David Bordwell's website on cinema   click for CV
    %62or%64%77e%6cl%40%77%69%73c%2e%65%64%75

Home

Blog

Books

Film Art: An Introduction

Christopher Nolan: A Labyrinth of Linkages pdf online

Pandora’s Digital Box: Films, Files, and the Future of Movies pdf online

Planet Hong Kong, second edition pdf online

The Way Hollywood Tells It pdf online

Poetics of Cinema pdf online

new! Chapter 6 | Film Futures pdf online

Figures Traced In Light

Ozu and the Poetics of Cinema pdf online

Exporting Entertainment: America in the World Film Market 1907–1934 pdf online

Video

CinemaScope: The Modern Miracle You See Without Glasses

How Motion Pictures Became the Movies

Constructive editing in Pickpocket: A video essay

Essays

A Celestial Cinémathèque? or, Film Archives and Me: A Semi-Personal History added September 2014

Shklovsky and His “Monument to a Scientific Error”

Murder Culture: Adventures in 1940s Suspense

The Viewer’s Share: Models of Mind in Explaining Film

Common Sense + Film Theory = Common-Sense Film Theory?

Mad Detective: Doubling Down

The Classical Hollywood Cinema Twenty-Five Years Along

Nordisk and the Tableau Aesthetic

William Cameron Menzies: One Forceful, Impressive Idea

Another Shaw Production: Anamorphic Adventures in Hong Kong

Paolo Gioli’s Vertical Cinema

(Re)Discovering Charles Dekeukeleire

Doing Film History

The Hook: Scene Transitions in Classical Cinema

Anatomy of the Action Picture

Hearing Voices

Preface, Croatian edition, On the History of Film Style

Slavoj Žižek: Say Anything

Film and the Historical Return

Studying Cinema

Articles

Book Reports

Observations on film art

Archive for the 'Narrative strategies' Category

What-if movies: Forking paths in the drawing room

Dangerous Corner (1934).

DB here:

We habitually indulge in what-if thinking. What if you hadn’t gone to that particular school, met those specific friends, lived in that particular place? Your future would have been very different, in ways you sometimes speculate about. Here is Brian Eno explaining how he found his career:

As a result of going into a subway station and meeting Andy [Mackay], I joined Roxy Music, and as a result of that I have a career in music I wouldn’t have had otherwise. If I’d walked ten yards further on the platform or missed that train or been in the next carriage, I probably would have been an art teacher now.

We think this way on a small time-scale too. If you had left that damned parking lot a little earlier, you wouldn’t have had the fender-bender you had down the street.

Just as flashbacks exploit our common-sense intuitions about memory, other narrative strategies tap our habit of what-if thinking. Some movies evoke alternative but parallel fictional worlds. Hong Sangsoo has explored this possibility with unusual energy although he’s less explicit about setting up the structure than these Hollywood films are. With his movies, sometimes you don’t realize you’re in a parallel-world plot until you notice repetitions of action with tiny differences. (We have entries on Hong here and here and here.)

The most recent what-if movie I know is Edge of Tomorrow, whose tagline and video release title, Live Die Repeat, sums up its premise. I thought it was an ingenious use of the format, although the ending left me puzzled. Earlier on this site I wrote about a more intriguing example, Duncan Jones’s Source Code. Sometimes I call these “multiple-draft” plots because they keep revising the action until it comes out right.

Some years back I wrote an essay, “Film Futures,” in which I analyzed the what-if, or “forking-path” narrative. That essay, revised for the book Poetics of Cinema, is now available on this site. It explores several examples: Kieślowski’s Blind Chance, Tykwer’s Run Lola Run, Wai Ka-fai’s Too Many Ways to Be Number One, and Peter Howitt’s Sliding Doors.

One Hollywood experiment in this vein was a film version of J. B. Priestley’s 1932 play, Dangerous Corner. I mentioned the play in the essay, but I wasn’t then aware that a film version had been made by RKO in 1934. (To add an extra sting to my ego, it was sitting in our massive collection of RKO movies on campus.) I learned of it just recently when it aired on Turner Classic Movies, a national treasure I have celebrated before. The film quickly showed up online at Rarefilmm, and probably elsewhere.

In the essay, my approach was to treat these films as a sort of genre. What conventions rule them? What motivates the forking-path format—a science-fiction device such as  a time machine, or fortunetelling, or something else? How do they tap our what-if thinking? Dangerous Corner lets me test my proposal on a new instance and offer a trailer for a new online essay.

As with any comprehensive narrative analysis, there are spoilers.

 

They have been here before

Darkness. We hear a gunshot and a woman’s scream. The stage lights come up and reveal some women in a drawing room listening to a radio play, “The Sleeping Dog.” Soon they’re joined by their male partners. An inadvertent remark by one of the women starts a cascade of confessions. The couples reveal a seething mass of illicit affairs, drug addiction, and repressed sexual desires.

As a result of the frenzy of truth-telling, the husband who set the process in motion lurches offstage and shoots himself. Darkness descends; a woman’s scream. When the lights come up, we are back in the drawing room. The broadcast play is at the same point as before. This time things go differently, and music from the radio fills the room as the couples enjoy a banal evening.

The action centers on three men who are partners in a publishing house. Robert is married to Freda, Gordon is married to Betty, and Charles is unattached. The young woman Olwen works at the firm, and Robert’s dissolute brother Martin is dead when the plot begins. The action centers around some missing bonds, which either Robert, Charles, or Martin stole a year earlier. Soon after the bonds went missing, Martin was found shot dead, an apparent suicide. He was assumed to have been the thief.

In the course of evening number one, all sorts of naughtiness are revealed. Martin, much loved by all, is revealed to have been a thrill-seeking drug addict with whom Robert’s wife Freda has been having an affair. Olwen is secretly in love with Robert. Gordon’s wife Betty is Charles’s mistress. Gordon in turn is in love with Martin, and we’re to understand they’ve had an intermittent gay affair. Martin has died not by suicide but by accident, when Olwen was struggling to escape from his attempt to rape her. In all, the three publishers’ private lives would suffice for a steamy best-seller.

The point of Priestley’s play is that revealing the truth is a risky business, like driving around a dangerous corner. He uses the forking-path format to suggest the harm of revealing things best kept hidden. Hence the radio play’s title, a reference to letting sleeping dogs lie. For Priestley, however, the parallel-worlds conceit was more than an artistic device. He insisted that Dangerous Corner not be regarded as a dream play but rather “a What Might Have Been.” It proceeded from Priestley’s deep interest in time, which he saw as not merely the linear, “once-and-for-all” track of daily life.

Our real selves are the whole stretches of our lives, and . . . at any given moment during those lives we are merely taking a three-dimensional cross-section of a four-, or multi-dimensional reality.

The same interest in time as split or looped is seen in his 1937 plays Time and the Conways and I Have Been Here Before.

The film version of Dangerous Corner makes some important changes. As you’d expect for a film of the 1930s, the gay plotline and the drug addiction are excised. The character of Charles (Melvyn Douglas) is made more virtuous. He is revealed as the thief, as in the original, but here he has stolen the bonds to help Betty pay off gambling debts. She is no longer his mistress but a friend he is protecting. In addition, Charles is shown pursuing Ann (the play’s Olwen), who turns aside his proposals of marriage. At the end of the film, she agrees to marry him, providing a romantic wrapup.

The first seventeen minutes of the film establish the Charles-Ann courtship and present portions of the play’s backstory. We witness the three men discovering the theft of the bonds, and soon we watch Charles’ discovery of the dead Martin. An inquest declares the death a suicide, and a year passes. Now begins the play’s opening situation, with the women in the parlor. But there’s no longer a radio play running; instead, they’re listening to music before they hear a gunshot. That turns out to be the result of Robert’s firing his pistol into the garden to show it off to Gordon and Charles.

Coming into the drawing room, the men pair up with the woman and banter with their guest, the novelist Maude Mockridge. (She’s in the play as well.) The radio becomes important when Gordon goes to it to tune in some dance music, but the tube fails and, as he puts it, “I guess we’ll have to talk.” From then on the film follows the general contours of the play, and I think it exemplifies the conventions of forking-path plots pretty well. What are those conventions?

 

Using the correct fork

In the essay I start by suggesting that the action in forking-path plots is understood to be linear. Within each track, there is a smooth progression of cause and effect. Both play and film obey this condition through the simple chronology of scenes, but it’s also controlling the puzzle of the missing bonds, which is eventually explained by detective-story logic. Charles admits to taking them, and in the film Betty further explains that he did it to cover the gambling debts she wouldn’t confess to Gordon.

Linearity is also reinforced by a simple either-or switch. In the film’s tell-all version, Gordon can’t play dance music on the radio because there isn’t a spare tube for the radio set. Then begin the exposures of all the peccadilloes. In the alternate-reality version, there is a spare tube in the drawer, and so the exposures can be averted. Tube there: certain things follow. Tube not there: other things follow. Each chain of events proceeds without break or further splitting.

The radio tube is part of the film’s use of a second principle, what my essay calls signposting. If we’re to understand that there are alternative plotlines, we need some clear markings. In the film, we get several.

First there is the “reset” moment in the second version, when we return to the women moving to the French windows and hearing Robert’s firing of the pistol into the garden. That is a pure replay. What follows reiterates the action we saw earlier: The men joining the women, the radio announcing the time, and the initial chat before the signal fails and Gordon looks for—and this time, finds—a fresh radio tube. He thoughtfully reiterates the split for us. Dancing with Betty, Gordon says: “If Freda hadn’t had that spare radio tube, there wouldn’t have been any dance music, and then—well, anything might have happened.” As I suggest in the essay, characters in forking-path plots often get quite explicit about the what-if premise.

