Is the use of kanji increasing in the Japanese writing system?

Dilhara Darshana Premaratne, College of Arts and Social Sciences, Australian National University [About | Email]

Volume 12, Issue 3 (Article 3 in 2013). First published in ejcjs on 17 February 2012.

Abstract

It is commonly believed that changes have been taking place in the Japanese writing system in recent years, with more kanji being used than kana when compared with the years before. If this is so, this could mean that words which were written in kana in the past are now being written in kanji. In other words, what this means is that the manner in which words are represented in the Japanese language is itself changing. This paper is based on a longitudinal study carried out to examine the changes in the Japanese writing system in relation to the use of kanji and kana. The study aimed to find out whether higher proportions of kanji have been used after the early 1980s, and if so, whether this increased use of kanji has had an impact on how words are represented in the language. The study examined the changes in script usage by using topic-controlled news texts published in the Japanese newspaper Asahi Shimbun from the 1940s to the 2000s.

Keywords: Japanese writing system, information and communication technology, kanji, kana, kango, wago.

Introduction

It is commonly believed that changes have been taking place in the Japanese writing system in recent years (Agency for Cultural Affairs 2009). The changes are particularly thought to be in relation to the use of kanji and kana, more specifically, in relation to the proportions of kanji and kana used in texts. It is generally supposed that more kanji have been used than kana in the past few decades when compared with the period before (Gottlieb 2000). If this is correct, this could mean that words which were written in kana in the past are now being written in kanji. In other words, what this means is that the manner in which words are represented in the Japanese language is itself changing.

It is possible that the spread of information and communication technology, particularly the widespread use of word-processors that came about in the early 1980s, contributed to changes in the use of kanji. The year 1984 can be taken as the beginning of the personal word processor age in Japan (Gottlieb 1998) as it marked an explosion in word-processor sales (Seeley 1990). One important change that has taken place due to the use of information and communication technology is the opportunity that the Japanese pubic now have to see more kanji in use than before (Agency for Cultural Affairs 2005). Another important change is the opportunity that people now have to use more kanji in their writing. This is because the word-processor has revolutionised Japanese writing by facilitating the process of converting kana to kanji. To produce a document using the word processor, words need to be input just in kana using the keyboard, which can then be converted to kanji merely by pressing the space bar. When more than one kanji come up on the computer screen, the user only needs to choose the one appropriate for the specific context. Consequently, the word processor has given people the opportunity to use kanji they have not committed to memory. In contrast, before the advent of the word processor, people could only use kanji that they were familiar with in their writing, unless they looked them up in a dictionary.

It is generally assumed that the use of the word-processor dramatically increased the use of kanji (Gottlieb 2000). Firstly, in addition to facilitating the production of kanji, the word-processor has the ability to produce a large number of kanji, far greater than what is required to produce ordinary texts. For example, the word-processor can print out all the kanji identified in the Japan Industrial Standards (JIS). The JIS list consists of two sets of characters—2965 characters in Level One and 3384 characters in Level Two. Although Level One contains basic characters, Level Two contains relatively uncommon kanji (Seeley 1990; Seeley 1994). Secondly, as kanji can be produced easily on the word processor (Gottlieb 2000) it helps to materialise the belief ingrained in people during school years that using kanji whenever appropriate is the correct way to write in Japanese (Umesao 1988). Finally, laziness on the part of users has also increased the use of kanji by making them accept as their first choice the kanji offered by the word-processor for native Japanese words usually written in kana (Tanaka cited in Gottlieb 2000). Tanaka observes that this could increase the proportion of kanji in a text to between 60 and 70 per cent.

In addition to the impact of information and communication technology, it is possible that the revision of the 1946 script policy has also contributed to an increase in the use of kanji. The revision which took place in 1981 is seen as a reversal of the former script policy (Gottlieb 1994). The Jōyō List which was introduced as part of the revision replaced the Tōyō List introduced in 1946. The Jōyō List not only had more characters than the previous list, but was also more liberal than the previous list.

Research on Script Usage in Japanese

Textual surveys help to investigate the changes that have taken place in script usage, because they provide information about the actual use of a script. Longitudinal studies are particularly valuable in this respect as they also provide information about trends in script usage. For example, they can indicate the time a new trend emerges or the time an existing one ends. In contrast, cross-sectional studies focus on a single point in time. They provide useful information about script usage by making comparisons across different texts written at a various times. However, they are unable to confirm whether any observed features in script usage existed in the past or would continue in the future. Although textual surveys provide useful information about script usage, not many studies investigated script usage by this method (Gottlieb 2000). Of these, only a very few used the longitudinal method of investigation.

