Title: A Quest for Knowledge

Abstract: Is more novel research always desirable? We develop a model in which knowledge shapes society's policies and guides the search for discoveries. Researchers select a question to study and how intensely to study it. The novelty of the question determines both the value and the difficulty of discovering its answer. We show that the benefits of discoveries are nonmonotone in novelty. Because of a dynamic externality, it can be optimal to incentivize research on more distant questions than myopically optimal to improve the evolution of knowledge. One reason is that the probability of a discovery and the novelty of a question are endogenously linked. They can be complements or substitutes depending on existing knowledge. We also analyze cost reductions and research awards as instruments to incentivize research. We find that while a benefit-maximizing funder's optimal funding mix depends on her budget, distant discoveries can only be incentivized through awards.