More unusual is the film’s use of intertitles as signposts. Robert, overcome with despair by the results of his relentless demand that everyone confess, runs to his room and shoots himself. The screen goes dark, with traces of smoke. Then we get this title:

At face value, this intertitle says that the stretch of time in which the characters exposed their private lives (the first “This”)  is fictitious, while the amicable, truth-concealing version we’re about to see (the second “This”) is veridical.

Is this a bit of hand-holding for an audience that wasn’t prepared for the forking-path device? It does have that function, and to our taste it’s probably too explicit signposting. But it’s more interesting than it appears, because it reverses an intertitle that appears in the film’s opening.

Before the action starts, we get this expository title.

Even without knowing how the film will develop, we are invited to imagine a two-part structure. After seeing the whole film, we can see that this puts the dual worlds on the same footing. The phrasing could be suggesting that the cascade of admissions we’ll see is what really happened, while the smooth social veneering of the final scenes was only an alternative possibility. At the climax, we’ll see the second title as a revision of this slightly puzzling one, which now might seem a bit of playful misleading.

But I think this first title is ambiguous in an intriguing way. The film really has three parts. First there are the 1933 events outside the drawing room, involving the robbery and Martin’s apparent suicide; these scenes aren’t in the play. Next block is the first, confessional versions of the 1934 evening. The third part is the alternative, calm version of the 1934 evening. From this angle, the first title is telling us that the 1933 section leading up to the crucial evening is “what really happened,” and that is indeed the case. We don’t get alternative versions of the robbery or Martin’s death. Accordingly, the sordid first iteration of the evening, the film’s second part, becomes “what might have happened.” In a weird way, the title is accurate about the first two chunks of the plot.

In other words, I’m suggesting that the titles fit the film’s structure as follows. The first title is in red, the second in green.

“This is a story of what really happened…”: The 1933 section.

“…and what might have happened.”/“This is what might have happened…”: First, scandalous version of 1934 evening events.

“…this is what did happen.”: Second, banal version of 1934 evening events.

Anyhow, two versions of the evening are quite enough for us. We can imagine more, but in films we never really encounter the radical plurality of multiple worlds. The physics of a true multiverse would offer indefinitely many variants, including ones in which any particular person doesn’t exist. In some worlds, Gordon would be married to Ann, Robert would court Betty, Charles would be a terrier, and so on. But—and here’s my third principle of such storytelling—the usual forking-path plot revisits essentially the same story world with the same characters, relationships, and settings, and most of the same actions. I call it the principle of intersection.

Intersection assures that we don’t get overwhelmed by having to meet a raft of new characters, figure out new settings or time periods, and generally reorient ourselves with each path we’re led down. Forking-path plots, from A Christmas Carol to It’s a Wonderful Life, keep things simple for us by changing very a few features of their rival worlds. Despite Priestley’s belief that each instant of our lives opens up many alternatives, that gigantic exfoliation of actions is very hard to dramatize and even harder for us to keep track of. So in both play and film, we have the same small world, with only a few differences. That works to the plot’s advantage, because those small differences—the radio works/ it doesn’t work, the cigarette box attracts attention/ it doesn’t—can be given enormous importance.

 

More rules

A fourth principle of forking-path plots involves the use of cohesion devices—elements of story or narration that smoothly link scenes. Cohesion devices are mid-sized examples of Hollywood’s love of continuity on every scale: cause and effect across the whole plot, foreshadowing of a later scene by an earlier one, hooks between sequences, and at the finest grain, shot-to-shot matching. Accordingly, each path in a forking plot ought to lead us along as easily as a normal plot would.

The essay mentions appointments and deadlines as common cohesion tactics. These aren’t very prominent in either the play or the film version of Dangerous Corner, chiefly because the forked action takes place in such a limited time span. But other devices, such as dissolves and fades, help the parts stick together. In the film a flashback dramatizes what Ann says took place on the night of Martin’s death in a way that fits tidily into the overall arc of the plot.

More interesting in the film are the parallels, a common byproduct of forking-path construction. At a macro-level, the two or three or however many paths are sensed as equivalent, variant versions that the viewer is invited to liken or contrast. And both our play and film reiterate the parallels among the couples; the visiting novelist Maude Mockridge says in the film that Charles and Ann should marry and complete the perfect symmetry set up by the “snug” pairings of the two other couples.

In the course of the first night’s exposures, hidden parallels are brought to light. Ann is mutely in love with Robert, who is mutely in love with Betty. Both the Freda-Robert marriage and the Gordon-Betty one are revealed as loveless. In the play, more parallels are piled on: Both Freda and Gordon have Martin for a lover, while Charles keeps Betty as his mistress.

Since parallels invite us to note contrasts, we can see that Phil Rosen (never accorded the status of an auteur) has somewhat altered things during the replay portion of the climax. Some variations in framing mark the second version as different from the first.

     

Mostly the purpose of the new angles in the replay is to highlight Charles and Ann, preparing us for their romantic alliance on the patio. The first version makes them small and far off-center left; the second emphasizes them much more, with the high angle enlarging and centering their flirtatious encounter.

     

My two last principles are also borne out in the film version of Dangerous Corner. One says that the last path traced presupposes the others. By that I mean it can take the earlier iterations as already read.

One result is that later versions of events can presented more briefly. When an alternative reality runs through bits we’ve already seen, we don’t need to see the full original version. This happens in the second fork of the film, which takes less than three minutes to get us to the crucial split—when Gordon discovers the fresh radio tube and tunes in the dance music. The first iteration of the evening’s events took almost five minutes to get us to the same point. This may seem a minor difference, but such intervals matter in a movie whose action consumes only sixty-two minutes on the screen.

Moreover, one time-saving passage shows the importance of the first iteration as setting up the situation. In the first version, Maude the novelist inquires about Martin, and she bends over a portrait photo of him.

     

This shot is less for her than for us, introducing the character whom we’ll see in Ann’s flashback. During the second version, we don’t get the full discussion between Maude and Freda, and we aren’t shown the photo. The narration assumes that after the flashback Martin is vivid enough in our memory. What replaces the insert of Martin is a reverse shot of Freda, describing Martin.

     

Now that we know she was Martin’s mistress, we’re in a position to appreciate how her dialogue and expressions conceal their relationship.

Similarly, the first version stresses the cigarette box that Ann inadvertently says she recognizes. The second version contents itself with a long shot, because now no one will question her about it.

     

The sense that the last path we see presumes the others has a more interesting side effect. Sometimes we get the sense that in the last go-round the character is mysteriously aware of the other paths she or he has taken. This happens in Sliding Doors and Run Lola Run, when each heroine seems to have learned from her experiences in the parallel worlds. And multiple-draft films like Groundhog Day and Edge of Tomorrow make this learning process essential to the action. The premise is illogical, but narratives often violate logic.

Something similar happens in the film (but not the play) of Dangerous Corner. Alone with Charles on the patio, Ann accepts his marriage proposal because the cigarette box she saw at Martin’s, now in Freda’s hands, made her realize she’s “been a fool.” We have seen the flashback in which she fought off Martin and accidentally killed him. The flashback isn’t in the play, and it’s especially interesting because presumably that scene really did take place, in both paths. That is, the evening orgy of confessions is one hypothetical alternative, but the manner of Martin’s death a year before, revealed in this path, is an actual event. It’s as if reliving Martin’s attack during the first path has made Ann appreciate Charles’s genuine love for her.

My seventh principle also bears on the last path we encounter. It’s a simple one: We take the final path as the correct one. Since endings are typically the place where all is revealed, we’re prepared to accept the last reiteration as what really happened. (This can be reinforced by a character’s learning curve, such as Ann’s.) Dangerous Corner is a good example.

I think we’re urged, by the second intertitle but also by the overall arrangement of the parts, to see that the quiet maintenance of civility predominated during that evening. The more sordid alternative, though given at much greater length, is what’s under the surface but what will never be acknowledged. The truth of the theft and Martin’s death, along with all the love affairs and animosities spilled out in the first version, will never be brought to light.

Other factors give the second version more weight. There are the shots I mentioned that stress the Ann-Charles relationship more, but for me the clinchers are more structural than stylistic. For one thing, both Ann and Charles play the role of raisonneur–the character(s) who explains the action and articulates central themes of the piece. In both the first path and the second, Ann echoes Priestley’s notion of our limited knowledge, saying that we prefer half-truths to the complete, factual account, the one that only God knows. And in both paths Charles warns against telling the truth, as it presents a “dangerous corner” that could lead to a smash-up. No other characters reflect so fully on the consequences of letting everything out.

Another structural factor involves the film’s beginning and the ending. The 1933 section starts with a scene in which Charles calls on Ann and once more asks him to marry her. Thanks to the primacy effect, this pair of characters becomes more salient in what follows than the married couples do. As a result, the epilogue, set on a terrace like the first one, harks back to the indubitably actual, pre-fork opening.

     

The fact that the movie ends with the creation of a couple, the uniting of the two characters who are the most self-aware and sympathetic throughout the film, reinforces the sense that this is the “real” outcome.

 

I hope you’ll read the whole essay. My purpose is to understand the dynamics of a small but increasingly common body of films. Forking-path films ask us to construct stories in unusual ways, but we quickly learn what guidelines to follow. Once the format exists, filmmakers take up the challenge of mastering it, stretching it, applying it to new material. (Edge of Tomorrow was sometimes considered Groundhog Day Goes to War against Aliens.) As filmmaking practice develops, we can track contemporary experiments and relate them to earlier efforts they are based on.

