REPRESENTATION

Cultural Representations and Signifying Practices

Edited by
STUART HALL

CHAPTER ONE - THE WORK OF REPRESENTATION

1 Representation, meaning and language
In this chapter we will be concentrating on one of the Keys processes in the ‘cultural
circuit’ (see Du Gay, Hall ET AL., 1997, and the Introduction to this volume) – the
practices of representation. The aim of this chapter is to introduce you to this topic, and
to explain what it is about and why we give it such importance in cultural studies.
The concept of representation has come to occupy a new and important place in the
study of culture. Representation connects meaning and language to culture. But what
exactly do people mean by it? What does representation have to do with culture and
meaning? One common-sense usage of the term is as follows: ‘Representation means
using language to say something meaningful about, or to represent, the world
meaningfully, to other people.’ You may well ask, ‘Is that all?’ Well, yes and no.
Representation is an essential part of the process by which meaning is produced and
exchanged between members of a culture. It does involve the use of language, of signs
and images which stand for or represent things. But this is a far from simple or
straightforward process, as you will soon discover.
How does the concept of representation connect meaning and language to culture? In
order to explore this connection further, we will look at a number of different theories
about how language is used to represent the world. Here we will be drawing a
distinction between three different accounts or theories: the reflective, the intentional
and the constructionist approaches to representation. Does language simply reflect a
meaning which already exists out there in the world of objects, people and events
(reflective)? Does language express only what the speaker or writer or painter wants to
say, his or her personally intended meaning (intentional)? Or is meaning constructed in
and through language (constructionist)? You will learn more in a moment about these
three approaches.
Most of the chapter will be spent exploring the constructionist approach, because it is
this perspective which has had the most significant impact on cultural studies in recent
years. This chapter chooses to examine two major variants or models of the
constructionist approach – the semiotic approach, greatly influenced by the great Swiss
linguist, Ferdinand de Saussure, and the discursive approach, associated with the French
philosopher and historian, Michel Foucault. Later chapters in this book will take up
these two theories again, among others, so you will have an opportunity to consolidate
your understanding of them, and to apply them to different areas of analysis. Other
chapters will introduce theoretical paradigms which apply constructionist approaches in

1
different ways to that of semiotics and Foucault. All, however, put in question the very
nature of representation. We turn to this question first.

1.1 Making meaning, representing things

What does the word representation really mean, in this context? What does the process
of representation involve? How does representation work?
To put it briefly, representation is the production of meaning through language. The
Shorter Oxford English Dictionary suggests two relevant meanings for the word:
1 To represent something is to describe or depict it, to call it up in the mind by
description or portrayal or imagination; to place a likeness of it before us in
our mind or in the senses; as, for example, in the sentence, ‘This picture
represents the murder of Abel by Cain.’
2 To represent also means to symbolize, stand for, to be a specimen of, or to
substitute for; as in the sentence, ‘In Christianity, the cross represents the
suffering and the crucifixion of Christ.’
The figures in the painting stand in the place of, and at the same time, stand for the
story of Cain and Abel. Likewise, the cross simply consists of two wooden planks
nailed together; but in the context of Christian belief and teaching, it takes on,
symbolizes or comes to stand for a wider set of meanings about the crucifixion of the
Son of God, and this is a concept we can put into words and pictures.
This is how you give meaning to things through language. This is how you ‘make sense
of’ the world of people, objects and events, and how you are able to express a complex
through about those things to other people, or communicate about them through
language in ways which other people are able to understand.

Why do we have to go through this complex process to represent our thoughts? If you
put down a glass you are holding and walk out of the room, you can still think about the
glass, even though it is no longer physically there. Actually, you can’t think with a
glass. You can only think with the concept of the glass. As the linguists are fond of
saying, ‘Dogs bark. But the concept of “dog” cannot bark or bite.’ You can’t speak with
the actual glass, either. You can only speak with the word for glass – GLASS – which is
the linguistic sign which we use in English to refer to objects which you drink water out
of. This is where representation comes in. Representation is the production of the
meaning of the concepts in our minds through language. It is the link between concepts
and language which enables us to refer to either the ‘real’ world of objects, people or
events, or indeed to imaginary worlds of fictional objects, people and events.
So there are two processes, two systems of representation, involved. First, there is the
‘system’ by which all sort of objects, people and events are correlated with a set of
concepts or mental representations which we carry around in our heads. Without them,
we could not interpret the world meaningfully at all. In the first place, then, meaning
depends on the system of concepts and images formed in our thoughts which can stand
for or ‘represent’ the world, enabling us to refer to things both inside and outside our
heads.

2
Before we move on to look at the second ‘system of representation’, we should observe
that what we have just said is a very simple version of a rather complex process. It is
simple enough to see how we might form concepts for things we can perceive – people
or materials objects, like chairs, tables and desks. But we also form concepts of rather
obscure and abstract things, which we can’t in any simple way see, feel or touch. Think,
for example, of our concepts of war, or death, or friendship or love. And, as we have
remarked, we also form concepts about things we never have seen, and possibly can’t or
won’t ever see, and about people and places we have plainly made up. We may have a
clear concept of, say, angels, mermaids, God, the Devil, or of Heaven and Hell, or of
Middlemarch (the fictional provincial town in George Eliot’s novel), or Elizabeth (the
heroine of Jane Austen’s Pride and Prejudice).
We have called this a ‘system of representation’. That is because it consists, not of
individual concepts, but of different ways of organizing, clustering, arranging and
classifying concepts, and of establishing complex relations between them. For example,
we use the principles of similarity and difference to establish relationships between
concepts or to distinguish them from one another. Thus I have an idea that in some
respects birds are like planes in the sky, based on the fact that they are similar because
they both fly – but I also have an idea that in other respects they are different, because
one is part of nature whilst the other is man-made. This mixing and matching of
relations between concepts to form complex ideas and thoughts is possible because our
concepts are arranged into different classifying systems. In this example, the first is
based on a distinction between flying/not flying and the second is based on the
distinction between natural/man made. There are other principles of organization like
this at work in all conceptual systems: for example, classifying according to sequence –
which concept follows which – or causality – what causes what – and so on. The point
here is that we are talking about, not just a random collection of concepts, but concepts
organized, arranged and classified into complex relations with one another. That is what
our conceptual system actually is like. However, this does not undermine the basic
point. Meaning depends on the relationship between things in the world – people,
objects and events, real or fictional – and the conceptual system, which can operate as
mental representations of them.
Now it could be the case that the conceptual map which I carry around in my head is
totally different from yours, in which case you and I would interpret or make sense of
the world in totally different ways. We would be incapable of sharing our thoughts or
expressing ideas about the world to each other. In fact, each of us probably does
understand and interpret the world in a unique and individual way. However, we are
able to communicate because we share broadly the same conceptual maps and thus
make sense of or interpret the world in roughly similar ways. That is indeed what it
means when we say we ‘belong to the same culture’. Because we interpret the world in
roughly similar ways, we are able to build up a shared culture of meanings and thus
construct a social world which we inhabit together. That is why ‘culture’ is sometimes
defined in terms of ‘shared meanings or shared conceptual maps’ (see du Gay, Hall et
al., 1997).
However, a shared conceptual map is not enough. We must also be able to represent or
exchange meanings and concepts, and we can only do that when we also have access to
a shared language. Language is therefore the second system of representation involved
in the overall process of constructing meaning. Our shared conceptual map must be
translated into a common language, so that we can correlate our ideas with certain

3
written words, spoken sounds or visual images. The general term we use for words,
sounds or images which carry meaning is signs. These signs stand for or represent the
concepts and the conceptual relations between them which we carry around in our heads
and together they make up the meaning-systems of our culture.
Signs are organized into languages and it is the existence of common languages which
enable us to translate out thoughts (concepts) into words, sounds or images, and then to
use these, operating as a language, to express meanings and communicate thoughts to
other people. Remember that the term ‘language’ is being used here in a very broad and
inclusive way. The writing system or the spoken system of a particular language are
both obviously ‘languages’. But so are visual images, whether produced by hand,
mechanical, electronic, digital or some other means, when they are used to express
meaning. And so are other things which aren’t ‘linguistic’ in any ordinary sense: the
‘language’ of facial expressions or of gesture, for example, or the ‘language’ of fashion,
of clothes, or of traffic lights. Even music is a ‘language’, with complex relations
between sounds and chords, though it is a very special case since it can’t easily be used
to reference actual things or objects in the world (a point further elaborated in du Gay,
ed., 1997, and Mackay, ed., 1997). Any sound, word, image or object which functions
as a sign, and is organized with other signs into a system which is capable of carrying
and expressing meaning is, from this point of view, ‘a language’. It is in this sense that
the model of meaning which I have been analyzing here is often described as a
‘linguistic’ one; and that all the theories of meaning which follow this basic model are
described as belonging to ‘the linguistic turn’ in the social sciences and cultural studies.
At the heart of the meaning process in culture, then, are two related 'systems of
representation’. The first enables us to give meaning to the world by constructing a set
of correspondences or a chain of equivalences between things - people, objects, events,
abstract ideas, etc. – and our system of concepts, our conceptual maps. The second
depends on constructing a set of correspondences between our conceptual map and a set
of signs, arranged or organized into various languages which stand for or represent
those concepts. The relation between 'things', concepts and signs lies at the heart of the
production of meaning in language. The process which links these three elements
together is what we call ‘representation’.

I.2 Language and representation

Just as people who belong to the same culture must share a broadly similar conceptual
map, so they must also share the same way of interpreting the signs of language, for
only in this way can meanings be effectively exchanged between people. But how do
we know which concept stands for which thing? Or which word effectively represents
which concept? How do I know which sounds or images will carry, through language,
the meaning of my concepts and what I want to say with them to you? This may seems
relatively simple in the case of visual signs, because the drawing, painting, camera or
TV image of a sheep bears a resemblance to the animal with a woolly coat grazing in a
field to which I want to refer. Even so, we need to remind ourselves that a drawn or
painted or digital version of a sheep is not exactly like a 'real' sheep. For one thing, most
images are in two dimensions whereas the 'real' sheep exists in three dimensions.
Visual signs and images, even when they bear a close resemblance to the things to
which they refer, are still signs: they carry meaning and thus have to be interpreted. In

4
order to interpret them, we must have access to the two systems of representation
discussed earlier: to a conceptual map which correlates the sheep in the field with the
concept of a ‘sheep’; and a language system which in visual language, bears some
resemblance to the real thing or 'looks like it' in some way. This argument is clearest if
we think of a cartoon drawing or an abstract painting of a 'sheep', where we need a very
sophisticated conceptual and shared linguistic system to be certain that we are all
‘reading’ the sign in the same way. Even then we may find ourselves wondering
whether it really is a picture of a sheep at all. As the relationship between the sign and
its referent becomes less clear-cut, the meaning begins to slip and slide away from us
into uncertainty. Meaning is no longer transparently passing from one person to
another…

FIGURE 1.1
William Holman
Hunt, Our English
Coasts (‘Strayed
Sheep’), 1852
So, even in the case of visual language, where the relationship between the concept and
the signs seems fairly straightforward, the matter is far from simple. It is even more
difficult with written or spoken language, where words don’t look or sound anything
like the things to which they refer. In part, this is because there are different kinds of
signs. Visual signs are what are called iconic signs. That is, they bear, in their form, a
certain resemblance to the object,
person or event to which they refer. A
photograph of a tree reproduces some
of the actual conditions of our visual
perception in the visual sign. Written or
spoken signs, on the other hand, are
what is called indexical.

FIGURE 1.2
Q: When is a sheep not a sheep?
A: When it’s a work of art.
(Damien Hirst, Away from the Flock,
1994).
They bear no obvious relationship at all to the things to which they refer. The letters
T,R,E,E, do not look anything like trees in the Nature, nor does the word ‘tree’ in
English sound like ‘real’ trees (if indeed they make any sound at all!). The relationship

5
in these systems of representation between the sign, the concept and the object to which
they might be used to refer is entirely arbitrary. By ‘arbitrary’ we mean that in principle
any collection of letters or any sound in any order would do the trick equally well. Trees
would not mind if we used the SEERT – ‘trees’ written backwards – to represent the
concept of them. This is clear from the fact that, in French, quite different letters and
quite different sound is used to refer to what, to all appearances, is the same thing – a
‘real’ tree – and, as far we can tell, to the same concept – a large plant that grows in
nature. The French and English seem to be using the same concept. But the concept
which in English is represented by the word, TREE, is represented in French by the
word, ARBRE.

