Rhythm in early Ming qin tablature  
 T of C 
Qin as
Qin in
/ Song
Analysis History Ideo-
Personal email me search me
Historically Informed Performance  /   Dapu  /   and the creative process  /   mode  /   glossary 首頁
Rhythm in Early Ming Qin Tablature 明初琴譜節拍
Originally presented at the 34th meeting of ICANAS, Hong Kong, 1993 1
  - Additional comments appended 2
- 附加更新的評論  

General introduction

Qin tablature did not transmit music with the aim commonly attributed to music in Western notation -- to express music composed and written down by one person in order to prescribe how other people should play it.4 Instead it was a description, often by a disciple, of how a master played a particular piece. This transmission method is said to date back over 2,000 years, with a shorthand version similar to what is used today going back over 1000 years. This short-hand tablature, unique to the qin, specifies tuning, hand positions and stroke techniques, but does not directly indicate note values.

This lack of direct rhythmic indication has led some people to argue that qin music should have no rhythm, or that historically it had no rhythm (see further). Others argue that rhythms were either not important, or were left entirely up to the player. However, in the vast majority of circumstances these opinions are almost certainly incorrect, or at least misleading: the reason it was not necessary to indicate note values was that the music was usually learned by copying a teacher, with the tablature functioning mainly as a memory aid for when the teacher was not around. If memory of the melody itself became uncertain, the tablature descriptions of exactly how the master played the piece could also serve as reminders of the original rhythms. As a result there was no need actually to write down the rhythm.

Of course, with the scholar's respect for history, these documents were preserved and later interpreted. In this way, qin tablature was seen as saving something old rather than revealing something new. In fact, qin music did continue to develop over time, and it is important to study the role of tablature in this development of qin music. However, the present paper focuses only on a related issue: the possibility and desirability of trying to play the tablature accurately.

The nature of this qin tablature makes the process of going from tablature to performance rather different from the task of reading modern Western notation. The Chinese term for learning qin melodies directly from the tablature is "dapu". This term is often translated into English as "reconstruct" or "recreate", but here and on the "dapu" webpage I try consistently to translate it as "realize": make real. In fact, the meaning of dapu/realize includes both of these other terms, "recreate" and "reconstruct":

It is not always possible completely to distinguish between these two concepts, as realization will always include both reconstruction and recreation. However, it is important to keep in mind the theoretical distinction. From my experience most players speak of dapu as reconstructing; but the results of most transcriptions indicate strong portions of recreation.

Qinpu compiler's intention: description or prescription?

The earliest surviving collection of tablature for qin melodies is the Shen Qi Mi Pu (hereafter SQMP, published in 1425 CE by the Ming prince Zhu Quan. Much of the music in the second surviving handbook, Zheyin Shizi Qinpu (hereafter Zheyin) is copied from Zhu Quan's SQMP (or from the same source as SQMP), and so it is appropriate here to consider what was Zhu Quan's intention in compiling the tablature. This is what he wrote in his preface:

"I personally have been taught 34 tunes. These all have their phasing indicated here. The playing methods, including ornamentations and the proper tunings, have in no way been concealed. They have been printed here in order to transmit them: later students can look at this tablature and obtain everything by themselves: they won't have to wait for a teacher who can pass it on to them....

"I have published them in order to pass them on to our generation, and see that later generations can also obtain them. The result should be they won't become lost to people who want to study them in the future. I have frequently added corrections, and not worked at it for simply a day or so, but rather have continued at this for 12 years...."

This quote suggests that Zhu Quan might have considered the versions he published to be prescriptive. This doesn't mean that he would have objected to people modifying them or playing other versions; he says nothing about that. But why spend so much time getting all the notes correct if it wasn't important for someone to play them as written?

In 1974, when I first began to study the qin in Taiwan, my teacher, Sun Yü-Ch'in, wanted me to play exactly as he did; on the other hand, if an advanced player changed something, he considered this quite acceptable. I believe this was an important part of the tradition, and assume that, while Zhu Quan may not have insisted that the versions he published were the only correct ones, and may indeed have intended that people should use this book as a basis for "new" music, he wanted at least to allow qin players the opportunity to study the old master's performances before developing new ones.

Whatever the intention, Zhu Quan tried as much as possible to present an accurate account of the way someone (in some cases, perhaps he himself) played or might play a piece. Unfortunately, we don't know which particular aspects he struggled so much over. If it was to decide whether a particular stroke should be second finger inwards or outwards, what significance does that have? If it was to clarify phrasing, what does that mean? To what extent was he trying to set parameters for interpreting rhythm? Some other early handbooks have so many mistakes that their music is, virtually unplayable. Thus the aim might simply have been to present one possible version.

Although Zhu never directly indicated how much variation he considered acceptable, he did have (or relay) this comment in the middle the piece Qiu Hong (Section 27).

"This section is very profound. One must receive personal instruction to avoid the problem of not playing smoothly and continuously, and from anxiety about getting the rhythm wrong. In this it is like section 15 above."

This sounds like someone trying to encourage a specific style of play, though aware this cannot completely be done through tablature. The most likely explanation for this is that Zhu was simply copying out some tablature which had been/was being used within a particular school to complement direct teaching.

Qin tablature as teacher

Students learned the music orally, copying their teachers exactly. As mentioned above, the music was probably written down originally as a mnemonic device. But when writing it down, why not also write down the note values?

My experience is that, generally, players remember melodies quite well. The tablature was a description -- a sort of choreography -- of the hands. Looking at this tablature would remind the player of the way he or she had learned a piece, and the melody would come back. I have found this to be the case when I have not played a piece in a long time.

My teacher Sun Yü-Ch'in told me it was not necessary to look at the tablature when I studied -- I only needed to copy him. When he himself was playing a piece he hadn't played in a while he might consult the tablature. All the pieces I learned had tablature, from a variety of sources. Where the version he was teaching me differed from the tablature, sometimes he would say there was a mistake in the tablature; other times he would say he had learned a somewhat different version.

