Twisted Tales
Cognitivism and Narrative Distortion

An essay for 'Media Technology and Culture: Theory and Practice'
MA in History of Film and Visual Media
David Mitchell
December 2002


The 9th Laterna Film Academy was held in Pécs, Hungary in September 2002. The subject, 'Representation Of Time And Space In Film - Film Theory and Practice Across Disciplines' was very germane to this essay, indeed several papers on the third day discussed Memento. Despite writing to the organisers and authors concerned, I have been unable to get copies of most of the papers presented (but see the Acknowledgements below). If I had, this essay might have taken a different course. For details of the conference program, see:



I should like to thank Professors Edward Branigan (UCSB), Gregory Currie (University of Nottingham) and Maureen Turim (University of Florida) for helpful suggestions and useful pointers.

I can imagine flashback films yet to be made that do not simply echo the brave transformations of the "new wave" but make of the act of remembering and retelling something as yet unexplored.

- Maureen Turim (1)

Writing in 1989, and looking back at films like Last Year at Marienbad (Resnais, 1961), Turim accurately predicted Christopher Nolan's film Memento (2000), a film which did indeed use flashbacks in a way 'as yet unexplored'. In recent years several films with unusual narrative structures have been both critical and commercial successes. If Memento is the most obvious example, Pulp Fiction (Tarantino, 1994), The Usual Suspects (Singer, 1995), Run Lola Run (Tykwer, 1998), and The Sixth Sense (Shyamalan, 1999) were all successful films (2) that made major use of narrative distortions. It seems that modern audiences have become more comfortable with narrative complexity. (3)

The academic study of films normally views them through the lenses of theory. This essay will adopt the opposite viewpoint, viewing theories through the lenses of film. In science progress is made by experiment; when theories are tested against extreme examples the robust theories survive while weaker ones break under the strain. How well do theories of narration explain or predict the audience response to a film like Memento with its reversed narrative?

There are of course many theories of narrative, but I shall restrict myself to the more recent class of theories influenced by cognitive science and exemplified by the work of writers like Bordwell, Branigan, Carrol and Currie. Such cognitivists 'reject dominant views in film studies: Saussurean linguistics, Metz’s semiotics, and psychoanalytic accounts of neurotic viewer responses to or ideological victimization by film'. (4) This rejection makes such theories more amenable to testing. As Bordwell points out:

The hermeneutic bent of film studies leads to the practice of describing texts in a metalanguage derived from a theoretical doctrine. But a description, even a moving or pyrotechnic one, is not an explanation. It does not show that the particular case manifests a general tendency. ... Cognitivism ... provides explanations rather than explications... Like all theorising, it asks the Kantian question: Given certain properties of a phenomenon, what must be the conditions producing them? It then searches for causal, functional, or teleological explanations of those conditions. (5)

These claims for a scientific basis and explanatory power imply that difficulties in providing explanations can be used to refine or replace the theory, just as the difficulty that Newtonian mechanics had in explaining the Michelson-Morley experiment led to its eventual replacement by Einstein's Theory of Special Relativity.

Before we begin, it makes sense to outline the cognitivist approach. As Bordwell and Carrol are at pains to point out in Post-Theory, perhaps the best compendium of cognitivism, 'you do not need a Big Theory of Everything to do enlightening work in a field of study'. (6) But Bordwell has also said that 'any theory of the spectator's activity must rest upon a general theory of perception and cognition'. (7) He goes on to say that cognitivism is:

best characterized as a stance. A cognitive analysis or explanation seeks to understand human thought, emotion, and action by appeal to processes of mental representation, naturalistic processes, and (some sense of) rational agency. ... Like feminism, cognitivism is not so much a Theory as a perspective which includes diametrically opposed theories. (8)

Dissatisfied with post-structuralist theories of narrative emphasising the film viewer's unconscious or ideologically coded responses to screened images, cognitivists instead analyse cinematic comprehension in terms of active viewers' ordinary psychological processes and strategies of problem solving. As Kevin Sweeney puts it:

Narrative film viewing, [cognitivists] claim, consists of the same sorts of top-down (conceptualising and inferring) and bottom-up (sensory, data-driven, automatic) psychological processes that perceivers use to understand events in the world around them. (9)

I began by listing five recent films that employ narrative distortion. The vast majority of successful films, from Hollywood or elsewhere, use the techniques of classical continuity editing (what Burch calls the 'Institutional Mode of Represention'). They present us with seamlessly edited narrative structures - a beginning, middle and end (in that order) - where flashbacks are clearly signaled as such and are directly tied to the narrative at both ends. As Metz says, 'a narrative is not a sequence of closed events, but a closed sequence of events'. (10)

Films which run counter to these principles are typically, and usually deliberately, less immersive due to breaks in the diegesis. They are often seen as disturbing or 'difficult'. For such films to be popular successes, the general audience, and not just the cinematically-aware critic, has to enjoy the process of disentangling them, of solving the cognitive problems they pose.