Studies that Examined Script Proportions

Some of the textual surveys conducted to determine script usage in texts focused only on script proportions. Of these, some examined the proportions of kanji used both in terms of Jōyō kanji and Non-Jōyō kanji, while some examined the proportions of all the scripts used. There are both short-term studies as well as longitudinal studies in this group.

Nomura (1980) examined the proportions of kanji used in twenty-seven magazines published for the general reading public in 1979. He divided the magazines into six categories based on subject matter, such as politics and economy, news magazines, and men’s and women’s magazines. He examined one hundred sentences per magazine from each of the six identified areas. He found that magazines within the same category had a similar proportion of kanji. Magazines in the area of politics and economics had the highest proportion of kanji while women’s and men’s magazines had the lowest proportions of kanji. This study shows that the proportions of kanji used in a text could vary according to the type of the text.

Ogura and Aizawa (2007) did a similar study on contemporary writing by analysing the proportion of Jōyō List kanji and Non-Jōyō List kanji used in magazines. This study is based on the data of an earlier study, the Linguistic Survey of Contemporary magazines conducted in 1994, that examined the patterns of kanji use in seventy contemporary magazines. Ogura and Aizawa analysed the usage frequency of Jōyō and Non-Jōyō kanji observed in the two thousand most frequent kanji identified in the previous study. They found that Jōyō kanji mostly fulfilled their function as the standard of kanji use in 1994, as only 1.9% of the kanji found in the most frequently occurring 2,000 kanji were Non-Jōyō kanji.

Although short-term studies reveal the proportions of kanji used in texts, they do not reveal the changes in the proportions of kanji used over time as they are conducted on a short term basis. Longitudinal studies are valuable in this respect as they shed light on trends in kanji use that take place in different periods of time.

Yasumoto (1963) is a longitudinal study that examined the use of kanji in novels published between 1900 and 1954. This was done by examining 1,000 characters each from a total of 100 novels written by different authors. The study found that the proportion of kanji used in the novels progressively declined during the period studied.

Miyajima (1988) is another longitudinal study that aimed to determine the change in kanji proportions in novels. The study examined kanji proportions in 100 novels at five year intervals from 1935 to 1985, examining 1,000 characters from each five year period under survey. The findings revealed that the proportion of kanji used in the novels declined between 1936 and 1960. However, Miyajima noted that kanji use did not decline after this point.

Miyajima (1989) is also a longitudinal study that aimed to determine the proportions of kanji used in a magazine between 1906 and 1976. This was done by examining samples taken from the monthly magazine, Chuo Koron, at 10 year intervals. The study found that kanji proportions in the texts declined up to 1956 and slightly increased from then onwards.

Tomoda (2005) examined the use of kanji in white papers between 1960 and 1997. This was done by selecting four white papers written about labour and examining the section written about the living conditions of workers in each white paper. It was found that the proportion of kanji increased between 1960 and 1997, but that the increase was relatively small. Tomoda noted that this trend was in accordance with the Chuo Koron study that pointed to an increase in kanji use after the 1950s. Tomoda also focused on the proportions of kana and other script types used in the examined texts.

Although the longitudinal studies discussed above provide useful information about trends in kanji use observed in the past, these studies are few in number. Also, the studies cover different periods of time and therefore, cannot be used together to draw valid conclusions about trends in kanji use over a continuous period of time. In addition, apart from Tomoda (2005), these previous studies examined kanji use only up to the early 1980s. As such, they do not provide any information about trends in kanji use in the late 1980s and after. Therefore, studies conducted on the same text type that cover the period before as well as after the early 1980s are necessary to determine the changes in script usage that are happening at present.

Studies that Examined Script Choice to Represent Words

A study conducted by the National Language Research Institute (NLR) in 1980 analysed the top 7000 words appearing in ninety contemporary magazines. With respect to nouns it was found that a higher proportion of kango was used to refer to abstract nouns, nouns related to products and appliances, and nouns related to daily activities. A higher proportion of wago was used to refer to objects found in nature and natural phenomena. With respect to verbs it was found that a higher proportion of verbs were wago verbs. This study shows that there is a variation in the way kango and wago are chosen for different parts of speech in a text, as well as for different types of words within the same part of speech (Miyajima ed. 1982).

