Expertise and Forms of Knowledge in the Government of Families

This paper examines the relationship between the state and the individual in relation to an aspect of mundane family life – the feeding of babies and young children. The nutritional status of children has long been a matter of national concern and infant feeding is an aspect of family life that has been subjected to substantial state intervention. It exemplifies the imposition upon women the ‘biologico-moral responsibility’ for the welfare of children (Foucault, 1991b).


Introduction
There is a paradox at the heart of the relationship between the state and the family in contemporary liberal states. On the one hand, there is a commitment to the limiting the role of the state and respect for the autonomy and privacy of individuals and families. On the other, there is a concern to influence and regulate social and economic life, so as to foster desired values including wealth, health and other kinds of well-being (Rose, 1992). This paper examines this tension using empirical data in relation to one area of mundane family life -the feeding of babies and young children. In particular, it considers the medicalisation of infant feeding and the ways in which women, who are made responsible for infant feeding, endorse, resist and refuse such medicalisation.
The empirical data are drawn from a longitudinal interview study of infant feeding practices, which followed a cohort of women from the later stages of pregnancy until their babies were two years old. Infant feeding is a fertile area in which to examine the relationship of the state to the family. The nutritional status of children has long been a matter of national concern. It is an aspect of family life that has been subjected to substantial state intervention. In the UK, for example, there is an elaborate state-sponsored apparatus for the dissemination of advice about optimal infant feeding. Before babies are born, mothers receive feeding advice at antenatal classes and clinics. Midwives have a statutory duty to visit and advise new mothers at home for the first ten days after birth. Subsequently, health visitors visit mothers and babies to provide advice, much of which focuses on infant feeding practices. They also run regular clinics for weighing babies for and advising mothers.

Government, biopower and discipline
Foucault traces the tension between autonomy and privacy in the regulation of social and economic sphere back to the juxtaposition of two separate, and potentially contradictory, developments in the seventeenth and eighteenth centuries. On the one hand, Enlightenment ideals of freedom and autonomy pointed towards the limiting state intervention in the private affairs of citizens. On the other, a new rationality of government emerged, which took the welfare of the population as its primary and ultimate goal and implied an active and interventionist role for the state: In contrast to sovereignty, government has as its purpose not the act of government itself but the welfare of the population, the improvement of its condition, the increase of its wealth, longevity, health etc. . . . the population is the subject of needs, of aspirations but it is also the object in the hands of the government. (Foucault, 1991a: 100). State attempts to influence infant feeding practices can be seen as part of this project of government. Adequate and appropriate nutrition in childhood is held to make a vital contribution to the health of the adult population, with all the implications this has for national security and welfare expenditure. In the words of a recent UK government report, Nutrition in the early years of life is a major determinant of growth and development and it also influences adult health. (Department of Health, 1994: 7) The new focus on the welfare of populations was associated with the emergence of 'bio-power' (Foucault, 1980a: 143). This term designates a range of technologies, knowledges and discourses used to analyse, control and regu-late human bodies and populations. These include the disciplines derived from the human sciences which, Foucault argued, Centred upon the body as a machine, its disciplining, the optimization of its capabilities, the extortion of its forces, the parallel increase of its usefulness and its docility. (Foucault, 1980a: 139) Bio-power also draws upon a new 'technology of population' (Foucault, 1991b) which offers Demographic estimates, the calculation of the pyramid of ages, different life expectations and levels of mortality, studies of the reciprocal relations of growth of wealth and growth of populations, . . . the biological traits of a population become relevant factors for economic management, and it becomes necessary to organize around them an apparatus which will ensure not only their subjection but the constant increase of their utility. (p. 278) Infant feeding is a prime example of these two aspects of bio-power. State interventions are directed towards maximising the life-long contribution of infants, through regulating their diets, to produce useful and healthy bodies. The promotion of breast feeding is a crucial element of the state programme. In a very real sense, breast feeding involves the optimisation of mothers' capabilities to provide the best possible nutrition for their babies through the 'extortion' of their bodies' forces.
Similarly, 'technologies of population' are central to the body of technical knowledge about infant feeding that is disseminated to mothers by the state via the health care system. For example, the UK Department of Health's Committee on Medical Aspects of Food Policy produced a series of reports and recommendations, based on expert evidence, designed to ensure optimal health and development (Department of Health and Social Security, 1974, 1988Department of Health, 1994). The recommendations are based on actuarial assessments of risks and benefits to babies' short and long-term health (Murphy, 2000). For example, exclusive breast feeding for at least four months is promoted because epidemiological studies have demonstrated statistical associations between formula 1 feeding and a wide range of respiratory and gastrointestinal diseases (Howie et al., 1990); allergies, including eczema, asthma and food intolerance (Virtanen et al., 1991); Crohn's Disease and cancer (Lawrence, 1995); Sudden Infant Death Syndrome (Ford et al., 1993); and impaired mother-infant bonding, confidence and self-esteem (Lawrence, 1995).
Telling mothers how they should use their bodies and how their babies should be fed could be interpreted as an illegitimate incursion into the privacy of family life and an assault upon mothers' autonomy and self-determination. Foucault considers how Enlightenment ideals of freedom and autonomy came Expertise and forms of knowledge to be subordinated to bio-power. The techniques of bio-power were not those of overt repression or coercion. Rather, they were disciplinary technologies whose aim was to produce a 'docile body that may be subjected, used, transformed and improved' (Foucault, 1991c: 136). As a consequence, the functioning of these technologies as mechanisms of social control, and their challenge this posed to Enlightenment ideals, were obscured. They operated as a series of 'quiet' or 'subtle' coercions (Foucault, 1991c: 137). Indeed their very invisibility enhanced their effectiveness.
Disciplinary technologies incorporate powerful discourses that shape how we understand ourselves and distinguish between the true and the false, the right and the wrong. They incorporate 'normalizing judgements' (Foucault, 1991c: 183) which qualify or disqualify people as 'fit and proper members of the social order' (Danaher, 2000: 61). Thus, individual behaviour is constrained through 'a set of standards and values associated with normality which are set into play by a network of ostensibly beneficent and scientific forms of knowledge' (McNay, 1994: 94). The most powerful discourses are those, such as medicine, that have a firm institutional base (Weedon, 1987).
Medicine plays a key role in responding to the imperative of promoting the health and consequent utility of the population. Foucault refers to the emergence of a 'medical police' (Foucault, 1991b: 171), as health became the object of surveillance, analysis, intervention and modification. The family became a key site of 'medicalisation' (Foucault, 1991b: 172), with the associated imposition upon women of a 'biologico-moral responsibility' for the welfare of children (Foucault, 1980a: 104): New and highly detailed rules serve to codify relations between adults and children . . . a whole series of obligations imposed on parents and children alike: obligations of a physical kind (care, contact, hygiene, cleanliness, attentive proximity), suckling of children by their mothers . . . Health, and principally the health of children becomes one of the family's most demanding objectives. (Foucault, 1991b: 172-3).
