Ming versions of Gaoshan and Liushui – An Experiment

Debut performance of the merged SQMP versions, Oct. 26. Photo by Shin'ichiro Hieida, composited by the author.

Personally, the piece Liushui (Flowing water) has been an important work in my repertoire – the Tianwenge version (Qing dynasty) being a level 9 piece recognized by the Central Conservatory, I have studied it 3 times (failing the first two times), and finally mastering it to perform at my debut into the guqin career at the 2005 Toronto Kiwanis Music Festival. While this piece is well-performed, well-played, and well-researched, I have always had the notion of wonder as to the ties with its sibling piece, Gao Shan (Lofty Mountains). This is even more so if one looks at earlier Ming records of the piece – they are almost parallel, save some parts that signify the unique symbolisms of the theme.

So I thought, if the piece was divided around the Song-Yuan period, why not try putting them back together? An experiment has then been planned, and is now underway.

***

Boya plays to Ziqi Lofty Mountains and Flowing Water

When it comes to the research of Ming scores, the Shenqi Mipu (Zhu Quan, 1425) is the first stop for reference. What better reference is there than John Thompson’s articles for Gao Shan and Liu Shui for this very score? After comparing the audio recordings by Thompson and his notes, I’ve come to the conclusion that the compilation of the two pieces back into one can take three approaches:

I. Put the pieces one next to the other, solo.
II. Paraphrase the pieces and merge them into a common tune with distinct themes, solo.
III. Put the pieces one on top of the other, in synchronized duet.

In order to retain the most amount of original score available, choice III seems best, but also requires two players of high calibre and some mutual coordination – something I wasn’t able to acquire until a few months ago, when I suggested the idea to my friend and another fellow of the Xi Kang Qin Society – Alex Tsang of Toronto.

For the purpose of this experiment, we have examined the scores and existing dapus/recordings of all major Ming dynasty manuscripts containing the two pieces. We began preliminary trials of merging the two scores from the SQMP. Before I go on describing the details involved, let me inform all readers on the nature agreed by us two on this dapu project:

– We agree that the nature of dapu is not to ‘recover ancient ways of playing’, but ‘to innovate on playing techniques based on existing material’,
– We agree that this merging project is unconventional, and may involve major modifications of our own, therefore dismissing all ‘historical informity’.
– It might not work.

On Oct.18, the first five sections (divided according per Johnson’s standard) of the SQMP version were successfully merged, to much of our surprise. Here are some notes on the qualities of the results (hopefully with audio to back up, when we’re done):

– Despite Gao Shan having 3 notes to start and Liu Shui only 2, they should be synchronized at the cuo. We think the first note sounds better with two notes than have the Liu Shui score come in later (as it starts on Gao Shan’s second note).
– On many occasions, the two scores play on the opposite ranges of the guqin’s scale. We considered this natural and act as contrasting harmony of one another.
– To ensure that these two pieces are indeed played together, phrases are divided on the two pieces, and we sync in with open notes – which is common and synonymous with one another on the ascent of the 3rd musical phrase of both pieces.
– The Gun/fo movement followed by the suo occur at different times for the two pieces, but should come together as one continual chain, with Gao Shan playing the Gun just as Liu Shui goes into the suo.
– On the end of Section one, the suo movement with two strings are played TOGETHER (as per Yao Gongbai’s recording, instead of Thompson’s separate treatment)
– Section 2: On 4th hui harmonics, the Liu Shui score should be main melody as Gao Shan plays accompaniment. We cut the last repeat of Gao Shan, deeming it redundant.
– Juan/diejuan of 2 strings should be ONE sound on each string (2 sounds total), instead of two (as per Thompson). This is from Cheng Gongliang’s recent analysis.
– Section 3: Gao Shan becomes main vocal, Liu Shui plays quietly. The two scores should end together.
– Section 4: Liu Shui plays on 1st hui harmonics, and is accompaniment as Gao Shan plays a new theme.

*Cut here*

On October 24th, a second session has worked the piece to the end, but much refinement is needed.

– By Thompson’s standard, Liu Shui had at least 16 sections and Gao Shan only 10; we determined in this session which parts were solo and which were played together (since Gao Shan would long be finished if we matched up). Besides, there were obvious sections that didn’t match each other, as each carried a distinct theme.
– Section numbering past 8 had to be reorganized according to theme instead of Thompson’s divisions.
– Again, key points in matchups are in the dajian’gou and other similar circular motions.
– A solo section for Liu Shui from the second last line of left open page 25, until the thrice-repeat plus jiangou on the 6th line of right open page 26.
– The two play together (Gao Shan on Thompson “section 8”) until Liu Shui’s nian movement on the first line of the left open page 26, then Liu Shui goes into solo again.
– Liu Shui’s solo does not end until the second repeat of the theme line and entering ritardando on the fourth line of left open page 27, and Gao Shan (“section 9” stops again at Liu Shui’s thrice-repeat plus jian’gou.
– Some confusion was left over whether a solo section was to be put for Gao Shan (“section 9”) due to our fatigue (working after dinner, and past 12:30AM), at just before the repeat Li phrases that conclude open page 27 of the Liu Shui score. Perhaps Gao Shan can repeat just once more on section 9? Or maybe not.
– If not, Liu Shui keeps playing solo. The two scores play their final movements together starting at Liu Shui’s Wang 汪 marker on the last page.
– To make sure the piece ends together, the identical finale must be synchronized (last line on right page of Liu Shui open page 28).

