Bach and cantabile heaven
Introduction
In a chapter entitled ‘Freedom of Response to Music’ from his book Music for all of us, Leopold Stokowski writes: ‘A few individuals realised [Bach’s] greatness during his lifetime — now millions all over the world respond deeply and powerfully to the inner spirit of this music. There is no question of right or wrong here — it is simply a question of individual response... It is interesting and instructive to hear the opinions of educated and experienced musicians about various kinds of music, but we must always be free to feel about music in our own way — with sincerity and spontaneity’.
The internet is a wonderful and extraordinary world. Google ‘goldberg variations’ if you will. At the time of writing there were 580,000 results (nearly 3 times as many as for Beethoven’s Missa Solemnis, and over twice as many as for his ninth symphony). Anyone requiring historical, analytical, cosmological, apocryphal, or just about any topical information and context about this work, can satisfy their every desire in cyberspace. It is therefore almost as daunting a task to write a humble note about these variations as it is to add to past performances and recordings of them. I can think of no other single keyboard work that has so much baggage attached to if for both the performer and the public. A personal account of my own feelings about and response to Bach’s unique and arguably greatest keyboard piece, and an equally personal view of its history, would seem the only course of action. Members of the audience are more than likely to be familiar with the historical facts and myths surrounding the work either by education, interest or (it seems) obsession. If by chance a first-time listener is undergoing an initiation with these notes, I sincerely hope that they may inform, entertain and, as Bach hoped his variations might, inspire delectation and spiritual refreshment in the reader by sharing with them my own enthusiasm and addiction.
Explorations of these variations’ outer and inner workings, the overall structure and the internal nuts and bolts of how the variations are put together, are numerous. They range from the simple and sublime to the convoluted and ridiculous. Simply put, the ‘Aria’ (32 bars in length) is followed by 30 ‘diverse’ variations, after which Bach gave the direction ‘Aria da Capo è Fine’ — ‘Go back to the beginning, repeat the Aria, and finish’. On the simplest level, the 32 sections (two arias and 30 variations) are in numerical harmony with the 32 bars of the Aria itself. The 30 variations compositionally form 10 groups of three: the third variation of each group containing a canon between two voices — the tonal distance employed increases progressively by one step per canon, from a unison (i.e. with the two voices starting on the same note) gradually increasing to an interval of a ninth. The astute mathematician will already have sniffed a rat. Surely with 10 groups of three, the last canon should be at the tenth: the final variation, by following through his self-imposed system, should have contained a canon at the tenth? In its place Bach presents us with a ‘Quodlibet’. A fine description of this odd musical ‘experience’, even involving the family Bach, is reported by Johann Nicolaus Forkel (1749–1818), J.S. Bach’s first major biographer: ‘For now they sang popular songs, the contents of which were partly comic and partly naughty, all together and extempore, but in such a manner that the several parts thus extemporized made a kind of harmony together, the words, however, in every part being different’. They called this kind of extemporary harmony a Quodlibet, and not only laughed heartily at it themselves, but excited an equally hearty and irresistible laughter in everybody that heard them. Why did Bach place such a mishmash of rough and rude songs here? Can analysis, and musicological research really help us to answer questions like these? Will we only find false treasure, or does the Emperor actually have new clothes?
The aria
Thinking positively and with musicological help, what do we know of Bach’s great variations? They were published at the end of 1741. Described as an ‘Aria with diverse Variations’ and written specifically for a harpsichord with two keyboards, it was to form the fourth and perhaps final part of Bach’s great cycle, the Clavierübungen. The standard translation of ‘übung’ as ‘exercise’ seems to me deeply inartistic — it conjures up images of bodybuilders who have arrived at their conception of bodily perfection by hours of mindless and routine iron pumping. Admittedly, in certain aspects of conservatoire training today there is definitely a parallel — but perhaps Bach deserves better. I offer ‘keyboard education’ as perhaps a more musical translation. Study of these publications was surely meant to bring about both a technical and musical growth. One myth recently dispelled (but still propagated) is that the Aria was written earlier, as it appears in the 1725 Clavierbüchlein for Anna Magdalena Bach. Present musicological research has declared however that it was copied into the book later, after the work’s actual publication.
