6
DISCOURSE ANALYSIS
IAN PARKER
_1
In this chapter we are concerned with a form of analysis that addresses the ways in which language is so structured as to produce sets of meanings, discourses, that operate independently of the intentions of speakers, or writers. Discourse analysis treats the social world as a text, or rather as a system of texts which can be systematically 'read' by a researcher to lay open the psychological processes that He within them, processes that the discipline of psychology usually attributes to a machinery inside the individual's head. Most texts convey assumptions about the nature of individual psychology. In the example we have chosen you will see that, despite first appearances, the text is closely linked to the concerns of the discipline.
Background
The Latin roots of the word 'text' are to be found in the activity of weaving, and the tissue of material that clothed us is now the model for the tissue of meaning that holds the social world together. The recent history of discourse analysis is woven into the history of transformations inside and outside psychology, which started in the late 1960s and early 1970s. Inside psychology, the 'turn-to-language' that followed the paradigm 'crisis', events we described in Chapter 1, opened the way for what we now recognize to be a 'turn-to-discourse'. Outside psychology, a turn
Discourse analysis    93
to language in German phenomenology and French post-structuralism gave issue to discourse theories that now enrich and dynamize studies of speech and writing in qualitative research.
The debates that prompted the turn to language were crucial for the development of qualitative research ín psychology, for they permitted psychologists to break from a positivist fetish for figures to an exploration of meaning. As we pointed out in Chapter 1, the new paradigm writers gave a warrant for doing research in a way that was, they claimed, both scientific and sensitive to the sense that people construct in their everyday lives. The type of research - a form of qualitative research - that was proposed by the new paradigm writers focused on the roles and rules that govern ordinary language in the different social worlds we inhabit. Some interesting work appeared in the wake of the Harré and Secord (1972) manifesto for this 'ethogenic' approach, looking at such social worlds as school classrooms and football terraces {Marsh et al. 1974).
Despite Harré's (1979, 1983) ambitious theoretical reworking of Goffman's writings to produce a systematic framework for social and individual psychology, there were not many applications of that 'ethogenic' approach, and it has now all but burnt out in social psychology. Its most important legacy has been the space it provided for others wanting to do research in a non-positivistic way; the studies collected in the Reason and Rowan (1981) book, for example, are presented as part of the 'new paradigm' but range from action research to personal construct theory (approaches we believe still to be important, as our chapters in this book testify). Harré himself has moved rapidly from ethogenics through social representations to what is now the cutting edge of the new paradigm movement, discourse analysis (e.g. Davies and Harré 1990).
One important conceptual problem that the ethogenic study of social vorlds foundered on was that of the diversity of meaning, the different :ontradictory ways of speaking that govern what we do (and who we can je). The figure that seemed to structure the way an ethogenic researcher looked at a social world was that of a jigsaw; here, each member carries L partial view of the whole, and the researcher gathers 'accounts' (through interviews, sometimes through the use of repertory grids) from different nembers to piece together what the underlying form of that world is really like. (It is no accident that the search for underlying structural forms was inimating 'structuralists' in other disciplines in France and then the English-speaking world in the 1960s and 1970s.) However, the jigsaw analogy vill not work, for conflicting representations of any social world enter iom the language used outside (a social world is never a closed system). Vleaning is continually changing (it is not static but dynamic), and language is composed of many 'languages' or discourses. Writers heavily influenced by the sociology of scientific Icnowledge
94    Qualitative methods in psychology
(looking at how science is socially constructed) and by conversation analysis (looking at the mechanisms of talk) and ethnomethodology (looking at the everyday making of sense) connected with these debates in the late 1970s. These are all approaches in sociology that privilege the 'ordinary' understanding people produce about the world over researchers' theories of what is going on. These writers made the point that rather than fetishize 'consistency', researchers into language should focus on variation, that a variety of what they called 'interpretative repertoires' constructed a sense of what was going on for members, and that language understood in this way functioned in the world rather than simply represented it (Potter and "Wetherell 1987). The emphasis on variability, construction and function was already a distinguishing feature of a powerful intellectual movement - 'post-structuralism' - outside psychology (Macdonnell 1986), though the terminology was different: instead of speaking of 'interpretative repertoire', for example, post-structuralists used the term 'discourse'.
