Images of Childhood in Mothers' Accounts of Contemporary Childrearing

This article examines how a sample of first-time mothers in the UK constitute childhood in general, and their own children in particular, in and through their talk about the mundane practices of childcare. The data analysed are drawn from a longitudinal qualitative interview study that followed a sample of mothers from late pregnancy until their babies were 2 years old. The analysis explores links between mothers' representations of childhood and the actualities of their childrearing practices. The extent to which both the representations of childhood and the practices of individual mothers are consistent or contradictory, and change or remain constant over time, is also examined.


Introduction
This article examines how mothers of young children constitute childhood in general, and their own children in particular, in their talk about childrearing. I use the term 'constitute' to underscore the way in which understandings of childhood and children are social constructs, located in particular forms of discourse. As such, concepts of childhood and of the child are not neutral, objective, natural 'givens' but are impregnated with theoretical assumptions (Jenks, 1996). They are 'knowledged into being' (Stainton Rogers and Stainton Rogers, 1992). Childhood is a social creation and, although 'truths' about childhood may appear to have the status of objective fact, in reality they reflect specific social and political preoccupations and concerns (Lawler, 1999). In a very real sense, children and childhood are formed, 'made up' or constituted through these theory-impregnated forms of discourse that operate within both the specialist disciplinary worlds of, for example, paediatrics, psychology and sociology, and also in the commonsense world. These forms of discourse are discernable in the ways in which children are represented in everyday talk. As Ribbens (1994) argues, such representations draw upon 'typifications' (Schutz, 1954), which are available

IMAGES OF CHILDHOOD IN MOTHERS' ACCOUNTS OF CONTEMPORARY CHILDREARING
This article examines how a sample of first-time mothers in the UK constitute childhood in general, and their own children in particular, in and through their talk about the mundane practices of childcare. The data analysed are drawn from a longitudinal qualitative interview study that followed a sample of mothers from late pregnancy until their babies were 2 years old. The analysis explores links between mothers' representations of childhood and the actualities of their childrearing practices. The extent to which both the representations of childhood and the practices of individual mothers are consistent or contradictory, and change or remain constant over time, is also examined.
in the sociocultural world for individuals to draw upon to make sense of the behaviour of others.
There is a substantial body of, largely feminist, scholarship focusing upon dominant ideologies and social constructions of mothering and motherhood (see, for example, Apple, 1995;Bernard, 1975;Everingham, 1994;Glenn, 1994;Hays, 1996;Hill Collins, 1990;Lawler, 2000;Miller, 2005). However, motherhood as a social category is formed on the basis of the prior social category of childhood (Lawler, 1999;Miller, 2005;Urwin, 1985;Ribbens McCarthy et al., 2000). As Lawler has argued, 'Motherhood is constituted on the basis of, and as response to "truths" about the nature of the child' (Lawler, 1999: 65). Ideologies of motherhood are therefore predicated upon knowledges of childhood, and, in particular, of the needs of children. The assertion that something is a 'need' has a powerful impact upon the audience. It implies a factual basis and induces a sense of responsibility or guilt in those who fail to respond (Woodhead, 1997). Within contemporary discourses around motherhood, children are formulated in terms of their neediness and mothers are positioned as the primary meeters of such needs (Everingham, 1994). Indeed, the adequate meeting of children's needs is one of the defining features of good motherhood (Duncan and Edwards, 1999). As a result, the needs of mothers are effaced, either by insisting that their needs are identical with the needs of their children or by denying that the mother's needs are needs in the true sense of the word (Lawler, 1999). Thus, children's needs are granted absolute priority and mother's needs claims are pathologized.
A recognition of the social construction of childhood is therefore of vital importance to discussions of motherhood. Nevertheless, as Lawler (1999) points out, relatively little attention has been paid, in such discussions, to the social constitution of the categories of 'child' and of 'children's needs' (but see Alwin, 1990;Ribbens, 1994). As a result, the ways in which knowledges about 'the child' contribute to the (largely non-coercive) subjection of mothers are overlooked. Data collected as part of a longitudinal qualitative interview study of mothers' choices about feeding their young children (Murphy, 1999(Murphy, , 2000(Murphy, , 2003(Murphy, , 2004a(Murphy, , 2004b offer an opportunity to explore the ways in which, in relation to this mundane area of childrearing, mothers construct their children and call up 'knowledges' about children's needs. This article examines how mothers represent children and childhood in their talk. It also considers how the representations (or motifs and images as Jenks [1996] terms them) were related to childrearing practices, particularly in relation to infant feeding. The longitudinal nature of the data collected also allows us to examine how both mothers' constructions of childhood and their childrearing practices change over time as their children move from babyhood to early childhood.

Shifting constructions of childhood
A number of authors have examined historical trends in the social construction of children and childhood (Ariès, 1962;Cunningham, 1991Cunningham, , 1995Hendrick, 1997;Pollock, 1983;Shorter, 1976;Stone, 1977). Using historical data, they have challenged the assumption that childhood is a naturally occurring, self-evident phenomenon that is always and everywhere the same. These authors have linked changes in the ways in which childhood has been represented to political and structural conditions. There is general consensus that constructions of childhood have undergone significant shifts between the Middle Ages and the present day.
On the basis of an overview of historical research on changing constructions of childhood carried out by such authors, Jenks (1996) identifies two dominant ways of talking and thinking about children: the Dionysian and the Apollonian. He argues that these are best seen as 'codes' or 'images arising from mythology' that may, at different times, be 'supported and reinforced by various religious beliefs, political ideologies and scientific doctrines' (Jenks, 1996: 70).
