Lexical Gender Made Simple: A Scalable Methodology for Gender Detection with Online Lexical Databases

Anonymous

Abstract: The evaluation of gender bias in Natural Language Processing relies on the use of gendered expressions, such as pronouns and words with lexical gender. Up until this point, researchers have manually compiled lists that record lexical gender for individual words. However, manual compilation leads to static information if lists are not periodically updated and categorization requires value judgements by annotators and researchers. Moreover, words that are not covered by the list fall out of the range of analysis.To address these issues, we devised a dictionary-based method to automatically detect lexical gender that can provide a dynamic, up-to-date analysis with high coverage. Our approach reaches 90 % accuracy in determining the lexical gender of words retrieved randomly from a Wikipedia sample, and when testing on a manually compiled list that the method aims to replace.