Many qin pieces survive in a number of versions, some so different that they require their own tablature to describe them; but even within performances attributed to the same tablature there can be significantly varying interpretations. If analyzed like Western classical music, some might easily be called "mistaken interpretations" -- the player thinking what he or she is doing follows the tablature, but at times clearly is not doing so. However, with qin music, in practical terms, it is not always easy to get a consensus on what is correct and what is incorrect. For example, in some cases it is considered that the deportment of the player at his instrument is more important than a fluent technique. This is particularly said of elderly players.

During my studies in Taiwan with Sun Yü-Ch'in, if I played something differently from the way he did, including the rhythm, he would almost invariably have me play that section again until I played it the way he did. When I asked him why the note values were not indicated in the tablature, he said it was because each player eventually must decide his own. To me this meant you could be flexible with these, but only after you had reached a certain level of technical skill and aesthetic understanding. Even then I have observed that in actuality most people change the note values very little.

To sum up, one might say that the relationship between the player and the tablature is in many ways comparable to that between a student and a teacher. How closely the student follows the teacher may well depend on whether the learner has already acquired a deep-rooted personal style. I still consider the tablature to be a teacher, or the reflection of a teacher, but have probably by now acquired something of a personal style. I would like to think the style is indebted to the exclusivity with which over the past 20 years I have been playing pieces from 15th and occasionally 16th century tablature.

Inconsistencies, errors and development

Since the first folio of SQMP had pieces found only in manuscript form, Zhu Quan must have found it difficult to make accurate corrections in places where the tablature was unclear or wrong. An example of a clear mistake would be for the tablature to say, "put the left ring finger in the 9th position then slide down to the 7th position." Since 9th to 7th is up, there must be a mistake either in the position or the direction.

Many more such mistakes occur in Zheyin. In particular, many passages which in SQMP are playable as written are in the Zheyin editions quite unplayable as written. These can easily be corrected by looking at the SQMP versions, but for pieces not occurring in SQMP, or which occur in versions different from SQMP, this is more difficult. This is particularly true as no later handbooks follow precisely the versions of pieces introduced in Zheyin.

The lack of these mistakes in SQMP perhaps indicates that Zhu Quan had to make corrections to the manuscripts he had in his possession. Particularly with regard to pieces in the first folio, which Zhu Quan had never heard before, quite possibly some of his corrections were "incorrect", and so he was in effect modifying old pieces. Over the centuries many pieces could have been continuously so modified.

In SQMP there is some difference in tablature style between pieces; this is usually reproduced in Zheyin. Van Gulik (Lore, p. 184) regrets these differences, saying Zhu Quan was perhaps being deliberately obscure. However, considering Zhu Quan's claim to have been copying earlier tablature, it would seem to be a more likely explanation that the differences reflect their varied sources, and hopefully the tablature differences may some day, with sufficient research, help to distinguish between some of the different regional and/or chronological sources for the music.

The occasional inconsistencies in tablature style within single pieces perhaps indicate "corrections", as described above, by someone using his own style. The music itself may or may not have been changed from the earlier version. But even though such changes might not have been intended to indicate a change in either notes or note values, just the possibility of such changes having actually occurred makes the task of recovering a playing style more complicated.

As old qin pieces appear in later tablatures in forms which vary from being almost identical to being very much different, one must consider again the intention of the compilers. Were they consciously changing something, or simply playing their own version? The changes may be small: different parameters within which note values should be selected; some new interpretation of a plucking method or the position on the string, and what the resulting note is in terms of do, re, mi and so forth. In other cases there may be completely new musical material. In either case, such changes may clearly indicate a changed sense of modality.

Freedom versus precision in guqin tablature

The choreographic aspects of qin tablature are very detailed. To write down the rhythmic detail in a music as rhythmically complex (some will say "free" or "flexible" are better words) as qin music would require very complex (or flexible) rhythmic signs. The first well-known effort to describe the rhythms comes in the handbook Qinxue Congshu (1910); if it had been easier to notate with precision more aspects of the rhythm, perhaps this might have been tried earlier.

On the other hand, perhaps one reason why this wasn't done was due not so much to the imprecise nature of the acceptable rhythmic parameters, as to the fact that the tablature does already give sufficiently detailed indications of note values.

Guqin music is one of the world's few surviving written instrumental music traditions -- perhaps the oldest -- but the music was probably never composed the way we generally consider Western written music compositions to have been. It wasn't prescriptive, at least not for most people. And the fact is that almost all players who make the qin a lifetime avocation at some stage change the music they were taught.

In sum, qin tablature was an attempt to describe the way a person played a piece -- though in reality it may have been describing a combination of the ways several people played it. It is quite within the qin tradition to use this written material as inspiration for one's own music, just as one might do so with the music taught by a living teacher. Some people might add that one of the marks of the true artist is his or her ability to create something new out of what he or she learned.

On the other hand, since there is no way to know with certainty how the music actually sounded, it is difficult to judge creativity in interpretation, or lack thereof, in terms of attempted faithfulness to the tablature. And in addition we do not know enough about the aesthetic standards of ancient China to be able to make a categorical statement about the relative regard for qin players who faithfully followed their teacher's style as opposed to those who changed it.

One of the reasons Western music is so vibrant is that, alongside contemporary music, we have music reconstructed and recreated from so many past eras. In this regard, whether or not the interpretations are "accurate" is not particularly relevant. Recreated ancient music, to be accepted, needs only to be accurate enough to allow the temporary suspension of disbelief. (See also Some Issues in Historically Informed Qin Performance.)

Determining the notes
(See also
Modality in Early Ming Qin Tablature)

This is usually quite straightforward. The tuning of the strings is given and positions are indicated with regard to the 13 studs (hui) which mark harmonic nodes on the strings (so that, with only three exceptions I know of, harmonics are always played directly opposite a stud). The instructions for playing a note are written in a cluster of symbols, mostly shorthand forms of Chinese characters. On top the figures tell what the left hand does and on the bottom what the right hand does. Thus the top will indicate which left hand finger should be placed in what position (and do what ornamentation); the bottom will say which right hand finger does what stroke and on what string this is all done.

