The Seven Doors – on Bartók’s Bluebeard’s Castle

Béla Bartók, who composed his only opera, Bluebeard’s Castle, in 1911. Public domain · Wikimedia Commons

I. Two Voices in the Dark

There is one opera that I carry with me in a particular way. When I have the score in front of me and am about to begin work on Bartók's A kékszakállú herceg vára – Bluebeard's Castle – I always feel something in my body before I feel it in my head. The work opens in darkness. No overture, no fanfare, nothing that lifts the audience into the theatre. Only a spoken prologue by an old bard, and then deep strings, deep F sharp minor; this is a space that has not seen daylight for a long time.

There is something about this opening that affects me every time. It is as if the work begins somewhere where there is no air, and demands that I and the orchestra breathe differently in order to begin at all. I shall return to this, because it is in fact the essay's core for me: what kind of leadership the work requires of the conductor, and how different this is from what I am used to from Verdi or from the belcanto tradition.

It is an opera with only two singing characters. Bluebeard, a bass or bass-baritone. Judith, his new wife, mezzo-soprano. No other voices. No chorus. About sixty minutes, in one act, without a break. The whole work takes place in the same room – the main hall of Bluebeard's castle, where seven doors stand closed in the wall – and the whole action is that Judith demands to open them, one after another.

I want to write about the work as I know it here and now. This is not an academic analysis. It is a conductor's attempt to understand something he has worked with over time, and which continues to give up new layers the longer he stays with it.

II. A Work Without Precedent

Bartók composed the work in 1911, when he was thirty years old. He had by then spent about six years working with Zoltán Kodály collecting Hungarian folk music – journeys out into the countryside with a phonograph, transcriptions of songs no one had written down before, a systematic effort to reconstruct the authentic Hungarian musical language from the sources themselves. It was a political-aesthetic project. Bartók held that the picture the world had of "Hungarian music" – Liszt's rhapsodies, Brahms's Hungarian Dances, the café music of Budapest – was a Viennese cliché about Hungary, not Hungary's own music. The real Hungarian music lay elsewhere, and he had to bring it out, bar by bar, from the peasants who sang it.

Into this work came a libretto. Béla Balázs, a young Hungarian poet and philosopher in Bartók's circle, wrote a short text based on the French Bluebeard legend, but turned it deeply inward. The text was originally dedicated to Kodály, but Kodály never composed it. Bartók took it up in 1911.

What Balázs had done was to rewrite the myth from fairy tale to psychology. The classic Bluebeard story is about a man who has killed his previous wives and intends to kill the new one as well, and about the woman who is rescued at the last moment. Balázs made it into something quite different. His Bluebeard is not a murderer in the conventional sense; he is a man with an inner world so vast and dark that the ordinary world cannot find room in him. Judith is not a wife who must escape; she is a woman who wants to know the man she has married, every room, or chamber, in his soul, and she is willing to pay what the knowing will cost.

The seven doors are then not physical doors. They are her way into him. What she finds behind each door is something about him – his cruelty, his power, his wealth, his beauty, his kingdom, his sorrow, his previous women. It is an opera that does not let an outer action unfold. The whole action is a structured uncovering of an inner territory.

Bartók found in this material something that suited his development as a composer. He had followed Debussy's Pelléas et Mélisande closely after the work had reached Vienna and Budapest – the symbolist operatic form, the finely tuned recitative, the atmospheric orchestration. Bluebeard's Castle clearly stands in that line of inheritance. But Bartók does something Debussy did not do: he combines the symbolist operatic form with Hungarian folk intervals – especially the characteristic pentatonicism he had found in the old Magyar songs – and with a modernist orchestration that neither Debussy nor any contemporary composer had carried so far. The result is a work that is neither pure Debussy nor folk music, but something entirely its own.

