Bunuel’s Net Work

Bunuel’s Net Work – Luis Bunuel

James Tobias

Performative Doubles in the Impossible Narrative of The Phantom of Liberty

Introduction: the Impossible Narrative of The Phantom of Liberty

The Phantom of Liberty (1974), cited by Luis Bunuel as one of his favorite films and his “most surrealist” work begins with a historical reference to the Napoleonic Wars, and proceeds to careen through a number of episodes which, although each is linked to those that precede and follow it, remain partial and fragmented.(1) We encounter dozens of characters through a well-designed but seemingly random narrative that introduces characters who seem both puzzled and puzzling,only to escape from the situation in which these characters find themselves and move on to a new one. In total, the experience we have depends more on impossibility than verisimilitude, and that quality has helped to inspire a range of divergent responses to the work. In an early review, Marsha Kinder notes the circularity brought about by the film’s conclusion,(2) while in a later critical study Susan Suleiman considers the continually displaced narrative as radically linear.(3) Suleiman also considers Kinder’s positing of a thematics of the “tyranny of convention” to be “too tragic.”(4) Still later, Linda Williams finds the film a surrealist statement on the oppositional structuring of law and desire, but considers Suleiman’s reading overly formalistic, while disallowing Kinder’s reading of the film as a moral statement by Bunuel. Moral statements shouldn’t be interpreted in such a strategic surrealist art.(5) The most fascinating aspect of all of these readings, though, is what they reveal about the film: they are all apt, although each author may successively disagree with or contradict the preceding critical review. In fact, the succession of contradictions which these readings of the film offer echoes the succession of contradictory episodes in the film itself.

It’s possible to frame the work in such a way as to make all of these responses coherent readings of the film. I will argue that The Phantom of Liberty works by presenting a series of episodes which we can regard not as linear narratives cut short but as potential narrative networks which, although only partially activated for our experience, can continue to operate and influence each other. Each episode presupposes the existence of a set of real world power networks. The narrative networks in which the characters make their moves in the film resemble these real world networks, so we understand that the film is realistic to the degree that we recognize reality.

How can material that is networked be narrative? Roland Barthes made one of the clearest attempts to place the figure of the network at the service of interpretation. In S/Z, Barthes applies five codes of denotation and connotation which, he says, all lexia fall into. The voices of a text are woven as these codes are networked to produce pluralities of meaning in a definitely segmentable literary topos which produces texts that are imperfectly reversible.(6) Rather than put the figure of the network at the service of interpretation, I’m more interested in how the network figure has been used in a variety of discourses that evolved concurrently with Bunuel’s work; S/Z is a crucial example of such a use. On the one hand Barthes’ use of the network figure relies on a propositional segmentation of verbal elements which is more appropriate for the construction of verbal subjectivities as opposed to cinematic ones, and on the other, his usage of networks doesn’t provide much leverage in terms of understanding the actions of the cinema as a network that produces massive bodies of spectators. Phantom calls for an analysis which can handle texts which are not definitely segmentable, and which might have sections which demand reading backwards as well. Most importantly, from his earliest films, Bunuel has always been aware of both the specificity of the cinema and the power structures that it deploys to produce imagining subjects. His films in general and Phantom in particular demand a fuller accounting than S/Z provides. How has the network figure come to be used across domains?

From originally referring to communication links established over a geographical space as in a broadcast network, “network” as a structuring motif has taken over from “circuit” for much of our thinking about complex elements of a biological nature, allowing us to speak of “neurotransmitters,” “receptors,” “input” and “output.” “Thought” and “soul” can be placed in a neural chain of command somewhere farther down the line than “retina” or “visual cortex.”(7) With the rise of cybernetic telecommunications, network has also come to mean the elements and links that carry data within and between computers,(8) leading to the use of the figure as a processing paradigm wherein networks function in a description of the knowledge and actions that produce, for example, intelligible speech.(9) The connectionists of the 1980s built “massively parallel” computers that would enable “intelligent machines,” and modeled schemes for the acquisition, storage, and processing of data as operations upon items existing in networks, this time configured after a neurological model.(10) Marvin Minsky’s polemic, Society of Mind, combines social, technical, and neurological senses of the network figure to argue that our brains are societies of elements assisting each other in a technical collaboration that results in effects that we call emotion and intelligence.(11)

“Network” can thus describe any structure that includes elements which communicate to each other in one sense or another. What constitutes an element and what constitutes a link become defining qualities with great implications. With the garrulous networks whose cumulative effect is intelligence in a “society of mind,” Minsky proposes an understanding of “the strangest mysteries of mind.” Minsky reproduces his concept of networked agents by making each page in the book an independent section which links to the others, and which are organized into sections which handle a specific function: memory, emotion, and so forth. Each page in the book itself is an element in a potential network of rationality which waits for the user to read, activate links, and begin to bring the network into being. On a macrolevel, the book reprises the work of the networked agents he describes as the builders of consciousness. It sets out an agenda for a cognitivist enterprise that will finally build mind out of brain by positing “agents” that communicate with each other in different sorts of groups acting hierarchically to perceive the world, make it into sense, and move the body into action. “Agents” are the “particles” that theories of artificial intelligence need to become robust and productive. And at the level of the book, the reader becomes an agent which can help Minsky to implement his society of mind.

