Statistical Analysis Of Lexical Density In Corpora

Lexical density is an important aspect in language teaching and lexical acquisition. This thesis aims to discover the regularity of lexical density in LOB, BROWN A and T4, which represent news style and technical English style respectively. Different from traditional methods, the research is conducted mainly on the basis of statistical theories and by means of statistical methods. The use of quite a few complicated computer programs facilitates the study. By drawing samples from each corpus and making comparison and contrast among them, the lexical density is calculated in a certain length of discourse Quantitative and contrastive analyses uncover the distribution of the number of new lexicons That is: by simple comparison, the number of new lexicons calculated in a certain length of texts is normally distributed, whereas by cumulative comparison, the number of new lexicons displays a descending tendency.This finding is of particular importance in language teaching practice Based on the normal distribution of lexical density, the teacher can make reliable assumption of the rough number of new words in a certain textbook so as to decide whether it is suitable for his teaching. Additionally, though the lexical density does not vary greatly in a certain length of texts, it decreases when a learner's vocabulary size enlarges. If the same level of lexical input is to be maintained, the size of the texts should be increased so as to keep the same probability of encountering new word. This would provide the teacher and the textbook compiler with useful information.