China comprises a vast land space and a multitude of different peoples and languages.
Nonetheless, the majority group known today as Han Chinese is regarded as having a single cultural system transmitted by a written language that comprehends the considerable diversity of Han populations nationwide. As discussed in the introduction, critical Han studies has sought to deconstruct the notion of a monolithic “Han Chinese” identity on the grounds that this elides or renders invisible the considerable diversity that exists in the population known as Han Chinese. In this study we continue the debate by examining the situation of folk traditions in the case of China’s Wu-language-speaking populations of the lower Yangzi Delta.
Once one turns to oral traditions, the distinctiveness of Chinese regional cultures becomes more obvious. However, scholars of Chinese civilization have focused on written works in classical Chinese (*wenyan* 文言), or on more popular works in a standard vernacular (*baihua* 白话), and only rarely on the so-called “regional cultures” (*difang wenhua* 地方文化) that reflect the actual lived experience of communities known today as Han Chinese. This relative neglect is particularly true of the rich folk culture of rural populations bordering Lake Tai and Lake Fen in the lower Yangzi Delta region, no doubt due to the geographic and linguistic complexity of the region. The Wu 吴 language grouping is complex and multiform. It prevails with numerous variants over large tracts of the provinces of Jiangsu and Zhejiang, together with the metropolis of Shanghai.
The discovery of the “folk,” their languages, beliefs, and cultural forms, dates back a century. It began as part of China’s early modernization movement of the 1920s but investigations came to an end during the ensuing period of foreign invasion, civil war, and political turmoil. Early socialist China (post 1949) saw a renewed interest in the popular culture of the common people, but scholarship in this field was aborted and largely destroyed during the Cultural Revolution (1966–1976). This situation began to change after party leader Deng Xiaoping came to prominence (post 1978). In the 1980s, the Chinese state began a massive effort to compile items of regional culture such as local opera, storytelling, songs, proverbs, and the like. In the case of the lower Yangzi Delta, folklorists collecting songs from villagers were surprised to discover song-cycles of epic length in the rural hamlets on the shores of Lake Tai and its tributaries. Chinese scholars were puzzled by the “anomalous” nature of their findings, which did not fit into existing paradigms about the historical development of Chinese civilization. Contemporary perceptions of the essential unity of the vast population designated as “Han Chinese” also obscured exploration into significant divergences among “Han Chinese,” of which the most significant is the difference between people living north and south of the Yangzi River.[^1]
Another thorny issue was how to deal effectively with stigmatized aspects of these songs such as “feudal superstitions” and erotic material. The local ethnographers were torn between two contradictory goals: on the one hand, the academic goal of accurately recording and publishing what the singers actually sang, and on the other hand, the perceived need to win approval from the political hierarchy. There were aesthetic problems as well. The ethnographers wished to develop exempla from oral culture that provided a satisfying reading experience and conveyed a clear message acceptable to the general public and the authorities (such as “an attack on the feudal marriage system”). However, the fragmented nature of the material did not easily lend itself to assimilation into the norms of written culture and the perceived need to arrive at a coherent narrative with an unambiguous message.
At a more practical level, ethnologists and local authorities found it difficult to record these Wu-language songs (*shange* 山歌, mountain songs or folk songs) in textual form because the Chinese character script has developed over the centuries to record primarily the language of the imperial court and the official examinations, as well as a standardized form of the vernacular used for entertainment literature. As a nonphonetic script, it was less accommodating of localized speech forms, including the Wu languages spoken in the lower Yangzi Delta. Song transcribers had to invent new orthographic conventions to record the words sung by singers in different parts of the delta.[^2] Chinese ethnographers labelled the folk songs of their region as Wu songs (*Wu ge* 吳 歌 ) in order to mark them as distinct from folk songs known elsewhere in China. The term “Wu” is redolent of historic associations. It refers to the ancient Wu kingdom of the delta (585–473 BCE), the ancestral founder, Wu Taibo 吴太伯, the Wu family of languages, and the people from the region. The reconstitution of the region’s *shange* into Wu songs was not simply a matter of conserving a traditional but now virtually defunct oral performance art for future generations. The issue was rather whether the oral culture of commoner farming communities of the immediate past could become a standard bearer of valued tradition in the delta region of the present day.
The discovery of lengthy song narratives comprising thousands of lines, together with the consequent publication of a large number of transcripts, is a significant event in the field of Chinese popular culture. These long narrative song-cycles from the lower Yangzi Delta (here called “folk epics”) are a new and largely unknown niche in China’s rich oral tradition.[^3] Unlike professional oral arts, such as storytelling and regional opera, which flourished in delta market towns and metropolises, the song narratives of villagers toiling in the paddy fields around Lake Tai circulated primarily in the scattered rural settlements hugging the shores of the water channels and major lake and river systems. While their zone of circulation was mostly limited to rural areas, the songs were enormously popular in their day. They offer unsurpassed insight into the everyday lives, supernatural beliefs, and folk ecologies of the farming populations of the Yangzi Delta in the immediate pre-contemporary period, and provide a rich storehouse of plots, themes, and motifs adapted by professional performers of operatic and storytelling genres and by writers and publishers of vernacular literature.
It was not until the early twentieth century that Chinese ethnologists began to collect the shorter form of Wu-language songs ([source:137]); the longer narrative forms remained largely undiscovered until the 1980s.[^4] The majority of Wu folk epics deal with events regarded by singers and their audience as “true stories” about actual people from the past. Many tales deal with the tragic fate of blighted lovers and focus on the social conflicts endemic to patriarchal society in pre-contemporary China. The ill-fated lovers contend with the disapproval of parents, a mismatch between the social classes, separation, sickness, persecution, and the wrath of local authorities. Their nights of “illicit” love are few; the stories linger rather on the stratagems of the lovers and the events that lead inevitably to the tragic end. Unlike the vernacular novels and stories about supernatural retribution that prevailed in the written culture of the last two dynasties, the Wu folk epics offer tales of lyrical intensity that, at least in their original state, do not seem to provide a definite “moral message” of the kind familiar to Chinese narrative traditions more generally. The most popular narratives contain powerful and moving song lyrics that were passed on by generations of communities bordering Lake Tai and Lake Fen.
