Anna Matamala (Universitat Autònoma de Barcelona)
Easy-to-understand language: practice, training and standards
This presentation will provide an overview of easy-to-understand language, ranging from Easy Language (or Easy-to-Read) to Plain Language. It will present some of the main challenges in terms of terminological choices and end user definition, and provide an overview of the current situation of easy-to-understand language training and practice across Europe. To this end, the results of a 3-year Erasmus + project funded by the European Commission (2018-2021) will be used: EASIT, Easy Access for Social Inclusion Project (pagines.uab.cat/easit). It is an innovative educational project that stresses the need to go beyond written documents and considers how easy-to-understand language could be integrated in audiovisual media and audiovisual access services.
The presentation will also highlight the need for standardisation and further research that supports current standardisation efforts. More specifically, it will refer to the Spanish standard on Easy-to-Read and to the ISO standard on making written text easy to read and easy to understand (ISO/IEC CD 23859-1), currently under development.
This introductory presentation will set the ground for other presentations in the workshop which will adopt a more technological approach, considering tools that support writing processes in easy-to-understand language.
Anna Matamala, BA in Translation (UAB, Universitat Autònoma de Barcelona) and PhD in Applied Linguistics (Universitat Pompeu Fabra), is an associate professor at UAB. Currently leading TransMedia Catalonia, she has participated and led many projects on audiovisual translation and media accessibility and has published extensively in international impact journals. She is involved in standardisation work at Spanish UNE and at ISO, where she has co-edited a standard on audio subtitles and is currently co-editing a standard on easy-to-understand language. More information: gent.uab.cat/amatamala.
Klaus Miesenberger (Johannes Kepler University Linz, JKU)
The H2020 Easy Reading software framework
The Easy Reading software framework, a browser extension for users with cognitive disabilities and a backend for care providers, support personalising web content in real time. This is achieved by:
- Adaption of layout and structure of web pages
- Annotation of web content with symbols, videos, …
- Conversion and Translation of content into a different language level, e.g. Plain Language or Easy-to-Read
The personalisation is only done on demand, allowing the user to remain and work within the original content. This fosters independent access, supports learning to cope with original content and keeps the user in the inclusive discourse to enable participation and contribution. The backend for care providers and web owners allows for customising and controlling how web content is displayed for individual users or for groups of users.
The Easy Reading tool, developed in cooperation with end users as co-researchers, allows for addressing cognitive accessibility at the highest level without separating the user from the original content.
The project has received funding from the European Union‘s Horizon 2020 research and innovation programme under grant agreement No.780529.
Dr Klaus Miesenberger is head of the Institut Integriert Studieren (www.jku.at/iis). His R&D and teaching activities in Computer Science encompass Assistive Technology, Accessibility, Design for All, Usability, Human-Computer Interaction, and Socio-technical Inclusion. He is founder/CEO of JKU’s Service Centre for Students with Disabilities. He is chair of ICCHP (International Conference on Computers Helping People with Special Needs, www.icchp.org), and founder/chair of the ICC (Summer school for disabled university beginners, www.icc-camp.info), of the Regional Competence Centre IT for People with disabilities (KI-I, www.ki-i.at), and of BookAccess (Accessible teaching/learning materials, www.bookaccess.at). He is Past-President and secretariat of AAATE (www.aaate.net), active member of www.ocg.at, www.ifip.org, www.acm.org and www.uniability.org, and expert for the European Commission and the Austrian Government for AT, Accessibility and elnclusion.
Susanna Laurin (company Funka, Sweden)
Norm-breaking research: creating understandable content for migrants
This presentation discusses Funka's projects related to simplified language. The focus will be on a project carried out together with a local government involving a population covering a large number of different languages and a high percentage of recently arrived migrants. In this project, Funka has investigated, tested and created guidelines on how to best create information that reaches these target audiences when translating into all native languages is not feasible. Other projects of Funka concerning simplified language involve the conversion of ordinary text from public sector websites into plain language using machine learning, and the communication of information and instructions from public sector agencies to disabled people, using multimodal means (text combined with video, audio and images).
Susanna Laurin has been Funka’s CEO and Chief Research and Innovation Officer since 2003. In the Web Accessibility Directive Expert Group of the European Commission, she chairs the subgroup focusing on monitoring methodology. She works on the development of EN301549 in the Special Task Force 536 of ETSI (European Telecommunications Standards Institute), chairs the Global Leadership Council of IAAP (International Association of Accessibility Professionals), and is joint representative to the EU for the UN-initiative G3ict (Global Initiative for Inclusive ICTs) and IAAP. She acts as Managing Director of the first local chapter IAAP Nordic and is expert and advisor for Swedish Standards Institute and Standards Norway. She served as technical expert for Mandate 376, that developed harmonised requirements for accessibility in public procurement in all EU member states.
Vincent Vandeghinste (Instituut voor de Nederlandse Taal / Dutch Language Institute)
Easy Language in Belgium - prospects for automatic simplification
In the summer of 2021, the Handbook of Easy Languages in Europe is released. According to Lindhom & Vanhatalo (2021), "the notion of Easy Language refers to modified forms of standard languages that aim to facilitate reading and language comprehension. Easy Languages have been adapted in terms of content, vocabulary, and structure to make them more readable and comprehensible. They are intended for people who have difficulties understanding standard language."
This talk presents the chapter about Belgium, that was co-written with Adeline Müller, Thomas François and Orphée De Clercq, zooming in on the needs for Easy Language, the past and current status of Easy Language and Plain Language, and their target groups.
A second part of the talk presents a number of writing aid tools for Dutch and some Easy Language resources and their linguistic properties, and how they could be used to train automatic simplification systems.
