Download This Chapter As A PDF
Please ensure you have
the latest Adobe Reader
Elements of Explanation in the Social Sciences
I will begin this chapter by going to two points in Flyvbjerg’s consideration of the difference between the social and natural sciences (points 1 and 4). These are related points about how the natural and social sciences differ, which can be a useful way to start to consider explanations and theories in the social sciences – and to look at the elements of accounts in the social sciences.
1. There is no reason to think that we could end up in social sciences with the kind of knowledge that is attainable in the natural sciences. This theoretical knowledge of the natural sciences is abstract and context independent. It allows us to formulate generalized laws, which can be applied to create accurate predictions.
Flyvbjerg notes the aspiration of social scientists to become “real scientists” like those who study natural objects. Yet the experience of the history of social science does not give us grounds for optimism. After “more than 200 years of attempts, one could reasonably expect that there would exist at least a sign that social science has moved in the desired direction, that is, towards predictive theory” (Flyvbjerg 2001: 32).
The problem is even more marked. There is no accumulation of agreed knowledge, as in the natural sciences. Instead:
… social sciences go through periods where various constellations of power and waves of intellectual fashion dominate, and where a change from one period to another, which on the surface may resemble a paradigm shift, actually consists of the researchers within a given area abandoning a ‘dying’ wave for a growing one, without there having occurred any collective accumulation of knowledge. (Flyvbjerg 2001: 30)
The basic reason for this situation is that it is impossible in the social sciences to develop the conditions for a theory of the kind that is required to predict. As Flyvbjerg lays it out (Flyvbjerg 2001: 38-39), such a theory must be:
· Explicit – can be understood by all
· Universal – apply in all times and places
· Abstract – can be described without reference to concrete examples
· Discrete – elements of the theory must be context independent
· Systematic – context independent elements are related by laws
Flyvbjerg uses arguments from Dreyfus and Bourdieu to show the impossibility of such a project for the social sciences. He summarizes this in point 4, which I will put as follows:
4. A theory in the social sciences must include phenomena, which are defined by their context. So there can be no context free theories in the social sciences, and hence no predictive ability for social theory.
According to Flyvbjerg, Dreyfus argues as follows. As soon as we attempt to de-contextualize everyday objects to develop an abstract and universal theory, we leave out the context in which humans actually “pick out … objects and events”. The theory can make good predictions only so long as there is coincidence between the elements the theory nominates and those that humans actually pick out. To get the conclusion that Dreyfus and Flyvbjerg want, the following must be added. In fact humans may pick out a whole different set of objects and events. It is impossible to predict which objects and events they will pick out as representing the elements nominated in the theory.
Reply to Flyvbjergs’ argument – context free generalizations in the social sciences
Let us consider this argument using a concrete example. We may have a theory about human uses of food. Yet a purely biological definition of food in terms of nutritional uses for the human body is very inadequate to deal with different ideas about food in different cultures. Although foods are mostly things that are nourishing, people may decide to eat something with no nutritional value as a sign of status or to add things to food for some other purpose – for the taste, as a mind altering drug, to promote virility. Further, among all the things that are available as food in a given situation only some will be culturally nominated as food. There is almost no way of predicting this in detail and certainly not comprehensively. Why not? Because, as far as we can tell, these cultural choices are arbitrary and creative. They are created by cultures as a whole – and also on a daily basis by people as they make choices. So because of this “food” as a theoretical item in the social sciences is defined by its context, as Flyvbjerg is arguing.
In this sense, Flyvbjerg is absolutely right that what people nominate as “foods” is very context dependent. However, what is in fact “context free” about food is that it is what people want and expect to eat. So what is common about food is not a set of events that can be specified – without talking about the intentions of human subjects and comparing those intentions across cultures. The common context free definition of “food” is a definition that refers not to observable events and a set of biological substances at all but at least in part to common inner states shared by the people who eat food. If we are attempting to arrive at a “context free” definition of food that does not refer to its common intentional basis, it is hard to get past the biological law that humans mostly eat things (an external observable act that we can compare in different cultures) that satisfy the nutritional needs of the body (as defined by a medical science which looks at chemical compositions and disease symptoms). This may well be true but it seems a little obvious and does not strike us as a great revelation of social science theory. It is really biology rather than social science.
So, what do social sciences generally do in this situation? Mostly, what they do is help the reader to get a sense of the cultural context in which certain choices are more likely. It is more likely for South Asian food to be curried. More detailed ethnography can tell us about the kinds of foods people are eating and the meanings they invest in these practices. So, in a way Flyvbjerg is absolutely right in saying that social sciences are not that interested in context free generalizations. Context free generalizations about food are mind numbingly obvious. It is the context specific realization of the concept of food that is interesting. On the other hand, what this example also reveals is that our concept of food; the one that we happily apply in different contexts is in fact context free and relates to common intentional states shared across time and space in different human cultures – there is no culture in which the concept of food (as what people want and expect to eat) has no meaning.
Does Bourdieu demonstrate the impossibility of cultural universals?
Flyvbjerg’s second argument is from Bourdieu. Bourdieu reviews Levi-Strauss’s analysis of “the gift”. What is interesting about this is that Levi-Strauss is attempting to create context free generalizations about the gift in all human cultures. He sees the gift as “reversible”. This means that people are expected to return an equivalent gift to the one they have received. This is presented by Levi-Strauss as a law of culture.
But Bourdieu argues that this cannot be the whole truth “of a practice which could not exist if it were consciously perceived in accordance with the model” (Bourdieu in Flyvbjerg 2001: 41). The “full truth of the gift” is two opposing structures. In one there is the reality of the reversibility of the gift. However, the temporal delay and the difference between a gift and its return, create the experience that the gift is not reversible. For example, if someone took your gift and gave it back to you pronto, you would experience your gift as having been refused. As Bourdieu argues, delay and difference are necessary for it to be experienced as a gift. In other words, the experience of irreversibility is necessary as well as the reality of reversibility. Moreover, to actually time this delay so it is regarded as apt (or not, if you want to insult someone) is a social skill and there are no rules for this behaviour. It is like a good stroke in tennis.
Accordingly, as Flyvbjerg notes, what seems from the outside, if we abstract from the context, as the same act (returning a gift), is actually different depending on the circumstances that have been developed socially. So a gift in one context may be an insult in another. So, he claims, there can be no context free generalizations about “gifts” in human cultures.
This is a peculiar argument. Flyvbjerg wants to use it to show that there can be no context free generalizations in the social sciences. Levi-Strauss presents a context free generalization – a gift is what has to be reciprocated. And Bourdieu knocks it down. Bourdieu shows that reciprocity as such is not sufficient to establish an act as a gift – context in the form of timing determines the aptness of reciprocity.
Yet to me what is odd is that Bourdieu actually uses evidence (in the form of ethnographies) to show that Levi-Strauss’ generalization is incorrect. This seems to fly in the face of Flyvbjerg’s claim that rival theories in social sciences cannot be debated by marshalling evidence. And then Bourdieu uses this evidence to suggest an alternative and more adequate generalization. As follows. I will write out what I think is the context free generalization implied in Bourdieu’s critique of Levi-Strauss. At least in the way that Flyvbjerg tells us the story of their debate!
Gifts have to be reciprocated but this reciprocity has to be masked by an appearance of disconnection between the first gift and its return – a seeming irreversibility. This masking effect is created by timing and in most cases by a difference between the item which is the first gift and the item which is its return.
This is actually another context free generalization about human cultures. Far from supporting Flyvbjerg’s contention, these two doyens of cultural theory seem to be debating which context free generalization about human culture best fits the ethnographic evidence.
Flyvbjerg could truly claim that the realization of this generalization in concrete instances cannot be specified in advance in a context free manner. In other words, what constitutes an apt time for return and an apt difference between gift and return are deeply and minutely specified by the cultural context. You cannot say – it is always 2 hours and 50 minutes or that in all cultures the gift of a yellow Toyota Hilux should be repaid with a pink Nissan Pajero. As well, Flyvbjerg is correct in saying that the ability to operate this system of gifts in a culturally appropriate manner is a practical skill, developed over years of experience; a skill that cannot be expressed in a set of rules that are invariable in a specific cultural context. A skill that is mostly exercised unconsciously and cannot be accurately or comprehensively rendered conscious by the practitioner.
