ISSN 0439-755X
CN 11-1911/B

中国科学院心理研究所

• 论文 •

主语表征预报思维指向状态的马尔科夫链模型

1. (1广州大学心理辅导中心, 广州 510006) (2广州大学数学与信息科学学院, 广州 510006)
• 收稿日期:2012-01-06 出版日期:2012-09-28 发布日期:2012-09-28
• 通讯作者: 熊健
• 基金资助:

教育部人文社会科学研究规划基金项目(11YJAZH106)。

The Markov Chain Model for Forecasting of Subject Representation to Thinking Point State

XU Peng;XIONG Jian

1. (1 Psychology Consulting Center, Guangzhou University, Guangzhou 510006, China) (2 School of Mathematics & Information Science, Guangzhou University, Guangzhou 510006, China)
• Received:2012-01-06 Online:2012-09-28 Published:2012-09-28
• Contact: XIONG Jian

Abstract: Acquiring quantitative data of linguistic and mental representations is not only a techno-keystone about the correlative between language and mentality but also a difficult point of research methodology about linguistic analysis. The mathematical model which, combining natural linguistic structure, provides quantitative analysis is of very important impact for understanding the psychological information delivered by language representation. The sentential subject, including mostly noun and pronoun, is the starting point of this sentence and the foundation of logistic thinking. The linguistic representation people use is a kind of random phenomenon, in other words, people have to use natural language to symbolize thinking process and its outcome. However, in a sentence, whether the noun or the pronoun will appear first is purely random. In linguistic representation of sequential sentences, the structure of the subjects is of Markov property. So we may adopt the Markov chain model to depict the random phenomenon of sentence subject representation. We collected 124 data sets from a general subject class, a selective course taken by 124 college students finished with a written report by each student. The source of the participants was widespread, including students from sophomore to senior, while their majors involve arts, science, and engineering, as well as sports, beaus-arts, and music, etc. Such a sample set fits the requirement for this research quite well. Each sample involved 5 topic sentences and 1 Thematic Apperception Test. There were 121 valid data, the valid rate of the data reached 97.58%. Based on the Markov property definition of sentence subject representation, the research aim of this project is to calculate the transition matrix of sentence subject representation in aggregate of language representation so as to observe the implicit status of the thinking point trend a person demonstrated through explicit natural language representation. Through this research, we found that the n/r-state (noun/pronoun) of sentence subject is a regular chain, no matter what it is initially, the sentence subject representation a person spoke will eventually move to the n/r-way. Excluding 13 invalid data, the 109 valid data show that the average transfer rank of subject n/r (r/n)-state reached 2.4 times. This transformation of sentence subjects displayed the regular chain law. The result of this research indicate the following facts: (1) the sentence subject representation noun (n) / pronoun (r) was a regular chain. No matter what is the initial state, it will approximate the n/r-way. The transition law of sentence subject representation reflects the thinking process. Individuals adjust its thinking according to the situation surrounding it. The subject representation of the first sentence reflects the randomness of the subjective and objective thinking trend. A person would present the thinking point of objective or subjective through some sentences which show that people care outside as well as selfhood. (2) The explicit natural language representation is the objective clue in observing implicit psychological change. Based on this, the possible implicit psychological change can be predicted (3) As a research methodology, Mathematic Model is capable of revealing the quantity and qualitative relation between language and mind representation effectively. Further research directions are as follows: (1) The language representation model will be optimized by the language volume modeling and the signal stability modeling in order to boost the external validity of application and popularization. (2) Take the age as mediator variable, the time series model of language representation and the discrete model of group discrepancy will be established, and further exam the formation of individual community mentality. Take the year as a mediator variable in addition, it will build a corpora of natural language to go a step further to observe and study the social psychological feature of an individual.