Download as PDF Download PDF

Nicolas Roeg

Fragile Geometries


Do you dream in colour or black and white is a question we might choose to paraphrase by musing over which filmmakers think in monochrome or colour? Where critic Paul Coates once astutely noted in The Story of the Lost Reflection that Truffaut’s best films were in black and white, Nicolas Roeg, who actually shot Truffaut’s first colour film Fahrenheit 451, is a director whose work cannot have been conceived without the use of colour photography.  Most obviously the case in the literary adaptation Don’t Look Now, where Roeg changed the thematic colour from Daphne Du Maurier’s melancholic blue to peripherally horrific red, much of the decadence and despair, the exoticism and visceral force of Roeg’s mise-en-scene, generally, comes from the use of colour as visual impact. Whether it is the pink reddish meat at the beginning of Walkabout, the sharp contrasts between the blue sky and the golden yellow sand in the same film, the orange hair of David Bowie in The Man Who Fell to Earth, or the tip that is Milena’s flat near the end of Bad Timing, Roeg is a colour hyperbolist. He creates tension not only through his fragmentation of screen space, through the fragile geometry that indicates time and space are subjective notions as much as fixed realities, but also in his colourism. Roeg’s famously fractured editing maybe appears so partly because of the colour contrasts coming out of that fracturing.

In an article in his book Negative Space, Manny Farber astutely observes that a number of directors absorbed Alain Resnais’ editing structures in the late sixties and into the seventies, and mentions Joseph Losey, John Boorman and Roeg. Yet though Losey’s Accident was of course in colour, and likewise Boorman’s Point Blank, the use of colour isn’t especially disjunctive and from this point of view, from the perspective of montage, both could have been made in black and white without greatly losing their impact. Roeg’s films would however lose much of their effectiveness.

To explain further we may think of David Bordwell and Kristin Thompson’s breakdown of four key aspects of editing – the temporal, spatial, the graphic and the rhythmic. Of the four it is of course the graphic that becomes especially pronounced in colour film, with the filmmaker capable of creating startling colour linkages, or colour contrasts. Bordwell and Thompson mention how Ozu will cut from one similar image to the next to create a smooth flow of graphic consistency, saying in Film Art “Ozu’s cutting is often dictated by a much more precise graphic continuity than we find in classical continuity style.” in Ohayu Ozu uses color…cutting from laundry on a line to a domestic interior and matching on a red shape in the upper left of each shot (a shirt, a lamp)”. Though such matching might not be as common in Ozu’s work as Bordwell and Thompson suggest, it offers a useful way into understanding Roeg’s ‘graphic radicalism’ and helps us distinguish what Roeg was doing from Resnais and Losey and Boorman. Few filmmakers were pushing graphic montage in narrative film further than Roeg.

Whether it is the opening scenes in Performance, Walkabout or Don’t Look Now, the consistent element lies in the graphic, for in relation to the spatial and the temporal each film works differently – the graphic is the common denominator. At the beginning of Performance the film cross-cuts between the two leading characters who have not yet met, and a judge, as we try to find our way into the film’s spatial and temporal organization. In Walkabout this is slightly less complicated as we can work out we’re watching a daughter and son at school and walking home. Don’t Look Now superficially creates the least problem of all, and functions spatially almost conventionally, as it cross-cuts with the kids playing in the garden and the parents sitting in the house. Yet, graphically, the film is Roeg’s most experimental as he works the idea of second-sight into an enquiry about film form. At one moment we notice the match between the girl walking across icy water, and the brother cycling over a pane of glass that breaks; at another the mother offers a particular hand gesture that is matched by the daughter’s similar one in the next shot. The daughter’s red coat is matched both by the flames in the fire place, and the red blotch on the image the father is working with.

This is graphic matching however not to generate smoothness, but to emphasise the fragile connection between things. Generally speaking graphic matches firm up the relationship between different shots, adds to the plausibility from one image to the next. Thus when Bordwell mentions Ozu’s graphic matches he does so to help explain the smoothness of the Japanese master’s aesthetic. In Don’t Look Now, the graphic matching brings out Roeg’s interest in the fractured interconnections between things. As he says, “all knowledge and all things are connected”, and yet at the same time Roeg also believes that his approach to cinema in films like Don’t Look Now lies in “making you aware of danger”.  (SFX, Aug. 1999). Hence the fractured aspect.