But what this comes back to is what I have referred to before in discussing the context free social concept of food. What is general or context free is really intentions and drives which are seen to be common in different societies. In other words, Bourdieu is saying that in different societies there is always a way of realizing the return of a gift by making it “too soon” and consequently causing the return to be perceived as a “refusal”. It is the ideas of “too soon” and “refusal” that are claimed to be context free while their realisation in specific societies and specific contexts in those societies cannot be specified in advance.
For me, the attempts of Levi-Strauss and Bourdieu to develop universals of culture imply a theory of human nature. It is supposed that given the context of our life on earth so far, human nature has produced some common cultural practices, which the ethnographer can distil by examining evidence from different cultures. At the very least, the formulation of this discussion depends upon the translatability between cultures of a set of terms and concepts which in itself implies a common human nature – for example “gift”; “the same gift or a different gift”; “an insult”. Their debate is an attempt to discover some cultural universals and both parties are equally implicated in this.
The universals that the two cultural theorists are considering are universals specified in terms of people’s intentions and understandings of the situation. In other words, what we call gifts are not just a set of objects, which are observably the same in all cultural settings. They are a set of objects, which have the same relationship to understandings and intentions that are the same in all these cultural contexts. For example a gift has to be something that the recipient is likely to want. It is these understandings and intentions, which provide the elements of any context free abstracted generalization about human culture. It is our knowledge of the similarity of intentional states across varied cultural contexts, which enables us to recognize a somewhat random and arbitrary set of objects as “gifts” and to talk about the conditions for reciprocity and irreversibility across a variety of cultures.
We could re-state Bourdieu’s theory in a way that makes some of his underlying claims about human nature more obvious.
People like and expect gifts to be reciprocated. However they prefer to believe that the return gift is an act of voluntary generosity. To achieve this effect, it is necessary to delay the return but not for so long that the original donor feels that their original gift has not been appreciated. It can also be a good idea to return a gift that is seen as equivalent in value but different from the original gift. Both these tactics create the appearance that the return gift is an act of voluntary generosity, while both parties are also in fact aware of the reciprocity of the transactions.
Is this a useful generalization? Perhaps. Certainly, Levi-Strauss and Bourdieu are going about their attempts to discover cultural universals in a way that makes sense – by comparing ethnographies for a variety of cultures to find common elements.
Why context free generalizations are not that common in the social sciences
The failure of social science to come up with – or in fact to be that concerned with – a set of context free generalizations about human social life can be stated in this way.
The only context free generalizations about human culture are statements that describe universal human responses to life on earth. Such generalizations are important in the social sciences but most of them are somewhat obvious. Instead, we mostly make use of ideas about human nature when we are explaining specific social situations. These ideas appear in explanations of specific cultural behaviour in ways that are usually below the radar of theoretical recognition. We share assumptions about the impact of human nature on people’s behaviour and these assumptions help us to make sense of the explanations of cultural practices that we are faced with when we read social science.
Where ideas about human nature do get debated openly is in the context of different explanations of fairly common cultural outcomes. For example, what are the origins of patriarchy – or social class – or religion? In these cases, different theories of human nature can become quite explicit and take an obvious role in different explanations of cultural outcomes. But mostly, what we are more interested in is the social meanings and likely outcomes of cultural practices that are quite specific. In making these explanations we constantly refer to aspects of human nature that we are taking for granted.
Predictions in the social sciences
Generalizations about human nature can in fact come up with very definite and reliable predictions. Societies will create means to secure supplies of food. But generalizations such as this are so obvious that they are not a great deal of use in the social sciences.
More specific predictions are impossible precisely because the relationship between aspects of human nature and social acts is very context specific. A particular cuisine that is socially significant in a particular cultural setting is an invention of culture that certainly is inspired, or even driven, by aspects of human nature but it is far from totally determined by these natural drives and capacities. A historical and specific social account is necessary to discuss cuisines and their social implications. Social sciences cannot generally predict accurately because humans create their social responses through a process of invention that we cannot predict. In some science fiction world we could perhaps predict how the nerve cells of 6 billions brains would respond to the myriad of inputs that are coming in from the universe at every moment. But this is hardly a realistic goal for social science!
What do we do when we do predict? Marx’s prediction of working class revolution is of course the most notorious and its failure has led to somewhat extreme claims that the social sciences cannot actually predict anything at all. What we are working with in cases like this are actually generalizations about the way specific cultural orders are operating at specific historical periods. So Marx describes the mechanisms of capitalism. Given that, he aims to show that the machine of capitalism is likely to undo itself in a specific manner. Right or wrong, there is nothing problematic about this analysis as social science meta-theory. For me, writing about this now, he is examining discursively created social constraints on actions that can be shown to be operating in a specific situation. He uses this analysis to make a prediction about what is likely to happen so long as these constraints continue to operate.
This kind of analysis, and the predictions that come with it, are a necessary and indeed inevitable aspect of any social investigation. It cannot be purged from the social sciences by a new meta-theory as Flyvbjerg imagines. What is a Balinese tooth filing ceremony? It is a set of activities. Examples of tooth filing ceremonies are similar to each other because they are constrained by the same discourse. Analysis like this carries the possibility of prediction with it. Given that these constraints are operating, what is the most likely outcome? For example, what is likely to happen next in the tooth filing ceremony given what has happened in it so far? When prediction fails, we can go back and revise the model, or we can start talking about unexpected cultural responses that caused a different outcome. None of this suggests that the original interpretation of the social constraints and the predictions associated with this were a silly exercise.
As indicated, all such predictions and interpretations of cultural events are predicated on common sense ideas about human nature as well as on investigations of how things are going in a specific situation. However, the description of social structures, social constraints and discourses, is necessarily a summary, an incomplete account of events. At every point in the real social world, individual social actors are creating events that cannot be predicted. It is this which makes prediction in the social sciences a completely different kind of animal from predictions in the natural sciences – where water boils at 100 degrees and so on.
Is there cumulative knowledge in the social sciences?
With this discussion as a background, what can we make of Flybvbjerg’s assertion that there is no “collective accumulation of knowledge” in the social sciences. Instead different paradigms come and go – or indeed co-exist. These changes and people’s preference for new paradigms reveal at the most “fashions” in social theory rather than any accumulation of knowledge.
To begin, with, going back to the distinction between “everyday facts” and “interpretations” that I considered in the previous chapter, it makes sense to say that there is actually an accumulation of social knowledge as far as “everyday facts” are concerned. These are statements about the social world that we can be fairly sure of and consequently there tends to be a fair bit of agreement about them. For example Willis’ famous ethnography “Learning to Labour”, is a study of working class high school students in a particular school in Britain. Much of what is revealed in this study would be seen as everyday facts and the publication of this research reveals these facts. Consequently they become part of an accumulation of social knowledge – for example what the interviewees said as quoted in the book; the kinds of activities that Willis reports taking place in the school. As well as trusting the honesty of Willis’ account, the reader compares the situations described to similar situations in their own experience. Flyvbjerg presents a lot of knowledge of this kind as he illustrates his own book. For example, his research into town planning decisions in Aalberg in Denmark. It is presented as a fact that the Chamber of Industry and Commerce secretly reviewed town-planning decisions before they were presented to elected officials on the Council. These are all everyday facts (Flyvbjerg 2001: 148). Or elsewhere when Flyvbjerg maintains that the groups that worked for adult suffrage for men had no vision that women should also get voting rights (Flyvbjerg 2001: 97).
