In studying benefits and costs of language diversity the use of computer models is rare. There may be various reasons for this situation, e.g. a) the complexity of the interpretation of language diversity, b) the difficulty in operationalizing factors and dimensions of diversity, c) the absence of realistic actor models, that is to say of actors that can communicate in a realistic way and d) the underperformance or restrictedness of existing software programs. They may be only a few of the reasons to abstain from using computers models to analyze, explain and predict language diversity. How valid these arguments may be, it is strange that for weather analyses and forecasts and for tide and streaming conditions many computer models have been developed and are successfully applied. Tides and weather are also complex, many factors and dimensions are also applicable, and software programs and computational power can always be enhanced. The only big difference that we can think about in modeling and comparing natural (tide and weather) and artificial (language and communication) systems is the presence or absence of a realistic model of plausible cognitive actors and the interaction of these actors. It is also possible to defend the line of reasoning that by definition humans systems and their social structures cannot be modeled in software. The existence of cognitive science and social simulation since the 60s of the last century to our opinion shows that the rejection of this kind of modeling is not justified. In our presentation we will propose a pseudo specification of plausible cognitive actors that have a memory, have goals, have communication instruments, i.e., a language, and are able to exhibit social behavior (Helmhout, 2006; Wijermans, 2008).