In Minsky, the network figure becomes internal to a machine in the same way as it is internal to a human, human intelligence turns out to be that of a “wonderful machine,” and the mind falls into line as the rational (and rationalized) output of a neurological brain. Mind is explained by this cognitive networking in the same way as biology has successfully explained “life.”(12) In the bargain, we have a model for our social world as well. Minsky understands liberty, that is, “freedom of will,” to be an imaginary construction, a phantom, if you will, that is necessary for sanity and social order:

Consider how our social lives depend upon the notion of responsibility and

how little that idea would mean without our belief that personal actions

are voluntary. Without that belief, no praise or shame could accrue to

actions that were caused by Cause, nor could we assign any credit or blame

to deeds that came about by Chance. What could we make our children learn

if neither they nor we perceived some fault or virtue anywhere? We also use

the idea of freedom of will to justify our judgments about good and

evil…. But if we suspected that such choices were not made freely, but by

the interference of some hidden agency, we might very well resent that

interference. Then we might become impelled to wreck the precious

value-schemes that underlie our personalities or become depressed about the

futility of a predestination tempered only by uncertainty. Such thoughts

must be suppressed.(13)

I quote this passage at length in order for the implications of network, knowledge, element, and link to become clear and their relation to power become unmistakable. Minsky posits the mechanical as but a well understood version of the human, and designs a cognitive science that finds networked agents internal to the brain, a science that will explain the mind just as other sciences–physics, biology, chemistry–have explained the external world. In this virtual world in Minsky’s mind’s eye (which he proposes as a program for the world of industry and research that will take many years to implement), communicating agents are unique but can possess multiple links and be hierarchicized in complex ways, while free will is a ghost of an idea that is necessary to motivate children to learn and virtuous citizens to refrain from evil. To question freedom of will in the name of self-knowledge is to risk the loss of “cheerfulness!” Not much to give up, in fact, if you are an agitating filmmaker out for a hearty laugh in a less than perfect world where mystery is more of a criterion for hope than is mechanization.

The networks that Michel Foucault describes in The History of Sexuality are all power-oriented. Foucault’s networks make no apology for wanting to be intelligent but not being able to manage the task. For Foucault, power operates in networks of “institutional devices” and “discursive strategies.” For example, sexuality is produced in these networks. Repression does not reduce or confine sexuality but helps to multiply it in new forms:

Speaking about children’s sex, inducing educators, physicians,

administrators, and parents to speak of it, or speaking to them about it,

causing children themselves to talk about it, and enclosing them in a web

of discourses which sometimes address them, sometimes speak about them, or

impose canonical bits of knowledge on them, or use them as a basis for

constructing a science that is beyond their grasp–all this together

enables us to link an intensification of the interventions of power to a

multiplication of discourse.(14)

Power-oriented networks might deploy or utilize the sorts of resource-oriented networks described above, as well as institutions such as schools, hospitals, psychiatrists’ offices, or confession cabinets. But Foucault’s networks of power cannot be reduced to networks of resources, and operate fundamentally differently. An element in a power network may have more than one name, and delivery of knowledge as an effect of power is never guaranteed and not always predictable. Foucault finds power at work in the family producing “multiple, fragmentary, and mobile sexualities.”(15) Sex can be transformed from what we might think of as a “link” between singular and “elemental” bodies(16) into “that which had to be confessed”: power gives many names to sex, and the difference between element and link begins to blur. It is for this reason that Foucault goes to great lengths to define how power operates, what its effects are, and thus how it can be discerned in action.

Bunuel, on the other hand, structures narrative in The Phantom of Liberty in a network of effects. The effect-oriented network works as a depiction of a power network; as in a power network, elements are no longer unique and links are no longer distinct from elements. But Bunuel shows that when it is only the reading of the work that makes power real, networks of power can be represented in disconcerting ways that call into question the presumed coherence of those networks’ operations. Networks of effects can have links that are disjunctive as well as conjunctive; reversal, juxtaposition, doubling, opposition and other operations work as well as input and output in producing knowledge. Resource networks, then, operate in terms of input and output between linked elements to fabricate and supply knowledge. Power networks multiply elemental identities to produce truth. Effect networks link multiplied elements through disjunctive operations to question the truth of knowledge. Minsky’s book is a network of effects that aims at convincing the reader of his program for the mechanization of consciousness in resource networks; Foucault’s book is a network of effects for identifying and understanding power networks; Bunuel’s film is a network of effects for causing a power failure.

The narrative network is the primary mode of cinematic operation in The Phantom of Liberty. Particular characters involved in particular configurations representing real world power networks constitute narrative networks which Bunuel introduces successively throughout the film, but which he quickly leaves in potentio as one set of elements and links leads to the next. The character’s moves in these narrative networks prove fragmentary and contradictory, revealing the fragmentary and contradictory nature of the power networks which Bunuel is setting up as extra-textual referents. And our own engagement with each character, each network, also is limited to a fragmented and contradictory experience, as the story moves through many narrative networks just as a data packet moves through a digital network.

Bunuel’s presentation of these narrative networks makes use of a critical notion of power networks which define and even produce individuals at the site of the reception of the messages that these networks carry. Much of Bunuel’s filmic work on power not only historically precedes that of Foucault, but provides a cinematic precedent. We can best understand The Phantom of Liberty as a special effect made visible where narrative technologies produce knowing viewers in the power network that cinema is. In this film, one particular path out of many potential paths is taken across narrative networks that Bunuel screens as a critic of the political and social structures that continue to rule the twentieth century. In a world where we imagine networks of power articulating possibilities of identification and movement, Bunuel constructs a personal vision of those networks of power in order to expose them as such.

A family network may be represented by the characters, their relationships to each other, and the situation within which they are found. But the family narrative may be surprised to find that the family network structure that it seems to depend on is somehow dramatically or comically hinged on the narrative movement of a character belonging to a different network, the maid, for example, whose relation to the family is one of class and labor. These links are enabled and can become critical, since characters having roles in different networks can lead to multiple networks which depend on one character. Bunuel, in a manner similar to Foucault in his analyses of power, intertwines functional structure and motivating content, ultimately portraying power as utilizing both but pretending to require neither. A shift in one may produce a shift in the other.

