This thesis addresses the phenomenon of depression from a dimensional perspective and focuses on the study of individual symptoms. In order to test the clinical model of Sydney Blatt depression (1974, 2004) two studies were carried out. The first was the cross-cultural validation of the Multidimensional Depression Echelle (EMD; Multiscore Depression Inventory; Berndt, Petzel, Berndt, 1980). This is a questionnaire to measure the severity of depression and 10 individual symptoms per 10 sub-scales. A French version of the instrument has been developed and statistical analyses have been carried out to determine the psychometric properties of the instrument in terms of linguistic equivalence and the validity and precision of the overall shape and the abbreviated shape. Language equivalence was demonstrated by using a sample of bilingual students from the Laval University undergraduate (N = 30). The construction validity confirmed the multifactorial structure of the overall shape (N = 278) and the abbreviated shape (N = 305). The fidelity analysis determined the internal consistency and temporal stability for the overall scale and its sub-scales in the overall shape, as well as the internal consistency for the abbreviated shape. Two sub-ladders of the abbreviated shape failed to prove temporal stability. Once the EMD was validated, a second study was carried out to test the clinical model of Blatt depression. This study analyses the relationship between the two depressive forms proposed by Blatt: anaclitic and introjective form, and three fundamental variables of the model: depressive symptoms, representations of objects and self-representations. The participants were 30 women and 14 men with diagnosis of depression or clinically significant depressive symptoms. The results support the hypothesis of symptomatic specificity for the introjective form, which is more closely associated than the anaclitic form with symptoms of irritability, low self-esteem and feelings of guilt. On the other hand, the anaclitic form has been associated with a single symptoms: cognitive difficulties. Our results bear very little on the association between depressive forms and representations of object and self-esteem. For anaclitic depression, no significant results were found and for the introjective depression, malicious representation of the spouse and poor relational quality were the only significant predictors. The results are discussed in relation to existing theory and research.