Report on Symposium on Nushu

     The “International academic symposium on the propagation of Hanzi漢字and China-Vietnam cultural exchange”, which I participated in, was held in Shenzhen city深圳市, China, December 19 to 21 , 2003, under the auspices of Qinghua University Graduate School清華大学大学院.  The symposium was divided into two sections: cultural exchanges between China and Vietnam and the propagation of Hazi漢字. The study of Chinese women’s scripts was the central focus of the latter section, in which I made a presentation of joint research with Sun Qi 孫埼 (a doctoral student at Waseda University) entitled “Some problems with the Nushu list”.

Some Problems with the Nushu List

1. Study of the eight kinds of lists

     The Chinese women’s script, Nushu 女書, has already fulfilled its historical mission and is on the brink of extinction. The local government has taken various measures to preserve the script, such as building a Nushu study hall to teach young girls the script.  However, since the environment where it was historically used has greatly changed, it is no longer possible to imagine that the script can retain its conventional meaning and usage at the local site.

     Consequently, what we must do now is to shed light upon how the relation of the script affected to the lives of women in the local site, how they were comforted and cheered up by the script and how they got joy from it, to universalize the works written in Nushu and to accurately hand down this valuable cultural property to posterity on the basis of a proper understanding of its essence.

     One of the ways to meet these needs is to develop and record a comprehensive understandig of the script.  In the past, researchers drew up some lists of the script as a first step in this process.  Here, I would like to assess whether these lists describe a correct and comprehensive picture of the script. I also consider more precisely the kind of list that is required at this stage of our inquiry.

     As far as I know, there are eight lists of Nushu script elements, contained in the following eight books:

(1) Gong Zhebing宮哲平 and Zhao Liming趙麗明(1986),“婦女文字和瑶族千家洞” [Women’s Script and the Yao Tribe Qianjia Dong ], 中国展望出版社

(2) 趙麗明 and 宮哲平(1990), “女書----個驚人的発現” [Women’s Script---A Marvelous Discovery], 華中師範大学出版社

(3) 宮哲平(ed.) (1991), “世界唯一的女性文字”, [The World’s Only Eomen’s Script], 台湾婦女新知基金

(4) Chang William Wei (1991), “We Two Know the Script: We Have Become Good Friends: Linguistics and Social Aspects of the Women’s Scripts Literacy in Southern Hunan, China”, Ph.D. dissertation, Yale University

(5) 江永県誌編集委員会編(1995) [Jiangyong Prefecture Journal Editorial Committee (ed.)], “江永県誌” [Jiangyong Prefecture Journal ]   

(6) Yang Yenli, Chen Qiguang and Zhou Shuoyi楊仁里・陳其光・周碩沂 (1995). “永明女書” {Yong Ming Women’s Script}, 岳麓書社 

(7) Zhou Shuoyi 周碩沂 (ed.) (2002), “Nushu Zidian女書字典” [ Dictionary of Women’s Script] , 岳麓書社

(8) Orie Endo 遠藤織枝 (2003), “中国女文字研究” [Study of Chinese Women’s Script], Ph.D. dissertation, Ochanomizu University Graduate School of Humanities and Sciences

     The following are the number of scripts recorded and the order of arrangement found in the above-mentioned books, along with my notes on making the lists:

(1) 670 script: arranged in two ways (according to stroke count and in phonological order)

(2) 1,120 script: arranged in two ways according to stroke count and in phonological order)

(3) 1,120 script: arranged according to the order of stroke count

(4) 1,535 script (including 719 alien scripts): arranged in phonological order 

(5) 1,731 script: arranged according to the structure of creating the component parts; phonology according to Baishui tu hua白水土話

(6) 1,570 script (including alien scripts): arranged according to troke count

(7) Over 1,800 script: arranged in phonological order in dictionary form

(8) 470 script: arranged according to the order of stroke count and phonological order; the scripts were written by He Yanxin何艶and the phonology is based on her pronunciation

2. The problem of alien script

    As seen above, the number of script elements is wildly various because the interpretation of alien scripts and dealing with them are different.  Other than lists (1) and (8), all the script elements are counted individually and given separate numbers.  In the case of (8), alien script elements are included in the same frame as the central scripts and treated as variations on them, and are not considered in the number.

    These scripts are certainly different in the numbers of dots and lines, and also different in type, but it is presumed that they represent the same sound and that original Chinese characters from which they are derived are the same. Their usage is not different. In the course of transmission, they have changed according to regional and individual differences between transmitters and learners. Dealing with such alien script elements as independent ones would only lead to increasing the number of script elements, and the real number of script elements used by traditional women would be different. Script elements with the same phonology and usage and which are only a little different in shape should be considered to be in the same category.

     For example, let’s look at the case of “難”().

