Anyone who knows me, knows that I'm fanatical about two things in the performance of Austro-German classical music, the question of repeats and of antiphonal placement of the violins. Both remain vexed questions, despite the fact that the historically informed performance (HIP) movement makes it clear that repeats ought to be observed and that the first and second violin desks ought to be divide antiphonally, that is, firsts on the conductor's left, seconds on the right.
Strangely, despite a half century or more of these insights, both precepts are honoured more in the breach than in the observance, especially in the non-German speaking world, and most certainly in the Anglophone and Francophone musical spheres. It's only recently (that is, in the last ten or fifteen years) that the Vienna Philharmonic, bastion of tradition if ever there was, has returned to antiphonal violin placement as the norm. This was the norm from the time of Haydn through to Richard Strauss, and is what composers would have expected.
At some point after the Second World War, the norm on both sides of the Atlantic became to lump the violins on the conductor's left. In one variation, cellos were out to the right, and violas tucked in behind then; in another version, the somewhat less barbarous of the two, the reverse. There are all sorts of theories as to why this happened, and no one (including conductors themselves as well as scholars of music) has yet furnished a fully satisfying explanation. The most plausible account seems to be that with the increased premium place on accuracy, conductors felt that the first and seconds, who often, though, crucially, not always, play in unions in tuttis, could hear each other better and therefore play together more effectively. Certainly, this helps explain why martinet conductors who insisted on precision at all costs, such as Herbert von Karajan, preferred this to the pre-war seating arrangement. The next generation, such as Claudio Abbado, continued this tradition. Only in recent times has it been realized, and here the HIP movement deserves much credit, that this does a great disservice to the composer's intentions. Listen to just about any movement of a Mozart or Haydn symphony, or a passage from a Mozart opera, or a Beethoven symphony, and you'll hear what I mean.
I've spoken to many distinguished conductors about the orchestral placement question. Those who do it as a matter of course obviously agree with me (and the scholarship). Some take.a pragmatic view. A few told me that they suggest antiphonal violin placement, but, if they are guest conducting and an orchestra prefers the alternative, they will defer to the orchestra's choice. Others point to acoustical problems in certain venues, which make it difficult to hear the seconds (which play inward, rather than out to the audience), if violins are divided. Likewise, some orchestra musicians have told me that the firsts and seconds prefer to sit together, to better hear each other.
It's clear that these are pragmatic, rather than artistic, considerations. Most I spoke to agreed that the optimal situation be antiphonally placed violins.
There are some related questions I won't dwell on, such as where to place the cellos and violas. Typically, in Berlin or Vienna, if violins are divided, cellos are placed centre left, and violas tucked in behind the seconds on the right. In Vienna, at the Musikverein, the Philharmonic places the double basses lined up against the back below the organ. In Salzburg, they place the basses to the back left, behind the cellos and the firsts. Some conductors, such as Italian maestro Antonello Manacorda, prefer to divide cellos from basses, thus placing the basses on the far right. This is extremely effective -- listen to his recent recordings of Beethoven symphonies for evidence.
Now to repeats. A traditional view -- espoused, for instance, by the late Sergiu Celibidache -- is that repeat marks in Classical and early Romantic repertoire were a Baroque hangover, a convention that can safely be ignored, because the composer didn't expect it to be followed. A variant of this is that repeats are only useful for new and unfamiliar works. We hear the Mozart symphonies so often, on this view, that who needs a repeat?
This rather cavalier view has been challenged by the historically informed performance school, who argue that composers knew exactly what they were doing, and, if a repeat mark was placed in a score, the composer intended it to be observed. As the late Georg Tintner once told me, not observing a repeat is akin to making a cut, something he was very much against.
Clear evidence that composers intended a repeat is in those cases where there is a lead back to the repeat. No one would have written those extra bars of lead back if the intention was to skip the repeat, if it were there merely as a convention.
It's now largely agreed that omitting exposition repeats is an egregious error, but it still does happen. How often do you hear the Beethoven Seventh, for example, with the exposition repeat in the finale omitted? Far too often. However, when it comes to second half repeats --which are common in Haydn and in Mozart -- many conductors today make the strange decision to observe the exposition (or first half) repeat but omit the second half repeat -- such as, for example, in the outer movements of the "Prague" symphony, or in the second movement of the G minor symphony K. 550, or in the finale of all three of the last symphonies.
Now, this is peculiar indeed, as it distorts the structure of the movement, to observe the first repeat and to omit the second. A lesser evil, which is what was done by an older generation of conductor, is perhaps to omit both repeats altogether. Far from ideal, but at least the symmetry of the structure is preserved.
On the repeats questions, most conductors I've spoken to agree that in principle they should be taken, and some follow this approach. Others cite pragmatic reasons for omitting repeats. Some examples include the fact that playing all the repeats might make a concert too long, or require that musicians be paid overtime, for example. It's sometimes said that orchestral musicians don't like to play repeats, so conductors omit them so as not to antagonize orchestra members. Playing all repeats in a long program might be fatiguing for musicians and audience alike. And so on and so forth.
Again, it's pretty clear that all of these are pragmatic rather than musical reasons. In an ideal world, there is no good reason to omit repeats, nor to lump violins together. The former simply pays disrespect to the composer's intentions. The latter leads to a thick, occluded orchestral sound, robbing us of the greater transparency and the clear delineation of the antiphonal interplay between firsts and seconds that the composer intended us to hear.
To rest my case, here's a recording of the second movement of the Mozart K. 550, where the composer marks both halves for repeat, and, of course, expects antiphonally placed violins. Observing the repeats gives the movement its necessary breadth and importance in the work. With both repeats observed, the movement clocks in at 15 minutes or more -- about the length of a slow movement in a Bruckner or Mahler symphony -- providing the necessary repose (albeit, a troubled repose) before we turn to the minuet and finale, which drag us down into the looming abyss. Divided violins bring the necessary transparency. This is a Deutsche Grammophon recording with the Vienna Philharmonic, conducted by James Levine, and recorded at the Musikverein under ideal conditions.
Listen for yourself.