In this way, what I call the “impossible narrative” becomes a requirement, if one is to even pose the question of portraying power in film fiction. Without apparent contradiction, the necessary yet invisible relations of power to power will not become visible, and so cannot become cinematic. In The Phantom of Liberty, each time Bunuel shows one networked structure hinged to another in a narrative movement, he does so with a twist which in the same moment makes it clear both that the structure is apparently the double of a real world power structure external to the narrative, and that the external structure that he is referring to is, in some sense, twisted. Because the narrative network performatively doubles real world structures of power through dramatic operations and thus is not simply a sign referring to power, undoing the double effectively becomes the imaginary undoing of the original. Still, the double can only resemble the original to the degree that you share Bunuel’s cultural background or can read his frames. The undoing of the double, ultimately, is your job. Bunuel’s job is to make your work a pleasure. You become involved in what may be termed a syndicalism of narrative work, but in Bunuel’s anarcho-syndicalist vision, you labor to produce The Phantom of Liberty.

A critical work as well as a narrative plaything, The Phantom of Liberty projects the viewer as a willing participant in the elaboration of power, but at the same time as it puts the viewer in the spotlight, it offers the chance of escape back into the dark, irreducible mysteries beyond power’s blaze. In a vivid example of the fact that a work that seems narrative may not only elucidate a critique of power but work to devoid power of itself, The Phantom of Liberty requires a disjunctive reading of power networks both interior and exterior to cinema; Bunuel’s spectator is created recognizing those power networks which not only produce cinema but which cinematic narrative reproduces. As viewing actions, the contradictions through which The Phantom of Liberty is structured make those networks discernible, and allow the viewer produced at their articulation to take pleasure at their expense.

The film shimmers with ambiguity. We are left with many possibilities for rendering the film coherent as we trip down the primrose path where The Phantom of Liberty leads us. Not the least of these possibilities is that in an originary experience of the film, we viewers are only a phantom of Bunuel’s imagination. This possibility becomes conceivable to us, yet patently impossible. Of course we exist, but how can we otherwise explain our experience of the hermetic nature of Bunuel’s passionate meaning-making? After all, somehow we are making some sense of this nonsense. Contrariwise, we can also imagine that Bunuel has only produced this work to satisfy the actions which we will take upon it. In this reading, he becomes the phantom, and if we imagine him at all, we imagine him only as an originary excuse which relieves us of taking full responsibility for the implications that the film sets forth. And yet, the film’s contradictions demand a response: what of our phantom, liberty? Bunuel’s “net work” posits both a creative, historically existing author, and a creative viewing subject, each of whom exist at cross-purposes to each other in a work which is not entirely reducible in terms of either figure’s perspective.

Looking at The Phantom of Liberty

In order to characterize the networks Bunuel is lacing together, we need to look closely at the film. We enter The Phantom of Liberty through a painting: The Executioners of May 3, 1808. The painting is by Goya, to whose thematics many have compared Bunuel’s. It depicts the execution of Spanish resistance fighters by Napoleonic invaders bringing French empire to Spain in the name of the professedly enlightened values of the French revolution. After suggesting we are entering into a painter’s retelling of a historical moment almost two hundred years old, the film moves into scenes of execution, with Spaniards shouting “Vivan las ca’enas!” or “Long live our chains!”(17) before dying. The resistance fighters choose autonomy as a monarchist state over the liberty offered by the French, at the cost of death.

But no sooner do we believe we are seeing a film referencing an important work of art which itself visualizes a defining moment in Spanish history than we are informed by a vivid red title that the film is based on the novel by Gustavo Becquer, whose text inspired Goya’s painting. And in fact, the film never re-poses actors in a literal re-staging of the Goya work, but uses it to contextualize a new version of a historical novel. Shortly we find that this new literary context is also a pretext. The film now no longer seems to be retelling the story of the Goya, as we move on to the story of a French captain who falls in love with the funerary statue of a beautiful woman. Framed in profile, the captain caresses her stony loveliness, but the frame suddenly shifts to a wider view where we see, unexpectedly, the statue of the woman’s father, a knight, lift his arm and smite him. For the moment, the family head has protected the honor of the young woman; the virgin saint is intact, and the foreigner repulsed. A woman’s voice begins a voice-over narration of the scene as the captain opens the tomb of a woman who turns out to be a saint, miraculously preserved.

Though the film settles in a narrated version of Becquer’s historical novella, the voiceover is halting, unprofessional. In fact, it is rather squeaky, and as she tells the story, she doesn’t seem to read well. Just as we are wondering what kind of historical narration this could possibly be, the woman comes to the word “paraphernalia” and cannot pronounce it. As our narrator struggles over the difficult word, the Romantic scene of the open casket slips to a view of two women, maids, sitting on a modern park bench. Our narrator has been reading aloud. Her companion, knitting beside her, must explain what “paraphernalia” means.

If her question arises from a scene of death and violation, the answer comes from critical comparison. The knitter calmly explains that “paraphernalia” has most often meant the dowry of a woman to be married, but in England, it refers to all of the household objects, even the sewing kits. This tranquil knitter, twin of the lace-maker often found in Bunuel’s films, seems even more omniscient than our narrating maid. She gives us a key to the film: in summing up the meaning of the word, she is summing up the materials and the strategies of the film: this film, above all, is the knitting together of paraphernalia. Paraphernalia is culturally based, gendered, subject to historical forces and operations of power. And yet, in her breadth of knowledge, she indicates that meanings are multiple and contingent, and culturally translatable, within limits. Just as she knits wool and meaning at the same time, the film has been knitting visual references, literary references, and historical figures. Just as she is binding up her production of meaning in a lacy network, so, might we think, is Bunuel. And should we protest such an immediate evocation of an enunciating author in the form of a speaking figure bestowing meaning on a confused reading of a text, we might also consider that Bunuel has already disguised himself to appear as one of the fallen Spaniards whom we have left for dead in the distant past.(18) As the first episode melts into the second, the film presents us with a questioning as to the effects of cinematic narrative gaining pre-eminence over competing arts such as literature and epic painting in the representation of our nations, our passions, and our cultural meanings. Then, in the slip of a mispronounced word by a faltering narrator, we are moved into a problematic of authorial enunciation and receptive sensemaking. Which counts more: the hesitating speech of the narrating maid or the omniscient explanation of the listener knitting away beside her?