Such obligations and associated normalising judgements fell most heavily on women.
Normalizing judgements, coupled with 'the panoptic gaze' (Foucault, 1991c: 195-228), are central to the mechanisms of social control characteristic of biopower (Sawicki, 1991). The effectiveness of Bentham's Panopticon, as a penal mechanism, lay in the consciousness it gave rise to in its subjects that they were constantly visible, even when not subject to external surveillance. For Foucault this represents the essence of disciplinary society. The most powerful effect of the panoptic gaze is the attitude of self-policing engenders in its subjects: He (sic) who is subjected to a field of visibility, and who knows it, assumes responsibility for the constraints of power; he makes them play sponta-neously upon himself; he inscribes in himself the power relation in which he simultaneously plays both roles; he becomes the principle of his own subjection (Foucault, 1991c: 202-3) Bio-power operates through these technologies of normalization and the panoptic gaze to produce self-regulating subjects. It is a power that 'insidiously objectifies those on whom it is applied, to form a body of knowledge about these individuals (Foucault, 1991c: 220). It operates through disciplinary practices which produce a 'subtle, calculated technology of subjection. ' (p. 221) Expertise and government at a distance In this paper, I consider how such disciplinary technologies operate in relation to mothers' feeding practices. I examine how such technologies are used to reconcile the state's commitment to respect the autonomy of citizens and the privacy of family life with its concern to regulate the health of the population. Medicalised expert discourses, with their associated normalising judgements, play a crucial role here. As Rose argues, expert discourses enable liberal governments to 'govern at a distance' (Rose, 1992: 192). The formal freedom of the individual is maintained but that individual is 'shaped, guided and moulded into one capable of exercising that freedom through systems of domination' (Dean, 1999: 163). Expert discourses that identify certain practices as healthy, and therefore legitimate, and others as unhealthy, and therefore illegitimate, play a crucial role in such systems of domination.
Attempts to govern mothers at a distance through expert discourses are well documented. Apple coined the phrase 'the ideology of scientific motherhood' (Apple, 1995: 161) to describe the paradoxical positioning of US mothers as both responsible for their families and dependent upon experts to exercise that responsibility. Similarly, Arnup draws on Canadian data to show how motherhood has been exalted while, at the same time, mothers have been reduced to 'slavish followers of detailed medical advice' (Arnup, 1990: 203). In the UK context, Rose observes that 'the family is simultaneously allotted its responsibilities, assured of its natural capacities, and educated in the fact that it needs to be educated by experts in order to have confidence in its own capacities. Parents are bound into the language and evaluations of expertise at the very moment they are assured of their freedom and autonomy.' (Rose, 1999: 208, emphasis added).
The state's promotion of 'healthy' feeding practices operates primarily through education and persuasion. As noted above, mothers have numerous encounters with health professionals which are occasions for the 'implantation of technologies of calculation' (Rose and Millar, 1992: 197) in mothers as they are educated and persuaded to 'exercise continued informed scrutiny ' (p. 195) of the potential consequences of their feeding practices. At such encounters normalising judgements about appropriate maternal behaviours are communicated and reinforced.
The advice given in these encounters is detailed and specific (Heritage and Sefi, 1992). It is formulated as a set of rules for healthy infant feeding. Mothers are advised to fully breast-feed their babies for at least four months. If the baby appears dissatisfied, mothers are told to increase the frequency of breast feeds, rather than offering formula milk. They are warned against introducing solid food before four months or delaying a mixed diet beyond six months. Advice about the nature and consistency of foods at various stages is highly prescriptive.
Rose has termed such professional practices 'government through freedom' (Rose, 1999). They act 'on the choices and self-steering properties of individuals, families, communities and organizations' (Rose, 1999: xxiii). In Foucault's terms, they rely upon a series of 'quiet coercions' to render mothers 'docile bodies' who 'choose' to align their feeding practices with expert recommendations. These quiet coercions incorporate normalising judgements. When mothers are warned of the dangers of formula feeding, the implicit message is that a good mother will breast feed. Expert expositions of the benefits of breast feeding imply that only mothers who put their own interests above their babies' would choose to formula feed. Thus, medicalised expert discourses seek to render women self-policing subjects.

Resistance
If mothers follow expert prescriptions on infant feeding to the letter the costs in terms of time, energy and opportunity can be considerable. Advice is premised on the assumption that mothers' exclusive goal in infant feeding is (or at least ought to be) enhancing children's physical and mental functioning and promoting their long-term health. However, as I have argued elsewhere (Murphy et al., 1998a;, this is only one of a number of projects mothers juggle as they decide how, when and what to feed their babies. Many have to balance their babies' needs with other obligations both at home and in external employment settings. Mothers are particularly concerned to secure a good night's sleep for themselves and their partners. They are keen to reestablish an ordered family life after the disruption of the birth. They also use food to control or distract babies, particularly in public settings. Mothers are often eager to encourage their babies to move on to 'the next stage' of feeding and see this as a measure of the child's progress and/or intelligence. There are, then, good reasons why women might resist state-sponsored attempts to govern their feeding practices. Foucault has frequently been criticised for over-estimating the effectiveness of disciplinary control in his treatment of the discursive production of docile bodies. He has been accused of producing a theory of power which is unidirectional, monolithic and irresistible (McNay, 1992). Feminists, including McNay and Ransom (1993), have argued that such an understanding of power places women in the role of hapless victim with no possibility of resistance or undermining the operation of disciplinary power. While it is true that in early works such as Discipline and Punish, Foucault paid scant attention to the possibilities of resistance, he distanced himself from such an understanding of power: The idea that power is a system of domination that controls everything and leaves no room for freedom cannot be attributed to me. (Foucault, 1984: 442) Foucault's later works balance his early focus on technologies of domination with an exploration of the 'technologies of the self' (Foucault, 1990). This idea complements Foucault's earlier work on the ways in which the subject is constituted as an object of knowledge with an analysis of the ways in which individuals come to understand themselves as subjects (McNay, 1994). It is through this concept that a theory of possible resistance is developed. Foucault argues that the process through which large-scale cultural patterns come to be demonstrated at the individual level is not one of straightforward imposition. Cultural patterns will not be perfectly reflected in individual behaviour. Nevertheless, individuals cannot simply act in any way they wish, I would say that if I am now interested in how the subject constitutes itself in an active fashion through practices of the self, these practices are nevertheless not something invented by the individual himself (sic). They are models that he finds in his culture and that are proposed, suggested, imposed upon him by his culture, his society and his social group. (Foucault, 1984: 441-2) Thus, although individuals and groups can resist dominant discourses by creating or reproducing 'counter' discourses, the availability and acceptability of such alternative discourses are always limited by social and cultural contexts (Lupton, 1995). Rose (1999) argues that while 'government through freedom' multiplies the points at which individuals are required to collaborate in their own government, it also multiplies the points at which they are able to resist such government. This interplay between government and resistance is central to this paper. Through an analysis of mothers' accounts of their infant feeding practices, I examine how mothers, who are made responsible in contemporary discourses around child nutrition, refuse, counter and challenge the demands that are placed upon them.