The Dionysian image of childhood is defined by an assumption that children enter the world with a bias towards evil rather than good. They are drawn to pleasure and self-gratification and lack self-control or sensitivity to the others' needs. This image is rooted in the Judaeo-Christian doctrine of original sin and rejects any notion that children are innocent beings (Stainton Rogers, 2001). It was characteristic of 17th-century Puritanism and subsequently reinvigorated during the Evangelical Revival of the late 18th and early 19th centuries with its emphasis on children as 'sinful polluted creatures' (Hendrick, 1997: 40). Hannah More, founder of the Sunday School movement, warned against the 'fundamental error' of thinking of children as 'innocent beings'. Rather, she argued, they are 'beings who bring into the world a corrupt nature and evil dispositions' (quoted in Robertson, 1976: 421).
The model of childrearing conventionally associated with this Dionysian image of the child is one of strict discipline and parental constraint (Stone, 1977). Left to their own devices, children would turn out badly and, for their sake, parents must instil good habits from an early age. The metaphors associated with this parenting style are those of horticulture rather than natural growth: 'of preparing good soil, of rooting out weeds, of training young shoots in the direction you want them to go' (Cunningham, 1995: 48). The emphasis is upon control, regulation and discipline (Stainton Rogers, 2001). Jenks (1996) describes this approach to childrearing as 'one that saw socialisation as almost a battle but certainly a form of combat where the headstrong and stubborn subject had to be "broken", but all for their own good' (Jenks, 1996: 71). Jenks (1996) contrasts this Dionysian motif of childhood with the Apollonian, which, he argues, is the preferred modern, western, 'public' way of regarding the child: Such infants are angelic, innocent and untainted by the world which they have recently entered. They have a natural goodness and a clarity of vision that we might 'idolize' or even 'worship' as the source of all that is best in human nature. (Jenks, 1996: 73) Cunningham (1995) places the emergence of this second image during the 18th century. The implications for childrearing were, in principle, profound. Apollonian children were to be encouraged to express their natural goodness. Rather than being shaped, controlled and constrained, as in the Dionysian model, they are to be nurtured, protected and enabled to develop their own innate potential. Adults were responsible for providing the environment for children's own natural goodness and individuality to flourish. 'The art of child-rearing became one of hearkening to nature, giving free rein to growth rather than bending twigs to a desired shape' (Cunningham,1995: 62). Here the metaphor is of the prolific wild garden where seeds flourish naturally rather than of the strict horticultural regime required for the Dionysian child. Conventionally, these two images of childhood are presented in historical sequence, with the Dionysian preceding the Apollonian. However, while a significant move towards a more tolerant, child-centred approach is discernible from the 16th to the 20th centuries, Cunningham (1995) argues that motifs of childhood conventionally associated with premodern times, may, in fact, persist in contemporary understandings of childhood. Indeed, one might argue that the Dionysian image of the child, with its associated demand for parental control and constraint, is clearly discernible in recent UK policy developments such as parenting orders, which may, for example, require parents to ensure that their children are at home at certain times or do not visit specified places.
Stainton Rogers (2001) suggests that these two motifs (or, as she prefers, discourses) of childhood coexist and underpin diverse policy orientations towards children. Similarly, Jenks (1996: 70) argues that although these two images are 'competitive to the point of incompatibility', they nevertheless exist in parallel at the same time. There is, therefore, a danger of rendering constructions of childhood at particular periods more distinctive or patterned than they really were. The understandings of what it means to be a child, which are current in any given culture, at any particular historical period, may be more complex, contradictory and ambivalent than some historical accounts allow.
Other concerns have been expressed about the interpretation and implications of research into the construction of childhood. We are warned against perpetuating a triumphalist and/or presentist history of childhood, in which changes in understandings of the nature of childhood are interpreted as evidence of progress from unenlightened past to enlightened present (Jenks, 1996;Steedman, 1990). We are also reminded of the importance of examining empirically the relationship between images of childhood and concrete childrearing practices (Pollock, 1983). As Cunningham (1995) points out, these do not necessarily stand in a one-to-one relationship.
These considerations are central to the analysis presented here. I examine the extent to which different understandings of childhood coexist in mothers' talk and how they relate to practices in one central, highly contested but mundane aspect of contemporary childrearing, that of infant feeding. I examine differences in how children are represented by mothers and how these differences relate to childrearing practices. I also consider how both representations and practices change and develop over time.

Material and methods
The data are from an interview study of the choices that first time mothers made in relation to feeding their babies and young children, which was funded by the UK Economic and Social Research Council (ESRC). The overall aim of the study was to improve our understanding of the decisions mothers make in relation to feeding their babies and young children. From the outset, the ways in which the mothers constructed their children and their needs, and the extent to which they presented their children as legitimate participants in decisions about feeding were central to the research. Issues around childrearing, particularly those of parental authority and control, were identified, in advance, as important areas for investigation in the interviews.
The interviews were conducted in and around Nottingham, England. A cohort of first-time mothers was followed from late pregnancy until their babies' second birthdays. Following approval by the Local Research Ethics Committee and the General Practitioners' Local Medical Committee, access to six general medical practices, with diverse occupational class profiles, was negotiated. The general practitioners wrote to the women on their birth registers, inviting them to discuss participation in a study whose aim was to 'improve our understanding of the choices which mothers make about what, when and how to feed their babies'. A leaflet explaining the study was enclosed and the researchers contacted the mothers who responded positively to this invitation to answer any queries and to negotiate formal consent.
