Engendered Emotion: Gender, Power,
and the Rhetoric of Emotional Control
in American Discourse
Catherine A. Lutz
From Harre, R, & Parrot, W.G., (1996) (Eds.) The emotions, Thousand Oaks, CA: Sage. pp. 152-170.
In Western academic discourse, emotions have begun to move from their culturally assigned place at the center of the dark recesses of inner life and are being depicted as cultural, social and linguistic operators. In the process, we can ask not only about the cultural foundations of things construed as emotional, but about the organising category of 'emotion' itself. One important aspect of that category is its association with the female, so that qualities that define the emotional also define women. For this reason, any discourse on emotion is also, at least implicitly, a discourse on gender.
As both an analytic and an everyday concept in the West, emotion, like the female, has typically been viewed as something natural rather than cultural, irrational rather than rational, chaotic rather than ordered, subjective rather than universal, physical rather than mental or intellectual, unintended and uncontrollable, and hence often dangerous. This network of associations sets emotion in disadvantaged contrast to more valued personal processes, particularly to cognition or rational thought, and the female in deficient relation to her male other. Another and competing theme in Western cultural renditions of emotion, however, contrasts emotion with cold alienation. Emotion, in this view, is life to its absence's death, is interpersonal connection or relationship to an unemotional estrangement, is a glorified and free nature against a shackling civilization. This latter rendition of emotion echoes some of the fundamental ways the female has also been 'redeemed', or alternatively and more positively, construed (Lutz, 1988).
In this chapter, I will explore how emotion has been given a gender in some sectors of American culture and, in the process, make two related arguments. First, I will demonstrate that local or everyday lay discourse on emotion explicitly and implicitly draws links among women, subordination, rebellion and emotion by examining interview conversations conducted with a small group of American women and men. In particular, I will explore a 'rhetoric of control' that frequently accompanies women's (and, to a lesser extent, men's) talk about emotion, and argue that talk about the control or management of emotion is also a narrative about the double sided nature - both weak and dangerous - of dominated groups. Talk about emotional control in and by women, in other words, is talk about power and its exercise. Second, I will argue that this and further aspects of local discourse are echoed and reproduced in many areas of social and natural scientific discourse that deal with the 'emotional female'. Finally, I will present a further, more syntactic analysis of the interview conversations that contradicts at least some of the stereotypical beliefs about the relationship between gender and emotion that these informants, as well as social science, have voiced. This analysis looks at the degree to which women and men might differentially use syntactic patterns that distance, disavow, or depersonalise the experience of emotion. The failure to find systematic differences can be taken as tentative evidence that cultural models that paint women as more emotionally expressive or more comfortable with a discussion of their own emotions remain surface models and do not organise discourse at more microscopic or out-of-awareness levels.
Gender, power, and the rhetoric of emotional control
Western discourse on emotions constitutes them as paradoxical entities that are both a sign of weakness and a powerful force. On the one hand, emotion weakens the person who experiences it. It does this both by serving as a sign of a sort of character defect (e.g., 'She couldn't rise above her emotions') and by being a sign of at least temporary intrapsychic disorganisation (e.g., 'She was in a fragile state' or 'She fell apart'). The person who has 'fallen apart', needless to say, is unable to function effectively or forcefully. On the other hand, emotions are literally physical forces that push us into vigorous action. 'She was charged up', we say; 'Waves of emotion shook his body'. Women are constructed in a similar contradictory fashion as both strong and weak (e.g., Jordanova, 1980), and I will present evidence from the interviews mentioned earlier that when American women and men talk about emotion, they draw on that similarity to comment on the nature of gender and power. This feature of the emotional and of the female produces frequent discussion in the interviews of the problem of controlling one's feelings. Such discussion is found in both men's and women's discourse, but much more frequently in the latter. I will show that this talk about control of emotions is evidence of a widely shared cultural view of the danger of both women and their emotionality. It is also talk that may mean different things to both the speaker and the audience when it is uttered by women and by men, and this factor will be used to help account for differences in the rate of use of this rhetoric of control. Although both women and men draw on a culturally available model of emotion as something in need of control, they can be seen as often making some different kinds of sense and claims from it.
The material I turn to first was collected in four extended interviews on emotion with fifteen American working- and middle-class women and men. All white, they ranged in age from the early twenties to the mid-seventies and included a bank teller, factory worker, college teacher, retiree, housing code inspector and stockbroker. Most were parents. The interviews were usually conducted in people's homes, and the interviewers included myself and a number of graduate students, most of them women. Each person was interviewed by the same individual for all four sessions, and although a small number of questions organised each session, every attempt was made to have the interviews approximate 'natural conversation'. Nonetheless, it is clearly important to keep in mind the context of the discourse to be analysed, as it was produced by a group of people who agreed on letter and phone solicitation 'to talk about emotion' for an audience of relative strangers who were also academics and mostly females.
Many people mentioned at one or several points in the interviews that they believe women to be more emotional than men. One example of the variety of ways this was phrased is the account one woman gave to explain her observation that some people seem inherently to be 'nervous types'. She remembered about her childhood that
the female teachers had a tendency to really holler at the kids a lot, and when I was in class with the male teacher, it seemed like he just let things pass by and it didn't seem to get his goat as fast, and he didn't shout at the same time the female may have in the same instance. . . . I think emotional people get upset faster. I do. And like with men and women, things that are sort of important or bothering me don't bother my husband. I think that's a difference of male and female.
One theme that frequently arises in the interviews is what can be called the 'rhetoric of control' (Rosaldo, 1977). When people are asked to talk about emotions, one of the most common sets of metaphors used is that in which someone or something controls, handles, copes, deals, disciplines, or manages either or both their emotions or the situation seen as creating the emotion. For example:
I believe an individual can exercise a great deal of control over their emotions by maintaining a more positive outlook, by not dwelling on the negative, by trying to push aside an unpleasant feeling. I'm getting angry and like I said, he's over being angry, more or less dropped it and he expects me to also. Well we don't have the same temper, I just can't handle it that way.
And in a more poetic turn, one person mused:
sadness ... dipping, dipping into that ... just the out-of-controlness of things.