pregnancies until two years after their babies' births. Since large scale UK infant feeding surveys (eg White et al., 1992) report that feeding practices vary with both occupational class and age, the sample was stratified in respect of these two variables. Information on the occupational class profile of National Health Service general medical practices within a ten-mile radius of Nottingham was obtained by combining data from the 1991 UK population census with information from the Family Health Service Authority (the government body then responsible for the provision of primary medical services). On this basis, ten general practices with contrasting occupational class profiles were selected. Access to the antenatal registers of these practices was negotiated and 144 women expected to give birth in a six month period were identified. Nine women were excluded at this stage because they had a foetal abnormality, a multiple pregnancy or would, in their GP's opinion, find the interviews distressing.
The GPs wrote to the remaining 135 women, enclosing an information leaflet, prepared by the research team. This outlined what would be involved in the study and stressed that participation was voluntary. The aim was described as 'finding out about women's views and experiences of feeding their babies'. Women were asked to return a slip indicating whether or not they were willing to discuss participation in the study. Twenty (15%) refused, 9 were untraceable at the address on their records, and 11 were excluded because of an imminent move out of the study area. A further five women were excluded by their GPs because of unexpected medical complications.
A quota sample of thirty-six mothers, stratified by age and occupational class, was drawn sequentially from the remaining 90 women (see Table One). Women's own occupational class was categorised using the UK Registrar General's classification (Office of Population Censuses and Surveys, 1980). Half the mother in each class grouping were above the mean age at birth of first child and half were below. Two mothers, one from class 3 and the other from class 4/5 were from minority ethnic groups. One was African Caribbean and the other was from South Asia. Following discussion, all 36 women agreed to participate and all completed the full schedule of six interviews over a twoyear period. Six qualitative interviews were carried out with each mother, by one of two female interviewers. One was conducted before the births and the remaining five at fixed intervals up to two years after the births, giving a total of 216 interviews. The interviews lasted between one and two hours and were audiotape-recorded and fully transcribed. At their request, interviews with three women were not taped but notes were taken and fully written up immediately after the interviews. In all six interviews mothers were invited to tell the story of how they had come to make the decisions they had about feeding their babies in the period since the previous interviews. A guide was developed for the each of the six interviews, identifying a range of topics, relevant to the baby's age, to be explored at some point during the interviews. These included how the baby was currently being fed, who had been involved in decisions about changes in the baby's food, the mother's feelings about how feeding was going, difficulties encountered, ideas about how feeding might change in the interval before the next interview, and so on. However, the interview style was conversational and the interviewers encouraged women to express their experiences, views and practices in their own terms, rather than being constrained by a pre-determined schedule.
For the initial stage of the analysis, a sub-sample of 12 women was selected, reflecting variations in age, occupational class, and feeding outcome. Transcripts for this sub-sample were subjected to detailed analysis by the author and two research associates. The analysis drew upon some 'foreshadowed' themes identified while planning the project. These included, for example, the domestic division of labour around infant feeding and the role played by professionals, lay networks and babies themselves in influencing infant feeding. Alongside this, the researchers also conducted a more inductive style of analysis, allowing analytic themes to emerge from the interview data. Such themes included, for example, different constructions of motherhood and of babies and the theme of expertise that is central to this paper.
Each of the three team members carried out a preliminary analysis of the data from four of the twelve women selected for this first stage of analysis. Through repeated readings and of the transcripts, they extracted data relevant to the foreshadowed themes and also identified potential new themes. Results from these preliminary analyses were incorporated into analytic memos, which were discussed in weekly analysis meetings. Other team members considered the relevance of emerging themes to their own transcripts and reflected upon the similarities and differences within and between the twelve sets of interview data. Overlap between themes was identified and, where appropriate, themes were collapsed in the interests of parsimony. A small number of themes was discarded at this stage because they were supported by only a small amount of data.
A coding framework was developed from these twelve case studies and then applied to the remaining data. The framework was amended to take account of data that did not fit. The revised coding frame was then applied to all the interviews. Data relevant to each theme were extracted and analysed, following the principles of the constant comparative method (Glaser and Strauss, 1967). Data were compared across different informants and across the same informants at different times. Comparisons were also made by occupational class and age groupings and by both intended and actual feeding practices.

Findings
In common with women interviewed for large-scale surveys (eg Foster et al., 1997), these women all reported breaking at least some of the expert-defined rules of infant feeding. For example, 83% introduced formula before four months. Ninety-two per cent gave solid food by the same age. In this very practical sense, women frequently and successfully resisted attempts to govern their behaviour 'at a distance' through technical expertise. Unlike women in childbirth (Graham and Oakley, 1986;Katz Rothman, 1982;Oakley, 1981Oakley, , 1984Oakley, , 1986Romalis, 1985;Kitzinger, 1992), these mothers were only rarely co-present with health professionals and this enhanced their freedom to depart from expert recommendations. In this physical sense, expert surveillance of feeding is episodic rather than continuous.
As reported elsewhere (Murphy, 1998a), many departures from expertdefined 'good practice' were designed to allow the women to juggle the various competing demands on their time and energy. They supplemented or replaced breast with formula milk to encourage the baby to sleep through the night or to ease the burden of being constantly 'on-call'. They introduced solid foods earlier than recommended to reduce the frequency of feeds and encourage progress. They gave snacks and treats to reinforce good behaviour or quieten a fractious or distressed baby.
The fact that, in practice, mothers do not always follow expert advice, does not mean that they are unaffected by the quiet coercion of normalising judgements or impervious to the panoptic gaze of bio-power. While mothers may depart from the medicalised advice of scientific experts, such departures do not take place in a morally neutral environment Murphy, 1999). The most powerful effect of the medicalised discourse is the attitude of self-policing it engenders in mothers as they let it play upon their selfassessments of their own feeding practices. Departures from expert advice raise questions about the legitimacy of the mothers' practices and this was reflected in the mothers' interview accounts. 2 For those who either had already engaged in feeding practices that ran contrary to expert advice, or who intended to do so, the interviews were an occasion when their practices and their moral status were both exposed to scrutiny and where, as a result, 'presentation of self' (Goffman, 1959) was likely to be an issue. 3 I shall show how mothers both engaged with and resisted the normalising discourse of medicalised scientific expertise and how they offered counterdiscourse within which their own feeding practices were entirely legitimate. This counter-discourse can be understood as a rhetorical strategy of resistance. They accorded particular salience to the notion of expertise in their accounts 4 while, at the same time, they subverted it. The women's resistance did not involve mounting a full-frontal attack upon expert knowledge. Rather, they actively endorsed the centrality of expertise while, at the same time, they redefined and relocated it.
