Digitality and the Persistence of Realism in Birdman

This article has been peer reviewed.

There is a scene in Alejandro González Iñárritu’s Birdman or (The Unexpected Virtue of Ignorance) (2014) in which two characters play a game of “Truth or Dare” while chatting on the rooftop of Manhattan’s St. James Theatre. “Truth or dare,” one character asks with an air of challenge. “Truth,” comes the other’s offscreen reply. “No,” the first retorts. “Truth,” the second responds. “No,” the first replies again. 

The following article contends that this moment in which “truth” is repeatedly asserted and negated encapsulates a key tension at the heart of the film. Specifically, I argue that this “truth”-“no” dialectic evokes Birdman’s central formal conceit: most of the film offers the impression of having been filmed in a single, continuous take. This extremely extended shot, in persisting without cuts, foregrounds the presence of the camera, the act of recording, and the idea of indexicality—the last of which refers to the sense of a material, analogical continuity between film image and pro-filmic reality, emerging from cinema’s historical basis in photography. Simultaneously, however, the film also introduces “impossible” effects within the technique that contradict the long take’s positing of “real” time and “real” space. Furthermore, these effects are connotatively digital: even if achieved using analog means, they reference concepts and capacities that have become associated with digital cinema. In other words, if Birdman’s long take seems to declare “truth,” articulating a historically salient conception of physical reality as expressed through the photographic film image, the spectacularly unreal moments comprise a contrapuntal “no,” unsettling the long take’s reality claim from within the image itself. Through this tension, Birdman thematizes and makes palpable the ways in which the digital transition—here loosely dated as beginning in the 1980s with the mainstream rise of computer-generated imagery (CGI) and extending into the present—subverts traditional film realism.

 In an article on Birdman’s long take and the digital media sphere, Celestino Deleyto and María Del Mar Azcona examine how the technique’s uniting previously incompatible spaces evokes the “lack of respect for ontological boundaries”¹ that characterizes the contemporary media experience, in which media multiplicity is experienced as “a single reality.”² My article also concerns how digitality disrupts prior conceptions of space and time, but I focus on digital cinema’s destabilization of traditional film realism. I begin by examining how Birdman’s long take references a film realism based in the idea of indexicality. Afterwards, I explore how the film’s introduction of spectacular digital effects within the long take creates a tension wherein the specific terms of digital cinema’s challenge to this realism are accentuated. In closing, I look at the role of narrative in mediating the viewer’s experience of filmic reality and, given this role, how Birdman’s narrative elements contribute to the film’s larger thematization of digital cinema’s “truth”–“no” interplay.

The Forceful Realism of the Long Take

With the exception of its prologue and epilogue, Birdman offers the impression that it was filmed using one unbroken shot. Even though the presence of masked cuts is deducible in moments of obscurantist staging—e.g., the camera conveniently ducking into a dark space, enacting a de facto fade-to-black—the prevailing sense is of a persistent continuity, a pointed and systematic refusal to cut. It is true that, tracking through the different floors, corridors, and dressing rooms of the St. James Theatre as preparations are made for a Broadway adaptation of Raymond Carver’s “What We Talk About When We Talk About Love,” Birdman’s long take reproduces the logic of découpage. As David Bordwell notes, the way the film stages its actions, such as using mirrors to generate a shot/reverse shot effect or deploying pans to “approximate a point-of-view shot,” emulates classical continuity editing, with the implication being that the long take is not as disruptive to the operations of narrative cinema as it may sound on paper.³ 

