Return to SCE Homepage
Return to Session Program

Autism and Representation
Case Western Reserve Univ
Cleveland, Ohio
28-30 October 2005



Kristen Loutensock
Berkeley, California

Do not cite without permission of author

"Filming the Self: Self-Narration and Film for Children with Autism"

Central to many theories of reception is the concept of ‘theory of mind' This theory, which broadly refers to our ability to understand the intentationality of other subjects, is understood and applied differently by disciplines ranging from cognitive science to linguistics. Striking among these, from a film studies perspective, is John Barker and Robert Gordon's hypothesis that an individual's theory of mind stems from their use of cognitive simulations to model how the other would react .1 This hypothesis, which posits that theory of mind stems not from a complex understanding of the other but from our ability to imagine how we ourselves would react to given situations, speaks in a number of ways to film spectator theories ranging from suture to coherence theory. However, Barker and Gordon's work, like that of many in the field of cognitive science, includes autism only as an 'outside' group, the 'non-normative' mind that helps define the norm. This use of autism as outlier seems to stem directly from the reliance of most cognitive theorists on the definitions of autism proposed by Uta Frith and Simon Baron-Cohen, both of whom use theory of mind in their definitions .2 This use implicitly denies the position of the spectator with autism. This exclusion is especially problematic given the number of individuals with autism who are avid film spectators, and who use the dialogue of feature films to 'speak the self.' It is this use of films in the construction of identity, particularly for subjects who are understood as non-normative, that has the deepest implications for expanding current theories of active spectatorship, for it redefines the way in which the non-normative subject is understood to be interacting with the film text. Further, if this theory is taken from the realm of spectatorship to authorship, it would follow that artists with autism make art not to engage with others in a grand form of ‘other-regarding pretending', but rather as part of a closed-loop sense of self. In other words, art made by those with autism, if those with autism are supposed to lack theory of mind, would be directed only at the self, with no place for an outside spectator. Yet recent works by filmmakers with autism do quite the opposite, building into their narrative and form the idea of a spectator to whom the self is being narrated. Film-making, then, can be regarded as a powerful tool for self-narration for children with autism not only because of the way it builds upon their use of film in the creation of an identity, but also because of the way the medium invites 'other-regarding pretending', problematizing normative notions of theory of mind

To begin with, it is necessary to trace the development of the term theory of mind and how it was incorporated into the diagnosis of autism. Although the roots of theory of mind lie in Jean Piaget's work on early childhood development in the 1920s ,3 it was not until the 1970s that his work was rediscovered by those working in the field of meta-cognition (or, thinking about thinking). However, it was not until the late 1970s that theory of mind coalesced as a field. The switch followed the publication of “Does the chimpanzee have a theory of mind” by David Woodruff and Guy Premack, who had performed a series of experiments with a female chimpanzee with a limited sign language vocabulary and other training. These experiments, which were meant to show whether or not the chimpanzee could make inferences, were performed using videotapes of human actors in various situations, such as attempting to get out of a locked cage. Then, the chimpanzee was shown four photographs of different solutions; in this example, a hose, an electrical cord plugged into a socket, a match, and a key. Since the chimpanzee pointed to the card with the correct solution, Woodruff and Premack reasoned, it had made inferences about the intentions of the human, and, therefore, possessed theory of mind .4

Following the publication of Woodruff and Premack's work, the philosophers Jonathan Bennett, Daniel Dennett, and Gilbert Harman independently proposed a modification of the experiment. Each of the three suggested that a more sophisticated test for theory of mind would test a subject on whether they could recognize that people sometimes made errors acting on the basis of false beliefs. 5 These ideas were used by Heinz Wimmer and Josef Perner in a classic experiment in which children between the ages of three and five were told a story as it was acted out by puppets. In the story, there were two puppets in a room, one of whom has a chocolate. There are two containers in the room, here represented by a box and a basket. Puppet A puts the chocolate in the basket, then leaves the stage While she is gone, puppet B moves the chocolate from the basket to a box, then leaves the stage. Puppet A then returns, and the child watching is asked where the puppet would look for her chocolate. The researchers found (and this experiment has often been replicated, with multiple variations, with a high rate of success) that children younger than four tended to answer that the puppet would look in the box, while children older than four tended to answer that she would look in the basket. 6