To achieve heterogeneity, a quota sample of 36 mothers, stratified by age and occupational class (see Table 1), was drawn sequentially from women who responded. Thirty-two women lived with a male partner, including two who lived with other members of his family. Three women lived with either one or both their parents and one lived alone. Twenty-three women were married. Two mothers, one each from the intermediate and lowest occupational class groups, were from minority ethnic groups (African Caribbean and South Asian). Fifteen had some post-compulsory education, and seven had attended college or university.
Six qualitative interviews were conducted with each woman in her own home, one before the birth and five at fixed intervals over the subsequent 2 years. The interviews were conducted by one of two researchers. The analysis presented here is drawn from the interviews at 2 months and 2 years. Interviews lasted between 1 and 2 hours. Interviews with 33 women were tape-recorded and fully transcribed. Detailed notes were taken of interviews with three women and fully written up immediately afterwards. The interview transcripts were made anonymous by substituting pseudonyms for the names of mother, baby, partner, health professionals and any others named in the course of the interviews. Transcripts were reviewed for any other identifying material and this was removed.
A subsample of 12 women, reflecting variations in age, occupational class and feeding practices, but not in ethnicity, was selected for preliminary analysis. Transcripts for this subsample were subjected to inductive analysis by the author and two research associates. Each investigator examined four transcripts in detail and produced written reports identifying emerging themes and categories. The emerging analysis was discussed in weekly meetings, related to relevant literatures, and developed into a coding framework, specifying operational definitions for each code. This framework was applied to the remaining interviews. Difficulties were discussed and the coding handbook amended before the revised coding was applied to the full data set. Although the analysis was primarily inductive, it involved the interplay between research questions identified in advance, 'foreshadowed themes' (Glaser and Strauss, 1967), previous theoretical and empirical work and the data, which is characteristic of much qualitative research (Emerson, 1983;Murphy and Dingwall, 2003). At the end of the study period, a printed summary of the project's results was sent to the participants who were invited to contact us with any comments or feedback. We considered but rejected the possibility of holding a face-to-face briefing for the study participants on the grounds that this might compromise their anonymity. While a number of the participants did subsequently contact members of the team, no substantive comments about the research findings were received.

Newborn babies
The child's state as he or she enters the world is central to both images of childhood discussed earlier. The Dionysian child is born with a bias towards evil. From birth, the Apollonian child is endowed with essential goodness. In this section, I explore the extent to which mothers' representations of childhood reflected one or both of these images and consider the implications for childrearing practices. Data are from the interviews conducted 2 months after the births.
At this early stage, nearly all mothers embraced the Apollonian motif of the innocent, self-regulating child. Babies could be trusted to identify their needs and not to make inappropriate demands. Many mothers described their babies as demanding and even coercive but treated their demands as entirely legitimate. Mothers sought to meet rather than resist their babies' demands. For example, one mother said: I just go along with him really. I've not tried that much to make him go this or that . . . if I think that milk is all he wants then I feed him. (Alison, Older, Intermediate, WB 1 ) The behaviour of these babies was not presented as wilful or deviant, as demonstrated in excerpts from interviews with Tracey and Rosemary: I mean really I was rooted to the sofa and she was wanting to feed all the time which I hadn't prepared for . . . I felt like I was stuck and in order that I ate I needed other people here to feed me. They were doing things like making me food and literally bringing it in and giving it to me so that I could carry on feeding her. . . . I would suddenly find that you know for 2 hours she was wanting to feed almost continuously in the afternoon. (Rosemary, Younger, Highest, WB) I was actually letting her feed every time she wanted but I spent almost the whole evening feeding her like I say from about 6 until 12 with only sort of the odd hour off and so I mean . . . that's quite difficult but I accept that's what she wants if that happens. (Tracey, Older, Highest, WB) The use of the word 'want' in these two excerpts is interesting. In relation to older children and adults, there is a sharp conceptual distinction between needs and wants. As discussed earlier, the identification of needs carries with it a strong moral obligation to meet those needs (Lawler, 1999;Woodhead, 1997). Needs are defined as unconditionally legitimate. Wants, on the other hand, are only conditionally legitimate, particularly when their satisfaction depends upon another's cooperation. They call for a judgement about reasonableness. There is no such distinction in this talk about very young babies. Here 'need' and 'want' are used interchangeably. To fail to meet either is described as 'cruel', as in the following excerpt: I didn't see any point that if she was, if after only 2 hours after a feed, if she lay crying, screaming and whatever else, I thought it would be cruel to leave her. . . . I don't agree with leaving her to cry if she is hungry. (Helen, Younger, Lowest, WB) This conflation of needs and wants in such talk reflects a widespread understanding of small babies as self-regulating and the strong message from many contemporary childrearing manuals that it is impossible to spoil a baby in the first year of life. For example, babies were trusted to know how frequently they needed to be fed: I fed her when she wanted it. . . . You can't sort of feed her when you're ready. You've got to feed when she's ready. (Daphne, Older, Intermediate, WB) Babies' demands may not be consistent from feed to feed or day to day but they can be trusted to balance out their requirements: She tends actually to have a 6 ounce the next feed and then the feed after she'll want a bit more and then she'll have 8 ounces again and then she'll go to 6 and she lets me know she's still hungry. She'll scream when you take the bottle away or if she's finished it. So I have to rush out and get her some more . . . if she started taking more all the time then I'll make a bigger bottle. (Annette, Older, Lowest, WB) The image of the child embedded in these descriptions is Apollonian. These babies are understood to be innocent, spontaneous and self-regulating. They are trusted to identify their needs and the mother's task is to interpret and respond. There is no suggestion here that the baby might be wilful, manipulative, self-indulgent or greedy. The implications for mothers are profound. As Lawler (1999) argues, mothers' own needs are effaced as they are positioned as primary meeters of their children's unconditionally legitimate demands.
