Or … maybe not so radical. It doesn’t change that much.
Some readers here will recall a small debate over at SV-POW! where I argued with Mike Taylor over the effective purpose and reason behind the recognition of the name Giraffatitan  (originally named as a subgenus of Brachiosaurus by Greg Paul ). In this work, Mike Taylor described the known material of Brachiosaurus altithorax (as recognized by Riggs ), and noted the distinctive qualities between Brachiosaurus altithorax and then-Brachiosaurus brancai, the latter a sauropod taxon recovered from the latest Jurassic of Tendaguru, Tanzania .
Some astute readers will recall my arguments, which amounted to this:
1. When two fossil species are each other’s sister taxa (they are closer to one another than either is to another taxon), then having or not having the same containing name (in this case, a genus) is effectively the same thing.
2. When an author makes a claim that hasn’t been overturned through phylogenetic inference (i.e., through phenetic or cladistic analysis), or is original, this claim can be followed. In this case, Janensch  named Brachiosaurus brancai on the premise that it was very similar to Brachiosaurus altithorax, a species included in the containing taxon Brachiosaurus.
Mike Taylor responded with arguments along the lines of:
1. Recognition of genera follows various criteria, and can include morphologic disparity (even when no phenetic or cladistic analysis has been made) or some other metric of variance that discriminates two taxa. In this case, Brachiosaurus brancai is very, very different from Brachiosaurus altithorax , and its nomenclature should reflect this.
2. Some phylogenetic analysis ( being an example where the two taxa separated from one another, one being closer to Titanosauria than the other) supports recognition of the two in separate containers.
I’m sure it will come as some surprise that despite arguing against it, I agree with Mike. But my reasons do not discard my arguments. While I may have lost the argument on its face, I believe I am correct on the details (but so is Mike). Much of the discussion hinged on the very definition of the word “genus,” a rank-name employed by Linnaeus (and, as I argued here and others have argued ad nauseam, completely meaningless on its own).
So I have a fairly radical idea (although I’m sure many have seen it coming): abandon the genus. This has fewer implications than you might think, even though it alters the way we communicate about taxa — e.g., we cannot refer to “levels” of taxa that should be used without coordination to the names of those taxa, but maybe that’s a good thing in itself anyways.
The Taxonomic Couplet
The first part of this abandonment is a re-treatment of the binomen: In classical Linnaean taxonomy, the binomen is the genus-species couplet all taxa have, so that each species has a genus containing it. Without a genus name, of course, Linnaean taxonomy explodes (as does many peoples’ heads and tempers). The solution is (unlike the praenomen of others) a treatment that a species name includes always a device that Linnaeus himself employed: The descriptive couplet.
A descriptive couplet is such a part of virtually all languages that it may be difficult to recognize what I am talking about without using one. We do use them, even taxonomically, when referring to “common names” of animals, often used as placeholders for taxa themselves, such as “white rhino” and “black rhino.” These delineate concepts into a adjective-noun couplet, and have been historically employed to describe subsets of things for as long as man has communicated (and we have records).
If I were to employ this device to, say, Sauropoda, it would result in each species of sauropod would receive its own couplet, some of them new, but some of them old, employing the old system of style of naming things to the new taxonomic couplet. In this fashion, one could actually have a Brontosaurus excelsus and an Apatosaurus ajax (as they are recognized as distinct species), but also the various species of Mamenchisaurus, Omeisaurus, Diplodocus, etc.
This then follows a form where the name is capitalized, save the second, “appositional” part of the name, and each containing taxon name is capitalized, as is currently done. While removing ranks can be extreme, it’s less dramatic in the details simply due to principles such as coordinated definitions , wherein each name has a definition, and synonymies are not tracked across rank by through definitions. So, if I were to recognize a Brontosaurus excelsus rather than leave it as Apatosaurus excelsus as it recognized by virtually all sauropod workers, I would formulate a definition for Brontosaurus excelsus that is equivalent to that of Apatosaurus ajax, and this allows me to compare these definitions across sauropods.
Definitions for species would be a tad different from definitions for other taxa, which follow a form where other taxa are included only) and include the form wherein the type specimen of the original species the new couplet contains is the primary internal specifier; exclusive specifiers would include the type specimens of all other species, and do not need to be explicit. One could then even take a specimen that falls into a species’ definition and make it a type specimen, include its own taxonomic couplet, and it would automatically gain all external specifiers each other species would possess.
This would, of course, require a slew of new taxa to be formed: Not only would each species of Sauropoda gets its own taxonomic couplet, this would cause taxa otherwise grouped together (the former “genus” being a clade containing the other species) to be no longer grouped. One would note that these taxa are grouped by taxonomy only in recognition of phylogeny, so that it is phylogeny that groups them. In this manner, then the clade containing the species shifts from a “genus” name to a blank label. Such a label may need (but not require) a name; but it certainly allows some wastebasket taxa (e.g., Cetiosaurus, Megalosaurus, Troodon) to shed their unnecessary content, and permit each of the valid taxa contained therein to stand on their own. All other doubtful taxa can be positioned within the new containing clade without a couplet, using the model [“Former name” former species name].
