Postmodernist researchers have been divided about the value of autobiographical writing in their research. One broad group has argued against the use of selfrepresentation, of a stable' self as narrator, in research, on the grounds that such a move is restrictive, essentialist and retrogressive. Another group has viewed the textualized self as necessary to replace the discredited modernist pretense of impersonal objectivity in research writing. This paper analyzes two recent uses of self-representation in curriculum research as indicative of the problems and challenges researchers face in attempting to secure a textual authority through personal evidence and testimony. The paper argues that there are major problems with any attempt to rely on a stable textualized self to achieve new kinds of textual legitimacy, and that any use of self-representation in research needs to be integrated effectively with other kinds of evidence.