To contents page

To Psychiatric Activism Page

Chapter 7

'The unreturning stylus':

Interviews with psychiatric patients

A conversation begins

with a lie. And each

speaker of the so-called common language feels

the ice-floe split, the drift apart

as if powerless, as if up against

a force of nature.

A poem can begin

with a lie. And be torn up.

A conversation has other laws

recharges itself with its own

false energy. Cannot be torn

up. Infiltrates our blood. Repeats itself.

Inscribes with its unreturning stylus

the isolation it denies.

- Adrienne Rich, 1978

In this chapter I present an analysis of interviews conducted in 1993 with a group of patients at a mental hospital in what is now the Gauteng area. The 120-bed hospital, which I shall call Valhalla, was organised into 'bottom' and 'top' wards. The bottom wards were for the more well-defined psychiatric conditions - depression, bipolar disorder, schizophrenia, dementia and mental retardation - while the top wards were for adolescents thought to have behavioural problems, those presenting as anorexics or bulimics, as well as a 'psychotherapeutic' ward for adults thought to have adjustment problems. The patients included in the study were all from this latter ward, commonly referred to as the 'personality disorder' or PD ward. These patients were chosen as subjects for the study not so much by design, but rather as a result of the process of gaining access to the hospital.

Briefly, I was assigned to the hospital as part of the compulsory national service for whites then in operation in South Africa. My appointment was as a clerk, and much of my time was taken up with administrative duties. However, I was also at times, especially in the first two years, given an opportunity to assist psychiatric registrars and psychology interns with their research work, and it was accepted that I could do some work on my own. On this basis I got to be on reasonably friendly terms with some of the key figures in the PD ward - the psychiatrist, the psychiatric registrar, the matron and several of the nurses - and they allowed me access to ward rounds, and (after a formal proposal had been submitted to the medical superintendent) to individual patients.

Valhalla hospital was at the time of the study still almost exclusively for whites, although legal constraints on admitting black patients had been lifted some time previously. In other respects also the hospital's demographics probably closely reflected that of similar hospitals for short-stay non-certified patients in 'western' countries. The mean age for patients in the PD ward was 31.59 (N=258, SD=9.9); and 37.53 (N=587, SD=19.92) for patients in the other wards. In common with most psychiatric institutions internationally, the great majority of patients in both the PD ward (70.3%) and the other wards (63.5%) were female. Almost a third (27%) of admissions were readmissions. The hospital did not cater for certified patients, although patients were occasionally certified and transferred to other institutions (mainly Sterkfontein and Weskoppies) in the area.

Although a state-funded institution, Valhalla was well equipped (pools, tennis courts, a creche, a nurses' residence, sports fields, spacious gardens) and it had an extremely favourable staff to patient ratio. A newspaper article described it as looking "like a five-star hotel with well-groomed gardens and lawns, a nine-hole golf course, two tennis courts and a swimming pool" (Weekly Mail & Guardian, 1997). The hospital officially catered for 120 patients, but in practice this rarely rose above 110. To cater for these patients' needs (as well as those of about 1000 outpatients) there were 30 or more full-time psychiatrists, psychiatric registrars, medical doctors, psychologists, psychology interns, psychiatric social workers and occupational therapists, more than 50 nursing staff, 15 to 20 administrative staff, and 80 'general assistants' (low-paid black workers). This clearly compares very favourably with other institutions in South Africa where there are on average 501 beds, 1 psychiatrist per 133 patients and 1 nurse per 31 patients. On an average weekday only 63% of South African psychiatric units have access to a doctor, 30% to a psychiatrist and 17% to a psychologist. Only 3% of patients have access to individual psychotherapy (Visser, Haasbroek & Bodemer, 1989).

Diagnostic categories at Valhalla hospital are detailed in Table 7.1, which is based on discharge data for an 18-month period. When patients are discharged from Valhalla, as when they are admitted, there is a flurry of official forms to be completed, such as the 'Clearance Certificate', which demands that "before discharge from this hospital, patients will obtain the necessary signatures from the undermentioned departments to the effect that they have been cleared", as well as forms to do with whether towels have been handed in, valuables collected, fees paid and so on. These forms all find their way into the patient file which, a day or so after the patient's departure, migrates back from the ward into the central registry where it awaits a possible readmission. Perhaps the most important part of this discharge process is the final diagnosis determined by the discharging physician. This diagnosis is used for various official purposes, such as for the Department of Health's annual statistics and for medical aid reimbursement purposes.

The information on discharge diagnoses presented in Table 7.1 was collected on a daily basis and reflects about 80% of admissions and discharges during the period. The remaining data were unavailable as a result of my sporadic absences from the discharge office and is unlikely to have resulted in any systematic sampling bias. These diagnoses are based on the anachronistic disease classification system then used by the Department of Health, and are replete with outdated terms such as organic, neurotic and reaction, but nevertheless represent the 'official' diagnosis with which the discharging physician in each case released a patient into the world.

As can be seen from Table 7.1, the most common diagnosis in both the PD and other wards was 'affective psychosis', with fully a third of all patients leaving Valhalla with this label. The next most popular diagnosis in the PD ward was 'adjustment reaction' (25%), while in the other wards it was 'special symptoms not elsewhere classified' (13.8%). The relative blandness of these diagnoses, as compared to the third most popular categories - Schizophrenia (10.7% of discharges from the non-PD wards) and Personality Disorder (9.6% of discharges from the PD ward) - may provide one clue as to why they were so frequently used, and is in itself a commentary on the difference between psychiatry and other medical disciplines.

Table 7.1 Discharge diagnoses of patients at Valhalla hospital


Other wards PD ward

Diagnosis N % N %



Senile organic condition 4 0.6

Alcoholic psychosis 3 0.4 2 0.8

Drug psychosis 9 1.3 1 0.4

Transient organic psychotic conditions 2 0.3

Other organic psychotic conditions 5 0.7


Schizophrenic psychoses 78 10.9 3 1.3

Affective psychoses 278 38.8 81 33.8

Paranoid states 4 0.6 1 0.4

Other non-organic psychoses 1 0.1 1 0.4

Psychosis with childhood origin 1 0.1


Neurotic disorders 6 0.8 6 2.5

Personality disorders 11 1.5 23 9.6

Sexual deviations & disorders 4 1.7

Alcohol dependence syndrome 3 0.4 4 1.7

Drug dependence 7 1.0 1 0.4

Non-dependent abuse of drugs 2 0.3 20 8.3

Physiological malfunction due to mental factors 8 1.1 1 0.4

Special symptoms not elsewhere classified 99 13.8 1 0.4

Adjustment reaction 33 4.6 62 25.8

Non-psychotic mental disorder due brain damage 14 2.0 2 0.8

Depressive disorders not elsewhere classified 68 9.5 6 2.5

Disturbance of conduct not elsewhere classified 14 2.0

Hyperkinetic syndrome of childhood 2 0.3

Specific delays in development 1 0.4


Mild mental retardation 2 0.3 1 0.4

Other specified mental retardation 3 0.4

Unspecified mental retardation 2 0.3


TOTAL 717 100.0 240 100.0


Note: Diagnoses with a frequency of more than 10% are printed in bold. In cases of readmission, the most recent diagnosis is given.

Thus one interpretation of the fact that Personality Disorder was only the third most common diagnosis in the PD ward (well behind 'affective psychosis' and 'adjustment reaction') is that it reflects the invidious position medical personnel are placed in when the vocabulary they have to use reads like a dictionary of insults rather than diagnostic categories. As Kleinman (1988) observes:

Dysthymia will strike many as only a technical euphemism for unhappiness, hysterical personality disorder as a medical shorthand for uncooperation from aggressive or attention-seeking females, who might regard both the term and the doctors who use it as paternalistic and unempathetic (p. 61).

More specific categories of personality disorder have an even less empathetic flavour, the three broadly recognised (e.g. Yates, Seileni, Reich & Brass, 1989) personality clusters being:

I. Paranoid, schizoid, schizotypal;

II. Histrionic, narcissistic, antisocial, borderline; and

III. Avoidant, dependent, compulsive, passive-aggressive.

No wonder that psychiatrists and psychologists, although they make free use of such terms in their research reports and formal communications (in informal settings I as often heard the expression 'vrot(1) personality'), prefer to use less damaging labels such as 'depression' when the diagnosis is meant for more general consumption.

Another more straight-forward interpretation of the relative rarity of 'personality disorder' diagnoses in the PD ward is that it confirms the sentiment repeatedly expressed by the psychiatrist in charge of the ward, namely that the ward was not for cases of Personality Disorder per se, but rather for all kinds of patients who were considered likely to benefit from psychotherapy and from the ward's therapeutic milieu. The evidence regarding diagnosis would thus appear to indicate that the patients in the PD ward were not necessarily considered less seriously ill than those in the bottom wards, but simply as more amenable to psychotherapeutic intervention.

In the next section I describe the subset of PD ward patients I interviewed in more detail, both in terms of diagnosis and of other biographical data.

'Subjects and sampling'

I conducted a total of 66 interviews with a group of 38 patients in the PD ward (38 first interviews, 23 second interviews and five third interviews). Of the 38 patients, 25 (65.8%) were female, which compares well with the 70.3% females already reported for the PD ward as a whole and the 63.5% females for other wards. The mean age of the patients I interviewed was 29.11 (with a range from 18 to 39), which is somewhat lower than the mean age of 31.59 for the ward as a whole and 37.53 for other wards.

Ten of the 38 patients were readmissions to Valhalla at the time that I interviewed them, while a further four of the first admissions were readmitted within four years after the initial admission. Thus despite the relative youthfulness of the sample, 14 (36.8%) have now been readmitted at least once. Two patients had a total of six admissions each to Valhalla and together the 38 patients had accumulated 62 admissions. Several had also spent time in other mental institutions.

Final discharge diagnoses making use of the Department of Health nosology described earlier were available for 24 of the 38 interviewees and are shown in Table 7.2. This can be seen to fairly closely match the overall figures presented in Table 7.1. In both cases Affective Psychosis and Adjustment Reaction are by far the largest categories, while Personality Disorder and Non-dependent abuse of drugs are the only other categories above 5%.

Table 7.2 Discharge diagnoses of interviewees and of the ward as a whole


Diagnosis N % Ward %


Affective psychosis 10 42 34

Adjustment reaction 7 29 26

Non-dependent abuse of drugs 3 13 8

Personality Disorder 2 8 10

Depressive disorders not elsewhere classified 1 4 3

Paranoid states 1 4 0

TOTAL 24 100 81


Note. The last column indicates the rounded percentage (from Table 7.1) for the PD ward as a whole. The percentages for the ward as a whole do not sum to 100 as only those diagnoses presented in the interviewee sample are listed.

The diagnostic labels assigned to patients were not however as unequivocal as these tables may suggest. In addition to diagnoses coded in terms of the official Department of Health nosology, the psychiatrist or registrar would normally also note his or her final diagnosis in an open-ended format in the patient file, thus allowing greater scope for the notorious indeterminacy of psychiatric diagnosis to emerge. The various discharge diagnoses assigned to the 38 interviewees are recorded in Table 7.3. However, even the relatively complex situation depicted in this table does not begin to do justice to the variety of diagnostic labels which may be used before a more definitive picture emerges and the patient is ready for discharge.