If the films I have mentioned all play narrative tricks, they do not all distort narrative in the same way. Run Lola Run tells three variations of the same story, one after the other, with embedded flash-forwards. Like Sliding Doors (Howitt, 1998) or Blind Chance (Przypadek, Kieslowski, 1981), it is an example of what Bordwell calls the 'forking path' plot, a reference to The Garden of Forking Paths, a short story by Jorge Luis Borges. Pulp Fiction tells three interlocking stories out of time sequence, and ends more or less where it began. Memento rearranges time in an even more drastic fashion, telling its story in reverse, as a series of flashbacks going back in time. The Usual Suspects and The Sixth Sense do not play tricks with time, but instead with plot. Apparently conventional, linear narratives, their distortion lies in last-minute plot twists that completely change our understanding of what we have been watching.

Distinguishing the films in this 'textual' way is not as useful as we might expect, because it does not focus on what, from a cognitive point of view, is the key point. As Hugo Münsterberg argued in 1916, (11) a film device should be defined in terms of its effects on the mind - on attention, memory, imagination, and the emotions. Looked at this way, all of the films achieve their effects by stimulating the spectator's memory and attention and by making the spectator perform specific imaginative acts. It is these mental processes that cognitivism must elucidate, and show how they can be made pleasurable.

Let me start with Pulp Fiction. Tarantino's film consists of three stories, preceded by a prologue (before the titles) and followed by an epilogue which is a replay of the prologue from a different point of view. What makes the film interesting from a narrative viewpoint is that although all five parts are related, both in terms of plot and characters, the temporal relationships between plot events are quite complex. The diagram below shows both the plot sequence and the story sequence.

Plot (Film Sequence)
10. HoneyBunny and Pumpkin in the diner
Vincent Vega and Marsellus Wallace's Wife
2. Jules and Vincent in car
3. shootout (1)
12. Butch and Marsellus in bar
13. Vincent and Jules hand over briefcase
14. Vincent buys heroin, shoots up, collects Mia
15. Vincent and Mia at Jack Rabbit Slim's
16. return to Marsellus' home, Mia's overdose
17. Vincent takes Mia to Lance
18. Vincent stabs Mia with Adrenaline
19. Vincent takes Mia home
The Gold Watch
1. Butch's childhood (dream)
20. just before Butch's fight
21. after the fight - Butch's escape
22. Marsellus swears vengeance
23. Butch's taxi ride to motel to meet Fabienne
24. Butch and Fabienne at night
25. (morning) Butch discovers watch missing
26. Butch goes to his apartment, kills Vincent
27. Butch drives off and hits Marsellus
28. Fight between Butch and Marsellus
29. Butch and Marsellus captured
30. Butch escapes and saves Marsellus
31. Butch collects Fabienne and rides off
The Bonnie Situation
4. shootout (2) - miraculous escape of V and J
5. Marvin's death in car
6. Jimmy's
8. flash-forward - Bonnie sees the body
7. the clean-up
9. Monster Joe's
11. Jules and Vincent in the diner
End Titles

Story (Temporal) Sequence

1. Butch's childhood (dream)

2. Jules and Vincent in car
3. shootout (1)

4. shootout (2) - miraculous escape of V and J
5. Marvin's death in car
6. Jimmy's
7. the clean-up
8. flash-forward - Bonnie sees the body
9. Monster Joe's

10. HoneyBunny and Pumpkin in the diner
11. Jules and Vincent in the diner

12. Butch and Marsellus in bar
13. Vincent and Jules hand over briefcase
14. Vincent buys heroin, shoots up, collects Mia
15. Vincent and Mia at Jack Rabbit Slim's
16. return to Marsellus' home, Mia's overdose
17. Vincent takes Mia to Lance
18. Vincent stabs Mia with Adrenaline
19. Vincent takes Mia home

20. just before Butch's fight
21. after the fight - Butch's escape
22. Butch's taxi ride to motel to meet Fabienne
23. Marsellus swears vengeance
24. Butch and Fabienne at night
25. (morning) Butch discovers watch missing
26. Butch goes to his apartment, kills Vincent
27. Butch drives off and hits Marsellus
28. fight between Butch and Marsellus
29. Butch and Marsellus captured
30. Butch escapes and saves Marsellus
31. Butch collects Fabienne and rides off

The Structure of Pulp Fiction

As can be seen, while each of the three stories is presented in a sequential fashion, to reconstruct the 'whole' story requires a complex re-ordering of events. For example, the final story we watch actually occurs between episodes of the first. In fact the anachrony is so complex that it is unlikely that a spectator, after a first viewing, would be able to disentangle them well enough to answer questions about the time relationships between the stories, such as the following:

Does the boxing match (in the story of The Gold Watch) occur before or after the scene at Jack Rabbit Slim's (in the story of Vincent Vega and Marsellus Wallace's Wife)?

The only clue that helps us answer this question is the brief presence of Vincent and Mia in the scene labeled 'Marsellus swears vengeance' which occurs soon after the boxing match in the second story. A very short conversation between the two at that point makes it clear that the visit to Jack Rabbit Slim's must have occurred much earlier. (12) Similarly we have little evidence for the temporal position or plot relationship of the prologue until we watch the epilogue (although a keen-eyed viewer may have noticed a brief rear view of Vincent Vega in the prologue as he goes off to the john).