These mothers rejected advice that failed to honour the babies' innocence or attempted to override the babies' self-regulation. Sometimes health professionals were criticized for failing to recognize the babies' self-regulating capacity. For example, Julia challenges the notion that her baby's unwillingness to feed was 'naughty': One or two of the midwives were saying, 'Oh you naughty baby, blah, blah, blah' and I thought, 'Well, how can she be naughty?' you know. That was sort of as if it was her fault. (Julia, Older, Highest, WB) Such 'naughtiness' would imply wilfulness, a suggestion that Julia derides. Mothers frequently described battles to protect their babies from officious interference, particularly from health professionals. For example, Diane described a midwife persuading her to make her baby feed: There was a very young midwife on that night and she said, 'You really must try harder than that Mrs X. . . . He'll be awake all night if you don't feed him more than that.' I said, 'Well he obviously doesn't want to feed, otherwise he would feed.' 'Try again', she said and she tried to latch him on. I mean obviously he didn't want to and she just bustled away. (Diane, Older, Highest, WB) This mother reasserts her confidence in her baby's capacity for self-regulation. Similarly, Helen rejected her health visitor's suggestion that she should limit her baby's intake of food: My old health visitor . . . 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.' (Helen, Younger, Lowest, WB) Here Helen confronted the suggestion that she was confusing wants with needs. The health visitor categorized the baby's consumption as illegitimate because it was more than babies 'need'. The mother rejects this advice and reaffirms her baby's self-regulating capacity.
These women insist that their babies should be protected from officious interventions from whatever source. Barbara and Rosemary described their commitment to being baby-led: Several mothers told salutary tales about being tempted to doubt their babies' capacity for self-regulation and/or innocence. Such stories always ended with the mother showing how such doubts had proved baseless and a reassertion of the babies' trustworthiness. Bryony described frustration with her baby's apparent stubborn lack of cooperation: I tried to latch her on and she took and I thought, 'Oh great you know, wonderful', but within like 30 seconds she'd dropped off, was crying she wanted a feed and I thought, 'Well, it's there, what's the matter with you?' You know and I was like not losing my temper with her but I was getting so frustrated because I thought, 'Well, what do you want, woman? It's there, come on, take it. Why are you dropping off, you know? You're hungry, there's milk there, why are you dropping off?' (Bryony, Younger, Lowest, WB) Up to this point, the mother's frustration appears justified. The baby's demand for food, followed by refusal to take it appears wilful. However, faith in the baby's innocence is restored as we discover that the mother had developed an undiagnosed breast infection making feeding difficult: I think what it was, me boob had swelled up that much . . . it was that hard she couldn't suck the nipple out enough so she just weren't interested.
Other mothers told similar stories of initial doubt followed by restoration of faith. They described how they had been tempted to impose structure but had since concluded that this was inappropriate. Christine commented upon her attempt to stretch the time between feeds: I tried to make him wait at least 3 hours . . . but he was just crying all the time and being miserable, so I thought, 'Why should I make him be miserable?' If he's hungry I feed him. He's only a baby; that's how I see it. (Christine, Younger, Highest, WB) Like Christine, many of the women linked their commitment to being babyled in their feeding practices to the special status of their children as 'only babies'.
Thus, most mothers drew on Apollonian motifs of innocence, trustworthiness and self-regulation. Any suggestion that their babies' behaviour was deviant and therefore sanctionable was dismissed. There were, however, three women whose talk constituted their babies in very different ways, reproducing elements of the Dionysian motif. They presented their babies not only as difficult and demanding but also as naughty, manipulative, out of control, impatient and uncooperative.
For example, one mother clearly interpreted her baby's behaviour as wilful: She'll spit it [dummy] out on purpose because she knows I'm gonna damn well stick it back in . . . sometimes she just, she has this thing. . . . I'll put the dummy in and she'll look at me and go 'plop' and spit it out and I'll ignore her and she'll try and get it back in, going like this [gestures with hands]. So I, because to me that's obvious she wants it, so I'll put it in and then, blow me, she spits it out again and I'll say, 'No, Joanne, I'm not doing it again' and I ignore her and she'll sob, sob, sob, trying to get it and I say, 'Why are you doing this?' and it drives me up the wall. (Dilys, Older, Intermediate, AC) This mother undermines any suggestion that this behaviour is accidental or inevitable. Spitting the dummy out is not interpreted as expressing need. The baby's intentionality is emphasized. She is described as spitting the dummy out 'on purpose'. The reference to looking at her mother before she spits the dummy out implies defiance and manipulativeness consistent with the Dionysian motif.
Sally also categorized her baby's demands as illegitimate: She's got a real temper on her when she's hungry. She just screams you know and won't calm down and I found when I was trying to breast feed her she wouldn't hold her mouth still for me to, to latch her on. . . . She just wouldn't calm down and wait for 2 minutes even though she had been asleep 1 minute earlier, you know, she was really, 'I want it now in my mouth. I don't want to wait'. (Sally, Younger, Highest, WB) This baby is presented as wilfully refusing to cooperate with attempts to feed her. Her impatience and bad temper are portrayed as unjustified given her mother's efforts. Sally, who was a part-time undergraduate student, went on to allege that her baby's demanding behaviour was a deliberate attempt to sabotage her other activities: I mean, just the other day she was really grumpy all day and I think it was because she knew I had to get my college essay done, you know.
