, Graves added, “this (new iPad) commercial says, ‘No, we’re going to take all the creativity in the world and use a hydraulic press to push it down into one device that everyone uses.’"

In a statement shared with Ad Age on Thursday, Apple apologized for the ad. The outlet also reported that Apple no longer plans to run the spot on TV.

“Creativity is in our DNA at Apple, and it’s incredibly important to us to design products that empower creatives all over the world,” Tor Myhren, the company's vice president of marketing communications, told Ad Age. “Our goal is to always celebrate the myriad of ways users express themselves and bring their ideas to life through iPad. We missed the mark with this video, and we’re sorry.”

Cupertino, California-based Apple unveiled its latest generation of iPad Pros and Airs earlier this week in a showcase that lauded new features for both lines. The Pro sports a new thinner design, a new M4 processor for added processing power, slightly upgraded storage and incorporates dual OLED panels for a brighter, crisper display.

Apple is trying to juice demand for iPads after its sales of the tablets plunged 17% from last year during the January-March period. After its 2010 debut helped redefine the tablet market, the iPad has become a minor contributor to Apple’s success. It currently accounts for just 6% of the company’s sales."

#metaglossia_mundus

"A new Norton Anthology, edited by Harvard’s Martin Puchner, reimagines the global literary tradition. 

by Nina Pasquini

RAINER MARIA RILKE’S 1908 poem “Archaic Torso of Apollo,” written in German, is notoriously difficult to translate. The poem, about a fragmentary statue of Apollo, is dense with enigmatic metaphors: the speaker describes the (imagined) eyes as ripening apples, the twisted loins as smiling, the marble as the glistening skin of a predator. For more than a century, translators have sought to balance faithfulness to the poem’s meaning with maintenance of its form.

The fifth edition of the Norton Anthology of World Literature, published June 1, includes five translations of the poem for readers to consider, including two contemporary ones: one by Wien professor of drama and of English and comparative literature Martin Puchner, and one by Google Translate. Throughout the anthology, “translation labs” such as this one offer several versions of translated texts, allowing students to see how various translators capture and lose different elements of the original, said Puchner, who serves as general editor of the anthology and who has written widely about world literature and other subjects.

The inclusion of the Google translation, he continued, was intended “to start a discussion about translation and machine learning and AI”—part of a broader effort to make the anthology relevant and accessible to contemporary students. Google’s translation is more literal than many others, misunderstanding some metaphors. But it also provides insights into the original German that more literary translations miss. The anthology encourages students to ask: What does this say about how machine learning models process language? What does this show about what poetry attempts to do?

THE NORTON ANTHOLOGY of World Literature attempts a formidable task: to curate, in six volumes, a representative collection of poetry, novels, short stories, folklore, and plays from across the world and from the beginning of written language (and in some cases from before that).

Conversations about what to include and exclude from anthologies have long been fraught, and attempts to distill literary traditions into a select few works have often focused on Western and male voices. But in recent decades, editors have made efforts to expand the scope of anthologies, including works by women and non-Western authors (see “A World of Literature,” September-October 2019). With the fifth edition of the anthology, Puchner said, those efforts have continued—not just by including new works, but by reconsidering how those works are organized.

The anthology’s six chronological volumes span literature from the ancient world to the twenty-first century. In previous editions, within each volume, literature was grouped by region, beginning with Europe and the Middle East and then moving eastward. This made sense for Volume A, which covers the ancient world: “The first written cultures are in the Near East, so that’s how the first volume opens, and we continued to follow that structure in Volume B,” Puchner said. But with this edition, “We realized that makes no sense [for the other volumes], and in a structural way centers Europe and the Middle East.” Accordingly, this edition varies the structure depending on the volume. For example, volume B, which covers the Middle Ages, starts in China, since “the Chinese Middle Ages happens before the high tide of the European Middle Ages,” Puchner said.

The new edition also expands the conception of what counts as literature. Traditionally, “Literature means letters, it means written text,” Puchner said. “But storytelling is a much broader activity that happened before writing arrived—and oral literature has continued to be an important way of communicating, even in many cultures that have writing.” This edition includes Yoruba and Native American creation myths that in some cases were written down only in the nineteenth or twentieth centuries. The chronological order in which texts appear was typically determined by when they were written—a method that doesn’t work for oral literature. So, to accommodate these additions, the fifth edition includes some thematic clusters, allowing the Native American and Yoruba myths to be placed alongside Mesopotamian creation myths, which were recorded much earlier.

These additions are now “right there in the beginning of the anthology and really, I think, change how it opens,” Puchner said. Later works in the anthology also draw upon these narratives: in the final volume, for example, a play by the contemporary Nigerian playwright Wole Soyinka, a Nobel laureate—Death and the King’s Horseman, about a horseman prevented from committing ritual suicide by colonial authorities—alludes to Yoruba creation myths.

AMID RECKONINGS about equity and representation in the academy, many critics have questioned the primacy of a “literary canon,” and whether such a concept can be truly inclusive of diverse voices. This logic has extended to literature survey courses, and as a result many universities “got rid of these kinds of mandated survey courses, and replaced those with distribution requirements,” Puchner said. At many schools, then, students can choose from a variety of literature courses that focus on specialized areas, but aren’t required to take courses that provide broad overviews of world literature.

ADVERTISEMENT

Survey courses and anthologies can certainly be criticized for what they exclude or for their seeming superficiality. But at their best, they situate readers in a lineage far longer and more expansive than a single cultural tradition or time period—allowing them to see how 20th-century plays draw on ancient creation myths, or how translations of a poem have transformed over time. As Cogan University Professor Stephen Greenblatt—who edits the Norton Shakespeare and the Norton Anthology of English Literature—said in a 2012 interview, they “remind people that the humanities carry the experience, the life-forms of those who came before us.”

Of course, many schools have retained world literature surveys, and the Norton Anthology, taught at more than 1,000 U.S. colleges, is the leading textbook for such courses. About half of those 1,000 schools are in the U.S. South, and many are public universities or community colleges. Instructors at these schools have said the anthology helps to achieve what Greenblatt described: it “[gives] perspective, showing students how our civilization has been shaped,” said Devona Mallory, a professor at Albany State University in Georgia. “You can see, through the literature, how some beliefs have changed over time, and how some have stayed the same.”

Among the works in the anthology, Lisa Roddy—an instructor in the University of South Alabama’s English department—especially enjoys teaching The Love Suicides at Amijima, an eighteenth-century Japanese play by Chikamatsu Monzaemon. “It’s sometimes hard to get students to think outside the context of their own moral codes,” she said. But when she asks her students to think about whether the lovers’ suicides are “honorable,” she encourages them to think about the samurai code, and how this society differed from their own. Amid such differences, some themes remain familiar. The anthology explains that The Love Suicides at Amijima was written during an era of increased commercialization in Japan. Students consider “how people in the play are defined by their professions and their commercial success. I think that really resonates with [students],” Roddy said, “because they are in this moment when they have to decide what they want to do with their lives,” balancing passion with economic considerations.

Beyond the academic benefits, the anthology—which typically retails at $92.50 per three-volume set—simply provides a relatively economical way to access such a breadth of literature. “Textbook affordability is a big issue,” said Verne Underwood, the chair of the humanities department at Rogue Community College in Medford, Oregon. “Norton has always been affordable with the way they bundle the books together.”

Anthologies and world literature survey courses are “always compromises,” Puchner said: trying to teach a representative array of world literature in one or two semesters will inevitably exclude certain works. But that doesn’t mean it isn’t worth trying, he continued—especially today, amid “the resurgence of a kind of ethno-nationalism in the U.S. and many other parts of the world.” Engaging with world literature is “the best vehicle I know for not just reading something a journalist writes about some part of the world, but to actually read the literature that’s part of the crucial DNA of different parts of the world,” he said. “And I think that’s a much deeper and more genuine—sometimes more difficult—way of engaging with other cultures.”"

#metaglossia_mundus

FAROOQ AHMED PEER June 6, 2024 12:00 am "The deep interest of scholars in the Rajatarangini continues without any break and even in modern times people are interested in its translation. Everybody is aware of the fact that Kashmiri literature has been enriched purely due to the efforts of the Kashmiri scholars and writers.

The galaxy of writers we have in Kashmir has contributed a lot to the Kashmiri literature and it is due to them that we have came to know about the political, social history of Kashmir. Dr. Basher Bashir is one such Kashmiri writer who has done a great service to Kashmiri literature.His translation of the 8th part (Taranga) of the Rajatarangini is an effort which even the common man can appreciate. (It is important to mention that he has done the translation of the first two parts of Kalhana’s book before the present translation of the 8th part of the chronicle).

 

Dr. Basher Bashir has translated the book into Kashmiri from the English translation of R. S. Pandits Rajatarangini. The translator does not know the Sanskrit language in which the book has been written by Kalhana, a fact which he confesses himself in the preface.

 

The book of Kalhana tells us the history of Kashmir up to 1150 A.D. In his translation, Dr. Basher Bashir shows that large portions of history of Kashmir deal with the events that took place in Kalhana’s own time. Kalhana describes what he himself experienced and witnessed on the political and social front of Kashmir. His information about the geography of Kashmir has become a source of information for Kashmiris who now know much about their valley.

