Аннотация:Modelling of social interactions by means of mathematical neuronal networks is considered promising, firstly, by imitating and predicting certain social outcomes, dealing with the information processing in social groups and working out a collective decision, and, secondly, by engineering of social groups and building social groups with higher processing capacity in the logic of the neuronal network efficiency. This fact makes it necessary to select special parameters which, on the one side, are essential to keep track in such models since they are totally connected with the fundamentals of the human social nature, and, on the other side, do not excess the model. Such parameters are likely to include speech and linguistic features of social communication.The research is focused on three linguistic factors which seem logical to be included into the sociomorphic neuronal network models: 1) the factor of comprehension or language proximity related to the vector proximity since linguistic variations in social groups are likely to have a continuous nature; 2) the factor of communicative malfunction and distortion; 3) the factor of ideological proximity. The article reviews the possibilities of considering these factors within the structure of neuronal network models.