Given the importance of the folk epics to delta populations and the relevance of these long song narratives to our understanding of oral-literate interactions in pre-contemporary China, the publication of volumes of Wu song-cycles in recent years is a welcome event. However, as with all attempts to capture in writing the elusive nature of an oral tradition, many questions remain. To what extent do these publications reflect the true state of these songs as originally sung in the paddy fields and along the waterways of the delta region? How were the original transcripts mediated in the process of editing and rearrangement? What compromises were made and what is the potential scholarly value of these transcripts? This study attempts to offer insights into the process of contemporary textualization of the Wu folk epics through detailed analysis of one well-known example, the song-cycle known today as “Fifth Daughter” (*Wu Guniang* 五姑娘). “Fifth Daughter” is regarded as one of the most outstanding of all the song-cycles discovered. Versions of the tale were sung all around the region of Lake Tai and Lake Fen and in the provinces of Jiangsu and Zhejiang. “Fifth Daughter” was one of the first of the lengthier songs to come to public attention and was one of the most contentious among scholars. Certain local folklorists still debate today whether “Fifth Daughter” is a genuine example of an orally-transmitted narrative or simply a number of separate short songs pieced together by contemporary folklorists to form an epic-length narrative. The controversial nature of the “Fifth Daughter” song-cycle and its significance as a now canonical example of Wu songs makes it a particularly revealing case study of the interaction between singer, ethnographer, and the Chinese state.
The collection of folk songs has long been an important cultural activity of the Chinese Communist Party. After the founding of the People’s Republic of China in 1949, professional writers under party guidance drew strongly from existing folk culture to rewrite and recreate folk songs, storytelling, and regional opera into new cultural forms that embodied the socialist agenda of the “new China.”[^5] However, during the “reform era” (that is, post 1978) a new, avowedly more “scientific” interest in cultural heritage superseded earlier attempts to eradicate “feudal poison” from popular performance culture. Nonetheless, as discussed here, political considerations remained to vex the next generation of ethnographers. This case study will thus illuminate the dilemmas of contemporary Chinese ethnographers and local authorities in attempting to preserve performance arts that once played a central part in community life but now face extinction in what has become one of the most affluent and economically advanced regions of China. As observed in this study, issues of preservation and publication are closely bound to the type of identity that regions choose to project in the competition for “ownership” of their regional heritage. Notions of what constitutes Han Chinese identity was a debated issue in the case of the long folk narratives of the delta. An important conclusion to this study is that the process of textualization of the Wu folk epics failed to take into account significant aspects of the spiritual beliefs of the farming populations of the delta. Here we highlight these purely regional understandings, often elided in constructions of national culture, which arguably formed the “mental texts” or implicit assumptions of the singers.
### Wu songs in the lake tai region ###
A thousand years ago, populations from north China fled war and disaster by migrating south to the marshy regions around the mouth of the Yangzi River. Here they intermarried with the indigenous population, known in ancient times as the Bai Yue 百越. Over the centuries, the original marshlands of the delta were laboriously converted into rice paddies, polders, embankments, and enclosed ponds, as a thriving agricultural economy was built around rice paddies, fish, and silk handicrafts.[^6] However, the same period was also characterized by the pressure of a growing population, deforestation, low holdings per capita, and a greater incidence of periods of drought and flooding ([source:133], 238–39; [source:119], 33). Until the modernization of recent decades, this region was “a land of rice and fish.” Even today in the regions bordering Lake Tai and Lake Fen, rivers and streams intersect each other in a chessboard pattern of interlocking waterways. Before the contemporary period, many townships and villages were almost completely surrounded by water. People travelled by boat or journeyed on foot along narrow paths bordering rivers or on narrow embankments above fields of rice paddies.
The Wu folk epics reflect the lives of the small landholders who eked out a precarious existence along the shores of Lake Tai and Lake Fen in the late nineteenth and early twentieth centuries. Usually all family members took part in the work in the fields. Women participated in the exhausting task of planting rice shoots, as well as threshing, weeding, raking, and operating the “dragon wheels” or pedal-operated irrigation wheel to water the rice paddy. These families also employed hired hands who went from home to home offering their services on a periodic basis.[^7] Women of the family would bring the midday meal to the men toiling in the fields. In “Fifth Daughter,” this is how the heroine, a young unmarried girl from a good family, happens to meet the hired hand, Xu Atian 徐阿天, who will become her lover.
As elsewhere in China, farming settlements were based around descent groups in the male line. Villages or groups of hamlets tended to have the same family name and identify as a kinship group. They formed tight-knit, self-governing communities. The family head was in charge of safeguarding the reputation of his family by enforcing a strict code of morality on the womenfolk in particular. In these remote water-locked villages, the operations of imperial governance appeared distant. For this reason it is only on rare occasions that one catches a glimpse of the institutions of imperial government in these folk epics. One example of this is the story about a doomed lover, Xu Atian, who is sent in chains to the Wujiang County *yamen* (office). A corrupt magistrate finds him guilty of a false charge of theft and condemns him to death. However, this song-cycle dwells primarily on the arbitrary power exercised by household elders and the brutal treatment they enact on their own flesh and blood. In the isolated settlements amid the wetlands of the delta region, one hears very little of the Confucian norms that, in theory, were meant to govern relations within the family and society at large.
In the original tradition, as reflected in original singer transcripts, the female protagonist is portrayed as the active initiator of the love-match. Original transcripts contain erotic segments depicting the young couple in the bath or in bed. It is precisely this that most offended gentry households in the pre-contemporary period and socialist authorities in more recent times. The story also contains a class aspect. In “Fifth Daughter,” a girl from the home of a small landholder falls in love with a hired hand employed by her older brother. This transgression of both social class and sexual norms can only end in catastrophe. Nonetheless, although it is the young couple who violate the social norms, it is the weak and gullible older brother and his jealous and brutal wife who are seen to be the real villains of this tale.
By the 1980s, an elderly woman called Lu Amei 陆阿妹 (1902–1986) was the only singer who could recall a relatively full version of the tale of “Fifth Daughter.” She came from the Fenyu Xiang region of Jiashan, on the southern banks of Lake Fen in modern-day Zhejiang Province. From a young age, her father, Sun Huatang 孙华棠, an acclaimed singer, taught her to sing mountain songs. By early adulthood, she became known for her talent at extempore performance. In the early 1950s she followed her husband north to Luxu, on the other side of Lake Fen. Lu Amei claimed to have first learned the song about Fifth Daughter from her father, who had in turn learned it from another singer of local fame, Yang Qichang 杨其昌, who lived during the Xianfeng era of the last dynasty (1851–1861).[^8]
On the basis of this claimed genealogy, it appeared that Lu Amei’s version has been in circulation in some form for around 150 years. The song transcripts sung by Lu Amei comprise some 2,800 lines, far more than any other singer of this song-cycle. During her lifetime, which bridged the republican era and the early decades of socialism, it was already difficult to sing songs about illicit love affairs. In fact, she became the butt of gossip precisely because of her skill in singing this tale:
> People would say, “When Lu Amei sang the tale of Fifth Daughter/All her children will be carried away by Fifth Daughter.” This startling rumor meant that Lu Amei as the singer was at the receiving end of innuendo and social pressure, to the point where fewer and fewer people dared to sing this tale and fewer and fewer people were familiar with the story.[^9]
> 有人谣传说‘陆阿妹唱仔五姑娘哉， 生的小倌都被五姑娘领走了’。这些骇人听闻
> 的流言， 使陆阿妹这样的歌手，也受着诋毁与压抑，以至后来， 传唱的人越来越少了，熟悉它的人 也越来越少了。
This statement implied that the spirit of Wu Guniang would punish those who sang of her scandalous life. Lu Amei herself became reluctant to continue singing “Fifth Daughter” as it was seen to lead to the premature deaths of her children. It was not until the early 1980s that, on the urgings of local folklorists, she agreed to sing her repertoire and allow recordings to be made. By this time she was an elderly woman, frail and somewhat deaf. Reports at that time indicate that she felt deeply embarrassed about some of the erotic material in the repertoire she had mastered as a young woman. This report comes from Qian Shunjuan, a female ethnographer who came to know Lu Amei in the 1980s.
