The Court of Appeals for the Federal Circuit (CAFC)—the appeals court in charge of virtually all patent cases—has been fraught with controversy since its creation in 1982. To its critics, the Federal Circuit engages in puzzling behaviors, out of step with its role as an Article III appellate court. The Federal Circuit shows little deference to district courts on questions of fact and to the Patent and Trademark Office (PTO) on technical issues. It surprisingly resorts to formalistic rules in an area of the law that requires flexibility to adapt to changing technological landscapes. These criticisms have become increasingly salient, leading to calls for an end to the Federal Circuit's exclusive jurisdiction over patent appeals. Several explanations have been put forth to account for these puzzling behaviors. Yet, none can fully explain the range of unique Federal Circuit conduct. Without a full explanation for Federal Circuit behavior, however, the debate over Federal Circuit jurisdiction will remain gridlocked. Drawing upon studies from the sociology of expertise, this Article provides a model of Federal Circuit decision-making that explains and predicts Federal Circuit behavior as a product of four distinct but interrelated expert community features: (1) epistemic control, (2) codification, (3) typecasting, and (4) inability to self-coordinate. The drive that expert communities exhibit for maximal control and autonomy over their knowledge base— referred to as epistemic control—explains why the Federal Circuit is less likely to defer to solutions proposed by other expert communities, such as the PTO, than would be expected of generalist courts. Those motivations also predict that expert communities such as the Federal Circuit will be more likely to defy non-expert superior generalists, such as the Supreme Court, than would be expected under traditional accounts of behavior in judicial hierarchies. The codification feature of expert communities gives a richer account than decision over flexible standards. It predicts that the Federal Circuit will resort to rules not only to simplify technical knowledge or control subordinate communities, but also to build external legitimacy and manage internal dissent. Normatively, this model offers a path out of the gridlock by revealing a framework to evaluate and design proposals for Federal Circuit reform. To minimize the distortive effects of typecasting in the context of a centralized court, while retaining the advantages of expertise, this Article proposes the use of advisory panels to house technological, sociological, and economic expertise. Additionally, the model has important implications beyond the Federal Circuit, as it provides a novel theoretical lens to analyze the behavior of other specialized courts.