Pulp Fiction met with great critical success; as well as an Oscar for best screenplay, the film won many other awards. That Tarantino's playful rearrangement of plot events was also enjoyed by audiences is clear enough from the success of the film at the box-office. While its unusual structure and post-modernist 'tricks' might explain its critical success (after all they gave critics a great deal to talk about), the cognitivist has a more difficult job explaining the audience's reaction. Is it possible to enjoy and understand the parts without making complete sense of the whole?

Tarantino has said many times that if he had presented Pulp Fiction as a novel, nobody would have expressed any surprise at its unusual temporal structure; that novelists play tricks like this all the time and that he enjoys transposing such literary techniques into films. (13) This does not directly help the cognitivist explain audience reactions - after all Pulp Fiction is a film, not a novel, and most films do not employ such techniques. Nevertheless Tarantino's remarks do suggest that he planned the film with every intention of allowing it to be 're-read' for further enjoyment. It is certainly true that the film is packed with references to other films and that the songs are deliberately arranged to comment on and highlight key plot events. Much of this design is not apparent on a first viewing. Just as a proper understanding of a novel often requires the text to be read in a non-linear fashion, turning back to re-read parts which only make sense in the light of later episodes, so Pulp Fiction repays repeated viewing by revealing more of its structure. (14)

That this complex layering of clues and references is only partly perceived on a first viewing does not mean that spectators will be troubled or upset by it. Indeed, the reverse is likely to be true. I claimed earlier that it was by stimulating our memory, attention and imagination that Pulp Fiction made its temporal complexity enjoyable. Acceptance, even enjoyment, of the lack of closure that such narrative complexity induces can be explained in terms of the anticipation of future closure, the sense that further viewings will make things clearer. We see and understand enough while watching the film to believe that there are things we missed; that if we had only paid enough attention we might have seen more. Pulp Fiction is after all a cult film, and one of the key features of cult films is their ability to encourage repeated viewings, to elicit pleasure over and over again. As Susan Feagin remarks:

Significant cognitive effort may be required to understand a story given the temporal disparities between narrative time and story time. Pulp Fiction ... no doubt enjoys some of its popularity because of the way one's understanding of the story develops along with one's understanding of the structure of the film. (15)

Carl Plantinga calls this 'reflexive criticism and appreciation', his fifth source of spectator pleasure. As he says:

Any technique that draws attention to itself, and away from the story, is thought to transgress that fundamental rule. Yet although this is a common rule of thumb, it is also one that is commonly ignored, as reflexive works become increasingly popular on both film and television. We also enjoy the intertextual pleasures of the text. (16)

Let us turn our attention to Run Lola Run. Each of the three stories it presents is straightforward enough, though the short embedded 'flash-forward' summaries of characters that Lola meets as she runs are an unusual innovation. There is no problem in understanding each story; it is their relationship that is confusing. Seeing the whole as more than just the sum of the three separate parts is the cognitive problem. What do the alternatives mean? Are there more that we are not being shown? Why do different things happen in each one? Which, if any, is 'true'?

As it happens cognitive 'explanations' of Run Lola Run, along with Sliding Doors (Howitt, 1998) and Blind Chance (Przypadek, Kieslowski, 1981) are discussed at some length in essays by David Bordwell, Edward Branigan and Kay Young in SubStance. (17) In his essay Bordwell claims that narratives are built not upon 'philosophy or physics but folk psychology, the ordinary processes we use to make sense of the world'. He makes a convincing case that cognitivism can explain our enjoyment of films such as these. As Branigan points out in his essay:

Bordwell demonstrates that what he calls "forking-path " plots in such films as Sliding Doors and Run Lola Run have certain fundamental properties that are quite familiar to us from classical narratives. For example, forking-path plots are well-marked, linear, developed, cohesive, unified with one another, ordered sequentially to make the final path a climax, and designed to pinpoint clear, contrasting parallels. (18)

Given the detailed and convincing arguments in the three essays, I do not intend to do more than highlight just one issue here. Bordwell describes in some detail seven 'conventions' used in such films; conventions that mean that the infinity of potential futures 'have been trimmed back to cognitively manageable dimensions' by means of strategies characteristic of certain traditions of cinematic storytelling. He gives as his seventh convention that:

All paths are not equal; the last one taken, or completed, is the least hypothetical one. (19)

This is surely a key point. The final path in Run Lola Run is the only one with a happy ending; the first ends with Lola being shot by a policeman and the second with Manni being run over by an ambulance. Ending the film with either the first or second story would do more than just turn the film into a tragedy, it would destroy the sense of closure and satisfaction we feel at the end. As Bordwell says:

Instead of calling these "forking-path "plots, we might better describe them as multiple-draft narratives, with the last version presenting itself as the fullest, most satisfying revision. Once more, this conforms to our propensity to weight the ending, to treat it as the culmination of what went before it...even if all of what went before couldn’t really have come before. (20)

There is of course another side to Run Lola Run. The techno-music soundtrack and the film's use of animation sequences give it the feel of a video game - with Lola as a variant of Lara Croft. Each episode starts as a kind of replay, as though Lola were pushing the restart button in the hope of getting a better score. Some of the film's popular success surely lies in this exploitation of the conventions of the video game genre. (21) I do not intend to expand on this here, but I will return to the subject of video games towards the end of the essay.