Discussion of the roles of mothers and professionals (in particular health visitors) was central to the women's talk. The longitudinal nature of the data collected permitted consideration of changes in the women's orientation to health professionals' expertise over time. In the next section, I describe and discuss how women engaged with technical expert knowledge in the antenatal interviews. The subsequent section considers the transformation of their talk in later interviews as they sought to establish the legitimacy of practices that departed from the prescriptions of scientific experts.

Mothers and expertise before the births
Antenatally, almost all mothers presented themselves as novices, with minimal knowledge of infant feeding. They looked to health professionals for expert advice about all aspects of feeding, actively aligning themselves with the ideology of scientific motherhood outlined above (Apple, 1994;. One woman said, With the health visitor and the midwife I'm not sure how that works. Actually who comes for what. I think one comes for so many days, but I think they're the people I'll ask if I don't think the baby's getting enough food, or you know if I'm not happy with it. Yes, it'll be those I'll ask I think rather than anybody else. (Older mother, occupational class 3) At this stage, knowledge about infant feeding was conceptualised as a set of prescriptive rules held by health professionals, by virtue of their expert status, and which mothers, as novices, must elicit. One woman described her 'ignorance' of infant feeding and her dependence upon experts to explain the 'rules' to her: I want to ask the midwife whether you can express 5 it [milk] and you can give the baby it you know cold. I don't know whether you can or you can't. Before their babies' births, most mothers constructed infant feeding as an expert domain, governed by scientific and technical knowledge, invested primarily in health professionals. Here the knowledge relevant to infant feeding was characterised as a set of readily formulated rules capable of being disseminated to mothers by technical experts. Health professionals were experts and mothers were novices. Professionals would communicate the 'rules' of infant feeding and mothers would implement them. One woman remarked, 'You do want somebody to tell you what to do' (Younger woman, occupational class 3).
The women anticipated conflict between advice from professionals and from lay advisers such as mothers and siblings. While they did not exclude the possibility of seeking advice from lay networks, they treated professional advice as authoritative. While relatives and friends might have knowledge derived from their own experience, health professionals were expected to have knowledge of universal relevance.
The trouble with your family and friends is that they haven't done it, or they have only done it once, they don't know all the troubles that you're suffering if they didn't suffer themselves, so it would probably be the midwife or the health visitor. (Younger woman, occupational class 4/5) Thus, at this stage, practical, experience-based knowledge was treated as inferior to professional, technical expertise. Another mother said, I'll definitely speak to someone professionally. I wouldn't sort of ask anybody I knew who'd done it . . . they would only know what they'd done.
(Younger mother, occupational class 3) Deference to professional expertise was characteristic of almost all the women, across all occupational groups and in both age groups, when they were interviewed antenatally. Boundaries between mothers and professionals were treated as relatively straightforward. There is little evidence of tension between maternal autonomy and the medicalised discourse here. The women presented themselves as consumers freely choosing to use the technical expertise of professionals. By aligning themselves with the medicalised discourse, these women presented themselves as responsible citizens making active, well-informed and legitimate choices about how their babies would be fed. In Foucault's terms, they can be understood as 'docile bodies', responding to the quiet coercions of disciplinary power. As we shall see, such easy endorsement of professional expertise did not survive when women moved from anticipating feeding their babies to actually doing it. After the births, accounts of expertise became more complex. For the moment, they were sufficient for the mothers' practical purpose of representing their feeding intentions as reasonable and morally defensible.
There were, however, three women (one in the lowest class group and two in the highest) who actively resisted the dominant medicalised discourse before the births. They directly challenged both the legitimacy of professional attempts to direct their feeding and the validity of the advice given. For example, one woman, who intended to formula feed, categorically denied both the factual basis and the legitimacy of expert scientific assertions that breast was better than formula milk.
There's the same nutrients in both [breast-milk and formula-milk]. They say breast-milk is better because it's yours and it's nature -there's definitely the same nutrients in both . . . I don't care what everybody else says, it's what I feel is better for my child. Whose child is it when all's said and done? It's what I feel is better for my child. (Younger mother, occupational class 4/5) In this excerpt, the mother attacked the medicalised expert discourse around breast feeding on several counts. On the one hand, she asserted her autonomy. She suggested that attempts to direct her feeding practices were overstepping the mark by arrogating her rights as the child's mother. She anticipated the counter-challenge that such autonomy should be exercised responsibly and that this required her to implement expert advice. She challenged such expert advice at two levels. The first involved a flat rejection of its content. 'There's definitely the same nutrients in both.' The experts were simply mistaken. The second involved arguing that the knowledge disseminated by professionals, was inappropriate to the task at hand. It involved applying rules, derived from technical knowledge, when what was required was a judgement based upon maternal feelings and intuition. 'It's what I feel is better for my baby.' By virtue of her relationship to the baby, she could and should be trusted to know what was in the best interests of the child. Thus she relocated relevant expertise in mothers rather than in professionals. She skilfully presented her intended feeding practices as legitimate, even though they did not match the prescriptions of professional experts. In doing so, she resisted the discourse of technical expertise and presented herself as a good Expertise and forms of knowledge mother exercising her right to make well-informed choices in her child's best interests. In the light of Foucault's caveat that, even where individuals produce counter discourses, they are constrained by the social and cultural context in which they are operating, it is notable that this woman's resistance to the medicalised discourse around infant feeding is, nonetheless, expressed in terms of the baby's needs rather than her own preferences or convenience. As Woodhead (1997) has argued, the construction of children as the bearers of needs has profound moral implications for mothers, in particular, who are identified primarily responsible for meeting those needs (Phoenix and Woollett, 1991). Good mothering is defined in terms of meeting children's needs (Lawler, 1999(Lawler, , 2000Ribbens McCarthy et al., 2000).
The three women who actively challenged the authority and expertise of health professionals, at the antenatal interviews, all intended to formula-feed. This decision placed them in moral jeopardy and raised questions about the legitimacy of their choices (Murphy, 1999(Murphy, , 2000. Their accounts were complex and elaborate compared to those of mothers intending to breast-feed, in line with professional advice. Formula feeding involves overt rejection of professional advice. It brings the mother into potential conflict with her co-actors (midwives and health visitors). It is not surprising, therefore, that such women actively clarified the boundaries between their own sphere of expertise and authority and that of health professionals.

Mothers and expertise after the birth of the babies
For most women, however, such active boundary clarification was delayed until after the births. When actors undergo social transitions, their field of relevances changes and, along with it, the degree of elaboration of their accounts (Schutz, 1964). Transitions, like crises, interrupt 'the flow of habit and [give] rise to changed conditions of consciousness and practices' (WI Thomas, cited in Schutz, 1964: 96). Accounts appropriate to one situation may be inadequate in another. Greater clarity and specificity about the standing of different actors and the boundaries between them may be needed. Matters inconsequential in one situation may become highly salient in another. Such transformations are particularly likely to occur when an individual becomes an active participant rather than an onlooker. In a theatrical metaphor, Schutz (1964) compared this to 'jumping from the stalls to the stage' (p. 97).