In the psychological literature, this test is said to establish that the children older than the age of four were able to recognize that the subjects in the story had limited consciousness and, therefore, could not see through walls or otherwise grasp events that they were not physically present for, regardless of whether the children themselves saw the events. What is not discussed is the fact that this experiment could also be read as an exercise in suture, particularly as it uses enacted events within a diegetic world. The children, therefore, can be seen as being tested on their ability to separate their consciousness from the diegetic world. The 'normative' response to a fictional world, then, is one of fragmentation, of following only the point of view that is dictated. This scenario does not allow for the spectator to project themselves into the fictional world in a new way, but forces them to choose one position, a position dictated by the text. Interestingly, it is this same 'over-identification' with the diegetic world that is present in the world of movie fan-fiction, where writers use parts of a diegetic world to create their own stories. For instance, a writer might choose to set a story in the narrative of Harry Potter , but to also insert their own characters and change or add on to existing plot points. This type of writing, which has exploded in popularity as the internet has become more accessible, has attracted many writers on the autism spectrum. However, it is also denigrated by many as disrespectful to the original authors, or as trash. What is ignored in this simple condescension is the way in which fan fiction can be seen, particularly for writers on the autism spectrum, as a way to create and modify a self.

To return to the puppet test, a variant of it was used by both Simon Baron-Cohen and Uta Frith in their much-cited works on what they term the psychological theory of autism Baron-Cohen, Helen Tager-Flusberg and Donald Cohen, in the introduction to Understanding Other Minds , note the first problems encountered by most researchers using subjects with autism to test theory of mind is that the diagnostic criteria of autism are focused on clusters of behavioral traits, and that there is a wide variance in measurable IQ and behavior even among those diagnosed as having the disorder. 7 However, at the same time that they mention these variants, the three also reassure the reader that “the diagnosis of autism is remarkable for the general agreement among clinicians, over decades and across nations.”8 This statement, which manages to confirm that a normal person with autism exists, is remarkable because it is not accompanied by any sort of description or further allusion to this normative individual. Frith, writing separately in the paper “Mindblindness and the Brain in Autism” in 2001 (8 years after Baron-Cohen, and nearly 20 years after their work together), gives a much more concrete definition of autism, one that focuses on neuropsychological and genetic causes while still ultimately resting on the same behavioral diagnosis as Baron-Cohen.9 This behavioral diagnosis in both cases draws from both the World Health Organization's International Classification of Disease and from the American Psychiatric Association's Diagnostic and Statistical Manual.10 The reliance on these two sets of behavioral criteria stems from their use in determining funding; without an ‘official' diagnosis, one that is ratified by one of these two bodies, an individual is not eligible for government-funded treatment or for inclusion in Special Education classes. Since these diagnostic manuals now include language directly tied to experiments done by Frith and Baron-Cohen in the early 1980s on theory of mind, language such as gross abnormalities in peer relations and marked lack of awareness of others, it is clear that there are economic as well as social issues implicit in this research.