Connection and fragmentation are the push me and pull me of Roeg’s aesthetic. He wants to make the disparate connect and the similar estranged. In The Man who Fell to Earth there is a sex scene between Rip Torn’s professor and a student, cross-cut with Bowie’s character watching a Japanese sword fight. Here Roeg creates a halfway house between a cluster of meaning and a clutter of imagery. This is perhaps intellectual montage as Sergei Eisenstein might have meant it, where an idea comes out of the juxtaposition of images, but it is as though Roeg is as interested in the clutter of images as the cluster of meaning. If Roeg is often seen as an irrational filmmaker does it not lie in the non-rational cutting between one image and another, and the graphic disjunction accompanying it?

If one thinks of Eisenstein’s Battleship Potemkin and the famous cross-cutting between the troops coming down the steps and the locals being mowed down, this is strictly speaking implausible in reality but intellectually and ideologically coherent. Commentators may have noted that it takes less than half the time to march down the stairs as it takes the troops in the film, but the ‘irrationality’ of time is compensated for by rationality of impact. The viewer knows that Eisenstein’s purpose is to hammer home the unremitting power of the troops and the helplessness of the crowd. At the same time, Eisenstein gives a rhythmic coherence to the shots as he cuts between implausibly linked spaces, but with a plausible purpose.

Roeg often removes the ideologically and intellectually coherent for the visually curdling and dramatically inexplicable. It is entirely justifiable that Roeg cross-cuts between Bowie and Torn – for soon Torn will be working for Bowie. It could basically have been an expositional cross-cut, the sort one often sees in a film where two characters who haven’t yet met are soon, say, to ‘meet cute’ in a romantic comedy, with the viewer knowingly aware that we don’t need to be told verbally about this forthcoming ‘appointment’ by some omniscient voice-over, but that we can work it out visually instead. At the same time the film might play on the inevitability of this meeting by the loneliness of each character, the sense that they’re both at a loose end the filmmaker will tie together. Roeg’s juxtapositions are much more problematic than either Eisenstein’s or the conventional filmmaker’s for a variety of reasons. The combination of his montage colourism, that we’ve already touched upon, his dramatic irrationality and what we might call his temporal distrust, all create the Roegian world.

Let’s look at the idea of temporal distrust, and link it to the sort of distrust Gabriel Josipovici addresses in much 20th century fiction in his book On Trust. In his introduction he quotes T. S. Eliot’s idea that “last year’s words belong to last year’s language/And next year’s words await another voice”. Eisenstein’s work is in this sense last year’s language: as though it could incorporate the radical confounding of space that so startled many of those initially viewing cinema, but that could be put back together again, almost humpty-dumpty like, by an intellectual and ideological coherence. Anybody watching the Odessa Steps sequence in Battleship Potemkin might not be able to offer a plausible mapping of the filmic space, but they would have little problem making sense of what that fracturing of space meant.