For first-time mothers, birth represents such a major transition. They move from anticipating caring for their babies to actually doing so. It is a time when they begin to confront the challenge of balancing the various projects associated with infant feeding which were described above. As we have seen, professional preoccupation with securing the short-and long-term health of babies is associated with a set of prescriptive rules for feeding babies. If fully implemented, these rules limit mothers' freedom to negotiate feeding practices which are flexible enough to allow them to juggle these various projects.
As we have noted, almost all the women interviewed for this study did deviate from these expert rules in some way. It is clear, therefore, that they can and do resist professional prescriptions. However, given the force of scientific expert discourses, such deviations risk the characterisation of mothers' practices (and possibly mothers themselves) as irresponsible and illegitimate. As we shall see, the women interviewed for this study actively resisted such characterisation. They offered a counter-discourse which endorsed the centrality of expertise, while, at the same time, redefining and relocating it. It is to the rhetorical strategies they used to accomplish this that I now turn.
After the births, the women redefined appropriate roles for mothers and professionals, with particular attention to the knowledge and expertise possessed by each group. Occasionally, this involved outright rejection of professional expertise and involvement in infant feeding. For example, one woman had stressed the importance of professional advice when interviewed antenatally, saying, It's always better to get a professional, professional answer than someone who thinks they know. (Younger mother, occupational class 4/5) Thus, before her baby's birth, she treated lay knowledge as potentially unreliable and self-deluding. However, two months after her baby's birth, the same woman rejected all professional involvement or advice. She said, I've just gone and done it (made decisions about what and how much to feed the baby) without asking the health visitor or the midwife. The only way you learn is by doing it yourself. Before I had him, I just sat there and read books but I haven't related to the books since I had him. I just went and done it.
Postnatally, this woman contrasted her own practical, experience-based and trustworthy knowledge with that of professionals who 'just read it out of books'. She suggested practical rather than technical expertise was required for successful feeding. In asserting the primacy of practical knowledge, she denied health professionals any legitimate role. By dint of practical experience, only she was qualified to determine how her baby should be fed. In her interview talk, she actively engaged in setting boundaries to the legitimate sphere of professionals.
At the other extreme, there was just one woman who continued to represent health professionals as holding a specialised body of technical knowledge that should determine her feeding practices in every detail. Two months after the birth, she described how she would submit every decision to professional scrutiny: When I think he needs to move on like I'll ask when it's best for him to go on to solids. I know I've got the book and everything but it's nice to get it Expertise and forms of knowledge from the doctor or the health visitor isn't it, so you're 100% sure you know that what you're doing is right. (Older mother, occupational class 4/5) Two months later, she described her baby's eating pattern: Yeah, he's on a tablespoon now because before he was on two teaspoons but his growth and weight was very dramatic when we last weighed him, so she [health visitor] says, 'Oh, just give him two teaspoons twice a day' and this time she weighed him, he's just gone up a pound, so she says, 'you can increase his food now'. She told me how much to give him.
The use of language normally reserved for pharmaceutical dosages and/or recipe books underlines the construction of infant feeding as a technical matter. This women relied 'absolutely' upon her health visitor's opinion 'because you know, this little lad Jonathan 6 is relying on me and I've got to rely on their expertise'. Health visitors had a monopoly of relevant knowledge. Feeding could be reduced to a set of rules which professionals would communicate to mothers and which mothers had a moral duty to implement. Here the primacy of technical knowledge dictated the legitimate boundaries between professionals and mothers. The mother's duty was to refer all feeding decisions to professionals for resolution. This mother presents herself as freely choosing to submit all her feeding practices voluntarily and conscientiously to the detailed prescriptions of health professionals, who act as conduits for the dissemination of scientific expertise. As such, she could be seen as an archetypal docile body produced by the disciplinary practices associated with the medicalised discourse around infant feeding. For her, there was no apparent need for elaboration of the account advanced at the antenatal interview or any re-negotiation of the boundaries anticipated then.
However, most mothers, once their babies had been born, fell somewhere between the two extremes outlined above. As they moved from anticipating feeding their babies to actually doing so, the accounts they advanced became more complex and sophisticated. Knowledge about infant feeding was no longer treated as undifferentiated. Rather, mothers identified different kinds of expertise, with different degrees of relevance to infant feeding practices. Practical knowledge, born of experience, was now favoured over technical knowledge formulated in textbooks: There's more behind you than a professional that doesn't know you . . . I mean my midwife . . . she hasn't got any children. So although they can read it in the books and she might see it everyday, from actually seeing it and doing it there's a lot of difference . . . because they're professionals at it, or supposed to be, it doesn't mean they're qualified. . . . There does seem to be a lot of people saying 'Well, we're the professionals at this and you do as I say, and I'm thinking, 'Sod off' (Older mother, occupational class 4/5). Professional, scientific expertise or, as one mother put it, the idea that there was 'any one right way to do things' (older mother, occupational class 1/2) was criticised on a number of grounds. The women pointed to its failure to adapt to the individuality of particular babies. One mother condemned the naiveté of such thinking: They think, you know that all babies are the same, and that they all grow at the same rate, but it don't work, especially with Mary. (Younger mother, occupational class 4/5) Whereas, antenatally, technical knowledge had been valorised, increasingly mothers highlighted the relative importance of practical, individualised knowledge. In doing so, they legitimated their own claims to decide how their babies should be fed, even when their decisions ran counter to the prescriptions of scientific experts. Expertise continued to be an important issue around which women's talk pivoted, but it was redefined and, as a result, relocated. Expertise relevant to infant feeding was now based on individualised knowledge of a particular baby. Such knowledge was, by definition, invested in the person who had day-to-day care of the baby.
This distinction between different kinds of knowledge echoes Mayall and Foster's finding that mothers prioritised 'experiential' knowledge, whereas health visitors emphasised 'book' knowledge (Mayall and Foster, 1989). This recalls Oakeshott's distinction between 'knowledge of technique' and 'practical knowledge' (Oakeshott, 1991: 12). Knowledge of technique can be formulated precisely as a set of rules. It can be written down and directly communicated to those who do not possess it. By contrast, Oakeshott argues, practical knowledge 'exists only in use, is not reflective and (unlike technique) cannot be formulated in rules' (p. 12). Oakeshott's definition of knowledge of technique certainly fits with the medicalised discourse around infant feeding. However, his definition of practical knowledge differs in subtle but significant ways from the practical expertise advocated by mothers in this study. In particular the practical expertise these mothers invoked was highly reflective. Indeed, it appears from these data that the process by which mothers arrived at the practical knowledge that underpinned their feeding was considerably more reflective than their earlier engagement with medicalised discourses.
