åè¬ä¼¯éæ°ï¼ã女çºãå¨åãããå¬åãç£ä¹ï¼äººèä¸çºãå¨åãããå¬åãï¼å ¶ç¶æ£çé¢èç«ä¹èï¼ã
The Master said to Bó Yú, âHave you applied yourself to the âZhÅu Nánâ and the âShà o Nánâ? Someone who fails to apply themselves to the âZhÅu Nánâ and the âShà o Nánâ is like someone standing facing a wall, is it not so?â
In early 2015, ÄnhuÄ« Dà xué å®å¾½å¤§å¸, the University of ÄnhuÄ«, obtained a cache of ancient manuscripts. Central to this cache was a copy of ShÄ« è©©, the Songs (henceforth Än Dà ShÄ«), which on the surface bear close resemblance to the songs recorded in the ShÄ«jÄ«ng è©©ç¶, Classic of Songs, of the Máo recension.1 With the exception of the ShÄ«, the great majority of the texts in the cache were unknown prior to their acquisition.
The materials are being arranged and published according to their contents and style in multiple volumes as ÄnhuÄ« Dà xué cáng Zhà nguó zhújiÇn å®å¾½å¤§å¸èæ°å竹簡 (henceforth ÄnhuÄ« University Manuscripts). Volume 1 is devoted to the ShÄ« (Songs). It was published in late 2019, nearly five years after ÄnhuÄ« University first obtained these manuscripts.2
The said manuscripts continue a sad trend in the acquisition of knowledge about China in the Warring States period (ca. 453â221â¯BC), in that they were not obtained by scientific means.3 This aspect brings with it the standard problems, and concerns, of working with unprovenanced materials,4 academically, methodologically, and ethically.5 Our own position is that, academically, methodologically, and ethically, the field would lose much more by disregarding the materials than by working with them.
1 The Än Dà ShÄ«
The Songs as recorded in the ÄnhuÄ« University Manuscripts only contain what the Máo recension calls âGuó fÄngâ å風, the âAirs of the Statesâ. The âXiÇo YÇâ å°é âLesser Elegantiaeâ, the âDà YÇâ 大é âGreater Elegantiaeâ, and the âSòngâ é âHymnsâ are not present.
The Än Dà ShÄ« never refer to any of the states generically as guó å, as in Máo, or bÄng é¦, as in *KÇngzÇ ShÄ« lùn ååè©©è«, but specifically by the name of the stateâor, as in the first twenty-five songs discussed in this book, as âZhÅu Nánâ å¨å, the âRoyal ZhÅuâ, and âShà o Nánâ å¬å, the âRoyal Shà oâ.
The slips of the Än Dà ShÄ« are all numbered consecutively at the tail of the recto, from 1 to 117.6 As twenty-four slips are missing,7 only ninety-three slips remain, carrying fifty-seven songs. Missing slips might account for the absence of songs from a given state, but cannot account for the fact that whole states are themselves absent in the Än Dà ShÄ« but present in Máo.8
The ÄnhuÄ« University Manuscripts contain the songs of *six states.9 They are, in the preserved sequence:
-
âRoyal ZhÅuâ å¨å, 11 songs;10
-
âRoyal Shà oâ å¬å, 14 songs;
-
âQÃnâ 秦, 10 songs;
-
âHóuâ* ç¦, is transmitted by Máo as âWèiâ é, with one exception (âWèiâ 1), 6 songs;
-
âYÇngâ* ç¬ (Máo: YÅng é), 9 songs;11
-
âWèiâ* é, with the exception of âWèiâ 1, is transmitted as âTángâ å, 9 songs.12
The relationships between âHóuâ ç¦ and âWèiâ é; and between âWèiâ é and âTángâ å, are rather obscure, as there is no clear distribution pattern for individual songs between these states when comparing the Än Dà ShÄ« with the Máo recension. An âexchangeâ of songs between âWèiâ and âTángâ seems possible, given their geographical positions (Fig. 1), as both were situated in western ShÄnxÄ«, with Táng north of Wèi. This leads us to speculate that âHóuâ (perhaps a toponym like âWángâ ç in Máo [lit. âkingâ; but usually understood as referring to the âRoyal Domainâ]) was in western ShÄnxÄ« as well, and refers to a place between Wèi and Táng.13 Regardless of whether this location for âHóuâ can be accepted, the fact that sixteen of the remaining fifty-seven songs in this selection of ShÄ« have an affiliation with western ShÄnxÄ« seems significant, and appears to indicate a preference.14
Figure 1
Map of the states of the Shījīng
Adapted from FÄng Yùrùn (ADâ¯1811â1883) 2017: 10â11When looking at the states and their distribution, geographically and within the Än Dà ShÄ«, two things are noteworthy. First, geographically we notice a peculiar western bias of the Songs. Not one of the eastern states and only one of the central states of the Máo recension is present in the Än Dà ShÄ«, seemingly a glaring omission.15
The sequence of states differs remarkably from the received âGuó fÄngâ, and this warrants comment. The position of âQÃnâ 秦 is particularly striking. While it comes eleventh in Máo, it is third in the Än Dà ShÄ«. Given a terminus ante quem of circa 300â¯BC (330â¯BC ± 30 years; see below) for the manuscript, this may well have to do with the rising political power of QÃn at the time (and thus possibly the omission of the eastern states?). Out of respect for tradition, the songs of the âRoyal ZhÅuâ and the âRoyal Shà oâ had to come first, but the approach to the sequence then shows itself to be more flexible. In this context it is perhaps also worth noting that the QÃn songs have nothing of the sometimes satirical and playful elements we see in many of the songs from other states, but instead impress upon their reader a rather belligerent tone. Was this how the QÃn wanted to be remembered? Or is it a reflection of QÃnâs status at the time, given further substance by having it appear near the top of the sequence, after the âTwo Nánâ äºå16 (âZhÅu Nánâ å¨å and âShà o Nánâ å¬å)? Then, the section here called âHóuâ ç¦ (six songs) is the fourth state listed in the Än Dà ShÄ« but, under the name âWèi fÄngâ é風, comes ninth in the Máo recension. âYÇngâ ç¬/é (seven songs) is listed fifth in the Än Dà ShÄ« but fourth in Máo. What is called âWèiâ é (ten songs) in the Än Dà ShÄ« is, with one exception, âTángâ å in Máo, where it is the tenth state.
While the sequence of the states differs between the Än Dà ShÄ« and Máo, the sequence of songs within a state is consistent in the âRoyal ZhÅuâ and the âRoyal Shà oâ. The sequence breaks down, however, in the other states: âQÃnâ 秦, âHóuâ ç¦, âYÇngâ ç¬ (YÅng é), and âWèiâ é.
With a total of fifty-seven remaining songs, the Än Dà ShÄ« has slightly more than one-third of the 160 songs of the âAirs of the Statesâ of the Máo recension.
The songs pertaining to particular states are consistently separated by thick black marks on the slips. A hook-shaped mark (
) is used, though inconsistently, to signal the end of each subsection. The Än Dà ShÄ« also contain entries which detail the number of songs from one state, such as in âZhÅu Nán, elevenâ å¨åååä¸; or âHóu, sixâ ç¦å
. The songs themselves, however, are never listed with a title. On two occasions, at the end of a subsection, the manuscript text lists the first song by name, as in âYÇng, nine: White boatâ ç¬ä¹ç½è, or âWèi, nine: Grass-cloth sandalsâ éä¹è縷. Because the first song is listed for these two, the editors of the Än Dà ShÄ« suggest that each song must have had a nameâalbeit not one mentioned on the slips.17 Methodologically the editors go one step further, as they suggest that the names of the songs recorded on the manuscripts must have been roughly the same as those of the Máo ShÄ«. Hence, they introduce each song by its Máo designation, a choice we find philologically unsound and methodologically problematic. As the Än Dà manuscript is incomplete, we cannot be certain whether âShà o Nánâ and âQÃnâ also listed a song title. However, we can ascertain that the basic way of referring to subsections was by the name of the state and the number of songs, not by including the titles of songs.
We certainly would not want to exclude the possibility that the songs, or at least some of them, might have had the same or similar names as their counterparts in Máo.18 But if they did, the manuscript witness shows they were not considered to be an integral element, and so they should not be imported into this version of the Songs. To assume so as a working hypothesis therefore strikes us as a poor choice. (It would be impossible, anyway, to test this assumption.) Methodologically it is also flawed, because it guides our expectations and therefore the way we read the songs. Instead, in the representation of individual songs, we simply list them numerically, in the sequence order in which they appear in the ÄnhuÄ« University Manuscripts, which in the âZhÅu Nánâ and the âShà o Nánâ is largely compatible with the Máo recension. The fact that the songs are counted in the manuscript might suggest that the order was more important to the textual community of the Än Dà ShÄ« than paratext.
Despite the macro-consistency of the âstatesâ, the Än Dà ShÄ« nonetheless only contain about one third of the âGuó fÄngâ in Máo. They are also presented in a different order, not only in how the states are organised relative to one another, but also in how some individual songs are arranged within and among the latter four ânon-royalâ (or âcommonâ) statesâthat is, âQÃnâ 秦; âHóuâ ç¦; âYÇngâ ç¬; and âWèiâ é. The composition of individual songs, too, can differ, with stanzas being moved vis-à -vis their received counterpart. There are four such instances among the first twenty-five songs, that is, the songs of the âZhÅu Nánâ and the âShà o Nánâ.19 Moreover, in one instance, âShà o Nánâ 14, the song shows an additional stanza to its Máo counterpart (Máo 25 âZÅu yúâ 騶è).
2 Attempting an âEmicâ Reading of the Än Dà ShÄ«
We offer the first book-length study of the first twenty-five of the songs of the Än Dà ShÄ«, the songs of the âZhÅu Nánâ and the âShà o Nánââthe old royal songs which constitute a gateway into the ShÄ«. Radiocarbon analysis suggests the manuscripts are roughly 2280 years old, calibrated to the year 1950.20 This furnishes a date of circa 330â¯BC for the manuscripts. Subsequent chemical analysis of the slips has confirmed their mid-Warring States pedigree, though the ink was not tested. Theoretically, this means we cannot with certainty exclude the possibility that recent forgers produced the texts on ancient bamboo slips. But we follow the lead of Chinese scientists and palaeographers and rule out, as a working premise, that recently-fabricated texts were produced on ancient slips.
The editors of Volume 1 of ÄnhuÄ« University Manuscripts do not hide their excitement about the texts collected in this volume, for they now have a Warring States-period copy of the Classic of Songs, ShÄ«jÄ«ng è©©ç¶, or so they claim.
To be the first to edit a Warring States-copy of the Classic of Songs is, of course, an enormous burden. The editors duly responded to this difficult task by providing a philologically highly competent, but conservative (or, perhaps, extremely radical?) reading of the Songs. It depicts the Songs as decidedly consistent with the received Máo recension.
As our goal is to carry out the first fully-annotated translation into English of the first twenty-five songs of this collection, our stakes are lower, and so our task is considerably easier. Let there be no doubt: we are guided by the same academic ethos of striving to attain the best possible reading of the text and to do it no harm. But as we determine our reading, we do not feel the same ancestral weight resting on our shoulders, and so the burden of tradition restricts us far less when making our choices as to what we believe constitutes the best, and most honest, reading of the text.
The eminent scholar Qiú Xīguī once remarked the following:
å¨å°ç°¡å¸å¤æ¸èå³ä¸å¤æ¸ï¼å æ¬åä¸æ¸çç°¡å¸æ¬åå³æ¬ï¼ç¸å°ç §çæåï¼åè¦æ³¨æé²æ¢ä¸æ°ç¶çã趨åãåãç«ç°ãå ©ç¨®å¾åãåè 主è¦æå°ç°¡å¸å¤æ¸åå³ä¸å¤æ¸æ義æ¬ä¸ç¸åä¹è說æç¸åï¼å¾è 主è¦æå°ç°¡å¸å¤æ¸åå³ä¸å¤æ¸ä¸å½¼æ¤å°æçãæ義ç¸åæå¾ç¸è¿çå說ææ義ä¸åã21
When comparing bamboo and silk manuscripts with transmitted ancient texts (including bamboo and silk versions and their transmitted counterparts), we must resist two incompatible tendencies, one is âthe urge to equate themâ, the other is âestablishing differencesâ. The former primarily conflates differences in the content of bamboo and silk texts and their received counterparts; the latter stresses differences between the graphs in bamboo and silk texts and received texts, while their meaning is actually similar or very close.
These, however, are not our choices. As we shall detail in our discussion of the songs, the disparities between the Än Dà ShÄ« and the Máo recension are such that they seem to reflect an internally consistent (but in each case different) logic, which determines certain choices made by the communities behind the two versions as they wrote out the songs. Variances between the versions are not selective, let alone random, but rather systematic, and so the Än Dà ShÄ« appears as an independent iteration of the Songs. Because the date of the manuscripts is circa 330â¯BC, as a thought experiment we treat the texts upon them as non-Máo. Methodologically, this means that when reading the Än Dà ShÄ« we are not bound by Máo, and so, unlike the editors of the ÄnhuÄ« University Manuscripts, we do not have to force them through Máoâeven though the Máo recension crucially informs our choices.
