Brahms: The Four Symphonies. Academic Festival Overture; Haydn Variations; Symphony No. 1, Andante sostenuto (first performing version, ed. Robert Pascall)
Charles Mackerras, cond; Scottish CO Telarc CD-80450 (3 CDs: 199:02)

Reviewed by Bernard D. Sherman. Reprinted by permission from Fanfare (November/ December 1997) Revised January 1999

Selected comparisons discussed in the review:
Complete symphonies, plus overtures and "Haydn" Variations:
Abbado/ Berlin Philharmonic (DG); Chailly/ Concergebouw (London); Furtwängler/Vienna and Berlin Philharmonic Orchestras (Music and Arts CD-941, 4 discs for the price of 3, including also the Second Piano Concerto);Kubelik/Bavarian Radio Symphony (Orfeo) Norrington/London Classical Players (EMI, 3 separate discs); Walter/Columbia Symphony Orchestra;
Individual symphonies:
Symphony No. 1:
Abendroth/Bavarian State Orchestra, 1956 (Tahra 141-142);
Symphony No. 2:
Beecham/London Philharmonic 1936, Mengelberg/Concertgebouw 1939 (Biddulph 057); Bohm/Vienna Philharmonic 1942 (Preiser 90922); Damrosch/New York Symphony Orchestra, 1928 (Biddulph 052)
Symphony No. 4,
Abendroth/London Symphony, 1927 (Tahra 102 and Biddulph 053)

ith only 49 members, the Meiningen Court Orchestra was miniature for its time, but that didn't keep Brahms away. He worked with the Meiningen regularly, and when given the chance to augment its string section for the premiere of his Fourth Symphony, he declined. On this set, Sir Charles Mackerras has re-created the size of the Meiningen and revived some aspects of its layout, with first and second violins on opposite sides of the podium. That left-right division of the fiddles has become rare since the Second World War, but Brahms counted on it in a number of string passages.

Brahms did not count on orchestras' being as compact as the Meiningen. When he conducted his Handelian Triumphlied he wanted as huge a group as he could get, and he seems to have preferred large groups for the German Requiem as well. What he liked about Meiningen may have been not its size but its precision (Brahms did once or twice express a preference for small groups for brand-new symphonies, but precision may be all that was about. And after premieres there's no evidence that he preferred small groups.) So size is not a true advance here. Besides, Sir Roger Norrington recorded Brahms with an orchestra almost this small using period instruments; Mackerras uses modern instruments except in those few cases where, in his judgement, period instruments make a difference. He uses Brahms-era leather-skinned timpanis, "Vienna" horns, rotary-valve trumpets and narrow-bore trombones not because Brahms did, but because Mackerras finds them more rhythmically incisive than their modern descendants. I consider this approach to period instruments healthy; to quote my book Inside Early Music, it's a good sign when period instruments are a choice, not a recipe.

Compared to modern brass sections, the period brass blend less with each other and with the rest of the orchestra. Their heterogenous sound, in tandem with the small (some will say undernourished) string sections, helps project important lines that many performances submerge. For that kind of clarity, however, what is crucial is not the instruments but the conductor's ears, and I suspect Mackerras could achieve similar clarity with a standard modern orchestra. The transparent texture, by the way, is not the work of the engineers, who used only two microphones and no mixing console. Audiophiles will plotz.

Mackerras's orchestra sounds something like Norrington's but is far more appealing. Mackerras's use of modern strings and winds helps, but the real difference lies in the style of string playing. The attenuating of long notes in Norrington sounds mannered. And Norrington's strings chastely minimize their use of such expressive devices as vibrato and portamento (sliding from note to note), while Mackerras's indulge more.

On the other hand, don't expect the Berlin Philharmonic. The Scottish strings don't equal the Berlin's virtuosity (or, for that matter, the virtuosity of the other sections of their own orchestra); but they also have a different sound ideal from the continuous vibrato of Herbert von Karajan. While orchestral string players certainly did make use of vibrato in Brahms's Austria and Germany, they used it more selectively than Karajan did. Mackerras has that in mind, even if he may not re-create the full extent of period fingering. In any case, I certainly don't find the strings an impediment to his interpretations.