The first score was submitted to Hungarian operatic competitions in 1911 and 1912 – both the Erkel Prize and the Rózsavölgyi Prize – and was rejected both times as unperformable. Bartók revised the work in 1912 and 1917, but it was not premiered until May 1918, at the Royal Hungarian Opera House in Budapest. Even then the opera received a mixed reception. It took several decades before the work found its place in the international repertoire.

Today it is recognised as one of the most important modernist operas of the first half of the twentieth century, and it is the only opera Bartók ever completed.

III. Who Leads Whom

My reading of the work – and it is a reading, not a truth – begins with a question about who leads whom in this opera. It is a question I have carried through all the work, and it is really from there that everything else opens for me.

I come myself from a tradition in which the conductor often has to learn to follow. In Italian belcanto – Bellini, Donizetti, early Verdi – the conductor's task is largely to carry the soloist. The singer breathes, phrases, expands a line, inserts a cadenza where the score has not expressly asked for it. The conductor listens, gives, lets go. The orchestra is more backdrop than architecture. I mean this without any diminishment – it is an art form in itself to be able to follow a great singer in such a way that the music breathes with him or her, and it can be very demanding in its own way. But the position is nevertheless fundamentally one of following. The singer is the one who carries the story through time; the conductor is the one who makes sure the orchestra is there when she arrives where she is going.

In Bluebeard's Castle this is, in my view, reversed. It took me a long time really to understand this. I had worked with the piece in the way I knew – as a conversation between two singers whom the orchestra should support – and something did not come out right. The lines did not sit where they should. The transitions from door to door were slack. Judith's questions did not become dangerous enough, and Bluebeard's resistance did not become heavy enough. Something was missing.

What I came to understand is that in this opera the orchestra is not backdrop. It is the building. The whole castle – with its seven rooms, with its acoustic worlds of sound, with its darkness and its sudden opening towards light – is the orchestra. The singers live inside that building. They move through it; they climb its stairs; they open the doors. But the building belongs to the orchestra. And that means the conductor and the orchestra have to lead. We have to be so steady, so architecturally clear, so calm in our own pulse, that the singers can float on top of what we do, almost like two figures on deep water where the current beneath is what carries them forward.

This is the specific discovery I have had with this work, and it has changed how I stand in the rehearsal room. I cannot follow. I have to lead. Not in an authoritarian way, not by drowning out the singers – but by building an acoustic space so firm that they have something to stand on.

I often think of Bernard Haitink's recording of the work. Just this is so clear in him: he never lets go of the orchestra's underlying pulse. He does not chase dramatic effects. He builds the castle, bar by bar, and lets the singers find their places inside what he builds. The result is a recording in which each door truly opens as a new room, precisely because the main room never releases its hold. For me it is a model of how this work wants to be.

This leadership is not without an ethical dimension for me. There is a certain irony in that I as a conductor have to lead so firmly in a work that is itself about what happens when someone tries to penetrate completely into another human being. Judith demands to see everything. She cannot bear that Bluebeard conceals anything. She wants every room, every tear, every previous woman in the light. She confuses knowing with possession – and this is what in the end turns her into yet another image inside the castle, no longer a living counterpart.

There is a philosophy I find it useful to have in peripheral vision here. Martin Buber distinguishes between two fundamental attitudes a human being can have towards the world: Thou and It. In the Thou-attitude one meets the other as an other, without grasping, without trying to explain completely. In the It-attitude one turns the other into an object one can have an overview of. Judith's tragedy, as I see it, is that her love is transformed along the way from Thou to It. She wants to know Bluebeard so completely that he can no longer be an other, only an open map of himself. And at the moment she has been given everything to see, she has lost the one she wanted to meet.

This is not a philosophy Bartók or Balázs had on the table. But it is a reading I feel I can have in the back of my mind without its forcing itself on the interpretation, and it helps me understand what kind of intensity the work must carry. It is not a gruesome fairy tale. It is a meditation on what it is to want to own the one you love, and what that costs.