We do not know much about the Máo recension. It was promoted as the authoritative edition by King Xià n of HéjiÄn (Liú Dé åå¾·) (r. 155â130â¯BC) but did not gain official recognition until quite late in the Western Hà n, somewhere between 1â¯BC and ADâ¯6, ousting the other major traditions, LÇ é¯, Qà é½, and Hán é.22 The Máo recension is often associated with Máo HÄng 亨 and Máo Cháng è, whose dates are uncertain.23 It is important to maintain a distinction between the Máo recension and the Máo commentary (i.e., Máo ShÄ« gÇxùn zhuà n æ¯è©©è©è¨å³). The transmission of the ShÄ«jÄ«ng is a convoluted topic, and beyond the remit of this book, but accounts are usually rooted in KÇngzÇâs compilation of âtheâ ShÄ« and its subsequent dissemination by BÇ ShÄng åå (507â? BC; courtesy name ZÇ Xià åå¤). An early medieval model of transmission that we follow relates that the âelderâ Máo, Máo HÄng, learned âaâ ShÄ« of the XúnzÇ (ca. 300â219â¯BC) tradition.24 Bernhard Karlgren (1889â1978), proposing that âelderâ Máo was a âdiscipleâs discipleâ of XúnzÇ, demonstrated a close affinity between citations of ShÄ« in XúnzÇ and the Máo recension. He further suggested that the Máo commentary, which he concluded was completed in the second century BCâânearer to 150â¯BC than to 200â¯BCââhad been strongly influenced by XúnzÇâs thought.25 Be that as it may, it remains uncertain whether what we now know as the Máo recension was actually in circulation before Máo HÄng obtained a copy of it, or whether he âestablishedâ a new recension from an older (and primarily sound-based) modelâor potentially a combination of the two. In this regard, it is perhaps noteworthy that of sixty-one instances of ShÄ« quotations in XúnzÇ, forty-eight agree with the Máo recension and only thirteen disagree.26
As we take the Än Dà ShÄ« as a non-Máo iteration of the Songs, methodologically we are not bound by the Máo recension, and thus by received canon. Rather, our reading sets out to understand the texts as written on the slips, insofar as this is possible but helped, of course, by modern understandings of the phonetics of Old Chinese and by advances in palaeography.27 We seek to understand how a certain community with a profound specialisation in the tradition afforded meaning to the Songs. In order to do so we use a two-stage approach. We first have recourse to broadly contemporaneous texts, principally the *KÇngzÇ ShÄ« lùn of the Shà nghÇi Museum collection of ChÇ manuscripts, which allows us to understand how at least one other textual community worked with the ShÄ«.28 Second, where possible, we systematically consult the Máo recension, as well as fragments of texts from other major Western Hà n traditions, to help us reach an informed decision as to what a given song actually says. We can surmise that Máoâs reading was at least partly based on pre-existing notions of matters pertaining to the Songs, albeit perhaps different in some details. It is thus instructive to see where the Än Dà ShÄ« differs from the major Western Hà n ShÄ« recensions, and, possibly, why. As we are accessing the long, philologically-immaculate tradition of ShÄ« exegesisâof which we are now a partâwe rely on the commentarial traditions of the Western (206â¯BCâADâ¯9) and the Eastern Hà n (ADâ¯25â220), as reflected mainly in the Máo commentary and the copious notes produced by Zhèng Xuán éç (ADâ¯127â200), so as to be clear about our own choices. We have supplemented this by consulting QÄ«ng Dynasty (ADâ¯1636â1912) scholarship on the ShÄ« and its Hà n reception, as reflected mainly in the work of Yáo Jìhéng å§éæ (ADâ¯1647âca. 1715), FÄng Yùrùn æ¹ç潤 (ADâ¯1811â1883) and Wáng XiÄnqiÄn çå è¬ (ADâ¯1842â1917); Bernhard Karlgrenâs âGlosses on the Book of Odesâ, which relied heavily on the scholarship of Chén Huà n é³å¥ (ADâ¯1786â1863), is a vital resource.
The goal of this exercise is to establish a hypothetical âemicâ perspective of reading the Songs through the Än Dà ShÄ«.29 It is hypothetical, if only because it can never be fully achieved. âEmicâ perspective considers an insiderâs account or perception.30 The approach is borrowed from linguistics,31 and relates to the study of a language or culture in terms of integral elements and respective functions, rather than through existing models or schemes that are necessarily external.32 In cultural anthropology, the emic approach describes a viewpoint obtained from within a social group.33 The application of this approach to ancient texts and how they were used is necessarily limited, but as Mikhail Bakhtin (1895â1975) has stated, behind the iteration and interaction of texts is always âa contact of personalities and not of thingsâ.34 Texts are the secondary products of the multi-faceted social realities of meaning construction. The primary actors are peopleâindividual or groupsâparticipating in a discourse.35
In attempting an emic perspective of reading the Än Dà ShÄ«, and so to cast light on how the community behind the Än Dà ShÄ« afforded meaning to the Songs, we do not claim to be metaphorically venturing back in time, like embedded anthropologists, to sit idly among the textual community that produced this iteration of the Songs, to witness their use, and to experience it for ourselves. But the manuscript at hand and the above-mentioned support material afford sufficient grounds from which to approach the text on its own terms and, on this basis, to reconstruct what its textual community may have wished to achieve by engaging with the songs. Remembering that this collection is a reflection of people and their concerns âand not of thingsâ, we are thus establishing a hypothetical reading of this particular iteration of the Songs as it circulated among certain conceptual communities during the Warring States periodâconceptual because they must remain our theoretical projections.36 To seek an emic reading therefore does not mean we are reading this manuscript in isolation, as though we had some benediction granting us access to it. We also do not treat it as though it came to us out of the blue. We consider it not possible toâand do not aspire toâtreat the Än Dà ShÄ« as separate from context of wider ShÄ« scholarship.37 Just as this manuscript iteration was produced in a given social setting, so too was the Máo recension, and the recensions of the other major (and minor) Western Hà n traditions.
It is important to stress that we do not, and never will, claim that we are reading âtheâ ShÄ«. Rather, our approach has always been that of a âgedankenexperimentâ where, methodologically, we provide a hypothetical reading of what certain conceptual communities during the Warring States period might have made of the Än Dà ShÄ«, as they were themselves not immediately guided by the Máo recension. (Most likely they were informed by the predecessors to Máo which, in turn, might at least partly have informed Máoâs conceptualisation of the songs too.) What we are reading is âaâ ShÄ« from the Warring States period.
3 Writing the Image Programme of the Songs
We detect at least two orthographic manners on the slips.38 Certainly in âZhÅu Nánâ, and for the most part also in âShà o Nánâ, the calligraphy is executed with great care. It is often exceedingly elegant, and sometimes a graph is written to be evocatively pictographic. The graph for horse, mÇ é¦¬, in âZhÅu Nánâ 3 (Máo 3) is a case in point, in that it actually shows a horse: 
. Interestingly, the writer of the songâwho is emphatically not the composer of the song; we chose the designation âwriterâ for the person who executed the calligraphy on the manuscript deliberately as it does not imply a directional relationships between the manuscript and the text, as âcopyistâ or âscribeâ does39âalso allowed for some variation in the calligraphy, as the last horse in the same song appears like this: 
. That is, the horse without the legs, so to speak:
è=ð§¶è³. ä¸æº.
å·®æè¤äººç®å¨è¡
éç®é«é¬æ(馬)çé» 6|
\æå¤åºç®å è¡¡é¹ã ¯ç¾éç®åµæ
(馬)
éº
æå¤åºéé¹ã ¯ç¾è¤
é 7||ç®ð£³ç£æ(馬)å¾ç£.
夫ç£å¡å¯ç¡ç£ 
What is especially interestingâand this differs decidedly from the Warring States-manuscripts that carry, say, philosophical textsâis that many of the graphs in the Än Dà ShÄ« consistently contain an extra layer of information that expresses the meaning of the word through the graph in ways other literary texts of the time do not seem to do.
For example, the graph representing âflying [birds]â 
for fÄi é£, is written as 
in stanza 1 of âZhÅu Nánâ 2 (Máo 2). The graph for âflyingâ is not normally indistinct, and so it is apparent the added signifier âbirdâ was not used for the purpose of reducing ambiguity.41 Rather, it strikes us as something playful, yet which served to evoke images through writingâor reflects what the writer had in mind as they produced a copy of the song. We see an attention to detail, by which the writer sought to have the graphs visually embody the wider meaning of the words they represent. One might also consider the onomatopoeia â*rih rihâ for the sound of birdsâ chirping, written as 
= (
) in the same stanza:
èä¹å¯.éäº
æµ´.
é¹ 3| | ⢠èè=é»é³¥äºï¼
ï¼
éäºæ¬æ¨äºé³´ï¼
ï¼=
How the kudzu spreads! extending deep into the valley.The leaves are plentiful, orioles are in flight [to it].Gathering on the trees with yellow blossoms, they tweet ârih-rihâ.
This extra layer of information in the graph means the writer was not just writing âto flyâ. By adding the signifier âbirdâ to it, the writer further evoked the image of orioles in flight. Similarly, the writer did not just write the sound â*rih rihâ to represent the chirping of orioles gathered in the trees, but continued to evoke the imagery of chirping through the addition of yet another âbirdâ signifier in the graph that wrote out their sound.42 Examples of this sort can be found throughout the âZhÅu Nánâ and âShà o Nánâ (and in much of the Än Dà ShÄ« more generally).
As they add to their meaning in significant ways, so too do the graphs on the slips often go beyond, or at least differ from, their modern equivalents in terms of meaning and connotation. Because this was done with such care, reflecting a strategy on the part of those who wrote out the songs, we have opted not to follow the standard form of text representation, where a direct transcription precedes the interpretative transcription using modern equivalents, as in é»é³¥äº
ï¼é£ï¼éäºæ¬æ¨äº (å
¶) é³´
=ï¼ååï¼.43 This would have meant distorting the text in such a way that we, for our part, would have had to invent a text which the writer of the Än Dà ShÄ« did not mean to produce. (For âZhÅu Nánâ 2, it would have meant taking some âbirdsâ out of the lyrics, so to speak.) Hence, we have simply maintained the direct transcription of what the writer wrote. With this in mind, we shall return briefly to the horse with- and without legs in âZhÅu Nánâ 3.
Subsequent to the first stanza, which serves as a form of preamble situating âZhÅu Nánâ 3 emotionally and contextually, three stanzas describe a progression of decline on the part of the travelling male as imagined, and voiced, by the lone female singer, his lover:
Ascending that high ridge, my horse () has turned dark yellow.
I now ladle into that rhinoceros horn cup, it is thus my pain shall last.Ascending that craggy height, my horse () is on the verge of collapse.
I now ladle from the bronze bucket, it is thus my yearning will last.Ascending that slippery slope, my horse () is at a crawl.
and as for my humble servant, there is nothing more to say!
The horse is the one repeating feature in the image programme of the song,44 and it serves to stand metaphorically for the pains of the male traveller as imagined by his suffering woman, who places herself in the narrative perspective of her lover. The progressive decline of the horse is emblematic for that of the male traveller and his companion. First, the horse has ascended a high ridge and the writer shows that its strong legs are moving forward. Second, the writer has changed the position of, and inverted, the horseâs once-strong, forward-striding legs to show the result of it being pushed to ascend a rocky height. By the third stanza, the travellerâs horse is literally finished. As it slows to a crawl, it appears as though it has no legs to stand on while trying to make its way up a slippery slope, a situation observed and emotively commented upon by the writer of the song.
Had this visualisation been one of a handful of examples, we would not emphasise it and declare it a chance coincidence. But this is not so. Close analysis of songs of the âZhÅu Nánâ and âShà o Nánâ shows that a layer of meaning was regularly added through careful execution of the writing, including the binomes, and that this was an important means of visually expressing what was occurring in the songs.45 There are of course many graphs in the Än Dà ShÄ« where this aspect is lacking, but key terms in the songs are marked with striking consistency. We here cite but a few examples:
âZhÅu Nánâ 1 (Máo 1) establishes an image programme centred on a male osprey seeking his prey as a metaphor for a male seeking a lithe and beautiful female of his desire. The female is referred to, sportively, as an âadversaryâ, that is, the object of his desire, and the graph writing this word chóu, is written 
, with the phonophore zÇo æ£ âjujubeâ (*tsûÊ), which the writer specified by adding a âdagger-axeâ æ signifier. This was a common way of writing the word in Warring States ChÇ script, and is the same word used in the LÇ (followed by Zhèng Xuán) and Hán recensions (written with the graph ä»). It forms a semantic bracket with the words lán-lán (*rôn-rôn) 𨷻𨷻 âenter recklesslyâ (into someoneâs space), before it, and liú æµ âflow around [like water]â, qiú æ± âseekâ, fú æ âsubdue; cause to obey; give upâ, cÇi é âpluckâ, yÇu æ âobtainâ, and jià o æ (æ) âhandpickâ, after it, to produce a reading that is decidedly different from what its counterpart, qiú é, âmate, companionâ, means in the Máo recension.
âZhÅu Nánâ 2 (Máo 2) describes a womanâs joy as she enters into marriage. Orioles (huáng niÇo é»é³¥ âyellow birdsâ), chirping, carry forth the happy news. We have already mentioned the added signifier âbirdâ for the orioles in flight, and in their chirping of the happy news. The Än Dà manuscript text also differs from the Máo recension in where the birds alight. The chirping orioles gather in the Än Dà ShÄ« on âtrees with yellow blossomsâ, quán mù æ¬æ¨, the colour yellow symbolising the female and serving as a metaphor for marriage, whereas in the Máo recension they alight, more prosaically, on âbushesâ, guà n mù çæ¨. While the soundâand thus the prescriptive sound mouldâof quán and guà n is the same (both are written with the phonophore guà n (*kôn) é âheronâ), the repeated occurrence of the graph ç later in the manuscript textâs lyrics (stanza 3), where it writes the word guà n in its primary meaning âpour out; rinseâ (contra Máo: huÇn æ¾£ âwashâ), demonstrates that great care was taken by the writer of this song when choosing the appropriate signifierâwithout compromising its sound mould.
âZhÅu Nánâ 3 (Máo 3) has the above-mentioned âhorseâ in emotive postures that change along with the progression of the song. The male subject of the song is imagined to be in the mountains in stanzas 2 to 4, and six of the fifty-one words are written with âmountainâ signfiers.46 When describing the manâs horse as it moves through the mountains, graphs with a âmovementâ signifier express exactly that movementâwhich differs from the graphs chosen in Máoâand with them the songâs image programme, while keeping the sound mould intact.47 There is also a discernible pun in the writing of the word shÄng âpainâ (Máo: å·) with the graph 
(觴) âcup of hornâ. The writer added an extra layer of signification to the word âpainâ by writing it, hardly unintentionally, with the very ale cup that the subject was going to use to prolong that pain. Reading 
as a simple phonetic loan for å· is, of course, possible. But this presents a much weaker reading, one that ignores the songâs image programme and thus misses the ingenuity of how the writing is employed; more fundamentally, it potentially nullifies further associations and the feelings it evoked when encountered in the song.