The biggest difference between Mackerras and Norrington is in the interpretations, and the Meiningen Orchestra plays a role here as well. One conductor associated with the Meiningen was Fritz Steinbach. According to Brahms's friend and biographer Max Kalbeck - perhaps not a reliable reporter - Steinbach "modeled" his conducting on Brahms's own and became the composer's favorite interpreter of his works. We do know of Brahms asking Steinbach to repeat a performance of the Fourth. We have no recordings of Steinbach, but we do have a 1933 typescript by his disciple Walter Blume, which purports to detail Steinbach's Brahms interpretations bar by bar. We can't know for sure whether the book is accurate, or whether it bears any resemblance to Brahms's style. But it opens up new ways of thinking about performance - ways that, as Walter Frisch has pointed out, differ even from the "freer" Brahms conductors of the recorded era, such as Furtwangler and Mengelberg; The the Meiningen style presented here involve far more local nuances of tempo and articulation than even these conductors would use. Norrington makes an occasional nod to the book, but Mackerras engages with it more fully.

Those with access to it will notice the influences right from the start of the Mackerras First. The most striking example comes in the second thematic group - a place where Norrington avoids almost any inflection. As Frisch has written, Steinbach would set off the four-note oboe motif in bars 133-34 and 137-38 with a little breath before and after it, and a slight broadening of the tempo during it. Mackerras follows suit [track 1, 4:17-4:24]. A case could be made that the score's notation, with its two-note slurs and its hairpin swells on these motifs, is trying to convey exactly this effect. But it hardly matters - the effect is so convincing that it justifies itself. Steinbach also asks that the ensuing horn/clarinet dialogue over this motif be "freely shaped," and so it is here. It is one of the high points of the set.

The exploration of the Steinbach manuscript yields many such high points. Some of them involve articulation, which is more detailed than in mainstream Brahms playing--the oboe motif in the First is a good example-- and I was almost always grateful for the added nuance. Only one of the articulations is hard for my ears to accept: the opening of the Second, whose horn melody is notated with a series of two-note slurs. Brahms once wrote to Joachim that in some cases he wanted the second note under a two-note slur to be played shorter than written; but could he really have wanted that in these melodious opening bars? Perhaps. Elsewhere in the score Brahms wrote some very long slurs, which suggests that he wanted something different here; and Steinbach appears to have articulated these opening slurs as Brahms describes it in the letter. My view, however, is that both Mackerras and Norrington overdo the effect - a typical case of modern performers exagerrating a historical finding. The second note under each slur is cut too short; it sounds a bit like a hiccup. What comes closer to the mark are recordings by Mengelberg (1939) and Walter Damrosch (1928); while these performances provide a slight breath between the slurs, the second notes under the slurs have a less marked cutoff. One can also hear some of the effect in 1940s recordings of the Vienna Philharmonic under Furtwängler (1945) and Bohm (1942); I take this as the tradition-based work of the Viennese horn player (since Furtwängler's post-war Berlin recordings are phrased in the modern way).

Still, it wouldn't be fair to accuse Mackerras of using the Steinbach book as a recipe, or of botching the instructions because of their unfamiliarity. Mackerras usually integrates the Steinbach ideas into a natural-sounding musical whole, so that they don't feel pasted on. Besides, he adds interpretive touches of his own - and most important, he ignores many of Steinbach's. In the First's finale, bar 395, Steinbach apparently emphasized the ben marcato by slowing just for that bar; Mackerras chooses instead to maintain the head of steam he has built up. In Fourth's first movement coda (and its preceding bars), the Steinbach book gives detailed instructions about lengthened upbeats, holding back and acceleration, and so forth - and Mackerras ignores it all. If you want to hear these Steinbach practices, you can, in Hermann Abendroth's recordings of the works; Abendroth, Frisch suggests, may well have had a direct connection with the Meiningen tradition, not one based on a book. Moreover, in Abendroth's 1927 London SO recording of the Fourth (now available on Biddulph), the extensive portamento in the opening phrases surely takes us closer to period style than anything by Norrington or Mackerras. Still, Abendroth can't approach the latter two for clarity of orchestral texture, even in his post-war recordings.

Brahms (or perhaps Joachim) pencilled a comment into the score of the Fourth saying not to accelerate in the coda in the opening movement. If that is what Mackerras is responding to here, he might still be accused of following interpretive recipes. But the accusation has no strength in bar 395 of the First finale and in many other places where Mackerras diverges from Steinbach; the clear explanation for these is that the instructions in the book make no musical sense to Mackerras, and he follows his instincts. Indeed, the musicology is not what makes this new set so notable. True, it has led Mackerras to effective interpretative nuances he would not have thought of otherwise; but they work only because Mackerras knows how to shape paragraphs and phrases, he has mastered the difficult art of transition, and he has a feeling for the ambiguity, range, and depth of emotion in Brahms. While Norrington often sounds as if he doesn't speak the language, Mackerras's fluency rarely fails.

