Textualism, extratextualism, and the fiction/nonfiction distinction in documentary studies

ABSTRACT

This article critiques existing textualist and extratextualist (intentionalist and reception-driven) approaches to capturing the ordinary understanding of the fiction/nonfiction distinction in philosophical and film scholarship on documentary and offers an alternative extratextualist approach dubbed institutionalism. I argue that textualist attempts fail because no textual element (presentational strategy, misrepresentation, staging, or indexicality) is necessarily either fictive or nonfictive. Intentionalism falls short because films can change their non/fictional status over time (e.g. phantom rides). Finally, reception-driven approaches confuse personal categorizations for public ones. The proposed institutionalism, by contrast, combines the strengths of moderate textualism and reception-driven theories (allowing for the changing status of documentary and nonfiction) with those of intentionalism (denying that some textual elements are necessarily fictive and others nonfictive) to capture the ordinary understanding of the fiction/nonfiction distinction.

KEYWORDS:

• Textual features
• intention
• reception
• institutionalism
• make-believe

Introduction

Documentary film scholars generally agree that the distinction between nonfiction and fiction is a matter of degree rather than some firm boundary. When focusing specifically on documentaries, understood as a subclass of nonfictions in the first instance, it is also common to say that documentaries regularly include fictive elements and exist ‘along a fact – fictional continuum, each text constructing relationships with both factual and fictional discourses’ (Roscoe and Hight 2001, 7).¹ I propose to call these approaches textualist, insofar it is the presentational strategies and/or the content that constitute fictive elements. Looking at some key accounts from this tradition, I argue that standard examples of these elements – discursive forms, misrepresentation, staging and re-enactment – are misplaced and that the boundary between documentary and fiction is much less porous than assumed. To outline this border more precisely, I build on alternative extratextualist (intentionalist and reception-driven) approaches and offer an extratextualist variant – an institutionalist approach which combines lessons from documentary studies and analytic philosophy.

When it comes to text, I understand presentational strategies to include mise-en-scène, editing, camera, sound and syuzhet. Content I construe as what is represented and made through these presentational strategies. Most importantly for us, this includes the image and audio track, profilmic, fabula and assertions. Moreover, although I espouse the neoformalist view that fabula is not in the text, but a spectatorial mental construct, in this framework fabula is still a textual feature because it cannot exist without a text. Similarly, reference, meaning and truth value as parts of making assertions are all mental phenomena as well because they demand an agent interpreting signs which bear content of these assertions. But they again remain textual features under this framework because the text is necessary to mentally construct the relevant reference, meaning and truth value.

Before proceeding, I also wish to emphasize that I am not interested in semantic policing but in the ordinary (narrow) understanding of fiction/nonfiction distinction. Lies, falsehoods and deceptions are all legitimately called fictions in ordinary language. It is perfectly proper to label Brexiteers’ claims that the weekly £350 m sent to Brussels will be spent on NHS instead as fictions. Yet this broad sense of fiction is easily distinguishable from the narrow (no less ordinary) one where, for example, superhero movies are fictions and family home videos are not. Narrowly construed, then, Brexiteers’ claims are lies and not fictions. It is in this narrow sense that the fiction/nonfiction distinction is understood here. Crucially, I argue that scholars interested in documentary should focus on this narrow understanding of the distinction rather than the broad one.

Textualism

Perhaps the most radical example of the textualist approach in documentary theory is Michael Renov who has argued that ‘all discursive forms – documentary included – are, if not fictional, at least fictive, this by virtue of their tropic character (their recourse to tropes or rhetorical figures)’ (1993, 7). They include: ‘character “construction”; poetic language; emotionalizing narration or musical accompaniment; “embedded” narratives; dramatic arcs; the exaggeration of camera angles, camera distance, or editing rhythms’ (1993, 198). Renov’s definition of fictive elements in terms of discursive (presentational) strategies is far removed from both broad and narrow ordinary understanding of fiction. Rephrasing a factual statement ‘Unfortunately, I missed my train’ into (poor) poetic language ‘I missed the train/what a shame!’ does not introduce any fictive elements in the ordinary sense. A presidential debate shot with cameras positioned at eye-line level does not become fictive if low angles are used instead either. This does not deny that, as Renov himself points out, ‘documentary has availed itself of nearly every constructive device known to fiction (of course, the reverse is equally true)’ (1993, 7).² Rather, it means that presentational strategies are neither fictive nor nonfictive on their own.

