To talk of Johannes Brahms is to enter a world where fantasy, melancholy and the contradiction of bluff appearances and unfathomable depths are crucial prerequisites. Violinist Gidon Kremer etches a telling metaphor. 'Look at virtually any picture of Brahms and you see a fat man,' he says. 'Of course, I can't say how much that influences your way of approaching his music, but too often Brahms is played in a rather "fat" fashion. One forgets that inside this ample person was a very fragile soul, and yet it can be difficult to transfer from what seems like mere physical weight to fragility, warmth and striving after the inexpressible. Brahms is too often considered a sort of throned monument, looking down on the romantic world; a mighty academician, the heavyweight champion of composing.'

Brahms's reputation for solid, beefy musical textures is based largely on his orchestral works - four symphonies, two Serenades, two piano concertos, two overtures (the Tragic and Academic Festival), the Violin and Double Concertos, Haydn Variations and the handful of Hungarian Dances that he orchestrated himself. It is a magnificent corpus of work, though there are still some who balk at its bulk. 'It's a very heavy world,' says violinist Frank Peter Zimmermann, ' and although I love to enter it and there's much of value in virtually every note - I opt for a slim-line style rather than the more opulent approach favoured by, for example, David Oistrakh.' For conductor Christoph Eschenbach, it's less a question of calories than of balance.' I think that Brahms was a splendid orchestrator,' he says, 'and that the people who reproach him are those same people who might also reproach Schumann. You could as easily criticize Beethoven and almost any symphonic composer for the same kind of thing. Mahler, however, knew the chances of a wrong interpretative conception, and so he wrote every tiny detail into his scores - having a flute play fortissimo in the same voice as the violins, for example.' Generally speaking, then, a good conductor will pre-empt a composer's anxieties about texture and balance by treating the problem as a priority. 'Yes. You must always delineate,' affirms Eschenbach, 'always identify what is most important in the texture. Of course everything is important but, like Schoenberg, you should think in terms of a "main voice" and a "subsidiary voice". It could only ever sound awkward with a bad score, but with Brahms...absolutely never.'

Riccardo Chailly takes up the Brahms/Schoenberg connection (the chosen couplings for his own Brahms cycle compound the point). 'This great conservative figure,' he says, regarding Brahms, 'who could handle sonata form with such immense ability, found his "mirror image" in Schoenberg. And it is indeed a paradox that the great "destroyer of the old" and "inventor of the new" felt his Brahmsian roots so strongly, could not ever think to cut them; that he felt the need, time and again, to frame his own new language with them.' Both Brahms and Schoenberg were, according to Chailly, 'vehicles that travelled between the past and the future'. It is therefore highly significant that both composers venerated - and were influenced by - the works of JS Bach.

Stephen Kovacevich, like Eschenbach, has experienced Brahms both from the keyboard and, latterly, from the conductor's rostrum. When we spoke during April he was preparing for a performance of the Fourth Symphony. 'I adore Brahms,' he exclaimed, 'but of the three great musical "Bs", he's the one whom musicians seem to enjoy the least. There's a kind of all-round indifference to the suffering that's inherent in his music. I hear performances like huge, featureless buildings, bland and gutless, even passionless: the most you can say of them is that they have a kind of generalized warmth. In fact, my feeling about most current Brahms performances is that they suggest a comfortable, overweight bank manager reluctantly refusing an overdraft - that's about as much suffering as they can muster.' 'With Brahms, there are many roads to Rome,' Kovacevich adds, 'and you can either opt for the weary, resigned, "perceived wisdom" route - or you can identify a sense of desperation in the music, the real passion in Bra hms, which is sometimes overlooked.'

Mighty confrontations

Brahms's two piano concertos are chalk and cheese, the First, in D minor, the phoenix born of a raging unfinished symphony but with an ineffably still core; the Second, in B flat major – which Brahms himself 'sent up' as 'a tiny, tiny piano concerto with a tiny, tiny wisp of a scherzo' - gnarled and granitic, a rugged, heart-warming masterpiece. 'If I'm in shape and playing part of the Second Concerto,' says Kovacevich, 'I find conductors telling me that they can't hear the orchestra - which of course thrills me, if only for "annoyance" value! And there are times when it's the other way around, when you can't be heard in the way that Brahms hoped you would be. It is a much more ambiguous work than the First; the first three movements are the greatest, whereas the finale is like a well-earned encore. The scherzo second movement is thrilling, and there's fury both in that and in the first movement; but there's also graciousness, voluptuousness and extraordinary tenderness.' Kovacevich talks of 'rubato on a very broad scale, though there 's very little of the kind of detailed rubato that you find in Chopin.' There are some hair-raising pianistic challenges, too. 'Anyone who has ever learned the Brahms Second Concerto dreads the time when they know they're going to have to play the middle of the scherzo (bar 215); it's a notorious passage which some people can't do reliably, and others can't do at all. I've probably played the piece about 50 times now, and I've only had one disaster. I don't know anyone who looks forward to the passage because not only is it difficult, it has to be played legato, and very, very quietly [the actual marking is sotto voce pp legato].' Kovacevich also mentions treacherous thirds in the finale, 'but as far as I know,' he adds wryly, 'nobody really tries them, so it's not a problem. I know one pianist who played them all - he's dead now - but he slowed up for them, drastically, and it sounded terrible!'