> Lu Amei’s own thinking became gradually liberated. At first she would not sing the segment “The brother’s wife catches the couple in flagrante,” which includes some “chaff” [explicit material]. It was not until Zhang Fanglan, Ma Hanmin [male folklorists], and others came to know her as if they were truly part of her family that she would first go secretly into her room to bow before the image of Guanyin [the Buddhist deity of compassion], and beg forgiveness from the bodhisattva. Only then would she come out from her room and sing this segment. ([source:131], 22)
> 陆阿妹的思想也是逐步解放的， 她开头也没有唱‘五姑娘’长歌中的‘嫂捉奸’这样的包含有些糟粕成分的片断， 直到张舫澜， 马汉民等和她相处得真像自家人一样， 她还要先悄悄跑到里间屋拜过观世音， 请菩萨宽恕， 再出来唱。
In the early 1980s folklorists in the Jiangsu town of Luxu, on the banks of Lake Tai, collected songs relating to the song-cycle of Fifth Daughter from local singers such as Lu Amei, Zhang Yunlong, Zhao Yongming, and Jiang Liansheng. Their songs were recorded on tape, transcribed in manuscript form, and then printed off in a mimeographed edition of two volumes in private circulation. It is this text, produced in 1983 and known as the Wu Guniang ziliao ben (here termed “Transcripts;” see Jiangsu sheng minjian wenxue gongzuozhe xiehui: Suzhou fenhui Bian 1983, hereafter [source:122]), that provided the basis for all later printed editions. On a trip to Luxu in May 2011, the authors of this study were gifted with a copy of “Transcripts,” and will be relying on this text to better understand the song-cycle as sung by Lu Amei and the later process of revising and editing.[^10]
The first textualization of the song-cycle, “fifth daughter” The “Transcripts” comprise 87 separate songs of which 76 are by Lu Amei. Altogether there are 2,900 lines of sung material in 790 stanzas.[^11] In order to retain the original flavor, the compilers adapted Chinese characters to record the Wu language expressions, although some compromises were made in the case of obscure Luxu idioms.[^12] The songs were elicited by folklorists Zhang Fanglan, Ma Hanmin, and Lu Qun, who visited Lu Amei in her home between April 1981 and May 1982. It is evident from the dates on the transcripts of individual songs that Lu Amei did not sing the song-cycle in any particular “logical” narrative sequence. As the ethnographers asked her further questions about the plot she would call to mind further elements of the story. Given that she had a comprehensive mastery of the formulaic material of the Wu songs of her area, and well-developed views on the characterization of the central protagonists, it is also possible that she came up with new story elements to meet the demands of her scholarly audience for a fuller account of the tale. This would be in keeping with the usual conventions of epic singers who have mastered the repertoire and are able to compose the performance to meet the needs of the audience ([source:118], 338–39). One can thus assume that the Wu Guniang song-cycle sung by Lu Amei contained both material inherited and adapted from previous singers as well as material she composed extempore to meet the requests of the folklore scholars.
The “Transcripts” additionally contains a lengthy interview with Lu Amei where she explains in prose the story material of “Fifth Daughter” in her own understanding ([source:122], 257–74). Lu felt a deep sympathy for the tragedy that enveloped the central characters and a strong reluctance to “conclude” the tale with what her listening folklorists regarded as a satisfactory plot resolution. She told the folklorists in April 1981:
> You can never sing to the end of “Fifth Daughter.” Even after ten nights and ten evenings you still can’t sing right to the end. If you do wind up singing to the very end then you’ll suffer misfortune. Fifth Daughter doesn’t want people to tell her story because then she’ll lose face. If you don’t get chest pains then your belly will hurt. People who continue to sing this song are becoming fewer and fewer. I don’t dare sing it rashly.[^13]
> ‘唱无尽格五姑娘，’ 十夜十黄昏也唱勿完， 先底于说唱完要触霉头格， 五姑娘勿肯人家郃伊格事体才讲出来，坍台格， 侵寻， 吼事（坏事）， 勿是心口疼， 就是要肚皮痛， 唱格人 越来越少了， 俉也不敢瞎唱格。
Selecting from the independent song segments in “Transcripts,” Zhang Fanglan, Ma Hanmin, and Lu Qun arrived at what is now known as “the first arranged edition.” According to Zhang’s account, it took the three men three months to complete the task. The three of them divided up the allocated material, with Zhang providing the beginning, Lu Qun the middle section, and Ma Hanmin the ending. Zhang was then responsible for combining the three sections so that the text read as a coherent whole. This was not an easy task as the song excerpts were not necessarily mutually consistent. The singer would occasionally go off on tangents, providing more material that never came to a conclusive ending. The team of folklorists took it upon themselves to devise a “logical” sequence for the story, to form narrative linkages between individual songs, and to provide a “satisfactory” ending. They were pressed for time. The main goal of the folklorists was to exhibit the artistic qualities of “Fifth Daughter” in time for a conference on Wu songs run by the local cultural authorities in December 1981. Their arranged text of “Fifth Daughter” was presented at the conference and subsequently published in the literary journal Zhongshan in 1982. The text was later published as a separate monograph by the publisher Jiangsu Renmin Chuban.[^14]
In interviews with the authors of this study in May 2011 and with Emily Zhang in March 2012, Zhang Fanglan explained the complex local politics of the time and how this influenced decisions made about how to edit the transcripts in order to provide a readable edition. Further correspondence from the early 1980s also lays bare the dilemmas of the ethnographers, publishers, and authorities. Below we discuss in turn three major issues that key players struggled with in the politics of the early reform era. The first problem was the “anomalous” discovery of lengthy song narratives belonging to a Han Chinese people as distinct from non-Chinese minority groups. The second issue was controversy over the authenticity of “Fifth Daughter” as a long song-cycle. The final issue was how to evaluate the “message” of this story of transgressive love and to render it into a form appropriate for a broad readership both within the delta and the nation.