According to a more moderate account by another key documentary theorist, it is the specific textual features such as the degree of fabrication in content that constitute fiction rather than discursivity in general: ‘The division of documentary from fiction, like the division of historiography from fiction, rests on the degree to which the story fundamentally corresponds to actual situations, events and people versus the degree to which it is primarily a product of the filmmaker’s invention’ (Nichols 2017, 8–9). For instance, whether we treat Nanook of the North (Robert Flaherty, 1922) as documentary or fiction, Nichols claims, rests on whether it is ‘a plausible representation of Inuit life [or] Flaherty’s distinct vision of it’ (Nichols 2017, 9). But Flaherty taking liberties with the facts does not mean he introduces any fictive elements. These liberties only point to a (deliberate) misrepresentation of Inuit life. To flesh this out, consider Bowling for Columbine (Michael Moore, 2002) where Moore uses editing to misrepresent how little time it takes to buy a gun. The misrepresentation is not a fictive element. Rather, this misrepresentation (and others like it) reveals that the documentary which presented itself as a plausible representation fails to be truthful under closer scrutiny. In other words, Flaherty’s distinct version of Innuit life is not something apart from its plausible representation but is precisely what is presented as the plausible representation (although this plausible representation is in fact a misrepresentation). In another variant of Nichols’ vocabulary, Flaherty is not presenting us with a world, but still with the world – he is just misrepresenting it. Much like with the Brexiteer slogan, misrepresentation (deliberate or otherwise) does not constitute fiction, nor does it introduce fictive elements. In short, textual features alone (presentational strategies and content) cannot determine whether something is fiction or not.

Another specific textual element often deemed to be fictive in documentary context is staging. Take, for instance, Springer and Rhodes’s introduction to a volume on docufictions:

For the authors in this collection, these intersections begin literally at the beginning of the cinema, blurring the lines between fact and fiction. Blacksmith Scene (1893), Edison and Dickson’s first publicly exhibited film, purports to show exactly that: three blacksmiths working and enjoying a beer. But of course the men captured on film were not professional blacksmiths but Edison employees, and the film was staged and shot at the Edison laboratory. (2005, 6)

Here staging is meant both as manipulation of mise-en-scène and deception. Undeniably, the Blacksmith Scene was not recorded at the blacksmith’s but in Edison’s Black Maria studio with a specific set design, props, lighting, character placement, work choreography, etc. in place. However, since the manipulation of the mise-en-scène is a discursive trope in Renov’s sense, it does not necessarily entail any fictive elements. Much like a presidential debate does not turn fictional simply because it includes a complex mise-en-scène of lighting, candidate placement, speaking rules, set design, etc. in front of a camera so is a recorded wielding and hitting of hammers and passing the bottle not fictional simply because it was staged in this sense. Matters do not change even when it turns out that the protagonists using these tools and enjoying a drink are not blacksmiths but actors. Precisely like with the Nanook example, then we are merely dealing with a (deliberate) misrepresentation. In other words, Blacksmith Scene remains nonfictional even in its deceptive aspects.

