Barristan - The arguments are complicated, as you'll see. Taxonomy is a human construct, but evolutionary history is not.
It is new, deeper understandings of the relatedness, evolutionary history, and fine details of genetics and morphology that lead to the name changes. The names themselves are just labels or signposts that reflect or guide us to those understandings. In this case I presented, Colobopsis & Dinomyrmex turn out to be less closely related to other "Camponotus" in the old sense than are some ants always thought to be in other genera, e.g., Echinopla. The taxonomic choice thus becomes to include Echinopla, etc. in an expanded Camponotus, or break the latter up into smaller groups that are each distinct branches on the camponotine evolutionary tree.
As for the Tetramorium situation, there is a philosophical difference about naming involved. The lumpers, Ward et al., believe all descendents of a recent common ancestor belong in the same taxon (like putting Echinopla and all the others in a single large Camponotus, which they chose not to do), while the splitters, the European group of ant taxonomists, believe that names should reflect the degree of morphological and behavioral divergence of species among these descendents, even if the divergent ones are evolutionarily right within the larger group. To their thinking, Anergates, etc. are so divergent from other modern descendents of the original Tetramorium that they deserve a unique label to indicate this level of change from the ancestral "plan".
Me, I happen to like good signposts, so I lean a bit toward the European's view. Indeed, have little doubt that most of them would support the changes in the camponotine taxonomy, even though its basis is phylogenetic monophylesis rather than anagenesis. (There's some terminiology many readers never imagined they'd need to look up.) Full disclosure, I am a way-down-the-list co-author on the Europeans' paper, though I only did English editing, not any of the the original writing.