This link between theory of mind research and the political and social construction of autism is one that becomes highly problematic when Frith and Baron-Cohen's methodology and assumptions are examined. The two collaborated in the early 1980s on a series of false-belief tests that, as the title of their article puts it, asked the question “does the autistic child have a theory of mind?'11 In this formulation, which is identical to that of Wimmer and Perner except for the substitution of ‘autistic child' for ‘chimpanzee', one can see the first way in which this testing set up children with autism as the non-human other mind in the same way that the chimpanzee was used This formulation, which seems shocking in its dehumanization of the individual with autism, still has currency for both psychologists and those with autism. There is both a theoretical work on theory of mind, entitled Introduction to Theory of Mind: Children, Autism, and Apes , as well as a primatologists account of her experince with autism entitled Songs of the Gorilla Nation: My Journey Through Autism. This dehumanization, while not the intent of the original studies, resulted in part because of the researchers insistence on a binary opposition between the normative (neurotypical) and non-normative (non-neurotypical or autistic) mind. This binary was established by Baron-Cohen and Frith in their original study, which used a control group of neuro-typical, non-developmentally disabled (as they put it, normal) children, with a verbal mental age of 4 (which corresponded to a chronological age of four). These children, along with a group of children with Down syndrome (who had a chronological age of 10 and a verbal mental age of four) and a group of children with autism (who had a chronological age of 10 and a verbal mental age of 9), were presented with the puppet test from above, using either puppets or adults to enact the test. On average, Frith and Baron-Cohen found that the majority of the ‘normal' children and children with Down syndrome passed the test in two trials, while less than 1/5 of the children with autism did.12 Interestingly, while these results are widely cited in philosophical and psychological writing on theory of mind, and usually held as proof that it is the ability to ‘mindread' that separates the normal child from the child with autism, both Frith and Baron-Cohen complicate this binary in their subsequent revisiting of the test, complications which suggest even more strongly the non-binary nature of narrative comprehension.

Frith follows her account of the original puppet test with an addendum concerning false photographs. Using the same methodology for selecting subjects, this study by Alan Leslie and Laila Thaiss, modified the puppet test by having a small area set up with a bed, chair, and toy cat. At first, the cat is placed on the chair, and the child takes a Polaroid picture of it. Next, the researcher removes the photograph from the child's sight, then moves the toy cat to the bed. Finally, the child is asked where the toy cat is in the photograph. In this test, it was found that the children with autism were markedly better at correctly answering questions about the position of the toy in the photograph than the other two groups .13 Frith takes this as further proof that children with autism are not able to mind read, as she claims that this demonstrates that “understanding of false photographs, but not understanding of false beliefs, was well within the comprehension of children with autism.”14 However, this variance, when considered together with Baron-Cohen's finding that phrasing the puppet question as “where does the puppet first think its marble is?” rather than “where does the puppet think its marble is?” resulted in a significant increase in the number of children with autism answering correctly, suggests that something larger is at work.15 This word of clarification would seem to suggest that the children with autism might in fact be identifying with both puppets, or even creating a larger narrative where the puppet goes to the second container in the room when the first is empty Similarly, the Polaroid test could be interpreted as a stronger grasp on cause and effect by the children with autism.

In fact, this pattern could be read as evidence that mind reading is not what is occurring; instead, it could be argued, as Robert Gordon and John Barker do, that what is occurring is closer to a type of simulation. In their argument, Barker and Gordon summarize that Baron-Cohen and Frith see the puppet test as evidence of the way in which normal children read the mind of another, ie. by making inferences as to the actions of another mind using their developmentally acquired knowledge that people act on the basis of mental models. This formulation, which is nearly identical to the concept of folk psychology, relies on the idea of the concept of a mental model or representation of the world-the child must have a theory to draw inferences from. What Simulation Theory does is remove this layer of the model; instead of taking the role of the puppet and then using their concept of belief to determine what the puppet would do, the child in the Simulation Theory would take the role of the puppet and then project their own beliefs onto it. Instead of thinking about what the puppet's model of the world would be, which requires a complex set of reasoning through generalizations and representations ranging from the generalization that someone outside cannot see what is happening inside to generalizations about searching behavior, the child using Simulation Theory would simply put themselves in the place of the puppet, mentally subtracting the information that would be blocked by the wall between the puppet and the room, then use the new facts to answer the question as the puppet.16

