Policy and Practice in the Anonymisation of Linguistic Data

What is anonymisation? This paper addresses this question, its relationship to linguistic data and its potential importance to corpus builders and users. It examines attitudes towards anonymisation such as hostility and disinterest and investigates relevant rights, responsibilities, and obligations. The paper then overviews and critiques methods of anonymisation and seeks to assess which items should be anonymised and which maintained. Finally, some troublesome and noteworthy cases are presented as evidence of the need for sensitive, realistic consideration of this issue. The paper was developed through consultation with researchers from the international community of corpus builders and users and, therefore, reflects the diversity of attittude and practice currently at large. It addresses this variability by finally proposing methods for systematic assessment of the need for anonymisation within individual corpora.