What Does Lulu Wear?

William Kentridge


I was invited by the Metropolitan Opera to direct a production of Alban Berg’s opera, Lulu. I turned down the invitation.  I didn’t have a solution for working on an opera that was four hours long. The charcoal animation technique I used was fine for projects which were fragmentary, which were short. It could take a month to animate a minute and here was four hours in the theatre needing to be imagined, to be put into images. I had already directed one theatre production, George Buchner’s Wozzeck, in which a woman is stabbed to death. I couldn’t imagine doing another project in which again the central woman is the subject of the violence of her lover. I couldn’t make sense of the music of Lulu, which does not have any clearly recognisable melodies. No audience member comes out of the performance humming a tune. I had no possible way of entry into the opera. 

I mentioned my rejection of the offer to two women friends who were musicians. I was castigated by both of them. “It’s the great opera of the twentieth century.” “But it’s so misogynist.”  “Don’t be ridiculous. Lulu is the most interesting character in opera. How could you turn it down?” “Oh,” I said, “of course yes, yes you are right. I misunderstood.”  

I had to recalibrate. There were obviously things in the opera that I had been both blind and deaf to. 

Some months later at the Museum of Modern Art in New York, I saw an exhibition of German expressionist prints. And in the hour of looking at the woodcuts of Emil Nölde and the drypoint prints of Max Beckmann and Otto Dix, it became clear how a production of Lulu could be made. The mixture of violence, sex, the world of Europe after the First World War, the brutishness of it, was all there in the prints. In front of me the whole production mapped itself out. I phoned the opera house and said I was ready to do the production of Lulu. I was informed that the production had been offered to a different director. Now of course all doubts disappeared and I really needed to make the opera.  I looked at it in the past tense and saw the production that I would have made, had I still been able to do the production. I think this past tense perspective is an important one, as it took away all the anxiety of thinking what is still to be done when the production was in front of me. Instead it was looking back, as if the production had been finished some time ago. I had used the energy of those German prints I had seen to infect and inflect the violence done on Lulu, the violence done by Lulu. In the production I had projected (I had made), I could see the woodcuts and music coming together, I could see the opera unfolding in a series of huge projections enfolding the stage. The production played itself out and I was more and more upset at not doing the production, or more precisely at not having done the production that was still to be made.

A few months passed and I was phoned by the opera house: “If you are interested in the production it is available, as the director who was doing it has decided to do a different project.” My bluff had been called and without hesitation, I had leapt into the project. Afterwards of course there were many months, in fact three years, of hesitation, regrets, ambivalences about the project, as the work unwound and the piece was made.


For a project like and opera production to take off, two factors are essential. The music or the libretto needs to connect to something in the outside world; or at least release associations to the outside world (that is, outside of the theatre and outside of the studio). A set of questions need to be present in the libretto but also to extend beyond the immediate story (footnote: in The Nose there was the question of utopian thinking and the absurd; in The Magic Flute, the limits of the enlightenment; in Ulisse, vulnerability and Fate).

Secondly, the project needs to meet a medium or technique or visual language, whether it is working with charcoal or scissors or cardboard or wood. This medium or technique informs the very way of thinking about the broader questions of the project. It is the medium of thinking about the project. It leads the production (and staging and the production of meaning).There is a meeting of the opera, the studio and the outside world.

It had been the woodcuts which seemed to me to be a language for the roughness, the violence, for the slash of the knife of Jack the Ripper. The slash of the woodcutter’s chisel or the chisel in the woodblock could meet both the violence of the text and the violence of the music. These woodcuts would be printed, photographed and projected onto the sets of the opera. The set would be made of panels and flat surfaces, walls, a ceiling, Chinese screens that could receive the projections.

But as a prelude to making woodcuts, I made ink drawings. An ink drawing, as opposed to a charcoal drawing, has clearly defined black and white. There is no mid-tone, there are no light greys, and if there are, they are achieved by cross-hatching or fine lines which are either on or off, black or white. This corresponds to the way of working with a relief-printing block: the surface is either cut away, or left intact to hold the ink. 

So the starting point was ink drawings. I worked on ink drawings with large Chinese brushes, working very roughly on multiple sheets of paper, usually sheets of paper from dictionaries or old encyclopedias, so that a drawing of a face might be made up out of four or six sheets of paper loosely pinned together, the corners meeting somewhere between the eyes and the nose. In many cases the drawing would be done more than once, either to get a likeness or to find the right brush mark or the right energy, and then an amalgamation would be made of the different sheets of paper, taking a mouth from one, half the head of another, changing the angle of the head slightly, tilting the head on the shoulders, adding an extra sheet of paper with more hair, adding a separate sheet with a dark background to make a cheek stand out. A breath across the sheets of pages would shift them. A new configuration would be made (the breath of the singers could move the projections). These drawings were the basis of woodcut prints. Once the sheets had been selected and realigned, this image would be translated onto a block of wood or linoleum, and the image cut out.