An arguable point is right at the outset, when he takes the "Un poco sostenuto" introduction of the First in a tempo exactly proportional to the ensuing Allegro (with the latter's dotted quarter at the same speed as the introduction's eighths). My doubts about this approach are by no means universally shared: some major Brahmsians - Toscanini, Karajan, Weingartner, Szell, Giulini - also maintain an exact tempo here, and the theorist David Epstein and composer/conductor Gunther Schuller argue that this is what Brahms intended. But I would argue that Brahms didn't intend this proportional relationship. [Click here for an explanation of why I think so] As to whether you prefer it, it's a matter of taste - to my ears the opening tempo feels too fast in some of the above performance, while clearly others disagree. In any case, Mackerras's speeds in the first movement (including the introduction) sound totally convincing even to me, and I prefer Mackerras in this movement to any of the recordings I listed above. To mention only Weingartner, his tempos are almost identical to Mackerras's, but Mackerras is more nuanced and expressive.

Mackerras is by no means dogmatic about using such "proportional tempos" - tempos that relate to each other in simple ratios like 1:1, 1:2 or 3:2. He does use some others in the First, but these are to good effect. Epstein has called for relating tempos this way across entire Brahms works, but I've argued (Early Music, August, 1997) that Brahms typically did not relate tempos of different movements this way. As often as not, though, he would relate different sections of a single movement by such ratios. In two such cases - the First's Finale and the "Haydn" Variations - Mackerras's ratios differ from Epstein's, and I prefer Mackerras in both of them.

To return to the First Symphony: in the Andante sostenuto, the Mackerras performance of the main theme is moving, though later in the movement I would have liked the tempo to be even more responsive and flexible. The finale's main theme is phrased gloriously. The lingering on the first upbeat note is just one more well-digested Steinbachism. Overall, Mackerras's First is one of the most remarkable in digital sound.

In keeping with the lilting opening discussed earlier, Mackerras's Second has the fastest initial tempo since Mengelberg (Mackerras proceeds at about 112 to the quarter note, Mengelberg at 118). Both are faster even than Beecham in 1936 or Furtwängler in 1945 (both of whom begin at 108; the typical tempo is 100-104). Mackerras is more insightful and moving than Beecham, who passes over too many great moments, and is more coherent than Mengelberg. Mackerras falls far short of Furtwängler, but his performance is similar in one respect: the tempo relationship of the dolce theme that arrives in the first violins in b. 44. Most conductors increase the tempo at this point, but doing so can create a problem later in the movement, when Brahms opens the recapitulation by combining the opening horn theme with this b. 44 theme, both reorchestrated. As Reinhold Brinkmann asks in his book on this work, Late Idyll, if a conductor takes the themes at different tempos early on, which tempo should prevail when they combine at the recap? Mackerras, Furtwängler, and a handful of others (Stokowski, Monteux, Walter, and Giulini), needn't choose, since they all give the two themes essentially identical tempos when they first appear. [For more on the 1945 Furtwängler performance, click here.]

Mackerras's tempos are less flexible in this movement than Norrington's, by the way, but his interpretation is more successful. Norrington does well with the movement's more extrovert moments, but disappoints in the lyrical or deep ones. An example is the mysterious timpani/trombone/wind passage that leads into that new violin theme in b. 44. Norrington is oddly dry in the transition and in the new theme, but Mackerras makes something special of them both; notice the warmth with which the new theme flows forth.

The Adagio's main theme, too, benefits from the use of old-fashioned portamento. And in the finale, Mackerras uses some effective Steinbach touches, especially in the second group (note the expressive swells), and in the coda where, starting in bar 386, he does a momentary pull-back (8:07).

Most conductors increase the tempo in a few places in this finale, the first one being bar 23 (where, after the sotto voce opening, the full orchestra crashes in), and the last one being the coda. But acceleration, like portamento, began to seem tasteless sometime in the middle of our century, so moderns tend to use it far less here and in other movements than did earlier Brahmsians. This reticence contributes, I think, to the reserved quality of some modern Brahms. Nobody will accuse Mackerras of that in this movement. He picks up the tempo at bar 23 by about 12 metronome points, which is almost as much as Furtwängler in 1945 (16 points). In the coda he accelerates eight additional points, and if he doesn't attain the sheer abandon of Furtwängler (who accelerates twice that amount), he still is joyous. Furtwängler 1945 remains the ultimate in this movement and, indeed, in the entire symphony; but the Mackerras reading earns a place among stereo recordings. (Some favorite digital recordings, by the way, are those of Claudio Abbado and the Berlin Philharmonic, and Riccardo Chailly with the Concergebouw.)