Like Sally, a third mother, Beatrice, represents her baby as wilfully uncooperative: A terrible lot of pressure was put on me to breast feed and it just wouldn't work. In presenting their babies as wilful and manipulative, these women cited health professionals' opinions, as in the following excerpts: Well she's obviously a very naughty little girl and she doesn't want it you know, give her a bottle. (Beatrice, Younger, Intermediate, WB) This last week and a half, the health visitor said, 'Oh, she's wrapping us round her little finger and every time she's crying a lot she's associating that with more milk.' (Sally, Younger, Highest, WB) These three mothers' talk reproduces many elements of the Dionysian motif. The babies are seen as having a bias towards wilful, manipulative and selfdestructive behaviour.
The marked contrast between the two groups of mothers discussed above reinforces the importance of recognizing the heterogeneity of understandings of childhood prevalent at particular periods (Cunningham, 1995;Jenks, 1996). While most mothers in our sample reproduced versions of the Apollonian child, which is argued to be the predominant motif in contemporary western societies, just three characterized their children in ways closer to the Dionysian motif. Of these three, one (Dilys) was the only African Caribbean woman in the sample. However, as the sample was not designed to explore differences between mothers from different ethic groups, it is not possible to speculate about whether this Dionysian image of childhood is more or less prevalent among African Caribbean mothers.
How, then, do these two different constructions of childhood relate to childrearing practices? For mothers who rejected the notion that newborn babies are capable of making illegitimate demands, the practical implications appeared straightforward. If babies' demands are, by definition, legitimate this removes any imperative to 'break the will' of a potentially anarchic baby. These mothers emphasized the importance of meeting babies' demands as fully and immediately as possible. As Christine put it: I just go along with what he wants. I just go along with him really. (Christine, Younger, Highest, WB) This reflects the tolerant, sympathetic, child-centred approach conventionally associated with the Apollonian motif.
For the three mothers who constituted their babies as fully fledged agents, capable of wilful manipulation and acts that are damaging to themselves and to others, the implications for childrearing practices were less straightforward. These mothers presented themselves as impotent in the face of their babies' wilfulness. The babies were portrayed as having the upper hand. This is evident in some of the excerpts reproduced in this section, where, faced by their babies' intransigence, the mothers made comments such as: Thus, although these women's constructions of their babies are consistent with the Dionysian motif, this was not linked to the repressive childrearing practices conventionally associated with it. Although frustrated and even incensed by their babies' behaviours, the mothers reported appeasing rather than resisting their babies' demands. As Cunningham (1995) has argued, ideas about childhood do not necessarily stand in a one-to-one relationship with the actualities of childrearing.
In summary, most, but not all, mothers of newborn babies constituted their babies in ways that reflected the Apollonian motif of the trustworthy, self-regulating, innocent child and eschewed repressive, controlling or disciplining approaches to childrearing. This reflects Murcott's (1993) finding that adult rules are suspended for babies 'on the grounds that they can do no other'. Murcott contrasts the social accommodation of babies with the treatment of adults whose capacity for bodily self-management is also compromised: the very old, the incontinent, the sick and the insane. She points out that, while these groups may be absolved insofar as their deviant behaviour is deemed unmotivated and involuntary they nonetheless risk constantly the ascription of spoiled identity (Goffman, 1963) in a way that babies do not. Murcott suggests that the differences between social response to babies and to other groups who share some of the incapacities of babyhood relate to the reversibility of babies' inability to follow the rules. While this is no doubt true, the data presented here suggest that another factor is also at work. For these mothers, the imperative to accommodate their babies' demands was grounded in the conviction that these demands reflected genuine needs on their babies' part. Unlike, for example, the very old or the insane, whose judgement is frequently called into question, these babies were trusted to self-regulate.
I now consider whether and how this view of babies as innocent, trustworthy and self-regulating had attenuated by the time their babies had reached their second birthdays.

Two-year-olds
Two years after the birth, major shifts are observable in the mothers' descriptions of their children. Children were now presented routinely as knowing and manipulative. Terms like 'naughty', 'mischievous', 'on purpose' and 'playing up' are now common. One woman described her child's exploitation of her grandparents: They're making a rod for their own back really because I wouldn't appreciate it. I mean I wouldn't like it if Jane was coming home and the first thing she said is 'Choc, choc' because I would feel like she's using me for what she can get and I feel that is what she does with my Mum and Dad. She's actually using them for what she can get, because children are like that. (Bryony, Younger, Lowest, WB) Bryony suggests Jane has demonstrated a calculating, instrumental attitude to grandparents' generosity. Their relationship is grounded in self-interest rather than the reciprocity of a loving child-adult relationship. The comment that, 'children are like that' indicates that Bryony sees such instrumentality as characteristic of all children. Here one can see hints of the Dionysian motif emerging.
