It is common practice to take surveys of the frequencies of kanji (Chinese characters) as the basis for selecting which kanji are to be introduced in textbooks for the teaching of Japanese as a foreign language, and in what order they are to be introduced. In this study we have sought to aid this process by developing an efficient system for surveying kanji usage in full-text newspaper article data bases. Using this system, we have conducted a large-scale survey on the texts of approximately 110,000 articles, covering a full year of the Asahi Shinbun, both morning and evening editions, producing a table of the frequencies of use of 4,476 kanji, 83 hiragana and 86 katakana (phonetic characters). Some of the results show that : (1) The 1,000 most frequently used kanji account for about 95% of the total kanji usage ; (2) The 1,600 most frequently used kanji account for about 99% of the total, while the nearly 3,000 remaining kanji only make up 1% of the total use ; and (3) There is a very high correlation in relative frequency of use between newspaper articles in 1966 and 1993. These results give an accurate picture of the actual kanji use in newspapers, and should be useful fundamental data for the teaching of Japanese as a foreign language, as well as for the teaching of kanji to Japanese children and in the field of natural language processing.

View full abstract

The purpose of this paper was to analyze eye movements of Japanese subjects viewing subtitled American movies. Condensed versions of four scenes totaling five minutes from two subtitled movies, "JFK" and "Malcom X", were presented to subjects who wore eye tracking instruments. Recorded eye movements were classified into four processes : (1) jump to the subtitled area, (2) jump to the head of the subtitle, (3) reading of the text and (4) jump to the head of the second line (for two-lined subtitles). Analysis of the movements during presentation of four kinds of subtitles (six characters in one line, 13 characters in one line, 13 characters in two lines, and 20 characters in two lines) determined that reading rate was lowered by two factors ; the number of lines and the number of characters. Results attest to importance in the design of simultaneous presentations of text and moving pictures.

View full abstract

In the near future, with advancements in the multimedia age, it is expected that an information-originating type of university campus will regularly broadcast its own televised programs, for instance 30 minutes every day. The objective of this paper was to devise a guideline for organizing 30 minute televised programs. The duration and programming organization of the NHK televised news program "Good Morning Japan" (Ohayo Nippon), was analyzed for 14 days for this purpose. The first 90 minutes of the 133 minute program is divided into three section of equal length. Analysis of each section showed the following findings in both structuring of duration distribution and programming organization. 1) Almost all newscast items were completed within two minutes and ten seconds. 2) The first newscast items (important newscasts) tended to be two-part separated broadcasts, the first part being a survey of relatively short length and the second being the main body with a longer time length shown later. 3) Newscast items were divided into two categories : regular and non-regular type programming. 4) Average total number and average total time of items were 29 and 71 seconds for non-regular programming, and 34 and 95 seconds for regular programming, respectively. 5) Both groups of regular and non-regular programming were alternately broadcast. Programming organization of both duration and the order of broadcasted items were greatly dependent upon the pattern of daily morning lives of televiewers. 6) Comparison of the duration of newscast items in the three sections showed that the 6 : 30-7 : 00 a. m. one had two typical trends : (1) relatively fewer items or relatively longer time lengths, and (2) relatively more weight for regular, as opposed to non-regular, programming. 7) An "advant", or brief overview of the TV program, is placed at the start of every thirty minute section. Average duration of the advant was two minutes and eight seconds in the 6 : 00- 6 : 30 section, two minutes and forty-eight seconds in the 6 : 30-7 : 00 section, and one minute and forty seconds in the 7 : 00-7 : 30 section.

View full abstract