### Did Han chinese populations have epic song traditions? ###
Chinese theorists, influenced by Marxist conceptions of the progress of history, held that Han Chinese communities had not produced long narrative songs of epic length. Qian Shunjuan is one of many local ethnographers who reported that the discovery of the long version of “Fifth Daughter” in 1981 “smashed the traditional view that the Han Chinese people have no long narrative poetry” ([source:131], 3). The notion that epic poetry belonged to a specific pre-civilizational age was first advanced by Lewis H. Morgan (1818–1881). Morgan proposed three stages of human development, from savagery to barbarism and, finally, civilization (Morgan 1985). Each stage was associated with a more advanced use of technology and more sophisticated modes of living. The second stage, barbarism, was said to include the emergence of poetry and mythology, analogous with the Greek Homeric age. Frederick Engels (1820–1895) took up Morgan’s general analysis, which was then adopted by Chinese intellectuals in the twentieth century and became a fixture in later Marxist historiography.[^15] In this line of thinking, epic songs were the provenance of “primitive” non-Han Chinese minority groups residing in the borderlands. Famous examples include the Tibetan epic “Gesar,” the Mongolian epic “Jangar,” and the Miao creation epic “Miluotuo.” In other words, in the official discourse of the history of Chinese civilization, orally-transmitted epics belonged to the high tradition of nonliterate peoples. The Hua-Xia people, on the other hand, who developed into the group known today as Han Chinese, invented a literate culture at least three millennia ago. So, in this line of interpretation, the proto-Chinese population entered advanced civilization at an earlier date than the borderland tribes and had no need for an epic song tradition.[^16] Closely related to this debate was the controversy over whether “Fifth Daughter” was really an epic length song narrative. To some it appeared to be rather the invention of enthusiastic ethnographers who had allegedly pieced together fragmentary song material to greatly extend the length and coherence of the original song. As we will discuss below, this controversy was ultimately resolved in favor of a consensus that songs of epic length were indeed indigenous to the Wu-speaking populations of the Yangzi Delta, even if this put them on a par with the “primitive” borderland populations. In the changed conditions of the 1990s it now became possible, indeed even a matter of pride, that the Han people had their own epic song tradition.
### The authenticity of “fifth daughter” as a long song narrative ###
In 1981 the Luxu folklorists, Zhang, Han, and Lu, did their best to come up with a rendition of “Fifth Daughter” that met the needs of the time. As the first of the Wu long song-cycles to be recorded and published in a national publication, “Fifth Daughter” caused a minor literary sensation. Leading folklore scholars praised the lyric beauty of the songs and the heartfelt message of the suffering of the commoners in “the old society.” But beyond Luxu, on the southern shores of Lake Fen, in the Jiashan region of northern Zhejiang Province, where Lu Amei had been born and raised, folklorists in her home region raised severe objections to the claim that the bundle of songs now associated with “Fifth Daughter” did in fact reflect an authentic tradition of an orally-transmitted song-cycle. These critics pointed out that in the Jiashan tradition of folk songs, there was indeed a narrative about “Fifth Daughter,” but it was known only as a short song structured around the annual calendar (a poetic form known as Twelve Month Flower Names 十二月花名), comprising around one hundred lines. According to local legend, this song was first composed by the tailor who made a shroud for Fifth Daughter. He sang it as he went from door to door and in this way the song came into circulation ([source:136], 79). Jiashan singers interviewed in the early 1980s said they could not possibly remember a song two thousand lines long like the one produced from the transcripts of Lu Amei and claimed the longer story was not known locally ([source:115], 328). The short song version was, moreover, sung across a much wider geographic region in places such as Pinghu, Jiaxing, and Qingpu. Further, a lot of the song material in Lu Amei’s rendition belonged to Wu songs of the area more generally and could have simply been added in by the singer to elaborate her song. Some critics went so far as to claim that the long version of Lu Amei published in Zhongshan had been pieced together from separate songs and substantially elaborated and revised by the folklorist editors in Luxu. In other words, it should not be considered an authentic example of a long narrative song (*changpian xushi ge* 长篇叙事歌) ([source:115], 328).
The Folk Culture Association of Suzhou, a major city in the delta, was keen to get at the truth of the matter and carried out an extensive investigation through visits to Wujiang, Qingpu, and Jiashan to meet and interview folk singers. When they listened to Lu Amei sing “Fifth Daughter” they were won over by the strength of her performance and concluded that the existence of a long, orally-transmitted song was indeed plausible. The emergence of an abundance of other long verse narratives in the delta region (close to forty) also convinced them of the veracity of the claim for the existence of “Fifth Daughter” in the long version.[^17] Duan Baolin, a professor at Beijing University and Chairman of the Folk Arts Association, had been told that “Fifth Daughter” was a fraud and called for a copy of “Transcripts.” Jin Xu subsequently sent off the mimeographed volumes, which then circulated among leading folklore scholars ([source:115], 332). After a period of discussion, senior Chinese scholars were won over to the view that “Fifth Daughter” existed in the long version, at least in the region of Luxu, where several singers could still sing this song-cycle in the early 1980s. Luxu was also the site of origin of numerous other long narrative Wu songs and historically had been a center of mountain song competitions in public forums. It was also known for its mountain song teams (*shange ban* 山歌班), comprising peasants with good voices who were hired by landlords to entertain the laborers toiling in the fields.[^18] In addition, the volumes of “Transcripts” at least provided sufficient proof of the prevalence of the long version in the Luxu region. In this way a scholarly consensus was reached that, at least in the rendition of Lu Amei, “Fifth Daughter” could be considered a genuinely long narrative song-cycle.[^19]
### Controversy over the textualization of “Fifth Daughter” ###
Although the authenticity of the long version of “Fifth Daughter” was more or less resolved, there still remained debate about the quality of the first published text by the Luxu ethnographers. This text became known as “the first arranged (*zhengli* 整理) edition.” The main reason was the perceived deficiencies of this version. Even scholars who agreed with the existence of the long narrative song held that the Zhongshan version had been edited and revised in such a way as to cast doubt on its claim to represent an oral text.[^20] Subsequently, Suzhou scholar Jin Xu produced what is now known as “the second arranged edition” (*zai zhengli ben* 再整理本). The second arranged edition comprised 1,700 lines in 431 stanzas, about a quarter less than the first arranged edition. Care was taken to include original Wu-language expressions with annotations to make the text accessible to a national readership.[^21] It was this edition that was included, with some pruning, in the important collection by Jiang (1989). Many years later, lengthy selections from “Transcripts” were published in another leading compendium of Wu songs ([source:115], 196–310). The latter included most of the material in the original “Transcripts” but left out some erotic material. Each song was listed separately and no attempt was made to knit the separate songs together as one finds in the first and second arranged editions. The disparities between these different versions of the song-cycle were now apparent to the scholarly observer who had access to all four versions.