As for indicating finger positions, the present decimal system (see row New in the chart below), in use since the 17th century, can do this with mathematical precision: 7.6 and 7.9 mean respectively 6/10 and 9/10 of the distance between the 7th and 8th positions. In theory this system could be used to indicate very precise tonal differences, e.g., 10.7 for a slightly sharpened E.

Since this was never done, the system used in SQMP (see row Old in the chart below) was in theory just as precise as the contemporary one. Thus the decimal position 7.6 was in SQMP almost always written 7 8, meaning "the correct place between the 7th and 8th positions"; 7.9 was rounded off to 8; and 7.3 was written 7-. When SQMP deviated from this system, using values found in row Alt of the chart below, it was usually on slides (see explanation below the table, since I could not fit all the various Alt figures within that one row).

Other early handbooks may be less precise. In particular Zheyin, although in pieces copied from SQMP it also follows that system, in pieces not in SQMP it normally used the values in row Alt., which can lead to much confusion. Thus, in pieces, or versions of pieces, occurring for the first time in Zheyin Shizi Qinpu, 7.6 may be written as 7 8, but it is more likely to be 8+ (above 8), 7 1/2, or even 7- (below 7), with perhaps all three occurring on the same page.

Standard positions on a qin string (items marked * have comments in the notes)

  C C# D D# E F F# G G# A A# B c c# d d# e f
0 (13.9*)
0 (wai*)



  1. in the modern repertoire wai is always 13.1; old handbooks usually call 13.1 wai, but may also call it 13, then use wai for 13.5 and even 13.9; but there is no consistency, so this is somewhat speculative. Sometimes a distinction between 13 and wai seems to occur in passages at 13.1 over several strings, with notes played by the 3rd finger using 13 and those by the 4th finger using wai. Could this be an example of a transcriber who may not have known about temperament or "correct" positions giving his impression of where the player seemed to put his fingers?
  2. "9-10" = "between 9 and 10" (9, 10 jian); likewise for "8-9", "7-8", "6-7", "5-6", and "4-5". SQMP quite consistently uses these Old positions as indicated, but other early handbooks (Zheyin in particular), generally use an Alternate figure, only one example of which is given in row Alt. Thus, in addition to expressing the modern 9.4 as "9-" (below 9, or 9 xia), it also indiscriminately uses "10+" (above 10, or 10 shang) and "9 1/2" (half 9, or 9 ban), all to mean this same position. The same variants are available for all the other "belows" on this line. Sometimes 8+ is even used for New position 7.9.
  3. All these three alternatives (as well as the Old figures) can also be found for positions between 8 and 9, 7 and 8, 6 and 7, 5 and 6 and 4 and 5.
  4. Although SQMP is generally rather consistent about using the Old system, see, for example, Gao Shan, measure 33 of my SQMP transcription, where 8+ clearly means 7.9, but plain 8 is used for 7.9 in the rest of the piece.
  5. "9-" = "below (i.e., to the left of) 9" (9 xia); likewise for 8-, 7-, etc.
  6. "7+" = "above (i.e., to the right of) 7" (7 shang); likewise for 6+, 5+, etc.
  7. Positions higher than 4.0 are extremely rare, except when playing harmonics.
  8. Ban (half), is particularly popular with sliding sounds. For example, pieces in which the modern 7.6 is written as 7-8 for fixed positions will often for slides write 7 ban.
  9. It is perhaps the inconsistent usage of the Old and Alt systems which led to the change from the potentially precise Old system to the more potentially precise New decimal system.

For an example of the potential of the "New" system, assume this table refers to the 3rd string, tuned to C, and that the 4th string is D. G is then 9.0 on the 3rd string, 10.0 on the 4th. Using the New figures, 10.8 on the 3rd string is Pythagorean E. If, as in Table 4, the 3rd string is tightened so that the harmonic at 11 becomes the Pythagorean E, then the stopped G on the 3rd string must be adjusted down to about 9.1 or 9.2 to give the same note as stopped G (10.0) on the 4th string. The New system in theory has the precision to indicate this, but not the Old system: "above" and "below" were almost exclusively reserved to indicate half tones, not musical commas.

The fact that such distinctions have never been found in the New system suggests that they were not important to qin players. In addition, the apparent precision of either system can be somewhat misleading. For example, in actual play if you sound the octave harmonic by lightly touching the 7th position (7.0), then right away push the finger down to form a stopped sound you will find that the pitch is slightly sharp, the degree of sharpness varying with the height of the bridge and the tension of the strings.

A further explanation of tablature is given with my book of transcriptions from Music Beyond Sound.

Determining note values

Since the tablature details tuning, finger positions and playing techniques, but lacks a direct indication of note values, the most difficult problem to someone trying to recover the music is in deciding how long to hold the notes. This can also be seen as trying to determine within what parameters might the person who transcribed the music originally have considered variations in note values acceptable without having to change the tablature.

If old qin tablature is used in an attempt to re-create old music, what are the parameters within which change is acceptable and yet the music can still be attributed to that tablature? What clues are there on how to interpret the note values?

These clues come mainly from the tradition as players have all learned it, but in addition they also come from clues inherent in the method of writing the tablature. Reconstructing rhythms from qin tablature is for me largely a search for possible structures. Clues discovered in this attempt include the following:

  1. There is a dominance of double and quadruple meters.
  2. Couplets are common: repeats, repeat with variation, pairs with related structure
  3. Rhythms may be suggested by punctuation and lyrics.
  4. Similarities and differences in earlier or later versions yield clues.
  5. Note values relate to ornamentation and fluency of finger patterns.
  6. Repeated or altered rhythmic patterns may be indicated by repeated or altered:
    a. Ornamentation patterns.
    b. Note/finger patterns;
    c. Musical contours/structures.
  7. Certain figures indicating multiple note patterns require fixed rhythms.
  8. Idiosyncrasies in the placing of notes may yield clues.
  9. Music should be appropriate to the stated theme.

A final check, once the transcription is completed, is to try to determine how the resulting music might have been transcribed 600 to 1,000 years ago, then compare this with the original tablature.