What happens however when the space becomes ‘distrustful’ partly because “we’re inventing what is the truth”, in Roeg’s words? Then a cross-cut like the one between Bowie and Torn becomes less the objectified account of two characters soon to meet, than the collision of two subjectivities. The cross-cut seems to offer not the objective filming of the present that gives us a perceptual assuredness in a future event, but more the opposite. It is a perceptual unease generated out of a montage that senses a future event rather than predicts it. In relation to trusting the image the way we might in a romantic comedy clearly anticipating a future meeting, the film plays fair with our perceptual faculties as it creates a high level of probability. Roeg instead seems to want us to trust our senses: the edge of ourselves, if you like, rather than our perceptual dead centre. Where Eisenstein de-centred us perceptually but re-centred us ideologically, Roeg is interested in playing with our peripheries, creating a constant sense of perceptual doubt. We cannot trust the image partly because time doesn’t belong to the world so much as it belongs to the self. It is a variation on Henri Bergson’s distinction between clock time and personal time, time as divisible units, and time as continuous flow. “Dive back into the flux”, Bergson tells us, “if you want to know reality, that flux which Platonism, in its strange belief that the immutable is excellent, has always spurned: turn your face towards sensation, that fleshbound thing which rationalism has always loaded with abuse.” The statement could almost be Roeg’s. As Bergson questions determinism, as he questions the idea that the same motives acting on the same person always arrive at the same result, one can note the determinism at work in the romantic comedy where the filmmaker cross-cuts and the viewer awaits the inevitability of the meeting. The term meeting cute might indicate contingency, but it contains within it the givens of determinism. The sort of distrust Roeg generates is anti-deterministic: it is the idea that each subjectivity is caught in the flux of life, not the mechanics of existence, and Roeg’s style captures it. Hence the temporal distrust. This also helps explain why Roeg says in interviews that we are constantly in danger, every moment of our lives. Roeg isn’t a staunch pessimist; more someone who sees that with life non-deterministic, then how can we create the certitude that at some moments we are very safe and others in great danger? Many filmmakers create determined worlds that put us consistently in danger, consistently in places of safety. Roeg wouldn’t trust such determinations.

Out of such ontological insecurity, out of seeing the world as constantly in flux, Roeg creates a certain dramatic irrationality, and this ties into the montage colourism. If we look at the first couple of minutes Performance, we can see how this works. In Performance Roeg opens with a supersonic jet crossing the sky, cuts to the pilot’s point of view, and then to a black Rolls Royce driving through the countryside, before crosscutting between the car and a couple making love. In such montage choices Roeg (and in this instance his co-director Donald Cammell) isn’t setting up narrative coordinates; he is generating a collagist chaos. Certainly later in the film we’ll make certain connections between the images we’re shown, but it is as if Roeg wants to work from a climate of confusion, an image structure that isn’t asking for greater cognitive effort, necessarily, but for one to be absorbed in the “fleshbound thing”. To apply the same rules to a Roeg film as one applies to the determinisms of a romantic comedy is to miss the significant differences: the more disjunctive the approach; the more suggestive the narrative. Obviously however Roeg is still working with narrative. The characters in the Rolls Royce are gangsters, and James Fox’s character, the man making love, is himself a gangster. The images aren’t entirely arbitrary. But that doesn’t mean they are entirely appropriate either.  Roeg’s films are like many loosely modernist works interested in what Gilberto Perez in The Material Ghost astutely sees as the appropriateness of arbritrariness in much modern cinema. What counts isn’t cognitive assertiveness, but speculative possibilities. As Perez says, “appropriateness rules out arbitrariness. Classical art dwells in appropriateness…” On the other hand, “The art of modernism, a precarious art, lies in finding the appropriateness of its arbitrariness, the appropriateness that will hold at bay, hold in suspension – maybe just for a moment, and that moment is the work of art – an arbitrariness that cannot be ruled out.” Is Roeg not a filmmaker interested in the balancing act between the appropriate and the arbitrary as he works with the appropriateness of clock time and the arbitrariness of time as flux? This seems to be the problem he returns to again and again.

To explore this further, we can look at Don’t Look Now not specifically as a fright film about second sight, but a problem of temporal imprecision as the expression of feeling. John Baxter is the grieving father of a daughter who has recently died in the freezing family pond, and Baxter and his wife Laura go to Venice to escape from it all as he works on some church frescoes. Yet what he can’t escape from are his own premonitions, one of which came to him moments before his daughter died. When he rushed out into the garden this wasn’t based on the empirical detail: a look out of the window, a shout from his son. It was no more than a feeling, and it is these feelings which keep coming back to him throughout the film as he envisages his own demise. From the perspective of clock time, the events make no sense: how can he have a response to an action that is still to take place? But from the angle of time as flux, why should he not? Perhaps it is true Roeg pushes this problem a little too far into the paranormal and the generic, but are there not certain environments, certain moments, where one senses more than clock time at work? We needn’t believe in second sight as such to know that to remain too rigidly within the limits of time as it is usually perceived, restricts one’s perceptual faculties. What finally works in Roeg’s film isn’t the plot about premonition, but the premonitory atmosphere he creates, the sense that out of the spaces we occupy we can read time that is greater than the present moment.

