We Are Depleting Our Research Subject as We Are Investigating It: In Language Technology, more Replication and Diversity Are Needed

We Are Depleting Our Research Subject as We Are Investigating It: In Language Technology, more Replication and Diversity Are Needed Ant´onio Branco University of Lisbon NLX-Natural Language and Speech Group, Department of Informatics Faculdade de Ciˆencias Campo Grande, 1749-016 Lisboa, Portugal antonio.branco@di.fc.ul.pt Abstract In this paper, we present an analysis indicating that, in language technology, as we are investigating natural language we are contributing to deplete it in the sense that we are contributing to reduce the diversity of languages. To address this circumstance, we propose that more replication and reproduction and more language diversity need to be taken into account in our research activities. Keywords: science ethics, science policy, language technology, human language diversity, replication of research results, lan- guage policy. 1. Introduction Natural language is a most extraordinary object of scientific inquiry lending itself to be researched at least as a referen- tial symbolic system, a socially effective type of behavior or a class of specialized mental activities, and hopefully one day as a principled unified combination of all its di- mensions. As it is at the core of what distinctively human nature may be, the approximately 7 000 human languages existing in our planet are a most valuable treasure trove for scientific inquiry on the human brain, mind and behavior, and for advancing our understanding of ourselves and find- ing better technological solutions that improve our life and heal us. While informed laypersons are aware of the dramatic con- sequences of the depletion of important natural resources, from energy to bio-diversity and including potable water, ozone layer among several others, they are much less, or not at all aware of the threat hanging over language diversity. Around one third of the world languages are at present vul- nerable to become extinct according to UNESCO (Moseley, 2010). In this paper we start by pondering on the impact that the very development of language science and technol- ogy at large is having on language diversity. We will then proceed with this analysis by narrowing the focus of our re- flection into a case study in the realm of language resources. With the present paper, we aim at fostering the debate and action on how language scientists and the research activities on language science and technology, including on the de- velopment of language resources, can be much more mind- ful of the language diversity issue and must bring it to the center stage of its mission. It is not only natural and human heritage but also their very own object of research that is being eroded as they are investigating it and because they are investigating it they way that investigation is being un- dertaken. 2. Monolingual vortex In the prevailing model for the promotion and funding of research, science progress is mostly driven by the soci- etal priorities identified in the different countries and en- tities supporting its development. The growth of scientific knowledge is thus asymmetric among different areas, dif- ferent disciplines inside a given area, different topics …

Content truncated to 500 words