When Lacan uses the term 'structure' in his early work of the 1930s, it is to refer to 'social structures', by which he means a specific set of affective relations between family members. The child perceives these relations much more profoundly than the adult, and internalises them in the complex (Ec, 89). The term serves as a peg upon which Lacan can hang his own views of the 'relational' nature of the psyche, in opposition to the atomistic theories then current in psychology (Lacan, 1936). From this point on, the term 'structure' retains this sense of something both intersubjective and intrasubjective, the internal representation of interpersonal relations. This remains a key point throughout Lacan's work, in which the emphasis on structure is a constant reminder that what determines the subject is not some supposed 'essence' but simply his position with respect to other subjects and other signifiers. Already in 1938, we find Lacan arguing that 'the most notable defect of analytic doctrine' at that time was that it tended 'to ignore structure in favour of a dynamic approach' (Lacan, 1938: 58). This anticipates his later emphasis on the symbolic order as the realm of structure which analysts have ignored in favour of the imaginary; 'social structures are symbolic' (Ec, 132).
In the mid-1950s, when Lacan begins to reformulate his ideas in terms borrowed from Saussurean structural linguistics, the term 'structure' comes to be increasingly associated with Saussure's model of language. Saussure analysed language (la langue) as a system in which there are no positive terms, only differences (Saussure, 1916: 120). It is this concept of a system in which each unit is constituted purely by virtue of its differences from the other units which comes to constitute the core meaning of the term 'structure' in Lacan's work from this point on. Language is the paradigmatic structure, and Lacan's famous dictum, 'the unconscious is structured like a language', is therefore tautologous, since 'to be structured' and 'to be like a language' mean the same thing.
Saussure's structural approach to linguistics was developed further by Roman Jakobson, who developed phoneme theory; Jakobson's work was then taken up by the French anthropologist, Claude LÈvi-Strauss, who used the structural phonemic model to analyse non-linguistic cultural data such as kinship relations and myth. This application of structural analysis to anthropology launched the structuralist movement by showing how the Saussurean concept of structure could be applied to an object of enquiry other than language. Lacan was heavily influenced by all three of these thinkers, and in this sense he can be seen as part of the structuralist movement. However, Lacan prefers to dissociate himself from this movement, arguing that his approach differs in important ways from the structuralist approach (S20, 93).
Alongside the references to language, Lacan also refers the concept of structure to mathematics, principally to set theory and topology In 1956, for example, he states that 'a structure is in the first place a group of elements forming a covariant set' (S3, 183). Two years later he again links the concept of structure with mathematical set theory, and adds a reference to topology (Ec, 648-9). By the 1970s, topology has replaced language as the principal paradigm of structure for Lacan. He now argues that topology is not a mere metaphor for structure; it is that structure itself (Lacan, 1973b).
The concept of structure is often taken to imply an opposition between surface and depth, between directly observable phenomena and 'deep structures' which are not the object of immediate experience. Such would seem to be the opposition implied in the distinction Lacan draws between symptoms (surface) and structures (depth). However, Lacan does not in fact agree that such an opposition is implicit in the concept of structure (Ec, 649). On the one hand, he rejects the concept of 'directly observable phenomena', arguing that observation is always already theoretical. On the other hand, he also rejects the idea that structures are somehow 'deep' or distant from experience, arguing that they are present in the field of experience itself; the unconscious is on the surface, and looking for it in 'the depths' is to miss it. As with many other binary oppositions, the model Lacan prefers is that of the moebius strip; just as the two sides of the strip are in fact continuous, so structure is continuous with phenomena.
The most important feature of structural analysis is not, then, any supposed distinction between surface and depth, but, as LÈvi-Strauss shows in his structural analysis of myth, the discovery of fixed relations between loci which are themselves empty (LÈvi-Strauss, 1955). In other words, whatever elements may be placed in the positions specified by a given structure, the relations between the positions themselves remain the same. Thus the elements interact not on the basis of any inherent or intrinsic properties they possess, but simply on the basis of the positions which they occupy in the structure.
In line with many other psychoanalysts, Lacan distinguishes three principal nosographic categories; neurosis, psychosis and perversion. His originality lies in the fact that he regards these categories as structures rather than simply as collections of symptoms. (N.B. Lacan prefers to speak in terms of 'Freudian structures' rather than 'clinical structures', but the latter term is the one which predominates in the writings of Lacanian psychoanalysts today.)
Lacanian nosography is a categorical classification system based on a discrete series, rather than a dimensional system based on a continuum. The three major clinical structures are therefore mutually exclusive; a subject cannot be both neurotic and psychotic, for example. The three major clinical structures together constitute all the three possible positions of the subject in relation to the Other; every subject encountered in psychoanalytic treatment can therefore be diagnosed as either neurotic, or psychotic, or perverse. Each structure is distinguished by a different operation: neurosis by the operation of repression, perversion by the operation of disavowal, and psychosis by the operation of foreclosure. Lacan follows Freud in arguing that the classical method of psychoanalytic treatment (involving free association and the use of the couch) is only appropriate for neurotic subjects and perverse subjects, and not for psychotics. Thus when Lacanian analysts work with psychotic patients, they use a substantially modified method of treatment.
One of the most fundamental axioms of psychoanalysis is that the subject's clinical structure is determined by his experiences in the nrst years of life. In this sense, psychoanalysis is based on a 'critical period hypothesis'; the first years of life are the critical period in which the subject's structure is determined. Although it is not clear how long this critical period lasts, it is held that after this critical period the clinical structure is fixed for ever and cannot be changed. Neither psychoanalytic treatment nor anything else can, for example, turn a psychotic into a neurotic.Within each of the three major clinical structures Lacan distinguishes various subdivisions. For example within the clinical structure of neurosis, he distinguishes two kinds of neurosis (obsessional neurosis and hysteria), and within the clinical structure of psychosis he distinguishes between paranoia, schizophrenia and manic-depressive psychosis.