T of C 
Qin as
Qin in
/ Song
Analysis History Ideo-
Personal email me search me
Qin songs  /  YFSJ, Qin melody lyrics 網站目錄
Cipai and Qin Melodies 1
Ci lyrics as applied to qin melodies 2
A qin setting of ci lyrics (complete)3  
As will be discussed on this page, very few "qin ci" (or "qinci": qin melodies having ci lyrics) survive from the Ming dynasty. This seems somewhat surprising, perhaps suggesting they were more amenable to an oral tradition. But since the majority of qinci that survive do so from the Qing dynasty, and since the focus of this web site is the Ming dynasty (and to a certain extent earlier), I have not reconstructed many of them. Those that I have done almost all survive either from one of the handbooks preserved in Japan beginning around 1676 (1676 ci list), or from one of the Songfeng Ge handbooks of 1677-87 (in particular Shuhuai Cao and Songsheng Cao).4 Those for which I have done recordings include the following:

  1. Zui Weng Yin (sometimes called Zui Weng Cao; trace; structure)
  2. Fenghuang Taishang Yi Chui Xiao (1609 and Japan; same lyrics, different melody)
  3. Chang Xiang Si (image at right; Japan; 1682/1687: two different melodies and lyrics)
  4. Qingping Yue (Japan; 1682/1687; 1687: three different melodies and lyrics)
  5. Lang Tao Sha (Japan and 1618; same lyrics but different music [1-string qin])
  6. Tanpo Huan Xi Sha (1682)
  7. Shui Diao Ge Tou (1687)

Those that I have transcribed but not yet recorded include the following:

  1. Dongfeng Qi Zhuo Li (only Japan)
  2. Libie Nan (only Japan)
  3. Yi Wangsun (only Japan)

Qin melodies with Ci lyrics: qin ci

In overly simple terms, ci poems/lyrics might be contrasted with shi poetry/lyrics, which had lines of regular length (most notably 4-, 5-, 7- and sometimes 6-character lines).5 In contrast, ci, one of the forms with lines of irregular length,6 originated during the Tang dynasty as song texts applied to existing melodies. This form reached its fullest development during the Song dynasty, but its popularity continued after that period. Nevertheless, throughout this period, and particularly during the Ming dynasty, their absence from surviving qin handbooks suggests that application of ci patterns to qin melodies/songs was not a major interest amongst qin players of that time.7

A qin melody having the name of a ci pattern does not guarantee that it has either ci lyrics or music that would fit lyrics in that ci pattern.8 In addition, there are almost no melodies from the Ming dynasty that have lyrics that fit any ci pattern.9 Those that do make no mention that their music can be applied to other lyrics in that pattern; in fact, the only such mention is for a melody not in ci form but one piece with lyrics in the basic pattern (7+7) x 2 (further below).

Thus, although one might speculate that during the Ming dynasty there must have been an oral tradition of setting ci lyrics to qin melodies, for examples of how this might have worked one must look to a few surviving handbooks from the early Qing dynasty, specifically the handbooks related to Jiang Xingchou, preserved in Japan, and the Songfeng Ge handbooks (1677-87) of Cheng Xiong (n.d.).

Traditional methods for creating lyrics following the pattern of lost ci songs

Since by the Song dynasty most of the original songs that apparently supplied the ci patterns would have been lost, new ci lyrics had to be patterned largely on two literary aspects of the poems, the number of characters (i.e., syllables) per phrase,10 and the tonal patterns of the individual characters.11 By itself this seems to suggest the original songs had purely syllabic settings; however, I am not familiar with academic discussions of this very important issue.

The most obvious characteristics of ci lyrics is that they generally do not have regular line lengths (e.g., four characters on each line). Ci likewise often have a different number of syllables (i.e., Chinese characters) on each line. Presumably, then, the melodies they originally accompanied would not all have had the same number of notes on each line. But what does this say about the original rhythms?

Unfortunately, in addition to all the original melodies being lost, the original pairing method is not known. Although people seem to assume it was one syllable for each note, as is often thought to have been the case with poems in the Shi Jing (Book of Songs) or with early Yuefu (Music Bureau) lyrics, there seems to be no hard evidence proving such a pairing method was ever used, at least not exclusively used.