People typically talk about controlling emotions, handling emotional situations as well as emotional feelings, and dealing with people, situations and emotions.
The notion of control operates very similarly here to the way it does in Western discourses on sexuality (Foucault, 1980). Both emotionality and sexuality are domains whose understanding is dominated by a biomedical model; both are seen as universal, natural impulses; both are talked about as existing in 'healthy' and 'unhealthy' forms; and both have come under the control of a medical or quasi-medical profession (principally psychiatry and psychology). Foucault has argued that popular views of sexuality - as a drive that was repressed during the Victorian era and gradually liberated during the twentieth century - are misleading because they posit a single essence that is manipulated by social convention. Rather, Foucault postulated, multiple sexualities are constantly produced and changed. A popular discourse on the control of emotion runs functionally parallel to a discourse on the control of sexuality; a rhetoric of control requires a psychophysical essence that is manipulated or wrestled with and directs attention away from the socially constructed nature of the idea of emotion (see Abu-Lughod and Lutz, 1990). In addition, the metaphor of control implies something that would otherwise be out of control, something wild and unruly, a threat to order. To speak about controlling emotions is to replicate the view of emotions as natural, dangerous, irrational, and physical.
What is striking is that women talked about the control of emotion more than twice as often as did men as a proportion of the total speech each produced in the interviews. To help account for this difference, we can ask what the rhetoric of control might accomplish for the speaker and what it might say to several audiences (see Brenneis, 1990). At least three things can be seen to be done via the rhetoric of emotional control: It (1) reproduces an important part of the cultural view of emotion (and then implicitly of women as the more emotional gender) as irrational, weak and dangerous; (2) minimally elevates the social status of the person who claims the need or ability to self-control emotions; and (3) opposes the view of the feminine self as dangerous when it is reversed, that is, when the speaker denies the need for or possibility of control of emotion. Each of these suggestions can only briefly be examined.
First, this rhetoric can be seen as a reproduction, primarily on the part of women, of the view of themselves as more emotional, of emotion as dangerous, and hence of themselves as in need of control. It does this first by setting up a boundary - that edge over which emotion that is uncontrolled can spill. A number of people have noted that threats to a dominant social order are sometimes articulated in a concern with diverse kinds of boundaries (whether physical or social) and their integrity (e.g., Martin, 1987; Scheper-Hughes and Lock, 1987). One of the most critical boundaries that is constituted in Western psychological discourse is that between the inside and the outside of persons; individualism as ideology is fundamentally based on the magnification of that particular boundary. When emotion is defined, as it also is in the West, as something inside the individual, it provides an important symbolic vehicle by which the problem of the maintenance of social order can be voiced. A discourse that is concerned with the expression, control, or repression of emotions can be seen as a discourse on the crossing back and forth of that boundary between inside and outside, a discourse we can expect to see in more elaborate forms in periods and places where social relations appear to be imminently overturned.
This rhetoric of emotional control goes further than defining and then defending boundaries, however; it also suggests a set of roles - one strong and defensive and the other weak but invasive - that are hierarchized and linked with gender roles. Rosaldo (1984) notes of hierarchical societies that they seem to evince greater concern than do more egalitarian ones with how society controls the inner emotional self and, we can add, with how one part of a bifurcated and hierarchically layered self controls another, The body politic, in other words, is sometimes replicated in the social relations of the various homunculi that populate the human mind, a kind of 'mental politic'. When cognition out reasons and successfully manages emotion, male-female roles are replicated. When women speak of control, they play the roles of both super- and subordinate, of controller and controlee. They identify their emotions and themselves as undisciplined and discipline both through a discourse on control of feeling. The construction of a feminine self, this material might suggest, includes a process by which women come to control themselves and so obviate the necessity for more coercive outside control.
There is the example of one woman in her late thirties; she talked about the hate she felt for her ex-husband, who began an affair while she was pregnant and left her with the infant, an older child and no paid employment.
So I think you try hard not to bring it [the feeling] out 'cause you don't want that type of thing at home with the kids, you know. That's very bad, very unhealthy, that's no way to grow up. So I think now, maybe I've just learned to control it and time has changed the feeling of the hate.
The woman here defines herself as someone with a feeling of hate and portrays it as dangerous, primarily in terms of the threat it poses to her own children, a threat she phrases in biomedical terms (i.e., 'unhealthy'). She replicates a view that Shields (1987) found prevalent in a survey of twentieth-century English-language child-rearing manuals; this is the danger that mothers' (and not fathers') emotions are thought to present to children. In addition, this woman's description of her feelings essentializes them as states; as such, they remain passive (see Cancian, 1987 on the feminization of love) rather than active motivators, a point to which we will return.
In other cases, people do not talk about themselves, but rather remind others (usually women) of the need to control themselves. These instances also serve to replicate the view of women as dangerously emotional. Another woman spoke about a female friend who still grieved for a son who had died two years previously: 'You've got to pick up and go on. You've got to try and get those feelings under control.' (The 'you' in this statement is a complex and multivocal sign (Kirkpatrick, 1987), and directs the admonition to control simultaneously to the grieving woman, the female interviewer, the speaker herself, no one in particular, and everyone in a potential audience.
A second pragmatic effect of the rhetoric of emotional control is a claim to have the ability to 'rise above' one's emotions or to approve of those who do. Women, more than men, may speak of control because they are concerned about counteracting the cultural denigration of themselves through an association with emotion. 'I think it's important to control emotions', they say, and implicitly remind a critical audience that they have the cooler stuff it takes to be considered mature and rational. It is important to note that, as academics, I and the graduate students who conducted the interviews may have been perceived as an audience in special need of such reminders. The speakers would have been doing this, however, by dissociating themselves from emotion rather than by questioning the dominant view both of themselves and of emotion.