Post-structuralist writers had recognized that social relationships and our sense of ourselves is not produced by one structure but that what we do and what we are is created, 'constituted', in such a way that conflict between discourses marks all symbolic activity. For Michel Foucault (1969), discourses are 'practices that systematically form the objects of which we speak' (p. 49), and he argued that 'we are difference, that our reason is the difference of discourses, our history the difference of times, our selves the difference of masks' (p. 131). These assertions are powerful challenges to the ways we understand ourselves to be undivided, consistent individuals, and in the analytic and discussion parts of this chapter í will elaborate upon and explore these ideas.
At this point I will restrict myself to a brief example of how discourses weave together to produce a text. Three unlikely examples of phrases may serve to illustrate the operation of single discourses: if you say 'my head hurts so I must be ill', you will be employing a medical discourse; if you say 'my head hurts so I cannot really want to go to that party', you will be employing some sort of psychodynamic discourse; and if you say 'my head hurts but not in the way that yours does when you are trying it on in the way women do', you will be employing a sort of sexist discourse (whether or not, reader, you are a man). In the real world, of course, things are more complex. Take the (admittedly unreal) statement 'I've got a migraine caused by your mother-in-law's nagging which makes me relive my mother's complaints when I was a child.' Here you may find at least the three discourses, and the task of a discourse analysis is, among other things, to tease apart the discourses that are at work.
Discourse analysis in psychology is now a well-established method, and various forms of discourse analysis have illustrated how texts are not as coherent as they first seem and how they are constructed out of cultural
Discourse analysis    95
resources. To take some examples: Hollway (1984) argues that heterosexual couple talk is held in place by 'male sexual drive', 'have/hold', and 'permissive' discourses; Gilbert and Mulkay (1984) describe how scientists use empiricist (resting on evidence) and contingent (relying on intuition) repertoires to account for their choice of theory to the scientific community; and Squire (1990) shows how social psychology is organized around detective, autobiographical and science fiction narratives.
Example
It is the Foucauldian form of discourse analysis that informs the reading I will present in this chapter. I will take up some of the problems with the approach, along with the reproaches of those who prefer more ethno-methodologically inclined styles of discourse research towards the end of the chapter. Among the advantages of Foucault's position are that we need make no assumptions about what the writer or speaker 'meant' to say, and that in his historical studies he has been preoccupied with the way in which discourses, 'practices', produce types of 'psychology' (Foucault 1961, 1976). I hope to illustrate these points through the analysis of our sample text.
Common sense psychology is reproduced through all the texts of the mass media and different competing forms of popular culture. Rather than taking a segment of transcribed interview material or conversation, I have chosen a text from among the litter of contemporary consumer packaging; my assumption is that the consumer buys the message in the text on the package when they buy the product. In this case, the advice that is provided as to how to use the item partakes of a wider system of regulating practices, practices which the discipline of psychology feeds. The instructions on a packet of children's toothpaste struck me as an innocent and intriguing text, and my analysis flows from my first suspicions about the function of this text when it was first found on a friend's bathroom sink.
The text
The front of the white toothpaste tube bears the legend, on three separate lines, 'MAWS', 'PUNCH &C JUDY TOOTHPASTE', 'Children's Toothpaste with Fluoride'. This bright multi-coloured print is framed by pictures of Punch and Judy, and there is already a multitude of meanings that cluster around this segment of text which could be explored, ranging from the whiteness of the tube, which signifies the whiteness of teeth, to the happy patterns addressing the reader as child, to the connotations of 'maw' as an animal or human stomach, perhaps. As consumers, even
96    Qualitative methods in psychology
before reading the tube we have read the cardboard packet, and perhaps we have stood in shops and chosen the packet from among many jolly toothpastes ('Postman Pat', 'Mr Men') targeted at children and parents, and parents addressed as if they were children. These acts of reading lead us into the text at the back of the tube. This is the text I will focus on here:
Directions for use
Choose a children's brush that has a small head and add a pea-sized amount of Punch ÔC Judy toothpaste. To teach your child to clean teeth, stand behind and place your hand under the child's chin to tilt head back and see mouth. Brush both sides of teeth as well as tops. Brush after breakfast and last thing at night. Supervise the brushing of your child's teeth until the age of eight. If your child is taking fluoride treatment, seek professional advice concerning daily intake. Contains 0.8% Sodium Monofluorophosphate
Analysis
It will be helpful for purposes of analysis, and for pedagogical reasons in this case, to structure the reading of the text through steps to discourse analysis which have been discussed elsewhere with reference to criteria that we may use to identify discourses (Parker 1992). It should be said that these steps conceal the feelings of muddle and confusion that will overwhelm a researcher approaching a text for the first time. As the process of analysis goes on, this feeling of bewilderment will be succeeded by a conviction that the analysis is banal. What could not be seen is now seen too clearly. It is worth bearing this in mind as I trace through my reading of the toothpaste text, and when you choose a text of your own to untangle. The steps in this analysis particularize and detail the conceptual and historical work of Foucault on the construction, function and variation of discourses as they pertain to the requirements of qualitative research in
psychology.
