SEX AND GENDER: A CRITICAL REALIST APPROACH Caroline New Much gender theorising, I shall argue, is in an impasse from which critical realism indicates a way out. In the heyday of second wave feminism, structural inequalities such as women’s under-representation in political and economic leadership, their lower pay, sexual objectification, and widespread violence by men against women, were seen as in need of explanation, the better to transform them. ‘Radical’ or ‘revolutionary’ feminists saw these inequalities as symptoms of the enduring differences between women and men, while liberal, Marxist and socialist feminists understood them as resulting from a motivated refusal to acknowledge the similarities. All saw womanhood as a basis for grounded solidarity – at least while the oppression of women persisted. To end it, different sources of power had to be mobilised. For radical feminism, the power of women, united across all differences to refuse cooperation with the patriarchy, could bring about a cultural revalorisation of the subordinate terms in the repeated dualisms of Western thought: man, woman; culture, nature; human, animal; reason, emotion. For liberal feminism, women united had to harness the power of the state to render its laws and practices consistent and genderequal. For Marxists and socialists, women had to unite with men to seize the state, and at the same time use female solidarity to ensure that the interests of half of the working class did not get forgotten. Similarities, differences. From the 1980s, feminist critics of second stage feminism argued that the call to ‘sisterhood’ obscured the differences between women, making middle-class white Western womanhood the default position, and membership of other oppressed groups mere ‘add-ons’ (Spelman, 1988, Mohanty, 1992). ‘Woman’ was a fragmented identity, not a unitary one. It was doubtful whether it was a usable basis for political solidarity, since so many oppressive relationships took place between women themselves. Feminist politics became increasingly fragmented, with single issue campaigns and coalitions replacing the broad sweep of a movement – corresponding to the ‘anti-foundationalist’ epistemological theory that allows an ontology of parts but forbids one of wholes, and views abstraction as a dangerous betrayal of concrete particularities. The politics of ‘recognition’ (of subordinated 1 identities) became increasingly important, and with it the idea of struggle as situated primarily in the discursive realm (Fraser and Honneth, 1998). But how could subordinated identities be revalorised while leaving hegemonic identities intact? ‘Woman’ was first deconstructed into fragmentary identities, and then, as we know, postmodern thinking transcended and rejected identities altogether. The very gender categories ‘woman’ and ‘man’ were themselves challenged. As irrealism (strong social constructivism) become increasingly fashionable, the social world was identified with the cultural realm, with discourse or ‘text’. Postmodern forgetting of the material has always been an uneven tendency. Sometimes material differences, including the inequalities I mentioned, are seen as the products of certain discursive regimes, sometimes as expressive of them (the former, but not the latter, is an unacknowledged causal thesis). In either case, while institutionalised inequalities between women and men persist, and much academic feminist work continues to underlabour for feminist activism, most gender theorising pays less attention to material forms of gender oppression and more to the categories in terms of which gender relations are described. In a poststructuralist classic, Weedon remarks that the discursive is the source of power, ‘exercised through the constitution of subjectivity within discourse and the production of social agents’ (Weedon, 1997: 163), and therefore assumes it is the effective locus for action. In contrast, feminism has historically used ‘explanatory critiques’(Collier, 1994), identifying false beliefs or unmet needs, explaining the mechanisms through which these are reproduced, criticising the states of affairs which allow the exercise of these mechanisms, and recommending action to end them. Where the ‘discursive turn’ has been influential, such critiques have been undermined and feminist activism weakened (New, 2003). The deconstructive politics of ‘disidentification’ of gender categories, for example through parody or mimesis (Braidotti, 1997) lack any strategy for structural change. We are not told how disidentification influences subjectivity, and thus agency. Unsurprisingly, the aspects of the gender order which receive most attention are the socio-cultural ones where such means of struggle have most chance of success (Butler, 1995). 