This was the cycle that set the pace for performances
of Beethoven which use modern instruments but which take into
account all that has been learnt by the use of period instruments,
of which Harnoncourt himself was a pioneer. It was greeted with
a great deal of enthusiasm around ten years ago; more recently
the feeling has been expressed that it is not holding up so well
to the test of time.
It is not normally my practice, when listening
to music I know so well, to follow with the score; later on the
oracle may be consulted over specific points. It quickly became
evident that in this Beethoven cycle specific points would be
so frequent that the score was indispensable. These are performances
that strike more for their small details than as a whole. So anyone
who reads my notes below and thinks I am being pernickety, seizing
upon niggling matters and ignoring the overall line, is invited
to look up my other Beethoven symphony reviews on the site (there
are quite a lot) and reflect that, if I adopt a different method
here, it arises from the nature of the performances themselves.
Symphony no. 1
Harnoncourt’s liking for rhetoric shows in his
creative treatment of note-values and rests in the introduction
to the first movement. This makes for a more throat-clearing effect
than those performances which proceed at an even tempo. At least
it avoids a drop in tension as the Allegro con brio starts, as
sometimes happens. The main body of the movement bowls along ebulliently
at a swift but not excessive tempo. Not actually memorable but
satisfyingly vital.
In the following movement one is struck by the
swift one-in-a-bar tempo and by Harnoncourt’s very deliberate
slurring of the pairs of notes which characterise the main theme:
the first and second note, the fourth and fifth, the seventh and
eighth and so on. It is a moot point whether Beethoven meant these
slurs as phrasing or, as most other conductors seem to
think, as bowing; that is to say a technical matter for
the violinist. The effect is that of a courtly minuet full of
little bows and graces. I am torn between finding it piquantly
charming and feeling that plainer interpretations have found more
depth in the music.
A very fast Menuetto (so-called, we all know
this is a true Beethoven scherzo) means slowing down for the trio.
I find it odd that Harnoncourt plays the wind chords in the trio
so smoothly, completely ignoring Beethoven’s staccato markings.
Several conductors have seen fit to repeat the first section of
the Menuetto when it returns after the trio; Harnoncourt repeats
the second section too, and I wonder what his authority is (several
of the cycles on period instruments also adopt this practice).
The finale is basically brilliant, but as early
as bar 8 of the Allegro molto e vivace I wondered why Beethoven’s
staccato markings were being made so little of. Admittedly the
coiled-spring Rossini-like staccatos which Toscanini taught us
to accept as the norm may be overdone in the opposite direction,
but this passage vocalises one of my big worries about these performances;
for all their speed, they can be strikingly deficient in actual
forward movement.
Symphony no. 2
In the Allegro con brio of the first movement
Beethoven has marked an unusual number of dynamic contrasts, sprinkling
liberally fortissimos, pianissimos and sforzandos all over the
score. Obviously these have got to be done, but you have to find
the music in them. Much of this is simply brutal, forcing an ugly
sound out of the orchestra. More bullish than ebullient.
The Larghetto shows that Harnoncourt can proportion
his fortissimos to the context in hand when he wants to, and much
of this flows quite nicely. I query whether the march rhythms
starting at bar 128 should dominate the texture when other instruments
have melodic phrases which most other conductors prefer to bring
out.
Another swift scherzo, superbly sprung, resulting
in a slower trio. In theory I agree that the trio should follow
immediately, without any pause, but when the reverberation of
the hall means that the first bar of the trio is completely covered
by the echo of the end of the scherzo, then common sense suggests
that a tiny pause would be in order. Here again we have both repeats
in the scherzo when it returns and if you’re not expecting the
second, your surprise will be all the greater since, thanks to
the reverberation, you won’t notice they’re playing it till they’re
about a bar and a half in.
Some terrific playing in the finale, which goes
at a real lick. The magical change to D minor on the rondo theme’s
second appearance is rendered null by the reverberation. For all
its vitality, I found this a pretty joyless, jack-booted reading.
Symphony no. 3
A swift first movement has an impressive sense
of continuity, although even Harnoncourt has to yield a little
in second subject territory and he makes a notable rallentando
on the three forte chords that close the exposition (and an even
bigger one at the end of the recapitulation). The effect, in the
context of such a tightly controlled interpretation, is incredibly
pompous, like an old gentleman waving his umbrella to stop a taxi.
