The proposed research explores a contextual perspective in multiple object situations. Specifically, I focus on the context effect created by an object associated with strong attitude on the subsequent evaluative judgment of a target object. Through five studies, I find that the context effect of a strong attitude object is greater for objects with neutral-strength attitudes than for the objects with very strong or very weak attitudes, which result in a curved relationship between the magnitude of the context effect and the strength associated with attitudes towards targets. In addition, I find that the direction of the context effect of a strong attitude object is determined by the valence of the attitudes towards the target objects: targets with positive attitudes become less positive, and those with negative attitudes become less negative. That is, the results of the context effects represent as a form of decreased extremity of the attitudes towards targets. More importantly, it is found that these differing magnitudes and directions of the context effect of an object with strong attitude finally result in evaluative space distortion. I trace the underlying process mechanism of theses effects and find that: 1) the divergent magnitudes of the context effects are the result of the differing level of comparison difficulty between a contextual object and target objects; and 2) the divergent directions of the context effects are the product of perceived uncertainty about the attitude towards target objects.

Ce travail propose une présentation et une défense du "sentimentalisme", c’est-à-dire de la théorie philosophique selon laquelle les émotions constituent notre principal accès aux valeurs, et la source de nos jugements évaluatifs. Après avoir brièvement retracé les origines historiques de ce courant (Shaftesbury, Hutcheson, Hume, et Adam Smith), j’examine la littérature psychologique et philosophique actuelle portant sur les émotions. Je propose une définition de la fonction des émotions, en soutenant que ce sont des "attitudes" corporelles et cognitives dotées d’une intentionnalité évaluative. Elles nous permettent ainsi d’apprécier la valeur des objets en fonction de leur pertinence pour nos états motivationnels, tels que nos désirs ou nos sentiments. Je procède ensuite à une analyse des conséquences épistémologiques et métaphysiques que l’on peut dériver de cette thèse. Je montre que les émotions sont susceptibles de jouer un rôle crucial dans notre connaissance des valeurs, en étudiant le débat relatif aux "conditions de correction" des émotions. Je propose de distinguer ici deux caractérisations des conditions de correction qui ne sont généralement pas explicitées dans les débats actuels, que j’appelle la conception "indépendantiste" et la conception "motivationnelle". J’argumente en faveur de la deuxième approche, et je défends la thèse selon laquelle les émotions constituent un accès fiable aux valeurs, lorsqu’elles sont basées sur des motivations qui sont elles-mêmes correctes ou appropriées. Selon cette approche, les valeurs doivent être comprises comme des propriétés réelles et relationnelles. Je m’intéresse enfin plus particulièrement au rôle que les émotions sont susceptibles de jouer dans les jugements moraux. Après avoir écarté les théories innéistes au sujet de la morale, je soutiens que les émotions sont des conditions nécessaires au développement de la moralité. / This dissertation is both a presentation and a defense of "sentimentalism", the philosophical theory according to which emotions constitute our main access to values, and the source of our evaluative judgments. After considering briefly the historical origins of this philosophical approach (Shaftesbury, Hutcheson, Hume, and Adam Smith), I review the current psychological and philosophical literature about emotions. I define the function of emotions, arguing that they are bodily and cognitive "attitudes" that possess an evaluative intentionality. They allow us to appraise the value of objects according to their relevance for our motivational states, like our desires or sentiments. I proceed then to an analysis of the epistemological and metaphysical consequences that we can draw from this thesis. I show that emotions are likely to play a crucial role in our knowledge of values, focusing on the debate about the "correctness conditions" of emotions. I propose a distinction between two characterizations of correctness conditions, which are generally not explicit in the literature, and which I name the "independence" approach and the "motivational" approach. I argue in favor of the latter and I defend the thesis that emotions constitute a reliable access to values, when they are based on motivations that are themselves correct or appropriate. According to this view, values should be conceived as real and relational properties. Finally, I consider the role which emotions may play in our moral judgments in particular. After having rejected nativist accounts of our moral competence, I claim that emotions are necessary conditions to the development of our moral life.