The translation shows that Kalhana was a knowledgeable man and his ideas about the people and their manners show his own personality. The translation informs that Kashmir was a place of kings and rulers, hence the intrigues and strifes were the part of the game. The translation of the chronicle makes it explicitly clear that Kalhana did not hold any office under any of the rulers he saw or witnessed.

On the contrary, his father Canpaka enjoyed high position in Harsha’s rule. The chronicle does not have any clue wherein it could be found that Kalhana was patronized by any ruler as a poet or historian and there are no references which can show that he wrote his compositions on the instructions of Jayasimha in whose rule he completed his book. He, however, describes the defects and follies of certain rulers and praises other rulers for their qualities.

The translator in his translation of the chronicle depicts Kashmiris in their real colour and at times one comes to know that Kashmiris lacked physical and moral courage, the fact which the other historians have also highlighted in their compositions.

The translation shows that Kalhana informs that besides merchants, traders and agriculturists, the Kashmiri society was full of teachers, astrologers, physicians, labourers, artisans, soldiers and carters, workers of water wheels and hand mills. The translation shows that there were divisions and sub-divisions among those classes of people according to their professions.

The translation of Rajatarangini gives an impression that Kashmiri women received a liberal atmosphere and education in their parent’s house. They could speak Sanskrit and Prakrit languages fluently. Kalhana’s chronicle proves that there was no seclusion of women, nor were they segregated or veiled in any way.

They were queens also and took active part in the affairs of the government. Even women of lower status took an active part in the affairs of the state and more importantly, were not discriminated on the basis of gender. The chronicle also states that there were instances of immorality prevailing among some classes of women.

The custom of dedicating girls (devadasies) to temples for singing and dancing was prevalent in Kashmir also. The translation also shows that the Kashmiris were highly superstitious like Indians. They believed in black magic, and sorcery affecting the lives of human beings. There was also wide spread practice of witchcraft in ancient Kashmir.

Dr. Basher Bashir through his translation shows that the Rajatarangini is an important historical record of Kashmiris, which gives a detailed account about the kings, masses and olden times of Kashmir. The translation of the book is a valuable contribution to the history of Kashmir as well as to the literature of Kashmir. It makes amply clear that Kalhana was not only a historian but also a poet par-excellence and his Rajatarangini is more than an account of the rulers of the Kings of Kashmir.

The translator through his translation shows that the chronicle serves moral lessons also which show that Kalhana was of the belief that the glories of the world were transitory and there must be a moral code of conduct in everything. In the 8th canto or part itself, he says, “Shadow is itself unrestrained in its path while sunshine, as an incident of its own nature, is pursued a hundredfold by nuance. Thus is sorrow from happiness a thing apart; the scope of happiness, however, is hampered by the aches and hurts of endless sorrow”.

The translation brings to our notice that the chronicle suffers from many drawbacks. There are no accurate dates of the rulers and one can not believe that the names of the rulers can be also authentic. The translation of the eighth part, consisting of 3,449 verses, shows a record of contemporary events but is full of obscurities and ambiguities in terms of the narrative and construction of certain ideas.

The obscurity and ambiguity in the narrative of the translation reflects ambiguity in Kalhana’s book itself. The narrative is such as if Kalhana was writing for the readers who were well acquainted with the Kashmir of his day. However, with regard to the topography of Kashmir, local references seem to be exact and clear.

In spite of the shortcomings, the translation of the Rajatarangini is an important historical commentary with regard to the past of Kashmir. It shows that Kalhana has given a wonderful picture of out native land with rise and fall of empires, the conflict of different races, cultures and civilizations. The translator deserves appreciation for his commendable effort and his services need to be utilized in a proper way at a proper place."

#metaglossia_mundus

"The company uses generative AI to achieve photo-realism and accuracy to provide greater accessibility for the Deaf community. 

Cate Lawrence05 June 2024

 

Signapse, a generative AI Sign Language translation software company, today announced the closing of a new seed funding round of £2 million with £1.5 million from investors and £500k from the UK Government. 

The company uses generative AI to achieve photo-realism and accuracy to provide greater accessibility for the Deaf community. 

Over the past two years, Signapse software has created 5,000 BSL train announcements daily throughout the UK, as well as delivering over 4,000 manual BSL translations for clients in the website and video markets. 

More recently, Signapse has delivered Deaf accessibility to transport companies such as LNER Train Stations in the UK, and CVG Airport in the US. 

“I couldn’t be more excited and proud about this next stage in our growth. It is a landmark opportunity, a once in a lifetime moment, to build a transformative product,” says Sally Chalk, CEO and Co-Founder at Signapse.

“This is an important chapter for Signapse, where we can expand our AI messaging from limited domains of discourse such as Transport to more open ended markets such as Website and Video Translation.”, says Ben Saunders, CTO and Co-Founder at Signapse.

This new round of funding will be used to improve Deaf accessibility through building unconstrained translation for British Sign Language (BSL). 

Specifically, the company will focus on:

• Expanding into the Video Translation market to improve Deaf accessibility globally
• Building unconstrained translation for BSL with the ability to translate from any English sentence into BSL using AI
• Improve the quality of Signapse AI-generated Sign Language Videos

Soulmates Ventures and Deeptech Seed Fund led the funding, with participation from other investors, including the Royal Association for Deaf people, Empirical Ventures, CEAS Investments, and FSE Group.

“We at Soulmates Ventures are proud and excited to be a part of the Signapse endeavour in building unconstrained generative AI sign language translation. Their innovative solution is reflecting the daily challenges and needs of the Deaf people.” 

“We are looking forward to supporting the Signapse team through our acceleration programme to amplify their distinctive real-time solution and enhance accessibility in transport, mass media, TV, streaming services and others for the Deaf community globally”, says Michal Sikyta, Investment Director at Soulmates Ventures."

#metaglossia_mundus

"Under-resourced languages incorporated into machine-translation model.

Machine-translation models use artificial intelligence (AI) to translate one human language into another — a worthy feat, given the potential for enhanced communication to break down the barriers posed by differences in language and culture. Yet most of these models can interpret only a small fraction of the world’s languages, in part because training them requires online data that don’t exist for many languages. The US technology company Meta has designed a project called No Language Left Behind (NLLB) to change that. Writing in Nature, the NLLB team1 presents a publicly available model that can translate between 204 languages, many of which are used in low- and middle-income countries..."

#metaglossia_mundus

Par Luc Ferry  Publié le 05/06/2024 "CHRONIQUE - La vérité, paradoxale, c’est qu’à force d’être radicalement opposées, la culture allemande et la culture française sont devenues complémentaires, voire inséparables.

Malgré trois guerres, dont chacune fut plus meurtrière que la précédente - 1870, 1914, 1939 -, rien n’y fait : le couple franco-allemand reste la poutre maîtresse de l’Union européenne. Comme le disait de Gaulle avec humour, « l’Allemagne et la France, c’est la viande ; les autres, c’est les légumes ! ». On parle beaucoup ces temps-ci des difficultés du couple franco-allemand, des soucis économiques de l’Allemagne, de la dette de la France, des désaccords sur les troupes au sol en Ukraine, etc., mais on s’empresse malgré tout de souligner la « coopération » et l’« amitié » (on évite quand même le mot « collaboration »…), qui nous lient par-delà nos divergences de vues. Même si, malgré les récents efforts de Scholz et de Macron, le couple n’est pas au mieux, il n’en reste pas moins qu’au regard du passé douloureux qu’évoquent ces trois dates l’amitié franco-allemande demeure aux yeux du monde comme une espèce de miracle, un exemple unique..."

#metaglossia_mundus

"Impulsionando a eficiência e precisão na comunicação corporativa, a solução oferece recursos de personalização e proteção de dados

Por Mayara Cardone06/06/2024 A DeepL, empresa líder em inteligência artificial linguística, anunciou hoje o lançamento do DeepL para Empresas, uma extensão de sua plataforma voltada para o setor empresarial. A solução foi desenvolvida para atender à crescente demanda por ferramentas de tradução e escrita de alta qualidade, oferecendo novos preços e planos, recursos de personalização, proteção de dados e suporte global para implementação e gestão. 

Com o objetivo de auxiliar as grandes organizações na superação dos desafios relacionados à adoção de inteligência artificial, a novidade já está disponível em todo o mundo. A plataforma é utilizada por mais de 100.000 empresas em mais de 60 países, incluindo metade das listadas na Fortune 500, para obter traduções de alta qualidade e melhorar a escrita. 

“As lideranças das empresas estão sob pressão para definir e executar estratégias de adoção de IA que sejam sólidas e gerem ROI, mas se deparam com um mercado saturado e repleto de modelos genéricos ou de código aberto que não atendem a necessidades específicas. Elas também precisam considerar com cuidado os requisitos de conformidade para proteger dados confidenciais e reduzir possíveis riscos associados à segurança cibernética”, comentou David Parry-Jones, Diretor de Receita do DeepL.