Underlying the above is also the assumption that qin music has motion tension and release. This is usually accomplished by setting up a musical expectation and then either fulfilling it or doing something unexpected. This could be theme and variation within a single piece, or it could mean playing a piece differently from the way it had been played before.

Detailed analysis of the melody Huaxu Yin

While reading the following discussion please refer to:

This study highlights the following rules, which I generally follow in doing any reconstruction.

1. There is a dominance of double and quadruple meters.

My fundamental assumption is that almost all qin melodies surviving from 15th century sources are based in duple or quadruple meter. This fits most traditional Chinese music of today (the main exception being some types of song) as well as the reconstructions I have done so far, but has not otherwise been scientifically tested. An examination of the active qin tradition also supports this idea.

This is not to say that the meter is always obvious. Often there seems only to be a sort of pulse. Often it may be quite in the background, at other times it is obviously rhythmic. For me there is considerable interest in the way the rhythm shifts -- perhaps going along regularly for a while, then maybe having a few extra beats put in, or removed. In addition, this structure is often hidden through free interpretation - especially if a melody has been played many times. However, when I first reconstruct a melody it is always quite rhythmic, and later even when the rhythm may not be readily apparent to listeners it is always still there in the back of my mind.

Some years ago the famous qin player Yao Bingyan wrote an article, in conjunction with his realization (dapu) of Jiu Kuang, saying some ancient Chinese music used triple rhythms. He based this largely on the argument that it sounded good and there were triple rhythms in Chinese poetry. Without a more convincing argument than this, for the time being I will continue assuming that the rhythms are mostly duple and quadruple.

2. Couplets are common: repeats, repeat with variation, pairs with related structure

This is also something very popular in Chinese poetry, so perhaps there is a connection. Huaxu Yin begins with a repeat: mm. 5-8 repeat 1-4. My own feeling is that the structure of mm.3-4 is related to that of 1-2, so that they also can be seen as a couplet. This is true also of mm. 9-10, which feel to me as a sort of question, and 11-12 which are its answer. Below there is some discussion of phrases which are repeated with variation.

3. Rhythms may be suggested by versions with text/lyrics.

Before determining rhythms one must decide on phrasing. Since the melodies in Folio I of SQMP originally had no punctuation, this is also related to the following section on comparing the tablature with that in later versions. Here it is discussed in conjunction with the SQMP melody Huaxu Yin (Huaxu Clan Prelude); Huaxu Yin was published again about 65 years later with lyrics added. In all it survives in five qin handbooks. Four versions (1425, <1491, 1539 and 1670) are virtually identical; the one from 1585 has the same lyrics as <1491, but the melody is very different. This pattern is rather typical for melodies in Folio I of SQMP, and quite possibly means these pieces were being preserved for their antiquarian value, and were not necessarily part of the standard Ming repertoire: pieces in the active repertoire tend to change more.

The first edition of SQMP came out in 1425. No copies of the first edition are known to have survived. What we have are a second edition from the Jiaqing period (1522-67) and a third edition from the Wanli period (1573-1620). These two editions are virtually identical, to the extent that the third edition apparently uses some original plates from the second. All indications are that the first was exactly the same. However, the last line of the first folio of the second edition has this comment added by an anonymous editor:

"The old tablature of the preceding pieces had no punctuation. Now during leisure time this humble person, using broad concepts to examine the sounds and flavors, has put marks at the ends of phrases. Everyone who understands music should examine it. Dragon collection's xinwei year. (This should mean 1511 or 1571; since the front of the book says it was published during the Jiaqing period (1522-67), perhaps the punctuation was made around or before 1511 then used during the later publication), summer, 4th lunar month; carefully recorded."

Huaxu Yin comes from Folio I of SQMP -- pieces for which Zhu Quan said, finding no players, he published the tablature without punctuation. The most widely published 3rd edition also has no punctuation. However, the 2nd edition does.

Perhaps earlier than this punctuation, however, is that of the version in Zheyin Shizi Qinpu. The surviving part of Zheyin Shizi Qinpu has 40 pieces, 28 of them identical to versions in SQMP, but with lyrics, added. These lyrics (which I sometimes refer to as "text" because it is not clear that all of them were intended for singing) had no punctuation, but anyone who knows classical Chinese can determine the punctuation rather accurately. And those less skilled in this can refer to the invaluable compilation by Zha Fuxi, Cunjian Guqin Qupu Jilan, which copies with phrasing the text for all the songs he and his helpers found in old qin handbooks. Having determined the phrasing of the words in Zheyin Shizi Qinpu, one must decide if it follows that in the SQMP.

This is quite a difficult question. After spending a lot of time studying Zheyin Shizi Qinpu, I still am not sure whether the text was added from elsewhere by the person who is supposedly only the editor, or whether he wrote at least some of it himself. This man, the self-styled Beyond-Sounds Immortal, is thought to have been Zhu Dianpei, a grandson of Zhu Quan who succeeded him as prince in Nanchang. The Beyond-Sounds Immortal calls Zhu Quan his Royal Ancestor. Generally analysts have said that the words were added because of some idea that all qin music should be sung, but that these pieces cannot in actuality be sung. My research indicates that perhaps some of them can be, but for other songs perhaps the editor or the Beyond-Sounds Immortal himself did not even understand the music being copied down.

The text is added according to a rather rigid formula: one word for each right hand stroke, two for the left hand pluck called dui qi, one for the left hand stroke tao qi; occasionally words were added elsewhere. In some places the music and text fit together pretty well -- Huaxu Yin is one such case. This doesn't mean one should sing along with Huaxu Yin, out loud. However, one can sing or recite the words silently, as one plays. In this, Huaxu Yin is unlike some other pieces. It is also not what one would expect from a random application of the formula. It is thus evidence that the Beyond-Sounds Immortal had some idea of the music, perhaps from doing the same kind of re-creation as we do today. Remember, Zhu Quan had already said no one played the music any more -- or at least he could not find anyone who could.

So Zheyin Shizi Qinpu has the same fingering but a slightly different version of the punctuation than does the second edition of SQMP.