Nomura and Yanase (1989) surveyed vocabulary in books for juveniles to find out the peculiarities of juvenile vocabulary. In terms of the type of words that were used, they found that the percentage of wago was higher than the percentage of kango. It was also found that the proportion of wago used was considerably higher than the proportion of kango, 64% and 30% respectively. They also reported that this was in sharp contrast to the wago and kango proportions observed in other research, 35% and 38% respectively in newspapers, and 44% and 47% respectively in magazines. Although the disparity between wago and kango was less in these studies, wago was used to a lesser extent than in the books for juveniles. This study shows that the proportions of kango and wago used in a text could vary on the basis of text type and intended audience.

Nomura (1998) confirmed the close relationship between the proportions of kango and wago in a text and the type of text and intended audience that was observed by Nomura and Yanase. Nomura examined words in Japanese as a second language text books and found that there were 74.9% nouns, 16.4% verbs, 2.3% adjectives and 0.9% adverbs. An examination of the distribution of kango and wago revealed that there was a higher proportion of kango than wago in the examined texts, 55.8% and 30.8% respectively. When compared with magazines it was found that kango was also higher than wago in magazines, 47.5% and 37.7% respectively. However, the proportion of kango in magazines was lower and the proportion of wago was higher in comparison to Nomura’s current study. When compared with children’s books it was found that wago was higher than kango in children’s books. It was also found that wago was again higher than kanji in spoken discourse, more specifically, in conversations of intellecutuals and on TV.

Nomura (1999) confirmed the findings in the NLRI study of 1980 in relation to the use of kango for certain types of nouns and verbs, depending on what the words refer to. A database of 22,000 contemporary kango created by Nomura was used in the study. Of these, the 3000 most frequent kango were examined. With respect to verbs it was found that the majority referred to actions/movements. With respect to nouns it was found that most of the selected words referred to abstract concepts, while only a few words referred to specific things.

Miyajima (1988, 1989) and Tomoda (2005) reported on the choice of kanji and kana to represent kango and wago, in addition to other word-related aspects. Miyajima (1988) did a longitudinal study of the distribution of kanji and kana in novels. The period examined was from 1935 to 1985. It was found that while a very high proportion of kango was written in kanji, a similarly high proportion of wago was written in hiragana.

Miyajima (1989) did a longitudinal study of the distribution of kanji and kana in magazines. It was found that only about one third of the words in the magazines was kango, and that words with three or more characters were increasing at the expense of one-character and two-character words. It was also found that there was no fixed rule in relation to the choice of kana and kanji in representing wago. In addition, it was found that a word may have several different written forms depending on how okurigana related to the word is represented. However, Miyajima found that the number of okurigana was becoming smaller, and that the Japanese orthography was stabilising in this respect.

In her longitudinal study of white papers, Tomoda (2005) found some policy related changes as well as some stylistic changes in script usage that contributed to an increase in kanji. For example, official changes to okurigana usage, and to the use of hiragana for adverbs reduced the number of hiragana and favoured an increase in kanji across the examined samples. Similarly, stylistic changes in the form of a shift towards greater kango use also contributed to the relative increase in kanji and indicated a move away from the earlier policy of using a plain spoken style.

Most studies examined the proportions of kango and wago in different types of texts and in different types of words in the same text type. However, only a few studies went on to examine which scripts were used to represent kango and wago in the texts. This kind of investigation is necessary to find out textual factors that determine the proportions of kanji and kana used to represent the two types of words. Studies that examine this on a longitudinal basis are particularly useful as they not only show what the current practice is in choosing kanji and kana to represent kango and wago, but also whether the practice has remained stable over the years. Although there are a few longitudinal studies that looked at this aspect, no single study examined it from the period before the first script reform right up to present times.

The present study therefore set out to find out on a longitudinal basis how kanji proportions in texts are influenced by the choices made by individual writers between kango and wago and also between kanji and hiragana to represent kango and wago. The same newspaper samples that were used to find out kanji proportions from 1942 to 2009 were used again for this second investigation. Although examining all words used in the samples would have produced the best possible outcome, the examination was confined to verbs only, for practical purposes.

Purpose of the study

The purpose of the current study was twofold:

Firstly, the study aimed to find out whether higher proportions of kanji have been used in texts after the early 1980s when compared with the period before.

Secondly, the study aimed to find out whether the practice of choosing kanji and kana to represent verbs has changed over the years.

In order to accomplish the above purposes, the study examined a series of newspaper samples extracted from various issues of the same newspaper from the 1940s to the 2000s. Newspapers were used in the study for a number of reasons.

