Flexibility & Drama (exposition only)

This applies the same method used on the previous page to the exposition.

Procedure. As before, each performance was divided up into sections, twelve in this case. Where before we were comparing 90 second segments, here they are 18 seconds.

Why does the segment-size matter? Theoretically, there's no difference between 'playing the second theme slower' and 'playing the second beat slower'; but if this were true in practice, we would not have a special term, 'rubato', specifically to describe small-scale fluctuation. These pages go in the opposite direction, comparing the variation between 8-16 bar 'themes' and 60-80 bar 'sections'. (As a guide, the movement as a whole falls into five roughly equal sections: Exposition, Development I, Development II, Recapitulation, Coda.)

How does this compare to the other chart? The central 'cloud' shares two important features. First, a gentle downward slope. Second, a definite (if irregular) lower limit.

But a very important difference is that all performances in general treat the exposition in a more 'dramatic' fashion, and pre-war performances most of all. (In statistical terms, the 'average flexibility' runs generally higher here.) Note on the math: because we used 12 points in the exposition and 10 points overall, we would expect the 'average flexibility' (middle of the 'cloud') to be 1.2 times greater here; but obviously it's far more than that.

Could this be random noise? To attempt to answer this question, we now combine the two charts, to see if any patterns emerge.