In addition, it’s worth knowing that a fairly sophisticated filmic treatment of the format appeared eighty years ago. In the essay, I find even earlier examples. And in a period when every movie seems at least 130 minutes long, it’s nice to encounter one that offers so much narrative complexity in about half that running time.

More generally, I think that probing this body of film shows the value of systematically studying narrative formats of any type. We’re used to talking about genre as a fluctuating body of conventions, but we should also study conventions that cross genres—conventions of story worlds, plot structure, and narration. These conventions can prod us to execute some unusual mental moves. Filmmakers are practical psychologists, and they’ve learned how to tease and tickle our minds. What-if movies are just one example of how norms and forms guide our understanding of story.


My first quotation from Priestley comes from Three Plays and a Preface (Harper and Brothers, 1935), ix. The second comes from Two Time Plays (Heinemann, 1937), ix.

In production, Dangerous Corner seems initially to have replaced Priestley’s what-if premise with more of a whodunit plot, while incorporating subjective sequences. According to a Variety review of an 83-minute cut before release, the explanations offered for Martin’s death by the various men are followed by something fairly unusual.

Surprise and twist, with increasing suspense, are accomplished through a shift from the factual elements to subjective processes on the part of the three women most closely related–as wives or sweethearts–with the suspected men. An innocent revolver shot precipitates the terrific speculations as each woman wonders if her man has killed himself in an adjoining room (Daily Variety, 13 September 1934, p. 3).

By the final release, however, the film had lost twenty minutes, and the result was the version we have. I haven’t seen versions of the screenplay, but I bet they’d be interesting. We’re left with other what-if questions, this time about the production of the movie itself.

Some of the basic concepts I employ here and in “Film Futures” are explained in the essay “Three Dimensions of Film Narrative,” online here. That essay is in turn applied to The Wolf of Wall Street in this blog entry.

 Dangerous Corner (1934).

Gone Grrrl

Gone Girl.

DB here:

How are contemporary movies, even this weekend’s releases, indebted to earlier traditions? Sometimes Kristin and I tackle this question. We’re not (I hope) trying to impress you as connoisseurs of esoteric knowledge. We’re definitely not trying to play down a movie’s originality by yawning and shrugging and murmuring “We’ve seen it all before.” Instead, as historians of film forms and styles, we’re interested in the ways that current movies reconfigure techniques that have been circulating across cinema history.

A lot of those techniques involve storytelling. So we’re obliged to study narrative conventions and innovations across the decades. And since cinema isn’t sealed off from other media, we’re curious about how films borrow narrative devices from other arts. The borrowing isn’t only one-way, of course; cinema has influenced storytelling in other media too.

Mystery and suspense stories have long attracted people who study narrative (“narratologists”) because such fictions depend almost completely on storytelling subterfuge. True, other genres can include false leads, or misleading ellipses, or questionable flashbacks, or strange point-of-view switches. But mystery-based plots require them in a way that science-fiction or romance plots don’t. In mystery stories, characters keep secrets from each other and the author must keep secrets from the audience as well.

Which brings us to Gone Girl.

Detective stories and thrillers are one-off demos of narrative trickery, so studying them can teach us something about how we understand stories. We’ve made a stab at showing this elsewhere on this site (see our codicil). But now comes a flagrant instance of narrative manipulation that has set people talking ever since Gillian Flynn’s novel was published.

That novel and the film she wrote and David Fincher directed throw into relief how popular narratives revise devices from earlier traditions. As narratologists in training, we’re keen as well to understand how the film creates its particular effects. I can’t answer all such questions here, but I offer some thoughts on the film’s storytelling strategies, with notes on how those adapt earlier ones.

Of course beyond this point there are spoilers for Gone Girl. I also heedlessly spoil Leave Her to Heaven, both book and film.

 

Two plots for the price of one

The domestic thriller usually involves a couple living together—a husband and wife, or in modern times a pair of lovers. The conflict might center on them or on a third threatening figure. A very common option focuses the plot on either a murderous husband or a murderous wife.

Gone Girl, both novel and film, starts with a classic murderous husband situation. Amy disappears. She has made her husband Nick discontented and angry, and he has started an affair with one of his students, Andie. Thanks to a crucial ellipsis in showing the first day, what Nick has been doing during Amy’s disappearance is skipped over. When Amy vanishes, apparently leaving a pool of blood behind, Nick is the prime suspect.

If your plot centers on a murderous husband, you have a choice. You can let the audience in on his plans, as in Rage in Heaven (1941), Conflict (1945), The Two Mrs. Carrolls (1947), and, much more recently, Safe Haven (2013). But Gone Girl doesn’t unequivocally show that Nick has killed Amy, or even plotted to kill her.

The second alternative for handling a murderous-husband situation is to keep it mysterious, chiefly by confining us to the wife’s range of knowledge. This ploy was made famous in Francis Iles’ novel Before the Fact (1932) and in Hitchcock’s adaptation Suspicion (1941). The premise is: I think my husband is trying to kill me. The 1940s crystallized this plot format in films like Secret Beyond the Door (1948) and Sleep, My Love (1948), and it survived for decades, in films from Sleeping with the Enemy (1991) to Side Effects (2013).

The first half of Gone Girl uses Amy’s diary entries to present the familiar arc of suspicion. The Dunnes’ marriage frays and she becomes increasingly frightened. Just before their fifth anniversary she buys a gun for self-defense. In her last entry she records her fear that he will kill her.

Nick looks pretty guilty at first, but with the whiff of doubt, another convention kicks in. That’s the one that film scholar Diane Waldman has called the “helper male.” If there’s another man nearby able to rescue the wife and play the role of a future romantic partner, then the husband is likely to be exposed as villainous. Examples are the Hollywood version of Gaslight (1944) and Sleep, My Love. If no helper is visible, then we’re likely to have a plot based on the wife’s misperception of the husband, as in Suspicion and Secret Beyond the Door. In Gone Girl’s first half, Amy seems to have no recourse to a helper male. This fact might dissolve some suspicion attached to Nick. Maybe, some viewers might ask, Amy was indeed abducted by a third party?

So much for the convention of the killer husband. A second type of domestic thriller, rarer than the first, centers on a homicidal woman. She shows up in the great Vera Caspary’s novel Bedelia (1945; made into a British feature in 1946) and in such films as Ivy (1947), A Woman’s Vengeance (1948), and Too Late for Tears (1949).

An especially shocking 1940s specimen of the killer wife is the cool, irresistible Ellen Berent in Leave Her to Heaven (novel 1944, film 1945). At first, Ellen wishes no harm to her husband Dick; she just wants to eliminate anybody with whom she’d have to share him. She lets his little brother drown, and, fearing that her unborn child will come between them, she flings herself downstairs and induces a miscarriage. Eventually, though, she turns her wrath on Dick. (Ellen’s most extreme tactic I’ll save for later, as it looks forward to Gone Girl.)

What is exceptionally clever about Gone Girl, again both novel and film, is that its second half replaces the murderous-husband schema with a revelation of Amy as a spider woman. Angry with Nick’s failure to sustain the role of the man she wants him to be, she has elaborately prepared an apparent murder that will lead police to suspect him. Here Flynn revives an old reliable of mystery plots, the faked death.

Amy has dovetailed three sets of clues for the police. There are clues she leaves in the treasure hunt, a little-girl game she obliges Nick to play every anniversary. This time, though, the hints point to uncomfortable aspects of their relationship. A second array of clues—imperfectly cleaned bloodstains, obviously faked signs of abduction, big credit-card purchases in his name—make Nick seem a lying killer. Then there is Amy’s faked diary, which she arranges to appear at just the moment that would harm him most. The diary entries initially coax us toward the Suspicion situation, seeming to provide a record of a wife’s growing apprehension of danger.

Knowing that a dead body will clinch the uxoricide case against Nick, Amy initially considers doing away with herself and letting the corpse be found in the river. Interestingly, this vindictive-suicide motif is the extreme tactic Ellen Berent pursues in Leave Her to Heaven. She poisons herself and sets up her sister Ruth, who’s in love with Dick, as her murderer. Like Amy, she has left a damning testament behind: a letter that will lead the police to arrest Ruth. Like Amy, Ellen has dropped a judicious trail of clues prepared well in advance.

So Flynn’s novel and screenplay shrewdly couple two thriller plot schemes, the murderous husband and the lethal wife. As an extra fillip, once Amy has been robbed by the desperate Jeff and Greta, she calls rich Desi Collings and convinces him of the threat Nick supposedly represents. In effect, Amy re-launches the lethal-husband scenario and recruits Desi as her helper male. Of course in most such plots, the helper male rescues the woman from peril. Here she is the peril.

 

Thank you, Mr. Griffith

So far I’ve talked mostly about what I called, in an earlier entry and in an online essay, the story world. But I couldn’t keep clear of two other dimensions of film narrative: plot structure and narration. I’ll talk about these more now.

I’ve indicated that Flynn’s novel breaks fairly neatly into two halves, splitting when Amy reveals that she hasn’t been kidnapped or killed. (“I’m so much happier now that I’m dead.”) The film, though, is a little less tidy.