Dr Vincent Vandeghinste is a researcher at the Dutch Language Institute (Instituut voor de Nederlandse Taal) in Leiden, the Netherlands, and at the Centre for Computational Linguistics and Leuven.AI (the Leuven Institute for Artificial Intelligence) at KU Leuven in Leuven, Belgium. His work focusses on inclusive language technology, on machine translation, and on linguistic infrastructures. He is teaching in the courses on "Machine Translation and Post-Editing" and "Language Engineering Applications" at KU Leuven.
Thomas François (CENTAL, IL&C, UCLouvain)
Natural language processing can support clear writing: the example of the AMesure platform
Recent research (e.g. Nord, 2018) revealed that, although several plain language guides are available to assist writers of administrative texts in their work, the guidelines provided are not always followed by writers, mainly because they are too vague and too numerous. To support the practice of clear writing and the application of all these guidelines, we have designed a web platform, AMesure, which aims at automatically identifying clear writing issues in administrative texts and provides simple writing advice that is contextually relevant (François et al., 2020). In its current state, the platform offers four main functionalities: (1) providing an overall readability score based on a formula tailored to administrative texts; (2) identifying, in a text, the linguistic phenomena that are assumed to have a negative effect on the comprehension of the text; (3) for these phenomena, proposing simplification advice from clear language guides; and (4) automatically generating simpler synonyms of difficult words.
This talk will briefly sketch various opportunities offered by the field of natural language processing to clear writing studies. Then, based on a recent survey of the practice of professional writers (Müller et al., 2021), it will report some current challenges and opportunities for clear writing and discuss how a platform such as AMesure could assist writers of French administrative texts to simplify their writing.
Thomas François is an Assistant Professor at UCLouvain in Applied Linguistics. He is currently leading a team of nearly 15 researchers and is coordinating several projects focusing on efficient communication in business contexts, automatic readability assessment, automatic text simplification, and computer-assisted language learning. He achieved his Ph.D. at the Centre for Natural Language Processing (CENTAL, UCLouvain) and has received the best Ph.D. Thesis award by the ATALA (Association pour le Traitement Automatique des Langues) in 2012. He then spent a one-year research stay at IRCS (University of Pennsylvania) as a B.A.E.F. and Fulbright Fellow. He became Assistant Professor at UCLouvain in 2018. The main projects he led are CEFRLex, a CEFR-graded lexicon for foreign language learning and AMesure, a platform to support simple writing (more information on http://cental.uclouvain.be).
Fernando Alva-Manchego (University of Sheffield)
Challenges in evaluation of automatic text simplification
Most current models for Automatic Text Simplification are data-driven: given a large dataset of parallel original-simplified sentences, models are trained to implicitly learn how to perform a variety of editing operations that aim to make a text easier to read and understand. However, how do we verify that an automatic output is actually 'simpler' than its original version? As is the case for many Natural Language Processing tasks, this should be done using both automatic and manual assessments. In this talk, Fernando will first present the results of a meta-evaluation of automatic metrics for Automatic Sentence Simplification, and will show how much the correlation between metrics and human judgements is affected by factors such as the perceived simplicity of the outputs, the system type, and the set of references used for computation. After that, he will present some preliminary results on a study of joint Translation and Simplification, and show how difficult it can be for lay users to manually assess simplicity. He will conclude with some recommendations and ideas for future work in evaluation of automatic simplifications.
Fernando Alva-Manchego is a Postdoctoral Research Associate at the University of Sheffield and a member of the Natural Language Processing Group. He holds a PhD from the University of Sheffield, and his thesis focused on Automatic Text Simplification with Multiple Rewriting Transformations. Currently, he works on Quality Estimation for Machine Translation as part of the Bergamot project (EU’s Horizon 2020) and co-leads a project on Readability-Controlled Machine Translation, sponsored by the European Association for Machine Translation. His research interests include Automatic Text Simplification, Readability Assessment and Evaluation of Natural Language Generation.
Daniela Zambrini (Freelance technical translator)
Less is more: keeping it short and simple with ASD-STE100
This presentation will focus on the purposes of Simplified Technical English (STE), illustrating the structure of the ASD-STE100 Specification and its advantages for translators and technical authors. ASD-STE100 is a controlled natural language which was developed in the early eighties (as AECMA Simplified English) to help the users of English-language aircraft maintenance documentation understand what they read. STE addresses difficulties in English comprehension that are related to complex sentence structures, confusing word forms, and ambiguous vocabulary. Although STE was created to improve aircraft maintenance documentation, its principles can dramatically improve the reading quality of technical documentation in any industry. Its advantages in the translation process are directly related to the reduced chance of risk and ambiguity.
Beware! The word “simplified” may somehow sound misleading. As the quote that is spuriously attributed to Nathaniel Hawthorne states: Easy reading is damn hard writing!
The presentation is structured as follows:
- Brief introduction on the history and structure of STE
- One word, one part of speech, one meaning
- Procedural and descriptive writing in STE
Daniela Zambrini has been working in the airline industry since 1987 and as a freelance technical translator for over twenty years. She specialises in aviation, defense, logistics and maritime topics (Italian to English and vice versa). Daniela grew up in a bilingual family and focused her studies on foreign languages, achieving her MA in Foreign Languages and Literature in 1992 at Rome University “La Sapienza”. In January 2013 she successfully completed a Postgraduate Master in Specialised Translation at the University of Pisa. In 2014, Daniela joined the STEMG as an Associate Member. The mission of the Simplified Technical English Maintenance Group is to control, develop and maintain the ASD-STE100 specification.