One of the pleasures in Robert Bernhardt’s performance of the Haydn 104th yesterday was his adoption of correct tempi.  Correct for Haydn, that is.  For instance, the slow introduction should not be that slow, it should be only about twice as slow as the allegro.  How do we know? Most obviously, the metrical relationship becomes clear, nothing is forced, and there are even audible motivic connections between the intro and the allegro with a tempo that moves along.  Less obviously, but equally important, there needs to be a cognizance of the very simple harmonies, which do not invite solemn contemplation.  Harmonic rhythm is the rate at which chords change. Haydn’s harmonic rhythm in the intro is reasonably slow, hence a none too dragging alla breve is indicated.  There is a similar situation in the Brahms First Symphony, by the way.  When Brahms writes “Un poco sostenuto” what does he mean? Is “A little bit sustained” a tempo indication? No.  It is in fact a direction for articulation.  Then what’s the tempo? You find the tempo by leafing a few pages ahead to the allegro, which is absolutely unproblematical.  The allegro is an easy tempo to find.  It’s comfortably in an non-subdivided two.  Subdivide each beat, and, voila! You have the tempo for the introduction.  The harmonic rhythm is faster in Brahms, therefore the whole complex of tempi should be slower than in the Haydn case. 

One of the real, honest to goodness difficulties with Schoenberg’s style is the combination of complex chords with fast harmonic rhythm.  Works like the First String Quartet and the First Chamber Symphony exemplify this.  I advise the sort of person who reflexively associates the difficulties with Schoenberg’s style exclusively with atonality to consider the harmonic rhythm.  The first quartet is a tougher nut to crack than the atonal but neoclassic third quartet.  Schoenberg’s tempos are often counterintuitive.

Which brings me to today’s topic.  Leaving aside technical stuff as discussed above, do composers have their own allegros, moderatos, adagios, etc.? Yes.  Consider Schubert’s “moderatos”.  These are slow.  I think often slower than Schubert’s andantes.  If this seems crazy, just look at the scores, especially of the piano sonatas.  Piano sonatas allow for quite a bit of leeway, because you don’t have to accommodate the co-ordination of an ensemble.  That’s why there is so much radical divergence in Beethoven’s piano sonatas as opposed to his symphonies.  By the way, I honestly believe that some very well known conductors adopt inappropriate tempi in Beethoven’s symphonies just to “make their mark.”  I don’t care much about the metronome debate in Beethoven, because an experienced musician can look at these scores and feel the naturalness of the correct tempi.  Where one doesn’t feel the natural tempo, you absolutely must consult Beethoven’s metronome marks.  He put them in there for a reason.

Medelssohn’s allegros are faster, in general, than the Classical masters.  Mozart’s prestos have to move, and fast! Shostakovich’s allegrettos are sometimes exponentially faster than Schubert’s allegrettos.  Well written music generally has a natural tempo, and the words allegro, andante, adagio, etc. are mood indicators, rather than firm indicators of so many beats per minute.  

Any conductor who has “fast” or “slow” tempos is a bad conductor.  A conductor needs tempi appropriate to this or that  composer and piece, not according to his own personal marketing angle.  Toscanini was a great conductor, not unthinkingly fast.  Otto Klemperer was a great conductor, and the slowness of his last recordings is probably due to his health issues, rather than some stylistic agenda. 

There is an incredible amount of distortion in the authentic performance movement, in regard to tempi.  Consider the crucifixus in Bach’s B-minor Mass in the Jon Eliot Gardiner recording.  It’s a joke.  You can wave tome after tome of scholarly exegesis in my face, but I prefer to rely on my own essential sanity, which tells me that the crucifixtion should not be a jolly dance.