Design and Implementation of Social Persuasive Ubiquitous Knowledge Systems to Revitalise Endangered Languages

Abstract

One language and its culture dies every two weeks. If nothing is done, half of the over 7105 languages spoken today could disappear by the end of this century. Only 600 or so languages with large numbers of speakers (more than 100,000) may survive. When a language is extinct, communities lose their cultural identity and practices tied to that language and its intellectual wealth. The preservation of endangered languages is a vital, but challenging effort as both preservation and revitalisation are heavily dependent on the usage of the language. The current state of language endangerment highlights the need to revitalise endangered languages for the survival of diverse cultures and knowledge. Current language revitalisation efforts, approaches, and systems have been fragmented and predominantly focused on language learning, while there has been limited research on models, frameworks, and/or architectures for the design and implementation of integrated language preservation and learning systems. An integrated and holistic approach needs to be adopted for the preservation, curation, transmission, and usage of endangered languages and the cultures embedded therein. The inherent strengths of exponential information systems and technologies have the potential to be a great help here. The vision of this research is to revitalise endangered languages for future generations by leveraging concepts, models, processes, frameworks, and architectures from social media, knowledge management, collaboration, persuasion, gamification, and ubiquitous systems and devices. The five key objectives of this research are to: (a) explore concepts, models, processes, and frameworks available in current literature and industrial systems that support language revitalisation; (b) propose concepts, models, processes, frameworks, and architectures for the capturing, curating, learning, and usage of endangered languages; (c) create and implement a socially driven persuasive ubiquitous knowledge system that helps manage fragmented knowledge, diversity, and richness of endangered languages; (d) validate all the aforementioned artefacts through various evaluation techniques; and (e) generalise the research artefacts to create a platform that can be easily configured and customised to preserve and save endangered and mainstream languages, and thereby, the associated cultural artefacts and culture. This research applied a generalisation driven design science research methodology, which consists of multiple interconnected iterative phases: observation, theory building, system development, evaluation and generalisation. The research resulted in a number of conceptual and system artefacts such as concepts, models, processes, workflows, frameworks, architectures, and implementations. The core implementation - Save Lingo platform helps to harness collective intelligence to revitalise endangered languages through four distinct processes: (1) Capture – words, phrases, songs, stories, and idioms in different dialects; (2) Curate – filter and approve content by language experts; (3) Discover – search and explore categorised content; and (4) Learn - creating dynamic learning modules to promote use of the language. Significant mechanisms were used to evaluate the generalisability of this research and whether its components would apply to other languages. Initially, the system was instantiated and validated in the context of Te Reo Māori, the native language of the indigenous population of New Zealand. The research artefacts were generalised to support Roman script-based languages. The architecture and systems were further enhanced and extended to support both Roman and Non-Roman scriptbased languages. We have implemented and validated the system to support multiple languages including Arabic, Hawaiian, Hindi, Taiwanese, Tamil, Telugu, Te Reo Māori, Vietnamese, and Urdu. Furthermore, the research artefacts were iteratively refined, evaluated, and peer-reviewed through their publication and dissemination in a journal article, three book chapters, and five conference presentations and proceedings.