The diversity of our intuitions about simplicity is matched only by the tenacity with which these intuitions refuse to yield to formal characterization. Our intuitions seem unanimous in favour of sparse ontologies, smooth curves, homogeneous universes, invariant equations, and impoverished assumptions. Yet recent theorizing about simplicity presents a veritable chaos of opinion. Here one finds arguments that simplicity is high probability, that it is low probability, and that it is not a probability at all. Indeed, the complexities of the problem of simplicity have led some to question the pos ... More

The diversity of our intuitions about simplicity is matched only by the tenacity with which these intuitions refuse to yield to formal characterization. Our intuitions seem unanimous in favour of sparse ontologies, smooth curves, homogeneous universes, invariant equations, and impoverished assumptions. Yet recent theorizing about simplicity presents a veritable chaos of opinion. Here one finds arguments that simplicity is high probability, that it is low probability, and that it is not a probability at all. Indeed, the complexities of the problem of simplicity have led some to question the possibility and the fruitfulness of trying to define the notion of simplicity that seems to be involved in hypothesis choice. This book tries to show that the simplicity of a hypothesis can be measured by attending to how well it answers certain kinds of questions. The more informative a hypothesis is in answering these questions, the simpler it is. The informativeness of hypotheses relative to questions is characterized by the amount of extra information they need to yield answers. The more additional information a hypothesis needs to answer a question, the less informative it is relative to that question.