The Aria itself has accumulated some strangely ingrained ‘traditional’ descriptions. First, more often than not it is described as a ‘Sarabande’. Next, that it is a ‘ground bass’. It is neither and both at the same time — perhaps one of the reasons Bach called it ‘Aria’. Looking at our Aria, we hear a left hand slowly arpeggiating a standard, instantly recognisable harmonic pattern of the day — the chaconne — recognisable just like a blues would be to a twentieth-century musician. If fact, and not without import I feel, Handel had published a ‘Chaconne’ in 1733 in the same key with exactly the same harmonic pattern and containing precisely twice as many variations. This flamboyant piece ends very strangely, almost hesitantly — with a canon at the octave. It seems reasonable to assume Bach was aware of Handel’s work. Bach in his Aria, over his much-extended ‘chaconne’, adds an extremely finely wrought, single-voiced cantabile of great range and with surprising melodic and rhythmic inflections. Only in the last six bars does the texture change to a more regular and more harmonically conceived texture. Although it would pass as a chaconne, other formal considerations disallow this title. The 32 bars are formally divided in two halves of 16, with repeat marks for each half, creating the appearance of one of the many standard Baroque dance forms. The sarabande by the mid-eighteenth century was a slow dance with three beats to a bar. Music written for the sarabande often contained important harmonic or rhythmic stresses on the first and second beats of the stronger bars in a phrase, which reflected something of the steps involved. Our Aria is indeed in triple-time and the rhythm in the melodic line demonstrates the first and second beat fingerprint associated with the sarabande. By mixing musical forms, Bach is from the outset trying to blur, mix, and complicate matters to challenge us.
The immediate challenge in performing Bach’s work is to find the character for the Aria. A Baroque composer, in choosing to write an instrumental chaconne in G major, could be expected to produce music with an outgoing, strong, perhaps fairly lively character (as Handel magnificently did). A G major sarabande could be slow, tender, thoughtful and flattering. Bach’s own G major French Suite contains a good example (although there is no ‘standard’ emphasis of the second beat here in the first bar). This is a difficult formal and musical mix to accommodate. The high tessitura of the right-hand ‘voice’ seems to me to demand a certain rhythmic support and movement underneath, rather than a static and hushed meditation. A certain inertia also affords a more natural transition into the first actual variation.
The variations
It seems important with any set of variations, particularly baroque or classical, to allow them to develop from the theme or harmonic pattern in an organic manner. Later sets of variations from Mendelssohn, Liszt and Rachmaninoff onwards display a much freer approach towards character and tempo modulation and juxtapositions. With this in mind it is amazing to note how many performances of Bach’s great variations, after a slow delivery of the Aria, suddenly race off at twice the speed for Variation 1! It is perhaps time, before further analysis and remarks, to acknowledge the influence that recording has had in our world’s reception, perception and performance of this work. Currently, our understanding and appreciation of this work, for players and listeners, is inescapably affected by Glenn Gould and his recorded performances on the piano. The only pre-Gould recorded performances in common circulation today are the two made by Wanda Landowska and that of Claudio Arrau. However, today’s collective consciousness on the matter seems not to trouble itself with these earlier approaches. Gould’s first studio recording of the variations from 1955 radically readjusted everyone’s perception of what Bach could sound like with its devastating speed, brilliance and Gouldian intellectual deconstruction. From that point on, the performance and recording of Bach on any keyboard, and the Variations in particular, was changed both for performers and listeners. Love it or hate it, Gould’s individual genius is impossible to ignore. The two highly individual studio recordings differ radically in many ways, giving us different examples of extreme tempo contrast between variations. Tempo relationship is one of the work’s most discussed problems and remains ultimately a subjective interpretative decision.
The question of tempo relationship between the variations is of primary importance to the overall success of this complex musical web. What information do we actually have to help us? The answer: very little. Of the 30 variations, only four have any heading or description that might affect a basic ‘tempo ordinario’. Two of these are in reality only character markings — ‘al tempo di Giga’ for Variation 7, and ‘alla breve’ for Variation 22. That leaves the Andante and Adagio given to two minor key variations, Variations 15 and 25 respectively. We are very fortunate to have this information. The markings for Variations 7 and 22 were added by Bach into his own copy of the first edition — which only turned up in 1974! Aside from these markings, the only other clues to tempo relationship are the time signatures. These range from the ‘alla breve’ cut-C (two minims to the bar) to the combination of 18/16 and 3/4 in variation 26. Further considerations that may affect tempo (such as the use of the minor key, chromaticism and the influence of dance-forms) I shall discuss later.
