Structuralism, in a broad sense, is the practice of studying phenomena as different as societies, minds, languages, literatures and mythologies, as total systems, or connected wholes – that is, structures – and in terms of their internal patterns of connection, rather than in terms of their historical sequence of development. It is the study of how the structures of these entities affect the way they function.
Structuralism, in the more familiar narrow sense, is an attempt to find a model for both the structure and functioning of societies, minds, etc in the structuralist model of language, developed in linguistics in the early twentieth century. Language is studied from many points of view, by philosophers, sociologists, literary critics, etc., all of whom make a point of having some axe to grind. Linguists study it for its own sake, to find out its intrinsic structure, and are sometimes roundly condemned for doing so. An influential early version of the structuralist model of language is to be found in the Course in General Linguistics (1916), by Ferdinand de Saussure, the basic principles of which are explained in this chapter. Attempts to apply the model to literature date at least from 1928, in the programme of Jakobson and Tynjanov. This was the start of literary structuralism.
1.1 The elementary concept of structure
Here is a possible definition of a structure:
A structure is a set of parts which are connected together
In this sense, a containerful of spares is just a set; but if they are connected together to make a car, that car will be a structure. Similarly, there was once a popular children’s engineering toy called Meccano, nowadays displaced by Lego. You bought a box of Meccano parts – small pieces of metal of various shapes, drilled with a regular pattern of holes so that they could be bolted together, which was called, with mathematical correctness, a Meccano set. From your set you took parts and bolted them together to make model cars, or cranes. These are structures. Adult engineers make prefabricated houses on this principle too; and polytechnic administrators make modular degree courses. The main advantage is the enormous number of different structures you can make, not all at the same time of course, out of the same limited kit of parts. This is also the main advantage of structuralist theories.
The general principle of structuralism is to extend this idea – of the standard kit of parts out of which you can make particular structures like cars and cranes – as far as it can possibly go; to extend it, in particular, outside the limits of the physical, or even biological worlds and into the contents of the mind.
How far is it possible to see objects as structures when we don’t actually construct them ourselves? Obviously it is possible. In this sense – which is definitely one of the senses employed in common speech – almost every material object in the world is a structure. A tree can be seen as a structure and so can a geological formation; meteorologists even think of clouds as structures. The whole of physical science, in this sense, is based on explaining the behaviour of the world in terms of its physical structure.
But structures do not have to be material objects. An argument can be a structure – the parts being two premises and a conclusion and the connection being that the conclusion follows logically from the premises. A mind can be a structure made up of ego, superego and id. A society can be a structure: made up, say, of a set of statuses or positions like father, son; employer, employee – with connections between them; or of connected classes: bourgeoisie and proletariat.
The Meccano analogy is useful here, though it is necessary to keep it as a conscious analogy and to wonder just how far it can be pressed. Anybody who has ever played with a Meccano set has grasped the essential principle of structuralism. (Though not, of course, some of the more complex derivative notions like those of ‘structural transformation’.) A structuralist, on the Meccano model, is a person who puts together some pieces from what he takes to be a pre-existing kit of parts (material or otherwise), to make a model of some aspect of the material, social or mental world. But if we are talking of societies, or minds, how far can we think of them as made up of parts that could possibly exist separately?
There is an equally common sense of structure which is slightly more abstract; here is a definition:
A structure is the set of connections between parts, in a set of parts which are connected together.
In this sense we should say that a car has a structure rather than that it is a structure; similarly that we build a crane with a certain structure out of our Meccano set; similarly that a tree, a hill, a cloud, an argument, a mind, a society have structures. In this sense of the word, to specify the structure of something, you specify the connections between the parts; you do not have to say what the parts are, beyond what is necessary to identify them. Hence different objects, made out of different parts, might have the same structure. (We can build a crane of the same structure out of Meccano or out of one of the numerous imitations.)
Claude Lévi-Strauss attaches a great deal of importance to this distinction between the more concrete and the more abstract sense of structure: he distinguishes them as structure and form respectively. In ‘Structure and Form: Reflections on a Work by Vladimir Propp’ (1960), he writes:
The supporters of structural analysis in linguistics and in anthropology are often accused of formalism. This is to forget that formalism exists as an independent doctrine from which structuralism – without denying its debt to it – separated because of the very different attitudes the two schools adopt toward the concrete. Contrary to formalism, structuralism refuses to set the concrete against the abstract and to recognise a privileged value in the latter. Form is defined by opposition to material other than itself. But structure has no distinct content; it is content itself, apprehended in a logical organisation conceived as property of the real.
For Lévi-Strauss, then, if he accepts the Meccano analogy at all2
, a model car made out of Meccano will be a structure; while the
specification in the instruction book of how the parts fit together will specify the form ofthat structure.
If we accept either of the notions of structure defined above it is possible to offer a definition of a very broad concept of structuralism:
Structuralism is the practice of studying phenomena as different as societies, minds, languages, literatures and mythologies as total systems, or connected wholes – that is, structures – and in terms of their internal patterns of connection, rather than as sets of isolated items and in terms of their historical sequence.
Under this heading we can include at least the following: accounts of social structure in sociology and anthropology, by anthropologists of the Anglo-American schools just as much as by Lévi-Strauss; accounts of mental structures from the diverse points of view of Freud, of Piaget and of Chomsky; accounts of human language of the kind due to de Saussure, Sapir and the whole tradition of modern general linguistics; accounts of literature from formalist points of view, ranging from Northrop Frye through the Russian Formalists and into Structuralism in the narrower sense which is explored in the next section; accounts of mythology and legend from a similar range of points of view; and some recent approaches to history. In this account of general structuralism language has no particular privilege – it is one object of structuralist enquiry along with many others.
