Abstract: Trademark scholars widely agree that our current system for evaluating what rights a trademark owner should have over others’ uses of their (or similar) marks is broken. Courts too readily find too many acts to be infringing even when they’re harmless or actually useful to consumers. Trademark practitioners, meanwhile, while often quite approving of broad interpretations of trademark law, widely recognize that our trademark registration system has significant practical problems. What we haven’t done is try to unite concerns over the expansion of trademark rights with concerns over the registration system and explain their relationship to each other. Registration offers some of the most challenging puzzles in trademark law. Consider: If the mark REDSKINS for a football team is disparaging and its trademark registration therefore invalid, can trademark law nonetheless protect the team against unauthorized uses of the term? This question became more than theoretical when a district court recently upheld the invalidation of the REDSKINS registrations, a ruling now on appeal and likely headed to the Supreme Court. Or suppose the PTO determines that, in the abstract, an applied-for trademark is likely to cause confusion with another previously registered mark. If the applicant decides to use the mark anyway, without a registration, should the PTO’s determination bind a federal court asked to determine whether the new mark, as actually used, causes confusion with that previously registered mark? The Supreme Court just decided this issue in a way that generated large-scale uncertainty about the new relationship between registration and infringement liability. These questions, and a number of others, highlight the need for renewed attention to trademark registration as such. Registration provides opportunities to limit trademark’s current structurelessness. Specifically, registration works best in a system that doesn’t aim to search out and extirpate every possible instance of confusion, instead recognizing multiple reasons that we might avoid fact-intensive confusion inquiries and either ban or allow certain market behaviors. Moreover, maintaining the registration system requires the investment of substantial government and private resources, which is currently almost irrelevant at the enforcement stage. Applicants and the PTO spend much time and effort crafting the equivalent of an exquisitely detailed origami crane; rather than considering the details, courts then ask the equivalent of “is this paper folded?” and move on. Not only is this a waste of resources, but it also leads courts to misunderstand the proper scope of a registration. There are a number of changes, ranging from small tweaks to sweeping statutory reforms and the rejection of the Supreme Court’s ahistorical conclusion that registration is a matter of factual accident rather than an important distinction between types of marks, that could improve the law to the benefit of trademark owners and nonowners alike.