PUBLICATIONS
The Latest Language News
Let us talk again about the latest language news. I have highlighted what I believe to be the most significant news.
Addressing the status of Jamaican “Patwah”
https://www.jamaicaobserver.com/letters/addressing-the-status-of-jamaican-patwah/
On October 11, 2023, The New York Times carried an article titled A Patois Revival: Jamaica Weighs Language Change as Ties to Britain Fray. Since this publication, which was later redistributed via other media outlets in different countries and languages, including French and Spanish, many Jamaicans both locally and in the diaspora have added their voices, once again, to either support or discredit the push. The majority of Jamaicans are bilingual, with a variation of the Jamaican “Patwah” being the dominant first language, while English is the secondary one. This stance has had several implications on pedagogic approach, curriculum development, and even exam results. This is why those of us who are the experts in first and second language acquisition have made repeated calls for bilingual education and for English to be taught from a second language perspective.
Research conducted by the United Nations Educational, Scientific and Cultural Organization underscores that children’s first language is optimal for literacy and learning throughout primary school. Children stand a better chance of developing their comprehension skills in their native language. In spite of this blatant evidence, many educational systems around the world insist on the exclusive use of one or more, or sometimes several, privileged languages. This means excluding the other language(s) that children speak, which ultimately leads to excluding the children who speak their native language(s). UNESCO has encouraged mother-tongue instruction in primary education since 1953 and has highlighted some of the advantages of mother-tongue education right from the start. With this set-up, children are more likely to enroll and succeed in school; parents are more capable to communicate with teachers and participate in their children’s learning; and rural children with less exposure to a dominant language will stay in school longer and repeat grades less often. In fact, one of the most important benefits is that bilingual or multilingual education tends to develop better metacognitive skills compared to monolingual children.
Promoting Jamaican “Patwah” does not mean English will further suffer. It will require the investment of adequate resources and the equipping of more personnel to help with the transition. Jamaica and its culture are known globally. Jamaican “Patwah” is now taught at York University, Canada; Harvard University, USA; and City College, Birmingham, UK; among other institutions.
Tech breathes new life into endangered Native American languages
Linguistics experts are turning to cutting-edge technologies to revitalize threatened Native American languages – and rejuvenate generations of Indigenous tradition – through new approaches such as children’s books and smartphone apps. In one such endeavor, three Native American women rack their brains as they gather around a computer, trying to remember – and record – dozens of Apache language words related to everyday activities such as cooking and eating. They are creating an online English-Apache dictionary, just one of several projects working to preserve endangered Indigenous languages in the United States. The women are working with Rapid Word Collection software, which uses an algorithm to search Apache text and audio databases for so-called forgotten words. The words are then defined, translated into English, and their pronunciation recorded, so the dictionary’s users will know how to say them properly. Teacher Joycelene Johnson and two of her colleagues validate the definition of the Apache word “kapas,” which means potato in English.
“The applications in the written language are good for (a) non-speaker – at least they’ll have a museum of it where they can go to for reference,” said Johnson, a 68-year-old who teaches Apache vocabulary and grammar. According to her, the bilingual school on her reservation has about a thousand students – but only one, an 11th grader, is fluent in Apache. Johnson spoke at just one of several workshops at the International Conference on Indigenous Language Documentation, Education and Revitalization at the University of Indiana. Representatives from around 40 Indigenous groups from around the world gathered in the college town of Bloomington just days after the United States – which counts about 6.8 million Native American residents, or about two percent of the population – marked Indigenous Peoples’ Day. Linguists, teachers, students, researchers, and Indigenous leaders spent the weekend brainstorming how exactly to rescue these vulnerable languages from the brink.
The software has “increased the efficiency in the workflow,” said Wilhelm Meya, the CEO of TLC and one of the ICILDER organizers – now an Indigenous community can build a dictionary from scratch within a year instead of 20. Those are still just a small fraction of the 400 to 500 Indigenous languages that were spoken in the two countries before the arrival of Europeans and their decimation of native populations some 500 years ago. Paula Hawkins, who teaches the Tahltan language – which is spoken in parts of British Columbia – said she is “really excited” to see an online dictionary, just as her parents helped create the first Tahltan print dictionary in the 1980s. But her colleague, 51-year-old Danielle North King, from the Chemehuevi, or Nuwuvi nation, fears that such projects impose a “Western way of writing” onto “an Indigenous way of speaking” – the vast majority of human languages are solely oral, with no writing systems. Indeed, Lakota Indigenous leaders denounced TLC last year, after the organization tried to copyright teaching material that included recordings from the nation’s elders.
Extinct language Kalkatungu revived with new generation of students in outback town of Mount Isa
Ninety-four-year-old Cecil Moonlight is the last living fluent speaker of the Kalkatungu language. Thanks to the efforts of Uncle Moonlight, several linguists, and Kalkatungu people who have worked across decades, the officially extinct language is set to be revived with a new generation. From the Kalkadoon lands in and around Mount Isa in north-west Queensland, Mr. Blackley, his wife Sheree Blackley, and their team have been working to revive the nearly extinct language of their ancestors. For several years they have been building a Kalkatungu language program, and with the help of linguist Belinda Keller are preparing a school curriculum that incorporates the language into the studies of Prep through Year 9 students.
