Repensando Sexo e Gênero - Cristine Delphy

Rethinking Sex and Gender

Christine Delphy

Up till now, most work on gender, including most feminist work on gender, has been based on an unexamined presupposition: that sex precedes gender. However, although this presupposition is historically explicable, it is theoretically unjustifiable, and its continued existence is holding back our thinking on gender. It is preventing us from rethinking gender in an open and unbiased way. Further, this lack of intellectual clarity is inextricably bound up with, on the one hand, the political contradictions produced by our desire as women to escape domination, and, on the other, our fear that we might lose what seem to be fundamental social categories.

What is common to these intellectual impasses and political contradictions is an inability (or a refusal) to think rigorously about the relationship between division and hierarchy, since the question of the relationship between sex and gender not only parallels this question, but is, in fact, the self-same issue.

What I want to do here is argue that in order to understand reality, and hence eventually to have the power to change it, we must be prepared to abandon our certainties and to accept the (temporary) pain of an increased uncertainty about the world. Having the courage to confront the unknown is a pre-condition for imagination, and the capacity to imagine another world is an essential element in scientific progress. It is certainly indispensable to my analysis.

From Sex Roles to Gender

The notion of gender developed from that of sex roles, and, rightly or wrongly, the person who is credited with being the founding mother of this line of thought is Margaret Mead. Put very briefly, it is her thesis (Mead, 1935) that most societies divide the universe of human characteristics into two, and attribute one half to men and the other to women. For Mead, this division is quite arbitrary, but she does not condemn it unreservedly. She sees it as having many advantages for society, culture and civilisation.

Mead herself does not deal with either the sexual division of labour or differences in the status of men and women. As far as she is concerned, the division of labour is natural, and the few comments she does make about it show that she attributes it to the different reproductive roles of males and females, and to differences in physical strength between the sexes. These are, of course, the ‘classic’ reasons used within both anthropological and ‘commonsense’ (including feminist) thinking. Mead also does not question the hierarchy between the sexes. She either ignores it, or considers it legitimate. Nor does she discuss the prescribed differences between the sexes, except within the very limited domain of ‘temperament’ (under which heading she groups abilities, aptitudes, and emotional personality).

For a long time, Mead’s analysis of prescribed differences was the major theme in the critique of sex roles—a critique that arose from a concern to defend the rights of individuals to express their individualities freely. In the process it was implied that ‘masculine’ and ‘feminine’ traits together constitute and exhaust the whole of human possibilities (see below).

Although the term is frequently accredited to her, Mead herself rarely uses the term ‘sex roles’ because she was not in fact concerned about these roles, still less with critiquing them. Her concern was rather the analysis and critique of feminine and masculine ‘temperaments’. In fact, the idea of sex roles was critically developed from the 1940s to the 1960s, that is, in the decades commonly considered to be a period when feminism was ‘latent’—through the work of Mirra Komarovsky (1950), Viola Klein and Alva Myrdal (Myrdal and Klein, 1956), and Andree Michel (1959, 1960). All these authors worked within a Parsonian sociological perspective, and saw a role as the active aspect of a status. Broadly speaking, ‘status’ was the equivalent of the level of prestige within society, and each status had roles which the individuals who held that status had to fulfil. This perspective is clearly sociological in the true sense of the word: people’s situations and activities are held to derive from the social structure, rather than from either nature or their particular capacities.

Thus, when these authors spoke of the ‘roles’ of women and men, they were already taking a large step towards denaturalising the respective occupations and situations of the sexes. Their approach was not actually opposed to Mead’s anthropological approach, but rather developed it in two ways:

1 They confirmed the arbitrary aspect of the division of qualities between the sexes, this time by an epistemological diktat: that is, by their postulate that everyone plays roles.

2 More importantly, they considered a social ‘role’ to be not simply the ‘psychological’ characteristics Mead had spoken about, but also (and principally) the work associated with a rung on the social ladder (a status), and hence a position in the division of labour.

The division of labour and the hierarchy between men and women therefore began to be accorded a cultural character, whereas Mead had considered them to be natural; and since they were cultural rather than natural, the authors stressed they were arbitrary. In addition, since the concept of sex roles also emerged within the framework of a feminist critique (even when the term feminist was not explicitly used), these authors all stressed that as the position of women was socially determined, it was changeable. Even though the concepts they used were Parsonian in origin, they questioned Parsons’s theory and its premise of harmony between the sexes; and Andree Michel, in particular, strongly criticised the containment of women within traditional roles, and also Parsons’s idea that this was good for women and for society.

