In the last decade of the 20th century a new field of language research emerged that has come to be known as ‘language documentation’ or ‘documentary lingusitics’ (Himmelmann 1998, 2002, 2006; Lehmann 2001; Austin 2010; Grenoble 2010; Woodbury 2003, 2011). In this paper we explore how it was defined in the seminal work of Himmelmann (1998) and others, including what were presented as significant characteristics that distinguished language documentation from language description, and how the field has changed and evolved over the past 20 years. A focus on best practices, standards, tools and models for documentary corpora appeared in the early years, which led later to more critical discussions of the goals and methods of language documentation. The paper examines some current developments, including new approaches to language archiving, and suggests that there are new opportunities for language documentation to adopt a more socially-engaged approach to languages and linguistic research, including better engagement with language revitalisation. There are also opportunities to work towards addressing what is currently a language documentation output gap through experimentation with new genres and and innovations in writing and publication.