So it cannot be the accumulation of knowledge of everyday facts that Flyvbjerg has in mind when he says there is no accumulation of knowledge in the social sciences. He must be talking about theoretical paradigms; interpretive accounts of social practices. But even here it is hard to see him as consistent. What is the debate between Levi-Strauss and Bourdieu founded on? Bourdieu cites the evidence of ethnography to undermine Levi-Strauss’s theory of the gift. Surely here is an attempt to accumulate knowledge by supporting a superior theory by citing relevant everyday facts.
For the most part, interpretive theories in the social sciences are attempts to make sense of a range of everyday facts that are collected by the researcher or well known to the reader already – or both. The interpretation is an attempt to state what are the underlying structures of social constraint that can be used to understand and bring all these disparate instants into a common interpretive scheme. This is capitalism. This is the disciplinary society. This is the intersection of patriarchal power and capitalist economic relations. This is a tooth filing ceremony. I will suggest in this and the next chapter that these kinds of explanations are attempts to discover underlying social powers or social forces, which are responsible for the everyday events we are citing. They always include assumptions about human nature and locate the events within a historical setting that is assumed or specified.
One would have to agree with Flyvbjerg that such interpretive schemes often co-exist without social researchers agreeing on which has priority. One would also have to agree that they come and go to a certain extent as a result of fashion – or at least without any clear evidence that an older theory has been abandoned because the evidence does not fit it.
Nevertheless, there is equally no doubt that theories are backed up by evidence. The difficulty of accounting for many obvious facts does harm a theory.
We could take the “materialist” theory of history as an example. Maybe there are a number of ways to interpret what Marx meant by this. In one interpretation it is the claim that what drives changes in society are people’s material interests – their interests in food, housing, physical well being. While this is an attractive simplification of world history and at least part of the story, social historians have undermined this through detailed research. A key study was E.P Thomspon’s “Making of the English Working Class” – the essential argument of this account is that the English working class was socially constructed as a piece of social invention by members of the group that came to see itself as the working class. It was not some automatic and inevitable response to common material conditions. This was one of a number of social histories that gradually undermined the materialist theory of history.
Or to take another example, a common folk theory of male homosexuality (that is also presented as medical science) takes it that male homosexuality is an “inversion” of gender and that homosexual desire is an innate condition of some small part of the male population. There is no doubt that most social scientists regard this theory as refuted by the evidence of the everyday facts revealed in studies of the classical world, and studies of the sexualities of traditional Melanesian cultures. These studies show that homosexual acts are universal across the male population in these cultures and they are not linked to gender inversion. The social construction of sexuality is currently the most popular theory to account for this field.
Why different paradigms coexist in the social sciences
Beyond this, there are some other useful ways to think of the unresolved contestations between paradigms in the social sciences. One is to say that different paradigms are different explanations of the same sets of everyday facts. So a dispute between two different theories may remain unresolved because both theories are good accounts of the facts; but they make different interpretations of the social constraints that are leading to these facts. The looseness of fit between social theories and the everyday facts is easily explained if it is in the nature of people as social actors to respond to social constraints in a somewhat unpredictable fashion. No wonder we can develop different accounts of the social constraints that are operating. Another possibility is that there actually are ways of mapping the different theoretical paradigms onto each other. There is not always as much contradiction here as one might assume. As I shall argue later, there are ways of mapping much of the Marxist conceptual language into the terms of poststructuralism and vice versa.
Perhaps we can make a useful analogy from the natural sciences. Newtonian physics and Einstein’s theory of relativity have very different conceptions of the physical forces that operate in the universe. Yet both can be used to explain sets of observations about the physical world – the everyday facts. Although these two theories do actually predict different observations, different everyday facts, these differences are not normally observable. They are very minor and hard to detect. It makes sense to think of social theory as always operating in conditions like this. However in social science there is no possibility of the kind of fine experiments, which might be used to determine which theory is the most true account of the nature of social reality.
Elements of explanations in social science
Events
Social events are what social scientists are mostly interested in explaining. Singular events such as the end of feudalism in Europe are often investigated but mostly, social explanations summarize a whole number of events regarded as being similar. For example the end of feudal society includes the freeing of the serfs and this process is itself defined in terms of a set of similar events, which together make up that process.
Social objects
People as individuals are social objects, but then again there are also larger scale social objects such as social classes, nations, ethnic groups, genders, and organizations. All of these things are taken to be real, although large-scale social objects may just be collections of individuals, or certain actions of individuals, theorized to be linked up in a certain way.
Powers
You can get nowhere in understanding social science explanations without realizing that various kinds of social powers are being considered. Powers are causal properties of things, which are defined and described in terms of their tendency to cause a particular kind of effect. That is, all powers are seen as “directed” to a certain kind of action. Explanations in terms of powers are the stock in trade of everyday accounts of physical events. For example if a brick is hurled through a window and the glass breaks, we could explain these events by saying that it is a well known fact that glass is fragile and bricks are heavy. To be fragile is to have a tendency to break easily. To be heavy is to be able to produce a multiplicity of effects – like breaking glass but also like sinking in water or being difficult to lift. More technical terms such as “force” also are powers and are defined in terms of tendencies to produce actions and to have directionality. Powers are independent of their manifestations. In other words, a power can be present without manifesting itself. They are quite real, just as much as objects or events. Powers are tied to their objects; their existence in an object does not depend on anything that other objects are doing, though how they manifest can obviously depend on what other objects are doing. They exist whether or not we are aware of them. The point of any kind of science is to discover them and work out how they operate (Molnar 2003).
Social scientists use explanations based on powers all the time. The most obvious ones, which it would be hard to ignore, are to do with the actions of people. When we say someone is acting, we mean that the event that we are talking about is informed by an intention. So bumping into someone is just an event if it is an accident. However, if it is intentional it is an action informed by the intention to bump. The intention being talked about here is a power – which is described here in terms of its likely outcome – to bump someone. Going beyond such an obvious case are actions, which are described in terms of intentions that are not named for the action in question. Here the action is seen as a means to realize the intention and what is being assumed is a set of beliefs about the relationship between the action and its intention. So someone buys a ticket to catch a train. Why are they buying the ticket? To catch the train. The intention to catch the train is seen as a cause of the action. The power is defined in terms of its direction – catching the train – but the action is a means to that end. The kind of sociology referred to as Verstehende sociology – based on an understanding of the meaning of action – is an explanation of actions in terms of intentions; or in other words in terms of causal powers.
For the most part we tend to think that people’s intentions are socially constructed. The social scientist investigates actions and finds the socially constructed intentions that lie behind them and the sets of beliefs that inform these intentions. This is certainly a large part of what is going on. However I want to advance the view that social scientists very often make reference to human nature as a set of causal powers. Human nature is composed of those intentions, which are common to all people and causally central to human action. Aspects of human nature are discovered through more or less explicit and more or less common sense cross-cultural research. We are now beginning to also develop various biologically based theories of human nature. Of course, much of what are argued to be biologically informed views of human nature are rejected by social scientists as ill informed by cross cultural analysis – they are merely wild hypotheses about how certain traits unique to Western culture are supposed to have been formed through the stringency of natural selection. Nevertheless there are certainly aspects of biological theories that social scientists do tend to accept as true and do inform their understandings. I will discuss different theories of human nature in a later chapter.
In social sciences, large-scale social objects are also invested with social powers. Classes may be considered to have a class interest. For example business people attempt to maximize their profits by tending to move investments to follow profit. In a way, this is clearly just a summary of a range of individual actions, but the patterned nature of this set of actions implies a social force that constrains action. It becomes evidence in social science for the existence of a social power, which we may want to portray as class interest; or we may want to look at it in reference to private ownership of the means of production, which in itself is a social force located in the economy as a whole. Like other powers, private ownership is defined in terms of what it can do – exclude others from using your property and distributing its products. Its directedness is apparent.
As argued above, the existence of social powers of this kind is “induced” not “deduced” from what we observe. What that means is that the more everyday facts, which we take as evidence for the existence of this social power, are never sufficient as to imply its existence. The social power in question is supposed as something that, if it exists, could explain a great number of the everyday facts that we do notice. This is the evidence for its existence. But what we can also acknowledge in this situation is that a number of rival accounts of social powers may be advanced to explain the same set of everyday facts.
