文教大学学術機関リポジトリ

(1)

and Its Significance in

Distinguishing Social Class : A Preliminary Study

David C. Heil

1 . Introduction

It is well known in the field of linguistics that how we speak reflects a mirror image of who we are. Social, ethnic, educational, as well as regional influences, inter alia, are all reflected in one's speech. Studies such as The Social Stratification of (r) in New York City Department Stores (Labov, 1972), It Bees Dat Way Sometimes (Smith-erman, 1977) and The Study of Nonstandard English (Labov, 1969) all give irrefutable support to this sociolinguistic claim.

The above studies were all carried out in the linguistic-rich environment of the U. S., that is, an environment full of cultural, social, and ethnic diversity. However, other studies have also been carried out in other countries around the world, documenting similar sociolinguistic phenomena, which suggests that these types of phe-nomena occur world wide. With this in mind, the question arises whether these phenomena also exist in Japan, a homogeneous country of predominantly middle-class people. Therefore, the aim of this paper is to examine the linguistic environment of Japan for signs of social stratification and then base a study on these signs.

In Section 2, various linguistic features and approaches for studying these features are considered and then selected for research. Section 3 presents the hypotheses, and Section 4 discusses the

(2)

— Dialectal

Usage

and

Switching

in Japan

and

Its

Significance

in Distinguishing

Social

Class

: A Preliminary

Study

—

methodology used. In Section 5, the results are given and discussed,

and Section 6 examines problems encountered in this study and

suggests ways for improvement. Finally, Section 7 presents

conclud-ing remarks.

2. Isolating Items for Study

To find data for this research, I first considered items that were

previously studied in other research projects, such as those mentioned

in Section 1. I considered the nasalized ga (Lel nga I) versus the

non-nasalized/ga/in the Japanese language, but after careful

exami-nation I found that this was virtually non-existent in the speech

patterns of those Japanese I tested. Although I recorded the speech of

subjects without their knowing in order to get the vernacular style of

speech, being a foreigner and especially not that fluent in Japanese,

I encountered predominantly only speech that was carefully

monitor-ed (also known as the observer's paradox), because, as the subjects

teacher or friend, or just because I was a foreigner, the subjects

tended to adjust their speech to my level or a level they associated

with a white foreigner. This accommodation of their speech caused

such data to be useless.

Another area considered was the twirling/r/found

prevalently in

the speech of Japanese Yakuza. This can also be heard in the speech

of many Japanese from a variety of educational and occupational

levels (mainly men) as a kind of macho or tough way of speaking,

used especially when angered. The twirled/r/seems

to be produced

from the blending of certain combinations of consonants and vowels

in rapid or angered speech. This type of data would also be difficult

(3)

to obtain, just because the subjects would have to be angry in order

to get good data. However, in the case of the Yakuza, I believe such

speech could be easily identified with that group as a legitimate

sociolinguistic feature, but, as I said before, this feature is not

restricted to only this group, thus, negating the legitimacy of the

results. [Such negating of research results has previously occurred

with other studies (Philips, 1980).] Besides, working with the Yakuza

could be difficult, if not hazardous, to one's health.

The problem of studying consonants, as well as vowels and

grammatical features, is compounded further by the influence of local

and regional dialects. The pronunciation of hi as shi in the Tokyo

dialect produces a variable which is not related to social status.

Vowels, for example, can be pronounced in numerous ways according

to the dialect (Miller, 1967). Variation in lexical items and grammar

usage can also be attributed to dialectal differences rather than social

status.

If you consider the enormous influx of people moving to Tokyo

from other urban, as well as rural, areas, the everyday commuting

distances (two hours by high-speed train) workers travel to go to

work, and the frequency with which workers are reassigned to other

areas, it becomes apparent how extremely difficult, if not impossible,

it is to carry out research along these lines, because it would be very

easy to mistake a dialectal feature for a feature related to social

class.

After considering the above arguments, the question still

remains : How can we determine social class from linguistic features

of speech? I decided to consult Japanese themselves ; first in groups

(4)

— Dialectal Usage and Switching in Japan and Its Significance in Distinguishing Social Class : A Preliminary Study —

(e.g. English classes and families) and then as individuals.