A solução oferece modelos especializados e adaptados especificamente para contextos de linguagem e idiomas, o que proporciona traduções mais precisas e reduz o risco de distorções e ambiguidades. Um estudo da Forrester de 2024 revelou que o uso do DeepL proporcionou um retorno sobre o investimento de 345% para empresas internacionais, reduzindo o tempo de tradução em 90% e gerando uma diminuição de 50% na carga de trabalho. "

#metaglossia_mundus

"A tecnológica Unbabel desenvolveu um grande modelo de linguagem otimizado para a tradução automática. Segundo a empresa, superou dois modelos da OpenAI nas tarefas de tradução.

Portuguesa Unbabel lança modelo de IA otimizado para as traduções

A tecnológica Unbabel desenvolveu um grande modelo de linguagem otimizado para a tradução automática. Segundo a empresa, superou dois modelos da OpenAI nas tarefas de tradução.

06 jun. 2024, 17:01  

A portuguesa Unbabel lançou o TowerLLM, um grande modelo de linguagem (LLM, na sigla em inglês), que foi concebido, treinado e otimizado para fazer tradução. Os LLM são modelos de inteligência artificial (IA), como o GPT-4o que alimenta o ChatGPT, e que foram treinados com enormes quantidades de dados para conseguir gerar informação para os utilizadores.

 

A Unbabel anuncia em comunicado que o TowerLLM, que foi gerado para ser multilíngue, “melhora significativamente a qualidade da tradução automática, reduzindo os erros e os custos”. A empresa menciona testes em que o modelo terá conseguido obter um desempenho acima dos modelos LLM GPT-3.5 e GPT-4o da OpenAI em tarefas ligadas à tradução. Também terá apresentado um “desempenho consistentemente superior” quando comparado com modelos da Google – não é especificado com qual versão de modelo de IA é que foi comparado.

 

O comunicado da Unbabel é acompanhado de alguns gráficos de comparação entre modelos, uma prática que se tornou habitual quando é revelada uma novidade deste género. Nestes exemplos foi aferido o desempenho do modelo em questões como a tradução entre pares (de inglês para português, para alemão, espanhol, francês, italiano ou coreano), tradução em vários domínios (financeiro, jurídico, médico e técnico) ou no processo de compreensão contextual.

A empresa explica que treinou o modelo com dados “públicos e internos” e que o TowerLLM foi “afinado com dados de tradução de elevada qualidade”, que foram filtrados e selecionados.

 

Na fase de lançamento, o modelo vai conseguir fazer a tradução automática em 18 pares de idiomas, conseguir reconhecer entidades nomeadas para localizar nomes, métricas e códigos (por exemplo, moedas, marcas ou localizações) e pós-edição automática para melhorar a tradução. Também é feita a promessa de correção de fonte para eliminar erros de gramática e ortografia antes da tradução. O modelo só vai estar disponível para clientes dos serviços da Unbabel.

“Apesar das dúvidas que surgiram na indústria, o TowerLLM demonstra, claramente, que os LLM são a melhor solução para a tradução automática”, diz João Graça, cofundador e diretor de tecnologia da Unbabel. “Investimos anos no desenvolvimento de LLM e, por isso, não estamos surpreendidos por ver o TowerLLM superar o GPT-4o. Este é o segundo LLM multilíngue que lançamos depois do COMETKiwi e estamos apenas a começar.”

O COMETKiwi foi lançado pela Unbabel em setembro de 2023: é um modelo LLM de código aberto desenvolvido para conseguir prever a qualidade de tradução."

#metaglossia_mundus

June 06, 2024 PATRICIA JONES "Receiving a notice to appear in court can be a stressful experience for anyone, but even more so for those who aren’t fluent in English.

Middle Tennessee’s exponential growth of the last decade has included an influx of non-English-speaking residents, but many courts outside of Davidson County still lack a reliable process for managing civil cases in which a litigant speaks a non-English language. Civil cases can include evictions, child custody, orders of protection, debt and bankruptcy.

Tennessee’s Administrative Office of the Courts maintains a database of certified and registered interpreters. While the clerks of courts in rural counties try to schedule interpreters for hearings, it is not their official job. Tennessee rules do not specify who is responsible for doing so. Complicating the matter is that many AOC-certified interpreters are located in Nashville and have to drive longer distances to rural courts.

This forces non-English speakers to try to navigate court hearings in a language they don’t understand — putting them at a significant disadvantage in matters that can affect their basic needs and stability. Though some people attempt a workaround by bringing an English-conversant family member or friend to serve as a translator, the legal terms and jargon used during court hearings can limit the effectiveness of this approach.

As an attorney at Legal Aid Society, I assist clients in eight rural Middle Tennessee counties where it’s not uncommon to see civil court hearings involving non-English speakers proceed without an interpreter present (in Nashville, it’s standard practice to have interpreters at both civil and criminal hearings).

While everyone should get a free interpreter if they are not fluent in English, the court does have discretion in determining whether an interpreter is needed. So, it is very important that the court recognize a person’s need for a court interpreter. While it’s legally unclear if civil courts are required to provide interpreters (they’re a Constitutional right for criminal matters), it’s certainly the right thing for them to be doing.

If you’re a non-English speaker with a court proceeding coming up, it’s important to understand how to advocate for yourself and have your best chance of having an interpreter present in court:

• As soon as possible after receiving your notice to appear (and definitely at least a week before your hearing), contact your court clerk and tell the clerk you need an interpreter for the day of your hearing. It’s best to do this in person and submit your request in writing so it’s documented and placed in the court’s files. Because the request will need to be submitted in English, see if a friend or family member can help you write the request and state clearly in the request that they assisted, so the court doesn’t assume that you wrote the request by yourself.

• If you don’t receive a response to your request, call the Administrative Office of the Courts, which is responsible for maintaining the list of interpreters, at (615) 741-2687.

• If you qualify for Legal Aid Society’s services, we might be able to provide legal assistance. (We regularly help non-English speaking clients file petitions for Orders of Protection, for instance.) Call our main number at 1-800-238-1443, where you can leave a message in your native language and receive a prompt callback. The Nashville Hispanic Bar Association (615-701-7957) also offers legal assistance to non-English speaking clients.

• If you show up the day of your hearing and there’s no interpreter present, ask for a continuance (the legal term for delaying proceedings until another day). Take an English-speaking friend or family member along to make the request, if necessary. Although the judge doesn’t have to grant your request, it will be documented in the court record, which is important. If you do receive a continuance and a translator still isn’t present on the rescheduled date, you can request another continuance.

When in court, non-English speakers often make a great effort to be agreeable, nodding their heads even if they don’t understand a question they’re being asked. But it’s important for people to be clear about their limitations so that judges recognize when a situation is unfair. If you don’t understand, say you don’t understand.

The best way to create awareness, and hopefully improve this current situation, is to begin documenting examples of this situation repeatedly taking place — creating a “paper trail” that can potentially strengthen your own case or lead to a future court case in which the process of appointing interpreters is challenged.

Patricia Jones is the managing attorney of Legal Aid Society of Middle Tennessee and the Cumberlands’ Columbia office."

#metaglossia_mundus

"By Kosidichimma Anyanwu

The year 2024 has seen landmark events in the growing knowledge and embracing of artificial intelligence (AI) around the world. Researchers, scientists, software developers, and several stakeholders have continued to announce advancements and innovations in AI technologies, including machine learning, natural language processing, and robotics, across laboratories on different continents. What we are experiencing is a surge in algorithm developments, increased computational power, and the expanding application of AI to advance various sectors, from digital media to healthcare, education, finance, business, real estate, governance, and so on.

This sudden shift and integration into daily lives and global systems is also raising smoke within the social and political environs, where there is an increasing need to revisit relevant ethical considerations and regulatory frameworks while continually monitoring societal impacts. However, throughout history, culturally unfamiliar evolutions in society, such as the resurfacing of AI, have attracted cynicism—and understandably so. Therefore, amid an ongoing technology shift, there is a salient urgency for government and stakeholders across all sectors to conscientiously engage with the current trends as a signal of collective control over their potential societal impacts.

 

 

While the largest nation in Africa, Nigeria, is grappling with multifaceted socio-economic challenges, it is important to acknowledge and reinforce its timely interventions in the scurrying global advancements in AI. Despite its domestic dips and spikes, Nigeria continues to contend as one of the largest economies in the world. Hence, it is rightly positioned to make a significant global impact in this field, given its largely underestimated advantages.

According to the OECD report, one out of every four Africans around the world and one out of every five persons of African origin is a Nigerian. Its population represents about 2.6 percent of the human population, accounting for a relevant global labour force and a large domestic market with potential to influence international economies. Nigeria also has the highest population of African immigrants in countries such as the United States and the United Kingdom, extending the country’s impacts to global economic powers.

 

Moreover, one peculiarity of Nigeria’s leverage over its global counterparts lies in its undocumented yet thriving informal sectors. The country has one of the most significant and booming informal economic sectors in the world. Officially, IMF growth estimates Nigeria’s GDP in purchasing power parity (PPP) terms at $1.116 trillion by the end of 2023. However, Nigeria’s informal economy is estimated at approximately $1.230 billion in size, representing 58.2 percent of overall GDP PPP levels. Consequently, World Economics’ Research’s updated database estimates Nigeria’s GDP to be $2.113 trillion—89 percent larger than the official figures using 2023 data, given the impact of the informal economy.