By the age of 2, all the children were presented as capable of wilful misbehaviour. Their intentionality was stressed. For example, Helen described her child's behaviour at mealtimes: When she does throw her food about I think she's doing it on purpose rather than having accidents. (Helen, Younger, Lowest, WB) The mothers emphasized their children's capacity for knowing right from wrong. Elaine remarked: He's a very active child but very switched on as well. Sort of quite naughty sometimes, even though he knows not to do something he'll do it. (Elaine, Older, Highest, SA) This idea that children are choosing to misbehave was central to many mothers' talk about their children. Carol talked about her child refusing food, behaviour that she designates 'naughty': She knows it's naughty . . . she'll push things away and she'll say, 'Grace naughty girl, Grace naughty' and she'll say, when we've eaten ours, 'Daddy's gone, Daddy's good, Mummy's gone, Mummy's good, Grace's not gone, Grace naughty.' So she knows. She really knows everything about it. (Carol, Younger, Intermediate, WB) Children were presented as wilfully manipulating their parents, as illustrated by the following two excerpts: She'll think, 'Oh I can get away with this' and she will, she won't have her Weetabix . . . he'll [child's father] say, 'Come on then Ruth, you're gonna have your breakfast', she'll wind him up . . . and she's laughing, thinking, 'Oh I've won this round. . . . I haven't got to have Weetabix after all.' (Dilys, Older, Intermediate, AC) He uses drink as a delaying tactic at bed now. I think he's cottoned on that we try to get him to drink all the time and . . . he has his milk with him into bed. He'll say, 'Anthony want milk', so I'll have the milk there and I'll give it to him and, 'No, Anthony want tea', so I bring the tea up and he has a sip, 'No Anthony want fizzy'. (Rita, Younger, Highest, WB) Such intentionality is presented as an emergent quality, linked to the child's growing maturity. For example, Alison comments that: He does know now and he's old enough to know that you know he should sit and eat his meal . . . so he knows exactly what is going on. (Alison, Older, Intermediate, WB) Here there is an implicit contrast between what he knows 'now' and his earlier more limited capacity to understand the 'rules' of appropriate behaviour. Elaine (Older, Highest, SA) describes her child as 'slowly learning to play the system'. Bryony suggests that her child's attempts to manipulate potty training indicate a stage she had reached: But then you see she'd got to the stage where she'd go and wee and then want a sweet and then she'd go back 2 minutes later and do a little dribble again you know but it got to where she'd come and she'd say 'I weed, sweeties' you know. I was like, 'oh no, you don't get sweeties, you don't wee just to get a sweetie.' (Bryony, Younger, Lowest, WB) Thus, contrary to some formulations of the Dionysian motif, wilfulness and naughtiness are not seen as inborn characteristics. Rather they are a developmental stage attained as the children become intentional actors.
As noted earlier, most mothers of newborn babies collapsed any distinction between needs and desires. This, too, had changed radically by the time the children were 2. A clear distinction emerged between demands and needs. This was evident in the mothers' concern that their children might demand unhealthy food or refuse food needed to maintain health. For example, Christine expressed concern that, if not restrained, her child would become obese.
I've asked her [childminder] . . . just to give him juice and to cut down on his milk now. . . . He's portly . . . I mean he would drink all day. He would if you let him I'm sure. . . . I am in control of what he eats so I've got to look out for him haven't I? (Christine, Younger, Highest, WB) The sense that children could be seen unequivocally as self-regulating had disappeared from the mothers' talk. This is not to say that individual children were never trusted to make good choices. For example, Dilys described how her child moderated her intake of sweet food: She tends to, she tends to just moderate her intake anyway. She seems to regulate it. She won't go barmy for jelly and ice cream. She'll have a bit of it, she'll have a taste. (Dilys, Older, Intermediate, AC) However, such trust in the child's ability to self-regulate was no longer unconditional. It had to be earned. It was based on observation of particular

C H I L D H O O D 1 4 ( 1 )
aspects of the child's behaviour. Simply because children self-regulated in one area, this did not mean that they could be trusted in other areas. This is illustrated by a further excerpt from Dilys's interview: My husband drinks bitter and lager and I like wine and I dread to think what we're gonna do with her. I just, because she insists on having a sip and you can't say, 'Oh here, have a sip' and then that pacifies her. It doesn't, she just wants more and more. . . .. I have some real tantrums when I won't let her have any more. (Dilys, Older, Intermediate, AC) Thus, although Dilys's child is trusted to moderate her demands for sweet foods, her demands for alcohol are seen as uncontrolled.
By 2 years, the mothers routinely talked about their children in ways that drew less upon the Apollonian and more upon the Dionysian motif of childhood. While they avoided the somewhat archaic language of 'evil', they nevertheless portrayed children as drawn to pleasure and self-gratification and lacking self-control or sensitivity to the needs of others (James et al., 1998). The children were all presented as intentional actors whose demands were potentially illegitimate.
As we have already observed, there is no necessary correspondence between the ideas that people hold about childhood and the actualities of childrearing practice. I now turn to consider to what extent the major shifts in constructions of childhood between 2 months and 2 years were reflected in descriptions of childrearing practice.