Although women may have less access to a view of themselves as masterful individuals, a common aspect of the cultural scheme that is available paints them as masterfully effective with others on joint tasks, particularly interpersonal or emotional tasks (social science versions of this include Chodorow, 1978; Parsons and Bales, 1955). This subtly alters the meaning of the rhetoric of control; knowledge of what the feelings are that ,need' control and of what control should be like is perceived and described as a social rather than an individual process. For example, one woman says: 'If you're tied in with a family, . . . you have to use it for guidance how you control your emotions.' This is the same woman whose central life problem during the interview period was coping with her husband's ex-wife and family, who lived across the street from her. The regular, friendly contact between husband and ex-wife has left her very unhappy but also unsure about what to do. The ambiguity over who ought to control or regulate what is evident in her description of an argument she had with her husband over the issue.
I was mad. I was mad. And I said, 'I don't care whether you think I should [inaudible word] or stay in this at all, it's too, and cause I'm going to say it.' And I said, 'How dare you tell me how I'm supposed to feel', you know. Bob [her husband] would say, you know, 'You got to live with it' or 'You got to do this' or 'How dare you tell me this, I don't have to put up with anything' or 'I don't have to feel this way because you tell me I have to feel this way'. You know, it was, in that case Robin is his ex-wife, 'and you have to just kind of deal with it', you know, 'all the problems that she presents in your own way'. And it was almost sort of like saying 'You're going to have to like it'. Well I don't. I don't, you know. And for a year and a half he kept saying, you know, 'You're going to have to like it, this is the way it's going to be, you're going to have to do this, you're going to have to have, be, act, this certain way', you know, act everything hunky-dory, and it wasn't, you know, and I was beginning to resent a whole lot of things. 1, 1, I resented him for telling me I had to feel that way when 1, I wasn't real fond of the situation. I didn't like it. When I would tell him that I didn't like it, it was 'It's your problem, you deal with it'. I didn't like that, that made me really angry because I was saying, 'Help me out here, I don't know how to deal with this'.
This woman is frustrated with her husband for failing to join her in a collaborative project of 'dealing with' her feelings of resentment. Here control is given away to or shared with others. This strategy of control is more complex and subtle than the simple self-imposition described in other parts of the transcripts so far; it aims to control both the emotions of the self and the attention and assistance of the other. Note also that she speaks of 'resenting' or 'not liking' (relatively mild terms of displeasure) the overall situation but is most incensed ('mad, mad, mad') about her husband's assumption that she ought not to feel a certain way. She asserts the right to 'feel' unhappy about her predicament but is clearly defining that feeling in the standard contemporary sense of a strictly internal and passive event. Nowhere in the interview does she explicitly state or appear to imply that she wants, intends, or ought to act in concert with those feelings. What is being controlled or dealt with, therefore, has already been defined as a relatively innocuous feeling rather than an action tendency.
Finally, the rhetoric of emotional control can also be employed in both idiosyncratic and 'reversed' ways that may intend or have the effect of at least minimally resisting the dominant view of emotionality, and thus of women. A few people, for example, spontaneously spoke about the problem of emotional control, thereby evoking the whole schema we have just been looking at. They went on, however, to define 'control' in a way that entailed relatively minimal constraints on emotional communication. One woman, a twenty-eight-year-old bank teller, said: 'Let me explain control. It's not that you sit there and you take it [some kind of abuse] and, you know, I think controlling them [emotions] is letting them out in the proper time, in the proper place.' Perhaps more radically, some women (as well as one of the gay men with whom I spoke) denied that they had the ability to control some or many of their emotions. One man in his twenties critically described a previous tendency he had to over-intellectualise problems and explained that he worked against that tendency because
It wasn't that I wanted to cut off my emotions, I just didn't, they would get out of control, and I found that the more I tried to suppress them, the more powerful they would become. It was like this big dam that didn't let a little out at a time, it would just explode all of a sudden, and I'd be totally out of control.
The question remains, however, of the validity of seeing these latter seemingly resistant uses of the rhetoric of emotional control as 'oppositional' forms (Williams, 1977) within that system. This is certainly a dangerous rhetorical strategy, caught as they (we) are within a hegemonic discourse not of our own making. The opposition to self-control will most likely be absorbed into the logic of the existing system and so come to equal not resistance but simple deficiency or lack (of control). A possibly oppositional intent may have collaborative outcomes to the extent that the denial of self-control is taken by most audiences as a deficit and a confirmation of ideas about women's irrationality.
The culturally constructed emotionality of women is rife with contradiction. The emotional female, like the natural world that is the cultural source of both affect and women, is constructed as both pliant (because weak and a resource for use by civilized man) and ultimately tremendously powerful and uncontrollable (Strathern, 1980). Emotionality is the source of women's value, their expertise in lieu of rationality, and yet it is the origin of their unsuitability for broader social tasks and even a potential threat to their children.
There are vivid parallels between this and the cultural meanings surrounding colonialism that Taussig (1984) and Stoler (1985) have described. Looking at early-twentieth-century colonists' views of the local Colombian labour force, Taussig describes their alternation between fear and awe of Indians who were perceived as dangerous and powerful figures, on the one hand, and disgust and denigration of their perceived weakness and lack of civilization, on the other. Taussig describes the process as one in which a ,colonial mirror' 'reflects back onto the colonists the barbarity of their own social relations' (1984: 495). In a (certainly less systematic or universally brutal) way, a 'patriarchal mirror' can be conceptualised as helping to produce the view of women as emotional - as dangerously 'eruptive' and as in the process of weakly 'breaking down'. A 'paradox of will' seems consistently to attend dominating relationships - whether those of gender, race or class - as the subordinate other is ideologically painted as weak (so as to need protection or discipline) and yet periodically as threatening to break the ideological boundary in riot or hysteria. Emotion talk, as evident in these transcripts, shows the same contradictions of control, weakness and strength. Given its definition as nature, at least in the West, emotion discourses may be one of the most likely and powerful devices by which domination proceeds.