It would be possible to explore the meanings of the shape and feel of the package in more detail (and I have already referred to the ways in which the colours of the letters signify that this is a product primarily aimed at children); to do this we would take the package as a 'text', and the first step to 'reading' it would be (a) to turn the text into a written form,_,This production of a written text, which would then be something more akin to a transcript, allows us to bring into focus connotations that normally just twinkle on the margins of our consciousness. We can then ask questions about what it means, for example, that the tube is smaller than standard tubes; it is important to note here that the smaller size not only 'reflects' the smaller size of the intended user and its smaller and
'^«^«s^ma^:imsmiš^mmĚU^M^i
Discourse analysis   97 fewer teeth, but reproduces the child as a smaller version of the adult. It is not necessary that the tube should be smaller (tubes targeted at much older people are not smaller because they also tend to have fewer teeth), and this variation in size alerts us already to the ways in which the text functions to create particular images of the child. Alongside the point that discourse analysis can be applied to visual texts, and must then be put into words, we should also be aware that the new written text will be something different, one created by the analyst and now read, as it were, second-hand.
It is not easy, or advisable, to engage in discourse analysis on one's own; it is always better (and this advice applies to some degree to all varieties of qualitative research) to work with other people. This is particularly important in the first stages of analysis, in a second step in which you should (b^free associate to the text. In the case of a piece of text that must be turned into written form It would be helpful to note the different ways in which it could be described together with other people and also to free associate with them at that point. The chains of connotations may appear bizarre, and it is tempting to disregard them. This would be a shame, for they could be useful: what is the significance, for example, of the chain that leads from Punch to child battering, to Judy as negligent mother, to the policeman, to the crocodile with the big strong teeth? We need not presuppose that the author of the toothpaste text, the designer of the package, intended that these meanings should be available to a user of the product in order to note that the Punch and Judy narrative is one that can act as a quite specific (negative) template for the care an adult may give to their child when brushing teeth.
If we are to consider the ways in which discourses, as Foucault (1969: 49) puts it, 'systematically form the objects' that are referred to in any text, we should now, as researchers, (c) sy^ematically; itemize .'objects' that appear in thisjext. A useful rule to follow here is to look for nouns. Where are they, and what could they signify? If we do this, we will then be in a better position to piece together the type of world that such a text presupposes, the world it calls once again into being each time it is read. There are:
•   'directions' (procedures for application of the product, for which this text specifies the correct application);
■ 'uses' (types of application, of which in this case it is implied that there is only one);
•  'choices' (actions presupposing a range of possible alternatives and the ability, comprising evaluation and agency, to select from that range);
•  'children' (the categories of being for whom certain types of 'brushes' are intended);
98    Qualitative methods in psychology
8 'peas' (objects of determinate size against which 'amount' can be
measured); 8 'Punch fie Judy' (puppet characters who exemplify bad parenting); 8 'Punch 6c Judy toothpaste' (specified make of toothpaste);
•  'teaching' (tutoring of others, including in the practice specified by these instructions);
8  'teeth' (with 'sides' and 'tops', identified surfaces requiring brushing);
•  'hand' (for the restraint of the child to accomplish brushing); 8  'chin' (part of the child to be grasped to restrict movement); 8  'head' (part of the child to be targeted for restraint);
•  'mouth' (part of the anatomy containing the teeth);
8 'breakfasts' (first meals after which first brushing should commence);
•  'night' (last part of the day, which should culminate in brushing of teeth);
8 'ages' (as markers of development, in which the age 'eight' figures here
as a significant marker); 8 'fluoride' (substance whose ingestion is implicated in the use of the
toothpaste); 8 'treatments' (regimes of health care); 8 'professionals' (categories of person charged with regulating treatment
and intake); 8 'advice' (mode of communication provided by professionals, distinguished
here from simple command); 8 'intake' (amount of substance deemed medically appropriate by professional); 8 '0.8% Sodium Monofluorophosphate' (specified amount of active substance).