This differentiation between different kinds of knowledge and the redefinition of the expertise relevant to infant feeding, as that which was grounded in practical experience of individual babies rather than that derived from scientifically-based expertise, can be understood as a rhetorical strategy of resistance. It allows mothers to claim legitimate control of their own feeding work and, at the same time exhibit conformity to the liberal imperative of expert-led practice.
Mothers also dismissed attempts to apply technical rules to infant feeding as unrealistic and impractical. They argued that good infant feeding practices Expertise and forms of knowledge recognise that practical contingencies necessarily modify rules. For example, one mother described her difficulty in implementing the 'rule' of universal breast-feeding: I felt so under pressure that week . . . a terrible lot of pressure was put on me to breast-feed and it just wouldn't work . . . she [the midwife] came back two days later . . . the look on her face, as if to say, 'You haven't tried hard enough'. . . . You can't starve a baby saying 'Oh you've got to have it'. (Younger mother, occupational class 1/2) This mother challenged professionals' advice as narrowly theoretical, based upon a set of supposedly generalisable rules, which ignored the reality of feeding her baby. Later she rejected advice that her baby was sleeping too long between feeds, saying, I mean there wasn't a lot we could do. I mean if she's not hungry, she won't have it.
She presented her baby as an active agent, capable of accepting or rejecting her mother's offers of food, rather than the passive recipient of expert-led feeding practices.
Certain kinds of technical knowledge were also seen as at odds with 'common-sense'. One woman criticised what she saw as a doctrinaire commitment to theoretical knowledge. She told an 'atrocity story' (Webb and Stimson, 1976) about her struggle to persuade the hospital midwives to modify their intransigent commitment to doctrinaire rules. Such atrocity stories are likely to be found 'wherever attempts are being made to control aspects of the life of some group by others whose justifications for such attempts are illegitimate' (Dingwall, 1977: 377). For reasons of space I can only reproduce a small portion of this story here. This mother described how, shortly after the birth, concern arose about this baby's low glucose levels, leading to a decision to supplement breast with formula milk. The midwives were reluctant to use a feeding bottle, because of a reported correlation between bottlefeeding and early termination of breast-feeding. Instead, they insisted on giving formula milk by cup. The mother described her ensuing dispute with hospital staff.
They categorically did not want to give her a bottle and I really wasn't in a fit state to argue the point because her blood sugars were very low and they were trying to get fluids into her and I would have thought the easiest way to get fluids into her would be by a bottle. So they kept trying to cup feed her and this went on for a couple of days. It was terrible. What didn't make sense was that they couldn't see that cup feeding was ineffective. Her blood sugars were not rising so when I sort of came round [following traumatic delivery] I said, 'Well, can she have a bottle because I'm sure she'll suck a bottle? And they said, 'No, no' and they continued with this cup feeding and I continued to say, 'Can she have a bottle?' and eventually she got a bottle and they wrote in my notes that they weren't responsible for offering me a bottle. (Older mother, occupational class 1/2) She went on to contrast the intransigent commitment to doctrinaire theories among these midwives with the common-sense realism of another, There was a night midwife, who'd had 28 years in the job or something and she said, 'I believe we're offering your baby a cup' and I said, 'Well I'd rather have a bottle and she said, 'Well, I'm glad you said that because I've never heard so much rubbish in my life. I've been a midwife for 28 years and I mean, babies, I've seen more babies go from breast to bottle to breast than I've had hot dinners and you know I don't know what all this fuss is about. She gave me a bottle, no problem. She laughed and said, 'Don't tell the day staff'.
Many mothers compared professionals' technical knowledge with their own special (and more useful) practical knowledge. One woman, when interviewed two months after the birth, explained why she no longer consulted her health visitor, To be quite honest with you, I didn't bother asking her anything because I felt she was pretty useless (laughs). I'm being honest. Obviously you know your own baby and she was having a bottle when she [the health visitor] came and she's always messed about and she's never, she's never taken a feed all in one go and the midwife turned round and she says, 'Well I don't think you should've put her on bottles because she's not taken to it very well'. She seemed to be, she wanted to bring her up more by the book than what suited. She used to say to me, 'Well you know you really shouldn't be giving her that much, she doesn't need it, and whatever else and I turned round and said, 'Well I think she does. She takes it. She needs it. She knows when she's had enough.' (Younger mother, occupational class 4/5) Here, again, the mother identifies practical expertise, adapted to the individual baby's needs, as central to successful infant feeding.

The location of practical and technical knowledge
In the previous excerpt, practical expertise was firmly located within the maternal domain. However, not all mothers who distinguished between technical and practical knowledge located the former exclusively in health professionals or the latter exclusively in mothers. Some questioned whether professionals did possess accurate technical knowledge. They acknowledged the existence of a secure body of technical knowledge about feeding but doubted the competence of some professionals to communicate it. For example, one woman described the 'ignorance' of some hospital midwives about breast-feeding. In what again amounted to an 'atrocity story', she described the practices of midwives who 'helped' her to initiate breast-feeding: What they tended to do was sort of to grab her head and your nipple and sort of shove them on. Which, as I say, I didn't really know any better, but I've since, on reading about it, I've since discovered that you know that's the worst thing you could do because they respond to the touch at the back of the head and turn towards it. (Older woman, occupational class 1/2) This woman endorsed the role of technical knowledge but did not necessarily attribute such knowledge to health professionals, who advocated 'outdated practices.' She presented her earlier naive reliance upon the expertise of the midwives as understandable in a novice mother. Now, however, she had progressed beyond an assumption that health professionals are experts by virtue of their status. She now saw herself as responsible for 'sifting' advice to determine both its validity and relevance. Rather than passing on sound, theoretical, expert knowledge, some professionals were now held to base their advice on idiosyncratic, personal opinion: The two midwives that visited me were again sort of 'I've got a piece of unique advice' (laughs) and they offered it. One of them said, 'try tea bags in your bra to toughen up your nipples.' The other said cabbage leaves or something. Pieces of advice that, which you know weren't really very useful at all, but fascinating. 7 In this case, professionals were presented as having a third kind of knowledge, which was neither technical nor practical, but akin to 'old wives tales'. This mother claimed that her own technical knowledge was superior to that of at least some of the professionals who were supposed to be advising her.
