Similarity, properties and concepts
This thesis argues that one can fruitfully think of Nelson Goodman's New Riddle of Induction as a reductio ad absurdum of a certain set of views of the relationship between similarities, on the one hand, and properties, concepts, or predicates, on the other. It argues that any view which takes similarities between particulars to be most fundamentally explained by those objects' sharing a property, satisfying a concept, or falling under a predicate leaves itself without the resources to provide a satisfying answer to a Goodmanian sceptic who proposes that inductive inferences using "grue" are equally as warranted as those using "green". I argue for an alternative view of similarity and inductive warrant which holds that the content of perceptual experience includes non-conceptual content the satisfaction conditions of which include that concept-independent similarities obtain. I argue further that it is only on the basis of that non-conceptual content that we are able satisfactorily to distinguish predicates like "grue" from those like "green." We must make such a distinction if we are to provide an acceptable account of inductive warrant. In the course of developing this view, I critique a range of mainstream, contemporary accounts of the relationship between similarities, concepts and properties, and of the role of perceptual experience in justifying empirical beliefs. Chapter 1 argues for a realist view of similarities between particulars which takes our concepts of properties to spring from our observations of those similarities. This view is contrasted with David Armstrong's universal realism, which is rejected. Chapter 2 argues that Goodman's approach to his New Riddle based on entrenchment fails, and argues that if and only if one embraces the view of similarity and concepts that I favor then the New Riddle can be reduced to traditional Humean concerns about induction. Chapters 3 through 5 discuss difficulties for Donald Davidson's approach to the New Riddle, his account of the justification of empirical belief, and his rejection of the very idea of a conceptual scheme, tracing each of these difficulties to Davidson's view that similarities must always be understood in terms of concepts under which particulars fall. Using John McDowell's Mind and World as an example, Chapter 6 argues that any account of perceptual justification of empirical belief according to which the content of perception is limited to conceptual content will fall into the New Riddle, while accounts which permit non-conceptual content can avoid this problem.