To finish or not to finish?
Stephen Pettitt Here's your starter for ten. What's the most famous unfinished piece of classical music in the world? Schubert's 'Unfinished' Symphony, his Symphony No. 8, of course, which is usually played as a twomovement torso, bereft of the Scherzo and finale which a symphony of its provenance would normally include. Usually, but not always. The latest man to attempt to fill in the missing parts is the Russian composer Anton Safronov, whose version the Orchestra of the Age of Enlightenment is due to play for the first time on 6 November at the Royal Festival Hall in London under the baton of Vladimir Jurowski, with performances both in the main evening concert at 7.30 p.m. and in a 'Night Shift' event beginning at 10 p.m.
From time to time all of us leave undone those things which we ought to have done. Composers are not exempt, but sometimes it's not really their fault. Given the choice, Mozart, I am sure, would much rather have got to the final bar of his Requiem than give the legend-writers fodder for exercising their creative talents through his premature demise. Likewise Puccini and Berg would have preferred not to have left the completion of their final operas, Turandot and Lulu, to others. On other occasions a piece's progress might be halted because of the necessity of dealing with that other certainty of life, tax. It would be showing unsound business sense to continue writing a piece that's not been commissioned if a nice fat commission — with, however, an early deadline — drops on to your doormat. That seems to be why Schubert didn't finish his Seventh Symphony. The composer laid it aside in order to concentrate on his opera Alfonso und Estrella, written for Weimar, and simply never got round to resuming work. And sometimes, of course, a piece is discontinued just because it doesn't seem to be working out.
But what to do with all of these incomplete pieces? Music is an architecture of time, and therefore, surely, if a piece isn't complete it cannot work, just as a building without one of its external walls cannot work. So we must attempt to fill in the missing bits. But then there's the opposing argument. How can we presume to be able to second-guess a great composer? Would we dare to complete an unfinished canvas by Titian or Picasso? And in any case hasn't the popularity of, for instance, Schubert's Unfinished Symphony in its truncated state proved that an incomplete structure can at least sometimes work?
Frankly, I'm not sure which side of this debate I am on. It rather depends on the work. As far as Schubert's Unfinished Symphony is concerned, I'd suggest that we do not hear the symphony at all but instead, perhaps, a sentimental, regretful comment on its unfinished state. With Bach's The Art of Fugue, however, I feel completely differently. It's touching when that last great fugue tails off to nothing, but in some strange way the mind provides its own resolution (and not in imagined sounds). Equally, other men's completions have proved satisfying.
And what about that Requiem mostly by Mozart? Constanze made sure that it was finished by someone, choosing Joseph von Eybler and Franz Siissmayr for the job, and with contributions also, maybe, from Maximilian Stadler. But a number of present-day scholars — Franz Beyer, Duncan Druce, Richard Maunder, the Haydn expert H.C. Robbins Landon, and Robert Levin — have presumed to do better, to get closer to the spirit and substance of Mozartian than Siissmayr and company did, reorchestrating this, discarding that, rewriting the other. I have a problem with this. While Siissmayr's orchestration might not be on the same level as Mozart's, he at least had the advantage of living and working alongside the great man. That gives him a head start over his late-20th-century colleagues, which in my view he maintains.
Yet living in a different epoch does not necessarily mean disqualification from filling in the missing parts if nobody has done it before and if you are honest about what you are doing. Two outstanding examples are Deryck Cooke's magnificent, utterly convincing realisation of Mahler's Tenth Symphony, a triumph of insight and deduction, and, more recently, the composer Anthony Payne's similarly brilliant work on the sketches of Elgar's Third Symphony. Enough of both works survives to tell us that both were looking to break new ground for their composers. Even so, there's no pretending by Cooke and Payne that their work represents what Mahler and Elgar would have done. The Mahler is described as 'a performing version of the draft', the Elgar as 'an elaboration on the sketches'.
Luciano Berio's completion of Schubert's extremely fragmentary Tenth Symphony goes a stage further, at least partly because it has to. Rather than ape Schubert's language, Berio uses ideas from the sketches and from other Schubert in a work whose structure, manner and language is pure Berio, drawing together original Schubert with his own 'connective tissue'. It is his analysis of and comment on Schubert rather than a speculative reconstruction. Hence its title, Rendering. It's honest, affectionate and creative in a way that completions 'in the style of' cannot be.
Schubert's Seventh Symphony is a different matter, since its outlines are complete, though it's mostly unorchestrated and there are passages with just melodies or just bass lines with the odd bit of counterpoint added. We know basically what the music was supposed to do, where it was supposed to go. But we don't know the colours Schubert envisaged, and neither do we know how much or how little Schubert might have changed things had he advanced to the 'filling out' stage. Another countermelody here? No countermelody there? What about a different harmonisation second time round? Decisions, decisions for those who attempt to finish it, among whom have been the conductor-composer Felix Weingartner in 1934 and the Schubert scholar Brian Newbould — an inveterate finisher of unfinished Schubert symphonies — in 1981.
And what about the Unfinished? There is, in fact, a third movement for the work, a Scherzo (with an incomplete trio) of which Schubert orchestrated only a few bars. So half of the restoration process is largely a matter of orchestration, a task attempted by many. And it turns out that there's a candidate movement for fulfilling the role of finale, the entr'acte from Schubert's incidental music for the playRosamunde. It happens to be in the right key — B minor, an unusual key for this composer — and shares the same instrumentation as the movements Schubert completed. Even more helpfully it's also in sonata form, and its mood fits well. The obvious possibility is that this movement began life as the symphony's finale but when Schubert abandoned the work he recycled it, perhaps modifying it on the way. Both Newbould and Gerald Abraham assumed this to be the case in their completions. But again the question must be asked: is conjectural restoration better or worse than no restoration at all?
As it happens Safronov has chosen to ignore the Rosamunde theory, instead composing a completely new finale based on a number of Schubert keyboard works. He's commendably honest about his intentions. It is, he says, simply 'an attempt to move into the mind of a 19th-century composer. It is an experiment.' Not, note well, an attempt to recreate that which cannot be recreated. We shall see just how well or otherwise it props up, fits in with or thrillingly contrasts with Schubert's broken architecture.
Previous page
