#48: Complexity in transition
Section 6: Antonin Dvořák – String Quartet Op.96, ii: Lento
The end of the Lento’s A’ section (b.16-19) presented an especially complex moment of ensemble interaction. We initially grappled with the Czech Quartet’s timing here by breaking it down into four independent strands, with each player establishing their counterpart’s precise rhythmic ‘modifications’, with help from Sonic Visualiser (and other reference points, such as a metronome). We were able to determine, for instance, that the violist plays the first three notes in b.17 extremely late, the final note early, and the next bar more simply. But this method quickly encountered hard limits, for in performance the relationships are so precisely balanced that such explicit diagnoses of timing variation, when ‘put into practice’ all at once, were simply incapable of capturing either the fine detail or — more importantly — the delicate character of the original.
We therefore explored other ‘intermediate’ tools, including treating one part as a detailed ‘pivot’, which was based directly on the Czechs’ rendition (rather than on a group of static references). While listening, we imagined a structure whereby one player — here the second violin — would learn to capture the timing very precisely by playing side-by-side with the original recording, such that the timing could become more embedded in physical motion. (There was a coincidental pleasure in determining that second violinist Josef Suk might make an appropriate pivot here, because we had already had the feeling that his playing had a streak of appealingly stubborn conviction!) Then the other parts would try to relate to that pivot more organically, and more responsively. As an approximation of the original, this was more successful; but we still felt that the exponential complexity that arose from all four strands actively interweaving would always evade explicit copying methods. In other words, there was a tension in the internal relationships — that resulted from every part embracing subtle irregularities in timing — which could never be replicated when those patterns were isolated from that context, and then ‘brought back together’.
Clearly, our intermediate, ‘re-presentational’ model was unlikely to have reflected their original conceptualisation. There was no real sense in the original of such a differentiation: of consistency surrounded by inconsistency. As a copying technique, however, this often proved useful, because it gave us ‘hooks’ with which to diagnose and track their conventions — at least to some extent.
