Curras + Baladi: Towards a Levantine Corpus

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 769–778 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 769 Curras + Baladi: Towards a Levantine Corpus Karim El Haff, Mustafa Jarrar*, Tymaa Hammouda, Fadi Zaraket University of Strasbourg, Birzeit University, American University of Beirut karim.el-haff@etu.unistra.fr, mjarrar@birzeit.edu, 1171779@student.birzeit.edu, fz11@aub.edu.lb Abstract This paper presents two-fold contributions: a full revision of the Palestinian morphologically annotated corpus (Curras), and a newly annotated Lebanese corpus (Baladi). Both corpora can be used as a more general Levantine corpus. Baladi consists of around 9.6K morphologically annotated tokens. Each token was manually annotated with several morphological features and using LDC’s SAMA lemmas and tags. The inter-annotator evaluation on most features illustrates 78.5% Kappa and 90.1% F1-Score. Curras was revised by refining all annotations for accuracy, normalization and unification of POS tags, and linking with SAMA lemmas. This revision was also important to ensure that both corpora are compatible and can help to bridge the nuanced linguistic gaps that exist between the two highly mutually intelligible dialects. Both corpora are publicly available through a web portal. Keywords: Arabic morphology, Annotated Corpus, Arabic Dialect, Levantine, Palestinian Arabic, Lebanese Arabic 1. Introduction The processing of the Arabic language is a complex field of research. This is due to many factors, includ- ing the complex and rich morphology of Arabic, its high degree of ambiguity, and the presence of several regional varieties that need to be processed while tak- ing into account their unique characteristics. When its dialects are taken into account, this language pushes the limits of NLP to find solutions to problems posed by its inherent nature. It is a diglossic language; the standard language is used in formal settings and in ed- ucation and is quite different from the vernacular lan- guages spoken in the different regions and influenced by older languages that were historically spoken in those regions. Indeed, Arabic speakers use those local varieties in day-to-day communication. We can distin- guish several families of dialects: Moroccan, Egyptian, Sudanese, Levantine, Iraqi and Khaliji (Gulf). Arabic dialects tend to diverge from Modern Standard Arabic (MSA) in terms of phonetics, morphology, syntax and vocabulary. Arabic content was mainly written in MSA. Recently, dialectal content has been increasing massively, espe- cially on social media. MSA is considered among the under-resourced languages by the NLP commu- nity (Darwish et al., 2021). Dialectal Arabic (DA) is even less resourced. The resource gap between MSA and the dialects implies a large margin of error when MSA tools are used against dialectal content (Zbib et al., 2012). Thus, it is important to build resources and tools to identify dialects in context and to treat Arabic content based on its unique dialectal identity. In this research, we focus on the Lebanese variety of Levantine Arabic, which is used in daily conversations and in the Lebanese media. It is spoken by about 6 mil- lion locals, and almost double that number in diaspora. The paper presents a morphologically …

Content truncated to 500 words