The third source is Fengxuan Xuanpin, dated 1539. This latter collection has exactly the same tablature and virtually the same punctuation as the 2nd edition of SQMP: which came first I cannot say. Once again, Zheyin Shizi Qinpu may be the earliest source for punctuation.

Differences in punctuation generally seem to come from one version sub-dividing what another version has as one phrase. There are three potentially significant places where the phrasing differs. These occur in measures 21, 46 and 74, where they are marked by an arrow.

My solution to the potential problem in such places is generally to make the note values such that both punctuations are simultaneously acceptable. There are also other places where the phrasing can be ambiguous. It is natural that this should sometimes be the case, and that this might result in differing punctuations.

Besides having the rhythm of the music fit the rhythm of the lyrics, I suggest this compromise between versions as an example of a rhythm suggested by a version to which text has been added.

4. Similarities and differences in later versions yield clues.

The previous section mentioned that the version of Huaxu Yin printed in 1539 had some minor differences in punctuation. Other than this (and the clues from the lyrics) there is nothing to be learned from examining later versions of Huaxu Yin. Most helpful in learning music from handbooks like SQMP are melodies that were more actively played, mostly pieces in the second and third folios like Meihua Sannong and Xiao Xiang Shui Yun. When I first started doing recreations I worked on these melodies because they are still part of the active repertoire, along with Liu Shui from Folio 1. Initially I made the rhythms of the old as much as possible fit the new. Later I made further revisions based on the different notes as well as different stroke techniques used in the earlier versions.

5. Note values relate to ornamentation and fluency of finger patterns.

6. Repeated or altered rhythmic patterns may be indicated by repeated or altered:

a. Ornamentation patterns.
b. Note/finger patterns;
c. Musical contours/structures.

7. Certain multiple note patterns require fixed rhythms.

These three categories of rhythmic clues often work together. Note values relate to both finger techniques and patterns of finger techniques. Some good examples can be seen in measures 1-12 of the transcription of Huaxu Yin.

Measures 1 to 4 have the hand moving to finger positions that are quite distantly spaced, and so not conducive to fast play. The first phrase is repeated. By custom the repetition should often, or generally, be slightly different from the original. More about that later.

Measures 9 to 12 suggests some fairly definite rhythms. Note the repetition of strokes on strings 4, 5, 6, 5, 4, 3 then 5, beginning first in the 4th position (mm. 8-10), then in the 2nd (mm. 10-12). In the first case it is juan the 2nd and 3rd fingers inward on 4 and 5, then li the 2nd finger outwards on 6 and 5, gou the 3rd finger inwards on 4, da the 2nd inwards on 3 and tiao the 2nd outwards on 5; in the 2nd after the initial juan on 4 and 5 there is gou the 3rd inwards on 6, li the 2nd outwards on 5 and 4, da the 4th inwards on 3 then tiao the 2nd outwards on 5.

Li normally requires running the finger quickly over two or more strings; with the other notes, the changes in finger strokes suggest separating the sounds.

Juan provides one of the few problems in actual note interpretation. It is a technique not found in handbooks after the Ming dynasty and is open to several different interpretations. In one the first string is plucked twice, then the second; in the other the strings are alternated, like another figure called quan fu.

Although both interpretations are found in old books, one place in Huaxu Yin has a quan fu (m. 37-8) set up in contrast to the juan (m.36-7); for this reason, throughout the piece I use the interpretation for juan which is different from that for quan fu. The juan across 3 strings in mm.68/9 also suggests this interpretation.

The four phrases from the end of m.12 to m.21 illustrate a similar point. In different positions the first three feature juan 4 and 5, li 6 and 5, tiao 6, da 2 or 3, then tiao 5; the fourth phrase changes tiao the 2nd finger outwards on 6 to pi the thumbnail inwards on 6. This suggests that the rhythm of the first three is identical, changing in the fourth. My understanding of the difference is as follows. After li 6 and 5 the player must player must pull the 2nd finger back in order to tiao 6, requiring a slight delay. However, after the li the thumb is just in the right position to pi 6, so in this case the note can be played more quickly. (The astute listener will note that on my recording will observe on the last note of m.15 I play pi instead of tiao 6, thus giving the second of the four phrases the same rhythm as the last.

Concerning li followed by pi, this at first seemed a bit awkward to me, perhaps because it is not a very common sequence today. The first several times I saw it in SQMP I tended to make it slow because of this awkwardness. However, the passages often seemed to suggest something faster, and in fact, once practiced, this is a very natural sequence and almost invariably a quick rhythm, as suggested here, fits.

Of course the particular fingering selected affects tone color and dynamics as well as rhythm. Does this mean the way it is used to affect color and dynamics might conflict with these interpretations based on "natural" fingering? For example, as mentioned, some initially awkward fingering must sometimes be played fast.

Generally speaking, on virtually every line the fingering gives strong indications of the rhythm, tone color and dynamics; it is only in special -- albeit important -- places where the fingering indicates something about tone color in a way that at the same time might give misleading or false information about the rhythm. Here it is only possible to discuss the issue in terms of individual phrases; it would require a discussion of overall construction to include examples of passages with seemingly awkward fingering which should indeed be played quite fast.

What else must be done is to study the relationship between finger techniques and rhythm in current practice. As mentioned, my teacher Sun Yü-Ch'in said not to look at the tablature: copy the teacher. However, being in a hurry to learn, at home I would look at the tablature, observing that most writing was idiomatic in the sense of the fingering being natural to the tempo. When it wasn't I would sometimes ask him why. I do not recall him ever saying that it was for a special tonal effect.

Of course, this perception may not necessarily be significant. And in any case, I have not studied it in detail. My initial assumption, though, is that if an odd fingering is used it will usually affect the rhythm as well as the tone color. My cursory examination of transcriptions in the Guqin Quji (1962) did not reveal to me any conflicts with this assumption. However, I believe this collection publishes tablature from a handbook alongside staff notation of an actual interpretation, without ensuring that the performer's fingering is actually as written in the tablature.

How can one put small patterns such as those suggested above into an overall structure? In Huaxu Yin the overall structure comes from a fairly regular four-beat phrasing (though it may be interpreted very freely), with occasional beats added or dropped; within that are the sorts of phrases repeated with variations, such as have been described above.