Individual ci lyrics are generally given the title of the original song from which they take their pattern. A well-known example of this would be the numerous eight-line ci said to be "to the tune Endless Longing" (Chang Xiang Si; the image above shows part of a typical setting). These generally have the syllabic pattern 3,3.7.5 repeated once. Apparently Chang Xiang Si was originally a song, but long after the melody was lost the lyrics and therefore the syllabic structure remained. The two surviving old qin melodies called Chang Xiang Si both have lyrics with this pattern.12

Qin songs: pairing lyrics and music 13

Assuming it is true that the lyrics of surviving ci poems follow the structure of ancient songs now lost, using the surviving ci lyrics to reconstruct the rhythms of any of the original melodies would seem to require the original song to have had two specific characteristics:

Generally speaking, neither of these characteristics is true of qin songs (if the length of a note is known it is through oral tradition, as qin tablature does not provide that detail). Here, although the standard pairing method within published qin songs always followed the pattern of one character (syllable) for each right or left hand stroke/pluck, the custom regarding left hand ornaments such as slides or multi-note right hand strokes such as glissandi (gun and fu), was more flexible. Thus the songs either did not or could not (particularly with regard to note duration) use these two characteristics in its pairing. Instead the pairing seems to be based on what might be called "structure points" or "musical accent points" ("dian") within each phrase ("ju").15 The dian, in a way, finesse the pairing problems discussed here.

Once information about the structure of a melody or a song is known, it should not be necessary actually to refer to the original melody or song. Indeed, the pairing method that seems actually to have been used in the Ming dynasty suggests that they might have taken a structure such as the 3,3.7.5 mentioned above and treated these numbers as though they represented not numbers of syllables per phrase, but simply as the aforementioned "dian". If one could create rules for determining such accent points within the music or its score, then these accent points might be used in the pairing of any related ci lyrics to that music. But how often did they actually use it in this way?

Looking at this another way, Ming dynasty tablature seems to pair lyrics with qin melodies following a formula that uses such dian to structure the rhythm in a way that allows flexibility: instead of aligning the lyrics to a note pattern, the qin tablature aligns them, as just outlined above, to the places in the tablature where there are certain types of stroke indications. Perhaps this is just a coincidence rather than intentional, but since qin tablature does not directly indicate rhythm, this pairing method really does finesse the issue of rhythm.

More specifically this method paired words and music more or less as follows (there were inconsistencies).

To combine lyrics with an existing melody, qin tablature applied one Chinese character for each right hand stroke or left hand pluck, but usually no characters for slides (always done by the left hand) and with some flexibility on right hand strokes such as glissandi. With simple right hand stroke passages this was quite straightforward: inconsistencies were largely with the slides and multi-stroke techniques. For example, although characters generally were not applied to left hand slides, occasionally they were;16 the tablature generally paired one character to a pluck by the left thumb but some tablature often paired two characters;17 and multi-stroke right hand techniques could be paired even more inconsistently, perhaps particularly when what seems to have been an originally instrumental melody had lyrics added to it.18 In addition, where the tablature includes the instructions "play again" it is not always clear whether the intention was to sing the phrase on the repeat.19 In certain places or times the people writing the tablature or commenting on it may have tried to apply consistent rules to this, but at present such rules can not be determined.20

This word to note method was also used in reverse: if existing lyrics were to be given a new melody, then for each character a plucked note could be applied; if one wished to have more notes than lyrics, then the extra notes would be in the form of slides or other left hand ornaments.

Although the value of this pairing method was sometimes debated,21 it seems generally to have been followed for the written versions of all guqin songs prior to the modern period. As a result, almost all qin songs seem to have been very word-intensive.

Clearly methods such as this could be used to allow people to apply lyrics to melodies that were "lost" in the sense that tablature existed but no one knew what the actual note values were, since no one had actually ever heard them: here they only had to have copies of the tablature. This may well have been the case for the earliest surviving qin melodies with lyrics, in Zheyin Shizi Qinpu (<1491; see the commentary as well as lyrics that can be applied to Shen Qi Mi Pu melodies). Someone experienced in writing ci lyrics could easily adapt its method (as described above) to the stroke numbers in old qin tablature.

As for pairing qin tablature to actual ci patterns, one might suspect that this would have been done quite commonly: the resulting melodies that fit a particular ci pattern could then be applied to any poems following that pattern. As yet, however, I have found only two examples of something like this actually being done:

Instead what one usually gets is the pairing of a variety of melodies (over time) to a single a set of lyrics.23

Singing qin songs: appropriate performance practice

(This discussion has been moved under Qin songs.)