Table 7.3 Biographical details and discharge diagnoses of interviewees


ID Gender Age Diagnosis


1 F 27 a, b & c. Bulimia

5 M - Dependent Personality Disorder

8 F 24 Major Depressive Episode

9 F 36 Substance abuse

10 F - Alcohol Dependence

11 F 25 a. Adjustment Disorder (ADJUSTMENT REACTION)

b. Depressed mood/ cannabis abuse

13 F _ a. & b. Borderline Personality Disorder

c. & d. Psycho-active Substance Abuse (AFFECTIVE PSYCHOSIS)

e. Anxiety State f. Deferred

14 M - Alcohol Abuse / Adjustment Disorder


17 F - a. Major Depression Single Episode (ADJUSTMENT REACTION)

b. Major Depression Recurrent / Bulimia (AFFECTIVE PSYCHOSIS)

18 M 23 Adjustment Disorder with Depressed Mood (ADJUSTMENT REACTION)

19 M 25 Atypical Depression / Substance Abuse (AFFECTIVE PSYCHOSIS)

20 F 18 a.. Adjustment Disorder b. Deferred / Bulimia c. Deferred

21 F - Adjustment Disorder / ?Anorexia

22 M 39 a. Dysthymic Disorder (AFFECTIVE PSYCHOSIS) b. Major Depression

23 M 20 Adjustment Disorder

24 F 25 a. Organic Hallucinosis b. Organic Delusional Disorder c. Schizoid Traits

27 F 25 Adjustment Disorder with Depressed Mood

28 F 26 Adjustment Disorder (AFFECTIVE PSYCHOSIS)

34 F - Panic Disorder / Major Depression / Dysthymia (ADJUSTMENT REACTION)

35 M 39 Alcohol Dependence (NON-DEPENDENT ABUSE OF DRUGS)


c. Substance Abuse / Schizophrenia

37 M 30 Paranoia / Delusional Disorder (PARANOID STATES)

38 F 33 Alcohol Abuse / Agoraphobia (NON-DEPENDENT ABUSE OF DRUGS)

39 F 27 a. Borderline Personality Disorder b. Depression

40 M 24 Bipolar Disorder rapid cycling

41 F 21 Borderline Personality Disorder

42 M 18 Substance Abuse / Major Depression

43 F - Adjustment Disorder (ADJUSTMENT REACTION)

44 F 21 a. Major Depression Seasonal (AFFECTIVE PSYCHOSIS)

b. Atypical Depression (AFFECTIVE PSYCHOSIS)

45 F 22 a & b Depression

46 F - a. Dysthymic Disorder b. & c. Major Depression (AFFECTIVE PSYCHOSIS)

d. Dysthymic Disorder (PERSONALITY DISORDER) e. Panic Disorder / Agoraphobia

47 M 35 Substance Abuse

48 F 26 Substance Abuse

49 F 22 a. & b. Dysthymic Disorder (PERSONALITY DISORDER)

51 F 20 a. & b. Substance Abuse (ADJUSTMENT REACTION)

52 F 23 Substance Abuse (AFFECTIVE PSYCHOSIS)

54 M 22 Substance Abuse / Anxiety Disorder (AFFECTIVE PSYCHOSIS)


Note. 1. ID refers to the interview number. 2. Where diagnoses have been labelled a, b, c etc. these indicate readmissions. Department of Health diagnoses are in capitals and brackets. A slash indicates alternative diagnoses

While it may therefore be useful to try and describe the subjects from whom the raw material for analysis was collected in terms of traditional biographical indices such as sex, age and diagnosis, it is also apparent that these matters are themselves part of the swirl of discourse which accompanies each patient in her passage through the institution, and that it is at best difficult (and at worst contradictory) to maintain a distinction between subjects on the one hand and the discourse which produces them as subjects on the other. It is likely that it is considerations such as these which have prompted Potter and Wetherell (1987) and others to question traditional notions of sampling. However, reporting on diagnostic and other details such as those in Table 7.3 does have a certain utility, if only in conveying an impression of how the producers (or conduits) of the discourses to be analysed were themselves positioned in the official discourse of the institution.

In trying to locate the transcripts analysed below, it would certainly have made a difference had the subjects all had iron-clad diagnoses of Personality Disorder, or if all were from one of the bottom wards, just as my own role in co-producing the discourse would have had to be differently regarded if my position in the institution had been different.

In addition to background details on the subjects, information on the conditions within which data was collected is crucial in making sense of any text (Potter & Wetherell, 1987; Van Dijk, 1987a) and this is discussed more fully below.

The interviewing process

I conducted interviews at irregular times when not engaged in other duties at the hospital. Nursing staff had agreed to inform me of all new arrivals in the ward, but in practice this almost never happened. In order to obtain subjects for the study I therefore visited the ward office as often as possible, identified recent admissions (maximum one night on the ward) from the patient register, and then asked a nurse on duty to call the patient for me. While this may appear to be a relatively straight-forward process, patients were often extremely difficult to track down at such short notice, and due to the nature of my position at the hospital I was unable to institute a more regular system. As a consequence roughly every third patient admitted to the ward was interviewed. I am not aware of any systematic bias resulting from this selection process.

Interviews were conducted in an office near the ward and were tape recorded. I informed patients of the scope (two hour-long interviews) and nature of the study ("a project investigating the language people use to describe their difficulties"), emphasising that participation was voluntary and that the ward staff would not be informed of the patient's decision. Two patients declined participation. One (a middle-aged man) had been brought to the hospital against his will and was under the impression that Valhalla was a commitment facility. His expressed intention was to escape, rather than to participate in hospital procedures, and he was some days later recorded as discharged, coded RHT(2). The other (a woman in her late twenties) started crying incessantly as soon as she came into the office and I did not feel able to continue with the interview. She was subsequently subjected to a lengthy series of ECT treatments, the outcome of which was described by ward staff as reasonably favourable.

My impression of the emotional state at the start of the interview of those patients who did participate was that it varied widely. Some seemed melancholy, others withdrawn; the majority seemed in reasonably good spirits at the time of the interview, although many spoke of a more general sense of gloom. One woman was in an overly exuberant mood, and one man seemed at first not to be talking at all coherently. Both subsequently participated in the interview without much difficulty.

Once patients had verbally agreed to being interviewed, they were asked to sign a consent form (Appendix A) which reiterated the points already made as well as giving assurances of confidentiality. I then asked for permission to switch on the tape recorder and start the interview.

Interviews were between 20 and 90 minutes in duration. A predetermined set of topics (derived from a small pilot study involving five unstructured interviews) were covered in a relatively set sequence, but with considerable scope for digression. The strategy was to introduce a topic, trying to get the interviewee to pursue that (using minimal encouragers, reflection, and follow-up questions), and only introducing a new topic when the interviewee had run out of steam. There were also a small number of questions which I put almost verbatim to each interviewee. The ideal interview structure (although no single interview fully achieved this) was as follows:

Reason for admission. I started off by asking each interviewee directly why she was at Valhalla, typically using a phrase such as: "The first thing that I usually ask people is if they could just tell me why they're here." [17/1](3)

Events leading to admission. In answering the previous question, many interviewees alluded to events in the weeks immediately prior to admission. If not, I asked about this, using a formula such as: "If you could maybe just tell me the past two, three weeks - what's been happening, how did it...?" [34/1]

Professional opinions. Interviewees would often mention one or other professional person (such as a GP, psychiatrist or psychologist) in talking about these events. If not, I would try to establish if they had had dealings with such a person and how this person had defined the interviewee's problem. E.g., "What does your GP think about your difficulty, how does he see it?" [10/1]

Lay opinion: Interviewees often dwelt on the process of coercion and persuasion that led up to admission. If they did not raise the topic spontaneously, I would ask directly what a family member or friend made of their situation. E.g. "What about a family member or a friend or something? I'm still kind of fishing for opinions of other people." [24/1]

Defining and managing depression. 'Affective Psychosis' (which in most cases appears to mean depression) was by far the most common diagnosis at Valhalla as a whole, in the PD ward, as well as in the sample of patients interviewed, and most interviewees mentioned depression in talking about the reasons for their admission. In addition, an earlier study involving patients at Valhalla hospital (Strong, 1987) found depression to be the most commonly given reason for admission to the PD ward. I therefore asked them to explain what they meant by the term. In the two instances in which depression was not mentioned by the patient, I introduced the topic by asking if they considered depression to be part of their present difficulties. I asked for the interviewee's definition of depression using a formula such as the following: "I wonder if you could describe depression to me as if I really had no idea [mmm], as you experience it?" [34/1] I also asked a further 3 standard questions relating to managing depression and the purpose of depression. The depression-related questions were not used in the current study.

Short-term prognosis. I closed the interview by arranging for a follow-up interview in two weeks' time and asking interviewees how they saw things changing in that period. E.g. "...if you could perhaps predict now how things will have changed for you between now and then." [43/1]

Two weeks after the first interview, subjects were re-interviewed. In part due to the practical constraints already mentioned, only 23 of the original 38 interviewees could be located for re-interview two weeks later. An additional reason for the attrition was early discharges (which occur in 20% of PD-ward admissions) and patients who are discharged with an RHT code.

The second interview format was very similar to that of the first.

Reason for admission. Interviewees were again asked to explain their admission to Valhalla. E.g. "To start off sort of the same as last time uh if you can this time just give me sort of for the record a summary of why you are here." [23/2]

Last two weeks. Interviewees were again asked to recount events over the past two weeks, but this time with reference to their stay at Valhalla. E.g. "If you can sort of fill me in about the last two weeks since I saw you last, what's been happening around here?" [23/2]

Depression. The same questions regarding depression asked in the first interview were repeated in the second interview.

Prognosis. Interviewees were again asked how they saw things turning out during the next two weeks. E.g. "So if we assume say two weeks from now, uh if you could predict how things will change for you between now and then." [23/2]

Finally, third interviews were conducted shortly before discharge with five individuals. These took the form of informal discussions and were not intended to be used as part of the formal analysis, but to provide background material. Similarly, although this is not an ethnographic study, I also drew on my experiences for background and to illustrate particular points.

The purpose of all interviews was to encourage interviewees to talk as freely as possible about the nature of the problem which brought them to Valhalla while still retaining a measure of comparability across interviews and ensuring that certain themes were covered in each interview. The structure of the interviews was not as obtrusively apparent as the above description perhaps makes it appear. The format is perhaps most similar to that used by Van Dijk (1987a) in his classic discourse analytic studies of racism, i.e., relatively free-ranging one-on-one interviews in which particular topics are either deliberately introduced or allowed to emerge 'spontaneously'. Importantly, I would in retrospect have preferred to conduct more informal interviews with less concern for comparability across 'subjects'. A complete sample interview is reproduced as Appendix 4.