As recidivist readers of this site know, Kristin has proposed that for decades Hollywood feature films have tended to break into several distinct parts that don’t fully correspond to the three acts of screenwriting manuals. The core structure for a normal feature involves four parts. Kristin labels these the Setup, the Complicating Action, the Development, and the Climax, with a brief epilogue tacked on. They’re determined by turning points that alter the goals that the characters pursue, and they tend to run twenty-five to thirty minutes or so.

Kristin argues in Storytelling in the New Hollywood that short films can delete a middle chunk, and long films can iterate one. For instance, she finds that Amadeus has two Development sections. Picking up on this, I proposed in The Way Hollywood Tells It that The Godfather (a very long movie) has not only two Developments but two Complicating Actions.

The film version of Gone Girl offers an interesting extension of the basic pattern. The film runs 144 minutes without credits. I divide up the first 126 minutes according to major turning points.

Setup. After the prologue close-up of Amy, she goes missing and Nick begins to conceal things from the police (roughly the first half hour).

Complicating action, in which the main character conceives a new goal. Now that public opinion casts Nick as the killer and Andie becomes another secret he must conceal, he must try to convince all he’s innocent. He fails. Boney summarizes the case against him at about 60 minutes in. Then he discovers the luxury goods stuffed into his sister Margo’s shed and he realizes that he’s been set up.

Development, in which backstory is provided, the protagonist confronts more problems, and many delays are set up. As Amy drives away from town and assumes a new identity, her Cool Girl monologue confirms for us that she has framed Nick. She hides in the motor court and strikes up an uneasy friendship with Greta and Jeff. While Nick engages Tanner Bolt as attorney and learns of Amy’s earlier framing of O’Hara, Amy calls Desi Collings for help. Nick has agreed to go on Sharon Schieber’s show.

Once we learn that Amy has faked her diary and loaded it with lies, we follow her stratagems after the first day. Gradually her life on the road syncs up with the progress of Nick’s situation, so that via crosscutting they eventually watch the TV coverage simultaneously.

Development sections tend to run a little long, and this one needs a chunk of backing-and-filling to explain Amy’s scheme. This part ends, I think, around the 104-minute mark, when Andie at a press conference confesses her affair with Nick while Amy accepts sanctuary at Desi’s lake house. Now Nick must take the initiative and fight back, while Amy must concoct a new plan for her new circumstances.

Climax: Here a plot culminates in success or failure, goals definitely achieved or not. Nick does well in his Sharon Schieber interview, but almost immediately the police discover the loot in Margo’s shed and confront him with the diary. He’s arrested. It’s his darkest moment so far. Meanwhile, Amy has become Desi’s prisoner. But Nick’s TV performance has convinced her that he’s ready to resume playing her ideal man. Accordingly, with typically surgical preparation, she kills Desi and returns home, announcing her escape from sex slavery. She comes back at about the 126-minute mark.

Were this a normal film, things would end here, with the couple restored; we’d need only a gloating tag showing humiliated police. Amy’s revelation of her scheme at 65:00 would then become a neat midpoint. But the film has almost twenty more minutes to run.

Is this section a protracted epilogue? Some viewers seem to take it as such, and to find it draggy, but it’s structurally necessary. I think it’s fruitful to see Gone Girl as having two climaxes.

 

Two climaxes for the price of one

The double climax in Gone Girl, I think, occurs because the film has a tandem structure from the start. The first hour alternates scenes involving Nick’s search for Amy with brief flashbacks triggered by shots of Amy writing in her diary. The diary supplies what we initially take to be exposition about their meeting, falling in love, getting married, losing their jobs, and moving to Missouri. These past scenes are sandwiched in between present-time scenes of the inquiry into Amy’s disappearance. The Griffith of Intolerance, not to mention the Christopher Nolan of The Prestige, would enjoy the extent to which book and film rely on large-scale crosscutting between protagonist and antagonist.

The duplex story lines lead to two turning points, one assigned to each major character. At 65 minutes, Nick’s discovery of the fancy purchases in Go’s shed changes his goal: He now must prove that Amy has set him up. At the same juncture, the revelation that Amy is on the road sets up her new goal of escape—at first, she thinks, to suicide but then to a life free of Nick, who seems safely en route to the death house.

Given the dual line of action, we have a first climax that puts Nick in jail and shows Amy killing Desi, which ends her flight. Her return provides a first phase of resolution for the overall action: Back home, she spins a new narrative for the media, one in which she “fought her way back” to her husband.

But we don’t have full resolution. Nick still has the goal of proving that Amy faked her disappearance, and now he must also show that she murdered Desi in cold blood. In addition, his rage against his wife’s frame-up threatens to make him the homicidal husband she painted in her diary. Will he be driven to kill her, as he sometimes indicates he’d like to? Meanwhile, Amy’s goal of reunion isn’t fully achieved. She still must evade punishment (Boney the cop is suspicious) and she must also persuade Nick to back up her rigged story of his abusive and spendthrift ways.

So a second climax shows Amy thwarting Nick’s goals. She blocks his efforts to reveal the truth and won’t allow a divorce. She retained Nick’s stored semen when they were trying to conceive a child, and she has impregnated herself. She will keep his son from him if he tries to make trouble. Nick accepts her frame-up and the couple become, as he says in their final TV interview, “partners in crime.” If you assume that Nick is the film’s protagonist and Amy the antagonist, we have that rare mainstream movie in which the antagonist wins.

Some critics have called the book and the film Hitchcockian, and film geeks will notice the midpoint giveaway as similar to that in Vertigo. More generally, the complicit couple exemplifies the transfer-of-guilt dynamic we find in Shadow of a Doubt, Strangers on a Train, and other of the Master’s films.

But this last quality isn’t unique to Hitchcock, or Flynn-Fincher. In Leave Her to Heaven, Dick becomes an accomplice to Ellen’s act of drowning his brother. He keeps quiet for the sake of the child he thinks she is carrying. As a result, the novel gives us a passage that could come straight from Gone Girl, on the page or on the screen.

And she said in serene and level tones: “But you lied to protect me, so—we share the guilt. That binds us together. We can never escape that now.” After a moment she added: “So we must go on together, wearing a mask for the world, being honest only with ourselves.”

 

Look who’s talking, or not

In both film and novel, Gone Girl’s large-scale plot patterns—the wedged-in diary entries, the ABAB attachment to characters, the cross-stitched timelines—are enhanced by choices about narration. I take narration to be not only voice-overs sound but the moment-by-moment flow of story information. That flow is regulated by cinematic techniques, orchestration of point of view, and kindred strategies.

Now we encounter some differences between film and literature as storytelling media. For instance, in the novel the parallel plotlines are rendered in first-person narration, alternating accounts from Nick and Amy. Amy’s fourteen diary entries are motivated as the sort of things one enters in a private journal, whereas Nick’s aren’t presented as him telling anyone his tale. When Amy’s diary is revealed as a hoax, she continues to recount events, and still in present tense, as if she couldn’t shake the habit. Nick, however, continues to tell us what happened in the past tense. This sort of difference is rather hard to achieve in film, unless you have two continual voice-over narrators. This option Flynn and Fincher decline, probably in the interests of clarity.

Both the alternation and the variation in tense have a long novelistic past. Dickens’ Bleak House (1853) switches between chapters in third-person narration in the present tense and chapters in first-person past. The device of alternating viewpoints was picked up in twentieth-century modernism (notably Faulkner’s books) and popular genres as well. A simple example is Philip MacDonald’s 1933 mystery, X v. Rex, which switches between first-person letters written by a serial killer to the police and third-person accounts of the efforts to track him/her down. Somewhat fancier is Anita Boutell’s Death Has a Past (1939), which takes a series of scenes transpiring across one week and sandwiches among them bits of a confession written by the killer afterward—“flashforwards,” in effect.

To go back to Leave Her to Heaven, Ben Ames Williams’ original novel alternates chapters filtered through the consciousness of husband, wife, brother, and sister, although all are treated in the third person. At about the same time, mystery writer Bill S. Ballinger gained notoriety for alternating chapters told from two characters’ viewpoints in two different time frames. Examples are Portrait in Smoke (1950) and The Tooth and the Nail (1955), the latter of which also toggles between first- and third-person discourse. More recently, in novels like Rebel Island (2007), Rick Riordan has intercut first- and third-person chapters.

Journals and assembled documents have been one standard way that classic novels have been organized, and shrewd writers have exploited many possibilities. I think of the moment in The Woman in White (1860) in which a long and engrossing account written by a character in peril is revealed, only at the end, as being read by her adversary. More specifically, a major Flynn trick—the discovery that the diary mixes reliable accounts of events with false ones—has one major precedent. It occurs in a famous 1938 mystery novel that, for once, I will not spoil by naming. Here the diary in the first section is intended to be found by investigators and to cover up the identity of a killer.

This isn’t to call Flynn unoriginal; she has come up with a new variant of these narrational conventions. The point is simply that once you work in the realm of the mystery thriller, you will probably be seeking out ways to mislead the reader, and some of those stratagems will have a kinship with your predecessors. More generally, all narrative traditions exfoliate. Storytellers are constantly experimenting with new ways to engage us, and it would be surprising if two authors, both bent on achieving similar effects, didn’t occasionally hit on similar devices. Just as Flynn welded together two domestic-murder plot premises, she recast traditions of shifting, unreliable narration.

In the novel, the two first-person accounts, Nick’s and Amy’s, restrict our knowledge to each one’s perspective. Which isn’t to say that each is transparent. Nick will sometimes report his dialogue with the speech tag, “I lied.” This teases us to wonder what he’s holding back from the cops. Interestingly, confessing lies makes him a more reliable narrator, because he’s confiding in us. This has the effect of making us trust his claims about wanting children, not beating up Amy, and so on. By contrast, Amy’s chipper narration seems completely open about her feelings, but many of those entries are revealed as part of her murder masquerade. Her unreliability, though, seems chiefly confined to the diary entries; once she’s on her own, she seems to be reporting her sociopathic reactions sincerely.