It is a matter of heated discussion, like so much in analytical and musicological life, as to how a time-signature and a subsequent change of time-signature suggest specific tempos and tempo-relationships in music of this period. In my opinion, and to quote Stokowski, ‘there is no question of right or wrong here’. Clues are often given musically: the last bars of one variation often contain the musical seed for the following. For example: the first three notes of Variation 1 are hinted at in the second beat of the last bar of the Aria — an important reason not to race off in Variation 1. (Landowska in her second recording manages this magnificently.) The rising-third pairs of semi-quaver notes heard in Variation 17 are already present in the third bar of the 3/8 section of the preceding ‘Ouverture’. The semi-quaver notes that end Variation 19 slip seamlessly into Variation 20. It seems desirable to me to render the relationships as harmoniously and logically as possible: the arrival very early on (Variation 2) of a duple meter (2/4, or two crotchets to the bar) seems effortless when keeping the same crotchet pulse. Likewise, from this variation into the first compound triple (12/8; that is, 12 quavers in a bar arranged as four beats, each of three quavers) of Variation 3 — the first canon — works easily by making the quaver note relationship equal. This way of working allows for a naturalness of continuation in the music which I feel has more place in Bach’s world than abrupt senseless changes. Invoking pre-Baroque musicological proportional ‘rules’ of the Renaissance to support highly complex tempo modulations seems to me misplaced. It is perhaps more to do with our own post-Schoenbergian struggle to find musical sense and control in our modern world.
Goldberg & numerology
I have, until this point, avoided the use of the name ‘Goldberg’ in connection with this work. Forkel tells a tale of a Count Kaiserling, whose sleepless nights were soothed by an amazing young keyboard talent, Goldberg. The Count, he reports, asked Bach to provide Goldberg with variations for him to add to his nocturnal repertoire. At the present time, the majority of musicological thinking considers this story to be untrue, or at least highly unlikely. Elsewhere in his biography, Forkel refers to the work as ‘The Great Variations’. Bach’s own son, Carl Philipp, describes them in his father’s obituary exactly as published: An Aria with 30 Variations for 2 manuals. I have my own theory, admittedly somewhat fantastic, about this story. Maybe it was an ‘urban myth’ generated by Bach himself with the aim of adding just one more layer to the number games which he enjoyed later in life.
Bach’s tuning system
Another happy coincidence and musicological discovery allows me to use Bach’s own tuning system for my performance. There have been many claims and counter-claims since Bach’s death as to his own preferred method of tuning. Most of them have little actual connection to Bach himself. Incredibly to me, there is still even a camp in favour of equal temperament. In a brilliant and refreshingly musical piece of thinking and musicology, Bradley Lehman seems to have discovered ‘the Truth’, or at least a Truth. His work was published in the journal Early Music.
Lehman’s ‘Bach’ tuning system is rooted in the seventeenth century (a so-called sixth comma, meantone tuning), but — as you would expect with Bach — simply and brilliantly adjusted to accommodate the remotest key areas. I am sure that Lehman’s idea will not receive universal acceptance, but I find it utterly convincing. Where else would you expect Bach to leave his tuning system but on the title-page of The Well-Tempered Clavier, Book 1?
Structure
The modern post-Gould tradition of fast and straight delivery that infects recordings of these variations is something that I have tried to address and redress in my performances. It seems inappropriate to the harpsichord, and intrinsically rooted to performance on the modern piano. I wanted to make the work feel like a musical whole — not a series of disconnected and unrelated exercises focused mainly on achieving high-speed, rhythmically inflexible mechanical fingers. I tried to give equal importance and musical value to the less obviously virtuosic variations and retain flexibility and humanity in the more acrobatic and overtly demonstrative parts. I always tried to keep a sense of the Aria in the musical background.
Long term planning for Bach was just as important as the smaller element in his music. The publication of his Clavierübung spans 15 years — from the appearance of the first Partita in 1726 to the Variations of 1741. Each part has its own formal planning but is inextricably linked to and part of the whole. Even the key relationships and order of the individual parts and whole were carefully worked out. Each part within the cycle also has some formal similarity to the others. One in particular can be found at the mid-points of each. Here Bach places an ‘Ouverture’ — a French Overture with characteristic dotted rhythms in the first of its two sections. Such long term thought and care seems highly surprising to us now. Our pace of life and unsettled world makes this kind of long view and organisation rare.
