Abstract
The media has been shown to be the public's single most important source of information about mental illness [1, [2]], and these sources are generally very negative [3, [4], [5], [6], [7], [8], [9], [10]]. This negativity is identified in both factual and fictitious materials [11, [12], [13]]. Detailed analyses of these depictions are a vital resource for mental health professionals seeking to reduce the stigmatisation suffered by their clients. There are published reports of mental illness depictions in Australasian print media [14, [15]] but we could find no analyses of the depictions of mental illnesses in Australasian television.
Media researchers [16, [17]] agree that all media products are fictions, ‘constructive interpretations’ [16].
People engage with these media stories, whether accepting, challenging or rejecting, using discursive resources of language, images and stories [18, [19], [20], [21]] available in their community. Further, there is evidence [22] that people do not consistently monitor the origins of these resources and that the use of the resources is frequently shaped by heuristics of availability and representativeness [23]. In these circumstances, the frequent exposure to the same understandings of mental illness can create self-validating impressions of the correctness of those understandings.
From the perspective just outlined, those who produce television drama are understood to draw upon the discursive resources with which they and, they assume, their intended audience are familiar [24]. This creates two interdependent, research tasks: to identify the discursive resources used in the producers' constructive interpretations [16]; and to explore the resources that viewers use when engaging with the media materials. There is already compelling evidence [25, [26]] that viewers use much the same resources as producers when making sense of media products.
Whether analysing productions, as in this study, or viewer responses, attention is directed to the language, images and stories used as the agents work to achieve their social goals. From suitable samples of material the patterned use of linguistic and rhetorical resources can be identified. As argued elsewhere [14], analyses of media materials must take cognisance of media practices. For fictitious materials, the primary drive is to entertain [24], to attract their audience in domestic settings where there are considerable distractions [27].
In television dramas, as in cinema, characters are created by appearance, words and actions as well as the responses of other characters to them. The impact of these features can be heightened or moderated by shot selection, setting, sound, lighting and other technical aspects of the production. These, in our argument, are discursive resources contributing to the broadcast depiction. Our analytic work on a sample of television dramas enabled us to identify patterns (we refer to them as themes) in the constitution of mental illness. Models for the work reported here are provided by studies of race relations [18, [28], [29]], gender [30] and scientific talk [31, [32]].
Method
Programs were selected from weekly listings in a program guide (The New Zealand Listener). Items were included when there were references to mental illness in characters or storylines.
Dramas sampled
The first author (CW) undertook the initial analysis of the sample material. This required systematic, repeated viewings from which a synopsis file was created for each item. These files recorded the date shown, duration of episode, channel, country of origin, episode title (if any), and a list of the core, central and relevant minor characters. It also included where the drama was set and a summary of each storyline. Programs differed in the number of interwoven stories and how these were handled.
Characters were classified as having a mental illness if they were designated as such by another character in the program [3]. Designations diff e r e d from a single revelation to relatively elaborated presentations of a psychiatric history, naming of an identified mental disorder, or classification as ‘mentally ill’. In drama, mental illness may be presented as a bald fact, or aspects of the attribution may be contested, with the contest providing a dramatic focus. Eleven of our sample programs listed in Ta b l e 1 (programs 1–2,5–9,11–14) involved 17 characters with an uncontested designation as mentally ill. Two other programs (3,4) were built upon contested mental illness in the context of brutal crimes. The last program (10) portrayed the central character (Janice) in a manner suggestive of a mental disorder, although she was not formally designated as having a mental illness. This created possibilities that were exploited for dramatic p u r p o s e s.
When characters were identified as having a mental illness, their actions and relationships in the storyline were summarised in a character profile following further critical viewing. This profile included the available information on their history, medication, diagnosis (labels used and by whom) together with displayed symptoms, behaviour, and summaries of their social interactions. Many of these characters were said to have a chronic mental illness. Some had been institutionalised frequently, others recently released into some form of community care. The most common diagnosis was ‘mental illness’ with more specific labels (schizophrenia, depression, and bipolar disorder) being assigned to only five characters.
Transcriptions and detailed descriptions of important scenes within the mental illness storyline were added to the character profile, together with descriptive adjectives and other labels applied to the character. From these profiles, we identified the themes common to the depictions of characters with a mental illness. A theme, also referred to as a pattern [29] or interpretative repertoire [20, [33]], is a cluster of mutually consistent discursive resources; words, images and storylines. Characterisations and plots typically draw on more than one theme. In practice, each theme will be instantiated in what the character says and does together with comments, responses and descriptions by other characters.