Newspapers are an excellent medium for the longitudinal comparison of texts. This is because texts written over a long period of time are available in newspapers. This allows the examination and comparison of text samples across different periods of time. Also, texts written on the same topic are available for study in newspapers over many years. This is very important for longitudinal studies. Controlling the topic or subject matter of texts is particularly important when comparing kanji use across different periods of time, because the topic can affect the proportion of kanji used in a text. In addition, newspapers are a good medium to test the changes in script usage as news texts are written by a wide range of writers.

Method

The study is based on news texts extracted from the Japanese newspaper Asahi Shimbun. The newspapers selected for the study covered an extended period of time from the 1940s to the 2000s. News texts were extracted from a single newspaper source instead of from a random collection of newspapers to control any impact of stylistic variations on character numbers.

The topic of the texts was controlled by selecting only news texts written about the single topic of the Japanese script reform. This topic was chosen as it was assumed that a considerable number of news texts written about script reform would be available as it has been a popular topic in newspapers over the years.

A total of 17 news texts written about the script reform were extracted from the Asahi Shimbun newspaper, to cover each decade from the 1940s to the 2000s. A random sample of 30% of the lines was chosen from each selected text, with a minimum of 10 lines chosen from shorter texts. Although examining the entire text would have been the ideal strategy, it was not practical when considering the number and the length of news texts investigated. Random sampling of lines was preferred to extracting lines from a continuous chunk of the text, because when random sampling is used, each line has an equal chance of inclusion in the sample. This was considered to be important as the number of kanji could be influenced by the organisation of the text. For example, a higher number of kanji could occur in the introduction of the text than in the body or conclusion of the text, and vice versa.

Kanji, kana and other types of script in the selected lines were extracted from each sample. Punctuation marks and other symbols were not extracted.

Results and Discussion

The purpose of the present study was to find out whether higher proportions of kanji have been used in news texts after the early 1980s, and if so, whether the practice of using kanji and kana to represent verbs has also changed. This was done in two steps. First, the study examined the proportions of kanji used in news texts from the 1940s to the 2000s. Then the study examined the practice of using kanji and kana to represent verbs.

Proportions of Kanji from the 1940s to the 2000s

This section looks at the proportions of kanji used during the period covered by the current study. The purpose of this analysis is to find out whether kanji use increased in the period after 1980. To this end, the proportion of kanji used in each news text was calculated and expressed as a percentage. This information is displayed in Figure 1.

Figure 1. Proportions of Kanji from 1942 to 2009

Premaratne, Figure 1

First kanji proportions in the period from 1984 onwards were examined. The examination revealed that high and low kanji proportions are mixed. For example, while the two highest kanji proportions of 60% and 61% were observed during this period, a comparatively low kanji proportion of 40% was also observed in 1993. In addition, two moderately high kanji proportions were also observed during this period, 44% in 1998 and 45% in 2001.

The data showed not only a considerable variation in kanji proportions in the period from 1984 onwards, but also a decline in kanji use across this period. The following kanji proportions were observed between 1984 and 2009: 1984- 60.6%, 1988- 59.7%, 1993- 40.3%, 1998- 44.3%, 2001- 45%, 2009- 42.11%. Although kanji proportions in the 1980s were found to be high, the comparatively low kanji proportions observed during the rest of the period suggest that kanji use has declined after the 1980s. Therefore, these results did not confirm the common belief that kanji use has been increasing since the early 1980s.

Next, kanji proportions in the period before 1984 were examined. This examination also revealed a mix of high and low kanji proportions similar to that observed in the post-1980 period. For example, while the two lowest kanji proportions of 28% and 21% were observed during this period, high kanji proportions of 50% and above were also observed in five of the eleven news texts that belong to this particular period.

Finally, kanji proportions within each decade were examined for each of the seven decades. This examination revealed that news texts with high and low kanji proportions occurred simultaneously in most of the decades surveyed: the 1950s when a news text with 53% kanji occurred with two news texts with 28% and 31% kanji; the 1960s when a news text with 36% kanji occurred with two news texts with 46% and 56% kanji; the 1970s when a news text with 55% kanji occurred with 37% and 38% kanji are examples. There was less variation within the 1980s, the 1990s and the 2000s. However, the variation in kanji proportions between the 1980s and the 1990s as well as between the 1990s and the 2000s is very pronounced.

