Interactions between learners are fundamental when implementing connectivist pedagogy. Essentially, the establishment of “connections” among students underpins the learning process. These connections can be interpreted as learners’ available pool of social capital, access to which is leveraged through a distributed learning environment such as a MOOC. This study applies linear mixed models to explore the factors associated with social capital for learners in two connectivist MOOCs. Using Coh-Metrix, a computational linguistics tool, we analyzed interactions distributed via Twitter, blogs and Facebook, and examined how learners’ linguistic characteristics are associated with their social capital distributed within a network. Our analyses show that the language used by learners is related to the creation of ties between them. We also observed the role of media, time and learner activity on the development of social capital. The findings suggest that pedagogical considerations are essential to help learners leverage access to potential social capital in a networked learning context.