Conventional accounts of the Dionysian motif associate this image of childhood with repressive and controlling childrearing practices intended to socialize, shape and even 'break the will' of the child. There were, indeed, indications of such approaches in the talk of some women. For example, Belinda described how she enforced the requirement that her child should eat up at mealtimes: It's a case of 'if you don't eat your lunch you're not having a load of crisps and things' because you know I tend to think well, she'll have to eat her tea. (Belinda, Older, Lowest, WB) Some mothers also described how they seek to discipline disobedient children: If the mood takes him, sometimes he's terrible and he won't eat his dinner and he just runs off and I don't give him anything to eat then. If he's naughty I take his meal off him and he doesn't get anything. . . . He's had nothing some days from early in the afternoon. He's refused or played with his dinner and I've taken it off him and then he's had nothing until the next day, drinks obviously, . . . I'm not bothered if he misses a few meals but I'm not, I don't want him to be naughty at mealtimes. If he messes with it, that's it. (Christine, Younger, Highest, WB) There is also some talk that focuses on avoiding or redressing bad habits. Sally described how she used physical chastisement: She doesn't do it any more but she started to have a little habit of when she'd had enough breakfast cereal, to bring the rest of the milk on to her tray and whiz her hand around in it but we've just about got rid of that now because if she does it she gets a smack, so. (Sally, Younger, Highest, WB) However, the childrearing practices described by the women were far from homogeneous. Although drawing on Dionysian motifs, many of the women distanced themselves from more controlling or directive approaches to childrearing. Their practices were much closer to the laissez-faire approach associated with the Apollonian motif. For example, Hilda described her approach to her child's diet: He gets what he wants . . . he gets his own way all the time. I mean it doesn't really bother me. . . . I don't see it as a problem. I just let him get on with it and I mean, as he gets older, he'll probably eat more, at regular times like when he gets to nursery and then school so I just leave it up to him. . . . Kids are kids. All kids are different. They'll eat what they wanna eat and drink what they wanna drink. (Hilda, Younger, Intermediate, WB) These mothers emphasize the self-determination they accord to their children: If she doesn't want it, I don't try to force it down. (Angela, Younger, Intermediate, WB) They give a variety of reasons for accommodating their children's wishes. In some cases, they suggest that while, in principle, they would like to exert more control, they are powerless to do so: In an ideal world I'd like her to sit down, real quiet, sit still, eat her dinner with a knife and fork or a knife and a spoon whatever and have no problems but it doesn't work out that way does it? (Helen, Younger, Lowest, WB) Dilys described her anxiety about her child's food intake, her attempts to control it, and her ultimate acceptance of the conditions under which her child was willing to eat: It worries me if she puts her foot down and won't have Weetabix because . . . she went through a phase, the health visitor called it hamster syndrome . . . rather than just not having any more food in they keep it in their mouth and puff their cheeks up so you can't get any more in. . . . It got me really annoyed and I was push, push, push like this to her cheeks. . . . I was going swallow, swallow but she wouldn't . . . my dad would have called it bloody naughty, he wouldn't have said it was hamster syndrome. . . . I know she's dictating to me but if it's, if she's gonna be happy and it means eating all her dinner then she can sit within a foot of the television. (Dilys, Older, Intermediate, AC) In some cases, the mothers judged that the effort required to enforce control over the children was disproportionate: He's mad for coffee. I try not to give him too much but it's a bit hard because he demands it and if he don't get it he has a tantrum. (Hilda, Younger, Lowest, WB) Some mothers described how earlier, more directive approaches had been revised in the light of experience. For example, Harriet described her changed practice: It can be very stressful you know . . . actually face to face you know, your sort of stress levels rise very, very rapidly I've felt when I was battling with her. . . . I try not to let myself now because . . . if I start to feel that way and you know she's obviously getting wound up I just drop it and just get out of the situation you know, just take her out of her high chair and just forget whatever I was trying to do with her you know and just sort of resign myself to the fact that you know. (Harriet, Older, Highest, WB) There is no suggestion here that the judgement of the child's behaviour as inappropriate has changed. A more permissive approach to childrearing has been adopted because the stress to both mother and child were judged excessive.
Similarly, Harmony described how she had stopped 'forcing' her child to eat: What she doesn't eat then I chuck away. I used to force feed her but I don't do that now. I don't, what she don't eat I chuck away. (Harmony, Younger, Intermediate, WB) A number of mothers attributed their reluctance to adopt a more directive style of parenting to concerns about their children's physical or psychological welfare. Carol described how she had oscillated between more controlling and more permissive practices in relation to her daughter's food intake: I'm finding it very difficult with her . . . I've tried to say . . . 'Right, she eats what we have, if she doesn't eat it that's fine, but she's not having something different' but then I just panic and worry because her food intake for that day has been next to nothing. (Carol, Younger, Intermediate, WB) Sarah described how her concern to ensure that her child drinks enough fluids forced her to accept his preference for coke: He's a big coke fan as well. . . . It'll be like, 'No, have some juice', 'No, coke' and you're like, 'Oh, so' but then you're back to that Catch 22. I'd rather him drink something than not at all. (Sarah, Younger, Intermediate, WB) Helen expressed anxiety that an overly directive approach to parenting might be counter-productive: We try all sorts, coaxing her, bribing her, forcing her but she's so strong-willed that you've got to walk gently else you'll push her further away. (Helen, Younger, Lowest, WB) Dilys described her concern that her child might develop a 'complex' if forced to eat something she did not want: If she puts her foot down and won't have her cereal and you know, I'll go, 'Right', but I won't force it into her because I don't want to give her a complex about it. (Dilys, Older, Intermediate, AC) This concern about potential psychological damage and future consequences of overriding her child's food preferences is also evident in Harriet's comments: I was actually having to sort of physically force it into her mouth and you know it was really upsetting me and upsetting her and you know we were going nowhere so I just, in the end, just decided right, well, ok, we're gonna feed you what I know you will eat and you know since then it's been a lot better . . . mealtimes are not a battleground now and I do worry about things like that you know because I, oh you know, what are we doing to them, what are they gonna be like in years to come, are they having problems with you know food, er, eating disorders and things, so I didn't really want to start getting into you know a crisis situation with her. (Harriet, Older, Highest, WB) Thus, many of these mothers combined a construction of their child that approximates to the Dionysian motif, with a reluctance to adopt the controlling, directive, even punitive styles of childrearing that have been associated with that image of childhood. This reluctance was grounded in concerns about physical or psychological damage and impotence in the face of children's recalcitrance.