The Third may be the finest performance in this set. Mackerras builds the opening theme to a thrilling high point, even making subtle use of the change in articulation - the first real staccato in the theme - that Brahms indicates at that point (0:19 seconds, bar 10). Mackerras sails into the development section after taking the crucial repeat (he takes those in the first two symphonies as well, though these are less important). At the transition back to the recapitulation, he finds the best possible tempo relationship - as Frisch explains in his book on the symphonies, a half note in the retransition should equal about a dotted half note in the recapitulation. Performances of this movement rarely have enough of the Dionysian for my taste; Sir Charles has plenty. In the Poco Allegretto third movement, Mackerras lengthens the first note a bit, as Steinbach did (and many old recordings do as well). The effect is eloquent, as is the liberal use of portamento and the general lack of hurry. All of it serves the movement's melancholy. Another nice touch is the increase in tempo for the unmarked "trio" section - something Steinbach calls for, "to bring out the scherzando character." Steinbach's ideas also enrich the finale. Mackerras builds its exposition to its true climax in the C-minor theme that begins at bar 75; part of that episode's power come from Steinbach inflections, such as giving extra emphasis to the upbeat note just before bar 75 (1:54) and to the cadence in bars 90-91 (2:16). In many other respects, Mackerras shows his own insight.

Mackerras's mastery of transition is evident in his Fourth. At bar 110 of the first movement, at the end of a mysterious pianissimo passage in the strings, a chivalric theme suddenly enters softly in the winds (2:57). A ritardando or luftpause is out of place, but something is needed, or else the transition sounds metronomic. That is just how Norrington sounds, because he does exactly the wrong thing - he takes the new theme imperceptibly faster than the previous passage. Mackerras, by contrast, takes the new theme imperceptibly slower and, more importantly, he doesn't hurry its half notes and rests; the effect is perfectly judged.

While the first movement is flexible, it doesn't start nearly as far below the main tempo as Abendroth, Furtwängler or Norrington, and it doesn't accelerate nearly as much in the coda. As we've seen, the latter case may be Mackerras's deference to a marking penciled into the score. It's worth adding that Brahms penciled a few calls for speeding or slowing into the score in the finale, but that Mackerras observes only some of them. Brahms's pencil markings counter the idea that the finale of the Fourth requires an unchanging tempo for its structure to be heard; they bolster the opposing view that the sonata-like proportions (described by Frisch) require some changes of tempo to be made audible. Mackerras's handling of tempo variation in this movement shows how.

The Haydn Variations are unusually fleet, but not rushed, and the Academic Festival Overture is right up Mackerras's alley. In the First, Mackerras includes both the familiar second movement and the very different version played at the premiere, as re-constructed by Robert Pascall. Hearing them both gives irreplaceable insight into Brahms's creative process, as we hear the familiar material used less convincingly in the first draft.

Mackerras's four performances are worth hearing in their own right. Besides, they could give conductors and orchestral players new ideas about what Brahms imagined when he orchestrated, at least with respect to balances and to brass articulation. Let me add that I am not claiming that period instruments are important in Brahms (I've come to believe that they usually are not) - nor, again, that small is more beautiful for Brahms orchestras - but rather that the sound some Brahms conductors favor today is too thick to let the music's polyphony emerge. Also, Mackerras may cause some musicians to reconsider the cliché of seeking profundity in Brahms through sheer weight and slowness. Like other true Brahmsians, he relies instead on passion and nuance in phrasing and articulation.

After Norrington's missteps cast doubt on period-instrument Brahms, and Gardiner's Requiem proved deeply flawed, Mackerras gives us our first convincing reason to think that historical investigation can contribute at least something to Brahms orchestral performance. Brahms lovers should hear him; the disc with the Third and Fourth would be the best place to start sampling if you don't want to invest in the complete set. For those who do want complete sets, Furtwängler on Music and Arts would be my own desert-island choice. And modern performances by Abbado (the Second and Third), Walter (the Second and Third), Kleiber (the Fourth), Chailly (the Second),and Kubelik (all of them, especially his Fourth) have been noteworthy. But Mackerras's is one of those rare sets I would recommend buying complete.
- Bernard D. Sherman

Back to home page

NOTES (in case you didn't click on these links from the main text):

Why I think that Brahms wanted the introduction to the First Symphony to be slower than the Allegro that follows. Return to the main text