Group interviews, as used in Levett's (1988) discourse analytic study have many advantages and may also in retrospect have been preferable. However, there is a particular reason why individual interviews may have been more appropriate in this case, namely the individual nature of the psychotherapeutic interventions used with these patients. Although there was a fair amount of group work, for instance in occupational therapy sessions and much talk about the beneficial effects of the ward milieu, patients clearly understood that, apart from medication, one-on-one therapy was the centrepiece of their treatment. Many patients came to Valhalla while in psychotherapy, often as long-standing clients of one of the numerous therapists in private practice in the Johannesburg area. Once in the ward they were interviewed at length in the course of obtaining a history, followed by weekly therapy sessions with a psychologist, psychiatrist or psychiatric social worker, and more frequent informal one-on-one chats with a nurse-therapist assigned to their case. Thus the interviews I conducted could to an extent be seen as analogous to the one-on-one therapy sessions in which interviewees were already accustomed to exploring their emotional and relationship difficulties, despite the fact that it was repeatedly emphasised that these interviews were purely for research purposes. These interviews were therefore arguably analogous to the kinds of confessional or therapeutic conversations commonly engaged in by those who have become part of the psychiatric system.


I transcribed the tape recordings as soon as possible after each interview, usually within a week. Following Levett (1988), Potter & Wetherell (1987) and Van Dijk (1987a), the transcription format was relatively straight-forward, registering the speaker (myself or the interviewee), indicating pauses by means of dots, and demarcating inaudible and comment sections with square brackets. More complex notations of paralinguistic features such as elapsed time, inflection, and overlap in speaking turns were not used.

Despite the relative simplicity of the transcription task, this nevertheless proved to be one of the most time-consuming aspects of the study, fully bearing out Potter & Wetherell's (1987) rule of thumb that the ratio of recorded to transcription time can be as much as 1: 10. The sheer labour involved in transcription is a constraint often mentioned by researchers attempting to analyse relatively large samples of verbal material, e.g. Levett (1988) who found transcription time to be 10 to 14 hours per tape. In the next chapter the utility of a data source not dependent on such extensive transcription input is investigated. For the present study I transcribed a total of just less than 100 000 words. The transcribed text was spell-checked, which aided greatly in correcting typos and misspellings and in standardising the spelling of unusual word forms. A word count before and after spell-checking revealed that around 300 pseudo-types (typos etc) were removed in the process.

Contractions such as don't, I've, could've were left as is, and not expanded to their original form. In most cases contractions involved two words only, although at least one colloquial form (dunno) sometimes used in the transcription consists of three words (do not know) in its expanded form. The text was annotated using the COCOA format as described in the previous chapter. Identifiers used were SPEAKER (with values SELF or INTERVIEWEE), INTERVIEWNO (FIRST or SECOND) and TOPIC. Values used for TOPIC related to the different sections of the interview as set out above.

Analytic strategy

The analysis consisted of qualitative discourse analysis, informed by quantitative indices as set out in the previous chapter. The analysis was concerned with highlighting recurrent patterns of talk about mental illness, chronicity and readmission, rather than to look for the causes of readmission in possible individual differences in language use. The purpose of the analysis was also not to highlight psychiatric patients' experiences of serial incarceration, but rather to reveal something of the architecture of the 'prison house of language' within which they were temporarily or permanently resident. Thus no attempt was made to test for significant differences among different diagnostic groups or between readmissions and first-time admissions.

The text analysed consisted, for the most part, of those sections of the transcripts representing the interviewees' speech. My own contribution was only included where it was necessary for making sense of sample extracts. Although the general tendency to exclude the interviewer's speech from discourse analytic studies can be seen as a weakness of the approach, this is mitigated in this case by the fact that the interviewer's contributions have been described in considerable detail above, so that the context within which interviewees spoke is relatively clear.

The analysis is presented in two parts. First, I analyse the entire first interviews, excluding the sections in response to the standard questions on depression. Second, I compare the first and second interviews in terms of the sections that follow after the standard question on why interviewees were at the hospital. This section, which is at the start of each interview and follows a more-or-less standard question from me, most closely approximates responses obtained under controlled experimental conditions and could therefore more legitimately be used for comparison than the rest of the interviews, which often differed markedly from the first to the second occasion(4).

Each of the two parts of the analysis were done in two phases. First, quantitative indexes and lexical nets (as described in Chapter 7) were calculated. Second, these were used as a starting point and backdrop for qualitative analysis. As discussed previously, I have throughout excluded my contribution to the conversation from the quantitative indexes, but in the qualitative analysis I have in places included it as part of illustrative extracts to help provide a context for what is said. I have also, towards the end of the analysis, included a number of extracts from third interviews for illustrative purposes, although these were not part of the main analysis.

Each section of qualitative analysis typically starts with a discussion and interpretation closely associated with a particular quantitative index, e.g., part of a lexical net. The procedure followed in constructing this qualitative analysis was as follows. First, I located all extracts containing the collocational pairs (or longer sequences) contained in the part of the net being analysed. Second, I examined these extracts with a view to understanding the typical functions of the collocational pairs or sequences and to identify types of talk that repeatedly occur in close proximity to the collocations. Third, I selected what appeared to me to be representative extracts and presented these as illustrative material together with the analysis. Finally, I in some cases extended the qualitative analysis to consider thematically, but not necessarily statistically, related issues. No attempt was made to systematically pursue different levels of analysis, such as topics, metaphors, story elements and rhetorical devices separately. Instead, whichever features appeared, from the quantitative data, to be prominent in the text were brought into the discussion.

Preliminary quantitative overview of the first interview corpus

The corpus of first interviews consists of 33 644 tokens, 2 729 types, 12 377 lexical tokens and 2 545 lexical types. The type-token ratio is 1:12.33 and the lexical density 36.79%. The 60 most common types, representing 18 943 tokens or 56% of the corpus, are shown in Table 7.4.

Table 7.4 Sixty most common types in the first interview corpus


I 2227 and 1257 to 988 the 774 a 643 that 623

of 522 it 519 was 514 you 512 uh 503 my 503

know 458 me 439 in 380 just 307 I'm 300 because 286

but 274 it's* 273 they 273 don't 260 ja 244 she 244

like 242 so 224 for 217 is 214 I've 214 not 206

at 202 with 187 very 182 what 175 had 169 think 166

he 165 have 163 on 161 been 158 no 155 well 154

this 154 then 153 about 146 got 141 all 139 do 136

here 134 be 127 get 127 go 119 there 116 really 115

one 113 if 112 out 110 time 109 that's 108 as 107


Note. Frequencies are shown next to each type. * it's = it is (its did not feature among the first 60 types)

As can be seen from Table 7.4, by far the most common type used is the pronoun I, which is almost twice as common as the next most frequent type. In fact, no fewer than 5 of the 60 most common types refer to the first person singular (I, my, me, I'm, I've) and, despite representing only 0.18% of types, together account for 10.95% of tokens. Every tenth word uttered by interviewees directly indicated the first person singular(5). Although this may seem a trivial fact, it underscores the extent to which interviews of this sort constitute what Foucault calls 'confessions' about the self.

Turning to lexical types (Table 7.5), several clusters can be identified. The only types that appear to be directly related to psychological problems are depressed (N=64(6)) and depression (N=43). A separate count revealed that 66% of interviewees gave depression as the primary reason for their admission, with an additional 15% mentioning suicide attempts or suicidal thoughts as the primary reason, giving a total of 81% (as opposed to 42% who at one time or another were assigned a diagnosis of depression or dysthymia). It is perhaps not surprising that interviewees should be unaware or avoid direct mention of the other common diagnoses (alcohol or substance abuse, adjustment disorder and personality disorder(7)). What is remarkable is that none of the many other more moderate 'lay' and professional terms for mental illness figured in the top 60 lexical types. An examination of the full list of lexical types revealed that types such as neurosis, neurotic, psychosis, psychotic and nervous breakdown are all absent. Types such as nerves (N=5), nervous (N-1), anxiety (N=6), and stress-related types (stress N=9; cope N=12; support N=4), are nowhere near as frequent as depression-related types. In these interviews depression appears to have almost completely supplanted other shorthand formulations as an explanation for being in a mental hospital, possibly because it connotes extreme mental anguish while avoiding any implication of insanity, but also because it has considerable medical and scientific legitimacy.

Table 7.5 Sixty most common lexical types


know 458 just 307 think 166 well 154

got 141 get 127 really 115 time 109

things 105 thing 100 people 93 say 88

said 86 now 82 feel 75 see 74

lot 69 come 68 work 65 depressed 64

back 63 actually 62 mean 61 help 59

two 58 sort 55 day 54 last 54

much 52 went 51 take 51 problem 51

always 51 getting 49 way 47 quite 45

even 44 life 44 problems 43 depression 43

years 42 good 41 bit 41 started 40

used 40 Valhalla 39 home 38 thought 37

feeling 37 better 36 again 36 talk 35

whole 35 look 34 three 33 understand 33


Note. Frequencies are shown next to each type.

In addition to depression and other illness-related words, another group of nouns could be read as specific and non-specific indicators of the difficulties patients describe themselves as having to deal with: things, thing, people, work, problem, problems, life, Valhalla and home. Another group of verbs and adverbs appear to refer to epistemological concerns: know, think, really, see, actually, mean, thought, understand and possibly feel and feeling. These will be discussed in the context of the collocational pattern among types in the corpus, reproduced in Appendix 3.

The span of 8 (4 on either side of the target word) used in calculating collocations is the same as that for Figure A2.3 (Appendix 2), as this appeared to represent a fair compromise between more extreme possibilities. This span is likely to highlight typical word pairs and phrases, as well as connections among words that tend to occur in fairly close proximity without being part of the same repetitive phrase. A minimum collocation frequency of 15 was set to ensure that identified collocations would occur on average at least once in every second interview. To limit the complexity of the web, the first 218 collocations to a minimum z-score of 11 were used. The collocational information in Appendix 3 was used to draw up the lexical net (as discussed in Chapter 6) shown in Figure 7.1. Before discussing Figure 7.1, however, a brief section of text is subjected to contextual markup to provide a foretaste of the types of redundancies that are identified by this method.

Getting into the text

Below is a more or less randomly chosen extract from the first interview with participant number 34, a young woman in her first admission to Valhalla:

M: Ja, maybe a bit of that, who knows?

I: I don't know. Because you know I explain it to my doctors but what, the thing is there, we, we get our tablets from Valhalla [ja]. Every time you go you just start getting used to a new doctor, you know doctor, then a new one comes [ja, ja]. And she'll have her [-s], you know her say, and then you get used to her, and then another one will come [ja, ja]. And there's this one who keeps on pressuring me I must go and work, I must go and work. And like last year I got an after-school job, you know looked after children after, in the afternoons [yes]. And these attacks, I started getting these attacks in class, you know [ag]. And at the end of this year, December last year, they said to me look you know they, look you know [inaudible]. Ja. That's embarrassing for me. I'd love to go and work. And I do, do you think I like sitting at home and things like that. It's embarrassing.

A bio-psychiatric interpretation of this woman's difficulties would probably depend on collateral information regarding the nature of the attacks and the treatment received, and if these suggested no organic or major psychiatric problem she would be consigned to the wastebasket of 'personality disorder'; a psychological reading would similarly probably want to place the woman as a histrionic personality with passive-aggressive and dependent traits; while a sociological reading would perhaps try to relate aspects of her form of post-deinstitutionalised existence to classic themes from older forms of institutionalisation (e.g., the ironic story of the long-time inmate having to educate the novice jail-keeper in the ways of the institution). A reading sensitive to issues of power would want to highlight the subtle (although ultimately perhaps self-defeating) ways in which the woman both colludes with and subverts the demands of institutional authority.