One of the aims of the present study was to find out whether higher proportions of kanji were used from the 1980 onwards when compared with the period before. Results showed that there was great variation in the occurrence of high and low kanji proportions in this period as well as in the period before. In addition, kanji use declined in the 1990s and the 2000s when compared with the 1980s. Therefore, it was not possible to conclude that the use of kanji increased from 1980 onwards.

Practice of using Kanji and Kana to Represent Verbs

This section looks at the proportions of all verbs (kango and wago verbs) represented in kanji, the proportions of kango and wago verbs, and the proportions of kango and wago verbs represented in kanji and kana. The purpose of these analyses is to find out whether verbs represented in kanji increased in the post-1980 period; whether kango verbs which are usually represented in kanji increased in the post-1980 period; and whether there is a close relationship between the proportion of kango verbs and wago verbs in the texts and the proportion of verbs represented in kanji in the texts.

Proportions of All Verbs Represented in Kanji

The proportions of all verbs represented in kanji were first calculated and expressed as percentages for each sample. This information is presented in Figure 2.

Figure 2. Proportions of Verbs Represented in Kanji in Each Sample

Premaratne, Figure 2

Figure 2 shows that samples with a high proportion of verbs represented in kanji and samples with a lower proportion of verbs represented in kanji are mixed across the period surveyed. Samples were divided into three categories as high, medium and low, according to the proportion of verbs represented in kanji. There are eight samples in which the proportion of verbs represented in kanji is 60% or higher. There are six samples in which the proportion of verbs represented in kanji is between 45% and 59%. There are three samples in which the proportion of verbs represented in kanji is less than 45%.

Figure 2 also shows that samples in the first category occur in each of the seven decades surveyed, while samples in the second category occur in six of the seven decades surveyed. This means that samples with a high and medium proportion of verbs represented in kanji occur simultaneously throughout the period surveyed.

Also, samples with a high proportion of verbs represented in kanji occur in the same decade with samples with a medium or low proportion of verbs represented in kanji. The variation observed within the same decade is most pronounced in the 1940s, 1950s, 1960s, 1980s and the 2000s. This can be considered as unusually high for samples taken from texts written on the same topic and in the same text type.

It is interesting to note that the variation in verbs represented in kanji mirrors the variation in the total proportions of kanji in the samples observed earlier (Figure 1).

Proportions of Kango and Wago Verbs

Analyses were then performed to find a possible explanation for the variation across samples that was observed in terms of high and low proportions of kanji used in the verbs. The literature showed that the proportion of kanji used in a text could be high when the proportion of kango used in the text is high. This is because kango verbs are generally represented in kanji (Miyajima 1988; Tomoda 2005).

First, proportions of kango and wago verbs were calculated for the period surveyed. The results of the analysis are given in Table 1 and Table 2.

Table 1: Average Proportions of Kango and Wago Verbs
Kango Verbs (%) Wago Verbs (%)
16 84
Table 2: Proportions of Kango and Wago Verbs in Each Sample
Year Proportion of Kango Verbs (%) Proportion of Wago Verbs (%)
1942 41 59
1946 36 64
1951 0 100
1956 6 94
1958 8 92
1962 15 85
1965 18 82
1968 44 56
1971 16 84
1974 20 80
1979 19 81
1984 35 65
1988 33 67
1993 14 86
1998 15 85
2001 10 90
2009 19 81

Table 1 shows that the average proportion of wago verbs in the samples was higher than the average proportion of kango verbs. Table 2 shows that the proportion of kango verbs used in each sample was lower than the proportion of wago verbs used in each sample.

Second, the average proportion of kanji used in all the verbs was calculated for the sample as a whole. This was found to be 57% which can be considered as fairly high. However, the average proportion of kango verbs used in the samples was very low at 16%. Therefore, it can be considered that in the samples studied, it was not the proportion of kango verbs that contributed to the high proportion of kanji used in the verbs as a whole.

The results of the analyses also showed that in terms of the proportion of wago verbs used in the samples, the current study is closer to other types of texts than newspapers. For example, the NLRI study on magazines conducted in 1980 found that wago verbs occurred in a higher proportion than kango verbs. Miyajima (1989) reported on a similar finding in relation to all words in magazines. This study found that only one third of the words were kango. However, some past research found that kango was higher in both magazines and newspapers (Nomura and Yanase 1989; Nomura 1998). What past research seemed to agree on is that wago is higher in children’s books and in books for juveniles (Nomura and Yanase 1989; Nomura 1998). Nomura and Yanase found that the proportion of wago used in juvenile books was as high as 64% while the proportion of kango used was much lower at 30%. When compared with juvenile books, the average proportion of wago found in the verbs in the present study is much higher at 84% and the average proportion of kango found in the verbs is much lower at 16%. Another area where past research found more wago than kango is spoken discourse among educated adults and on TV (Nomura 1998). Comparison of the findings in the present study with that in past research seems to suggest that in terms of kango and wago proportions found in verbs, the present study is similar to some magazines, children’s books, books for juveniles and spoken discourse among intellectuals.