Gunther Schuller (in The Compleat Conductor, pp. 280-81) and David Epstein (in various writings) argue that Brahms wanted the introduction to the first movement of the First to be taken in a tempo exactly proportional to the ensuing Allegro - with the dotted quarter of the latter at the same speed as the eighth of the former - a policy Mackerras adopts. When you take it this way, it projects a sense of continuity - a sense that the underlying tempo has not changed. Why do I question that? After all, Brahms didn't want the introduction to be very slow, as is shown by his adding the words "Un poco" to the original marking, "Sostenuto." But to Brahms, the capitalized word Sostenuto suggested something about tempo, namely a slower tempo than the main one (there is good evidence for that in his correspondence and his metronome marks). If Brahms had hoped to get conductors to take the introduction at the same tempo as the Allegro, it would have been more typical of his notational habits to have indicated that more directly. He could have begun the Allegro section by noting the explicit ratio of eighth note=dotted quarter, a type of notation he used in, for example, the third movement of the Second Symphony. But the actual markings he used-- Un poco sostenuto, followed by Allegro-- would indicate to most musicians of his day a somewhat contrasting pair of tempos, not an identical pair.

Another bit of evidence: later still, Brahms had his publisher change the tempo at the end of the first movement to "Meno allegro," instead of the original marking, "Poco sostenuto." His reason? I quote: "people misunderstand that ["Poco sostenuto" marking] and take the same tempo as in the introduction." This almost certainly means that that they were slowing down too much at this point-- thus the change to the more moderate sounding "Meno allegro." The implication, obviously, is that the introduction had been audibly slower than the Allegro.

To defend the idea of the unaltered introductory tempo, Epstein or Schuller might, hypothetically, argue that what Brahms meant by taking "the same tempo as in the introduction" was that conductors were keeping the tempo unchanged at the end, instead of slowing down. But this interpretation is so implausible that I doubt that either of them would in fact propose it. After all, if Brahms had been trying to get conductors to slow down more at the end, it is hard to believe that he would have changed the original marking to the faster-sounding "Meno allegro." And why would he complain about conductors taking the same tempo as the introduction, if the Allegro was also at the same tempo--would he not mention the Allegro instead?

[Norrington, by the way, goes faster in the introduction than the Allegro, which makes no musical sense, and (for what it's worth) has no historical justification at all.] Return to the main text

Link to my review of Schuller's
The Compleat Conductor.

Link to my study of Tempo and Proportions in Brahms, which discusses Epstein's theories in detail.

2. Furtängler's 1945 Brahms Second
Furtwängler's 1945 performance of the Second with the Vienna Philharmonic leads many writers to rave unabashedly. Most insightful is Walter Frisch's detailed analysis in his extraordinary book Brahms: The Four Symphonies (pp. 181-185). I agree with him and others that this recording, taped minutes before Furtwängler fled to Switzerland to escape a Nazi plot on his life: it is the greatest Brahms Second on record, and surpasses Furtwängler's postwar recordings.

After Frisch, there's no need to say more, but I will nonetheless add a few comments. In the main text, I spoke of the dolce violin theme at b. 44 of the first movement, and of how only a handful of conductors do not increase the tempo for it. But when I mentioned Furtwängler I was being a little approximate. In 1945 (though not in 1952), Furtwängler actually lets the tempo broaden here - the theme starts at about quarter=100, when the movement had opened at about 108. In practice, the tempo sounds more or less unchanged, but letting the theme breathe in this way results in its sounding more melting and profound than in any other recording. It also gives more voice to the movements' underlying dark side, a focus of Reinhold Brinkmann's book. Since Furtwängler's opening has some of the previously discussed lilt as well as the dark undertow, his is the most emotionally complete performance I've heard. (The treatment of the slurs in the opening theme, by the way, satisfies me more than any other: they are neither glossed over in a generalized legato, nor are they strongly punctuated.)

Another moment worth mentioning is the phrasing - both beautiful and insightful - of the opening of the second movement. This theme begins idiosyncratically on the fourth beat of the bar, an offbeat that is by convention the weakest, least stressed beat. I agree with Gunther Schuller's analysis of the implications (in The Compleat Conductor, pp. 291-294): the movement's meter should be ambivalent at first. Furtwängler gives a sense of this. Consider, for example, the last note, B, in the opening sentence (bar 3, penultimate eighth note in the cello line). This note is not only the last in its phrase, it is also brings us the movement's first tonic chord in root position. Thus most conductors give it extra emphasis and also insert a little breathing space before starting the next phrase. In 1945, however, Furtwängler does neither of these things. Instead, he keeps moving through the final B and on to the following phrase. And with good reason: that final B is also on the fourth beat of its bar. By heeding the implication of that placement - that the note should not be stressed - Furtwängler undermines the sense of resolution suggested by that first tonic root chord. Return to main text