All such readings would tend to focus on the semantically rich portions of the account - getting tablets from Valhalla, having to explain things to a succession of new doctors, being pressured into going to work, the embarrassment of an attack, sitting at home - and largely ignore the apparently unremarkable linguistic commonplaces which link the 'purple patches'. It is precisely these normally almost invisible redundancies which are foregrounded when the extract is subjected to contextual mark-up based on the collocational information in Appendix 3:

I don't know it to my doctors butour tablets from valhalla every time you go you just start getting used to a new doctordoctor then a new one comesyou know her say andand then another oneme I mustwork and like last yearknow looked afterin the afternoons andgetting these attacks in class youthe end ofto me look youme i'd love to go< I do do like that it's

The collocational data used for the markup in a sense represent those patterns of speech closest to the surface. It will be shown, however, that these verbal gestures can be used as a good start for understanding what goes on in 'psychological' conversations. Figure 7.1 is a lexical net of the most prominent connections between words in the interviews, incorporating all the collocations listed in Appendix 3.

Despite the tragic circumstances that the interviewees find themselves in, there is something almost comical about the textual universe shown in Figure 7.1, with all the embarrassing little indispensabilities of everyday talk (sort-of, and-then, last-year, that's-why) set out for display. As a (sort-of) sentence generating machine, it can be used to crank out any number of little confessions: My-mother-and-I-don't-want-to-talk, She-told-me-to-say-what-I-don't-understand, And-then-I-would-have-wanted-to-go-home... At the same time as facilitating discourse, however, this 'technology of the self' sends it along certain predetermined paths - misquoting Foucault, these are 'stammered, imperfect words with very fixed syntax'.

Figure 7.1 Lexical net of the first interview corpus

Getting to know you

At the centre of the net in Figure 7.1 is of course the first person singular, with a myriad possibilities radiating from this cardinal point: I know, think, suppose, mean, feel, felt, want, can, can't, didn't, don't, and so forth. However, the most prominent collocate in Appendix 3, you know, does not involve the first person, but the second - which is connected to the first only through the verb know. Some extracts help to give the flavour of the ubiquitous you know sequence:

[17/1](8) so I DIDN'T really THINK YOU KNOW it necessary

[20/1] I DIDN'T WANT to live any more and uhm YOU KNOW AND THEN

[44/1] to discover why YOU KNOW I FEEL the way I do

[48/1] and the post mortem and the YOU KNOW all the other things

[54/1] phoned just at the wrong time YOU KNOW

[48/1] and the police investigation and YOU KNOW

With contextual markup reflected in font size, you know screams out from every page of the transcript, as in the following short extract:


M: I see, so it's all mixed together?

I: yes and even just to be in theknow even just to see what those people go through you know that youthrough such hardrealise that you youryou know

You know, although I have not seen it referred to in the discourse analytic literature, is very common in spoken discourse and has been extensively studied in corpus linguistics. It has, amongst other things, been referred to as a 'verbal filler', 'fumble', 'softening connective', 'cajoler', 'compromiser', 'hedge', 'plea for cooperation' and 'conversational greaser' (Holmes, 1986) and may have any of these functions depending on the context. However, Holmes maintains that you know has a core meaning which leads speakers to choose it over other possible 'fumbles' such as sort of or kind of (present in the bottom left corner of the net). This core is an allusion to the relevant knowledge of the addressee in the context of the utterance - i.e. "you know the kind of thing I mean".

You know may not be more common than usual in the present interviews. Holmes (1986) reports that it accounts for 1.3% of a small (30 000 words) New Zealand corpus of transcribed informal conversation, while it comprises just over 1% of the present corpus.

Irrespective of whether you know has a special meaning in these interviews over and above its usual role in interpersonal interactions, it does serve as a reminder that not only myself as the interviewer, but also interviewees were actively engaged in fabricating a sense of mutual understanding, in charging the conversation with what Adrienne Rich called 'its own false energy', and in fostering the illusion of a 'so-called common language'.

At the same time you know is also a subtle form of resistance to the tyranny of the confessional. While on the one hand it affirms the existence of a shared understanding between speaker and listener and perhaps even implies the listener's privileged claim to understanding ("you know more than I do"); on the other, by allowing for certain things to remain unsaid, you know is also an effective information-withholding device.

Two other collocational pairs at the centre of the lexical net, I don't and I know, which are often part of the longer I don't know sequence, form an apparent mirror image to you know but in some ways may have an almost identical function. Extracts featuring I don't know as a more-or-less overt knowledge-withholding device are listed below:

[1/1] I DON'T I DON'T really KNOW, really

[11/1] I DON'T KNOW. I DON'T KNOW how TO answer that question, it confuses me

[10/1] I DON'T KNOW what you mean


[8/1] YOU KNOW it's very difficult to tell the truth sometimes. I DON'T KNOW.

[19/1] I DON'T KNOW IF I CAN be more detailed than that

[46/1] Anyway I DON'T KNOW

These sorts of utterance could of course have many functions (and perhaps often have them simultaneously) - withholding knowledge, incitement to instruction, stalling. The interview is an intricate dance around the precious commodity of knowledge, which starts with my assuring the interviewee of the confidentiality of the material they are about to divulge and ends with snippets of that material being publicly released (as for example in this dissertation). This trading and withholding of confidences occur in the minutiae of sentence construction, as is visible in the lexical net, as well as on a larger scale with the sharing or otherwise of factual information and self-understanding. In a sense the binary tension between revealing and withholding is what keeps the interview, and perhaps the hospitalisation, going.

Apart from the merely linguistic constraints visible in the lexical net, there are of course many other kinds of rules concerning how the flow of information from private to public should be managed. When these are transgressed the interview grinds to a halt and the therapeutic value of the hospitalisation is brought into jeopardy. One patient, for example, scandalised the therapeutic staff by taking photographs around the ward and insisting on tape-recording her therapy sessions (something habitually done by trainee therapists). When she resisted gentle persuasion to stop these activities, legal sanctions against bringing cameras and tape recorders into state mental institutions were invoked.

Although I consciously tried not to encourage interviewees to reveal sensitive information (partly because I was fearful of upsetting them and partly because I did not want to be seen as usurping the role of therapist), I felt inordinately pleased when such information was volunteered. One young man, for example, signalled early in the first interview that he could potentially make an important confession:


M: Ja [laughs]. Uh, how do you predict will things have changed for you by then?

I: Well, as I'm not here out of my own will, uh I've got to try my best, I can only gain something, I can't lose anything [ja]. So I can only gain something, and go out, and give the world another bash, you know [ja]. See how it goes. Maybe, maybe this place would have helped me, and maybe I don't tell them anything.

M: How do you mean by that?

I: Maybe I don't want to tell them. Maybe, maybe it's too ... personal for me to tell.

M: Right. I don't know to what extent they'll try and dig it out of you.

I: Maybe they haven't got a spade! [laughs]

By the second interview, he seemed much more willing to accept the therapeutic value of confession:


Well I get a lot of things that have been bothering me from maybe when I was in my teen, when I was in my teens, and getting that out, all my worries, and all that, and why I'm really here, because I tried to commit suicide and all that ... And I think [thera] therapists [inaudible] lot of things out that I've wanted to get out for a long time, but I just haven't had the ... right person to talk to about it, in a way, you know ...

When I saw him again shortly before discharge, he finally presented me with the gift of his confidence. I felt both pleased and quite ashamed of having so easily slipped into the role of psychological father confessor:


I: The main thoughts that occupied my mind before trying to commit suicide ... to do with my problems at work and financial problems ... Uh I was still worried about my parents' divorce as well as other things that had been bothering me since my teens ... Mmm ... Like ... What, can I just say something else?

M: Yes.

I: [inaudible] ... Since my teens. What bothered me was ....

M: Do you find it hard to talk about?

I: In a way ja. What bothered me was, my relationship with my friends ... and ... getting girlfriends. Like all my friends had girlfriends and not me [mmm].

Despite resorting to therapeutic tricks such as "do you find it hard to talk about?", I at times thought myself more straight-forward and authentic than the hospital's therapeutic staff, many of whom seemed to have cultivated a certain clinical distance. Of course this show of openness (and sometimes collusion) on my part, could equally be construed as just another technique for eliciting disclosure, different in kind but not in effect from that produced by more accomplished therapeutic agents. The following extract both describes the kind of disclosure elicited by therapists and in itself epitomises the type of collusion I sometimes managed to involve interviewees in:


Uh ... my therapist is this very distant person who tries to outstare me all the time [laughs]. She's totally non-committal. She'd make an excellent politico [M laughs]. She nods knowingly all the time. And she could actually wear a mask or, you know, that type of thing. Come in wearing a, wearing a bloody, you know visor and that type of thing. And then she just nods sagely occasionally and, and, OK it's good if, if I'm really worked up and I can go in there and she asks me how I am and I give her like twenty minutes of how bad I am [laughs] and then tell her well you asked, type of, type of thing ...

Unlike you know, I don't know and similar epistemological devices at the centre of the net (they don't know, I don't understand, they don't understand) in fact explicitly extend the problematic of shared knowledge beyond the dyadic interaction with the interviewer to all such figures of potential medical or psychological authority, at one and the same time acknowledging this authority and signalling a more or less tacit resistance to it:

[49/1] I DON'T KNOW I DON'T FEEL depressed

[8/1] I DON'T KNOW what they're gonna do, I DON'T KNOW.

[11/1] I DON'T necessarily KNOW what they mean but I take it to heart AND I CAN go and sit and cry and get totally depressed about it.

Thus I don't know often functions with they don't know to construct a domain of privacy which both should and should not be penetrated by the gaze of official knowledge. The inner self should be sacrosanct:


I DON'T like people knowing things that I DON'T WANT them TO KNOW. So in other words now, I HAVE nothing secret, nothing. AND I THINK it is very important for a person TO HAVE some secrets. So now I'M GOING TO HAVE TO make up another secret so that THEY DON'T KNOW that.

And however hard they may try to know, they never will understand what is really going on with me:

[17/1] THEY DON'T live, THEY DON'T UNDERSTAND my world. THEY DON'T, THEY DON'T, you see

[22/1] I just feel I'm probably A BIT tired from work, A BIT irritable ...and well I [inaudible]. And thing, people around you DON'T UNDERSTAND [yes], THEY can't cope with it, THEY DON'T, THEY DON'T really WANT TO GET involved.

[39/1] THEY can't UNDERSTAND it. They tell me YOU KNOW, how, but why, YOU KNOW, the same old story [ja] AND THEN my mom's attitude now is just TO GET better.

[45/1] THEY DON'T UNDERSTAND moodiness, or how you can, stupid things can worry you where, where it's nothing of consequence to them.

Where "they" don't, can't or shouldn't know me, "I" don't necessarily understand them or myself either:

[5/1] I DON'T UNDERSTAND things people do and say, why they do them, what happens TO ME, why things happen TO ME [mmm] AND ... I tend TO BE over-sensitive AND I [mmm] take things ... too seriously, I DON'T KNOW.

[8/1] Ja, it's it's I DON'T UNDERSTAND why. That's the only thing I DON'T UNDERSTAND.