As mentioned, the rhythm of the words in Zheyin Shizi Qinpu version seems to fit fairly well the rhythm suggested by the fingering. Look at the text of the song so far, and note that the stressed syllables come on off-beats.

This brings us to the first ambiguous phrase from Huaxu Yin. As mentioned above it occurs in mm.20-21, which in Zheyin Shizi Qinpu ends Section 1 and begins Section 2.

M.22 begins a tiao, da, tiao sequence played three times, again suggesting a repeated rhythm, though I sometimes use a dotted rhythm on one note or the other -- perhaps depending on whether I am thinking of the words. Then in m.24-5 the juan on 2 and 3 combined with a cho slide causes some problem. Cho indicates that while plucking the string with the right hand, the finger that stops the string slides audibly into the designated position. This is a bit awkward. I would be inclined to extend that note, except for the prevailing rhythm, perhaps also influenced by the text. Thus you see the hold sign written over my transcription -- though I have noticed I often miss this out when I play.

Measures 26-7 have another repeat with variation. First the 4th finger on 3 slides up from outside the 13th position to the 10th position then returns. This is followed by a zhuang from this 13th position. Zhuang is a slide up, then back; to me the implication is that this time it is faster than the same movement written out just before in full. Hence my rhythm.

Measures 28 and 31 each have a note given the ornament called by some people nao and others rou. The interpretation of the rhythm here is based on the necessity of holding these notes long enough for the ornament.

The figure at the end of m.32, chang suo, indicates a nine note sequence which might be an example of a multiple note pattern requiring a fixed rhythm -- but different people have different interpretations of what the rhythm actually is. I learned two possibilities from Sun Yü-Ch'in. The one written out here, and one (see mm.13-4 of my transcription of Xiao Xiang Shui Yun) calling for two eighth notes (forming an upbeat) followed by four quarter notes, two more eighth notes then a final quarter (or longer) note.

In my SQMP transcriptions I almost always select the first pattern. This seems supported by the fact that where chang suo occurs in the text of Zheyin Shizi Qinpu it almost invariably, as here, has nine characters alongside it arranged into a four plus five rhythm that fits the first pattern above.

The end of m.34 to m.36 repeats the li pi pattern of line 3. Then the sequence from the end of m.36 to m.38 has a phrase and repeat sign open to differing interpretation. This is not strictly a rhythmic problem: the actual notes played is also open to question. The instruction with the repeat is, "Twice do; after do quan fu." This sort of instruction does not occur today, and so perhaps this is why some interpreters ignore it or interpret it to mean do the repeat, then repeat it again changing the juan to quan fu. I interpret it to mean "do twice, the second time changing juan to quan fu." I selected this interpretation because it allowed me to be more consistent in interpreting later occurrences of this repeat-with-change instruction. It comes five more times in the piece.

There are four more occurrences of repeat-with-change instructions. My interpretation of those in mm.40-42 is, "do twice, add nao"; for mm.65-6 and 67-8 I have "do twice, the second time adding da 4 (which I add at the beginning of the repeat on the assumption that if it came after the repeat it would have been written separately, not as part of the repeat); and I interpret "mm.66-7 as "do twice, the second time go (only) as far as gou 6, add nao." Others interpret all these differently, usually repeating the phrases exactly then adding something more at the after the repeat instead of during it; as stated, this does not seem logical to me. My main problem is in mm.40-42, where I think perhaps someone forgot to indicate where the nao should be added; at different times I add it at various places, depending on my feeling.

To me it is especially important to note that when a phrase is repeated it is often given a different rhythm the second time. If the fingering is not changed for the repeat, the person writing the tablature might simply write "do twice". However, I have noted that many (though not all) players like to change a rhythm even when the tablature simply says to repeat a phrase.

8. Idiosyncrasies in the placing of notes may yield clues.

There are no examples of this in Huaxu Yin. An example from the first page of Da Hujia in SQMP can be seen at right. The instructions (after gou 5) are gou (no stroke is indicated, so by default continue to use the stroke indicated in the previous note) 4, 5, 6 and 7, but 4 and 5 and placed together, as are 6 and7. This could simply reflect something about the printing method, but because Zhu Quan was so careful about his manuscript I have decided to consider the possibility that this was intended to give some rhythmic indication.

9. Music should be appropriate to the stated theme.

The obvious need for the music to be appropriate to the stated theme creates a complex problem. First, Zhu Quan himself stated in his preface that he changed some names that were inappropriate, but he doesn't say which ones he changed. Next, players in China generally comment on this aspect but say little about the technical details discussed here, so emphasis is here placed elsewhere, in part to fill this gap.

Finally, it is not at all clear what the established cliches were in Chinese music at that time. We only have general comments on particular finger techniques having such and such a symbolism.

(10. The music is fundamentally rhythmic, but the rhythms are often freely interpreted.)

This comment, which basically re-phrases what is written in #1 above, has been added since I initially wrote the paper (see further comment.)


This paper has largely concerned my effort to play the notes in ancient Chinese qin tablature exactly as written. The crucial matter of comparing my versions with those of others has been difficult, in part because, as stated above, published transcriptions usually give the notes performed alongside the original tablature, without indicating whether the performer has in fact changed some of the fingering. Another factor is that traditional players have proven very reluctant to discuss specific details of their dapu.

Qin music published in the early Ming dynasty has musical structures combined in such a way as to give the music sense and motion. The above gives a preliminary idea of how I think my search for relationships between finger technique and rhythm have led me to some obvious structures which allow the melodies to make sense to me. Analysis of a longer piece would bring out some of the other structures.

Not all of the musical decisions are based on such logic. Examples of this are the phrases from Huaxu Yin that in my transcription are put into bars of three beats. Here I can think of no specific argument against this (as I might find for an extended passage in triple rhythm), and it sounds appropriate to me, so I tend to play it that way.