âZhÅu Nánâ 5 (Máo 5) centres around locusts and their various sounds and movements as a metaphor for the continuity of a lordâs line, coming forth in multitudes. Unlike in Máo, the writer preserves an evocative and striking intimacy between âlocustsâ and âpeopleâ, equating them through intersecting parallelisms in the writing of the graphs. The first two lines of each stanza are about the locusts, while the third and fourth lines are each about the lordâs descendants. But where the song talks about locusts, it uses graphs with humans-related signifiers; where it talks about people, the writing uses graphs with the insect signifier. This parallelism begins with the songâs very first word, zhòng âmultitudesâ, and ends with its final word, shéng/mÇn âto spring up ceaselesslyâ. In the opening sentence, the one about âlocustsâ, the graph used to signify the word contains âthree peopleââ
(ç¾)âwhereas in the songâs final sentence, about the lordâs âdescendantsâ, the graph used to write the reduplicative binome modifying them is composed with an âinsectâ signifier, 
= (with the reduplication mark repeating, in fact âdoublingâ, the âinsectâ: 
.)
âZhÅu Nánâ 7 (Máo 7) centres on the image of a rabbit net å
(*lhâh-tsha)âthe materiality of its composition, its construction, and its placement in several key positionsâas part of a phonetically-driven metaphor for the martial man æ¦å¤« (*maÉ-pa) and his disposition in properly serving his lords, the GÅng and the Hóu. Similar to how the imagery of humans and locusts intersects in âZhoÅ« Nánâ 5, in this song the martial man and the rabbit net are braided together using reduplicative binomes written with a âsilkâ signifier that is absent in the Máo recension.
In âShà o Nánâ 6 (Máo 17), the song is sung in the voice of a woman who suffers from ongoing harassment, and litigation, by a man she presumably wishes to divorce. When the song arrives at the part about litigation in stanzas 2 and 3, two specialised forms in Än Dà â𤶬 and 𧫷, which correspond to the word sù é âbeckon, invite, urge onâ (but also ârapid, quickâ) in Máoâadd meaning through the use of âsicknessâ and âspeechâ signifiers. While the same phonophore makes it certain that all three graphs write the same word, we propose that 𤶬 captures the ârapidâ onset of illness, with the underlying sense that this is akin to forcing litigation upon the female defendant. As for 𧫷, which is listed in the ShuÅwén jiÄzì as a Warring States gÇwén form of sù é, the addition of the signifier âwords, speakâ suggests yet another specialised form, and thus signify âto speakâ or âcall forth to speakâ, in this case at trial.
âShà o Nánâ 13 (Máo 24) celebrates the offspring of King PÃng (r. 770â720 BC) and the Hóu of QÃ, as one of the royal princesses is married to an elite family in the east. The writer of the song visibly intended to sublimate the image of flora and fruit in abundance through a sustained use of graphs written with floral signifiers, and as a way of adding meaning to the metaphor of beautiful royal females and their many anticipated offspring. Of the thirty-seven graphs still legible on the slips, eight are written with the floral signifiers: âgrassâ å±®/è¸; âtreeâ æ¨; and âplant (of grains)â 禾. In the songâs final stanza, the metaphor changes from royal females and their descendants as beautiful flora and abundant fruit to a silk fishing line. This is also reflected in the songâs written instantiation as three of the stanzaâs remaining eleven words, including the songâs final word, âgrandchildrenâ, sÅ«n å« (*sûn), are written with a âsilkâ signifier.
These few examples (we discuss these and other cases in more depth in the philological annotations of the songs) show that the written graphs in the Än Dà ShÄ« serves to develop its image programme, and thus the songsâ meanings. This is not an argument for a fundamentally written tradition of the Songs (or as it is often put, their âwritten natureâ). But it shows how the writer of the Än Dà ShÄ« either expressed the feelings which the ShÄ« invoked in them as they wrote down their version, or how they voiced, in written form, what they felt the Songs expressed. In any case, the writer(s) made the Songs more profound through the careful execution of written graphs. It thus appears that by around the mid-fourth century BC, both the phonetic texture and image programme were integral elements in ShÄ« iterations, at least in this case, thus demonstrating extensive knowledge of tradition, and transmission, on the part of the writer of the Än Dà ShÄ«.
The ShÄ« lend themselves to such forms of meaning construction through the written graph in the build-up of an image programme in ways other genres, with the exception of the Yì (Changes), might not.48 In so many ways the ShÄ« are more evocative due to metre, regulated words per line, sound moulds, and the traditionâs emphasis on prompting feelings through sound and imagery, which might explain this facet. Many of the examples we have cited add meaning to the songs through the use of signifiers associated with nature, like âgrassâ, âwoodâ, âmountainâ, âavianâ, âanimalâ, and âinsectâ. This is not without precedence though.49 The writer of this iteration of the ShÄ« placed importance on how to write key terms, including binomes (sometimes by using semantically-charged phonophores), so as to add meaning to the songs, or to reveal a connection between the songs of a given state. The Än Dà ShÄ«, with its discernible care regarding how sounds were written out, enables us to revisit the value of the written graphs in other iterations of the ShÄ«, including the Máo recension, and in other genres.50
We would like to stress, once more, that the aspect we have just described says nothing about the nature of the Songs in general.51 But it shows the limitations of using isolated, external, sourcesâfor instance, contemporaneous philosophical textsâto then conclude that the Songs must have been exclusively oral. Contemporaneous texts might follow the Songsâ phonetic textures, but specific semantic elements such as those in Máo are, at best, secondary. Rather, we posit that the different ways philosophical texts relate to the Songs demonstrates how different conceptual communities during the Warring States period would have used the Songs for their own ends. We thus gain information about conceptual user communities and their requirements, not about the Songs themselves. Likewise, the writer of the songs in Än Dà ShÄ« executed the calligraphy with such care that it produced an extra layer of significance, but this does not imply a characteristic of the Songs during the Warring States period in general. Rather, it shows that in this written instantiation of the Songs, the process of writing and the written graphs were important to the textual community that produced them (or, just possibly, to one writer). Writing was a tool used to evoke images to further stimulate the reader, and to enhance the meanings of the songs in ways they saw fit. This shows an aspect of the Songs that has, until now, been abstruse. By moving beyond philosophical texts we have gained insight into how the Songs were used by real communities in the Warring States period.
In summary, we have already noted that as our working hypothesis, our gedankenexperiment, we consider the Än Dà ShÄ« to be non-Máo. Yet, with this proposition in mind, we do not insist it must have been pre-Máo. Methodologically, this choice to treat the Än Dà ShÄ« as non-Máo enables us to read the songs of the ÄnhuÄ« University Manuscripts on their own terms, as given representation on slips by a community who afforded meaning to the Songs. By comparing the various representations of the Songs in texts from the Warring States period, as well as those in the Máo recension, we conclude that by this time the ShÄ« had formed a matrix, through their form, structure, and phonetic value. Conceptual communities followed this matrix, but nonetheless had some liberty when choosing how to inform in the precise semantic content of the ShÄ«, within sound moulds that in turn conformed to the expectations of the various textual communities of the time.
4 Bringing to Life the Sound Moulds of Shī Production
The Än Dà ShÄ« were not mindlessly penned. Certainly for the âZhÅu Nánâ and the âShà o Nánâ, the calligraphy was executed with care and served as an additional and meaningful element of signification. This includes the addition of signifiers or phono-semantics to certain graphs, and often binomes, so as to add clarity to what was said in the song and to further evoke its imagery; to comment on it in an âinstructiveâ way; or as a means to connect songs and themes into larger clusters of signification. In this regard, the discovery of a Warring States instantiation is nothing short of a momentous event for early ShÄ« studies because, with the Än Dà ShÄ«, we realise for the first time how the songs were not just received only through an appeal to the ears, but also to the eyes: a literate community gave voice to songs through a writing-supported text performance in which the written word was not secondary, but played a significant, and to some extent primary, role.52
Of course, this observation does not deny the vital aural element of the songs. They are, after all, songs. In this book we describe how sound moulds function in a ShÄ« tradition where units of signification were not just created through rhyme, but were distinctive phonetic textures that bound the songs into tight units, and yet may be filled with diverging but meaningful content, as revealed by comparisons between the Än Dà ShÄ« and the Máo recension. Whether these sound moulds demonstrate the oral prevalence of the songs is not for us to say. Rather, we believe that sound moulds confirm the aural significance of the songs, carried not just by rhyme but by distinctive phonetic textures, and that they produce containers for the songsâ contents. These contents need to be made explicitly, something we believe would not work well without a written element. The Än Dà ShÄ« manuscript thus shows how ShÄ« worked on the ground during the Warring States period, that is, within different textual communities who responded to the sound moulds of individual songs according to their own biases and requirements.
Sound moulds carry a songâs words, and thus they define the individual songs of the ShÄ« phonetically. But as our term âmouldâ suggests, during the Warring States period they also served as containers that could be filled, to an extent, with semantically malleable content deemed appropriate by the community in question. As long as the sound of the chosen graph did not violate the prescriptive mould into which it was placed,53 the writer, who, by virtue of writing the song was also commentating on itâand thus laying claim to itâwas at some liberty to apply a groupâs consistent reading of the song.
Huáng DékuÄn and Edward Shaughnessy have already commented on linguistically variable word choices in the Än Dà ShÄ« vis-à -vis the Máo recension, where the phonetic value of a word remains intact while rendering a different meaning.54 Both commentators focus on the example of âShà o Nánâ 14 (Máo 25 âZÅu yúâ 騶è). The Än Dà ShÄ« writes cóng hÅ« å¾
*dzoÅ-*hâ âafter themâ or zòng hÅ« 縱
ârelease themâ where Máo has zÅu yú 騶è (*tsro-*Åwâ). The latter is notoriously difficult to interpret.55 Although the difference between the two versions is phonetically sound, Huáng DékuÄn insists that this does not necessarily present an innocent phonetic loanâthe words may just as well show two different readings of the song.56
While we agree with this assessment, we go one step further. Spurred by our reading of âYÇngâ ç¬ 1 (Máo 45 âBÇi zhÅuâ æè âCedar boatâ), we realised that such variances need not be selective, let alone arbitrary, but apply to the expressions of ShÄ« more profoundly, going beyond the odd graph or binome in a song.57 Something quite systematic occurred during the Warring States period when communities voiced ShÄ« through the written word. In the case of âYÇngâ ç¬ 1, we see meaningful variance that stretches consistently over entire lines, such that they alterâsometimes dramaticallyâthe overall character of the song, while keeping intact its phonetic value.
The âTwo Nánâ show similar characteristics. A case in point is âZhÅu Nánâ 3 (Máo 3 âJuÇn Ärâ å·è³ âCockleburâ).58 Stanza 3 reads as follows:
|
é 7||ç®ð£³ç£æ馬å¾ç£. |
Ascending that slippery slope, my horse is at a crawl. |
|
æ |
and as for my humble servant, there is nothing more to say! |
夫ç£å¡å¯ç¡ç£ 
The Máo version has the following:
|
éå½¼ç ç£ï¼æ馬çç£ã |
I was ascending that flat-topped height, but my horses became quite disabled. |
|
æåç¡ç£ï¼äºä½åç£ã |
And my servants were [also] disabled. Oh! how great is my sorrow!59 |
In the even-numbered line (2), the Máo recension has tú ç (*dâ) âfatigued, illâ where the Än Dà ShÄ« writes tú å¾ (*dâ) âon foot, walkâ. Phonetically they are identical and so one might dismiss this difference as a simple phonetic loan. However, both renderings each provide systematically stable readings. In the Än Dà ShÄ« we see the steady decline of the horse, to the point where it can no longer walk (it is âat a crawlâ), while the Máo recension lays stress on how the horse is increasingly unwell. This situation continues into the next lines where the iterations focus on different aspects in accordance with their image programmes. In one case it is the âhumble servantâ of whom âthere is nothing more to sayâ. In the other case, it is the âservantâ who, just like the horse, is âdisabledâ, leading to the self-pitied exclamation âhow great is my sorrow!â To ignore such differences means to miss the opportunity of seeing how the sound moulds allowed different communities to construct meaning variously.
As these examples show, a song might display different word choices, but aural correspondence remains largely intact between their various iterations, be it Máo, Än Dà ShÄ«, or another. As we discuss more fully in our annotations to the songs, this is especially true of the keywords and the rhyming words,60 although there are also instances of disparity that are perhaps best explained as reflecting linguistic features of ChÇ regional dialects, and how local writers accommodated them.61 A comparison between the Máo and Än Dà versions sees significantly less lexical variation in the songs of royal states, the âTwo Nánâ, than in the songs of the other four âcommonâ states. This fidelity indicates an intimate reception of the royal songs and, for late Warring States-Western Hà n users of the ShÄ«, a reverence for their exalted pedigree.62 That Máo and the Än Dà ShÄ« both adhere to the same sound moulds also suggests the tradition they follow was much older. And that by the time when the Än Dà ShÄ« was voiced in writing, these moulds were fundamentally stable. The ShÄ« are characterised by an aural primacy, but that does not extend to semantic fixity.63 Taking note of the productive sound moulds in ShÄ« iterations might not just prove useful when discussing new finds of manuscript texts in the future; it might also serve as a suitable tool when revisiting the major traditions of the Western Hà n period, the foremost of which is Máo.
5 Sounding the Image Programme of the Songs
Within a song, accumulation and layering of sound moulds produces a âphonetic textureâ regarding how certain elements are linked. And as with written forms, phonetic textures can evoke meaning and contribute to mounting image programme of a song.