Footnotes (Shorthand references are explained on a separate page)

1. Cipai and Qin Melodies
First, the character 詞 ci should not be confused with 辭 ci, as in 楚辭 Chu Ci (Songs of Chu) or 歸去來辭 Gui Qu Lai Ci. As for a possible 辭牌 39539 has no such entry. On the other hand, although ci patterns are called 詞牌 ci pai, the use of the word ci does not guarantee there is/was an associated cipai (e.g., 冷玉詞 Leng Yu Ci.

Online lists of cipai include:

What I have not yet found is one that lists them by structure (e.g., [7-7-7-3] x 2 for one with four phrases of 7+7+7+3 characters, repeated).

2. 琴詞 Qinci: ci lyrics as applied to qin melodies
Although an internet search for "琴詞" (qinci) yields a number of results, I have not yet found any published studies on this subject. It is my own belief that short qin songs such as qinci must have an ancient history, but that perhaps their nature or milieu made them more compatible with oral tradition.

In this light it is particularly interesting that the Song dynasty's Yang Zuan, who is of great importance to the qin, was himself also noted for his ci studies (see in particular this online article [Chinese with English summary at end]).

3. Image: Setting for qin of the song Chang Xiang Si
This setting, discussed further separately, is from Hewen Zhuyin Qinpu, published in Japan before 1676, probably having been brought there from China. The symbols by the Chinese characters tell Japanese how to pronounce the characters in Chinese. In these first two double lines the setting has one character for each qin note except for the slides on the fourth cluster of each line of tablature.

4. Qing dynasty qinci
Although a number ci songs from qin emerged in the early Ming dynasty, after this very few seem to have been published.

5. Poetry with lines of regular length
A general study of Chinese poetic forms is beyond the scope of this website. A good introduction is Cai Zongqi, How to Read Chinese Poetry, New York: Columbia University Press, 2008 . This webpage (唐人琴詩) has a large collection of Tang dynasty poems.

Examples of qin songs set to lines poetry with regular length includes:

More could be added if sections of longer melodies were included separately. For example,

  1. Section 6 of the 1525 melody Kang Qu Yao begins with a song having the pattern (4+4)x2.
  2. Section 8 of the 1525 melody Gu Jiao Xing ends with a song having the pattern (5+5)x2.

More to be added. Of course people can simply create their own melodies for lyrics that inspire them, but melodies such as these also provide essential guidelines to how this was done in the past.

6. Poetry with lines of irregular length
Besides the 詞 ci there are also 賦 fu (as in Qian Chibi Fu, etc.) and 騷 sao (as in Li Sao).

7. Apparent rarity of qin melodies/songs using ci patterns
It is also possible that such pairing was common in informal gatherings or other events where the results were not written down.

To my knowledge the largest collection of ci melodies in any qin handbook are the ones in the Shu Huai Cao of 1682 and Song Sheng Cao of 1687. Their contents are listed in this table of contents.

8. Qin melodies with the names of ci patterns but that do not fit the pattern
These include the following:

  1. Feng Ru Song (compare the structure of 1511 with that of 1687)
  2. Shuilong Yin (no lyrics; no apparent relationship to the ci pattern of this name

Such qin melodies are not all songs. Note also that it seems to have been more common, at least in publications, to have different melodies set to the same lyrics; see, for example, Xiang Si Qu.

9. Finding poems that use specific ci patterns
The most convenient way I have found so far of finding other poems that use these ci forms is to go the to McGill/Harvard-Yenching Ming Qing Women's Writings website - I don't know of a comparable site for all writings that include male writers. To do the search, go to the search page, then under Section II "Browse a List" select "Poem Title" and then type in the ci pattern name, e.g., feng ru song (quotation marks are not necessary and it does not seem to be case sensitive, but it seems to be important that each syllable be written as a separate word.

10. Syllable counts in ci poetry
The Wiki entry mentions about 800 ci patterns including tonal variations (see next footnote), but does not say how many patterns can be identified purely by syllabic count.