The term ‘sex roles’ then remained in use for a long time, until the concept of gender, which derived directly from it, appeared in the early 1970s. If we take one of the first works directly on ‘gender’, Ann Oakley’s Sex, Gender and Society, published in 1972, we find the following definition:

‘Sex’ is a word that refers to the biological differences between male and female: the visible difference in genitalia, the related difference in procreative function. ‘Gender’, however, is a matter of culture: it refers to the social classification into ‘masculine’ and ‘feminine’. (Oakley, 1985, p.16)

Oakley’s book is devoted partly to a critical account of recent research on the differential psychology of the sexes: to innate and acquired elements of aptitude (‘talents’ in Mead’s terminology) and attitude (‘temperamental’) differences between women and men, and partly to an account of what anthropological research can teach us about the division of labour between the sexes. According to Oakley, psychological differences between the sexes are due to social conditioning, and there is no research that allows us to infer any biological determinism whatsoever. She also says that while a division of labour by sex is universal, the content of the tasks considered to be feminine or masculine varies considerably according to the society.

Oakley’s use of the concept of gender thus covers all the established differences between men and women, whether they are individual differences (studied by psychologists), or social roles, or cultural representations (studied by sociologists and anthropologists). In addition, in her work the concept of gender covers everything that is variable and socially determined—variability being the proof that it is social in origin. She says: The constancy of sex must be admitted, but so too must the variability of gender’ (op. cit., 1985, p. 16).

However, the facets that are missing from Oakley’s definition, although they were already present in the work on sex roles, and which have become central to feminist positions and been developed subsequently, are the fundamental asymmetry (Hurtig and Pichevin, 1986) and hierarchy (Delphy, 1980; Varikas, 1987) between the two groups, or roles, or sexes, or genders.

Sex and Gender

With the arrival of the concept of gender, three things became possible (which does not mean they have happened):

1 All the differences between the sexes which appeared to be social and arbitrary, whether they actually varied from one society to another or were merely held to be susceptible to change, were gathered together in one concept.

2 The use of the singular (‘gender’ as opposed to ‘genders’) allowed the accent to be moved from the two divided parts to the principle of partition itself.

3 The idea of hierarchy was firmly anchored in the concept. This should, at least in theory, have allowed the relationship between the divided parts to be considered from another angle.

As studies have accumulated showing the arbitrariness of sex roles and the lack of foundation for stereotypes in one area after another, the idea that gender is independent of sex has progressed. Or rather, since it is a question of the content, the idea that both genders are independent of both sexes has progressed, and the aspects of ‘sex roles’ and sexual situations that are recognised to be socially constructed rather than biologically determined, have grown. Everyone working in the field has certainly not drawn the dividing line between what is social and cultural and what is natural in the same place—but then it would have been astonishing if they had. It is right that the question should remain open.

What is problematic, however, is that the on-going discussions around this question have presumed epistemological and methodological paradigms that should actually have been questioned. We have continued to think of gender in terms of sex: to see it as a social dichotomy determined by a natural dichotomy. We now see gender as the content with sex as the container. The content may vary, and some consider it must vary, but the container is considered to be invariable because it is part of nature, and nature ‘does not change’. Moreover, part of the nature of sex itself is seen to be its tendency to have a social content/to vary culturally.

What should have happened, however, is that the recognition of the independence of the genders from the sexes should have led us to question whether gender is, in fact, independent of sex. But this question has not been asked. For most authors, the issue of the relationship between sex and gender is simply ‘what sort of social classification does sex give rise to? Is it strong or weak, equal or unequal? ’ What they never ask is why sex should give rise to any sort of social classification. Even the neutral question ‘we have here two variables, two distributions, which coincide totally. How can we explain this co-variance? ’ does not get considered. The response is always: sex comes first, chronologically and hence logically—although it is never explained why this should be so.

Actually, whether or not the precedence gets explained does not make much difference. The very fact of suggesting or admitting the precedence of sex, even implicitly, leads to one being located, objectively, in a theory where sex causes or explains gender. And the theory that sex causes gender, even if it does not determine the exact forms gender divisions take, can derive from only two logical lines of argument.