This process of collage and deconstructing and reconstructing the images is a secondary process, because it was done in the service of a woodcut yet to be made as a fixed image. This secondary process, with all its instability and provisionality, became the heart of the project. At a certain point I understood that the woodcuts themselves were not necessary. The ink and the multiple pages were where the project had its logic. And it met the themes of the opera of Lulu in the form of the instability of an image, which seemed to me to correspond in the opera and in the plays by Wedekind to the instability of the objects of desire. 

The impossibility of fixing who Lulu was, both in herself, but particularly in relation to the others, to the men, had a correspondence with the instability of the image, the flexibility of it. Different sheets could be blown away from each other, a different face made, a different body made, Lulu could be remade by the different forces around her. The opera is both about the refashioning of Lulu, and about her resistance to that refashioning. Lulu cannot be the woman that the men wished her to be, which is a stable single object of their fascination and desire; and the men cannot be the men that Lulu wants them to be, able to accept her as she is, able to follow the vicissitudes of her multiple desires, and accepting of those desires.

In opera terms this always ends in incomprehension and disaster, which on the opera stage is represented by death. So each of the men in her life dies and in the end of course Lulu herself dies. But the heart of it for me was the possibility of fragmentation, not as just decoration and not as illustration, but as part of the engine of the production. These ink drawings that could be projected on the stage, either the full size of the proscenium or at a much smaller scale, became both part of the scenery inside of which the opera would be performed and also part of the narration, and a commentary on it. 

Filling the stage with portraits of Lulu, with drawings of men and women, drawings of objects, has a grounding in the opera. A key protagonist in the opera is an artist, who is Lulu’s first lover and becomes her second husband. In the first scene the artist is busy making a portrait of Lulu. Even though he dies at the end of the second scene of the first act (he slits his throat when he discovers that Lulu still has her old lover Dr. Schön), his portrait of Lulu is present throughout the opera. This is called for in the libretto. There are references to the portrait in each scene.

In our production this portrait is a sheaf of pages of painted ink – in other words the projections from the wall come down into the hands of the singers. When they look at the portrait they look at pieces of paper arranged in various ways. We the audience see what they are seeing as projections behind them. Ink paintings of the portrait of Lulu continue even into London where Lulu is killed.

The first impulse of the production, the German woodcuts, have given way to ink drawings, and the ink brought with it several other associations that fed back into the themes of the production. You have the ink blot, the idea of a Rorschach test, of a blotted piece of paper with a dot of ink inside it. A Rorschach test, an inkblot test, is a space in which to see the response of the viewer. This connects to the psychological understanding of the characters, as written by Wedekind and expanded on by Berg, and the associations of desire that come from this Rorschach test. And of course their images are always enantiomorphs, mirror images: butterflies, skulls, and pelvises, things which are divided in the center. And looking at them is always about recognition. This is both recognition within the opera – what do the men project onto Lulu, the Lulu they wish she would be? And recognition across the orchestra pit – what do we recognize (of ourselves) in the characters, actions, and emotions engendered by the music?

The set on which the projections appear is made up of flat surfaces, the walls of the rooms described in the libretto (an artist’s studio, a bedroom, a dressing room in a theatre, a London garret), Chinese screens, furniture.

So the images became doubly shattered, first in the disjointed drawings made of imperfectly aligned sheets of paper, and secondly by the set itself. We see each image slightly differently depending on our position in the auditorium, and we have to do the work of completion and reconstruction of the image. This is a collaborative work between what comes to the viewer from the stage and what is projected by the viewer’s understanding onto these images.

Within the making of the images there is the collaboration of the makers of the opera. In the year or two before the production, the creators of the set design, lighting design, costume design and the video editor and constructor, convene (generally in workshops held in my studio in Johannesburg) to work out the balance of colour (in costumes, set, lighting) and focus (how much light is on the singers, how much projection is across the stage, how much light is on the stage?). We work with stand-in performers to work out how each scene could conceivably be staged. Here (on a model of the set) is where we compared woodcut projections to projections of ink drawings, and where the fragmented figure of Lulu could be moved under a camera while we watched the projection in the model.

Then of course there is the collaboration in rehearsal, with the conductor and singers; when much of what had been worked out in advance has to be abandoned, where new logics are revealed. This is where the promises and premises of the preparatory work are laid bare.


The erotic is like the comic. It’s not something you decide, it’s something you recognise. You don’t say, “Oh this joke is funny,” and then instruct yourself to laugh. You either laugh or you don’t. You don’t instruct yourself as to what is erotic, the erotic catches you. It catches you unaware, attracts itself into places that you yourself may not have access to. But with an opera (as in a comedy), and an opera about desire, eroticism and obsession, these questions need sober exploration. 