As to the First Piano Concerto, Kovacevich quotes David Cairns, who once said that 'after the D minor Concerto, Brahms put up the shutters'. A telling phrase, indeed, 'because there's something bare-faced and confrontational about the piece; it's a work where Brahms puts all his cards on the table.' Kovacevich feels that the huge opening tutti (or at least the first eight or so bars of it) benefits from selective re-orchestration. 'It always disappoints me when it's played the way Brahms wrote it,' he says, 'because when some of the parts are doubled - letting the violists take some of the theme, for example - it does make a difference.' Kovacevich also has strong views on how the opening should be conducted, whether in 'two' or 'six'. 'Six is absolutely terrible,' he says emphatically, 'and two gives you almost no control over flexibility. When I was a kid, I played the concerto under Sir Adrian Boult. He had a friend who had played it under Brahms's baton and although he didn't remember anything about the tempo, he did recall that Brahms beat the first, the third, the fourth and the sixth beats. It seems a logical solution to me, but whenever I suggest it to conductors, I'm greeted by a patronizing smile.' In Kovacevich's absence, I suggested this novel solution to Sir Charles Mackerras and the Brahms scholar Malcolm MacDonald, both of whom concurred that it might work well, MacDonald adding 'especially if you wish to underline the waltz-like nature of the music'.

The First Piano Concerto and the Violin Concerto share similar physical proportions, although the later work is richer in lyrical ideas. Frank Peter Zimmermann talks about a 'healthy D major world, with no shadow which is perfect for a stringed instrument that has to play against a huge orchestra'. He eulogizes about the return of the principal theme just after the first-movement cadenza (bar 527). 'I think that even the greatest pianists are jealous of that moment!' he says, 'It really is out of this world. And I still get a tingling in the back of my neck whenever I hear the oboe solo at the beginning of the second movement, even though it doesn't Involve the violin.' Zimmermann sees both the Violin Concerto and the later Double Concerto as predominantly 'symphonic', 'not the work of a composer who wants to write a solo concerto for a certain player - like Tchaikovsky did, or Lalo'. He recalls discussing the work with the late Nathan Milstein,' who admitted struggling with it, because it is not naturally written for the violin'. Violinist Hugh Bean makes a similar claim and yet it was Joseph Joachim, the pre-eminent virtuoso of the day, who worked with Brahms on the solo part. Zimmermann opts for the Joachim cadenza because 'Joachim wrote so many passages anyway, it seems somehow to be part of the concerto.'

Although respectful towards Joachim's place in musical history, Gidon Kremer cannot suppress the suspicion that 'being himself a violinist, Joachim maybe tried, here and there, to influence a composer for his own purposes of – let's put it very bluntly - of showing off, or for the convenience of the instrument, but not always in step with the composer's original intentions. So I'm not sure that Joachim's influence was altogether positive, although in many respects he understood Brahms and we can learn from the relationship just how important it is for a performer to be in touch - or enter into dialogue - with the composer.' Kremer confesses that, for his latest recording of the Violin Concerto, whenever or wherever he suspected Joachim's influence (specifically with regard to phrasing), he would discuss the passage in question with conductor Nikolaus Harnoncourt and try to understand 'Brahms's first impulse - which I would trust much more'. Zimmermann speculates on how the Violin Concerto might have emerged had, say, Sarasate taken Joachim's place: 'it probably would have sounded rather like Lalo, and I can tell you that a lot of violinists would have appreciated it that way! As it is, the Violin Concerto is a very "northern" piece that demands an emphatic style of playing.'

When it comes to the Double Concerto, a twilit essay rich in poetic allusions, Kremer notes that 'when Harnoncourt showed me the manuscript, he pointed out material that was significantly different from the published version. For example, the whole coda of the concerto differs from what is usually played. There are also some alterations in the middle episode of the finale.' All these variants are incorporated into the recording that Kremer and Lukas Hagen have made under Harnoncourt's direction.