Some of the wilder claims for the importance of structuralism – the claim, for example, that it is the key movement of the modern period – probably have this very broad conception at the back of them. So also has the metaphysical structuralism of which Jacques Derrida (1967b) offers a critique. ‘… for it would not be difficult to show’ (he says, without actually showing) ‘that a certain structuralism has always been philosophy’s most spontaneous gesture’ (p. 159). And, ‘It would be easy enough to show that the concept of structure and even the word “structure” itself are as old as the episteme – that is to say, as old as Western science and Western philosophy …’ (p.278).
It is worth asking, however, whether one is saying anything worth while about these very diverse thinkers and bodies of thought by labelling them all ‘structuralist’. I think in fact that one is: one is saying something extremely simple about them and making a very weak claim, but it is a genuine claim and it is true. The claim is that in all these domains it is necessary to recognize significant structural patterns necessary to recognize significant structural pattern – where a structural pattern is defined as a pattern of connections between parts. One cannot simply reduce these domains to unordered sets of elementary parts and treat holistic properties as common
properties of each individual part, or even as general properties of the whole collection.
Here are three examples of this weak structural property, ranging from the banal to the slightly contentious.
(1) The properties of a building are not those of its bricks; you can build a square building forty feet high out of bricks nine inches long and rectangular.
(2) The properties of a passage of text are not those of the words in it; its meaning is not the meaning of any individual word, nor of all the words heaped together, but of a particular linear structure built out of them; and is dependent also on the structural patterns in the language from which the words come.
(3) The properties of a social institution cannot be reduced to properties of the people who take part in it. The Prudential Assurance Company is not identical with any or all of its employees, or with its articles of association; its purposes and functions are not necessarily identical with theirs; there is no collective consciousness to have purposes and functions; and yet the company does exist, and does have purposes and functions of a quite straightforward kind which depend on its own structure and the legal and social structure of society.
Is this claim too weak to be of any interest or importance? Many would assume that it is. Piaget, for example, in his book, Structuralism (1968), (which is less an introduction to the field than an attempt to re-found it) adopts a much stronger definition. As is common in this literature, it is unclear. According to him:
a structure is a system of transformations … the structure is preserved or enriched by the interplay of its transformation laws, which never yield results external to the system, nor employ elements that are external to it. In short, the notion of structure is comprised of three key ideas: the idea of wholeness, the idea of transformation and the idea of self-regulation.
Piaget thus replaces the simple concept of a structure with the much more complex concept of a self-regulating system. It is not clear precisely what he means, since he never defines the key concept of transformation. But the obvious definition is:
a transformation is a change from one structure to another
where the word ‘structure’ bears the simpler sense defined earlier.
Piaget, of course, is interested in what might be called the self-generation of complex structures out of simpler ones; something which is important in itself, especially in biology, where this is a
characteristic mode of physical development. It had been Piaget’s life-work to show that it was a characteristic mode of cognitive development, too; in his empirical studies of cognitive epistemology
(Piaget and Inhelder 1963), he had attempted to show how the capacity to recognize basic logical structures develops in children.
Piaget tries to show the importance of his concept in every field from mathematical logic to the social sciences. It is an idea to which it will be necessary to return when considering the possible philosophical consequences of structuralism. But it remains a complex and derivative concept, often most confusing in just those central areas of structuralist thought, like linguistics, where such notions as ‘transformation’ are most precisely defined.
Thus generative grammar is concerned with tree-structures, which are, as the name suggests, structures made out of parts of speech connected together like the twigs and branches of a tree. A ‘transformation’ here means a mapping from one tree-structure to another.3
We need the concept of structure in order to define the concept of transformation, not the other way round! There seems little point in defining a sense for ‘structuralism’ that doesn’t fit any model of linguistics very well and is better adapted to cybernetics.
I return therefore, to the very weak claim made above: that is that all these things that the structuralists consider contain significant patterns of connection between their parts; and it is these that need to be studied rather than – or at least as well as – the substantive elements of which things are made. Curiously enough, even from a claim as weak as this we can deduce a number of worthwhile consequences. Some of them are obvious; some are more subtle; but they are all capable of leading to confusion if they are forgotten; and they have often been forgotten in structuralist writing. By spelling them out at the start of my account, before getting on to structuralism in the narrower and more important sense of the linguistics-based model, I hope to avoid these confusions later on.
1.2 Structure and real existence
My first point is a blindingly obvious one.
Except in pure mathematics, which I discuss immediately below, every structure is a structure of something. If we use the word ‘structure’ in a concrete sense, so that we would say a Meccano model of a car is a structure, then we mean that it is a structure of – that is,
made of – Meccano parts. If we use it in the abstract sense, then we mean that the structure of the Meccano model is the pattern of connection of the parts. We can specify this structure – in the abstract sense – without actually making it; this might be done by a diagram in the Meccano set’s owner’s manual. But if we want to make the actual model, we need more than a structural diagram to make it from. We need actual Meccano parts.
You cannot make a real object out of an abstract structure alone; only out of a structure of parts which are themselves real. As I said, the point is obvious when you are talking of material things like Meccano models. But it...