Deconstructing Paranoia: An Analysis of the Discourses Associated with the Concept of Paranoid Delusion

David J. Harper PhD Thesis June 1999

Back to title page


This thesis focuses on 'paranoia'(1) and attempts to describe and examine some of the assumptions upon which this concept draws as well as analysing some of the consequences of its use. Paranoia is a very broad term, one which functions within both popular culture and the field of psychiatric and mental health practice and so this analysis ranges over material drawn from both realms.

Since this study is aimed at understanding the meanings of paranoia and its implicit assumptions, the conceptual and methodological resources which have seemed most relevant here have been hermeneutic and post-positivist, in particular social constructionism and discourse analysis. Early work in this area in psychology had to struggle to be recognised as valid amidst the dominance of mainstream positivism. However, interest in qualitative and social constructionist approaches has risen (eg Danziger, 1997; Neimeyer et al., 1994) with increasing numbers of qualitative texts in psychology emerging (eg Banister et al., 1994; Hayes, 1997; Richardson, 1996). Qualitative methods have started to make inroads into clinical psychology (eg Good & Watts, 1989; Orford, 1995; Turpin et al., 1997) and even psychiatry (Buston et al., 1998; Smith, 1996) -- although some warn of a danger of assimilation and the tendency for more radical and critical approaches to be less popular with the mainstream (Boyle, 1998). There have also been special issues on qualitative research of journals like the Journal of Community and Applied Social Psychology (1994), Journal of Mental Health (1995), Changes (1996) and Clinical Psychology Forum (1998). Even the predominantly positivist British Journal of Clinical Psychology has published social constructionist and qualitative papers (Clegg, 1993; Clegg et al., 1996). There are a wide variety of cogent arguments for qualitative and social constructionist research (see the texts referred to above) but, despite this, debates continue about validity (see, for example, Michael Morgan, 1996, 1998 and Sherrard, 1997, 1998). Boyle (1997b, 1998) has queried whether these debates are entirely intellectual and epistemological but may be also emotional and ideological. She argues that resistance to qualitative and social constructionist approaches may also be due to: the loss of the authority and status of the researcher in these traditions; the loss of control over subject matter; the revelation by these approaches of important gaps in psychological theory; the threat of the breaking down of disciplinary barriers; the threat to individualistic explanations; and the threat of contentious and controversial topics. She suggests that qualitative researchers may have spent a disproportionate amount of time convincing positivist researchers of the validity of their work with intellectual arguments thus obscuring these emotional and ideological debates. She concludes that this has put qualitative researchers on the defensive since they spend more time warranting their approach than more traditional researchers. Following this, I will argue that choice of method is driven by both the aims of a particular study and its epistemological framework. Justifications of the validity of these approaches and methods are now somewhat hackneyed and will not be repeated here -- readers are referred to the texts noted above if they wish to familiarise themselves with these debates.

Social constructionist work is useful because it does not attempt to set a priori a fixed meaning on concepts, nor does it attempt to compare concepts with a notion of an independently existing reality. As a result it becomes a useful theoretical resource from which to analyse the ways in which a concept emerges historically and the different ways in which it functions in different domains in culture. There is no one definition of social constructionist theory since it is in reality more of a framework and workers within this tradition constitute a broad church. Gergen (1985) has argued however, that there are four assumptions implicit in most social constructionist work: a radical doubt in the taken-for-granted world; the viewing of knowledge as historically, socially and culturally specific; the belief that knowledge is not fundamentally dependent on empirical validity but is rather sustained by social processes; and that descriptions and explanations of phenomena can never be 'neutral' -- the inverted commas here (and elsewhere) aiming to bring a taken-for-granted concept into question -- but constitute social action which serves to sustain certain patterns to the exclusion of others.

Burr (1995) has noted that social constructionism is different from traditional positivistic psychology in a number of ways. First, it is anti-essentialist in that it does not believe there are innate discoverable psychological essences like 'personality' or 'cognitions' or even 'emotions'. Second it is anti-realist in that it does not believe in a naïvely objective reality 'out there' which can be directly perceived. This should not be interpreted to mean that social constructionists deny reality. Rather that 'reality' is socially constructed and there are different attempted solutions to this issue within social constructionism as we will see later in this introduction.

