Produced by François Eckert 

Recorded June 2007 in St Silas, Kentish Town, London
Producer/engineer François Eckert (Sonmaître, Paris)
Post-production by Julia Thomas, Finesplice UK

Design and cover photography by John Haxby

Harpsichord: two-manual instrument by Ian Tucker (after Ruckers/Hemsch)
Pitch: 415 / Temperament: Niedhardt (1724)
Keyboard technician: Edmund Pickering

Sarabanda con Partite BWV 990

Though listed both in volume 42 of the Gesamtausgabe der Bachgesellschaft and in Roitzch's supplement to the Peters edition of Bach's keyboard works (revised in 1904 by Max Seiffert), the authorship of this charmingly beautiful set of variations upon a Sarabande has been hotly contested. Considerable concern has been voiced that the form and musical style of the work differ from other Bachian models, leading to speculation that it may have been composed by any number of slightly lesser-known North German composers. Although two sources explicitly attribute this piece to Johann Sebastian (two manuscripts bear the ascription "Da mio J. S. Bach"), they cannot be entirely trusted since all of the available sources derive from an unknown original source.

Weighing up all the evidence it seems impossible to assert with any real confidence that Johann Sebastian was the author of this work. However, if another source should come to light in the future, strengthening the association, it would be in many ways immensely satisfying as it would provide us with an important precedent for the ‘Goldberg Variations'. For here, as in BWV 998, is an ingenious set of variations upon the underlying harmonic structure of a simple Sarabande. Some have tried to suggest that the piece was composed in the mid-seventeenth century (Eichberg 1975), but, given the complexity of the figuration and at times highly developed harmonic idiom in some of the more expressive movements, a later date would seem more appropriate. Whatever the case, this is a superbly attractive set of variations upon the most simple and beautiful of dance-inspired themes - without doubt the work of a highly-skilled composer and absolutely worthy of its current (albeit peripheral) status in Bach's keyboard oeuvre.

One unusual feature of the variation set is the inclusion of a suite at the end, occupying the final four partitas (numbers XIII-XVI). After the seemingly climactic fanfare-like partita XII this comes as something of a surprise and marks a significant departure from the norm compared with mainstream developments in the variation traditionof the early eighteenth century. However, the inclusion of the four dance movements (Allemanda, Courante, Aria Variata and Giguetta) lends the work a certain charm and there is something recognisably conclusive about the rustic and up-beat gigue which brings the work to a close. The use of delicate style brisé textures throughout leaves little doubt that these variations were conceived for the harpsichord rather than the organ.

Goldberg Variations BWV 988

The work which nowadays we refer to simply as the ‘Goldberg Variations' was originally published in 1741 as the fourth part of Bach's monumental Clavierübung under the much lengthier title: "Keyboard practice, consisting of an aria with diverse variations for the two-manual harpsichord". The more popular title came later and owes its existence to the writings of Forkel, Bach's earliest biographer. The story goes that Bach wrote these variations for his young student, Johann Gottlieb Goldberg, who was at the time working in the services of Count Keyserlingk and, it would seem, was often called upon to provide musical diversions well into the night when his employer's insomnia became too much. Those who doubt the basis of Forkel's anecdotal recollections relating to the genesis of this extraordinary work have pointed to the unlikelihood of the thirteen-year old Goldberg being able to actually play the variations, although we do know that he went on to become a celebrated keyboard virtuoso. One thing we can be sure of however is the fact that J. S. Bach did visit the court in Dresden in 1741 at the invitation of Count Keyserlingk and, whilst BWV 988 may well have been composed before such a visit, it is entirely plausible that on this occasion Bach presented a special copy of the work to the Count.

Whatever the impetus behind the composition of such a massive set of variations, Bach was revisiting a form that he had not explored for some considerable time. Perhaps when Bach decided once again to embrace the tradition of the encyclopaedic variation set (a tradition reaching well back into the sixteenth century) he was primarily fascinated by the pedagogical benefits of such an undertaking. What better way for a composer to exhibit his compositional prowess whilst offering along the way insights into all manner of compositional techniques and innovative developments in keyboard technique? There are of course many aspects of the work that continue to fascinate and astound performers and listeners alike but perhaps the single most important achievement of the ‘Goldberg Variations' in purely compositional terms is the remarkable multiplicity of styles and forms surveyed by the composer over the course of the thirty variations. In fact, given the scale and sheer audacity of the work, it's almost as if Bach set himself the challenge of showing that almost any genre - including the most complex of contrapuntal forms such as canon and fugue - can be realized above a simple recurring underlying harmonic structure.

The underlying structure in question is of course that of the aria, which takes the form of an exquisitely beautiful sarabande. The graceful and flowing melody of the right hand almost seems to disguise the fact that it is the harmonic progression underneath which is the integral feature of the whole set. One can be forgiven for not grasping the full potential of this bass line upon first hearing - especially given the pleasant distraction of the right-hand melody - due to its sheer length. Thirty-two bars employing one principal change of harmony per measure (sixty-four if you count the repeats) constitute the underlying harmonic theme - far longer than most variation set bass lines. Four descending bass notes lead us quickly to the dominant chord of D major and commentators have been quick to find precedents for such a simple idea in the works of others such as Johann Christoph Bach and Handel. However the real beauty of invention lies in the scale of the bass line - a binary structure built upon two perfectly balanced sections of sixteen measures.