What does Lulu wear? One of the decisions made by the costume designer Greta Goiris, was to avoid making the obvious. To try to find a distance from predictable, cliched dress conventions of seductive clothing (fish net stockings), and to try to see if the erotic and the desirable could force its way through in unexpected ways. 

In our production, Lulu always wears men’s clothes – her first husband’s coat, the jersey of her second husband the painter, the starched shirt of Dr Schön, her 3rd husband, and the oversized jacket and shoes of Alva, even when she is a streetwalker in London. Only in the Paris party scene does she have a dress – and this is made of paper, a prop left over from the painter’s studio.

It became a question of what the music and the character and the libretto pushed forward towards the audience, and what they meet it with. Whether the skin of Lulu against the stiffness of the starched shirt makes its own attractions, or whether an oversized coat that was ready to fall off her body had its own eroticism or simply hid it – these were questions we were happy to test out on stage.  

But the question of tracking the lineaments of desire, what it is that makes for desire, is always complicated, unpredictable and often beyond our direct knowing. What is it about the twitch of the muscles at the edge of a particular mouth that becomes irresistible? Does it have to do with the nature of that movement itself or an old memory of another mouth with a similar muscle and an unattainable desire for that first mouth? There’s always a way in which the desirable and the focus of our obsessions is rooted in so many hidden and unknown parts of ourselves. Within the question of obsession and the obsessive desire that men have for Lulu, there are these questions. Lulu’s desirability and the obsession Dr Schön and Alva have with her has to do with her indifference. The fact that she will always be more or different from what they hope. 

Scene 3 of Act 1, in the dressing room of a theatre. Dr Schön comes to demand that Lulu behaves properly in front of his fiancée. He starts the scene in full control of it, ordering her about, shouting at her, reducing her to an object of his violence. And then at a certain point, instead of being cowed by his violence and his demands, she becomes indifferent, “Do what you like,” she says, “ I don’t care, I’ll do whatever you want but it makes no difference to me.” It is her indifference, her unreachability that Schön cannot bear. And he changes from the person of giving instructions, to a supplicant. 

“What should I do in these circumstances, what do I have to do to take you out of that position of being indifferent?” And she of course says, well what you have to do is write a letter to your fiancée absolving her of the vows and staying with me. Write her a letter saying you are sitting at the side of a woman who has you in her power. And Dr Schön writes the letter saying, “This is the worst day of my life,” and Lulu says, “I’ve never felt so good in my entire existence.” 



What is the interpretation that a singer brings to Lulu? I am wary of an interpretation being imposed rather than discovered in the process of work. Who Lulu is, one discovers in the rehearsals and staging. In the first scene of the opera, the artist chases Lulu up a ladder. In our production Lulu pushes him away. But the questions that arise are: How is it done? Does she meet him halfway? Does she allow him to follow her up the staircase?  If she puts her foot on his chest and pushes him away – does she pause before she pushes him away? Is it mock violence hiding real violence? Does she wait to see his response and then climb higher up the ladder? Does she turn her back on him and climb up the ladder irrespective of what he’s doing? It is in these minute shifts and movements and micro-decisions, that the character is revealed, it is in these movements that she is enticing, or disdainful, where the power lies. This is what makes the character of Lulu, and all the characters, rather than abstract decisions: that she was abused as a child, she was not abused, she likes the life that she has, she doesn’t like the life that she has. These are all matters for discovery in the production. We have fixed parts of the production: the music itself, the words of the libretto. Then we have ambiguities and possibilities for the piece to find its particular shape: in the drawings, in the projections, in the nature of the acting, in the specifics and minute specifics of the acting by the singers – and of course in the singing itself, the aggression, the restraint, the lyricism, the hardness produced by the singer, conductor and orchestra.  


Lulu is a nasty opera. Each time one hopes for a redemptive moment in one or another character, one’s hopes are dashed. When one thinks that Lulu is finally going to release herself into voluptuous love, she says to Alva, who has declared his love for her as he buries his head in her breasts, “Isn’t this the sofa where your father bled to death?” One of the wonders of the opera is our complicity as audience in Alva’s response: “It doesn’t matter, it doesn’t matter, I want to stay in this moment.”

The moral failings and bad decisions in the opera are made by the men in the opera. Lulu is very clear about who she is and the clearness and straightness of her needs and desires, and her lack of any sentimentality. The men always hide their brutish realpolitik in an assumption of love and gentleness. One of the difficulties of doing the opera was wondering how I would spend three years in the company of such nastiness, of characters one could not like.   It is a very domestic opera. Every scene happens inside a room; but it is an opera of the 1930s: an opera of domestic realpolitik, of tough negotiation, of no sympathy, of cruelty and violence. It becomes a mirror of the nastiness, cruelty and inhumanity of Austria and Berlin in the realpolitik of the 1920s and 1930s.