I concur with Bordwell’s assessment that Birdman’s long take, in remaining tethered to narratively motivated actions and events, facilitates viewer absorption. However, this particular long take is very long, seeming to run continuously for over 100 minutes. As such, even if the film’s narrativized staging is immersive on a moment-to-moment basis, the shot’s extreme length consistently guides attention to the presence of the camera and the act of filming.  Through underscoring these, the long take also highlights the idea of indexicality and the conception of reality it embodies. Originating from Charles Sanders Peirce’s theory of signs and resonating with Bazin’s realist film theory, indexicality names the idea that the film image is a material trace of the pro-filmic scene—a physical, mechanically generated imprint of actual, light-based inputs.⁴ Also operative within Bazin’s account, however, is Peirce’s category of the icon, which works based on a logic of visual resemblance: a shot of a tree is recognizable as such because the imaged tree resembles actual trees. Films do not need to be iconic to be indexical (a photographic image is still a material trace of pro-filmic reality even if it is visually indecipherable), nor do they need to be indexical to be iconic (hand-drawn animation regularly depicts familiar, real-world objects). That said, in Bazin’s account, the two are inextricably joined; his film realism hinges not merely on the fact of filmic capture, nor on visual resemblance, but on cases in which visual resemblance is experienced as being conditional upon filmic capture, and, hence, as being indexically and analogically linked to physical reality. A comparative reading of two of Bazin’s most famous essays, “The Ontology of the Photographic Image” and “The Evolution of the Language of Cinema,” reveals the theorist’s investment in both indexicality and iconicity—both “an image of the world [that] is formed automatically, without the creative intervention of man” (indexicality),⁵ and films that “bring the spectator into a relation with the image closer to that which he enjoys with reality” (iconicity).⁶ For Bazin, a shot of a tree both looks like a tree and is a material trace of the tree itself. Through iconicity, the idea of indexicality is expressed; conversely, the idea of indexicality continually frames the viewer’s experience of iconicity, positing the image as being “of” physical reality itself. Such is the case with Birdman’s long take, which, by dint of its sheer length, seems to insist upon this tight correspondence between iconicity and indexicality, and, by extension, between image and reality. 

This reality, which is indexed by the film image and iconically resembles the viewer’s own, is imbued with the indeterminacy, contingency, and excess of “unshaped nature itself.”⁷ By foregrounding the fact of the camera and the act of filming, Birdman‘s long take accentuates the physical reality that is captured and hence these qualities, which underpin the film’s impression of virtuosity, its technical impressiveness as a formal stunt.⁸ The longer the shot runs, the more the camera’s presence asserts itself, and the more the camera’s presence asserts itself, the more apparent becomes the risk of failure and accident that shadows every feat of coordination between pro-filmic bodies and the camera—the “real-life” choreographic planning and effort that was manifestly needed to move the camera in this way at this moment around these bodies arranged just like so. When, for example, the camera follows behind a character only to peel away and follow another when they enter the frame, the camera’s sustained recording attunes the viewer to the careful orchestration required for the second actor to enter at a specific point in the camera’s original trajectory, and for the camera to pivot smoothly from its original subject to the new one. Crucially, this sense of choreography emerges from the long take’s  spatiotemporal unity, the way the technique expresses the “real-time” duration during which any number of things could have gone wrong but did not. Successful execution feels all the more satisfying for having defied the odds; conversely, the “odds” themselves are foregrounded by successful execution, since the pleasurable experience of virtuosity is also an experience of contingency, of mishaps averted. In other words, the impression of formal and choreographic virtuosity in Birdman not only draws on but further highlights the contingent physical reality indexed by the long take.⁹

Given the long take’s insistence upon the idea of indexicality and, by extension, the contingency and indeterminacy of physical reality, Birdman’s introduction of “impossible” effects has not only stylistic but metaphysical implications, underscoring the way the emergence of CGI unsettles prior conceptions of the film image and the reality to which it points. 

Digital Ruptures

Introducing fantastical phenomena within an aesthetically photographic image is, in itself, nothing new. Since cinema’s earliest days, analog special effects such as multiple exposure and optical printing have enabled the illusion of flesh-and-blood actors interacting with otherworldly creatures and occurrences, something CGI has taken to new heights of visual seamlessness. Julie Turnock, writing on the development of American mainstream special effects, has identified an overwhelming tendency toward photorealism, which she defines as a seamless compositing of flesh-and-blood actors and digitally constructed elements such that all appear to be occupying the same, photographed physical space.¹⁰ This stylistic approach extends into films that depict fantastical phenomena. As Stephen Prince notes, the special effects landmark Jurassic Park (1993), despite its far-fetched subject matter, deploys a “correspondence-based model of cinematic representation” in which viewers, drawing on their own “audiovisual experience of three-dimensional space,” are invited to experience human actors and computer-generated dinosaurs as existing within the same physical and spatial field—a case of “referentially fictional” images that are “perceptually realistic.”¹¹ Especially given the present proliferation of CGI-heavy blockbusters, the coexistence of the mundane and the outlandish within the same photorealistic frame has become commonplace in contemporary cinema.