 

In April 2024, the Nigerian government announced its commitment to leveraging the power of AI systems. This aims to foster inclusivity by integrating indigenous knowledge systems and language bases through the launch of its multilingual large language models (LLMs). The country’s initiatives, including the formation of the Nigeria AI Collective, the launch of a multilingual LLM, and substantial investments in computing infrastructure, have far-reaching implications, particularly in the areas of content diversity, an inclusive digital economy that considers informal sectors, and international trade expansion.

Given factors including its demographic strength, economic diversity and size, and potential for comparison with global economic powers, its move into developing LLMs is a pivotal achievement in the international AI field that cannot be ignored.

Content diversity

The development and deployment of Nigeria’s multilingual LLM stands as a pivotal moment for content diversity in AI. Traditional AI models often focus on high-resource languages, predominantly those spoken in the Global North, thus marginalising many languages and dialects spoken in Africa and other parts of the world. Nigeria’s LLM, trained in five low-resource languages and accented English, challenges this paradigm by ensuring these languages are represented and preserved in the digital age.

 

This inclusion has profound implications. First, it enriches the AI training datasets, leading to more robust and versatile AI systems capable of understanding and processing a wider array of human languages. This diversity in language representation can drive innovation in AI applications, ranging from more accurate language translation services to culturally relevant content generation. It democratises access to AI technologies, empowers local communities, and ensures that the benefits of AI are more equitably distributed.

Furthermore, this move can inspire other nations to prioritise their linguistic heritage, fostering a more inclusive global AI ecosystem. As Nigeria leads by example, other countries with diverse linguistic landscapes may follow suit, contributing to a richer, more varied digital content repository worldwide.

Modelling an inclusive digital economy

Nigeria’s AI initiatives have significant potential to reshape the global digital economy by promoting inclusivity. The establishment of the Nigeria AI Collective and the enhancement of the National Centre for Artificial Intelligence and Robotics (NCAIR’s) capacity with support from global technology leaders like Cisco underscore a commitment to achieve this robust goal.

 

The goal of inclusivity is highly attainable if the stakeholders maintain a strict ecosystem that is open to participation, including AI enthusiasts, students, startups, and civil society organisations (CSOs), apart from elite researchers or tech giants. Such inclusivity ensures that the benefits of AI are accessible to a broader segment of the population, including stakeholders who operate outside the purview of formal economic sectors.

 

 

Nigeria will signal support for grassroots innovation at a global level by giving local researchers and companies access to cutting-edge computing infrastructure. The democratisation of AI technology may result in the creation of solutions for regional problems in a variety of fields, including banking, agriculture, healthcare, and education. Nigeria can therefore provide an example for other countries by concentrating on creating AI with moral guidelines specific to its socioeconomic environment.

Nigeria can manage the possible downsides of AI, like employment displacement and privacy problems, in a way that puts its citizens’ welfare first by developing its ethical frameworks. This strategy can be used as a model by other countries, contributing to the global development of a more just and equitable digital economy.

International trade expansion

The multilingual capabilities of Nigeria’s LLM also have significant implications for international trade. Language barriers often pose substantial challenges in global commerce, particularly for businesses and entrepreneurs in non-English-speaking regions. Nigeria’s LLM can promote smoother trade relations and negotiations by facilitating more effective communication across linguistic boundaries. This increases local Nigerian enterprises’ access to global markets.

 

To increase their reach and accelerate economic growth, entrepreneurs and small and medium enterprises (SMEs) can interact with overseas partners and customers more effectively. Additionally, as international companies try to get a foothold in Nigeria’s expanding market, the improved communication capabilities may draw in foreign collaborations and investments.

Nigeria is poised to improve its trading prospects and make a positive contribution to a more integrated and cohesive global economy by overcoming linguistic barriers through investments in AI systems.

Nigeria’s pioneering move into AI, particularly through the development of multilingual LLMs and the establishment of a comprehensive AI ecosystem, holds key global implications. Alongside changing its socioeconomic landscape, Nigeria also sets an example for other countries by promoting variety in content, building a digital economy that is inclusive of all people, and increasing international trade. Whether we pay attention or ignore it, the world is indeed observing Nigeria’s AI journey.

Despite the optimistic outlook on Nigeria’s adoption of AI and LLMs, there are significant concerns that continue to warrant careful consideration. From a potential threat to human labour when not meticulously implemented to contradictory ethical standards often dictated by the Global North, which may not align with the unique socio-cultural contexts of countries like Nigeria, and so on.

These arguments emphasise the necessity of approaching the adoption of AI with a balanced approach that considers both its potential benefits and the associated risks.

 

Dr Kosidichimma Anyanwu, a tech and media expert, writes from Ireland."

#metaglossia_mundus

"South African startup Vambo AI is empowering users to use advanced AI in their native tongues, currently supporting 11 African languages.

June 6, 2024

This Startup’s AI Supports 11 African Languages And Counting

South African startup Vambo AI is empowering users to use advanced AI in their native tongues.

In April 2023, Chido Dzinotyiwei and Isheanesu Misi founded Vambo AI to create AI that “understands the languages of people in the rest of the world”. Its platform currently supports 11 African languages.

“The reason why most of the world has not used AI – or explored its revolutionary power for education, economic empowerment, and more – is because it is not built in languages they understand,” Dzinotyiwei said, according to Disrupt Africa.

A Multilingual AI Platform

Vambo AI offers an array of functionalities, including writing, searching, translating, and transcribing in 11 languages, including Arabic, KiSwahili, IsiZulu, and French. 

This innovation stems from Dzinotyiwei’s earlier venture, Vambo Academy, a language-learning startup launched in 2019. 

The insights gained from this project highlighted a gap in the market for AI platforms catering to African communities’ linguistic diversity. 

Traditional platforms like Duolingo and Babbel often overlook the complexities of African languages, a gap Vambo AI is set to fill.

Localized Solutions For Global Impact

Vambo AI’s impact extends beyond individual users to businesses. Businesses can leverage the startup’s API and developer tools to create localized solutions tailored to various regions. 

“Approximately 95% of African languages lack digital representation, which has dire consequences for economic participation, language preservation, and societal development,” says Dzinotyiwei. 

Vambo AI addresses this by offering tools that facilitate language learning and enable content generation and communication in African languages.

Achievements And Future Plans

Despite being a young startup, Vambo AI has bootstrapped its way to success with some support from AWS credits. Its business model includes API licensing and B2B enterprise solutions. 

The platform has evolved from an MVP with four languages to one that now supports 11 languages.

The startup has garnered attention from various organizations and participated in prominent accelerators, such as the Injini Mastercard Foundation EdTech Fellowship and Grindstone Africa. 

These opportunities are paving the way for Vambo AI’s expansion plans, which include developing better models, adding more languages, and enhancing features to meet the needs of its diverse client base."

#metaglossia_mundus

"Canadian poet-translator George McWhirter stood alone at a lectern Wednesday night to accept the Griffin Poetry Prize he earned with his friend and longtime collaborator Homero Aridjis for Self-Portrait in the Zone of Silence. 

George McWhirter has been translating Homero Aridjis's work since 1987; 60% of prize given to translator

Nicole Thompson · The Canadian Press · Posted: Jun 06, 2024 12:05 AM EDT | Last Updated: June 6

Canadian poet-translator George McWhirter stood alone at a lectern Wednesday night to accept the Griffin Poetry Prize he earned with his friend and longtime collaborator Homero Aridjis.

The prodigious Mexican poet couldn't attend the ceremony in Toronto where his collection, Self-Portrait in the Zone of Silence, won the $130,000 award because he was recovering from intestinal surgery.

"I'm so sorry he is not here to accept the prize for his work, which is expansive," McWhirter said in his acceptance speech. "It covers everything that you could think of in Mexico. It covers the people, the politics, the history and the dreams and the myths. And it all floats through his poetry."

Jurors praised the collection, which uses mythical imagery to examine emotional realities, for presenting "a rounded human being engaged with our total experience."

McWhirter, who formerly headed up the creative writing department at the University of British Columbia, has been translating Aridjis's work since 1987. But his translation of poetry goes back as far as his writing does, he said.

 

"Even when I was in high school, I started when I was just 16," McWhirter said in a brief interview after the ceremony. "I learned to write poetry by translating poetry."

The Griffin gives 60 per cent of the prize winnings to the translator and 40 per cent to the original poet, in recognition of the art of poetry in translation.

With Aridjis, McWhirter said, the translation process is collaborative — and not just between the two of them.

Aridjis's wife, Betty Ferber, is the third pillar of their team.

"She is the great second-guesser, and she is rigorous, and a great proofreader," McWhirter said in his speech.

While Aridjis speaks English, McWhirter said it "doesn't reach as far" as his Spanish.

"Homero is many poets in one," McWhirter said earlier in the night. "He writes about many Mexicos and many Mexicans."

Aridjis is president emeritus of PEN International, a writers association, and also served as Mexico's ambassador to Switzerland, the Netherlands and UNESCO.

Numerous runners-up

The runners-up are A Crash Course in Molotov Cocktails, translated by Amelia Glaser of the United States and Yuliya Ilchuk of Ukraine, from the original Ukrainian by Halyna Kruk; U.S. poet Jorie Graham for To 2040; Ann Lauterbach of the United States for Door; and Ishion Hutchinson of Jamaica for School of Instructions.