There is the expected thrashing at accents, but for all its busy-ness
the performance gives less of an impression that it is getting
somewhere than many others that build it up more patiently. It’s
a very modern hero and one wonders if Harnoncourt is suggesting
that Beethoven had a hidden agenda, rather on the lines of some
of Shostakovich’s depictions of Stalin; praising Napoleon to his
face while (for those in the know) sneering at him behind his
back. But if this had been so, Beethoven would not have needed
to scratch out the dedication.
The opening of the Marcia funebre will be the
stuff of an original instruments man’s dreams. The strings are
shorn of all vibrato, and instead of building up a long legato
line, as incorrigible romantics from Weingartner to Toscanini
and Klemperer have done, the long notes are allowed to fade away.
It’s rather impressive. Another section which gets an interesting
new look is the fugato starting at bar 114. One of the conductor’s
tasks, according to Wagner, was to bring out the melody. Mindful
of this duty, conductors such as Klemperer have brought a rare
luminosity and transparency to this passage by giving each line
its exact weight and guiding the ear towards the part which carries
the argument forward. Harnoncourt evidently believes that the
conductor’s duty is to bring out the sforzatos, stabbing at them
without trying to relate them to their context. The passage acquires
a new look since the familiar melodic lines are obscured by a
series of sforzatos arriving from various parts of the orchestra.
It’s fascinating in a way, and if you think it’s what Beethoven
wanted you’re welcome to it. What I do find impressive, though,
is the way Harnoncourt holds his tempi steady in the various episodes;
many conductors, Weingartner in primis, find it necessary to move
forward.
In the Scherzo Beethoven made one important change
during the repetition after the trio; the insertion of a few bars
in 2/4 time. On account of this he had to write the whole lot
out again instead of merely writing "da capo". So what
does he do about the repeats? The short first section, the repeat
of which was written out in any case, is maintained, while that
of the longer second section is not. There are similar cases elsewhere
in Beethoven’s work and they all point in the same direction;
short first section repeats are maintained on the repetition after
the trio, long second section repeats are not. The scherzo is
so swift that the tempo has to slacken between bars 128 and 150;
that apart it is superb. The delayed upbeats Harnoncourt applies
frequently in the trio are just as mannered and irritating in
their way as was Furtwängler’s romantic dawdling in its later
stages, and perhaps less musical.
The finale, for all its speed, has little Beethovenian
drive (or is that a romantic concept I should try to forget?)
and sounds rather segmented. Given the conductor’s ideology, I
would have expected a less protracted treatment of the Poco Andante.
Harnoncourt makes some interesting comments on Beethoven’s metronome
markings during an interview in the booklet, and of course we
don’t expect slavish observance of them, but surely their relative
values tell us something? This Poco Andante is scarcely faster
than the Marcia funebre, yet Beethoven marked the latter 80 and
the former 108, which is a big difference. By the time we get
to the great horn statement of the "Prometheus" theme
this is the same old romantic Beethoven we have always known.
Symphony no. 4
An impressively mysterious introduction. Maybe
neither Harnoncourt nor his players really believe in the zippy
pace he sets for the Allegro vivace since the tempo keeps dropping
back, picking up and dropping back again until by the time the
exposition is repeated they have settled down to a perfectly normal
speed. Thereafter things go very nicely though I must point out
a couple of oddities. At bar 81 all instruments are marked fortissimo,
but the theme is in the lower strings. If each instruments plays
at his fortissimo all we will hear are the trumpets and
drums hammering away on just two notes, which isn’t very interesting.
The normal practice is to mark down the fortissimos on the heavier
instruments and mark up those on the weaker ones in order to bring
out the melodic and contrapuntal interest of the passage. Indeed,
in the old days that’s what people thought a conductor was there
for. Evidently Harnoncourt doesn’t agree. Another oddity is his
accenting the staccato string octaves from bar 121 in groups of
three, for which my score gives no authority at all and reduces
the music to mere pattern-making, followed by a pompous rallentando
in bars 133-4. A pity; as I said, much of this is very good, and
Harnoncourt keeps his sforzato accents here in reasonable proportion
to the context.
The Adagio has the long melodies very beautifully
played, but the accompanying figure is very jerkily done, deliberately
intrusive, rather as though somebody is cheekily playing a polka
in the background which has nothing to do with the matter at hand.
I suppose nothing in the score actually says it mustn’t be played
like this, and if you like it ...
Virtually every performance I’ve heard of the
scherzo suffers from a tendency to separate the rising and falling
phrases between the wind and strings, creating a stuttering effect.
Harnoncourt avoids this pitfall and this movement is extremely
vital and brilliant. Good, too, that he plays the trio only Un
poco meno Allegro, as Beethoven asks, and not Molto,
molto meno Allegro, as so often happens. This is another case
where Beethoven writes out the return of the scherzo – and eliminates
both repeats.
