This artic1e proposes that intellectual property's c1ose relationship to property stems from the role that information costs play in the delineation and enforcement of exc1usion rights. As theorists have emphasized, the nonrivalness of information causes exc1usive rights to be more cost1y in terms of forgone use than in the law of tangible property. But if intellectual property does not solve a problem of allocation, it can play a role in allowing those who find and develop information to appropriate the retums from their rival inputs. It is on the cost side that exc1usion emerges as a possible shortcut: exc1usive rights in information are simple, indirect, and low-cost devices for solving the problem of appropriating the retums from these rival inputs. Building on a framework that identifies exclusion and govemance as complementary strategies for defining property rights, the article derives some propositions about which factors can be expected to push toward and away from exc1usion in delineating entitlements to information. The role that exc1usion plays in keeping the system of entitlements over information modular－allowing information to be hidden behind metaphorical boundaries－is both its strength and its weakness. Because exc1usion is both more cost1y and potentially more beneficial as interconnected information becomes more valuable, it is an empirical question whether we would expect more exclusion－and whether it would be desirable. The artic1e uses this information-cost theory to explain some of the basic differences between the more tort-like copyright regime and the more property-like patent law. The information-cost theory also has implications for suggestive sources of empirical evidence on the structure of entitlements, such as rules within business organizations. Intellectual property, like property in general, can be seen as (at best) a second-best solution to a complex coordination problem of attributing outputs to inputs.