Abstract

In this dissertation, I critically investigate the usefulness of impossible worlds in the analysis of various modal concepts. I argue that we have reason to adopt analyses that appeal to impossible worlds of a very conservative variety, but that we should reject analyses that appeal to radically impossible worlds, or worlds where some classical contradictions are true. ;I begin by presenting and motivating two theories of possible worlds. These views are two of the most prominent and widely accepted conceptions of worlds, the first being the conception most famously endorsed by David Lewis , the second being the conception most famously endorsed by Alvin Plantinga . I next present and motivate two theories that posit both possible and impossible worlds. The first is a view suggested by Takashi Yagisawa , the second is a view often attributed to William Lycan . These latter two views are extensions of the former two views in that they take the existing Lewis and Plantinga frameworks and add impossible worlds of the same basic kind . ;I then present and critically discuss what I take to be the criteria that must be satisfied by any theory of worlds in order for it be counted as adequate. Each of the four worlds theories under consideration are evaluated with respect to these criteria; viz., plausibility of ontology, ability to accommodate core modal concepts, and ability to provide analyses of non-core modal concepts. I argue extensively for the claim that, given the particular failures of each of the four views with respect to these criteria, none should be regarded as adequate. ;In the concluding chapter, I suggest and endorse a new sort of worlds view: one that extends the Plantingean conception, but adds impossible worlds of an ultra-conservative variety. This theory of metaphysically impossible---but logically possible---worlds, I argue, comes closest to satisfying all of the criteria for adequacy