The Poststructuralist challenge to “humanist” social science
In the rest of this chapter, I will be looking at some aspects of recent poststructuralist theory and its claim to have developed a set of radically new concepts to describe society – a new set of social objects which are in flat contradiction with the kind of concepts hitherto used in the social sciences, those of Marxist and “humanist” sociology. This meta-theory gap is one of the roots of Flyvbjerg’s claim that there is no cumulative development of knowledge in the social sciences. In other words, without any new accumulation of contradictory facts, poststructuralism introduces a radically new way of explaining what is going on in the social world. This new meta-theory has taken the world of social science by storm and calls into question the idea that earlier social science can be seen as “explaining” the same social object which is now considered by poststructuralism. The concepts that I will map will be discourse, agency, ideology and hegemonic masculinity. I have two aims. One is to show that supposedly contrary concepts can actually be mapped onto each other. The other will be to defend some of the concepts often used by earlier sociologists, which have been meta-theoretically challenged by poststructuralism.
Discourses and Subjects
I will begin this analysis by talking about Weedon’s description of poststructuralism, which creates a good summary of key writings of Foucault. She begins with the idea of “subjectivity” which is someone’s sense of themselves. Subjectivity is socially produced through “discourses”. These two terms are thus related.
Subjectivity is used to refer to the conscious and unconscious thoughts and emotions of the individual, her sense of herself and her ways of understanding her relation to the world. (Weedon 1987: 32)
Weedon argues that traditional discourses of “liberal humanism” suppose a fixed essence at the centre of subjectivity. However poststructuralism:
… proposes a subjectivity which is precarious, contradictory and in process, constantly being reconstituted in discourse each time we think or speak … the individual is always the site of conflicting forms of subjectivity. (Weedon 1987: 33)
This is because the subjectivity of individuals is always created within socially produced discourses. These discourses are incoherent in the sense that they offer “competing ways of giving meaning to the world” (Weedon 1987: 24).
Discursive fields consist of competing ways of giving meaning to the world and of organizing social institutions and processes (Weedon 1987: 35)
So a discourse is not just a set of understandings. It is crucially also a set of actions, which have been undertaken with those understandings in mind. Actions taken within the framework of a discourse in fact organize social life in institutions and processes. This way of looking at social action is made sense of through the concept of “subject position”. Individuals construct their subjectivity by taking up subject positions within discourses. We can think of a discourse most helpfully by comparing it to a game with specific rules and positions. For example the game of Rugby League with positions such as front row forward or hooker and so on. Or the game of chess with positions such as Black Knight, White Pawn and so on.
A good example of the use of this framework is Cranny-Francis’ discussion of the “Red Riding Hood” discourse (1992). The folk tale, Red Riding Hood offers various subject positions, which the reader can take up. The Wolf, Red Riding Hood, the Mother, the Grandmother, the Hunter. Of course one important site of this discourse is the tale of Red Riding Hood itself, but Cranny-Francis also argues that it is present in other forms, such as the popular film of Batman. It is also manifest, she argues, in the way people organize and live their lives. The male reader is invited by the discourse to take up one of two subject positions – the Wolf or the Hunter.
The boy is positioned to see in himself as a man two different kinds of character, both of which are violent. There is the admired patriarch, the hunter, the male authority figure, who will protect women – apparently from the wolf, though actually from themselves (that is, from any transgressive expression of their own sexuality). The other character is the animalistic, uncontrollable beast within the man, who preys on women. (Cranny-Francis 1992: 82)
These are the only subject positions made available to men by this discourse. Both these figures use violence with the implication that it is impossible to be a real man and not to be violent at all. The narrative implies that the sexually active male – the Wolf – is a dangerous anti-social figure, while the other option is to be in control of masculine sexuality – the position of the Hunter. So the discourse of Red Riding Hood is seen by Cranny-Francis as offering a number of subject positions for readers of the narrative and readers may be influenced to take up these subject positions and so to create their “subjectivity” (or what we would most usually call their personality) through this. Cranny-Francis argues that through this discourse (in all its manifestations) men are urged to see their sexuality as innately predatory and dangerous; an urge which they can barely control. So men who take up this subject position of the Wolf interpret their own behaviour through this discourse. There was an example relevant to this in the Australian media. Diane Brimble was taking a holiday on a cruise ship, and one of the men she met drugged her to ensure that she would be a passive recipient of sex. She became unconscious and instead of calling for medical help, he and his friends left her to die from an overdose. We could say that this man was taking up the subject position of the Wolf – a man whose “natural biological sexual urges” could not be controlled, regardless of the harm wrought by this behaviour.
Weedon expresses this concept of the relationship between subjects and discourses in a number of ways and as I will suggest later, these are not always consistent. Discourses:
… offer the individual a range of modes of subjectivity. (Weedon 1987: 35)
In other words, different discourses offer individuals different subject positions and within the one discourse there are also a range of subject positions offered to people. In poststructuralist theory:
… the structure and function of the position of the subject within discourse is the precondition for the individual to assume historically specific forms of subjectivity within particular discourses. (Weedon 1987: 31)
The important point here is that subjectivity is never naïve; there is never an essential “I” behind all the various historical forms of subjectivity. Each form of subjectivity is a subject position within a historically created discourse. In this and the next chapter I will argue there are some respects in which this cannot be the case, but it is certainly a key idea of poststructuralist theory. Some aspects of this position have already been argued for in the previous chapter. Perception, as claimed above, is never naïve – it is the acquisition of belief. Because of that it is always informed by other beliefs, which come out of historically located experience. If this is true of perception it is certainly true of conscious thought in general. In another statement of this relationship, Weedon writes:
… it is language, in the form of conflicting discourses which constitutes us as conscious thinking subjects and enables us to give meaning to the world. (Weedon 1987: 32)
Does Weedon mean by this that there is nothing more to conscious thought than what we can express in sentences? This seems unlikely if we think of trying to describe the shape of an inkblot. We certainly have beliefs about the shape of an inkblot and these beliefs may be conscious, although they are very hard to put into words except in the most summary and incomplete fashion. Yet it is certainly language and thoughts in language that “give meaning” to the world.
An important aspect of the poststructuralist position is that discourses offer competing ways of understanding the world. As a result of this, subject positions and subjectivity can shift depending on what discourse or subject position is being adopted. For example, men may shift between the subject position of the wolf and that of the hunter depending on the circumstances. A good example of this idea of shifting subject positions informed by different and contradictory discourses is Hudson’s article on femininity and adolescence (date ). Hudson argues that adolescent girls are invited to take up subject positions in two different discourses with implications for behaviour that are quite contradictory. On the one hand there is the discourse of “adolescence” which suggests that young people are in the grip of hormonal changes, which lead them to wild excesses of sexual experimentation and rebellion against older authority figures. This is the discourse of adolescence and the subject position is “adolescent”. The other discourse is the discourse of “femininity”. This has quite different implications. The girl who is becoming a woman is meant to express femininity by compliant and deferential behaviour where authority figures are concerned and to indicate a commitment to marriage and moral motherhood by sexual restraint and “dating” as a series of quasi-marital romantic commitments leading to marriage. Hudson suggests that individual adolescent girls swing between these two subject positions and often feel in a situation of conflict in relation to the demands of these different discourses.