In the first line of argument, biological sex, and particularly the different functions in procreation of males and females that it provokes, necessarily gives rise to a minimal division of labour. I would include in this line of argument, with its naturalist premises, most contemporary anthropological accounts, feminist as well as patriarchal, from George Murdock (1949) to Martha Moia (1981) by way of Gayle Rubin (1975) with just a few notable exceptions, such as Mathieu (1991) and Tabet (1982)—in this volume. It fails to explain satisfactorily: first, the nature and the natural reason for this first division of labour; and second, the reasons it is extended into all fields of activity; that is, why it is not limited to the domain of procreation. It therefore fails to explain gender other than by suppositions that reintroduce upstream one or more of the elements it is supposed to explain downstream.

The second line of argument sees biological sex as a physical trait which is not only suitable, but destined by its intrinsic ‘salience’ (in psycho-cognitive terms) to be a receptacle for classifications. Here it is postulated that human beings have a universal need to establish classifications independently of, and prior to, any social practice. 2 But these two human needs are neither justified nor proven. They are simply asserted. We are not shown why sex is more prominent than other physical traits that are equally distinguishable, but which do not give birth to classifications that are (1) dichotomous and (2) imply social roles which are not just distinct but hierarchical.

I call this latter line of argument ‘cognitivist’, not because it is particularly held by the ‘Cognitivists’, but because it presumes certain ‘prerequisites’ of human cognition. The best known academic version of such theories is that of Levi-Strauss, who, while not a psychologist, bases all his analyses of kinship, and (by extension) human societies, on an irrepressible and pre-social (hence psychological) need of human beings to divide everything in two (and then into multiples of two). Levi-Strauss (1969) was very much influenced by linguistics, in particular by Saussure’s phonology (Saussure, 1959), and he devised by analogous construction what the social sciences call ‘structuralism’.

A rather more recent version of this thesis has been presented by Derrida (1976) and his followers, who say that things can only be distinguished by opposition to other things. However, while Saussure is concerned purely with linguistic structures, Derrida and his clones want to draw philosophical conclusions about the importance of ‘differance’. These conclusions themselves incorporate presuppositions about the conditions for the possibility of human knowledge, hence about the human spirit, which are very similar to those of Levi-Strauss. Saussure’s theory had no such ambitions, and its validity in its own field of reference—linguistics—should not be taken as a guarantee of its applicability elsewhere. We may agree things are only known by distinction and hence by differentiation, but these differentiations can be, and often are, multiple. Alongside cabbages and carrots, which are not ‘opposites’ of each other, there are courgettes, melons, and potatoes. Moreover, distinctions are not necessarily hierarchical: vegetables are not placed on a scale of value. Indeed, they are often used as a warning against any attempt at hierarchisation: we are told not to compare (or to try to add) cabbages and carrots. They are incommensurable. They do not have a common measure. Therefore, they cannot be evaluated in terms of being more or less, or better or worse than one another.

Those who adhere to Derrida’s thesis thus fail to distinguish between the differences on which language is based, and differences in social structures. The characteristics of cognition, in so far as they can be reduced to the characteristics of language, cannot account for social hierarchy. This is external to them. They therefore cannot account for gender—or they can do so only at the expense of dropping hierarchy as a constitutive element of gender.

Hence, neither of the two lines of argument that might justify a causal link from sex to gender is satisfactory. The presupposition that there is such a causal link thus remains just that: a presupposition.

But if we are to think about gender, or to think about anything at all, we must leave the domain of presuppositions. To think about gender we must rethink the question of its relationship to sex, and to think about this we must first actually ask the question. We must abandon the notion that we already know the answer. We must not only admit, but also explore, two other hypotheses: first, that the statistical coincidence between sex and gender is just that, a coincidence. The correlation is due to chance. This hypothesis is, however, untenable, because the distribution is such that the co-incidence between so-called biological sex and gender is ‘statistically significant’. It is stronger than any correlation could be which is due to chance.

Second, that gender precedes sex: that sex itself simply marks a social division; that it serves to allow social recognition and identification of those who are dominants and those who are dominated. That is, that sex is a sign, but that since it does not distinguish just any old thing from anything else, and does not distinguish equivalent things but rather important and unequal things, it has historically acquired a symbolic value.

The symbolic value of sex has certainly not escaped the theoreticians of psychoanalysis. But what has entirely escaped them is that this should be one of the final conclusions of a long progression: the point of arrival and not of departure. Unfortunately, this blind spot is one that many feminists share with psychoanalysts.