I.3 Sharing the codes

The question, then, is: how do people who belong to the same culture, who share the
same conceptual map and who speak or write the same language (English) know that
the arbitrary combination of letters and sounds that makes up the word, TREE, will
stand for or represent the concept ‘a large plant that grows in nature’? One possibility
would be that the objects in the world themselves embody and fix in some way their
‘true’ meaning. But it is not at all clear that real trees know that they are trees, and even
less clear that they know that the word in English which represents the concept of
themselves is written TREE whereas in French it is written ARBRE! As far as they
concerned, it could just as well be written COW or VACHE or indeed XYZ. The
meaning is not the object or person or thing, nor is it in the word. It is we who fix the
meaning so firmly that, after a while, it comes to seem natural and inevitable. The
meaning is constructed by the system of representation. It is constructed and fixed by
the code, which sets up the correlation between our conceptual system and our language
system in such a way that, every time we think of a tree, the code tells us to use English
word TREE, or the French word ARBRE. The code tells us that, in our culture – that is,
in our conceptual and language codes – the concept ‘tree’ is represented by the letters
T,R,E,E, arranged in a certain sequence, just as in Morse code, the signs for V (which in
World War II Churchill made ‘stand for’ or represent ‘Victory’) is Dot, Dot, Dot, Dash,
and in the ‘language of traffic lights’, Green = Go! And Red = Stop!
One way of thinking about ‘cu1ture’, then, is in terms of these shared conceptual maps,
shared language systems and the codes which govern the relationships of translation
between them. Codes fix the relationships between concepts and signs. They stabilize
meaning within different languages and cultures. They tell us which language to use to
convey which idea. The reverse is also true. Codes tell us which concepts are being
referred to when we hear or read which signs. By arbitrarily fixing the relationships
between our conceptual system and our linguistic systems (remember, ‘linguistic’ in a
broad sense), codes make it possible for us to speak and to hear intelligibly, and
establish the translatability between our concepts and our languages which enables
meaning to pass from speaker to hearer and be effectively communicated within a
culture. This translatability is not given by nature or fixed by the gods. It is the result of
a set of social conventions. It is fixed socially, fixed in culture. English or French or
Hindi speakers have, over time, and without conscious decision or choice, come to an
unwritten agreement, a sort of unwritten cultural covenant that, in their various
languages, certain signs will stand for or represent certain concepts. This is what
children learn, and how they become, not simply biological individuals but cultural
subjects. They learn the system and conventions of representation, the codes of their

6
language and culture, which equip them with cultural ‘know-how’ enabling them to
function as culturally competent subjects. Not because such knowledge is imprinted in
their genes, but because they learn its conventions and so gradually become ‘cultured
persons’ – i.e. members of their culture. They unconsciously internalize the codes which
allow them to express certain concepts and ideas through their systems of representation
– writing, speech, gesture, visualization, and so on – and to interpret ideas which are
communicated to them using the same systems.
You may find it easier to understand, now, why meaning, language and representation
are such critical elements in the study of culture. To belong to a culture is to belong to
roughly the same conceptual and linguistic universe, to know how concepts and ideas
translate into different languages, and how language can be interpret to refer to or
reference the world. To share these things is to see the world from within the same
conceptual map and to make sense of it through the same language systems. Early
anthropologists of language, like Sapir and Whorf, took this insight to its logical
extreme when they argued that we are all, as it were, locked into our cultural
perspectives or ‘mind-sets’, and that language is the best clue we have to that
conceptual universe. This observation, when applied to all human cultures, lies at the
root of what, today, we may think of as cultural or linguistic relativism.
One implication of this argument about cultural codes is that, if meaning is the result,
not of something fixed out there, in nature, but of our social, cultural and linguistic
conventions, then meaning can never be finally fixed. We can all ‘agree’ to allow words
to carry somewhat different meanings – as we have for example, with the word ‘gay’, or
the use, by young people, of the word ‘wicked!’ as a term or approval. Of course, there
must be some fixing of meaning in language, or we would never be able to understand
one another. We can’t get up one morning and suddenly decide to represent the concept
of a ‘tree’ with the letters or the word VYXZ, and expect people to follow what we are
saying. On the other hand, there is no absolute or final fixing of meaning. Social and
linguistic conventions do change over time. In the language of modern managerialism,
what we used to call ‘students’, ‘clients’, ‘patients’ and ‘passengers’ have all become
‘customers’. Linguistic codes vary significantly between one language and another.
Many cultures do not have words for concepts which are normal and widely acceptable
to us. Words constantly go out of common usage, and new phrases are coined: think for
example, of the use of ‘down-sizing’ to represent the process of firms laying people off
work. Even when the actual words remain stable, their connotations shift or they acquire
a different nuance. The problem is especially acute in translation. For example, does the
difference in English between know and understand correspond exactly to and capture
exactly the same conceptual distinction as the French make between savoir and
connaitre? Perhaps; but can we be sure?
The main point is that meaning does not inhere in things, in the world. It is constructed,
produced. It is the result of a signifying practice – a practice that produces meaning, that
makes things mean.

I.4 Theories of representation

There are broadly speaking three approaches to explaining how representation of
meaning through language works. We may call these the reflective, the intentional and
the constructionist or constructivist approaches. You might think of each as an attempt

7
to answer the questions, ‘where do meanings come from?’ and ‘how can we tell the
“true” meaning of a word or image’?
In the reflective approach, meaning is thought to lie in the object, person, idea or event
in the real world, and language functions like a mirror, to reflect the true meaning as it
already exists in the world. As the poet Gertrude Stein once said, ‘A rose is a rose is a
rose’. In the fourth century BC, the Greeks used the notion of mimesis to explain how
language, even drawing and painting, mirrored or imitated Nature; they thought of
Homer’s great poem, The Iliad, as ‘imitating’ a heroic series of events. So the theory
which says that language works by simply reflecting or imitating the truth that is
already there and fixed in the world, is sometimes called ‘mimetic’.
Of course there is a certain obvious truth to mimetic theories of representation and
language. As we’ve pointed out, visual signs do bear some relationship to the shape and
texture of the objects which they represent. But, as was also pointed out earlier, a two-
dimensional visual image of a rose is a sign – it should not be confused with the real
plant with thorns and blooms growing in the garden. Remember also that there are many
words, sounds and images which we fully well understand but which are entirely
fictional or fantasy and refer to worlds which are wholly imaginary – including, many
people now think, most of The Iliad! Of course, I can use the word ‘rose’ to refer to
real, actual plants growing in a garden, as we have said before. But this is because I
know the code which links the concept with a particular word or image. I cannot think
or speak or draw with an actual rose. And if someone says to me that there is no such
word as ‘rose’ for a plant in her culture, the actual plant in the garden cannot resolve the
failure of communication between us. Within the conventions of the different language
codes we are using, we are both right – and for us to understand each other, one of us
must learn the code linking the flower with the word for it in the other’s culture.

The second approach to meaning in representation argues the opposite case. It holds that
it is the speaker, the author, who imposes his or her unique meaning on the world
through language. Words mean what the author intends they should mean. This is the
intentional approach. Again, there is some point to this argument since we all, as
individuals, do use language to convey or communicate things which are special or
unique to us, to our way of seeing the world. However, as a general theory of
representation through language, the intentional approach is also flawed. We cannot be
the sole or unique source of meanings in language, since that would mean that we could
express ourselves in entirely private languages. But the essence of language is
communication and that, in turn, depends on shared linguistic conventions and shared
codes. Language can never be wholly a private game. Our private intended meanings,
however personal to us, have to enter into the rules, codes and conventions of language
to be shared and understood. Language is a social system through and through. This
means that our private thoughts have to negotiate with all the other meanings for words
or images which have been stored in language which our use of the language system
will inevitably trigger into action.
The third approach recognizes this public, social character of language. It acknowledges
that neither things in themselves nor the individual users of language can fix meaning in
language. Things don’t mean: we construct meaning, using representational systems –
concepts and signs. Hence it is called the constructivist or constructionist approach to
meaning in language. According to this approach, we must not confuse the material
world, where things and people exist, and the symbolic practices and processes through

8
which representation, meaning and language operate. Constructivists do not deny the
existence of the material world. However, it is not the material world which conveys
meaning: it is the language system or whatever system we are using to represent our
concepts. It is social actors who use the conceptual systems of their culture and the
linguistic and other representational systems to construct meaning, to make the world
meaningful and to communicate about that world meaningfully to others.
Of course, signs may also have a material dimension. Representational systems consist
of the actual sounds we make with our vocal chords, the images we make on light-
sensitive paper with cameras, the marks we make with paint on canvas, the digital
impulses we transmit electronically. Representation is a practice, a kind of ‘work’,
which uses material objects and effects. But the meaning depends, not on material
quality of the sign, but on its symbolic function. It is because a particular sound or word
stands for, symbolizes or represents a concept that it can function, in language, as a sign
and convey meaning – or, as the constructionists say, signify (sign-i-fy).

I.5 The language of traffic lights

The simplest example of this point, which is critical for an understanding of how
languages function as representational systems, is the famous traffic lights example. A
traffic light is a machine which produces different coloured lights in sequence. The
effect of light of different wavelengths on the eye – which is a natural and material
phenomenon – produces the sensation of different colours. Now these things certainly
do exist in the material world. But it is our culture which breaks the spectrum of light
into different colours, distinguishes them from one another and attaches names – Red,
Green, Yellow, Blue – to them. We use a way of classifying the colour spectrum to
create colours which are different from one another. We represent or symbolize the
different colours and classify them according to different colour-concepts. This is the
conceptual colour system of our culture. We say ‘our culture’ because, of course, other
cultures may divide the colour spectrum differently. What’s more, they certainly use
different actual words or letters to identify different colours: what we call ‘red’, the
French call ‘rouge’ and so on. This is the linguistic code – the one which correlates
certain words (signs) with certain colours (concepts), and thus enables us to
communicate about colours to other people, using ‘the language of colours’.

But how do we use this representational or symbolic system to regulate the traffic?
Colours do not have any ‘true’ or fixed meaning in that sense. Red does not mean ‘Stop’
in nature, any more than Green means ‘Go’. In other settings, Red may stand for,
symbolize or represent ‘Blood’ or ‘Danger’ or ‘Communism’; and Green may represent
‘Ireland’ or ‘The Countryside’ or ‘Environmentalism’. Even these meaning can change.
In the ‘language of electric plugs’, Red used to mean ‘the connection with the positive
charge’ but this was arbitrarily and without explanation changed to brown! But then for
many years the producers of plugs had to attach a slip of paper telling people that the
code or convention had changed, otherwise how would they know? Red and Green
work in the language of traffic lights because ‘Stop’ and ‘Go’ are the meanings which
have been assigned to them in our culture by the code or conventions governing this
language, and this code is widely known and almost universally obeyed in our culture
and cultures like ours – though we can well imagine other cultures which did not
possess the code, in which this language would be a complete mystery.