11. Tonal patterns in ci poetry
This is now explained in terms of the 平仄 ping ze (or pingze; 9371.37; 2/922), the level (平 ping) and oblique (changing; (仄 ze, of which there are three: 上 shang, 去 qu and 入 ru) tones that formed the patterns of much traditional Chinese poetry. The origin of this system is not clear (the earliest reference in 2/922 is from 陳鵠 Chen Hu of the Southern Song dynasty (in 9371.37 is the late Ming 嘨餘譜 Xiao Yu Pu). After the Song dynasty ce took on other meanings and only ping ze was used for the tones.

Most important to the present discussion is first whether these tones can help in trying to reconstruct the early melodies, and then whether the nature of this system sheds any light on whether or not the original songs had syllabic settings.

To my knowledge there are no answers to this question. A few melodies follow the ci syllabic counts if one considers dian rather than notes. It does not seem that any qin songs actually have ci tonal patterns.

12. Everlasting Longing (長相思 Chang Xiang Si): qin melody and ci pattern
Zha Guide 35/--/510 lists this title in five handbooks, but within this there are basically two surviving qin songs. These are discussed in more detail separarately.

13. Pairing words and music
As discussed here (with specifics), lyrics are paired by the method of one character for each dian, apparently defined as any right hand and certain left hand strokes. This pairing method seems to have led to a rather odd situation where lyrics can be paired following this formula but there is no apparently relationship between the lyrics and the musical phrases. Some examples of this are discussed in connection with the earliest handbook to pair lyrics and music (Zheyin Shizi Qinpu; see also the related footnote).

14. Note values in the original melodies
As for what the ci structures might say about musical structures, unfortunately I do not have a deep familiarity with this subject and so do not know if anyone has ever tried to analyze it. For example, I have not seen any arguments actually made that the pairing assumed that each syllable had one note of equal or fixed length, that each syllable had a fixed number of notes, or any other permutation on such structured pairing. Nor have I seen arguments that stated simply that if a poetic structure such as 3,3,7,5 fit a certain melody, any poem with that structure should fit that melody, acknowledging that nevertheless this told us nothing about the actual structure of the melody.

15. Dian and Ju
In the same way that early Chinese literary text often had no punctuation, early qin tablature quite likely also had none. For example, there was no punctuation for many of the melodies in the earliest surviving edition of Shen Qi Mi Pu, Folio 1. I do not know what term might have been used for such musical punctuation. "Dianju 點句" (12/1350: dots that stop the continuation of "ju", phrases) is a Chinese term for punctuation that seems to date from the Song dynasty. A more common term is "biaodian 標點" (4/1269: put up marks [dots] to indicate phrases), also with references dating from the Song dynasty.

In this context ju were clearly the phrase markings, but can one find "dian" in qin tablature? 49060 點 does not seem to include a discussion of dian with regard to ci poems. Ci poets, as suggested above, apparently assumed the musical settings of the original poems on which they based theirs were done one note per syllable; in this case there would be no need for a concept/term such as dian to show how the words fit to the notes of the melody. But if, as suggested above, the original melodies might have had more than one note per syllable, would not such a term have been necessary? The lack of such a term then seems to suggest either that the original melodies did set one note per syllable, or that by the time the ci forms were set the original melodies were lost.

With qin songs each character can be considered as a sort of point in the melody. And in at least one qin handbook, Yingyang Qinpu (which has no melodies with lyrics), the phrasing is emphasized by putting the beginning of each phrase on a separate line, and stating at the beginning of each section of every melody how many ju and how many dian it has. These dian seem to correspond to the points on old tablature at which lyrics could be paired, but the aim here seemed to be only to help the student learn the melodies correctly. If the aim had been to show how lyrics could be paired, then it would have been more helpful to say how many dian there were in each ju.

16. Inconsistencies within left hand slides
There are no right-hand slides. Left hand slides include simple up, simple down, multiple up (which could be indicated, e.g., as from the 7th position [hui; stud] go to the 8th then 9th positions, or from the 7th position do a double slide to the 7th position), multiple up, up and down (shang xia), up-down (e.g., zhuang), etc. In some tablature none of these slides have characters paired. Occasionally all do (a note plucked followed by two slides up will be given three characters). Often this is inconsistent. Inconsistency here might suggest that the lyrics and music were once independent of each other, then matched later.

