The same rule holds good in the matter of key distribution. When the sonata form was young, the key of its second subject was fixed by an almost unalterable convention: if the movement was in a major mode, it was the dominant, if in a minor mode, it was the relative major. Hence the audiences of Haydn and Mozart always expected the same key system, and were hardly ever disappointed. But Beethoven, from the outset of his career, broke through this traditional arrangement, and so began by surprising his hearers, and ended by making their intelligence co-operate with his own. Take, for instance, the first movement of the Hammerclavier Sonata. The first subject is in B flat, and the transition after modulating to its dominant F, proceeds with a vehement and emphatic assertion of the new key, as though Beethoven intended to revert to the customary usage, which, it must be remembered, he often follows. But the very emphasis makes the hearer suspicious. It is not in Beethoven's manner to underline his keys with so much flourish and ostentation: perhaps, after all, appearances are deceitful, and he is only throwing us off the scent. Then our uncertainty is artfully intensified by an interpolation of the opening theme, which, at this stage of the movement, is the last thing in the world that we expect; and immediately after it comes a modulation to G major, and a presentation of the second subject in that key. The anticipation of this event is an exercise of critical sagacity not dissimilar to that afforded by a novel of Balzac or a play of Shakespear. In the famous scene of Madame Marneffe's confession, we are half-cheated into believing that the woman's repentance is real, though we know that its reality is rendered impossible by all laws of characterisation. When Lear decides between his three daughters, we feel that Cordelia's coldness of manner has raised a false issue which the subsequent development of the drama will correct. In short, the true function of structure, whether it be in literature or in music, is to set before us two competing impulses and bid us reflect upon them.
But it may be urged that a musical composition can only surprise or baffle on the first occasion: after that we remember what is coming, and can foretell the end as readily as the composer himself. This view pays an undeserved compliment to the capacities of human nature. The average listener does not really hear a work of any complexity the first time that it is performed in his presence: he apprehends more or less of it according to the degree of his ability or experience, but there will certainly be effects that escape his notice, and, if the composition be truly organic, those effects will be vital to the appreciation of the whole. Indeed, we have here one of the most obvious tests of a great work. We grow tired of a trivial melody or a shallow fantasia, for it tells us its whole secret at a single hearing: but we may spend our lives over Bach's Fugues or Beethoven's Symphonies without ever hoping to exhaust their limitless reserve. Again, we are not such creatures of pure logic that an effect once produced in us is incapable of repetition. We may know our Shakespear by heart, and yet be moved by the humour of Falstaff and the pathos of Imogen, by the subtle questionings of Hamlet and the frenzied self-accusations of Othello. So in listening to great Music we often allow ourselves to be carried away by the impulse of the moment: we forget that we know what is going to happen, or expect it in a new mood and from a new standpoint. There are many avenues by which the sense of novelty can be approached, and among them not the least important is that of our own imagination. No doubt this influence would be seriously impaired if we were to hear the same passage day after day and hour after hour, but this, of course, we are never called upon to do. With the present range and variety of our musical literature, an effect that is genuinely striking may be weakened by familiarity, but can hardly be ever wholly obliterated.