In the 1950s and 1960s, linguists Barry Blake and Gavan Breen recorded endangered Indigenous languages, including the Kalkatungu language. Thanks to their work, Mr. Blackley and his team were able to build a language program with assistance from the Queensland government’s Local Community Engagement Through Co-design model, which was developed on the basis that local, Indigenous-led solutions were more effective at closing the gap than band-aid government programs. After several years of running the Kalkatungu language program locally, many Kalkatungu people are semi-fluent.
However, there were some gaps in the pronunciation of certain words, which triggered a trip to Alice Springs to meet with the last-surviving fluent speaker of Kalkatungu language, Uncle Moonlight. “The Kalkatungu language was officially extinct because there were less than 12 fluent speakers all those years ago when, luckily, Gavan Breen and Barry Blake recorded them,” Mr. Blackley said. In 2024, students in Prep all the way through to Year 9 at Spinifex State College in Mount Isa will learn the full Kalkatungu language as part of their curriculum. Mr. Blackley said the revival of his people’s language would go a long way to empowering local Kalkatungu culture.
Indigenous Indonesians use Korean letters to save dialect
https://news.yahoo.com/indigenous-indonesians-korean-letters-save-033526878.html
In an eastern Indonesian village, schoolchildren scrawl the distinctive circles and lines of Hangul script on a whiteboard, but the language they are learning is not Korean. It is their own Indigenous Cia-Cia tongue. The language of the Cia-Cia ethnic group in southeast Sulawesi province’s Baubau has no written form, and the syllable-based tongue does not readily translate to the Latin alphabet often used to transcribe Indonesia’s national language. But the Korean Hangul script, developed in the 15th century, shares a syllable-based system that has made it an unusual tool in the effort to preserve and transmit the language of the approximately 80,000 Cia-Cia people. “In Latin words, for instance, there’s no agreed way to pronounce the sounds ‘pha’ or ‘ta.’ But after I learned Korean, it turns out there are Korean characters for the sounds,” 48-year-old teacher Abidin, who goes by one name, told AFP. “They are not exactly the same, but they’re similar.”
Cia-Cia’s language had no surviving written form until 2009, when Hangul was introduced after a cultural exchange between Baubau city and Korean scholars. The decision was the outcome of a concerted push by South Korean linguists, who visited to tout Hangul’s sound-based system as the perfect fit. In ancient times, leaders of the community committed the language to scraps of paper and wood using non-standard symbols that were never passed down and withered away. But now Cia-Cia names dot the city’s schools, streets, and government institutions, rendered in Hangul. The tongue is also taught to students from elementary to high school using Hangul symbols, though it remains largely a spoken rather than written language.
Even the spoken form of Cia-Cia faces pressure from the dominance of Bahasa Indonesia and other regional languages, said Ilyas, a local elder, who goes by one name. “Many words have been lost due to the influence of Indonesian and other regional languages. This has been happening for about 20 years,” the 50-year-old said. Baubau is the only place in Indonesia to use Hangul, and while the South Korean push for the script is not backed by Seoul, its nationalist tint could blur the community’s identity, said Periangin-angin.
What’s the difference between Machine Translation Quality Evaluation and Estimation?
Evaluating the quality of machine translation output is difficult, and assessing MT quality at scale is a major challenge for companies looking to translate large quantities of text with short turnaround times. Language service providers and academia have been reliant on in-depth analysis of machine translation engines based on a “gold standard” segment. Machine translation quality has been measured using a reference metric, such as BLEU, or WER. This is known as “machine translation quality evaluation” and, despite its shortcomings, continues to be the modus operandi in research for testing MT systems. Human-first machine translation evaluation co-exists with a machine-first, hands-off approach to assessing machine translation output, known as “machine translation quality estimation.”
Table: SlatorSource: ModelFront
| Machine Translation Quality Evaluation | Machine Translation Quality Estimation | |
| Measured by: | Humans, using a metric (BLEU, WER, etc.) | Machines, via a system or model |
| Measured using: | Rules and/or manual labeling | Artificial intelligence |
| Outcome: | Read-only score to measure MT model | Score showing which segments need post-editing |
| Used on: | Segments in a test set | New segments in production |
MTQE differs from machine translation quality evaluation in that the machine assesses the quality of MT segments, and it does not rely on manual scoring. At a practical level, MTQE is already integrated into translation management systems to assess the quality of a translation as part of the translation workflow. A TMS applies a translation memory to a given text, it applies machine translation, and then MTQE kicks in, giving the MT segments a score from 0-100. This drives innovation and efficiency at the beginning of the translation process, as the system decides which segments require an expert-in-the-loop.
On the topic of MTQE at SlatorCon, Conchita Laguardia, Senior Technical Program Manager at Citrix said, “The traditional way of looking at quality management is always very language-based. What you’re asking [the industry] now is actually to trust a machine to tell where the MT has gone wrong.” Language service providers and localization buyers can access this technology through tools such as KantanQES from KantanAI, through a TMS such as Smartling or Phrase, or via a connector to MTQE specialist providers, namely TAUS or ModelFront. More recently, ModernMT by Translated has rolled out MTQE functionalities, as well as RWS with RWS Evolve. With the growing use of large language models in the translation industry, Unbabel has also released the first open-source LLM specifically fine-tuned to predict translation quality, as a development from its predecessor, OpenKiwi, opening up the possibility for further development in MTQE models.
Machine Translation Specialist, R&D
Artem Eidel
Artem Eidel is a Machine Translation Specialist in R&D. His responsibilities include creating and maintaining MT models, evaluating MT quality, and selecting suitable MT providers for different language pairs and subjects.