9
Let us stay with the example for a moment, to explore a little further how, according to
the constructionist approach to representation, colours and the ‘language of traffic
lights’ work as a signifying or representational system. Recall the two representational
systems we spoke of earlier. First, there is the conceptual map of colours in our culture
– the way colours are distinguished from one another, classified and arranged in our
mental universe. Secondly, there are the ways words or images are correlated with
colours in our language – our linguistic colour-codes. Actually, of course, a language of
colours consists of more than just the individual words for different points on the colour
spectrum. It also depends on how they function in relation to one another – the sorts of
things which are governed by grammar and syntax in written or spoken languages,
which allow us to express rather complex ideas. In the language of traffic lights, it is the
sequence and position of the colours, as well as the colours themselves, which enable
them to carry meaning and thus function as signs.
Does it matter which colours we use? No, the constructionists argue. This is because
what signifies is not the colours themselves but (a) the fact that they are different and
can be distinguished from one another; and (b) the fact that they are organized into a
particular sequence – Red followed by Green, with sometimes a warning Amber in
between which says, in effect, ‘Get Ready! Lights about to change.’ Constructionists
put this point in the following way. What signifies, what carries meaning – they argue –
is not each colour in itself nor even the concept or word for it. It is the difference
between Red and Green which signifies. This is a very important principle, in general,
about representation and meaning, and we shall return to it on more than one occasion
in the chapters which follow. Think about it in these terms. If you couldn’t differentiate
between Red and Green, you couldn’t use one to mean ‘Stop’ and the other to mean
‘Go’. In the same way, it is only the difference between the letters P and T which enable
the word SHEEP to be linked, in the English language code, to the concept of ‘the
animal with four legs and a woolly coat’, and the word SHEET to ‘the material we use
to cover ourselves in bed at night’.
In principle, any combination of colours – like any collection of letters in written
language or of sounds in spoken language – would do, provided they are sufficiently
different not to be confused. Constructionists express this idea by saying that all signs
are ‘arbitrary’. ‘Arbitrary’ means that there is no natural relationship between the sign
and its meaning or concept. Since Red only means ‘Stop’ because that is how the code
works, in principle any colour would do, including Green. It is the code that fixes the
meaning, not the colour itself. This also has wider implications for the theory of
representation and meaning in language. It means that signs themselves cannot fix
meaning. Instead, meaning depends on the relation between a sign and a concept which
is fixed by a code. Meaning, the constructionists would say, is ‘relational’.
Provided the code tells us clearly how to read or interpret each colour, and everyone
agrees to interpret them is this way, any colour will do. These are just colours, just as
the word SHEEP is just a jumble of letters. In French the same animal is referred to
using the very different linguistic sign MOUTON. Signs are arbitrary. Their meanings
are fixed by codes.
As we said earlier, traffic lights are machines, and colours are the material effect of
light-waves on the retina of the eye. But objects – things – can also function as signs,
provided they have been assigned a concept and meaning within our cultural and
linguistic codes. As signs, they work symbolically – they represent concepts, and

10
signify. Their effects, however, are felt in the material and social world. Red and Green
function in the language of traffic lights as signs, but they have real material and social
effects. They regulate the social behaviour of drivers and, without them, there would be
many more traffic accidents at road intersections.

I.6 Summary
We have come a long way in exploring the nature of representation. It is time to
summarize what we have learned about the constructionist approach to representation
through language.

Representation is the production of meaning through language. In representation,

constructionists argue, we use signs, organized into languages of different kinds, to
communicate meaningfully with others. Languages can use signs to symbolize, stand
for or reference objects, people and events in the so-called ‘real’ world. But they can
also reference imaginary things and fantasy worlds or abstract ideas which are not in
any obvious sense part of our material world. There is no simple relationship of
reflection, imitation or one-to-one correspondence between language and the real world.
The world is not accurately or otherwise reflected in the mirror of language. Language
does not work like a mirror. Meaning is produced within language, in and through
various representational systems which, for convenience, we call ‘languages’. Meaning
is produced by the practice, the ‘work’, of representation. It is constructed through
signifying – i.e. meaning-producing – practices.
How does this take place? In fact, it depends on two different but related systems of
representation. First, the concepts which are formed in the mind function as a system of
mental representation which classifies and organizes the world into meaningful
categories. If we have a concept for something, we can say we know its ‘meaning’. But
we cannot communicate this meaning without a second system of representation, a
language. Language consists of signs organized into various relationships. But signs can
only convey meaning if we possess codes which allow us to translate our concepts into
language – and vice versa. These codes are crucial for meaning and representation.
They do not exist in nature but are the result of social conventions. They are a crucial
part of our culture – our shared ‘maps of meaning’ – which we learn and unconsciously
internalize as we become members of our culture. This constructionist approach to
language thus introduces the symbolic domain of life, where words and things function
as signs, into the very heart of social life itself.

2 Saussure’s legacy
The social constructionist view of language and representation which we have been
discussing owes a great deal to the work and influence of the Swiss linguist, Saussure,
who was born in Geneva in 1857, did much of his work in Paris, and died in 1913. He is
known as the ‘father of modern linguistics’. For our purposes, his importance lies, not in
his detailed work in linguistics, but in his general view of representation and the way his
model of language shaped the semiotic approach to the problem of representation in a
wide variety of cultural fields. You will recognize much about Saussure’s thinking from
what we have already said about the constructionist approach.

11
For Saussure, according to Jonathan Culler (1976, p. 19), the production of meaning
depends on language: ‘Language is a system of signs.’ Sounds, images, written words,
paintings, photographs, etc. function as signs within language ‘only when they serve to
express or communicate ideas … [To] communicate ideas, they must be part of a
system of conventions …’ (ibid.). Material objects can function as signs and
communicate meaning too, as we saw from the ‘language of traffic lights’ example. In
an important move, Saussure analysed the sign into two further elements. There was, he
argued, the form (the actual word, image, photo, etc.), and there was the idea or concept
in your head with which the form was associated. Saussure called the first element, the
signifier, and the second element – the corresponding concept it triggered off in your
head – the signified. Every time you hear or read or see the signifier (e.g. the word or
image of a Walkman, for example), it correlates with the signified (the concept of a
portable cassette-player in your head). Both are required to produce meaning but it is
the relation between them, fixed by our cultural and linguistic codes, which sustains
representation. Thus ‘the sign is the union of a form which signifies (signifier) … and
an idea signified (signified). Though we may speak … as if they are separate entities,
they exist only as components of the sign … (which is) the central fact of language’
(Culler, 1976, p. 19).
Saussure also insisted on what in section 1 we called the arbitrary nature of the sign:
‘There is no natural or inevitable link between the signifier and the signified’ (ibid.).
Signs do not possess a fixed or essential meaning. What signifies, according to
Saussure, is not RED or the essence of ‘red-ness’, but the difference between RED and
GREEN. Signs, Saussure argued ‘are members of a system and are defined in relation to
the other members of that system’. For example, it is hard to define the meaning of
FATHER except in relation to, and in terms of its difference from, other kinship terms,
like MOTHER, DAUGHTER, SON and so on.
This marking of difference within language is fundamental to the production of
meaning, according to Saussure. Even at a simple level (to repeat an earlier example),
we must be able to distinguish, within language, between SHEEP and SHEET, before
we can link one of those words to the concept of an animal that produces wool, and the
other to the concept of a cloth that covers a bed. The simplest way of marking
difference is, of course, by means of a binary opposition – in this example, all the letters
are the same except P and T. Similarly, the meaning of a concept or word is often
defined in relation to its direct opposite – as in night/day. Later critics of Saussure were
to observe that binaries (e.g. black/white) are only one, rather simplistic, way of
establishing difference. As well as the stark difference between black and white, there
also the many other, subtler differences between black and dark grey, dark grey and
light grey, grey and cream and off-white and brilliant white, just as there are between
night, dawn, daylight, noon, dusk, and so on. However, his attention to binary
oppositions brought Saussure to the revolutionary proposition that a language consists
of signifiers, but in order to produce meaning, the signifiers have to be organized into ‘a
system of differences’. It is the differences between signifiers which signify.
Furthermore, the relation between the signifier and the signified, which is fixed by our
cultural codes, is not – Saussure argued – permanently fixed. Words shift their
meanings. The concepts (signifieds) to which they refer also change, historically, and
every shift alters the conceptual map of the culture, leading different cultures, at
different historical moments, to classify and think about the word differently. For many
centuries, western societies have associated the word BLACK with everything that is

12
dark, evil, forbidding, devilish, dangerous and sinful. And yet, think of how the
perception of black people in America in the 1960 changed after the phrase “Black is
Beautiful’ became a popular slogan – where the signifier, BLACK, was made to signify
the exact opposite meaning (signified) to its previous associations. In Saussure’s terms,
‘Language sets up an arbitrary relation between signifiers of its own choosing on the
one hand, and signifieds of its own choosing on the other. Not only does each language
produce a different set of signifiers, articulating and dividing the continuum of sound
(or writing or drawing or photography) in a distinctive way; each language produces a
different set of signifieds; it has a distinctive and thus arbitrary way of organizing the
world into concepts and categories’ (Culler, 1976, p. 23).
The implications of this argument are very far-reaching for a theory of representation
and for our understanding of culture. If the relationship between a signifier and its
signified is the result of a system of social conventions specific to each society and to
specific historical moments – then all meanings are produced within history and culture.
They can never be finally fixed but are always subject to change, both from one cultural
context and from one period to another. There is thus no single, unchanging, universal
‘true meaning’. ‘Because it is arbitrary, the sign is totally subject to history and the
combination at the particular moment of a given signifier and signified is a contingent
result of the historical process’ (Culler, 1976, p. 36). This opens up meaning and
representation, in a radical way, to history and change. It is true that Saussure himself
focused exclusively on the state of the language system at one moment of time rather
than looking at linguistic change over time. However, for our purposes, the important
point is the way this approach to language unfixes meaning, breaking any natural and
inevitable tie between signifier and signified. This opens representation to the constant
‘play’ or slippage of meaning, to the constant production of new meanings, new
interpretations.
However, if meaning changes, historically, and is never finally fixed, then it follows
that ‘taking the meaning’ must involve an active process of interpretation. Meaning
has to be actively ‘read’ or ‘interpreted’. Consequently, there is a necessary and
inevitable imprecision about language. The meaning we take, as viewers, readers or
audiences, is never exactly the meaning which has been given by the speaker or writer
or by other viewers. And since, in order to say something meaningful, we have to ‘enter
language’, where all sorts of older meanings which pre-date us, are already stored from
previous eras, we can never cleanse language completely, screening out all the other,
hidden meanings which might modify or distort what we want to say. For example, we
can’t entirely prevent some of the negative connotations of the word BLACK from
returning to mind when we read a headline like, ‘WEDNESDAY – A BLACK DAY
ON THE STOCK EXCHANGE’, even if this was not intended. There is a constant
sliding of meaning in all interpretation, a margin – something in excess of what we
intend to say – in which other meanings overshadow the statement or the text, where
other associations are awakened to life, giving what we say a different twist. So
interpretation becomes an essential aspect of the process by which meaning is given and
taken. The reader is as important as the writer in the production of meaning. Every
signifier given or encoded with meaning has to be meaningfully interpreted or decoded
by the receiver (Hall, 1980). Signs which have not been intelligibly received and
interpreted are not, in any useful sense, ‘meaningful’.

13
2.1 The social part of language
Saussure divided language into two parts. The first consisted of the general rules and
codes of the linguistic system, which all its users must share, if it is to be of use as a
means of communication. The rules are the principles which we learn when we learn a
language and they enable us to use language to say whatever we want. For example, in
English, the preferred word order is subject-verb-object (‘the cat sat on the mat’),
whereas in Latin, the verb usually comes at the end. Saussure called this underlying
rule-governed structure of language, which enables us to produce well-formed
sentences, the langue (the language system). The second part consisted of the particular
acts of speaking or writing or drawing, which – using the structure and rules of the
langue – are produced by an actual speaker or writer. He called this parole. ‘La langue
is the system of language, the language as a system of forms, whereas parole is actual
speech [or writing], the speech acts which are made possible by the language’ (Culler,
1976, p. 29).
For Saussure, the underlying structure of rules and codes (langue) was the social part of
language, the part which could be studied with the law-like precision of a science
because of its closed, limited nature. It was his preference for studying language at this
level of its ‘deep structure’ which made people call Saussure and his model of language,
structuralist. The second part of language, the individual speech-act or utterance
(parole), he regarded as the ‘surface’ of language. There were an infinite number of
such possible utterances. Hence, parole inevitably lacked those structural properties –
forming a closed and limited set – which would have enabled us to study it
‘scientifically’. What made Saussure’s model appeal to many later scholars was the fact
that the closed, structured character of language at the level of its rules and laws, which
according to Saussure, enabled it to be studied scientifically, was combined with the
capacity to be free and unpredictably creative in our actual speech acts. They believed
he had offered them, at last, a scientific approach to that least scientific object of inquiry
– culture.
In separating the social part of language (langue) from the individual act of
communication (parole), Saussure broke with our common-sense notion of how
language works. Our common-sense intuition is that language comes from within us –
from the individual speaker or writer; that it is this speaking or writing subject who is
the author or originator of meaning. This is what we called, earlier, the intentional
model of representation. But according to Saussure’s schema, each authored statement
only becomes possible because the ‘author’ shares with other language-users the
common rules and codes of the language system – the langue – which allows them to
communicate with each other meaningfully. The author decides what she wants to say.
But she cannot ‘decide’ whether or not to use the rules of language, if she wants to be
understood. We are born into a language, its codes and its meanings. Language is
therefore, for Saussure, a social phenomenon. It cannot be an individual matter because
we cannot make up the rules of language individually, for ourselves. Their source lies in
society, in the culture, in our shared cultural codes, in the language system – not in
nature or in the individual subject.
We will move on in section 3 to consider how the constructionist approach to
representation, and in particular Saussure’s linguistic model, was applied to a wider set
of cultural objects and practices, and evolved into the semiotic method which so

14
influenced the field. First we ought to take account of some of the criticisms levelled at
his position.