Runners-up receive $10,000, with 40 per cent going to the original writer and 60 per cent to the translators.

North by Northwest17:26Griffin Poetry Prize nominee George McWhirter on translating the works of Homero Aridjis

Vancouver-based poet George McWhirter is co-nominated for the Griffin Poetry Prize for his translation of "Self-Portrait in the Zone of Silence," a collection of poems by acclaimed Mexican poet and activist Homero Aridjis.

Until recently, the Griffin handed out separate awards for international and Canadian poets, but prize benefactor Scott Griffin announced in 2022 that the award would consolidate the categories — which were worth $65,000 apiece — into one global purse for the best book of poetry published or translated into English.

Last year's winner was U.S. poet Roger Reeves, for his book Best Barbarian.

McWhirter is the first Canadian to win since the change.

N.L.-based authors get recognition

Wednesday night's festivities also included a reading by Maggie Burton, who won this year's Canadian First Book Prize for her debut collection, Chores, which she said took 10 years to write.

A St. John's city councillor, violinist, mom of four and poet, Burton plays many somewhat unconventional roles. It's difficult, she said, to balance the obligations of daily life with her creative process, which is a recurring theme in her collection.

"I'm so happy any time someone says the poems resonate with them," she said in an interview ahead of the readings.

"Anyone who experiences tension between various opposing ideals that are listed in my book ... it's really meaningful when people can respond to it in that way. That makes it all worth it."

 

An attentive audience of poetry lovers and publishing industry insiders packed into Toronto's Koerner Hall for the readings, which were punctuated by laughter and applause.

"It is wonderful to see so many human beings wanting to listen to poetry in this desperately unpoetic time," Lauterbach, one of the runners-up, said ahead of her reading.

Those human beings were rapt as Newfoundland-based poet Don McKay received the $25,000 Lifetime Recognition Award.

He won the Griffin Poetry Prize in 2007 for Strike/Slip, a work jurors at the time praised as "a playful yet resonant microcosm, charted with virtuosity and love."

He was a finalist for the prize two other times, and has twice won the Governor General's Literary Award for Poetry.

"Thank you so much," he said as he accepted the award. "I'm deeply moved. Nevertheless, I'm going to read an essay."

That reading garnered a standing ovation."

#metaglossia_mundus

"As we dethrone TM discounts, we welcome a future where technology and human expertise combine to break new ground in translation.

It’s Time To Say Goodbye To Translation Memory Discounts

Joao Graca

Forbes Councils Member

Forbes Technology Council

COUNCIL POST| Membership (Fee-Based)

Jun 6, 2024,08:30am EDT

Updated Jun 7, 2024, 09:05am EDTJoão Graça is the co-founder and CTO of Unbabel.

Consumers demand that content is available in their spoken language and adapted to their culture—but that’s challenging for businesses to deliver efficiently given the skyrocketing volumes of content they create across countless touchpoints. A long-time approach to making translations more efficient and consistent was translation memories (TMs). But what are TMs and how has this affected the prices charged by translation organizations?

 

Translation memories were first introduced in the ‘90s to leverage repetitive content in translations. The idea was that if a translator had already translated a sentence, this sentence could be reused the next time it shows up. Simple. This would lead to faster turnaround times and more consistent quality. It became obvious, though, that if recurring translations were reused, there was good reason to pay translators less, and as a consequence, businesses started asking for discounts for TMs matching. So, if X% of a document are TMs, then that X% shouldn't be charged for.

 

This has been the operating model in the translation industry for the last 30 years. However, there are several drawbacks to TMs that are making this method extinct.

 

The problem? Translation memories can lead to inaccurate or inconsistent translation by not taking into account the surrounding context, which happens all the time. For instance, how should you disambiguate a pronoun that shows up in a previous sentence? This leads to some evolution in the form of in-context TMs, where the system tries to match not only the source sentence but also the previous one or two source sentences at the same time to capture more context. But although this helps, it doesn't completely solve the fundamental problem of context.

When TMs are used by different translators, this compounds inconsistency because every translator has their own style. Even if both TMs are correct, the translation will look off. A similar problem arises when carrying TMs over between different translation projects because instructions, guidelines and tone differ from project to project. As if this wasn’t enough, the static nature of TMs makes them hard to update and maintain whenever there are new or updated terminology entities or when some stylistic element of the customer changes over time. The risk here is that TM content goes out of date quickly, making it even harder to achieve consistent quality.

If you combine the inherent problems of using TMs with the fact that translators don't always produce perfect quality, it severely undermines the premise of a translation memory. TMs don't produce perfect quality and shouldn't necessarily be reused. On the contrary, often, they compound errors and incur the extra cost of human review. Nonetheless, this is the current state of an industry that lacks an automatic and scalable way to assess quality.

TMs also fell behind because of advancements in machine translation (MT). Because TMs are only created after a translation is provided by humans, they miss the opportunity to capture good quality MT. In the old days, when MT was poor quality, this made sense, but as it has progressed, this approach no longer holds.

 

MT quality has come a long way, and for some content types, it performs just as well as human translators. Following the same principle, a translator shouldn't spend time reviewing correct MT and by equal measure, a customer shouldn't pay for human input on any content that MT correctly translates.

It’s at this point that quality estimation (QE) algorithms become very interesting. QE has shown remarkable performance in evaluating the quality of translation, both when performed by machine translation and by professional translators.

QE’s value is that it unlocks the possibility of discounting high-quality segments regardless of whether it's translated by machine or human, while at the same time, it wouldn't discount poor quality TMs. When quality is the benchmark, the input doesn’t matter, whether it’s new or previous (i.e., TMs) content. When machine translation is combined with quality estimation, businesses can drive more automation efficiency in translation while only paying for work requiring human involvement when it makes sense—a much more flexible and transparent pricing arrangement that aligns incentives between providers and customers. It’s time for QE discounts.

But there’s more. I also urge translation companies to opt for flexible pricing, where customers only pay for what they use per translation segment. For example, if they only need MT, they pay the MT cost for that segment; if they need a human editor, then they pay for that cost; and if they need a subject matter expert, they pay for that service. This type of flexible pricing is arbitrated by QE and flexible workflows. This is a clear value alignment with what customers require, starting with AI first and then only relying on human input where needed to achieve the desired quality.

As we dethrone TM discounts, we welcome a future where technology and human expertise combine to break new ground in translation. It’s time to say goodbye to TM discounts and embrace an AI-first approach and flexible pricing.

Joao Graca

João Graça is the co-founder and CTO of Unbabel. Read João Graça's full executive profile here."

#metaglossia_mundus

"By Maja Carrieri and Johannes Kratzheller, WMO SecretariatThe widely diverse staff at WMO hail from about 90 countries. They enrich the working environment by bringing a vast array of cultures and languages into the Organization, which contribute to WMO’s ability to serve Members worldwide. The understanding of issues and opportunities on-the-ground stems partly from the common shared experiences of WMO staff and its constituents and beneficiaries. Further enhancement of diversity, multilingualism and openness at WMO is key to its continued success.

Multilingualism and Diversity at WMO – Bringing the Secretariat Closer to Members and Regions

MAGAZINE ARTICLE

07 June 2024

Further enhancement of diversity, multilingualism and openness at WMO is key to its continued success.

By Maja Carrieri and Johannes Kratzheller, WMO Secretariat

The widely diverse staff at WMO hail from about 90 countries. They enrich the working environment by bringing a vast array of cultures and languages into the Organization, which contribute to WMO’s ability to serve Members worldwide. The understanding of issues and opportunities on-the-ground stems partly from the common shared experiences of WMO staff and its constituents and beneficiaries. Further enhancement of diversity, multilingualism and openness at WMO is key to its continued success.

Multilingualism 

United Nations Secretary-General António Guterres emphasized that “multilingualism helps guarantee the effectiveness of our multilateral system”. This is true across the United Nations, and especially so in the life-saving work facilitated by WMO and its partners. Coordination of global cooperation in meteorology, climatology, hydrology and other related environmental fields, as per the WMO mandate, requires multilingualism. At WMO, the exchange of observational data, the provision of expert training, the delivery of weather forecasts and alerts and the calls for action on the state of the climate require listening and communicating with diverse audiences in languages they understand. Multilingualism is key to reaching and connecting with those audiences. 

The Organization communicates effectively with its diverse membership in the six official languages of the United Nations system, fostering trust, engagement and solidarity across linguistic borders. The more Members feel heard and valued, the stronger and more productive the relationships between them and with the Organization.

WMO has been investing in multilingualism over the years. Teams of linguists around the world work hand-in-hand with Secretariat professionals behind the scenes to enable experts in meteorology, climatology, hydrology and environmental sciences to understand each other and to be understood by the public and stakeholders. 

Each of the 193 WMO Member States and Territories selects one of the six official languages – Arabic, Chinese, English, French, Russian and Spanish – as their preferred language for official correspondence. Translation and interpretation are provided in these languages at all WMO constituent body meetings. By ensuring that all communication, including important documents, are available in the official languages, the WMO demonstrates commitment to inclusivity and respect for linguistic diversity. 