When a foreigner, as I had done, directly asks a Japanese in front of other Japanese whether some form of social stratification exists in Japan, the usual response is a quick, and very confident no. The same confident no is also received when asked whether they could tell a person's social or educational status from just listening to a person's speech. This is to be expected since the middle-class consciousness of the Japanese after WWII, in a land where the group is valued more than the individual and where the nail that sticks up gets hammered down, has been nurtured and re-enforced by many of the dominant social forces (i.e. large companies) of Japanese society (Nakane, 1970). In fact, an opinion poll taken in 1977 showed that 90% of Japanese considered themselves to be middle-class and no more than 5% claimed low-class (Woronoff, 1980). Certainly, now, the percent-age of Japanese who claim low-class status has reduced to an insignif-icant level in light of Japan's continued economic growth. How-ever, we should remember that these statistics are based on opinion and not fact. Recent surveys of the ten richest men in the world have several Japanese names on their list, presenting some evidence con-trary to these opinion polls.

Despite the lack of any conclusive support for class distinction from group discussions, I have, however, encountered in private interviews many differing opinions, mainly from people with high educational backgrounds (e.g. university professors and business men), that give some evidence from a linguistic perspective to support the existence of social stratification in Japan. Examples of people referring to taxi drivers as low-class (or low-level) which is the

(5)

translation of shita no hito or hikui hito, or overhearing Japanese complain about the incorrect usage of polite standard forms of Japanese (e.g. using irrasharu-honorific to go-for oneself ; or itadaku-humble to eat 1 receive-for someone else) give support for promising research topics.

To isolate those linguistic elements that are related to social identification in a culture that considers themselves to be predomi-nantly middle-class is very difficult because the distinction between classes has gradually disappeared over the years with the introduc-tion of a naintroduc-tional educaintroduc-tion program, mass media, mass transit, etc. To discover possible angles for researching such a topic, we need to take a step back and examine those features of Japanese before such institutions became common in Japan, and then use these elements to examine modern Japanese. To do this, we can't very well go back in time, and we also can't use the literature from past history because this material was primarily written by and for those of the upper classes. One thing we do have available today is numerous T.V. programs depicting the life of Japanese in earlier periods of Japanese history. These T.V. shows are often called Samurai shows by many foreigners in Japan and are very similar in design to American westerns.

The Samurai program I examined was Toyama no Kinsan. This program is about a high ranking government official during the Edo period (1603-1868), who doubles as a commoner to carry out law and order. In this program, one can easily tell the difference between the upper class and the common class from speech. The main difference is that the upper class uses standard, more formal Japanese ;

- 87 —

(6)

— Dialectal Usage and Switching in Japan and Its Significance in Distinguishing Social Class: A Preliminary Study —

as, the common, lower class uses dialectal, informal Japanese. If we apply this principle to Japan of today, we can easily tell the difference between, say, a university professor and a construction worker. The

professor will use polite, formal forms of standard Japanese much more than the construction worker, thus, identifying the professor as the one having higher social status. One may argue that the construc-tion worker might make more money than the professor, thus, giving him/her more social status. Here, it becomes necessary to define how social status is determined. Social class is based not only on money ; it is based on one's education, one's responsibility in the community, and other factors. The main factor in determining social class is how society views a person's profession. For example, a foreign diplomat from an extremely poor country visits Japan. The diplomat probably makes much less money than most Japanese construction workers, but he is obviously considered higher class than the construction

worker because of his education, position, etc. and is treated as such. Along the same lines, the professor is considered to be in a higher class than the construction worker.

To expand on the model further, one can visit different ranking department stores and notice similar differences in the speech of their workers reflecting the status of the store. In Takashimaya depart-ment store, an upper middle-class store, one hears the workers speaking in very polite standard Japanese. In Daikuma, a middle middle-class discount store, one hears standard Japanese spoken, but not with the same degree of politeness as is spoken in Takashimaya,

along with traces of the local dialect or other informal speech. Here, we can easily identify the difference in status and prestige of the two

(7)

stores by their employee's language.

This approach seems to be relatively valid for simple observa-tions. The next step, as will be presented in Section 3, is to develop a hypothesis or hypotheses related to these linguistic features, that is, the use of standard Japanese and the use of dialectal Japanese.