These objects are organized and reconstituted in this text through particular ways of speaking, and it will be"Kelpful'frbm now on in the analysis (d)"to refer to these ways of speaking as objects, our objects of study, the discourses. The identification of the objects that are referred to in the text Has just brought us to the edge, to the point of being able to identify the discourses that hold them together. Before we can move beyond that point to the part of the analysis where the discourses will start to take on a life of their own in our reading we should (e) systematically itemize the 'subjects' (the categories of person) who appear in this text, and (f)~Tečon-struct^as a. device to explore differential rights to speak within discourse, what each type of person may say within the framework of rules presupposed by the text. To take the fifth step, then, some of the objects I have already identified are also sentient beings, the 'subjects'. They are:
8 'children' (the categories of being for whom certain types of 'brushes' are intended);
Discourse analysis   99
8 'professionals' (categories of person charged with regulating treatment and intake).
In addition to these two evident categories - and leaving aside the agency attributed to 'Punch &C Judy' in popular representations for the moment - there is a third category of subject, that addressed by the text: 8 'parent' (category of person for whom directions are intended, and the nature of this subject is constituted through the three points in the text in which the reader is addressed as the owner - through the index 'your child' - of the child for whom the product is intended). We can now, as the sixth step, reconstruct the rights and responsibilities of this most important subject of the text and the network j?f relationships that are reconstituted which position this parent, tEereader, in relation to the 'child' and the 'professional'. First, in relation to the child, the parent must choose for it, teach it, stand behind, restrain and brush its teeth (both sides and tops), perform this duty twice a day at specified times, and supervise the activity (which here implies the increasing self-direction of the child in the task) until a specified age (at which point, it is implied, the child can carry on without supervision). Second, in relation to the 'professional', the parent must seek advice, and follow prescriptions concerning treatment and intake. Third, in relation to the addressor (the 'subject supposed to know', to have written the text, to be speaking to the reader), the parent must follow the directions, and, as a part of the directions, seek advice, if necessary, from a 'professional'. This circuit of responsibilities positions the addressor in alliance with the 'professional' in the instruction to seek advice (but with deference, also, in the attribution of rights to the 'professional' to determine appropriate daily intake).
One of the functions of the text, as of any text, is to bring to life (again, for"visTiÖw as researchers) a network of relationships, and as we move on to link this network together~around the objects the text refers to we can start (gJ~toThap the different versions of the social world which coexist in thFtext. Äs we do this we are coming closer to identifying discrete ways oFspeaking that are at work in this text. The_instructions require the reader to behave in a rational way. They are worded in such a way as to presumeTthat the reader is in permanent charge of a child (from every breakfast to every night). They call for agreement with the idea that the child develops in a particular way up to a particular point (age eight) and they also assume that the reader is willing to consult professionals about the health of the child.
Note that the category of the 'child' here is not gendered (it could be a boy or a girl). Not many years ago, it would have been likely to have been referred to as 'he'. This contrast in ways of specifying gender also draws attention to broader cultural assumptions that appear in texts at unlikely
100   Qualitative methods in psychology
moments. Consider, for example, the difference between the addressee for this text, whom we have taken to be a parent (from the designator 'your child'), and the addressee who would be in charge of the child in many other cultures outside this text's frame of reference, an addressee who could well be an older sibling. "We are thus arriving at some pictures of relationships at work here: rational rule-following, parental, developmental and medical.
Each of these ways of organizing the world carries with it rules for reproving those who fail to adhere to it: to break from rationality and rule-following will lead to claims that the reader is stupid or dangerous; to refuse parental responsibilities invites accusations of irresponsibility; to reject the idea that the child follows a normative developmental route and that teaching should be geared to it may lead to one being labelled as selfish and complicit in delinquency; and to challenge the call to consult medically qualified professionals is often to be viewed as deviant and anti-scientific. These possibilities are enumerated here as a step injwhich we (h) speculate as to how each of these patterns would deal with objections to these instructions and the cultural rules hidden within them. I have suggested how such defensive procedures might be played out after I have listed what we are increasingly taking to be the four key sets of statements, but the relationship between steps (h) and (g) in the process of analysis is messier, and it is also useful to ask how 'imaginary' authors of statements in the text would respond to those who contradicted them. This technique can help us arrive at separate discourses.