A third difference is that social constructionists assert that all knowledge is bound by time and culture and thus grand theories which attempt to explain phenomena in an ahistorical and culture-free manner are ultimately flawed. Theories should be explicit about their local and provisional status. Fourth, as Burr puts it 'our ways of understanding the world come not from objective reality but from other people, both past and present' (1995, p.6) through language. Language is then, not a peripheral matter, but is central to the way we view the world. Fifth, language is seen as constitutive rather than as merely descriptive. Sixth, social constructionism rejects the false opposition where explanations are either searched for within the individual (eg in their 'attitudes') or within social structures, but focuses on interactions between people and on social practices. Finally, social constructionism focuses on processes rather than static entities like personality traits. As Burr notes 'knowledge is therefore seen not as something that a person has (or does not have), but as something that people do together' (1995, p.8, emphasis in original).

Given the important constitutive power of language, an attempt to understand the meanings associated with paranoia needs to be sensitive to the myriad ways in which paranoia is constructed through language. Paranoia is brought into being through a vast array of different types of discursive events: from the referral letter from a General Practitioner (GP) about a patient who is felt to be paranoid, to the script for a movie like Conspiracy Theory (dir. Richard Donner, 1997), to the newspaper article about conspiracy theorists, to the psychiatric diagnostic interview, to the entry about 'persecutory delusions' in a person's nursing notes. Such an approach is a far cry from the traditional ways in which language is studied in psychiatry as a potential sign of pathology (eg Oxman et al., 1982). This is not to fetishize language since language is not everything and what is equally important is how concepts like paranoia become enshrined in certain kinds of institutional spaces and practices (eg admission to hospital or the prescription of neuroleptic medication) which are anything but simply linguistic. However, language provides one way into an analysis of this realm. It makes sense to try to analyse the ways in which paranoia is constructed by examining a range of different texts since, to some extent, the concept could actually be said to be constructed in and through language in psychiatric interviews, case conference discussions, ward-rounds, letters to GPs and so on and and this is what I have sought to do in this thesis.

Having introduced my epistemological framework and research aims, I now want to say something about method. In much social constructionist and post-positivist work, method is problematised (see, for example, Burman, 1996; Hollway, 1989). I have discussed these issues, where relevant, in the chapters and in chapter 7 where I reflect on the research. For the most part, although I have used Discourse Analysis (DA) as a method throughout the thesis, its structure and focus has been driven both by the particular demands of the different textual materials used and my aims in each chapter. My methodology covering my analysis of the interview material is discussed briefly in the introduction to Part II but this could only be written after the analysis and it would therefore be misleading to include a 'method' chapter near the start of the thesis as if I could have written it before. Moreover, my methodology for analysing the texts in Part I (which is empirical but in a different way from Part II) evolved over time. As a result, what I include here are the analytic concerns which drove the analysis. I am persuaded by the notion of the hermeneutic circle which sees analysis as a circular, iterative and recursive process of interpretation rather than notions of a method which is then applied to an issue. Newman & Holzman make a similar argument:

Practice is not our method. Rather, method is what we continuously practice. To us, method is not something applied to something. In our view, it is, to paraphrase Marx, an activity 'for itself' as opposed to 'in itself'. In Vygotsky's language, method is a 'tool and result' as opposed to a 'tool for result'.(1996, p.ix)

Within social constructionism DA has become an increasingly used approach for examining the constructive aspects of language. I have noted how the concept of paranoia is constructed. Parker (1992) has argued that a discourse can be defined as 'a system of statements which constructs an object' (p.5). If we wish to examine a conceptual object like paranoia then, we need to elucidate the system of statements which surround and construct it. Stenner & Stainton Rogers (1998) have suggested that by regarding jealousy as 'a discursively constituted manifold of understandings it becomes germane to explore the multiplex through pattern analysis' (p.71). They use Q methodology but DA is another approach which can help with this. In reality DA is an umbrella term for a number of different approaches to language. Within this approach a growing body of work has focused on language use as a form of social action.

My use of DA here draws on a tradition in British social psychology influenced both by work broadly in the fields of linguistics, hermeneutics and ethnomethodology (cf Potter & Wetherell, 1987) and by Feminism, Marxism, critical sychology (eg Fox & Prillitensky, 1997) and the writings of Michel Foucault (Burman & Parker, 1993; Parker, 1992; Levett et al., 1997).