2.2 Critique of Saussure’s model

Saussure’s great achievement was to force us to focus on language itself, as a social
fact; on the process of representation itself; on how language actually works and the role
it plays in the production of meaning. In doing so, he saved language from the status of
a mere transparent medium between things and meaning. He showed, instead, that
representation was a practice. However, in his own work, he tended to focus almost
exclusively on the two aspects of the sign – signifier and signified. He gave little or no
attention to how this relation between signifier/signified could serve the purpose of what
earlier we called reference – i.e. referring us to the world of things, people and events
outside language in the ‘real’ world. Later linguists made a distinction between, say, the
meaning of the word BOOK and the use of the word to refer to a specific book lying
before us on the table. The linguist, Charles Sanders Pierce, whilst adopting a similar
approach to Saussure, paid greater attention to the relationship between
signifiers/signifieds and what he called their referents. What Saussure called
signification really involves both meaning and reference, but he focused mainly on the
former.

Another problem is that Saussure tended to focus on the formal aspects of language –
how language actually works. This has the great advantage of making us examine
representation as a practice worthy of detailed study in its own right. It forces us to look
at language for itself, and not just as an empty, transparent, ‘window on the world’.
However, Saussure’s focus on language may have been too exclusive. The attention to
its formal aspects did divert attention away from the more interactive and dialogic
features of language – language as it is actually used, as it functions in actual situations,
in dialogue between different kinds of speakers. It is thus not surprising that, for
Saussure, questions of power in language – for example, between speakers of different
status and positions – did not arise.
As has often been the case, the ‘scientific’ dream which lay behind the structuralism
impulse of his work, though influential in alerting us to certain aspects of how language
works, proved to be illusory. Language is not an object which can be studied with the
law-like precision of a science. Later cultural theorists learned from Saussure’s
‘structuralism’ but abandoned its scientific premise. Language remains rule-governed.
But it is not a ‘closed’ system which can be reduced to its formal elements. Since it is
constantly changing, it is by definition open-ended. Meaning continues to be produced
through language in forms which can never be predicted beforehand and its ‘sliding’, as
we described it above, cannot be halted. Saussure may have been tempted to the former
view because, like a good structuralist, he tended to study the state of the language
system at one moment, as if it had stood still, and he could halt the flow of language-
change. Nevertheless it is the case that many of those who have been most influenced
by Saussure’s radical break with all reflective and intentional models of representation,
have built on his work, not by imitating his scientific and ‘structuralist’ approach, but
by applying his model in a much looser, more open-ended – i.e. ‘post-structuralist’ –
way.

15
2.3 Summary
How far, then, have we come in our discussion of theories of representation? We began
by contrasting three different approaches. The reflective or mimetic approach proposed a
direct and transparent relationship of imitation or reflection between words (signs) and
things. The intentional theory reduced representation to the intentions of its author or
subject. The constructionist theory proposed a complex and mediated relationship
between things in the world, our concepts in thought and language. We have focused at
greatest length on this approach. The correlations between these levels – the material,
the conceptual and the signifying – are governed by our cultural and linguistic codes
and it is this set of interconnections which produces meaning. We then showed how
much this general model of how systems of representation work in the production of
meaning owed to the work of Ferdinand de Saussure. Here, the key point was the link
provided by the codes between the forms of expression used by language (whether
speech, writing, drawing, or other types of representation) – which Saussure called the
signifiers – and the mental concepts associated with them – the signifieds. The
connection between these two systems of representation produced signs; and signs,
organized into languages, produced meanings, and could be used to reference objects,
people and events in the ‘real’ world.

3 From language to culture: linguistics to semiotics

Saussure’s main contribution was to the study of linguistics in a narrow sense.
However, since his death, his theories have been widely deployed, as a foundation for a
general approach to language and meaning, providing a model of representation which
has been applied to a wide range of cultural objects and practices. Saussure himself
foresaw this possibility in his famous lecture-notes, collected posthumously by his
students as the Course in General Linguistics (1960), where he looked forward to ‘A
science that studies the life of signs within society … I shall call it semiology, from the
Greek semeion “signs” …’ (p. 16). This general approach to the study of signs in
culture, and of culture as a sort of ‘language’, which Saussure foreshadowed, is now
generally known by the term semiotics.
The underlying argument behind the semiotic approach is that, since all cultural objects
convey meaning, and all cultural practices depend on meaning, they must make use of
signs; and in so far as they do, they must work like language works, and be amenable to
an analysis which basically makes use of Saussure’s linguistic concepts (e.g. the
signifier/signified and langue/parole distinctions, his idea of underlying codes and
structures, and the arbitrary nature of the sign). Thus, when in his collection of essays,
Mythologies (1972), the French critic, Roland Barthes, studied ‘The world of wrestling’,
‘Soap powders and detergents’, ‘The face of Greta Garbo’ or ‘The Blue Guides to
Europe’, he brought a semiotic approach to bear on ‘reading’ popular culture, treating
these activities and objects as signs, as a language through which meaning is
communicated. For example, most of us would think of a wrestling match as a
competitive game or sport designed for one wrestler to gain victory over an opponent.
Barthes, however, asks, not ‘Who won?’ but ‘What is the meaning of this event?’ He
treats it as a text to be read. He ‘reads’ the exaggerated gestures of wrestlers as a
grandiloquent language of what he calls the pure spectacle of excess.

16
In much the same way, the French anthropologist, Claude Lévi-Strauss, studied the
customs, rituals, totemic objects, designs, myths and folk-tales of so-called ‘primitive’
peoples in Brazil, not by analysing how these things were produced and used in the
context of daily life amongst the Amazonian peoples, but in terms of what they were
trying to ‘say’, what messages about the culture they communicated. He analysed their
meaning, not by interpreting their content, but by looking at the underlying rules and
codes through which such objects or practices produced meaning and, in doing so, he
was making a classic Saussurean or structuralist ‘move’ from the paroles of a culture to
the underlying structure, its langue. To undertake this kind of work, in studying the
meaning of a television programme like Eastenders, for example, we would have to
treat the pictures on the screen as signifiers, and use the code of the television soap
operai as a genre, to discover how each image on the screen made use of these rules to
‘say something’ (signifieds) which the viewer could ‘read’ or interpret within the formal
framework of a particular kind of television narrative (see the discussion and analysis of
TV soap operas in Chapter 6).
In the semiotic approach, not only words and images but objects themselves can
function as signifiers in the production of meaning. Clothes, for example, may have a
simple physical function – to cover the body and protect it from the weather. But
clothes also double up as signs. They construct a meaning and carry a message. An
evening dress may signify ‘elegance’; a bow tie and tails, ‘formality’; jeans and trainers,
‘casual dress’; a certain kind of sweater in the right setting, ‘a long, romantic, autumn
walk in the wood’ (Barthes, 1967). These signs enable clothes to convey meaning and
to function like a language – ‘the language of fashion’. How do they do this?
The clothes themselves are the signifiers. The fashion code in western consumer
cultures like our correlates particular kinds or combinations of clothing with certain
concepts (‘elegance’, ‘formality’, ‘casual-ness’, ‘romance’). These are the signifieds.
This coding converts the clothes into signs, which can then be read as a language. In the
language of fashion, the signifiers are arranged in a certain sequence, in certain relations
to one another. Relations may be of similarity – certain items ‘go together’ (e.g. casual
shoes with jeans). Differences are also marked – no leather belts with evening wear.
Some signs actually create meaning by exploiting ‘difference’: e.g. Doc Marten boots
with flowing long skirt. These bits of clothing ‘say something’ – they convey meaning.
Of course, not everybody reads fashion in same way. There are differences of gender,
age, class, ‘race’. But all those who share the same fashion code will interpret the signs
in roughly the same ways. ‘Oh, jeans don’t look right for that event. It’s a formal
occasion – it demands something more elegant.’
You may have noticed that, in this example, we have moved from the very narrow
linguistic level from which we drew examples in the first section, to a wider, cultural
level. Note, also, that two linked operations are required to complete the representation
process by which meaning is produced. First, we need a basic code which links a
particular piece of material which is cut and sewn in a particular way (signifier) to our
mental concept of it (signified) – say a particular cut of material to our concept of ‘a
dress’ or ‘jeans’. (Remember that only some cultures would ‘read’ the signifier in this
way, or indeed possess the concept of (i.e. have classified clothes into) ‘a dress’, as
different from ‘jeans’.) The combination of signifier and signified is what Saussure
called a sign. Then, having recognized the material as a dress, or as jeans, and produced
a sign, we can progress to a second, wider level, which links these signs to broader,
cultural themes, concepts or meanings – for example, an evening dress to ‘formality’ or

17
‘elegance’, jeans to ‘casualness’. Barthes called the first, descriptive level, the level of
denotation: the second level, that of connotation. Both, of course, require the use of
codes.
Denotation is the simple, basic descriptive level, where consensus is wide and most
people would agree on the meaning (‘dress’, ‘jeans’). At the second level – connotation
– these signifiers which we have been able to ‘decode’ at a simple level by using our
conventional conceptual classifications of dress to read their meaning, enter a wider,
second kind of code – ‘the language of fashion’ – which connects them to broader
themes and meanings, linking them with what, we may call the wider semantic fields of
our culture: ideas of ‘elegance’, ‘formality’, ‘casualness’ and ‘romance’. This second,
wider meaning is no longer a descriptive level of obvious interpretation. Here we are
beginning to interpret the completed signs in terms of the wider realms of social
ideology – the general beliefs, conceptual frameworks and value systems of society.
This second level of signification, Barthes suggests, is more ‘general, global and
diffuse…’. It deals with ‘fragments of an ideology… These signifieds have a very close
communication with culture, knowledge, history and it is through them, so to speak, that
the environmental world [of the culture] invades the system [of representation]’
(Barthes, 1967, pp. 91-2).

3.1 Myth today

In his essay “Myth today’, in Mythologies, Barthes gives on another example which
helps us to see exactly how representation is working at this second, broader cultural
level. Visiting the barber’s one day, Barthes is shown a copy of the French magazine
Paris Match, which has on its cover a picture of ‘a young Negro in a French uniform
saluting with his eyes uplifted, probably fixed on the fold of tricolour’ (the French flag)
(1972b, p. 116). At the first level, to get any meaning at all, we need to decode each of
the signifiers in the image into their appropriate concepts: e.g. a soldier, a uniform, an
arm raised, eyes lifted, a French flag. This yields a set of signs with a simple, literal
message or meaning: a black soldier is giving the French flag a salute (denotation).
However, Barthes argues that this image also has a wider, cultural meaning. If we ask,
‘What is Paris Match telling us by using this picture of black soldier saluting a French
flag?’, Barthes suggests that we may come up with the message: ‘that France is a great
Empire, and that all her sons, without any colour discrimination, faithfully serve under
her flag, and that there is no better answer to the detractors of an alleged colonialism
than the zeal shown by this Negro in serving his so-called oppressors’ (connotation)
(ibid.).

Whatever you think of the actual ‘message’ which Barthes finds, for a proper semiotic
analysis you must be able to outline precisely the different steps by which this broader
meaning has been produced. Barthes argues that here representation takes place through
two separate but linked processes. In the first, the signifiers (the elements of the image)
and the signifieds (the concepts – soldier, flag and so on) unite to form a sign with a
simple denoted message: a black soldier is giving the French flag a salute. At the
second stage, this completed message or sign is linked to a second set of signifieds – a
broad, ideological theme about French colonialism. The first, completed meaning
functions as the signifier in the second stage of the representation process, and when
linked with a wider theme by a reader, yields a second, more elaborate and ideologically
framed message or meaning. Barthes gives this second concept or theme a name – he

18
calls it ‘a purposeful mixture of “French imperiality” and “militariness”’. This, he says,
adds up to a ‘message’ about French colonialism and her faithful Negro soldier-sons.
Barthes calls this second level of signification the level of myth. In this reading, he adds,
‘French imperiality is the very drive behind the myth. The concept reconstitutes a chain
of causes and effects, motives and intentions … Through the concept … a whole new
history … is implanted in the myth … the concept of French imperiality … is again tied
to the totality of the world: to the general history of France, to its colonial adventures, to
its present difficulties’ (Barthes, 1972b, p. 119).
For another example of this two-
stage process of signification, we
can turn now to another of
Barthes’s famous essays.
Barthes suggests that we can read
the Panzani ad as a ‘myth’ by
linking its completed message
(this is a picture of some packets
of pasta, a tin, a sachet, some
tomatoes, onions, peppers, a
mushroom, all emerging from a
half-open string bag) with the
cultural theme or concept of
‘Italianicity’ (or as we would say,
‘Italian-ness’). Then, at the level
of the myth or meta-languageii, the
Panzani ad becomes a message
about the essential meaning of
Italian-ness as a national culture.
Can commodities really become
the signifiers for myths of
nationality? Can you think of ads,
in magazines or television, which
work in the same way, drawing on
the myth of ‘Englishness’? Or
‘Frenchness’? Or ‘American-
ness’? Or ‘Indian-ness’?