For the conductor this gives a concrete task. Each door must truly open a new room – there must be a dramatic difference in sonority, in tempo and pulse, in orchestral density between door and door – and at the same time the main room, the castle's own dark F sharp minor, must never be entirely gone. We must be able to hear, even in the fifth door's radiant C major, that we are inside the same building. If that thread breaks, the opera becomes a succession of tableaux. If it is held, it becomes a journey.

IV. Four Doors That Open

Four moments in the work make this clear for me.

The first is the prologue and the opening of the castle itself. Bartók begins, as I mentioned, with a spoken prologue by an old bard – a short text that invites the audience into the story and asks, in Balázs's own words, where the stage is: outside us or inside us. It is a gesture that frames the whole work. This is to be an opera about something within.

When the music then begins, it is with low strings and woodwinds in dark F sharp minor. Bartók establishes from the first bar the castle's own sound – closed, damp, without light. It is this tonality he returns to throughout the whole work; it is the castle's own voice, an acoustic space from which the seven doors open out and to which they close back.

For me as a conductor this is the most important moment in the whole work, after the ending. If the castle is not established as a sufficiently compressed, closed space in the first minutes, none of the later door-openings will have any meaning. It is by contrast with the castle's darkness that each door's world receives its significance. I must therefore keep the opening low, dense and unusually concentrated. The orchestra must be so quiet and so focused that the audience feels they have been let into a place where they ought not to be. If we do not achieve that here, we lose much of what the work can give.

It is also here that my leadership begins. Already before the first note I must have fixed the tempo so precisely that I can carry it without having to negotiate it with anyone. This is not music for negotiation. This is music where a particular underlying pulse must be felt through all the sixty minutes that follow. If I do not myself have the pulse in my body before I raise the baton, I cannot ask the orchestra to have it.

The second is the first door. Judith insists on opening it; Bluebeard protests, but gives way. She opens. Behind the door is the torture chamber. The blood is already there – the first red colour that will recur throughout the whole work, the blood-motif, a sharp chromatic element that Bartók sets against the castle's dark modality.

This first motif is important because it establishes the dramatic form the work will follow. Each door will have its own sonic space, its own harmonic world, its own emotional temperature. And each room will have blood in it somewhere – it is the red thread that binds the castle's rooms together. Judith sees it the first time and reacts with a shudder, but she does not stop. She demands to open the next ones.

It is here, in the transition from the first to the second door, that I as a conductor first have my own leadership tested. The transitions in this work are always the conductor's task to carry. The singers have their lines, but it is the orchestra that must move us from one room to the next – often through short, concentrated orchestral passages where the tempo and the sonority must change while the underlying pulse is held. These are the places where good work with the castle in the opening pays off. If the underlying pulse is in place, I can safely let the sonority and the tempo shift within it. If the underlying pulse drifts, everything drifts.

The third is the fifth door. This is the great moment, the work's visual and aural climax. Judith has already opened four doors – torture chamber, armoury, treasury, garden – and she has seen blood in each of them. When she opens the fifth, she looks out over Bluebeard's kingdom, from his highest tower. It stretches as far as the eye can see.

Bartók answers with one of the most spectacular orchestral revelations in twentieth-century music. Full orchestra, with organ which Bartók expressly specifies, in a radiant C major. The woodwinds blaze. The brass breaks through. It is a moment where the music almost burns through the score.

Judith does not react with joy. She reacts with restraint. "Beautiful – but I see blood on the clouds," she sings. Even at Bluebeard's greatest self-revelation – this is my kingdom, this is me, all this I possess – she cannot overlook that blood is there too. Her insight grows sharper, not warmer. She demands to see more, deeper.

This is the most technically difficult moment in the work for me as a conductor. The orchestra must give everything, and at the same time, already as I stand there at the golden C-major flowering, I must know that I am to lead us back down from it, into Judith's low, dark response. The transition from full orchestra to muted vocal line over a few bars is one of the most exposed dynamic shifts in the whole operatic literature. The dynamic shaping must be almost surgical, and the tempo must be held, so that Judith's objection does not simply become a muted postlude but truly a counter-answer that has room to stand. Here I have to hold the orchestra in rein like a rider holding back a horse – firmly, but without tightening.