This is of course a horror generic mainstay. It constantly works with the possibility of terror not based on the givens of the situation as in an action film, where the cross-cutting is usually and promptly actualized in action, but in potentialised spaces. In such potential spaces for terror, the filmmaker usually then actualizes them after creating the suspense of potentiality. Horror more than any other genre works with this temporal imprecision as characters have strong emotional reactions without always knowing quite what they are reacting to. Time is often out of joint in this sense in the horror film, as characters anticipate events not especially as premonitory, but certainly where reaction precedes action: here feeling frightened precedes the reason for that feeling of fear. This is all central to Don’t Look Now, but where one senses most horror films work with the generic form that contains this problematic of cause and effect within it, Roeg turns to horror in Don’t Look Now as the problem coincides with one that fascinates him in all his films.

Thus we might also note that while Don’t Look Now is a horror; the editing strategies Roeg adopts are much closer to his work generally than the horror film specifically. The difference lies partly in horror cinema’s interest in creating anticipatorily empty spaces, as opposed to Roeg’s fascination with fracturing screen space. When Roeg says that danger lies everywhere, and also insists that “we try and discipline ourselves to think that we are not living in a science fiction environment, but we are, the whole time”, he is invoking a different genre from the horror movie, and of course the genre he would tackle in his next film, The Man Who Fell to Earth, but whether sci-fi, horror, gangster film or love story, the genre disappears into film form. Roeg for example does not generically fill these spaces for us, as one would expect with a horror. He makes all the spaces he films potentialised, as though any actualization, any shock or horror within the shot, cannot quite eradicate the amount of potentialisation he has given it. Now as we’ve proposed, the action film creates very little potentiality but a high level of actualization, with the horror film creating potential spaces that are then actualized at key moments. In Roeg’s films however the actual cannot quite dissipate the potentializations created. It is why we talk of the temporal incoherence of Roeg’s worlds, that temporal distrust.

To explain further let’s take a closer look at the scene just after John and Laura have made love. They go down stairs into the foyer where the restaurant is and announce that they’ve decided to eat out for dinner. This is end of season Venice, with the hotel soon to close, and the restaurant kept open for the odd diner. The impression given is that the Baxters were the only people likely to eat there that night, and now the staff are at a loose end. Roeg holds on the shot for a few seconds as the manager looks deflated, and then cuts to a shot of a room in the hotel covered in sheets. The film slowly dissolves to black, and then the next shot is of the waters of Venice, before the film follows the Baxters trying to find their way to the restaurant. Initially they can’t find it as Roeg utilises a number of cuts and zoom shots to emphasize this sense of spatial dislocation. At one moment Baxter stands next to the canal and rubs his hand against a wall and says, “I know this place”. John might think he is looking simply for a restaurant, but Roeg indicates that he might be looking at his own future demise. Moments later while he’s talking across the canal to Laura, John hears a strange noise and someone opens their shutters, shouts down, and we notice a small figure crossing a path in the distance, wearing a red coat. At the beginning of the film his daughter had been wearing a red rain coat when she drowned, and one might notice at the end of the scene someone coming out of a building with their daughter in their arms, wearing red. In such a scene Roeg creates a sense of unease by refrains, non-sequiturs and inexplicable scene shifts, and yet one senses nothing so strong as a formal pattern as we try and work out the connections between things. Why does Roeg stay with the hotelier so long, and why then cutaway to the disused room? Why does he then fade out the shot, and why does he show a couple of empty shots as the couple try and find their way onto the main street?

Roeg creates nothing quite like a formal pattern, then, but at the same doesn’t simply tell a story, and one of the biggest problems in relation to much formalist criticism is the insistence to see a pattern where there isn’t necessarily an obvious pursuance of narrative. For example Edward Branigan writing on Ozu, or Bordwell writing on Bresson, must see a pattern to the form.  Talking about the presence of soy sauce bottles in a scene from Ozu’s An Autumn Afternoon, Branigan in Screen magazine notes “thus objects (and persons) in Ozu are seen not as subordinate to, or reflecting, a theme or ideological stance, but a complex organisation of space that reveals not meaning in the ordinary sense, but geometry and symmetry in a state of tension.” Bordwell reckons, meanwhile, in Narration in the Fiction Film, that “without leisure to linger, the viewer may also attribute ineffability to the ungraspable differences generated by the parameter and cannot keep them all in mind at once.”