Because the novel restricts us to the two main characters, we can’t know what’s happening outside their ken. Something similar happens to the attached viewpoints in Leave Her to Heaven, in which Dick and Ruth only gradually learn how his wife Ellen plotted to make her death seem to be murder. In adapting the novel to film, screenwriter Jo Swerling respected the novel’s systematic attachment to characters to a surprising extent.. First we are mostly with Dick, but at crucial points we’re given blocks of scenes organized around Ellen. Unlike the book, the film version shows us her executing her plan to implicate Dick and Ruth in her death. Here, for instance, Ellen puts arsenic in the sugar bowl that Ruth will use to sweeten Ellen’s coffee.

The film Gone Girl doesn’t give Nick a pervasive narrating voice as the book does, and we aren’t wholly restricted to his range of knowledge. On several occasions we watch the police making discoveries that he’s not aware of. Crucially, we see them find the diary some days before they tell him about it. As so often happens, mainstream film introduces some unrestricted narration, which yields suspense rather than surprise. Amy narrates the eight diary entries which introduce flashbacks, some of which turn out to be false ones. But once she finishes her Cool Girl monologue after the big turning point, we don’t, I think, hear her voice-over again.

Nick’s voice-over enters only twice, with carefully symmetrical effect. The film’s first shot is a close-up of Amy turning toward us as a hand strokes her head. Nick’s voice talks of wanting to split open her skull, unspool her brain, and find the answer to married folks’ perennial questions. “What are you thinking? What are you feeling? What have we done to each other?”

Most broadly the shot has the effect of rendering this beautiful woman mysterious. It also suggests a violent impulse in an uncaring, obtuse male—the sort of scenario that would lead to the murderous-husband plot that hovers over the film’s first hour. The final shot of Gone Girl repeats the close-up, and Nick asks his questions again but adds, “What will we do?” Now we know how devious that brain is, and how much justified anger the man speaking may be feeling toward her. Perhaps, after the movie ends, he will be ready to kill her. The two shots, unanchored in story time, bracket the movie with the central duality that the plot and the narration will enact: a potentially murderous man and an innocent-appearing but lethally dangerous woman.

 

There’s much more to be said about the film and the novel. I haven’t touched upon Flynn’s subject matter—what we might call Yuppies 2.0, the brain-entitled Net-enabled cool kids—and her theme of marriage as a struggle to play the role your partner cast you in. Issues like these have ingeniously set readers talking about marriage’s putative dark side and how men can feel “picked apart” by dazzling, demanding women. For my money, the presence of this brilliant, beautiful Crazy Lady, another legacy of the 1940s, favors Nick’s side of things. He’s a weasel but not as dangerously nuts as his wife. Your mileage may vary, for reasons I attribute to Hollywood’s perennial urge to cover every bet on the board.

We’d also want to consider the novel’s style, which offers caffeinated versions of Product Placement Realism and Vivid Writing, in both male and female registers. The screen version loses this aspect of the novel, and we get instead Fincher’s calm, polished direction. Too many shots of vehicles pulling up to buildings, I suppose; but if everybody laid out scenes as slickly as Fincher does we wouldn’t complain so much about our movies. I admired in particular the virtuoso sequence accompanying Amy’s Cool Girl monologue. Her eloquent rant runs underneath a crisp replay of her scheme and then supplies a montage of her trip and her change of identity, with glimpses of female types illustrating her diatribe. This passage is a good example of the crackling pace of the movie, which, thanks to smooth hooks and concise exposition, rushes along at just the speed of our comprehension.

Consider as well the economy of a single moment, when Greta and Jeff steal Amy’s money. At one level, the couple serves purely a plot function, one typical of a Development section: new problems, often overcome, serve to delay the climax. (Another plot construction would have had Amy simply lose her money and call Desi right off, shortening the movie by quite a bit.) The confrontation in the motel also serves a thematic function, contrasting Amy’s sheltered life with another social level she both detests and fears, as well as giving us another couple to compare with Amy and Nick. (Interestingly, in the Greta-Jeff pairing, it’s the woman controlling the man.) The moment I have in mind, the instant when Greta slams Amy’s face against the wall and says, “I don’t think you’ve really been hit before,” fulfills even more purposes. The whack (a) shows that Amy’s new identity is easier to see through than she thinks; (b) reminds us that her diary reports of being beaten by Nick are false; and (c) engenders a certain sympathy for her, invoking the classic woman-in-peril situation. This tight packing of implication and emotion into an instant is characteristic of classical studio storytelling. It exemplifies the unfussy efficiency celebrated by Otis Ferguson and Monroe Stahr.

My purpose here has simply been to indicate that we can usefully understand any plot as a composite of possibilities that surface in other plots. In a way, I’m revisiting ideas floated in my Shklovsky/Sondheim entry months ago. Again, this isn’t to put down Gone Girl as formulaic. Instead, it’s to suggest that any narrative we encounter in the mass-market cinema (and probably in other forms of filmmaking) is part of an ecosystem, a realm offering niches for many varieties—including hybrids.


Thanks to Kristin, David Koepp, and Jeff Smith for conversations about Gone Girl.

For comprehensive accounts of mystery and detective fiction, visit Mike Grost’s encyclopedic website. Relevant to today’s entry is my web essay, “Murder Culture,” and my entry comparing Safe Haven and Side Effects. Film Art: An Introduction includes a section on various forms of the thriller in Chapter 9. And don’t get me started on the relation between Gone Girl and one of my favorite domestic thrillers, Double Jeopardy.

The distinctions among story world, plot structure, and narration that I use here are explained in two other items: the chapter, “Three Dimensions of Film Narrative” in Poetics of Cinema, available on this site; and the analysis of The Wolf of Wall Street illustrating that argument.

Diane Waldman outlines the role of the helper male in her article “‘At last I can tell it to someone!’ Female Point of View and Subjectivity in the Gothic Romance Film of the 1940s,” Cinema Journal 23, 2 (Winter 1984), 29-40.

For more on four-part plots, go to Kristin’s entry here and the essay “Anatomy of the Action Picture.” I’ve argued in another entry that popular novels can be built upon the four-part structure Kristin outlines, and it’s interesting to compare them with film versions that follow the same template. My examples were The Ghost Writer and The Girl with the Dragon Tattoo (Fincher version). I think that Flynn’s novel is built on a five-part armature roughly comparable to that of the film. As for double-climax plots, I argue in The Way Hollywood Tells It that In Cold Blood is another instance of this structure.

A more intricate use of diary narration in film is Nolan’s The Prestige, which derives from the even more intricate deployment of it in Christopher Priest’s original novel. We discuss the film in Chapter 7 of Film Art: An Introduction, and there’s a bit about the novel here.

Why so much emphasis on the cat in the film of Gone Girl? I suspect it’s at first a place-holder for the vanished Amy. Nick strokes the kitty as he strokes Amy’s head in the opening and closing shots. Seeing the cat tiptoe near broken glass (“You haven’t got a clue, have you?”) provokes Nick to solve the final clue in the treasure hunt. One shot, after Amy has returned, includes both wife and cat perkily welcoming Nick to breakfast. The cat is oddly matched eventually by the robot dog that Amy orders in Nick’s name. As a result, when Ellen Abbott gives Nick a robot cat, we have matching mechanical pets. Sort of like the human couple at the end.

Gone Girl.

Cinematic storytelling: A podcast on narrative

The Wolf of Wall Street (2013).

DB here:

Michael Neelsen is a filmmaker and consultant based here in Madison. In a discussion on ReelFanatics Michael and I consider some ideas about cinematic storytelling. My allergies gave me some Clintonesque hoarseness, and there are some things I’d rephrase better if I were writing them down, but maybe you’ll find something of interest there.

A couple of blog entries are relevant to our conversation: one on The Wolf of Wall Street and another on American Hustle. The first of these links to a general analysis of film narrative originally published in Poetics of Cinema and available elsewhere on the site. Our discussion of suspense and surprise harks back to other entries too, in particular those about Hitchcock and the bomb under the table (here and here). In the podcast I mention the Godard film Adieu au langage as well because I was then working on this blog entry.

You can also visit Michael’s company site StoryFirst.

Thanks to Michael for an enjoyable discussion, and for sharing it via podcast.

ADIEU AU LANGAGE: 2 + 2 x 3D

Adieu au langage (2014).

Godard’s Adieu au Langage is the best new film I’ve seen this year, and the best 3D film I’ve ever seen. As a Godardolater for fifty years, I’m biased, of course. And I might feel that I have to justify taking a train from Brussels to Paris to watch it (twice). But the film seems to me superb, and it gets better after several more (2D) viewings.

People complain that Godard’s movies are hard to understand. That’s true. I think they provide two different sorts of difficulty. He lards his dialogue and intertitles with so many abstract (some would say pretentious) thoughts, quotations, and puns that we’re tempted to ask what he is implying about us and our world. That is, he poses problems of interpretation—taking that to mean teasing out general meanings. What is he saying?

I think that this type of difficulty is well worth tackling, and critics haven’t been slow to do it. Scholars have diligently tracked the sources of this image or that barely-heard phrase. Adieu au langage provides another field day; there are movie clips, some quite obscure, and citations (maybe some made-up ones) to thinkers from Plato and Sartre to Luc Ferry and A. E. van Vogt. Ted Fendt has discovered a massive list of works cited in the film, and even his list, he acknowledges, is incomplete.