The finale is a triumph of spick and span orchestral
playing. If you think this music has spiritual qualities too,
you’ll have to go elsewhere, but it’s certainly vital. At bars
305, 307 and 309 Harnoncourt evidently feels that Beethoven hasn’t
given him enough accents to jab at, and provides some of his own
(at any rate, they aren’t in my Hawkes Pocket Score; perhaps more
recent scholarship has found them). Despite my reservations, this
is the best performance so far.
Symphony no. 5
A thrustful, urgent first movement – not a hint
of an unmarked romantic rallentando in the four-note motto theme,
naturally – but also finding time for a very clearly phrased second
subject, more meaningful than is often the case. It is noticeable
that in this movement Harnoncourt balances the instruments so
as to bring out the melodies, exactly as I complained he did not
in parts of no. 4. What an unpredictable man he is!
It is a measure of the amount of tempo variation
we usually hear in the Andante con moto that while you will perhaps
find Harnoncourt rather fast at the beginning (it seems a minuet),
many other passages sound "normal" and a few even seem
slow. Harnoncourt’s steadiness obviously makes it all the more
telling when Beethoven really does go into a faster tempo for
a few bars near the end. I don’t think I’ve ever heard a better
account of this movement.
In the first edition and most subsequent printed
editions, the scherzo and trio are not repeated – the trio leads
directly into the mysterious pizzicato reprise of the scherzo
and hence to the finale. Some evidence has been found that Beethoven
originally wanted the scherzo and trio to be played twice. The
first to record it like this was Pierre Boulez, a record that
apparently existed solely to make this point, so uninterested
did the conductor seem in the rest of the symphony, to the extent
of omitting the repeat in the finale, which in the context seemed
quite perverse. Whether the repeat is needed or not, you can hardly
regret hearing such an urgent performance as Harnoncourt’s twice
over. The trio is fractionally slower but very fine all the same..
The mysterious reprise and the link to the finale
generate a good deal of tension. When the finale itself bursts
in it has the crudity of a village festival. Harnoncourt may argue
that Beethoven’s art is so all-embracing as to find space for
a spot of honest-to-God banality, and he may be right. He also
makes more than most conductors of the dolce marking on
the second subject. One query: when, as in bars 22-25, there are
sforzatos on the offbeats, conventional wisdom says that we have
to accent the beats too, otherwise we will hear not syncopations
but a displacement of the beat and bar-lines. In baroque music
it may often be right to create this sense of displacement, but
I am not so sure that the practice continued into the classical
period. This is a particularly clear example but I’ve been a little
perplexed by several others in all the symphonies up till now.
Still, recommendable fifths are few and far between,
and this certainly is one.
Symphony no. 6
I braced myself for an upfront arrival in the
country. You could have knocked me down with a feather when I
heard the gentle opening! At 13:07 we are in Klemperer (1957)
territory (13:04), but the Klemperer conceals a host of subtle
tempo modifications while Harnoncourt is absolutely steady. The
long crescendos and diminuendos are superbly controlled as are
the dynamic gradations between piano and pianissimo on the one
hand, and forte and fortissimo on the other. Nothing is allowed
to disturb the serene, sublime atmosphere – sforzatos are carefully
related to their context.
The Scene by the Brook is swift – this time Harnoncourt’s
11:59 compares with the 11:56 of Keilberth, who is exactly on
Beethoven’s metronome mark – but more than any other swift reading
I know, this one succeeds in maintaining a mood of total serenity.
After a while it actually comes to sound slow. Like Weingartner
and Keilberth, Harnoncourt separates the three-note motives in
the accompaniment at the beginning. He also gives exactly the
right weight to the various syncopated notes, for example the
horns from bar 7, so they register but do not intrude.
After so much calm the Merrymaking of the Country
Folk has a welcome vitality and the storm is powerful at quite
a broad tempo – which is what Beethoven asked for. Quite a number
of conductors have noted that Beethoven’s metronome marking for
the finale is only a notch faster than that for the Brook, and
have made a memorably poetic moment out of the transition from
the storm. Usually, however, they feel the need to move on a little
later. Harnoncourt maintains his slow tempo, returning to the
mood of Olympian sublimity with which he began, if anything winding
down still further towards the end.
In some moods one might wish for a more bracing
approach, but when you want the calmest, most serene Pastoral
imaginable, here it is, a remarkable achievement, and who would
have expected it from this source?
