How documentaries mark themselves out from fiction: a genre-based approach

ABSTRACT

Assessments of the truthfulness or otherwise of documentaries are best understood as genre conventions which vary historically. Genre conventions are shared between audiences, filmmakers and institutions. Beliefs about the acceptable use of fictional techniques in documentary storytelling, particularly in television, are subject to occasional public controversies. The move from photographic to digital processes underlay one such controversy at the end of the last century. This was particularly the case around the so-called ‘docu-soaps’ on television, but public doubts about the truthfulness of documentary filming meant that many filmmakers developed new approaches. The result is that both public and professional documentary beliefs and practices have changed. Where once observational filming was seen as the bedrock of authenticity, newer approaches have developed a growing emphasis on the assessment of the ‘documentation’ of past events. They frequently gather and interrogate footage and other visual information from very diverse sources. They are seen as evidence in explicit attempts to reconstruct those events.

KEYWORDS:

• Documentation
• observation
• genre
• evidence
• narrative

The truthfulness or otherwise of ‘documentary’ is a historically variable matter. One period’s stunning truth becomes the next period’s inadequate account. To apply the concept of ‘truth’ to items of documentary footage (let alone to entire documentary texts) is therefore difficult without recourse to knowledges outside that footage. An understanding of the slippery relationship between documentary and truth requires two linked approaches: the application of genre theory, and an overall understanding of documentary footage as the evidence of particular and circumscribed encounters. This understanding is, in effect, a philosophical ‘set-up’ or way of thinking that concentrates on the idea of ‘evidence’ rather than the idea of ‘observation’. I am using the term ‘set-up’ here as a convenient term for the shifting habitual beliefs preferences and practices that underpin genres. This set-up is the result of always ongoing negotiations about the conditions of meaning, which can be summed up in the word ‘genre’.

Documentary is a genre of filmmaking if genre is defined, as Mittell and other modern scholars do, as ‘clusters of discursive processes running through texts, audiences, and industries via specific cultural practices’ (Mittell 2001): This definition of genre has largely replaced older definitions, mainly drawn from literary studies, which see genre as an exclusively textual phenomenon, where textual characteristics are sought in order to classify individual works in a genre category. Mittell’s definition sees genre as a social process of meaning construction involved in both the production and reception of texts (Mittell 2013). So a genre study requires attention to the historical and institutional context for which works were produced, as well as their subsequent usage.

Genre clues can usually be read from visual materials. Genre is often implicit, but particularly with documentary it is worth noting that television often signals genre explicitly. An upcoming programme can be described as ‘a documentary’ or ‘a comedy’, whereas this is less common with movies. A genre approach to documentary enables us to understand the truthfulness of documentary as a constant negotiation between filmmaker, institution and viewers. This crucially means that the truthfulness or fictionality of any documentary cannot be judged from examination of the film text alone. Instead, the assessment of truthfulness depends on a negotiation between different beliefs and perceptions. Filmmakers anticipate probable reactions to their films; users (viewers) bring their own understandings of what it is like to be filmed, and their assumptions about filmmaking practices. All deploy their own conceptions derived from their previous experience of examples of the genre. Media institutions regularly debate whether a documentary (or an entire sub-genre of documentary) is ‘true’ or ‘authentic’ or ‘factual’. The truthfulness of documentaries is decided in the conflict between these knowledges and beliefs, which are differentially held by groups and individuals. ‘Truth’ is habitually required of documentary, but it is a characteristic to which filmmakers and users alike aspire, rather than anything that is inherent in the footage. Whether that aspiration has been realised is constantly up for debate.

Several influential books on documentary (Bruzzi 2006; Ellis 2011; Nichols 1991, 2017) have adopted such a social constructionist approach. Their definitions of the documentary genre hinge on the human relationships which generate the documentary footage and structure the texts into which the footage is organised. Nichols seeks to classify different kinds of documentary inquiry. Bruzzi uses the concept of performance to analyse the increasing practice of inscription of (rather than effacement of) the presence of filmmakers into their filmmaking practices. Ellis charts the evolution of documentary from a belief in observation to increasing user scepticism which results from the impact of digital practices. This article seeks to develop that argument to highlight the shifting nature of conventions of truthfulness within the documentary genre.