As society locates the sign that marks out the dominants from the dominated within the zone of physical traits, two further remarks need to be made.

First, the marker is not found in a pure state, all ready for use. As Hurtig and Pichevin (1986) have shown, biologists see sex as made up of several indicators which are more or less correlated one with another, and the majority are continuous variables (occurring in varying degrees). So in order for sex to be used as a dichotomous classification, the indicators have to be reduced to just one. And, as Hurtig and Pichevin (1985) also say, this reduction ‘is a social act’.

Second, the presence or absence of a penis3 is a strong predictor of gender (by definition one might say). However, having or not having a penis correlates only weakly with procreative functional differences between individuals. It does not distinguish tidily between people who can bear children and those who cannot. It distinguishes, in fact, just some of those who cannot. Lots of those who do not have penises also cannot bear children, either because of constitutional sterility or due to age.

It is worth pausing here, because the ‘cognitivists’ think sex is a ‘prominent trait’ because they think physical sex is strongly correlated with functional differences, and because they assume that the rest of humanity shares this ‘knowledge’. But they only think biological sex is a ‘spontaneous perception’ of humanity because they themselves are convinced that it is a natural trait that no one could ignore. To them it is self-evident that there are two, and only two, sexes, and that this dichotomy exactly cross-checks with the division between potential bearers and non-bearers of children.

To try to question these ‘facts’ is indeed to crack one of the toughest nuts in our perception of the world. We must, therefore, add to the hypothesis that gender precedes sex, the following question: when we connect gender and sex, are we comparing something social with something which is also social (in this case, the way a given society represents ‘biology’ to itself)?

One would think that this would logically have been one of the first questions to be asked, and it is doubtless the reason why some feminists in France (for example, Guillaumin, 1982, 1985; Mathieu, 1980; and Wittig, 1992) have been opposed to using the term ‘gender’. They believe it reinforces the idea that ‘sex’ itself is purely natural. However, not using the concept of gender does not mean one thereby directly questions the natural character of sex. So economising on the concept of gender does not seem to me the best way to progress.

‘Sex’ denotes and connotes something natural. It was, therefore, not possible to question ‘sex’ head on, all at once, since to do so involves a contradiction in terms. (‘Naturalness’ is an integral part of the definition of the term. ) We had first to demonstrate that ‘sex’ is applied to divisions and distinctions which are social. Then we had not only to separate the social from the original term, which remains defined by naturalness, but to make the social emerge.This is what the notions of first ‘sex roles’ and then ‘gender’ did. Only when the ‘social part’ is clearly established as social, when it has a name of its own (whether it be ‘sex roles’ or ‘gender’), then and only then could we come back to the idea we started with. We had first to design and lay claim to a territory for the social, having a different conceptual location from that of sex, but tied to the traditional sense of the word ‘sex’, in order to be able, from this strategic location, to challenge the traditional meaning of ‘sex’.

To end this section, I would say that we can only make advances in our knowledge if we initially increase the unknown: if we extend the areas that are cloudy and indeterminate. To advance, we must first renounce some truths. These ‘truths’ make us feel comfortable, as do all certainties, but they stop us asking questions—and asking questions is the surest, if not the only way of getting answers.

Division, Differences and Classifications

The debate on gender and its relationship to sex covers much the same ground as the debate on the priority of the two elements—division and hierarchy—which constitute gender. These are empirically indissolubly united, but they need to be distinguished analytically. If it is accepted that there is a line of demarcation between ‘natural’ and socially constructed differences, and that at least some differences are socially constructed, then there is a framework for conceptualising gender. This means, or should mean, recognising that hierarchy forms the foundation for differences—for all differences, not just gender.

However, even when this is accepted as an explanation, it is not accepted as a politics, nor as a vision of the future, by feminists. It is not their Utopia. All feminists reject the sex/gender hierarchy, but very few are ready to admit that the logical consequence of this rejection is a refusal of sex roles, and the disappearance of gender. Feminists seem to want to abolish hierarchy and even sex roles, but not difference itself. They want to abolish the contents but not the container. They all want to keep some elements of gender. Some want to keep more, others less, but at the very least they want to maintain the classification. Very few indeed are happy to contemplate there being simply anatomical sexual differences which are not given any social significance or symbolic value. Suddenly, the categories they use for analysis, which elsewhere clearly distinguish those who think difference comes first and hierarchy afterwards from those who think the contents of the divided groups are the product of the hierarchical division, become muzzy, and the divergence between the two schools fades away.