Determinist and agentic versions of poststructuralism
Before going on to the question of the role of the subject in poststructuralism we should review the kinds of statements of Foucault’s that have been the basis for this position. I will just mention a few examples. For example in “What is an Author” Foucault imagines a scenario in the future in which an author’s role as originator of the text is no longer socially significant. In such a context the questions that would be asked would be:
What are the modes of existence of this discourse [the one that is instanced in the text in question]? Where has it been used, how can it circulate, and who can appropriate it for himself? What are the places in it where there is room for possible subjects? Who can assume these various subject functions? (Foucault 1984: 120)
Earlier, he says that we should stop asking questions about how “a free subject” can activate a language and make it confirm to designs of their own. Instead we should consider questions such as:
How, under what conditions, and in what forms can something like a subject appear in the order of discourse? What place can it occupy in each type of discourse, what functions can it assume, and by obeying what rules? (Foucault 1984: 118)
In other words, a discourse sets up certain rules and these specify the “subject functions” or what have been called “subject positions” and thereby create “room” for possible subjects. Clearly there are also rules about “who can assume” these positions – for example that it is men who can take up the positions of wolf and hunter. The implication is that subject positions change historically with discourses and so people’s subjectivity is historically variable. The very idea that there is an author who is a unique individual who creates a new idea, is a historically specific construction. It does not represent any kind of universal truth about the way humans operate to create culture. Instead, what is universally true is that human subjectivity is constituted very differently in different discourses. In “Truth and Power” he writes that he does not want to posit a subject:
… which is either transcendental in relation to the field of events or runs in its empty sameness throughout the course of history. (Foucault 1984: 59)
In other words, there is no essential “I”, some kind of neutral mental arbiter which is running the show in all people, across different periods of history. Instead, people’s internal subjectivity, their self, is totally determined by historically specific discourses. A relevant statement on the way discourses offer competing forms of subjectivity is the discussion in the “History of Sexuality” of forms of resistance to power:
Are there no great radical ruptures, massive binary divisions, then? Occasionally, yes, but more often one is dealing with mobile and transitory points of resistance, producing cleavages in a society that shift about, fracturing unities and effecting regroupings, furrowing across individuals themselves, cutting them up and remoulding them, marking off irreducible regions in them, in their bodies and minds … the swarm of points of resistance traverses social stratifications and individual unities. (Foucault 1980: 96)
We can take this statement as referring not just to types of resistance but also to discourses. Discourses cut across individuals because different discourses offer the individual different types of subjectivity. So the implication is that there is no unified central mental entity that makes decisions and chooses and invents things – people are merely the locations in which different discourses formed in history play out and express themselves.
One of the most tricky issues with poststructuralism is to work out whether people who write within this framework mean us to believe that people’s subjectivity is determined absolutely by discourses or alternatively that people have some control over the discourses and subject positions that they take up. There is much in Foucault’s writing that supports both these interpretations and consequently it seems that different writers in this tradition either replicate this ambivalence or fall more heavily on one side of it than the other. What is also not entirely clear in Foucault’s own writing is whether he is asking us to believe that the determinist vision is a picture of how social reality actually is – or does he want to present it as a kind of useful thought exercise. As a corrective to “humanist” visions which act as though the subject is an individual, independent of social influence?
The determinist version of poststructuralism
It is actually quite difficult to get one’s head around a vision of social life which we “dispense with the constituent subject” or “get rid of the subject itself” and only look at the subject as it is constituted “within a historical framework” (Foucault 1984: 59). My best analogy for this vision is that of corks bobbing on waves in the ocean. The individual is merely a point through which social discourses operate – a nodal point of discourses. The individual subject has no more power to impact on these discourses than a cork bobbing on the ocean. Discourses that contradict each other are like waves coming from different directions in the ocean. They impact together on the cork to push it one-way and then another. As Geras remarks, this determinist vision is an “overweening social ontology”:
… absolutely everything – subjects, experience, identity, struggles, movements – has discursive ‘conditions of possibility’, while the question as to what might be the conditions of possibility of discourse itself does not trouble the others so much as to pause for thought … subjects are no kind of social ‘origin’, not even in the limited sense of being endowed with powers that render an experience possible! (Geras 1990: 102)
On the analogy of waves and the ocean, all we can do is to say that different waves have different weights and in fact move corks to a different extent. Their weight can be inferred from the behaviour of the corks. Where waves themselves come from or how a new wave is started is outside of the scope of social science.
Agentic versions of poststructuralism
In an agentic version of poststructuralism, at least some kind of subjectivity running in “empty sameness” throughout history is presupposed. At least there is a choice maker, whether conscious or not, that decides which discourse and subject position to invoke out of a number of socially available options. This is the “agent”, which is inherent in subjectivity and has agency. Then what may be also added is the view that these choice makers and social agents are actually the inventors of new discursive elements that have never existed before they were invented by these agents. If we are using an example from Foucault’s own historical writing, there are those subjects who are the first to invent and use the new discourse of homosexual identity as a natural condition and hence one deserving civil rights – the new discourse that challenges the medical discourse of inversion. An invention and intervention, created by “agentic” subjects. As I shall suggest in the next chapter, even elements of a theory of human nature can creep into poststructuralism when it actually starts to explain social change in detail, and explain why subjects made certain kinds of choices rather than others, why certain kinds of social invention were taken up. With this kind of interpretation of poststructurialism, the metaphor of the bobbing cork and humans as nodal points for the interaction of discourse is abandoned. We have human agents as the inventors of discourses and the ones who make choices about what discourses to adopt.
But of course all of this reduces the uniqueness of poststructuralism as meta-theory and ends up creating it as a set of meta-theoretical elements that can be mapped onto humanist understandings of the subject that are shared by sociologists such as Marx and Weber, as well as more recent writers.
How this debate plays out in poststructuralist writing - Weedon
What I will argue is that this “agentic version” of poststructuralism is present as a theme even in poststructuralist writings that also present discourses as the only relevant social object for understanding social existence. I shall trace this ambivalence in Weedon’s explanation of poststructuralism and again in Foucault’s own discussions. My point is really that the meta-theory of determinist poststructuralism is much easier to put in grand programmatic phrases than it is to actually sustain over pages of description of social life.
A typically ambivalent passage from Weedon is the following:
… it is language, in the form of conflicting discourses which constitutes us as conscious thinking subjects and enables us to give meaning to the world and act to transform it. (Weedon 1987: 32)
Here, discourse functions like a kind of programming which sets us up as a thinking machine. However once these programs are running “we”, in other words, we agentic subjects, are able to use them to transform the world. Using these programs “we” are able to give meaning to the world and transform it. However in a more determinist vein she writes that we are:
… subjected to the regime of meaning of a particular discourse and enabled to act accordingly. (Weedon 1987: 34)
Here it feels as though the only way we can act is that which has been programmed by the discourse. Another passage certainly creates the vision of the individual as a bobbing cork:
… poststructuralism proposes a subjectivity which is … being reconstituted in discourse each time we think or speak … the individual is always the site of conflicting forms of subjectivity. (Weedon 1987: 33)
Here our subjectivity is the product of discourses. When we think or speak, these discourses are occasioned at the site of the individual and constitute their subjectivity at that point in time.
However, later a more agentic version of the meta-theory is enunciated:
The individual who has a memory and an already discursively constituted sense of identity may resist particular interpellations or produce new versions of meaning from the conflicts and contradictions between discourses. (Weedon 1987: 106)
Here, one has a sense of the individual with various capacities, programmed by discourses in the past but capable of producing “new versions of meaning” by noting the contradictions between discourses. The implication is that there is in fact a human subject, an inventor and choice maker, that actively and in all periods of history, shapes the up take of the discourses that are socially available and creatively participates in the invention of new discourses.
Foucault as determinist poststructuralism
There is the same ambivalence in Foucault’s own writing relevant to this. Much fits with the determinist vision. For example the passage from the History of Sexuality:
… one is dealing with mobile and transitory points of resistance … furrowing across individuals themselves, cutting them up and remolding them … the swarm of points of resistance … (Foucault 1981: 96)
Here it is not individuals who resist. The individuals are sites at which points of resistance take place as discourses furrow through them. This fits with the statement in “Truth and Power” where Foucault does not want to posit a subject that is “transcendental in relation to the field of events” (1984: 59). There is no individual who resists in the context of a variety of different discursive situations, since the individual is just the site at which points of resistance take place. Similarly in “What is an Author?” he states:
… it is a matter of depriving the subject (or its substitute) of its role as an originator, and of analyzing the subject as a variable and complex function of discourse. (Foucault 1984: 118)
So we are not to think of the subject as the creator of new discursive elements, or indeed as the one that chooses between discursive positions. Instead the subject is a “function” of discourse. This is the determinist reading.
