Double play

Comparing composers' first and final thoughts is always fascinating. But for anyone who knows the definitive Sibelius Five, this original 1915 score will be a sequence of surprises, the kind that stir up important questions about the nature of the creative act.

It's astonishing to think that Sibelius was (or seemed to be) happy to let the Fifth be heard in this first version. So many of the work's great imaginative touches seem half-formed - an idea only half-glimpsed through thick mist - or, more devastatingly, they simply aren't there at all. Take the symphony's opening: that quiet but potent horn motif just has to be the germ, the seed, from which everything is to grow. And yet in 1915 it wasn't there: the symphony began with the woodwinds' "answer". You can argue that this was an even more radical conception - the moment of fertilisation happens in the dark, beyond hearing. But if you know the 1919 score, it's hard not to feel the loss.

There are more big surprises to come. What about that breath-taking moment of breakthrough, when the Scherzo emerges from the first movement like an impossibly speeded-up film of a mountain rising from the sea? Again, it isn't there: instead the music stops, just as something is beginning to change - there's a pause, then the Scherzo picks up where the first movement left off. It's intriguing, but it's a long way from the elemental thrill of 1919. And as for the finale... No, I won't spoil the fun any more; enough to say that the shocks continue right to the end.

Differences between the familiar and the 1892 En Saga are not so fundamental, but textures, voices, even whole passages, can be quite different, and the work feels less coherent; Sibelius clearly had things to learn about musical narrative too. There's little point in pretending that either score is a valid alternative to the version we know and (I'd say, rightly) love, but with performances and recordings of this quality, it's the perfect chance for anyone who cares about creative processes to probe more deeply. An important release - though, for me, it only makes Sibelius's achievements in those final scores all the more astonishing. SJ

The first thing you anticipate are those far-away horns. But they're not there. It's amazing what a shock to the system that is: a sound so much a part of symphonic literature simply wiped from the page. But in the beginning, there were no horns; Sibelius hadn't heard them yet. We're working backwards here from the final version that we know and love to this, the first-born Fifth, a masterpiece in progress. And between the two scores are privileged insights into the workings of a great and unpredictable musical mind.

This 1915 original is music of the elements in search of form: Sibelius, the explorer, the visionary, seeking and finding, feeling and experiencing, but not yet fully reconciled with Sibelius, the symphonist. Sounds and sensations go straight down on to the page. That's what makes these first versions so fascinating: the basic elements are all in place, but roughly assembled, more discursive, in many ways more extraordinary, but not better, than their later counterparts. So the original Fifth is considerably longer than the final revision, the radical re-fit. There is an extra movement: the ingenious fusion of the first and second (by way of that long, exhilarating accelerando from "sunrise" to the explosive, dancing coda) is clearly implied but not yet realised. Separated, the two movements make for a daring musical juxtaposition, the one entirely dependent on the other. Together, they make sense.

As for that extraordinary finale - add 200 extra bars. You might even be sorry to see a few of them discarded. But the memorable woodwind theme we know as the soaring counterpoint to Sibelius's tolling horns doesn't emerge until halfway through. Indeed, it is never heard in counterpoint. So there's one inspired development of the later version. Another is the dramatic pay-off: those notorious spaced chords of the final bars - the triumph of time over timelessness. So, in most respects, the original will have presented even more of a challenge to its first listeners.

Likewise, the original En Saga - wild and wilful, rough and primitive, indisciplined but intriguing, and - even more so than the symphony - a quite different (and much finer) piece in its final form. We should be grateful to BIS for these first recordings. Important as documents, but quality goods in their own right. Listening to either work will never be quite the same again. ES