The voice was at the centre of Brahms's music. At its beginning and end, too. Of the first seven completed works he allowed into the public domain, three, significantly, were sets of songs. His last composition, the Vier ernste Gesänge ('Four Serious Songs'), was also to be for voice. The preoccupation with song was ever present, always crucial, occasionally defining; the voice perpetually exploited and explored in more than 200 songs for solo voice, duet and solo quartet, and in a myriad of works for accompanied and a cappella choir, the volume of which it is impossible to do justice to in a single article. Suffice to say that the sense of the sung voice was instinctive and integral to Brahms's own sense of himself as both man and musician. 'The ideal', he once remarked, 'is the folk-song.' He might well have added, 'and the voice that creates it, the sublime'. More than the symphonies and the chamber music, it is the vocal music that most explicitly reveals the head and heart of a composer whose true, full and complex identity has been obscured by time and reputation. Where the autobiographical subtext of his instrumental music was frequently disguised, habitually denied, the provocation and purpose of the vocal music is often gloriously, unashamedly, immediately apparent. It is a repertoire that is honest, direct and intellectually unfettered.

Unequal, too, it has to be said, in terms of quality. Yet, in a curious way, it is those obvious occasional weaknesses that give the whole its strength, make it so tantalizing, so endearing, so much more truthful and reliable an expression of the man who composed them. Punctuating the life and career as they do, the compositions for voice provide a revealing series of thumbnail portraits of Brahms that belie the lingering image of him as being ill-equipped or disinclined to deal with the complex reality of his own emotional responses to the people and events that impinged upon his life. Certainly, for pianist Graham Johnson, the Lieder provide for an altogether fresh and vital perspective on the composer.

'In Brahms's case, there is a distinct feeling of the song texts being chosen because they mirror his mood of the moment. He looked to poetry as a type of echo of his own mental state. An analysis of the poems that he set very often finds an emphasis on thwarted love; on love that always seems in the final analysis to never triumphantly find its end but is always cut off at the last moment. Like the man in Von ewiger Liebe who thinks himself not good enough for the girl he adores, or like In Waldeseinsamkeit, which is about a man in the forest with his loved one and his jittery hands long to do something with her but in the end don't.'

For Johnson, the resonance of particular narratives within the Lieder clearly has little to do, per se, with the pictures of poetic yearning they paint, but in what those pictures represent in terms of the moment in which Brahms set them. In the Lieder, the autobiographical context often has greater currency because it is that much closer to the surface, that much more apparent, that much more honest and open. Their emotional utility to Brahms takes on a far more crucial, far more intimate role than does the bulk of the orchestral and instrumental repertoire. 'They're about empathy, about expecting people to read his mind, about being unable to articulate his feelings, unable to ask for his needs to be met. There is an inscrutable quality where he goes to his songs and writes about these preoccupations and others - a very extraordinary preoccupation in songs about mothers and daughters. He seems always to be moved by these colloquies, some of them very dark, as in Mädchenfluch. The very first song we know - Liebestreu - inaugurated this; a conversation between a mother and daughter in which the younger woman declares, "I will love him despite what you say, mother!".'

The songs describe Brahms's tentative emotional life: his passion for Agathe von Siebold, his infatuation with Hermine Spies, his yearning for Elisabeth von Herzogenberg, his adoration of Rosa Girzick, his devotion to Julie Schumann and, above all else, his deep and abiding love for Clara Schumann. Though the texts he chose to illustrate his feelings are of often questionable literary merit, the eloquently candid music they inspired more than compensates. Recognizing the diminished importance of the word in Brahms's Lied, Eric Sams noted in his erudite and seminal survey of the composer's songs (BBC Music Guide: 1972) that, 'Other composers set the words, Brahms uses them to set the tone, or the scene, of his own experience.'

The early songs may not be anchored in actual, lived experience but they clearly connect with the rich emotional imagination of the young Brahms. Liebestreu (Op 3 No 1), with its tragic E flat minor key signature, the lyrically tender Wie die Wolke (Op 6 No 5), the unapologetically melodramatic Treue Liebe (Op 7 No 1) and the troubled rather than tortured Nachtigallen schwingen (Op 6 No 6) all point to the burgeoning sense of emotional dislocation that was to affiict him throughout his life. Each new song represented a further step along the journey that was, in Graham Johnson's opinion, 'to take the Lied back to Vienna for the first time since Schubert's death'. Self-evident in the two other sets of songs that appeared in the 1850s the Eight Songs and Romances, Op 14, and the Five Poems, Op 19 - were the many influences acting upon the young composer: Schubert and Schumann of course, but also Mendelssohn and Robert Franz and the highly coloured idioms of traditional folk-songs and tunes.