But we could also take note of some of the ambivalence of this passage. Does he just mean that we should stop paying attention to that which has been obsessive in literary criticism – the author as originator of new creative ideas and instead pay attention to the author as someone who is producing discourses which are socially available at large – the author’s text is just one instance of a more broadly socially constituted way of thinking? In other words, these could be seen as two sides of reality and it is high time we started paying attention to the latter. This is an eminently sensible suggestion but does not create a mystique that something truly strange and profound is being demanded.
As well, what is stated in this essay and also in other works is the idea that the concept of a “subject” is something, which is unique to the modern age – something that can “appear” in discourse (Foucault 1984: 118). That, for example, it does not exist in, mediaeval society – where people were very loose about attributing authorship, compared to the period of high capitalism that followed. This is an intriguing idea and is itself ambivalent in its meaning.
Does Foucault mean that people in other social orders were not particularly concerned about the creative individual genius of the author and that they perhaps believed that creation was in fact from the gods or came from some such source outside the author? Were they not so silly as we are in our culture to attribute everything to the individual who actually puts pen to the page? None of this is in any sense problematic as social history or anthropology. Or does he literally mean that in other cultures individuals did not act, did not create and were never seen as doing so? This seems so unlikely as to be hardly thinkable in any language. Ethnography and history are full of accounts of the actions of individuals and it is clear from these accounts that their societies recognized them as actors, along with a whole range of other beings. Again, one would have to say that it is the ambivalence of Foucault’s statements that creates the sense of something truly profound being put. The reader uses their understanding of the first (more obvious) sense of the statement to put clothing on the idea while they are entranced by the possibility that something much more surprising is possible – that in other cultures individuals are not seen to act and do not act in the sense in which we understand that.
Foucault as agentic poststructuralism
Yet of course Foucault himself cannot actually maintain the truly radical meta-theoretical idea that there are no subjects and that people never act. He constantly implies the opposite. For example two pages after the preceding grand declaration we get:
What are the modes of existence of this discourse? Where has it been used, how can it circulate, and who can appropriate it for himself? What are the places in it where there is room for possible subjects? Who can assume these various subject positions? (Foucault 1984: 120)
So who is doing all this appropriating and assuming if not the subject, transcendental in relation to the various discourse positions, which they assume. Interpreted like that, the statement makes perfect sense as an encapsulation of the concept of discourse and the way discourses relate to active, choosing subjects. Discourses are sets of ideas, which we can see circulating in society at large. They are like a game in that they offer various kinds of options for people – to take up positions and play parts within the vision of the social world that they set out. They also specify the kinds of people who are to be allowed to take up the subject positions that they allocate – men, women, the poor or whatever. But at the end of the day, individuals also make some decisions about what to do with these resources – whether consciously or not.
This agentic version of poststructuralism is also present in other statements about how discourses should be interpreted in relation to power. Foucault often invites us to think of discourses as resources that can be used by people (subjects) to engage in social conflicts and to gain or lose power. For example history:
… is intelligible and should be susceptible to analysis down to the smallest detail, - but this is in accordance with the intelligibility of struggles, of strategies and tactics. (Foucault 1984: 56)
Some conclusions
As may be obvious from the above, I tend to think of determinist poststructuralism as a parlour game in which we pretend not to know the things about people which make sense of the behaviour of discourses and the production of discourses by speaking subjects. We pretend not to understand that people create discourses and make choices about taking up discourses and we pretend not to understand that these decisions are influenced by various aspects of human nature that are understood or taken for granted in other social science. I find agentic poststructuralism more plausible and believe that any extended account of real social life ends up by smuggling in agentic meta-theory despite the best intentions of social determinism. As I will go on to claim, you can see talk about social objects and talk about individuals as two contradictory but mutually necessary ways of understanding social behaviour – as being parallel in physical sciences to talking about electrons as particles or waves.
This analysis has two implications. Firstly, that the meta-theory of poststructuralism cannot be as radically different from that of standard “humanist” social science as some of its proponents would like. Much about it can be mapped onto these other metatheories. The second is that poststructuralism is a very useful set of meta-theoretical ideas because it helps us to look at standard social science questions with new some tools.
The multiplicity of the subject?
What is probably another issue is the poststructuralist view that the subject is not “unified” and not always “conscious”. Weedon talks about this a lot and so does Foucault. I want to separate this from the issue of whether poststructuralism is “agentic” or “determinist”. I do not see any contradiction between an agentic version of poststructuralism and a view that the subject is multiple (and for the most part unconscious) and that the sense of subjective unity is an illusion.
For example, as an approach to these issues that mirrors some writing in poststructuralism we can take Dennett’s philosophical writings on the nature of the brain and consciousness. Dennett (1993) presents what he calls a “functionalist” view of consciousness. He argues that the brain throws up conscious thoughts as a means of broadcasting information from one part of the brain to another. As a result, conscious thought never comes from just the one place in the brain, and nor do actions. The brain operates through parallel and complementary processes of thought, creating and revising versions of what is going on at any one time. These processes go on constantly and are only made “conscious” at certain points and under certain conditions. There is no central place of consciousness, which orders other events in the brain. A motley crew of brain processes, operating quite independently to a large extent, produce actions, including actions of conscious thought.
On the other hand, Dennett argues that part of what the brain does in all this is to create a false illusion of mental processes, which is like a short hand useful summary, which relates all these conscious elements together and processes them into a narrative of the individual conscious human subject. Dennett seems to think that this narrative is necessary or inevitable for the human being to negotiate life – it is not really an invention of culture in the way Foucault suggests. On the other hand, there is no doubt that a process of this kind is likely to be heavily influenced by cultural knowledge and assumptions, since what we are talking about here is really a false belief, a false theory of how the mind operates.
It is not out of the question that the experience of unified conscious subjectivity is in fact a theory that arises within a social context. Which could be patriarchy, where men act as heads of families or it could be class societies, where a ruling king or ruling class is the centre of social life. It is certainly not a “necessary” piece of brain hardware that must always be operating for day to day thought to take place. We abandon it by meditation or taking hallucinogenic drugs; states of consciousness in which the unified subject of liberal humanism is asleep on the watch. This territory is of course the field of a variety of postmodernist theories; Lacan, Deleuze and Guattari, Grosz and so on.
Dennett himself quotes some poststructuralist writing with approval and clearly believes that his own view fits with the abandonment of the unified subject that we find in poststructuralism. While these two views certainly share some territory, there is a key difference between Dennett’s view and some poststructuralism. Dennett’s theory does not imply that the subject is just a nodal point through which discourses operate. To that extent he does not take on a determinist poststructuralism. It is clear that he is quite happy with the idea that the brain (in parts) initiates ideas, becomes conscious, reasons and takes up options. He maintains that different parts of the brain constantly take on the role of creating and initiating conscious and unconscious thought and action. These outcomes arise out of a vast variety of complex processes going on in parallel, and the individual brain has a huge part to play in these outcomes, they are not just produced as an effect of external social forces.
I am happy to acknowledge Dennett’s ideas and their parallel expression in poststructuralism as a contribution to social sciences that is quite novel. The disunity of the subject and the unconsciousness of much thinking are discoveries of the modern period that have to be taken on board in social science. The reader might well have to read the scientific evidence that Dennett presents to be convinced. I note that if this account is true it is in itself an aspect of a theory of human nature!
Discourses and Ideologies
My view is that one of the things that is going on when different fashions of social theory come and go is that these different fashions are not necessarily incompatible and incommensurable, as Flyvbjerg argues. On the contrary, it is often relatively easy to translate between theories and map different perspectives onto each other. Sectarian conflicts and mutual antagonism between theoretical camps can sometimes be better explained socially rather by than assuming that there is a radical meta-theoretical incompatibility.