2 When social structures and their effects are misleadingly reduced to discourse, or situated within it, causal explanation is rendered impossible. This is the impasse of postmodern gender theorising. It is not seen as an impasse, because of a deep suspicion of causal argument, of abstraction, and of generalisation. Flax, for instance, rejects all ideas of human nature, since lest they be used as ‘the basis for causal, deterministic (as opposed to interpretative) explanations of human behaviour or culture’ (Flax, 2002: 1042). No distinction is made between causation, where the causal powers of A, when exercised, result in a tendency for B to happen in certain contexts (while if counter tendencies are operating B may not happen), and determinism, where an instance of A is inevitably followed by B. Yet beneath much postmodern gender theorising (even by those who most fête contingency, such as Butler) lie post-Lacanian narratives of sexual difference as a basic (determining) organising principle of discourse – at least so far in human history. Unless subjectivity is treated as a mere epiphenomenon of discourse (which would rule out resistance to the ‘phallogocentric order’), these must be causal narratives. Whatever brings about change is a cause, and in such accounts enduring discursive mechanisms produce and form gendered subjectivity. So much for the cul-de-sac into which postmodern feminism has led us. I now go on to argue that critical realism offers a way forward which is neither foundationalist nor determinist. A missing concept in postmodern feminism is ‘ontological stratification’. Critical realism and the ontology of sex and gender Not only is the world differentiated in itself, independent of human thinking: its properties and powers are also stratified. The levels in question, I will go on to show, are not neat and static like the layers in a sandwich cake, but co-present and interpenetrating (New, 1996, Dyke, 1988). The upshot is openness and complexity. My argument in the rest of this essay will be that sexual difference is real, and salient in all imaginable human societies; and that it is a causal condition for social gender orders, though it is compatible with many and various forms of these. This position is not unique to critical realism (see, for instance, Connell, 2002), but it is rendered far more compelling by the use of critical realist concepts of stratification and emergence. 3 If this case is so compelling, why has it failed to compel? I can hazard an informed guess at a part of the answer. It would be hard to exaggerate the harm that biological determinism, as a legitimating ideology, has done and continues to do to women. Without any concept of ontological stratification, the admission that sexual difference has causal powers to enable and constrain, seems akin to nineteenth century arguments that women in higher education would lose their breasts and shrivel their wombs (Sayers 1982). If traumatic historical events have determinate effects on national consciousnesses decades or even centuries later, small wonder that many feminists still prefer to work on representations, eschewing their referents as imaginary or as wholly docile hangers on. Transitive and intransitive dimensions A key distinction in critical realism is between the transitive and intransitive dimensions. Things and processes exist in the world, ‘active independently of their identification by human beings’ (Bhaskar, 1989:11). If we exemplify this with reference to sex, we can say that human beings are (almost all) sexually dimorphic, female and male, whether and however they conceptualise this difference, and this dimorphic structuring is active, causally powerful, enabling different reproductive roles and certain sexual possibilities and pleasures, and ruling out others. The (fallible) claim I have just made exists in the transitive dimension, but it refers to the intransitive dimension. The effects of the properties of bodies to which I am referring are distinct from the effects of the concepts I am using to do so. The transitive dimension is that of knowledge, of explanation, and of the conceptual and cognitive raw materials which enable these. Its contents – ideas, taxonomies, theories, ways of seeing the world - are themselves real and causally powerful, and can become objects of knowledge. Although our conceptual work takes place within and may reproduce or modify the transitive dimension, the structured contents of that dimension are also external to human subjects and constrain and enable what we can do. As such the transitive dimension has an intransitive aspect, as do the practices and institutions of particular ‘discursive regimes’. The power of discourse to bring the natural world into line is well exemplified by Judith Butler’s discussion of how babies with ambiguous genitalia are surgically 4 adjusted to fit the social requirement that all citizens be readily identifiable as either male or female (Butler, 1993, Preves, 2001). But such practices do not justify Butler’s further conclusion that sexual difference is merely conceptual, for if bodies were did not have particular characteristics (in the intransitive dimension) such surgical alteration would not be possible. Butler collapses the intransitive into the transitive dimension. For