The engendering of emotion in science
Demonstrations of the political, moral, and cultural bases of Western science have been made convincingly in a number of natural and social fields (e.g., Asad, 1973; Fausto-Sterling, 1985; Haan et al., 1983; Sampson, 1981). In like fashion, it can be argued that the sciences of emotion have been, in a significant sense, a product of their social context. In particular, the academic literature on emotion can be considered a form of political discourse on gender relations because of the marked associations between the two domains. That literature thus arises out of and re-enters a field of power struggles for the definition of true womanhood. As Haraway (1986) has said of American primatology, it can be seen as 'politics by other means', and in the case of emotions, it is most centrally a politics of gender by other means. By examining several examples of studies of emotion, we will see that much research over the years in biology, psychology, sociology, sociolinguistics and other fields has been implicitly based on everyday cultural models linking women and emotionality, and that this research moves from the assumption of these cultural premises to their 'proof'. Most striking about these studies is the number that naturalise the purported gender differences by attributing them to biological or necessary and universal features of the female role in physical and social reproduction. I will briefly examine several areas of research, including the analysis of pre-menstrual syndrome and mood, sex differences in the recognition of facial expressions of emotion and in aggression, and studies of the affective components and concomitants of motherhood. Feminist critiques of a number of these latter fields have been intensive, and I will draw on them while extending the analysis of the domain of emotion.
Studies of the relationship between mood and hormonal changes have focused on women's (rather than men's) cycles and in the process have discovered the hormonal disease of pre-menstrual syndrome. This syndrome is characterised by both physical pain and mood disturbances and has been attributed by the biomedical research community to hormonal imbalances in the women who suffer from it. The syndrome has been used to explain a host of emotions ranging from irritability and mood swings to depression, anxiety and panic attacks. A number of feminist critiques (Archer and Lloyd, 1985; Fausto-Sterling, 1985; Gottlieb, 1987; Whatley, 1986) have pointed out the weakness of the evidence for this syndrome. Assessment of women's mood is usually based on retrospective self-report via questionnaires (one popular version being titled the 'Menstrual Distress Questionnaire'), which allow women to draw on cultural knowledge about the relation between gender, emotion and hormones. Conversely, studies that disguise the purposes of the questionnaire show no significant pre-menstrual mood changes. The putative therapeutic effects of hormone injections are taken as primary evidence of the female hormonal basis for mood changes, but these studies have not been 'double-blind'. As Whatley argues, this biomedical discourse on emotions and gender may 'cause us to ignore the fact that our pre-menstrual mood changes ... may also correlate more closely to a monthly cycle of low bank balances than of hormonal fluctuations' (1986: 183). Moreover, the emotional symptoms of pre-menstrual syndrome can be seen as a discourse on both the good and the deviant woman, on the necessity of her emotional suffering and the abnormality of, especially, her anger or irritability (Gottlieb, 1987), both common symptoms attached to the syndrome. Normative academic and clinical work on pre-menstrual syndrome focuses on the emotionality of women as both common and yet as a 'symptom' in need of a cure. This research draws on the entrenched cultural view of emotions as sited in females, as natural in essence (like but independent of the 'naturalness' of females), and as irrational or pathological when they occur.
This line of research follows from and reinvigorates the cultural model in which women are more emotional than men because they are more tied to the biological processes that produce emotion. Wombs, menstruation, and hormones 'predict' emotion. A more tacit part of the cultural logic connecting women and emotion may arise from the view of women as biologically inferior both because they menstruate and because they are smaller, weaker and lack a penis. When viewed as a form of physical chaos or 'breakdown', emotion is one other form of biological weakness suffered by women.
A number of people in the interview study just described spontaneously articulated related ideas about the relationship between women, hormones, emotion and pathology. In several cases, they referred to research as the authoritative source of their assertions, although my argument is that the relationship between everyday and scientific ideas about women and emotion is dialectical rather than an idea system imposed hegemonically on a previously blank or very different lay model. According to one woman, a forty-eight-year-old telephone operator, 'women have been known to have different reactions to the same situation at different times of the month. And that's been a study. I've seen where some women can be downright dangerous, they could be potential killers.'
Another field in which some attention has been paid to sex differences is the study of facial expression of emotion. In one sociobiological account, female emotionality is a product of evolution. Babchuk et al. (1985) interpret studies showing that women are better able than men to read facial expressions of emotions in infants. In their view, this is the result of women's long history of being the primary caretakers of infants and the reproductive value of using these facial cues to detect infant distress. This argument is implausible on many grounds, not the least of which are the redundancy in infants of facial expression and other cues to discomfort, and the theoretically at least equal value of facial expression recognition skills for the prehistoric males, who, in many evolutionary accounts, were engaged primarily in defending the female and infant against threatening and dissembling outsiders. In addition, one of the central studies that demonstrates female superiority in decoding facial expressions of emotion (Hall, 1978) has been re-analyed and shown to account for less than 4 percent of the variance between individuals in facial expression recognition skills (Deaux, 1984, cited in Shields, 1987).
Despite its obvious problems, this account of the evolution of facial expression identification is a story with some power, as it draws on entrenched cultural narratives about women, motherhood, children and love. Here, the first premise is that women are more attuned to emotion in themselves and others. Unlike the pre-menstrual syndrome studies, however, female emotionality is celebrated here, with emotions taking on their positive sense of the interpersonally engaged, the unalienated. Women's emotionality becomes a skill and an asset. It is significant that the sociobiological account focuses on the use of that asset to detect distress (rather than, for example, threat). Distress, of course, calls for nurturance, whereas other facial expressions (in either infants or adults) might call for flight or defense, but only the former behavior is normative for women and mothers.
Another line of research, on sex differences in aggression, also draws on cultural views of emotion and women. This happens, first, because aggression, at least in the Western cultural view, is seen as retrospectively predictive of anger (Montagu, 1978). Anger is the one emotion that is exempted in everyday discourse from the expectation that women feel and express more emotion than men. It is in fact every emotion but anger that is disapproved in men and, conversely, expected in women (Hochschild, 1983). This gender stereotype has been shown to have been thoroughly learned by American children as early as the pre-school period (Bimbaum et al., 1980, cited in Shields, 1987). A recent, widely accepted, and often cited set of studies makes the parallel claim to have demonstrated a relationship between levels of the 'male' hormone testosterone and aggression. Fausto-Sterling (1985) demonstrates the weakness of the evidence for this claim and questions why it has been taken up so enthusiastically by so many.
