The term autoethnography was described by Reed‐Danahay as ‘a form of self‐narrative that places the self within a social context’. We outline autoethnography’s main characteristics, situate it in relation to the so‐called ‘crisis of representation’, and describe five loosely configured categories of autoethnographic practice. We argue that all types of authoethnography dissolve to some extent the boundary between authors and objects of representation, as authors become part of what they are studying, and research subjects are re‐imagined as reflexive narrators of self. Dismantling the author/represented boundary in this way has implications for how researchers understand their objects of research and ethnographic knowledge itself. The study discusses the relevance for geographers of the various categories of autoethnographic practice and of a broader autoethnographic sensibility.