17. Inconsistencies within left hand strokes
Left hand strokes consist largely of "covering" (罨 yan): bringing the left thumb down on a string while keeping the ring finger in place; and plucking up (搯起 taoqi): while holding down a position with the left ring finger plucking up or in with the thumb. Generally the technique taoqi seems to be identical to the one called 對起 dui qi (earlier called duian). However, the <1491 handbook (again see the commentary) generally attached one character to taoqi but two meaningless characters such as 的那 de na to duiqi. Later tablature was more inconsistent here, with the meaningless syllables used less and generally either one or no characters applied to the duiqi or taoqi.

As for genuine multiple stroke techniques such as 搯撮三聲 taocuo sansheng, these could be assigned a widely varying number of characters, while sometimes there is only a vague indication that one should repeat syllables just sung; rarely are no characters at all assigned.

18. Inconsistencies within right hand strokes
These are mainly with the symbols that represent multiple notes, such as a 滾拂 gunfu. A gunfu consisted of a gun downward over several strings, usually all the way to the 1st, then a fu from the 1st or 2nd string back up to the original string. The most common gunfu began at the 7th string, so it would consist of a total of 13 or 14 notes played in rapid succession. These were often paired inconsistently. For example, they might be assigned just one character for the whole gunfu, two characters (one for gun and one for fu), but sometimes it was three, four or even more. There are somes guns to which seven characters are applied (here usually the tablature usually, instead of "gun 7 to 1", says "gun 7 6 5 4 3 2 1). In my opinion this usually suggests that someone was taking a purely instrumental melody and applying lyrics.

19. Inconsistences in repeats
A major characteristic of many qin songs is not only that when there is the instruction "repeat" (再作 zai zuo) there is no indication of whether the lyrics should be repeated, in many cases there are no instructions saying where the repeat begins. In the absence of specific commentary on this, one can only speculate as to whether this suggested flexibility within the qin song idiom, or siply confusion on the part of the transcriber.

20. In tablature for the period I have studied, i.e., until the end of Ming dynasty, left hand plucks generally are assigned characters but slides generally are not. However, both of the two surviving examples from before the Song dynasty, Gu Yuan and Huang Ying Yin are syllabic settings, with the former adding characters to slides, and the latter having no slides.

21. Debating the one character per stroke pairing (see also Zha Fuxi's Differentiating Qin Songs)
Considering the significance of this issue, there does not seem to be a lot of debate about qin songs. Comments worthy of mention include the following:

  1. Yan Cheng, Comment in an afterword (QQJC VIII/162; original text)
  2. Speaking of Having or Not Having Lyrics (有詞無詞說 Youci Wuci Shuo) in Lü Hua (XXI/384-6; 1833; .pdf 1.4MB)
  3. A question and answer from: Questions and Answers about the Qin (琴學問答 Qinxue Wenda) in Qinxue Conshu (XXX/371; 1910; .pdf 94KB)

It should be emphasized that I have not made a thorough study of this and so the present analysis, including this listing of articles, is clearly very much incomplete. (Is this discussion of Saishang Hong and kunqu relevant?)

Perhaps the most notable Ming dynasty criticism of this method for pairing lyrics to music was the one by Yan Cheng listed here and discussed in QSCM, Chapter 7a2. In spite of such criticism, though, I have not yet found a handbook that pairs multiple stroke techniques to each syllable, or any other specific evidence that it ever happened (though I am sure it must have at least been tried). The closest I can find to an exceptions are some melodies in Xilutang Qintong. This handbook uses the same pairing method, but for a number of melodies it has lyrics with only one or two sections of a longer melody instead of with all the sections.

22. Existing melodies intended for use with specific poetic meters
These two melodies are unrelated to each other. Their structures do not fit any named ci pattern but are common for classical poetry.

23. Melodies that have been used for more than one set of lyrics
Although the common custom was to pair a variety of melodies (over time) to a single a set of lyrics, and as yet I have found only the two mentioned examples (from 1573 1590) of melodies designed to be used with many sets of lyrics, I have also found at least one occurrence where an old melody is re-fitted with new lyrics. The handbook Kumuchan Qinpu (1893) has a version of the melody Gui Qu Lai Ci in which for the original lyrics the handbook has substituted Buddhist lyrics; it calls the piece Lianshe Yin.

Appendix: Content of
Shu Huai Cao (1682) and Song Sheng Cao (1687)

To my knowledge these two handbooks have the largest collection of ci melodies in any existing qin handbooks

Content moved

Return to Qin Poetry and Song or to the Guqin ToC.