I confess myself less interested in interpretive difficulties. I don’t go so far as my friend who says, “Godard is a poet who thinks he’s a philosopher.” But I do think that he uses his citations opportunistically, scraping them against one another in collage fashion. In particular, I think that by having characters quote, quite improbably, deep thinkers, he’s trying for a certain dissonance between the abstract idea and the concrete situation.

What situation? That brings us to the second sort of difficulty. It’s often rather hard to say just what happens, at the level of plot, in a Godard film. From his “second first film,” Sauve qui peut (la vie) (1980), “late Godard” (which has lasted over thirty years, much longer than “early Godard”) has made the story action quite hard to grasp. Oddly enough, most reviewers pass over these difficulties, suggesting that story actions and situations that we scarcely see are fairly obvious. (Reviewers do have the advantage of presskits.)

The brute fact is that these movies are, moment by moment, awfully opaque. Not only do characters act mysteriously, implausibly, farcically, irrationally. It’s hard to assign them particular wants, needs, and personalities. They come into conflict, but we’re not always sure why. In addition, we aren’t often told, at least explicitly, how the characters connect with one another. The plots are highly elliptical, leaving out big chunks of action and merely suggesting them, often by a single close-up or an offscreen sound. Godard’s narratives pose not only problems of interpretation but problems of comprehension—building a coherent story world and the actions and agents in it.

We ought to find problems of comprehension fascinating. They remind us of storytelling conventions we take for granted, and they push toward other ways of spinning yarns, or unraveling them.

Case in point: Adieu au langage.

Since the film will be appearing in the US this fall, under the title Goodbye to Language, I want to encourage people to see this extraordinary work. But I’m also eager to talk about it in detail. So here’s my compromise, a four-layered entry.

I’ll start general, with some sketchy comments on some of Late Godard’s narrative strategies. In a second section I make some speculative comments on Godard’s use of 3D. No real spoilers here.

Then I’ll offer an account of the opening fifteen minutes. If you haven’t yet seen the film, this section might be good preparation. But part of experiencing the film is feeling a bit at sea from the start, so this section might make the film more linear than it would appear on unaided viewing. You decide how much of a preview you want.

The last section briefly surveys the overall structure of the film, and it is littered with spoilers. Best read it after viewing.

Spoilers notwithstanding, nothing stops you from eyeing the pictures.

 

Ecstasy of the image

Film Socialisme (2010).

Much in Adieu au langage is familiar from other Godard films. There are his nature images–wind in trees, trembling flowers, turbulent water, rainy nights seen through a windshield–and his urban shots of milling crowds. All of these may pop in at any point, often accompanied by fragments of classical or modern music. Again he returns to ideas about politics and history, particularly World War II and recent outbreaks of violence in developing countries. His standard techniques are here too. The film begins before, and during, the credits, which appear in brusque slates often too brief to read. Music rises, often just enough to cue an emotional response, before being snapped off by silence or an abrasive noise.

In his narrative films, as opposed to the collage essays like Histoire(s) du cinéma), we get scenes, but those are handled in unusual ways. He tends to avoid giving us an establishing shot, if we mean by that a shot which includes all the relevant dramatic elements. He often has recourse to constructive editing, which gives us pieces of the space that we are expected to assemble. Although Godard’s early films relied on this a fair amount, it became pronounced in his later work, where he tweaks constructive cutting in unusual ways. I discuss one example here.

Often we get an image of one character but hear the dialogue of an offscreen character. And the shot of the lone character may hang on quite a while, so that we wait to see who’s speaking. By delaying what most directors would show immediately, Godard creates, we might say, a stylistic suspense. I can’t prove it, but I suspect the influence of Bresson, who said to never use an image if a sound will suffice.

When Godard doesn’t give us unanchored close-ups or medium-shots, he may do something more drastic. A signature device of his later work is the shot which stages its action in ways that make the characters hard to identify. He may shoot in silhouette (Notre musique, 2003).

More outrageously, he may frame people from the neck or shoulders down (Bresson again?) and make us wait to discover who they are (Éloge de l’amour).

Such decapitated framings are disconcering, since orthodox cinema highlights faces above all other body areas. When we can’t access facial expressions, then the dialogue, gestures, postures, and clothes become very important. Godard can, of course, combine these strategies (below, Éloge de l’amour; also the Film Socialisme image above). In this shot, the man standing in the background is an important character but we never see him clearly.

Godard’s opaque “establishing” shots may be very condensed and laconic; he jams in a lot of information, partial though it is. In one shot of Adieu au langage, a dog approaches a couple on a rainy night and the woman urges her partner to take him in. All we see, however, is the man gassing up the car (and we don’t see him all that clearly).

We hear (dimly) the dog’s whimpering and the woman’s plea, but we see neither one.

Godard frets and frays his scenes in other ways. He creates ellipses, time gaps between shots that may leave us uncertain. What happened in the interval? How much time has passed? He also interrupts the scene through cutaways to black frames, objects in the scene, or landscapes; the scene’s dialogue may continue over these images, or something else may be heard.

At greater length, the scene can open up onto a digression, a collage of found footage, intertitles, or other material that seems triggered by something mentioned in the scene. In Film Art: An Introduction, we argued that one alternative to narrative form is associational form, a common resource of lyrical films or essay films. Godard embeds associational passages in his narratives, the way John Dos Passos embedded newspaper reports in the fictional story of his USA trilogy. Sometimes, though, the associations are textural or pictorial. At one point in Adieu au langage, Godard associates licked black brushstrokes on a painting with churned mud and the damp streaks on the coat of the dog Roxy.

          

By fragmenting his scenes, Godard gets a double benefit. We get just enough information to tie the action together somewhat, and our curiosity about what’s happening can carry our narrative interest. But the opaque compositions and the bits and pieces wedged in call attention to themselves in their own right. Blocking or troubling our story-making process serves to re-weight the individual image and sound. When we can’t easily tie what we see and hear to an ongoing plot, we’re coaxed to savor each moment as a micro-event in itself, like a word in a poem or a patch of color in a painting.

But those images and sounds can’t be just any image or sound; they hook together in larger patterns that sometimes float free of the plot, and sometimes work indirectly upon it. The best analogy might be to a poem that hints at a story, so that our engagement with the poetic form overlaps at moments with our interest in the half-hidden story.

Where, some will ask, is the emotion? We want to be moved by our movies. I suggest that with Late Godard, we are mostly not moved by the plot or the characters, though that can happen. What seizes me most forcefully is the virtuoso display of cinematic possibilities. The narrative is both a pretext and a source of words and sounds, forms and textures, like the landscape motifs that painters have used for centuries. From the simplest elements, even the clichés of sunsets and rainy reflections, the film’s composition, color, voices, and music wring out something ravishing.

We are moved, to put it plainly, by beauty–sometimes exhilarating, sometimes melancholy, often fragmentary and fleeting. Instead of feeling with the characters, we feel with the film. For all his exasperating perversities, Godard seeks cinematic rapture.

 

3D on a budget

The smallest set of electric trains a boy ever had to play with? Photo: Zoé Bruneau.

Most of the 3D films I’ve seen strike me as having two problems.

First, there is the “coulisse effect.” Our ordinary visual world has not only planes (foreground, background, middle ground) but volumes: things have solidity and heft. But in a 3D film, as in those View-Master toys, or the old stereoscopes, the planes we see look like  like cardboard cutouts or the fake sections of theatre sets we call flats or wings (coulisses). They lack volume and seem to be two-dimensional planes stacked up and overlapping. Here’s an example from a German stage setting of 1655, with the flats painted to resemble building facades.

In cinema, the thin-slicing of planes seems to me more apparent with digital images that are rather hard-edged to begin with. (3D film was more forgiving in this respect.) Sometimes the flat look can be quite nice, as in Drive Angry (2011). In this action sequence, the planes prettily drift away from one another, with no attempt to suggest realistic space.

Apart from the coulisse effect, there’s the problem that the 3D impression wanes as the film goes along. I’ve long thought it was just me, but other viewers report perceiving the depth quite strongly at the start of the movie and then sensing it less after a while, and maybe not even noticing it unless some very striking effect pops up. Part of this is probably due to habituation, one of the best-supported findings in psychology. Maybe, as we get accustomed to this fairly peculiar 2.5D moving image, it becomes less vivid.

More than our perceptual habituation might be at stake. Filmmakers may reduce depth during certain scenes to save money on postrproduction effects. Some gags in A Very Harold and Kumar 3D Christmas (2011) rely on old-school, smack-in-the-eye, paddle-ball depth, but much of the middle of the film doesn’t employ it. By tipping up the glasses and checking how much displacement is in the image, I’ve been surprised to find that remarkably long stretches of 3D films have little or no stereoscopy.

My impression is that Adieu au langage has overcome the problems I mentioned. Granted, many of the shots have sharply-etched images that emphasize the thinness of each plane. But other shots have unusual volume. Several factors may contribute to this. Unusual angles sometimes give foreground elements a greater roundness. This happens in the low-angle tracking shots created by the toy-train rig shown above.

In addition, the relatively low resolution of some of the images avoids creating hard contours.The wavering blown-out softness may enhance volume.

Perhaps as well the slight tremors of the  handheld camera mimic one of the factors that yield volume for our normal vision: the very slight movements of our head and body. Such shots shift the aspect enough to suggest the thickness of things.