It is the discourses that 'form the objects of which they speak', and not authors who speak through the text as if the text were a kind of transparent screen upon which the writer's intentions were displayed. Our 'imaginary authors', then, are our own creations (as, indeed, are the discourses to an extent, but I will return to that issue below), and we use them to emphasize the variation, the contradictions in the text. It is helpful to focus on this contradiction and concordance between voices in the text, and to spend a little time doing this in two further steps of the analysis: (i) identifying contrasts between ways of speaking; andjj) Ídentifyirj,g,p_oints where these ways of speaking overlap. In this case the concern with instruction, supervision and professional rights locks together alarmingly, and I shall discuss this further in our discussion (under the heading 'Repercussions of the reading'). I will also want to pursue the issue of how distinctions between the discourses could be magnified, and how 'the child' of the parent constructed here and 'the child' of the medical professional differs. "We can also note at this point how the serious tone of the directions for use contrasts with the frivolity of the Punch and Judy Imagery, but also how that imagery works then to confirm the position of the parent and the professional as guarantors of serious guardianship.
Wp ran nnw make some comparisons with other texts (k) to assess how
Discourse analysis    101
these ways of speaking address different audiences. It may also be possible to find expressions of the discourse in which it seems to fold around upon itself and comment on how important it is to speak that way. Although we may find it useful to look at instructions on other toothpastes, we are moving now beyond this type of text to look at how the patterns of rňeäňíhg that are apparent within it also operate elsewhere. Instructions such'as these are already assessed, for example, by the Campaign for Clear English, which draws attention to and praises clarity and rationality (linking these two qualities) in official documents; the ways in which parents are addressed in conservative political discourse are often explicit about the importance of the family as foundation of civilized society; the discussions of education in debates on the relationship between schooling and family values are closely tied to the claim that there are distinct identifiable stages of intellectual and moral development; and, with the increase in popularity of 'alternative' medicine, scientific and professional standards are emphasized as bulwarks against charlatanism and in the defence of correct medical terminology (and those who have the right to use it).
"We have now reached the point at which we_(l)._choose an „appropriate terminology to label the discourses. This is one way of structuring a reading of the text. I have tried to make this reading plausible, and you may disagree. In your analysis of other texts, you should also write your report in a spirit of polemic and debate. The collapsing of rationality and rule-following under the heading 'rationalist', the labelling of the terms which invoke parental duties as 'familial', the linking of themes of development and education together as 'developmental-educational' and the use of the rubric 'medical' to include the reference to professionals, daily intake and use of the chemical terminology are, in part, operations applied for convenience, tidiness in presentation, but I will have to justify these choices later, in the discussion. To summarize, so far, I can identify four discourses:
•   ''rationalist - in which the ability to follow procedures ('directions for use') requires choices of implement and judgement of amount ('small head' and 'pea-sized amount') and is predicated on recognition of appropriate authority in health care (following 'directions' and seeking 'professional advice');
•   'familial1 - in which ownership ('your child') runs alongside supervision and continuous care (the assumption that the child is present each breakfast and 'last thing at night') and is framed by the image of bad parenting (the figure of 'Punch 6c Judy');
•  'developmental-educational' - in which the teaching of the child (parental activity) precedes supervision (the child's still tutored but self-governed activity) and then reaches an identifiable stage as a developmental milestone (the 'age of eight');
102    Qualitative methods in psychology
• 'medical' - in which the process of using the toothpaste is necessarily linked to hygiene (brushing after meals), professional supervision ('fluoride treatment') and the specification of ingestion and chemical composition of substances ('daily intake', '0.8% Sodium Monofluorophosphate').
Repercussions of the reading
The analysis in this type of study differs markedly from the 'results' section of an experimental report, in which the different measures are tabulated and the significance level is identified. The analysis section of the report is longer (and it shares this characteristic with many other types of qualitative research we describe in this book). The analysis is also more 'discursive' in the sense that it traces the reasoning by which discourses were located in the text (though this characteristic is exaggerated in this chapter because I am not only describing an analysis but also recapitulating a series of steps to educate a reader in technique), and in so far as the unravelling of the text into discrete discourses necessitates a discussion of associations, cross-connections and contradictions between groups of terms and their everyday uses. This is a reading of one case example. It is not necessary to read twenty different tubes of toothpaste, though it may be interesting to do so.