her, although bodies may in some sense exist outside of discourse, attempts to discover or describe their characteristics are constructing what they claim to represent. Since the transitive dimension is also intransitive, by studying the properties of its contents we may come to understand their interrelationships, histories, powers and effects. Among these properties are those of being good or bad abstractions, theories, or representations in relation to their referents. ‘The intransitivity of both beliefs and meanings’ co-exists with ‘their susceptibility to scientific explanation and hence critique, in a spiral (which reflexively implicates social science as a moment in the process that it explains’ (Bhaskar,1989: 22). The sociology of knowledge is a valid field, but it does not entail ontological relativism. The stratification of reality The intransitive dimension is not flat: reality is stratified. It is the twin ideas of ‘levels’ and ‘emergent properties’ that give critical realism its ‘depth’. Most frequently cited is Bhaskar’s distinction between the actual, the empirical and the real (Collier, 1994: 44), although this is just a beginning way of characterising the complexity of ontological stratification. The ‘actual’ refers to things and events in their concrete historicity, only some of which will ever be known or experienced by human beings. The empirical is a subset of the actual, consisting of those phenomena which are experienced. Experience must here be understood not as given but as socially produced from the actual: ‘in this sense it is the end, not the beginning of a journey’ (Bhaskar 1998: 43). The actual and the empirical are, of course, both real, and as such are contained within the third domain. But the domain of the real also includes things which are real in a different way. These are the structures of objects (e.g. the internal relations of their constituent parts, such as the interrelationship of family members) and the 5 emergent properties to which their structuring gives rise (e.g. the powers and susceptibilities of a family, which are not the same as those of its members). Since these powers of structures, when exercised, may bring about certain effects, we often speak of ‘generative mechanisms’. This metaphor needs to be used with caution. The relationship between cause and effect is not ‘mechanical’ in the sense of predictable and automatic, since in an open world many things are always happening at once. Mechanisms, we shall see, operate at many levels in the social world, and are variously related. One may form the ground for the possibility of another, higher level mechanism, which emerges in certain circumstances as a set of powers, which may begin to be exercised only when yet another mechanism is triggered. Whether a generative mechanism does generate the effects it could, may depend on the current state of something else again. In poststructuralist and deconstructionist feminist texts, we often come across a warning against ‘universal causal claims’ (e.g. Hawkesworth, 1997: 654). According to anti-foundationalist rhetoric, to strive to explain on too big a scale is itself oppressive. This fear seems to rest on a philosophical error. Universal causal claims, if worth their salt, are about the generative powers of mechanisms, not about any (Humean) constant conjunction of cause and effect. The training of men in military camps aims to mobilise certain human psychological powers – to split off feelings – and liabilities – to lose access to split off feelings. If these mechanisms are exercised, and if the context in which this happens allows their causal powers to be realised, there will be certain outcomes at the level of events, which could include heroic defence of comrades, massacres and rapes. But counteracting mechanisms can certainly exist. We are talking about tendencies here, and claims about tendencies cannot be falsified by counter instances, which are inevitable in an open system. ‘Actualism’ is the view that accepts the reality of things and events, but denies that there are ‘underlying structures which determine how the things come to have their events, and instead locates the succession of cause and effect at the level of events: every time A happens, B happens’ (Collier, 1994: 7). Footbinding was experienced by millions of women in China between the mid tenth and early twentieth centuries (Feng, 1994). It is sometimes explained in terms of events: ‘The emperor Li Yu ordered his concubine to bind her feet’ at one point in history, while at another ‘Mao 6 Tse Tung forbade the practice’. No doubt the emperor’s original order (if it really happened) was preceded by other happenings in his personal life which might shed light on his sexual predilections, but even these cannot be explained by stringing events on a Humean chain without reference to underlying mechanisms. For one man’s demand to give rise to a fashion, and that fashion to become a widespread and long lasting institution, many things had to be the case. The emperor’s