Proceedings of the 4th Celtic Language Technology Workshop within LREC2022

These proceedings include the programme and papers presented at the 4th Celtic Language Technology Workshop (CLTW 4), co-located with the Language Resources and Evaluation Conference (LREC) in Marseille, June 2022. In Classical times, Celtic languages were found across a wide swathe of modern Eurasia. Today, they are spoken in regions of the UK, France and Ireland, as well as in emigrant communities in Argentina and Canada. The modern languages are: Breton, Cornish, Irish, Manx, Scottish Gaelic and Welsh. Although the hereditary communities of these languages are small compared to those of most other European languages, they continue to have a vibrant presence in their traditional areas as well as in urban centres. While Irish is the only Celtic language that has official EU language status (since 2007), Welsh, Gaelic and Manx have co-official status. Breton and Cornish also have some limited status in their home regions. That said, all Celtic languages face the same issue in lacking natural language processing (NLP) resources to ensure continued technology support in the digital era. Until recently, the Celtic languages lagged behind in the areas of NLP and applied language technology. Consequently, research and resource provision for this language group was poor. In recent years, however, some Celtic languages have been able to benefit from improved provisions for under-resourced languages in academia and the tech industry. Some now also have dedicated research teams working on language and speech processing technologies and related resources. The CLTW community and workshop, inaugurated at COLING (Dublin) in 2014, provides a forum to help connect these researchers and their associates to one another, to disseminate cutting-edge work and to raise the profile of Celtic language technology, more generally. The accepted papers cover an extremely wide range of topics, including: computer-assisted language learning (CLL); automatic speech recognition (ASR), handwriting recognition; speech synthesis; syntactic parsing; part-of-speech tagging; NLP with mediaeval languages and coreference resolution. We thank our invited speaker, Prof Kevin Scannell of Saint Louis University. We also thank our authors and presenters for their hard work, and workshop attendees for their participation. We are also very grateful to our programme committee for reviewing and providing invaluable feedback on the work published.