In terms of the variations themselves three different cycles appear to be running throughout the course of the work. First of all there are the character pieces (consisting of dance forms and cantabile arias) followed by the virtuoso variations (revealing the likely influence of Domenico Scarlatti's recently-published Essercizi) employing all sorts of highly original keyboard techniques, and then the sequence of canons that neatly frame each group of three movements. This pattern is almost entirely consistent with the exception of the first and last three variations (leaving the overall symmetry of the  work completely intact). There is also a strong binary structure to the work (mirroring the theme itself) with an important hiatus occurring after the fifteenth variation before the exuberant French Overture (Variation 16) opens the second half of the work.

In many ways the cycle of nine canons forms the backbone of the Goldbergs and what makes them so intriguing is the fact that the interval of canonic imitation increases by one step each time. The first canon is instantly recognisable as such with its imitation at the same pitch (canon at the unison), but as we move progressively through the canon at the second, canon at the third etc. the canons begin to create unusual - often mesmeric - sound worlds which at times might seem to verge on the austere. The level of complexity in terms of the chromatic voice leading in some of these canons is quite simply breathtaking.

From the most introverted and hauntingly beautiful variations (Variation 25) to the most exuberant and extroverted displays of virtuosity (Variations 28 and 29) the Goldbergs are as rich in variety as it is in innovation. One of the most surprising and delightful features of the work is the Quodlibet (Variation 30) that occurs just before the return of the aria. The evocative title denotes what is essentially nothing more than a light-hearted and entertaining nonsense piece, of the type improvised by the Bach family at their annual gatherings. Here the texture almost gives way under the weight of invention, containing no less (and perhaps even more) than six distinct melodic fragments woven into a sort of musical patchwork quilt. Two of these melodies are known to have been popular folk songs and the texts are intriguing. The first of these, Kraut und Rüben, formed the basis of Buxtehude's thirty-two part partite entitled La Capricciosa, giving perhaps some insight into the extent to which Bach revered this earlier variation set and the music of his old mentor.

Kraut und Rüben Cabbage and beets
haben mich vertrieben, have driven me away;
hätt' mein' Mutter Fleisch gekocht, had my mother cooked meat
so wär ich länger bleiben I might have stayed longer.

The second folk tune used in the Quodlibet worthy of mention appears to have belonged to a tune used as the final dance at a wedding party and it has been described as a sort of wanderer's melody. The text seems entirely appropriate - indeed beautifully poetic - at this climactic point in the work just before the return of the graceful aria.

Ich bin so lang nicht bei dir g'west

I have been away from you for so long a time

Aria Variata "alla Maniera Italiana" BWV 989

Given the doubts already discussed about the authenticity of BWV 990, the Aria Variata stands as the only truly authentic work of the composer's early years in this form, with the exception of the chorale partitas. It is rather more serious and introspective in its style, largely due to the choice of the minor key and the chorale-like chordal nature of the somewhat melancholic opening aria. Lest there be any misunderstanding the term ‘aria' should be interpreted here - as in the Goldberg Variations - as a short theme in binary form which establishes the underlying harmonic structure and phrasing patterns of the subsequent variations. It is once again the harmony that forms the real basis of the variation treatment and not the melodic line - a style of variation cultivated by Italian and Southern German keyboard composers towards the end of the seventeenth century. In the case of this work the aria seems to have been composed by Bach himself and amongst the most attractive features are the poignant chromatic harmony (including a bold sideways step through G minor to D minor in measure 8) and the rich embellishments (taken from the version in the Andreas Bach Book) that here form the basis of the repeats.

The variations are almost entirely written in two-part style with the occasional addition of some harmonic filling out. Even in the climactic final variation - a virtuosic tour de force - employing running semiquavers in both hands, the two-part texture is maintained. The style here is similar to that of Bach's early manualiter fugues and also bears a strong resemblance to the figurations employed in two final variations found in Pachelbel's Hexachordum Apollonis. A great deal of the musical material employed  in the earlier variations suggests the influence of composers such as Sweelink and Scheidt and the insistent use of a number of rhythmic fragments such as the figura corta (dancing figure) in Variation 1 is certainly more in keeping with earlier styles of keyboard writing.

One of the slight mysteries of this work is the large left-hand spreads in the opening and closing movements. The spacing of some of the more expansive chords allows for the right hand to come into play in order to render the harmony complete whilst in other places there seems to be no other alternative than for the left hand to stretch an octave and a fifth - something beyond the scope of most hands! In view of this dilemma it has been suggested that the piece was conceived with the pedal harpsichord in mind (certainly a possibility) or perhaps it was written to exploit the possibilities of a special type of keyboard (Krebs's version of the piece is entitled ‘all'manual Italiana'). In the absence of either of these possibilities it is left to the modern performer on the harpsichord to use his or her imagination in order to decide how best to realise these wide spacings in performance.