FIGURE 1.3 ‘Italian-ness’ and the Panzani ad.

4 Discourse, power and the subject

What the examples above show is that the semiotic approach provides a method for
analysing how visual representations convey meaning. Already, in Roland Barthe’s
work in the 1960s, as we have seen, Saussure’s ‘linguistic’ model is developed through
its application to a much wider field of signs and representations (advertising,
photography, popular culture, travel, fashion, etc.). Also, there is less concern with how
individual words function as signs in language, more about the application of the

19
language model to a much broader set of cultural practices. Saussure held out the
promise that the whole domain of meaning could, at last, be systematically mapped.
Barthes, too, had a ‘method’, but his semiotic approach is much more loosely and
interpretively applied; and, in his later work (for example, The Pleasure of the Text,
1975), he is more concerned with the ‘play’ of meaning and desires across texts than he
is with the attempt to fix meaning by a scientific analysis of language’s rules and laws.
Subsequently, as we observed, the project of a ‘science of meaning’ has appeared
increasingly untenable. Meaning and representation seem to belong irrevocably to the
interpretative side of the human and cultural sciences, whose subject matter – society,
culture, the human subject – is not amenable to a positivistic approach (i.e. one which
seeks to discover scientific laws about society). Later developments have recognized the
necessarily interpretative nature of culture and the fact that interpretations never
produce a final moment of absolute truth. Instead, interpretations are always followed
by other interpretations, in an endless chain. As the French philosopher, Jacques
Derrida, put it, writing always leads to more writing. Difference, he argued, can never
be wholly captured within any binary system (Derrida, 1981). So any notion of a final
meaning is always endlessly put off, deferred. Cultural studies of this interpretative
kind, like other qualitative forms of sociological inquiry, are inevitably caught up in this
‘circle of meaning’.
In the semiotic approach, representation was understood on the basis of the way word
functioned as signs within language. But, for a start, in a culture, meaning often depends
on larger units of analysis – narratives, statements, groups of images, whole discourses
which operate across a variety of texts, areas of knowledge about a subject which have
acquired widespread authority. Semiotics seemed to confine the process of
representation to language, and to treat it as a closed, rather static, system. Subsequent
developments became more concerned with representation as a source for the
production of social knowledge – a more open system, connected in more intimate ways
with social practices and questions of power. In the semiotic approach, the subject was
displaced from the centre of language. Later theorists returned to the question of the
subject, or at least to the empty space which Saussure’s theory had left; without, of
course, putting him/her back in the centre, as the author or source of meaning. Even if
language, in some sense, ‘spoke us’ (as Saussure tended to argue) it was also important
that in certain historical moments, some people had more power to speak about some
subjects than others (male doctors about mad female patients in the late nineteenth
century, for example, to take one of the key examples developed in the work of Michel
Foucault). Models of representation, these critics argued, ought to focus on these
broader issues of knowledge and power.

Foucault used the word ‘representation’ in a narrower sense than we are using it here,
but he is considered to have contributed to a novel and significant general approach to
the problem of representation. What concerned him was the production of knowledge
(rather than just meaning) through what he called discourse (rather than just language).
His project, he said, was to analyse ‘how human beings understand themselves in our
culture’ and how our knowledge about ‘the social, the embodied individual and shared
meanings’ comes to be produced in different periods. With its emphasis on cultural
understanding and shared meanings, you can see that Foucault’s project was still to
some degree indebted to Saussure and Barthes (see Dreyfus and Rabinow, 1982, p.17)
while in other ways departing radically from them. Foucault’s work was much more
historically grounded, more attentive to historical specificities, than the semiotic

20
approach. As he said ‘relations of power, not relations of meaning’ were his main
concern. The particular objects of Foucault’s attention were the various disciplines of
knowledge in the human and social sciences – what he called ‘the subjectifying social
sciences’. These had acquired an increasingly prominent and influential role in modern
culture and were, in many instances, considered to be the discourses which, like religion
in earlier times, could give us the ‘truth’ about knowledge.
We will return to Foucault’s work in some of the subsequent chapters in this book (for
example, Chapter 5). Here, we want to introduce Foucault and the discursive approach
to representation by outlining three of his major ideas: his concept of discourse; the
issue of power and knowledge; and the question of the subject. It might be useful,
however, to start by giving you a general flavor, in Foucault’s graphic (and somewhat
over-stated) terms, of how he saw his project differing from that of the semiotic
approach to representation. He moved away from an approach like that of Saussure and
Barthes, based on ‘the domain of signifying structure’, towards one based on analyzing
what he called ‘relations of force, strategic developments and tactics’:

Here I believe one’s point of reference should not be to the great model of
language (langue) and signs, but to that of war and battle. The history which bears
and determines us has the form of a war rather than that of a language: relations of
power not relations of meaning …
(Foucault, 1980, pp. 114-5)

Rejecting both Hegelian Marxism (what he calls ‘the dialectic’) and semiotics, Foucault
argued that:

Neither the dialectic, as logic of contradictions, nor semiotics, as the structure of

communication, can account for the intrinsic intelligibility of conflicts. ‘Dialectic’
is a way of evading the always open and hazardous reality of conflict by reducing
it to a Hegelian skeleton, and ‘semiology’ is a way of avoiding its violent, bloody
and lethal character by reducing it to the calm Platonic form of language and
dialogue.
(ibid.)

4.1 From language to discourse

The first point to note, then, is the shift of attention in Foucault from ‘language’ to
‘discourse’. He studied not language, but discourse as a system of representation.
Normally, the term ‘discourse’ is used as a linguistic concept. It simply means passages
of connected writing or speech. Michel Foucault, however, gave it a different meaning.
What interested him were the rules and practices that produced meaningful statements
and regulated discourse in different historical periods. By ‘discourse’, Foucault meant ‘a
group of statements which provide a language for talking about – a way of representing
the knowledge about – a particular topic at a particular historical moment. … Discourse
is about the production of knowledge through language. But … since all social practices
entail meaning, and meanings shape and influence what we do – our conduct – all
practices have a discursive aspect’ (Hall, 1992, p. 291). It is important to note that the

21
concept of discourse in this usage is not purely a ‘linguistic’ concept. It is about
language and practice. It attempts to overcome the traditional distinction between what
one says (language) and what one does (practice). Discourse, Foucault argues,
constructs the topic. It defines and produces the objects of our knowledge. It governs
the way that a topic can be meaningfully talked about and reasoned about. It also
influences how ideas are put into practice and used to regulate the conduct of others.
Just as a discourse ‘rules in’ certain ways of talking about a topic, defining an
acceptable and intelligible way to talk, write, or conduct oneself, so also, by definition,
it ‘rules out’, limits and restricts other ways of talking, of conducting ourselves in
relation to the topic or constructing knowledge about it. Discourse, Foucault argued,
never consists of one statement, one text, one action or one source. The same discourse,
characteristic of the way of thinking or the state of knowledge at any one time (what
Foucault called the episteme), will appear across a range of texts, and as forms of
conduct, at a number of different institutional sites within society. However, whenever
these discursive events ‘refer to the same object, share the same style and … support a
strategy … a common institutional, administrative or political drift and pattern’
(Cousins and Hussain, 1984, pp. 84-5), then they are said by Foucault to belong to the
same discursive formation.
Meaning and meaningful practice is therefore constructed within discourse. Like the
semioticians, Foucault was a ‘constructionist’. However, unlike them, he was concerned
with the production of knowledge and meaning, not through language but through
discourse. There were therefore similarities, but also substantive differences between
these two versions.
The idea that ‘discourse produces the objects of knowledge’ and that nothing which is
meaningful exists outside discourse, is at first sight a disconcerting proposition, which
seems to run right against the grain of common-sense thinking. It is worth spending a
moment to explore this idea further. Is Foucault saying – as some of his critics have
charged – that nothing exists outside of discourse? In fact, Foucault does not deny that
things can have a real, material existence in the world. What he does argue is that
‘nothing has any meaning outside of discourse’ (Foucault, 1972). As Laclau and
Mouffeiii put it, ‘we use [the term discourse] to emphasize the fact that every social
configuration is meaningful’ (1990, p.100). The concept of discourse is not about
whether things exist but about where meaning comes from.

The idea that physical things and actions exist, but they only take on meaning and
become objects of knowledge within discourse, is at the heart of the constructionist
theory of meaning and representation. Foucault argues that since we can only have a
knowledge of things if they have a meaning, it is discourse – not the things-in-
themselves – which produces knowledge. Subjects like ‘madness’, ‘punishment’ and
‘sexuality’ only exist meaningfully within the discourses about them. Thus, the study of
the discourses of madness, punishment or sexuality would have to include the following
elements:
1 statements about ‘madness’, ‘punishment’ or ‘sexuality’ which give us a certain
kind of knowledge about these things;
2 the rules which prescribe certain ways of talking about these topics and exclude
other ways – which govern what is ‘sayable’ or ‘thinkable’ about insanity,
punishment or sexuality, at a particular historical moment;

22
3 ‘subjects’ who in some ways personify the discourse – the madman, the hysterical
woman, the criminal, the deviant, the sexually perverse person; with the attributes
we would expect these subjects to have, given the way knowledge about the topic
was constructed at that time;
4 how this knowledge about the topic acquires authority, a sense of embodying the
‘truth’ about it; constituting the ‘truth of the matter’, at a historical moment;

5 the practices within institutions for dealing with the subjects – medical treatment
for the insane, punishment regimes for the guilty, moral discipline for sexually
deviant – whose conduct is being regulated and organized according to those ideas;
6 acknowledgement that a different discourse or episteme will arise at a later
historical moment, supplanting the existing one, opening up a new discursive
formation, and producing, in its turn, new conceptions of ‘madness’ or
‘punishment’ or ‘sexuality’, new discourses with the power and authority, the
‘truth’, to regulate social practices in new ways.

4.2 Historicizing discourse: discursive practices

The main point to get hold of here is the way discourse, representation, knowledge and
‘truth’ are radically historicized by Foucault, in contrast to the rather ahistorical
tendency in semiotics. Things meant something and were ‘true’, he argued, only within
a specific historical context. Foucault did not believe that the same phenomena would
be found across different historical periods. He thought that, in each period, discourse
produced forms of knowledge, objects, subjects and practices of knowledge, which
differed radically from period to period, with no necessary continuity between them.
Thus, for Foucault, for example, mental illness was not an objective fact, which
remained the same in historical periods, and meant the same thing in all cultures. It was
only within a definite discursive formation that the object, ‘madness’, could appear at all
as a meaningful or intelligible construct. It was ‘constituted by all that was said, in all
the statements that named it, divided it up, described it, explained it, traced its
development, indicated its various correlations, judged it, and possibly gave it speech by
articulating, in its name, discourses that were to be taken as its own’ (1972, p.32). And
it was only after a certain definition of ‘madness’ was put into practice, that the
appropriate subject – ‘the madman’ as current medical and psychiatric knowledge
defined ‘him’ – could appear.
Or, take some other examples of discursive practices from his work. There have always
been sexual relations. But ‘sexuality’, as a specific way of talking about, studying and
regulating sexual desire, its secrets and its fantasies, Foucault argued, only appeared in
western societies at a particular historical moment (Foucault, 1978). There may always
have been what we now call homosexual forms of behaviour. But ‘the homosexual’ as a
specific kind of social subject, was produced, and could only make its appearance,
within the moral, legal, medical and psychiatric discourses, practices and institutional
apparatuses of the late nineteenth century, with their particular theories of sexual
perversity (Weeks, 1981, 1985). Similarly, it makes nonsense to talk of the ‘hysterical
woman’ outside of the nineteenth-century view of hysteria as a very widespread female
malady. In The Birth of the Clinic (1973), Foucault charted how ‘in less than half a
century, the medical understanding of disease was transformed’ from a classical notion

23
that disease existed separate from the body, to the modern idea that disease arose within
and could be mapped directly by its course through the human body (McNay, 1994).
This discursive shift changed medical practice. It gave greater importance to the
doctor’s ‘gaze’ which could now ‘read’ the course of disease simply by a powerful look
at what Foucault called ‘the visible body’ of the patient – following the ‘routes … laid
down in accordance with a now familiar geometry … the anatomical atlas’ (Foucault,
1973, pp. 3-4). This greater knowledge increased the doctor’s power of surveillance vis-
à-vis the patient.
Knowledge about and practices around all these subjects, Foucault argued, were
historically and culturally specific. They did not and could not meaningfully exist
outside specific discourses, i.e. outside the ways they were represented in discourse,
produced in knowledge and regulated by the discursive practices and disciplinary
techniques of a particular society and time. Far from accepting the trans-historical
continuities of which historians are so found, Foucault believed that more significant
were the radical breaks, ruptures and discontinuities between one period and another,
between one discursive formation and another.