The problem faced by all the editors and arrangers was how to present orally-transmitted song segments in a readable form for the purpose of scholarly preservation and also for the promotion of the longer Wu songs as a regional treasure. This involved more than just coming up with a coherent storyline. The editors had to work out how to evaluate the central protagonists, the true nature of the central conflict of the story, and how to resolve the conflict in a way that would be well received by both the political authorities and regional communities. In other words, a moral message was required. It seemed clear to these folklorists that “Fifth Daughter,” in common with the other *siqing* 私情 type of narratives (tales of illicit love), conveyed a message about the evils of arranged marriage under the “feudal” system of pre-contemporary China. Virtually all Chinese commentators justify the publication of these narratives on the grounds of their denunciation of “anti-feudal marriage.”[^22] However, the original material also contained some troublesome scenes, such as the songs where the heroine took the initiative in seducing the hired hand, scenes of lovemaking that were deemed too graphic for publication, and, most troubling, an undeveloped plot resolution where the villains of the piece, the older brother and his unsavory wife, were not seen to be thoroughly punished for their evil deeds.
A comparison with “Transcripts” shows the scope of changes made by the editors of the first arranged edition. The 790 stanzas in “Transcripts” were compressed to only 583 stanzas. This edition also contained new material not found in “Transcripts”, although the latter claimed to be the original recordings of the singers. In other words, the editors made insertions of their own, particularly to the ending. Song segments indicating what appeared to be alternative plot lines were excluded from the arranged edition. In other words, whereas “Transcripts” included the potential for multiple plot developments, even including elements that were mutually contradictory, the editors of the arranged text chose material from among disparate story material to form a single cohesive plot line.
One major point of difference lies in the treatment of the villainous sister-in-law. As recorded in “Transcripts,” Lu Amei provides a much lengthier treatment of this character, including her reputation for promiscuity before marriage to Yang Jinda, her dominance of her hen-pecked husband, her interest in extramarital liaisons, and her cruel treatment of the two sisters.[^23] Only a small portion of this material was included in later editions. Lu Amei’s treatment of this theme opened up the potential for a plot line not explored by later editors. For example, in one song segment, Fifth Daughter voluntarily tells her sister-in-law that Atian has sent her a love token, presumably to seek her permission for her own affair. The sister-in-law urges her to accept and then spells out her own designs on Atian. This sets up scope to develop a consenting relationship between one man and two women, a not uncommon plot type in Wu songs.[^24] Polygamy was a feature of pre-contemporary Chinese marriages, but a relationship between an unmarried man and two women who were sisters-in-law was regarded as abhorrent not least because it upset the hierarchies of seniority between the wife of the older brother (*saosao*) and his younger sister (*meimei*). In this case class was involved as well because the husband was a landowner and the offending male a hired hand.
Wu songs commonly treat the subversion of traditional kinship hierarchies. For example, a married man taking on the younger unmarried sister of his principal wife was a favored theme of several Wu song-cycles, and the phenomenon was not uncommon in historical practice.[^25] In Wu songs dealing with a polygamous relationship, the focus of interest is on the competition between two women, one younger than the other, for the attention of the male. In the song sung by Lu Amei this conflict is expressed in terms of a choice between two types of common foods: “Of the wild flowers daubed with dew, only those fresh in bud are good to eat/But with sugarcane from Tangxi, it is the mature stalks that are sweet.” The final couplet resolves the issue: “Atian rejected the mature stalk on the sugarcane/ He only wanted the red caltrop with the fragrant fresh stamen.”[^26] This song was not included in later print renditions, perhaps because it placed in jeopardy the perceived moral virtue of Fifth Daughter.
Another major point of difference is the plot resolution, particularly the issue of whether the villains are thoroughly punished. The songs recorded in “Transcripts” present little information about what happens to the major characters. One song by Lu Amei relates a story about the execution of Xu Atian by the local magistrate.[^27] A brief four-line stanza further relates that the Yang homestead was burnt to the ground.[^28] Finding this unsatisfactory, the editors of the arranged versions set up a version of the story that provided a more elaborate ending and punishment for the evildoers. The editors of the first arranged edition inserted sections that pointed the blame at the magistrate for condemning Atian to death and presented a tragic ending where both ultimately die. The editors of the second arranged edition set up an ending where the family accidentally sets fire to their own residence and the villainous older brother and his wife die in the flames. In other words, both sets of editors, finding “Transcripts” too skimpy and contradictory, borrowed from the tradition to craft an “appropriate” ending that punished the evildoers. From the song segments in “Transcripts,” it appears that a singer could choose from one or another version as the occasion demanded, which accounts for numerous contradictions in the songs of Lu Amei.[^29] In addition, there were personal reasons why Lu Amei was reluctant to provide an elaborate ending.
The open-ended nature of the original song-cycle reflects local beliefs about the powers of the spirits of those who die a violent death. Lu Amei told folklorists that it was unlucky to completely finish off the tale of Fifth Daughter because she did not come to a good end (*bu guangcai* 不光彩). The singer did not want to linger over the final events of her life in case the spirit of Fifth Daughter was offended and she herself met with ill fortune. In contrast, the editors of the later editions played down the role of the sister-in-law in bringing about the catastrophe, preferring to blame instead the patriarchal powers of the older brother and the venality of the legal system in the imperial era. In this way they could frame “Fifth Daughter” as less a story about an illicit love affair than an attack on “feudal marriage” in the “old society.” However, in so doing, they distorted the understandings of the singers and their audiences, who preferred to frame disaster in terms of heavenly destiny, human frailties, and, as we shall discuss later, the need to appease the spirit of Fifth Daughter.
### The mental text of the singer ###
The numerous discrepancies between “Transcripts,” representing the songs of actual singers at specific sessions, and the arranged editions compiled by folklorists, are readily explicable when one considers the differing nature of the oral and written epic form. These differences are well known in the history of the recording of oral epics and their later reshaping to form a literary tradition.