However, derivative though Brahms's early songs may have been in some respects, there were original elements to be found in them, and innovations just around the corner. Graham Johnson again: 'What he did which Schubert never did, was to show an interest in folk-song. That is certainly the thing that separates him from almost all the other great Lieder composers. Schumann had shown the beginnings of an interest, Wolf wasn't interested at all, and Schubert had had the talent to write his own folk-songs if he so wished. Brahms developed a quasi-scholastic but largely romanticized attitude to folk-song. He made his own liberal adjustments to Zuccalmaglio's 'traditional' songs (Op 14 Nos 6-8] and in his own folk-song sets [most conspicuously the Zigeunerlieder] he actually composed original things and passed them off as traditional. It wasn't yet the age of Kodály and Bartók!' But even in early Brahms there is more to be found than merely musical sleight of hand.

'The big thing he did which was different to Schubert was that he experimented, in his first youthful cycle - his only really major cycle, the Schöne Magelone, Op 33 - with the Romance form, producing a type of strophic song with very highly developed piano part and a huge narrative sweep. I don't see it as having anything in common at all with the Schubert of Die schöne Müllerin and Winterreise. Each of the songs is enormous and there are many word repeats. There's nothing quite like them in Schubert.'

Nor anything quite like them in the rest of Brahms. For all 'the experimental energy, brilliance and confidence' Johnson finds in the Magelone songs, he also recognizes that they set themselves apart by virtue (if such it is) of their 'sheer difficulty and grandiosity'. With hindsight, the dark-hued seriousness that characterizes them can be seen as quintessentially Brahmsian. When published in 1865, the first two books of the cycle brought some considerable success to Brahms. Today they are considered a partial failure; overly ambitious and insufficiently discriminating in their choice of suitable text for setting, the latter criticism one which Johnson forcefully refutes. 'There's absolutely nothing wrong with the poetry of Ludwig Tieck! Schubert, Schumann and Wolf may all have ignored him but he was a very important figure indeed in German literary history, so at least Brahms can be applauded for filling a very big gap with Die schöne Magelone.'

Johnson also comes to Brahms's defence when the question of his predilection for setting poetry of a sub-standard quality is raised. 'The literary cupboard of Brahms shows an immensely well-read man with perhaps leanings towards certain tastes in literature which are no longer considered fashionably central today, but which were at the time. He went back to the eighteenth century and set Holty and expanded on Uhland where Schumann had only set Frühlingsglaube. He set a certain amount of Goethe, six very fine songs by Heine and two or three very fine Ruckert settings. There are two settings by a poet who is otherwise ignored, Detlef von Liliencron - his settings of Auf dem Kirchhofe and Maienkätzchen are absolute jewels. The Op 32 songs offer extremely fine settings of an important German poet, Platen, only set previously by Schubert and ignored by the rest of the Lied composers. And he set Morike long before Wolf began to set him.'

Johnson's complementary roll-call continues with, 'people like Hoffman von Fallersleben, no bad poet; a certain amount of Eichendorff; Emanuel von Geibel (whom Wolf was to set later in the Spanisches Liederbuch); Daumer, another poet important at the time whose Liebeslieder waltzes, based on Persian originals, are very erotic; not to mention Candidus, reasonably famous in his own time, Keller and Schack.'

Schack, incidentally, was also to be later set by Richard Strauss, the mention of whose name prompts Johnson to observe en passant, that, 'in terms of Lieder production, Strauss's literary taste is much more open to question than that of Brahms.'

The baritone Jose van Dam, having recently recorded the Op 32 Lieder and the Vier ernste Gesänge (Forlane, 11/96), also happily volunteers himself as a witness for the defence when Brahms's literary acumen is called into question. 'You have the same thing in Pelléas et Mélisande, where the text is simple and not so rich but with Debussy's music it acquires an extra dimension, one that gives it substance and sense. With Brahms it's the same thing: the text is sustained by the music.'

The music also provides cohesion and coherence, argues van Dam, where superficially none might appear to exist. 'It's difficult to find the feel and the focus of Op 32 because it's not a cycle, it's a suite. In truth, the words tell less than half the story, it's the music that provides the detail and the depth of understanding. 'And it's in the music that the key to unlocking and deciphering the cryptic emotional mass that underpins the vast majority of Brahms's Lieder is to be found. To Graham Johnson the songs paint the portrait of 'a wounded individual who wore his tribulations, as Auden once said, like a rose. There seems to have been some major heart-set at the very centre of his psyche that we will probably never get to the bottom of.'