3. Developing Hypotheses

Before developing hypotheses for research, we need to determine what resources are available to the researcher, and then adjust the hypotheses so they can be adequately proved or disproved. As a foreigner with low Japanese ability, my resources for doing research of this sort are extremely limited. I was definitely going to require some translation work. Here, I wanted to limit this to cut down on costs as well as on time. Another important factor was that I was doing this research by myself, which, again, limited my time. A final, but very important factor, was that prosodic speech features such as intonation, or paralinguistic features like voice quality, though very distinguishable in Japan, could not be studied simply because no practical techniques are available for such research (Trudgill, 1974). With these points in mind, I came up with two hypotheses that took into consideration the points discussed in Section 2 and were also provable with the resources available to me. The hypotheses are as follows :

1. The amount of standard Japanese used in one's daily speech will reflect one's social class, i.e. lower-middle, middle-middle, and

upper-middle.

2. The degree to which one switches to standard Japanese in — 89 —

(8)

various status related situations will be higher in those Japanese

with high status positions.

Both of these hypotheses can be tested relatively easily. The first one could be tested by just asking individuals what percentage of their daily speech consists of standard Japanese, and then correlating it with his/her occupation and educational background. The second one could be tested by presenting different scenarios correlating to the vertical structure of Japanese society, and asking which type of speech one uses ; either standard Japanese, one's local dialect, or the interlocutor's dialect.

The methodology for testing the hypothesis is described in more detail in the next section.

4. Methodology

The methodology for collecting data was that of a questionnaire. Data for hypothesis 1 would be gathered by directly asking the subjects how much of their daily speech consisted of standard Japanese and how much consisted of their dialect (see Appendix). Although this approach depends a lot on the subjects' subjectivity, it is a very quick means of collecting data ; and because the results are averaged among the other subjects' data within a particular group, it should give fairly accurate figures for each of the three groups.

Since the use of a dialect is usually associated with members of a group that are relatively isolated among their peers within a specific region, usually associated with lower class groups, its usage should be most predominant in the low-middle class group, because

(9)

this group associates the least with members of the middle-or upper middle classes and has much less mobility socially as well as territor-ially. Therefore, this group has the least need for a language to communicate with outsiders (i.e. standard Japanese) or polite forms (also to be found primarily in standard Japanese). Likewise, the upper-middle associates with the language of education, the language of government, etc. and will accordingly speak it (standard Japanese) more. The speech of the middle-middle class should range somewhere between the low and the upper. Therefore, the results of this question-naire should reflect a higher use of standard Japanese in the upper-middle class than the upper-middle-and low-upper-middle classes.

For the second hypothesis, a questionnaire was also used. How-ever, the nature of each question in this section was much more indirect. There is a vertical structure in Japanese society that affects how one speaks. Depending on one's position, a Japanese can either be in a high, an equivalent, or a low position with respect to his/her peers in the group he/she belongs. Therefore, questions simulating these three positions were presented to the subjects in the question-naire (see Appendix...). Questions 8 and 9 represent a low position for the interlocutor and a high position for the subject ; questions 10 and 11 represent a middle or equal position for the interlocutor and the subject ; and 5,6, and 7 represent a high position for the interlocutor and a low position for the subject. The amount of dialectal speech should be the highest in situations in which the subject is in an upper or middle position with respect to the interlocutor.

Numerous distractors were added, and these distractors were designed to provide information for other research topics, e.g. sex

(10)

— Dialectal Usage and Switching in Japan and Its Significance in Distinguishing Social Class : A Preliminary Study —

and language.