Third, the proportions of kanji and kana used to represent kango and wago verbs were calculated. The results are displayed in Table 3.

Table 3: Average Proportions of Kango and Wago Verbs Represented in Kanji and Kana
Kango in Kanji Kango in Hiragana Wago in Kanji Wago in Hiragana
100% 0 39% 61%

Table 3 shows that while all kango verbs are represented in kanji, some wago verbs are represented in kana and some in kanji. Table 3 also shows that a higher proportion of wago verbs are represented in kana.

Fourth, an analysis was performed to find out whether it was the script chosen to represent wago verbs that resulted in the high proportion of verbs represented in kanji in the texts. The results of this analysis are presented in Table 4.

Table 4: Proportions of Wago Verbs Represented in Kanji and Kana in Each Sample
Year Proportion of Wago Verbs Represented in Kanji (%) Proportion of Wago Verbs Represented in Hiragana (%)
1942 50 50
1946 22 78
1951 70 30
1956 6 94
1958 4 96
1962 41 59
1965 21 79
1968 40 60
1971 39 61
1974 50 50
1979 50 50
1984 69 31
1988 100 0
1993 38 62
1998 53 47
2001 65 35
2009 40 60

Table 4 shows that more writers in the examined samples used kana to represent a higher proportion of wago verbs. More specifically, the proportion of wago verbs represented in kana is higher than 50% in nine of the seventeen sample studied. However, a considerable number of writers also used kanji to represent a higher proportion of wago verbs. More specifically, the proportion of wago verbs represented in kanji was higher than 50% in five of the seventeen samples, with one sample representing all wago verbs in kanji. Also, in three of the samples, an equal proportion of wago verbs (50%) was represented in kanji and kana.

The above observations suggest that the script chosen to represent wago verbs in a text makes a significant contribution to the proportion of kanji used to represent verbs in that text.

Finally, samples that have a high proportion of wago verbs represented in kanji and samples that have a low proportion of wago verbs represented in kanji were examined carefully to see whether they occur at particular periods in time. The result of this analysis is displayed in Figure 3.

Figure 3. Proportions of Wago Verbs Represented in Kanji

Premaratne, Figure 3

Figure 3 shows that although a fairly high number of writers chose to represent a higher proportion of wago verbs in kanji, samples that have both high and low proportions of wago verbs represented in kanji are scattered throughout the surveyed period. The eight news texts which have 50% or a higher percentage of wago verbs represented in kanji occur in each of the seven decades, except in the 1960s. Similarly, the nine news texts which have less than 50% of wago verbs represented in kanji also occur in every decade, except in the 1980s.

In addition, the data show that samples with higher and lower proportions of wago verbs represented in kanji are found within the same decade. This variation within decades is pronounced in each decade in the surveyed period, particularly so in the 1940s, 1950s, 1960s, 1980s and the 2000s. For example, one of the two samples in the 1940s has 50% of wago verbs represented in kanji while the other has only 22%; one of the three samples in the 1950s has 70% of wago verbs represented in kanji while the other two have only 4% and 6%; one of the two samples in the 1980s has 100% of wago verbs represented in kanji while the other has only 69%; and one of the two samples in the 2000s has 65% of wago verbs represented in kanji while the other has only 40%. It is interesting to note that the same five decades were found to have a great variation across samples in terms of high and low proportions of all verbs represented in kanji. The fact that variation within decades is common to both the pre-1984 and post-1984 period suggests that news text written in the latter period did not have a consistently high proportion of wago verbs represented in kanji.

The findings of the present study are based on samples extracted from news texts written on the topic of script reform. Therefore, further longitudinal research needs to be done on different topics in different media in order to determine changes in kanji proportions that have taken place over the years.

The findings of the present study in relation to the representation of kango and wago words are based on an examination of the verbs in the chosen samples. Therefore, further research needs to be done on other parts of speech as well in order to fully comprehend the issues related to the representation of kango and wago words.