[44/1] Uh AS FAR AS I UNDERSTAND it I suffer depression. I'M NOT sure whether it's endogenous or whether it's reactive.

[48/1] I HAVE attempted to UNDERSTAND within myself ...I FEEL very very deeply

Thus what emerges from the epistemological tangle at the centre of the net (you-know -I-don't-know-they-don't-understand-I-understand-they-understand) is the modern (private) self, both colluding in and denying the intersubjective moment of mutual understanding (you-know), both wishing to be and resisting being understood - the sadly misunderstood, uncomprehending subject of truth:

[1/1] I really DON'T KNOW, YOU KNOW. I really DON'T KNOW. I'm sure that's why I'm here. Because I'M NOT sure how to cope.

The blues

Growing under this canopy of general incomprehension are of course numerous more specific ailments, but the only one sufficiently common to be caught up in the lexical net was the sequence I-was-very-depressed(9). This sequence rarely appeared as such, but in many cases the I was pair occurred in relatively close proximity to depressed, depression or very depressed. As already mentioned, depression appears to be the preferred way, in this group of patients, of accounting for being admitted to a mental hospital. Even where other issues such as drug taking are mentioned, they tend to be subsumed under the depression label:

[1/1] uh I WAS drinking far too much, uh I WAS VERY uh ... DEPRESSED

This was also sometimes the case even where the stereotypical I was and very depressed pairs did not occur:

[11/1] I'm here because I'm depressed and I I'VE BEEN taking cocaine.

[36/1] I smoked dope with him [mmm]. AND THEN, I DON'T KNOW, I just, then I started getting depressed again, I'm really depressed.

However, once again a subtext of resistance is audible behind the pat answers regarding depression as the cause for hospital admission, frequently in association with knowledge-withholding devices such as I suppose and I dunno discussed earlier:

[8/1] Uh, well its I DUNNO I I SUPPOSE I FEEL pretty ... depressed

[16/1] Uh, I DON'T really, I SUPPOSE depression.

[34/1] Ag I SUPPOSE I'm suffering from depression or something, I DUNNO, I just DON'T FEEL well.

[49/1] I'M HERE for a depression I THINK. [You think?] Ja [laughs]. I DON'T KNOW. I DON'T FEEL depressed.

To further distance the self from this diagnosis, even while proclaiming it, the attribution of depression is also often imputed to professional others:

[5/1] I DON'T KNOW ...Uh ......I KNOW from WHAT THEY SAY, from what psychiatrists have said TO ME, uh ...THEY just SAY ...uh I suffer from depression ...I DON'T KNOW

[52/1] Uh well I haven't been officially told but I guess for depression.

[8/1] So they think that by coming here I'M GONNA I'M GONNA come to the hospital. They gonna TELL ME that I'm depressed and uh they gonna make me all better. Well that's how they think it's gonna happen [mmm]. YOU KNOW I DON'T THINK THEY quite UNDERSTAND that it's ...I DON'T UNDERSTAND it myself, but ...I DON'T KNOW, JA. I DON'T KNOW what they think. I just ...I just FEEL that they're doing this TO ME. THEY want ME TO be here, it's it's a good thing. I DON'T KNOW, it's strange.

Depression, it would appear, is something which is inherently difficult to recognise and diagnose correctly. Either it is hidden so deeply in the inner core of the subject that others fail to notice it, or professionals flounder around looking for the correct way of diagnosing and treating it:

[11/1] To a person who doesn't understand depression, and never FELT it [yes] - they thought I WAS just plain lazy. And I wasn't - I WAS really feeling depressed [ja] and my job was starting to suffer cause I wasn't [inaudible]. My bosses were starting to get upset. That's basically it.

[39/1] AND THEN I went for blood tests, AND THEN I went to my psychologist. And we were just talking, YOU KNOW. But I still couldn't get out of the depression. I WAS getting deeper into depression [yes]. And out, when people looked at me, I looked fine [ja, ja]. I looked fine.

[46/1] Uh this time I'm here BECAUSE I WAS VERY suicidal and VERY destructive. I used TO GET VERY angry with A LOT OF things. Uh originally I WAS here for a major depression, but now it's a dysthymia something, I'M NOT sure what.

Thus depression, as the major symptom of a misunderstood and impossible to understand patient is itself mysterious:

[17/1] I've obviously been DEPRESSED, VERY DEPRESSED about four months back, really depressed. But I, it's, it's, you see I DON'T UNDERSTAND BECAUSE I. Uh does depression uh encompass the feeling of, obviously the feelings of like [inaudible] and all that, but I actually force my mind not TO WANT TO care, not TO WANT TO think, not TO WANT TO do, not TO WANT TO BE. Is that depression?

[20/1] Ja, uhm, when I WAS, it is funny like when I WAS really depressed, I, I DIDN'T really feel depressed. I WAS running all over the place getting anxious and, and doing crazy things. AND THEN I took anti-depressants AND THEN I felt better. And now, afterwards, I'M just FEELING low. I'M NOT even being crazy, I'm just low [ja].

[44/1] Uh as far as I UNDERSTAND it I suffer depression. I'm not sure whether it's endogenous or whether it's reactive. I THINK it's A BIT OF both [ja], AND I become extremely suicidal. I DON'T HAVE a great love for life as it is [ja]. And that's basically why I'm here. And also to discover why YOU KNOW I FEEL the way I do about MY LIFE and my circumstances.

I was although associated with very depressed is also frequently used in other contexts. Sometimes it is used to introduce long-standing problems ([10/1] it's been going on since I WAS a child) or more immediate precipitating conditions ([52/1] IT WAS JUST I JUST DIDN'T do anything I couldn't work), but most often it signals legitimation of the current admission in terms of previous psychiatric treatment:

[16/1] ja uh I WAS at X-clinic ... I actually overdosed after I WAS AT THE X

[28/1] I WAS in the X-hospital before I came HERE

[36/1] well I I WAS HERE in 1987 and I WAS HERE FOR 6 months

[39/1] uh I WAS HERE August September last year I'VE GOT temporal lobe epilepsy

[43/1] uhm the reason why i'm here WAS I WAS discharged from X-hospital

[46/1]] uh originally I WAS HERE FOR a major depression ... this time I'M HERE BECAUSE I WAS very suicidal

It is interesting that even in the context of a superficial attempt to explain the reasons for admission to a mental hospital, previous admissions to the same or other institutions should play such a conspicuous role. Many of these interviewees seemed clearly embedded in a discourse of perpetual patienthood, as is exemplified in the following extract:


WELL uh I'M VERY DEPRESSED [mmm]. AND it's been GOING ON since I WAS a child [ja] and it's just over the years I'VE GOT to a point now where I MEAN I FEEL I need help and they have TRIED TO treat ME over the years and nothing seems to help [really?]. ja.

In my experience many patients do more than merely drift along in a state of serial chronicity, but often, at the very least, express an awareness of the ironies of their condition. For example, I asked one patient in an informal conversation how long she had been at the hospital during her previous admission. Her response was: "The first time ... two months. The second time ... my whole life." Another patient told me a funny (and perhaps somewhat rehearsed) story about a party for ex-patients that he had attended. At 6 o'clock everybody suddenly became restless and the next moment wherever you looked the little yellow packets in which the hospital's outpatient department dispenses medication appeared.

Mommy daddy me

From the epistemological never-never land of You-know-I-don't-know-they-don't-understand-I-don't-understand-that-I-was-very-depressed-here-before, into which the psychologised self is perpetually both readmitting and discharging herself, it is a relief to turn (reading the lexical net from left to right now) to the more clearly demarcated territory of the family (my-mother-my-brother-my-sister-my-husband-my-family-and-me). As we have been told so often, there is a 'stigma' attached to being admitted to a mental hospital, and in its first incarnation the family indeed appears in these interviews as actually or potentially scandalised, or as uncomprehending and sceptical:

[23/1] Well ... Well the first time MY MOTHER, MY FATHER, and my mother's couple of friends knew about it ... Well most of my mother's friends knew about it, and my father's. But uh the second time I DON'T even THINK they wanted to tell their friends about it

[24/1] So I've told them its something to do with my drugs and they've got TO just GET the drugs back into balance. So I lied TO my grandmother and my aunt. MY MOTHER I DON'T KNOW what she's thinking, she's in X. MY FATHER thinks it's like protective custody.

[34/1] Like MY MOTHER, LIKE YOU KNOW if she comes out and I tell her I'm feeling sick, she says ah just forget about it, YOU KNOW, stop thinking about it, YOU won't feel LIKE it. I said I'M NOT thinking about it [ja]. Because I USED TO BE a very active person.

Not only does the patient present her illness as an embarrassment for the family, but it also as an occasion for expressing concern (echoing recent professional concerns about the 'burden' mentally ill people place on their families) that she may be negatively influencing this otherwise healthy family:

[5/1] I get VERY DEPRESSED and that, and upset [mmm]. And because they love me it affects them to ...a very great extent to ...YOU KNOW they get ...I just affect them A LOT and I'll land up pulling them down with me [mmm] which is not fair [mmm].

[17/1] I've driven MY SISTER crazy, YOU KNOW. Because LIKE, instead of doing something constructive AND active AND positive, I DON'T. I dwell on it, YOU KNOW [ja].

There is also the older theme of the family, not as the victim, but as the cause of the mental illness, creating the conditions for all manner of psychological damage: Bereavement ([11/1] when I WAS little MY BROTHER DIED; [38/1] I WAS crazy about MY FATHER and he died), abuse ([43/1] MY HUSBAND can be very abusive and at times violent with me), family break-up ([23/1] MY MOTHER and FATHER are divorced), and impoverishment ([17/1] MY FATHER and MY MOTHER, YOU KNOW, they have this very poor SORT OF background). This is the Freudian family, generator of traumatic childhood memories (irrespective of the objective significance of what happened) that come flooding back and threaten to engulf the present:

[50/1] AND I CAN remember since I WAS small WE WERE living on a plot uh MY BROTHER actually said TO ME that, I WAS seven or five, I DUNNO I DIDN'T even start school yet, I WAS five, MY BROTHER said TO ME THAT I WAS sleeping with a little kaffir(10) ... I MEAN THAT hurts, YOU DON'T realise ... how much, it's a bottleneck, everything just goes in, and it hurts. It's hurting now, more, than I THINK it's hurt that time [ja].

However, far more prominent in the interviews than any of these stock images is that of the medicalised family:

[34/1] [What's you mother think of the, of the whole ...?] Ag she's also under the doctors [laughs]. She's, also been here. [Is it?] Ja. [You mean she's also got difficulties of her own?] Ja, ja. And MY other SISTER, and MY other SISTER. We all, basically, you know suffer with our nerves [ja]. All of us.

Even if not themselves patients, the families that emerge from these interviews approximate the ideal families of the psychiatric literature to a remarkable extent, working in close partnership with professional caregivers for the benefit of their errant members.

[5/1] Uh, I was just ...well I WAS I WAS staying down in X [mmm] and ...uh ...I DON'T KNOW ... just ...I lost my job and everything went wrong and ... uh my mom found out and she came down to X and fetched me and SAID it ... SHE thinks its time I should do something [mmm]. Between her and MY BROTHER they convinced me well to come back with them AND TO ... try AND GET professional help [mmm] AND ... I agreed TO COME back with them and I ... went to see some people and I'VE BEEN referred here.