We find a variety of strategies by which meaning is evoked through sound. Consider âShà o Nánâ 8 (Máo 19 âYÇn qà léiâ æ®·å
¶é ââ¯âQinâ sounded the thunderâ). Here, just as in other songs of the âZhÅu Nánâ and âShà o Nánâ,64 we can see how the writer in question filled in a prescriptive sound mould using just enough variation to show meaning. We argue that this does not show a degree of indifference to the semantic content, but rather corresponds with how the writer (or the community of which the writer was a part) saw it fit to detail the content of the song. The song is highly repetitive, with just the last word of lines 2 and 4 changing. The sound of the song is carried by the dominant vowels *-e-, -É-, and -a-, which characterise all three stanzas. Only two elements break away from the phonetic pattern, the repeated sounding of the thunder, 
äº
(*Éin *gÉ *rûi), marked by *-i- and the pharyngealised *-û-, and the âso quaking is the lordâ 
=å (*tÉn-*tÉn *kun), which takes up these sounds, albeit weaker and not pharyngealised. These two sounds thus form a phonetic bracket which links the two choses words (or âimagesâ) into a meaningful unit, a technique we encounter in many other songs of the âTwo Nánâ.65
âShà o Nánâ 8 contains two consistent rhymes, which is consistent with Máo. It is in the voice of the narrative âIâ of a female speaker, lamenting her lordâs departure as soon as the thunder sounds. The song is highly repetitive and, as such, it is predictable. So too is its message. We consider this not to be a coincidence, but rather suggest that it is part of the songâs image programme. As the thunder (*rûi) sounds (*Éin), the narrative âIâ of the female voice knows her lord (*kun) will be prompted into action (*tÉn). As the speakerâs experience predictable, so too is the repetitive structure of the song and the texture of its sound moulds.
Something different occurs in âZhÅu Nánâ 8 (Máo 8 âFúyÇâ è£è¢ âPlantainâ), a song in four stanzas which describes the monotonous work of plucking the plantain. The rhythmic and repetitive nature of the song, with its continued stress of glottal stops, beautifully captures this monotony. Just like âShà o Nánâ 8, on a phonetic level the song thus reduplicates its semantic content:
*tshÉÌÉ-tshÉÌÉ puÊ-lÉÊ, *bâk-Åan rhyme word-*1 tÉ*tshÉÌÉ-tshÉÌÉ puÊ-lÉÊ, *bâk-Åan rhyme word-*1 tÉ;*tshÉÌÉ-tshÉÌÉ puÊ-lÉÊ, *bâk-Åan rhyme word-*2 tÉ*tshÉÌÉ-tshÉÌÉ puÊ-lÉÊ, *bâk-Åan rhyme word-*2 tÉ;*tshÉÌÉ-tshÉÌÉ puÊ-lÉÊ, *bâk-Åan rhyme word-*3 tÉ*tshÉÌÉ-tshÉÌÉ puÊ-lÉÊ, *bâk-Åan rhyme word-*3 tÉ;
Or one might consider âShà o Nánâ 6 (Máo 17 âXÃng lùâ è¡é² âWalking in the dewâ). This sings of a woman who suffers ongoing harassment by, presumably, her former husband. In highly accessible visual terms, which parallel her husbandâs actions with how a sparrow forces its way through a roof, the song describes how the man encroaches upon female space. This threat is captured, phonetically, by the *-ôk and -ok endings of the rhyme which reproduce the hacking sound of the âbeakâ (jiÇo è§ [*krôk]) and thus invokes a sense of threat.
6 Receiving the Shī
As both the visual and aural information are relevant when constructing the image programme of a song, questions arise as to how ShÄ« were transmitted and received. Based on the consistency and attention to detail in how an image programme was established, through carefully executed writing within the existing sound moulds of ShÄ« production, we disagree with interpretations that regard the written song lyrics of this selection of ShÄ« as minimal notations of sounds, heard through second-or third-party recitation. A contemplated transcription from self-recitation strikes us as conceivable for thus instantiation, as might written transmission from a pre-existing source text.66 Particularly applicable to our reading is the model of a writing-supported text performance. This describes the enabling, and execution, of a complex utterance (text) through the support of the written word.67 The enabling of a textâs iteration through a writing-supported text performance implies certain constraints, genre-specific or otherwise. In the case of the ShÄ«, this would be the sound moulds within a prescribed phonetic texture. Such frameworks ensure that the expectations held by the different groups as to what constitutes a valid instantiation would be met, within respectively defined boundaries: a text is accepted within a given genre; an idea is considered good and an argument sound within a particular discourse; and with the ShÄ«, a song is received, and voiced in writing, within the accepted confines of its sound moulds and phonetic texture.
7 The Significance of the Royal ZhÅu and the Royal Shà o
For the most part, the songs of the âZhÅu Nánâ and the âShà o Nánâ display a marked stability vis-à -vis Máo. Even though at times the order of the stanzas differs, as mentioned, structurally complete songs are generally characterised by their consistency. One reason for this stability was almost certainly the permanency of the sound moulds and phonetic texture they produce. Thanks to the acquisition of the Än Dà ShÄ« we can conclude that by the time of the mid-fourth century BC, at the latest, phonetically the songs of the âZhÅu Nánâ and âShà o Nánâ were largely stable. Not one of the songs of âZhÅu Nánâ or âShà o Nánâ is particularly long, and each has a clear rhyme scheme that further eases reproduction in different contexts, which certainly helps their stability in the long term. That the image programmes of many of these songs differ, sometimes spectacularly, from Máo should not come as a revelation. Different versions simply speak to different communities and their specific needs. In those cases where we see a different sequence in the order of the stanzas compared with Máo, we often feel that the sequence in an Än Dà song is intuitively more appealing than that in Máo, as it generally follows a clear progression, temporally or otherwise, sometimes following a prelude that sets the songâs mood or context. It is often in Máo, not in the Än Dà ShÄ«, that this progression appears to be interrupted.68
But what about the macro-consistency of an entire state (guó/bÄng)? How is it that the sequence of the first twenty-five songs, the songs of the âRoyal ZhÅuâ and the âRoyal Shà oâ, is so stable between the Än Dà ShÄ« and Máo, much more so than for any of the other states?
To answer this question, we scrutinised the image programme of each song and searched for possible connectors from one song to the next. In many cases, though not all, we were able to discern connectors that produce a âdistant readingâ of the songs, which ties them together in a stable sequence across a state. Connectors may vary in type. They include correspondences in the narrative voice between two or more songs, images that run across the different songs, calligraphic connectors (as we call them) that tie two or more graphs together across songs, and distinctive sound patterns such as phonetic brackets or coded phrases, which are marked by phonetic exceptions that are repeated across the certain songs. But more often than not we found a combination of these and other literary devices that produced a macro-matrix of ShÄ«-organisation within the states.69
The other reason for stability lies in the special status of âZhÅu Nánâ and âShà o Nánâ as a conceptual pair. Pre-Hà n texts usually refer to the âZhÅu Nánâ in conjunction with âShà o Nánâ. Certainly by the mid-Spring and Autumn period they were considered a unit and, at least structurally, a comparatively stable one. This is because unlike any other state, the songs of the âRoyal ZhÅuâ and the âRoyal Shà oâ are ultimately linked to the two paragons of Chinese civility, ZhÅu GÅng and Shà o GÅng, if only by tradition rather than fact. Moreover, the âZhÅu Nánâ and the âShà o Nánâ formulate a normative claim to carry the songs of old royal states, as marked by their programmatic title nán å.
The tight organisation of the songs within the two states and their special, normatively royal status may well have worked in tandem, one driving and strengthening the other. Mututal enforcement might have lent them a special place, not just in the collection of âThe Statesâ, the âFÄngâ, but in the ShÄ« more generally.70 One can only hypothesise whether the two states were also passed on as one unit, separate from the other states.71
This is why we chose to focus closely on this conceptual pairing from antiquity, examining their shared features and stable sound moulds, and highlighting their importance to the different conceptual communities of the Shī.
Accordingly, we review the âTwo Nánâ of the Songs during the Warring States period from an emic perspective, and relate them to what we can reconstruct from the transmitted literature. We shall take a take a smallâbut informedâleap of faith and claim that, in the sixth century BC, the word nán å, when used in combination with yÇ é , referred to the songs of the âZhÅu Nánâ and the âShà o Nánâ. This does not imply there was a precise match between sixth-century âNánâ and the first twenty-five songs of the Máo recension and the Än Dà ShÄ«. But a substantial overlap, certainly with regard to the sound moulds of the songs, however many they were, seems most likely.
8 What Does âNánâ Mean?
We have no intention of revisiting the wealth of interpretations regarding to the meaning of the two-word combinations âZhÅu Nánâ and âShà o Nánâ. Too much has been written about it already, as nearly every commentator who works on the ShÄ« feels obliged, by tradition, to take a position on the issue, which nonetheless remains controversial. The problem is the function and meaning of nán å and its relationship to the state-names âZhÅuâ and âShà oâ. No other subsection carries this, or any other, designation, instead being only known by state name.
Chen Zhi has, in more than fifty pages, engaged in a detailed discussion of the dispute concerning nán, and we could not have written this section without his important contribution to the topic.72 The 1983 discovery of a set of bells, the *Shèngliù-zhÅng, in DÄntú, JiÄngsÅ«, is vital as they show an interesting overlap with the song âGÇ zhÅngâ é¼é âBeating the drumsâ of the âYÇâ section of Máo (Máo 208).73 The Än Dà ShÄ« further informs the way we understand the concept of âNánâ, which differs from existing scholarship.
ZhÅu and Shà o are state-names; this is certain. But ever since (at least) the Western Hà n, their exact location has been debated. ZhÅu and Shà o were probably neighbouring states with an east-west relationship. With the exception of QÃn, which was farther west, they were to the west of other states. To locate the two, then, partly depends on how one understands the date of the songs, which is a controversial issue, and how one understands nán å.
As Hà n commentators suggested, we think of nán as a directional termââsouthâ, or âsouthernâ vis-à -vis the ânorthernâ positions of ZhÅu and Shà o.74 Since it is well-accepted in Western ZhÅu oracle bone and bronze studies that ZhÅu was located just south of Mount Qà å²å±±, it seems justified to locate it broadly there, and this in turn implies that Shà o was somewhere nearby.75 We have placed them accordingly on our map (Fig. 1). We thus disagree with hypotheses which locate them between modern-day Nányáng in Hénán province and JÄ«ngzhÅu in HÇbÄi, or which suggest that âZhÅu Nánâ was in Luòyáng, and that Shà o, accordingly but awkwardly, was to its west.76
The notion that the designation nán means âsouthernâ has led to a push in scholarship to locate ZhÅu and Shà o yet further south. As one can see from the map, no state occupied the expansive stretches of territory ranging from the ancient locations of ZhÅu and Shà o all the way south to the JiÄng river, and southeast into the western bank of the RÇ and the Huái river valley.
As is apparent from the Än Dà ShÄ« manuscripts, when compared with *KÇngzÇ ShÄ« lùn, the ShÄ« were a well-structured entity by the time of the Warring States period, with loose ends perhaps, but ends that have no bearing here.77 Texts such as the archaeologically-obtained ZÄ« yÄ« show that, by the fourth century BC, ShÄ« were referred to by section, not just individual songs, which further confirms their stable organisation at the time.78 The ZuÇ zhuà n, admittedly retrospectively, considers the âTwo Nánâ to have had organisational stability by the sixth century BC, and assumes it to be common Warring States knowledge that they were a conceptual unit.79 It is evident, therefore, that the dates of the individual songs must be substantially earlier, else it would be difficult for a community to conceive, understand, and refine these traditions. It seems not exaggerated, therefore, to push their composition, broadly and with at least one exception,80 to some time during the Western ZhÅu period (1045â771â¯BC). We cannot however rule out the possibility that more songsânot just âShà o Nánâ 13âwere composed in the early Eastern ZhÅu (770â256â¯BC), but we are here arguing on the basis of sound moulds and phonetic textures, not the precise semantic rendering of the songs.
The Máo âPrefaceâ asserts that the âZhÅu Nánâ and âShà o Nánâ were âtiedâ to ZhÅu GÅng Dà n and Shà o GÅng Shì.81 This association, if only in its reception, seems uncontroversial.82 It reinforces the notion of the songs having a royal âairâ.83 As crucially established in the Warring States text *KÇngzÇ ShÄ« lùn, the appearance of âShà o Bóâ (i.e. Shà o GÅng Shì/Shà o KÄng GÅng) in âShà o Nánâ 5 (Máo 16 âGÄn tángâ çæ£ âSweet pear treeâ) confirms this association as an emic reading during the latter half of the first millennium BC.84 It might also be the remaining clue needed to associate the songs of âZhÅuâ with ZhÅu GÅng Dà nâalbeit indirectly, since his name does not occur anywhere in them.85
The Máo âPrefaceâ further understands the songs collected under âZhÅu Nánâ as ânorthernâ songs, that is ZhÅu compositions, and takes the word nán å as a noun meaning âsouthâ. However, this was in reference to the âsouthernâ states that were orientated, spatially, to the âsouthâ of ZhÅu. The songs collected as âShà o Nánâ were explained as the ââ¯âFÄngâ-songs of regional lords, taught by the former kingsâ.86 It is left unstated whether this means the songs were composed in outlying states to the south of Shà o and made their way north, or whether, as with ZhÅu, they were composed in Shà o. We feel that the latter is the more likely scenario.