In each instance, formal rigour replaces the denotative and connotative levels: the story and the theme; what the film shows and what it means. But maybe it is better to think not of formal rigour, but formal suggestiveness, so that the deviation from a perceptual expectation doesn’t lead to finding a pattern in the carpet, but a plausible explanation of the feeling. If Roeg’s films create a high degree of perceptual unease in the viewer, the critic’s job is to find the intuitive sense that helps explain this feeling. Great filmmakers work with what we might call affective problematics; they try to find ways in which to make us feel without relying on Pavlovian conventions. In the scene at the hotel and by the canal, Roeg could have shown that Baxter was simply having a premonition, and left it at that. It would have served the plot function of John becoming increasingly aware of his second sight. In such an approach form can serve the plot function, but Roeg’s interest would seem inside out: the plot function, the generic expectation, serves the affective problematic. When Roeg says “I think we try and discipline ourselves to think that we are not living in a science fiction environment, but we are, the whole time. It must be rooted somewhere in our souls because there’s no imagination without experience”, this is basically the director exploring his affective problematic.  Making sense of Roeg’s work seems to require balancing the generic, narrative expectations and the formal innovations. If one looks too readily for formal patterning it can bury the affective problematic just as readily as a filmmaker can do the same to his own work by working too solidly within genre expectation.

This returns us to several aspects we’ve raised: Roeg’s graphic radicalism, his spatial potentializations, and how they fit into the affective dimension. If we can assume, taking into account Roeg’s own comments and our own speculations, that Roeg is fascinated by the attempt to make meaning out of chaos, and yet retain the sense of chaos in this meaning, then we can see how graphic radicalism and spatial potentializations can generate this problematic. There is another scene in The Man Who Fell to Earth a little like the one quoted above. Here Bowie falls ill in a lift and Candy Clark’s character puts him into bed. At the same time Roeg cross-cuts with Torn again, on this occasion in bed with two women, but we may notice that this is obviously covering two different evenings as each duplicates the other’s actions, commenting on Torn’s penis. While Torn is lying in bed though as the women attend to him; Bowie is also being attended by Clark – the former sexually; the latter considerately, medically. Now when Roeg first cuts from Clark and Bowie to Torn and one of the girls, we might initially wonder whether it is a cut to another scene, or momentarily wonder whether, because of a semi-matching cut from Clark to the young woman, it is Clark naked getting into bed with Bowie. Moments later we again might wonder whether the first woman is the same as the second, only this time she is wearing glasses, before we work out that this is yet another of Torn’s lovers. In each instance Roeg demands momentary shifts in perspective, so that instead of the relative matching achieving a higher degree of harmoniousness (as we find for example in the graphic matching in Ozu’s work), Roeg generates a high level of disconcertion where a match from one shot to the next creates a perceptual quickening that accelerates rather than alleviates unease.

For example in Don’t Look Now, just after Laura realises her daughter is dead, Roeg cuts from her scream in the garden in England, to the sound of a drill in Venice – and audio matches the scream with the drill. This obviously resembles a similar audio cut in, say, Once Upon a Time in the West, but where in Leone’s film the gun going off audio matched to the whistle on the train is a knowing gag, however horrific (a young boy has been shot), it creates a certain comfort in the form. Roeg searches out matches, whether visual or aural, that generates disquiet. Sometimes he will even do it by the sheer overloading of the match. Near the beginning of Eureka, the central character played by Gene Hackman is lying in the snow and the moon is more or less matched across a series of images with a block of ice, a sliver of the moon reflected in a crystal ball, and a glass ball showing a Hackman figurine inside it.  All of these images create not a sense of cinematic ease (as in Once Upon a Time in the West), nor graphic harmony (as in Ozu), but adds to Roeg’s interest in making the world not what it seems as we try to make sense of the associative editing.