Staging can also be understood as re-enactment. Because of the extensive use of re-enactments, The Act of Killing (Joshua Oppenheimer, 2012) – a documentary in which perpetrators of Indonesian Killings of 1955–1956 re-enact their crimes – is regularly cited as a mixture of nonfiction and fiction. As Nichols puts it:

Usually, documentaries embed reenactments as acknowledged reconstructions (fictional representations) of historical but originally unfilmed events within a larger context of nonfiction representation. But this need not be the case, as The Act of Killing amply demonstrates in a befuddling, disturbing, and illuminating manner. […] Befuddlement arises when a clear distinction between fictional and documentary representation fails to materialize. (Nichols 2013, 25)

Here Nichols treats re-enactments as fictional representations. But this cannot be the case for a couple of reasons. The first is that whether something is a re-enactment is a textual feature and as such is neither necessarily fictive nor necessarily nonfictive. Furthermore, Nichols understands re-enactment as a recorded staged representation of a previous event. Yet recorded staged representation can easily be nonfictional. For example, I want to relate to my friends the latest claims made by Donald Trump during a presidential debate. In the first version, in an email I send to my friends I write down what Trump said. To the best of my recollection this is a reasonable reconstruction of the event and as such a nonfictional propositional representation of the same. In the second version, I re-enact Trump’s speech i.e. stage its performance through mise-en-scène. I grab a make-shift podium, put on a wig, stand in front of my friends, and say Trump’s lines. This time it is a staged representation, but nonfictional, nevertheless because again it is presented as a plausible reconstruction. The representation remains nonfictional in the final version as well where I record the re-enactment and send it to my friends via Facebook. In fact, as the discussion of Nanook of the North and Blacksmith Scene reveals, I could also be completely misrepresenting what Trump said but this would only make the re-enactment a deliberate misrepresentation and not fiction.

Furthermore, most of the re-enactments in The Act of Killing cannot be fictions either precisely because they are presented as veridical accounts of the killings i.e. plausible representations in Nichols’ vocabulary. For instance, Anwar, one of the main perpetrators, restages in detail how he used metal wire to ‘optimize’ the executions. The audiences are supposed to either confirm the plausibility of these representations by believing in them (in line with official Indonesian history) or deny them (from a critical humanist standpoint). Undeniably, the re-enactments are repeatedly misrepresented as a heroic act and as such certainly entail blatant fabrications, but they nevertheless make a claim to how things happened. It is also true that other re-enactments in the film are often highly stylized using visual tropes from gangster films, among others. But again, as the discussion of discursivity has demonstrated, these are textual features which cannot determine on their own whether something is fiction or not.

The last textualist feature I wish to address is indexicality. Unlike the preceding ones, indexicality is a standard marker of nonfiction. The most recent iterations of this view allow for non-indexical documentaries like animated ones (Nichols 2017, 12).³ But they do purport that if there is an indexical link between the image and its relevant content, then we are dealing with a documentary (Nichols 2017, 24–28). Phrased in analytic philosophy parlance, indexicality is a sufficient but not necessary condition for documentary. The problem, I propose, is that indexicality standardly understood does not even extend to most non-animated documentaries and nonfictions, analogue and digital alike.

Since Peter Wollen (1969), indexicality has generally denoted an automatic causal link and some form of contact between the photograph and its object. Much like a pawprint is an index of a dog imprinted through its weight so is a photograph of Sun an index of Sun imprinted through its light. But what holds for photographs of light-emitting objects like Sun does not hold for a much broader class of objects including dogs (and typical subjects of documentaries) which only reflect light. In other words, while in the case of the pawprint dog is in direct contact with the sand, in the case of the photograph the dog is not the source of light but an object against which the light bounces off. This means there is merely indirect contact between the photograph and the dog via light. Things get even further removed once it is recognized that audiences virtually never watch the original negative print of a given documentary film. In the case of analogue films, the screened print would have gone through at least one photographic printing where a new source of light is shone through the original. In this step, even the indirect contact where the relevant light touched both the object and the photograph is lost. (In digital cinema, copying does not even involve the image but its representation in the form of binary voltage states). Put succinctly, even if original negatives are indices of objects further prints are not (Slugan 2017).⁴

