D1.17 Report on the greek language

Abstract

This report is part of the European Language Equality (ELE) reports series that seeks to not only delineate the current state of affairs for each of the European languages covered, but to additionally – and most importantly – identify the gaps and factors that hinder further development in Language Technology (LT). Identifying such weaknesses lays the groundwork for a comprehensive, evidence-based, proposal of required measures for achieving Digital Language Equality in Europe by 2030. The report at hand sketches the state of affairs for the Greek language, the official language of two EU member states, Greece and Cyprus. Following a brief introduction to the history, prominent linguistic features, writing system and dialects of Greek, the report focuses on the presence of Greek in the digital sphere; this section discusses the progressive digitisation of the Greek and Cypriot societies, the slow but steady prevalence of Greek-based digital tools and applications that replace previously dominant English-based ones, and the availability of Greek digital public open data. It gives an overview of the status of language resources, and of tools, services and applications concerning Greek; it presents and discusses the variety of language resources for Greek, either intended for human users or supporting LT systems: corpora (general and domain-specific, synchronic and diachronic, monolingual, bilingual and multilingual, parallel, written, spoken and multimodal corpora), lexical/conceptual resources (e. g. computational lexica and online dictionaries, terminological lists and glossaries, thesauri), language models and LT tools and services for the processing of Greek, either for written text or for speech, provided by the R&D community or by the local industry and found at different levels of maturity. The national policies concerning Artificial Intelligence (AI) and LT are briefly reviewed, in particular with respect to the prevalence of language-centric AI or LT. The Cyprus Strategy for AI and the central role reserved for AI in the Greek Digital Transformation Bible indicate that national policy makers have full understanding of its importance. However, the fact that a vision and specific plans for supporting LT are missing from strategic documents indicates that the fundamental contribution of LT to achieving ubiquitous human-centred AI has not been adequately recognised yet. On the positive side, the establishment of a research infrastructure dedicated to Language Resources coupled with a sister infrastructure dedicated to the Humanities is evaluated as a critical facilitator of LT development. The presence of AI and LT is also gradually attested to in the academic domain: almost all Greek Universities offer courses (mainly at the postgraduate level) on Natural Language Processing and Language Technology, either as autonomous courses, or coupled with Artificial Intelligence, Big Data and Data Science. With respect to entrepreneurship in the LT and AI domain in Greece and Cyprus, the local industry is small but active; it consists mainly of SMEs offering services both in the country and abroad. The report concludes that technological support for Greek has progressed in the past decade, while digital language resources have both increased in volume and improved in quality and variety. A critical factor for the availability of resources and tools for Greek has been the creation of Language Resources Infrastructures that cater for storage, curation, and distribution of datasets and technologies/services, properly described with metadata and accompanied by clear and explicit licensing terms. Despite this progress, when compared to the so-called big languages, Greek is obviously disadvantaged. Prominent among the challenges impeding the development of LT for Greek, is the fact that LT is not included in the overall language policies or in the AI strategies of Greece and Cyprus, while the recognition of the significance of language-centric AI is still lacking. Lack of continuity in research and development funding is an additional progress hampering factor. In conclusion, there is a desperate need for a large, coordinated initiative focused on overcoming the differences in language technology readiness for European languages.