However, this model does require that the child is able to juggle multiple contexts and separate facts-for-the-puppet from facts-for-the-child. Simulation Theory, then, requires not only that the child project their own knowledge and feelings onto others but, to truly simulate, then recognize their own point of view in the world Barker and Gordon, using Baron-Cohen's notion that children with autism don't follow the lines of sight of others in the same way that normal children do, extrapolates that this lack is what prevents children with autism from realizing that they share a world with others .17 It is at this point that Barker and Gordon's lack of experience with children with autism, and reliance on the clinical descriptions given by Baron-Cohen and Frith, seems most troubling. What is striking, though, is the way in which Barker and Gordon's language here mimics that of many theories of film spectatorship, particularly those dealing with resistant spectatorship. What they seems to be saying is that the attention focusing devices of the world (or film) are not responded to by the individual with autism, causing them to focus on elements that do not create a sense of coherence or suture with a normative conception of time and space. Instead, it seems that this model along with the exceptions to the puppet test show that the spectator with autism, in the world, in these tests, and, possibly, as a film spectator focuses on other objects and other points of view, then constructs a model that seems to both incorporate and expand on the diegetic world as it is presented.

Intriguingly, this idea of the person with autism not lacking a theory of mind, but instead having a theory of mind that is strikingly different than that of the so-called neurotypicals, is one that was discussed at length in an online mailing list for individuals with autism. One of the participants in the list brought up Frith's work, and stated that this work seemed to suggest that people with autism lacked “logical reason and empathy”, a claim the participant disputed As others on the list added their thoughts, the most common criticism of the work done on theory of mind were that the tests themselves seemed to be testing verbal ability, attention, and information processing rather than a ‘theory of mind', and, more strikingly, that their own experiences as children with autism were not that they were unaware of the consciousness of others, but, rather, that they believed certain individuals (usually adults) had magical powers allowing them to understand more of the world. 18 This belief, which results from a different interpretation of the behavior of others, is not proof of an absence in the belief in a shared world, or in a lack of understanding of social codes, but, rather, a different interpretation of their own childhood experiences. Whether these experiences are affected by neuropsychological differences in sensory processing, or by some other mechanism, is here beside the point, for these differences in interpretations point out a more fundamental problem with theory of mind. The theory of theory of mind rests upon the assumption that there is a universal understanding of the world, the understanding possessed by those who are 'neurotypical', and, therefore, automatically assumes that those who don't adhere to it are non-normal.

Setting aside this question of the theory of mind, and returning to the idea of simulation and film, it is helpful to note that many so-called neurotypical individuals, can, conceivably, be thought of as having developed their understanding of the world at least partially through fiction. It is not surprising, then, to learn that many individuals with autism speak about the ways in which they have used moments from film to attempt to understand the behavior of others. Temple Grandin, in her autobiographical work Thinking in Pictures mentions several times the “CD-ROM memory of videotapes” she has in her head, and the way in which she models her responses to others on responses contained within this memory. This memory contains virtual videotapes of film and tv situations, fictional representations of the world that allow Grandin to access a 'neurotypical' understanding of how a narrative can be read and to proceed accordingly19 While Grandin does not mention her own enjoyment of this process, several other individuals have noted that it is one of pleasure. In his film My Life as an Autistic Boy , Kyle Priebe constructs a narrative of his life using a combination of animated still photographs, voice-over narrations, film stills, and music (much of which is taken from the soundtracks of popular films). In one segment, Priebe mentions the movies that he creates and plays in his own head. In a separate segment, he also discusses the way in which he created a friend for himself from the film Mars Attacks! , a friend that he used to re-enact scenes from the movie in his head in a less frightening way. Other individuals I have worked with report similar uses of film, wherein they shift elements of narrative, insert their own subjectivity, and then replay the films in their heads. While similar to Grandin's use of film, this use expands upon it, for not only does it create a databank of 'simulations' of the world, but it also becomes part of the internal narrative of the individual, and, therefore, of the self. Further, Priebe's film as a whole uses this assemblage of photographs, voice-over, and music to narrate his life story. His voice-overs consistently address an audience outside of himself, and refer to not only what he was told about his early life, but also what he remembers. Most striking, however, is the way in which self-consciously narrates himself as an ‘autistic boy', returning multiple times to the ways in which felt autism affected his life.