Birdman

Birdman distinguishes itself from other CGI-laden films by introducing such spectacular effects within the context of an exceedingly long take. Despite the technique continuously highlighting the indexical connection between the film image and the captured reality, Birdman interjects phenomena into the image that are referentially impossible. These range from time jumps—hours and even days of narrative time pass within a few real-time seconds of the long take—to displays of telekinetic power, to a moment in which a massive, winged metallic creature lands on a building in broad daylight. 

Crucially, these effects are not just spectacular, but spectacularly digital—that is, they exhibit a digital logic that contradicts the “real” time and “real” space asserted by the long take. By “digital logic,” I am referring to effects that manifestly could have been achieved only through CGI—e.g., the spectacle of the winged monster, the sight of protagonist Riggan levitating high above the streets of Manhattan, and a moment when the camera seems to pass through an impossibly small space. That said, I am also referencing those effects that, even if literally achieved or achievable by practical/analog means, are connotatively digital in our contemporary historical moment, given the way they evoke specifically digital concepts and capacities. The film’s displays of telekinesis are an example of this connotative digitality. These effects are attainable using practical methods, such as a hidden string to show an object moving seemingly without any physical cause, or, in the film’s most low-tech example, simply clicking a television remote offscreen to simulate Riggan shutting off a monitor with his mind. That said, I would also argue that telekinesis emblematizes digitality’s creatio ex nihilo logic. Based in manipulable ones and zeroes, the digital image lacks the causal, material continuity with physical reality that characterizes photographic indexicality. Rather, in requiring even physical inputs to be first converted into “machine-readable notation”¹² before being assembled into photorealistic images, the digital image replaces a logic of analogical causation with one of arbitrary creation. Given the mediating step of digital conversion, digital cinema is no longer materially tethered to physical reality. Rather, it more closely approximates the blank-canvas logic of a medium like painting, in which the realm of the mental supersedes that of the physical, since, in theory, whatever can be imagined can be painted.¹³ As D.N. Rodowick observes, whereas photographic cinema 

grounds its sense in a special kind of indexical logic in which the judgments of correspondence [between image and referent] are anchored in physical reality and reference to the past […] the key point of reference [in digital cinema] now will be to mental events—not physical reality molded to the imaginary, but the free reign of the imaginary in the creation of images ex nihilo that can simulate effects of the physical world (gravity, friction, causation) while also overcoming them.¹⁴

Telekinesis was depicted in films long before the age of CGI, but it acquires a specifically digital ethos today given its parallels with the logics and affordances of the new technology (along with the fact that onscreen depictions of superpowers like telekinesis tend to be realized, and are popularly understood as being realized, through the heavy mediation of CGI in procedures such as digital wire removal). Within the digital regime, images are generated seemingly ex nihilo, since the microscopic bits and speed-of-light microprocesses that constitute the digital image are, in Shane Denson’s words, radically “discorrelated” from human phenomenological experience—i.e., falling below the threshold of human perception and alien to the ways in which we phenomenologically make “sense” of the world.¹⁵ As such, when computer code causes something perceivable to “happen” at the level of the image, it has a magical quality to it, seeming to appear out of thin air; this conjuring-from-seemingly-nothing evokes Wendy Chun’s concept of “sourcery,” or the feeling of omnipotence experienced by the software programmer who can, on command, bring things into being as if by wizardry, since the source code itself remains black-boxed away from human perception.¹⁶

Birdman

Riggan’s telekinetic displays demonstrate a digital logic by referencing both the shift to “mental events” and digitality’s “magical” quality. Telekinesis begins with the mind, and yet it generates visible consequences sans any discernible, physical medium linking cause and effect, suggesting the presence of non-phenomenal forces that manifest at the human scale as if by sorcery. The iconicity-indexicality of traditional film realism hinges on the notion that the visible image is materially continuous with physical reality and, conversely, that physical reality is visible and capturable on film—an idea that Birdman’s long take accentuates. Telekinesis, however, embodies a contrary logic in which a conception of reality as continuous, physically visible, and visibly physical is sidelined in favor of the connotatively digital categories of the mental and the magical. Thus, when telekinetic effects occur within the long take, there emerges a tension in which both the long take’s reality claim and digitality’s challenge to it are thrown into bold relief.