It's that enigma, that incessant emotional dyspepsia deep within the songs that has made Brahms so unpalatable for so many for most of the present century. Françoise Sagan's inquiry 'Aimez-vous Brahms?' may seem less provocative, less controversial now, but there remains a lingering, overly cautious suspicion, a fact Graham Johnson is keenly aware of.

'He's not the easiest composer to produce an all-evening programme of. All-Wolf and all-Schubert are possible, but all-Brahms can leave a certain feeling of depression in people's minds. This is because, like all great composers, he writes himself into his music and because he was a very lonely, very unhappy, depressed person, that communicates itself to the audience and gets into their bloodstream. I would say he is the most controversial song composer I know in the sense that those who love the music love it very much and those who don't positively hate it.'

Soprano Barbara Bonney , who like van Dam finds herself increasingly turning away from opera, convinced that 'Lieder is my future', takes a refreshingly feisty stand against 'the traditional assumption of Brahms as dark and "woofy" with all the keys in flat'. And a surprisingly fruity one, too. 'Every composer has their own flavour and for me Brahms is like a banana milkshake: you treat yourself with it, you know it's good for you, it tastes delicious and it's filling.' (By comparison, she offers , 'Mozart is more like lemonade: astringent, clean cut, clear and it makes your teeth tingle and makes you smile.' Appetizing stuff.) The evening before we spoke, Bonney had sung a trio of Brahms songs for the first time in public and had found in them 'lots of amazing melodies' and an array of 'colours, lights, atmospheres and a real, almost Grieg-like lyricism. He has a very folky, natural, almost Mahlerian approach and his simple, heartfelt sincerity is not too dissimilar to Mozart.'

Her enthusiasm can't be explained away as just the fervour of the newly converted. 'Von ewiger Liebe I've taught a lot and delved into quite deeply for some time, and I find a real sweetness in it that some would deny Brahms. Every time I've worked on it with students they've screamed their socks off with it and that made me go back to listen to Elly Ameling singing it, and she does it so beautifully, never overdone, never operatic, just completely natural. People do tend to forget that Brahms does have this very sweet side to him.'

Bonney has much to look forward to as her investigations of the Lieder progress. Brahms enjoyed and exploited the soprano voice to the full. Though the tessitura of most of the soprano songs is, she admits, 'high and very much Schubert an, they remain incredibly singable'. On the whole. Already she is finding exceptions to her own rule. 'When you've sung a whole evening of the Liebeslieder and the Neue Liebeslieder it's like singing Brünnhilde. They're damn hard to sing, especially for soprano. You 've got three people underneath you vying for attention and you have to carry the melody practically all the time. After a while it kills you!'

The daunting scale and scope of the two Liebeslieder collections 33 songs for SATB and four-hand piano (Opp 52 and 65) - may aspire to the condition of a song-cycle, but they are atypical and do so by default rather than design, Brahms preferring to work on the considerably more intimate level of private expressions of grief than public displays of grieving suggested by the extended song-cycle. If his own circumstances coloured that preference, so too did his reluctance to step deliberately into Schubert's shadow.

'He probably felt that with Die schöne Müllerin and Winterreise, and perhaps with Schumann, who wrote almost everything in cycles - Liederkreis, Frauenliebe und -leben, Dichterliebe - that the song-cycle had been done,' Graham Johnson offers in explanation of the lack of other Brahms cycles. 'Of all composers, he was the greatest musicologist of his time and so was very aware of what had gone before. He knew the Schubert songs and never went near his territory. The few times he set the same poems as Schubert - Die Liebende schreibt (Op 47 No 5), Minnelied (Op 71 No 5) - Schubert had not set them so wonderfully, so perhaps he thought he had something to say there? Die Mainacht (Op 43 No 2), which Schubert had set extremely simply and strophically, Brahms set more expansively, but he never went near any poems that he thought Schubert had set definitively. He was much less audacious than Wolf in that respect.'

That Brahms chose the Lied form as the arena in which to take his leave from composition has a significance that cannot be overstated. When he published the 49 Deutsche Volkslieder in 1894, he alerted Clara to the fact that a theme from the last of the songs, Verstohlen geht der Mond auf, also occurred in the slow movement of his Op 1 Piano Sonata and that the repetition was no mere coincidence. 'It was in fact designed to say something,' he wrote in a letter to Clara, 'intended to represent the snake that bites its own tail, and thus to express symbolically the idea that the tale is over, the circle completed.' Almost completed, but not quite. On May 7, 1896, his 63rd birthday, he revealed rather modestly that, 'I had amused myself by writing a few little songs' and produced, like a rabbit out of a conjurer's hat, the Vier ernste Gesänge, his first original songs in a decade and his penultimate opus. Mimicking the form of a baroque solo cantata, they articulate, for Graham Johnson, 'huge, philosophical reflections on the nature of life and death', while to José van Dam they assume the form of an extended prayer. 'Brahms would not have described himself as religious, but he was none the less convinced of God deep within himself and you can feel that in the Four Serious Songs. When I sing them I feel as if I am a preacher delivering a sermon in church not from God but from deep within the human heart. Perhaps that is the same thing? Perhaps that is what Brahms would have wanted us to believe?'