The analysis has applied itself to the task not only of reading the text in question but also of following, 'steps', and this is certainly not the lightest and most engaging way of presenting the material. The presentation of discourse analysis marks it as a variety of qualitative research which, unless measures are taken to the contrary, tends to conceal its reflexive aspects (a characteristic it shares with observational studies and some personal construct approaches and which differentiates it sharply from action research and feminist work). We should be clear, then, that the reading I have presented here is my response to the text and that the discourses are as much our creations as they are 'objects' existing independently of us. Our encounter with these discourses as they manifest themselves in this text is not an encounter with something new to us but rather with something very familiar; for the history that bears the discourses as 'objective' phenomena is also the history that bears us as 'subjective' beings. An advantage here is that discourse analysis makes public 1 its sources in a reading. Our subjectivity as a historically produced and contingent form of matter is, then, an important research tool for the decoding of forms of language.
My discussion of the analysis of the toothpaste text must unfortunately be restricted here to an overview of the types of points I would want to cover in a lengthier study. The discussion in discourse analytic research may extend the analysis through (m) a study ofjwher^jmd when these _d^ccmrsej..dejeic^ed_and (n) a description of how they have operated to
Discourse analysis    103
naturalize the things they refer to; that is, how they 'form the objects of wTnHľtfiéy sp&^ in such a way that it appears perverse and nonsensical to question that they are really there. These two tasks underlie questions I would want to pose concerning the role of the discourses in the life of institutions, power relations and the transmission of ideology".
ín this text the discourses clearly reinforce the institutions of the family and of medicine. Foucault's way of analysing the history of discourse has been applied to the family, and the role of state and welfare practices in shaping the internal structures of the family has been closely connected with the images the medical profession has distributed over the years in prophylactic advice against bad parenting (Donzelot 1979). Punch and Judy operate in this history as a contradictory sign of familial relationships, for while they are used to illustrate the moral dangers of neglect and child abuse they also function as subversive carnivalesque emblems of revolt against the authorities. The extension of analysis into the discussion is already following here a step (p) in which the discourses' role in reproducing institutions is examined alongside a step (p) in which the discourses that_subvert those institutions are explored.
Such institutions do not simply structure social life, they also constrain what can be said, who can say it and how people may act and conceive of their own agency and subjectivity. Wherever there is power there is resistance (Foucault 1975), and the analysis of institutions could be extended to look atjq) who would benefit and who would be disadvantaged by such discourses, and so also (r) who would want to support or who to discredit tnešé"waýs öf talking. The powers that are accorded to the parent and the medic would appear to single these figures out as subjects who are wielding power over the child. We should take care, however, not to treat this exercise of power as deliberate, or to neglect the ways in which those who exercise power are also enmeshed within it (Foucault 1975, 1976). The figure of the parent with regard to the medic is contradictory, for example. Discourse analysis is concerned with the ways in which meaning is reproduced and~Säns"fprmed in texts, and when such reproduction and transformation concerns institutions and power relations we are led inevitably to a consideration of the role of ideology. At this point I can link together some of the discourses I have described in the analysis to show (s)  , how~they "also entail other discourses.which enjoy,power, and . (t) how j ík^L£§E^9^0c^PÍ_clia^eilS(?ÄPinínant conceptions^as^tq what can change   i and whatjnaybe possible in the future. I can only suggest here that it may   ; be illuminating to trace connections between images of rationality in the adult, accounts of child development and conceptions of the family as the 'normal' arena for the care of the child; not only medicine, but psychology as well slips in here naturally as an institution concerned with hygiene, or mental hygiene. The descriptions of psychology as an apparatus, a 'psy-complex' which emerged aloneside mprlirin^ ;* +J-° i—*■ —-
104   Qualitative methods in psychology
relevant here (Rose 1985, 19S9). The text seems, in this light, to encapsulate an image of psychology and cognate disciplines as practices obsessed with surveillance and control. The discussion could not move much further forward without trespassing into the disciplines of sociology and history, without an account of psychology itself as an institution suffused with power and ideology.
Assessment
I will discuss some of the limitations of the approach I have adopted here, and the criticisms that may be levelled against the reading by other writers in psychology sympathetic to discourse analysis, before moving on to consider briefly some deeper problems with this type of work.