Back in the park, the maid’s ward, Veronique Foucauld, and her girlfriends are approached by a sinister man. We instantly suspect him of being a child molester, so cautiously yet insidiously does he watch and then approach the children. He shows them what we believe are pornographic postcards, which he gives to Veronique, who rejoins the maid to bicycle home. The narrative trots along with them. Madame Foucauld, scandalized by the postcards her daughter delivers, consults with her stressed husband. Disgusted, then aroused, Mr. Foucauld recalls the couple’s romantic vacations as they thumb through the pictures, never quite mentioning any explicit detail. Madame hushes him, upset that Veronique may overhear their suggestive reminiscing, but the idea of being overheard by the child only contributes to her heightening excitement. She kisses him madly. As their ardor cools, we see the contents of their photographic aphrodisiac: well-known tourist monuments, which now again appear, in their eyes, revolting. He rips one to pieces: the curvaceous domes of Sacre Coeur simply “go too far.”

The maid must be fired for allowing Veronique to receive these dirty pictures, while the pictures, now spent, are returned to the girl so she may trade them for pictures of insects. Family networks, networks of class and status, and networks of nationhood are all invoked and interlinked, and skewed just enough to become contiguous. Each structure is called up through a narrative that begins to happen, only to shift into a different one. The way these networks are invoked is far from realistic. Each one appears through a series of operations that question its coherence and independence. Bunuel has begun a movement through a set of narrative networks which seem to resemble the real world power networks we know, but which differ in revealing ways.

In an important sequence that follows, we see an insomniac Mr. Foucauld set his alarm as he notes the time. He turns out the light, and the camera pans to his wife. The camera pans back, and after a few moments, he turns the light back on, and sits up to get a drink of water. He is astonished to find that hours have passed. Has he been asleep? Or is he still asleep? Might this be a dream of his wife? After we watch Mr. Foucauld watch a nocturnal parade of visitors comprising a chicken, a secret agent, a postman who delivers a letter, and an emu, all within moments of film time but apparently taking hours according to the chiming of the clock in the story, we hear a voice which says, “That’s sufficient,” and we find ourselves in a doctor’s examining room. The husband has not actually been dreaming or having strange visitors, but retelling the strange visions of the night before. For Mr. Foucauld, the experience is strange but real, for the doctor it must be a dream, while for us it is a diegetic device that moves the narrating forward, a narrated retelling that intensifies Mr. Foucauld’s anxiety, and a cinematic distraction in its own right. A reconciliation between these points of view seems impossible. To prove the experience actually happened, he hands the doctor the letter he received, which seems to prove something, but what?

A nurse comes in to solve the problem. Leading the camera and the doctor into the next room, she hands the doctor yet another letter, detailing the precarious health of her aging father. She wants to take a few days off to visit him. The doctor takes this letter for the real thing, grants her request, and returns to his distraught patient. The nurse’s solution is simple: the camera finds her driving through pounding rain, as the narrative branches yet again to follow her into another set of networks.

Critical Configurations

Accounting for the sudden shifts and radical openness of this film has typically been understood by configuring the elements of the film as linear episodes which progress through syntagmatic elaboration or metonymic shift. Susan Suleiman’s article, “Freedom and Necessity: Narrative Structure in the Phantom of Liberty,” allows us to understand the limits of a “linear” reading of the film, and what a “networked” reading offers. Suleiman divides the film into 12 narrative sequences.(19) “Constructing the sequences is a way of arriving at the linear-logical progression of a narrative.”(20) For Suleiman, there must be a specific point in which one episode ends, and another begins.(21) She delineates episodes by locating a “Subject” in each one; a new subject indicates a new episode. Thus, episode leads to episode, as sentence leads to sentence. Since the film avoids closure in any one episode by moving into a different story, The Phantom of Liberty subverts narrative codes by obeying them in such a way that no coherent story can emerge. The film is a maximally free narrative still obeying narrative form; Bunuel is expressing the freedom “not to make sense.”(22) But two aspects of segmentation work against the film’s construction.

On the one hand, a set of scenes may have more than one subject; on the other, they may belong to more than one “episode.” Are the strange events that follow the camera’s move away from Madame Foucauld’s sleeping face definitely events that happen in a diegetic reality focussing on her husband, whom Suleiman names this episode’s Subject? The cinematic language is ambiguous enough here to make us wonder whose dream this is, especially if we remember the dream-within-a-dream weave of The Discreet Charm of the Bourgeoisie (1972) that makes that film so puzzlingly rich. Part of the key to understanding the ambiguity here is in understanding the potential for reading film sequences both forward and backward. Robert Eberwein identifies two modes, a forward and a “retroactive” mode of dream narrative.(23) Certain scenes may be motivated in forward mode: Mr. Foucauld goes to bed and has strange visions. They may also be motivated in a retroactive mode: the night visitors appear not as an actual happening or dream but as a retelling which occurs in the doctor’s office. Segmenting from both directions, we get two readings of a set of scenes, and the episodic nature of those scenes becomes indeterminate in terms of where each begins and ends.