This is not to deny that there is a heuristic value to using textual features in distinguishing fiction and nonfiction. Keeping track of real-life people vs actors, naturalness vs staging, everyday events vs fantastic stories, on-location shooting vs spectacular sets, live action vs animation, trace vs CGI can go a long way. But it falls short as a more robust categorization strategy. Films and TV shows like This Is Spinal Tap (Rob Rainer, 1984) and Parks and Recreation (NBC 2009–2015) make extensive use of documentary aesthetics but are fictions, nonetheless (Roscoe and Hight 2001). Conversely, as Annabelle Honess Roe (2013) has argued, despite animation’s traditional association with fiction film there are numerous animated documentaries, including The Sinking of the Lusitania (Winsor McCay, 1912) and Tower (Keith Maitland, 2016). It is even possible to conceive a nonfiction film indistinguishable from A Trip to the Moon (Georges Méliès, 1902) and a fiction empirically identical to The Arrival of a Train (Lumière brothers, 1895). In the first case, the film – A Recording of a Trip to the Moon – is a single-take documentary recording of one particular screening of Méliès’ film. In the other example, a gallery projects the Lumière brothers’ film as The Arrival of a Train in Freedonia instructing visitors to make-believe train pulling into a train station of an imaginary country resembling France c.1900.

Intentionalism

Given that textualist accounts produce too many counterexamples I propose taking an extratextualist stance. This is usually the route taken by analytic philosophers interested in film. Most notably, Noël Carroll (1997) has proposed that documentaries are films of presumptive assertion while Carl Plantinga (2005) has defined them as asserted veridical representations. Crucial for both are the authorial intentions:

[In the film of presumptive assertion] the filmmaker intends that the audience entertain the propositional content of his film in thought as asserted. (Carroll 1997, 186)

When a filmmaker presents a film as a documentary, he or she not only intends that the audience come to form certain beliefs, but also […] that the use of motion pictures and recorded sounds offer an audiovisual array that communicates some phenomenological aspect of the subject, from which the spectator might reasonably be expected to form a sense of that phenomenological aspect and/or form true beliefs about that subject. (Plantinga 2005, 111)

For Carroll and Plantinga, then, films are documentaries if their authors intend them to be seen as making truth claims about what is represented. Importantly, although the two can coincide, the intentions in question are not what audiences infer from the film (textual features), but what the authors actually had in mind when producing the film (extratextual traits). The trouble with intentionalism is that it does not allow for films originally intended as nonfictions which are now treated as fictions (Slugan 2019a, 170–173).

If we take producers’ promotional strategies as revealing the intentions behind filmmaking, then we must concede that there are films like The Great Train Robbery (Edwin S. Porter, 1903) and Tracked by Bloodhounds (Selig Poliscope Company, 1904) which were originally promoted and intended as veridical representations but today count as fictions (Slugan 2019b). Consider their contemporary advertisements:

[The Great Train Robbery] has been posed and acted in faithful imitation of the genuine ‘Hold Ups’ made famous by various outlaw bands in the West (Edison Film Catalogue, 1903, no. 201)

Tracked by Bloodhounds Or

A Lynching at Cripple Creek

NEGATIVE ACTUALLY MADE AT COLORADO’S GREAT GOLD CAMP AND OF ACTUAL OCCURRENCE. (Selig Polyscope Company Catalogue, ‘Tracked by Bloodhounds’, Supplement 17, 1905, 1)

Whereas The Great Train Robbery is admittedly presented more as an illustration or a re-enactment of how a hold-up looks like rather than a reproduction of a specific one, Tracked by Bloodhounds is claimed to be an actual recording of a specific genuine event. A more recent example of such intentions is, of course, The Blair Witch Project (Eduardo Sánchez, Daniel Myrick, 1999). Given that all these films are currently understood as fictions, extratextualist approaches based on intentions also fail to deliver a criterion for distinguishing nonfiction from fiction ordinarily and narrowly construed. Intentionalists could dig their heels in and still claim that it does not matter how we categorize things now i.e. that the current categorization is mistaken precisely because it neglects original intentions. But a theoretical account should not simply define the distinction at will – the approach is only as good as it fits the ordinary understanding of what it tries to explain.