A similar argument can be made about the film’s temporal manipulations, which, from a technical standpoint, are achievable through analog means (time-lapse photography for the couple moments that accelerate through night into the next day) and often do not require anything digital at all (many of the time jumps are coded into the image at the level of script and staging, such as a moment in which the camera tracks from a character remarking about a Broadway star’s forthcoming arrival to the star standing just a few meters away, suggesting that hours have passed). That said, in the way they violate the presumed spatiotemporal unity of both the film image and the physical reality indexed by it, these time jumps thematize the fact that filmmaker manipulation can, in the digital age, occur within the space of a single image, rather than simply between images. Since the analog film image was understood to be a direct, mechanical registration of physical reality, it was tacitly assumed that reality’s spatiotemporal unity was reproduced in the image; once a shot was taken, its contents became fixed and ontologically sacrosanct. Different shots could be freely edited together in sequence, but the space within the frame was relatively off-limits. As Berys Gaut notes, however, with the rise and increasing public awareness of digital visual effects’ capacity for image-manipulation, “the traditional dichotomy noted by film theory between the relative difficulty of manipulating an image and the ease of editing breaks down: both image- and sequence-manipulation are easy in digital cinema.”¹⁷ It is the capacity for intra-frame editing that inspired Manovich, in a different context, to attribute to digital cinema what he calls “spatial montage.”¹⁸

If Birdman’s telekinetic displays along with the even more obvious, spectacularly “digital” moments like the scene with the winged monster, which enact a digital logic by clearly actually having required sub-phenomenal computer code to be “magically” realized onscreen—disrupt traditional film realism’s iconicity-indexicality conflation (i.e., the notion that reality “is” what is visible in the image, and vice versa), the time jumps transgress the film image’s presumed spatiotemporal unity, shifting ontological weight from the “real-time” recording of footage during production to the post-hoc image adjustments that occur in post-production. Once again, all these violations of cinematic indexicality register more forcefully because they occur within the context of the long take; they are the “no” to the long take’s “truth,” generating a tension that attunes viewers to the specific ways in which the digital turn unsettles the regime of photographic cinema.

Narrative Worlds

Earlier, I mentioned how it has become common for fantastical phenomena to coexist with more mundane elements within the same, photorealistic image. This naturalization is facilitated by not just seamless compositing but also narrative continuity. Since well before the digital turn, viewer absorption into fantastical worlds has hinged on not only the convincing representation of a coherent onscreen space, but also the narration of this world through continuity editing techniques, actors’ performances, and script-level elements that help construct the impression of a unified diegesis. Photographic realism plays a key role in generating the illusion of spatial continuity between dinosaurs and humans in Jurassic Park, but so does the convincing interaction of elements—e.g., a character believably reacting in fear as a Tyrannosaurus Rex appears in frame—and the a priori narrative establishment of this world as one in which dinosaurs and humans could coexist. Such narrative framing is the reason that films whose special effects do not appear seamless to contemporary eyes, such as the original King Kong (1933), can continue to be experienced as diegetically immersive. Along these lines, even if a long take were populated with fantastical occurrences, the convergence would unlikely register as especially disruptive if these phenomena are presented as existing within the diegesis. 

The reason Birdman’s effects feel disruptive is that the film narratively frames most of them as being unreal (or at least of uncertain reality), thereby challenging the long take’s photographic realism on two fronts: not just the referential impossibility of the effects vis-à-vis the viewer’s ordinary, phenomenological experience of physical reality, but the presentation of such effects as fantastical even within the diegesis. Throughout the film, Riggan’s telekinetic displays are implied to be figments of his worsening psychosis, brought on by personal insecurities and the pressures of mounting his Broadway debut; these hallucinations reach a climax in the scene involving the winged creature—precisely when Riggan is at his emotionally lowest and when his stress has been compounded by a night of heavy drinking. Other characters seem unable to see what he sees: they do not react proportionately to the would-be stupefying spectacle of what Riggan witnesses, as the Jurassic Park characters do to the dinosaurs. Repeatedly, the film shows a spectacular effect, then questions its diegetic reality moments later. During an emotional breakdown in which Riggan is seen telekinetically hurling objects around in his dressing room, the camera pans right to show another character peeking in, after which it pans back left to reveal Riggan anticlimactically tearing apart a wad of paper with his hands, suggesting the earlier telekinesis may have been a hallucination. 

