ICE - WATER - STEAM
Thoughts on the Aria from Bach's Goldberg Variations
Ice - Water - Steam: Thoughts on the Aria from
J.S. Bach's Goldberg Variations
© Peter Tuite
(1)
As Bach himself writes in the preface to the ‘Goldberg Variations’ this is music “prepared for the soul’s delight”. And as a Lutheran, I think Bach would have taken this to mean that the music was a pious offering to the greater glory of God, and no mere trivial entertainment. We, as a modern audience, must bear this in mind, for despite its undoubtedly attractive qualities, this is music with a serious moral purpose: it is for the betterment of one’s soul.
The work’s title – ‘Goldberg Variations’ – is attributable directly to J. N. Forkel through his recounting of the now-famous anecdote. Everyone knows this one: The count Keyserlingk asked Bach if he could write some pieces “of a character so gentle and somewhat merry” so that his court musician Johann Gottlieb Goldberg might perform them at nighttime in order to ease his insomnia. Although we cannot say it with absolute certainty, the likelihood is that the story may not be true. For a start, the period of its composition was almost without doubt sometime between 1739 and 1740; whereas we know that Bach did not visit the Count until November 1741. It is also quite improbable that Goldberg could have performed them at this time as he was only born in 1727 – making him barely 12 years of age. What seems far more credible is that Bach presented the variations to Keyserlingk when he visited – a visit he made not least to see his son Wilhelm Friedman Bach – and it was not until later that Goldberg started to perform the pieces at the court for his employer. One suspects that the whole story was then probably recounted to J. N. Forkel some years later by Wilhelm Friedman, no doubt with some additional imaginative flourishes – or perhaps with some altering of the chronology of events. In fact I think it certainly seems possible that the piece was written for, if any virtuoso, then Wilhelm Friedman himself. Bach had already written numerous pieces for his son including the six sonatas for organ and perhaps even the first book of the Well Tempered Clavier; A far more credible story then than the one Forkel propagated, though colorful it undoubtedly is.
But what of the Aria from this gigantic work? For a start it’s actual title of Aria with 30 diverse variations itself presents a kind of puzzle. Namely, that if the 30 movements are really variations on this Aria or theme, then why is it that the theme is never heard again, or paraphrased or even hinted at until it is simply repeated verbatim at the very end (repeated, but not written out, as Bach writes Aria da Capo e Fine). Of course the variations are on the ground bass and not on any theme – but still the question remains why does Bach call this first piece an Aria? Especially when there are numerous other examples where Bach uses the term more accurately; such as in the Aria Variata or in the Chorale Partitas. It seems not unreasonable to suggest that this Aria may be a type of generic Sarabande Tendre, with its leisurely pulse, slow harmonic rhythm, harmonies made from full triads, various emphases on the second beat of the bar, a singing melody and no up-beat. Interestingly though, the fleshing out of the rhythmic texture by using an increasing number of semiquavers and indeed of a greater complexity generally suggests the character of a Sarabande Doublée – good examples of which are found in Bach’s A minor and D minor English suites.
The Aria from the ‘Goldberg Variations’ though also shares something in common with the Aria from the G major French suite. Apart from sharing the same key, the two bar phraseology of the Goldberg can be revealingly compared to the four-bar phraseology of this Aria from the French Suite, which also melts into a semiquaver movement towards its end. In both pieces the slurs are designed it seems to prevent a potentially perky way of playing the shorter notes implying, thus, a very long line in the RH which the unique notation of bars 27-30 confirms as a kind of Sostenuto. The long appoggiaturas also conform to the neo-galant style of the Sarabande as found in say Bach’s own D-major Partita.
