This paper describes morphological properties of 4 ``informal'' Japanese auxiliaries which are often used in written-Japanese. Alternative expressions of those auxiliaries for ``formal'' use are also mentioned. It becomes possible with using such morphological properties to analyze several kinds of ``informal'' Japanese expressions, and to increase accuracy of analyzing written-Japanese sentences which includes such expressions.

A questionnaire was administered to 47 Japanese university students. Results and comments are summarized as follows:
1) In face-to-face situations, the confrontational type only appeared when the speaker referred to a small object, with both confrontational and fusion types appearing when a large object was the referent. We therefore hypothesized that the fusion type will only appears when the apeaker refers to a larger object.
2) It was also found that in face - to - face situations, the speaker in effect 'drew his hearer to him' psychologically when the state was transferred from the confrontational type to the fusion type.
3) Only the fusion type appeared in side - by - side situations, regardless of whether the referent was large or small. We also found that KO, SO and A divided the territory into three equal parts in such situations, with little attention actually being paid to the hearer.

Some adjectives lack adverbial usage, some others are used only as adnominal. ``IPAL Dictionary of Basic Japanese Adjectives for Computer'' describes these usages in detail. Those adjectives which have three usages according to this dictionary differ in the frequency of each usage. We counted the usages of adjectives in the data of 90 magazines used in the survey by the National Language Research Institute and found these tendencies.

Comparing these two kinds of sentences from statistical points of view, no statistically significant difference at the 5% level could be found in the usage rate of kanji, mean sentence length, number of sentences per entry, or distribution of parts of speech.

The purpose of this paper is to reveal the structures of query languagesand how they are used. I assume a database to be a collection of a data which is composed of pairs of an attribute and its value. Based on this view, five ways to retrieve a piece of data using these elements and Japanese ``basic query unit'' expressions which correspond to these ways are determined. I assume a query sentence to be composed of a predicate and a basic queryunit which is equivalent to a subject phrase, or its paraphrase. Querysentences are also classified into eight classes based on their predicates.Expressions into which basic query units are syntactically embedded are explained based on the basic query units used and the structures these have. Query expressions which are used depending on whether a value is specifiedor not specified are also considered.

        [Characteristic Semantic Areas]
Japanese: Relations; Time; Unit; Countries and Cities; Languages; Art; Living; 
German: Place;
French: Positions; Functions; Clothes;
Spanish: Unit; Race; Countries and Cities;

Three structuralized 4-byte code methods are given. One is a structure composedof a 3-byte integer that is the Daikanwa Jiten index number, base-16 expressed by base-94, with a l-byte decimal point section. This can accommodate codes for data IO that include accesses to the Daikanwa Jiten. Next there is a structure that combines the existing 2-byte code, this can accommodate Kanji code for the preservation of data. The third method converts to 4-bytes using codes where the two 8th bits of each digit are '01', this can accommodate the computers internal code.

Among these three elements, it is found that the statistics of the characters immediate befor commas is most useful to characterize or distinguish authors.

In this paper stochastic characteristics of morpheme chain generated by typographical error is discussed. We also propose a typographical error detection rule based on the morpheme chain probability of both correct and erroneous sentences.

We use 8,000 spoken sentences taken from 86 dialogs. There are 583 occurrences of particle ellipsis whose headword modifies a single verb. 455 modify the verb immediately follows the case element, while 128 of them have other elements in - between.

If the case element modifies the immediately following verb, the omitted particle is `wo' for the most of the time. More specifically, there are 214 examples of omission of `wo', 156 for `ga', 80 for `ni', 4 for `de', and 1 for `to'.

When the valency pattern of the modified verb contains the 'wo' case, and when the sentence does not have an element explicitly signaled as the case `wo', it is most probable that the missing case is `wo'. Next come cases that represent the destination of action, such as `ni' and `he' (we can confirm this when the head noun has the semantic feature `locative').

Other than these, `ga' or `ni', whichever closer the last element in the valency pattern of the modified verb, is most likely the missing case.

When the case element is distant from the predicate verb, the omitted particle is 'ga' for the most of the time. `ga' accounts for 60 examples, `wo' is 40, `ni' is 13, `de' is 3, and others are 12. `Others' include cases that do not appear in valency patterns. When a noun is not strongly related to a predicate, it has more topical role and it is difficult to determine its case.

The Characteristics of the theme, in contrast with Japanese popular songs are:

        There are many songs which describe nature.
        There are songs to educate the people.
        There are many theme songs of movies and television plays.

        Personal pronouns were used much more often. It is due to the
lack of ellipsis and many personifications.
        ``Love'' between men and wemen is rarely portrayed.
        There are many words which refer to nature.
        There are many more words which refer spring than to other
        seasons.