Previous studies have relied on semantic categories in existing thesauruses, however, such a conventional method inherit shortcomings. For example, redundancy and absence of categories are observed, because categorization of general vocabulary applied in existing thesauruses does not share the specific perspectives necessary for categorization of verbs.

The present study described verb valence as the researcher constructed original semantic categories. Based on such categorical descriptions of valence, the study identified the semantic categories of NPs and constructed a set of semantic categories critical to the description of verb valence. The study also examined the statistical relationship between the numbers of verb concepts, for which valence was described, and the number of the necessary semantic categories of NPs.

Next, the ``principal-component-analysis method'' was applied to analyze the results of the investigation, 12 magazines were classified into eight style types, and then the hypothesis was examined. The stylistic characteristics of women's magazines are more ``conversational with C:/Documents\ and\ Settings/sirai/quotations -- non-logic lyricism type.'' On the other hand, it is clear that the stylistic characteristics of men's magazines are more ``textual with less quotations -- logic-narrative type''. Generally speaking, the articles in women's magazines are written in a sensitive style, whereas the ones in men's magazines can be interpreted in a logical style. Therefore, it can be proved that the first hypothesis is right.

1. Quantitative study comes first before anything.
2. Select corpus suitable for your research theme.
3. Check the design and contents of the corpus before using. ( The following recommendations are for the vocabulary investigation. )
4. Check whether the length of the counting unit of the corpus is suitable for your research theme.
5. Check whether the width of the analyzing unit of the corpus is suitable for your research theme.
6. Consider about the amount of the corpus suitable for the purpose, the period, the budget, and the labor of the research.
7. Understand the content of the processing of the analyzing program.

1. Kêsu is most frequently used as an appositive head noun modified by a clause.
2. For this usage, in expressions where the modifying clause means "a bad occurrence (which has already happened)" and the predicate of the main clause means "many" or "happens", kêsu is used much more often than synonyms rei or baai.
3. It is thought that the high frequency of this kind of expression in newspaper texts is the reason why kêsu has become a basic word in newspaper vocabulary.

In the analysis, hybrid honorific functions have been found as well as conventional single honorific functions. The analysis has also revealed that Japanese native speakers vacillate in their judgments of a variety of honorific functions.

On the other hand, the correspondence analysis has divided the honorific functions into four groups (kenjogo-sonjogo, sonkeigo, bikago-teineigo, and neutral one) and the vacillation of the honorific functions has been found in each functional group. Most honorific nouns belong to specific functional groups, whereas there are some honorific nouns which waver among them, that is, they do not follow any pattern that determines under which group they can be classified.

In this investigation, I counted the number of loanwords used in editorials from the Japanese newspapers Asahi Shimbun and Yomiuri Shimbun. I placed the data plots on a time axis, and applied non-linear regression (y = K¡¿(1¡Üa e-bx), called a ``logistic curve'').

The results of this study are as follows: The data plots agree well with the logistic curve. Thus, the logistic curve is one of the theoretical curves that express well the increase in the number of loanwords from the Taisho era to the present day.