This specific issue is not a problem for Nichols (2017) who explains that the understanding of documentary also changes over time through the influence of producers and distributors, filmmakers, films and audiences. Producers and distributors influence how the film is going to look like usually by exercising pressure to use existing conventions. These may be subverted by filmmakers. Films which are initially outliers may come to popularize new conventions. Audiences, finally, have expectations in line with conventions. This resolves the problems plaguing intentionalists and proposes one explanation of how films can change their non/fictional status over time. But Nichols’ approach remains textualist because what matters most is the relationship of these four agents to changing textual conventions.⁵ John Ellis (2021), similarly, speaks of a ‘social constructionist approach’ but then goes on to focus on a range of textual features including patterns of misrepresentation, editing, sound, camera and mise-en-scène to explain ‘How Documentaries Mark Themselves from Fiction’. Such theories, therefore, are still vulnerable to the textualist problems which motivated extratextualist solutions in the first place because for them there will always be some textual element which is deemed fictive or nonfictive at a given point in time. I, by contrast, argue that re-enactments, for instance, were not necessarily fictions neither today nor around 1900. Indexicality defined as automatic causal link and contact, similarly, covered only a very small subset of nonfiction films from their first appearance until present day.

Reception-driven approaches

A different extratextualist approach heralded by some documentary scholars like Dirk Eitzen (1995) and Vivian Sobchack (1999) proposes that it is spectatorial experience that defines the status of a film:

On first viewing (for people who do not know the film’s secret), No Lies [Mitchell Block, 1973] is labeled as a documentary, perceived as a documentary, and interpreted as a documentary. For all intents and purposes, it is a documentary. (Eitzen 1995, 94)

One viewer’s fiction may be another’s film-souvenir; one viewer’s documentary, another’s fiction. (Sobchack 1999, 253)

Put succinctly, if a viewer experiences a film as non/fiction, then it is non/fiction. The strength of this view is that it also allows for historical change of non/fiction, but the problem is that it falls afoul the ordinary understanding of these categories. It is true that spectators can, say, view The Favourite (2019) as a nonfictional recording of whatever was in front of the camera during the shooting of Yorgos Lanthimos’ film. But although the spectators may experience it as nonfiction, The Favourite is still ordinarily understood as fiction. Reception-driven approaches simply replace ordinary public categorization with personal experience. But for categorization purposes the point is to understand how public categories come to be understood as ordinary rather than change the definition by fiat. Arguably, idiosyncratic film readings can bring about change in the film’s categorization much like specific films can break conventions of non/fiction and establish new ones. None of this, however, happens in isolation and no single element is the defining one.

Institutionalism

My proposal – which I dub institutionalism – combines the strengths of moderate textualism and reception-driven theories (allowing for the changing status of documentary and nonfiction) with those of intentionalism (denying that some textual elements are necessarily fictive and others nonfictive) while eliminating their shortcomings.⁶ First, borrowing from analytic philosophers, I argue that the distinction between nonfiction and fiction is not a matter of making truth claims as opposed to deceiving, but a question of whether audiences are supposed to imagine something or not. Second, building on the documentary tradition, I propose that a non/fictional status of a film may change over time. Third, diverging from both the analytic and documentary tradition, I claim that it is not the change in textual conventions or idiosyncratic readings that are crucial for this temporal fluidity of a film’s non/fictional status, but the negotiation between the agents of production, filmmaking, promotion, exhibition and reception on whether to imagine something or not. I will address these points in turn.