Here we will draw on the studies of Lauri Honko, noted for his studies on the textualization of epic forms. Honko stresses the importance of the creativity of particular individuals in the formation of long narratives. These can be illiterate singers who have outstanding ability to compose songs extempore and also literate editors immersed in the tradition who are at once scribes and creators capable of composing epic-length songs themselves. One example of the former is the illiterate singer, Anne Vebarna, who when provided with a “schematic plot” based on folklore was able to compose extempore a narrative song called “Peko” that came to be regarded as the national epic of her people, the Setu in Estonia ([source:118], 338). In another example of her versatility, Vebarna provided two radically different versions of a single story, one ending happily, and the other tragically ([source:118], 337). In the case of Lu Amei, she provided a suite of songs on the theme of “Fifth Daughter” for over a year to the folklorists who visited her. While many were indeed recalled from past performances and can be found in the songs of other singers, some may have been composed afresh based on her own imaginative adaptation of songs in that tradition. This does not render them less “authentic” than the songs she shares with other singers. Rather, one can see these songs as creative elaborations that demonstrate the potential of the repertoire to produce songs of epic length and power if circumstances permit ([source:116], 22–23).
The case of the folklorist editors can also be considered as another act of creation, this time away from the oral epic to what Honko calls the written “traditional-oriented epic” ([source:116], 7). Honko defines the latter as those “tradition-bound epics which have been molded, if not created, in the hands of performers, scribes, and editors. They possess ‘anterior speech’ in the form of oral epic registers internalized not only by their performers but also by their scribes and editors, but they are not direct documents from oral performances” ([source:116], 7). The Finnish epic, Kalevala, is of particular relevance here. The written epic known today is largely the work of one man, Eric Lönrott (1802–1884), who collected narrative songs from eastern Finland, near the border with Russian Karelia, a region known for its epic poetry. Lönrott became immersed in the poetic tradition to the point where he was able to subordinate the collected material to his own internal vision and create what became an original folk epic that resonated with the Finnish public. As Honko puts it, Lönrott had a self-developed “mental text” that guided his own creation of the Kalevala. This “mental text” comprised “the storyline of the narrative in question, standard descriptions of events, repeatable expressions, phrases, and formulas familiar from the performances of other singers” ([source:117], 14). While the “mental text” contains many repeatable stereotypical elements shared by singers within a tradition, Honko is insistent that this “multiform” material “must be ordered by a single mind” to create a successful long epic ([source:117], 15).
Honko’s idea of a “mental text” that must pass through the mind of an individual singer is relevant to both the oral composition of Lu Amei, who expanded on the tradition passed down to her to create her own unique version of a familiar narrative, and also to the reworkings of her recorded song-cycle by successive local ethnographers, who edited her text along the lines of their own particular “mental texts” of the narrative as a whole. In creating the Kalevala, Lönrott expunged obscenities, smoothed out localized expressions that would puzzle his readers, and filled in obvious plot gaps ([source:117], 18). Similarly, the folklorists discussed here, Zhang Fanglan, Ma Hanmin, Lu Qun, and Jin Xu, exercised their own judgment in reshaping the material from the original singers to make a text that would pass muster with the cultural authorities and be deemed suitable for a national readership. It is the first edition that contains the most deviance from the original set of transcripts. This is no doubt due to the political exigencies of the time and reflects the difficulties of establishing that Han Chinese communities did indeed have narrative song traditions of epic length, and that these songs had “progressive” content that should be widely celebrated. The editors of the second edition were more aware of the pitfalls of an overly-enthusiastic adaptation. To avoid criticism of a lack of authenticity, they sought to return the written form of “Fifth Daughter” to one that was more firmly based on the original songs. Nonetheless, one could argue that both editions remain at some distance from the “mental text” that clearly existed in the mind of Lu Amei, with all its complexity and contradictions. Of crucial importance is the “mental text” of the singer, Lu Amei, specifically the set of popular beliefs that lay behind her construction of the story of “Fifth Daughter.” Our sources here are the song transcripts attributed to her and the lengthy oral interviews recorded in “Transcripts.” There were several aspects that informed her “mental text”: first, her belief in the historicity of the tale of Fifth Daughter and Xu Atian, and in her ongoing existence as an aggrieved spirit; second, her belief in the efficacy of mountain songs in mediating between the world of the living and the dead; and third, the way she associated a purely local incident with one of the master narratives told all over China.
Lu Amei and other Lake Fen locals strongly believed in the historical veracity of the story. Lu claimed: “When I was a child I heard the story from my grandmother and I myself have seen Fifth Daughter’s combs and furniture.”[^30] Locals at Fenyu in Jiashan held that Fifth Daughter grew up in Fangjiabang on the shores of Lake Fen in the home of Yang Jinda. Descendants of the Yang family had continued to carry out ancestral rites using spirit tablets representing deceased family members in a side room of the original residence until the early 1950s. The Yang residence was eventually demolished in 1959 but some remnants could still be seen in 1981, including a boat shed, the ox herder’s hut, and stone benches. Xu Atian was believed to have come from Yao’an cun, about 1.5 kilometers north of Fangjiabang. The original Xu family residence was not demolished until the 1980s. Descendants of the Xu family believed that Atian did indeed steal the spirit tablets of Fifth Daughter from the Yang residence and set them up in his own home. They claimed that Atian had his legs broken by the Yang family and ended his life as a beggar. His spirit tablet had been venerated by the family until the political movements of 1959.[^31]
The belief in the historicity of the tale, the existence of sites relating to the couple for a period of over a hundred years, and the continuation of ancestral rites of veneration by descendants would have conditioned the reception of the song-cycle surrounding Fifth Daughter and Xu Atian and surely contributed to its longevity from the mid-nineteenth century to the present. This song, like many others in the Wu song repertoire, deals with the premature and brutal death or disfigurement of one or both young lovers. In the context of delta life in the nineteenth century, scandals of this sort were the common substance of gossip and storytelling ([source:129]). However, premature and violent deaths were also a threat to the community because of the wandering souls of the aggrieved dead. Other folk epic traditions find their inspiration in striking acts of violence that disturb the social equilibrium. In Indian folk epics, for example, the spirits of those who died unnaturally are feared as a likely cause of misfortune. The community engages in ritual acts of propitiation, which in turn spurs the development of song and story-cycles ([source:124], 108–12).
Some claim that there is evidence of a cult to a woman whose situation paralleled that of Fifth Daughter close to the Lake Tai region in the early nineteenth century. In the neighboring region of Jinze in Qingpu a temple was erected to a woman from Luxu called Chen Third Daughter. She was said to have engaged in a secret love affair, and when this was exposed her father had her drowned in the river. When her spirit manifested itself after death, locals erected a statue of her, only to see it destroyed by a Qingpu official in 1826 ([source:131], 18–19). Locals tend to link the story of Chen Third Daughter with the song-cycle of Fifth Daughter. While this cannot be attested, the existence of cults focusing on those who died with grievances is likely to have added further strength to the tale of “Fifth Daughter.” These folk beliefs retained their persuasive power well into the twentieth century and help to account for the lack of an explicit ending of the song-cycle. Lu Amei herself felt a strong reluctance to “conclude” the tale with what the Chinese folklorists regarded as a satisfactory plot resolution. As mentioned previously, she told the folklorists she did not wish to offend the spirit of Fifth Daughter: “Fifth Daughter doesn’t want people to tell her story because then she’ll lose face.”
