What capacities enable linguistic interactions? While several proposals have been advanced, little progress has been made in comparing and articulating them within an integrative framework. In this paper, we take initial steps towards a connectionist framework designed to compare different cognitive models of social interactions. The framework we propose couples two simple-recurrent network systems (Chang, 2002) to explore the computational underpinnings of interaction, and apply this modeling framework to predict the semantic structure derived from transcripts of an experimental joint decision task (Bahrami et al., 2010; Fusaroli et al., 2012). In an exploratory application of this framework, we find (i) that the coupled network approach is capable of learning from noisy naturalistic input but (ii) that integration of production and comprehension does not increase the network performance. We end by discussing the value of looking to traditional parallel distributed processing as flexible models for exploring computational mechanisms of conversation.