There are two important goals to do research on language-extinction problem with computer simulation. One is to describe the process of language getting extinct from the earth in auxiliary; the other is to in auxiliary describe the endangered state before language extinction and make early-warning. At present, there are many researchers focusing on the research on language-extinction problem with computer simulation with different purposes. But what researchers want to achieve tends not to be the same. The research work about the second purpose is less than others. When we take the second purpose in our consideration, it's natural to combine the linguists' definition of language endangered state with computer simulation. It's difficult for us to determine the language's endangered state relying only on a static standard. That is also the main reason why the existing language extinction simulation model can't be compatible with the simulation of the early-warning of endangered state before language extinction. Based on the views of academics in the field of linguistics, this paper gave dynamic assumptions of language endangered state in three different aspects--the proportion of losing mother-tongue population, the proportion of mother-tongue speakers' age range and the ability to use language. Then, we designed the dynamic characteristics of the early-warning simulation model of endangered state before the language-extinction with multi-agent. And we dispatched the mentioned computer simulation model in a discrete event simulation manner. The work on computer simulation did in this paper confirms the subjective viewpoint of endangered language definition in the field of linguistics and reveals the significance of language evolution simulation modeling in the efforts made in rescuing the endangered language

Abstract-In recent years, computing technology has been involved in the field of synchronic phonetics. However, in the research of diachronic phonetics, modern computing technology is always hard to intervene. Human speech system involving a high degree of complexity and many factors, is a typical complex system. The traditional phonetics research methods can not adapt to the characteristics nonlinear complexity of human speech system. Social characteristics in human speech system should be thought as research variables, and non-traditional human speech system evolution research model based on complex system should be applied to diachronic phonetics. First, many basis of this non-traditional research method are given. Then, non-traditional research model for human speech system evolution is proposed. The above work offers some effective intelligent supports for research of diachronic phonetics.