Furthermore, multilingualism enhances the Organization's outreach efforts by ensuring that critical information reaches stakeholders in a language they understand. WMO’s meticulous translation of relevant constituent body documentation and publications, training materials and policy documents maximizes their impact and facilitates broader uptake by National Meteorological and Hydrological Services (NMHSs), stakeholders and the public. Most of these are available through the revamped WMO e-Library, launched in 2023.The e-Library provides unlimited access to digital publications on meteorology, climatology, hydrology, and related fields. It has a six-language interface, offers advanced search and the filtering of possibilities for ease of research in publications that go as far back as the time of the International Meteorological Organization.

WMO pushes this effort beyond the prescribed United Nations scope of languages by supporting translation of public outreach materials in many more languages. For example, “We care for our climate” was published in the six official United Nations languages, but also Croatian, Dutch, Finnish, Greek, Papiamento, Persian, Turkish, etc., in cooperation with NMHSs. A recently published children’s book series that uses simple terms to explain what to do in case of natural hazards such as floods, earthquakes and landslides was also published in several languages (see Figure 1). Several videos have also been adapted to local cultural contexts and language (see Figure 2).

Figure 1. WMO works with its Members to make public outreach materials available in many languages, far beyond the United Nations six languages

Figure 2. The WMO Ocean Buoy awareness is available with voice-over narratives in Fijian, Solomon Island Pijin, Caribbean Creole and English, all with English subtitles

Diversity and inclusion into the future

WMO plays a lead role in the United Nations Early Warnings for All initiative, shoring up gaps in Earth observations, data sharing and processing, and by providing staff capacity development to NMHSs to ensure that everyone everywhere has access to Multi-hazard Early Warning Systems (MHEWS). When danger looms, people at risk must receive timely early warnings in a language or dialect that they understand. The most vulnerable are children, women, older people and those with disabilities, who often do not receive warnings or do not understand the warnings received. WMO must take advantage of every opportunity to improve capacity warning services so that no one is left behind. Two immediate tools at hand are staff recruitment and proactive harnessing of new technologies, such as artificial intelligence (AI).

When it comes to recruitment, diversity and inclusion have a true meaning in the WMO context. The development and promotion of meaningful, timely weather, climate and water-related information services should also aim to secure equal treatment to all and for all, which goes far beyond communication in official languages. The WMO Secretariat promotes such an approach through recruitment that places an emphasis on diversity of staff and consultants, gender-wise and language-wise.

Just a few years back, the use of technologies for early warnings might have seemed like science-fiction, but emerging and developing language technologies have brought it into the realm of possibility. It is by remaining at the forefront of the latest technological developments in the field of languages that WMO has continued to deliver an increasing number of translations every year. Today, speech-to-text programs and captioning are commonplace. The WMO Secretariat moved from “English only” information documents to multilingual documents thanks to computer-assisted and machine translation tools.

This new paradigm opens the door for future solutions that can overcome language barriers, far beyond the official languages. Large language models can be trained for WMO terminology and the benefits of AI can already be observed in some areas. Linguists are the custodians of translation quality who will harness AI to the benefit of multilingualism in the WMO community of experts and users."

#metaglossia_mundus

"Invité de l’émission de Richard Gaitet, le traducteur du russe mais aussi de l’anglais et du breton, entre autres, revient sur sa passion et ses écrits personnels.

ARTE RADIO – À LA DEMANDE – PODCAST

Il commence par taper très vite, sans trop réfléchir. Dit que « quand on lit, les yeux glissent, [alors que] quand on traduit, les yeux plongent ». Celui qui se confie ainsi, c’est André Markowicz, dernier invité en date de Richard Gaitet et de son podcast « Bookmakers ».

Dans le premier épisode, il est question d’enfance – et donc beaucoup de sa mère, sa grand-mère et sa grand-tante. De sa rencontre avec le linguiste et dissident russe Efim Etkind, qui invite l’adolescent à rejoindre son cercle de traducteurs : « La traduction, c’est une passion, une raison de vivre, dit André Markowicz. Le geste premier du traducteur est celui du partage. Il aime et il a les moyens de faire aimer les autres. Ensuite viennent les questions techniques, les impossibles, les affres. »

Les questions techniques, il y viendra à l’invite de Richard Gaitet, et aussi sur ce qui est l’un des chefs-d’œuvre de la littérature : Eugène Onéguine (1833), d’Alexandre Pouchkine, dont André Markowicz s’est attaché à restituer le son. Pour le reste, dans une traduction, il faut que « tout soit là, ajoute-t-il, mais que veut dire “tout” ? La traduction, c’est l’art de la perte. D’où cette question : qu’est-ce que tu ne peux perdre en aucun cas ? »

 

Plonger au cœur des œuvres

A l’épisode 2, il sera beaucoup question de Fiodor Dostoïevski, qu’André Markowicz commence à traduire après sa rencontre, en 1990, avec Hubert Nyssen, le fondateur des éditions Actes Sud. De Dostoïevski et notamment des trois principaux motifs (le poids, la puanteur et la marche) qui constituent et structurent Crime et châtiment – et c’est passionnant, comme à chaque fois qu’il parlera cuisine, c’est-à-dire qu’il fera entendre les vers et nous fera plonger au cœur des œuvres.

Au troisième et dernier épisode, Richard Gaitet invite André Markowicz à parler des quatre recueils de poèmes et du récit autobiographique en vers et sans point (L’Appartement, Inculte, 2018) qu’il a écrits. Et aussi du journal qu’il tient depuis dix ans sur Facebook, et qui constitue la matière première des épais volumes de sa série Partages, parue chez Mesures, maison d’édition indépendante qu’il a créée avec Françoise Morvan, l’indispensable compagne avec qui il a notamment traduit le théâtre d’Anton Tchekhov et celui de Nikolaï Gogol.

 

Soulignons alors que celui qui essaie de ne pas avoir de certitudes, préférant ponctuer ses phrases de « j’ai l’impression » ou encore de « si je comprends bien », participe à la mise en scène de ses traductions, remaniant souvent le texte à l’épreuve du plateau. Ajoutons in fine que ces trois épisodes donnent largement à entendre les textes (en version originale et en français). Qu’il y est aussi – et c’est heureux – surtout question de poésie, et de silence aussi.

« Bookmakers », podcast de Richard Gaitet, réalisé par Charlie Marcelet (Fr., 2023, 3 x 45 min environ). Disponible à la demande sur Arte Radio et sur toutes les plates-formes d’écoute habituelles."

#metaglossia_mundus

"280 nominations pour le Prix Cheikh Hamad pour la traduction

Vendredi 7 Juin 2024

Le Prix Cheikh Hamad pour la traduction et l'entente internationale a reçu pour la dixième édition 280 nominations provenant de 35 pays, indique le comité des nominations.
Cette année, le comité a reçu 195 candidatures de 17 pays arabes et 85 de 18 pays non-arabes, a déclaré Hanan Al-Fayyad, porte-parole et conseillère médiatique du Prix, précisant que la liste des participants à cette édition comprend des individus et des organisations impliqués dans la traduction.
Elle a expliqué que le nombre de nominations dans le domaine du livre unique a atteint 209, réparties entre 164 personnes et 45 organisations, contre 71 nominations dans le domaine de la réalisation, réparties entre 38 individus et 33 organisations.
Les œuvres nominées couvrent diverses disciplines des sciences humaines, notamment la littérature et les études islamiques, la philosophie et les sciences sociales, les sciences politiques et l'histoire.
Le français a été choisi comme langue principale à côté de l'anglais dans la catégorie des livres uniques, après avoir été choisi en 2017, tandis que le baloutche, le tatar, le hongrois et le yoruba ont été choisis comme des langues secondaires dans la catégorie des réalisations, le but étant de "stimuler les efforts de traduction de et vers l'arabe dans les pays arabophones et d'encourager la continuité de l'acculturation entre les sociétés de ces pays et les sociétés arabes", selon Mme Al-Fayyad.
Le prix s'adresse aux individus et aux institutions dans le domaine de la traduction pour honorer les réalisations et les efforts de longue haleine dans le domaine de la traduction et faire valoir l'apport des acteurs dans ce domaine au renforcement des liens d'amitié et de coopération entre les peuples, l'établissement de passerelles de communication entre les nations, et à la diffusion de la diversité, du pluralisme et de l'ouverture.
Lancé par le Qatar en 2015, le Prix vise à promouvoir une culture de la connaissance et du dialogue, et la culture arabe et islamique, à développer l'entente internationale et à favoriser l'acculturation féconde entre l'arabe et les autres langues du monde."

#metaglossia_mundus: https://www.libe.ma/280-nominations-pour-le-Prix-Cheikh-Hamad-pour-la-traduction_a145202.html

 

"Ensuring users find the information they need quickly and efficiently is paramount to a successful search experience. When users find what they’re looking for quickly and effortlessly, it translates into a positive experience.

Furthermore, the ranking position of relevant results also plays a crucial role — the higher they appear, the more valuable they are to the user. This translates to increased user engagement, conversions, and overall website satisfaction.

This article explores the key metrics used for evaluating Search Relevance and Ranking, empowering you to optimize your Search Engine and deliver a superior user experience..."

