This paper explores how 11 Canadian doctoral candidates performed as researchers in the final doctoral oral examination in handling questions that they identified as "difficult to answer". Drawing from communities of practice theory, the study views the defence as an examination in which a novice researcher (the doctoral candidate) demonstrates "knowing" to a group of experienced researchers (the defence committee) in order to be approved of the membership of a scholarly community. The doctoral candidate's researcher identity is considered as composed of three aspects: thinking about oneself as, performing as and being thought of as a researcher. It was found that many (41%) of the questions were difficult because of the questioners' different perspectives on the dissertation; in answering them, the candidates balanced "knowing" and "not-knowing" by navigating across research areas/fields, methodological and epistemological borders. All the defences were successful, which indicate that these candidates were all thought of as competent researchers; yet, only those who thought about themselves as novice researchers felt satisfied with their performance.