As I said earlier, The Usual Suspects andThe Sixth Sense utilise last-minute plot twists. For these twists to come as a surprise the films must lead us down a wrong path, so that we expect a different outcome right up until the last minute. Branigan outlines the cognitive explanation for this:

Filmmakers employ the psychology of the everyday in order to aid spectators in comprehending a narrative. Filmmakers also employ this psychology against spectators when it is important that something not be seen or fully understood during the telling of a story (e.g., to create mystery or surprise), or when the spectator must understand in a new way (e.g., in a metaphorical way or through a sudden revelation), or when something disturbing or traumatic must be reconfigured by the text or repressed. As spectators, we make mistakes in making inferences because we are systematic in drawing inferences and authors count on that. (22)

In both films we spend almost two hours carefully constructing a fabula from the evidence placed before us, only to have it suddenly torn down and replaced by something very different. Surely this should cause us to feel cheated, lied to or tricked? (23) The cognitivist must not only explain how we are misled, but also why both films have been wildly successful with audiences. How do such films avoid engendering negative reactions at the point where all is revealed?

In one sense, The Sixth Sense plays fairer than The Usual Suspects, which may make it simpler to explain. Shyamalan is very careful to present us with the truth, and nothing but the truth, even if it is not the whole truth. Throughout the film we never see or hear anything which conflicts with the idea that Malcolm (Bruce Willis) is a ghost, so we cannot claim later that the film lied to us. We are misdirected because the events we are shown have another possible interpretation - that Malcolm is still alive. By restricting the set of events we see, Shyamalan lets us choose the explanation which is (initially at least) not only more probable but also less tragic. Because we did the choosing and Shyamalan has played fair, when we do realise the truth we do not say 'Shyamalan cheated me, he lied, it's a trick'. We look back for evidence that he has done these things and realise he has not. We chose the version we did, not because he made us do it but because it was the one we wanted to be true.

Nor can we claim that he withheld clues. In fact along the way we are given many clues which could lead us to the other version. Thus at one point the boy, Cole, says to Malcolm, 'they [ghosts] only see what they want to see; they don't even know that they're dead'. Why do we not see this as applying to Malcolm himself? Why don't any of the clues lead us to modify the fabula we are constructing until we are forced to accept the truth, as the ring rolls across the floor? There must be an explanation, a cognitivist explanation, for this failure too. (24)

To me, it seems clear that the reason we ignore such clues is that we have become like Malcolm. We 'see what we want to see', we 'don't even know that [he's] dead'. What Shyamalan does so well is to present us with what Malcolm sees, and no more, aligning us with him. (25) Writing in Passionate Views, Gregory Currie denies that we 'imagine ourselves directly related to the characters and events of the story'. We are, he says, 'cognitively focused ... but in being interested in and concerned about those characters and event, the reader does not imagine being located at the places and times where the characters and events are located'. (26) In an earlier book, he lays out a detailed theory of imagining, based upon the idea that people use their 'own mind to simulate the mind of another'. (27)

As Berys Gaut remarks, 'while psychoanalytically inspired theories have responded positively to these claims [of identification] but treated them in a hyperbolic fashion ...those who draw on analytical philosophy and cognitive science generally have little time for such psychoanalytical constructs of spectator responses'. (28) Gaut goes on to attempt 'to rehabilitate the notion of identification for cognitive theories of film, to show that the notion does not suffer from the deep conceptual confusion alleged against it and to demonstrate that it has explanatory power in accounting for spectators' emotional responses to films'. He points out that there are various 'aspects of the character's situation one imagines oneself in' and that while we might ignore the character's physical properties, we might share some psychological ones - affective say, or motivational or epistemic for example. In this he seems to me at least partly successful. Although we do not imagine, while watching The Sixth Sense, that we are Malcolm, we do suffer from the same perceptual, conceptual confusion. This not only goes much further than Smith's 'alignment' but also beyond Currie's 'simulation', because the confusion is ours. (29)

The circumstances surrounding our eventual realisation of the 'truth' help explain our enjoyment of the process. While the fact that Malcolm is dead might seem tragic, the apparent painful alienation between him and his wife turns out to be false. Quite the reverse - even after his death she finds it as difficult to let him go as he has the idea that he is still alive. Coupled with the manner in which Cole becomes reconciled with his 'gift' and with his mother, the ending seems both inevitable and satisfying.

We undergo a similar revelatory process in The Usual Suspects. At the heart of film's last five minutes is a short 'realisation' sequence, lasting less than eighty seconds, which is designed to tell us the secret behind the film we've been watching; that the story Roger 'Verbal' Kint has been telling Agent Dave Kujan is in fact untrue. The flashback scenes, such as the car park robbery that goes wrong, may have seemed as real and truthful as the interrogation scenes that happen in the film's present, but they were not - they could never happened as we saw them. (30)

Singer gives us few clues (31) to this deception until the end of the film, so how does he avoid the risk of disappointing the audience; of making them feel they've been tricked and that the film hasn't played fair? Stage Fright (Hitchcock, 1950) is celebrated for its false flashback, a flashback heavily criticised at the time for lying to the audience. Here Singer would seem to lay himself open to much graver charges, but he cleverly uses the realisation sequence not only to reveal the deception, but also to provide an explanation that absolves him from blame.