Limits
There is, as I pointed out in the introduction to this chapter, a strand of work closely tied to ethnomethodological studies in the sociology of scientific knowledge which was a conduit for entry of discourse analysis into psychology (Potter and Wetherell 1987). The concern with everyday accounting for action takes priority over the researcher's perspective, and the rhetoric that people employ is privileged over any evaluation a psychologist could give. The analytic process í have described in this chapter has been subjected to criticism from writers in this tradition, and the concerns have been to do with the tendency to reification, the ways in which the analysis presupposes what it pretends to discover, and the use of common sense knowledge in the elaboration of the categories that are eventually 'discovered' (Potter et al. 1990). It could be argued, perhaps, that the four discourses simply do not exist as if they were invisible girders that held language together, and that it does not do justice to the subde strategies people engage in to make sense, to pretend that experts can detect what is really happening. "Worse than that, there is also an element of mystification in that I have pretended to tell you what is really there as if you did not know in the first place; I have only re-presented to you common sense notions of rationality, the family, developmental stages and medicine. I would agree that there is a problem with the assumption, an assumption that guides much positivist research, that the psychologist knows best, but that does not mean that there should be no critical perspective on the ways in which language is used. The discourses are not really there hidden away awaiting discovery; they are indeed produced through analysis, but they do then give a coherence to the organization of language and tap institutional structures of power and ideology in a way that a simple appeal to common sense reasoning could never do. As I noted before, I am influenced in this view by Foucault's work and the post-structuralist
Discourse analysis    105
tradition. As a counterpart to these criticisms, however, I should also consider those that come from the other side, from writers who may argue that I have been too cautious, not too far from but too close to the ethnomethodological strand of work. It could be argued, for example, that a fifth discourse, 'self-regulation' is also at work in our toothpaste text, and that this discourse, like the others that were described but more so, could only be drawn out by using a prior theoretical historical framework (Foucault's). I have indeed used theory to produce this reading; perhaps, for some, not enough. Criticisms have already been directed at the ethnomethodological strand (Bowers 1988), and I could now guess how they may be adapted to apply to the analysis I have presented here.
Although I have described how notions of rationality are reproduced in this text, it could be argued that I have presented an image of the meanings in the text as fairly static, that I have slipped back into standard structuralist f styles of analysis, which do not really have much to say about resistance / or the desire of 'readers' and 'writers' in the process of resistance. Al-thougn^lHTTe^dÜTgTs^supposed to focus on variation, there is too little analysis and discussion of contradiction and free play in the text. The text pretends to be a serious document, but it is bounded by'fun (in the figures of Punch and Judy), and analysis needs to work more thoroughly with the idea that subjectivity is always split, anarchic. Such analysis requires, perhaps, the use of psychoanalytic ideas (Hollway 1989; Parker 1995).
Critiques
It would not be difficult to predict the objections that would be levelled against the analysis by more traditional psychologists, and a clear expression of the hostility of the experimental tradition to this type of work has already appeared (Abrams and Hogg 1990), It is not clear in what sense this text is representative of instructions on children's toothpaste tubes, and no attempt was made to compare the text with those that may appear on, for example, Postman Pat or Mr Men toothpaste. The reading I have presented is only my opinion, and I have made no attempt to validate it against other forms of analysis, or even to discover whether the procedure 1 used is reliable when applied to other texts. I do not even know whether parents read the instructions, let alone whether the instructions actually determine the behaviour of parents. I have drawn on accounts of the 'psy-complex' that are speculative at best, and I assumed that a reader will simply and unthinkingly accept and implement the instructions in the way I assume them to function. Such complaints could be addressed to many of the examples of qualitative research we describe in this book, and they would apply to all studies of discourse that fail to use quantitative methods. This analysis, experimental psychologists would say, is a travesty of scientific inquiry.
106    Qualitative methods in psychology
From the other side, however, we must note the dissatisfaction with discourse analysis of some researchers who have no such qualms about abandoning 'science'. This last set of criticisms chimes in with those who are still inspired by the radical political aspects of post-structuralist and psychoanalytic theory. One overview of problems from this position identifies thirty-two problems with discourse analysis (Parker and Burman 1993), but we have space here only to note that they include the problem of treating language (texts, discourses) as more powerful than other material constraints on action, and the fantasy that the researcher can pull out a toolbag and apply it to any and every text without reflecting on the effects of an analysis. There is more variability in human action and experience than that expressed in language; as researchers we construct our own image of the world when we reconstruct 'discourses'; and we have some responsibility for how our analysis will function.
We have traced the analysis of sets of statements that course through a tiny text and tracked the ways in which the discourses carry in their wake sets of assumptions about the nature of social relationships, relationships that the discipline of psychology has in the past both investigated and endorsed. Psychology which operates in these ways has traditionally relied on the rhetoric and practice of quantification and observation. In contrast we have presented an analysis that is also a critique from the standpoint of qualitative research and those who are usually subjected to the professional gaze.