While Lu Amei apparently tried hard to appease the spirit of Fifth Daughter by according her dignity through her singing, some of her neighbors believed she had indeed suffered adverse consequences due to her involvement in this story-cycle. At least one woman interviewed believed that unfortunate events in the life of Lu Amei could be attributed to her singing the tale of Fifth Daughter. In an interview in April 1982 a neighbor, Zhou Jiama, attributed the early deaths of Lu Amei’s daughters to her involvement in transmitting the song-cycle of Fifth Daughter:
If you sing of the scandalous events about Fifth Daughter she will be angry. At worst she could make you sleep for over a score of days, or at best, you could be sick in your heart. Of those who listen to the story of Fifth Daughter right to the end, there’s not a single one who doesn’t get heart pain. Why is that? It’s because Fifth Daughter’s angry qi 气 [life force] blows into the hearts of those listening to mountain songs.… Lu Amei raised seven girls, but all were taken away by Fifth Daughter. Amei now prays to the Buddha because she fears Fifth Daughter will cause trouble.[^32]
As well as trying to accord “face” to the spirit of Fifth Daughter, another way that Lu Amei constructed her own version of the tale was through the lens of a famous narrative about the ill-fated lovers, Liang Shanbo 梁 山 伯 and Zhu Yingtai 祝英台, which enjoyed national as distinct from regional circulation. In one of Lu’s songs, Fifth Daughter declares that she is just like Zhu Yingtai, who fell in love with the handsome Shanbo but whose affair with him ended in tragedy.[^33] The association made here between Fifth Daughter and Zhu Yingtai allowed Lu Amei to relate a local scandal about a sexual transgression to the much-loved figure of Zhu Yingtai, who sacrificed her own life for that of her lover. It is also an example of how a tale known only to a particular region gathers strength and resonance from association with a national master narrative.
### Conclusion ###
In the early twenty-first century the Wu songs, including “Fifth Daughter,” were accorded national status as important items of cultural heritage. In 2004 China joined the UNESCO Convention for the Safeguarding of Intangible Cultural Heritage. In the following year the city of Suzhou won approval for Wu songs to be added to the Chinese national register ([source:128]). The township of Luxu has set up a Wu Song Pavilion for the exhibition of Wu song artifacts and the teaching and performance of Wu songs (see figure 1). In some schools, Wu songs have been added to the local curriculum. Yet, in this highly developed region of China, very few young people know of the rich cultural heritage of their farming forebears along the waterways of Lake Tai. The printed versions of the folk epics remain to be passed on to the next generation as examples of the oral traditions of the recent past. However, as discussed here, the true spirit of the original song-cycle can only be found after an intensive exploration of what the singers actually sang and the “mental text” that was their inspiration. This study thus delineates the limitations of the textualization of this particular song-cycle but more importantly points to the need for a detailed understanding of the social and political context shaping such “folk” re-creations in the contemporary period. In this case, controversy over what constitutes Han Chinese ethnicity played a role in shaping debate over the recognition of the song traditions of the region. Essentially, notions of Han Chinese as the “advanced civilization” gave way to a more nuanced assessment of the different ethnic and regional modalities that underlie the supposed unitary basis of the vast population designated Han Chinese. This is the major achievement of the local ethnographers and scholars of the region. In textual form, the Wu folk epics are the product of the efforts by local folklorists, in negotiation with the Chinese state, to transform a previously stigmatized regional genre into a cultural form worthy of national recognition as a cherished part of “Chinese” civilization.
- Anne McLaren wishes to express her sincere appreciation to the Chiang Ching-kuo Foundation for International Scholarly Exchange for providing fieldwork and research assistance funding for this project (“The Cultural Heritage of the Lower Yangzi Delta”). Both authors wish to thank Chen Qinjian of East China Normal University for his inspiring influence on our projects in Chinese popular culture over many years. We also wish to express our gratitude to members of the Culture Bureau of Luxu, Jiangsu Province, and members of the Cultural Center, Fenxiang village, Dingzha, Zhejiang Province, for their assistance with this project. Zhang Fanglan of Luxu provided us with his enthusiastic assistance during several visits to Luxu.
### Notes ###
[^1]: Traditionally, northerners were predominantly cultivators of grain and southerners cultivators of rice. Linguistic and cultural diversity is much more pronounced south of the Yangzi River than in the north. It is believed that southern Chinese populations have assimilated many of the customs and cultural forms of the originally non-Chinese peoples who resided south of the Yangzi in antiquity. This could potentially include song traditions ([source:113]). Archaeological finds in recent decades have proven the existence of multiple sites for originary proto-Chinese civilizations, including some predynastic civilizations in the lower Yangzi Delta. As Edward Friedman notes, these new findings challenge conventional understandings of a northern originary site of Chinese culture and call for the inclusion of what he calls a “southern narrative” in national reconstructions ([source:114], 42).
[^2]: This is a similar problem to that faced by non-Mandarin communities of Chinese origin residing outside China. Shu-mei Shih notes the complexities of the “multiple orthographies” required to write down Sinophone literatures and argues that “Sinitic [written] languages … therefore pose important challenges to the fictive construction of Han ethnolinguistic homogeneity” ([source:134], 10). A famous early ethnologist, Gu Jiegang (1893–1980), writing in 1920, pointed out the difficulty of recording expressions in the Wu language: “There are truly so many words that one cannot explain or even write down and so many things one cannot verify” (preface to _Wuyu jilu_, republished in [source:137], 679). In order to publish his collections of Wu songs it was necessary to draw on unconventional vernacular forms in local circulation and provide glosses of local usages. The same dilemmas confronted delta folklorists in compiling the song collections of the 1980s.
[^3]: We use the term “epic” here simply to refer to the narrative quality and length of these songs. The term should not be confused with epics on martial themes, such as Greek epics, or with the Chinese translation of “epic” as “poem on a historical theme” (*shishi* 史诗). Here we follow the broader definition of the term “epic” as defined by Blackburn et al. in the case of Indian village epics, where the term “heroic” is expanded “to include female and non-martial heroes” ([source:111], 5). Blackburn et al. use the term “romantic” to refer to the latter type of oral epic. The romantic epics deal with sexual relations that form a threat to the social order. They suggest that this “romantic” type of oral epic should be considered as “a cultural variant of an international genre” ([source:111], 5).
[^4]: The only substantial Western study of delta folk songs is by the ethnomusicologist Antoinette Schimmelpenninck ([source:132]). This work is based on recordings of songs conducted during fieldwork carried out between 1987 and 1992. It does not deal with the longer song-cycles discussed here although their existence is noted.