4.3 From discourse to power/knowledge

In his later work Foucault became even more concerned with how knowledge was put
to work through discursive practices in specific institutional settings to regulate the
conduct of others. He focused on the relationship between knowledge and power, and
how power operated within what he called an institutional apparatus and its
technologies (techniques). Foucault’s conception of the apparatus of punishment, for
example, included a variety of diverse elements, linguistic and non-linguistic –
‘discourses, institutions, architectural arrangements, regulations, laws, administrative
measures, scientific statements, philosophic propositions, morality, philanthropy, etc. …
The apparatus is thus always inscribed in a play of power, but it is also always linked to
certain co-ordinates of knowledge … This is what the apparatus consists in: strategies of
relations of forces supporting and supported by types of knowledge’ (Foucault, 1980b,
pp. 194, 196).

This approach took as one of its key subjects of investigation the relations between
knowledge, power and the body in modern society. It saw knowledge as always
inextricably enmeshed in relations of power because it was always being applied to the
regulation of social conduct in practice (i.e. to particular ‘bodies’). This foregrounding
of the relation between discourse, knowledge and power marked a significant
development in the constructionist approach to representation which we have been
outlining. It rescued representation from the clutches of a purely formal theory and gave
it a historical, practical and ‘worldly’ context of operation.
You may wonder to what extent this concern with discourse, knowledge and power
brought Foucault’s interests closer to those of the classical sociological theories of
ideology, especially Marxism with its concern to identify the class positions and class
interests concealed within particular forms of knowledge. Foucault, indeed, does come
closer to addressing some of these questions about ideology than, perhaps, formal
semiotics did (though Roland Barthes was also concerned with questions of ideology
and myth, as we saw earlier). But Foucault had quite specific and cogent reasons why
he rejected the classical Marxist problematic of ‘ideology’. Marx had argued that, in

24
every epoch, ideas reflect the economic basis of society, and thus the ‘ruling ideas’ are
those of the ruling class which governs a capitalist economy, and correspond to its
dominant interests. Foucault’s main argument against the classical Marxist theory of
ideology was that it tended to reduce all the relation between knowledge and power to a
question of class power and class interests. Foucault did not deny the existence of
classes, but he was strongly opposed to this powerful element of economic or class
reductionism in the Marxist theory of ideology. Secondly, he argued that Marxism
tended to contrast the ‘distortions’ of bourgeois knowledge, against its own claims to
‘truth’ – Marxist science. But Foucault did not believe that any form of thought could
claim an absolute ‘truth’ of this kind, outside the play of discourse. All political and
social forms of thought, he believed, were inevitably caught up in the interplay of
knowledge and power. So, his work rejects the traditional Marxist question, ‘in whose
class interest does language, representation and power operate?’
Later theorists, like the Italian, Antonio Gramsci, who was influenced by Marx but
rejected class reductionism, advanced a definition of ‘ideology’ which is considerably
closer to Foucault’s position, though still too preoccupied with class questions to be
acceptable to him. Gramsci’s notion was that particular social groups struggle in many
different ways, including ideologically, to win the consent of other groups and achieve a
kind of ascendancy in both thought and practice over them. This form of power Gramsci
called hegemony. Hegemony is never permanent, and is not reducible to economic
interests or to a simple class model of society. This has some similarities to Foucault’s
position, though on some key issues they differ radically. (The question of hegemony is
briefly addressed again in Chapter 4.)
What distinguished Foucault’s position on discourse, knowledge and power from the
Marxist theory of class interests and ideological ‘distortion’? Foucault advanced at least
two, radically novel, propositions.
1 Knowledge, power and truth
The first concerns the way Foucault conceived the linkage between knowledge and
power. Hitherto, we have tended to think that power operates in a direct and brutally
repressive fashion, dispensing with polite things like culture and knowledge, though
Gramsci certainly broke with that model of power. Foucault argued that not only is
knowledge always a form of power, but power is implicated in the questions of whether
and in what circumstances knowledge is to be applied or not. This question of the
application and effectiveness of power/knowledge was more important, he thought,
than the question of its ‘truth’.
Knowledge linked to power, not only assumes the authority of ‘the truth’ but has the
power to make itself true. All knowledge, once applied in the real world, has real
effects, and in that sense at least, ‘becomes true’. Knowledge, once used to regulate the
conduct of others, entails constraint, regulation and the disciplining of practices. Thus,
‘There is no power relation without the correlative constitution of a field of knowledge,
nor any knowledge that does not presuppose and constitute at the same time, power
relations’ (Foucault, 1977a, p. 27).
According to Foucault, what we think we ‘know’ in a particular period about, say, crime
has a bearing on how we regulate, control and punish criminals. Knowledge does not
operate in a void. It is put to work, through certain technologies and strategies of
application, in specific situations, historical contexts and institutional regimes. To study

25
punishment, you must study how the combination of discourse and power –
power/knowledge – has produced a certain conception of crime and the criminal, has
had certain real effects both for criminal and for the punisher, and how these have been
set into practice in certain historically specific prison regimes.
This led Foucault to speak, not of the ‘Truth’ of knowledge in the absolute sense – a
Truth which remained so, whatever the period, setting, context – but of a discursive
formation sustaining a regime of truth. Thus, it may or may not be true that single
parenting inevitably leads to delinquency and crime. But if everyone believes it to be so,
and punishes single parents accordingly, this will have real consequences for both
parents and children and will become ‘true’ in terms of its real effects, even if in some
absolute sense it has never been conclusively proven. In the human and social sciences,
Foucault argued:

Truth isn’t outside power. … Truth is a thing of this world; it is produced only by
virtue of multiple forms of constraint. And it induces regular effects of power.
Each society has its regime of truth, its ‘general politics’ of truth; that is, the types
of discourse which it accepts and makes function as true, the mechanisms and
instances which enable one to distinguish true and false statements, the means by
which each is sanctioned … the status of those who are charged with saying what
counts as true.
(Foucault, 1980, p. 131)

2 New conceptions of power

Secondly, Foucault advanced an altogether novel conception of power. We tend to think
of power as always radiating in a single direction – from top to bottom - and coming
from a specific source – the sovereign, the state, the ruling class and so on. For
Foucault, however, power does not ‘function in the form of a chain’ – it circulates. It is
never monopolized by one centre. It ‘is deployed and exercised through a net-like
organization’ (Foucault, 1980, p.98). This suggests that we are all, to some degree,
caught up in its circulation – oppressors and oppressed. It does not radiate downwards,
either from one source or from one place. Power relations permeate all levels of social
existence and are therefore to be found operating at every site of social life – in the
private spheres of the family and sexuality as much as in the public spheres of politics,
the economy and the law. What’s more, power is not only negative, repressing what it
seeks to control. It is also productive. It ‘doesn’t only weigh on us a force that says no,
but … it traverses and produces things, it induces pleasure, forms of knowledge,
produces discourse. It needs to be thought of as a productive network which runs
through the whole social body’ (Foucault, 1980, p. 119).
The punishment system, for example, produces books, treatises, regulations, new
strategies of control and resistance, debates in Parliament, conversations, confessions,
legal briefs and appeals, training regimes for prison officers, and so on. The efforts to
control sexuality produce a veritable explosion of discourse – talk about sex, television
and radio programmes, sermons and legislation, novels, stories and magazine features,
medical and counseling advice, essays and articles, learned theses and research
programmes, as well as new sexual practices (e.g. ‘safe’ sex) and the pornography
industry. Without denying that the state, the law, the sovereign or the dominant class
may have positions of dominance, Foucault shifts our attention away from the grand,
overall strategies of power, towards the many, localized circuits, tactics, mechanisms
26
and effects through which power circulates – what Foucault calls the ‘meticulous
rituals’ or the ‘micro-physics of power’. These power relations ‘go right down to the
depth of society’ (Foucault, 1977a, p. 27). They connect the way power is actually
working on the ground to the great pyramids of power by what he calls a capillary
movement (capillaries being the thin-walled vessels that aid the exchange of oxygen
between the blood in our bodies and the surrounding tissues). Not because power at
these lower levels merely reflects or ‘reproduces, at the level of individuals, bodies,
gestures and behaviour, the general form of the law or government’ (Foucault, 1977a,
p.27) but, on the contrary, because such an approach ‘roots [power] in forms of
behaviour, bodies and local relations of the central power’ (Foucault, 1980, p.201).
To what object are the micro-physics of power primarily applied, in Foucault’s model?
To the body. He places the body at the centre of the struggles between different
formations of power/knowledge. The techniques of regulation are applied to the body.
Different discursive formations and apparatuses divide, classify and inscribe the body
differently in their respective regimes of power and ‘truth’. In Discipline and Punish,
for example, Foucault analyses the very different ways in which the body of the
criminal is ‘produced’ and disciplined in different punishment regimes in France. In
earlier periods, punishment was haphazard, prisons were places into which the public
could wander and the ultimate punishment was inscribed violently on the body by
means of instruments of torture and execution, etc. – a practice the essence of which is
that it should be public, visible to everyone. The modern form of disciplinary regulation
and power, by contrast, is private, individualized; prisoners are shut away from the
public and often from one another, though continually under surveillance from the
authorities; and punishment is individualized. Here, the body has become the site of a
new kind of disciplinary regime.
Of course this ‘body’ is not simply the natural body which all human beings possess at
all times. This body is produced within discourse, according to the different discursive
formations – the state of knowledge about crime and the criminal, what counts as ‘true’
about how to change or deter criminal behaviour, the specific apparatus and
technologies of punishment prevailing at the time. This is a radically historicized
conception of the body – a sort of surface on which different regimes of
power/knowledge write their meanings and effects. It thinks of the body as ‘totally
imprinted by history and the processes of history’s deconstruction of the body’
(Foucault, 1977a, p. 63).

4.4 Summary: Foucault and representation

Foucault’s approach to representation is not easy to summarize. He is concerned with
the production of knowledge and meaning through discourse. Foucault does indeed
analyse particular texts and representations, as the semioticians did. But he is more
inclined to analyse the whole discursive formation to which a text or a practice belongs.
His concern is with knowledge provided by the human and social sciences, which
organizes conduct, understanding, practice and belief, the regulation of bodies as well as
whole populations. Although his work is clearly done in the wake of, and profoundly
influenced by, the ‘turn to language’ which marked the constructionist approach to
representation, his definition of discourse is much broader than language, and includes
many other elements of practice and institutional regulation which Saussure’s approach,
with its linguistic focus, excluded. Foucault is always much more historically specific,

27
seeing forms of power/knowledge as always rooted in particular contexts and histories.
Above all, for Foucault, the production of knowledge is always crossed with questions
of power and the body; and this greatly expands the scope of what is involved in
representation.
The major critique leveled against his work is that he tends to absorb too much into
‘discourse’, and this has the effect of encouraging his followers to neglect the influence
of the material, economic and structural factors in the operation of power/knowledge.
Some critics also find his rejection of any criterion of ‘truth’ in the human sciences in
favour of the idea of a ‘regime of truth’ and the will-to-power (the will to make things
‘true) vulnerable to the charge of relativism. Nevertheless, there is little doubt about the
major impact which his work has had on contemporary theories of representation and
meaning.