Selective List of Chinese Characters
(Numbers are page numbers in the original article)

chang suo 長鎖, 8
cho 綽, 8

da 打, 7
Da Hujia 大胡笳, 5
dapu 打譜, 1
dui qi 對起, 6

Emaciated Immortal 臞仙, 1

Fengxuan Xuanpin 風宣玄品, 6

gou 勾, 7
guqin 古琴, 1
Guqin Quji 古琴曲集, 7

Huaxu Yin 華胥引, 5

juan 蠲, 7

li 歷, 7
Liu Shui 流水, 9

Master Cinnabar Mound 丹邱先生, 1
Meihua Sannong 梅花三弄, 8

nao/rou 猱, 8

pi 劈, 7

qin 琴, 1
Qiu Hong 秋鴻, 2
quan fu 全扶, 7

Shen Qi Mi Pu 神奇秘譜, 1
Sun Yü-Ch'in 孫毓芹, 2

tao qi 搯起, 6
tiao 跳, 7

Xiao Xiang Shui Yun 潚湘水雲, 8

Yao Bingyan 姚丙炎, 6

Zha Fuxi, Cunjian Guqin Qupu Jilan, 查阜西,存見古琴曲譜輯覽, 5
Zheyin Shizi Qinpu 浙音釋字琴譜, 3
Zhu Dianpei 朱奠培, 3
Zhu Quan 朱權, 1
zhuang 撞, 8

Go to Dapu; Return to GuqinToC or to analysis.

Footnotes (Shorthand references are explained on a separate page)

1. Rhythm in Early Ming Qin Tablature (明初琴譜節拍)
Adapted from John Thompson, Rhythm in Shen Qi Mi Pu, in
Asian Music with special reference to China and India - Music Symposia of 34th ICANAS (publication details)
Hong Kong, HKU Centre of Asian Studies, 1997, pp.40-72
(ICANAS is the International Congress of Asian and North African Studies; the conference itself took place in Hong Kong from 22-28 August 1993.)

2. Additional comments on Rhythm in Early Ming Qin Tablature (明初琴譜節拍,更新的評論)
First I should reiterate what I wrote on my dapu page about the value of doing independent reconstructions and then comparing versions. This is particularly true with regard to the issue of structures and rhythms.

Since I wrote the above article I have heard or read many arguments about how guqin music should be rhythmic or should have no rhythm at all. This issue is very complex, in part because of a lack of relevant and clear historical information, in part because many people who wrote about the qin were more interested in its philosophy than in its actual music.

At the heart of this problem, particularly in historical writing, is distinguishing between rhythm (節奏 jiezou) and beat (拍 pai). In the Western notation system music the rhythms are usually specified (e.g., 3/4 [three beats per measure] or 4/4 [four beats per measure]); these may be more complex and there are sometimes instructions to "play freely". The opposite of playing freely is to "keep the beat". "Beat" can thus be seen as an element of rhythm, but it is also separate: one can have a beat without having a rhythmic structure.

Here then are three historical examples of writing on the issue of beat and rhythm.

  1. The first one comes from an essay by the Jin/Yuan dynasty mathematician 李冶 Li Ye (also known as 李治 Li Zhi, 1192-1279; Wiki). It comes from the chapter "Anecdotes 2" (逸文二 Yiwen 2) in his work 敬齋古今注 Jingzhai Gu jin zhu. The section ends as follows:

    諸樂有拍,惟琴無拍,祇有節奏。節奏雖似拍而非拍也。 前賢論今琴曲已是鄭衞。若又作拍,則淫哇之聲有甚於鄭衞者矣。 故琴家謂遲亦不妨。疾亦不妨。所最忌者惟其作拍。
    而能改齋漫錄論胡笳十八拍,引謝希逸琴論,云。 平調明君三十六拍。胡笳明君二十六拍。清調明君十三拍。閒絃明君九拍。蜀調明君十二拍。吳調明君十四拍。杜瓊明君二十一拍。 七曲皆言拍。果是希逸語否。在琴操其實不當言拍。止可言幾奏也。今琴譜載大小胡笳十八拍或十九拍者,乃後世琴工相傳云爾。【 卷八。十六段。】

    All music has beat ("pai), only qin has no beat, it has only rhythm ("jiezou). Although rhythm is similar to beat, they are not the same. The sages of the (recent) past already referred to qin songs of their time as being (as decadent as those from) Zheng and Wei. If beat is now also added, then those perverse sounds become even more so than those of Zheng and Wei. That is why qin players say that (playing) slow is not a problem, and fast is also not a problem. The thing to avoid most is keeping (a strict? an unchanging?) beat.

    On the other hand, Leisurely Notes from Nenggai Studio (能改齋漫錄 Nenggaizhai Manlu, essays by 呉曾 Wu Ceng, ca. 1162), in discussing (the melody) Hujia Shibapai (Reed Flute in 18 Pai), draws on the Qin Lun of Xie Xiyi (Xie Zhuang; discussed here, but under the melody title Wang Mingjun, not Hujia 18 Pai) as mentioning ("pai" in these titles):

    Pingdiao Mingjun 36 "pai"
    Hujia Mingjun 26 "pai"
    Qingdiao Mingjun 13 "pai"
    Jianxian Mingjun 9 "pai"
    Shudiao Mingjun 12 "pai"
    Wudiao Mingjun 14 "pai"
    Duqiong Mingjun 21 "pai".

    All seven songs have "pai" in the title, (but) is it really Xie Xiyi's conclusion (that pai here means "beat")? In qin melodies one should not talk about “pai” (as meaning "beats"), but only how many "playings" (i.e., sections). In today’s tablature one can find Long or Short versions of Hujia in 18 pai or 19 pai. It is later qin artisans who passed on these names (i.e., "pai" meaning "section".).

    That essay was somewhat vague, but it does not seem to contradict my own understanding that qin music generally has rhythmic structures (here: "jiezou"), but that the rhythms themselves ("pai" or "beats") can be quite flexible.

  2. For another example from classical writing, this linked story, though not mentioning rhythm or beat directly, tells of Zhang Dai using sticks to keep the beat when he played [perhaps also to learn the beat when he was learning a piece?]. As for the argument that there should be no rhythm, it seems largely to be based on the fact that qin tablature does not directly indicate rhythm. My own teacher said the reason for this was that the rhythm was up to you. In fact, the rhythms of the student generally follow closely those of the teacher(s).