(2)
And what of interpreting the Aria; how might it be performed? Perhaps a first step to answering this question lies with how one might view the Aria in relation to the work as a whole. Is it part of a continuous underlying structure, where it and the other thirty variations form part of an organic whole? Or is it simply an independent theme, with its own character and personality, which is ultimately independent of the other so-called personalities in the piece? I think it difficult to sustain the argument that there are variations on a theme; after all, where is the theme repeated or even referenced for that matter? It appears to follow from this rhetorical question that therefore what we have in the Goldberg Variations is a set of independent pieces, thirty experiments on the ground-bass which are explored in a series of discreet and different genres. And indeed, giving a marked character to each piece – including thus the Aria – was not without precedent in Bach’s writing: for example, in the Aria Variata (which sources suggest was subtitled by Bach “alla maniera Italiana” or “in the Italian style”). And why in the Italian style? It was probably for the simple reason that each of the movements in this piece use different tempos and tempo-signs (This practice in variation writing was generally unfamiliar in Germany at the time, but learnt by Bach from the last movement of Corelli’s Violin Sonata op. 5, known popularly as La Folia, and hence, the reference to the Italian manner. Perhaps in the Goldberg Variations, Bach takes this principle to a new extreme).
Such a view of the work – as a kind of pot-pourris of various little genre pieces - was expressed by the young Glenn Gould in his first, and now famous, recording of the piece in the 1950s. As he put it regarding the Aria, “in short it is a singularly self-sufficient little air which seems to shun the patriarchal demeanor to exhibit a bland unconcern about its issue, to remain totally un-inquisitive as to its raison-d’être”. One can indeed hear in Gould’s performance a ‘bland unconcern’ for the Aria’s tempo and characterful relationship to the rest of the work. Gould thus strives for an unaffected reading of the Aria; unpretentious, uninhibited and un-self-conscious, whilst maintaining a carefully calculated and tasteful modesty in its projection.
And I think consequently, numerous attributes stand out: the performance has a ‘natural sounding’ tempo; the trills are all unmeasured; the sound-world has a crystalline clarity; the phrasing has a generous sweep about it; and yet the whole projection of the sound remains intimate and unforced. There is almost no sustaining pedal either and the una corda pedal is unused, lending the sound a direct, simple and unaffected air. The rhythms are also unfussy, with little rubato – though it is mistake to think that there is no rubato, as Gould carefully allows the music to breath in relation to the particular phrase in question. The whole effect is to produce an Aria which is beautiful but nonetheless small in scale like a Japanese Haiku; and of course this, as Gould states, thus gives nothing away as to the size of what is to come. The crucial point being though that Gould is free to characterize the piece as he sees appropriate to its properties alone, without need to tie his interpretation to a larger unit.
(3)
Having made an argument for seeing the Aria as part of a collection of individual gems though, one can also clearly and legitimately see the Aria as part of something unified and indeed highly organized. Apart from the obvious observation that the variations are grouped into threes, one could also attempt to argue for some sort of unifying principle which connects all the movements together. Thus, one’s performance of the Aria would have to be considered in light of the rest of the work. And so unsurprisingly some scholars have duly, even bravely attempted exactly this. For example there is the cosmological allegory as put forward by Humphries in 1984/85. The thrust of this is that nine canons in the work relate to the nine spheres of Ptolemaic Cosmology. Humphries puts forward the idea that there are three cycles present in the work: (1) the aforementioned canon cycle, whose time signatures provide all possible parings of the digits 2, 3 and 4; (2) the planetary cycle which consists of 12 movements (nos. 1, 2, 4, 7, 10, 13, 16, 19, 22, 25, 28 and 30); (3) the virtuoso cycle which consists of nine movements (nos. 5, 8, 11, 14, 17, 20, 23, 26, 29); and so nine of the variations are said to correspond to nine of the planetary bodies, etc.
Another I think equally rather imaginative theory is the one often called the Rhetorical hypothesis. The idea here is that Bach’s score serves as a kind of defense against Scheibe’s criticisms and the model used is arguably similar to Quintilian’s Institutio Oratoria (Quintilian was a 1st century rhetorician whose work was still taught in some 18th century German schools). My principal concern though lies not in probing the merits or weaknesses in such postulates; others more eminent than I have done so with great rigor. Rather I am more interested in the fact that such theories illustrate above all else a deeply entrenched desire on behalf of their creators to unify the work as a whole.
