This paper proposes a new and testable view about the nature of species and other evolving lineages, according to which they are feedback systems. On this view, it is a mistake to think gene flow, niche sharing, and trait frequency similarities between populations are among variables that interact to cause some further downstream variable that distinguishes evolving lineages from each other, some sort of "species cohesion" for example. Instead, gene flow, niche sharing, similarities between populations, and other causal variables feed into each other-instances of these at earlier times help cause instances of these same variables at later times. And any lineage-identifying cohesion just is the recurrence or cycling of these feedback relations within metapopulations over generations. Such cohesion can then be represented as variable M within multi-dimensional variable spaces, where values of M vary dynamically with the frequency and magnitude of feedback relations. Related conditions for being a species or other evolving lineage are then clarified. To argue for the development and testing of this view, the paper shows how it improves upon others.