In describing each theme, we have sought to provide sufficient detail so a reader might recognise instances of the use of these resources in programs that they view. Examples of specific actions or qualities that are broadly representative of the theme are included to aid such recognition. For two reasons we have chosen not to provide a single detailed example. First, as indicated above, talk is only one of the means through which television dramas create characters and there would not be space to present the relevant dialogue within the pertinent context. Second, because we could at most provide a single, detailed example, we would face a difficult task to establish its representativeness for the reader.
Methods
The 14 programs portrayed 20 characters (11 men, nine women) with a diagnosis of mental illness. Of these, 16 played a central role in their particular program. For 15 of these 20 characters, the mental illness was identified entirely through generic symptoms and labels. Specific diagnoses were given in five programs; schizophrenia, programs 1, 12, 13; depression, program 7; and bipolar disorder, program 11. The following 10 themes were identified in the depiction of mental illness and are presented in the order of frequency of deployment. In each program, ‘mental illness’ is instantiated through an interplay of two or more themes. Contributions to this interplay may come from other characters or the target character's actions in different settings, as shaped by the needs of the story in the context of the television program.
1. Dangerousness–aggressive (15/20 characters)
The character is physically violent to self or others, speaks and/or moves in aggressive ways, or threatens others with physical and/or emotional harm.
Typical actions include self-mutilation, suicide, homicide and emotional outbursts. When the character is also criminal they can appear bad, evil, or frightening. Such characteristics are often established or implied in the first sequence in which they appear (8/15 characters). Personal characteristics associated with this depiction are spiky or unkempt hair, affect that is inappropriate in kind or quantity, actions that are socially inappropriate, such as standing too close to others, or erratic gaze in interactions.
Five characters were depicted attacking or killing others (programs 3, 4, 6, 10, 14) and three of these were women (programs 3, 10, 14). Four characters, two men (programs 6, 13) and two women (programs 2, 7), were depicted as being violent to themselves either by self-mutilation or suicide.
2. Simple/childlike (14/20 characters)
Characters lack comprehension and performance of everyday adult roles, appearing lost and confused.
Typical actions include speaking in grammatically simple sentences, in a childish voice, and breaking into children's songs such as: ‘Incy, wincy spider’ (14). As with children, their actions show a lack of awareness of social conventions. Other characters are relatively more important in this theme as their talk about or interactions with the person with mental illness can mark the latter as childlike; making decisions for them, talking down to them, and explaining concepts in exaggeratedly simple terms. Personal characteristics include unkempt appearance, such as disorganised or poorly matching clothes, which is given particular meaning for the character through the broader program context.
3. Unpredictability (12/20 characters)
The critical element is that the viewer is unable to anticipate or account for behaviour except through making an attribution of unpredictability.
Unpredictability involves the display of behaviour that is erratic, without adequate justification, or socially unjustified within the context in which it occurs. Examples include: knocking on elderly people's doors in the early morning; sudden mood changes; and abusive or violent outbursts. Theft or other kinds of unreliable or untrustworthy behaviour that have not previously been signalled may be shown. Predictably no consistent, commonly associated personal characteristics were identified for this theme.
4. Failures/unproductive (12/20 characters)
The character is shown, or described, as failing at tasks or being otherwise unproductive in the real world, rather than simply not belonging (theme 5: Asocial).
Portrayal of failure may be signalled by a lack of identifiable employment, ineffective interactions with family or other, potentially supportive people, and by living alone in run down surroundings. Other characters, such as the core cast, provide a strong contrasting group of productive, effective individuals. The failure to communicate effectively can be used to provide an ‘explanation’ for the erratic community involvement constructed with theme 5 (Asocial).
5. Asocial (9/20 characters)
Asocial depictions present characters as rootless, or as not belonging, by showing them as rarely participating in everyday social interactions within the community.
This depiction draws together resources used to present the ‘transience’ [5] that is part of news media depictions of mental illness. These characters usually lack close relatives or friends and are rarely shown or described as having any positive relationships in the community. Such loners effectively have no social identity outside of their mental illness. Their social isolation, which may appear to be desired by the character, together with the limited social skills displayed, can reduce the character to a cipher, a mere plot device. The asocial theme, used in conjunction with Dangerousness–aggressive and Unpredictability makes a character appear less amenable to social control and hence more threatening, whereas its use in conjunction with Failures/unproductive and Vulnerability positions the character as a victim for whom we may be encouraged to feel sympathy.
6. Vulnerability (9/20 characters)
Vulnerability is portrayed as helplessness, including the inability to control their lives which are dictated by others: landlords, neighbours, the wider community, doctors, the police, or the media.