The production is an interior art deco production. Fragments of the outside world are brought in: there are momentary revelations of Göring, of the Hitler youth, of a gas mask, intimations of violence to come. But essentially it happens within the different sheets of drawings, their movement, their fluttering, their instability. In the scene set in a Parisian party (left incomplete by Alban Berg at his death and orchestrated by Cerha), money and sex and the outside world meet, with blackmail, with murder planned, with speculation on the stock market, with a stock market crash, bringing all these toughnesses and harshnesses together. We look at it and understand this too is who we are.

Is it a misogynist opera? Of course it is. Is it a misogynist opera? Of course it is not. The possibility of the opera is in its very ambiguity, in the riddles that are not solved, in the meeting of the production and the viewer at some middle point in which the viewer brings their hearing, their viewing, their understanding, their sets of associations, to meet the opera.  If one says an opera is boring or it is misogynist, one has to understand that it is always yourself you are talking about.  The opera is the same.  The notes and the music is the same and so a different interpretation by a viewer has to do as much about themselves as the opera. And when you say it is boring, you really mean, “I find it boring.” Or to say it is misogynist is to say, “I find it misogynist.” It is not to say one can never talk about a work of art. But we have to understand that we are also always talking about the viewer. And the interest in doing the opera is not just to see the opera outside of ourselves, but to allow it to reflect back on us, the makers, or us, the viewers of it. And it is about Lulu’s obsession, but Lulu’s obsession is always a mirror – in what it meets, in what it doesn’t meet, in what we recognize and what passes us by – of our own obsessions and desires. 


When I was three years old, I was asked what I wanted to be when I grew up. I am reported to have answered that I wished to be an elephant. When I was fifteen, asked the same question, I replied that I wanted to be a conductor of operas. This longing was equally impossible. It was pointed out to me that to be a conductor of operas one had to read music. “Oh,” I said, “yes of course.” I had to recalibrate. In the end directing operas was as close as I would get to either of those desires.

The first time I heard Lulu, I struggled to find a sense or pleasure in the music. By the 4th listening, the music changed. Now after countless listenings, every note seems essential and vital.

But it is an ongoing conundrum of opera, that it relies on both newness and familiarity. The obsessive-compulsive disorder of opera-goers: they have to repeat endlessly the same activity, going to La Bohème not once but twenty-seven times, see not one production of The Marriage of Figaro, but one every year, as if they had to read the same 23 novels every year for the rest of their lives, is one of the problems of opera.  The audience says, we want something familiar; but show us something new. Which leaves a very narrow band of neither the too familiar, nor too new. 

But what is behind the obsessive attraction of opera? Part of it must certainly be the direct expression of strong emotion; that love, hate, revenge, desire are there in their most direct forms, a directness of expression one would be ashamed of, or embarrassed by, or shy to proclaim in an ordinary theatre, but which in opera one accepts and celebrates, buoyed by the music and by the extraordinary qualities of the singing voice.

These are the universal attractions of opera. But what is an artist doing in the opera house? Not as a designer. I am a director, making productions in which the images I make, which are projected, are part of the narrative and part of the meaning of the piece as it unfolds. For the set-design itself, I work with a set designer, Sabine Theunissen. I have a bad three-dimensional imagination (even the sculptures I make are essentially two-dimensional images that happen to be visible only in three -dimensional constructions). Opera is the making of a drawing in four dimensions, and working with the drawings changed in scale, moving through time, with performers and music as part of them. Allowing a whole structure to be constructed as one would construct a drawing, allowing an openness and unknowingness to exist both in the years of preparation and in the six or eight weeks of rehearsal. 

It has to do with scale. The operas offer a place in which a five-story high projection of a brush mark becomes possible. The scale of the images also corresponds to the scale of the music: a non-domestic scale. An orchestra at full volume is not a domestic instrument. And an opera house, with its seating for two or three thousand, is not a domestic setting. The singers recede in size, but the projections which can enlarge to the full height of the proscenium, can be seen and can engage the audience even in the very farthest seats of the auditorium. 

The projections operate in terms of scale and visibility, but also more importantly as a version of the Greek chorus commenting on the action and what is happening to the singers in front of it: either illustrating their thoughts or at right angles to those thoughts, or refuting the words that they are singing. 

And of course the heart of all opera: that unstable landing point of desire engendered by the human voice triumphant, in the music; in which we can’t situate precisely what in us is moved – it is more than just tears or joy or sexual desire – but an energy of what it is to be alive that we feel through all our senses; the provocation and satisfaction of an enigmatic longing. 

William Kentridge


Published in Spanish translation in M. Borja-Villel and S. Liaño Gibert, William Kentridge: Basta y Sobra, Museo Nacional Centro de Arte Reina Sofia: Madrid, 2017.