Political arguments about biodiversity and the preservation of endangered species generally assume we know what a species is. Yet answering the question of what constitutes a "good" species has long been a confusing and controversial exercise. Within ornithological circles, the debate over the "species question" has often been described as being between "lumpers" and "splitters," between those who group separate but very similar populations of birds into a single species and those who divide such populations into full species. Generally, the lumpers employ what is known as the biological species concept, which until recently was the dominant approach to species classification. Its proponents define a species as a group that is reproductively isolated from other groups, which means that there is no substantial interbreeding in the wild between members of the groups. Splitters, on the other hand, tend to use the increasingly popular phylogenetic species concept, which defines a species as a "diagnosable" population, among which there is a pattern of ancestry and descent. A phylogenetic species is a population in which members share a distinctive, genetically traceable feature that distinguishes it from other populations.

The late Charles G. Sibley, a prominent ornithologist and one of the fomenters of a controversial revolution in avian taxonomy, could be called a splitter. He used a process known as DNA-DNA hybridization—which compares DNA from different species—to determine the relationships of the various families of birds. From his studies he concluded that many earlier classifications of the relationships of bird families were wrong. For instance, he said that vultures found in North and South America were more closely related to storks than to European vultures, and that loons and grebes, which many taxonomists had argued were closely related, were not.

Sibley's work has not been widely accepted. "What the DNA data can give you is an approximation of how different the genes of two isolated populations are," one critic has written, "but how you interpret those differences is basically arbitrary, as arbitrary as any decision made in any species concept." Sibley might not have disagreed, at least not entirely. In 1996 he wrote that "Evolution produces all degrees of genetic differences between populations of organisms, but, for practical reasons, we must limit the number of degrees we choose to 'recognize' by names. We assign names of populations and proposed definitions of taxonomic categories based on various criteria, but the 'species concept' is slippery because there are so many examples in nature of populations that refuse to fit our limited set of definitions and names."

Whatever the merits of each position, the species question undoubtedly has political and economic stakes. For example, increasing the number of species would probably increase the number needing protection as well.