The echoes of the lay view in the scientific are followed by the echoes of the scientific view in the lay on this point as well. A professional woman in her forties in the interview study commented on the association between aggression and gender: 'So far the research shows that, yes, little boys are inherently more aggressive than little girls. . . . I think it bothers me that there's a sex link with aggression. There are a couple of sex-linked ones that bother me but ... but I can't do anything about it.'
A number of studies that use the cultural logic of engendered emotion focus less on physiological differences to account for emotional ones than on universal functions and roles. In particular, they draw on the notion of women's reproductive role and the nurturing role and emotions that supposedly naturally accompany it. From ethological bonding theory (Bowlby, 1969) to some schools of feminism (e.g., Ruddick, 1980), focus is placed on the natural or inevitable emotional concomitants of motherhood (rather than fatherhood), including particularly the positive emotions of love, caring, and attachment. Bowlby follows the prevailing cultural emphasis on women's emotional qualities when he focuses on the emotions of women and their children. He wants to explain the intensity of the bond between mother and infant, and roots that explanation in an instinctual need for attachment in the infant and fear of separation. Feelings of love for the child on the part of the mother are naturalised (cf. Scheper-Hughes, 1985), and disastrous consequences are chronicled should the infant fail to receive sufficient quantities of mother love. These two facets of Bowlby's approach provide the carrot and stick of natural instinct and psychological harm to the child as reasons for continued emphasis on the need for emotionality in women.
Ruddick (1980), on the other hand, identifies 'resilient good humour and cheerfulness', 'attentive love' and 'humility' as among the central features of maternal virtue that follow from (rather than precede) the task of parenting and, by frequent correlation, the task of being female. From these perspectives, women are more deeply embedded in relationships with others (with the mother-infant bond as the primary example and the primary cause). This interpersonal engagement with others is what produces emotion, which is here defined as responses to others with whom one is involved. From the perspective of feminism, male individualism is antithetical to the experience of emotion (see also Chodorow, 1978).
The differences between these two perspectives on mothering and emotion are, of course, crucial. Bowlby-style bonding theory naturalises the connection between women and affect through evolutionary theory and is continuous with earlier theorising about the elevated moral status of women achieved through their divinely assigned and naturally embedded mothering skills. Feminist theory most often identifies the social division of labour rather than nature as the ultimate source of such emotional differences. Interestingly, however, both kinds of discourse on emotion elevate women (the first to a domestic pedestal, the second to self-esteem and/or the ability to resist patriarchy) by focusing on positive emotions such as love and by using 'emotion' in its positive Romantic sense of connection and disalienation.
Yet another view of the cultural view of women as emotional is found in the Parsonian normative construction of family roles, in which women are the 'expressive expert' and men the 'instrumental expert' (Parsons and Bales, 1955). These competencies are seen as an outcome of the domestic market spheres in which the genders differentially participate. Compare this notion, however, with the contradictory view of women's emotional impact on the family noted in the interview example and the child-rearing manual themes described earlier. The point may be that women are expected to be experts in noticing and attending to the emotional needs of others (also per Bowlby), not their own, which are rather objects of control or suppression because they, unlike the emotions of other family members, are defined as dangerous.
Hochschild's (1983) important feminist revision of Parsons and Bales's scheme paints emotion less as a skill than as a form of labour. Women are socially assigned a much heavier burden of emotional labour than are men. Hochschild's ideas contribute to a breaking down of the dichotomy of emotion and thought; they can also extend the notion of women's double day of domestic and wage labour as women are required to contribute both emotional and cognitive labour in both paid and unpaid spheres. In this and other feminist analyses, gender and emotion are related through the relations of production. For Hochschild, emotion is a personal resource that women must self-exploit more than men. It nonetheless remains a psychophysical fact, socially manipulated, rather than a discursive practice that constructs women as more emotional than men.
In sum, social science disciplines women and their psyches. It constructs emotion as an individual and intrapsychic phenomenon and evidences the same concern as lay discourse with the emotionality of women - its frequency, its intensity, its virtues as an emblem of female gender identity, but most of all, its danger and implicitly the need for its control.
I now return to the question of how these cultural notions about the emotionality of women, articulated in scientific discourse, are related to everyday discourse. The rhetoric of control that we first looked at was shown to reflect, in multiple and complex ways, relations of power between men and women, and to reflect them in ways that can be said, in large measure, to reproduce the 'emotional female'. By looking closely at some more microscopic aspects of the interview talk, however, we can see that gender differences are minimal, a fact that may speak to the gaps and fissures in the construction of a hegemonic discourse.
In two of the series of interviews, people were asked, first, to describe recent experiences with each of several common emotions and, second, to talk about how they feel about their work and family lives. In an analysis of a sample of 286 randomly selected interview statements that include direct reference to emotions, I have focused on the degree to which the statement 'personalises' the emotion experience - that is, on a variety of ways emotions, even as they are discussed, can be distanced from the self. It might be expected that women would use more personalising and immediate syntactic forms if they operate following the cultural model in which women are more emotionally expressive and have a more emotional self-identity.
Personalisation, or a non distancing discursive strategy, was indexed by four speech patterns (see Table 7.1), which will now be discussed.
First, the present tense rather than the past or conditional tense (e.g., 'I get [or am] angry whenever someone talks to me that way' compared with 'I was very angry'), is used. Tense obviously does several things to the meaning that audiences can make of a statement about emotion. First, it can move the emotion experience farther away from or closer to the self or another in time. Second, it can either generalise or particularise the experience; the use of the present tense, for example, can often include the implication that the emotion is habitually experienced by the subject. On both of these counts, the stereotype would lead us to expect more use of the present tense by women speakers. In fact, there is no difference between male and female speakers in the interview sample in the use of the present tense. If anything, men as a group make slightly (insignificantly) more use of it.
