Several things have happened since I started this series of articles about my learning and study of Schubert’s penultimate piano sonata. The first is that after a long period of reflection and time away from the piano following my disappointing Fellowship diploma result, I started playing the sonata again – and with a very different mindset from last year – and the second that I have been commissioned to write a series of articles about the sonata for ‘The Schubertian’, the journal of the Schubert Institute (UK). This allowed me spend more time analysing and studying the music away from the piano, in conjunction with some serious reading and listening, all of which has thrown new light on the music for me and enabled me to approach it differently, and (based on my teacher’s comments at my latest lesson), more successfully.

The following article on the first movement of the sonata combines simple analysis with some personal thoughts on practising, interpretation and musical narrative.

“The performer only has to follow Schubert on his journeys and recognize its various stations”

Andras Schiff

The first movement has a symphonic sweep in its generous breadth, and an appreciation of Schubert’s orchestral writing is essential in approaching this movement, in terms of its textures, implied instrumentation and narrative flow. Much is made of Schubert the spinner of beautiful lieder melodies in his piano music, as elsewhere, but in the later Piano Sonatas we find Schubert the composer of tautly textured string quartets and large-scale orchestral writing. To better appreciate and respond to these aspects of the sonata, my “further listening” on disc and in concert has included the ‘Great’ C major Symphony, D944, the String Quintet in C, D956 (which Schubert composed during the final months of his life), and the String Quartet in D minor, D810, as well as the late piano music (the Impromptus) and of course the other two sonatas which comprise the final triptych. Such listening has proved invaluable in my understanding of Schubert’s distinct soundworld and idioms, and incorporating these sounds and textures into my own interpretation of the piano sonata, alongside my personal “vision” of the work, has enabled me to create a performance which is, I feel, three-dimensional and rich in orchestral detail.

FullSizeRender
Autograph score of the first draft of the first movement in which the opening sentence is more simplified than the final version

The movement opens with a majestic six-bar sentence comprising declamatory chords in the treble which leaves us in no doubt that this music is in the key of A major: there are no less than 14 A’s in the top voice in this sequence. For me, this opening sentence has all the grandeur and poise of the introductory sequence in a Bach keyboard Partita, and its “stand alone” quality is made all the more impactful by what follows it.

Next a sequence of descending arpeggios, marked piano, which could have come straight out of an impromptu with its idiosyncratic intimacy. The contrast between this section and the opening is striking and is the first example, of many, of Schubert’s musical and emotional volte faces which pepper this work. In bar 8 another cyclic motif is introduced, the short-long “ta-tah” rhythmic gesture (here a crotchet followed by a minim) which feels like a musical intake of breath in surprise or wonder, almost an “ah ha!”. This motif will also appear in various reincarnations throughout the work. It requires precise articulation and little or no pedal so as not to blur the transition from the first note/chord to the second, to retain an element of surprise, and to ensure the upward harmonic movement in the bass is clear; this is particularly important when the motif appears again as a quaver and minim separated by a rest (for example, at bars 23 or 28/9). These apparently tiny details create remarkable breathing spaces and suspensions in the music and clarify the structural expansiveness and improvisatory character that pervade this movement (and also the finale). The suspension from bar 6 is further reiterated in bar 13 and is not fully resolved until the A major harmony of bar 16, where the opening sentence and successive arpeggiated figure interact in another new idea whose bass line reflects the opening sentence in its ascending chords.

In just 50 bars Schubert gives us so much material. Varied, dramatic and contrasting in both weight and pace, it is far more than one would expect to find in the exposition of a traditional Classical-era sonata, and here, as elsewhere throughout the sonata, we see Schubert’s rich inventiveness, his desire to explore new ideas and his use of the piano sonata form as a vehicle for vivid experimentation and wide-ranging emotional impact. The challenge is how to integrate all these ideas while also retaining the improvisatory/evolutionary character of the music. In fact, I prefer to follow Schiff’s suggestion – that one simply follows, responds to, and trusts Schubert’s wanderings, moment-by-moment: as mentioned previously, the cyclic motifs serve as unifying elements in the work, drawing these varied strands together.

