This paper deals with the “24 hour survey of linguistic behavior” first carried out in the early 1950s. Theoretical considerations are attempted within the framework of social anthropology and social psychology. The paper discusses technical problems, and actual results are not shown. In Chapter 11 only the important results are summarized. Readers wishing to know concrete numerical results of this kind of investigation should refer to the official reports written in Japanese (Kokuritsu Kokugo Kenkyûjo 1951).
Technical developments since the time the research was carried out have been enormous. A wire recorder was used initially and the next year replaced by a tape recorder. Use of tape recorders is commonplace now and video-recorders are being used widely in several fields of sociolinguistic research, but this sort of thorough 24 hour investigation is rarely attempted. This kind of study was surely beneficial to the development of discourse analysis studies carried out later (by Fujio Minami) in Matsue city.
How many words are used by the Japanese in daily life? What kind of words are frequently used? How much time is spent reading newspapers and writing letters? These questions have never been answered. There are various methods to obtain scientific data about the actual state of language life in order to resolve various problems encountered. The 24 hour survey is the most complete and, in a sense, the most primitive method.
The 24 hour survey is a method for recording continuously a segment of language usage from beginning to end, just as it occurs, for a specific person for a whole day. Such a survey is not implausible and with nobody having tried it before, many scholars have shown a keen interest. This kind of total and continuous observational method is used in the ecological investigation of life forms, when the object does not exhibit systematic changes. When there are systematic changes, intermittent observations are sufficient, for example in the study of the flowering process of a morning-glory or the growth of a tadpole to a frog. No systematic changes in the language life over 24 hours have been recognized yet, so continuous recording must be used. These methods are referred to as “longitudinal methods” by social anthropologists and social psychologists in the USA, but even there, a continuous and all-embracing survey such as ours has not yet been carried out.
Generally there are two kinds of surveys. One is the “test method” which observes the reaction to a given stimulus. For instance, by showing an informant the picture of a radish and asking him what it is, the investigators can check whether his answer is pronounced [daikoN] or [de:koN]. The other