authority and symbolic role had to be such that he could set fashions. Women had to have feet with certain powers and liabilities, such that if most of the toes were broken and bound under when the child was about 3, the foot would remain small. Human sexuality, in particular male sexuality, had to be culturally malleable, such that a small foot and a mincing gait in females could become eroticised. Beliefs about sexual difference had to legitimise such a practice. Power relations between men and women, and adults and children, had to be such that the female relatives would carry out the initial act and the girl child would permit it. The economy had to allow the disabling of most middle and upper class women. These and many other structures (or mechanisms) at different levels had to exercise their emergent powers in particular ways and particular relationships to each other. Events and things cannot be attributed to one particular level of reality, while mechanisms may be. People, for instance, are themselves structures with emergent properties. People cannot sensibly be characterised as ‘physical’ or ‘chemical’ or ‘biological’ or ‘psychological’ or ‘social’. Some of their emergent properties (not necessarily instantiated in any one individual) could be characterised in this way – we might call the capacity for empathy a psychological property, and the capacity to produce eggs a biological property, and the causal powers emerging from membership of a particular group social properties. All of these mechanisms exist simultaneously and in non-additive interactions, enabling and constraining and otherwise affecting each other. Thinking again of footbinding, we can see that the musculoskeletal structure of the human foot and the diachronic processes of development are ontologically prior to the practice. They in no way determine it, but they make it possible. Determinate psychological structures of sexual development emerge, I would argue, out of the structuring of human nervous systems (the possibilities they allow) on the one hand, and the symbolic systems of human cultures on the other. If this is right, such an emergence out of two co-existing structures which are 7 themselves at very different levels shows that ontological stratification is not straightforwardly hierarchical (cf Collier, 1994: 132). In other words, humans have evolved as a species in which sexual feelings can become attached to symbols and markers of gendered power relations (exactly how this comes about has been variously theorised, with, for instance, psychoanalysis and humanistic psychology proposing different sorts of mechanism). The reality of sex Having introduced and illustrated some critical realist concepts and distinctions, I can now go on to sketch a critical realist theory of gender. Postmodern feminists sometimes gesture at a favourite target, ‘the base/superstructure distinction’, the invalidity of which is supposed to be patently obvious. For Rita Felski, much poststructuralist thought itself ‘remains caught within a surface/depth opposition’ (1997:17). A critical realist theory of gender would indeed identify structures at lower levels (the ‘base’ or ‘depth’), and in a non-reductionist way seek to explain how these condition possibilities at higher levels (the ‘superstructure’ or ‘surface’). It would not make the mistake of assuming that the second terms were mere expressions of the first, or that the causal relationship was invariably one way. The much criticised ‘second stage’ feminist distinction between sex and gender is a good place to begin. For some postmodern feminists, sexual difference is most certainly real, but must be understood as discursive rather than material (Felski, 1997). Repetitive gender practices give rise to ‘constitutive constructions’ of the body as material. Extradiscursive sexual difference is an illusion without which we could not live or think. ‘”Sex” is part of a regulatory practice which produces the bodies it governs’ (Butler, 1993: 1). Whence these surprisingly constant dualisms? We have to turn to the ‘law of the father’ in Lacanian psychoanalysis to understand this ‘base’. For other postmodern feminists (and Butler too is sometimes in this camp) bodies are, indeed differentiated. They vary in the amount of subcutaneous fat, width of the pelvis, distribution of hair, shape of internal and external genitalia, number of X and Y chromosones and so on. Such differences are not stratified, nor otherwise theorised. This is a realist conception, but an empiricist, actualist one. The categorisation of such differences into ‘male’ and ‘female’ is seen as arbitrary, comparable to dividing 8 humanity according to whether they had attached or unattached ear lobes, ignoring or operating on ambiguous cases, and organising society around this distinction. The distinction between ‘males’ and ‘females’ is entirely socially constructed, since there is no dimension on which these two groups can be reliably distinguished. Women have moustaches, men get breast cancer, and the