Godard maintains the sense of depth in a tiny ways. For instance, he discovers that the crackling snow on a TV monitor can yield shimmering depth in the manner of Béla Julesz’s random-dot stereograms. Julesz sought to show that 3D vision wasn’t wedded to perspective cues or the identification of recognizable objects–a conclusion that ought to appeal to the painterly side of Godard.

Production stills indicate that Godard shot the film with parallel lenses. Instead of creating convergence by “toeing in” the lenses during filming, he and his crew played with the images in postproduction to control planes and convergence points. What they did exactly, I don’t know, but the results yield, for me at least, some strong volumes and a continual impression of depth that doesn’t wane.

I wish I could analyze the film’s 3D technique more exactly, but I don’t know enough about the craft of stereoscopic cinema or Godard’s creative process. What this film shows, however, is that 3D is a legitimate creative frontier. In the credits, as usual Godard brusquely lists his equipment, from the high-end Canon 5D Mark II (and Canon is proud to be associated with him) to small rigs like GoPro (in 3D) and Lumix. What is clear is that filming in 3D can be pictorially adventurous with cameras costing a few hundred dollars.

 

Nature, the ultimate metaphor

Now I’ll concentrate on the first few minutes, at the risk of potential spoilers.

The narrative in Adieu au langage is sketchy even by Godardian standards. Normally he gives us some characters in a defined situation (though it takes a while for us to grasp what that situation is), and a series of more or less developed dramatic scenes that advance a sort of plot. In Passion (1982) a movie director recreates famous paintings on film while a factory owner, his wife, and a worker get embroiled in his project. Detective (1985) carries us through a stay of several people at a luxury hotel. Je vous salue Marie (1985) gives us not one but two plots (Adam and Eve, Joseph and Mary). Éloge de l’amour follows a young writer in his exploration of art dealing and commercial filmmaking.

Adieu au langage doesn’t give us a plot even as skimpy as these. Instead, Godard builds his film out of a bold use of ellipsis and a strict patterning of story incidents. The ellipses are exceptionally cryptic. We must, for instance, eventually infer, on slight cues, that a couple has been together for at least four years, and that the man has stabbed the woman. We learn, with almost no emphasis, that both of the women have ties to Africa–hence the footage of street violence and the recurring question of how to understand that continent.

These very vague plot elements are arranged in a rigorous pattern. This patterning will seem very schematic in my retelling. But it’s not obvious when you see the film. Godard wraps his film’s grid in digressions, sumptuous imagery, and, of course, striking 3D effects.

To get a sense of both the firm architecture and the wayward surface, let’s look at the opening. The first fifteen minutes of Adieu au langage introduce in miniature what the rest of the film will be doing.

A montage of citations before the credits is followed by a fuzzy image of a neon sign. Now we get a sort of overture. Frantic video shots of a crowd under attack and running to a fire are followed by a clip from Only Angels Have Wings and a close-up of the dog identified in the credits as Roxy. That’s followed by a black frame dotted with points of white light. That image will become a little clearer later (stylistic suspense again). Then a title superimposes the numeral one in red with the word, “La Nature.”

What ensues, after a shot of a ferry approaching a pier, is a fairly disjunctive scene. A booksellers’ table stands across the street from the Usine a Gaz, a cultural center in Nyon, Switzerland. People casually gather there: a redheaded woman (Marie), a young man in a sweater who seems to be the bookseller, a woman on a bicycle (Isabelle), and the older man Davidson (later identified as a professor), here seen from the rear.

The cockeyed low-angle framing might make you think that this is Godard’s Mr. Arkadin, but it suggests footage from a camera or cellphone simply left tipped on some surface behind the table. In that respect it would make manifest the line in Éloge de l’amour: “The image, alone capable of denying nothingness, is also the gaze of nothingness upon us.”

Soon Davidson is sitting in the street commenting on Solzenitsyn’s Gulag Archipelago as a “literary investigation.”

A question he asks Isabelle behind him leads to a punning exchange about the thumb (pouce) that we use on our phones, which leads to a question about Tom Thumb (Poucette), a pun on “push” (pousser), and the suggestion that digital icons are like Tom’s trail of pebbles to the giant’s castle. The little skein of associations knots in a remarkable shot of two pairs of hands tickling their mobiles while another person’s hands examine books.

As the men swap phones, a car coasts through the shot in the background.

This scenic fragment, suppressing faces that would help us identify characters is characteristic of Godard’s approach in the whole film. He isolates gestures and surroundings, letting sound suggest the scenic action; and often the most important narrative action—here, the arrival of the car carrying a gunman—is a minor element in the frame.

So far, we’ve seen one of Godard’s strategies for hiding his story action: ellipsis. Time is skipped over (Davidson behind the table/ in a chair/ then perhaps behind the table), and bits of scenic action are omitted. There is also the opaque framing that impedes character recognition. What about digression? ? We’ve had one example in the Tom Thumb dialogue, but digression can be more overt. Godard can insert shots that have only a tangential narrative connection to the action.

The Godardian digression usually develops in a spreading web of associations that takes us on a detour. Here, one trigger seems to be the mention of Tom Thumb’s Ogre; another is the video display on the phones. These bits lead to a montage about Hitler, who, a woman’s voice reflects, left behind the belief that the state should handle everything. In a polyphony with the woman’s voice reflecting on Hitler, we get Davidson reflecting on how Jacques Ellul foresaw a good deal of the contemporary world. The associational links spread further, to images of the French revolution, crowds hailing Hitler, crowds at the Tour de France, and finally flowers and a voice reiterating a question at the scene’s start: How to produce a concept of Africa?

Now we’re back to the street, with the car pulling up. A chair that may have been Davidson’s is now empty. A man in a suit, the husband, emerges and lights a cigarette, looking off left. A woman, Josette, is in close-up—evidently the target of his look.

     

Since a black-and-white shot of Josette, head bent, was inserted in the Hitler montage, it’s possible that hers was the voice reciting the argument about the enduring trust in state authority. Perhaps she is reading? In any case, no sooner has a drama of sorts started than we get another digression. Marie reads aloud to us from a book held by the sweater boy. Again, the subject is state power and its inability to acknowledge its violence.

Domestic, not state-sponsored, violence is next on the agenda. A long shot shows the husband stalking up to Josette and berating her in German. The Usine sign is a big help in anchoring the action in the space we’ve seen, and Isabelle’s bike is visible on screen right.

Josette hangs stiffly on his arm, passively resisting and saying, “I don’t care.” He rushes out left. Gunshots are heard, and she jerks in spasmodic response. People rush through the frame. (We’ll never learn exactly what happened offscreen, though later there’s a hint that someone was shot.)

After the car has turned around and left in the way it came, Josette walks stiffly out of the frame. The man in the background who was startled by the husband’s abuse walks to the empty chair and pauses for a time to stare at it.

Cut to leaves floating on water, with hands washing and a man’s voice off saying: “I am at your command.”

So far, so Godardian. The narrative gist is that a woman has fled her husband, refused to return to him, and been approached by a different man who offers to join her. But the flow of images and sounds has made that gist very obscure, obliging us to absorb some fairly ravishing images and to listen to words, noises, and music as they form jagged, interruptive patterns.

And now something very unusual happens. Godard re-plays the events of “1-Nature” in a different location and time of year, using some new characters and some old ones.

A new section, “2” supered on “Metaphor,” appears. Its opening images run parallel to the overture that led up to “1.” After a shot of a swimmer (echoing the previous image of water), we get newsreel footage of combat and fire, and another film extract, this one from Les Enfants Terribles. A shot of Roxy along a river bank is followed by one of a hand opening and closing as a woman’s voice speaks of the “return” of language and a title repeats her insistence that she has made an image.

As at the start of “1,” the ferry comes toward the pier. And now we’re back with Davidson, now sitting along the edge of the water, again reading. His position and the tipped angle suggest a mirror-image of the earlier shot of him near the book table.

A link to the previous scene is provided when Marie and the sweater boy come to Davidson and say they’re going to America. The boy will study philosophy (obligatory quote from Being and Nothingness follows). Their conversation is interrupted by the arrival of the husband again, who shouts and fires his pistol. The shot announcing him is another skewed mirroring: his earlier entrance is inverted–again, as if a mobile phone’s camera had fallen.

The young couple flee and a woman, Ivitch, steps in to talk with Davidson. She shouts at the husband in German, “There is no why here!” ( a line that gets explained later in the film) and tells Davidson to ignore him. This moment offers a variant of the close shot of Josette when the husband had approached.

Ivitch asks Davidson, who evidently has been her professor during the previous term, questions about fighting unemployment by killing workers and about the difference between an idea and a metaphor.

In a new angle, Davidson meditates about images. As if to confirm the professor’s hunch that images murder the present, the husband lunges into the frame and yanks Ivitch out. We now get a shot in which the two cameras diverge: the left eye stays on Davidson, the right one pans over to Ivitch and the husband overlooking the lake. This offers a dense composition akin to that of the book-table shot, with figures piled on one another. The superimposition below is somewhat faithful to what we see, but it can’t convey your temptation to close one eye, then the other, in creating your own shot/reverse-shot editing.

The husband paces around Ivitch, points the pistol, and hollers in German that she’s a dirty whore. She replies as Josette had: “I don’t care.” She walks back to Davidson on the bench, and shortly the husband strides back to the car waiting in the background. Davidson returns to Ivitch’s question about metaphor and then points out two kids playing with dice. These exemplify “the metaphor of reality.” Cut to the kids rolling three dice.