     In Book (3), this character is registered both at 6 strokes and 7 strokes.  Under the 6-storoke entry, “難” is allocated to “” nu51 and under the 7-stroke entry, two scripts, “難” and “奴”, are allocated to “” nu51. Both scripts look alike in type, but “” is slightly rounded in such a way that its number of strokes decreases and it was, thus, regarded as a different script.  As the roundness of scripts is caused by a particular writer’s way of writing, I have doubts about dealing with this variant as a different grapheme.  Given that the two entries both have the same phonological form, I consider them to be variants of the same grapheme.

    A further source of confusion is that only “難” is allocated in “” under 6- strokes and two Hanzi漢字, “難”and ”奴”, are allocated in “” under 7 strokes.  It is likely only by chance that “難” was used for “” and both “難” and ”奴” were used for “” in the material employed by the person who made the list.   Regarding “” and “” as different graphemes resulted in maintaining that the scripts allocated are also different.  Considering the converse situation, we also cannot say that “奴” is not used – or could not be used - for “”.  

    In such cases, it would be a realistic way of dealing the problem to register either “” or “ “ as the standard form and to register the other script grapheme as a variant one, and then allocate “難” and ”奴”.

3. The problem of phonological aspect

     In Book (7), “難” is allocated both to noi41 and noi33 on the grounds that , although its form is the same, its phonology is different.  I cannot understand why the same script is dealt with as two separate scripts even though their style, phonology and usage are not different.

     In the case of “” (門), too, each list is subtly different in phonology and tone, as for  (1) me51, (3) mei51, (4) me42, (5) mai41, (6) mai41, (7) mai41 and (8) mai42 (Chengguan yin城関音), mi42 (Shangjiang xuyin上江墟音).  Mei51 and mai42 are different in phonology and tone, and mai41 and mai42 are different in tone.  These subtle differences may have been caused by �@ differences in the speakers’ local pronunciations of specific graphemes or differences how they were taught, �A differences in the linguistic environment of a particular pronunciation, such as whether the script element is pronounced as in isolation, in a phrase, or in a song, and �B differences in recorder’s way of grasping the sound.  Except for Book (8), compilers have not specified whose pronunciation in which district the phonological rendering is based on, so that the meaning of the differences in phonology or tone between the lists is unclear.  Moreover, it cannot be judged whether the differences are significant or insignificant and simply  due to an accidental difference in pronunciation or recording.  It is essential to specify all the conditions mentioned above in making a reliable and accurate list.

    Generally, each character has its own sound, so that information on both its style and its pronunciation is required.  However, in the Nushu lists made to date, information on pronunciation is not described effectively, and there is considerable inconsistency from list to list, or even within lists. As seen in (7) above, there are many lists registering different characters based on subtle differences in phonology or tone as a result of judging the differences in phonology or tone too strictly.  Lists should be adjusted by assessing phonological and tone similarities comprehensively and widely, rather than being overly swayed by subtle differences that may derive from simple regional variation in pronunciation or inter-recorder differences in hearing/judgement.

4. Comparison of script style

    Showing the scripts written in a specific San Chao Shu三朝書(Third Day Missives) first, I will assess how the same graphemes are depicted in the lists in (3), (4), (5), (7) and (8). 

San Chao Shu is a cloth bound booklet which is given to a bridegroom from his bride’s mother as a present on the third day of marriage.  It is filled with songs of her sorrow of parting with her daughter leaving the village for marriage.

    The graphemes written in San Chao Shu are fine and thin; they are  and characteristically long and narrow, and the slope from upper right to lower left is strong.  In the contrast, the grapheme “a” in (3) has little slope.  The grapheme “b” in (4) is thick and squarish.  The graphemes “d” in (5) is not long and narrow, either.  All graphemes in (7) are written with thick lines, having quite a different distinctive quality.  Also “b” is not similar in style.  The graphemes “a”, “b” and “d” in (8) are very similar to the style of San Chao Shu.

    The style and distinctive quality in each list is different as just described because of the individuality of the author of each list shapes their distinctive character.

    To hand down these figures of the scripts to posterity, we should correctly transmit their style as written by women at their various local sites and properly reflect their distinctive qualities.  The characters recorded in (3), (4), (5) and (‘7) are not those written by transmitters; but those in (3), (4) and (7) were written by each author of the list, and those in (5) were written by a woman close to the editor.  Only the characters recorded in (8) are written by the transmitter herself.

     Script lists should be made by transmitters as long as even a few of them are still living.  Even basing a list on the writing of a deceased transmitter, it should be possible to make lists by using high-resolution digital reproductions rather than hand-produced copies of the graphemes in their original materials as the basis of the character list.  Without doing so, it would be too easy to make lists using characters written by researchers themselves.

5.  Summary

    I have reviewed the problems of Nushu lists from the three points of dealing with script variants, and phonological variability, and styles.  In all three of these areas, the ways of dealing with characters varies from list to list, and numerous contradictions and confusions are found.  At present, therefore, it is difficult to get a comprehensive picture of Nush.  However, if the situation is left as it is, we will be leaving an unnecessary legacy of confusion to future generations of scholars.