Second, another element of a personalising strategy might include the use of syntactic patterns that more directly portray the speaker as the experiencer of the emotion. Statements were coded as portraying the experiencer as the self, as another person (male, female, or gender unspecified), or as leaving the experiencer unspecified (e.g., 'It was a very strong feeling of hate' or 'And that developed a certain amount of hate toward that individual because of the fact that he . . .') or the emotion as an abstract entity with no particular experiencer (e.g., 'Well, hate and frustration usually go hand in hand, I would say' or "'Love" would be, I think, a good catchall phrase because . . .'). The category of self was further broken down by whether the self was portrayed as subject or object of the emotion experience (e.g., 'I'm very anxious about it' compared with 'It's making me angry just talking about this'). The belief in women's emotionality might lead to the expectation that women would more often portray the self (particularly the self as subject rather than object) as the experiencer of emotion, whereas men would portray the other as the experiencer or leave the latter ambiguous.
Table 7.1 Personalisation in syntax
1 Present tense
'I get [or am] angry whenever someone talks to me that way.'
'I was very angry.'
2 Experiencer of the emotion discussed
Self, as subject of emotion experience
'I'm very anxious about it.'
Self, as object of emotion experience
'It's making me angry just talking about this.'
Other person (male, female, or gender unspecified)
'My father was very annoyed with me for going into that field.'
'It was a very strong feeling of hate.'
'And that developed a certain amount of hate toward that individual because of the fact that he . .
None - emotion as an abstract entity
'Well, hate and frustration usually go hand in hand, I would say.' "'Love" would be, I think a good catchall phrase because . .
3 Cause or elicitor of emotion
'They were angry at me.''
'I just kind of giggled and made her even angrier.'
'I hate her because she was mean enough to tell me that.' 'I'm deathly afraid of dentists.'
'The most anxious moment I had ... was my first performance with the ... Choral Society.'
'I hate going out unless I really have to.'
'He loves books'.'
'Lots of little things are frustrating.'
'I can't talk anymore, I start screaming to begin with, when I'm really angry.'
'I [or she] wasn't angry.'
In the interview sample, it is not significantly more common for women, in their discussions of emotion, to focus on the experiencing self as the subject versus the object of the emotion, nor is it more common for men to leave the experiencer unspecified or abstract. In addition, neither women nor men are more likely to portray others as opposed to the self as the experiencer of the discussed emotion. Women and men speak more alike than differently in this sample when discussing the experiencer of emotions.
Third, statements about emotion usually contain an implicit or explicit etiology, that is they specify the cause (usually by specifying the object) of the feeling. Personalising strategies might include identification of either the self or, secondarily, another person as the ultimate cause of the emotion (rather than the use of syntactic patterns that obscure or fail to identify the cause). Statements were coded as portraying the cause as either the self (e.g., 'They were angry at me' or 'I just kind of giggled and made her even angrier'), another person (e.g., 'I hate her because she was mean enough to tell me that' or 'I'm deathly afraid of dentists'), an event (e.g., 'The most anxious moment I had ... was my first performance with the ... Choral Society' or 'I hate going out unless I really have to'), an object (e.g., 'He loves books'), or as leaving the cause unspecified (e.g., 'Lots of little things are frustrating' or 'I can't talk anymore, I start screaming to begin with, when I'm really angry') (cf. Shimanoff, 1983). Given the associations between gender and affect I noted earlier, we might expect that women more than men would see other people as intimately involved in their own emotion experience and themselves as evoking emotion in others, rather than seeing events as triggering emotion in themselves or failing to specify a cause. The latter strategy can be associated with the view of emotion as nonsensical, irrational, or without ascertainable cause. In fact, there are no significant gender differences in the use of personal versus impersonal causal attribution, nor do women use self versus other attributions more than men.
Finally, a number of statements about emotion in the interviews are essentially denials of emotion in the self or the other (e.g., 'I [or she] wasn't angry'). The stereotype might lead us to expect more negation in general from men and more negation of particular kinds of female-linked emotions (which include most emotions except anger) by men and of male-linked emotions (notably anger) by women. Here again, women's and men's speech are indistinguishable in terms of the proportion of emotion states that are negated as they are discussed.
The absence of extensive differences might be attributed to the special nature of the people interviewed, all of whom agreed beforehand to talk with a stranger about emotion. The results are consistent, however, with a study of gender differences in emotion language used by Shimanoff (1983), who did a similar analysis of the tape-recorded natural conversations of a number of American college students and married couples, and found few differences in male and female conversations that included reference to emotions. The results are also consistent with the trend in studies of psychological and linguistic sex differences in general, which have tended to show far fewer differences than researchers both expected - on the basis of cultural stereotypes about distinctive male and female styles of thinking, behavior, and speech - and then often found in self-fulfilling fashion. The absence of differences is more significant given the syntactic nature of the evidence examined; Shibamoto (1987) has concluded that gender differences that are not a response to audience expectations about particular gender identities are more likely to be found in syntactic patterns of use because they are typically outside of our awareness and hence of our easy manipulation, unlike semantic patterns such as those having to do with the notion of 'control' we examined earlier.
In all societies, body disorders - which emotion is considered to be in this society - become crucial indicators of problems with social control and, as such, are more likely to occur or emerge in a discourse concerning social subordinates. Foucault has made the claim that power creates sexuality and its disciplining; similarly, it can be said to create emotionality. The cultural construction of women's emotion can thus be viewed not as the repression or suppression of emotion in men (as many lay people, therapists, and other commentators argue) but as the creation of emotion in women. Because emotion is constructed as relatively chaotic, irrational, and antisocial, its existence vindicates authority and legitimates the need for control. By association with the female, it vindicates the distinction between and hierarchy of men and women. And the cultural logic connecting women and emotion corresponds to and shores up the walls between the spheres of private, intimate (and emotional) relations in the (ideologically) female domain of the family and public, formal (and rational) relations in the primarily male domain of the marketplace.
Rubin has remarked of sexuality that 'There are historical periods in which [it] is more sharply contested and more overtly politicised' (1984: 267). Emotionality has the same historical dynamism, with shifting gender relations often appearing to be at the root of both academic and lay struggles over how emotion is to be defined and evaluated. In other words, the contemporary dominant discourse on emotions - and particularly the view that they are irrational and to be controlled - helps construct but does not wholly determine women's discourse; there is an attempt to recast the association of women with emotion in an alternative feminist voice.
