Background: Research coproduction is advocated as an approach to produce more impactful evidence, by valuing a diversity of expertise and integrating knowledge users into research processes. Yet, extant literature finds that trying to bridge boundaries between different types of knowledge can also cause collaboration challenges and present barriers to success in coproduction. Aims and objectives: To study how researchers understand and manage knowledge boundaries in coproduced health research, or 'integrated knowledge translation' (IKT) as it is referred to in Canada. Methods: Data were collected from: 1) semi-structured interviews (n=20) with researchers leading different IKT projects across Canada; and 2) participant observation and document analysis for an in-depth case study of one IKT project. Data were combined and analysed using situational analysis, a modified grounded theory approach to visually map patterns of discourse along salient axes of controversy. Findings: We describe four key discursive positions participants take concerning knowledge boundaries in IKT: to recognise and handle, respect and clarify, blur and integrate, or challenge and embrace. These are plotted relative to two salient axes: the degree to which participants viewed boundaries as a problem, and the degree to which they believed boundaries should (or could) be challenged. Discussion and conclusion: The four discursive positions identified will help those doing coproduced research to critically reflect on their own position(s) regarding boundaries in collaborative research, and strategically discuss, select, or switch discourses as needed to support their goals.