plethora of supposed expressions have no essence behind them. For instance, Hawkesworth asserts that gender narratives (she would include accounts of sexual difference in these) are ‘totalising fictions that create a false unity out of heterogenous elements…’ which have ‘little to offer in a postmodern world that understands the body, sex and sexuality as socially constructed’ (1997: 651-2). Such thinkers are arguing that sex is like ‘race’, and should similarly be embellished with scare quotes. ‘Race’ has been exposed as a series of constructions from disparate elements with no intrinsic significance. In social realist terms, ‘race’ is a bad abstraction, theoretically misleading (Sayer, 1992, Danermark et al. 2002). Differences in skin pigment are real enough, and so are the other markers of ‘race’ at the level of the empirical. However, at a deeper, structural level, discrete races do not exist. The genotypical differences between them are less significant than the differences within these groups, counter to scientific racism. ‘Race categories have far more to do with politics than they have to do with science’ (Carter, 2000:3).The argument against the ontological validity of race classifications is that they do impose a spurious unity on groups of humans who are actually diverse – and do so for the worst of reasons. Carter, a social realist, argues that because there are no such things as races, the sociological use of the term to refer to beliefs in and discourse about races, and social relations and practices understood in terms of race, is highly misleading; preventing analysis and explanation of what is going on (Carter, 2003). Analogously, poststructuralists argue that all talk about gender should be detached from ideas about sex, or alternatively that sex should be seen as a creation of gender. ‘What is really going on’ is entirely within discourse. However, sex is not like ‘race’. Human beings have evolved as sexual creatures, and the markers of sex are not unrelated, nor of equal weight. 9 Approximately 99.6%… of all humans are born biologically male or female. Because several processes are involved in the sexual development, the laws do not operate with one hundred percent predictability… hypospadias, the condition in which the male urethral opening extends along the underside of the penis, is… made explicable when the common tissue origin of all genitalia is acknowledged. The condition reflects a shortage of the hormones necessary to convert the bipotential fetal genitalia into a male penis; in other words, hypospadias is a penis with a female urethra… Thus, the phenomenon of intersexuality is a reflection of the causal structure of sex, a real structure that intimately links male and female bodies while explaining differences. It is not the effect of a “resistance” to discourse or a lawless chaos or even a continuum… (Hull, forthcoming, my emphasis). Genetic sex, Hull points out, cannot (at this time) be altered. It is at a lower level, ontologically speaking, than its phenotypical effects (to which environmental factors also contribute). Postmodern writers massively exaggerate intersexuality and misrepresent sexual attributes as continuous rather than as distributed dimorphically, despite the variations and overlaps on any one dimension. Do these variations mean that sexual difference is not real? Once again postmodern feminists have higher standards than anyone else for categorisation. Hawkesworth maintains that females and males are not ‘natural kinds’ because there is no set of properties possessed by every member of each of these groups (1997). From a realist point of view, ‘natural kinds’ are so called because they tell us something about the causal structures of the world. Causally important properties are contingently clustered, but in such a way that the presence of some properties renders the presence of others more likely – because there are common underlying properties that tend to maintain the clusters of features (Keil, 1989:43). Biological kinds can never meet the essentialist criteria postmodern thinkers implicitly require (Boyd, 1992). Biology is messy and complex, and its regularities take the form of tendencies rather than laws. In the case of sexual difference, these tendencies are strong, ‘the genotypic and 10 phenotypic division of bodies into two sexes crosses species and millenia’ (Hull, forthcoming). Sexual difference, then, is a ‘good’ abstraction. Pace deconstructionists, it brings together characteristics that are internally connected, and the connections in question are substantial, not merely formal (Danermark et al 2002). Sexual difference is also socially significant. This significance is an emergent property of human societies, given the sorts of (sexed, mortal) beings humans are. While ‘race’ could be exploded and lose its credibility as a means of classification, this could not happen to sexual difference. It is relatively easy to imagine a world without ‘races’ and racism. It is hard to imagine a world where sex difference was