In these depictions, vulnerability is displayed by the character being subjected to harassment, accusations, manipulation or exclusion for no reason other than their mental state. When assailed in these ways, as when evicted from accommodation, or subjected to abuse (sexual, physical, emotional), the character is shown as an ineffectual individual whose testimony or interpretation of the event is discredited by mental illness. Vulnerability may be signalled by constrained body language, a low level of muscle tone, and elements of theme 2 (Simple/childlike). A vulnerable, incompetent person may, if sufficiently stressed, become dangerously aggressive and this was employed in conjunction with theme 7. (Dangerousness–incompetence) to ‘explain’ why the character had not responded more effectively before becoming stressed.
7. Dangerousness–incompetence (6/20 characters)
Marginal competence in self-care while living in and interacting with the community is presented as endangering the character and/or others.
Typical actions signalling incompetence include slow, excessively deliberate movement and patterns of speech, attention to detail, and being easily distracted. Close-up shots of list-making and self-talk can also be deployed in the depiction. Incompetence may appear initially as an inability to make simple decisions, to count money or to take responsibility. It may be presented through careless or reckless driving, failure to turn off a stove properly, or in interactions with other members of the community but, in this pattern, unlike 2 (Simple/childlike), it creates a threat to the patient and others in their community.
8. Untrustworthy (6/20 characters)
Characters may be shown as unworthy or incapable of behaving in trustworthy ways: shown or described as manipulative or misleading, for example by misrepresenting their own abilities or situation.
Viewers may see, or hear, the patient thieving, lying or taking advantage of another's trust. This sometimes occurred in conjunction with theme 4 (Failures/unproductive), or in isolation, to underpin their asocial or outcast status. Body language, particularly failure to make eye-contact, can be used in common sense fashion to emphasise untrustworthiness.
9. Caring/empathic (6/20 characters)
The character with mental illness is shown or described as expressing love, kindness, compassion and/or joy in the context of a concern for others and their wellbeing.
At these times, the character appears to relate in a natural, unforced and/or effective manner which may include a rather childlike directness. Examples include feeding a cat, buying chocolate for another and cuddling a baby. The theme has strong associations with 2 (Simple/childlike). In our sample, this theme appeared unexpectedly, infrequently and briefly, and could be inconsistent with earlier actions. None of the storylines drew substantially on this theme.
10. Social outcast (4/20 characters)
Ordinary characters actively rehearse the common stereotypes of mental illness such as dangerousness, unpredictability and incompetence in rejecting those with a mental illness.
The active rejection differentiates this theme from 5 (Asocial). Where such scapegoating is occurring those with mental illness can be referred to as dirty, evil and bad. Active rejection may be presented by the initiating characters as protecting vulnerable members of the community such as children. The initiators are likely to include threats against those with mental illness, the place in which they live, and those seen as their supporters. One storyline (program 9) was substantially shaped by the discursive resources of this theme.
Discussion
Our data show that the principal discursive resources for portraying mental illness in our sample of television dramas were outstandingly negative. First, these discursive resources enable depictions of mental illness as a threat to society and society's norms. In 10 of the 14 programs, such threats were realised in actions that were the focus of police investigations, paralleling the association between serious crime and mental illness in news reports [5, [8]]. Themes such as Dangerousness–aggressive, Unpredictability, Dangerousness–incompetence, and Untrustworthy were central to the instantiation of such threats as arising from the character's mental illness. Such portrayals were widespread in the sample material and our concern is these depictions of unpleasant, aggressive and antisocial actions are sourced in the character's mental illness. By sourcing the ‘threatening’ behaviour in the character's mental disorder, the depictions generate or sustain fear and mistrust further alienating other members of society from those who have a mental illness. Drama producers utilise these themes to create dramatic tension, to signal mental illness in a character, and as providing an accessible motivation for otherwise improbable actions.
Compared to previous reports our findings give less emphasis to violent acts. For example, Signorelli [6] reported high levels of violence to self and others by those characterised as having a mental disorder. Our sample depictions involve a relatively low level of violence and, with one exception (program 3), the violence shown is not extreme. This may reflect the absence of items from the United States in our sample.
Second, we identified ways in which persons with a mental illness are constructed as abnormal and unattractively different from the viewer. Themes particularly contributing to this construction include Failures/unproductive, Asocial and Social Outcast. The discursive resources of the Vulnerability theme may contribute to this negative portrayal or to creating a more sympathetic victim. Hyler et al. [7] created a category of ‘narcissistic parasites’ for those they described as ‘self-centred attention seekers’ that has many similarities to this construction. Characters depicted as unattractive do not engender sympathy or understanding, the person is masked by the fore grounding of their disorder, they remain two-dimensional, and do not develop in a manner with which the audience can identify or relate to. The lack of an acceptable identity renders these sufferers as other than ‘human’ as understood within daily experiences.
