Mozart's surviving sketches fall into two distinct groups: true sketches, in which he drafted ideas or worked out some sort of compositional problem, and fragments. The latter are usually pieces that he fully intended to compose and either abandoned or set aside, to be finished at some later date. His usual method of composing was, after preliminary sketching, to write out the main melodic line of the work in what was eventually to be the score, sometimes also filling in details of texture or harmony or subsidiary parts when they provided the continuity; he would then go back and finish the job, a task we know he regarded as relatively routine.
Most of the 30-odd pieces on this fascinating pair of discs are ones he left unfinished, some after just a few bars, others with entire movements complete, but most with about half a movement written at least in outline. Each one is a separate case, of course; but speaking very generally, I would say that most of these are movements he decided not to complete, usually for some internal reason though perhaps sometimes because the occasion for the particular piece had passed. I would guess that the longest single movement here, the solemnly beautiful Adagio, K580a, usually read as for english horn and strings (because that seemed to explain the transposition) but now understood to be for clarinet and three basset-horns, might have been completed had Mozart lived longer (the whole melodic line exists but the accompaniment stops halfway through); others, like the delectable piece for clarinet, basset-horn and string trio, K580b, were clearly abandoned because Mozart realized how awkward the medium was to write for in his kind of style. Some, for example the B flat clarinet quintet movement, are beyond explanation: a truly beautiful piece, of which Mozart wrote the exposition, then three more bars, and just stopped. He must have perceived something, some inherent flaw, that escapes more ordinary people. Or perhaps, given more time, he would have in fact pursued it. (It is worth reflecting that, had he died a year earlier, the B flat Piano Concerto, K595 would have come down to us only as a fragment; he had begun it some years back, but left it unfinished until the end of 1790, when evidently he needed it.)
Most of these pieces are also recorded in the Philips Complete Mozart Edition. There, generally speaking, they are completed, by Erik Smith or some other sensitive Mozartian. Here, with the exception of certain pieces where Maximilian Stadler's early completion cannot readily be distinguished from Mozart's own work, we are given what Mozart wrote and no more. Where he left the textures skeletal, so do we hear them; where he broke off, we have silence. From the musicologist's point of view, all this is splendid, although I am not sure that musicologists (not anyway a substantial proportion of the record-buying public) need much more than the printed score anyway, and that is readily available in the New Mozart Edition. The result is a shade austere. Often, as in the 'english horn' Adagio, we have long stretches of unaccompanied clarinet, or in some of the chamber works lengthy passages of solo violin. Often it would be easy to reconstruct the accompaniment, but of course that would involve a contribution that wasn't Mozart's, and no one could be certain that they would make the same compositional decisions as him (indeed the one certainty is that no one would make precisely the same ones). But these bare textures make, at best, enigmatic listening. In the Rondo, K581a, probably originally intended as the finale of the Clarinet Quintet, there are six bars of lonely semibreves for the clarinet, obviously jotted down by Mozart to remind himself of the harmonic scheme that the other parts would fill out: heard alone they are absurd. But the chaste purpose of the recording remains.
The works not in the Philips set should perhaps be briefly described. Most are very short. There is a puzzling part-piece for cello and piano (a medium almost unknown at this date) of some two minutes. There are two fragments for two pianos, presumably designed for a B flat Sonata that never came to fruition. All those are generally thought to date from about 1782. There are two keyboard and violin pieces, each lasting about a minute, one from the early-middle Viennese years, the other much later. I have already referred to the Rondo that didn't get into the Clarinet Quintet (though some of it did find a place in Cosi fan tutte); there is also the opening of an Allegro for two clarinets and three basset-horns, a piece that has always intrigued me as I can't imagine what kind of work Mozart was contemplating (perhaps he couldn't, either). Lastly there are two attempts at a first movement of a String Quintet in E flat, one probably from about 1785, the other, which is very brief, apparently belonging to the time of his last String Quintet, K614.
The performances are very adequate for the purpose; there is some particularly sensitive clarinet and violin playing. It does however seem to me a pity, when producing a set partly directed at the specialist and the historically-minded listener, not to use period instruments. The two-piano and duet music, and that for keyboard and violin, would have gained considerably. Still, I did find the set intriguing to listen to and am sure that anyone who loves Mozart would do so too if he or she does not already know some of these pieces; it may in any case be instructive to hear these 'purist' versions alongside the completed, or more nearly completed, ones, though I should add that there is only some 12 minutes of music here that is not available in some form in the Philips set.'