The species has been argued to be the most essential part of a biological nomenclature, in as such as the species represents a taxon equivalent to a form of lineage or population . When it is, containing and collecting species becomes a simple matter of naming the containers (nodes) and the lineages along each container to another (branches). It seems odd, then, that in the Linnaean system, the species is uncapitalized, as though it (the important entity in the binomial) were a secondary portion of the name. One would think it would be the other way around.
If instead we treat the species as the entire bionomial, and collapse the concept of ranks, we gain a more useful (and simplistic) system: A species name, of which the first of two distinct words is capitalized, contained in a series of clades which are all capitalized. The PhyloCode even treats all other taxa with italics, as is conventional only for “genera-species” couplets in the Linnaean system, which allows us to explicitly segregate the vernacular form from taxonomic nomenclature. In this manner, one can refer to sauropods, a sauropodan, and Sauropoda. It is clear one is a collective of various members of the clade Sauropoda, the next is a specific member, and the last is the clade itself. Italicization and capitalization of all clade names makes it clear these are all taxa, and the only word in this string not capitalized is the appositional part of the taxonomic couplet, the former “species.”
Of course, we can try this, too: Brachiosaurus Altithorax. The problem with this model, making the species “just a clade,” is that it treats the species like it could treat the “genus,” a name that stands on its own and can compete for synonymy in the same form as other taxonomy. Note: under either the Linnaean system or PhyloCode, no single taxon name is the same as another; while the Linnaean system only cares about taxa as far as the “family” rank, the process and rules have been applied further broadly, all the up to the “Dominion” rank (with the only exception that the bacteriologic, botanic, and zoologic codes can use the same clade name). Within Dinosauria alone, place names for productive regions are often shared among taxa, so that nigeriensis, mongoliensis, or madagascariensis are shared among many taxa as species names for genera. This practice is at odds with the premise that the species is more important than the “genus,” despite historical value being placed more strongly on the species.
So consider this chain:
Sauropodomorpha Sauropoda Eusauropoda Neosauropoda Macronaria Titanosauriformes Brachiosauridae Brachiosaurus altithorax
Here, you see a simple progression of clades, each containing the other in some fashion, until the end. The point at which the uncapitalized taxon is reached, is the end. There are no subtaxa (well, not as such). How do we apply this system, then, when there are systems of taxa including various other ranks, such as supergenera, subgenera, superspecies, supspecies, tribes, etc.? See below.
1. As Mike argued in the old debate last year on Giraffatitan brancai, changing the system we have now, using ranks, is difficult due to the lack of any other workable system in regards to how we treat taxa.
2. Similarly, the system above, allowing each species to essentially be a “genus-species” couplet, would require a massive nomenclatural undertaking, and create thousands of new names in a short span of time.
Both of these concerns are well-founded. But it is interesting to note that of these, the PhyloCode  already represents a rankless system for clarifying the taxonomy of species, although it treats them the same as the Linnaean system: uncapitalized, and secondary to a containing clade. So far.
So in response to the second issue, compatibility is limited so far: Zoological taxa are far too numerous to support a massive immediate influx of new clade names, one for each species. This is also a major rejection of the very system I argue for due only to the fact that it promotes secondary species requiring names, but in groups such as Diptera, Protista, or Aves, there may be tens or hundreds of “species,” sub-varieties of some sort. The “easy” answer to this is to create a clade within species, a “subspecies” and simply subsume those taxa down a container from their typical points. It’s not necessary, as various systematists have regarded a variety of clades between the classic “genus” and the “established” species, and one can simply elevate these taxa as successive containers. Thus, this system may handle them by splitting them as they have been done already, and in this case, the classic use of their “genera” would be discarded as I argued above already.
As for the level of applicability, we may simply find ourselves requiring a brief set of rules for enactment: 1) A cut-off point for retroactive enactment, only for extant taxa, as far back as the first cladistic analysis of a biological system. (In this case, publication of Willi Hennig’s Phylogenetic Systematics seems a good point .) 2) Application to all fossil taxa, and not extending initially to parataxonomic systems (ootaxa, ichnotaxa).
Hopefully, a test of this would require a broad analysis of its impact, as well as application to a single group (such as Sauropoda), preferably following a phylogenetic analysis and a morphologic similarity analysis of each holotype of each species, whether it’s been synonymized with another taxon or not.
 Taylor, M. P. 2009. A re-evaluation of Brachiosaurus altithorax Riggs 1903 (Dinosauria, Sauropoda) and its generic separation from Giraffatitan brancai (Janensch 1914). Journal of Vertebrate Paleontology 29(3):787-806.
 Paul, G. S. 1988. The brachiosaur giants of the Morrison and Tendaguru with a description of a new subgenus, Giraffatitan, and a comparison of the world’s largest dinosaurs. Hunteria 2(3):1-14.
 Riggs, E. S. 1903. Brachiosaurus altithorax, the largest known dinosaur. American Journal of Science 15:299-306.
 Janensch, W. 1914. Übersicht über der Wirbeltier fauna der Tendaguru-Schichten nebst einer kurzen Charakterisierung der neu aufgeführten Arten von Sauropoden [Overview on the vertebrate animal fauna of the Tendaguru Beds with a short characterization of the many kinds of Sauropoda]. Archiv für Biontologie 3:81-110.
 The Phylocode. Preface. Online: http://www.ohio.edu/phylocode/preface.html
 Hennig, W. 1966. Phylogenetic Systematics. (University of Illinois Press, Urbana.)