I should not have said that: Measuring contextually inappropriate language in different social settings

Abstract

Everyone has found themselves in a scenario where they think to themselves should I say that? The goal of this project is to investigate how context affects language appropriateness using computational methods. Involved in this is building a dataset of language-relationship-appropriateness sets for use in our work and for further research. We also are interested in what social information our model captures, and how it shows that appropriateness in common language is dependent on social context.