Lyrics which mention place-names of the south are often used as evidence that the word nán denotes the spatial orientation, âsouthâ, which, in the geo-political setting of the time means âthe southern statesâ.87 This strikes us as somewhat difficult to reconcile with the songs. The names of three southern riversâJiÄng æ±, Hà n æ¼¢, and RÇ æ±âoccur in just three of the twenty-five songs,88 while two songs with an overlapping theme (âZhÅu Nánâ 4 and 9) sing of the tall trees âin the southâ. That just four of twenty-five songs sing of the south is hardly convincing evidence to warrant the argument that the songs are from âthe southâ.89 That a song mentions a distant location, or perhaps even uses words with a southern tinge, does not reveal it to be a âsouthernâ composition. Since from at least the time of King Chéng of ZhÅu å¨æç (1042/35â1006â¯BC), the ZhÅu engaged with the southern regions through its military and through hunts, but also socio-economically and culturally,90 including the relocation of élite northern lineages, some of them royal, to conquered lands in the south. It should come as no surprise that ZhÅu élites sung of âthe southâ, even though it never served as a uniform concept during the Western ZhÅu period. Based on later descriptions, to them it was a land of strange things,91 infused with fecundity, and blessed with lush, moist terrain. Singing of the south might therefore show, in a âFÄngâ-style of composition, the differences between different cultural spheres, while revealing, from a royal perspective, hegemonic claims over it. The âYÇââalso old royal songsâoverwhelmingly express the latter, and usually in a distinctively strong tone. From the Sòng dynasty (ADâ¯960â1279) onwards, scholars studying the ShÄ«jÄ«ng therefore began to suggest that the word nán meant âsouthern musicâ;92 but this, too, remains tied to the vocabulary of a minority of the songs.
Commensurate with this, some scholars argue that the graph å depicts the image of a suspended bell.93 The problem with this hypothesis is that there are no early inscriptions in which the graph can be found writing this word. It occurs in oracle bone and bronze inscriptions, but only for different words, most frequently as a rebus to write the word âsouthâ and, less frequently, as a phonetic loan to write the word âpigletâ (hù è±°). In trying to make a case for reading nán å as âsouthern-style musicâ or a type of musical instrument (a âbellâ), scholarship about the Songs calls attention to the âYÇâ-song âGÇ zhÅngâ and the inscription on the *Shèngliù-zhÅng, as well as a passage from the âWén wáng shìzÇâ æçä¸å chapter of the LÇ jì. We read these sources quite differently and propose instead that, in all three instances, the word nán å refers neither to a type of music nor to an instrument, but to âFÄngâ-songs of ZhÅu and Shà o.94
The inscription on the bell (*Shèngliù-zhÅng) is presented in Figure 2 as a rubbing. Below we provide a transcription and translation to further the discussion.
å¯çæ£æï¼ååä¸äº¥ï¼èçä¹å«ï¼å°æ¥
ä¹åè
ï¼æå¥åéï¼ä½éé¾¢éï¼ä»¥äº«äºæå ç¥ã
éæ¯æï¼å å¯åéï¼ä½éé¾¢éï¼æå°ï¼ä»¥ï¼å¤å°ï¼ä»¥ï¼åï¼ä¸é³´åªå¥½ï¼æå°ï¼ä»¥ï¼æ¨æå¿ï¼å®=å·³=ï¼å=å«=ï¼æ°¸ä¿ç¨ä¹ã
It being the kingâs first month, first auspicious [day], DÄ«nghaì, the grandson of King ShÅ« and son of FÇ of Xún-ChÇ, Shèngliù, selected his finest metal to cast harmonious bells, in order to present offerings to our former ancestors. Having selected two pure metals, which indeed are fine metals, harmonious bells were cast. We use them to intone [songs of] the Xià (> YÇ) and the Nán, and they (the bells) have a balance of sounds that are truly pleasant, which thus cause our hearts to be joyous, all the way until the end. [May] sonâs sons and grandsonâs grandsons safeguard and use them in perpetuity.
The line âæå°ï¼ä»¥ï¼å¤å°ï¼ä»¥ï¼åâ corresponds with â以é 以åâ in the song âGÇ zhÅngâ of the âYÇâ section in Máo (Máo 208), and it is perfectly clear that the two lines say the same thing. The ShÄ« line informs our reading of xiÃ å¤ in the bell-inscription as é . This is not only based on their phonetic proximity (å¤: OCM *grâÉ; é : OCM *ÅrâÉ), but also by the *KÇngzÇ ShÄ« lùn, which writes the word yÇâas in the âYÇâ-songs of the Máo ShÄ«âas å¤. The progression of the inscription makes it apparent that the line in question speaks to how the set of âharmonious bellsâ (é¾¢é) were used, not that the donor was informing his descendants of the specific types of instruments that had been cast (or, even less so, that it meant the descendants ought to sound the bells, in general). In any case, the use of å¤(é ) for a type of bell is unattested, which severely weakens, if not totally undermines, a reading of å in its so-called primary sense.95
The Máo commentary understands the word nán å in âGÇ zhÅngâ as a type of âsouthernâ-style dance, but this too is not persuasive as it also takes no account of yÇ é . Some scholars have therefore suggested to read xiÃ å¤ in the inscription as it is written, and to take the pair of words å¤ and å as referring to the âmusic of the central [states] and southern [states]â.96 We acknowledge the plausibility of this reading, but cannot resolve to accepting its vagueness. Instead, like Yáo Jìhéng, we take é and å in âGÇ zhÅngâ as referring to songs of the ShÄ«, namely its âYÇâ and âNánâ songs.97 We think this also applies to the bell inscription.
This reading allows us to go one step further in speculating about nán. It seems significant that an élite family in sixth-century BC JiÄngsÅ«, a southern state (âShÅ« èâ on the map, Fig. 1), used bells to set the âYÇâ and the âNánâ to music. This indicates more than just the cultural dominance of the ZhÅu across a wider area: it confirms the conceptual stability of the core sections of the ShÄ« at the time. We know of the ZuÇ zhuà n story, retrospectively dated to 544â¯BC, which relates how an official from the southern state of Wù, Jì Zhá, upon visiting the northern state of LÇ to observe (and learn) ZhÅu music, witnessed a selection of ShÄ« songs in performance.98 But what else happened during his stay? Did he just listen to a single concert and leave? Was he able to learn it and take it back with him, or even request a copy of the sheet-music and lyrics? Upon his return, would he not have reported his achievement and relate to his lord what he had learned? Would his lord then not have called his music masters so they could learn from him?99 That Jì Zhá requested to observe ZhÅu songs demonstrates knowledge of these ShÄ« songs in the south. Shèngliù and his family too knew of the songs. He commissioned a large set of bells to share his joy in them with his ancestors, some of whom, being kings, would also have known them. Our reading therefore suggests that members of the aristocratic élite who were resident in âthe southâ at around 500â¯BC knew of the âNánâ, and had them performed to music.
In summary, we recognise that the âNánâ songs of âZhÅu Nánâ and âShà o Nánâ display striking organisational consistency between Än Dà ShÄ« and Máo: âZhÅu Nánâ and âShà o Nánâ come first in both versions; the sequence of these songs is identical; and their sound moulds are fully stable. This is unlike any other state recorded in the Än Dà ShÄ«, and suggests that the two states and their songs were afforded special status by at least the fourth century BC. Transmitted literature moreover confirms that by the sixth century BC, âZhÅu Nánâ and âShà o Nánâ were regarded as a conceptual pair. Texts of the time would not normally speak of either âZhÅu Nánâ or âShà o Nánâ in isolation, and later texts would express this ancient notion as the âTwo Nánâ äºå. As confirmed moreover by *KÇngzÇ ShÄ« lùn and ZÄ« yÄ«, the ShÄ« were structurally stable, in four sections, by the second half of the first millennium BC.100 The âNánâ held a special status, and the consistent evidence for that status leads us to infer that the inscription of the *Shèngliù-zhÅng speaks not of nán-music, but of putting to music the ShÄ« songs of the âNánâ and the âYÇâ.
This reading is validated by the close parallels between the inscription and song 208 of the Máo recension. In particular âYÇâ, âSòngâ (of the ZhÅu), and the âTwo Nánâ represent, time-honoured royal songs. As such, it is little wonder that they were named, generically and as a pair, in a bell dating to the late Spring and Autumn period. It shows that âNánâ in âZhÅu Nánâ and âShà o Nánâ cannot mean music, let alone a musical instrument. These were geographical designations, and their special status is indicated because no other state carries âNánâ as a qualification.
ZhÅu and Shà o were northern states. Their regional qualification as âsouthâ (nán) strikes us as unconvincing, unless it means âthe songs south of ZhÅu and Shà oâ, and so, âthe songs (of ZhÅu and Shà o) extending southâ (to the ZhÅng nán mountains, lit., âEnd of the South Mountainsâ). This reading might seem plausible, but it is difficult to reconcile with the special place afforded to the âNánâ in the ShÄ« as a conceptual pair, especially when juxtaposed with the old royal songs of âYÇâ. As such, we suggest reading nán in a normative sense as âsouth-facingâ. Conceptually, nán signifies the position of a ruler and reaffirms the royal status of the songs. In claiming royalty, the âNánâ thus affirm the asserted cultural dominance of the ZhÅu and Shà o over the south. This reading, insofar it is accepted, resolves contradictions that other readings are unable to avoid.
9 The Significance of the Än Dà ShÄ«
As mentioned, correspondences with ShÄ« in other Warring States texts abound, particularly in philosophical texts. But rather than being isolated support material for arrangements in other contexts, the Än Dà manuscript is the earliest extant iteration of the Songs. This is significant. The manuscript offers a unique glance on the standing of ShÄ« during the Warring States period. Again, this does not imply that the Än Dà manuscript shows âtheâ ShÄ« of the time. It may not even be representative. But it was certainly one instantiation of the ShÄ«, and so shows how at least one conceptual communityâthe Warring States textual community involved with the production and reception of this manuscriptâengaged with the text.
Perhaps it is to be expected that by the fourth century BC the Songs also circulated in writingâat least its core parts, such as the âTwo Nánâ that served as the conceptual normative royal pair as early as the second half of the first millennium BC.101 These carry, among others, the single most influential song of the entire history of China: âGuÄn jÅ«â (âZhÅu Nánâ 1). After all, this was the time when Chinaâs manuscript cultures matured,102 when ever more texts were put into writing, and when philosophical cogitation became a written exercise, at least in part, as complex transmitted texts such as the XúnzÇ or the manuscript texts from a tomb at GuÅdià n, HúbÄi, forcefully show.103 Traditions of Yì and ShÅ« were written down,104 as was historical knowledge.105 We see short philosophical aperçu,106 long disquisitions on music and dance,107 prayers,108 and literary compositions with verse.109 Why should the ShÄ« form a striking exception and not exist in written form? They would not. And yet it is only now, with the physical actuality of the Än Dà manuscript, that we can be more assertive about the Songs and their use during the Warring States period. The earliest extant collection of the Songs forces us to rethink some of the assumptions we may have long held dear, but which can no longer be maintained under closer scrutiny.
Prior to the acquisition of the Än Dà ShÄ«, the earliest copy of ShÄ« was from the first half of the second century BC (no later than 165â¯BC).110 But this copy, excavated from the Western Hà n tomb of Xià hóu Zà o at ShuÄnggÇduÄ«, is extremely fragmented and poorly preserved. It has limited value as a text. The Fùyáng éé½ ShÄ« shows how the Songs were organised during the Western Hà n,111 in the sense that individual songs in this manuscript are recorded with titles and word counts.112 The macro-organisation of the Än Dà text confirms this tradition of collecting songs under the rubric of states, arranging them in a sequence, and tabulating their total. That the first song in two of the Än Dà âstateâ sub-sections are named further implies that other songs in the manuscript text may have had names too.113 Alternatively, these names might refer to the respective groups of songs in a section, perhaps in a given order, but because âZhÅu Nánâ does not list the title of its first song, the songs association with that section do not appear to have relied on the name of its initial one.114 This complements the picture gained from *KÇngzÇ ShÄ« lùn, which slightly pre-dates the Än Dà ShÄ«, and which identifies its songs by title.115
While the Än Dà ShÄ« clearly differsâsometimes spectacularly soâfrom the received Máo recension in terms of its image programme, and often even semantically, the organisation of the songs and the close matches of their corresponding sound moulds suggests that both versions originated from another, and likely earlier but definitely independent, source. The *KÇngzÇ ShÄ« lùn, which confirms the existence of ShÄ« in sections during the latter half of the fourth century BCââBÄng FÄngâ, âXià (that is, YÇ)â, and âSòngâârecords fifty-eight songs, a meare 19 per cent of the 305 said to have been selected by KÇngzÇ and included in Máo. Amongst these, there are thirty songs from twelve states, with one song of uncertain affiliation.116 As mentioned, the Än Dà has songs from six states; fifty-seven in total. Twelve songs of *KÇngzÇ ShÄ« lùn overlap with the Än Dà . Conservatively we can therefore say that during the Warring States period at least these twelve songs were conceived as belonging to the tradition named ShÄ« by Máo, and by at least two different communities. More likely, however, is the scenario in which there were no fewer than seventy-five ShÄ« âSongs of the Statesâ, and perhaps as many as 103 ShÄ« songs circulatingâalso in written formâin the fourth century BC. Whetherâand if so, whyâthe Än Dà ShÄ« thus presents a âselectionâ from a larger corpus of ShÄ« is open to question. The *KÇngzÇ ShÄ« lùn strongly suggests this larger compendium closely resembled the Máo recension in organisation, and may also have contained many of the songs recorded there. Whether they also resembled each other linguistically, we do not know. However, what is clear from *KÇngzÇ ShÄ« lùn glosses is that the main themes of the songs correspond closely. We are therefore confident enough to assert the following:
During the latter half of the fourth century BC, ShÄ« were understood widely as partitioned into the sections known from the Máo recension; the Än Dà ShÄ« represents a selectionâreasoned or notâfrom a larger corpus of ShÄ«; third, there were also written instantiations of ShÄ«-selections circulating during the second half of the fourth century BC.
