Purpose - Throughout its history, information retrieval has struggled to handle contradictory needs of system oriented and user-oriented research. Information retrieval has gradually, starting in the 1960s, moved toward handling the needs of the user. This paper aims to consider the way boundaries toward the user and user-oriented perspectives are drawn, renegotiated and re-drawn. Design/methodology/approach - The central concept of relevance is seen as a boundary concept, complex and flexible, that is continuously redefined in order to manage boundaries. Five influential research papers from the 1960s and early 1970s are analysed in order to understand usage of the concept during a period when psychological and cognitive research tools began to be discussed as a possibility. Findings - Relevance does not only carry an explanatory function, but also serves a purpose relating to the identity of the field. Key contributions on research on relevance seems to, as a by-product, draw a boundary giving legitimacy to certain theoretical resources while demarcating against others. The strategies that are identified in the key texts are intent on finding, representing, justifying and strengthening a boundary that includes and excludes a reasonable amount of complexity associated with the user. Originality/value - The paper explores a central concept within information retrieval and information science in a new way. It also supplies a fresh perspective on the development of information retrieval during the 1960s and 1970s.