socially so insignificant that there was no sex-gender order. Means of reproduction will probably always be a salient feature of human societies, variously institutionalised but always deeply meaningful in their symbolic orders. Perhaps if humans lived to be two hundred, and their adult sexual and reproductive life lasted for only one eighth of that life span, gender arrangements might become far less significant and gender identity a temporary phenomenon, to be shrugged off at the end of the reproductive period as some workers seem to shrug off their occupational identities on retirement. It might even happen that sexual difference remained important in the second hundred years because it affected vulnerability to different illnesses, but otherwise had no effect on social practices. But even such far fetched thought experiments do not make credible a social world without sex (and therefore gender) as a partial basis for social organisation. Sex and gender A strength of the sex-gender distinction is that it recognises the stratification of reality. Sexual difference is real, and partly extra-discursive. Its material effects are relatively (though not entirely) constant between cultures and over time. Our differently sexed bodies enable different reproductive roles for females and for males, different sexual possibilities (famously, multiple orgasms for females), and by the same token limit what we can do. Little boys discover that they have no wombs in which to grow babies, that they will never have breasts with which to feed them. Little girls discover they cannot urinate standing up without getting wet legs. 11 However, we need to distinguish between the causal powers of sexed bodies on the one hand and sexual dimorphism on the other. Sexual dimorphism is a feature of the world facing all human societies. Recently, sociologists have distinguished between that material difference in bodies (sex) and cultural understandings of it and social organisation around it (gender). One broadly realist writer on gender expresses it in this way: ‘Gender is the structure of social relations that centres on the reproductive arena, and the set of practices (governed by this structure) that bring reproductive distinctions between bodies into social processes’ (Connell, 2002:10, emphasis in the original). Or, as he puts it informally, gender is about how human societies deal with human bodies – specifically reproductive difference. Yet, influenced in many ways by poststructuralism, he rejects ontological stratification. Gender refers to the bodily structures and processes of human reproduction. These structures and processes do not constitute a ‘biological base’, a natural mechanism that has social effects. Rather, they constitute an arena, a bodily site where something social happens. Among the things that happen is the creation of the cultural categories ‘women’ and ‘men’ (and any other gender categories that a particular society marks out) (op. cit. 48). I find it hard to understand categorisation as taking place in or at ‘a bodily site’ rather than referring to it. I suspect that, like poststructuralists, Connell is wary of stratified accounts because he (rightly) fears biological determinism. In fact, recognising sexual difference as a ‘natural mechanism’ need not be determinist or reductionist. Accounts which see gender as an expression or reflection of sexual dimorphism are obviously wrong, for the reasons Connell and many others have given. Gender arrangements are eminently culturally variable. But what they retain (and Connell himself insists on this, against the radical contingency asserted by poststructuralists such as Hawkesworth and Butler) is that they refer to sexual difference. ‘What makes a symbolic structure a gender structure, rather than some other kind, is the fact that its signs refer, directly or indirectly, to the reproductive relationship between women and men’ (Connell, 2002:38). It seems that Connell does not want to see the symbolic system as a ‘social effect’ of its referent, presumably because reproductive difference 12 and sexual dimorphism are compatible with many different forms of the gender order. And yet, I shall argue, sexual dimorphism is a causal condition of the gender orders we know. Sexual difference is, indeed, a natural mechanism, in that the different but related causal powers of female and male sexed bodies, when brought into interaction in specific ways have emergent powers – to reproduce. If impregnated through penetrative sex with male humans, female humans may give birth to children and feed them from their bodies (and in technologically advanced societies may have children in other ways too, which are also related to sexual difference and its properties). The meanings and implications of this, as much gender research has shown, depend on social and cultural context (where ‘context’ refers to other co-acting mechanisms). In a highly technological society with safe artificial feeding and readily available contraception, reproductive difference will