Third, as noted by Wahl [9], mental retardation is often confounded with mental illness in film and television depictions. We identified the themes Simple/childlike, Vulnerability, and Dangerousness– incompetence, through which sufferers of mental illness are relegated to a separate, inferior, somewhat ‘simple’ group, who are unable to cope with ‘real life’. In this construction, they are shown to operate with lower levels of comprehension, talent and intellect than ‘normal’ people. Such portrayals can provide ‘explanations’ for behaviour that is powerless, vulnerable or incompetent but the lack of differentiation is stigmatising of both those with mental retardation and those with a mental disorder. In addition, this type of depiction makes the possibility of recovery appear remote.
The paucity of resources for depicting mental illness in a positive way in these dramas is striking. The sole completely positive theme, Caring/ empathic, was only deployed occasionally, in brief sequences and was not central to any character. Despite thorough examination, we found no character depictions that drew on positive discursive resources to create qualities such as being productive, creative, or providing leadership.
Specific indicators or symptoms sustaining a diagnosis were rarely provided in our sample. For most programs, a psychiatrist would be unable to confirm or provide an authoritative diagnosis for the disorder on the basis of the evidence provided. Diagnostic labels, when present, appear in conjunction with the generic ‘mental illness’ label and the behaviour shown is rarely differentiated by the appropriate clinical symptoms. Such generalised depictions encouraged viewers to draw on prior understandings of mental illness to understand the character and to anticipate their role in the story. This process of generalising from existing knowledge about ‘mental illness’ to the character and the particularities of the character's behaviour can confirm the prior under-standing.
Our findings differ from those of Hyler et al. [7] in that a greater proportion of our themes are primarily negative. Four of Hyler's categories of mental illness depictions in film were clearly negative (‘homicidal maniac’, ‘seductress’, ‘narcissistic parasite’, and ‘zoo specimen’) and these are quite different from those identified in this study possibly reflecting differences between film and television. A film has more control of its viewers, who leave the everyday world for an entertainment experience. Television enters the home and competes with familiar demands and people for the viewer's attention [27]. Producers and writers anticipate such competition and present their story in ways that are accessible even to the distracted [24]. This requires stories to be simpler and more clearly drawn. Typically, drama series are given conventional settings and favour realist styles of storytelling.
Our study also differs in selection and analysis from that of Hyler et al. Without providing any rationale, Hyler names 28 films that appeared between 1904 and 1989. Describing a similar period, Winnick [34] identified 150 films containing mental illness depictions. Hyler et al. provide no description of the steps by which they performed their analysis, suggesting that their categories may be summaries of their impressions of central characters from a peculiar selection of films. Our emphasis was to identify the discursive resources used in constructing a number of characters, recognising that resources may be used in differing contexts to create differing portrayals. We acknowledge that the pre-eminence of the Dangerousness–aggressive, Unpredictability, Dangerousness–incompetence and Untrustworthy themes could have arisen because the majority (eight of 14) of our sample were police dramas. This might be considered a distortion except for the strong association between mental illness and serious crime in the news media [5, [8]].
Our findings must be considered in the light of several limitations. First, our sample is not complete. When describing serials, the program listings used in selection often assumed familiarity with the ongoing storylines and typically did not make explicit reference to mental illness. Several United States programs with relevant storylines were omitted as post-broadcast attempts to obtain a copy of the episodes were unsuccessful. This absence of United States programs may bias the depictions analysed towards realism and restraint. Despite this, our findings are largely consistent with previous analyses of media depictions of mental illness.
Second, we chose to unpick the depictions to identify collections of discursive resources deployed in the portrayals rather than counting occurrences of specified categories of behaviour. We recognise that increasing the size or nature of the analysed sample may reveal other themes or allow identified themes to be deconstructed providing more detailed delineation of the discursive resources employed. In addition, because of limits to space, we have not been able to provide readers with transcript material that would allow them to check our readings and analyses.
Third, readers will recognise that many terms involved in conveying the themes use words that are, in part, subjectively defined and we have reported no measures of reliability for the judgements involved. This weakness is being addressed in current studies with a range of materials which suggests that analytical judgements have an acceptable level of reliability across readers/viewers.
Despite the identified limitations, our discourse analytic techniques have enabled us to identify clusters of discursive resources deployed in prime-time drama constructions of mental illness. We have shown how a rather undifferentiated ‘mental illness’ is associated with dangerousness and rather unattractive, less than human persons through these resources. Our concern is that these dominant depictions encourage people to shun patients and to oppose facilities to help them [35]. We are also concerned that these resources undermine efforts being made by patients, their families, and carers to understand and respond to the disorder in a positive, effective way. Clearly, there needs to be further studies of visual depictions of mental illness both to test the generality of our findings and to show how viewers interpret the depictions.
Footnotes
Acknowledgements
This study was supported by a Faculty of Medicine and Health Science, The University of Auckland, summer studentship to Claire Wilson.