Ultimately, narrative flow becomes indeterminate and partial, yet not incoherent. The fluid transitions of this film allow a bidirectional reading to be taken to the extreme. One far-fetched possibility would be to read the film backward from the scenes of the zoo as the highly disturbed hallucination of an ostrich with a richly developed sense of social conscience. Alternatively, most of the film might be a Surrealist cut-and-paste rewrite of Becquer’s novella that begins with an interrupted reading of the word “paraphernalia”. The possibility of multiple segmentations suggests the possibility of multiple subjects. The dream retold in the doctor’s office might seem most appropriately assignable to Mr. Foucauld reading either forward or backward. But the camera’s pan from Mme. Foucauld to her husband before we see his visions makes it conceivable that she is their dreamer. Locating his dream within hers explains why Mr. Foucauld can hand the doctor a “real” letter: even the doctor’s office sequence is part of her dream. But then where does her dream end?

The uncertainty of the constituting subject of the sequences of the film contributes much to its narrative instability. Is the episode in the park only the maid’s turn as Subject as Suleiman suggests, even though the child Veronique commands much of the screen time and seems to motivate her parents’ anxiety in the scenes that follow? Actually, it is the child molester who initiates the tension in this scene. Perhaps he is the Subject, although he disappears quickly. Perhaps an episode is merely the after-effect of the degree to which we are willing to assign an entire set of scenes, or even a single camera move within one shot, an overarching significance. Since the man in the trench coat never becomes a Subject in Suleiman’s reading, she neglects to say that an episode might center on his role in the child’s experience. Likewise, we may neglect to remember that Bunuel himself has spoken of the Surrealists’ pleasure in exposing pornography to children.(24) Perhaps we viewers are infants to whom Bunuel is simply showing screens upon which to project our own pictures, so the pervert may be an important enunciating Subject. To say that there is a single Subject per episode effectively marginalizes the pervert as a speaking subject, which might consequently prevent us from wondering, along with Bunuel, why the provider of provocative postcards is more of a threat than the parents who display sexual behavior on the living room couch.

Global theories of narrative notwithstanding, in terms of cinematic language, episodic boundaries and centers in The Phantom of Liberty are not well defined. The narrative is carefully crafted not only to avoid style-setting segmenting between episodes,(25) but to accentuate the lack of overdetermined segmentation that especially characterizes this film among all Bunuel’s films. Each plot twist seems wholly unpredictable yet natural in terms of the cinematic language that they are phrased in. Crucially, the transitions between episodes are indistinguishable from the episodes themselves. When, for example, we realize that Mr. Foucauld has handed off the narrative to the nurse, her story has already started. Her story begins not with the cut that finds her in the car; but with the door in the doctor’s office through which we have fluidly been transported from Mr. Foucauld’s angst. This careful overlapping of one story with another is effected in terms of the way we read movement within the frame. Bunuel is careful to allow the expectation of returning to Mr. Foucauld to linger by using continuity to make the shift between episodes instead of strong cuts or other cinematic statements of severance. Narrative coherence results from a smoothly twisting chain of perverse relationships, since the film perversely refuses to distinguish, finally, between episode and transition. Instead of seeing the shift between Mr. Foucauld, the doctor, and the nurse as a line formed between two successive episodes, it’s more productive to see them as co-existing narrative networks that share complex elements of time, character, setting. The camera is not cutting between two episodes: it is panning between the visible effects of two, or more, narrative networks.

Bunuel’s knitter, or her twin, the lacemaker, is a sign of the power networks which he will be interrupting in addition to being a signifier of narrative coherence-in-complexity. Finally, she performs as a signature identifying his body of work.(26) Sign of her times, she leads us to Foucault. I read Foucault’s analytics of power as an attempt to render power visible and thus, addressable as such. Networks deliver truth to us in the form of a world of institutions and discourse in order to form us into subjects at power’s command; Foucault’s “genealogy” of sexuality helps us identify power’s role in the formulation of our obedience to law in all its forms.

Both Foucault and Bunuel link sexuality to other domains of power, such as politics, statehood, and history. Foucault relies on a historicizing power of language, while Bunuel takes advantage of the participatory dynamics of cinema. Since Bunuel’s film sets its critique of power against a vastly different set of expectations of reception and interpretation from those operative for Foucault’s text, the performative specificity of these texts becomes apparent. Cinematic fiction offers possibilities off-limits to the historian. Bunuel is able to go further than Foucault in structuring resistance to power, because he works through narrative discourses relying on voyeurism, fear, and humor. In fact, Foucault is even slyly evoked by Bunuel’s text in the anxious Mr. Foucauld: the pronunciation is the same.

The History of Sexuality for Foucault, like The Phantom of Liberty for Bunuel, comes late in his career as a historian amid mounting critical acclaim, after long periods of controversy. Foucault’s history sometimes reads almost as a primer for getting at the subtle and surreal dynamics of the film, to the extent that Foucault’s explanation of sexuality in the family neatly backs up Bunuel’s depiction of the Foucauld family dynamics. “Modern society is perverse, not in spite of its puritanism or as if from a backlash provoked by its hypocrisy; it is in actual fact, and directly, perverse.”(27) Foucault on the family network’s production of sexuality:

Was the nineteenth-century family really a monogamic and conjugal cell?

Perhaps to a certain extent. But it was also a network of pleasures and

powers linked together at multiple points and according to transformable

relationships. The separation of grown-ups and children … the relative

segregation of boys and girls, the strict instructions as to the care of

nursing infants …. the attention focused on infantile sexuality, the

supposed dangers of masturbation, the importance attached to puberty, the

methods of surveillance suggested to parents … the presence … of

servants: all this made the family … a complicated network, saturated

with multiple, fragmentary, and mobile sexualities. [T]his apparatus …

was less a principle of inhibition than an inciting and multiplying


Bunuel’s Foucauld family is perverse in exactly these ways. Bunuel’s joke on the homosexual, anti-psychoanalytic Foucault is to name the father after him, cause him to see bewildering visions which seem imaginary and yet have real effects, and then have a doctor tell him to go see a psychoanalyst if he wants to talk about dreams.(29) For each operation that Foucault offers as a feature of power, there is a corresponding reversal or dispersal in Bunuel.