The exposition’s second subject, introduced at bar 55, feels more like a traditional second subject: scored in the dominant (E major), it is a simple lyrical melody accompanied by string quartet textures. At the end of this section, one might expect a double-bar and an indication to return to the opening of the music to repeat the exposition, but instead Schubert introduces a turbulent extended chromatic passage leading to a climax of descending arpeggios which recall those from near the opening, followed by a declamatory, orchestral section (mm. 105-111). Despite the sforzando markings, I resist the temptation to give this passage a really full-bodied Beethovenian forte: this is, after all, Schubert not Beethoven, and I feel his dynamics are often psychological rather than purely physical, here suggesting an intensity of feeling rather than sheer volume. Greater emphasis in the bass helps to reinforce this. This section ends on a suspension: a whole-bar rest of complete silence before a passage based on the second subject. This pause needs to feel absolute, with a sense of “listening into silence” (Brendel), to create a magical contrast with the gentle lyricism of the passage which follows and the close of the exposition. The decision then is whether to repeat the exposition (which I do), or proceed straight to the development. Whatever the decision, there is no doubting the impact of the extraordinary modulation in the second-time bar (m 129) where the music moves into C major and the development is heralded by gentle pulsing quavers.

The development section, like the exposition, is not a development in the strict sense of the classical sonata structure: Schubert hardly develops the preceding material at all and instead a last-minute idea from the close of the exposition (mm 121-122) becomes the main motif. The section begins in C major but quickly oscillates between C and B major, which, together with the register in which it is played, creates an ethereal, almost hypnotically suspended atmosphere. The effect is further enhanced by moments of ambiguity in the modulations, where the music hovers momentarily in a minor key, as if passing, albeit fleetingly, into another realm entirely. The LH chords suggest a string accompaniment and should not be too “chugging”. Nor should this section be obscured by too much pedal. I aim for a Mozartian clarity here with little or no pedal: the overall effect should be “heavenly” and dreamlike.

A dramatic descent into the lower registers heralds the further development of this theme in the minor key. Now the atmosphere shifts again, to intimate and passionate, the pulsing quavers remaining as a unifying element. These continue as the music moves into the preparation for the recapitulation, now firmly in the dominant, as if the drama and darkness of the minor section is already long forgotten.

The recapitulation is traditional: it remains in the home key of A major, while the second subject is presented in C major. But at bar 219 the third motif from the opening is stated in A minor and an octave higher. This new variant echoes, momentarily, both musically and emotionally the minor-key sections from the dreamlike sequence of the development before the music moves into warm F major.

The coda restates the opening sentence, but in a much more hesitant manner: marked pianissimo, it is interrupted by whole-bar rests with fermatas, while the left hand imitates pizzicato strings. It feels like a nostalgic reminiscence of the opening sentence and to enhance this, I believe there is justification for broadening the tempo and lingering over certain elements, such as the appoggiaturas in bars 337 or 343, to create a wistful, hymn-like atmosphere. The movement closes with gentle ascending arpeggios, mirroring those from the opening section, the penultimate of which is in B flat major ending on an augmented sixth, which creates a sense of uncertainty, before the final gentle A major arpeggio closes the movement. The question of how to pedal these arpeggios is ambiguous. In both my Henle and Barenreiter scores, there is a pedal mark at bar 349 only. Any pedal I use here is for atmosphere: I like a slight wash of pedal, particularly with the B-flat major arpeggio, but too much will obscure the contours of this sequence. The softly-spoken, somewhat uncertain end to this expansive movement has the curious effect of setting the scene for the second movement without actually pre-empting it at all. When it comes, the Andantino seems distant and alien, so utterly different in character from what has gone before.

