Why defining "birds" precisely is hard
(A reply to @lyxthen-reblogs that got too long and is now its own post)
A long time ago (in the 1700s), we didn't really have any idea of how birds came about - evolutionary theory itself would have to wait another century! And, we didn't have knowledge of extinct species either, or even of the fact extinction was a thing. Carl Linnaeus, when setting up the first taxonomical classification of life, grouped modern birds in the class Aves. Mammals were grouped in Mammalia, reptiles, amphibians and cartilaginous fish in Amphibia, bony fish in Pisces, arthropods in Insecta and all other animals in Vermes.
This first classification was pretty crude, and, around 1820, scientists like de Blainville and Latreille began distinguishing reptiles from "batrachians" as separate classes. De Blainville, pointing out similarities between reptiles and birds, labelled the former as "ornithoid" (bird-like) while amphibians were "ichthyoid" (fish-like). In 1825, Latreille fully separated amphibians (Batrachia, later Amphibia) from reptiles (Reptilia).
The first major turning point for taxonomy came in the next decades, as many fossils of now-extinct creatures were unearthed. In 1842, Richard Owen coined the term Dinosauria, then uniting the recently discovered Megalosaurus, Hylaeosaurus and Iguanodon.
But it wouldn't be until the 1860s that Darwinian evolution would highlight the flaws in the earlier understanding of separate classes. In 1863, Thomas Henry Huxley would suggest uniting birds and reptiles, creating the class Sauropsida the next year. Huxley was the first to suggest birds evolved from dinosaurs, comparing the recently-discovered Archaeopteryx (1861) with Compsognathus. As cladistics didn't exist back then, no attempt at precisely extending the definition of "bird" to extinct forms was made, even though Archaeopteryx was usually called "the first bird" (Urvogel).
Unfortunately, this hypothesis would be shelved for a whole century, leading to little progress happening in terms of understanding bird evolution. It wouldn't be until the 1960s and the Dinosaur Renaissance that the links between birds and dinosaurs would be rediscovered, with birdlike theropods like Deinonychus being unearthed. This would really accelerate with the discovery of extremely well-preserved feathered dinosaurs, starting with Sinosauropteryx in 1996.
With numerous fossils showing steps of a gradual dinosaur-to-bird transition, the question of defining the "first bird" came to be asked again. To try to answer this, Jacques Gauthier coined the clade Avialae in 1986 as all dinosaurs more closely related to modern birds than to deinonychosaurs. This included Archaeopteryx, which other authors used for an alternate definition of Avialae: "the smallest clade containing Archaeopteryx and modern birds".
Still, the conflict didn't end there. Fundamentally, there were many ways to extend Aves (as defined from modern birds) to past ancestors, and, in 2001, Gauthier and de Quieroz identified four. Avemetatarsalia, defined as any archosaur closer to birds than to crocodilians (including all dinosaurs and pterosaurs!). Avifilopluma, defined as any archosaur possessing feathers homologous to bird ones. Avialae, redefined as any dinosaur able to fly (and their flightless descendants). And finally, Aves or Neornithes, the crown-group (the last common ancestor of modern birds, and its descendants).
The issues were many. Avifilopluma became mostly useless as a definition as ornithischians, then pterosaurs, were found to possess filaments homologous with bird feathers. Virtually every bird-line archosaur (with the possible exception of the little-known aphanosaurs) could likely fit in this clade, and its content was too uncertain to be reliably used.
Meanwhile, Avialae had (and continues to have) three distinct definitions. Notably, the ability for flight itself proved to be a poor definition, as bird relatives (Maniraptora, including Deinonychus, Velociraptor, Oviraptor, and many other bird-like theropods) likely evolved flight several times, from the four-winged Microraptor to the bat-winged Yi qi. Truly, most maniraptorans were extremely bird-like: wings had evolved much earlier than flight itself, with even dinosaurs like Velociraptor sporting fully feathered wings despite being unable to fly.
So, what was left? The crown group Neornithes, a vaguely defined Avialae, a more extensive Maniraptora, the stem group Avemetatarsalia, and lots of confusion. Usually, Aves is taken today as referring to either Neornithes or Avialae, although Avifilopluma/Avimetatarsalia are also in use (for instance, the Sinosauropteryx discoverers used Avifilopluma, and considered it a bird).
But none of these definitions are inherently better or worse. They are all different ways of extending a definition made for modern creatures to have it apply to past ones.