Revisions and repeats

There are parallel textual revelations nestling among the pages of Brahms's four symphonies. For his Telarc cycle with the Scottish Chamber Orchestra, Sir Charles Mackerras included, as one of his CD fill-ups, a reconstruction of the First Symphony's original slow movement. 'It's quite extraordinary,' he claims. 'The material is the same, but it's in a different order, rather like a strange dream of music you already know.' Sir Charles has consulted the recollections of Fritz Steinbach, successor to Hans von Bülow as conductor of the Meiningen Orchestra and the man for whom Tovey wrote his famous analyses of Brahms's orchestral works. Steinbach left detailed descriptions of precisely how he thought Brahms's symphonies should be performed.' They involve a lot of changing tempos, flexibility and hesitating on the upbeat,' Sir Charles tells me, 'You know the sort of thing: drawing out the upbeat before you actually begin the melody [he hums the beginning of the big string melody from the last movement of the First Symphony, elongating the first bar]. Steinbach recommends that you start quite a few of Brahms's melodies in that way, with a hesitation, lingering a little, then accelerating slightly until you reach your main tempo four or five bars later. That of course was a Wagnerian principle, and I was quite surprised to find that it should be applied to Brahms as well.'

Steinbach advised against playing the symphonies' first-movement exposition repeats. Apparently, Brahms himself never played them, and Mackerras views the da capo marking as a sort of period formality, although he does actually include the repeats in his Telarc recordings. He finds the Second Symphony's exposition repeat the most congenial, a view shared by Christoph Eschenbach. 'Many people shy away from the Second's repeat because it makes the movement very long, ' he says, 'but the truth is that only then does the movement assume its proper shape and depth. If you note what happens during the development section: it is highly dramatic, not at all "pastoral", and Brahms's lead-back to the beginning is extremely elaborate.' But what about the First Symphony? Does he usually play the repeat there, too? 'No, I don't really believe in it,' he says candidly. 'In fact, I'II tell you the truth about that. I don't take it in concert and I only recorded it for the sake of completeness. I don't respond to the sudden switch from E flat minor to C minor; there's no proper transition and it doesn't work, musically. In any case, the movement's structure is so monolithic, it drives on, and Brahms knew it. I think he included the repeat sign because Beethoven had done so before him, and he felt the weight of tradition on his shoulders.'

When I spoke to him some five years ago, Bernard Haitink found the First's repeat ' boring', although he pledged to reconsider the issue for his then-forthcoming Boston sessions [as it happened, he didn't actually change his mind]. Riccardo Chailly, on the other hand, views the C minor work's first-movement repeat as crucial. He responds especially to that 'strange modulation. I feel that the repeat adds to the shape of the movement,' he tells me, 'balances it with the finale, proportion-wise, and I wouldn't dream of leaving it out.' Chailly views Brahms's repeats more as 'pre-developments' than mere carbon copies of what went before. 'If you omit the repeat, you reach the development-proper too early; bul if you repeat with a slightly different tone - not falling into the trap of mechanical repetition - then you strengthen the argument. It's rather like when you repeat a spoken phrase, altering the tone of your voice: each time you hear it, it's almost like opening your ears afresh.' Chailly considers the 'nonchalance' of older conductors, most of whom fail to repeat Brahms's first-movement expositions, 'a little bit unforgivable'.

As to chosen tempos, Stephen Kovacevich believes that because Brahms's orchestra was smaller than those we know today, his speeds would have been quicker. 'Mannheim and Vienna would probably have witnessed different tempos, even under the same conductor.'

Pillars of symphonic greatness

All the musicians to whom I spoke place considerable value on all of Brahms's orchestral output. Bernard Haitink views the two genial, multi-movement Serenades as 'bridges between the chamber and symphonic aspects of Brahms's musical personality'. He once told me that 'I often think of them as studies for the symphonies. Both works are difficult to perform, to communicate to an audience - but they are such lovely works. They are really musicians' pieces - the First in particular, with its long slow movement - and I think they are perfect examples of music that is better listened to and contemplated at home than in the concert hall. Still, I can never understand why more of my colleagues don't perform them.' 

The Tragic Overture is commonly valued for its disciplined structure and keen sense of narrative, the Academic Festival Overture for what Malcolm MacDonald terms its 'irrepressible sense of fun' and the Variations on a Theme of Haydn for (again I quote MacDonald) an 'instinctive mastery of thematic transformation'. However, the symphony cycle is central to any appreciation of Brahms's orchestral output. Although some 20 years in the making, the First Symphony went on to achieve immense popularity. Eschenbach makes special mention of that moment in the slow movement (between bars 38 and 45), 'where the C sharp minor oboe solo gives way to the clarinet solo which takes over and leads into A flat major - a fascinating moment for me, a bit like two people talking to each other about separate poetical subjects that somehow intermingle.' Mackerras stresses the importance of dividing violin desks to the left and right of the podium. He quotes the interplay of violin lines in the first movement of the Fourth (as from, say, bar 18), and in the First, the ivy-like intertwining of violin desks (bar 22) prior to the big horn tune.