[53/1] I THINK MY MOTHER has been helpful. She ... YOU KNOW she just organises things for me, she'll always phone doctors and she had SORT OF spoken to people and she's found out about a certain psychiatrist and got me into hospital and uh ... that KIND OF thing.

Not only do families assist patients to get connected up with the psychiatric system ([8/1] Ja well uh, MY BROTHER phoned the crisis centre [mmm] and they sent us here), they also cooperate closely in the process of diagnosis, acting (although not always successfully) as interpreter between psychiatrist and patient:

[22/1] I KNOW they went to see ... a psychiatrist, MY BROTHER AND MY mom, just TO GET an idea of what it WAS, AND I CAN'T remember what the psychiatrist said. And they're still confused [mmm].

[36/1] No they, they, the psychiatrist at the X-hospital who saw me after I took my first overdose SAID that HE doesn't think, told MY MOTHER THAT he DIDN'T THINK I WAS schizophrenic, AND THAT they had actually made up a wrong diagnosis of me [ja], two years ago.

As before, patients also describe themselves as resisting, often passively, the process of physical and mental incarceration through which the medicalised family tries to guide them:

[17/1] I DIDN'T WANT TO COME HERE [ja]. Uh, when my family, MY BROTHER, tell me [inaudible] you're sick and you need some help [ja]. And didn't believe THAT myself, I WASN'T interested in helping myself AT ALL. So that's why I'm here [laughs]. I'll do whatever I HAVE TO, but I'M NOT REALLY into it.

[24/1] She made me phone MY FATHER. AND uh tell him, YOU KNOW, THAT I should be HERE. I MEAN she was very for me coming, more than I WAS [oh is it].

When cooperation is given, it may be reluctantly, and to achieve extrinsic rather than intrinsic objectives:

[8/1] But uh ...I DUNNO, I DUNNO what they gonna do, I just. Look I'M GONNA give it a bash [ja] and if they, if the therapy starts helping. Because I'VE GOT to I'VE GOT to I'VE GOT to come right, YOU KNOW. I'M GOING overseas so, I'VE GOT TO GET my head together. And you see this is a condition. You see MY MOTHER will not send me overseas unless I get this whole thing done TO ME [oh]. So IF I, IF I pull out of HERE, I won't go.

At other times, the family itself comes to be presented as just another institution in the patient's listless migration from one facility to another:

[36/1] I DUNNO. Unless I CAN live off MY MOTHER for the rest of MY LIFE, but I DON'T THINK I CAN. Or else I could GO TO X-Hospital, and just. But you see I DON'T WANT TO GO TO X-Hospital either [ja]. I DON'T WANT TO do anything.

The final member of this extended family to rise to the level of visibility in the lexical net is of course the psychiatrist (my psychiatrist), and we would expect her to play an even more active role in getting the patient incarcerated than the rest. But in stead we find her arguing, ineffectually, against entering or remaining in the hospital and having her ideas about the appropriate form of institutionalisation thwarted at every turn:

[16/1] And HE SAID OK Monday you must go [i.e., leave the hospital]. And I said no I'M NOT ready, Wednesday. And HE SAID uh OK. Wednesday I said no I'M NOT ready I'll go on Friday. HE SAID Friday that's it, finished. So Friday I went out. Uh, the weekend wasn't very good. An enjoyable Monday. The day was quite a good day. AND THEN the early hours of Tuesday morning, about half past two IN THE morning I took an overdose, about two hundred tablets. AND uh THEN I was in, YOU KNOW I woke up from a coma in X-hospital [ja], pipes coming out of me, all those machines on. AND THEN uh, the psychologist came, or psychiatrist, and SHE SAID uh Valhalla.

[19/1] I raised the idea of Valhalla in therapy. Over the last TWO WEEKS I raised WHAT my state was AT THE MOMENT, and if he thought IT WAS a good idea. And ultimately HE SAID, yes, but uh ...uh CAN I express it, only in response to explicit questions.

[21/1] [What's the psychiatrist think of coming here?] He's very opposed to institutions.

[39/1] [So the work felt it would be a good thing to go to Valhalla?] Ja, MY PSYCHIATRIST didn't. [Is it?] Ja, because she thought I WAS doing fine. She sees me, but then the work see me every day.

The oracle speaks

With one exception, I found it impossible to ascribe more than syntactic functions to the remaining branches of the lexical net, i.e., the they-say-what-do sequence, the want-to and wanted-to sequences, the have-been sequence, the I'm-scared and I'm-going sequences, and the independent sub-net around the type-of-the sequence. The exception is the he-said-she-said-tell-me-told-me pattern near the top left of the net. My expectation was that this would typically involve blow-by-blow accounts of arguments, and in a few cases this was so:

[11/1] AND I started accusing him of being with someone [mmm], of which the maid THE next DAY SAID that HE was. And HE SAID HE wasn't and that was basically the final straw.

Much more commonly, he-said-she-said-told-me-tell-me signalled a pronouncement by a professional or lay counsellor regarding the true nature of the patient's illness, often followed by a ritual endorsement by the patient.

[11/1] No, she, I told her about MY LIFE and SHE SAID could understand WHY I WAS DEPRESSED BECAUSE I HAD just come out OF A relationship where I HAD, where he, I found him in bed with somebody. And I freaked out because I was rejected [mmm].

[17/1] Uh, HE SORT OF SAID that uh ...Well HE, HE, HE SAID THAT I obviously haven't made any improvement or any, I've just got worse and worse. And uh thought that my environment has been ... constantly [inaudible] and reinforces the whole negativity in me [mmm]. And the fact THAT I WANT TO stay alone and keep myself isolated is not good. AND I I know he's right.

The ideal form of the oracular he-said-she-said pronouncement appears to be that of a deep insight (not necessarily by a professional) into the workings of the patient's mind:

[11/1] Ja, but this has been GOING ON like AS FAR AS I CAN remember. I remember an instance WHEN I WAS twelve years old AND ice skating AND I bought somebody some lip gloss. And SHE SAID you don't HAVE TO buy my friendship, YOU KNOW [mmm]. THAT I remember, THAT I remember as clear as daylight.

[28/1] YOU KNOW HE SAID TO me I KNOW exactly what's going through your mind, it's a place for the mad.

The pronouncement following he-said-she-said-tell-me-told-me, when it involves suggestions regarding treatment, seems almost invariably to elicit assent, but of a rather passive or even lukewarm variety:

[8/1] She TOLD ME that she's uh helped me enough now and I'VE GOT TO GET MY LIFE together, YOU KNOW. So I said cool, that was it.

[16/1] And SHE SAID no SHE recommended Valhalla, so I said OK, anything that'll make me better.

[34/1] Well SHE SAID they'll send me off for a few days TO GET treatment, and they at least can put me on the proper treatment [ja], YOU KNOW. Because IT'S getting worse, I DUNNO.

[39/1] So down I WANTED TO commit suicide, which I HAD a blade, and I WAS contemplating. So I went AND told MY PSYCHIATRIST [ja], and uh ... SHE SAID I must hand the blade in, which I DIDN'T. I handed the blade in here, in case I might need it, YOU KNOW.

The help-me sequence, which is part of the same cluster on the net as he-said-she-said-tell-me-told-me is thematically also strongly related, often taking the form of a formulaic expression of hope that hospitalisation may be of some use:

[01/1] [Is there anything good about being here?] Like I said, I SUPPOSE Valhalla being better equipped to HELP ME.

[02/1] They're professional people and they ... can HELP ME more than what an outsider can.

[04/1] They couldn't get.. uh.. what is really my, my illness [mmm] and uh ... that is why she had to TELL ME that day they they will bring me here at Valhalla [mmm] since I heard Valhalla is maybe more advanced in medication [mmm] to HELP ME with uh I [inaudible]

As with other psychological and discursive traps of this sort, interviewees were not necessarily unaware of the ambivalence of the position they were being spoken into. One interviewee, for example, spoke at length of how his mother would always help him ([05/1] She she also took ME to people for HELP, she never gave up ... right until she died she ... always TRIED TO get HELP for ME, YOU KNOW), while at the same time suggesting that:

[05/1] Instead of helping us, she's always DONE everything FOR us [mmm], instead of helping us with it [mmm], YOU KNOW. IF I went to her with HELP, instead of helping ME, she would do it, and consequently I ... It comes to me going out on my own now, and I DON'T KNOW how to do it [mmm].

He also spoke wryly of how the same pattern would be repeated with a friend:

[05/1] He would help me a lot ... You know where my mother left off he just carried on [mmm]. I told him not to. I said rather help me with something [ja], don't do things for me [ja, ja].

At other times, interviewees complained bitterly that the form in which the help was being offered was not sufficiently directive:

[02/1] THEY DON'T say much, YOU KNOW, THEY just ... THEY DON'T TELL ME what to do, which doesn't help much. I DON'T KNOW what to do myself.

[08/1] She [a previous therapist] was, she was OK, but she sort of betrayed my trust, which I won't go into. [Did she go back to the family?] She SORT OF she she helped me for like four years and then she didn't she like stopped helping me. She TOLD ME that she's uh helped me enough now and I've got to get my life together, YOU KNOW. So I said cool, that was it.

We are back, therefore, in the land of the singular individual, who cooperates, does not cooperate, half-heartedly cooperates, with sovereign power; the private self whose inner being may be glimpsed only in moments of intersubjective understanding and oracular revelation.

[5/1] I DON'T KNOW what's wrong with myself. I'm just hoping they can TELL ME [mmm]. And I'm just hoping they can change it or TELL ME why I'm doing, what's wrong [mmm].

[54/1] And that's everybody, that's absolutely everybody has TOLD ME that, YOU KNOW you've been through therapy because your face is wet [laughs]. So I'm A BIT apprehensive about that.

Occasionally the ideal of the intense intersubjective moment is realised, when patient, therapist and official transcript (each in their allotted place in the hierarchy) all converge on the same truth, and the self is finally realised as a mobile form of chronicity:


Well yesterday morning he spoke TO ME from about a quarter to 10 to after 11 o'clock. Obviously he'd had my file from the other hospital [ja], but he'd read it thoroughly, because he was asking me questions, but HE could TELL ME the months and the years. Where I could maybe just say the year, and he could say yes THAT WAS October or September or whenever. So, I felt as if he ... he knew what was GOING ON. AND THEN immediately he pinpointed the three areas where I HAD HAD problems ... Uhm so I FELT that he was very aware of what was GOING ON.

Before and after

Admission to a mental hospital, like so many rituals of modern life, is an overtly interventionist step. Although the aims of hospitalisation are now often more modest than they once were - typically to stabilise medication, forestall suicide, or offer respite from unbearable circumstances rather than to embark on intensive therapy - the patient is nevertheless supposed to emerge a different, and hopefully a better, person. In getting a proposal to do this research project accepted by the hospital's management committee, I therefore found myself using phrases such as the "influence of hospitalisation on patients' speech patterns" and "the way in which hospitalisation affects patients' use of language". It is difficult not to slip into causal and uni-directional formulations of this sort when dealing with diachronic material, which may be one reason why non-positivist approaches such as discourse analysis are (with the notable exception of Foucaultian analysis) so rarely used to analyse changes over time (Levine, 1996).

