With the increasing importance of computer-mediated discussions for supporting knowledge management activities, the role of discussion representation in knowledge acquisition can no longer be neglected. In this research, we investigate the potential of semantic network discussion representations. Specifically, we compare linear discussion representations to semantic network discussion representations at different levels of conceptual facilitation restrictiveness in terms of their effects on knowledge acquisition. Based on a field experiment, we demonstrate that semantic network discussion representations enable the acquisition of more complex and better integrated knowledge structures than linear discussion representations. Our empirical results also show that conceptual facilitation restrictiveness hinders knowledge acquisition. The theoretical and empirical implications of these findings are discussed.