In narratively coding its spectacular displays as being phantom psychological phenomena while nonetheless placing them within the long take alongside other more “real” occurrences as if all were inhabiting the same physical space, Birdman underscores the tension between photographic and digital image regimes. The long take, in highlighting the idea of cinematic indexicality, seems to confer an equal level of physical reality to all captured bodies, objects, and spaces. By narratively coding some of these as unreal—hallucinatory figments lacking basis in the physical reality of the diegesis—but maintaining the continuity of the long take across all, Birdman thematizes digitality’s challenge to photographic film realism, emphasizing that, today, what is “real” (materially continuous with photographed physical reality) can seamlessly appear within the same, photorealistic shot as what is “unreal” (conjured “after the fact,” in digital post-production). As mentioned, this juxtaposition of “real” and “unreal” has a long history in cinematic special effects. In Birdman, however, the narrative framing of this juxtaposition as a matter of reality (along with the digital logic of the effects themselves, and the fact that the plot point of hallucination resonates with digitality’s shift to “mental events” à la my earlier argument about telekinesis) is connotatively linked to our contemporary historical moment, in which digital media’s capacity to imperceptibly manipulate photorealistic images has been the source of much anxiety and debate—as seen, for example, in the public discussion surrounding deepfakes and cosmetic applications like Facetune.¹⁹ 

Given the film’s narrative as well as effects-level thematization of digitality, the spatiotemporal continuity of the long take, though on the one hand a forceful expression of filmic indexicality, also ends up having the opposite effect. In insisting on the photorealistic continuity between more quotidian, narratively “real” visual elements (flesh-and-blood actors, everyday objects, etc.) and more spectacular, narratively “unreal” ones (especially those that not only evince a digital logic but manifestly could have been achieved only through CGI), the long take causes the latter to “contaminate” the former, rendering the ostensible indexicality of the more mundane elements suspect. The long take, in continually foregrounding the spatiotemporal unity associated with the photographic film image, ultimately “unites” all elements as being equally “unreal,” equally reducible to sub-phenomenal bitmaps. Once one part of a shot manifests clear signs of digital mediation, the digital basis of the rest of the shot becomes accentuated, since, in order to be seamlessly composited with the more spectacularly digital elements, they themselves had to first be converted into binary code. 

Had diegetic unity been maintained through narrative framing that posits these spectacular elements as empirically possible within the world of the film, this ontological contamination would have been downplayed. Because Birdman narratively codes these fantastical phenomena as unreal, however, the tension between “real” and “unreal,” and the porous border between them, remain foregrounded. As a result of this contamination, the more mundane image elements acquire the quality of what Bernard Stiegler terms the “analogico-digital image”: an image that looks photographic but is known or assumed to be digital, hence inviting the viewer to approach it with a skeptical and searching eye that attempts to discern seams in the photorealistic illusion.²⁰ Resonating with broader, cultural anxieties about the veracity of digital images, this posture of doubt is the most powerful and unsettling consequence of Birdman’s collision of photographic realism with the logic of digitality, unmooring the spectator from any definite relation to a reality beyond the image.

And yet, the long take, as the film’s structuring formal conceit, remains at the forefront of the viewer’s experience, continually insisting upon the reality of the physical world even as increasingly apparent digital mediation rejects this claim again and again. Although it cannot be said to unequivocally denote a spatiotemporally unified physical reality and, in fact facilitates the digital contamination of all image elements, the long take nonetheless projects a desire for this reality, a persistent attempt to reach it.²¹ “Truth” is answered by “no,” but “no” is also answered by “truth.” Ultimately, Birdman is not about one or the other but the dialogue itself, expressing the cultural tensions surrounding questions of reality in the age of digital media. In closing, it is worth noting that, though the “truth”-“no” dialectic has been the subject of this essay, there is also a third term in the game: “dare.” Situated opposite “truth” but also not a complete negation in the manner of “no,” “dare” is an invitation to test boundaries. It names not a destination, but an intentional movement past prior limits. Although the “truth”-“no” interplay expresses an ongoing tension between traditional film realism and the digital turn, Birdman’s enigmatic ending and the additional term of “dare” also encourage reflection and curiosity about where this tension might lead us, and what might exist beyond it altogether.

Endnotes