Much work on refinement and looking for other places for sections that can be matched up needs to be done. More practice perhaps on Sunday (26th).

*Cut*

Performance on 26th was satisfactory, but still many mistakes were present (we panicked in the middle). Here are some major modifications for this performance (which may need to be resolved again):

– Gao Shan section 8 (per Thompson) is used twice, with the second repeat at Liu Shui’s Da to Chuan marker (大-川), and fade out at its natural end (which should result about one or two musical phrases after these markers, but just before going onto a new musical theme.
– Matchup in sections past Liu Shui 10 and Gao Shan 7 are essentially messy. More synchronizing markers are needed. Dynamics also need to be fine tuned for these sections.
– We tend to rush later on, as part of the guqin (mal)practice to increase speed over time. This needs to be controlled.
– Increased attention is required at the end, especially the visual communication between the players to sync.
– Differentiation in school/tradition must be accounted for, as each person treats musical phrases differently. A concensus must come to being.

We will work on the Xilutang Qintong version after we refine this SQMP version a bit more. Maybe we can perform this a few more times, first being a presentation towards other TorGuqin members.

*Cut*

On Oct.31, We worked on the XLT version, to which we were able to generally finish all in one night (Amazing!). Here are some points about it:

– John Thompson noted that XLT is “quite different” to SQMP, but fails to elaborate further. Compared to the SQMP version, both pieces are slightly shorter, as parts of the pieces’ repeats are truncated and fingerings redone (some get simplified, while others got more complex).
– Alex notes that while the piece is overall the same in structure, Gao Shan’s line in section 2 has been moved up an octave, playing pressed notes at 5.6 instead of 9th hui. To our context, could this mean that the focus during the performance then would shift MORE towards Gao Shan? Since Liu Shui is playing harmonics then, that would technically move the pitch gap closer to just one octave rather than two.
– There is a “climax” theme known as section 6 on both pieces. After much arrangement and research, we have concluded that Gao Shan 5 and Liu Shui 8 are unavoidably unique and must be played solo in their respective occasions. If one wants to play these TWO main themes together (which sort of works, but would eliminate the hearability of the uniqueness to the whole thing), one will have to move the Gao Shan theme until AFTER the “climax” section (To the time of Liu Shui 8), or vice versa. We prefer it the way as it is, to preserve the score’s originality.
– Section 8 of Gao Shan equals Section 9 and 10 of Liu Shui in length. Matchup of this compared to the SQMP was much easier.
– It can be said that this piece is potentially more ‘original’ than the SQMP version due to its higher degree of similarity (hence adaptability) of the two. SQMP’s equivalent of XLT’s section 8 was much larger (to which Thompson divided into multiple sections, since the musical themes were separately distinct). Perhaps Zhu Quan obtained an “altercated version” after the piece’s split?
– Liu Shui 8 (its unique theme) is, aside from its heavy sliding theme, sounds quite similar to our modern day Tianwenge version of Liu Shui’s 72 rolls and bubbles (gun-fo). The theme of rolling water is eternal, but perhaps as Thompson suggested, the ‘rapids of Sichuan’ are indeed a different inspiration than ‘the broad flows of the Yellow River’ – or in the Xilutang case, ‘the many meandering rivers into the great Yangtze River’. It is possible that Wang Zhi just got these scores from some old Song manuscript (as he did for many pieces), but also possible (perhaps?) that he has added some personal flavour into it (note the almost-intentional way of complication in this piece compared to Liu Shang, XLT’s version of Drunken Madness).
– Perhaps it is the simplicity of the XLT version’s layout, but matching the timing in the finale came much easier (less notes for Liu Shui) than the SQMP. Maybe it’s just that we suck (and our tempos are messed up), or maybe the treatment in this version has helped us indeed in quicker understanding.
– Except for one situation in the XLT score, Thompson’s treatment of playing each string separately in a chain movement (suo) will completely throw off synchronization. The Guan Pinghu recording of (“old”) Liu Shui made a correct judgement in playing those two strings together.

Perhaps I’ll get Alex to write more on this report. We will practice these further, and hopefully write up a matched score (without five-line staff though, that would be enormous work for us). This report will also, hopefully, be revised into a more formal essay for publish, preferably in some academic conference.

*Cut*

During the yaji on Nov.9, two recordings were made. This performance was not as well-performed as the previous Amateur Musician Concert one, but one can get an idea of how it works.

See Angle 1 (Facebook)
See Angle 2 (Google Video)