4.5 Charcot and the performance of hysteria

In the following example, we will try to apply Foucault’s method to a particular
example. Figure 1.4 shows a painting by André Brouillet of the famous French
psychiatrist and neurologist, Jean-Martin Charcot (1825-93), lecturing on the subject of
female hysteria to students in the lecture theatre of his famous Paris clinic at La
Salpêtrière.

FIGURE 1.4
André Brouillet, A clinical lesson at La Salpêtrière (given by Charcot), 1887
Brouillet shows a hysterical patient being supported by an assistant and attended by two
women. For many years, hysteria had been traditionally identified as a female malady
and although Charcot demonstrated conclusively that many hysterical symptoms were
to be found in men, and a significant proportion of his patients were diagnosed male
hysterics, Elaine Showalter observes that ‘for Charcot, too, hysteria remains

28
symbolically, if not medically, a female malady’ (1987, pp. 148). Charcot was a very
humane man who took his patients’ suffering seriously and treated them with dignity.
He diagnosed hysteria as a genuine ailment rather than a malingerer’s excuse (much as
happened, in our time, after many struggles, with other illnesses, like anorexia and ME).
This painting represents a regular feature of Charcot’s treatment regime, where
hysterical female patients displayed before an audience of medical staff and students the
symptoms of their malady, ending often with a full hysterical seizure.
The painting could be said to capture and represent, visually, a discursive ‘event’ – the
emergence of a new regime of knowledge. Charcot’s great distinction, which drew
students from far and wide to study with him (including, in 1885, the young Sigmund
Freud from Vienna), was his demonstration ‘that hysterical symptoms such as paralysis
could be produced and relieved by hypnotic suggestion’ (Showalter, 1987, p. 148). Here
we see the practice of hypnosis being applied in practice.
Indeed, the image seems to capture two such moments of knowledge production.
Charcot did not pay much attention to what the patients said (though he observed their
actions and gestures meticulously). But Freud and his friend Breuer did. At first, in their
work when they returned home, they used Charcot’s hypnosis method, which had
attracted such wide attention as a novel approach to treatment of hysteria at La
Salpêtrière. But some years later they treated a young woman called Bertha Pappenheim
for hysteria, and she, under the pseudonym ‘Anna O’, became the first case study
written up in Freud and Breuer’s path-breaking Studies in Hysteria (1974/1895). It was
the ‘loss of words’, her failing grasp of the syntax of her own language (German), the
silences and meaningless babble of this brilliantly intellectual, poetic and imaginative
but rebellious young woman, which gave Breuer and Freud the first clue that her
linguistic disturbance was related to her resentment at her ‘place’ as dutiful daughter of
a decidedly patriarchal father, and thus deeply connected with her illness. After
hypnosis, her capacity to speak coherently returned, and she spoke fluently in three
other languages, though not in her native German. Through her dialogue with Breuer,
and her ability to ‘work through’ her difficult relationship in relation to language, ‘Anna
O’ gave the first example of the ‘talking cure’ which, of course, then provided the
whole basis for Freud’s subsequent development of the psychoanalytic method. So we
are looking, in this image, at the ‘birth’ of two new psychiatric epistemes: Charcot’s
method of hypnosis, and the conditions which later produced psychoanalysis.

The example also has many connections with the question of representation. In the
picture, the patient is performing or ‘representing’ with her body the hysterical
symptoms from which she is ‘suffering’. But these symptoms are also being ‘re-
presented’ – in the very different medical language of diagnosis and analysis – to her
(his?) audience by the Professor a relationship which involves power. Showalter notes
that, in general, ‘the representation of female hysteria was a central aspect of Charcot’s
work’ (p. 148). Indeed, the clinic was filled with lithographs and paintings. He had his
assistants assemble a photographic album of nervous, a sort of visual inventory of the
various ‘types’ of hysterical patient. He later employed a professional photographer to
take charge of the service. His analysis of the displayed symptoms, which seems to be
what is happening in the painting, accompanied the hysterical ‘performance’. He did not
flinch from the spectacular and theatrical aspects associated with his demonstrations of
hypnosis as a treatment regime. Freud thought that ‘Every one of his “fascinating
lectures”’ was ‘a little work of art in construction and composition’. Indeed, Freud
noted, ‘he never appeared greater to his listeners than after he had made the effort, by

29
giving the most detailed account of his train of thought, by the greatest frankness about
his doubts and hesitations, to reduce the gulf between teacher and pupil’ (Gay, 1988, p.
49).

5 Where is ‘the subject’?

We have traced the shift in Foucault’s work from language to discourse and knowledge,
and their relation to questions of power. But where in all this, you might ask, is the
subject? Saussure tended to abolish the subject from the question of representation.
Language, he argued, speaks us. The subject appears in Saussure’s schema as the author
of individual speech-acts (paroles). But, as we have seen, Saussure did not think that the
level of the paroles was one at which a ‘scientific’ analysis of language could be
conducted. In one sense, Foucault shares this position. For him, it is discourse, not the
subject, which produces knowledge. Discourse is enmeshed with power, but it is not
necessary to find ‘a subject’ – the king, the ruling class, the bourgeoisie, the state, etc. –
for power/knowledge to operate. On the other hand, Foucault did include the subject in
his theorizing, though he did not restore the subject to its position as the centre and
author of representation. Indeed, as his work developed, he became more and more
concerned with questions about ‘the subject’, and in his very late and unfinished work,
he even went so far as to give the subject a certain reflexive awareness of his or her own
conduct, though this still stopped short of restoring the subject to his/her full
sovereignty.
Foucault was certainly deeply critical of what we might call the traditional conception
of the subject. The conventional notion thinks of ‘the subject’ as an individual who is
fully endowed with consciousness; an autonomous and stable entity, the ‘core’ of the
self, and the independent, authentic source of action and meaning. According to this
conception, when we hear ourselves speak, we feel we are identical with what has been
said. And this identity of the subject with what is said gives him/her a privileged
position in relation to meaning. It suggests that, although other people may
misunderstand us, we always understand ourselves because we were the source of
meaning in the first place.

However, as we have seen, the shift towards a constructionist conception of language

and representation did a great deal to displace the subject from a privileged position in
relation to knowledge and meaning. The same is true of Foucault’s discursive approach.
It is discourse, not the subjects who speak it, which produces knowledge. Subjects may
produce particular texts, but they are operating within the limits of the episteme, the
discursive formation, the regime of truth, of a particular period and culture. Indeed, this
is one of Foucault’s most radical propositions: the ‘subject’ is produced within
discourse. This subject of discourse cannot be outside discourse, because it must be
subjected to discourse. It must submit to its rules and conventions, to its dispositions of
power/knowledge. The subject can become the bearer of the kind of knowledge which
discourse produces. It can become the object through which power is relayed. But it
cannot stand outside power/knowledge as its source and author. In ‘The subject and
power’ (1982), Foucault writes that ‘My objective … has been to create a history of the
different modes by which, in our culture, human beings are made subjects … It is a
form of power which makes individuals subjects. There are two meanings of the word
subject: subject to someone else’s control and dependence, and tied to his (sic) own
identity by a conscience and self-knowledge. Both meanings suggest a form of power

30
which subjugates and makes subjects to’ (Foucault, 1982, pp. 208, 212). Making
discourse and representation more historical has therefore been matched, in Foucault, by
an equally radical historicization of the subject. ‘One has to dispense with the
constituent subject, to get rid of the subject itself, that’s to say, to arrive at an analysis
which can account for the constitution of the subject within a historical framework’
(Foucault, 1980, p. 115).
Where, then, is ‘the subject’ in this more discursive approach to meaning, representation
and power?

Foucault’s ‘subject’ seems to be produced through discourse in two different senses or

places. First, the discourse itself produces ‘subjects’ – figures who personify the
particular forms of knowledge which the discourse produces. These subjects have the
attributes we would expect as these are defined by the discourse: the madman, the
hysterical woman, the homosexual, the individualized criminal, and so on. These figures
are specific to specific discursive regimes and historical periods. But the discourse also
produces a place for the subject (i.e. the reader or viewer, who is also ‘subjected to’
discourse) from which its particular knowledge and meaning most makes sense. It is not
inevitable that all individuals in a particular period will become the subjects of a
particular discourse in this sense, and thus the bearers of its power/knowledge. But for
them – us – to do so, they – we – must locate themselves/ourselves in the position from
which the discourse makes most sense, and thus become its ‘subjects’ by ‘subjecting’
ourselves to its meaning, power and regulation. All discourses, then, construct subject-
positions, from which alone they make sense.
This approach has radical implications for a theory of representation. For it suggests
that discourses themselves construct the subject-positions from which they become
meaningful and have effects. Individuals may differ as to their social class, gendered,
‘racial’ and ethnic characteristics (among other factors), but they will not be able to take
meaning until they have identified with those positions which the discourse constructs,
subjected themselves to its rules, and hence become the subjects of its
power/knowledge. For example, pornography produced for men will only ‘work’ for
women, according to this theory, if in some sense women put themselves in the position
of the ‘desiring male voyeur’ – which is the ideal subject-position which the discourse
of male pornography constructs – and look at the models from this ‘masculine’
discursive position. This may seem, and is, a highly contestable proposition. But lets us
consider an example which illustrates the argument.

5.1 How to make sense of Velasquez’ Las Meninas

Foucault’s The Order of Things (1970) opens with a discussion of a painting by the
famous Spanish painter, Velasquez, called Las Meninas. It has been a topic of
considerable scholarly debate and controversy. The reason I am using it here is because,
as all the critics agree, the painting itself does raise certain questions about the nature of
representation, and Foucault himself uses it to talk about these wider issues of the
subject. It is these arguments which interest us here, not the question of whether
Foucault’s is the ‘true’, correct or even the definitive reading of the painting’s meaning.
That the painting has no one, fixed or final meaning is, indeed, one of Foucault’s most
powerful arguments.

31
The painting is unique in Velasquez’ work. It was part of the Spanish court’s royal
collection and hung in the palace in a room which was subsequently destroyed by fire. It
was dated ‘1656’ by Velasquez’ successor as court painter. It was originally called ‘The
Empress with her Ladies and a Dwarf’; but by the inventory of 1666, it had acquired the
title of ‘A Portrait of the Infanta of Spain with her Ladies in Waiting and Servants, by
the Court Painter and Palace Chamberlain Diego Velasquez’. It was subsequently called
Las Meninas – ‘The Maids of Honour’. Some argue that the painting shows Velasquez
working on Las Meninas itself and was painted with the aid of a mirror – but this now
seems unlikely. The most widely held and convincing explanation is that Velasquez was
working on a full-length portrait of the King and Queen, and that it is the royal couple
who are reflected in the mirror on the back wall. It is at the couple that the princess and
her attendants are looking and on them that the artist’s gaze appears to rest as he steps
back from his canvas. The reflection artfully includes the royal couple in the picture.
This is essentially the account which Foucault accepts.

FIGURE 1.5
Diego Velasquez,
Las Meninas,
1656
Las Meninas shows the interior of a room – perhaps the painter’s studio or some other
room in the Spanish Royal Palace, the Escorial. The Scene, though in its deeper recesses
rather dark, is bathed in light from a window on the right. ‘We are looking at a picture
in which the painter is in turn looking out at us,’ says Foucault (1970, p. 4). To the left,
looking forwards, is the painter himself, Velasquez. He is in the act of painting and his
brush is raised, ‘perhaps … considering whether to add some finishing touch to the
canvas’ (p. 3). He is looking at his model, who is sitting in the place from which we are

32
looking, but we cannot see who the model is because the canvas on which Velasquez is
painting has its back to us, its face resolutely turned away from our gaze. In the centre
of the painting stands what tradition recognizes as the little princess, the Infanta
Maragarita, who has with her an ‘entourage of duennas, maids of honour, courtiers and
dwarfs’ and her dog (p. 9). The courtiers stand behind, towards the back on the right.
Her maids of honour stand on either side of her, framing her. To the right at the front are
two dwarfs, one a famous court jester. The eyes of many of these figures, like that of the
painter himself, are looking out towards the front of the picture at the sitters.
Who are they – the figures at whom everyone is looking but whom we cannot look at
and whose portraits on the canvas we are forbidden to see? In fact, though at first we
think we cannot see them, the picture tells us who they are because, behind the Infanta’s
head and a little to the left of the centre of the picture, surrounded by a heavy wooden
frame, is a mirror; and in the mirror – at last – are reflected the sitters, who are in fact
seated in the position from which we are looking: ‘a reflection that shows us quite
simply what is lacking in everyone’s gaze’ (p. 15). The figures reflected in the mirror
are, in fact, the King, Philip IV, and his wife, Mariana. Beside the mirror, to the right of
it, in the back wall, is another ‘frame’, but this is not a mirror reflecting forwards; it is a
doorway leading backwards out of the room. On the stair, his feet placed on different
steps, ‘a man stands out in full-length silhouette’. He has just entered or is just leaving
the scene and is looking at it from behind, observing what is going on in it but ‘content
to surprise those within without being seen himself’ (p. 10).