Such is the spell of your emotional world that it very nearly blinds us to the greatness of your craftsmanship.
Franz Liszt on Franz Peter Schubert

 


Select bibliography

Brendel, Alfred, ‘Schubert’s Last Sonatas’, in Music, Sense and Nonsense: Collected Essays and Lectures (London: The Robson Press, 2015)

Fisk, Charles, Returning Cycles: Contexts for the Interpretation of Schubert’s Impromptus and Last Sonatas (Berkeley, Los Angeles and London: University of California Press, 2015)

Montgomery, David, Franz Schubert’s Music in Performance. Compositional Ideals, Notational Intent, Historical Realities, Pedagogical Foundations (Hillsdale, NY: Pendragon Press, 2003)

Schiff, Andras, ‘Schubert’s Piano Sonatas: thoughts about interpretation and performance’, in Brian Newbould (ed.) Schubert Studies (Aldershot and Burlington: Ashgate, 1998)

A personal journey through Schubert’s penultimate piano sonata (read previous posts here)

With a good deal of reading, of both the score and books and articles on this sonata and Schubert’s piano music in general, and listening, and thinking, by late November 2014, it was time to embark on some serious note learning……

As noted in an earlier post, Schubert’s late piano sonatas are large-scale works: their first movements alone, with exposition repeats intact, can last as long as an entire sonata by Beethoven, and this “heavenly length” can pose problems for the performer in terms of stamina (it takes me around 43 minutes to perform the D959 in its entirety, with repeats), retaining a clear sense of the cyclic elements which recur throughout the work, and appreciating the overall narrative of the work. From my reading of the score, and other materials, I had concluded that the second movement, the infamous Andantino was the most “difficult” (though this is all relative when considering such a large piece of music!). This is the movement which provokes the most discussion and theorising amongst pianists, musicologists, critics and audience members, many of whom believe this movement is the clearest indication we have of Schubert’s emotional and mental instability, probably due to his advanced syphilis. Musically, it feels like an aberration in the overall scheme of the D959, which is generally warm-hearted and nostalgic in its character and prevailing moods, and it is unlike anything else Schubert wrote. “Its modernity is incredible even today” (Andras Schiff, Schubert Studies). It has a “desolate grace behind which madness lies” (Alfred Brendel), the lyricism of the outer sections providing a dramatic foil to the savage intensity of the middle section. Its position in the overall structure of the sonata creates a striking contrast between the majesty and expansiveness of the opening movement and the quirky, playful Scherzo which follows it. In my own practical approach to this movement, I decided to ignore much of the psychobabble and work with what is given in the text.

The movement is in straightforward ABA (ternary) form, the A section reprised with a more intricate left-hand accompaniment and a haunting triplet figure in the treble.

The middle, B, section unfolds initially like a Baroque fantasia (bars 73-86), with descending diminished seventh arpeggios which take the music into C-minor. Gradually all the elements speed up (Schubert indicates this through note sub-divisions, striking modulations and volume of sound) and the music continues to build with increasing savagery via extreme registers and the use of trills to sustain tension, eventually arriving at C-sharp minor and culminating in dramatic fortissimo chords (bars 120/121). A short recitative-like section follows, interrupted by dramatic chords, before a serene passage reminiscent of the G-flat major Impromptu (D899/3). The A material returns at bar 159.

The opening A section combines a barcarolle bass line with a right-hand melody redolent of several of the Heine songs and ‘Der Leiermann’ from Winterreise, while its expressive qualities and character relate to the song ‘Pilgerweise’, also in F-sharp minor. Some pianists like to treat this movement as a barcarolle with a storm in the lagoon (the middle section). Daniel Barenboim has called A section “a melancholy folksong”, a description which I particularly like: the lilting style of a folksong is implied by the recurring bass figure and the simple melody from which is unleashed the turbulent and chaotic middle section.

A rather chilly, tread-like quality in the bass is created through the use of staccato on the first note and the slur on the second and third notes, with the third note lighter (although this is not indicated specifically after bar 2, we can safely assume that this is what Schubert intended throughout). I found it helpful to think of this in terms of a cello or bass pizzicato figure: it needs resonance but should also be balanced with the right-hand melody. I don’t sustain the staccato note with the pedal here, and indeed the pedal is used sparingly throughout this section. The repeated use of falling seconds and a limited range, together with largely understated dynamics, create a feeling of stasis and melancholy contemplation. With so many repetitive elements in this section, it is necessary to create contrasting musical colour (for example between bars 1-8 and 9-12). At bar 19 the music moves into A Major, one of those magical Schubert moments where the mood seems at once warmer and yet even more poignant because it is expressed pianissimo. I like to use the una corda pedal for this pianissimo passage, and the corresponding passage at bars 51-54 to create a sense of other-worldliness.