Although the comparison between the way patients talk about the reasons for their admission shortly after being admitted, versus two weeks later, was meant to show how discursive patterns may have changed in that time, one should be careful not to ascribe this simply to the 'influence' of the hospital environment. The change or lack of change is also a function of the analytic technique of collocational analysis. While patients may have undergone profound changes at other levels, the stock collocations they draw on may in many cases be typical of those that occur in all forms of informal talk and may therefore be impervious to rapid change; conversely, the technique may be particularly susceptible to rapid oscillations and any differences found may be unrelated to the two weeks of hospitalisation.

More important, however, than uncertainty about the stability of the technique used, is the probability that patients do not merely passively absorb influences and then betray the signs of these in their speech, but actively negotiate particular discursive positions to contend with changed circumstances at the beginning and some way into hospitalisation.

At the immediate level of their relationship with me, for example, the ostensibly standard stimulus to which interviewees were asked to respond (a request to explain why they were at the hospital) had a very different meaning on the first and second occasions. In both cases it was a fairly disrespectful question, although I of course attempted to soften this through my body language and the way in which I introduced the question. However, the question was disrespectful for different reasons on the two occasions. On the first occasion it was an intrusive question by a stranger relying on his position of relative power and expertise to oblige the interviewee to construct some form of reasonable response. On the second occasion the question was posed by someone with whom a degree of trust had been established, and who was thus even more strongly positioned to expect an answer. This was again demeaning to the interviewee - not so much because it was an overly intrusive question, but because asking a question that had already been answered implied that the interviewee was an experimental subject whose responses would not be taken at face value.

I therefore tried to bear in mind the pitfalls of an overly simplistic stimulus-response interpretation in comparing the two sets of responses, but will not discuss this in detail again, and in what follows apparently causative and uni-directional remarks should be read as being made in the context of the above caveats.

Preliminary quantitative overview of the reason for admission corpora

The corpora(11) used for the before-after comparison were much smaller than the first interview corpus thus far analysed, namely 1252 tokens and 1756 for the first and second corpora respectively. This is only 3.7% and 5.2% the size of the first interview corpus. The number of types used is also much smaller - 394 and 459 for the first and second corpora respectively, which is 14.4% and 16.8% of the first interview corpus. The type-token ratios are far larger - 1:3.17 and 1:3.83, as opposed to 1:12.33 for the first interview corpus - indicating greater vocabulary richness (i.e., more different words per number of running words). However, as discussed in Chapter 6, higher type-token ratios can also be a consequence of smaller corpus sizes.

The 20 most common types in the three corpora are shown in Table 7.6. As can be seen from the table, there is considerable stability among very high-frequency types such as I, and, to and the, irrespective of whether the whole interview or only the part dealing with reason for admission, and whether the first or second interview is considered. Of the 20 most common types in the first interview corpus only two do not occur among the first 30 types for the reason for admission corpora. This is an indication that much of the basic materials used to construct meaning in these interviews remain relatively invariant across different situations.

Table 7.6 Rank order of common types in the first interview as a whole and in sections of the first and second interviews dealing with reason for admission


Whole Interview Reason for Admission

_____________ _________________

First Interview First Interview Second Interview

_____________ _________________________________

1. I 1 1

2. and 2 2

3. to 3 4

4. the 5 5

5. a 6 8

6. that 9 10

7. of 11 9

8. it 10 7

9. was 7 3

10. you 22 30

11. uh 4 6

12. my 14 13

13. know 24 15

14. me 28 12

15. in 18 29

16. just 21 14

17. I'm 8 19

18. because 19 17

19. but 26 31

20. it's 16 33


Note. Numbers are ranked positions, with 1 being the most common type, 2 the second most common and so on.

Lexical types in the first reason for admission corpus (Table 7.7) also closely resemble those found in the larger first interview corpus (Table 7.4) of which it is a part, with similar clusters of depression-related types (depression, depressed, suicide, commit), diffuse problem-related types (things, people, life, drugs) and epistemological types (know, think, suppose, mean, basically). Because of the smaller corpus size there are also a number of unusual types (such as poison) that may have occurred in one or two interviews only. In common with the larger first interview corpus of which it is a subset, types clearly related to anxiety or stress (with the possible exception of handle) is absent from the set of most frequent types.

Table 7.7 Forty most common lexical types in the first reason for admission corpus


depression 12 just 10 know 9 well 9

things 8 get 7 depressed 6 think 6

feel 5 time 5 people 5 life 5

find 4 sure 4 suppose 4 first 4

lot 4 scared 4 better 4 problem 4

said 3 take 3 way 3 bit 3

suicide 3 drugs 3 handle 3 search 3

mean 3 rationalise 3 tried 3 basically 3

took 3 got 3 commit 2 therapy 2

away 2 poison 2 come 2 Valhalla 2


Note Numbers indicate the frequency with which each type occurs.

The situation with regard to the second reason for admission corpus is somewhat different (Table 7.8). Although depression-related types (depression, depressed, suicide) are still very prominent, other forms of illness-talk is also evident, notably anxiety and a number of stress-related types (cope, coping, and possibly pressure), suggesting that depression may be rivalled by other discursive formulations after two weeks of hospitalisation. This is discussed in greater detail below. Perhaps not surprisingly, there is also evidence of more treatment-related talk (therapy, medication, treatment) as interviewees began to settle into the hospital routine. Epistemological types (know, dunno, basically, actually, really) remain prominent.

Table 7.8 Forty most common lexical types in the second reason for admission corpus


just 19 know 18 time 17 got 11

depression 11 again 9 lot 8 therapy 8

working 7 get 7 depressed 6 sort 6

medication 6 life 6 said 6 well 6

better 5 quite 5 pressure 5 work 5

treatment 5 back 5 dunno 5 last 5

cope 5 sent 4 basically 4 actually 4

suicide 4 anxiety 4 took 4 now 4

big 4 too 4 coping 4 really 4


Note Numbers indicate the frequency with which each type occurs.

As a further comparison between the first and second reason for admission corpora, the collocational information in Appendix 5 was used to draw up the lexical nets in Figures 7.2 and 7.3. From Appendix 5 and Figures 7.2 and 7.3 it is again clear that there is a considerable degree of robustness in the collocational patterns found, with many of those identified in the first interview corpus again evident in the first and second reason for admission corpora. Examples include the ubiquitous you-know and I-don't-know sequences (with some breaks in the link, possibly due to the smaller corpus sizes), the I'm-here-because sequence and the proliferation of links fanning out from the cardinal point I. There are also various unique features to each net. Due to the small corpus sizes, several of these proved, on investigation, to come from a single interview or from a small number of interviews. Examples include the it's-the-search-for pattern at the top of Figure 7.2 which comes from a single interview in which the interviewee repeated a phrase containing these words many times over ([48/1]), and the reference to medication in Figure 7.3, which comes from two interviews ([16/2] and [44/2]) in which the interviewees discussed issues of medication at length when asked about the reason for their being at Valhalla.

Figure 7.2 Lexical net of the first reason for admission corpus using a span of 8 and minimum collocation frequency of 5

Figure 7.3 Lexical net of the second reason for admission corpus using a span of 8 and minimum collocation frequency of 5

Apart from such spurious variations, a number of differences between the first and second reason for admission corpora were identified which appear to be more substantial. These are discussed below.

Being there

One of the patterns in Figure 7.2 which is not present in Figure 7.3 is the I-was-there-and sequence at the bottom left of the figure. Almost without exception this sequence is employed in the first reason for admission corpus to talk about admissions immediately preceding the current one, often comparing the previous admission favourably or unfavourably with the current admission:

[16/1] DEPRESSION - UH, feeling down, I had an overdose [is it?]. Ja, UH, I WAS at X-Clinic and uh there they just [inaudible] you with the drugs [ja]. AND uh, THERE WAS no therapy [ja], AND I actually overdosed after I WAS at the X-Clinic [really]. AND [inaudible] said, no, the drugs aren't gonna do you any good.

[28/1] I WAS in the X-hospital before I came HERE, but it definitely wasn't the place for me TO BE, YOU KNOW. Because the patients there are different cases completely.

[47/1] UH I'M HERE BECAUSE I first went to X Lodge, which is a centre for people that abuse alcohol and drugs, AND I WAS unsuccessful THERE BECAUSE I TRIED TO commit suicide and they decided [inaudible] to send me to Valhalla, which I sincerely doubt WAS, WAS the right alternative.

Only a single case could be found in the second reason for admission corpus which made use of the I-was-there pattern, again to talk about an admission immediately preceding the current one:

[44/2] I WAS encouraged TO GO AND see a psychiatrist, AND I WAS put on medication, which worked initially, AND THEN it petered, well I DUNNO what happened, but I got depressed again AND I attempted suicide, AND I WAS admitted to the Gen, AND I WAS THERE. AND after discharge I WAS on other medication AND I GOT DEPRESSED on that as well AND THEN I came here ...

That talk about institutions from which they had just been discharged should figure more prominently immediately on admission than two weeks later is perhaps understandable in that interviewees were still very much in the midst of the decision to be transferred to Valhalla. By the second interview they would perhaps have become more preoccupied with issues around the current hospitalisation rather than the series of events leading up to it. There is some evidence, however, that interviewees were still equally concerned with pre-admission events, but were talking about them differently. This is discussed below.

Working hard; stressed out

Where a typical pattern in the first interviews when responding to a question about the reason for admission is to invoke the I-was-there sequence in relation to immediately preceding admissions, by the second interviews, a broader perspective is attained by invoking the I-was-working sequence to talk about work pressures preceding admission. This sequence is quite common in the second reason for admission corpus, but completely absent from the first. Some examples:

[11/2] Ja, BECAUSE I couldn't handle my life, IT WAS a bit TOO heavy, I WAS WORKING, I WAS WORKING, I used to work in the X industry. And my workload WAS TOO much, AND in order to keep up WITH that I used to take cocaine - but it didn't, I in fact slowed down, to such an extent that I almost stopped, totally. That's it.

[23/2] OK, I tried to commit suicide for my second time, and mostly because of work, pressure of, BECAUSE I WAS WORKING with a guy I didn't really like, he was giving me A LOT OF hassles and that

[54/2] Oh ja, I'M HERE as the result OF A suicide attempt ... Uh ... ja ... under, under, I'VE BEEN under A LOT OF pressure. I'VE BEEN working under a LOT OF pressure for, for quite a while. Uh ... AND I, I, while I WAS WORKING I had been SORT OF very bad, severely mugged, AND I felt that my nerve had sort of snapped.

What is being drawn on here, of course, is what may be termed stress discourse, in which pressures are presented as building up inexorably until something snaps. Talk about stress is not completely absent from the first reason for admission corpus, twice surfacing in the vicinity of the I-was-there pattern (which, as discussed above, more commonly signals talk about preceding admissions):

[24/1] I'M not quite sure. I SUPPOSE the main thing IS THAT everyone, all my support was going away, and THERE WAS no way that, I WAS not coping. I WAS also feeling very depressed.

[52/1] THERE WAS just ... I just DIDN'T do anything. I couldn't work. I de-registered for my course at varsity ...