Discussion
I have examined contemporary mothers' constructions of babies and young children and the ways in which such constructions relate to one area of childrearing practice. Focusing on talk about infant feeding has enabled us to look at the concrete reality of the mother's practices in one specific area of childrearing and the ways in which these do or do not articulate with their understandings of what it is to be a child. Feeding is one of the 'nitty-gritty issues of daily life' (Ribbens, 1994: 145) where underlying cultural assumptions about the nature of childhood are likely to be played out. It is not the only such issue and, indeed, other issues including sleep and cleanliness might raise different issues for the mothers. Nevertheless, in studying infant feeding, it has been possible to examine apparent contradictions between the ways in which some mothers construct their young children and the childrearing practices they feel comfortable in adopting.
The contrasting representations of childhood identified by other authors are evident in these data. Apollonian themes of natural goodness and innocence are observable, particularly in the early interviews where most mothers present their babies' demands as characterized by necessity and, therefore, as unconditionally legitimate. Similarly, stripped of the archaic language of original sin and demonic forces, the mothers' talk in later interviews reproduces elements of the Dionysian motif identified by Jenks (1996). While the mothers might hesitate to use the term 'evil' about their children, they do, at 2 years old, present them in ways that fit Jenks's definition of the Dionysian motif: '[the child] loves pleasure, it celebrates self-gratification and it is wholly demanding in relation to any object, or indeed subject, that prevents its satiation' (Jenks, 1996: 71).
As we have seen, while most mothers construct their newborn babies in Apollonian terms, the talk of just three mothers is closer to the Dionysian motif. The difference between these two groups of mothers does not lie in the objective behaviours displayed by their babies. Both groups of mothers described their babies as demanding and even coercive. The difference is in the ways in which the mothers interpret such behaviours. As sociologists of deviance have long argued, the designation of an action as deviant or otherwise does not depend upon the outcome of that action. Rather, like childhood, deviance is socially constructed and depends upon the interpretation of the action in question (McHugh, 1970). It is only where an action is deemed both avoidable and intentional that it is likely to be treated as illegitimate. By definition, needs are unavoidable. Where mothers interpret their babies' demands as the reflection of genuine need and reject the possibility that their babies might be acting manipulatively, they are unlikely to treat such demands as illegitimate.
These data reaffirm the importance of recognizing the complexity and ambivalence of the images of childhood that are current within particular cultures at particular periods. Different constructions of childhood coexist (Jenks, 1996). This is demonstrated, in particular, by the data presented in relation to newborn babies. Although most mothers presented their babies as innocent, self-regulating and trustworthy, in keeping with the Apollonian motif, three women constructed their children very differently. These babies were seen in ways which more closely drew on Dionysian ideas of children as 'wilful material energy' (James et al., 1998: 10). While no clear differences were identified between women of different occupational or ethnic groups, this may attributable to the small sample and should not, therefore, be taken as evidence that no differences exist. In particular, the very small numbers of women from ethic minority groups could be masking cultural diversity in both constructions of childhood and approaches to childrearing. This is an issue that could usefully be examined in future research.
These women do not, of course, construct childhood ex nihil. Rather, as Schutz (1954) suggests, common-sense talk, of the kind analysed here, draws upon the typifications of childhood that are available in their sociocultural worlds. The 'truths' about childhood that are reflected in such typifications are strongly influenced by specialist discourses, particularly within medicine and the 'psy' professions (Lawler, 1999;Rose, 1996). The health care professions, childcare publications and familial and neighbourhood networks all act as conduits through which these specialist discourses are transmitted and come to form part of common-sense ideas about childhood, childrearing and the responsibilities of motherhood (Miller, 2005). However, as we have seen, women are capable of resistance also. Some of the women were, for example, very dismissive of midwives and health visitors' attempts to challenge mothers' representations of their newborn babies as innocent and self-regulating. Mothers cannot, therefore, be understood simply as 'cultural dopes who unselfconsciously mimic' the ideas and childrearing methods of others (Hays, 1996: 75).
Historical analyses of images of childhood have emphasized the centrality of the child's state at birth. Thus the Apollonian child is born with a bias towards good and the Dionysian with a bias towards evil. The analysis presented here questions such continuity between the constructions of babyhood and of later childhood displayed by contemporary UK mothers. These mothers distinguish between their newborn babies' trustworthiness, innocence and self-regulation and their later bias towards self-gratification, insensitivity and lack of self-control. The only exceptions to this are those three mothers who drew on Dionysian motifs when their babies were newborn and carried these forward into the later interviews. For most mothers, the move from Apollonian innocence to Dionysian self-indulgence is seen as a developmental stage.
These data also underline the disjunction between images of childhood at particular historical periods and the childrearing practices that may or may not be linked to them. It is clear that the relationship between images of childhood and childrearing practices is asymmetrical. Where mothers drew on Apollonian images, the implications for childrearing were relatively straightforward. Mothers are to interpret and meet their babies' expressions of need. Where Dionysian images predominate and babies and children are constituted as intentional agents whose demands may go beyond their legitimate needs, the links between image and practice are attenuated. Some mothers do report childrearing practices that approximate to the model of 'strict discipline and parental constraint' (Stone, 1977) that has been linked to this image of childhood. However, many reject this view of socialization as a 'battle' (Jenks, 1996: 71) either because they view it as unenforceable or because they fear the short and long-term consequences of attempting to apply it. This article has examined the images of childhood drawn upon by a group of UK mothers of babies and very young children and the ways in which these relate to mundane practices of childrearing. In doing so, it has substantiated many concerns about interpretations of historical trends in images of childhood that treat these as more homogeneous, more consistent and more unequivocally linked to childrearing practices than they necessarily are. It also suggests that constructions of childhood are emergent and linked to stages of development that children are deemed to have reached at particular ages. This suggests that it would be worthwhile to extend the analyses presented here by examining the images and practices associated with older children in the UK and with both younger and older children in other countries.