getfileattachment
Page 3 of the Andantino with my annotations
Other details worth noting throughout this section (bars 1-32) are the inner voices in bars 7, 15, 16, 25, 29 and 31 (and then at bars 39 and 57), and the ornaments which should be played on the beat (though many celebrity pianists prefer to do otherwise, admittedly to beautiful effect). For example, in bar 15, the A sounds with the E sharp on the beat and the turn at bar 23 begins on the note above, but need not be pedantically with the bass C sharp. (See David Montgomery Franz Schubert’s Music in Performance for more on ornaments.)

At this point I feel it’s important to mention the overall tempo of the movement. It is marked Andantino, and needs a sense of forward propulsion. Despite this, some pianists tend more towards Adagio, and at this speed there is, in my opinion, a very fine line between the music sounding meditative or funereal, or even boring, which I do not feel is appropriate. I have aimed towards a metronome mark of quaver = c90 bpm. A quick browse through Spotify reveals quite a broad range of tempi, with some versions of the Andantino coming in at well over 8 minutes (Schnabel, Pollini, Uchida, Perahia) and others at around or under 7 minutes (Lupu, Schiff, Goode).

Here is Uchida

And Lupu

 

And so on to the B section, which leads to the most passionately and extraordinary part of the movement and indeed the whole sonata. It is typical of Schubert to create sections in the music which are vividly contrasting yet also complementary: the A sections are reflective in their lyrical subject while the middle section completely destroys this frame of reference, only for it to return at the reprise of the A section. It is the strong contrast between the A and B sections which makes this movement so arresting and so powerful.

The bridging passage begins at bar 69, and is preceded by three bars whose dark, descending chords mirror in their reverse movement the chords which form the opening sentence of the sonata (and a figure which recurs elsewhere). I like to create a sense of mystery in bars 69-72 with a wetter pedal effect and a little rubato to suggest improvisation as the music unfolds. The main difficulty I encountered in the entire B section was maintaining a sense of the underlying 3-in-a-bar pulse and clarifying the different note hierarchies, while also continuing the improvisatory/fantasia feel. In order to achieve this, I drilled the section strictly with the metronome for several weeks, a tedious but necessary task for once the note hierarchies and subdivisions were well learnt, I could let the music break free, particularly in bars 102-122, to create a rising sense of hysteria. 

A clear sense of pulse is required through bars 124-146, as the recitative section takes over. After all the “busyness” of the previous page, I like to create a sense of the music being restrained once again, with the contrasting disruption of the FFz chords. At bar 147 the music arrives in C-sharp major in a passage which seems directly drawn from the G-flat Impromptu. At bar 159 the A section returns, this time with the more elaborate LH figure and the triplet figure in the RH, which should have the quality of a separate, “other” voice. Throughout this section, it is important to retain a sense of the opening melody and a similar lightness in the LH to that in the opening bass figure (note the demi-semiquaver rests at the end of each bar). Bars 177-182 the RH accented E’s sound as tolling bells above the melody, and once again I like to use the una corda pedal here to give a more ethereal quality and to create contrast with the forte chords in bars 185/66 and the descending figure in bar 187. The movement closes with dark, arpeggiated chords which echo those at bars 65-68, and which are transformed into sparklingly joyful spread chords in the Scherzo which follows. I try to keep these in tempo until bars 198/9 at which I introduce a rit. to signal the close of the movement. The dynamic landscape here is very quiet and muted, and I feel una corda is perfectly acceptable in these closing statements.

It took me three months of fairly consistent work to bring the movement to a point where I felt confident enough to perform it for others (for friends at home). I then “rested” it for some weeks while I turned my attention to the first movement, the subject of the next article.

It was Schumann who coined the phrase “heavenly length”, specifically in relation to Schubert’s Great C major Symphony, D944, though this tag has now become synonymous with all of Schubert’s late music.