This is especially clear in the debate on values. Feminist (and many other!) theorists generally accept that values are socially constructed and historically acquired, but they seem to think they must nonetheless be preserved. There are two typical variants on this position. One says we must distribute masculine and feminine values throughout the whole of humanity; the other says that masculine and feminine values must each be maintained in their original group. The latter view is currently especially common among women who do not want to share feminine values with men. I am not sure whether this is because they believe men are unworthy or incapable of sustaining these values, or because they know men do not want them anyway. But we might well ask how women who are ‘nurturing’ and proud of it are going to become the equals of unchanged men—who are going to continue to drain these women’s time? This is not a minor contradiction. It shows, rather, that if intellectual confusion produces political confusion, it is also possible to wonder, in a mood of despair, if there is not, behind the intellectual haze, a deep and unacknowledged desire notto change.

In any case, both variants of the debate show an implicit interpretation of the present situation, which contradicts the problematic of gender. On the one hand, there is a desire to retain a system of classification, even though (it is said) it has outlived its function of establishing a hierarchy between individuals—which would seem to indicate that people do not really think that gender is a social classification. On the other hand, there is a vision of values which is very similar to Margaret Mead’s, which can be summarised as: all human potentialities are already actually represented, but they are divided up between men and women. ‘Masculine’ plus ‘feminine’ subcultures, in fact culture itself, is not the product of a hierarchical society. It is independent of the social structure. The latter is simply superimposed upon it.

Hierarchy as Necessarily Prior to Division

This last view is contrary to everything we know about the relationship between social structure and culture. In the marxist tradition, and more generally in contemporary sociology whether marxist or not, it is held that the social structure is primary. This implies, as far as values are concerned, that they are, and cannot but be, appropriate to the structure of the society in question. Our society is hierarchical, and consequently its values are also hierarchically arranged. But this is not the only consequence, since Mead’s model also allows for this.

Rather, if we accept that values are appropriate to social structures, then we must accept that values are hierarchical in general, and that those of the dominated are no less hierarchical than those of the dominants. According to this hypothesis, we must also accept that masculinity and femininity are not just, or rather not at all, what they were in Mead’s model—a division of the traits which are (1) present in a potential form in both sexes, or (2) present in all forms of possible and imaginable societies. According to the ‘appropriateness’ paradigm (i. e. the social construction of values), masculinity and femininity are the cultural creations of a society based on a gender hierarchy (as well as, of course, on other hierarchies). This means not only that they are linked to one another in a relationship of complementarity and opposition, but also that this structure determines the content of each of these categories and not just their relationship. It may be that together they cover the totality of human traits which exist today, but we cannot presume that even together they cover the whole spectrum of human potentialities. If we follow the ‘appropriateness’ paradigm, changing the respective statuses of the groups would lead to neither an alignment of all individuals on a single model, nor a happy hybrid of the two models.

Both the other sorts of conjecture presuppose, however, that these ‘models’ (i. e. the ‘feminine’ and the ‘masculine’) exist sui generis, and both imply a projection into a changed future of traits and values that exist now, prior to the change in the social structure.

To entrust oneself to this sort of guesswork, which moreover is totally implicit, requires a quite untenable, static view of culture. Even if it was progressive when Margararet Mead was writing just to admit that cultures varied and that values were arbitrarily divided between groups, this view is no longer tenable. It assumes the invariability of a universal human subject, and this has been invalidated by historians’ studies of ‘mentalities’, and by the social constructionist approaches inspired (even if generally unwittingly) by the marxist principles discussed above.

This vision of culture as static is, however, fundamental to all the variants of the notion of positive complementarity between men and women (even if those who hold such views do not recognise it). 4 They all presuppose that values precede their hierarchical organisation (as in Mead’s model), and this stasis can only lead us back to ‘nature’: in this case, to human nature.

Such a point of view, and only such a point of view, can explain why Mead was afraid that everyone would become the same, which was counter to nature. The fear that a generalised sameness, or absence of differentiation, would be provoked by the disappearance of what is apparently the only kind of difference that we know (for this viewpoint ignores all other sorts of variance)5 is, of course, not new; though currently the fear that the world will align on a single model often takes the more specific form that the single model will be the current masculine model. This (it is said) will be the price we shall have to pay for equality; and (it is said) it is (perhaps) too high a price. However, this fear is groundless, since it is based on a static, hence essentialist, vision of women and men, which is a corollary to the belief that hierarchy was in some way added on to an essential dichotomy.