Currie uses Stage Fright as an example when distinguishing between 'controlling' and 'embedded' narrators. (32) He maintains that film, unlike literature, cannot have 'controlling' narrators, since the images we see are not in fact produced by the embedded narrator. (33) I would claim that The Usual Suspects is an exception. For what exactly is it that we have been watching during the narrated flashbacks, if not the truth? Singer's answer is that the diegetic world we have been watching in those episodes is not some past objective reality, but the present subjective reality inside Kujan’s head, as he visualises the story that Kint is telling. It is not Singer's camera but Kujan's mind, cleverly manipulated by Kint, which has been lying to him, and to us. Kint is controlling the narrative.

Singer shoots the sequence in such a way as to emphasise that what we are watching are Kujan's thoughts. Skilful camera movement and editing make us see what is going on in Kujan’s mind as he realises how he has been fooled. Our involvement in his realisation means we end up feeling what Kujan feels - a sudden rush of insight. The effect is to make us think, as Kujan does, that it is Kint who has tricked us, not Singer. We end up saying 'Yes, how clever, Kint made it all up to fool me (Kujan)' rather than 'No, how annoying, Singer made it all up to fool me (the spectator in the audience)'.

Though the sequence lasts less than eighty seconds, it contains almost sixty shots. A detailed analysis lies outside the scope of this essay, but one recurring feature is the use of short close-ups zooming in on Kujan's face. These slow zoom shots, centred on Kujan's eyes, seem designed to draw us into his brain. (34) The sequence may last over a minute but Kujan's realisation takes just a few seconds, so Singer uses two specific cinematic devices to alter our perception of time. At the start we see slow-motion shots of the cup falling and shattering. This dilation of time mirrors Kujan's heightened senses, (35) the way his brain is racing so fast that he sees the world in slow-motion too. At the end, as Kujan rushes out of the room, Singer reverses the time-dilation effect by following Kujan with a whip pan so violent the room is reduced to a blur. The psychological effects of these devices seem to me to go far beyond what Currie's theory would allow. (36) Again it seems clear to me that more than 'alignment' or 'simulation' is at work here.

One interesting feature of both films is that they bear repeated viewings. Of course, we may be watching the film again to look for further clues or things we missed, (37) but as Richard Gerrig puts it:

one great puzzle of the experience of fiction is why some emotional reactions predicated on the absence of knowledge endure even when knowledge is present. (38)

Gerrig goes on to provide a possible cognitivist explanation. He sees the anomalous suspense we feel as a cognitive illusion, generated by our minds because the underlying processes assume that each experience is new. I find his argument somewhat unconvincing, not least because, as he points out, there are 'works which do not tolerate re-experience' (or at any rate, fail to elicit the same emotional response).

Finally let us turn to Memento. The principal character, Leonard Shelby, has suffered a brain injury in a brutal attack during which his wife was raped and murdered. He has anterograde amnesia; unable to form new long-term memories, his memory of the present fades after a few minutes. (39) Nolan helps us to understand Shelby's condition by putting us in a similarly disadvantaged position - he presents the narrative in reverse, as a series of flashbacks going back further and further into the past. In a TV interview with Elvis Mitchell shown on the Independent Film Channel he explains his intention:

The story is told in as subjective way as we were able to do. ... We really tried to put the audience in the head of the protagonist and make them experience some of his confusion, uncertainty and paranoia. (40)

In another interview he said:

[We told] the story backward so that it denied the audience the information the protagonist is denied (41)

Watching the story unfold in reverse - seeing results before causes - certainly makes for an intense cognitive experience. The success of the film demonstrates the existence of a sophisticated audience keen to engage with such narrative challenges. But once again the cognitivist must face the question - why do we enjoy the experience?

In The Narrative Construction of Reality the psychologist Jerome Bruner asserts that:

narrative organizes not just memory, but the whole of human experience - not just the life stories of the past, but all of one’s life as it unfolds... narrative is an instrument of mind that constructs our notion of reality ... the experience of life takes on meaning when we interact with it as an ongoing story, as our story. (42)

I would claim that in order to enjoy the experience, we must take part in it; we must 'interact with it as an ongoing story, as our story'. In their paper discussing four types of 'dysnarrativia', states of narrative impairment experienced by individuals with discrete focal damage in different regions of the brain, Young and Saver make a key point:

Studies of the forms of neurologic disintegration of human narrative capacity carry an additional important implication for interpreting autobiographic literature. These conditions reveal that narrative framing and recall of experience is a dynamic, variable and vulnerable process. .... Modern neuroscience has demonstrated that retrieving memories is not the simple act of accessing a storehouse of ready-made photos in a stable neural album, preserved with complete fidelity to the moment of their formation. Rather, each act of recall is a re-creation, drawing upon multiple, dynamically changing modular fragments to shape a new mosaic. Numerous consequences follow for literary interpretation, of which we will mention just one. All memories are suspect, at the neural level. Fidelity-stable recall and self-interpretation of the past is not a property of the human brain and mind. The varied subjectivity of literary autobiographic productions has its root in the inescapable subjectivity of the brain’s narrative and memory system. (43)

This 'inescapable subjectivity', rather than simple 'alignment', is surely at the heart of what happens to us while watching Memento, or the revelatory scenes in The Usual Suspects orThe Sixth Sense. In an almost literal sense, we 'incorporate' the narrative we are watching. (44) Because the problems we have to solve in understanding the film are so closely related to Leonard's, when the film eventually reveals his dark secret we shudder because we understand him not only on a rational level, but in a deeper, emotionally empathic way too.