First, the “negative relation”: “Where sex and pleasure are concerned, power can “do” nothing but say no to them; what it produces, if anything, is absences and gaps; it overlooks elements, introduces discontinuities, separates what is joined, and marks off boundaries. Its effects take the general form of limit and lack.” Rather than a negative relation, Bunuel portrays positive yet indeterminate relations. When the little girl is shown the “dirty” postcards in the park, she knowingly chuckles along with the pervert, humoring him. The pervert’s success continues as the postcards are enjoyed by everyone in the family.

Second, power insists on ruling, and dictates its law to sex: “Which means that first of all sex is placed by power in a binary system: licit and illicit, permitted and forbidden.” Through language, the hold on sex is maintained by acts of discourse that articulate a body of law. “The pure form of power resides in the function of the legislator; and its mode of action with regard to sex is of a juridico-discursive nature”. Bunuel defracts the rule of law through multiple enunciations within and beyond language. Bunuel accesses and redeploys massive networks both intertextual and corporeal–painting, literature, postcard, cinema–to have various sexualities appear as themselves without speaking the law. The pervert speaks via an exhibition of postcards, and passes on the pleasure without being disciplined. Generally, the film interrupts its sexually active members to turn to stories of police intransigence and repression. But while sexuality links up with a “juridico-discursive” legislative system in Phantom, that link works to visualize the relations of power, not to order sexuality through jurisprudence.

Foucault’s “cycle of prohibition” means that sex must renounce itself. “Power constrains sex only through a taboo that plays on the alternative between two nonexistences.” Could these twin nonexistences be figured as the stone statues guarding the rosy body of the dead virgin saint that the French captain lusts for in The Phantom of Liberty? Possibly, but we never find out, as we leave the captain to join the maid. There is always another alternative, and we must move on to try them. Bunuel structures taboo but shows the noncausal effects that surround it in the form of narratives leading elsewhere, away from prohibition, into another story.

Fourth in Foucault’s features of power is “the logic of censorship.” “The logic of power exerted on sex is the paradoxical logic of a law that might be expressed as an injunction of nonexistence, nonmanifestation, and silence.” Bunuel speaks sex, but in such a way as to make it both obvious and uncensorable.

Foucault’s fifth feature of power is the uniformity of the apparatus. “Whether one attributes it to the form of the prince who formulates rights, of the father who forbids, of the censor who enforces silence, or of the master who states the law, in any case one schematizes power in a juridical form, and one defines its effects as obedience. Confronted by a power that is law, the subject who is constituted as a subject–who is “subjected”–is he who obeys.” Not only is the sex as it is visualized in Bunuel fascinating, it is also patently ridiculous. By parodying the ways in which people obey the rule of power over sex, Bunuel makes it hard to portray the sexualities which he brings to the screen. Further, they are plural, existing in a set of networks which you see but will not likely ever experience. Try as you might, it would be exceedingly hard to obey the Bunuelian communication of sex as it wanders all over France in The Phantom of Liberty.(30)

Erotic desire and oppressive political systems are indeed linked in this film, as Linda Williams has pointed out. For Williams, this film begins as a celebration of the pursuit of erotic, political, and narrative transgression, but ends as a recapitulation of our “entanglement” within the “structure of oppositions that give rise to this pursuit in the first place.”(31) The film is a surrealist shock statement as Un Chien Andalou (1928) was almost fifty years before. She notes that the film does “achieve the same rupture with identification and with belief in the fictive unity of the work” which identifies it as a continuation of the surrealist project. What’s different in The Phantom of Liberty is an emphasis on metonymy that ensures the smooth transitions by which this film flows, as opposed to the metaphor ruling in Un Chien Andalou. Instead of relying on visual metaphors to make the link across a cut, Bunuel has now developed a rhetoric of metonymy to do that work instead.

But to find these transitions, Williams also must segment the film. Instead of Suleiman’s twelve segments, Williams finds eight. There’s no episode focused on the maid, for example; she simply belongs to the Foucauld family. Because Williams is less concerned with a theoretical convergence of language in literature and film and more interested in a circulation of the “figures of desire” of eros and thanatos, her episodes are defined around sex or death, instead of Suleiman’s Subject-protagonist. The maid here is part of a transition from a historical scene of erotic transgression between nations to a modern one within the family.

This accounting works by excluding even more figures from the meaning that we as spectators can produce. What of the servants at work in virtually every Bunuel film, and one servant above all in the later films: Muni, the wonderful character actress who is so overworked by Bunuel in the role of the maid?(32) Returning to her earlier themes of tyranny and convention in her reading of the film in Blood Cinema, Marsha Kinder focuses on this role. Kinder notes that it is the maid who leads us out of a situation where a military man is going to rape a dead woman into one where two women discuss property rights and women’s place in differing cultures. This shift happens through a female voiceover which at the same time as it challenges the dominant convention of giving the narrating voiceover to a male figure, leads to a serious discussion of property and law. The transition here effectively challenges gender conventions and class knowledge, Kinder points out, shifting us into the present day by “introduc[ing] a class discourse about punishment, power, and property.”

Kinder goes on to point out the switching and doubling that go on throughout the film. Extending the idea of paradigmatic variation that Suleiman has used to establish a “radical linearity” in the film, Kinder finds an “array of paradigms” being recombined in the film and reads it as a collection of stories competing for control of the narrative.(33) She gets at how the film mirrors our world by representing competing discourses. Perhaps that mirroring relationship can be characterized by observing the kinds of operations that Bunuel is performing in linking disparate erotic and political narratives (as Williams stresses) at the same time as he shows them to be conventions of form (as Suleiman indicates). Yet, how is it that paradigms can be switched in the way Bunuel allows? More importantly, how is it that we can make the switches that Bunuel sets up? Instances of paradigms are elements … and we are back to the network of effects.