Within a gender framework such fears are simply incomprehensible. If women were the equals of men, men would no longer equal themselves. Why then should women resemble what men would have ceased to be? If we define men within a gender framework, they are first and foremost dominants with characteristics that enable them to remain dominants. To be like them would also be to be dominants; but this is a contradiction in terms. If, in a collective couple constituted of dominants and dominated, either of the categories is suppressed, then the domination is ipso facto suppressed. Hence, the other category of the couple is also suppressed. Or to put it another way, to be dominant one must have someone to dominate. One can no more conceive of a society where everyone is ‘dominant’ than of one where everyone is ‘richest’.

It is also not possible to imagine the values of a future egalitarian society as being the sum, or a combination, of existing masculine and feminine values, for these values were created in, and by, hierarchy. So how could they survive the end of hierarchy?

This vision of a society where values existed as ‘entities’, prior to their being organised into a hierarchy, is, as I have said, static and ultimately naturalist. But it is also not an isolated idea. It is part of a whole ensemble of ideas which includes: first, commonsense and academic theories of sexuality that involve a double confusion: a confusion of anatomical sex with sexuality, and of sexuality with procreation; and second, a deep cultural theme to which these theories themselves refer back: namely that each individual is essentially incomplete in so far as he or she is sexed. Emotional resistance and intellectual obstacles to thinking about gender both originate from this: from the individual and collective consciousness.

This is what I previously called ‘a set of confused representations turning around a belief in the necessity of close and permanent relations between most males and most females’ (Delphy, 1980). I wanted to call this set (of representations) ‘heterosexuality’, but it has been suggested that it would be better called ‘complementarity’. Its emblem is the image of heterosexual intercourse, and this gives it a social meaning and an emotional charge which is explicable only by its symbolic value. It could, therefore, equally be called a set of representations of ‘fitting together’.

It would be interesting to develop this reflection further in relation to two main sets of questions: first, how this whole set of ideas forms a view of the world as a whole which is more than the sum of its parts—which possesses a mystical and non-rational character (a cosmogony); and second, how this cosmogony informs and determines the explicit and implicit premises of much scientific research— including feminist research and lesbian research.

Imagination and Knowledge

We do not know what the values, individual personality traits, and culture of a non-hierarchical society would be like, and we have great difficulty in imagining it. But to imagine it we must think that it is possible. And it is possible. Practices produce values: other practices produce other values.

Perhaps it is our difficulty in getting beyond the present, tied to our fear of the unknown, which curbs us in our Utopian flights, as also in our progress at the level of knowledge—since the two are necessary to one another. To construct another future we obviously need an analysis of the present, but what is less recognised is that having a Utopian vision is one of the indispensable staging-posts in the scientific process—in all scientific work. We can only analyse whatdoes exist by imagining what does not exist, because to understand what is, we must ask how it came about. And asking how it came to exist must involve two operations. The first I described earlier when I said that we must admit we do not know the answers when we think we do (Descartes’s famous ‘suspension of judgment’). The second operation is admitting, even if it is contrary to the evidence of our senses, that something which exists, need not exist.

In conclusion, I would say that perhaps we shall only really be able to think about gender on the day when we can imagine non-gender. But if Newton could do it for falling apples, we should be able to do it for ourselves as women.

Notes

1 An earlier version of this article, Tenser le genre: Quels problemes? ’, appeared in Marie-Claude Hurtig et al. (Eds) Sexe et genre: de la hierarchie entre les sexes, 1991, Paris, Editions du Centre National de la Recherche Scientifique.

The present version was translated (by Diana Leonard) and appeared first in 1993 in Women’s Studies International Forum, 16(1), pp. 1-9.

2 See, for example, Archer and Lloyd (1985), who say gender will continue because it is a ‘practical way of classifying people’.

3 This is ‘the final arbiter’ of the dichotomous sex classification for the state, according to Money and Ehrhardt (1972, quoted by Hurtig and Pichevin, 1985).

4 There is, however, no single meaning to complementarity. The paradigm of hierarchy as the basis of division also implies complementarity, although in a negative sense.

5 This would mean that I would only talk to a male baker since I would no longer be able to distinguish a female baker from myself.