Christoph MilerKirsten AlgeraErnst van der Hoeven2025-05-142025-05-142023-10-01https://www.alexandria.unisg.ch/handle/20.500.14171/122659This essay exlores the histories, techniques and limitations of automatic writing correction systems used by software such as Microsoft Word or WhatsApp. Ostensibly an innocent tool to eliminate typos, autocorrect turns out to be a normative force that is forever overwhelmed by the idiosyncratic and unpredictable way humans use language.en-USautocorrectcomputational linguisticlarge language modelslanguage predictionbook section