The exact relationship between this manuscript and its historical owner remains unknown. We have no information about the provenance of the manuscript, and, at the time of writing, only a preliminary report about what else was in the ownerâs possession has been published.117 That someone owns a ShÄ« manuscriptâamong other textsâat a minimum suggests an interest in the subject matter, whether that was generated by the manuscript, or the manuscript was acquired for that reason. But it does not mean that the person necessarily âstudiedâ the ShÄ« to acquire an advanced level of proficiency in the text or its tradition. If we suppose the punctuation marks in the manuscript were introduced by the holder, then we have every reason to think that the owner and user of the manuscript were the same person, with the implication that they read at least parts of the text closely. But what if the punctuation marks, or some of them, were not introduced by the owner? In that case, we posit, there were ShÄ« texts circulating as âreaderâs editionsâ, however preliminary they might have been. This too is significant; it indicates an interest in, and market for, ShÄ« manuscripts for reading with textual support provided. But was the owner simply reading the ShÄ«? Or were they having them intoned for whatever performative purpose?118 Or else, was this manuscript just something someone planned to read but never actually did? Was the text of the manuscript an attempt at establishing a new ârecensionâ? Is it a verbatim copy of a pre-existing one? Or is it an idiosyncratic rendering of the ShÄ« in writing? We can only muse on these and other questions.
Thus, the Än Dà ShÄ« manuscript returns us to the heart and mind of studying the ShÄ«. It forces us to rethink questions we had long thought settled (or perhaps we had merely thought as futile to revisit). From an emic perspective, centred on Warring States use of the ShÄ«, it makes us rethink how the ShÄ« worked on the ground, within a community; it makes us rethink how writing and the sound texture can be complementary and serve a songâs image programme, and how they fit within the sound moulds of ShÄ« production; it made us rethink, as we were writing a book about the songs of the âZhÅu Nánâ and the âShà o Nánââthe old royal songsâabout what å nán actually means.
Since at least JiÇ Kuà è³éµ (ADâ¯30â101), scholars have paid close attention to variations in ShÄ« recensions.119 Before the Än Dà , this mainly focused on texts from four ShÄ« âschoolsâ or traditions during the Western Hà n, but for which only Máo is complete and has any real value as a text.120 Although the Än Dà ShÄ« only contains songs from the âstatesâ, the kinds of variation seen in this manuscript increases our knowledge and understanding of ShÄ« traditions pre-QÃn, for which, aside from names and affiliations, we knew very little.121
Important questions remain. A controversial issue is perhaps what the Än Dà manuscript actually is. What was its purpose? To whom did it speak? Here we can only speculate. It seems clear, however, that the precise execution of the graphs served as an additional layer of meaning, and demonstrates the writerâs close familiarity with the textâa text perhaps passed on from teacher to student, with explicitness expressed in the written word, in many ways exaggerating, but never ridiculing, the image programme of the songs. The result is a unique mode of ShÄ« instruction, embedded into the text proper, that transmits principles of interpretation and conveys how the community for whom the text was written understood the deeper levels of its meaning. The Än Dà ShÄ« therefore signifies the ShÄ« reception of a particular conceptual community. We reiterate that this does not mean it was the dominant, let alone sole, articulation of the Songs as a recension within that community. Just like a good reader of poetry today can violate the rhythm of a poem as they stage it, an attentive student in antiquity may have seen different connections with the songs than they learned from their teacher, and passed them on accordingly. Yet variant texts still remain part a community. Different versions of the ShÄ« mightâand in all likelihood didâexist within a given community, with the Än Dà manuscript articulating just one of its instantiations.
10 The Songs of the Royal ZhÅu and the Royal Shà o: Conventions
We offer translations of the first twenty-five songs of the Än Dà ShÄ«, the songs of the âZhÅu Nánâ and the âShà o Nánâ; the old royal songs. For reasons outlined in our Introduction, we do not provide an interpretative transcription of the graphs, but instead render them as closely as possible to what was produced on the bamboo. For purposes of comparisonâand this is for the modern reader, not due to our methodologyâwe also reproduce the text of the Máo recension. We only translate the songs of Än Dà ShÄ«. Indeed, we only translate the extant text of the Än Dà manuscript. If a slip is corrupted due to material loss or other factors, we do not attempt to reconstruct the missing text on the basis of Máo.
We take the edition of ÄnhuÄ« University Manuscripts as our base text and justify in our comments where we deviate from the choices made by its editors. The translations are accompanied by a brief description of the song as it occurs on the slips, together with a portrayal of its phonetic texture, and a short summary of its image programme.
We also provide the necessary philological apparatus from which to defend our reading. We do this per stanza, as we see them. We close each song with a brief commentary.
Where, as in âZhÅu Nánâ 9, a song is listed as one song in Máo but appears as two in the ÄnhuÄ« University Manuscripts, we differentiate them with a superscript âAâ and âBâ, to maintain consistency with subsequent songs in Máo. Note, however, that this is for convention, not because we believe in a primacy of the Máo recension.
Some further remarks on our representation of features of the manuscript: a forward-slash or reverse solidus (\) indicates the beginning of a slip broken at the top; a back-slash or solidus (/) marks the end of a slip whose tail is missing; a vertical line (|) marks the beginnings and ends of complete slips; a full-stop in bold with a larger point-size (.) reduplicates the reading mark found at the right side of the slip; reduplication marks are provided as an equals sign (=); and the marker for the end of a song is reproduced as a black square (
).
When we refer to the Songs as a cultural institution, to which various user communities could relate in their own ways, we write ShÄ« (i.e., capital letter but not italicised). When we refer to a particular recension or instantiation of this cultural institution by a particular textual community, we write ShÄ« (as in Máo ShÄ«, or Än Dà ShÄ«, i.e., italicised capital letter). When we refer to an individual song within a recension or particular instantiation, we do not capitalise or italicise âsongâ.
ÄnhuÄ« Dà xué cáng Zhà nguó zhújiÇn å®å¾½å¤§å¸èæ°å竹簡 1. The manuscripts were purchased by an unknown buyer, possibly on the Hong Kong antiquities market, and then donated to ÄnhuÄ« University. We have no precise information as to where in China the manuscripts were found.
The circumstances thus bear regrettable similarities to the two famous Warring States manuscript collections from Shà nghÇi Museum and Tsinghua University. The former collection of ChÇ manuscripts was acquired in 1994. It contains some 1,200 inscribed bamboo strips, which the Shà nghÇi Museum started publishing in 2001 (Shà nghÇi bówùguÇn cáng Zhà nguó ChÇ zhú shÅ«). The Tsinghua collection was purchased in the summer of 2008 in Hong Kong. It contains around 2,500 bamboo slips, and features ShÅ«-type texts of which some have close counterparts in the transmitted Shà ngshÅ«, Yì ZhÅushÅ«, as well as in annalistic texts. QÄ«nghuá Manuscripts: QÄ«nghuá Dà xué cáng Zhà nguó zhújiÇn.
Goldin 2013: 156n6 introduced the neologism of âunproveniencedâ to distinguish between place and owner of non-scientifically obtained manuscripts. However, unlike ancient Egyptian collections that may have been passed on for generations, or even centuries, the Chinese case is not assisted by this differentiation and so we refrain from using it here.
A number of projects which are currently ongoing deal explicitly with the issue of unprovenanced texts and their academic use. See, among others, Brindley and Flad; Meyer, Metcalf, and Rota.
Complete slips measure on average 48.5â¯cm by 0.6â¯cm. They were connected by three binding straps and show prepared grooves on the verso. The total number of graphs per slip ranges from twenty-seven to thirty-eight, depending on the length of a song and its layout.
These are: slips 18, 19, 23, 24, 26, 30, 56â58, 60â71, 95â97.
ZhÅ« Fènghà n 2020: 62 suggests, with reference to Huáng DékuÄn 2017a, that there were also âXiÇo YÇâ å°é âLesser Elegantiaeâ songs in the Än Dà ShÄ«. Revisiting Huáng DékuÄn 2017a we could find no such reference, as the article confirms there are no other items from sections of the received ShÄ«jÄ«ng (âXiÇo YÇâ; âDà yÇâ 大é âGreater Elegantiaeâ; or âSòngâ é âHymnsâ) in the later volumes of ÄnhuÄ« University Manuscripts. We can therefore only posit that ZhÅ« Fènghà n was referring to the materials classed as âChÇcà typeâ by Huáng DékuÄn.
The editors (pp. 1â2) propose that songs from âsomeâ states were recorded on slips 60â71, in total approximately 380 words; while their speculation seems justified, we have no means to validate or contradict it. In response, and when referring to the sequence of the statesâ songs after âQÃnâ (3), we add an asterisk* ahead of the stateâs name to indicate the possibility that it was one place further removed from âZhÅu Nánâ (1).
These are all songs of the Máo recension. However, we felt compelled to split one song, 9, conceptually into 9A and 9B. The split song is not included in the total count of the âZhÅuâ songs or the total number of songs in the manuscript. Song *10 (almost certainly a version of Máo 10 âRÇ fénâ æ±å¢³ âRaised banks of the RÇ Riverâ) is missing. Slip 20 records âZhÅu Nán, elevenâ.
Slip 99 records a total of nine songs, only seven of which remain; as ten songs from this state are included in Máo, we cannot be sure which two are missing.
Slip 117 records âWèiâ as having a total of nine songs, but the text has ten. This updated number is included in the total count of fifty-seven songs.
âHóuâ, thus, perhaps refers to the area of QÅ«wò-HóumÇ æ²æ²ä¾¯é¦¬; this agrees, in principle, with the findings of Xià Dà zhà o who proposes that âHóuâ referred (self-referentially) to the state of Jìn æ. See also Xià Dà zhà o 2018: 119â125 and Xià Dà zhà o 2020: 5â15.
See further MÇ YÃnqÃn 2020.
âYÇngâ ç¬ (Máo: YÅng é) is the location of old capital of ShÄng; âWèiâ é is in western ShÄnxÄ«, etc. ShÄng oracle bone inscriptions refer to the area later called YÅng é (perhaps subjectively) as âcentralâ.
Although the designation âTwo Nánâ äºå for the âZhÅu Nánâ and the âShà o Nánâ became popular only during the Hà n dynasty, the two were treated as a conceptual pair in the early literature. See, for instance, LúnyÇ âYáng Huòâ 17.10; ZuÇ zhuà n âXiÄngâ 29.2; LÇ shì ChÅ«nqiÅ« âYÄ«n chÅ«â é³å 2; thrice in the YÃlÇ (see footnote 71).
ÄnhuÄ« University Manuscripts 1: 2.
We say more on this later in the Introduction.
They are âZhÅu Nánâ 3 and 5; âShà o Nánâ 8 and 11.
Huáng DékuÄn 2017a: 56â58.
Qiú Xīguī 2012a: 339.
We refer to Western Hà n ShÄ« traditions of LÇ, QÃ, and Hán as âmajorâ, and other, lesser-known ones, as âminorâ, thus drawing into parallel Jao Tsung-iâs conceptualisation of the Yì æ (Changes), as attempted in Jao Tsung-i 2009, vol. 4: 17.
Kern 2011: 21.
Lù JÄ« é¸ç (third century AD), Máo ShÄ« cÇomù niÇoshòu chóngyú shÅ« æ¯è©©èæ¨é³¥ç¸è²éç, 70: 21.
Karlgren 1931: 18â20; he used the term âXúnzÇ-colouredâ. Although we accept this view, it must be acknowledged that the pedigree of the LÇ tradition can also be traced back to XúnzÇ, and the Hán ShÄ« wà i zhuà n frequently quotes him as an authority.
Karlgren 1931: 26â35. The tabulation does not include repetitions and quotations from lost songs.
We found the best system is to reconstruct OCM (for âMinimal Old Chineseâ) after Axel Schuessler 2007, 2009.
*KÇngzÇ ShÄ« lùn is published in vol. 1 of the Shà nghÇi Museumâs Warring States Manuscripts. We use the title *KÇngzÇ ShÄ« lùn as a convention, not because we believe that the text actually had this title or that KÇngzÇ was the author. Following LÇ XuéqÃnâs (2002) slip order, it seems that KÇngzÇ was not the writer but rather was cited as an authority. In our opinion, the text was produced as an instruction to the ShÄ«, likely during the Warring States period; see also Kern 2015: 186. Reading the *KÇngzÇ ShÄ« lùn and Än Dà ShÄ« against one another is a novel approach to Warring States ShÄ« that was not available before the recovery of the Än Dà ShÄ«. *WÇ xÃng äºè¡ from tomb 1, GuÅdià n, is also valuable, but relying on it is risky as in the *WÇ xÃng the Songs are not the principle focus of the elaboration, but rather serve to support a specific philosophical argument. One must take account of the different needs and goals of the community that produced a given text and their use of the Songs. (More on this point below.) The version of the ZÄ« yÄ« ç·è¡£ texts found at GuÅdià n, and the ChÇ manuscripts in the Shà nghÇi Museum collection, were compiled from three separate sourcesâShÄ«, ShÅ« æ¸, and master sayings, zÇ yuÄ åæ°âand forged into archival units that store a learning repertoire that was relevant to a particular community (Meyer 2021: 74). But in these units the sources work in a co-ordinated fashion with one another, not hierarchically, and the ZÄ« yÄ« does not articulate ShÄ«-instruction. Whether this archival focus served argumentative ends is unclear. We therefore do not use it as a primary tool for an emic reading, as we do the *KÇngzÇ ShÄ« lùn.
In an important article, Kern 2010 discussed the difficulties in establishing the Shī pre-Máo, for which he had to rely predominantly on palaeographic materials commenting on the Shī, not the Shī itself.
Pike 1954.
In linguistics, for instance, Nöth 1995: 183 defined the âemic unitâ as an âinvariant form obtained from the reduction of a class of variant forms to a limited number of abstract unitsâ. The distinction between emic and etic in linguistics was coined by Kenneth Lee Pike. See in particular Pike 1943.
Cf. Hays 2012: 6 for the notion of establishing an âemicâ reading of Egyptian hieroglyphic material. See also the seminal essay of Harris 1976.
