The taxon Ceratosauria was coined by Marsh (1884) in a paper on dinosaur classification in Nature (31:68-69). It was a redundant taxon, including the single genus Ceratosaurus and family Ceratosauridae, which Marsh coined in another paper published in the same year (American Journal of Science 27:329-341). Gauthier (1986) correctly cited the former paper for authorship of Ceratosauria when he brought the taxon back into common use (also Kuhn 1967). Some subsequent authors, however, have incorrectly cited Marsh’s later paper that same year (e.g. Wilson et al. 2003; Tykoski and Rowe 2004).

Rowe (1989) defined Ceratosauria as “taxa more closely related to Ceratosaurus nasicornis”. Birds were implied, but not stated, as an external specifier of a stem-based definition, and so the definition is incomplete. A complete definition must be a statement that includes a definitional type and a minimum number of specifiers (Sereno 1999, 2005). Implied taxa, no matter how reasonable the implication may be, are insufficient. Neither Gauthier (1986) nor Rowe (1989) provided a complete phlogenetic definition for Ceratosauria.

The first complete phylogenetic definition for Ceratosauria was provided by Rowe and Gauthier (1990), who provided a node-based definition that listed eight species as internal specifiers. Rowe et al. (1997) and Padian et al. (1999) subsequently referred in error to this initial definition as stem-based. Tykoski and Rowe (2004) further complicated the defitional history of this taxon by attributing a stem-based version of the definition to Rowe (1989), whose definition was incomplete.

Holtz and Padian (1995) coined the first complete stem-based definition for Ceratosauria, doing so in an abstract and using Ceratosaurus and birds as specifiers. Sereno (1998) used Coelophysis and Neornithes, and later Padian et al. (1999) used Ceratosaurus and Neornithes. Given the emerging consensus that Ceratosauria as originally conceived by Gauthier (1986) is paraphyletic, use of Coelophysis as a specifier is not recommended. Wilson et al. (2003) adopted Padian et al.’s definition, replacing Ceratosaurus with Ceratosaurus nasicornis. The active definition is a first-order revision replacing Neornithes with Passer domesticus.