Arbuli sunnu: a Sicilian-Italian Parallel Treebank

The Natural Language Processing (NLP) community has recently begun to engage with endangered languages and dialects which encode culturally different perspectives and local knowledge. Regardless of the usefulness and applicability of NLP tools for such languages, creating resources for dialects increases our knowledge of them, encourages the community to study them further, and supports the preservation of an important heritage. As part of this endeavour, we are focussing on Sicilian, a dialect spoken in Sicily, with a rich cultural history. Sicilian preservation is crucial to maintaining Southern Italy’s linguistic diversity. In this paper, we present the first release of a novel treebank called Sicilian3bank. On the one hand, to improve the usability of this resource and provide access to non-Sicilian speakers, all sentences are linked to their translation into Italian, resulting in a 1:1 parallel resource. On the other hand, by applying the Universal Dependencies format, a widely used standard for the annotation of treebanks, we pave the way for data-driven cross-linguistic research. We hope that this work can serve as a basis for further linguistic research and computational applications for the Sicilian dialect

CAPPELLO Caterina Maria;  D'ALÌ Sabrina;  GUGLIELMETTI Mario;  DI NUOVO Elisa;  BOSCO Cristina; 

2025-12-09

CEUR-WS.ORG

JRC142844

1613-0073 (online),   

https://ceur-ws.org/Vol-4112/16_main_long.pdf,    https://publications.jrc.ec.europa.eu/repository/handle/JRC142844,