This is markedly different from the identity narrated by George Corra in George. Of course, part of this difference arises from the fact that the main narrative of George is not a self-narrative, but rather an editing together by George's father Henry Corra. This difference of narration can be most clearly seen towards the end of the film, where Uta Frith speaks extensively between scenes of George (often scenes which illustrate a point she makes about the nature of autism) and interviews of an adult man with autism. By editing together authoritative interviews with Frith, interviews which write over through commentary even the articulate interviews with the adult with autism, with scenes of George in which he does not have the camera and with close-ups of George freezing and turning to black and white, the film positions George as someone who cannot represent himself At the same time, the sequences in the film during which we are presented with the footage shot by George, or shown editing choices he makes (choices which most often result in long repetitions of phrases George finds amusing), allow for George to self-narrate. Many of these sequences begin with George directly addressing the camera and informing the spectator of the date, time, and weather conditions These sequences, which so vividly recall the verbal sparseness of the chronicles cited by Hayden White in The Content of the Form. These chronicles, which supply only a year and some bits of information such as “hard winter. Duke Gottfried died” are seen by White as lacking what the modern reader views as narrativity.20 However, like George's choices to repeat through editing words he finds amusing, or to ground his interactions with the camera by anchoring them in chronological time and meterological conditions, the chronicler's focus on dates and details is narrativity, just not a narrativity that follows normative models of temporal progression and psychological and emotional grounding of events. Instead, what can be read in both the chronicles and in George's sequences is self-narration that is non-normative yet still addressed to an audience of others and still part of the process of identity formation.

This self-narration problematizes Paul John Eakin's hypothesis that “narrative is not merely a literary form but a mode of phenomenological and cognitive self-experience” that is grounded in “the acquistion of narrative competence during a particularly rich phase of childhood.”21 Eakin goes on to make the assertion that if “self-narration is the sine qua non of identity, we should pause to consider its exclusionary implications for those individuals-many autistics, among others-who never master narrative in the first place.”22 What is clear in Eakin's argument is that there are rules to self-narration, rules about what to tell (the truth) and how to tell it (with affect). Yet these rules, which rest upon an understanding of the normal, Piagetian transitions from ‘ego-centered' monologues to syncretic understanding, ignores the rich possibilities not only of narrative, but of identity as well. Are we to understand that modern, Western standards of ‘truth' and temporal progression are necessary for identity? Although both White and Eakin are discussing written forms of narrative, it is likely that they would hold filmic self-narration and narration to similar standards, standards which, again, rest on normal temporality and affect. What this does, though, is create an extension of theory of mind that insists not only on a single, correct way to simulate the world, but also on a single, correct way to represent the self for others. Both parts rely on an ideal of normality in the way in which space and time is perceived and represented in the creation of the self. What must be remembered, however, is that there is no normative self any more than there is a normative body or a normative sexuality By insisting on only one correct way of simulating the world and creating an understanding of others, theory of mind automatically creates the need for an 'other'. This other has taken the form of the autistic subject, who, through the work of cognitive theorists, psychologists, neurologists, and even literary critics, has been placed on the outside of social interactions, self-representation, and engagement with narrative. Yet this is a move that denies the possibility of a multiplicity of ways of engaging with the world, and which further isolates the individual with autism.

1 Barker, John and Robert Gordon. “Autism and the Theory of Mind Debate.” In Philosophical Psychopathology , eds. George Graham and G. Lynn Stephens. Cambridge , MA : MIT Press, 1994. 172-173.

2 See for instance the definitions in Uta Frith's article “Mind Blindness and the Brain in Autism”, in Neuron v. 32(2001): 969 and Simon Baron-Cohen's book Mindblindness: An Essay on Autism and Theory of Mind , Cambridge, MA: MIT Press, 1995, pgs. 69-83.