In 1992, the distinguished theorist of visual culture WJ Mitchell argued that

Photographs … were comfortably regarded as casually generated truthful reports about things in the real world … But the emergence of digital imaging has irrevocably subverted these certainties, forcing us to adopt a far more wary and more vigilant interpretive stance … Today, as we enter the post-photographic era, we must face once again the ineradicable fragility of our ontological distinctions between the imaginary and the real. (Mitchell 1992, 225)

Mitchell took the position that the photochemical response of film to light had traditionally underpinned the truth claims of photography. Further, it was clear that any tampering with this initial relationship was usually detectable. In 1992, it seemed that the digital with its infinite malleability, would provide no such working certainty to underpin everyday imaging practices. From our perspective, there can be all kind of objections to this argument. We know that photochemical processes are by no means neutral: film stock was racist, balanced for white skin tones. We also know that digital tampering leaves its traces, both in the metadata (where it is provided) and in the details of image construction itself. But this is not the point here. Mitchell was arguing about the ‘comfortable regard’ that photographs ‘casually generated truthful reports’. He was arguing about the genre convention of photographic documentation and its truthfulness, about generally held beliefs rather than the precise configuration of photographic technologies.

Mitchell’s warning was prescient. A few years later, the scepticism of the ‘post photographic era’ had become widespread. It began to undermine the predominant generic belief in documentaries as observational records. In 1999, the truthfulness of television documentaries became a matter of public dispute in the UK, exemplified in the front-page headline of the populist Daily Mail newspaper in the UK: ‘Can we believe anything we see on TV? Ellis (2005) identifies this as a crisis in the credibility of documentaries and charts the response of institutions and individual filmmakers. Dover (2004) examines important evidence of the way that ordinary TV viewers participated in this debate, the sure sign of a crisis in the generic conventions of documentary.

The crisis was resolved by the move towards more subjective and authored documentary practices analysed by Bruzzi and others, along with some high-profile institutional changes and the occasional career casualties along the way. The result is an uneasy truce in which the issue of factual filmmaking and truth maintains a higher profile than was once the case. This uneasy truce means that documentaries are sometimes dragged into wider controversies about the status of ‘factual’ TV (a far wider genre than TV documentary). In 2020, the suicides of three former participants in ITV’s reality game show Love Island revived the question of ‘exploitation’ and factual TV, leading to reforms in the practices of finding participants and guaranteeing their mental safety (Broadcast 31 July 2019, 5 September 2019). This issue of duty of care on the part of factual programme makers also engulfed the daytime talk series The Jeremy Kyle Show which was taken off air permanently amid fulsome apologies from the broadcasters (Broadcast 15 May 2019). The issue of ‘duty of care’ for participants haunts documentaries, and some documentary practices were dragged into the controversy. The new broadcaster guidelines in the UK which resulted from this problem have changed TV documentary practices.

Elsewhere, wildlife documentaries have come under increasing scrutiny for their truthfulness. Controversies about the levels of intervention into the ‘wild’ that are sometimes involved have even engulfed the pinnacle of wildlife documentary practice, the BBC blockbuster series fronted by Sir David Attenborough. In one case, footage from remote cameras of a polar bear settling in to hibernate was combined with zoo-shot footage of a polar bear giving birth ‘under the snow’ (Broadcast 12 December 2011). The combination of the two kinds of footage enabled a story to be told, but it was not signalled as having transgressed any genre conventions. The controversy it generated was particularly interesting because it became the vehicle for a questioning of the use of fictional devices in documentary filmmaking. The binary oppositions between factual and fiction, documentary and drama, manipulation and observation, are major primary definers of the documentary genre. Up to this point, wildlife filmmaking had been one of its most enthusiastic proponents of lack of manipulation. Generally, the most prized wildlife footage is that which involves the least human intervention. The purer forms of wildlife filming base themselves on the principles of observation and the minimum use of artifice in generating their footage. However, conventions of storytelling figure heavily in how the programmes are assembled. The more observational the footage, it seems, the more that techniques of narrative organisation are needed to make sense of it. But its truthfulness as footage lies in the guarantee (and/or the belief) that there has been no human intervention into the lives of the animals.

In the wider area of factual filmmaking in general, there is a constant slippage between the techniques of fiction and the practice of filming, particularly in popular television and entertainment factual series. Factual TV work like The Only Way is Essex or Keeping Up with the Kardashians occupy a borderland between the factual and the fictional. They take the outward appearance of documentary-style filming and combine it with the pre-planning and scripting of a fiction series. The bond between the two is cemented by improvisational performances. Such series are the symptoms of a larger process: the constant social renegotiation of the fiction/non-fiction divide. This divide underpins the idea of ‘documentary’ itself. One of the bedrock definitions of ‘documentary’ is the absence of fiction and the dominance of fact or truth. However, the boundary between the two is permeable and shifting, as the work of Rouch and the Cinema Verite school amply demonstrate by foregrounding the interaction of camera and participants. Such foregrounding is entirely absent from the mock-observational practices of popular factual TV.