More evidence for this view of our involvement is to be found by looking at Betrayal (Jones, 1983), based on a play by Harold Pinter. The film tells its story as a succession of flashbacks going further and further back in time. At each step narrative information presented earlier (though later in story time) is revealed as duplicitous. Faced with narrative contradictions, viewers are prompted to reassess their earlier impressions in the light of later but chronologically earlier scenes.

So much is like Memento. Though the film's screenplay was nominated for an Oscar, audiences found the film difficult to understand or enjoy. (45) The explanation must surely lie in the fact that in this case the device does not help us to understand or get involved with one particular character. Instead it creates an oscillation between different possible narratives, with our sympathies being switched between characters as we learn more about them. (46)

Another example that helps us understand the nature of involvement with a character is the reception accorded The Lady in the Lake (Montgomery, 1946). The film uses a subjective camera throughout - the only time we see the hero, Philip Marlowe, is when he looks at himself in a mirror. On the surface this might lead us to expect that a spectator would experience a strong sense of identification. This does not seem to be the case. Audiences at the time found the experience 'cramped' and 'claustrophobic' and the film did not do well either at the box office or with the critics. Discussing the film Torben Grodal says:

The whole film, except the narrator sequences, is shot using 'subjective camera'; the effect, however, is not an intense 'subjective' identification with the protagonist but, on the contrary, a feeling of alienation, because there is no objective model, a body, on whom to anchor feelings of identification (and there are not - as in real life - any body-sensations to anchor the objective model of the self). The 'subjective' camera view cannot therefore be experienced with complete cognitive and emphatic identification by the viewer: it is experienced as the view of an alien. (47)

There are other disadvantages. As Clarke and Mitchell put it:

... as you can only see what is in front of you, it is difficult to generate and build suspense. Films often generate suspense by showing the viewer something that the protagonist cannot, and emphasise objects and actions significant to the story by showing them in close-up. Both of these techniques are impossible when the only view of the action is through the eyes of the character. (48)

It is interesting to compare these adverse effects with the popularity of the subjective viewpoint in computer games such as Doom, Quake or MYST. As Jesper Juul remarks:

Nevertheless there seems to be agreement that the view-from-within in 3d games increases identification. This is possibly because a view panning autonomously is an unknown (and uncomfortable) experience, but to look out from a set of eyes and be able to control the direction of your gaze is a well-known cognitive-physical experience. It simply corresponds to a basic experience of the world. (49)

In recent years the burgeoning market for 'narrative' computer games has resulted in considerable research into how to make such games more pleasurable from a narrative standpoint. (50)

Our survey of the impact of these five films on cognitive theories of narration has, I believe, shown that while cognitivism is relatively robust, there are areas where current cognitive explanations are weak or lacking. Films such as The Usual Suspects, The Sixth Sense or Memento seem to rely on a deeper form of involvement than most cognitivists are currently prepared to accept. It may not be necessary to do as Berys Gaut suggests, and 'rehabilitate the notion of identification', but there is certainly work to be done. There also does not seem to be an adequate explanation of 're-experience', either in terms of the persistence of emotional responses when the truth is already known, or of the rules governing which films bear re-experiencing and which do not.

I hope that this essay has also shown that the notion of performing 'experiments', of probing the strength of a theory by the use of specific, extreme examples, is a productive one not just in science, but for film theory too.

Web References:

The following websites are particularly relevant to this essay:

David Bordwell's website:
Film and Philosophy (Society for the Philosophic Study of the Contemporary Visual Arts):
Film Philosophy Portal:
9th Laterna Film Academy (Pécs, Hungary, September 2002):
Journal of Moving Image Studies:

Films mentioned in the text:

Betrayal, directed by David Jones, written by Harold Pinter (UK, 1983)
Blind Chance (Przypadek), directed and written by Krzyszstof Kieslowski (Poland, 1981)
The Lady in the Lake, directed by Robert Montgomery, written by Raymond Chandler (US, 1946).
Memento, directed by Christopher Nolan, written by Jonathan Nolan (US, 2000)
Pulp Fiction, directed by Quentin Tarantino, written by Quentin Tarantino and Roger Avary (US, 1994)
Run Lola Run (Lola Rennt), directed and written by Tom Tykwer (Germany, 1998)
The Sixth Sense, directed and written by M. Night Shyamalan, (US, 1999)
Sliding Doors, directed and written by Peter Howitt (UK, 1998)
Stage Fright, directed by Alfred Hitchcock, written by Whitfield Cook, (UK, 1950)
The Usual Suspects, directed by Bryan Singer, written by Christopher McQuarrie (US, 1995)



