To contents page

To Psychiatric Activism Page

Chapter 9

Conclusion: The prison house of language

- One difficulty, said Stephen, in aesthetic discussion is to know whether words are being used according to the literary tradition or according to the tradition of the marketplace. I remember a sentence of Newman's in which he says of the Blessed Virgin that she was detained in the full company of the saints. The use of the word in the marketplace is quite different. I hope I am not detaining you.

- Not in the least, said the dean politely.

- No, no, said Stephen, smiling, I mean...

- Yes, yes: I see, said the dean quickly, I quite catch the point: detain.

- James Joyce, A portrait of the artist as a young man

Squire (1990) suggests that different forms of social psychology can be understood as operating in terms of three kinds of popular narrative: Detective story, autobiography, and science fiction. Traditional empirical social psychology, intent on discovering the hard facts about individual and society, is structured as a detective story; alternative social psychology, of the sort which wishes subjects to express their views in their own words and engages in informal and participant forms of data-gathering, models itself after autobiography; while discourse-oriented social psychology, which is interested in collective as much as individual patterns of meaning, classes itself as science fiction.

Discourse psychology is science fiction not only because of its still marginal position in academia, but also because its professed interest in deconstructing and re-constructing the technologies of its parent discipline - and in grafting foreign apparatuses (borrowed from philosophy, sociology, linguistics, literary studies) onto the existing machinery for social-psychological knowledge production - often serves as a pretext for extra-scientific commentary on society and subjectivity.

In this dissertation I have taken apart various knowledge machines which have been, or could be, deployed in relation to the question of mental illness - quantitative and individualising empirical enquiry (Chapter 2), Whig history (Chapters 3 and 4), biopsychiatry (Chapter 5), and discourse analysis itself (Chapter 6) - and have used the parts to fashion fanciful new technologies of my own. Concurrently, I have presented a commentary on the form of subjectivity implied by recidivism, arguing with Foucault that the sporadic physical incarceration and liberation of the 'mentally disturbed' person occurs in the context of discourses of modernity which normalise and accommodate aberrant behaviour.

Like the rest of us, except perhaps more so, the recidivist is made the subject of an implacable scientific determinism while at the same time having agency thrust upon her. Thus her oscillation between freedom and incarceration is merely one instance of the fundamental duality which is constitutive of modernity. I have tried to demonstrate this duality historically, in the disciplinary texts of modern psychiatry, in the moment-by-moment vacillations of the confessional, and in psychiatry's technologies for discursive surveillance. Thus I have followed Foucault (1967) in claiming that "language is the first and last structure of madness, its constituent form; on language are based all the cycles in which madness articulates its nature" (p. 8).

As with science fiction, however, the danger with the kind of text I have put together to warrant this claim is that it will end up "too far from conventional representations of reality to be taken seriously as an alternative to them, but also close enough to these representations in some ways, to be co-opted into them" (Squire, 1990, p. 44). In particular a text such as this dissertation may become co-opted by what Michael (1991) refers to as the modern axis of clarification-stabilisation-practicality, i.e., it may start using words as currency in the academic marketplace to answer questions such as: How can we understand readmission and its relationship to mental illness? How can we use this knowledge as a basis for further research? How can we practically intervene to make things better? I have tried not to engage with such questions, instead addressing the prior question as to the status of the knowledge-making procedures used to constitute the objects on which modernist discourses fasten. Although I have also tried throughout to speak of my own procedures, even while allowing them to operate, some further reflection is required.

Methodological commentary

The techniques I have used in this dissertation could all broadly be classed as forms of discourse analysis. In earlier chapters I constructed a reading of the history of psychiatry designed to show how the twin projects of scientific progress and humanist reform are accompanied, as modernist projects are, by an extension of disciplinary power into, and from, the furthest recesses of the social world. I tried to show how psychiatry, which speaks continually of objective knowledge and of liberation from the distorting effects of power, stands mocked by the figure of the recidivist patient, who can never finally be grasped by or uncoupled from the system of knowledge and power of which it is an effect.