Feminist treatments of the question of emotion (e.g., Hochschild, 1983; Jagger, 1987) have tended to portray emotions not as chaos but as a discourse on problems. Some have contested both the irrationality and the passivity of feelings by arguing that emotions may involve the identification of problems in women's lives and are therefore political. Talk about anger, for example, can be interpreted as an attempt to identify the existence of inappropriate restraint or injustice. Sadness is a discourse on the problem of loss, fear on that of danger. By extension, talk about the control of emotions would be, in this feminist discourse, talk about the suppression of public acknowledgement of problems. The emotional female might then be seen not simply as a mythic construction on the axis of some arbitrary cultural dualism but as an outcome of the fact that women occupy an objectively more problematic position than does the white, upper-class, Northern European, older man who is the cultural exemplar par excellence of cool, emotionless rationality. According to a feminist analysis, whether or not women express their problems (i.e., are emotional) more than men, those women's audiences may hear a message that is an amalgam of the orthodox view and its feminist contestation: 'We (those) women are dangerously close to erupting into emotionality/pointing to a problem/ moving toward a social critique.'
Abu-Lughod, Lila (1986) Veiled Sentiments.. Honour and Poetry in a Bedouin Society. Berkeley: University of California Press.
Abu-Lughod, Lila and Lutz, Catherine A. (1990) 'Introduction: emotion, discourse, and the politics of everyday life', in C.A. Lutz and L. Abu-Lughod (eds), Language and the Politics of Emotion. Cambridge: Cambridge University Press.
Archer, John and Lloyd, Barbara (1985) Sex and Gender. Cambridge: Cambridge University Press.
Asad, Talal (ed.) (1973) Anthropology and the Colonial Encounter. New York: Humanities Press.
Babchuk, Wayne, Hames, Raymond and Thompson, Ross (1985) 'Sex differences in the recognition of infant facial expressions of emotion: The primary caretaker hypothesis', Ethology and Sociobiology, 6: 89 101.
Bimbaum, D.A., Nosanchuck, T.A. and Croll, W.L. (1980) 'Children's stereotypes about sex differences in emotionality', Sex Roles, 6: 435-43.
Bowlby, John (1969) Attachment and Loss, Vol. 1. London: Hogarth Press.
Brenneis, D. (1990) 'Shared and solitary sentiments', in C.A. Lutz and L. Abu-Lughod (eds), Language and the Politics of Emotion. Cambridge: Cambridge University Press. Chapter Six.
Cancian, Francesca (1987) Love in America: Gender and Self-Development. Cambridge: Cambridge University Press.
Chodorow, Nancy (1978) The Reproduction of Mothering. Berkeley: University of California Press.
Deaux, K. (1984) 'From individual differences to social categories: Analysis of a decade's research on gender', American Psychologist, 39: 105-16.
Fausto-Sterling, Anne (1985) Myths of Gender.. Biological Theories of Women and Men. New York: Basic Books.
Foucault, Michel (1980) The History of Sexuality, Vol. 1. New York: Vintage.
Gottlieb, Alma (1987) 'American PMS: A mute voice'. Paper presented at the annual meetings of the American Anthropological Association, Chicago.
Haan, Norma, Bellah, Robert, Rabinow, Paul and Sullivan, William (1983) Social Science as Moral Inquiry. New York: Columbia University Press.
Hall, J. (1978) 'Gender effects in decoding non-verbal cues', Psychological Bulletin, 85: 845-75. Haraway, Donna (1986) 'Primatology is politics by other means', in Ruth Bleier (ed.), Feminist Approaches to Science. New York: Pergamon Press. pp. 77-118.
Hochschild, Arlie (1983) The Managed Heart: Commercialisation of Human Feeling. Berkeley:
University of California Press.
Holland, Dorothy and Quinn, Naomi (eds) (1987) Cultural Modes in Language and Thought. Cambridge: Cambridge University Press.
Jagger, Alison (1987) 'Love and knowledge: Emotion as an epistemic resource for feminists',
Ms. in possession of author. Department of Philosophy, University of Cincinnati.
Jordanova, L.J. (1980) 'Natural facts: A historical perspective on science and sexuality', in Carol MacCormack and Marilyn Strathern (eds), Nature, Culture, and Gender. Cambridge: Cambridge University Press. pp. 42 69.
Kirkpatrick, John (1987) 'Representing the self as 'You' in American discourse'. Paper presented at the annual meetings of the American Anthropological Association, Chicago.
Lutz, Catherine (1988) Unnatural Emotions: Everyday Sentiments on a Micronesian Atoll and Their Challenge to Western Theory. Chicago: University of Chicago Press.
Maher, Vanessa (1984) 'Possession and dispossession: Maternity and mortality in Morocco', in H. Medick and D. Sabean (eds), Interest and Emotion. Cambridge: Cambridge University Press. pp. 103-28.
Martin, Emily (1987) 'The ideology of reproduction: The reproduction of ideology'. Paper presented to the Upstate New York Feminist Scholars' Network, September.
Montagu, Ashley (ed.) (1987) The Learning of Non-Aggression. Oxford: Oxford University Press.
Parsons, Talcott and Bales, Robert (1955) Family, Socialization, and Interaction Process. Glencoe, IL: Free Press.
Rosaldo, Michelle Z. (1 984) 'Toward an anthropology of self and feeling', in R. Shweder and R. Le Vine (eds), Culture Theory: Essays on Mind, Self and Emotion. Cambridge: Cambridge University Press. pp. 137-57.
Rosaldo, Renato (1977) 'The rhetoric of control: Ilongots viewed as natural bandits and wild Indians', in B. Babcock (ed.), The Reversible World.. Symbolic Inversion in Art and Society. Ithaca, NY: Cornell University Press. pp. 240-57.
Rubin, Gayle (1984) 'Thinking sex; Notes for a radical theory of the politics of sexuality', in Carol S. Vance (ed.), Pleasure and Danger: Exploring Female Sexuality. Boston: Routledge and Kegan Paul.
Ruddick, Sara (1980) 'Maternal thinking', Feminist Studies, 6: 70 96.