  1. Maureen Turim, Flashback in Film: Memory and History (Routledge, 1989), 246. (back)
  2. As illustrations of their success, Pulp Fiction won the Oscar for best screenplay in 1994, The Usual Suspects won in 1995, The Sixth Sense was nominated in 2000 and Memento in 2002. Run Lola Run won best film and best director in the German Film Awards and was Germany's entry for best foreign film Oscar in 1999. (back)
  3. For an interesting discussion of the increasing use of avant-garde techniques in mainstream films, see Curt Hersey, 'Diegetic Breaks and the Avant-Garde', Journal of Moving Image Studies, Volume 1, No. 2, Spring 2002. (back)
  4. Cynthia Freeland , 'Cognitive Science and Film Theory', American Society for Aesthetics, Panel on Cognitive Science and the Arts, Santa Fe, October 31, 1997. (back)
  5. David Bordwell, 'A Case for Cognitivism', Iris No. 9 (Spring 1989), 11-40. (back)
  6. David Bordwell and NoÎl Carrol, Post-Theory: Reconstructing Film Studies, (Madison: University of Wisconsin Press, 1996), 29. (back)
  7. David Bordwell, Narration in the Fiction Film, (Madison: University of Wisconsin Press, 1985), 30. (back)
  8. Bordwell and Carrol, op. cit., xvi. (back)
  9. Kevin Sweeney, 'Constructivism in Cognitive Film Theory', Film and Philosophy Vol. 2, (Society for the Philosophic Study of the Contemporary Visual Arts, 1994). (back)
  10. Christian Metz, Film Language: A semiotics of the Cinema, (New York: Oxford University Press, 1974), 24. (back)
  11. Allan Langdale,Hugo M¸nsterberg on Film: The Photoplay ó A Psychological Study and Other Writings (New York: Routledge, 2002), 135. (back)
  12. It seems likely that at least a day separates the two events, since Mia appears to have recovered completely and both episodes take place at night. None of the three stories show us anything of what occurred between the two episodes. (back)
  13. See for example the Charlie Rose TV interview with Tarantino that appears on the Pulp Fiction DVD. (back)
  14. The DVD for Pulp Fiction contains a feature that allows a subtitled 'Trivia Track' to be switched on. This track reveals just how densely the film is packed with references. As an example, in the scene labelled 'Butch drives off and hits Marsellus', three things happen in quick succession:
    1. the song on the car radio mentions 'Captain Kangaroo', a reference to the kangaroo ornament the watch hung from
    2. the song lyric reaches the words 'It's good to see you' just as we watch Butch catch sight of Marsellus
    3. Marsellus walks in front of Butch's car in a direct quote from the scene in Hitchcock's Psycho where Marion Crane's boss walks in front of her car as she is trying to escape (back)
  15. Susan Feagin, 'Time and Timing', in Carl Plantinga and Greg Smith (eds), Passionate Views: Film, Cognition and Emotion, (Baltimore: Johns Hopkins University Press, 1999), 173. (back)
  16. Carl Plantinga, 'Movie Pleasures and the Spectator's Experience: Toward a Cognitive Approach', Film and Philosophy Vol. 2, (Society for the Philosophic Study of the Contemporary Visual Arts, 1994). It is worth pointing out that Pulp Fiction provides an excellent illustration of all five sources of spectator pleasure. (back)
  17. See David Bordwell, 'Film Futures', Edward Branigan, 'Nearly True: Forking Plots, Forking Interpretations - A Response to David Bordwellís Film Futures', and Kay Young, 'That Fabric of Times - A Response to David Bordwellís Film Futures', all in SubStance 97, Vol. 31, No. 1, 2002, 88-118. (back)
  18. Edward Branigan, op. cit., 105. (back)
  19. David Bordwell, op. cit., 100. (back)
  20. David Bordwell, op. cit., 102. (back)
  21. For a detailed examination of the film as a video game, see the thesis The Cinematic Video Game - Lola Rennt, by Benjamin West (Washington University, St. Louis, Missouri, 2000), available at (back)
  22. Edward Branigan, op. cit., 106. (back)
  23. Some of the initial reviews of The Usual Suspects did indeed take this tone. See Ernest Larsen, The Usual Suspects, (BFI Publishing, 2002), 54-55. (back)
  24. Daniel Barratt's paper on 'Twist-Blindness' (see bibliography) would appear to provide possible answers. Unfortunately I have only been able to locate an abstract, rather than the full text. (back)
  25. I use 'alignment' here as defined by Murray Smith - 'the way a film gives us access to the actions, thoughts, and feelings of characters'. See Murray Smith, Engaging Characters: Fiction, Emotion, and the Cinema, (Oxford: Clarendon Press, 1995). (back)
  26. Gregory Currie, 'Narrative Desire, in Carl Plantinga and Greg Smith (eds), Passionate Views: Film, Cognition and Emotion, (Baltimore: Johns Hopkins University Press, 1999), 193. (back)
  27. Gregory Currie, Image and Mind: Film, Philosophy, and Cognitive Science, (Cambridge University Press, 1995). See particularly section 5.2, 144-152. (back)
  28. Berys Gaut, 'Identification and Emotion in Narrative Film', in Plantinga and Smith, op. cit., 200-201. (back)
