Participatory Language Technologies as Core Systems for Sustainable Development Activities

Introduction and purpose: Language is the medium by which people interact with all aspects of their worlds, whether economics, health, the environment, or technology. In both development programs and technology, however, language is usually given secondary consideration, if any at all. As a result, people who do not speak a major language are excluded from full participation in development programs and from technologies such as ICTs that could enhance their economic and social circumstances. In Africa, for example, where only a small minority speaks English or French, few development programs have the resources to devote to the most basic of language considerations, such as translating health information into local languages. Language technology can be a fast and cost-effective way of overcoming knowledge and communication gaps that underlie many other aspects of the development agenda. Design and methods: The most efficient way to address language development is through public tools and vocabularies that can be reused, revised, and repurposed for multiple domains. We discuss a universal multilingual dictionary that is designed to build a parallel vocabulary of core concepts across languages, with a special focus on languages with few existing resources. The lexicons are built in close cooperation with local partners. Much attention is paid to a data structure that will enable downstream technologies. Further, a system develops domain-specific terminologies through a participatory process, so that complicated concepts can be communicated clearly and consistently. Data is made available to the public for free, with strong efforts to develop systems for access via least-cost technologies with the widest reach along the bottom of the pyramid. Results: When successful, a focus on core language development can improve the outcomes of many other projects. In health, for example, translation is often too expensive and too difficult, because basic resources such as dictionaries do not exist and technical terms do not have adequate local-language equivalents. For the one-time cost and effort of building the lexicons and terminologies, in conjunction with the free tools being created to access those vocabularies, the infrastructure opens for cheap and rapid translation of health material. Similarly, students are able to use the lexicons to access knowledge that has previously been blocked behind linguistic barriers, reducing future language-based inequalities. Participatory data collection methods mean that vocabularies continue to grow in response to the expressed needs of particular linguistic and development communities. The perpetual accessibility of the public resource means that development programs can make sustainable use of the data in multiple languages, for multiple purposes, with no further investment. Conclusion: Language is a hidden aspect of the development equation; language technology in itself does not cure a disease or put food on a table. However, whether communicating agricultural techniques, delivering government services, or performing numerous other activities that fall under the rubric of development, attention to developing language technologies for underserved language populations can be the difference between working together and talking past each other – the difference between failing to communicate and succeeding in expressing the path toward accomplishing common goals.

Excluded Linguistic Communities and the Production of an Inclusive Multilingual Digital Language Infrastructure

Martin Benjamin

The consequence of linguistic digital exclusion is the inability of billions of people to access vital knowledge and economic resources that contribute to prosperity in an era of globalization. However, rectifying linguistic inequity is mostly absent from development discourse and the agendas of governments and agencies that undertake development activities. Most efforts to produce content for excluded languages depend on the haphazard occurrence of a commercial, academic, or programmatic purpose for an activity in a given language at a particular moment. The Kamusi Project seeks to address the digital linguistic divide by engaging communities in the systematic collection of codified data for any language – linguistic information that can be used in many kinds of advanced knowledge and technology resources. This paper explores assumptions about participants’ motivations and behaviors that underlie the project’s methods, including participation in online games and interactive mobile apps intended to elicit speakers’ knowledge of their own languages in ways that can be shared by others. While the Kamusi system aims to welcome all, disparities may continue to exclude those without substantial time, network access, equipment, digital experience, or literacy, leaving international members of a diasporic language group as its most active contributors. Further, smaller and more remote languages have, by definition, fewer potential participants and less access for participation, thus perpetuating their inability to jump the digital divide. Without external support for the time and effort necessary to gather linguistic knowledge, even the most carefully constructed tools will fail for thousands of languages spoken by millions of people, including many languages near extinction. This paper raises, without definitively resolving, the social challenges of a multilingual digital infrastructure platform that has the technical capacity to document every word in every language, but can only approach accomplishing this objective through the involvement of those who have the least access to taking part.

2015

Crowdsourcing Microdata for Cost-Effective and Reliable Lexicography

Martin Benjamin

Lexicography has long faced the challenge of having too few specialists to document too many words in too many languages with too many linguistic features. Great dictionaries are invariably the product of many person-years of labor, whether the lifetime work of an individual or the lengthy collaboration of a team. Is it possible to use public contributions to vastly reduce the time and cost of producing a dictionary while ensuring high quality? Crowdsourcing, often seen as the solution for large-scale data acquisition or analysis, is fraught with problems in the context of lexicography. Language is not binary, so there may be no one right answer to say that a word “means” a particular definition, or that a word in one language “is” the same as a particular translation term. People may misinterpret instructions or misread terms or make typographical or conceptual errors. Some crowd members intentionally add bad data. Without a payment system, incentives for participation are slim; micro-payments introduce the incentive to maximize income over quality. Our project introduces a public interface that breaks lexicographic data collection into targeted microtasks, within a stimulating game environment on Facebook, phones, and the web. Players earn points for answers that win consensus. Validation is achieved by redundancy, while malicious users are detected through persistent deviations. Data can be collected for any language, in an integrated multilingual framework focused on the serial production of monolingual dictionaries linked at the concept level. Questions are sequential, first eliciting a lemma, then a definition, then other information, according to a prioritized concept list. The method can also be used to merge existing data sets. Intensive trials are currently underway in Vietnamese, with the inclusion of additional Asian languages an explicit objective.

2015

Excluded Linguistic Communities and the Production of an Inclusive Multilingual Digital Language Infrastructure

Martin Benjamin

2015