Elements in networks of effect belong to more than one network at once; they have multiple names. Bunuel enables paradigm switching by building narratives as networks waiting for activation, and allowing us to make the new links. The links that Bunuel sets up are made through our recognizing double or multiple figures, such as the two police commissioners who unite to put down a demonstration at the close of the film. A multiple figure might be Muni’s role as maid or laborer in many films. Or we recognize a reversal when, later in the film, a defecating party stands in for a dinner party. Bunuel is not simply negating or rearranging terms; he did not, for example, show a vomiting party. He is reversing the terms that provoke physical feelings of disgust at the same time as they expose the feelings and customs associated with our most fundamental cultural and architectural discourses. Or we might recognize a juxtaposition such as monks seated before an SM scene. Finally, we might see an opposition, such as the doctor denying the authenticity of the letter that the dream postman has left Mr. Foucauld, followed by his granting the nurse a leave by virtue of the letter from her father. These narrative actions are the disjunctive operations by which Bunuel’s network of effects so smoothly accomplishes its work.


None of these narrative actions is simply a formal choice: they are choices made so that the connections we make reveal links between networks which we normally think of as separate, and so find ourselves surprised at the structures that determine so much of our lives. Technically, the continuity, the framing, and raise en scene, apart from the pseudo-random narrative connections that are followed, are a double figure in themselves: they twin the most typical types of filmic conventions. Finally, episodes are repeated with variation as the family is doubled and reappears, or as a chief of police confronts and collaborates with his double. Confronted ourselves with a text that demands connections be made, we make links backward and forward over sound and image, perhaps even laterally into other films. We are free to travel through these narrative networks as we will.

Formulating The Phantom of Liberty as an impossible narrative structured through narrative networks allows us to contextualize previous critical approaches, which now become specific viewer-configured readings of the film. More importantly, this formulation suggests that Phantom might well function as a decoder text for other impossible narratives that have emerged. With impossible narratives we are faced with well-formed works which re-assert the author as an artist in the guise of, well, anything or anyone the author wishes to be, but if not in a specific personage, at least the range of issues or styles that the film screens. These works also function as contradictory audiovisual effects involving the spectator in a deep production of knowledge so that the author becomes irrelevant except as a credited source of the film operating as an actor in a theatre of agency. Faced off in a confrontation with the interests of authority, our recourse is to consider that while the social may function as an imaginary, both our bodies and our representations are real. Our response is to reconsider what our bodies mean by watching in the first place. Our reconsideration performs a replacement of ourselves for the filmic experience we have just produced. We replace the author and the film as the site of agency, as agency is finally a cinematic effect, and after all, it is we who live, or enliven, the cinema. It took Bunuel decades to perfect his approach to film, to film that works en masse to undo the uniformity of masses. If The Phantom of Liberty is still surrealism, it is a surrealism that ironically has required the encoding of cinematic convention at a macrolevel and requires at least a transcultural agreement on those conventions before the fact that they are being overturned can become apparent. Cinema does not invent or prevent empire. Power accommodates and applies cinema. But the masses that cinema produces can become power’s undoing.

Bunuel himself suggests that he avoided circularity in this film because “if it closes itself into a circle, it’s not liberty, it’s death.”(35) Once we find ourselves awake in this cinema of hallucinations, confusions, dreams, and provocations, we are free to make them meaningful in ways that Bunuel could not have predicted. But we cannot make them meaningful in ways that restore the power networks that Bunuel is exposing to an unchallenged position, unless we literally forget what we have seen or refuse to interpret it. The usual relationship of surveillance and subject formation has been reversed: power has become visible, and we are safe in the dark to appraise its effects.


(1.) Aranda, Francisco, Luis Bunuel: A Critical Biography (New York: Da Capo Press, 1975), 249; and, Bunuel, Luis, My Last Sigh (New York: Vintage Books, 1984), 249

(2.) Kinder, Marsha, “The Tyranny of Convention in The Phantom of Liberty,” Film Quarterly (Summer 1975), 25. Kinder notes that the sounds of gunfire and scenes of violent repression which begin the film are echoed in the final scenes, cueing a “circular framework” since the visuals do not depict the same scenes: the film ends as zoo animals presumably watch gunfire directed at contemporary student demonstrators in France, while it opened with Goya’s “The Third of May” which depicts the execution by firing squad of Spanish patriots in Toledo during the Napoleonic Wars.

(3.) Suleiman, Susan, “Freedom and Necessity: Narrative Structure in The Phantom of Liberty,” Quarterly Review of Film Studies (Summer, 1978), 289.

(4.) Ibid., 295: “I would hesitate to affirm [Kinder’s thematics of tyranny] so strongly; the humor of the film, although it can be read as very bitter, can also be read as detached. The word “tyranny” evokes too tragic a mood, I think, for this film.”

(5.) Williams, Linda, Figures of Desire (Berkeley: University of California Press, 1981). “Because tyranny and slavery exist, there is a dream of freedom; because objects are beyond our reach, we desire them; and because stories have continuous linear movements, storytellers want to escape them. In each case the tyranny of law creates the desire for a phantom freedom. The phantom of liberty is born in the existence of chains” (178). Williams wraps up the film with repression creating the desire to escape it. Liberty doesn’t pre-exist tyranny, but is tyranny’s effect on desiring subjects. We cannot interpret here any moral statement about our “condition” or the state of our world: to read the film in terms of an animal revolt and human repression of it is to impose “moral categories” upon the film (177).

(6.) Barthes, Roland, S/Z, Miller, Richard, trans., (New York: Hill and Wang, 1974), 18-21.