Conrad 2006: 46 notes âthe emic approach investigates how local people thinkâ, which implies how they conceptualise the world and their place in it, and which actions they take from there.
Bakhtin 1986: 162.
Meyer 2021: 113.
For a detailed discussion of âconceptual communitiesâ see Meyer 2021.
See Kern 2010 for a related observation.
Note that we are not making a claim for multiple writers, only that we have detected at least two orthographic manners. We acknowledge the possibility that a single writer could write in different styles, and that graphic variants occurring in the same manuscript text might carry with them a complex history of transmission (for instance from abroad) or conversion (for instance into a local script). Stated briefly, and with an awareness that more detailed study is required, basic handwriting analysis discerns two distinct forms of æ, signifying the high-frequency word wÇ âmy, mine; I, usâ; and two different styles for å, signifying the high-frequency word yòu âright handâ > yÇu âhaveâ æ, among notable examples. See ÄnhuÄ« University Manuscripts 1: 224â225, 293â294.
The French term âscripteurâ would also be apposite, but we see no advantage over âwriterâ.
We provide a translation of the song on p. 14 below.
See also âZhÅu Nánâ 2, stanza 1, annotation 3 (p. 69).
For more information, the reader may wish here to consult âZhÅu Nánâ 2, stanza 1, annotation 5 (pp. 70â71).
Richter 2013: 48 supplies a prescriptive account of such transcription format.
We borrow the idea of âimage programmeâ from Schwartzâ 2018 studies of the YìjÄ«ng. We understand the precise semantic, lexical content in a songâs lyrics to be an image, and the intersection of images in a songâs lyrics as an image programme. We have chosen the word âprogrammeâ because of its association with systematic application and system-based learning. A songâs images form a series of instructions that control the operation of the songâs meaning. Note that semantic content refers not only to the words signified by graphs but also, and sometimes crucially, to the use of signifiers in the structure of the graphs and their significance.
This observation is on a surface level diametrically opposed to Kernâs (2010: 45) finding that âthe individual graphs in these binomes are utterly irrelevant; as has long been notedâ (Kern is citing Kennedy 1959: 190â198 and Knechtges 1987: 3â12), âsuch descriptive rhyming, alliterative, or reduplicative binomes cannot be decoded based on the meaning of each characterâ. But in his studies Kern was largely working with different commentarial traditions. While we admit that the graphs chosen had to agree to the productive mould, as we call it, of the ShÄ«, and thus they were guided by prescriptive rhyming and alliteration principles set by this mould, we cannot exclude the possibility that the various textual communities chose them with great care and for the meanings they represent, and that their choices may well conflict with those of other groups. On this point, the reader should please consult our discussion of yÄo dà è¦ç¿ in âZhÅu Nánâ 1, stanza 1, annotation 3 (pp. 54â56).
In the Máo recension, eight words are written with graphs using a âmountainâ or âhillâ signifier.
The first instance is a rhyming binome in Än Dà ShÄ«: 
éº, which corresponds to Máo: huÄ«-tuà èºé¤. Even though the word is the same, the difference between the two is that in the Än Dà ShÄ« the second syllable of the binome is written with a âmovementâ signifier, adding further signification and depticting the movement of the âhorseâ, whereas in Máo it is written with a âhillâ signifier, adding another layer of signification to the âmountainâ images serving as the background for the manâs travails, and thus responding to the different image programme in Máo. The second instance is the use of å¾ in Än Dà ShÄ«, again written with a âmovementâ signifier, which corresponds to ç in Máo, written with a âsicknessâ signifier. Thus, the Än Dà text maintains consistency in describing the horseâs belaboured movement with a âmovementâ signifier. (See âZhÅu Nánâ 3, stanza 3, annotation 2 for more on the âsicknessâ signifier in the written instantiation of the song lyrics in Máo and the other major Western Hà n recensions.) The graph ç in Máo, writing the word jÅ« ârockyâ, and thus maintaining consistency with its mountain-related image programme, is written with a âwaterâ signifier in Än Dà ShÄ«: ð£³. We read 𣳠as it is written, âslippery slopeâ, primarily because it is just as sensible, but also considering the image programme of the horse under duress at this point in the song.
Note that this observation is not necessarily at odds with the findings of Galambos 2006, but it draws different conclusions. Based on covenant texts from the fifth century BC, but confirmed by wider selections of manuscript texts, Galambos traces a general instability of a graphâs signifier in manuscript writing. While such instability is apparent in the Än Dà ShÄ«, signifiers were not chosen without consideration; see, for example, the comments of Shaughnessy 2021: 24â25 on the consistent use of âmouthâ signifiers to write out the song lyrics of âYÇngâ 2 (Máo 46 âQiáng yÇu cÃâ çæè¨ âOn the outer wall there is three-horned vineâ). (The lesson of the song, if read through the *KÇngzÇ ShÄ« lùn [slip 28], is an admonition to be âbe tight-lipped so that others do not know of [oneâs] wordsâ [æ å¯èä¸ç¥è¨]; not doing so is cause for gossip which, in turn, leads to âhumiliationâ, âperpetuationâ, and âdisgraceâ that cannot be âbundled upâ, âremovedâ, or âbrushed awayâ, as per Máo; we discuss the song in Meyer and Schwartz forthcoming 2023â2024.) See also Schwartz 2018b: 1189â1190, who discusses signifier exchange in the Yì traditions in a specific case study centred on trigram Gèn. Xià Hányà (Edward Shaughnessy) 2012 also pointed out possible correlations between the ShÄ« and Yì traditions.
As KÇngzÇ put it: [Through the study of ShÄ« one] âbecomes more knowledgeable about the names of birds, animals, plants, and treesâ å¤èæ¼é³¥ç¸èæ¨ä¹å; LúnyÇ âYáng Huòâ é½è²¨ 17.9.
While the addition, or accumulation, of signifiers writing a single grapheme is not at all new to the study of Chinese palaeography, the systematic use of certain signifiers that add meaning, either to a discrete, literary text, or across a genre produced at approximately the same time and in the same geographical location, is an issue of Warring States-Western Hà n studies of excavated materials that only became more widely recognised (but not necessarily accepted!) after the discoveries of the MÇwángduÄ« (1973) and GuÅdià n (1993) manuscripts and the ZhÅngshÄn (1977) bronzes. Discussing the MÇwángduÄ« *WÇ xÃng manuscript, for instance, Csikszentmihalyi 2004:169 notes a peculiar way of writing shÄng è² âsoundâ as shèng è âsagaciousâ which, he hypothesised, was not just a phonetic loan but an âunstated philosophical argumentâ. Scholars who work with excavated Warring States inscriptions and manuscripts are aware of an interpretive tradition that sees significance in the relatively large number of graphs composed with a xÄ«n å¿ âheartâ signifier in the ZhÅngshÄn bronze inscriptions, and in the philosophical manuscript texts produced during the latter half of the fourth century BC; see Liú Xiáng 1996; Páng PÇ 2000, 2011; and Liú BÇojùn 2020. Jao Tsung-i 2005 calls attention to the frequent use of words in the *KÇngzÇ ShÄ« lùn written with a âheartâ signifier, and associates them, collectively, with âthe range of different feelings one has when reading the ShÄ«â. Many such observations have thus far remained isolated, and have not received due attention because of the intense focus on phonetic writing over the past years.
Martin Kern and Edward Shaughnessy have launched a sometimes heated debate about whether the Songs were âoralâ or âwrittenâ in nature. We do not wish to partake in it. As this written instantiation of the Songs shows, there were clearly also written songs circulating among groups who may equally have been guided by oral primacy.
The model of a writing-supported text performance is laid out in Meyer 2021: 15. We further comment on this later on in the Introduction.
It is at this stage not possible to state how close a match was required for a chosen word to be phonetically acceptable, as this may also have differed from one Warring States-period community to another. However, a close reading of the Än Dà text against the texts of the Hà n traditions, of which Máo is dominant, indicates at least minimal adherence, at the word level, to a pre-existing rule about the agreement of the vowel sound(s).
Huáng DékuÄn 2018; Shaughnessy 2021.
See our notes on âShà o Nánâ 14, stanza 1, annotation 1 (193).
Huáng DékuÄn 2018: 73.
âYÇngâ 1 in the Än Dà version is composed in *two stanzas, each stanza in seven lines, each a line of four words. Whereas line 3 of each stanza in Än Dà writes 
ç®å
©
âSubmerging are those two wild ducksâ, the corresponding line in Máo is 髧彼å
©é«¦ âFalling down are those two tufts of hairâ (said of a young man). A comparison between the two shows that the phonetic value is largely stable, but that there is a drastic difference in the selection of its written words; these selections, in turn, significantly affect the songsâ overall meaningânot in their aurality, but in how the written instantiations express and produce meaning. Employing philological methods to try and equate the two sentencesâand thus the two versions of the songâis unwarranted, and doing so negates the ingenuity of the ShÄ« communities that produced them. We discuss the song in more detail in Meyer and Schwartz 2022/2023.
On this song see also above, âWriting the Image Programme of the Songsâ.
Legge 1961: 8.
Of the ten songs of the âZhÅu Nánâ extant in the Än Dà ShÄ«, seven have exactly the same rhymes as those in Máo; these are âZhÅu Nánâ 1â5, 9, and 11. Songs 6 and 8, each composed in three stanzas, show rhyme disparity only in the final stanza, but the disparity in âZhÅu Nánâ 6 is minimal and the vowel is still the same (the rhyme in Máo is in *-in; in Än Dà , it is in *-î/-in). Due to the inconclusive identification of one of the rhyming words in stanza 2 of song 7 (Máo 7) we are unable, at present, to use this in our tabulation; rhymes in the other two stanzas however, like those in songs 6 and 8, agree with the sound moulds in Máo. Song 10 (= Máo 10) is missing, as is stanza 1 in song 11. Thus, of the total number of stanzas in the Än Dà âZhÅu Nánâ, words in prescriptive rhyming moulds agree with Máo in twenty-eight of the twenty-nine instances, or 97â¯% (note that this count parses âGuÄn jÅ«â, âZhÅu Nánâ 1, in three stanzas, as per the Máo recension, and not in five, as per Zhèng Xuánâs reading.) See also Chéng Yà n 2020.
Huáng DékuÄn 2018: 71.
The songs of âZhÅu Nánâ and âShà o Nánâ are said to be âwhere sages areâ in JiÄo Gòngâs ç¦è´ (Western Hà n) Yì lÃn ææ; see Shà ng BÇnghé (ADâ¯1870â1950) 2005: 7.499. Songs of the âcommonâ states are never referred to in such terms.
Note that the point about the aural primacy of the Songs must not be confused with deliberations about whether the ShÄ« were primarily oral or written. Rather, we assert that the purpose of writing the ShÄ« during the Warring States period would not be for âmemorisationâ, but rather for interpretation: for providing semantic fixity to the already fixed aural forms.
We discuss the various cases in our philological commentary on the songs.
Take for instance âZhÅu Nánâ 7 (Máo 7 âTù jÅ«â å ç½ âRabbit netâ), a highly formulaic song of three stanzas. The songâs ârabbit netâ stands allegorically for the martial man, trapping the GÅng and the Hóu. The association of the rabbit net with the martial man is made explicit phonetically: å ç½ *lhâh *tsa rhymes with æ¦å¤« *maÉ *pa, creating the crucial link between them, thus marking them as a unit.
JiÇng Wén 2021 reaches this conclusion.
Note this model differs from Nagyâs 1996: 40 concept of âtextualisationâ, which he understands as âcomposition-in-performanceâ.
This observation would confirm a later date for Máo, suggesting some form of conceptual obstruction in the transmission of the relevant songs from one community to another. However, this is not the concern of this book and so we leave it as a footnote to be taken up at a later time.
We believe this is what is also addressed in the following entry in the ZuÇ zhuà n âXiÄngâ 16.1: æ侯è諸侯宴äºæº«ï¼ä½¿è«¸å¤§å¤«èï¼æ°ï¼ãæè©©å¿ é¡ããé½é«åä¹è©©ä¸é¡ãèåæï¼ä¸æ°ï¼ã諸侯æç°å¿ç£ãã âThe Hóu of Jìn and the many princes were feasting at WÄn when the Hóu made the high officers dance. He said: âThe Songs ought to be sung according to their type. The song by GÄo Hòu of Qà is not in typeâ. Xún YÇn was angered and said: âit is clear, the many princes are not of one mind with usââ¯â. Durrant at al. 2016: 1039, however, translate the speech as âThe ode sung has to match the right order. The ode by Gao Hou of Qi did not match the right orderâ, to which they note (462): âthe âright orderâ ⦠refers to the correspondence between the music and the dance or to the connection between the performance and the proper intent (ZZ 33.573)â. This âmay also include political hierarchy and ritual proprietyâ. This is not tenable. Rather, confirmed by the matrices we see from the Än Dà ShÄ« and *KÇngzÇ ShÄ« lùn, we believe that the songs had to be âin typeâ as they were performed and could not be disarrayed.
See JÄ«n Róngquán 2012.5: 165â168.
Throughout the YÃlÇ (âXiÄng yÇn jiÇ lÇâ é飲é 禮 14; âXiÄng shè lÇâ éå°ç¦® 11; âYà n lÇâ ç禮 21) the songs of the pair are thus referred to as the âproper songsâ (æ£æ). In ShÄ«-traditions, starting no later than its representation in the Máo âPrefaceâ, they are referred to as the âproper windsâ (æ£é¢¨), as opposed to the other thirteen statesâ songs as âdeviating windsâ (è®é¢¨); the Máo âPrefaceâ glosses the meaning of âFÄngâ-songs paronomastically as âwindââmeaning âinstructionââand, employing bi-directional wordplay, understands the songs of the âTwo Nánâ, the proper, royal songs, as having a transformative power, while the ânon-royalâ songs are understood to be satirical (fÄng è«·); see Máo ShÄ« gÇxùn zhuà n 1.6â7; MÇ Ruìchén 1.9â10. Although the songs of âBÄ«nâ è±³, a state located north of ZhÅu and Shà o, are associated with pre-dynastic ZhÅu ancestors, they are still considered ânon-royalâ in the ShÄ« tradition more broadly (FÄng Yùrùn 2017: 29), and their nature as âpureâ âFÄngâ-songs (as opposed to a hybrid style containing âYÇâ elements) is contested (FÄng Yùrùn 2017: 48â49). Since these songs do not occur in the Än Dà ShÄ«-selection, we shall not address them further in this book.