The first tradition is exemplified by recent work in rhetorical psychology and DA which suggests that talk is often varied and contradictory and this indicates that it is oriented to perform certain functions although this is not necessarily a planned or deliberate process (Potter et al., 1993). When talk is analysed it is possible to elucidate abstract discourses (Parker, 1992) or 'interpretive repertoires' (Potter & Wetherell, 1987). In contrast, the second tradition frames:

analysis explicitly in terms of the Foucauldian concept of discourses as networks or complexes of power-knowledge relations, together with the poststructuralist concept of deconstruction. The latter is the technical term for seeking out taken-for-granted meanings; for example the oppositional positions that are absent in texts.

Henwood & Pidgeon (1994, p.234)

These two approaches draw on the concepts of: variation and contradiction in accounts; how accounts are constructed and constituted; and how those accounts function and are linked to power (Parker, 1997a). The first approach focuses more on variation, construction and function whilst the second focuses more on the related notions of contradiction, constitution and power (Parker, 1997a).

In contrast to a traditional psychological view, language here is not seen as descriptive of the world but rather as constitutive and is viewed not as a path to finding out about something else (eg about 'cognitions' or 'pathology') but as something worthy of study in itself because of its effects since in Rose's (1990) words language makes 'new sectors of reality thinkable and practicable' (pp. 105-106). As I have argued, such a standpoint is extremely useful when studying phenomena like psychiatric categories such as 'paranoia' which are produced almost entirely within language. As Wetherell (1994) has noted, the discursive examination of such categories thus moves from using them as an explanatory resource (eg 'she is doing that because she is depressed') to regarding them as a topic to be explored (eg 'what interests are served by the concept of "depression" in this context?'). There are now a number of social constructionist, discursive and poststructuralist approaches to similar categories: anorexia (Malson, 1997); anxiety (Hallam, 1994); delusions (Harper, 1992); depression (Stoppard, 1998); jealousy (Stenner & Stainton Rogers, 1998); and multiple personality disorder (Gillett, 1997).

Edwards & Potter (1992) have described a number of ways in which speakers use rhetorical strategies to make what they are saying appear factual -- often when there is some disagreement about the facts and when the speaker has a stake in the outcome. Speakers position themselves and others by using such discourses tactically (although, again, this is not necessarily an intentional process) and are positioned by those same discourses since certain linguistic resources impose constraints on what can be said. Hepburn (1998) suggests that the identification of such strategies is part of the project of deconstruction.

It is important to emphasise however that discursive approaches are not homogenous and there are wide differences of view on central issues like reality, the ideological functions and effects of accounts, the notion of interest and politics. For Burman & Parker (1993), Parker (1992) and Willig (1998) reality is an important focus since power relations have real effects on people. Adopting a critical realist stance as they do means seeing reality not as an independently existing objective conception in a naïve sense but as a domain of institutional power relations. In contrast, Edwards & Potter (1992) are less concerned with 'reality' and more with how different accounts are made to seem real. Curt (1994) on the other hand worries about the desirability of competing accounts all claiming to be real, proposing the need for a critical polytextuality. Parker (1992) claims that DA does not necessarily entail examining the ideological effects of discourse but argues that it should do. He asserts that ideology should be understood not as a belief system or as something with a truth value but as a description of relationships and effects. On the other hand Potter et al., (1990) argue against what they see as a tendency to reify discourses as ideological objects, asserting that this misses the variability of use in specific contexts. This point is developed further by Wetherell & Potter (1992) when they emphasise the importance of a focus on ideological practice and ideological effects rather than ideologies per se.

The relationship between interest and talk is a central topic in discursive psychology (Edwards & Potter, 1992) in that 'the invocation of stake and interest is understood as a rhetorical strategy but not as a motivational explanation for rhetorical moves' (Madill & Doherty, 1994, p.269, emphasis in original). Here, however, I will be using the notion of interest differently to point out that the effect of certain accounts in certain contexts is to serve particular ideological interests. This means moving beyond the text and entails a political commitment on the part of the researcher (Burman & Parker, 1993; Harper, 1996c; Parker, 1992; Willig, 1998). In exploring what kind of effects such accounts have, varieties of DA which are politically committed are able to describe the specific ideological effects different explanations have at different times.

