Cette étude examine le développement psychosexuel d’un échantillon de 60 sujets condamnés pour crime contre les personnes et présentant un trouble borderline de la personnalité. Les stades du développement psychosexuel sont étudiés à l’aide de la McGill Object Relations Scale (Dymetrysym et al., 1997). La structure des corrélations entre les scores aux 6 stades de développement est analysée selon un modèle de dépliage (unfolding model). Les résultats supportent l’hypothèse d’une organisation temporelle des stades le long d’une seule dimension latente, avec une exception notable concernant le stade anal, ce qui corrobore la forte probabilité d’une fixation-régression à ce stade, liée à la particularité clinique des sujets d’étude.

Introduction. Psychosexual development is generally assessed clinically, that is to say, in a qualitative manner (Piper and Duncan, 1999). Dymetryszyn, Bouchard, Bienvenu, De Carufel, and Gaston (1997) recently proposed a more quantitative approach based on the overall maturity of the subject’s object relations. Background. Object-relation maturity is quantified using a profile that defines a score for each stage of an individual’s psychosexual development : oral-narcissistic, oral-objectal, anal, phallic pre-oedipal, oedipal, and genital. The McGill Object Relations Scale (MORS) (Dymetryszyn et al., 1997) was adapted (Combalbert, Vautier, Bourdet-Loubère, Favard, & Bouchard, 2002) and then used to obtain the quantitative scores. The relationship between overall object-relation maturity and psychosexual development is complex. Psychosexual development is considered here to be a constant latent dimension. Accordingly, the higher the level of development, the greater the individual’s possibility of establishing the object relations that correspond to more evolved stages. Inversely, the lower the level of development, the more the individual is forced to rely on object relations corresponding to earlier stages of development. However, the connection between the object-relation maturity profile and psychosexual development is made even more complex by the existence of the well-known phenomenon of fixation-regression. Objectives. This article attempts to determine how fixation-regression is reflected in the structure of the profiles of a sample of individuals whose clinical diagnosis suggests probable regression to the anal stage. Methods. The subjects chosen had borderline personality disorder coupled with either perversion or psychopathy. The data were modelled using an unfolding model. MORS was used on 60 criminal subjects who had been charged with, or convicted of, crimes against persons. The nosographic diagnosis was based on the psychiatric assessment of the prisoners. DSM IV diagnostic criteria were used as a reference to confirm or refute the pronounced diagnosis. Only subjects who met at least five criteria for borderline personality disorder were included in the sample. To assess the reliability of the individual scores, two expert judges blindly scored the 60 protocols. Results. The results obtained with an unfolding model support the hypothesis that the observed ordinates for the oral-narcissistic, oral-objectal, phallic pre-oedipal, oedipal, and genital stages obey a law of synchrony that complies with the postulate of a continuum in psychosexual development. The second main result was the fact that the observed ordinates for the level corresponding to the anal stage could not be described by the model. Furthermore, this finding did not come from a possible scoring error, since inter-judge reliability was 0.93 for that level. Discussion. These results are encouraging for the utilization of MORS as a technique for quantifying object relations. This study on a sample of individuals with a high probability of anal stage fixation-regression suggests that this type of phenomenon can be objectified by means of MORS profiles. However, the present study has some limitations of its own. The main methodological drawback pertains to the subjective aspect of the protocol scoring. The two judges had a great deal of joint experience in scoring other protocols on the MORS grid. This could explain the particularly high inter-judge correlations, compared to the within-class correlations published by Dymetryszyn et al. (1997). Conclusion. One cannot rule out the possibility that our analyses would have had different outcomes had other independent judges scored the 60 protocols. Moreover, our results cannot be generalized without replication. A critical test would be to replicate our analyses on similar data, and to examine the unfolding model’s behavior on data from a sample of ordinary adults diagnosed as not having strong fixation-regression tendencies. If the synchrony laws defined by the unfolding model are correct, the anal stage should be included in the arc formed by projecting the six stages on the factorial plane of the principal component analysis. Fundamental research based on the psychosexual theory of the libido and supported by statistical data can objectify the vicissitudes of psychosexual development (such as fixation-regression) or developmental differences across groups of individuals. This makes a strong case for pursuing MORS-based investigations.