We can explore this idea of mapping by looking at the differences and overlaps between concepts of “ideology” and “discourse”. Marx introduced his use of the concept of ideology to refer to ideas which deny the material realities of history and by doing this constitute a masking of social processes. This masking aids ruling classes and expresses their interests – they flatter themselves that they are the authors of ideas that are conceived without being tainted by material needs.
This use of the term by Marx has been broadened in two ways, Firstly, the term is now used to cover any kind of system of masking and false beliefs, which supports ruling class interests, not just beliefs that deny the material realities of history. This usage has also been broadened to consider a variety of systems of power and inequality, including race and gender as well as class. While this concept of ideology always assumes that ideologies are false ideas, it has also been realized that ideology cannot be effective with a set of completely implausible ideas. As Gramsci and Stuart Hall have pointed out, for an ideology to become “common sense”, there must be something about social experience, which makes an ideological view of things work. Let us take this as the current standard definition of the term, against which we can compare the concept of “discourse”.
The reader will be aware that this current sociological usage of the term, “ideology” is somewhat different from the way the same term is commonly used in political science or everyday journalism – where an “ideology” is just a set of ideas about society, with the usual implication that they are extreme and obsessive. We will just ignore this latter usage for the purposes of this comparison.
Ideologies are false
In writing in the social sciences, to say something is an ideology is always to make some claim about what the author thinks society is actually like and to counter pose that truth to the false claim made in the ideology. The term “discourse” has no such implications. Discourses always lay claim to truth, of course, but to talk about something as a discourse has no implications about whether these ideas are in fact true or false.
As I have explained, Foucault argues that it is important to divest ourselves of the view that true statements have any necessary relationship to power in society – being by their nature subversive because the powerful always try to hide the truth. This may not be the case. I am happy to take this on as an insight that has come with poststructuralism and as a necessary corrective to an enlightenment myth of progress that has gone with some Marxist thought. But it does not necessarily mean that the concept of ideology is now obsolete and should be discarded. Let us agree that not everything said to support ruling powers is false (and therefore an ideology). Let us agree that not everything that opposes ruling powers is necessarily true (that all statements that have a radical political effect are also good science). But there is nothing in this to suggest we are mistaken in using the concept of ideology to refer to ideas that are false and also support ruling powers. The definition of the term can quite well withstand this corrective to the empirical claims made by earlier Marxists.
Thoughts and actions
A second difference sometimes invoked to distinguish “ideologies” from “discourses” relates to the kinds of social objects included under the terms. While the term “ideology” refers to a set of ideas, when people think of a “discourse” they are also thinking of the actions, which are carried out with those ideas in mind. The discourse of “motherhood” is both a set of ideas about how mothering should be done and also the actions that people engage in as they realize these ideas – the actions of mothers and babies, fathers and so on.
Foucault is careful to label discourses in terms of sets of activities, not just sets of ideas that go with them. In the “Birth of the Clinic” it is the techniques of observation and the primacy accorded to the visual, which are as much the signifiers of medical discourse as any set of ideas. So this concept allows a certain amount of looseness of fit between ideas and activities.
Yet in practice this distinction between ideology and discourse is hard to maintain. There is always a way of stating in words what these activities have in common and that set of words could be taken as the central idea that animates the activities in question – the idea “that one should observe in order to understand”, as a key aspect of the new medical discourse at the heart of the birth of the clinic. At the same time we might have to agree that Foucault is right in that actions can be linked together as aspects of the same discourse – without the participants being consciously aware of ideas which link their actions. “Dressage” is something which Foucault as the social scientist can discern as a similar set of practices in diverse settings from the military to prisons – but it is not an idea which is being put forward as such to explain these strategies at the time. So we can think of a “common idea” as a mental object which links practices – without such an object necessarily being conscious.
So, one can feel that this distinction between ideology and discourse (as ideas versus actions) can be overemphasized. As we shall see, the concept of hegemonic masculinity as invented by Connell, prioritizes practice in a way that is very similar to Foucault’s discussion of discourses – while it also relies heavily on the Marxist concept of ideology.
At any rate, what we have been doing in this discussion is using terms which are understandable by both parties to explain differences in the way the contrasting concepts of ideology and discourse are being used. We are mapping these concepts onto each other. This is not a case where incommensurable paradigms come and go with the mad logic of a fashion and nothing to connect them. It is very far from this.
Ideologies support a ruling group
What remains as a key difference is that an ideology (as this term is used in sociology at any rate) always refers to a set of ideas that supports a ruling group; whether that is the ruling class as in Marx’s original use of the term, or men as a ruling group, as in feminist theory and so on. By contrast a discourse can have any relation to power structures. Foucault, as noted, is keen to point out that the same, or at least a related, set of ideas can actually be used tactically by different elements in society with very different relationships to power. One can see this as a key insight of poststructuralism and a genuine advance in social knowledge, which the concept of “ideology” tends to obscure. The concept of “ideology” has tended to presume that a particular set of ideas is necessarily locked into some particular part in power struggles.
On the other hand, given that we have taken this new poststructuralist insight on board, the concept of ideology does not in fact become obsolete. What we have is two kinds of ways of demarcating the world of ideas.
Mapping the concepts onto each other
In naming something as a “discourse” we are attending to the particular sets of ideas and actions that define it – it is the discourse of “moral motherhood” or the discourse of “Red Riding Hood” or the discourse of “gender inversion” or “dressage” or whatever. These practices linked by common patterns of thought are the common thread that links different instances of the discourse in question, even when the relationship of these ideas to power structures may change.
In naming a set of ideas as an “ideology” it is the relation of ideas to power structures that we are attending to. We are talking about “capitalist ideology” or “patriarchal ideology”. When we use the term “capitalist ideology” we are not expecting that all the various ideas that may back up capitalism on different occasions are the same or even that they hang together coherently. We can reasonably expect that the actual content of these ideas may vary as particular kinds of social changes take place and new ideas are found useful to support capitalist power structures. We do not expect “capitalist ideology” to always consist of the same set of ideas in every time and place where capitalism is dominant as a social structure. So the label “capitalist ideology” may be applied to a range of discourses that are quite different on different occasions. What these discourses have in common is their common relationship to a structure of power and inequality.
This mapping also has some consequences in terms of what kinds of things can be labelled as discourses or ideologies. The discourse label is more inclusive. We can have feminist and patriarchal discourses – or sets of ideas which support feminism or patriarchy. But we cannot have “feminist ideology” while we can have “patriarchal ideology”. Why? Because women are not the dominant group in society, so ideas that back up their struggles to gain power cannot be labelled “ideologies”. This is all about labels; there is nothing in any of this which implies that these meta-theories cannot be mapped onto each other to describe the same social realities in different ways.
These thoughts about ideology and discourse are not just my own. The ready overlap of these meta-theories leads to types of expression that make some poststructuralists shudder. A good example is Cranny-Francis’ writings on the myth of Red Riding Hood (1992) that I have explained above. She usually refers to this myth as a “discourse” and uses the language of subject positions that goes with that. In this context she uses the term “patriarchal discourse” to describe these ideas because they are ideas which support patriarchy. Yet without any sign of embarrassment or hesitation she also talks about the same myth as a “patriarchal ideology”. In other words, she is implying that this discourse is one of those discourses that also happen in fact to be patriarchal ideology – a set of ideas that supports patriarchy. While this may seem to be just a turn of phrase, it is interesting that there are various points where she follows the concept of ideology by making claims that the ideas of this myth are in fact false. For example she thinks it is false that men are innately violent by comparison to women and have sexual urges that are innately difficult to control in comparison to the sexual desires of women. She thinks men can liberate themselves from social constraint and lead happier lives by becoming aware of the falsity of this common myth, which has become part of the Red Riding Hood “ideology”. So her adoption of the term “ideology” here is not just a turn of phrase; she joins it to some of the most central aspects of the use of the term in Marxist sociology. For poststructuralist purists, this may suggest that Cranny-Francis is not a real poststructuralist at all. In my mind, this switching usage is not at all improper, it is a sign that she has correctly understood that these different framing concepts map on to each other quite readily.