(7.) Hubel, David H., Eye, Brain and Vision, (New York: Scientific American Library, 1988), 24. Hubel’s “circuits” of brain cells appear as networked structures in his diagrams which illustrate “neurotransmitters,” “receptors,” “input,” “output.” His figures echo Marvin Minsky’s positing of networks of communicating agents that produce the effects we understand as “intelligence,” “emotion,” somewhere after perceptual processing ends, “memory,” “thought,” and “soul” begin. See Note 11.

(8.) Malamud, Carl, Stacks: Interoperability in Today’s Computer Networks, (New York: Prentice Hall, 1992), 198, Malamud’s elegant “star” topology reminds us of the highly representational nature of the real-world networks that process knowledge. The more mundane “World Wide Web” is a meta-network enabled by a widely agreed upon specification for the transfer of data between computers of different types over a plurality of redundant links (the Internet) that, in this country at least, was designed as a military medium that would ensure the transmission of sensitive information in the event of a nuclear war. Networks with redundant links are “fault-tolerant,” resistant to systemic failure.

(9.) Slocum, Jonathan, editor, Machine Translation Systems, (Cambridge: Cambridge University Press, 1985), 76. Here, the network figure is intended to model our knowledge of language, not the neural processes themselves. Knowledge is held in a network form that is processed via rules of grammar into linguistic surface structures. Elements are linked by specific names which related those elements in a propositional equation which must describe the real-world nature of the referent of the word in question.

(10.) Cottrell, Garrison W., A Connectionist Approach to Word Sense Disambiguation (San Mateo, CA: Morgan Kaufman Publishers, Inc., 1989), 28. A connectionist approach that sets out an explicit metaphor between neural processing and the machine processing of language. Ambiguous meanings are eliminated by “training” wherein the network “learns” how to suppress certain elements to produce the “correct” parse.

(11.) Minsky, Marvin, The Society of Mind (New York: Simon and Schuster, 1986), 17, 168. Minsky arranges “agents” in “societies,” networks of linked elements. He avoids the word network, arguing implicitly that networks are societies. Instead of societies being networked, networks are socialized. Minsky seeks to understand humans as “wonderful machines”, but cautions that “machine” has negative connotations that we are better off forgetting.

(12.) Ibid., 19.

(13.) Ibid., 307, emphasis added.

(14.) Foucault, Michel, The History of Sexuality: An Introduction (New York: Vintage Books, 1978), 30.

(15.) Ibid., 46.

(16.) Ibid., 61: “In Greece, truth and sex were linked, in the form of pedagogy, by the transmission of a precious knowledge from one body to another; sex served as a medium for initiations into learning. For us, it is in the confession that truth and sex are joined, through the obligatory and exhaustive expression of an individual secret. But this time it is truth that serves as a medium for sex and its manifestations.”

(17.) de la Colina, Jose, and Turrent, Tomas Perez, Objects of Desire (New York: Marsilio Publishers, 1992), 216. Others have interpreted this line as “Long live the fallen!” and the film’s English subtitle translates it as “Down with Liberty!” Bunuel himself gives the line as I have included it above in de la Colina and Turrent’s text.

(18.) Bunuel, 249. The director also states that this film is one of his favorites; The Phantom of Liberty in any case is a signature work by the director, including many of his trademark images (sewing paraphernalia, or insects, for example). Bunuel claimed it as his “most surrealist film” in Aranda’s Luis Bunuel: A Critical Biography.

(19.) Suleiman, 282. A sequence is defined as “a limited series of events” after a proposal from Roland Barthes.

(20.) Suleiman, 283.

(21.) Suleiman, 288. Suleiman diagrams the film as a linear progression from one episode, built around a “Subject,” to another.

(22.) Suleiman, 289.

(23.) Eberwein, Robert. Film and the Dream Screen (Princeton: Princeton University Press, 1984), 140-141.

(24.) Bunuel, 121.

(25.) Bunuel describes his reshooting a scene because the excessively stylized way he had designed the shot with his cameraman was “vulgar.” He also describes working out a very stylized movement with his cameraman, with the preparations ending in laughter, and them proceeding to shoot the shot in a much simpler fashion. Cf. Aranda, 198, on the shooting of Viridiana, for a comparison between Bunuel’s purposefully indistinct framing and montage and the spectacular film language of Eisenstein.

(26.) Her figure recalls Madame DeFarge in Dickens’ A Tale of Two Cities, who not only is a recurring figure in the unwinding of Dickens’ narrative but finally turns out to be a subversive figure working to overthrow the French monarchy.

(27.) Foucault, 47.

(28.) Ibid., 46.

(29.) Part of Foucault’s project in the History of Sexuality was to see psychoanalysis as one of the ways that sexuality is produced through confession. The History of Sexuality was originally published in French in 1976, while the Phantom of Liberty was released in 1974.

(30.) Foucault, 83-85.

(31.) Williams, 178.

(32.) Muni plays a maid in Diary of a Chambermaid, Belle de Jour and other films, while in The Discreet Charm of the Bourgeoisie she plays a working class neighbor to the prosperous drug-dealing couple at whose home the film begins.

(33.) Kinder, Marsha, Blood Cinema, (Berkeley: University of California Press, 1993), 332.

(34.) For example, Lynch’s Lost Highway (1997) or Soderberg’s Schizopolis (1997).

(35.) Colina and Turrent, 218.

(36.) The ability of a mystery projected by another to clarify a vision of one’s own world is a constant theme in Bunuel: cf. My Last Sigh, 88.

James Tobias is currently pursuing a doctorate in the School of Cinema and Television, Division of Critical Studies at the University of Southern California. Interaction designer and artist, he was previously a member of the research staff at Interval Research, a multimedia think-tank located in Palo Alto, CA.

COPYRIGHT 1998 University of California Press

COPYRIGHT 2000 Gale Group