We observe that major works about modelling and simulation within social science, especially for social, organizational and cultural influences on opinion information spreading over a population, do not use specification languages to describe their models. These models are specified in the shape of math formulas and then directly coded using classical programming languages. The specification language can be a missing link. For instance, the DEVS formalism (Discrete EVent system Specifications) being general enough to represent dynamical systems, can provide an operational semantics applicable to this domain. These models independent from implementation are easily reusable. In this article, we recall first the use of discrete modelling approaches in the social influence. Then we present models for human information treatment and propagation using DEVS and Cell-DEVS (Cellular DEVS). Finally, we present a simulation transposed from epidemic models to the impact of information on individuals using CD++, a simulation tool for DEVS and Cell-DEVS.