Just as mothers did not necessarily present technical knowledge as the exclusive preserve of professionals, neither did they suggest that only mothers possessed practical knowledge. They differentiated between professionals with and without practical knowledge. One mother contrasted different professionals: If I do decide to introduce a few more solids in the near future I don't think I will really go and talk to her [health visitor] because all I'm gonna get is, 'Well you shouldn't do it yet. If she was more like my midwife. I mean she, she'd say, she's been a midwife for years and she's had three children and she'd say, 'Well, I'm supposed to tell you and you're supposed to do it like this, but she'd say, 'I did it like this and that worked'. And she says, 'The advice I'm supposed to give you is rubbish'. She says, 'It don't work and I don't agree with giving it', but she says, 'I've told you but I advise you to do this'. She was really good in that sense that she's been a mum herself. She knows that things don't work that midwives tell you and health visitors. (Younger mother, occupational class 4/5) Professionals who made suggestions, while accepting that such advice might not work in practice, were particularly valued. One mother said, I've got a lovely health visitor now and she just say[s], 'Well, yeah if that's what you're happy with you do that. Other mums do this. My new one [health visitor] seems to ask me more what I'm doing and why I'm doing it and she seems to accept the fact that I've got me own little ways and she thinks they're alright. Obviously, as long as we're happy she's quite happy to go along with it. (Younger woman,occupational class,4/5) This data extract illustrates the significance to mothers of retaining control over feeding work. Professionals' contributions were tolerated, and even welcomed, where mothers' autonomy and their freedom to manoeuvre were preserved. One problem with technical knowledge is its categorical nature. When knowledge is formulated as a set of rules, rule breaking becomes a central issue (Murphy, 1999). Such rules can be seen as an attempt to control mothers' feeding work practices under highly indeterminate working conditions, where feeding work may have to be juggled with other responsibilities (Murphy et al., 1998a). Practical knowledge is more flexible and can be formulated and re-formulated by health professionals and mothers alike.

The shifting relevance of practical and technical knowledge
Mothers did not simply dismiss technical knowledge. Rather, they sought to re-draw the boundaries between circumstances in which technical knowledge was or was not useful. It was particularly valued when something abnormal occurred. Thus, for example, one mother who criticised much professional advice about infant feeding as insufficiently adapted to the needs of her particular baby, nevertheless described her response when her baby vomited: She was sort of feeding okay and then, after you'd put her down in her Moses basket she'd puked up everywhere and then she'd started mucking about with her feeds and she just kept puking up all the time and we didn't have a clue what was the matter with her so we phoned the health visitor up and she came round I mean you don't know when it's your first do you? (Older woman, occupational class 3) She distinguished between areas which demanded practical expertise, sensitive to the characteristics of her individual baby, and those which required technical knowledge. While no longer a novice in 'normal' infant feeding, she nevertheless needed direction from technical experts when something she defined as problematic occurred.
This redefinition of boundaries between health professionals and mothers was highlighted in the final interviews, two years after the births, when mothers were asked about professional input if and when they had another baby. 8 In reply, almost all contrasted their previous inexperience with their current situation. One said, I don't think they [health visitors] would [have much involvement] because I think second time around you probably know a bit more yourself and I didn't think that really I got much help the first time when I didn't know anything. So now, I think, knowing a little bit more than I did first time around I don't think that I'd really ask them for any advice at all unless I came across a particular problem that I didn't have with Caroline. (Younger mother, occupational class 1/2) Experience allowed this woman to dispense with professional advice. It will only become relevant again if something occurs outside the boundaries of that experience.
Another specifically related her reluctance to a preference for practical, experience-based knowledge rather than scientific, rule-based knowledge. I wouldn't necessarily take her [health visitor's] advice. I mean I'd be more inclined to take advice from those that have been through it . . . now she [the health visitor] has her own children but obviously she also has to adhere to certain code of practice. Well, that's neither use nor ornament to me.' (Younger mother, occupational class 3) In this case, the mother suggests that the health visitor probably does have useful knowledge drawn from her own personal experience of child-rearing. However, the mother cannot access this practical knowledge because the health visitor will be constrained by her professional position to ignore this and only communicate rules that are derived from scientific expertise.
These mothers now differentiated clearly between two roles. First, there was the mother's role in making routine feeding decisions and identifying any problems. Second there was the professionals' role in responding to motheridentified problems. The relegation of health professionals to the role of responding to mother-defined problems has been reported in other studies (McIntosh, 1986;Mayall and Foster, 1989;Heritage and Sefi, 1992). This division of labour allowed mothers to exercise autonomy in their everyday feeding practices while consulting professionals about problems they could not solve using the practical and technical expertise developed in feeding their first child. These women presented mothering as highly dynamic. They explicitly discussed their transformation from novice to expert status. One woman described this transformation, I've got to the stage now where I'm more confident in evaluating advice and then considering whether or not I want to take it. (Older mother, occupational class 1/2).
The women talked of reliance on professional advice about infant feeding as a 'career stage' they had outgrown. They now presented themselves as competent, independent agents who evaluated advice from a range of sources, and then made and implemented decisions about appropriate action.

Discussion
This paper examines the ways in which medicalised discourses are incorporated into the state's attempt to govern mothers' feeding work 'at a distance' and the ways in which mothers deal with this attempt. In relation to this area of mundane family life, mothers are invested with the moral and practical responsibility for making prudent choices on behalf of their babies, ensuring their present and future health and welfare. They are granted formal autonomy insofar as, except in extreme cases, they are free from the direct or coercive interventions of external agents. However, as we have seen, mothers' feeding choices are not unconstrained. They are made in the context of a powerful medicalised scientific discourse about how children ought to be fed. Such expertise is central to the apparatus of rule (Rose, 1992). The fact that this discourse has been translated into a set of categorical rules, which mothers can be seen to uphold or break, limits the mother's room for manoeuvre. The 'choice' (Murphy, 1998b) to feed one's baby in ways contrary to expert advice raises questions about the legitimacy of mothers' practices.
Like most mothers, those interviewed for this study frequently and routinely broke the 'rules' of infant feeding specified by scientific experts and disseminated through the state-sponsored apparatus of health professionals. As such, they actively resisted attempts to control their working practices. They sought to set boundaries to the activities of others including health professionals. However, such boundary setting was problematic for the mothers, given the powerful medicalised discourse to which they were exposed. Claims to expert knowledge are an extremely effective means by which the ascendancy of one group over another can be established.
Initially, most mothers appeared content to be 'governed at a distance'. They endorsed their own subordination to the dictates of scientific expertise. However, as they passed from anticipating feeding their babies to actually doing so, they began to assert the legitimacy of their own control over their feeding work. The centrality of expertise to contemporary liberal government Expertise and forms of knowledge is underscored by the extent to which these mothers appeared constrained to mount a rhetorical defence against the claims of others to determine how their babies should be fed.