Useful reading
Burman, E. and Parker, I. (eds) (1993). Discourse Analytic Research: Repertoires and Readings of Texts in Action. London: Routledge.
HoUway, W. (1979). Subjectivity and Method in Psychology: Gender, Meaning and Science. London: Sage.
Potter, J. and Wetherell, M. (1987). Discourse and Social Psychology: Beyond Attitudes and Behaviour. London: Sage.
Squire, C. (1990). 'Crisis what crisis? Discourses and narratives of the "social" in social psychology', in I. Parker and J. Shotter (eds) Deconstructing Social Psychology. London: Routledge, pp. 33-46,
References
Abrams, D. and Hogg, M.A. (1990). 'The context of discourse: let's not throw out the baby with the bathwater'. Philosophical Psychology, 3(2), 219-25.
Bowers, J. (1988). 'Essay review of Discourse and Social Psychology'. British Journal of Social Psychology, 27, 185-92.
Davies, B. and Harré, R. (1990). 'Positioning: the discursive production of selves'. Journal for the Theory of Social Behaviour, 20(1), 43-63.
Donzelot, J. (1979). The Policing of Families. London: Hutchinson.
Discourse analysis    107
Foucault, M. (1961). Madness and Civilization: a History of Insanity in the Age
of Reason. London: Tavistock. Foucault, M. (1969). The Archaeology of Knowledge. London: Tavistock. Foucault, M. (1975). Discipline and Punish. London: Allen Lane. Foucault, M.  (1976). The History of Sexuality,  Volume 1, an Introduction.
Harmondsworth: Pelican. Gilbert, N. and Mulkay, M.  (1984).  Opening Pandora's Box: a Sociological
Analysis of Scientists' Discourse. Cambridge: Cambridge University Press. Harré, R. (1979). Social Being: a Theory for Social Psychology. Oxford: Basil
Blackwell. Harré, R. (1983). Personal Being; a Theory for Individual Psychology. Oxford: Basil
Blackwell. Harré, R. and Secord, P. (1972). The Explanation of Social Behaviour. Oxford:
Blackwell. Hollway, W. (1984). 'Gender difference and the production of subjectivity', in J.
Henriques, W. Hollway, C. Urwin, C. Venn and W. Walkerdine (eds) Changing
the Subject: Psychology, Social Regulation and Subjectivity. London: Methuen,
pp. 227-63. Hollway, W. (1989). Subjectivity and Method hi Psychology: Gender, Meaning
and Science, London: Sage. Macdonnell, D. (1986). Theories of Discourse: an Introduction. Oxford: Blackwell. Marsh, P., Rosser, E. and Harré, R. (1974). The Rules of Disorder. London:
Routledge and Kegan Paul. Parker, I. (1992). Discourse Dynamics: Critical Analysis for Social and Individual
Psychology. London: Routledge. Parker, I. (1995). 'Discursive complexes in material culture', in J. Haworth (ed.)
Psychological Research: Innovative Methods and Strategies. London: Routledge. Parker, I. and Burman, E. (1993). 'Against discursive imperialism, empiricism and
constructionism: thirty two problems with discourse analysis', in E. Burman and
I. Parker (eds) Discourse Analytic Research: Repertoires and Readings of Texts
in Action. London: Routledge, pp. 155-72. Potter, J. and Wetherell, M. (1987). Discourse and Social Psychology: Beyond
Attitudes and Behaviour. London: Sage. Potter, J., Wetherell, M., Gill, R. and Edwards, D. (1990). 'Discourse - noun, verb
or social practice'. Philosophical Psychology, 3(2), 205-17. Reason, P. and Rowan, J. (eds) (1981) Human Inquiry: a Sourcebook of New
Paradigm Research. Chichester: Wiley. Rose, N. (1985). The Psychological Complex. London: Routledge and Kegan Paul. Rose, N. (1989). Governing the Soul. London: Routledge. Squire, C. (1990). 'Crisis what crisis? Discourses and narratives of the "social" in
social psychology', in I. Parker and J. Shotter (eds) Deconstruďing Social Psychology. London: Routledge, pp. 33-46.
Acknowledgements
Thanks are due to Deborah Marks for her helpful comments on an earlier draft of this chapter.