3 See particularly Piaget's work on language, which posits a normal series of stages the child passes through in the acquistion of language, from ego-centrism to syncretism of understanding and reasoning. A good overview can be found in his work The Language and Though of the Child , originally published in English in 1926.

4 Premack, Guyand David Woodruff. “Does the Chimpanzee have theory of mind?” Behavioral and Brain Sciences. Vol. 1 (1978):515-526.

5 All three modifications were published as part of a number of commentaries by a number of biologists, primatologist, linguists, psychologists and philosophers in the same issue of Behavioral and Brain Sciences in which “Does the chimpanzee have a theory of mind?” was published. See: Jonathan Bennett, “Some remarks about concepts”, p. 557-560; Daniel Dennett, “Beliefs about beliefs”, p. 568-570; Gilbert Harman, “Studying the chimpanzee's theory of mind”, p. 576-577, all in Behavioral and Brain Sciences , vol. 1 (1978).

6 Wimmer, Hans, & Josef Perner. (1983). “Beliefs about Beliefs: Representation and Constraining Function of Wrong Beliefs in Young Children's Understanding of Deception.” Cognition , 13, 103-128.

7 Baron-Cohen, Simon, Helen Tager-Flusberg and Donald Cohen “An Introduction to the Debate.” In Understanding Other Minds; Perspectives from Autism , ed. Simon Baron-Cohen. Oxford : Oxford U. Press, 1993. 3-9.

8 Ibid., 4.

9 Frith, Uta. “Mind Blindness and the Brain in Autism.” Neuron. V. 32(2001): 969.

10 Baron-Cohen, Simon, Helen Tager-Flusberg and Donald Cohen. “An Introduction to the Debate.” In Understanding Other Minds; Perspectives from Autism , ed. Simon Baron-Cohen. Oxford : Oxford U. Press, 1993. 4. The authors, in their section “What is Autism? A Note on Diagnosis,” summarize a move from a diagnosis of autism based on behavioral traits in the 1987 edition of the ICD-9 and the 1980 ed. of the DSM-III to one based more on developmental features in the 1987 DSM-IIIR. Frith, Uta. “Mind Blindness and the Brain in Autism.” Neuron. V. 32(2001): 969 Frith, however, presents the ICD-10 and DSM-IV as both using behavioral criteria in diagnosis.

11 Baron-Cohen, Simon., Alan Leslie and Uta Frith (1985). Does the autistic child have theory of mind? Cognition, 21, 37 - 46 .

12 Ibid.

13 Frith, Uta. “Mind Blindness and the Brain in Autism.” Neuron. V. 32(2001): 972.

14 Ibid.

15 Baron-Cohen, Simon. “The development of a theory of mind, and its dysfunction.” Understanding Other Minds. Ed. Simon Baron-Cohen Footnote on page 60.

16 Barker, John and Robert Gordon. “Autism and the Theory of Mind Debate.” In Philosophical Psychopathology , eds. George Graham and G. Lynn Stephens. Cambridge , MA : MIT Press, 1994. 163-177; see in particular page 173.

17 Ibid.

18 J. Blackburn, K. Gottschewski, Elsa George, and Niki L, "A discussion about Theory of Mind : From an Autistic Perspective," Proceedings of Autism Europe's 6 th International Congress , Glasgow 19-21 May 2000, in print. Available on-line at

19 Grandin, Temple . Thinking in Pictures and Other Reports from My Life With Autism. New York : Vintage Books, 1995. 137.

20 White, Hayden. “Narrativity in the Representation of Reality.” The Content of the Form. Baltimore : Johns Hopkins Press, 1987. 6-7.

21 Eakin, Paul John. “Breaking Rules: The Consequences of Self-Narration” Biography. 24.1(2001);115-116.

22 Ibid, 122.



Return to Top of Page

Return to SCE Homepage
Return to Session Program