In the same way that research cannot be theory-free so also it cannot be apolitical. As Wilden has pointed out 'all knowledge is political and the refusal of politics is a political act' (1972: xxii). Indeed one of the motivations for those adopting social constructionist approaches has been that such theories allow for a better representation of human diversity in contrast to more normative positivistic accounts. However, social constructionism and DA are not necessarily anti-oppressive -- Walkup (1994) has shown how social constructionist arguments can be used by reactionaries. Moreover, some have claimed that a political analysis which is too broad and committed does violence to texts and misses the point that politics and power are in the detail (Widdicombe, 1995).

Here I take a varied approach: at times I find it helpful to focus at the level of the rhetorical strategies individual speakers use (eg to position another person or group or in constructing a point of view as plausible) whilst at other times I find it helpful to focus at the level of cultural texts like movie scripts or to examine some of the ideological consequences of accounts. At other points my analysis will draw on notions developed within what has come to be termed positioning theory (Davies & Harré, 1990) -- there is a fuller description of this in chapter 2. My intention here is not to offer a cognitive or intentionalist account of texts and talk about paranoia but, in describing the effects of these accounts I will be arguing that they serve certain ideological interests. The debates about the virtues of realism versus relativism debate (see Parker, 1998 for a recent portrayal of the different positions) are, to some extent, clichéd and, in some ways can be seen as a transformation of the theory/practice debate (ie on what grounds can we privilege one account over another?). As a result I clarify my position in the context of this debate in chapter 8. I will attempt to detail some of the interests which have helped to shape my analysis in chapter 7 and in chapter 8 I will advance some of the implications of my research for political interventions in mental health.



A deconstructive approach to madness necessitates a critical understanding not only of the history of psychiatry but also of the history of cultural experiences of mental distress and of the cultural meanings of reason and unreason.

Parker et al., (1995, p.56)

In the introduction I noted that I have used a number of different varieties of DA to examine a range of textual materials. Methodology should be contextual, driven by the need to be sensitive to different kinds of analytic material. Although DA has most often been used in the analysis of interview material or discussions recorded 'live' (eg Burman & Parker, 1993; Hollway, 1989; Potter & Wetherell, 1987) discursive approaches have increasingly been used to examine a wider range of texts including those of popular culture (Parker, 1997a, 1997b) and material already in the public domain like parliamentary debates and newspaper reports (Edwards & Potter, 1992).

In the first part of this thesis I build on this wider use of DA to examine a number of structured textual sources in the public domain concerning paranoia. This is in order to explore three different contexts: historical (ie where did this concept come from and what were important influences in its construction?); cultural (ie how does popular culture sustain the notion of paranoia?); and professional (ie how do professionals construct the concept?). Each of these contexts construct paranoia in similar and different ways. Although these contexts intersect, I will be focusing primarily on each in turn whilst attempting to remain sensitive to these inter-connections.

I will be using textual materials and methodological resources differently in each chapter. In the first chapter, I develop an analysis of Aubrey Lewis's account of the history of paranoia, published in a psychiatric journal. In the second, I draw on a number of different sources from popular culture including cartoons, TV series', newspaper stories, films and so on. In the third, I analyse a number of different contributions to the professional literature on paranoia. As a result work in rhetorical and cultural psychology as well as cultural and literary studies informed my use of DA here. Of course, the chapters in this first part are recursively linked to the later chapters analysing the interviews since they were written and re-drafted whilst I was preparing interview questions, interviewing participants, transcribing and analysing the interviews. Indeed, readers may find it helpful to examine appendix 15 for an outline of steps in the analysis of the interview material to get an idea of my method for Part I since I followed similar steps. First, I collected different texts (Lewis's paper and other historical material, examples of 'paranoid' culture and commentaries, examples of professional literature for chapters 1-3 respectively) which I then read and re-read. I noted different discursive effects: themes; discourses (Parker, 1992); positions taken up; rhetorical strategies (Edwards & Potter, 1992); implied oppositions (Billig et al., 1988; Parker, 1992) and so on. Then I discussed various ideological consequences of these discursive features, using other theoretical material as an analytic and explanatory resource.

1. Throughout this dissertation, concepts like this are, at times, placed in inverted commas to reflect the fact that they are problematic. However, I feel it would only distract from my argument if I were to continually highlight all problematic terms in this way -- a particularly distracting side-effect of some postmodernist writing.

Back to title page