How to approach a work of the scale of the Sonata in A, D959? In common with the other late sonatas (including the Sonata in G, D894), it is a big work, with a first movement which can last as long as an entire mid-period Beethoven Sonata, if the exposition repeat is included (and I believe it should be).

The D959 is indeed long: the entire work takes around 40 minutes to play, and for both performer and audience there is a sense of traversing an epic landscape. One’s duty as performer/interpreter is to find connections, within the individual movements, and the work as a whole, in order to lead the listener on a unique journey deep into Schubert’s musical landscape. Schubert uses motivic and structural signposts throughout the four movements to enhance this sense of a journey (for example, the opening measures of the first movement are reprised in the closing bars of the finale, and there are many other cyclic elements – of rhythm, melody, articulation, and even character/emotion – which connect the four movements ). The cyclic elements also enhance the sense of a tightly organised structure: this sonata may be long, but it is not rambling (though some performers may make it so!). The first and the final movements are almost identical in length (c.13 minutes each if one observes the exposition repeat in the first movement). These edifices bookend the middle movements which are also of an equal length – c.7 minutes each (roughly half the length of the first and final movements) if one observes all the repeats in the Scherzo. Some scholars have pointed to a “golden ratio” or “golden section” in Schubert’s structural organisation. Whether this was conscious on the part of the composer is not known, but from the performer/interpreter’s point of view an appreciation of this almost perfectly-balanced structure is important in creating the sense of a distinct structural and narrative arc that runs through the entire work.

In terms of embarking on the learning of all these notes, I decided to treat the sonata as four separate works, learning each movement as a stand-alone piece while also remaining alert to the cyclic elements within each movement and the whole work through regular study of the score away from the piano and listening to recordings of the complete sonata. From the outset, I had a clear timescale in mind: I wanted to have the entire sonata in the fingers (learnt but by no means finessed) by summer 2015, and having a clear focus enabled me to meet my target, almost to the day. What has interested me throughout the learning process is the fact that the notes themselves are not that difficult, and most of the time the writing lies comfortably under the fingers and hand. What has been far more difficult is achieving a convincing rendering of Schubert’s unique compositional voice – and maintaining this throughout c40 minutes of music. Too many interpreters treat Schubert as Beethoven’s “enfeebled twin”, but as the pianist Paul Lewis has noted “Schubert is more internally stormy than Beethoven, which perhaps make it all the more powerful…….. Schubert almost never provides the answers. There are always more questions than answers…….it’s a reflection of what we find in life“.

Schubert interpretation is riven with some doubtful traditions which developed in the nineteenth century, many of which have been subsumed into standard performance practice today. The best commentary on the music can be drawn from the autograph scores and contemporary evidence.

Schubert should never be stereotyped or over-interpreted: such treatment can straitjacket the music, reducing it to a simplistic highlighting of elements which the performer may feel are obviously “Schubertian” (beautiful melodies, nostalgia, poignancy, depression, emotional volte-faces, rage, joy etc) rather than allowing oneself the freedom to appreciate that this music is so much greater than the sum of its parts.

Here is Inon Barnatan in the final movement of the D959, a reading I particularly like for his clarity and appreciation of the articulation, coupled with a real sense of the joy inherent in this music

Not a “top ten”, merely a selection of performances nominated by readers of this blog and followers and friends on Twitter and Facebook. Please feel free to leave further suggestions or your favourite recording in the comments section of this post.

1. Sviatoslav Richter

2. Clara Haskil

3. Paul Badura-Skoda

4. Claudio Arrau

5. Grigory Sokolov

6. Mitsuko Uchida

7. Daniil Trifonov

8. Clifford Curzon

9. Wilhelm Kempff

10. Maria Yudina

Daniel Barenboim, musical polymath, is in town for a four-concert Schubert Project residency in which he will traverse all 11 of Schubert’s completed piano sonatas. Prior to the first concert, he unveiled a brand new piano – one with his name on it. The Barenboim piano was conceived and developed in a collaboration between Barenboim and Belgian piano maker Chris Maene, with the cooperation of Steinway. What makes this piano different from the modern concert grand is that it is straight strung, and Barenboim used a Liszt piano as the inspiration for his eponymous instrument. It is said to offer a greater variety of colour, transparency and clarity across its range. Audiences at Barenboim’s Schubert recitals will have the opportunity to hear for themselves this new piano in action.

