Introducing the Autshumato Integrated Translation Environment

Proceedings of the 13th Annual Conference of the EAMT, pages 190–196, Barcelona, May 2009 Introducing the Autshumato Integrated Translation Environment Hendrik J. Groenewald Centre for Text Technology (CTexT) North-West University Potchefstroom, South Africa handre.groenewald@nwu.ac.za Wildrich Fourie Centre for Text Technology (CTexT) North-West University Potchefstroom, South Africa wildrich.fourie@nwu.ac.za Abstract Translation is an indispensable process for socioeconomic and cultural development in a multilingual society. The use of translation tools such as translation memories and machine translation systems can be beneficial in supporting the human translator. Unfortunately, the availability of these tools is limited for resource-scarce languages. In this paper, we introduce the Autshumato Integrated Translation Environment that facilitates a comprehensive set of translation tools, including machine translation and translation memory. The Autshumato Integrated Translation Environment is specifically developed for translation between the official South African languages, but it is in essence language- independent and can therefore be used to translate between any language pair. 1 Introduction South Africa is a culturally diverse country, with a large number of different ethnic groups. This rich cultural diversity is the main reason for South Africa having no fewer than eleven official languages. The South African Constitution (Republic of South Africa, 1996) guarantees equal status to each of the eleven official languages. Translating between these eleven languages is an immense task, and unsurprisingly, English often acts as the primary language for government communication. The problem with English being the lingua franca is that the other 10 official languages are marginalised. © 2009 European Association for Machine Translation. An example of this is the fact that the South African government does not have the capacity to produce parliamentary records in all eleven official languages. Multilingual parliamentary records are one of the most important resources for machine translation (MT) projects elsewhere in the world. One such a project is the EuroMatrix project, a statistical and hybrid machine translation project for translating between European languages (see www.euromatrix.net). The EuroMatrix project uses the parliamentary records of the European Parliament, available in all 23 official languages of the European Union, as one of their main resources for the development of machine translation systems. The parliamentary records of the South African government are only recorded in English on both provincial and national level. The result of this is that a large number of South African citizens are denied access to important government information in their home language, since English is the mother tongue of only 8.2% of the South African population (Van der Merwe and Van der Merwe, 2006). Another negative aspect is that researchers and developers working on the development of machine translation systems for South African languages do not have access to as large amounts of data as their European colleagues. The South African Government is aware of the importance of elevating the status and promoting the equal use of all eleven official languages on all levels. As a result of this, the government is involved in a large number of initiatives and projects to promote multilingualism. One such as …

Content truncated to 500 words