Written by Luisa Antunes
On 8 November 2022, STOA held a workshop on the research and development environment of language technologies in the context of EU multilingualism. This event follows STOA’s 20-year commitment to scientific developments in language technologies, including the STOA workshops in 2013 and 2017. A 2017 STOA study led to the European Parliament’s 2018 resolution on fostering language equality in the digital age.
At the beginning of the event, STOA Panel member Jordi Solé (Greens/EFA, Spain) stressed the importance of protecting multilingualism in the European Union, as a reflection of EU’s sociocultural diversity, which rests on more than 90 national, regional and minority languages on the European continent.
In the event’s first session, the main conclusions of the first part of the EU-funded European Language Equality (ELE) project were discussed. As main project coordinator, Professor Andy Way introduced the ELE project and its strategic agenda to achieve full digital language equality in Europe by 2030. The next steps in the project include developing large, open-access language models, fostering data availability and investing in computer infrastructure. Ahead of the event, Professor Way was interviewed by the European Science-Media Hub.
Dr Maria Giagkou presented the main results of the first stage of the ELE project. There is considerable inequality in technological support between the over 90 languages spoken on the European continent. English, French, German and Spanish dominate the EU official language sphere, with the digital gap between English and these languages increasing respective to 10 years ago. At the extremity of official languages, Irish and Maltese have the smallest amount of dedicated data, tools and technological resources. Across the continent, the best-supported regional languages are Basque, Catalan, Galician and Welsh. More details are available in the project’s deliverables.
The status of Irish, Bulgarian and Basque language technologies was then presented as EU examples and Professor Georg Rehm provided recommendations for achieving full language equality. Overall, there is a need to increase awareness of the value of open-data sharing, educational programmes and long-term political support strategies. Recommendations include assigning multilingualism to the portfolio of a Commissioner, working towards reducing the technological gap between English and other languages, and strengthening the interoperability of existing infrastructures.
This first session was followed by a panel discussion involving Jordi Solé (Greens/EFA, Spain) and Yana Toom (Renew, Estonia), policy analysts from the European Parliament and the European Commission and European language technology experts.
Yana Toom (Renew, Estonia) introduced the Minority SafePack citizens’ initiative, which had called for the protection and promotion of cultural and linguistic diversity in the EU, in particular for the protection and use of regional and minority languages in public administration and services. The European Commission did not take up the citizens’ initiative, considering the legislation and policies already in place as sufficient. Yana Toom commented that there is a lack of political will to foster language diversity in the EU, a view also shared by Jordi Solé, who pinpointed the need for a single Commissioner dealing with Multilingualism, as between 2007-2010 following the EU’s latest enlargement. This point was also underlined by Davyth Hicks, from the European Language Equality Network.
June Lowery-Kingston, working in language technologies at the European Commission, illustrated the impact of language technologies for social cohesion and economic growth. One example is the Commission’s innovative eTranslation project, which covers all EU languages and is also adding further languages based on the social, political and economic needs of Member States, currently Arabic, Chinese, Japanese, Russian, Turkish and Ukrainian, for instance.
Language technologies are also important for an inclusive communication that is accessible to all EU citizens. Magdalena Pasikowska-Schnass, policy analyst at the European Parliament, focused on sign languages, as well as minority languages such as Karaim, Ladino, Tartar and Yiddish languages. The Horizon 2020-funded project SignON is developing automatic translation tools to and from sign language.
Data availability is also important for advancing digital language equality. The Commission is working on the creation of the Language Data Space. Barbara Plank stressed the importance of access to data, as well as computational infrastructure and funding for natural language processing (NLP) research. Professor Way noted the issue of balancing privacy (e.g. General Data Protection Regulation) needs – a complex scenario further complicated when dealing with video and audio data.
The European Language Equality project is now moving on to its second stage, which is more ambitious in terms of the languages studied, tools explored and end goals, in view of ensuring full digital language equality in the EU by 2030.