Dvořák, whom Brahms had helped and encouraged as a young composer, might have found van Dam's speculations simply that: pure conjecture. 'Such a man, such a fine soul,' he rather despairingly remarked of Brahms in 1896, 'yet he doesn't believe in anything, he doesn't believe in anything.'

The only comparable analogue in Brahms is the mighty Ein deutsches Requiem. Published in 1868 some 11 years after it was first begun, it also employed baroque devices and married, too, that blend of the sacred and the secular peculiar to Brahms which exalted love, and elevated it, in his eyes at least, to be the highest and holiest of virtues.

Conductor Richard Hickox, who has recorded the work on Chandos (1/92), readily concurs with Grove's assertion that the Requiem is 'not only Brahms's greatest vocal work, but also the central work of his career'. ' It is', he agrees, 'the centre of his soul and his psyche; an intensely personal, deeply moving setting and emotionally wrenching. The art of conducting a great performance of the Requiem is to find a connection between the seven movements. I hate performances when there are huge gaps. I only ever allow the choir to sit once and that's for the fifth movement when they accompany the soprano solo. The rest I get them to sing while standing. I did it recently with Bryn Terfel and the Cleveland Orchestra and two members of the chorus fainted. It's quite a test of stamina!'

And of imagination, too. To José van Dam, for whom 'a Requiem is the most humbling and important music any composer will ever write', the joy of the work is that it is neither an exercise in pulpit-thumping posturizing nor pummelling hair-shirt punishment. The message is one of hope, of life. Thus, Brahms's juxtaposition of baritone and soprano solo with chorus is significant. 'For Brahms, the chorus represents our shared, universal experience, the baritone is the individual experience and the soprano is the heaven-sent angel. For me, this is a real Requiem with an honest spirituality informing it.' Surprising, then, to recall George Bernard Shaw's vehement antipathy to the work. 'There are some sacrifices which should not be demanded twice from any man,' he once opined in a rare moment where Shavian wit and wisdom went their separate ways, 'and one of them is listening to Brahms's Requiem.'

Richard Hickox adroitly alights on the work's structural strengths to explain its enduring appeal. 'It's a timeless piece. It has a lot to do with Handel - a fairly baroque manner and all those archaic fugues - yet it also has a very modern feel to it. The harmonic language is definitely of its time and the overall view has an intense romanticism about it. One incredible thing about Brahms's choral music, what makes it so special, is the spacing of the chords. If you think of the second movement where it starts in octaves but without the sopranos so you get the darkness through the prominence of the bass, or in the fifth movement when the chords are spaced very closely together and it becomes transparent, the actual texture of the choral writing is immensely effective.'

Brahms had combined vocal and instrumental music previous to the Requiem - the Op 12 Ave Maria, the Op 13 Begräbnisgesang, the Op 27 setting of the thirteenth Psalm for the unlikely combination of three female voices, organ, piano and strings, and the Op 30 Geistliches Lied had all preceded it - but it was the first work to combine solo voices, mixed chorus and full orchestra. Its success was eventually to be as great as its conception and execution was ambitious. And just as no other Lied composition he produced compared with the Vier ernste Gesänge, so, too, the Requiem is without parallel or compare in the rest of Brahms's choral works. Though the Schicksalslied - acclaimed by Hickox as 'an absolute masterpiece in ten minutes' - is occasionally referred to as the 'little Requiem' and its E flat major key is one of the central building blocks of the larger work, in truth it more closely resembles the work that immediately preceded it, the Alto Rhapsody wedding present for Julie Schumann, and the later Goethe setting, Gesang der Parzen. Again, Hickox highlights their shared approach to chord spacing as evidence of a forward-moving musical intelligence some would deny Brahms. 'That marvellous, magical moment in the Alto Rhapsody when the men's chorus come in, in that last C major section, and the unique glow throughout the Op 89! Texture is the thing for me which distinguishes Brahms and where he moved vocal writirg on.'