In moving from historical material to contemporary texts, I have continued with the same kind of analysis, namely to bring into visibility the discursive structures presumed to underlie the surface appearances of psychiatric practices - in this case the personal interview and the psychiatric case history.

Inevitably, the structuralist approach yielded the sorts of objects it is intended to, namely repetitive patterns of psychiatric writing and speech. These include depictions of patients as liberated into a 'community' which increasingly resembles the institution (Chapter 3); preemptive appeals to imminent scientific breakthroughs (Chapter 4); and ultimately circular references to previous encounters with institutional psychiatry to account for the current hospitalisation (Chapters 7 and 8). When such structures are made visible it is easy to start believing that they operate to reduce diversity and to limit the scope of what it is possible to say and do, while in fact they are productive of diversity and are precisely the kinds of mechanisms through which it becomes possible to say and do anything at all.

In that the analysis itself operates in and through language (although it also draws on an artificial language of statistical conjunctions), it can only point to those features of language which are already in the process of being supplanted by more nuanced formulations, much as humanist critiques of psychiatry are forever doomed to fasten onto a few remnants of its scandalous past rather than the productive realities of its present.

Thus a structuralist approach must, to recall Cox's (1989) words quoted earlier, result in 'salt-flats of abstraction' from which all the 'beautiful specifics of culture' have been stripped. So with reference to the psychiatric case histories, for example, we are left with no way of knowing that patient #378 has in the past three weeks shown a tendency to fall over his own feet, that patient #1018 believes that Satan has taken over her boyfriend's personality, or that patient #1465 is a political detainee who suffers from echolalia and suicidal thoughts and has been tearing holes in his clothes. It may be that meaning is contained not in frequent words and stereotypical formulations, but in everything that is infrequent, atypical and silent.

I have tried to show how the structuralist scaffolding on which for example psychiatric case histories appear to be built also forms a surface of visibility against which such singularities stand out all the more clearly, but in giving an account of the material for even minimally 'scientific' purposes it is impracticable to show more than a handful of such instances. Ironically, the case histories themselves appear to be caught in exactly the same tension, attempting in one gesture to give an account of patient histories and to account for them - to tell extraordinary stories while simultaneously showing how they should be understood.

The numerical techniques developed to assist with the task of structuralist analysis, while certainly not providing an escape from this impasse, do offer possibilities for further experimentation. I have, for example, in the analyses presented in this dissertation consistently used the strongest sets of connections among words. This amounts to following an aggressively structuralist strategy. It would be instructive for future projects of the same kind to work with the kinds of semantic units that are produced when less powerful (but still statistically significant) connections among lexical items are plotted. Another possibility would be to exclude sequences of immediately adjacent words from the lexical nets so that longer-range connections become more prominent.

Lexical nets are the result of a statistical projection of word distances in a one-dimensional string to a presumed many-dimensional discursive space (as reflected in the collocation matrix) and back to a two-dimensional printed page. For the first projection I used z-scores, while the second projection is done manually. In both cases, other approaches would be worth considering, for example using the mutual information statistic to form the collocation matrix and multi-dimensional scaling to collapse this multidimensional space back to two dimensions.

One of the justifications offered for the use of quantitative indices (Chapter 6) is that they offer a backdrop against which the relative importance of particular discursive features can be assessed, unlike many discourse analytic studies which fail to show how prominent identified features are. Although frequency counts, lexical nets and the like did prove useful for this purpose, it is important to maintain a critical distance from the idea of such contextualisation. The nets did show how certain linguistic features fit into patterns found in the text as a whole, but the choice of what is defined as a 'whole' text remains that of the analyst. Thus in future analyses it may be useful to see if it is possible to draw up a combined net for interview transcripts and case histories together, or perhaps to combine these with similar material from other hospitals, or with standard psychiatric texts. Conversely it may be worth segmenting the texts in terms, for example, of admitting physician, sex, or diagnostic category and constructing separate nets for each cell.