Sampson, E.E. (1981) 'Cognitive psychology as ideology', American Psychologist, 36: 730-43. Scheper-Hughes, Nancy (1985) 'Culture, scarcity, and maternal thinking', Ethos, 13: 291-317. Scheper-Hughes, Nancy and Lock, Margaret (1987) 'The mindful body: A prolegomenon to future work in medical anthropology', Medical Anthropology Quarterly, 1: 6-41.
Shibamoto, Janet (1987) 'The womanly woman: Manipulation of stereotypical and non stereotypical features of Japanese female speech', in S. Philips, S. Steele and C. Tanz (eds), Language, Gender, and Sex in Comparative Perspective. Cambridge: Cambridge University Press. pp. 26-49.
Shields, Stephanie A. (1987) 'Women, men and the dilemma of emotion', in P. Shaver and C. Hendrick (eds), Sex and Gender. Newbury Park, CA: Sage Publications. pp. 229 50.
Shimanoff, Susan (1983) 'The role of gender in linguistic references to emotive states', Communication Quarterly, 30: 174-9.
Smith-Rosenberg, Carroll (1972) 'The hysterical woman: Roles and role conflict in 19th-century America', Social Research, 39: 652 78.
Stoler, Anne (1985) 'Perceptions of protest: Defining the dangerous in colonial Sumatra', American Ethnologist, 12: 642-58.
Strathern, Marilyn (1980) 'No nature, no culture: The Hagen case', in Carol MacCormack and Marilyn Strathern (eds), Nature, Culture, and Gender. Cambridge: Cambridge University Press. pp. 174-222.
Taussig, Michael (1984) 'Culture of terror - Space of Death: Roger Casement's Putumayo report and the explanation of torture', Comparative Studies in Society and History, 26: 467-97.
Whatley, Marianne (1986) 'Taking feminist science to the classroom: Where do we go from here?', in Ruth Bleier (ed.), Feminist Approaches to Science. New York: Pergamon Press. pp. 181-90.
Williams, Raymond (1977) Marxism and Literature. Oxford: Oxford University Press. Woolard, Kathryn (1985) 'Language variation and cultural hegemony: toward an integration of sociolinguistic and social theory', American Ethnologist, 12: 738-48.
An earlier version of this chapter was presented on the panel 'Emotion and Discourse' at the annual meeting of the American Anthropological Association, Chicago, 18-22 November 1987. The draft has benefited greatly from the comments of Lila Abu-Lughod and Steven Feld. The research on which this chapter is based has been conducted with grants from the State University of New York Foundation and the National Institute of Mental Health and with the help of many people. Kathryn Beach, Robin Brown, Paula Bienenfeld and Walter Komorowski assisted in interviewing and transcription, and the expert analytic work of Angela Carroll and Marion Pratt helped give the chapter its form.
 The actual process by which these models of gender and emotion are acquired is a fascinating but unexplored question. We might expect that it includes, in part, the child's reasoning from the culturally assigned authority and control of the male teacher to a lack of emotion, the latter perhaps already having been learned to require 'strength' and 'control' to master - in other words, to generalise from the dominant position of males to a presumed lack of emotion (a process that might also have occurred in her teachers' views of themselves).
 The method used in looking at the transcripts draws on recent developments in the ,cognitive' study of cultural meaning. These focus on the analysis of extended and relatively natural conversations for the cultural knowledge or cultural models (Holland and Quinn, 1987) evident, if not always explicitly stated, in them. By looking at such things as syntax, metaphor, or the prepositional networks underlying the sensibility of sentence order, it is possible to draw inferences about the kinds of models individuals are using or, perhaps more aptly, to draw inferences about the kinds of inferences listeners can make about what the speaker has left unsaid but likely wants understood.
 There are 180 instances in those parts of the women's transcripts analysed so far, and 85 instances in the men's, with each set of transcripts being of approximately equal length.
 1 have found Woolard's (1985) analysis of the nature of hegemonic and oppositional forms of language use very productive in formulating what I have to say here.
 Martin (1987) has examined the American discourses on reproduction and women's bodies and has rigorously uncovered the contradiction between a view of uterine contractions during childbirth as involuntary and a view of the woman as in fact in control of the labour process. The women she interviewed about their birth experiences spoke very similarly to the women described in this chapter about their sense of control over the physical process and over their cries of pain and pleasure during labour and birth. She notes a class difference, however, with middle-class women speaking with more approval of control than working-class women. We might then expect men also to express more concern with and approval of control of emotion, which is not the case here. This is certainly a problem worthy of more study, particularly a delineation of what kinds of control of which domains appear to emerge from what kinds of experience within hierarchical systems.
 Acknowledgement of one's emotionality may mean very different things to female and male audiences. Women may announce to each other shared identity and solidarity, while asserting difference, submission, or defiance when making similar statements to men.
 Abu-Lughod's (1986) study of the Awlad 'Ali represents the most detailed and eloquent example of how, in another cultural system, the particular kinds of emotions allocated to and voiced by women articulate with other aspects of their ideological and social structural positions.
 This group of studies obviously follows in the tradition of centuries of expert explanations of hysteria. Although there have been many versions of the explanation (such as one nineteenth-century account that diagnosed its origins as an empty womb and a childhood where the restraint of emotion was not taught [Smith-Rosenberg, 1972]), they have been organised around the connection between female physiology and mood.
 Shimanoff (1983) found that male and female speakers did not differ in the number of affect words they used, in the tense, valence, or source (similar to the notion of 'elicitor' used here) of statements about emotion. She did find, however, that males made more reference to their own emotions than to those of other people when compared with females.
 The resurgence of interest in emotion in the late 1970s and 1980s across the social sciences may in part be the result of the feminist movement's revalorization of all things traditionally associated with women (Margaret Trawick, personal communication). Changing gender relations may also be at the root of the reinvigoration of a long-standing Western discourse on the value of emotional expression; the current debate pits expressionists, for whom healthy emotions are vented ones, against those who would dismiss the latter as 'self-indulgent' or 'immature'. This debate no doubt draws in a complex way, in each concrete context in which it occurs, on the gender ideologies and conflicts of the individual participants.