5.2 The subject of/in representation

Who or what is the subject of this painting? In his comments, Foucault uses Las
Meninas to make some general points about his theory of representation and specifically
about the role of the subject:

1 ‘Foucault reads the painting in terms of representation and the subject’ (Dreyfus
and Rabinow, 1982, p. 20). As well as being a painting which shows us (represents) a
scene in which a portrait of the King and Queen of Spain is being painted, it is also a
painting which tells us something about how representation and the subject work. It
produces its own kind of knowledge. Representation and the subject are the painting’s
underlying message – what it is about, its sub-text.
2 Clearly, representation here is not about a ‘true’ reflection or imitation of reality. Of
course, the people in the painting may ‘look like’ the actual people in the Spanish court.
But the discourse of painting in the picture is doing a great deal more than simply trying
to mirror accurately what exists.
3 Everything in a sense is visible in the painting. And yet, what it is ‘about’ – its
meaning depends on how we ‘read’ it. It is as much constructed around what you can’t
see as what you can. You can’t see what is being painted on the canvas, though this
seems to be the point of the whole exercise. You can’t see what everyone is looking at,
which is the sitters, unless we assume it is a reflection of them in the mirror. They are
both in and not in the picture. Or rather, they are presented through a kind of
substitution. We cannot see them because they are not directly represented: but their
‘absence’ is represented – mirrored through their reflection in the mirror at the back.
The meaning of the picture is produced, Foucault argues, through this complex inter-

33
play between presence (what you see, the visible) and absence (what you can’t see,
what has displaced it within the frame). Representation works as much through what is
not shown, as through what is.
4 In fact, a number of substitutions or displacements seem to be going on here. For
example, the ‘subject’ and centre of the painting we are looking at seems to be the
Infanta. But the ‘subject’ or centre is also, of course, the sitters – the King and Queen –
whom we can’t see but whom the others are looking at. You can tell this from the fact
that the mirror on the wall in which the King and Queen are reflected is also almost
exactly at the centre of the field of vision of the picture. So the Infanta and the Royal
Couple, in a sense, share the place of the centre as the principal ‘subjects’ of the
painting. It all depends on where you are looking from – in towards the scene from
where you, the spectator, is sitting or outwards from the scene, from the position of the
people in the picture. If you accept Foucault’s argument, then there are two subjects to
the painting and two centres. And the composition of the picture – its discourse – forces
us to oscillate between these two ‘subjects’ without ever finally deciding which one to
identify with. Representation in the painting seems firm and clear – everything in place.
But our vision, the way we look at the picture, oscillates between two centres, two
subjects, two positions of looking, two meanings. Far from being finally resolved into
some absolute truth which is the meaning of the picture, the discourse of the painting
quite deliberately keeps us in this state of suspended attention, in this oscillating process
of looking. Its meaning is always in the process of emerging, yet any final meaning is
constantly deferred.
5 You can tell a great deal about how the picture works as a discourse, and what it
means, by following the orchestration of looking – who is looking at what or whom.
Our look – the eyes of the person looking at the picture, the spectator – follows the
relationships of looking as represented in the picture. We know the figure of the Infanta
is important because her attendants are looking at her. But we know that someone even
more important is sitting in front of the scene whom we can’t see, because many figures
– the Infanta, the jester, the painter himself – are looking at them! So the spectator (who
is also ‘subjected’ to the discourse of the painting) is doing two kinds of looking.
Looking at the scene from the position outside, in front of, the picture. And at the same
time, looking out of the scene from the position outside, in front of, the picture. And at
the same time, looking out of the scene, by identifying with the looking being done by
the figures in the painting. Projecting ourselves into the subjects of the painting help us
as spectators to see, to ‘make sense’ of it. We take up the positions indicated by the
discourse, identify with them, subject ourselves to its meanings, and become its
‘subjects’.

6 It is critical for Foucault’s argument that the painting does not have a completed
meaning. It only means something in relation to the spectator who is looking at it. The
spectator completes the meaning of the picture. Meaning is therefore constructed in the
dialogue between the painting and the spectator. Velasquez, of course, could not know
who would subsequently occupy the position of the spectator. Nevertheless, the whole
‘scene’ of the painting had to be laid out in relation to that ideal point in front of the
painting from which any spectator must look if the painting is to make sense. The
spectator, we might say, is painted into position in front of the picture. In this sense, the
discourse produces a subject-position for the spectator-subject. For the painting to work,
the spectator, whoever he or she may be, must first ‘subject’ himself/herself to the
painting’s discourse and, in this way, become the painting’s ideal viewer, the producer

34
of its meanings – its ‘subject’. This is what is meant by saying that the discourse
constructs the spectator as a subject – by which we mean that it constructs a place for
the subject-spectator who is looking at and making sense of it.
7 Representation therefore occurs from at least three positions in the painting. First of
all there is us, the spectator, whose ‘look’ puts together and unifies the different
elements and relationships in the picture into an overall meaning. This subject must be
there for the painting to make sense, but he/she is not represented in the painting.
Then there is the painter who painted the scene. He is ‘present’ in two places at once,
since he must at one time have been standing where we are now sitting, in order to paint
the scene, but he has then put himself into (represented himself in) the picture, looking
back towards that point of view where we, the spectator, have taken his place. We may
also say that the scene makes sense and is pulled together in relation to the court figure
standing on the stair at the back, since he too surveys it all but – like us and like the
painter – from somewhat outside it.
8 Finally, consider the mirror on the back wall. If it were a ‘real’ mirror, it should
now be representing or reflecting us, since we are standing in that position in front of
the scene to which everyone is looking and from which everything makes sense. But it
does not mirror us, it shows in our place the King and Queen of Spain. Somehow the
discourse of the painting positions us in the place of the Sovereign! You can imagine
what fun Foucault had with this substitution.
Foucault argues that it is clear from the way the discourse of representation works in the
painting that it must be looked at and made sense of from that one subject-position in
front of it from which we, the spectators, are looking. This is also the point-of-view
from which a camera would have to be positioned in order to film the scene. And, lo
and behold, the person whom Velasquez chooses to ‘represent’ sitting in this position is
The Sovereign – ‘master of all he surveys’ – who is both the ‘subject of’ the painting
(what it is about) and the ‘subject in’ the painting – the one whom the discourse sets in
place, but who, simultaneously, makes sense of it and understands it all by a look of
supreme mastery.

6 Conclusion: representation, meaning and language

reconsidered
We started with a fairly simple definition of representation. Representation is the
process by which members of a culture use language (broadly defined as any system
which deploys signs, any signifying system) to produce meaning. Already, this
definition carries the important premise that things – objects, people, events, in the
world – do not have in themselves any fixed, final or true meaning. It is us – in society,
within human cultures – who make things mean, who signify. Meanings, consequently,
will always change, from one culture or period to another. There is no guarantee that
every object in one culture will have an equivalent meaning in another, precisely
because cultures differ, sometimes radically, from one another in their codes – the ways
they carve up, classify and assign meaning to the world. So one important idea about
representation is the acceptance of a degree of cultural relativism between one culture
and another, a certain lack of equivalence, and hence the need for translation as we
move from the mind-set or conceptual universe of one culture or another.

35
We call this the constructionist approach to representation, contrasting it with both the
reflective and the intentional approaches. Now, if culture is a process, a practice, how
does it work? In the constructionist perspective, representation involves making
meaning by forging links between three different orders of things: what we might
broadly call the world of things, people, events and experiences; the conceptual world –
the mental concepts we carry around in our heads; and the signs, arranged into
languages, which ‘stand for’ or communicate these concepts. Now, if you have to make
a link between systems which are not the same, and fix these at least for a time so that
other people know what, in one system, corresponds to what in another system, then
there must be something which allows us to translate between them – telling us what
word to use for what concept, and so on. Hence the notion of codes.
Producing meaning depends on the practice of interpretation, and interpretation is
sustained by us actively using the code – encoding, putting things into the code – and by
the person at the other end interpreting or decoding the meaning (Hall, 1980). But note,
that, because meanings are always changing and slipping, codes operate more like social
conventions than like fixed laws or unbreakable rules. As meanings shift and slide, so
inevitably the codes of a culture imperceptibly change. The great advantage of the
concepts and classifications of the culture which we carry around with us in our heads is
that they enable us to think about things, whether they are there, present, or not; indeed,
whether they ever existed or not. There are concepts for our fantasies, desires and
imaginings as well as for so-called ‘real’ objects in the material world. And the
advantage of language is that our thoughts about the world need not remain exclusive to
us, and silent. We can translate them into language, make them ‘speak’, through the use
of signs which stand for them – and thus talk, write, communicate about them to others.
Gradually, then we complexified what we meant by representation. It came to be less
and less the straightforward thing we assumed it to be at first – which is why we need
theories to explain it. We looked at two versions of constructionism – that which
concentrated on how language and signification (the use of signs in language) works to
produce meanings, which after Saussure and Barthes we called semiotics; and that,
following Foucault, which concentrated on how discourse and discursive practices
produce knowledge. I won’t run through the finer points in these two approaches again,
since you can go back to them in the main body of the chapter and refresh your
memory. In semiotics, you will recall the importance of signifier/signified,
langue/parole and ‘myth’, and how the marking of difference and binary oppositions
are crucial for meaning. In the discursive approach, you will recall discursive
formations, power/knowledge, the idea of a ‘regime of truth’, the way discourse also
produces the subject and defines the subject-positions from which knowledge proceeds
and indeed, the return of questions about ‘the subject’ to the field of representation. In
several examples, we tried to get you to work with these theories and apply them. There
will be further debate about them in subsequent chapters.
Notice that the chapter does not argue that the discursive approach overturned
everything in the semiotic approach. Theoretical development does not usually proceed
in this linear way. There was much to learn from Saussure and Barthes, and we are still
discovering ways of fruitfully applying their insights – without necessarily swallowing
everything they said. We offered you some critical thoughts on the subject. There is a
great deal to learn from Foucault and the discursive approach, but by no means
everything it claims is correct and the theory is open to, and has attracted, many
criticisms. Again, in later chapters, as we encounter further developments in the theory

36
of representation, and see the strengths and weakness of these positions applied in
practice, we will come to appreciate more fully that we are only at the beginning of the
exciting task of exploring this process of meaning construction, which is at the heart of
culture, to its full depths. What we have offered here is, we hope, a relatively clear
account of a set of complex, and as yet tentative, ideas in an unfinished project.

i
Soap opera: gênero de programa popular nos EUA, consiste numa série de ficção dramática ou cômica
caracterizada principalmente por duas características: o uso de personagens estereotipados e as tramas
longas, que se estendem por muitos episódios. Apesar das semelhanças, é importante não confundir as
soap operas com telenovelas. Soap operas podem ter décadas de exibição contínua, enquanto novelas
duram apenas alguns meses. No Brasil, o programa Malhação pode ser considerado uma soap opera,
pois está no ar desde o início dos anos 90 ininterruptamente. (N. T.)

ii
Metalanguage: a second-order language, in which one speaks about the first.

iii
In their work New reflections on the Revolution of our Time (1990), Ernesto Laclau and Chantal Mouffe
argue that physical objects do exist, but they have no fixed meaning; they only take on meaning and
become objects of knowledge within discourse. The round leather object which you kick is a physical
object – a ball. But it only becomes ‘a football’ within the context of the rules of the game, which are
socially constructed. It is impossible to determine the meaning of an object outside of its context of use.
A stone thrown in a fight is a different thing (‘a projectile’) from a stone displayed in a museum (‘a piece
of sculpture’).

37