The dominance of expertise in contemporary liberal societies can be seen as a problem for mothers who wish to retain control over their own feeding work in the face of contrary claims by other groups. Like other family matters (Morgan, 1985;Ribbens, 1994), infant feeding has increasingly been subject to the assertion of the primacy of medicalised knowledge (Apple, 1995;Apple, 1994). Those who engage in 'risky' behaviours (whether on behalf of themselves or of others for whom they are responsible) are in moral jeopardy (Lupton, 1993;Greco, 1993;Carter, 1995;Murphy, 2000). Where risk is defined within and through expert scientific discourses, mothers who engage in 'risky' feeding practices are faced with the challenge of maintaining their identities as 'good mothers'. The mothers in this study dealt with this potential threat to their identities by subverting the notion of expertise. Rather than challenging the centrality of expertise, they endorsed it, reinforcing the idea that infant feeding was an expert realm, in which only experts could legitimately decide how babies should be fed. In doing so, however, they creatively redefined and relocated expertise. They defined the expertise relevant to infant feeding as that to which they, as the children's mothers and everyday carers, had privileged access. They established themselves, rather than health professionals, as the primary bearers of relevant knowledge. This reorientation of the discourse around expertise allowed women to claim legitimate control over their feeding work which, in turn, allowed them the flexibility to juggle the demands of feeding their babies and their other obligations at home, in public and in external work settings.
By subordinating technical knowledge to practical knowledge, these women could preserve and promote their own control over their work while maintaining their identities as good, responsible mothers. By privileging practical and individualised knowledge, they presented themselves rather than professionals as 'experts'. As they jumped 'from the stalls to the stage' of motherhood, development of this counter-discourse of expertise allowed them to combine flexibility in their working practices with claims to responsible, prudent, knowledge-based motherhood. Trumping professionals' claims to technical knowledge with their own claims to practical knowledge permitted mothers to present their infant feeding practices as legitimate and attempts at professional control as illegitimate. It is noteworthy, however, that none of the counter-discourses around infant feeding offered by the women challenged the core assumptions that mothers bear the primary responsibility for ensuring the adequate nutrition of their infants and that good mothers prioritise their babies interests. As Foucault suggests, the acceptability of counter-discourses is always constrained by the social and cultural contexts in which they are developed.
In many ways, the women's resistance to the discourses of professional technical expertise is best seen as embodying rather than challenging the rationality of contemporary liberalism. Within this rationality, the responsible and moral actor is not the one who conforms blindly to expert recommendations. Rather (s)he is expected to subject such recommendations to evaluation and questioning, operating as an informed consumer. In defending their feeding practices as legitimate, the women make use of the subject position of entrepreneurial citizen that is characteristic of contemporary liberalism. Much of their talk is directed towards demonstrating that they have acted responsibly by subjecting the advice offered to critical appraisal and that their actions are the prudent outcome of such deliberations.
This paper is concerned with the dynamics of everyday power relations and the regulation of the private sphere in a contemporary liberal society. It focuses upon the regulation of women's bodies and, as such, is part of a longstanding and tradition of feminist research and scholarship which examines the ways in which areas of women's lives (including fertility, pregnancy, childbirth and motherhood) are annexed and dominated by predominantly male experts (see, for example, Ehrenreich and English, 1977;Rich, 1976;Gordon, 1977;Oakley, 1981Oakley, , 1984Scully, 1980;Corea, 1985;Kitzinger, 1992). These writers show the myriad ways in which women have been constrained to collaborate in their own subordination, not least by being made to feel guilty about risking their babies' health if they do not conform to advice from medical experts. Given the numerical dominance of men among the cadres of scientific experts and obstetricians (Pringle, 1998), these attempts to annex and control women's bodies can be understood as a manifestation of patriarchal domination of women.
Such analyses of the medicalisation of women's lives have highlighted the crucial role played by medical and scientific expertise in perpetuating the subordination of women in contemporary society. They have pointed to the ways in which medicalised discourses have limited women's choices and to their social and political implications. This paper is a contribution to that tradition of research. However, more recently, a number of feminist authors have begun to identify some of the limitations of orthodox medicalisation critiques and to point to the ways in which they could usefully be informed by a Foucauldian perspective. For example, Lupton (1997) argues that traditional medicalisation critiques tend to portray patients as somewhat helpless and passive victims of medical dominance and fail to do justice to their capacity for challenging and resisting such dominance. Similarly, Sawicki (1991) points to the way in which radical feminists' 'demonisation' of medical technologies, and the men who design and apply them, has distracted from women's capacity for struggle and resistance. A Foucauldian perspective, such as that adopted here, with its understanding of power as relational rather as a possession of the dominant class (Sawicki, 1991) is better suited to the exploration of women's resistance that is the focus of this paper.
Furthermore, some orthodox feminist critiques have tended to treat medicalising practices and discourses as patriarchal inventions that have then been cascaded down to the micro-level of society. By contrast, Foucauldian analyses operate with an 'ascending analysis of power' (Foucault, 1980b): One must rather conduct an ascending analysis of power starting, that is, from its infinitesimal mechanisms, which each have their own history, their own trajectory, their own tactics, and then see how these mechanisms of power have been -and continue to be -invested, colonized, utilized, involuted, transformed, displaced, extended etc by even more general mechanisms and by forms of global domination. (Foucault, 1980b: 99) As we have seen, midwives and health visitors play a major part in disseminating medicalised discourses around infant feeding to mothers. Given the role such dominant discourses play in the patriarchal disciplining of women, and the extent to which they distract from the material conditions under which women carry out their feeding work, it is perhaps ironic that two largely 'female' occupations are so heavily implicated in their transmission. This is reflected in some of the highly critical remarks mothers made about health visitors and midwives that are reported above. However, it is important to note that midwives and health visitors do not themselves stand in a position of exteriority to the dominant discourses they transmit. Rather they, like the mothers they advise, are caught up in the disciplinary technologies to which they contribute. Of course, again like the mothers themselves, they are capable of resisting dominant medical discourses and there is evidence in the women's accounts that many do in fact do so. Indeed, as we have seen, some of the mothers categorise health professionals as those who reproduce prescriptive medicalised discourses (unhelpful) and those who share their practical experiences of child-rearing (helpful). Furthermore, mothers appear to recognise that, even when at a personal level, individual midwives and health visitors may question elements of the medicalised discourse, they are, nevertheless, constrained by their contractual obligations to transmit them to mothers.
The data from this study illustrate the tension within contemporary liberalism between a respect for the autonomy and privacy of individuals and a concern to regulate social and economic life. 'Governing at a distance', through the choices of individuals who have been made responsible for themselves and for those under their control and protection, is a key means by which the state attempts to deal with this tension. Such 'governing at a distance' occurs at a number of levels, including the mundane activities of childrearing and family life. Expert knowledge is an important means through which the state seeks to regulate the choices of individuals while respecting the limits of government. Such expert knowledge is a powerful constraint upon the choices that individuals make. The recipients of such advice are not, however, mere 'docile bodies'. Those targeted by such advice may not be able to ignore it, but they can and do develop sophisticated and creative strategies for resisting attempts, based on such claims to scientific expertise, to determine how their work will be carried out. In so doing, they turn the discourse of expertise to their own advantage, re-establishing some freedom to manoeuvre in the process.

University of Nottingham
Received 30 October 2002 Finally accepted 1 September 2003