The image echoes Godard’s segment of 3 x 3D, where he puns on “D” as dés, or dice. The kiddies’ shot literalizes the metaphor: trois dés, 3D.

Finally we see Ivitch behind a grille, looking up, then down as we hear the ferry’s horn off. A man’s hand comes in from the left, voice off: “I’m at your command.”

This action repeats the end of the “1” section, but differently: There we saw Gédéon when he inspected Josette’s chair, and heard him say the same words over the leafy water shot. Here both the words and the face of the speaker, Marcus, are offscreen.

Again a woman is threatened by her violent husband and a man emerges to replace him. Again that action is occulted by verbal digressions, dislocated framings, and major characters–here, Marcus–not introduced in a normal fashion. Once more the separate pieces of the scene, straining to cohere, are pulled apart just enough to register as individual instants of beauty, shock, puns, metaphors, or just peculiarity.

Godard’s prospectus for Adieu au langage indicated: “A second film begins. The same as the first.” This describes, laconically, what we’ve seen in the first fifteen minutes. That parallel structure is laid out again with astonishing, yet mostly hidden, rigor in the film as a whole.

 

Two plus two

Maximal spoilers here.

Over the last thirty years or so, we’ve had plenty of films that replay sections of their stories. Sometimes that dynamic is motivated as time travel, as in Source Code or Edge of Tomorrow—“multiple draft” narratives that let characters, as in Groundhog Day, revisit situations until they master them. Sometimes the repetition has been motivated through varying point of view, so that we see the same action again, but from a different character’s perspective. Examples would be Go, Lucas Delvaux’s Trilogy, and Ned Benson’s recent Disappearance of Eleanor Rigby. Once in a while we get films that present the events as repeated but significantly and mysteriously different. This is what happens in some Hong Sang-soo films, such as The Virgin Stripped Bare by Her Bachelors, as well as in Lee Kwangkuk’s Romance Joe.

In all, this is a minor but important convention of modern screenplays. The replay plot is common enough for screenwriting guru Linda Aronson to consider it separately in her book The 21st Century Screenplay. Trust Godard to take this emerging norm and fracture it.

The opening I’ve just considered invites us to see the film as split into two storylines. Godard has explored duplex construction before, in Éloge de l’amour (with the second part in color video) and Film Socialisme (with its third “movement” appended to two long sections). Yet Adieu au langage offers something different.

Here we have multiples of two: a prologue bookended by an epilogue, the two opening parts that are mirrors of each other, and then two long sections that are uncannily symmetrical. Those sections continue the stories sketched out in the opening section. Each plotline bears the same title as before, but now presented in different graphics (the number and the words are not superimposed, but presented in separate title cards). What’s remarkable is the precise parallels and echoes set up between the pair of tales.

The couples were cast with resemblances in mind, and this affinity is expanded through rather precise doubling. Nearly every scene in the plotline of Josette and Gédéon has its counterpart in the one featuring Ivitch and Marcus. Two nude scenes, two toilet scenes, two bloody-sink scenes, two mirror scenes, two movie-on-TV scenes. There are parallel sequences of driving in the rain, of a woman fleeing into a forest, of Roxy wandering in the woods, of helicopters crashing, of men dying in fountains. As we saw in the early 1/2 segments, the shots’ framings often echo one another.

     

     

Godard has laid bare the device in the second story, when Marcus and Ivitch and Marcus talk in front of a mirror.

Marcus: Look in the mirror, Ivitch. There are both of them.

Ivitch: You mean the four of them.

Rather than exact repetitions, we get repetition with variation. One couple takes Roxy in, the other (perhaps) does not. The first couple abandons Roxy on a pier in summer; in the second part, the pier in winter stands empty.

     

Most remarkably, the parallel scenes of the long section “2/Metaphor” proceed in almost exactly the same order as in “1/Nature”. Evidently Godard shot the bulk of the first story well before he shot the second. It’s as if the first film became the script for the second. In any event, the two long parts mirror one another with unusual precision. This geometrical structure recalls the “grid” organization of Vivre sa vie, but it’s not announced as boldly. Godard refuses to mark the parallel scenes in normal ways–with titles, or musical motifs. The labeling of the sections, 1 and 2 in the intro, 1 and 2 in the longer stretches, are sufficient for this laconic filmmaker.

Just as Godard blurs the shape of individual scenes through digression and opacity, so he hides the tabular structure of the film behind interruptions, landscape shots, and above all the charmed wanderings of Roxy, who more or less takes over the last portion of the second part. In addition, certain images from the second part echo or condense images we’ve seen before. The blood-filled fountain at the end of the second tale echoes both the bloody sink of the first one and the floating-leaf fountain in the prelude, while the clasping hands seem to consummate the gesture begun in the grille shot. These hybrid images can only make the strict double-column scene lineup more difficult to notice.

The fact that the exceptionally exact parallels and orderings of the two parts aren’t remarked upon by critics (I began to sense them a little during my second pass) is a measure of how successfully Godard has camouflaged the film’s anatomy. What shall we call this tactic? Distant counterpoint? Barely discernible rhymes?

Second film, or two films (short and long) times two: We’re free to see the characters as couples running uncannily in synchronization, or as the same couple in two guises, or as two stories in parallel universes. More likely, though, Godard is distressing and disheveling the emerging conventions of replay plotting.

And yet the ending of “the second film, same as the first” isn’t quite the whole story either. Godard has always enjoyed setting up rigid structures and then spoiling them–cutting off the arc of a melody or chopping a shot that could have been breathtaking. So he cracks his elegant 2 + 2 structure by giving us an epilogue and a third couple.

Images recur: crowds on the streets, Roxy snuggling on a sofa, a TV (but this time with two empty chairs). We glimpse a man reading, but mostly we see one hand painting with water colors while another is writing in a journal. Godard’s familiar dichotomy between image and word is here tied to the harmony of an unseen, but clearly heard, man and woman making art in tandem. The male voice seems to be Godard’s; I can’t say whether the female voice belongs to his partner Anne-Marie Miéville, but the woman seems to understand Roxy best. She can even access his thoughts. (“He’s dreaming of the Marquesa Islands.”) Yet this couple has another parallel, shown a little earlier: Percy and Mary Shelley, a poet and a novelist, the latter seen finishing Frankenstein in the forest. This is at least one farewell to language, but it also implies that creativity binds a couple together.

Roxy Miéville, as he’s called in the credits, haunts the film. He checks out streams, train platforms, and tree roots. He is never seen in the same shot with the main characters; his link to them is tenuous. His ramblings suggest freedom, sensory alertness, and a trust in immediate experience that perhaps the people can’t attain. The final images after the credits show Roxy wandering off in the distance and then bounding eagerly back to someone who stands, of course, offscreen.

 

Godard: The youngest filmmaker at work today.


Many thanks to Robert Sweeney and Richard Lorber of Kino Lorber, a bold company that still believes in art films. It will be releasing Goodbye to Language on 29 October (not September as I erroneously stated in an earlier version of the entry.) Later the film will appear on Blu-Ray 3D. Thanks also to Marc Silberman for help with German translation and to Ben Brewster for advice on stage wings.

For an interesting memoir of the filming of Adieu au langage, see Zoé Bruneau’s En Attendant Godard (Paris, 2014). The photo of the camera train is drawn from p. 93 of her book.

An excellent evocation of the fizz of word and image in Adieu au langage is offered by James Quandt in Artforum (also in the September print edition). Some other stimulating appreciations of the film are Scott Foundas in Variety, Daniel Kasman for MUBI, and Blake Williams in Cinema Scope. A useful description of the film is by Jean-Luc Lacuve on the site of the Ciné-club de Caen.

Too bad the GoPro Fetch, a harnessed camera for dogs, wasn’t available for Roxy to use.

To get a sense of how complex Late Godard is at the level of narrative comprehension, see Kristin’s essay “Godard’s Unknown Country: Sauve qui peut (la vie),” in Breaking the Glass Armor: Neoformalist Film Analysis. I analyze strategies of storytelling in Godard’s 1960s films in Chapter 13 of Narration in the Fiction Film. She wrote about Film Socialisme on the blog here. For a discussion of Godard’s very fussy compositions, try this entry. I consider multiple-draft narratives more generally in the essay
“Film Futures” in Poetics of Cinema.

P.S. 30 Sept: Since Adieu au langage screened at TIFF, VIFF, and elsewhere, a great many critical responses have accumulated. Thanks to the assiduous passion of David Hudson, you can track them all at Fandor. My initial posting should have mentioned two more enlightening discussions of the film: Kent Jones’s Cannes thoughts and the heroic display of Godardiana assembled by Craig Keller at Cinemasparagus.

P.P.S. 15 October: The beat goes on. Ted Fendt’s astonishing list of “Works Cited” in the film, which I added to the body of the above entry, deserves another link here. And the ever-expandig Mubi deserves our thanks for making it available.

P.P.P.S. 29 October: And more, of course. Background on the production process from Fabrice Aragno for Filmmaker; David Ehrlich’s sensitive discussion on The Dissolve; and a story on NPR, with interviews with Héloïse Godet, Vincent Maraval, and (gulp) me. Thanks to Pat Dowell for asking me to participate.

P.P.P.P.S. 2 November: If you haven’t had enough, I posted another entry on the film.

P.P.S. 13 November 2014: Geoffrey O’Brien’s enthusiastic appreciation of the film not only illuminates it but conveys the excitement of seeing it.

David Bordwell
top of page

comments about the state of this website go to Meg Hamel.