Nevertheless, popular beliefs about the truthfulness of documentaries are inevitably affected by the goings-on of factual TV and even reality TV. These popular forms habitually borrow the clothes of documentary, adapting its practices and imitating its styles. The genre approach emphasises that such activities will inevitably affect documentary practices, particularly in regard to difficult concepts like ‘truthfulness’. The community of beliefs (and the divergence of beliefs) about documentary that exist between viewers, filmers and the institutions of TV and cinema are constantly being renegotiated and renewed as a result.

The relationship is inherently unstable because documentary uses the techniques of fiction. At various moments in the history of documentary, different combinations of fictional techniques have been adopted in documentary practice, both to generate material and to structure comprehensible and engaging texts. Until observational techniques became widespread in the 1960s, it was entirely acceptable for filmmakers to stage scenes with their participants. For Nanook of the North Flaherty even used his participants to play fictional versions of their forebears (Grace 1996). By the time of the 1930s British documentary movement, it was routine for amateurs to ‘play themselves’. Such practices are no longer tolerated as truthful. One of the controversies during the crisis of 1999 related to a participant in a ‘docu-soap’ about learning to drive who was directed to recreate waking in a panic in the middle of the night. The event had happened, but without the presence of a camera. Elements of direction of participants are virtually inevitable, but cases such as this defined a far more restricted use of the practice than before.

Instead of direction, there is currently an increasing use of casting to find participants who require less direct input into their performances. The practice of finding participants involves many judgements which are not available for scrutiny by the end users of documentaries. We cannot know who was considered and turned down. All we see are the people who do appear on screen. Documentary makers carefully select those who will appear for the exemplary nature of their situation, for how well they can appear on camera, for having a reasonable level of articulacy, for their physical type. The documentary makers’ decisions are usually scrutinised by those who commission or provide finance and support, through the pervasive use of audition and pilot material. This choosing of participants constitutes a process of casting which is almost as explicit as that for reality TV programmes or for fiction. This process of casting is an essential process for documentary making in the current conjuncture. It is nowhere made explicit to viewers, but its existence as a practice does not, at the moment at least, disturb the conventions of truthfulness that govern the contemporary documentary.

The presence of readily accepted fictional techniques is even more pronounced in the construction of documentaries as texts. All documentaries organise the evidence of past encounters into coherent texts. We do not view rushes; we view edited texts. The separation in time and space between viewers and the original events has to be bridged by using the structures of narration. The separation in time not only enables but also requires the retrospective organisation of meaning. This is what fiction does. A fictional story always knows its own ending and is constructed towards delivering it. A documentary is not tremendously different. The ethical aim may be to inform or to shock rather than to entertain, but at the level of narrative techniques there is surprisingly little difference.

Documentaries use the elision of time within and between sequences. The same conventions of editing are applied: there is no difference between the fictional edit of a person going through a door and a factual edit of the same footage. Actions are elided in both. Equally, documentaries compensate for the fact that their viewers were not present at the original events by providing multiple points of view on those events. The fluidity of cinematic point of view is central to the way that cinematic fictions tell their stories. The art of fiction direction lies in providing (or refusing to provide) the best possible point of view on an action at a particular point, and then swiftly substituting another best view (Bordwell 2012). Increasingly, documentaries are able to operate in the same way, taking advantage of pre-existing footage (e.g. surveillance camera footage) or archival material, as well as the cheapness and ubiquity of cameras compared to earlier in the genre’s history.

Many contemporary documentaries make use of further elements of fictional technique. An important forerunner of this tendency was Errol Morris’s Thin Blue Line (1988) which pioneered a style that has since become almost a cliché. It constructed a sense of unfilmed past events through the marshalling of significant details in tight closeups. These were not quite reconstructions, but they were used to underpin the often conflicting or partial accounts of participants in the events. It would be tricky to characterise such footage as ‘documentary’, yet the film as a whole certainly is a documentary. Another key element in Thin Blue Line is a hypnotic score by Philip Glass. This music insistently reframes the contributions of interviewees into an entirely different mood register. Both are techniques of fiction, but they have since become acceptable documentary practice. Until the time of Thin Blue Line, observational documentaries had tended to establish their credentials as factual by refusing such techniques. Such are the changing conventions in the borderlands between fact and fiction in the genre conventions of documentary. The genre negotiation establishes which fiction techniques are acceptable at any given period. Filmmakers, viewers and institutions are involved in (sometimes antagonistic) dialogue about which fiction practices are acceptable and which ‘cross the line’. This negotiation conditions the techniques used in both shooting and textual construction.