One specific example of Bach’s skill in formal and musical design in the ‘Goldbergs’ is in his placing of the three minor key variations (15, 21 and 25). He judges them perfectly, giving them a special atmosphere and importance. Variation 15 comes directly before the mid-point. It contains a tortuous inverted canon at the fifth. It is also has a binary time signature (2/4). It is the first variation to really challenge the ears of the performer and listener, leading us into less normal territory. In Variation 21 (numerologists see the significance of 21 being the sum of 3x7), dark chromaticism is added to the ‘free’ bass-line. Again the triple meter of the Aria is rejected. The two canonic parts snake their way around each other at the uncomfortable interval of a 7th. At Variation 25 we reach the absolute extreme nihilistic far-point of our journey — musically, psychologically and emotionally. Here the harmonic contortions are extraordinary and time seems to stop — it is the most totally warped, ‘anti-matter’ version of the original Aria. The ‘adagio’ marking that Bach added in his own hand seals its fate. Ironically, this indication literally means ‘at ease’.
Cantabile heaven
In teaching his students, as we read from Bach’s title-page for the Inventions and Sinfonias of 1723, he demanded a number of things: to be able to play clearly two or three-voiced compositions — understand the compositional ideas (‘inventiones’) and character — and ‘above all’ achieve a true ‘cantabile’ from the instrument. This final and most important requirement is hardly one that seems offered or even mentioned today. More concern has been given to highly-articulated systems in playing the harpsichord which seem, as I understand things, contrary to its nature historically and aesthetically. It is a sad statistic from radio stations that the least desired and most switched-off type of music is that for solo harpsichord. This cannot solely be the fault of the instrument itself. It can, was, and should be able to make music in a noble and rich way. A heavenly cantabile and resonant approach on the instrument may switch a few more listeners on to the harpsichord.
It seems to me that the ‘Goldbergs’ are unique in Bach’s output, as they absolutely demand to be delivered on an appropriately voiced harpsichord, and played with a correct and historical sound aesthetic for the instrument. Using a piano, organ or synthesizer can (depending on the player) be interesting, but the Variations are ultimately symbiotic with the sound-world of the harpsichord. There is a wonderful scene in the Monty Python film The Meaning of Life in a hospital where the seemingly stupid doctors are asking for the ‘...machine that goes “ping”...’ The harpsichord is, unfortunately, the ultimate ‘machine-that-goes-ping’. Since its revival at the beginning of the twentieth century, it has suffered from jokes as to its dry mechanical qualities.
This cannot be historically correct, and it is our duty as players to try and convince the public otherwise. Only allowing French repertoire to be concerned with sound and insisting that Italian (the country of the voice and opera) and German repertoire (born of a total absorption of French lute music) be delivered in an ‘organistic’, detached manner is to deny the harpsichord its birthright. It firmly belongs within the ‘resonant’ family of instruments — with the lute and harp. We must try to make it the ‘machine-that-can-sing’, or at least ‘ring’.
With the ‘Goldbergs’, the unbroken influence of piano (and organ) teaching from the first half of the twentieth century still makes itself very evident. A radical reassessment of and fresh insight into the historical harpsichord to my mind never really took place. Gould’s explosion onto the scene further complicated things. ‘Organ’-istic ideas of articulation and the stiffness of mid-twentieth-century piano interpretations of Bach, continue to pervade and influence modern harpsichord playing. The ‘Goldbergs’ must take the prize for the work most recorded with an utterly dry, fast, short and rhythmically rigid approach.
The Harpsichord
My harpsichord is a copy of a 1638 Ruckers instrument with two manuals. ‘Authenticists’ might question such a choice. The search for a ‘Bach’ harpsichord produced something of a craze for the ‘Mietke’ in 1985. This was prompted by presentation at the Boston Early Music Festival in America of a copy of such an instrument. Its ‘German-ness’ and the maker’s connection to Bach inspired a number of recordings on such copies. One company even managed to record the ‘Goldbergs’ on an example of a real Mietke that was patched up for the job in Berlin. Unfortunately, without going into detail here, questions of scaling and stringing of the original Mietkes (and therefore the copies) were not thoroughly addressed. The first copies were strung in brass, and so has pretty much every copy made since. This appears to be incorrect. Such treatment in Baroque-violin terms would be like stringing a provincial German violin with completely the wrong kind of gut (or maybe even to use metal), thus creating a different tension and sound. Personally, I have always found the sound of these modern Mietkes musically unconvincing. A fine Ruckers copy offers a delivery of polyphony and lute-like sound that for me aids Bach’s music in every way.
One hugely important aspect I am keen to present in my performance is the presentation of a harpsichord voiced in a truly correct manner with the correct plucking material. Tragically, the use of real quill is something that is rarely adopted by harpsichordists in their instruments. It affects the touch and sound of the instrument enormously, making it much more sensitive, flexible and colourful in its range of attack, but quieter and more difficult to play. The heavily-voiced plastic plectra which one encounters on most modern instruments sound and feel brutal in comparison. Even the most carefully-voiced plastic will not give the same experience as an instrument voiced in quill. My instrument was thoroughly re-voiced in quill (seagull feathers) for this performance. The difference in touch and volume is certainly surprising. It immediately requires a different conception of sound. There is a softer-edged beginning to the sound that seems to coax the note into life rather than forcing it. It is absolutely more vocal — once again allowing Bach greater access to his desired ‘cantabile’.