  3. The third text comes from a passage in "琴曲音調節奏考 A Study on the Tone and Rhythm of Qin Music”, a chapter in the final volume of Yuguzhai Qinpu (1855). The whole chapter (as well as one later in that volume called "依譜鼓曲合節真詮 The True Commentary of Following Musical Score and Playing Melody With Jie (Rhythm)" has been translated by Peiyou Chang and put here on her website. The excerpt below is slightly adapted from there; she herself added the boldface.

    “"琴曲原具有自然之節",初試照譜彈,未易得成奏,習熟乃合節,漸至於神妙,此則存乎人。節即是板拍,《國語》所謂,「木以節之」也。琴曲譜不定板拍者,恐因定 固執而失入神化,非為初學者之但得其節奏而己也。然至入神化,亦莫不由於節奏而致矣。... 總以上下來去句讀,得呼吸氣息之自然為妙
    The original nature of qin music is that it has an inherent rhythm. When first attempting to play directly from the notation, it is not easy to achieve a complete performance. Through repeated practice and familiarity, the rhythm gradually becomes harmonious and eventually reaches a state of profound beauty. This process relies on the individual. The term 'rhythm' refers to the beat, as mentioned in the 'Guo Yu,' which states, "Use wood to beat the rhythm." Where the notation of qin music lacks an indication of beat, it may be due to the fear of becoming rigid and losing its spiritual transformation. However, for beginners, it is important to grasp the rhythm before anything else. Nevertheless, even in the realm of spiritual transformation, it is invariably achieved through rhythm… Overall, the movement up and down, back and forth, pauses and stoppings have to achieve a natural sense of breathing, which is the key.
    The second chapter translated by Peiyou from that volume mentions the role of a teacher in transmitting a correct rhythm. This aligns with my understanding that "it is important to grasp the rhythm before anything else" reflects the fact that a primary function of a qin teacher is to ensure that the learner can follow the rhythm as understood by the teacher.

    Here it is interesting to compare a teacher to a metronome. Some people claim that the rigidity of a metronome makes it inimical to good qin play. But does the fact that most people change very little the rhythms they learn from their teacher mean that to play freely one should never study from a teacher? Likewise, if one uses a metronome when learning a piece must this inhibit them from later playing freely, or does it actually help provide a basis from which to play freely? In other words, if the metronome does lead to rigid interpretations is this the fault of the metronome or the fault of the player?

More generally, such examples reinforce my opinion that perhaps I should add another rule regarding rhythm (call it Rule 10): Qin music is fundamentally rhythmic, but the rhythms are often freely interpreted. This largely just emphasizes what was written in rule #1 above, but it is so basic that it needs re-stating. In other words, the music is fundamentally structured, including the rhythm, but variations within these structures give life to the music. Thus the staff notation versions of my reconstructions (dapu) are usually written in fairly strict meter (usually 2/2), and have phrases of regular length (and/or specific ways of varying the length). However, the more often I play a piece, the freer the rhythm tends to become in some (or many) parts, while in other parts the rhythm remains steady.

Sometimes music may sound as though it has no rhythm; in such cases the music usually also seems to have no structure. Other times music may sound at first as though it is not rhythmic, but it does sound as though it does have some structure. In such cases, the music usually has a rhythmic structure, but the rhythms are freely interpreted. Thus, understanding the structure of a melody, especially its rhytmic structure, makes it easier to interpret the music freely.

In other words, if music has rhythmic structure this not only allows players to be freer than they would be if they have had to memorize a musically unstructured piece, it also allows the listener to appreciate better the variety within the music. The listener hears, and more easily notices (perhaps instinctively), the differences between different performances.

When this "free interpretation of strict rhythms" method is used for qin songs one can easily hear the freedom allowed by this approach. A traditional Chinese attitude towards such songs (or to duets in general) says that it is not so interesting if the two (or more) lines of music are performed in unison. It is more interesting when the music instrument and the voice play the same melody with the same rhythm, but each has its own interpretation of the melody: sometimes they are in unison, but at other times they are not together. This method, common in many types of traditional Chinese music, is called heterophony.

(Listen to this example; the song is called Wenjun Cao)






當這種「有着嚴格節奏的自由的演奏」方法應用於琴歌時,就可以明顯地聽到這種方法所带來的自由。 中國傳統上對這類歌曲(或對二重唱)的態度認為,如果同時演奏兩行(或更多行)音樂,就沒那麼有趣了。 當樂器和聲音以相同的節奏演奏相同的旋律,但它們各自對旋律有自己的解釋時,將會更有趣味:它們有時是一致的,有時又不在一起。 這種方法在許多類型的傳統中國音樂中很常見,稱為 支聲音樂)。


4. Qin melodies: "compositions" or "creations"? (see also separate article)
There seems to be a prevalent belief that calling qin melodies "compositions" elevates their worth. Though common amongst Chinese as well as Westerners writing in English, what this use of the word "compose" reveals is a bias towards the aesthetics of the Common Practice Period in Western music. Personally I think it more accurate to refer to early qin melodies as "creations".

In theological terms this distinction of the origin of qin melodies might be compared to a discussion about whether to think of God as a "creator", giving people free will, or a "composer", whereby everything is predestined or preordained. History suggests that qin melodies were created by people as they played them, and that the resulting musical structures were largely instinctive. The creator of any particular melody quite likely played it in different ways at different times, perhaps requiring students to play the melody in a specific way, but not thinking about whether or not people in the future would ever change the piece.

The irony of this is: someone trying to do an "Historically Informed Performance" of an old qin melody is in this way quite likely acting more like a composer than was the original creator (or were the original creators) of the piece. The process of looking for structures inherent in the music could be compared to the processes of composers who deliberately manipulate musical structures to give form to their compositions.

If a reconstruction of an ancient melody is successful, we can believe we are hearing how it originally sounded. This, it should be emphasized, does not mean that within the guqin tradition this original version should be considered sacrosanct, never to be changed again.