The current documentary genre set-up is based on the concept of evidence rather than that of observation. This set-up has two aspects. The first is the foregrounding of aspects of textual construction which provide evidence of how the film was made. The second is a more forensic attitude to the documentary as ‘documentation’ of events, which at its furthest extent has led to the comparison of different forms of documentation of events in order to discover a more complex truth behind competing narratives. Morris’s film is an early attempt at the latter approach, made before the proliferation of personal video recording devices, CCTV, satellite surveillance and the other technologies that have transformed the possibilities of documentary. Morris’s use of devices derived from fiction provides the necessary new organisational principles for this endeavour.

The devices that Morris adopts were consciously avoided by the earlier observational approach to documentary. Graef and Stewart’s Police: A Complaint of Rape (1983) was as influential as Morris’s film in that it provided vivid evidence of appalling police procedures. Around this filmic document, shown on BBC1, the feminist campaigns for better treatment of rape victims could finally achieve overwhelming purchase in the public debate. The film is a classic of observational documentary, characterised by its refusal of many of the devices of fiction. There is no music and no narration. There is hardly any shifting of point of view. The woman given a relentless grilling by unsympathetic and mainly male officers is seen entirely in back view. The evidence presented for the viewer is solely what the team’s observational cameras had captured on 16 mm film. Nothing else is admitted: no maps of the places that she describes, no close-up of the statement that she is forced to sign, no interviews with third parties or campaigners. There is, famously, not even an intervention by the crew themselves when the victim appeals for their support. It remains a relentless document of bullying, and was a successful film precisely because of its refusal of almost all the devices of fiction.

Yet today, now that the conventions of the documentary genre have changed, it is beginning to gather aberrant readings. When TV documentaries like 24 Hours in Police Custody use multiple cameras in a fixed rig format as well as roving cameras, CCTV footage, retrospective interviews and so on, the very purity of the observation concept produces misunderstandings. The fact that the victim’s face is never seen leads some student viewers to regard her as untrustworthy, to side with her interrogators in thinking that she is lying. This reading could be seen as ‘misunderstanding’ the premise of observation which suppresses any consideration of the documentary encounter. But if it is a ‘misunderstanding’, then it is underpinned by a shift in generic values. Nowadays it is expected that documentaries should give an account of themselves, to demonstrate the nature of their evidence-gathering activity.

Observational filmmaking relied on repressing aspects of its own production: not only by minimising the presence of the crew at the moment of filming, but also by its attempt to ‘put the viewer into the space of the filming’, to provide a surrogate presence for the viewer at the action. Observational documentaries pushed to its practical extreme the belief that film allows its viewers to ‘see it with your own eyes’. This represses the fact that documentaries are always accounts of past events, of past encounters between filmers, individuals and their actions. This is why observational documentaries, in their editing, refuse many of the devices of fiction and even of expository documentaries. Sometimes this goes to extreme lengths. Don’t Look Back (Pennebaker 1967) is faced with the problem of having to identify many of its large cast of characters around Bob Dylan. A brief commentary (or even superimposed captions) would have been the obvious solution to a more conventional filmer, but it would have interposed a disturbing element into the observational relationship to which the film aspires. So Pennebaker inserts scenes whose entire rationale seems to be that, during the scene, individuals are identified by name. Nowadays, in a different generic set-up, this choice seems excessive. At the time it was ground-breaking.

The new set-up of the documentary genre emphasises those aspects of filmmaking which were repressed by the classical observational project. It foregrounds the nature of the encounter between filmers and their subjects. Frequently, this is dramatised as the filmer’s quest for knowledge, which of course admits doubt and even deception on the filmer’s part. It also exploits the temporal distance between those past encounters and the current encounter of viewer and text. Editing becomes a major element rather than a necessary minimum. The film text is constructed around different kinds of filmed and visual evidence, marrying observational footage with retrospective interviews, surveillance footage, ‘user generated content’ and any other material that can be found.