However, the prototypical stress-related type, cope, is completely absent, while similar types such as coping (N=1) and stress (N=2) are very rare. By contrast there appears to have been an infusion of stress-talk between the first and second interviews. Interviewees still refer to depression to explain why they have been admitted, but now very frequently do so in conjunction with stress, making use of the I-wasn't and to-cope-with sequences at the bottom right of Figure 7.3:

[16/2] Uh to try and ... [interruption at door]. TO GET TO COPE WITH the DEPRESSION, or TO UH, or TO maybe find ways of not having it at all.

[08/2] Uh well, as I SORT OF mentioned before IT WAS DEPRESSION, UH YOU KNOW I JUST lost control, ja. I DON'T want to go into detail, it's JUST repeating everything I said, BUT UH, BECAUSE I feel a lot better about a LOT OF things, BUT UH ja IT WAS just VERY severe depression, just losing control OF MY OF MY MY entire thoughts, ja, and my ability TO COPE in life.

[19/2] UH ... well, DEPRESSION UH ... inability TO COPE WITH therapy AND conduct a day-to-day life, uh periodic SORT OF ... intermittent periods OF of intensity. Uh, a desire to SORT OF engage in A process of therapy full time [mmm]. [inaudible]

[46/2] Uh ... THIS TIME I'm at Valhalla, it's the fifth time I'VE BEEN HERE, is BECAUSE I WAS VERY suicidal AND I WASN'T coping with anything at home. Everything WAS like TOO big AND too, I DUNNO, everything was just beyond me, AND I WASN'T coping WITH anything ... Uh so my psychologist decided IT WAS time for me to come back again ... Uh I DUNNO.

[28/2] UH, a bit of DEPRESSION, the majority of it stress. After the car accident - final touch to basically five years OF A build-up: separation, divorce ...uh pressure from family side ... So I DUNNO ...I, I DUNNO, I REALLY DON'T KNOW. I'm going have TO just COPE WITH it, that's all.

It is remarkable that stress discourse as described by Young (1980) and Pollock (1988) could appear with such vigour where it had been almost completely absent two weeks previously, and doubly remarkable that it should almost invariably be used in the same breath as talk about depression. Where patients arrived at Valhalla talking about depression and suicide when asked about their reason for being there, two weeks later this talk had become hybridised to include coping with the pressures of life, therapy and depression itself.

Some of the dynamics which may have contributed to this process are the following: Firstly, stress as an explanatory construct and stress management techniques were explicitly taught in groups run by occupational therapists and social workers at the time of the study, and several patients are likely to have picked up the jargon from this source. Secondly therapy, although very highly valued in the ward, was (often without this being made explicit to patients) of a 'supportive' nature, and warnings were continually issued to inexperienced therapists to avoid doing any 'uncovering' work in the hospital milieu (the phrase often used was to 'keep the lid on'). This kind of support is of course in many ways seen as a temporary replacement for the 'social support' which the patient is presumed to lack - and from social support it is a short discursive journey to the idea of stress. Thirdly, at the root of the official censure of 'uncovering' therapy is a fear that patients may 'disintegrate' and become unsuitable for early discharge. Much of the acrimony between patients and staff, and among staff, thus centred around the issue of readiness firstly for weekend leave and then for discharge. Is the patient still too fragile to cope with being out of the ward? Will there be enough social support at home? These were the questions which concerned staff on a daily basis, and which inevitably impacted on patients' perceptions of the nature of their difficulties.

I find it impossible to say whether sinking into depression or drowning in stress is preferable, nor even if rapid discursive shifts from the one predicament to the other is harmful or beneficial. It is likely, however, that both kinds of discourse fulfil similar functions and are merely inflections of a more fundamental form of speech which invokes causal and psychological language to both give an account of, and resist (at many different levels simultaneously) being held accountable for, itself:


M: Uh first thing is if you could just sort of for the record again summarise why you're here.

I: Why I WAS here? [laughs] O God.

M: Can't remember?

I: No! [laughs] Ja, I DON'T KNOW. I SUPPOSE I JUST got sick. My nerves cracked in or something [both laugh]. Whatever. I DON'T KNOW. Pressure, I DON'T KNOW. Something like that.

M: OK. I can't remember what you said last time.

I: I know [laughs]. My doctor sent me here.

Breaking free

If, as Rose (1990) and others have argued, the purpose of psychotherapy and other forms of moral orthopaedics is to "restore to individuals the capacity to function as autonomous beings in the contractual society of the self" (p. 227-228), it is clear why an individual such as interviewee number 34 quoted above should be thought to be in need of help. Even after two weeks of psychiatric hospitalisation the patients I spoke to continued to resist imperatives to be accountable, both in the immediacies of an interview with me and (it would appear) in the larger context to which the interview referred. Had I again conducted interviews with the same patients shortly before discharge, a more purposeful and optimistic discourse may have started to become evident. In each of the five cases where I did have an opportunity to speak to patients a third time, there seemed to be an expression of greater optimism, willingness to take responsibility and an eagerness, as Rose (1990) puts it, to "make a project of our biography [and] narrativize our lives in a vocabulary of interiority" (p. 254):


M: Could you maybe expand a bit on that. You know I'm not quite sure how it helped you.

I: Ja [inaudible] actually a lot better, about myself uh, in certain situations. I mean take for example over the weekends [ja]. I mean I hadn't spoken to [my sister] for a long time. And uh, I just, I feel a lot better, and uh I, I can actually cope with most of the problems that I come up with [yes]. Maybe now and again I'll feel a little bit - stressed about it, but it doesn't get into a serious depression [ja].

However, even as these individuals took on the shackles of freedom, they maintained a connection to the institution:


Ja. The therapy at Valhalla has been useful, especially with the nurse therapist who is my age and has been through similar experiences. I think it's basically up to you to do something about it. It's true, I mean most of us cause it ourselves. Uh at the moment I'm on an aggression management course learning to deal with problems as soon as I can instead of bottling it up. I hope to leave here with a more positive outlook and able to deal with the emotions I'm going through ... I expect to start working from here and to move into my own flat.

Whether it is partial hospitalisation, attendance at an outpatients' clinic, rehospitalisation, seeing a private therapist, or simply keeping an eye on one's self, the former psychiatric patient, like the rest of us, is well advised to work at being a better person, while always maintaining contact with the therapeutic system in case of failure:


M: OK uh is there anything more that you want to say sort of in ... making things clear to me?

P: Uh I think it's very important to keep having therapy, and keep in contact with, with people who know about it and can help you [ja]. Because, a lot has to do with your attitude [ja]. And, and even when you, when you're better you, you've got to still work at it. I think you've got to try harder than most people to, to keep yourself up.

M: Ja, so it can easily happen that you just let it go if you don't have therapy and keep contact.

I: Ja, you've always got to try and be optimistic uh and uh keep a positive attitude [ja].

As Rose (1990) puts it:

It [psychotherapeutics] promises to make it possible for as all to make a project of our biography, create a style for our lives, shape our everyday existence in terms of an ethic of autonomy. Yet the norm of autonomy secretes, as its inevitable accompaniment, a constant and intense self-scrutiny, a continual evaluation of our personal experiences, emotions, and feelings in relation to images of satisfaction, the necessity to narrativize our lives in a vocabulary of interiority. The self that is liberated is obliged to live its life tied to the project of its own identity (p. 254).

Conclusion: No looking back

To claim, as I have done in this chapter, that psychiatric patients use particular forms of speech recurrently is at once to invite comparison with the language-use of 'ordinary' people, and from there to ask whether psychiatric patients' peculiar speech is somehow implicated in causing or maintaining their recidivist tendencies. Following such a line of enquiry one might argue that the analysis demonstrates (although tentatively, as there was no non-psychiatric control group) that patients appear to be at the mercy of their language in the same way as they are at the mercy of their illness, being sucked, again and again, into self-defeating linguistic patterns just as they are sucked into repeated hospitalisation.

However, language-use of necessity entails drawing on a system of conventions such as a common lexicon and syntax, idiomatic phrases, stock images, culturally recognised storylines, scripts, discourses, and so on, so that psychiatric patients can hardly be thought of as unique in this regard. One would have to prove that psychiatric patients, and particularly recidivist patients, are particularly stereotypical (or perhaps erratic) in how they apply the building blocks of language - or, more crudely, that they tend to use different building blocks (a peculiarly biased vocabulary, aberrant semantic forms) entirely. One may even propose new kinds of therapy in which individuals are supported and accompanied as they become differentially subjectified by psychiatric discourses, moving perhaps from robotic repetition to 'ownership' and elaboration, and finally to reflection and deconstruction.

This kind of thinking is seductive, but tangential to the purposes of this dissertation. The challenge is not to discover variables, however subtly-defined, that differentiate recidivists from non-recidivists - the sane from the mad - but briefly to disrupt the circuitry of power and knowledge of which this text is one relay, and to do it in such a way that the fluencies of what goes without saying take on a more staccato and machine-like quality. To present what passed between myself and the people I interviewed as part of a strictly regulated production line is not, however, to imply either that we were coerced by the power of psychiatry or that it would be possible to draw up a blueprint of the truth-factory in which we were labouring. Power and knowledge do not radiate formally from the centre, but seeps back from the extremities of social interaction - a physical examination, a psychometric test, an interview.

At the furthest end-points of these capillaries of power are perhaps those moments when linguistic redundancies such as those I have recorded here are passed between individuals to bring about the appearance of a common understanding. As Foucault (1967) recognised, there is no escape from these commonplaces of expression, as we now no longer have access to "all those stammered, imperfect words without fixed syntax in which the exchange between madness and reason was made" (p. xii-xiii). What is more, these micro regimes of true discourse are continually subsumed, as in Whorf's vision, into ever larger patterns of social intercourse, so that all of the social world of mental illness may at times be reflected in everyday phrases such as you-know, I-suppose, I-was-there. Just as it is no longer useful to speak of being outside the institution, there is also now no outside to the language of psychiatry.

As Sophie, "a 27-year-old former office secretary" six weeks into her first admission at Valhalla, recently confided to a journalist:

I believe there's a light at the end of the tunnel ... I know I have a very low self- esteem and lack of confidence. The point of me being here is to get better so that I never come back (Weekly Mail & Guardian, 1997).

1. 1 Afrikaans for rotten.

2. 2 Refused Hospital Treatment

3. 3 The numbers in square brackets identify the interview from which the extract was taken. The code /1 indicates that it was a first interview.

4. 4 It could be argued, however, that in limiting the comparison to this standard part of the interview I have been unduly influenced by ideas of 'spontaneous' versus induced speech. This point is taken up again later in the chapter.

5. 5 Of course words like you and one are also often used to indicate the first person singular.

6. 6 Unless explicitly stated otherwise, all frequencies in this chapter refer to number of types or tokens, not number of subjects.

7. 7 Personality appears 7 times, although inspection of the transcripts revealed that 2 of these were not in conjunction with disorder, while the remaining 5 came from an interview with a single patient (a nurse), who repeatedly identified herself as suffering from personality disorder.

8. 8 Figures in square brackets refer to interviewee number (before the slash) and interview number (after the slash).

9. 9 This is so even though the parts of interviews specifically relating to depression had been excluded.

10. 10 Derogatory term for African.

11. 11 I labeled these the first and second reason for admission corpora.

To contents page

To Psychiatric Activism Page