Some would argue that such experimentation would be pointless, as lexical nets and similar devices are simply overly literal interpretations of the idea of repetitive patterns (of which discourse analysts speak so repetitively), displaying as they do connections between mere words, while it seems more probable that God would have stocked our linguistic aquaria with larger, more shadowy creatures (phrases, sentences, natural meaning units), and that it is in the elaborate mating rituals which occur between these higher-level elements that we may witness the spawning of human subjectivity. Lexical nets, one might argue, are too fine-grained and too flimsy to catch anything bigger than the amorphous plankton of the discursive ocean.

However, while it seems evident that no mechanical device could be used to 'understand' all the levels of meaning encoded in language, the analyses in Chapters 7 and 8 suggest that there may be some utility in employing such a device to assist with the initial scanning of texts. Although, in truth, there may be nothing special about the little snatches of repetitive talk and writing identified in Chapters 7 and 8, and any connection between such micro events and larger systems of meaning and power may be entirely coincidental, it is also true, as Zipf (1935) observed that:

In concluding this introduction to a field of possible scientific enquiry, we may well be reminded that the actual speech-gestures, together with their meanings and patterns, are but accidents when compared to the close-knit relationships of the stream of events in the total universe of behavior (biological, psychological, sociological) in which these accidents occur. Yet, in their recurrences, these accidental speech-gestures have found acceptable use by human groups as time-saving representants of the larger universe of experience. A record of the recurrence of these gestures constitutes in fact the chief and almost the only record of human experience available for empirical study (p. 309, original emphasis).


While writing this dissertation I received a copy of a letter from a certified psychiatric patient at one of the facilities catering for such people in the Gauteng area. The letter was addressed to heads of government, legal authorities, international organisations, and similar possessors of sovereign power, asking them to intervene on the patient's behalf as he was being held against his will by callous and uncomprehending psychiatrists despite being of sound mind. As evidence of his sanity, the patient attached a photographic brochure of extremely finely wrought furniture which he had manufactured before being detained. After asking for advice from friends and colleagues I turned the letter over to a group of mental health workers concerned with human rights issues in psychiatry.

They addressed several letters to the superintendent of the hospital and to officials in the department of health, but by the time they received a reply the patient (who had been hospitalised on several previous occasions) had already been released.

It is in any case doubtful if the postmodern alternative (Michael's, 1991, transgression-accelerated turnover-consumption of spectacle axis) doomed to some form of coherence, however much stammered and colourless green ideas sleep furiously, however much pastiche, flagrant plagiarism, as again in Finnegan's case, doubtful if any escape is possible, whether voluntary or o dedi a dada orzoura detention house-arrest in the prison ich bin confus (see).

My wud! The warped floor of the lair and soundconducting walls thereof, to say nothing of the uprights and imposts, were persianly literatured with bursts loveletters, telltale stories, stickyback snaps, alphybettyformed verbiage, ahems and ahahs, imeffible tries at speech unsyllabled, you owe mes, eyoldhyms, fluefoul smut, fallen lucifers, counterfeit franks, best intentions, curried notes, upset latten tintacks, painful digests, once current puns, quoshed quotatoes, messes of mottage, unquestionable issue papers, seedy ejaculations, to which, if one has the stomach to add the breakages, upheavals, distortions, inversions of all this chambermaid music one stands, given a grain of goodwill, a fair chance of actually seeing the whirling dervish, Tumult, son of Thunder, self exiled in upon his ego, a nightlong a shaking betwixtween white or reddr hawrors, noondayterrorised to skin and bone by an ineluctable phantom (may the Shaper have mercy on him!) writing the mystery of himsel in furniture.

To contents page

To Psychiatric Activism Page