Chen Zhi 2007: 193â244.
Chen Zhi 2007: 197â198; 210â214.
Máo ShÄ« zhèngyì 1.19â20.
Wáng YìnglÃn (ADâ¯1223â1296) 2011: 1.183â185, 191â192; Chen Zhi 2007: 194â196. Cf. Máo 262 âJiÄng Hà nâ æ±æ¼¢ âThe JiÄng and Hà n Riversâ, which says, âOn the banks of the Këang (Jiang) and the Han, the king had given charge to Hoo (Hu) of Shaou (Shao): âOpen up the whole of the country; make the statutory division of my lands there ⦠as far as the southern seaââ¯â, and then, once he had accomplished the charge and was bestowed the gift of a libation cup and aromatic ale in the kingâs presence, the king declared, âAnd [I] confer on you hills, lands, and fields. In [Kâe-] (Qi)chow shall you receive investiture, according as your ancestor received hisâ. (Legge 1961: 553f). (Note that Karlgren 1950: 233â234 comes to a different understanding of this passage.) The âancestorâ mentioned here, if following Leggeâs translation, might refer to Shà o GÅng Shì. An early Western ZhÅu jade dagger recording that âTà i BÇoâ å¤ªä¿ (i.e., Shà o GÅng Shì) received a command (likely from King Chéng of ZhÅu) to âvisit the southern statesâ was discovered in this area in 1902; see LÇ XuéqÃn 1997: 135â141.
Wáng XiÄnqiÄn 2020: 1.1â2, citing a LÇ ShÄ« explanation, and ShÇ jì; see also Chen Zhi 2007: 221â222. In Western ZhÅu bronze inscriptions the ZhÅuâs eastern capital at Luòyáng was called XÄ«n yì æ°é âThe New Settlementâ, Chéng ZhÅu æå¨ âCompleted ZhÅuâ, and Wáng chéng çå âThe City of Kingsâ, the latter of which just might be the referent for the âstateâ-name âWángâ ç (lit. âKing(s)â) in ShÄ« traditions. The songs of âWángâ, absent in the Än Dà ShÄ«-selection, are dated in the Máo tradition to the beginning of the âEasternâ ZhÅu (circa 770â¯BC), when King PÃng moved the ZhÅu seat east to Luòyáng; see Máo ShÄ« gÇxùn zhuà n 4.117. The hypothesis that âZhÅu Nánâ was in, or subsumed, Luòyáng is diminished if âWángâ was also located in this area, although again it depends on how one dates the songs.
The mention of an unknown song in *Táng Yú zhÄ« dà o in tomb 1, GuÅdià n, is a case in point.
See the discussion in Meyer 2021, Ch. 2.
ZuÇ zhuà n âXiÄngâ 29.
âShà o Nánâ 13. See our discussion on pp. 203â204.
Máo ShÄ« zhèngyì 1.8. See also the discussion in Xià ng XÄ« 2016: 109â110.
The fact that LÇ shì ChÅ«nqiÅ« âYÄ«n chÅ«â é³å 2, which partly reflects Warring States ideas, draws the same connections as the Máo âPrefaceâ suggests that the latter inherited a long-established understanding of the songs, shared by different communities.
In this respect it is noteworthy that the Máo commentary to jÅ« jiÅ« é鳩 âospreyâ, in the first line of song 1, glosses it wáng jÅ« çé. This is of course just another name for the bird, not a Máo invention. But what it shows is that Máo is instructing the reader of the gloss in how to understand the binome é鳩: it is a metaphor for the âkingâ, and thus begins the âroyalâ songs! (ééé鳩ï¼èä¹ï¼ééï¼åè²ä¹ãé鳩ï¼çéä¹ âGuÄn-guÄn jÅ« jiÅ« is evocation. GuÄn-guÄn is a harmonious sound. âOspreyâ is the âking-fisherââ¯â.) Furthermore, having identified the âospreyâ as a hidden metaphor for âkingâ, the âgood girlâ æ·å¥³, as the next step in Máoâs hermeneutic reading, was identified as no one other than the queen (åå¦). The Máo commentary thus took what was almost certainly a pre-existing understanding of âGuÄn jÅ«ââhowever historically correctâas something composed during the time of King Wén. The commentary thus revealed linguistic evidence embedded within the songâs lyrics to justify itself. In the Máo tradition, as expressed by the Máo âPrefaceâ, the songs of the âNánâ as a conceptual unit are thus said to pave âthe way of a proper beginningâ (æ£å§ä¹é) and serve as âthe foundation of the royal transformationâ (çåä¹åº); Máo ShÄ« gÇxùn zhuà n 1.8.
See our discussion on pp. 155â156.
The word zhÅu å¨ occurs but once, in âZhÅu Nánâ 3, and means either the state or the kingdom.
Máo ShÄ« gÇxùn zhuà n 1.8.
Wáng YìnglÃn 2011: 1.187â195. There is of course a larger problem: what precisely do we mean when we speak of the âsouthern statesâ during the Western ZhÅu? Following the parameters suggested by LÇ XuéqÃn 2010c and 2010d (also 1997: 138), whose studies are based on contemporary bronze inscriptions and supported by later historical sources (i.e., ZuÇ zhuà n âZhÄoâ 9), âsouthern statesâ might refer to any number of polities, from modern ChéngdÅ«, SìchuÄn to the west to modern Cháoxià n, ÄnhuÄ«, in the east, and even as far south as northern Húnan.
âZhÅu Nánâ 9 mentions the JÄ«ang and Hà n; (Cf. Máo 204 âSì yuèâ åæ âFourth monthâ: æ»æ»æ±æ¼¢ï¼ååä¹ç´ âSo torrentially flow the JiÄng and the Hà n, regulators of the southern statesâ.) âZhÅu Nánâ 10 mentions the âbanks of the RÇâ; âShà o Nánâ 11 mentions âtributaries of the JÄ«angâ. Hà n and RÇ, however, do not necessarily imply a southern location, as both rivers extend into ZhÅu strongholds.
To date, ShÄ« scholarship, starting (at least) in the Sòng, reads the word hé æ²³ âRiverâ in âZhÅu Nánâ 1 as referring to the Yellow River; see Wáng YìnglÃn 2011: 1.187.
Take for example the late Western ZhÅu (King XuÄn) inscription on the *XÄ« JiÇ-pán å ®ç²ç¤ (YÄ«n ZhÅu jÄ«nwén jÃchéng 10174) that mentions economic issues stemming from southern lineages in the Huái river valley failing to provide ZhÅu markets with silk.
For instance, an early Western ZhÅu bronze discovered in Xià ogÇn, HúbÄi, in 1118, *ZhÅng-fÄngdÇng ä¸æ¹é¼ (YÄ«n ZhÅu jÄ«nwén jÃchéng 2751), which records the capture of a live phoenix in the mountainous area of modern-day ZÇguÄ«, also in HúbÄi, and its delivery to King Zhà o. The king was at the time northwest of modern-day SuÃzhÅu, again in HÇbÄi, preparing for war against the âviolating Tiger territoryâ, presumably further south and outside of ZhÅu control (*ZhÅng-zhì ä¸è§¶, YÄ«n ZhÅu jÄ«nwén jÃchéng 6514); see LÇ XuéqÃn 2006: 210â219.
This argument appears to stem from a fictional account collected in the LÇ shì ChÅ«nqiÅ« âYÄ«n chÅ«â 2, that ends by stating âZhÅu Nánâ and âShà o Nánâ songs were influenced by southern âsoundsâ; see Zhèng Zhìqiáng and ZhÅu YÇng 2004.6: 82â87; Chen Zhi 2007: 198â200; 243; Wáng YìnglÃn 2011: 1.186.
Chen Zhi 2007: 201â208.
We will not take up a discussion of the passage collected in the LÇ jì as we feel James Leggeâs translation has already determined that nán has this meaning. The excerpt, in âWén wáng shìzÇâ 20.625, reads: è¥é¼åãæ¥èª¦å¤å¼¦ï¼å¤§å¸«è©ä¹ç½å®, which Legge translates as â[in autumn and winter â¦] the assistants regulated by the drum (the chanting of) the Nan. In spring they (i.e., royal children and young men) recited (the pieces), and in summer they played on the guitar, being taught by the grand master in the Hall of the Blindâ. (Note that Legge reads ç½å® as âHall of the Blindâ, but the sentence should more likely end with zhÄ« ä¹ âitâ; GÇ zÅng ç½å® starts the next sentence in the BÄijÄ«ng University 1999 edition). Zhèng Xuánâs commentary (20.625â626) understands the word å as âmusic of the southern tribesâ and cites Máo 208 as evidence; he then understands the words sòng 誦 âreciteâ and xián 弦 âstring instrumentâ in the following sentence as referring to the ShÄ«.
Chen Zhi 2007: 212 cites Zhèng Xuán, who says a word signified with the graph é meant a âdrumâ, which we find dubious; either way, it is not a bell, which is the object carrying the inscription.
Chen Zhi 2007: 211.
Yáo Jìhéng (ADâ¯1674âca. 1715) 1961: 11.228.
ZuÇ zhuà n âXiÄngâ 29 (544â¯BC).
For a comparable example in Hán FÄizÇ âShà guòâ åé, see Chén Zhì 2016: 9.
*KÇngzÇ ShÄ« lùn, which is no longer intact, mentions only three sections for certain: the âBÄng FÄngâ é¦é¢¨ âAirs of the Statesâ, âDà Xià â 大å¤, viz., âDà YÇâ 大é âGreat(er) Elegantiaeâ, and the âSòngâ è¨, viz., é âHymnsâ. However, citations of ShÄ« in Warring States versions of ZÄ« yÄ« confirm the existence of âYÇâ-songs in two sections, âlesserâ and âgreaterâ, that is the âXiÇo YÇâ and the âDà YÇâ, as per Máo; see for example Shà nghÇi 1, slip 18.
LúnyÇ âYáng Huòâ 17.10; ZuÇ zhuà n âXiÄngâ 29 (544â¯BC).
See Meyer 2011 (2012), Krijgsman 2016.
Meyer 2014; 2011 (2012).
On the Yì traditions during the Warring States, see Shaughnessy 2014, Schwartz 2018a and Schwartz 2018b. On the Shū traditions, see Kern and Meyer 2017, Meyer 2021.
See Pines 2020 on *Xì nián 繫年 from the Qīnghuá manuscripts.
See Harbsmeier 2015 on *YÇ cóng èªå¢ from the GuÅdià n manuscripts.
See Cà i XiÄnjÄ«n 2017: 166â183 on ZhÅu GÅng zhÄ« qÃn wÇ å¨å ¬ä¹ç´è from the QÄ«nghuá manuscripts.
See LÇ XuéqÃn et al. 2017: 136â138 for a transcription and commentary on *Zhù cà ç¥è¾ from the QÄ«nghuá manuscripts; see Schwartz 2015.
See Kern 2019 on Qà yè èå¤ from the QÄ«nghuá manuscripts.
Preliminary reports state that from July 2014 to January 2015 archaeologists excavated sites at Xià jiÄtái å¤å®¶å° and LiújiÄtái å家å°. Tomb 106 at Xià jiÄtái, which dates to the Warring States period, is reported to have included songs of Bèi fÄng é¶é¢¨ (see HúbÄi rìbà o, 28 January 2016). No transcription of the songs has been published yet, so we cannot comment on this reported instantiation of ShÄ«.
Hú PÃngshÄng and Hán Zìqiáng 1988.
The Fùyáng Shī also contains some significant graphical and lexical variation unknown from other recensions, and so attests to the diversity of early Shī.
How closely these names, if they existed, might have corresponded with Máo, is yet another matter.
Whether the formula âstate, number of songs, song titleâ, as in âYÇngâ and âWèiâ, represents the creation of a retrospective anthology, by claiming certain songs for some states, is something we explore in our discussion of the âcommon statesâ.
The fact that *KÇngzÇ ShÄ« lùn never speaks of âNánâ but lists the songs of the âZhÅu Nánâ and the âShà o Nánâ along with other âFÄngâ songs by title, concurs with our hypothesis that nán during the Warring States period cannot refer to a type of music, but rather that in ShÄ« it is a normative, directional, term.
The *KÇngzÇ ShÄ« lùn mentions one state, Bèi å (é¶), and this implies that other songs also had state affiliations, as per the fifteen states in Máo. The *KÇngzÇ ShÄ« lùn does not, however, mention songs from the states of Wèi é, QÃn 秦, and BÄ«n è±³. As mentioned, songs from Táng were classified as songs of Wèi in Än Dà , and this leaves open the possibility, if reading the *KÇngzÇ ShÄ« lùn through the Än Dà ShÄ«, that at least two songs mentioned in the *KÇngzÇ ShÄ« lùn could be classified as songs of Wèi. Perhaps more important is the elevation of the QÃn songs immediately after the âTwo Nánâ in Än Dà , but their complete absence in the *KÇngzÇ ShÄ« lùn. Songs of BÄ«n do not occur in either manuscript text.
Huáng DékuÄn 2017. This includes texts about ChÇ æ¥ history (440+ slips, in two groups), thought (276+ slips, in nine groups), physiognomy (22 slips), divination (11 slips), and more poetry (50/51 slips, in two groups).
While the general assumption is that text production and text reception in antiquity was a male activity, we cannot conclude with certainty that females had no access these texts.
Wáng YìnglÃn 2011: 9.
Wáng XiÄnqiÄn 2020; Chéng Yà n 2010.
Wáng YìnglÃn 2011: 160â162.