Nous étudions la récurrence et la postérité dans la métaphysique analytique contemporaine d’un certain nombre d’arguments élaborés par Bertrand Russell afin de répondre à une question d’ontologie fondamentale : comment les éléments de l’être sont-ils reliés ? Comment peut-on répondre à cette question sans tomber sous le coup d’une régression bradleyenne ? Interroger le statut des régressions bradleyennes (régressions vicieuses ou régressions inoffensives ?), déterminer la nature des relations (particuliers ou universels ? ; relations internes ou relations externes ?), ou bien encore adopter une perspective plus épistémologique, celle des vérifacteurs, sont ici les voies explorées. Nous étudions plus particulièrement le réalisme immanent de D. Armstrong, la théorie des tropes de K. Campbell, le nominalisme de la ressemblance de G. Rodriguez-Pereyra et le réalisme modéré de D. W. Mertz. / The broad objective of this study is to examine the recurrence and the posterity of some arguments drawn up by Bertrand Russell to answer one of the main ontological question: How the elements of the being are related ? How can we answer this question without raising a bradleyan endless regress? Several strategies are assessed here: answering these questions by questioning the status of the bradleyan regresses (vicious or harmless regresses?), defining the nature of the relations (particulars or universals?, internal or external relations?) or using the epistemological perspective of the truthmakers. This study is focused on D. Armstrong’s Immanent Realism, K. Campbell’s Theory of Tropes, G. Rodriguez-Pereyra’s Resemblance Nominalism, and D. W. Mertz’s Moderate Realism.

According to natural class trope nominalism, properties are natural classes of tropes, where the "naturalness" of natural classes is taken to be primitive and unanalyzable. In this thesis I defend natural class trope nominalism from two objections: i) that the naturalness of natural classes is analyzable, and ii) that natural class trope nominalism cannot account for certain modal facts (namely, that there could have been more or fewer tropes of any given type), an objection raised by Nicolas Wolterstorff. I defend natural class trope nominalism from (i) indirectly by presenting several putative analyses (namely, those of D. M. Armstrong, Keith Campbell, and Gonzalo Rodriguez-Pereya) of natural classes and arguing that they are all deficient, thereby undermining the claim that natural classes are analyzable. Douglas Ehring has recently defended natural class trope nominalism from (ii) by developing a counterpart theory for types of tropes. However, counterpart theory is not universally accepted. So I present three non-counterpart-theoretic alternatives. The natural class trope nominalist can meet Wolterstorff's objection a) by positing existent, but uninstantiated, tropes, b) by accepting modal realism, and c) by accepting a thesis called 'transworld property exemplification'. / October 2005

According to natural class trope nominalism, properties are natural classes of tropes, where the "naturalness" of natural classes is taken to be primitive and unanalyzable. In this thesis I defend natural class trope nominalism from two objections: i) that the naturalness of natural classes is analyzable, and ii) that natural class trope nominalism cannot account for certain modal facts (namely, that there could have been more or fewer tropes of any given type), an objection raised by Nicolas Wolterstorff. I defend natural class trope nominalism from (i) indirectly by presenting several putative analyses (namely, those of D. M. Armstrong, Keith Campbell, and Gonzalo Rodriguez-Pereya) of natural classes and arguing that they are all deficient, thereby undermining the claim that natural classes are analyzable. Douglas Ehring has recently defended natural class trope nominalism from (ii) by developing a counterpart theory for types of tropes. However, counterpart theory is not universally accepted. So I present three non-counterpart-theoretic alternatives. The natural class trope nominalist can meet Wolterstorff's objection a) by positing existent, but uninstantiated, tropes, b) by accepting modal realism, and c) by accepting a thesis called 'transworld property exemplification'.

According to natural class trope nominalism, properties are natural classes of tropes, where the "naturalness" of natural classes is taken to be primitive and unanalyzable. In this thesis I defend natural class trope nominalism from two objections: i) that the naturalness of natural classes is analyzable, and ii) that natural class trope nominalism cannot account for certain modal facts (namely, that there could have been more or fewer tropes of any given type), an objection raised by Nicolas Wolterstorff. I defend natural class trope nominalism from (i) indirectly by presenting several putative analyses (namely, those of D. M. Armstrong, Keith Campbell, and Gonzalo Rodriguez-Pereya) of natural classes and arguing that they are all deficient, thereby undermining the claim that natural classes are analyzable. Douglas Ehring has recently defended natural class trope nominalism from (ii) by developing a counterpart theory for types of tropes. However, counterpart theory is not universally accepted. So I present three non-counterpart-theoretic alternatives. The natural class trope nominalist can meet Wolterstorff's objection a) by positing existent, but uninstantiated, tropes, b) by accepting modal realism, and c) by accepting a thesis called 'transworld property exemplification'.