Moderate pluralism is a popular position in contemporary philosophy of biology. Despite its popularity, various authors have argued that it tends to slide off into a radical form of pluralism that is both normatively and descriptively unacceptable. This paper looks at the case of biological species classification, and evaluates a popular way of avoiding radical pluralism by relying on the shared aims and norms of a discipline. The main contention is that while these aims and norms may play an important role in the legitimacy of species classifications, they fail to fend off radical pluralism. It follows from this that the legitimacy of species classifications is also determined by local decisions about the aims of research and how to operationalize and balance these. This is important, I argue, because it means that any acceptable view on the legitimacy of classification should be able to account for these local decisions.