To divide the subjects answering the questionnaire into the social classes of lower-middle class, middle-middle class, and higher-middle class, the following educational standards were applied.: For lower-middle class, a high school education or below ; for middle-middle class, an education above the high school level up to and including a vocational school education (Senmon Gakko), a junior college educa-tion, and a four year university education ; and for high-middle class, any graduate or post undergraduate (e.g. medicine, engineering, and law) education. It should be pointed out that it is very possible for someone categorized in the lower-middle class group or the middle-middle class group to speak Japanese at a level equivalent to that spoken in the high-middle class group. In addition, it is also possible for someone in the lower-middle and the middle-middle class cate-gories to have the status and responsibilities equivalent to those in the high-middle class category. Likewise, it is possible for someone categorized as high-middle class to have speech patterns and social status and responsibilities equivalent to the low-middle and middle-middle class levels. However, as is true in other sociolinguistic studies, it is generally safe to say that for each respective category as a whole the overall averaged speech patterns of that group should show stratification in the speech between the respective groups as already described. For this study, low-middle class should speak more dialectal speech than the middle-middle and high-middle classes, and standard Japanese should be more prevalent in the speech of high-middle class than in the lower high-middle classes.

(11)

5. Results

One hundred thirty questionnaires were mailed or given out with a self-addressed envelope to Japanese all over Japan. Of this number, only 60 responded.

The subjects were divided into three categories : 1) Japanese from outside the Tokyo area (Tokyo, Yokohama, Saitama, and Chiba) living in the Tokyo area ; 2) Japanese from outside the Tokyo area living in their native area ; and 3) Japanese from the Tokyo area living in the Tokyo area. Data was sufficiently collected for category 1. For category 2, there was insufficient data to evaluate the lower-middle group ; however, there was sufficient data for the other two groups. For the third category, there was some confusion about how to interpret the differences between standard Japanese and one's own dialect. Since the dialects of the Tokyo area are very similar to standard Japanese, many of the subjects referred to their dialect as standard Japanese. This made the data inconsistent and unusable. The results for hypothesis 1 were first calculated for each question and then averaged for all the questions for that position according to sex. Finally, the two averages were averaged together to get the overall average for a position. The results are listed below :

Percent of dialect usage for category 1 (Japanese from outside the Tokyo area living in the Tokyo area.

Men

Lower-Middle Middle-Middle Upper-Middle

23% 12% 10%

(12)

-— Dialectal Usage and Switching in Japan and Its Significance in Distinguishing Social Class: A Preliminary Study —

Women

38% 13% 0%

Total

30.5% 12.5% 5 %

Percent of Dialect Usage for Category 2 (Japanese from out-side the Tokyo area living in their native area.

Men

11% 8 . 5%

Women

41% 57% *

* One subject gave an unusually high percent and the number of subjects was only 3, raising doubt about the accuracy of this data. Excluding the one subject, the data would be 40%.

Total

26% 32.75% *

* Not conclusive due to insufficient _{data . Excluding the one subject,} the data would be 24.16%.

In general, the results for category 1 clearly show stratification according to dialectal speech usage. This is to say that the upper-middle group uses standard Japanese much more than the other two

(13)

groups. However, the second category doesn't show this for the group as a whole, though the men in this group did show stratification to some degree. The women, however, gave results that disproved hypothesis 1 and gave results that showed that the opposite was true. However, the data for the women's upper-middle in category 2 is suspicious. First of all, it was difficult to find enough subjects for the upper-middle group, and one subject who was a full professor of an Osaka University stated that she would even teach in the Osaka dialect, which is inconsistent with the other University professors who answered this questionnaire. Most professors would teach in standard Japanese even if they were teaching outside the Tokyo area. Thus, there is reason to doubt this data. However, there are possible explanations and one in particular. Through this research, I have noticed that some dialects, like Osaka-ben, have higher status and are received more positively than some other dialects. In the case of Osaka-ben, its higher status is probably due to the great economic success the area has enjoyed, and, thus, no one would object to it being used as a language for education. However, I doubt if Osaka-ben would be used as the language for academic, written use. In any case, this area needs to be investigated further to fine tune the

questionnaire to produce accurate data.

The results for the second hypothesis are given below. Here, too, the results are averaged in the same way as described above.

Percent of standard Japanese usage for category 1 (Japanese from outside the Tokyo area living in the Tokyo area).

(14)

Upper Position Scenario (Questions 8 and 9)

Men

75% 38% 66%

Women

63% 88% 100%

Total

69% 63% 83%

Middle Position Scenario (Questions 10 and 11)

Men

87.5% 66.5% 100%

Women

60% 82.5% 100%

Total

73.5% 74.5% 100%

Lower Position Scenario (Questions 5, 6, and 7)

Men

62.5% 75% 50%

(15)

Women

63.75% 77% 100%

Total

63.75% 76% 75%

To prove hypothesis 2, the above percentages should show an increase in the columns from left to right ; that is, the percentages for each row should be lowest in the lower-middle column and highest in the upper-middle columns. However, these results do not show this pattern. In fact, these results support the theory that Japanese are indeed all middle-class.