  29. Edward Branigan presents an interesting alternative reading of The Sixth Sense - as a 'forking path' narrative - in 'Nearly True: Forking Plots, Forking Interpretations - A Response to David Bordwellís Film Futures', op. cit., 110-111. (back)
  30. The soundtrack must be lying, since Kobayashi cannot be the name of the Pete Postlethwaite character, but almost all of the sights we see in flashback could have happened. Perhaps all Kint had to do was invent the dialogue! (back)
  31. There are some clues throughout the film. It is Kint, not Keaton, who fires the first fatal shot in the car park robbery. For a seemingly timid man whose left hand appears useless and whose right could not work a cigarette lighter earlier in the film, the speed, precision and resolve he shows is surprising. We are also told more than once that he is a confidence trickster with a talent for telling stories (and his nickname after all is 'Verbal'). (back)
  32. See Currie, Image and Mind, op. cit., 265-270. He takes issue with Seymour Chatman's approach in Story and Discourse: Narrative Structure in Fiction and Film, (Ithaca: Cornell University Press, 1978). (back)
  33. He cites Peeping Tom (Powell, 1960) as a rare example of a film which shows us images actually produced by one of the characters. (back)
  34. Truffaut, quoted in Peter Graham, The New Wave (Viking 1968), says 'The cinema becomes subjective when the actor's gaze meets that of the audience'. (back)
  35. This is a curious, almost counter-intuitive, psychological effect. In TV series such as The Six Million Dollar Man and Kung Fu slow-motion was used for a similar purpose, to emphasise how fast things are happening (back)
  36. In his discussion of time and space in films, Currie does not mention zooms at all, and dismisses the effects of under- and over-cranking in a sentence or two. See Currie, op. cit., 104-106. For an extended critique of Currie's book, see Erin Krause, 'Pay No Attention to the Man Behind the Curtain: Film, Cognition, and Experience', The Dualist, Vol. 4, 1997, (California: Stanford University). (back)
  37. Ira Nayman discusses the enjoyment of this kind of 're-experiencing' in his article 'The Man Who Wasn't There ñ Narrative Ambiguity in 3 Recent Hollywood Films', Creative Screenwriting, March/April 2001, 57-60. His examples are The Usual Suspects, The Sixth Sense and Fight Club. (back)
  38. Richard J. Gerrig, 'Re-experiencing Fiction and Non-Fiction', Journal of Aesthetics and Art Criticism 47, No.3 (Summer 1989), 277. (back)
  39. For a highly relevant account of the effects of various forms of brain injury on our ability to create and understand narrative, see Kay Young & Jeffrey Saver, 'The Neurology of Narrative', SubStance 94/5, Vol. 30, No. 1, 2001, 72-84. (back)
  40. This interview appears on the US version of the Memento DVD. (back)
  41. Renfreu Neff, 'Interview with Christopher Nolan', Creative Screenwriting, March/April 2001, 48. (back)
  42. Critical Inquiry 18.1 (1991), 1-22, quoted in Young and Saver, op. cit., 75. (back)
  43. Young and Saver, ibid., 79. (back)
  44. So-called 'false memory syndrome', where the patient is unable to distinguish memories of real events from those implanted through mechanisms such as hypnotic suggestion, would seem to operate along similar lines. (back)
  45. The original stage play was received more enthusiastically. This opens up an interesting set of questions (beyond the scope of this essay) about the differences, in a cognitive sense, between watching a play and watching a film. (back)
  46. See Kevin Sweeney, 'Constructivism in Cognitive Film Theory', Film and Philosophy Vol. 2, 1994 for a detailed discussion of Betrayal from a cognitive viewpoint. (back)
  47. Torben Kragh Grodal, Moving pictures - a new theory of film genres, feelings, and cognition, (Oxford: Clarendon Press, 1997), 115. (back)
  48. Andy Clarke and Grethe Mitchell, 'Playing with Film Language', Visual Narrative Matrix Conference, Southampton, November 1999. (back)
  49. Jesper Juul, A Clash between Game and Narrative - a thesis on computer games and interactive fiction, (MA Thesis, University of Copenhagen, 1999), available at The emphases are mine. (back)
  50. See for example, Kevin Brooks, 'Programming Narrative', IEEE Symposium on Visual Languages, Los Alamitos, California, 1997, R. Michael Young, 'Creating Interactive Narrative Structures: The Potential for AI Approaches', AAAI Spring Symposium on Artificial Intelligence and Interactive Entertainment, (California: Stanford, March 2000), Tom Kemper, 'Instant Re-Players - From Sports Fans to Video Game Players: A Cognitive History', MIT Communications Forum. and J. Yellowlees Douglas and Andrew Hargadon, 'The pleasures of immersion and engagement: schemas, scripts and the fifth business', Digital Creativity Vol. 12, No. 3, 2001. (back)

Back to Film Studies Page