[^5]: This phenomenon began in “liberated areas” well before 1949 but extended over the whole country in the early decades of socialist rule. An example much studied in the West is the tale of Liu Sanjie ([source:126]).
[^6]: For studies of the history and rural economy of the Yangzi Delta, see [source:133], [source:119], and [source:125]. Li stresses the interconnected nature of Lake Tai and its tributary systems and the need for constant waterworks to prevent drought and floods ([source:125], 447–62).
[^7]: Huang believes that around one third of the peasant population was engaged in some form of hired work from the fourteenth to the mid-twentieth centuries ([source:119], 9).
[^8]: See “Transcripts,” vol. 1, 1 ([source:122]). For a fuller biography of Lu Amei see ([source:131], 14–38; 147–57).
[^9]: See the postface by Jin Xu ([source:123], 206).
[^10]: We would like to thank Zhang Fanglan for presenting us with this valuable set of transcripts, for his hospitality in inviting us to his home, and in taking us on a tour of the Wu Song Pavilion (*Wuge guan*) in Luxu. We are greatly indebted to him for his frank discussion about editorial and publishing decisions made in the 1980s.
[^11]: For details on the performative aspects of Wu Songs see [source:132] ; [source:141], 92–100; [source:123], 4–5.
[^12]: If no standard Chinese characters were available for an expression they would use homophonic characters based on the sounds of Mandarin; see “Transcripts,” vol. 1, 6–12 ([source:122]). Jin Xu notes that in some cases the transcribers substituted more standardized forms of the Wu regional language for obscure Luxu expressions; see postface in [source:123] (207). This was done for the sake of reader comprehension.
[^13]: “Transcripts,” vol. 2, 257 ([source:122]).
[^14]: It was also republished later in Jin (2004) on the grounds that while “flawed” it was of major historical significance. For these publication details see [source:123] (157).
[^15]: For a discussion of the introduction of European notions of epic into China see [source:139]. He notes that the perceived lack of European-style epic poetry in Chinese (actually Han Chinese) literature was seen as a sign of inferiority by famed intellectuals such as Liang Qichao (1873–1929), Wang Guowei (1877–1927), and Hu Shi (1891–1962); see [source:139] (132–33). Chinese scholars today accept the statement by Karl Marx that the epic form belongs to a less developed stage of the evolution of human society (according to [source:139], 136). In his reappraisal of the literature, Yin seeks to extend Western epic paradigms to include a wider range of themes beyond the heroic. Nonetheless, he limits his discussion of epics in China to the non-Han minority groups and does not include the long narrative songs discussed here.
[^16]: In a parallel discussion, De has critiqued the adoption of the Morgan-Engels framework in Chinese scholarly understandings of kinship and marriage. He notes that in the attempt to arrive at “an objective, universal, linear development” for the Chinese race ([source:112], 560), Chinese scholars seek out signs of ancient primitivity in the marriage and kinship customs of contemporary non-Han minorities, which are thus deemed to be antecedent cultures to the dominant Han ([source:112], 575).
[^17]: Zheng ([source:141], 81–84) lists thirty-nine separate long narrative songs in his important study of the delta folk epic. Besides “Fifth Daughter,” the Luxu region is known for a number of other lengthy song-cycles such as “Zhao Shengguan.”
[^18]: On *shange ban* and public song competitions see Zheng ([source:141], 140–61). For Luxu song competitions see Jin ([source:123], 6–7). Singers who participated in these events acquired extensive skills in extempore composition. The prevalence of both song competitions and mountain song teams in Luxu greatly stimulated the development of the longer form of *shange*. By contrast, Jiashan was not known for its song competitions or singing teams and did not develop long narrative songs.
[^19]: However, there are still cadres who express scepticism about the existence of a long version of “Fifth Daughter,” as we found out on a visit to Taozhuang Culture Station on 26 May 2011 (interview with Zhang Songlin).
[^20]: Not everyone was in favor of the bowdlerized version. Wu Gongjian ([source:138]) condemned the excision of erotic material, arguing that the original promoted a type of “modern” love based on personal choice. See also Qian’s allusions to the inserted ending found in the first arranged edition ([source:131], 36).
[^21]: See the editorial notes on this tale ([source:121], 265–66).
[^22]: For an example see Jin, postface ([source:123], 206). He further explains that although the erotic sections were not considered to be as bad as that of the Jin Ping Mei (the famous erotic novel of the Ming period), nonetheless, it was difficult [impolitic] to include the more explicit material ([source:123], 206).
[^23]: This is narrated in separate song segments such as “The matchmaker pays a visit” and so on; see “Transcripts,” vol. 1, 22–31 ([source:122]). One finds three times more material on this topic in “Transcripts” than in the edited versions.
[^24]: Lu Amei spoke about this aspect of the story in an interview: “It is not that Xu Atian sought to have an affair with her, it was Yang Jinda’s wife [the evil sister-in-law] who wanted to have an affair with him.” See “Transcripts,” 258 ([source:122].
[^25]: This song type is categorized as “Going to get the younger sister” jie ayi
接阿姨 ; see [source:127].
[^26]: “The sister and the saosao
[older brother’s wife] form an amour with the one lover” 姑嫂结识一个郎; “Transcripts,” vol. 1, 123–24 ([source:122]. Not included in other editions.
[^27]: See “Transcripts,” vol. 2, 223 ([source:122].
[^28]: This is also by Lu Amei; see “Transcripts,” vol. 2, 224 ([source:122]). A song by Qian Afu included in “Transcripts” relates the whole story outline in around 113 lines. It presents a “happy” ending where the magistrate punishes the brother and his wife and allows the young couple to marry; see “Transcripts,” vol. 2, 231–41 ([source:122]). This highlights the flexibility of the tradition and the distinctiveness of the version presented by Lu Amei.
[^29]: Lu Amei sang a song about Xu Atian’s execution shortly after the affair was discovered. However, she also sang another song about how Atian comes to steal the spirit tablets of Fifth Daughter after her death, which means that he was still alive. In a “readerly” text these endings appear contradictory.
[^30]: Interview with Lu Amei in 1981; see “Transcripts,” vol. 2, 257 ([source:122]).
[^31]: For reports on these sites see [source:136], 78; [source:131] (17).
[^32]: “Transcripts,” vol. 2, 271 ([source:122]).
[^33]: “Transcripts,” vol. 2, 166–67 ([source:122]). Liang and Zhu were also held to be historical figures and reported grave sites can be found in various regions of China. One such site is at Ningbo, in coastal Zhejiang, within the Wu-speaking zone. For a study of this story-cycle over the centuries see [source:120].