It is widely agreed that there is no unified species concept that that can be used to both describe and categorize the entire diversity of life. There has been extensive debate about whether the concept species accurately captures a real feature of the natural world (de Queiroz, Dupré, Ereshefsky, Hey, Mishler). I try to move beyond the existing debate by arguing that the most important task for delimiting species is to understand the processes that generate diversity. A focus on identifying the processes crucial to creating differences in lineages is important because the causal processes underlying species integrity (or lack thereof) do not act uniformly act across the diversity of life (Hart; Padial and de la Riva; Sobel). This necessitates us to be species pluralists because different kinds of species kinds are produced by the differential interaction of causal processes. I further suggest that this kinds-of-kinds pluralism does not require one to be an anti-realist about species, and that the species category is united by being the collection of entities that are produced by the lineage-splitting processes that act on populations.