The Second Symphony is more problematic in that a whole host of past commentators have misread it as a predominantly 'sunny' piece, Brahms's 'Pastoral', a sort of extended bland interlude set between the epic statements of the First and Third. And yet Malcolm MacDonald quotes a letter where Brahms advises Elisabeth von Herzogenberg that 'you only have to sit at the piano, put your small feet on the two pedals in turn, and strike the chord of F minor several times in succession, first in the treble, then in the bass (ff and pp), and you will gradually gain a vivid impression of my "latest"'. Eschenbach recalls the first movement's 'epilogue', 'music of the soul, that is after the horn solo (from, say, bar 491), and the epilogue itself reflects on where Brahms has taken you, and how deep you have travelled.'

When it comes to dynamics, Bernard Haitink advises caution. 'The first movement's second subject is marked piano dolce, and sometimes it is played forte cantabile, as if the conductor wants to use the music to show off his expressive powers. Then there's the beginning of the second movement, which is marked poco forte; one should not play it like Tchaikovsky - and I say that as someone who loves Tchaikovsky. It is a completely different sort of music, and yet there are conductors who play it like the last movement of the Pathetique Symphony.' Riccardo Chailly, on the other hand, perceives the Second as 'permeated by darkness and shadows, even from the beginning which, although it has an 'apparent' serenity, is set in a low register, oscillating between crescendo and diminuendo. And think of the beginning of the finale, where the opening theme is announced sotto voce, very sinister - and then, suddenly, you have this explosion, before you go back to the sinister mood.'

In the Third Symphony, Bernard Haitink underlines the value of taking Brahms at his word. 'The very opening - two massive pillars,' he says, his arms spread wide, 'major and minor chords, a motto for the entire piece and a gateway to what will follow. Some conductors impose a crescendo on them; but Brahms knew exactly when he wanted a crescendo and when he didn't - here he wrote a simple forte, which means he wanted two bold statements.' Mackerras characterizes the wistful poco allegretto third movement as 'sort of sad scherzo', whereas Chailly notes that the Andante second movement 'is all about agogics, the cello melody especially, the crescendo-diminuendo on the minim (in bar 4), how and where you should locate the so-called mini-climax in side that crescendo - should it be before the new downbeat, which is very often done, or should the climax be on the note?'

Chailly has rehearsed and reappraised this passage extensively, but even the equivocal vicissitudes of Brahms in 'melancholy mode' are child's play when viewed against the monumental interpretative challenges posed by Brahms's last and greatest symphony, the Fourth. Haitink, who views the work as 'an incredible piece', guards against what he terms 'a sort of grubby, grey colour' that some conductors bring to Brahms. He cites the opening of the Symphony's first movement. 'If one is not careful, it can sound very pedestrian. Yet it is a truly dramatic movement that follows a sure course, right from the opening statement of what I can only describe as a whole process of life. I can never escape the golden-brown complexion of this music.'

However, it is the finale, an epic passacaglia, that habitually gives rise to the most heated controversies amongst interpreters (not to mention critics). Haitink insists on unity of tempo which, he adds, 'is quite different from uniformity'. Chailly laments Brahms's lack of indications, especially regarding the flute solo at bar 97. 'I try not to change the tempo too much,' he explains, 'but it is important to prepare for that 3/2 passage, relax a little, so that by the time you arrive at the flute solo, you have already established the proper mood. It wouldn't do to keep the tempo too steady, because when we return to 'Tempo I' and those despairing vertical string chords (bar 132) cry out, the "shock" effect would be lost.'

Eschenbach advises against taking the finale too quickly. 'If you do, then by the time you reach the flute solo, you slip into Adagio or something, and that's no good at all.' Conversely, Stephen Kovacevich thinks that many conductors take the finale too slowly, and that the flute solo is usually played too quickly. Pondering the whole question of interpretation, he recalled playing under Pierre Monteux who, on this particular occasion, had been reminiscing about his days as a violist. The great conductor had actually taken part in a performance of a Brahms quartet in front of the composer. 'Monteux recalled how, after the performance, Brahms responded with a mere "very good", though he added one significant qualifying remark. "I must tell you that another quartet. came here last week," said Brahms, "and they played the same piece quite differently - but that too was very good!"'

This article originally appeared in the July 1997 issue of Gramophone. To find out more about subscribing to Gramophone, the world's leading classical music magazine, please visit: gramophone.co.uk/subscribe

Brahms – the complete guide

Part 2: chamber and instrumental music

In the second of a three-part series, Michael Quinn talks about Brahms's chamber and instrumental music with leading artists... Read the article

Part 3: vocal music

In the third of a three-part series, Michael Quinn talks about Brahms's vocal music with Barbara Bonney, Marcus Creed, José van Dam, Richard Hickox and Graham Johnson... Read the article