Summary of Results and Discussion

The present study found that general kanji use fluctuated during the period surveyed. High and low kanji proportions occurred though out the period, often occurring within the same decade. Therefore, it was not possible to conclude that higher proportions of kanji have been used consistently in the period after 1980.

It was also found that the proportion of wago verbs was higher than that of kango verbs across the period surveyed. This was an important finding as past research found that kango occurred in a higher proportion than wago in newspapers. However, as past research also found that wago was higher than kango in children’s books, books for juveniles, some magazines, and spoken discourse among intellectuals, the present study suggests that at least for the topic of script reform, the examined samples are closer to these text types than newspapers in this respect. Tomoda (2005) noted a shift away from the use of wago in the texts she examined which she attributed to a shift away from the plain spoken style. Therefore, it can be said that the greater tendency to use wago in the present study suggests a preference for the plain spoken style in writing. These findings however, are based on an examination only of the verbs that occurred in the samples.

Another interesting finding was that there was no indication that more kango was used after 1980. This is because high and low proportions of kango verbs occurred in both the pre-1980 and post-1980 period. This means that writers of the examined samples did not show a greater preference for kango after 1980 than before.

The present study found that all kango verbs were represented in kanji. This confirms the finding in past research that the practice of representing kango is well-established. This could means that a text that has a large proportion of kango will also have a large proportion of kanji in the text. However, the study showed that when the average proportion of kango verbs in the sample was low, the proportion of kanji used in verbs was still relatively high. This suggests that the proportion of kango verbs in a text is not the only contributing factor to the proportion of kanji verbs in that text.

The present study also found that unlike with kango verbs, no established practice could be seen in relation to the representation of wago verbs. While some wago verbs were represented in kana, some were represented in kanji. Although a higher proportion of wago verbs was represented in kana than in kanji in the majority of samples, a fairly high proportion of wago verbs was represented in kanji. This suggests that the proportion of wago verbs represented in kanji in a text can be a contributing factor to the proportion of all verbs represented in kanji in that text.

Conclusion

The use of kanji in the examined samples did not show a great increase in the post-1980 period because a mix of high and low proportions of kanji use was observed in both periods, and very often even within the same decade. The use of kanji in verbs also showed a similar pattern.

With respect to the representation of kango and wago verbs, the study showed that while all kango verbs were represented in kanji, the representation of wago verbs was not rule-governed or custom-bound. This is because some writers used more kanji to represent wago verbs while some used more kana to represent wago verbs. The proportion of kango verbs used in texts did not increase after 1984, so the use of kango verbs did not result in an increase in the number of kanji used in verbs during this period.

There was much variation shown in the representation of wago verbs. Although more writers represented a higher proportion of wago verbs in kana in the examined samples, a large number of writers also represented a higher proportion of wago verbs in kanji. Some writers represented an equal proportion of wago verbs in kanji and kana. These findings show that even in a text that has a small proportion of kango verbs, the total proportion of kanji used in verbs could still be fairly high.

The variation observed in the representation of wago verbs raises questions. Past research suggests that the variation is due to a stylistic difference, because more wago words are represented in kanji in more formal writing such as newspapers, while less wago words are represented in knaji in children’s books and books for juveniles.

However, the present study suggests that the variation could be due to an individual preference of the writer. This is because in the present study, the variation was observed in texts published within the same newspaper and written on the single topic of script reform.

Similar to the proportion of kanji used in all words in the texts, the proportion of kanji used to represent wago verbs remained fairly stable over the years, except on rare occasions when it soared very high or dipped very low. More importantly, the increased and decreased proportions were common to both the pre-1980 and post-1980 periods. This was in spite of the fact that the representation of wago is not bound by custom which allowed writers to use more kanji to represent wago verbs in the post-1980 period if they so wished.

The present study found that there was great variation in the script used to represent wago verbs and suggested that the variation was due to the individual preference of the writer. However, this situation could change if the representation of wago becomes more standardised.

Standardisation of this nature could occur due to various reasons such as the increased use of information and communication technology, changes in the modes of communication, new education policies and so on. For example, Japanese people could start writing more wago words or all wago words in kana when text-messaging, as it is quicker than converting kana to kanji. When the practice of text-messaging becomes more and more widespread, using kana to represent wago words could become standard practice, as speed is of high importance in this mode of communication. With time, this practice could filter into other types of word-processed texts such as newspapers.