Symphony no. 7
No slackness about the tricky dotted rhythms
in the main body of this movement, which is left to make its point
bluntly but strongly. Though the booklet describes the second
movement as Allegretto the actual feeling is closer to Beethoven’s
first thought, Andante. An impressively grave reading, with no
running away in the more lyrical sections. A fast and brilliant
Scherzo has a trio which avoids the romantic dawdling which used
to be common and which would be quite intolerable in a performance
which observes all repeats.
Thus far the performances is mainly a catalogue
of pitfalls that have been avoided; exemplary but not as incandescent
as some (hear the live Beecham from Switzerland on Aura). One
would be grateful for this, but unfortunately the finale falls
into a pitfall of its own. Scrupulously bringing out the sforzatos
on the second fourth-note and then the fourth eighth-note (the
latter are usually lost), Harnoncourt has failed to notice that
thereby the swirling theme in the strings – which is the principal
theme of the movement – goes unheard. Anyone who knew the symphony
only by this recording would be unaware that the finale had a
theme at all – it is reduced to "sound and fury signifying
nothing". In view of the many excellent versions around I
don’t see how I can recommend one that gets a whole movement as
wrong as this.
Symphony no. 8
By the standards of period instruments-influenced
performances this has some fairly relaxed tempi (timings are longer
than Norrington, for example), or perhaps it is Harnoncourt’s
carefully controlled phrasing which makes them seem so. The first
movement seems more majestic than urgent and is very appealing.
Consistently with the other performances, off-beat accents are
allowed to create the effect of displaced bar-lines rather than
a syncopation. In the case of bars 70-72 I don’t see how mere
staccato dots can justify turning the passage inside-out compared
with how we usually hear it.
The second movement is light and graceful and
the minuet flows at a good tempo – just as well since we get both
repeats on its return after the trio, about which I’ve already
had my say. The trio is introduced by a surprisingly romantic
ritardando and what follows is delightfully affectionate. With
a finale notable for the beautiful playing of the lyrical second
subject as well as for its overall drive, this adds up to a highly
recommendable version. I should point out that, if you listen
at a neighbour-friendly volume, the dynamic range is so wide that
you might scarcely hear the four note figure which drives the
development of the first movement along, and you might not even
notice the finale has started until the woodwind enter.
Symphony no. 9
The metronome markings of this symphony have
given the original-instruments brigade a field day, since they
vary between the impossibly fast and the unusually slow. Harnoncourt
seemingly ignores the question and produces a fairly traditional
performance, clear and well shaped but without any great aspirations.
It is nice to have it spelt out so clearly which chords, in bars
149-150 of the first movement, are forte and which are fortissimo,
and I shall never again be indulgent towards the conductor who
doesn’t notice or can’t be bothered. There are similar points
all through. I’m pleased to report that he does not give
the repeats when the scherzo is repeated after the trio – for
this relief much thanks.
The slow movement contains a few idiosyncrasies.
On the upbeat to bar 8, why dig in as if there were an accent?
A little later, at bars 15-17, is there any reason why the accompanying
quavers in the strings should be staccato? And if there is, why
do the wind not play in the same way when they have the same music
a few bars further on? But to tell the truth, Beethoven shows
in bars 56-58 that he has sufficient musical knowledge to write
staccato dots when he wants them (did we doubt it?).
The finale is a clear-textured, level-headed
affair, apart from a very exaggerated "poco adagio"
shortly before the voices enter. The soloists are good and the
last solo quartet makes more sense than it often does. But in
the last resort I was underwhelmed, and that’s the last thing
I want when I hear this of all symphonies.
I’m afraid this very mixed bag seems to suffer
from too much of a "historical" approach. It is as if,
for Harnoncourt, the "Eroica" can only be a stepping
stone between nos. 2 and 4, not a revolutionary argument only
superficially related to its own period. And the Ninth, by the
same token, is just the next step up from the Eighth, a bit bigger
and better, but not a great leap into the unknown. Thus two of
the most epoch-making symphonies ever written emerge belittled.
The listener who learns his Beethoven from this set may get the
idea that the composer progressed logically from the First to
the Fifth, touched the sublime in the Sixth and thereafter rather
lost his way (as also in the recent Aimard/Harnoncourt set of
the concerto, where the greatest heights of sublimity are touched
with the Fourth).
Clearly, this is not a cycle I can recommend
in its entirety. It is also being issued on separate discs on
the Elatus label. You should certainly hear 5 and 6, maybe also
4 and 8, and even the Ninth. Or you might look up David Wright’s
reviews of 3, 4 and 5 on the site and reflect that, if we critics
can’t find a little more consensus of opinion than this, you’ll
just have to forget us and make up your own mind.
Christopher Howell