Unsurprisingly, it was a full house at the Royal Festival Hall and there was a distinct buzz of anticipation and reverence ahead of the start of the concert. Sitting in the rear stalls didn’t really offer myself and my concert companions a chance to examine the piano in detail. The piano remained firmly closed, lid down, until a few moments before Barenboim took to the stage, and was closed up again during the interval.

The jury is still out on whether the Barenboim piano was noticeably different to a modern Steinway, and any clarity and crispness of articulation, or nuanced dynamics are surely the result of the pianist’s technique, not the piano: one would expect an artist of Barenboim’s calibre to make even the most beat-up church hall piano sound lovely.

The theme of the first of Barenboim’s Schubert concerts was the key of A, as he presented three sonatas from different periods in the composer’s life. The D537, in A minor, was composed by Schubert when he was 20 and is the earliest surviving completed piano sonata, though it was not published until 1852 as the Op. post. 164. It begins with a dotted motif followed by filigree semiquaver broken chords. It’s emotionally charged and already demonstrates Schubert’s skill in unexpected harmonic shifts which colour the music. The middle movement, in warm E major, is genial and nostalgic, with a theme that would be heard later in the concert (Schubert “exported” it as the Rondo theme of the final movement of the D959). Yet, typically of Schubert, the mood shifts during the trio, a chilly march in A minor. The finale has a Beethovenian cast, with a dash of Haydn’s wit, yet already full of Schubert’s trademark unexpected harmonic shifts and emotional volte-faces.

I think many of us were trying to hear whether the piano really sounded that different instead of concentrating on the music, but the opening Sonata was presented with energy, though not always entirely convincingly, and keen sense of Schubert’s tonal palette, especially in the final movement. The middle movement, whose theme was reprised later in the D959, began genially enough, but the middle section had an ominous tread, for which the bass notes of the piano were suitably rich and dark.

The first A major Sonata of the evening is known as “the little A major” and was the most genial of the three sonatas presented in the concert.  Barenboim created a sense of intimacy in the first movement, but again one had the sense he wasn’t entirely convinced by it himself. It continued into the ethereal slow movement, whose pianissimos were, at times, barely a whisper. The finale was lyrical and good-natured, the opening theme played with a songful elegance, though I felt he pushed the tempo a little too much for my taste so that some of the lyricism was lost.

After the interval was “my’ Sonata, the penultimate of Schubert’s piano sonatas, the D959 in A major, which I have (perhaps recklessly) set myself the task of learning. I was extremely curious to hear Barenboim’s take on this big work, not least whether he could carry the narrative of the first movement right through to the closing sentence of the finale. My difficulty with hearing other people’s versions of this sonata is that they often conflict with my own, which can make me the most pedantic of listeners. I spend a lot of time with this Sonata. To say I eat, drink and breathe it might be excessive, but I often find myself waking in the night and playing it through it my head. The opening statement, a chorale-like sentence, lacked real nobility and drive and the propulsion towards the suspension at the end of the passage was lost in some curious pulling about of the tempo. There were one or two rocky moments as some of the triplet figures were lost – and this issue reappeared in the finale, where perhaps Barenboim was tiring (this is a big work – lasting around 40 minutes, even without the exposition repeat in the first movement), and overall I felt the movement lacked power and drive.

The slow movement, about much has been said, written and surmised, a melancholy folksong with a storm at its centre, lacked cohesion and there were some serious memory issues towards the end. The movement seemed relentless rather than revelatory. However, the scherzo was bright and crisp, with some sensitive highlighting of the melodic line in the trio section. The finale seemed rushed, the triplets often losing clarity, but the sections in the coda where the music stalls, as if to take a long breath, to reflect on what has gone before, were perfectly paced and the closing statement, a recapitulation of the opening sentence from the first movement, was emphatic. The standing ovation which followed was as much for Barenboim the man, the demigod, as for the performance and the new piano.