New education policies could also have an impact on the way writers represent wago words. For example, a new initiative was implemented in 2009 to re-introduce the practice of handwriting characters into the secondary school curriculum as it has been declining in recent years (Agency for Cultural Affairs 2009). As the initiative induces secondary school children to learn to write a large number of characters by hand, it is likely that they would begin to use kanji to represent wago words that they now represent in kana. This could gradually increase the number of wago words represented in kanji in Japanese society.

References

Agency for Cultural Affairs, 2005, Committee Discussion 1, viewed June 5 2008, <http://www.bunka.go.jp/1kokugo/pdf/kanji_shouiinkai170913_siryou3.pdf>

Agency for Cultural Affairs, 2009, Shin Joyo Kanji Hyo in kansuru shian (The Draft New Joyo Kanji List), viewed May 26 2009, <http://www.bunka.go.jp/oshirase_koubo_saiyou/2009/pdf/shin_kanjihyoshian.pdf>

Gottlieb, N, 1994, “Language and politics: The reversal of postwar script reform policy in Japan,” The Journal of Asian Studies, vol. 53, no. 4, pp. 1175-1198.

Gottlieb, N, 1998, “Keeping up with the Tanakas: The social construction of word processing technology in Japan,” Japanese Studies, vol. 18, no. 2, pp. 153-163.

Gottlieb, N, 2000, Wordprocessing Technology in Japan: Kanji and the keyboard, Curzon, Richmond, UK.

Miyajima, T, (ed.) 1982, Zusetsu Nihongo, Kadogawa Shoten, Tokyo.

Miyajima, T, 1988, “Kanji no shorai sonogo (Re-examination of “The Future of Kanji”),” Gengoseikatsu (Language Life), vol. 436, pp. 50-58.

Miyajima, T, 1989, “Changes in the language of a magazine,” in S Mizutani (ed.), Japanese Quantitative Linguistics, Studienverlag Dr. N. Brockmeyer, Bochum, pp. 164-174.

Nomura, M, 1980, “Shukanshi no kanjiganyuritsu (Chinese character content of weeklies),” Keiryokokugogaku< (Mathematical Linguistics), vol. 12, no. 5, pp. 215-222.

Nomura, M, 1998, “A study of a collection of words extracted from textbooks for Japanese as a foreign language students,” Koza Nihongo Kyoiku, vol. 33, pp. 95-114, viewed 29 May 2010, <http://dspace.wul.waseda.ac.jp/dspace/bitstream/2065/3331/1/30229_a.pdf>

Nomura, M, 1999, “A statistical study of basic kango words,” Bulletin of Centre for Japanese Studies Waseda University, vol. 12, pp. 21-54, viewed 28 May 2010, <http://dspace.wul.waseda.ac.jp/dspace/bitstream/2065/2763/1/KJ00000178752.pdf>

Nomura, M. and Yanase, T, 1989, “Statistical structure of vocabulary in books for juveniles,” in S. Mizutani (ed.) Japanese Quantitative Linguistics (pp. 14-30). Bochum: Studienverlag Dr. N. Brockmeyer.

Ogura H & Aizawa, M, 2007, “Gendai zasshi nana jushi ni okeru kanji no shiyojittai to joyokanjihyo: kokugoshisaku he no kopasu katsuyo ni muketa kisochosa (Kanji use in seventy contemporary magazines and Joyo Kanji: A preliminary study for the application of text corpora to Japanese language policy),” Nihongo kagaku (Japanese Linguistics), vol. 22, pp. 147-159.

Seeley, C, 1990, A History of Writing in Japan, E.J. Brill, Leiden, The Netherlands.

Seeley, C, 1994, “The Japanese script and computers: The JIS character codes and their periphery,” Japan Forum, vol. 6, no. 1, pp. 89-105.

Tomoda, T, 2005, “Change in Script Usage in Japanese: A Longitudinal Study of Japanese Government White Papers on Labor,” electronic journal of contemporary japanese studies, pp. 1-20, viewed December 2008, <http://www.japanesestudies.org.uk/articles/2005/Tomoda.html>

Umesao, T, 1988, Nihongo to Jimu Kakumei, Kumon Shuppansha, Tokyo.

Yasumoto, B, 1963, “Kanji no shorai (The future of kanji),” Gengoseikatsu (Language Life), vol. 137, pp. 46-54.

About the Author

Dilhara Darshana Premaratne, PhD, is currently a Research Assistant in Higher Education in the College of Arts and Social Sciences at the Australian National University. His PhD research was in the area of script reform in Japan and China. His research interests are language change, language policy and planning, contrastive rhetoric and tertiary teaching and learning.

Email the author

Back to top