Here are the results for category 2.

Upper Position Scenario (Questions 8 and 9)

Men

58.33% 91.66%

Women

77.33% 61%

Total

67.83% 76.33%

(16)

Middle Position Scenario (Questions 10 and 11) Men

87.5% 100%

Women

83% 87.5%

Total

85.25% 93.755%

Lower Position Scenario (Questions 5, 6, and 7) Men

75% 87.5%

Women

94% 87.5%

Total

84.5% 87.5%

In this category, from looking at the totals of each scenario, there is evidence that the upper-middle group uses standard Japanese more than the middle-middle group. However, the difference is not that significant to support hypothesis 2. In addition, the other data in

(17)

this paper has not overwhelmingly supported hypothesis 1 or 2. In fact most data has supported the myth that Japanese are middle class. However, after conducting the research and examining the data I have, to my embarrassment discovered many problems with my research and have devised ways to make it more reliable.

6. Conducting Better Research

Although my research failed to prove whether a social stratifica-tion exists in Japan or not, much was learned about how to conduct better research in Japan about determining social class of Japanese from their speech. First of all, I attempted to survey Japanese from too many different locations in Japan. This in itself created numerous problems because the attitudes towards one's dialect when contrasted with standard Japanese varies tremendously. When confronted with standard Japanese, some Japanese have a positive impression of their dialect and speak it openly, while others will hide their dialect completely and give no trace of where they are from. There are still other issues that complicate dialect usage and switching to standard Japanese. In addition to the positive or negative image a dialect may have in the eyes of its user, Japanese also switch to enhance commu-nication. The more complicated the dialect or, in other words, the more the dialect differs from standard Japanese, the more the speaker will switch to standard Japanese when faced with someone from any outside area. Thus, switching may occur more frequently in an area with a complicated dialect than in an area with a dialect that is similar to standard Japanese. This switching is not a result of social class but rather a result of trying to improve communication,

(18)

rendering the data collected in such situations useless.

Another problem that arose was that many believed that their particular dialect was standard Japanese. When I surveyed Japanese from Yokohama, most were aware that their dialect was called Yokohama-ben, but were not aware of exactly how it differed from standard Japanese. Almost all of these subjects answered the ques-tionnaire as if they spoke only standard Japanese, negating the results.

Still, the question of whether dialects have a polite or elevated form equivalent to similar forms in standard Japanese could also complicate such research further. If one could express

in their own dialect the same degree of, for example, intellect or social status, perhaps switching to standard Japanese would be decreased, rendering the hypotheses useless.

Another obstacle was the rather lengthy questionnaire. It was approximately seven pages long and had some questions that seemed overly difficult or ambiguous to the participants.

To overcome many of the problems mentioned above, the rule of simplicity should be followed much more diligently. One way to do just that would be to conduct the research in a specific region in Japan and solicit subjects who are living in and are from only that particular area. In this way, the different attitudes about the dialect, either positive or negative, would influence the data in a consistent manner, and, thus, the data would be more reliable. In addition, perhaps two studies could be conducted, for example, one in an area in which the people enjoy a positive image of their dialect face to face with standard Japanese and another in an area that doesn't. Such a comparison could not only prove the hypotheses but could reveal the

(19)

different attitudes Japanese may have about Japanese dialects.

Another possible approach would be to conduct research in an area that has a dialect that differs greatly from standard Japanese, e. g. Kagoshima-ben. In this way, we could learn not just about social stratification and its relation to language but also about how Japanese from this region switch between their dialect and standard Japanese when faced with people from outside their area.

7. Conclusion

In this paper, sociolinguistic research on Japanese within Japan was conducted. The question of how to distinguish from speech the difference in social status between, for example, a highly educated medical doctor and a construction worker were discussed. Two hypotheses were developed to answer such a question and became the focus of this research. The hypotheses are as follows

1. The amount of standard Japanese used in one's daily speech will reflect one's social class, i.e. lower-middle,

middle-middle, and upper-middle. (The more standard

Japanese is spoken the higher the social class.)