Documentary narratives have become narratives of the interactions of filmers as varied as Nick Broomfield, Kim Longinotto, Michael Moore and Iris Zaki. Broomfield began as an observational filmer with films like Juvenile Liaison (1975), shot by Joan Churchill. By the time of Chicken Ranch (1983), Broomfield began to experiment with approaches that foregrounded the filming process itself, eventually settling into an approach where he played the role of a hapless sound recordist. This strategy unsettled difficult interviewees such as the South African neo-Nazi leader Eugene Terreblanche in The Leader, His Driver and the Driver's Wife (1991). The viewer is in on the disguise as Broomfield constructed his films as his personal journey to find a truth (complete with maps in this case). Michael Moore emerged with Roger and Me (1989), which featured a lot more of the ‘me’ than the ‘Roger’ as Moore conducted an angry doorstepping campaign against General Motors and its closure of plants in his hometown of Flint, Michigan. The film has been picked apart by critics for its deceptive stance (see for example Bernstein 1994), and for its manipulation of the chronology of events for greater dramatic effect (Jacobson 1989). Yet it remains a touchstone for the emergence in feature-length cinema documentaries of an autobiographical approach that puts the filmers themselves front and centre in their films. Moore’s approach is aggressive and self-aggrandising where Broomfield’s is off-key and self-deprecating. Moore’s rather male approach suits a campaigning style but, as many critics have pointed out, leaves no space for collective or individual action by viewers.

Kim Longinotto’s work is often presented as ‘cinema verité’ as it uses a number of observational strategies and seeks to reduce the distance between viewer and the past events that she has recorded. On closer examination of her films, it is clear that she is just a committed to her representing her subjects and their situation as Michael Moore. The auto workers of Flint are largely absent from Moore’s film, however. The women in Longinotto’s films like Rough Aunties (2008) know that they are being filmed by a sympathetic filmer whose aim is to portray their lives and mission. They occasionally speak directly to camera, or more accurately to the woman behind the camera. Longinotto deliberately leaves in those moments where her presence becomes obvious, providing a distinct authorial presence, clearly communicating her sympathies and humanity.

Iris Zaki has adopted a radical approach that she has named ‘the abandoned camera’ (Zaki 2018). Her films are a series of conversations between Zaki and her subjects who are found by chance in a structured situation (hotel reception, hairdressers). She is in front of the camera, not behind it. Exploiting affordances of digital filming which were unavailable to 16 mm documentarians, she sets up the camera and lets it run as she is free to converse with her subjects. For Women in Sink (2015) the conversations take place as Zaki is washing women’s hair in a salon in Haifa. These are genuine dialogues rather than interviews, a strategy that Zaki takes further with Unsettling (2018) where she sets up her cameras outside a café in a Jewish settlement on the West Bank and simply invites the inhabitants to talk to her, a radical from Tel Aviv.

All these approaches have one thing in common. They are explicit that they are documenting past encounters by showing to different degrees the artifices employed in their production. Some, like Zaki, expose the conditions of their production, displaying their technological arrangements as a necessary part of understanding the filmic encounters. All inscribe the filmmaker as interlocutor or as investigator. They ensure that viewers understand the filming relationships by incorporating material about the subjects and their consents. They often use commentary or even speech to camera to express the filmer’s doubts and misgivings about the enterprise of filming they have undertaken. This subversion of authorial and textual authority can be extended to the point where the accidents and missteps of filming are occasionally foregrounded. For the modern documentarian, the spoiled interview or the refused encounter can be as rich a source as a successful interview. This is sometimes necessary as documentarians are more and more frequently refused co-operation in their endeavours. The access to police procedures granted to Roger Graef for his 1983 Police series would be refused point blank by the public relations professionals who have proliferated since this series was made. The access that can be secured is for fixed rig filming (e.g. 24 Hours in Police Custody) where, as Littleboy (2013) points out, the physical investment in equipment installation together with the ‘embedded’ nature of the filmers alter the balance of power in favour of the institution being filmed.