And again we can see this self-same desire illustrated most clearly in Glenn Gould’s equally famous performance of the ‘Goldberg Variations’ in 1981; a performance which unexpectedly turned into a valedictory statement, where Gould’s career was bookended by seminal recordings of Goldberg, mirroring the Aria’s bookending of the piece itself. Gould’s vision for this second take on Goldberg was radically different from his first. Here he worked out an entire system of tempo and rhythmic relationships in such a way as to connect each piece to the whole. Thus unity is achieved through a kind of underlying pulsation which runs continuously through the entire performance. As a result, his take on the Aria is no accident of mere eccentricity as some critics have mistakenly and rather ignorantly surmised. But rather it is a product of a whole intellectual model which Gould constructed in relation to the work. And so hence one finds the following striking characteristics: the extraordinarily slow tempo; the measured trills; the spacious conception of its design. Gould tightly controls every aspect of the Aria’s rhythm and tempo as it forms part of his overarching rhythmic plan for the work in its entirety. This is why the Aria is played so slowly; this is why the trills are so measured. One is reminded of the observation that certain phenomena despite their outwardly different appearance can nonetheless share deep underlying properties of form – like ice, water and steam; or like the musical score, the musical performance and the grooves on a record. These diverse variations in this recording share a similar characteristic. Gould at this stage in his life saw the ‘Goldberg Variations’ not simply a rag-bag of little gems but as a continuous musical structure, where each has the relationship of the whole to the parts and the parts to the whole.
(4)
One of the striking features of Gould’s Bach playing, and perhaps one of its most influential aspects, is I think his use of staccato, detacher sound. Although not used here in this performance of the Aria, it is found throughout his Bach playing. And it raises the whole issue of imitating the Harpsichord (which is one of the original reasons for it been used so extensively). The argument goes something like this: Bach wrote his keyboard pieces largely for the Harpsichord; yet I am playing this particular keyboard piece on the piano; so in order to be more “stylistically sensitive” I should try and imitate some of the Harpsichord’s textures and sounds. Now although the merits of this argument could really be debated ad infinitum, I think what is most interesting about it, is that it ignores much of the vast array of sounds with which the Harpsichord is in practice capable of making. To put this more succinctly: we think of the Harpsichord, and we think detacher lines. Whereas, in fact, we should think of a far larger and more complex vocabulary of sounds: the problem is thus not so much that pianists often try to imitate the harpsichord in Bach playing, but rather that they do no justice to the Harpsichord’s true capabilities – its true range of sounds and articulations.
If we take, for example, the recordings of Karl Richter, we can immediately get a strong sense of how varied and subtle the instrument can be. Although Richter was not really that interested in what later became known as the ‘Period Performance’ movement which took off in the 1960s, he was nonetheless a passionate keyboardist who unapologetically used then-modern harpsichords for his renderings of Bach’s music. In his recording of the ‘Goldberg Variations’ for Teldec in 1958, one hears the Aria presented in the most melodious of terms; yet it remains austere and beautifully measured. One is struck by the sheer variety of color and phrasing. What I think Richter achieves in fact is a sense of elegant beauty: there is a great freedom in trills and ornaments unlike in Gould’s 1981 recording whilst the use of a healthy rubato in the phrasing adds to the rhythmic flow of the music; there is the tremendous grasp he shows of musical gesture; and whilst there is undoubtedly rhythmic rigor in the playing the performance is by no means chained to any temporal over-arching idea. Rather, Richter demonstrates that many pianists’ preconceptions of Harpsichord sound are not only untrue, but such preconceptions only serve as an unfortunate and perhaps even pernicious influence on their interpretations. If we listened more to performers like Richter, I think maybe we would expand our vocabulary rather than impoverish and diminish it.