Historically-informed performance
The twentieth-century early-music movement came to fruition (perhaps coincidentally) at around the same time as Gould’s first ‘Goldberg’ recording. From our position half a century later, it is fascinating to compare the methods that were developed in different areas of Europe to play old music on old instruments using old techniques. A quick scan through the recordings of Bach, made by the Leonhardt Consort, Concentus Musicus Wien, La Petite Bande, Musica Antiqua Köln, The Academy of Ancient Music, and The English Concert, quickly show wildly differing results. The positive realisation gained by this comparison is that performance of music still remains subjective. It is ‘interpretation’. Once this was almost a dirty word in certain early-music camps, with its resonances of ‘romanticism’. Happily, this ‘Freedom of Response to Music’ is something that has become accepted again. Each of the ensembles mentioned in the comparison above presents a very clear picture of their individual interpretative involvement, and we are the richer for it. Standardisation, as Stokowski puts it, ‘is essential to machines, but fatal to music’.
The impositions of ‘standardising’ formulas and dogmas of articulation are likewise fatal to the individual musical ideas and lines (and therefore cantabile). Upbeats do not always have to be played short. Authentic fingering can be used to aid articulation if it makes sense for the short and long-term musical line. A mindless imposition of paired slurring/fingering is not always desirable. No spoken language abides by such dogma. Why should music?
Another dangerous trend is currently in fashion — discovering and imposing ‘dance-forms’ on patently non dance-form music (such as arias from passions and cantatas). Not every piece of Baroque music in 3/4 is a minuet, sarabande or polonaise. Sometimes a 3/4 is just a 3/4. Even if certain characteristics of a dance are present, this still does not imply or require a rigid tempo. The best dancing is also expressive and flexible. Some of the variations do seem to have characters closely akin to certain dances: 4 (Passepied), 11 (Gigue), 12 (Polonaise). However, I would still only claim these as subjective associations. These characters can be very useful in trying to find an appropriate tempo and registration for those particular variations, but musical elements such as texture, concentration of fast notes, chromaticism and clarity should always be important considerations.
The ‘Goldbergs’ & me
With all these thoughts then, and at the age of 41, I came to make my first recording of Bach’s Great Variations. It was only four years ago that I first played this extraordinary work in public. Until then I had not felt compelled or ready to perform them. However, once captured by their magic, they quickly became an obsession. Any recording is only an aural snapshot of musician’s hopefully ever-developing ideas. Performing this piece ‘live’ is an invaluable learning experience and extraordinary musical journey. It is one I hope to be making, exploring different paths, for many years to come.
Only by personal exploration is it possible to discover answers — if there are answers to be had at all — to the questions posed by Bach’s unique keyboard work, ‘Aria with 30 diverse Variations’. Musicological rumination and analytical dissection can aid us, but we must ultimately find our own way. Hopefully one which will, in performance, give the listener the delectation and spiritual refreshment that Bach wished.
Why a ‘Quodlibet’ for the last variation? Google again, entering ‘goldberg quodlibet discussion’. There is a veritable treasure trove of entertaining information, gossip and ‘virtual-Bach’ help in finding your own answer. For me, I think Bach was perhaps acknowledging God in Heaven as the single Creator of perfection. I’m sure Bach could have provided the perfect concluding variation (in 4 voices and including a canon at the tenth) to crown his most perfect keyboard creation. In throwing together four secular ‘arias’, albeit in a most cunning way, he shows himself to be earthly, humble and human — achieving perfection only in our eyes and providing us with cantabile heaven.
Going back to where we began, I would like to offer this quote from Stokowski’s chapter ‘Freedom of Response to Music’:
‘The music itself will create in us the emotions of which we are capable. Sincerity in response to music is as much to be valued as sincerity in every form of life. Freedom of thought about music and freedom of reaction to music are indispensable to true culture. Just as each of us must be free in this, so we must insist on everyone else being free. Tolerance and generosity of spirit are absolutely essential to art of all kinds — in us as individuals — in our community — our nation — and the whole world.’
I can only agree wholeheartedly.
This formed the programme note for Richard Egarr’s performance of the Goldberg Variations on 19 January 2007.