C ONSENSUS , SALIENCE AND NECESSITY IN SOCIAL REPRESENTATIONS - TECHNICAL NOTE *

The results of 7 studies show that among several (quantitatively) very salient traits of an object of social representation, some are (qualitatively) necessary, the others, not. This is made apparent by using a particular kind of questioning, the logic of which is described. The necessary traits are considered as constituting the central core of the social representation. Social representations, with or without consensus? This problem is too often formulated in quantitative terms. Of course, this is important, but insufficient. Unlike the theory of typicality, which contents itself with measuring a salient trait in social objects' study (which have a different nature than the objects the typicality studies. Cf. Semin, 1989), our results strongly suggest the existence of a character of necessity, which is suitable for some salient traits, but not for all of them. This qualitative problem is necessarily studied through numerical results but, when an extreme position gathers 96% of the subjects, a qualitative interpretation can be defended. We present here a type of empirical data which does not seem to have been much sought after other than in our laboratory. The fact that this type of data answers the questions we 1 have set ourselves within the context of a structural approach to social representations is not the point: it seems to us that this type of data can be perfectly reproduced (if used with care) 2 and should therefore be taken into consideration by any theory of social representations.