be differently understood from one in which formula and condoms are not available, or in which they are not affordable or are culturally forbidden. Gender orders involve social representations of sexual difference. They cannot be read off from the real causal powers of differently sexed bodies (powers which are of course affected by gendered practices) (Sayer, 2000). Sexual difference is just one among many mechanisms co-acting to produce the many varied gender orders, and probably best seen as a kind of ‘background mechanism’ or causal condition. In the Humean ‘constant conjunction’ sense of cause, sexual difference does not ‘cause’ the institution of marriage or any particular historical form of it, nor does it ‘cause’ sexual regulations forbidding or permitting homosexuality. But in a critical realist sense, sexual difference is a mechanism contributing to the development of such institutions and such rules. Just because it is a ‘basic’, lower level mechanism, its workings are compatible with many different ways of regulating reproduction and sexuality. To call it a ‘base’ is not, as Connell and others seem to fear, to attribute a greater, overwhelming determining power to sexual difference, or to claim that it cannot be affected by mechanisms at a higher level. It is simply to situate it at a lower level – at the bottom of the enormous culturally elaborated superstructure. In sum, sex is ontologically prior to gender, and is one of the many mechanisms the workings of 13 which shape gender orders. It does not determine gender, nor is gender reducible to sex. Levels of sex and gender To conclude, sexual difference – the differentiation of most humans into males and females, and of a minority into intelligible intersexed forms of human beings is basic in two senses. Firstly, there are real differences between the emergent properties of male and female bodies, their physical capacities and vulnerabilities. For fear of being misunderstood, let me repeat that we are talking about tendencies here, not inevitabilities – the tendency of males to be more muscular, of females to menstruate and lactate in certain circumstances, the greater average vulnerability of the male foetus and so on. Other mechanisms, such as gendered differences in the distribution of nourishment, or amounts of physical exercise, can stop these tendencies resulting in the outcomes mentioned. Secondly, sexual difference is the referent for the many ways of conceptualising sexed bodies, representations which legitimise various gender orders. Sexual difference can be understood in many different ways. Women’s bodies can be seen as similar to men’s bodies (but with inverted genitals), or as completely different (Helliwell, 2000). Rape can be seen as the inevitable result of sexual difference, or as unthinkable (New, 2003). These different representations are social constructions, but …we should take the metaphor of construction seriously: attempts at construction are only successful if they take adequate account of the properties of the materials that they use... these properties are at any particular time relatively independent of the constructors: they are not merely a product of wishful thinking… Social constructions may fail or be only partially successful (Sayer, 2004: 18). The ‘relatively independent’ properties in this case are those of differently sexed bodies, and the relationship between them. Bodies are malleable, but only up to a point. Just because sexual difference is basic, it does not constitute automatic grounds for solidarity between males or between females. At a higher level are the social and 14 cultural structures of the gender order, which are both emergent from the biological features of human populations, and (in their concrete forms) historically determined (Connell 1987). Within gender orders, across time and across the many cultures of the world, women are normatively assigned the bulk of the work of ‘producing people’. What that means for (the various sorts of) women, for men, for children varies tremendously, as do the beliefs which explain, legitimise and govern such norms. Sexual difference is among the causal conditions for every form of the sexual division of labour, whether egalitarian or oppressive, segregated or gender-flexible, so to explain the particular forms it takes we have to look to the many other mechanisms involved – but that does not mean that sexual difference just drops out, as a ‘tedious universal’. Emerging out of both human physiological and social structures, is human psychology. Our selfhood emerges from the possibilities implicit in our central nervous systems, realised in our practical engagement, as embodied beings, with the physical and social world (Archer, 2000). As beings with certain capacities and needs, we respond to what we find in our lives, by conceptualising (as well as by feeling and acting). We draw on the various resources around us to make sense of gender (among other things) when it is offered to us as an explanatory category. Our responses, and others’ responses to our responses, create new situations requiring further responses. In this iterative process we gradually conceptualise and experience ourselves as gendered: gender becomes part of what Archer calls our ‘inner conversation’ (ibid). How this happens in our personal biographies influences our desire, or lack of it, for change in the gender order. Feminists, like all activists, want change. And some changes are possible (while others are not) at each of the levels I have discussed, including the relatively enduring one of biological sexual difference. For activists, two crucial questions are: ‘What is possible?’ And ‘What do we want?’ Whatever changes are wanted, to bring them about requires strategy and tactics, since political activity is itself stratified. To have a chance of formulating effective strategy, we need to know as much as possible about the relevant causal mechanisms at work at all levels, and the relationships between them. Causal claims are, indeed, bids for power: the power to make possible and wanted changes, which comes from an adequate understanding of real causal 15 relationships. Since they are fallible, such claims are inevitably risky, but if we refuse to make them we give up a key dimension of agency. To restrict feminist activism to one level of social reality, or to rely on actualist ideas of causation, is rather like trying to impose the rules of ludo on your game of chess. 16 REFERENCES Archer, M. (2000) Being Human: The Problem of Agency. Cambridge: Polity. Bhaskar, R. (1989) The Possibility of Naturalism. Hemel Hempstead: Harvester Wheatsheaf. Bhaskar, R. (1998) ‘Philosophy and Scientific Realism’. In M. Archer, R. Bhaskar, A. Collier, T. Lawson and A. Norrie (eds) Critical Realism: Essential Readings. London, Routledge. Boyd, E. (1992) ‘Constructivism, Realism and Philosophical Method’. In J. Earman (ed.) Inference, Explanation and Other Frustrations. Berkeley: University of California Press. Braidotti, R. (1997) ‘Comment on Felski’s “the Doxa of Difference”: Working through Sexual Difference’. Signs, 23, 1, 21-40. Butler, J. (1993) Bodies that Matter. London: Routledge. Butler, J. (1995) ‘Contingent Foundations’. In S. Benhabib, J. Butler, D. Cornell and N. Fraser, (eds) Feminist Contentions: A Philosophical Exchange. New York: Routledge. Carter, B. (2000) Realism and Racism: concepts of race in sociological research. London: Routledge. Carter, B. (2003) ‘What race means to realists’. In J. Cruickshank (ed) Critical Realism: The Difference it Makes. London: Routledge. Collier, A. (1994) Critical Realism: An Introduction to Roy Bhaskar’s Philosophy. London: Verso. 17 Connell, R. W. (1987) Gender and Power: Society, the Person and Sexual Politics. Cambridge: Polity. Connell, R. W. (2002) Gender. Cambridge: Polity. Danermark, B. Ekstrom, M. Jakobsen, L. and Karlsson J. Ch. (2002) Explaining Society: Critical realism in the social sciences. London: Routledge. Dyke, C. J. (1988) The Evolutionary Dynamics of Complex Systems. Oxford: Oxford University Press. Felski, R. (1997) ‘The Doxa of Difference’. Signs, 23, 1, 1-21. Feng, J. (1994) The Three-Inch Golden Lotus. Honolulu: University of Hawaii Press. Flax, J. (2002) Re-entering the Labyrinth: Revisiting Dorothy Dinnerstein's The Mermaid and the Minotaur. Signs, 27, 4, 1047-1055. Fraser, N. and Honneth, A. (1998) Redistribution or Recognition? A PoliticalPhilosophical Exchange. London: Verso, 1998. Hawkesworth, M. (1997) ‘Confounding Gender’. Signs, 22, 3, 653-685. Helliwell, C. (2000) ‘”It’s only a penis”: rape, feminism and difference’. Signs, 25, 3, 789-815. Hull, C. L. (forthcoming) The Ontology of Sex. London: Routledge. Keil, F. C. (1989) Concepts, Kinds and Cognitive Development. Cambridge, Mass.: MIT Press. Mohanty, C. (1992) ‘Feminist Encounters: Locating the Politics of Experience’. In M. Barrett and A. Phillips (eds.) Destabilising Theory: Contemporary Feminist Debates. Cambridge: Polity. 18 New, C. (1996) Agency, Health and Social Survival. London: Taylor and Francis. New, C. (1998) ‘Realism, Deconstruction and the Feminist Standpoint’ in Journal for the Theory of Social Behaviour, 28, 4, 349-372. New, C. (2003) ‘Feminism, Deconstruction and Difference’. In J. Cruickshank (ed) Critical Realism: The Difference it Makes. London: Routledge. Preves, S. E. (2001) ‘Sexing the Intersexed: an Analysis of Sociocultural Responses to Intersexuality’. Signs, 27, 2, 523-556 Sayer, A. (1992) Method in Social Science: A Realist Approach. London: Routledge. Sayer, A. (2000) Realism and Social Science. London: Sage. Sayer, A. (2004) ‘Restoring the Moral Dimension: Acknowledging Lay Normativity’. Paper delivered at the British Sociological Association Conference, York, April 2004. Sayers, J. (1982) Biological Politics: Feminist and Anti-Feminist Perspectives. London: Taylor and Francis. Spelman, E. V. (1988) Inessential Woman: Problems of Exclusion in Feminist Thought. Boston, Mass.: Beacon Press. Weedon, C. (1997) Feminist Practice and Poststructuralist Theory. Oxford: Blackwell. 19