The situation has changed since the days of the strict observational approach to documentary exemplified by the films discussed earlier like Police: A Complaint of Rape and Don’t Look Back. The proliferation of media and image recording technologies has encouraged a greater awareness, and a greater suspicion, of the problems of being filmed. The circulation of moving image material has been revolutionised so that documentary material can circulate far beyond its first intended context (see Ellis 2011). Filming that foregrounds the circumstances of filming is one response to this crisis. The other is to exploit for documentary purposes the very affordances that this new situation has provided. In the digital era, there is always already other visual or aural documentation. Taking advantage of this, documentaries can be constructed using personal photos and recordings; selfies and social media posts; press photos; satellite mapping information; recorded songs, music and sounds from streaming sites; archive footage and sound recordings (some even drawn from earlier observational documentaries); drawings, paintings and prints. Such material is commonly found in more expository documentaries, but it is often neglected in reflection upon them because it is used for illustrative purposes. It was once disparaged among documentarians as ‘wallpaper’, with a lesser status than the material generated by those filmers themselves. However, the subaltern tradition of the ‘archive documentary’, whose antecedents go back as far as Esfir Shub’s Fall of the Romanovs (1927), was based on considering ‘documentary’ as a compilation and assessment of all forms of documentation of an event, rather than a practice that privileges the specifically shot filmic elements overall others.

This approach regards documentary as concerned with documentation. Fortified by the new affordances of digital image production, it now provides a filmmaking practice that widens the activity of reconstruction and setting of accounts against each other that was once used by Morris in Thin Blue Line. The contemporary work of the Forensic Architecture group assembles a dizzying number of audiovisual sources to assess conflicting accounts of an event. The film Killing in Umm al Hiran (2019) is the linear film version of a work that also existed as a gallery piece shortlisted for the Turner Prize in 2018 (https://www.tate.org.uk/whats-on/tate-britain/exhibition/turner-prize-2018/forensic-architecture). It uses eye-witness camera material, satellite surveillance footage and photos (including thermal imaging), police footage, news footage, on-site reconstruction of the action, mapping and modelling, all to discredit the official version of the shooting of an innocent Bedouin by Israeli police engaged in a raid on their community. The victim was accused of having shot dead a policeman. The various elements of evidence were brought together and married around the timeline provided by the video footage made by a journalist on the ground. This footage by itself reveals very little visual evidence. But it does provide a timeline (thanks to the timecode on all digital footage) and distinctive sounds that can be used to anchor all the other forms of evidence into an account pieced together from multiple points of view. Much of the footage was acquired by Forensic Architecture from public sources, everything from news bulletins to the police’s own press conferences. The collation of this material was supported by fieldwork which constructed a detailed 3D map of the location of the events, and reconstructed the way in which the victim’s car had travelled. This approach regards documentary precisely as documentation, documentation that is interrogated using its own digital metadata (timecodes, geolocation) and further digital analysis of both image and sound. In effect, it is using the new affordances and the excessive evidentiality of digital filming to overcome ‘the ineradicable fragility of our ontological distinctions between the imaginary and the real’ that Mitchell foresaw as its consequence.

To approach documentary as a genre is to reveal the shifts that have taken place over the past seventy years and more. It shows that the observational problematic has been replaced by an evidential approach. Observational filming is alive and well, of course, but it now constitutes just one form of evidence among many potential sources. Observational approaches are open to challenge and doubt in a way that was hardly ever the case before the 1990s. A genre-based approach to the issue of the truthfulness of documentaries reveals that the truthfulness or fictionality of any documentary cannot be judged from examination of the film text alone, but also requires attention to the historical and institutional context for which it is produced. This approach also emphasises that documentaries attempt to present evidence of encounters in the past through mechanisms that are often closely allied to those of fiction (narrative construction, time elision, multiple points of view, insertion of mood inducing material, sound mixing etc.). In order to be perceived as ‘documentary’ such films have to inscribe into themselves the markers of their authenticity as records of past encounters. Since the crisis in documentary credibility in 1999, the documentary genre has shifted away from the once common definition of documentary as ‘observation’ towards a different emphasis, one that sees documentary as evidence of past encounters.

Disclosure statement

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

Additional information

Notes on contributors

John Ellis

John Ellis is currently Professor of Media Arts at Royal Holloway University of London. He has worked both as a TV producer and a Media Academic. In 1982, he was one of the first independent producers commissioned by Channel 4 in the UK, making their Visions series through Large Door. The company continued to produce documentaries until 1999 and Ellis also served as vice-chair of the independent producers’ association PACT. His publications include Visible Fictions (1982), Seeing Things (2000) and Documentary: Witness and Self-Revelation (2012). He has led several research projects into television, culminating in the ADAPT project for the European Research Council (2013–2018) which studied how television used to be made using an innovative method outlined in our book Hands On Media History (2018). He chairs Learning on Screen which runs the UK’s largest streaming archive of digitised UK TV, Box of Broadcasts for UK education, and is an editor-in-chief of VIEW, the online journal of European TV history.