Branco Weiss Fellow Since
2020
Research Category
Computational social sciences
Research Location
Center for Brain and Cognition, Pompeu Fabra University, Barcelona, Spain
Background
There are over 6,500 languages spoken or signed in the world today, yet a handful of large and influential languages -English most prominently in the present- has served as the de facto prototype for how to describe, model and measure language. Research on linguistic diversity has revealed that, along with common properties shared by all languages, there are countless dimensions on which languages can differ from each other, ranging from the readily apparent (such as the vocabularies or the sound or sign repertoires they use) to the subtle and complex (like morphosyntactic strategies for reference tracking in discourse). These differences compromise the capacity to produce truly universal language-based resources, including communication technologies, medical therapies and educational assessments.
Details of Research
Dr. Blasi’s research goal is threefold. First, to quantitatively characterize the mismatch between the (often economically and politically dominant) languages in which language-based resources are produced and the ones where they are applied, across scientific and technological fields and aided by large databases of scientific publications and patents. Second, to deploy computational models to inform, plan and estimate the cross-linguistic generalizability of any given resource as dependent on the languages in which it was conceived and evaluated. Third, to promote a critical dialogue across society, academia and industry on the mostly silent but ubiquitous consequences of this systematic bias induced by large and powerful languages in the production of resources.