#metaglossia_mundus

"If you have used Google lately and been lucky—or unlucky—enough to encounter an answer to your query rather than a bunch of links, you have been subjected to something called AI Overviews. This is a new core feature that Google has been rolling out, a move widely anticipated since the company's experiments with its LaMDA large language model in 2021, and since OpenAI's ChatGPT artificial intelligence chatbot rocketed to prominence in 2023.

by Chirag Shah, The Conversation

If you have used Google lately and been lucky—or unlucky—enough to encounter an answer to your query rather than a bunch of links, you have been subjected to something called AI Overviews. This is a new core feature that Google has been rolling out, a move widely anticipated since the company's experiments with its LaMDA large language model in 2021, and since OpenAI's ChatGPT artificial intelligence chatbot rocketed to prominence in 2023.

 

This feature is yet another addition to the increasing number of add-ons and tools being integrated into search engines like Google. Some of the notable examples include knowledge graph-driven knowledge panels, which are used to populate relevant factual information in an infobox next to search results, and featured snippets, which are blurbs excerpted from a search result and provided before the link to that page.

But what's different about AI Overviews is that they are not simply extracted from relevant sources but generated behind the scenes by Google's generative AI technology. The company's goal is to give you a personalized, on-demand answer instead of a standard set of documents or even an answer box matching your query.

This seems almost magical and potentially useful in many situations. After all, people use search engines primarily to find answers and not lists of documents. But there's more to the picture.

My colleague Emily Bender and I have written about what search engine users need, want and have. We have shown that they want not only information but also the ability to discover, learn and question what they find. In other words, users have a wide range of situations and objectives, and compressing them down to a set of links or, worse, a single answer is problematic.

 

Bad advice

These AI features vacuum up information from the internet and other available sources and spit out an answer based on how they are trained to associate words. A core argument against them is that they mostly remove from the equation the user's judgment, agency and opportunity to learn.

This may be OK for many searches. Want a description of how inflation has affected grocery prices in the past five years, or a summary of what the European Union AI Act includes? AI Overviews can be a good way to cut through a lot of documents and extract those specific answers.

But people's searching needs don't end with factual information. They look for ideas, opinions and advice. Looking for suggestions about how to keep the cheese from sliding off your pizza? Google will tell you that you should add some glue to the sauce. Or wondering if running with scissors has any health benefits? Sure, Google will say, "it can also improve your pores and give you strength".

While a reasonable user can understand that such outrageous answers are likely to be wrong, it's hard to detect that for factual questions.

For example, while researching the faith of U.S. presidents, Google's AI Overviews gave the incorrect answer that Barack Obama is a Muslim. This misinformation was widely circulated and debunked years ago, but Google regurgitated it with no good way for users to learn that it is misinformation.

What about a student using Google for homework and asking which countries in Africa start with the letter K? While Kenya does meet this criteria, Google's AI Overviews incorrectly answered that there are no such countries.

Google has acknowledged issues with AI Overviews and said it has addressed them. But the concern remains: Can you really trust any answers you receive through this service?

How to avoid AI answers

There are alternatives. You can always go back to the traditional Google search with its 10 blue links. Click on "More" in the menu—All, News, Images, Maps, Videos and More—directly below the search field at the top of the Google search page and select "Web."

You can then do what you have likely done for decades now—sift through some of the top results, visit a few of those sites and decide for yourself. It does take a little work, but it gives you back the ability to examine multiple sites and evidence to support or refute something. More importantly, you leave open the possibilities for learning, discovery and serendipity.

AI Overviews is like fast food that gets delivered through a drive-through window—it's quick, hot and convenient, but not the healthiest choice. Going through Google's traditional search results is like examining a menu in a sit-down restaurant and placing an order that will take awhile to make it to your table. You can ask your server questions about those items and even request some changes to the restaurant's offerings. It's prepared with more care, customization and control, but also takes longer and may cost more.

These aren't the only methods of finding information, however. There are alternatives to Google's search engine, including specialty search tools.

 

For scholarly needs, Google Scholar, Semantic Scholar and CORE are helpful places to look for research papers and citations. Looking for medical information? Try PubMed, ScienceDirect and OpenMD. For legal needs, some services include Fastcase, Caselaw Access Project and CourtListener.

Concerned about privacy? Check out DuckDuckGo, Startpage and Swisscows. If you still want AI-generated answers, some of the alternatives to Google's AI Overviews and rival Bing's Copilot are You.com and Komo, which provide more transparency about the data they collect about you, provide greater privacy and also offer ways to opt out of having your data collected for training their AI models.

A balanced information diet

Perhaps you can't afford to eat out at a nice restaurant or prepare every meal from scratch every time, but it's important to avoid ending up going through a drive-through for all your nourishment. After all, you are what you eat, and in a similar vein, you are how you search.

It's easy to fall for sensational headlines and bite-size news that lack context. But you don't have to let that define you. You can expand the scope of how you search. It's OK to hit the drive-through every now and then and go for AI Overviews, but it's important to also find other more wholesome ways to fulfill your needs—for food and for information.

Provided by The Conversation "

#metaglossia_mundus

"Artificial Intelligence leaders are struggling with hallucinations. Governments are struggling with how to regulate risks. By Joshua Stein

June 6, 2024

Artificial Intelligence leaders are struggling with hallucinations. Governments are struggling with how to regulate risks.

“How many rocks should I eat, Google?” The search engine answered, “Eat one small rock per day.” That’s not intelligent. Google’s new AI-powered Overviews also repeated conspiracy theories about former President Barack Obama, identifying him, falsely, as the United States’ first Muslim President. Obama is a Christian. These startling errors led Google to vow improvements.

The debacle underlines the challenge of balancing AI innovation with safety. Companies face commercial and competitive pressure to rush to market. They are cutting corners, rolling out these technologies before testing them. Governments are rushing to reign them in, but are slower to act than private companies. Other solutions exist for technology companies themselves to take. One way to minimize and correct errors is to engage and empower users.

AI systems make things up and have biases. Google Overviews couldn’t distinguish between credible and unreliable sources; the advice to eat a rock came from the satirical publication The Onion. OpenAI recently began training its next-generation artificial intelligence software, just after several senior safety researchers quit, accusing the company of prioritizing profits and growth and promoting a culture of “recklessness and secrecy.” OpenAI then hastily appointed a new safety team.

The fast pace of innovation and the promise of vast riches incentivize speed. Many technology companies quickly bring products to the public, then identify problems and address them afterward. Instead of catching bugs before the release, companies offload the process of identifying bugs onto users, fixing them when users complain. Facebook’s pre-2014 motto “Move fast and break things” is a common attitude.

Governments are responding, though slowly. The EU’s AI Act, passed at the end of May, will only take effect in stages, starting at the end of 2024, and running through 2027. It takes time to build and organize resources to enforce regulations. Companies also need time to comply. But, in the meantime, those same companies will continue to release products without those regulations in force.

 

The US isn’t even at the regulatory stage. While the Biden administration rolled out the AI Bill of Rights, this “blueprint” lays out recommendations. It isn’t legally binding. The Biden administration has clarified how existing laws deal with AI. For example, the Justice Department can hold companies accountable for discrimination, even when that discrimination is the result of AI. A US AI Act requires action from Congress.

Another option is to turn to the courts. Lawsuits around AI copyright and competition will influence how companies develop products. Newspapers have sued Open AI and Microsoft alleging that the technology companies have been “purloining millions” of copyrighted news articles without permission or payment to train their artificial intelligence chatbots. US, UK, and EU regulators are also investigating AI leaders including OpenAI, Microsoft, and Nvidia, for antitrust violations.

But US courts, including the Supreme Court, so far have avoided sweeping rulings on key Internet issues.

Without effective regulation, technology companies must act by engaging their users as content moderators. In October 2023, YouTube launched a program to fight disinformation by making money available for training and development in short-form video content. Google and OpenAI should allow users to flag bad results and blind spots. Sites such as Wikipedia and Reddit succeed by allowing users to update information and engage in moderation. Both have moderators and algorithmic tools that flag articles that are being maliciously edited.

Google’s mea culpa blog post addresses some of these concerns. The company now restricts “nonsensical queries” and certain areas of sensitive material, such as politics and public health. This is a useful stopgap. But Google shouldn’t have rolled out a product without adequate testing. Google search remains the Internet’s most powerful engine.

AI safety is challenging. The technology, still in its infancy, is accelerating. Safety requirements and regulations are failing to keep up. Expect more hallucinations and mea culpas.

Joshua Stein recently completed a postdoctoral fellowship at the Georgetown Institute for the Study of Markets and Ethics. His work focuses on ethics, technology, and economics."

#metaglossia_mundus

"Les Glossaires bilingues de l'administration française, conçus et produits par l'association Maison de la Sagesse - Traduire, ont été au cœur de la table ronde « Traduire ou migrer d'une langue à l'autre » organisée par le ministère de la Culture le 22 mai 2024. Entretien.