I have already argued that the ordinary narrow understanding of fiction does not include misrepresentation, deliberate or otherwise. But to gain a firmer grasp of the non/fiction distinction we should also briefly conceptualize what fiction is. According to a dominant current in analytic philosophy, fiction is construed as ‘mandated imagining’ (Walton 1990; Stock 2016).⁷

Imagining is an as if stance which may include actions (imagining swimming), objects (imagining an orange), experiences (imagining hearing), propositions (imagining that p), etc. Unlike competing accounts of fiction like ‘willing suspension of disbelief’, imagining is not internally contradictory, because imaginings are voluntary. By contrast, it is impossible to willingly stop not believing in something. If we do not believe that there is a rock band Spinal Tap, we cannot will ourselves into suspending that disbelief any more than we can wish ourselves into stopping not believing that Jeremy Corbyn is the Prime Minister of the UK. At most, we can refrain from behaving according to what we do not believe in. But abstaining from actions stemming from beliefs is distinct from suspending the content of those beliefs. In other words, when engaging fictions, we do not willingly suspend our disbelief about the content of representation but imagine that content.

Furthermore, unlike ‘belief’ which is true or false, imagining is ‘mandated’ or not. We can imagine both things that are true (Joe Biden as US President) and false (Hilary Clinton as US President). But mandate means something else. That imagining is mandated entails that although one is always free to imagine anything, a work is fictional only so long as everybody is supposed to imagine the same thing in engaging the work. When watching This Is Spinal Tap I am free to imagine that I am a rock star but I am not supposed to imagine it. What I am supposed to imagine is (among other things) a five-group rock band Spinal Tap as is every other spectator. By contrast, with Bowling for Columbine viewers are only supposed to believe what is represented, and not imagine it.⁸

Arguably, audiences are also supposed to believe certain things in many fiction films and supposed to imagine specific things in some nonfiction films. Historical dramas like Apollo 13 (Ron Howard, 1995) often pride themselves on accuracy and invite audiences to believe a plethora of information provided there. But even if everything presented in that film were true, the audiences are still supposed to imagine all those things on top of believing them, sufficient to call even a hypothetical historically perfectly accurate Apollo 13 fiction.⁹

The case with nonfiction is somewhat different. When watching especially evocative nonfictions like Shoah (Claude Lanzmann, 1985) it could be argued that audiences are supposed not only to believe what is represented but also imagine (parts of) it. Next to believing the testimonies of Holocaust survivors like, for instance, Filip Müller who recounts burning gassed bodies in the incinerator, it makes sense to say that audiences are also invited to imagine the bodies piled up ‘like stones’ he speaks of. Here, admittedly, we are dealing with fuzzy borders between fiction and nonfiction but in a different and much more limited fashion from how the boundary is usually presented in documentary theory. Where earlier radical textualism spoke of discursive tropes like staging which are present in virtually all documentaries including Shoah and more moderate theory focuses on the deceptive features of content which can be found in a range of documentaries, my proposal focuses on what is supposed to be imagined on top of only believed and as such identifies a much smaller subset of hybrid films. If Shoah is in this fuzzy category it is not because Lanzmann regularly prearranged the mise-en-scène where survivors made their testimonies (Abraham Bomba, another survivor who worked as a barber, for instance, speaks while cutting his friend’s hair). Rather, it is because of the mandate to imagine. In other words, although important, the mandate to believe does not define either fiction or nonfiction. The mandate to believe is only a necessary but not a sufficient trait of nonfiction.¹⁰ It is the mandate to imagine that defines fiction and the absence of this mandate to imagine that defines nonfiction.

But who or what determines what the audiences are supposed to do i.e. where do these mandates to believe or imagine come from? I have already argued that it is not to be found either in textual features, authorial intentions, or individual reception. The first, because nonfiction can use any elements usually associated with fiction and vice versa. The second, since a film can cross the non/fiction boundary over time despite authorial intentions behind it. The last because fiction and nonfiction are communal categories, not individual experiences. Instead, the mandate is a product of a negotiation between production, promotion, exhibition and reception factors. To flesh this out consider the case of phantom rides – films shot with a camera positioned in front (and on some occasions back) of a moving vehicle, most often a train, popular c.1900.