The fourth is the seventh door. After Judith has seen the blood in all the previous chambers – the torture, the weapons, the treasure, the garden, the kingdom, the lake of tears behind the sixth door – she insists on the last. Bluebeard refuses. She demands. He opens.

What she finds is not what the Bluebeard legend promises. It is not a pyramid of slain wives. It is three women, clothed in jewels and crowns, alive, but unable to move or speak. They are Bluebeard's three previous women – the woman he met in the morning, the woman he met at midday, the woman he met in the evening. Judith becomes his fourth: the woman he met at night. She is crowned, robed, and goes into the darkness together with them. Bluebeard stands alone. The last door closes. The castle's darkness returns.

Bartók composes this scene as a long, slow orchestral descent. It is not a dramatic climax – the climax was the fifth door. It is a reconciliation with the inevitable. The previous wives do not sing; they are statues, or almost non-voices. Judith's last melodic line is resigned, not fought for. The castle closes around her.

For me as a conductor this moment is what I build the whole work towards. When I stand before the orchestra in the opening and establish the castle's dark F sharp minor, it is because I know we are to end there again after sixty minutes that have broken it open six times and each time returned. The ending works only if the opening works. And everything I have done as a leader through the work – the steady underlying pulse, the architectural transitions, the concentrated door-openings – is so that this ending receives its full weight. The last door that closes must give the audience the feeling that something irreparable has happened, and that we were all of us inside it.

The ending gives no release. Bluebeard has revealed himself completely, and the price was that Judith is gone for him now – not physically killed, but transformed into a memory, an archetype. He is alone with seven doors that are all now closed again, and with four women who are his inside the castle's darkness.

V. Coda

Bluebeard's Castle is an opera that resembles no other. Only two characters, one set, about sixty minutes, no chorus, no outer action. It resembles more a long song cycle for two voices and orchestra than a conventional opera. And yet: within its seven doors it holds one of the deepest meditations on the relationship between two human beings that music theatre has produced.

My reading is that the work is about the attempt to know another human being completely – and about how that attempt always ends, not in happiness or murder or escape, but in a form of mutual isolation. Judith demands to see everything; Bluebeard shows her everything; and when everything has been seen, they have lost each other in the same movement by which they have found each other. It is perhaps the most honest opera about love and knowing that has been written, and that honesty is dark.

For me as a conductor the work is one of operatic history's most demanding tasks, not in a conventional sense, but in an architectural one: to build seven distinct sonic worlds inside a closed whole, and to lead the audience into and out of them without losing the greater arc. I have never worked with any other piece where the conductor's own leadership is felt so physically as here. It is not a work where I can let go. I must be there in every bar, build the castle with the orchestra beneath me, and let the singers find their paths through what I have built.

It is an experience I find meaningful, almost in an ethical way. In belcanto I give my attention to the singer, and that is right there. In Bluebeard's castle I give my attention to the space, so that the singers can be in the space. These are two different ways of being present as a conductor, and this work has taught me the second in a way no other opera has.

For the listener it is a matter of entering the castle together with Judith. If one does, one discovers that the castle is more familiar than one might have thought – that these are one's own doors, one's own darkness, one's own will to know and the cost of knowing.

That is Bartók's great gift to the operatic form: he showed that an opera can be about the inside of a human being as precisely as it can be about historical actions and outer drama. It is a possibility that has since been exploited by many composers, but rarely so deeply as here.

And for me it is an invitation to a particular kind of leadership. I cannot follow the singers through this castle. I must build it around them and around us, with the orchestra as the walls and with my own pulse as the building's underlying pulse. When it is allowed to be like that, the singers can almost float on top of it – like two figures on deep water, carried by the current beneath. And then the work functions as Bartók wrote it: not as a conversation, but as a journey through a space that does not let us go until the last door is closed.