Although Richter was not part of it, the Period Performance movement which took off in the 1960s went on to have a huge impact on Bach playing. I should state straight away that I do not wish to get into the muddy waters of the debate over the merits and validity of its approach; others more eminent than have I done an exemplary job of exploring these rather thorny issues. Instead, I am more interested in simply investigating how such views affect and ultimately perhaps even determine a musical performance. But so said, the thorny issues do not stop there. For applying the fruit of Period Performance research to performing on modern concert instruments is a knotty problem and one which requires careful navigation and maybe even a little compromise.
There have been noble attempts at solutions to these problems and perhaps one of the more noteworthy of them can be found in Paul Badura-Skoda’s book Interpreting Bach at the Keyboard. It has to be said there are many excellent observations made in this fascinating work. To name just a few: He makes an excellent case against the prominent over-use of the auxiliary pralltriller as opposed to more stylistically contiguous and correct main-note pralltriller. The argument is in two parts: firstly that it was hardly ever used during Bach’s time and secondly that it often interrupts the melodic line, rather than ornamenting it (as it ought to). He makes some excellent observations about the Goldberg Aria: for instance there are numerous examples of places where main-note rather than auxiliary pralltrillers should be used and there are also some interesting places cited where overly delayed resolutions of appoggiaturas can cause difficulties with voice leading. An excellent example of such an informed approach to performing the Aria can be found in Murray Periah’s recent recording with Sony. It is a model of sensible, well thought through and considered playing which strives to balance knowledge of performance practice with the realities and possibilities of performing Bach on a modern concert grand. (Interestingly Periah has had made a comprehensive study of Schenkerian analysis and applies Schenkerian ideas to performing Bach, though it is difficult to say however, exactly how such analyses affect the performing decisions.) Periah’s interpretation radiates the kind of confidence that comes from knowing one’s subject and as such puts forward an untroubled, ‘objectively’ beautiful air.
(5)
An entirely different response to the challenges of performing Bach at the piano in a stylistically informed way can be found in Willem Kempff’s DG recording from 1970. His solution to ornamentation in the Aria for example is to simply abandon it altogether in the rather mistaken view I think of seeing the piece in its present state as a kind of doublee version of a simpler sarabande. What we get therefore is some kind of stripped down, skeletal version. It is undoubtedly fascinating and possesses a kind of austere beauty but I suspect that something quite crucial gets lost in this spring clean; for surely much of the Aria’s wondrous beauty stems in part from the flourishes Bach so imaginatively provides.
But perhaps one of the most baffling interpretations of the Aria that I have yet heard is the one recorded by Alexis Weissenberg in the 1970s for EMI. It displays an incredible take on the meaning of its title, Aria, which Weissenberg takes to mean as a kind of ornate Handelian operatic aria; hence the RH part is highly projected and free in its phrasing and shaping and sounds almost like an other-worldly yet equally flamboyant castrato. The ornamentation is highly elaborate as Weissenberg applies an 18th century French Rococo aesthetic to matters whilst the use of rubato is also quite striking. And although one might query the validity of Weissenberg’s historical and even stylistic accuracy, in some ways its sheer beauty, imagination and even courage renders such complaints slightly irrelevant.
In an interview with the South Bank Show, the theatre director Jonathan Miller put forward the idea that his productions were not so much grand interpretations as simply ‘thought experiments’. He argues that a production is no more than a hypothesis as to what the work might be; or what it might contain for that matter and points to the word ‘play’ as suggestive of the light and shifting nature of interpretation. Performances are hence like experiments which sometimes work and sometimes fail. Although one may query the idea, it hardly negates the importance of educating oneself in matters of stylistic concern. All too often one hears performances of the ‘Goldberg Variations’ that play foot-loose and fancy-free with the text and its implicit stylistic boundaries. And although no one seriously disputes the importance of individual imagination, tripping-lightly along the borders of stylistic abandon is a practice best undertaken when most informed.
Although, my sense is that there are certain important scholarly considerations that cannot be easily swept aside, my ultimate conclusion is this: that whatever one does in this piece – however one interprets the Aria – one must confront a whole range of issues; and only if the performance grapples with these issues (in the deepest sense) does that performer have any chance of producing an interpretation that endures.