Gender Discourses and Hegemonic Masculinities
What I want to argue in this section is that the concept of “hegemonic masculinity” used by Connell in “Gender and Power” (1987), and much used by sociologists since, is clearly and intentionally tied into the concept of ideology. Yet much about it is quite compatible with talk about “discourses of gender” which was equally prevalent soon after this book first came out.
In “Gender and Power”, Connell comes to a definition of hegemonic masculinity in the context of discussions of “sexual character”. In other words, by implication hegemonic masculinity is a personality type. It is a version of masculinity that has purchase in a whole society but resembles “patterns of face-to-face relationships within institutions”. For example, a private school in which boys who were successful at sport – the Bloods – were in ascendance. These Bloods bullied boys who were unsuccessful at sport but also necessary for the success of the school because of their academic success – the Cyrils. In the school, the version of masculinity represented by the Bloods has dominance, although it is not constituted in isolation but in reference to the Cyrils. Similarly in society at large hegemonic masculinity defines itself in relation to subordinated masculinities and in relation to women.
The public face of hegemonic masculinity is not necessarily what powerful men are, but what sustains their power and what large numbers of men are motivated to support. (Connell 1987: 185)
So fantasy figures of ideal masculinity, such as Rambo, are actually quite unlike the powerful men who rule society from boardrooms! Nevertheless it is this ideal of masculinity which has broad support. So what is hegemonic masculinity, as distinct from its “public face”? Most importantly, it refers a set of “practices” that maintain men’s power over women.
In this sense hegemonic masculinity must embody a successful collective strategy in relation to women. (Connell 1987: 186)
As I will consider shortly, Connell argues that no simple or unitary strategy is possible. A mix of strategies is necessary.
In some ways, the concept of hegemonic masculinity can be seen as presenting hegemonic masculinity as a discourse. For example, hegemonic masculinity is on the one hand a set of ideas – the public face of hegemonic masculinity. But it is also the actions or practices that are taken with this set of ideas in mind. As Connell remarks:
The power relations of the society become a constitutive principle of personality dynamics through being adopted as personal project, whether acknowledged or not. (Connell 1987: 215)
Which we could read as follows. The person takes up a subject position as masculine within the discourse of hegemonic masculinity and in doing this becomes committed to a project of oppression of women. As with Foucault’s concept of discourse, hegemonic masculinity is a resource in relation to power struggles or strategies in society.
However the key reason why we could not see “hegemonic masculinity” as a discourse is that it is named in reference to its place in a power structure, that of patriarchy. It is that set of practices and ideas – whatever that may be – which supports men’s power. In that sense the term “hegemonic masculinity” can be seen as linked to the term “patriarchal ideology”.
Clearly, Connell wants to foreground the practical content of this social object; not its content as a set of ideas. But as noted above, ideas are embodied in action and actions embody ideas. This is merely a matter of emphasis.
The effect of Connell’s conceptualisation is to allow the specific ideational content and even the particular strategies and actions that we can call “hegemonic masculinity” to be diverse in any one particular historical setting and changing in relation to history as a whole. While different social orders can be recognized as patriarchal and having a form of hegemonic masculinity in relation to this, the actual content of hegemonic masculinity is diverse.
Connell creates a pertinent example of this in “Masculinities”. Connell counter poses forms of masculinity organized around direct command – exemplified in military officers or business managers and those organized around technical knowledge – exemplified in professions and science.
The latter have challenged the former for hegemony in the gender order of advanced capitalist societies, without complete success. They currently coexist as inflections or alternative emphases within hegemonic masculinity. (Connell 1995: 165)
Because of that, if we were to label the different versions of hegemonic masculinity in relation to the ideas and ideals of their public face, we would end up talking about different discourses of masculinity. Connell is quite clear that there are a variety of hegemonic masculinities and gives a number of examples. A general statement is as follows:
I emphasize that terms such as ‘hegemonic masculinity’ and ‘marginalized masculinities’ name not fixed character types but configurations of practice generated in particular situations in a changing structure of relationships. (Connell 1995: 81)
As examples, Connell cites the invention of “the homosexual” as a type of subordinated masculinity in the eighteenth and nineteenth century, the development of new kinds of hegemonic masculinity that accompanied bureaucratisation and displaced the aristocratic masculinity of the old regime in Europe, the differences between working class masculinity based in physical prowess, machismo and contempt of office work versus professional masculinities, valorising emotional reserve and basing claims to power in technical expertise and theoretical knowledge. In relation to women, he refers to different types of “compliant” and “resistant” femininities and their historical origins and roots – for example the versions of resistant femininity carried by “spinsters, lesbians, unionists, prostitutes, madwomen, rebels and maiden aunts, manual workers, midwives and witches” (Connell 1987: 188). All of these varieties of sexual character could be also considered as “discourses” brought into play in various social struggles for power in relation to the gender order as it changes through time.
Poststructuralist takes on “masculinity” and “femininity”
It is interesting to note that at the same time that Connell was formulating concepts of hegemonic masculinity and emphasized femininity in a way that separated this conceptual system from poststructuralism, Davies, another Australian feminist researcher, was using poststructuralist analysis in tandem with concepts formulated by Connell. The historical variability of discourse is taken as a given within poststucturalist theory and Davies writes:
An assumption of poststructuralist theory is that maleness and femaleness do not have to be discursively structured in the way that they currently are. (Davies 1989: 12)
She assumes that current social constructions of maleness and femaleness are discourses:
… sex and gender are at one and the same time elements of the social structure, and something created by individuals and within individuals as they learn the discursive practices through which that social structure is created and maintained. (Davies 1989: 12)
Like Connell, she assumes that gender is constructed by society around the reproductive capacities of bodies:
The crucial distinction we need to make is between male and female reproductive capacity and the masculine and feminine subject positionings that have been usually made available on the basis of that reproductive capacity. (Davies 1989: 12)
Like Connell, she believes that there are versions of masculinity and femininity, which we can think of as different discourses of gender:
Of course taking oneself up as a boy or a girl is not a unitary process. How one ‘does’ masculinity or femininity with one’s parents, say, may differ profoundly from how one ‘does’ masculinity with one’s friends, or from one friend to another. (Davies 1989: 2)
In her research she substantiates this by looking at the way different girls in preschools took up different versions of femininity – the rough tough princesses, the sirens, the home girls etc. These are different feminine positionings (Davies 1989: 118). She has a similar analysis of boys who take up the superhero or intellectual positions and discourses. Drawing on early formulations by Connell, she adopts the concept of “hegemonic masculinity” within this poststructuralist meta-theoretical context:
Children as well as adults are all members of a society which celebrates hegemonic (dominant, powerful) masculinity. Hegemonic masculinity is an idea of masculinity (as well as something practiced by men) … (Davies 1989: 14)
Overlaps and mapping
What I draw from all this is that Flyvbjerg and others exaggerate the extent to which the social sciences have been fractured by the poststructuralist turn. To begin with, it is certainly true that the claims of determinist poststructuralism imply a radical break with the social science that Weedon and other poststructuralists call “liberal humanism”. Yet these determinist positions are not even maintained consistently by poststructuralist authors or by Foucault himself. I argued a similar case in relation to the relativism also claimed for poststructuralism in the previous chapter. When we turn to the interpretation of poststructuralism that I have called agentic poststructuralism, we find that it is not that hard to map some of its new and interesting insights and tools onto the terrain of supposedly “humanist” social science, especially that formulated in recent years. I have explored this idea by looking at the ways in which the concepts of “ideology” and “hegemonic masculinity” can be mapped within a poststructuralist framework and vice versa. In the next chapter I will look at two examples of typical social science research and talk about the kinds of explanatory concepts and categories, which are being used. I will finish up the chapter by arguing that accounts of social change, whether humanist or poststructuralist often refer by implication to human nature, albeit a human nature located within specific historical settings. | 