2 . The degree to which one switches to standard Japanese in various status related situations will be higher in those

Japanese with high social class.

Data to prove or disprove both hypotheses was collected through a questionnaire. Data for hypothesis one was collected by asking individual Japanese what percentage of their daily speech consisted of standard Japanese and then correlating this to their occupation

(20)

— Dialectal Usage and Switching in Japan and Its Significance in Distinguishing Social Class: A Preliminary Study

and level of education. Hypothesis two was tested by presenting different scenarios, and asking which type of speech one uses ; either standard Japanese, one's local dialect, or the interlocutors dialect.

The results failed to provide conclusive evidence that a social stratification based on dialectal and standard Japanese usage exists in Japan. On the contrary, the evidence suggests that Japanese, at least with respect to dialectal and standard Japanese switching patterns, are indeed middle class, as many Japanese openly attest to.

In this preliminary study, however, problems unique to Japanese dialect research were discovered and examined. These problems severely question the accuracy of the data collected. Possible solu-tions for overcoming these problems were suggested. Discovering problems in a preliminary study such as this provided great insights into the complexity of conducting linguistic research in Japan as well as prevented inefficient use of financial resources and time, showing the utility of a preliminary study.

It is hoped that by conducting such research in the future, for-eigners can gain a better understanding of just how language is used in Japan and provide valuable information for helping more people understand Japanese language and culture better.

References

Labov, W. 1972. The Social Stratification of (r) in New York City Department Stores.

Chapter 2 in W. Labov, Sociolinguistic Patterns. phia : University of Pennsylvania Press.

Labov, W. 1969. The Study of Nonstandard English. Washington : — 102 —

(21)

Center for Applied Linguistics.

Miller, R. A. 1967. Dialects. Chapter 4 in R. A. Miller, The Japanese Language. Chicago : The University of Chicago Press. Nakane, C. 1970. Japanese Society. Tokyo : Charles E. Tuttle. Philips, S. 1980. Sex Differences and Language. Tucson : Department

of Anthropology, University of Arizona

Smitherman, 1977. It Bees Dat Way Sometime : Sounds and Structure of Present-day Black English. Chapter 2 in (author unknown)

Talkin and Testifyin. (Publisher Unknown).

Trudgill, P. 1974. Language and Social Class. Chapter 2 in P. Trudgill, Sociolinguistics : An Introduction to Language and Society. London : Penguin Group.

Woronoff, Jon. 1980. Happiness is a well-regulated Society. Chapter 8 in J. Woronoff, Japan : The Coming Social Crisis. Tokyo

Lotus Press, Ltd.

Appendix

* The questionnaire is too long (7 pages) to be published in this Journal. Only questions relevant to collecting data for hypothesis two are provided.

Questionnaire-Part 3

5 . You go to a hospital in your home town and you notice the doctor is using standard Japanese. What dialect do you use to speak to the doctor?

Standard Japanese Your dialect — 103 —

(22)

6 . You go to a hospital in your home town and you notice the doctor is speaking your dialect. What dialect do you use to speak to the doctor?

Standard Japanese : Your dialect

7. You go to a hospital in your home town and you notice the doctor is speaking a different dialect than your own. What dialect do you use to speak to the doctor?

Standard Japanese ; Your dialect ; The

doctor's dialect

8. You go to the local supermarket in your home town. You ask the clerk where something is. What dialect do you use?

Standard Japanese ; Your dialect

9. You are living in an area that uses a different dialect than yours. You go to the local supermarket and ask for something you can't find. What dialect do you use?

Standard Japanese ; Your dialect ; the

local dialect

10. You are working in your home town. It's lunch time and you are sitting by yourself. A man comes up to you and introduces him/ herself saying in standard Japanese that he is working for the same company in a different town. What dialect do you use to talk to him?

Standard Japanese ; Your dialect •

11. The same situation as 10, but only this time you are talking to several of your friends in your local dialect, which the man happens to overhear. What dialect do you use to talk to him in the presence of your friends?

Standard Japanese ; Your dialect — 104 —