Cet article propose une traduction française de l'échelle d'évaluation de la pensée, du langage et de la communication [Scale for the Assessment of Thought, Language and Communication (TLC)] de Andreasen. Les qualités métrologiques de cette version sont évaluées sur une population de 107 patients schizophrènes selon les critères du DSM IV et comparées à celles retrouvées dans la littérature. Les résultats montrent une très bonne fidélité interjuges, un profil de réponse aux différents items et une analyse factorielle en composantes principales très proches de ce qui est retrouvé dans la littérature avec la version anglo-saxonne avec en particulier, lors de l'analyse factorielle avant rotation, un facteur principal autorisant au calcul du score global, et après rotation, trois sous-scores factoriels représentatifs respectivement de la désorganisation, la production verbale et la structure verbale. La traduction de la TLC de Andreasen que nous proposons ici semble donc avoir des qualités métrologiques suffisamment proches de celles retrouvées dans la littérature anglo-saxonne pour pouvoir être couramment utilisée en France.

This article proposes a french translation of Andreasen's Thought, Language and Communication (TLC) scale (Andreasen, 1979). This scale is widely used in current literature and remains a reference due to the fact that it has made it possible to establish a consensus with regard to formal thought disorders and has contributed to the operationalisation of the concept of dissociation. This scale consists of 18 items. Each item is clearly defined through the use of clinical examples, rated from 0 to 4 as a function of the intensity of the disorder (absent, slight, medium, severe, extreme). The interview conditions are also stated : free interview of minimum 10 minutes followed by a more structured interview. Some items of the TLC are taken directly from the SANS and SAPS. Their translation has been taken over from french translations already validated by Lecrubier and Boyer (1987). The others were translated within the department and have been verified by a native English speaker. The entirely of the translation has been verified by Andreasen. The metrological qualities of this french translation have been studied in a population of 107 schizophrenic patients who fulfilled all the DSM IV criteria : 73 males and 34 females, mean age 33.4 9 years, in or outpatients, all under neuroleptic treatment and all evaluated by an experienced clinician. Thirty one patients have been filmed to assess the interjudge reliability. The results indicate a high level of interjudge consistency (interclass correlation coefficient 0.96). The global score was 17 9.4. In the factorial analysis before rotation we observe a main factor that makes it possible to calculate a global score. The results of factor analysis of the TLC variables after rotation yield five factors that have an eigen value greater than 1. These five factors explain 66 % of the variance. All items have a weight greater than 0.45. The first factor includes Poverty of content speech, Tangentiality, Derailment, Incoherence, Illogical thinking, Loss of goal and Perseveration. It reflects thinking disorganisation. The second factor includes Pressure of speech, Circumstantiality, Self reference and Poverty of speech (negative weight). This factor reflects verbal production. The third factor is composed of Clanging, Neologisms, Word approximation and Echolalia. This factor reflects verbal structure. The fourth factor is only composed of Stilted speech and the fifth one composed of Distractible speech and Blocking. These data have been compared to those reported in the literature : Andreasen in 1979 (113 patients : 32 suffering from manic disorder, 36 from depressive disorder and 45 schizophrenic disorder) and in 1986 (194subjects: 94 controls, 25 suffering from manic disorder, 25schizoaffective disorder and 50schizophrenic patients), Harvey in 1992 (115schizophrenic patients) and Peralta in 1992 (142schizophrenic patients). Response levels for each item of the TLC french translation were very close to those found in the english versions. Differences in scores can be explained by clinical differences between populations studied. Factorial analyses also correspond well to such versions. In particular, after rotation, the three factorial subscores found representative of disorganisation, verbal production and verbal structure respectively are closed to those proposed by the english versions. In conclusion, the translation of Andreasen's Thought, Language and Communication (TLC) scale (Andreasen 1979) that we propose here therefore appears to exhibit metrological qualities sufficiently close to those reported in literature to permit its generalised use in France.