Nowadays, phantom rides are classified as nonfictional recordings of traversed vistas (Abel 2005). However, if promotion and reception at the time of their original appearance is consulted it becomes apparent that their contemporaries were not only expected to but also did engage them by imagining themselves travelling through space at then incredible speeds. Put in Walton’s terms, they constituted fictions because they mandated specific imaginings.

Consider, for instance, the newspaper reports of one of the first phantom rides – The Haverstraw Tunnel (Biograph, 1897) premiering in London in October 1897 – that Biograph was only too happy to include in its later bulletins:

If you desire a novel experience go to the Palace Theatre any evening at a quarter to ten and travel (in imagination) on the cow-catcher of the locomotive of a West Shore (American) Express through the Haverstraw Tunnel. (quoted in Niver 1971, 35)

Hitherto the audience merely watched moving objects, but in this recent addition the onlooker, by the aid of a little imagination, can fancy himself sitting an [sic] the bogie tracks of an engine travelling at the rate of sixty-five miles an hour with the landscape simply dashing towards him. (quoted in Niver 1971)

As we can see, the reports from Daily Chronicle and Music Hall regularly invite the viewers to imagine themselves riding on the cowcatcher. Importantly, the migration of reports from local newspapers to producer Biograph Bulletins also demonstrates the feedback loop between audience reception and promotion, constitutive of negotiating mandates. In this case, the press suggests to audiences certain types of imaginative engagement which are picked up by the film producers and included in their future promotional material which further regulates audience engagement thus reinforcing the mandate.

The negotiation between the reception and promotion of another Biograph film appearing in March 1899 – Brooklyn Bridge – demonstrates how in a space of only a few years these mandates could easily dissipate and how a film sharing the same key textual features – the presentation of an image from in front of a rushing vehicle – no longer mandated any imaginings:

The view starts with a swift rush down the incline from the station on the New York side, shows the Brooklyn trolley cars in motion on either side and pedestrians passing back and forth on the platform in the centre, then takes a plunge in the iron archway which covers the tracks of the electric line for about two-thirds of the distance. […] After spinning through the archway at an almost dizzy speed, the car suddenly rushed out […]. (quoted in Niver 1971, 42)

This picture bewildered the spectators last week, for people who have crossed the bridge one hundred times have never hitherto realized the immensity of the structure. (quoted in Niver 1971.)

In the review from Mail and Express, it is not the spectator but the car (or, at most, the view) that dives into the archway at incredible speed. For the Philadelphia Inquirer reporter, it is not imaginary travel that confounds the spectator but the scale of the bridge. Put differently, even if the film was initially intended as inviting imaginings on par with The Haverstraw Tunnel, Biograph did not shy away from reports emphasizing the illustrative power of Brooklyn Bridge instead and again made them part of their promotional strategy. In doing so, the mandates for imagining set up less than two years prior were dismantled, securing the present-day categorization of phantom rides as nonfiction.¹¹

In lieu of conclusion, it is useful to compare this institutionalist approach to competing views, textualist and extratextualist alike. Under more radical textualism, the unique and novel camera placement in phantom rides could be understood as a discursive trope introducing fictive elements. For intentionalists the intentions behind The Haverstraw Tunnel secure its fictional status. For reception-driven theorists each personal viewing determines the film’s status anew. But in all cases, this contradicts the present-day general understanding of phantom rides as nonfictions.

A more moderate textualism could explain how phantom rides migrated from fiction into nonfiction by arguing, say, that the convention of camera placement on a rapidly moving vehicle stopped being a marker of fictionality over time and became a sign of nonfiction instead. According to the institutionalist framework proposed here, by contrast, it is not whether some textual convention is assigned to fiction or nonfiction that is crucial but negotiating whether the whole film should be engaged primarily as a veridical representation or (also) as a prop for imagining in the first place. In other words, moderate textualism starts from textual features and declares whether they are fictive or not at any point in time. In that sense, it is not only that a film’s status changes over time but what counts as documentary – at one time it is this cluster of textual features and at another time a different cluster of such features. Winston (2013, 25), for instance, produces a table of varying list of traits associated with documentary over the years – classic, direct cinema, Kino Pravda, cinema verité, and post-1990. Institutionalism, by contrast, starts from the definition of fiction as mandated imagining, and then proceeds to determine for a given film and time, whether this mandate is operational or not.