Space, meanwhile, is rarely waiting to be filled through dramatic expectation; more by associational possibilities. It makes sense that Roeg would have chosen all the books in Milena and Dr Linden’s apartment in Bad Timing, and that “props have a life of their own.” Roeg as we’ve proposed utilises space as a dramatic multiplicity rather than as narrative singularity. Not only in the way he potentializes them as he refuses categorically to dramatize the space utilised, but also by creating so much meaning with the spaces he films. The use of color adds to this epistemological clutter, as especially in Performance and Bad Timing, the flats become mosaic like in their potential meaningfulness. Just as earlier we noticed that Roeg potentializes spaces without actualizing them, contains them within a certain type of dramatic emptiness, so he also does the opposite with a similar effect: to disorientate. The props in Roeg’s films are not simply part of the spatial appropriateness necessary for most film drama – a space realistic enough for no one to question its plausibility, but at the same not dense enough to cause us to look away from the dramatic action. Yet it is the density that so often fascinates Roeg. Frequently he will use the zoom to animate a prop, to give it an obscure significance, evident in various books Milena and Alex read in Bad Timing, the wide-angled shots of the brightly lit objects Jenny Agutter’s character lays out for the picnic near the beginning of Walkabout, the fragile geometery book on the table in Don’t Look Now. This is basically the animation of objects, but not as the prop that becomes a narrative device, like a candlestick holder suddenly wielded but as a speculative possibility. When Roeg asks, what is the first thing you do when you go to someone’s house: you go straight to the bookshelves; he is talking of the way we create speculative meaning around the lives of the people we know. This seems to be the density of meaning Roeg searches out, and yet this is a meaning so deliberately packed with possibilities that meaning cannot readily be extracted. Yet this doesn’t arrive at the meaningless a critic often assumes when meaning isn’t singular enough; its density is an invitation to speculation. Perhaps it is almost as absurd criticising a Roeg film for its richness of meaning, as it would be to attack a friend for the signs of meaning that they give off in their furniture, books, eating habits etc. “I think that what people call a usual setting”, Roeg reckons, “is always a fragile idea”. He is talking here specifically about the exotic locations he often uses, but he also opens it up into a comment about identity and society. “What comes first?” “Obviously identity comes first”; yet many filmmakers create a mise-en-scene that is dramatically meaningful but not especially personally so.

Roeg then is a meaningful filmmaker, but at the same time by working with layered meaning he also toys with the possibility of the meaningless. The film’s meaning isn’t readily extracted as a viewer objective, more a creative activity. Indeed when Roeg and his scriptwriter Paul Mayersberg were asked years after it was made what Eureka meant, and why it was badly received when it was released in the early eighties, they re-entered the work speculatively, as if trying to make sense of it themselves. A filmmaker working with the three act structure of narrative form, and the sort of editing so expertly outlined by Pudovkin in his book Film Technique, where he mentions montage as the building blocks of screen time and space, wouldn’t need to re-enter the work in the same way. Roeg’s editing practises that seem constantly to be generating radical graphic matches and contrasts, and the spatial organization that refuses often fulfilling the potentialization of that space, or so clutters it that meaning cannot singularly be extracted, offer dreams and nightmares in colour. But each dream, each nightmare, may register a little differently with each viewer. As Roeg says “I think I always somehow, subconsciously sensed that nothing is what it seems.” Few filmmakers, whatever Roeg’s limitations, have captured in editing and mise-en-scene the significance of that comment. Roeg, who may be significant for no more than six films – Performance, Walkabout, Don’t Look Now, The Man who Feel to Earth, Bad Timing and Eureka – all made within a fifteen year period, might not believe in the singularity of the self, but he certainly affirmed the singularity of cinema as Roeg revelled in the freedom he possessed for a small period of time. As Alexander Walker so nicely put it in Hollywood, England, Roeg was a director “fascinated by the “seemingness” of things,” revealing ”the treacherous world of appearances out of which a society is composed.” Would a filmmaker more stable in their perception of reality have been likely to so deliberately find a form which to destabilise that social assumption?


©Tony McKibbin