Publié le 7 juin 2024

La rencontre a tenu toutes ses promesses. Barbara Cassin, de l'Académie française, philosophe et philologue, Constantin Sigov, philosophe et professeur à l'université de Kiev, et Pouria Amirshashi, directeur du Campus francophone en Seine-Saint-Denis, ont échangé le 22 mai 2024 leur vue sur la question (fondamentale au moins depuis Babel et sans doute bien auparavant!) de la pluralité des langues et de la relation entre les peuples : « Traduire ou migrer d'une langue à l'autre ».

L’échange sera précédé d’une présentation des Glossaires bilingues de l’administration française – Pour une compréhension réciproque, publiés par l’association des Maisons de la sagesse – Traduire sous l’égide de Barbara Cassin. Ces glossaires bilingues sont disponibles en arabe, persan et soninké. Ils ont bénéficié du soutien de nombreux partenaires : la Délégation générale à la langue française et aux langues de France du ministère de la Culture, au premier chef, mais aussi le ministère de l’Intérieur, la Délégation Interministérielle à l’accueil et à l’intégration des réfugiés, les éditions le Robert, la Maison des Sciences de l’Homme et Paris Nord. Une ambition hautement généreuse, comme seule peut l’être l’hospitalité, ce trait de civilisation à prendre très au sérieux.

De quoi s’agit-il ? D’aider à la compréhension mutuelle des personnes qui arrivent et des agents administratifs qui reçoivent. Quel est l’obstacle ? La méconnaissance des deux langues, sans nul doute, mais aussi, surtout de deux conceptions de l’administration. Barbara Cassin et Danièle Wozny se sont intéressées à la confrontation, à l’occasion d’un simple formulaire à remplir, de deux représentations différentes de tant de formes culturelles et juridiques qui nous touchent de si près : la naissance, le mariage, la filiation, mais aussi les droits du travail, de la santé, du commerce…

Mis à disposition des principaux acteurs de l’accueil, les étrangers aussi trouvent dans ces Glossaires l’explication de nos usages administratifs, tandis que les agents celle des usages des pays d’origine de leurs interlocuteurs. Ils seront numérisés, afin de faciliter à la fois leur consultation et leur enrichissement permanent. Ils constitueront la référence d’un programme de sensibilisation des acteurs locaux en charge de ces questions (services de l’accueil, administrations publiques, associations et ONG).

Danièle Wozny, vice-présidente de l’association Maisons de la sagesse-Traduire, a dirigé l’élaboration de ces Glossaires dans leur moindre détail. Retour d’expérience.

 

Danièle Wozny, pourquoi ces Glossaires bilingues de l’administration française ?

L’accueil administratif des migrants est nécessaire, tout le monde en convient, et il est souvent une expérience éprouvante, d’un côté comme de l’autre. C’est le cas quand les primo-arrivants ne comprennent pas ce que les agents de l’administration leur demandent et que ces derniers se trouvent très embarrassés de constater que leurs interlocuteurs ne peuvent pas leur répondre. Or, ce blocage est le plus souvent, et tout simplement, le fait de différences culturelles entre les uns et les autres – et nous ne nous en apercevons pas.

C’est pourquoi notre idée a été d’élaborer ces Glossaires pour fournir un outil susceptible de débloquer ce genre de situation : expliquer les mots et les notions de l’administration aux étrangers, en leur précisant, aussi simplement que possible, comment et pourquoi ces formes administratives et juridiques sont là, dans la culture française, d’une part, et , d’autre part, expliquer aux personnes qui accueillent les raisons culturelles de l’incompréhension des personnes qui arrivent.

Nous pensons que, grâce en partie à nos Glossaires, l'administration peut devenir ce qu’elle devrait être en réalité, à savoir hospitalière. Les traducteurs qui se sont plongés très concrètement dans cet univers de l’administration nous l’ont confirmé : l’administration est une garantie de démocratie. Pourquoi, en effet, prendre la peine d’enregistrer l’état civil des gens, sinon pour acter leur citoyenneté et dresser des listes électorales sans lesquelles on ne peut pas organiser de scrutin légitime ? Faciliter l’accueil des étrangers, lors des premiers contacts avec l’administration, c’est aussi faire preuve d’une forme d’hospitalité propre à la démocratie.

C’est aussi pourquoi, au risque de surprendre, je ne suis pas favorable à une simplification excessive du vocabulaire administratif, qui rognerait toutes les spécificités culturelles qui nous différencient. Pour qui sait regarder, en effet, l’administration est une leçon d’histoire formidable, l’histoire infiniment variée de la construction de l’état de droit, même si le millefeuille est parfois indigeste !

 

Comment votre projet a-t-il été accueilli par les intéressés ?

Lorsque nous avons réuni une toute première fois le personnel de l’Etat civil d’Aubervilliers, ses policiers municipaux, et les membres des associations locales, la pertinence des Glossaires n’était une évidence pour aucun d’entre eux ! Heureusement, nous étions venues avec un ami Malien, auquel nous avons demandé d’expliquer pourquoi si nombreux sont les Africains nés le 31 décembre. Ce fait est difficile à comprendre pour de nombreux agents de l’administration et la méfiance s’installe des deux côtés.

Quand un Malien souhaite obtenir un passeport, en effet, il s’adresse à l’administration de son pays, et, dans l’ignorance de sa date de naissance, on inscrit le 31 décembre. Mais pourquoi n’a-t-il pas été déclaré à sa naissance ? La raison peut en être technique (pas de service d’état civil dans son village) ou culturelle : la date de naissance d’un nouveau-né doit, pour protéger ce dernier, rester secrète. Et le prénom peut également faire question. En pays tamoul, par exemple, le prénom d’un enfant ne peut survenir qu’après coup, lorsque la famille a fait connaissance avec le nouveau-né, que les rites ont été accomplis, que le porutham (accord du nom avec l’être de l’enfant) a été trouvé. Et cela ne peut se faire en trois jours.

Une fois que nous avons pu faire comprendre à chacun l’importance des différences culturelles, notre projet a pris, aux yeux de nos partenaires d’Aubervilliers, un sens tout différent. Eplucher en commun les formulaires administratifs pour en éclaircir les termes un par un, dans cette optique à la fois linguistique et anthropologique, et dans l’esprit d’une rencontre mutuelle, n’a plus fait question. Quant aux traducteurs, nous avons pris soin de les choisir parmi ceux qui ont vécu eux-mêmes la difficile expérience des premiers formulaires à remplir lors de l’arrivée en France. Ils sont devenus ainsi des passeurs de langues et de cultures.

 

Il arrive que certaines notions n’existent pas dans une autre culture. Comment vous y êtes-vous pris ?

La différence entre faute légère, faute grave, faute lourde, par exemple, qui est déjà un casse-tête en droit du travail, l’est encore plus lorsqu’il s’agit de traduire. Or, en cas de licenciement, les droits qui découlent de chacune de ses situations sont différents ! La notion de foyer fiscal, autre terme difficile, est devenu en arabe personnes vivant sous un même toit, mais cette périphrase ne peut valoir en soninké, une culture où les familles peuvent compter 60 à 70 personnes ! Les mondes de la finance, du commerce, de la médecine…, nous ont apporté encore d’autres différences intraduisibles, qu’il a fallu décrypter et expliquer.

Et il y a enfin la question de notre rapport au temps : vous savez qu’on demande à un demandeur d’asile de respecter, dans son récit à l’OFPRA, la « chronologie des faits ». Or, les étrangers ont souvent des difficultés à pratiquer la concordance des temps en français. L’absence de concordance des temps dans un récit ne désigne pas nécessairement un manque de cohérence. Par exemple, il faut avoir en tête que le mot boukra, qui signifie « demain » en arabe dialectal, n’est pas à prendre au sens littéral et désigne le futur, à échéance plus ou moins longue. Les usagers de ces Glossaires sont confrontés à une reconfiguration assez radicale de leur vision du monde, non seulement dans l’espace mais dans le temps !

Les traducteurs ont donc été amenés à créer des néologismes. Mais l’essentiel, encore une fois, a été pour chacun d’entre eux d’évaluer tant de subtiles différences, de ne pas succomber aux jugements de valeur dans des contextes difficiles (je pense à l’Iran) et de voir comment mettre à profit ces différences pour enrichir la réflexion de chacun.

De ce point de vue, le travail sur le glossaire Français-Soninké nous a valu des discussions passionnantes avec des réfugiés et des traducteurs. Le soninké connaît globalement quatre dialectes entre lesquels existe une intercompréhension quasi-totale. Le processus de traduction du glossaire a été l’occasion de longues discussions entre les traducteurs des quatre dialectes soninkés sur la richesse de leur propre diversité.

 

Au fond, l’écriture des Glossaires, et leur publication, sont déjà en elles-mêmes un geste d’accueil et d’intérêt pour l’autre…

Il faut comprendre que du côté des journalistes africains notamment, et cela se voit sur les réseaux sociaux, le Glossaire, qui traite la langue soninké et la langue française, la culture soninké et la culture française à dignité égale, prend une dimension symbolique voire politique. J’ai rencontré à la Bourse du travail de Saint-Denis, dans une salle bondée où je venais présenter notre Glossaire, une personne très respectée au sein de la communauté malienne qui m’a dit cette phrase, que je n’ai pas tout de suite comprise : « Madame, bravo ! C’était par-là qu’il fallait commencer. » Sans doute le plus beau des remerciements !"

#metaglossia_mundus