More specifically, for moderate textualists it is the producers’, distributors’, exhibitors’, filmmakers’ and audiences’ views on relevant textual elements – camera placement, primarily – that determines whether phantom rides are fictions or not. For an institutionalist, textual elements are secondary in determining non/fictional status. For an institutionalist, phantom rides could use a completely different set of textual features like placing a camera next to the locomotive driver or facing a passenger in the dining car. Focusing on textual features would not solve any of her problems because at any given time, no textual element is either necessarily fictive or nonfictive. Rather, the institutionalist is interested in the power of the extratextual context that can shift the default engagement with a specific film – one involving belief and the other (additional) imagining – with ease. Do the filmmakers intend audiences to believe what is represented in the film or (also) imagine specific things by using the film as a prop? Does the promotional material emphasize the film as a recording or an illustration of certain events or are the audiences supposed to imagine these events (as well)? Does the exhibition venue or platform provide any specific instructions or cues on whether to trust what is presented or does it (also) invite imaginings? Do the audiences treat the film as making truth claims or do they engage it imaginatively (as well)? And above all, how are all these potentially competing positions between the agents of production, promotion, exhibition and reception negotiated?

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work has been supported in part by Croatian Science Foundation under the project UIP-2020-02-1309.

Notes on contributors

Mario Slugan

Mario Slugan, Lecturer in Film Studies at Queen Mary University of London, works on the intersection of film theory, history and philosophy. He has published in Early Modern Visual Cultures, Projections, and Film and Philosophy, and has authored three monographs – Montage as Perceptual Experience (2017), Noël Carroll and Film (2019) and Fiction and Imagination in Early Cinema (2019). He is also co-editor of the special issue of Apparatus titled Fiction in Central and Eastern European Film Theory and Practice (2019) and New Perspectives on Early Cinema (2021, forthcoming).

Notes

1 Cf. Corner (1996, 43), Gaines (1999, 1), Rosenthal and Corner (2005, 11), Chapman (2009, 15–16), and Winston (2013, 6). Here, ‘fictive’ is a property of a textual element, whereas ‘fictional’ is a trait of the whole text which may include both ‘fictive’ and ‘nonfictive’ elements.

2 For a critique of the idea that all narratives are fictional see Winston (1995, 113–119) and Carroll (1996).

3 Earlier versions like Currie’s (1999) account of documentary as trace excludes animated documentaries.

4 If contact is removed as one of the criteria of indexicality and we insist only on the object’s automatic trace, then the photograph is also an index of whatever preceded that object in an automatic casual manner. Say we photograph remains of a forest fire caused by lighting. We would then be forced to say that this is a documentary photograph of not only the charred remains but also of the lighting though the lighting was never in front of the camera.

5 Plantinga (2005) does allow that what counts as veridical asserted representation may change over time, but intentions remain crucial.

6 Nichols (2017, 11–13) speaks of “an institutional framework” which includes producers, distributors and exhibitors whereas I use the term “institutionalist” in a wider sense of bringing together agents of filmmaking, production, distribution, exhibition, and reception. In his earlier work, Carroll (1983, 24), similarly excluded audiences when speaking of films being “indexed” as fictions or nonfictions by “[p]roducers, writers, directors, distributors, and exhibitors”.

7 An influential dissenter further develops her theory in this issue (Friend 2021).

8 This is another difference from reception theorists for whom categories hinge on what spectators actually do, whereas under this framework the categories rest on what the viewers are supposed to do.

9 For an institutionalist account of fiction in film see Slugan (2023).

10 The mandate to believe covers believing both in assertions and in how the recording looks and sounds that Plantinga (2005) distinguishes.

11 See Slugan (2019a) for more details.