     While the transmitters live in good health, it should be a matter of the first priority to correctly record their scripts and phonology.  On the occasion of such recording, we should consider carefully how to deal with script variants and phonological variability across speakers and specific localities, calling upon as much collaborative expertise as possible.  One effective way to deal with phonological issues would be to record the voices of the existing transmitters on CD-ROM, which then could be attached to the list.

[Other presentations]

     The summaries of the presentations of Zhao Liming Prof. 趙麗明 of Qinghua University and his students of the graduate and undergraduate schools as follows: 

(1) “Guanyu nushu ziku de jianli関干女書字庫的建立”（Wang Rongbo & Mo Jingqing王栄波・莫静清) 

    After extracting 447 basic Nushu graphemes, the author combined them with their corresponding Hanzi 漢字) one by one, and encoded them.   Different graphemes with the same meaning are distinguished by linking them to similar Hanzi.  As a result, if you input the correct Hanzi, you can retrieve its corresponding Nushu grapheme, and you will be able to write a composition in Nushu.  

(2) “Nushu chansheng yu cunzai de shehui jichu chutan  女書産生与存在的社会基礎初探” (Wei Qingyuan韋慶媛）”

    The author studied the social environment in the area where Nushu was transmitted.  The area was a society closed to outsiders, but internally there was relatively open women’s society, which acted advantageously on the transmission of Nushu.  Although women’s social position was low in this area, they rebelled against the Hanzi culture by creating scripts in which they could express their own sorrow and hatred.  In the area, Han (漢族) and Yao (瑶族) people lived together, and the women there had, under the influence of Yao culture, more liberty than those in Han area of Zhongyuan中原.  They had a festival only for women.  In such a daily life, they learned Nushu and sang its songs.  Thus, Nush was born in such a very unusual district and from there, spread gradually.

(3) “Nushu diyi rencheng daici ‘wo’ kaojiu 女書第一人称代詞‘我’考究” (Xie Xuan謝玄)

     In the works of He Yanxin何艶新, two characters are used for “I” (我) .

The author studied how these two characters are used distinctively based on analysis of their phonology and of their use in the context of Nushu songs. 

(4) ”Nushu de shuzuhua yanjiu 女書的数字化研究“(Sun Xiaomei孫暁梅, Wang Zhentian, et al.王振天他)

     The authors plan to classify the Nushu culture and studies into the categories of “research papers, researchers, Nushu works, Nushu transmitters, the peripheral culture of the central region where Nushu was propagated, and the natural environment of the region where Nushu was propagated.”  Then, for each item, they will build a database equipped of the available data, including scripts, photographs, and voices. 

    For example, in the case of the Nushu transmitters Gao Yinxian高銀仙, Yang Huanyi陽煥宣and He Yanxin何艶新, they are constructing such a database so that they will be able to observe the connection of their life stages and events with the songs at the same time by classifying their individual data into “infancy, girlhood, time in love, marriage, after marriage, old age and after death.”

(5) “Cong ＜Luoshi nu＞kan Gao Ynxian, Yi Nianhua Nushu zupin de chayi従＜羅氏女＞看高銀仙、義年華女書作品的差違” ( Yang Hua & Liu Shuangqin楊華、劉双琴)

     Based on the texts of “Luoshi nu羅氏, written by 女Gao Ynxian 高銀仙and Yi Nianhua義年華, the authors analyzed differences in their use of characters.  They found that Gao Ynxian had used 77 variant graphemes in the work, while Yi Nianhua had used only 20.  That is, Gao’s way of writing is highly individual, while Yi’s is normative.  

     Additionally, both of them incorrectly wrote the same phrase, and comparing the difference in their mistakes, the authors found that Yi chose more rhymed words and logical words than Gao.  That may be due to differences in the education they received.   

(6) “Nushu zifu tixi de xingcheng 女書字符体系的形成” (Zhao Liming趙麗明)

     Nushu was born from Hanzi 漢字, and the author classified the Nushu script elements by the degree of their change from Hanzi . That is a supplement to the results of his 20 years of research, as mentioned below:

1. Scripts changed by transplantation  

1.1 About one-fifth of those basically retain the original form.

1.2 About one-fourth have been changed.

1.3 About half retains their original basic outline.

2. Scripts created by using basic symbols

3. Scripts with a common pronunciation written by the same character

     Above, I have made a summarized report of the presentations on Nushu at the symposium.  As may be expected from a symposium organized by Qinghua University Graduate School, its undergraduate and graduate students were active participants and many made presentations.  However, most of the research focused on how to process the data using a computer, so that they might be called computer applied research rather than Nushu study per se.

     There was no presentation indicating that its author would engage in a full-fledged study of Nushu to squarely tackle some of its basic problems.  As the comparison of differences between graphemes may be the easiest entry-point part for such a full-scale study, it would be wise to begin with that.  Researchers could face a number of the basic problems in the process of that study, and there is an ample reason to believe that the study would develop in interesting directions.  I would like to call upon someone among the younger generation of Nushu scholars to tackle this task -- which is most basic to serious Nushu study – seriously in the future.