There were other large -scale choral pieces deserving of mention, each of varying degrees of success. The Triumphlied, Op 59, his jingoistic paean of praise to Wilhelmine prowess; the classically inspired Nänie, Op 82, with its Brahmsian lyric par excellence - 'Even the beautiful die. Yet to be a song of lament in the mouth of a friend is glorious' - and Rinaldo, the flawed dramatic cantata which employed the same instrumentation as the Alto Rhapsody and which made first use of a male chorus, are the three that most demand attention. But it's a measure of the neglect that Brahms has endured and continues to suffer that even deep into the present century it was possible for all of the interviewees for this article to confess to variations on the theme of, as one actually said, 'Never heard it; never played it;' for all of these works and considerably more.

One other area remains to be considered, however briefly: the 13 motets, 46 a cappella songs and 20 canons Brahms wrote in the 40-year interval between the Op 22 Marienlieder of 1859 and the Three Motets, Op 110, of 1889. Unlikely as it may seem in passing, the motets offer one of the most stirring examples of the quiet revolution practised by Brahms throughout his compositional life. Marcus Creed (who recorded the Opp 29, 74, 109 and 110 motets with the RIAS Chamber Choir - Harmonia Mundi, 5/96) argues that in the motets Brahms fused baroque and nineteenth-century devices together with a seamlessness none of his contemporaries could possibly have matched. 'The sacred motets, because of their contrapuntal nature, are harmonically more progressive than many of the folk-song-based secular motets. And being contrapuntally strict they allowed for developments and directions to emerge which were quite individual. In the later motets there's much more reaction to the texts he chose, more freedom in the way he sets them. Ich aber bin elend (Op 110 No 1) is musically much more expressive than it would have been years earlier; the way he nearly uses plainsong in there, combining block chords with what is almost a chanting device while at the same time, in the same movement, using a free fugal system, offers clear evidence of a musical mind that was very far from ossifying.'

Richard Hickox fully supports Creed's creed. 'They did influence choral writing from the moment they were written. From the unaccompanied motets you can make a direct link going up to Schoenberg's Friede auf Erden. ' So why the comparative neglect now? 'There's always an assumption that things fall into neglect for very good reasons, but I'm not sure that's wholly true in relation to Brahms's vocal music, especially if the choral pieces we are familiar with are any indicator of what he can do with the voice. Much as I love the Schubert Masses (and he's obviously the master of the song), I don't think he wrote for choral forces nearly as effectively as Brahms. We've had marvellous cycles of the symphonies and the Requiem along with them, but we haven't heard nearly enough of the rare repertoire.'

A humbling complaint. The problem, however, rests as much in Brahms's nonchalance as in our own neglect; specifically in his scant disregard for his place in history and in his refusal to bend himself or his music to the pressure of peers or elders or to conform to their salutations or their sarcasm. Brahms carried neither torch nor baton to pass on. 'His style was so individual', Marcus Creed agrees, 'that he didn't automatically begin with the presumption of developing what he had inherited. Nor did it make it possible for later composers to pick up where he had left off with any ease.'

In a very real sense, then, the nineteenth century didn't end when Brahms finally set down his pen in 1897, but rather at the moment he first picked it up some 50 or so years earlier. In deliberately refusing the entreaties of Liszt and Wagner's insurgent Weimar School, and in just as consciously stepping away from the inheritances of Beethoven and Schubert, Brahms stepped into territory which was largely uncharted and in which, for most of his life, he was the sole dweller. But if Brahms didn't belong in his own time, and if our present age has found it difficult to accommodate the full magnitude of his achievements, what does the future hold for him? Are we to add to Anton Rubinstein's much quoted catalogue of his apparent failings - 'Insufficiently at ease for the salon; insufficiently fiery for the concert-hall; not rough enough for life in the country; for life in the city not sophisticated enough' - the accusation that for posterity he was - is - insufficient in some other and equally spurious but unforgiving way?

The last word is left to Graham Johnson who, while recognizing 'something Edwardian and doomed about Brahms's Zeitgeist' (a diagnosis the composer himself would no doubt have resignedly agreed with), forcefully insists that, 'We have to rediscover Brahms again, take him out of mothballs, appreciate that we're no longer threatened by all those fusty morals and assumptions that he lived through and suffered through. Younger people who have escaped the leaden legacy with which we've burdened Brahms are unselfconsciously falling in love with his music again. They 're hearing it for the first time and for the first time are hearing what Brahms wrote about himself and not what others wrote about him. That's the way it ought to be. Perhaps that's the way it will be?'

This article originally appeared in the September 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 1: orchestral music

In the first of a three-part series, Rob Cowan talks leading conductors and musicians about the special appeal of Brahms's orchestral works... Read the article

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