Why avian systematics are no longer scientific?
Note: this is a draft pamphlet not intended for scientific publication. Names, dates, grammar, logic, political correctness, overall accuracy and other secondary details have not been checked.
In the late 19th and early 20th centuries, insect collecting (especially butterfly collecting) was almost as popular as birdwatching is today. A German company Staudinger came to dominate the commercial trade in insect specimens, and, eventually, in literature on the subject. Soon, its field guides and overview books became the main authority on systematics of Lepidoptera. To please the army of amateur collectors, its authors merged many difficult-to-identify species of Geometridae and other small moths, while splitting virtually every individual variation of large butterflies into a separate species. It took decades to correct those errors, and a few of them still linger since some groups haven't been revised ever since.
The present-day situation in avian systematics is similar, except the tendency is only to oversplit, and more than one mechanism is at work. Below is a brief overview of those mechanisms.
1. Species level
1.1. Birdwatching bias
There are probably hundreds of birdwatchers in the world for every professional ornithologist today, and the two categories increasingly mix. It's almost impossible to find a professional who doesn't maintain a life list. Since everybody wants to have a long life list, a strong bias towards splitting easy-to-see species exists. "Splitting" papers are immediately recognized without necessary skepticism, while "lumping" papers are often ignored. It's been conclusively shown years ago that Northern Parula is a subspecies of Tropical Parula, and not the most distinct one. I'm yet to see it reflected in any checklist or field guide. However, splitting of Bicknell's Thrush was immediately incorporated everywhere, even though there is substantial controversy concerning its validity.
1.2. Species concept bias
PSC is a poorly formulated species concept that essentially allows re-naming any subspecies, race, ecomorph, or isolated population as a full species. PSC species descriptions are non-falcifiable and so have nothing to do with science. No wonder many experts have never accepted it. However, many recent papers propose PSC splits either openly or while claiming them to be BSC splits. This refers to all recent splits of Cabo Verde taxa, virtually all Comoros Islands splits, and many Wallacea splits, among others.
1.3. One-character splits
Recently, many proposed splits were based exclusively on differences in vocalizations. In some groups, such differences are genetic and do carry some taxonomic importance, but there is usually no proof that they lead to reproductive isolation or are caused by anything more serious than a single-allele change. A large number of such splits have been recommended (and instantly accepted) for Strygidae, Troglodytidae and many Neotropic suboscines. Another recent trend is using mtDNA differences for splitting species. This is wrong for a number of reasons, one of them being that even a single hybridization event in distant past, no matter how insignificant for the species' phenotype and evolution, can alter mtDNA of a subset of a population and make it ripe for frivolous splitting.
1.4. Low-IQ splits
Some species in early stages of diverging exist as a number of distinct populations with complicated system of clinal variation and hybridization zones. If typical birds from these populations are easy to distinguish in the field, the whole system is likely to be split into a number of species simply because any other way of classifying it would be difficult to understand and describe in field guides. This has happened with larger gulls and stonechats, and is about to happen with wagtails.
1.5. Political splits
Some taxa are split to increase concern over their conservation, or to obtain financing for research (which is often the same), or for other obvious personal reasons. Recent examples of the former include many splits of Sunda Islands taxa, refusals to accept lumping of certain woodhopoe species despite overwhelming evidence that they are merely color morphs, and the proposed splitting of San Joaquin Valley population of Le Conte's thrasher into a new subspecies with no supporting evidence at all. I am not going to give any examples of the latter to avoid offending the perpetrators.
2. Genus level
2.1. Lost perspective splits
A level of systematic splitting in any major taxon is proportional to the amount of research directed at it. Birds are the most-studied large group of organisms, so they are oversplit relative to other groups. The way this mechanism works on genus level is usually the following. An expert begins to study a certain genus in detail. It soon becomes apparent that one distinct species or species group is a sister taxa to all others (which is usually the case for any genus). The expert proposes splitting it into a new genus. Since he is the leading (and often the only) expert on that particular group, his proposal is immediately accepted. But in the remaining genus, there is again one species (or species group) which is more distinct than others. If unchecked, this chain reaction can continue until each species is in its own genus. Bringing the whole thing back to general standards requires a major revision and a certain determination. Otherwise, very uniform groups continue to be split indefinitely. It has been going on for decades in hummingbirds and is now happening with sunbirds, thrushes, chats, and tits, among other groups.
2.2. Shape splits
Theoretically, all features should be considered taxonomically equal unless proven otherwise. In practice, some are more equal than others. In mammalogy, differences in skull and especially teeth structure have traditionally been assigned too much taxonomic importance. In ornithology, too much attention is given to bill shape and size. It's been repeatedly shown that the size and shape of the bill can change much faster than any other characters, sometimes within a single-digit number of generations. Still, in many cases differences in bill shape are automatically considered sufficient for splitting genera. Good examples include flamingoes, Spoon-billed Sandpiper, and hummingbirds.
2.3. Trojan horse splits
Cladistics insist that all taxa must be monophyletic. Recent advances in molecular systematics have shown that many large genera are paraphyletic in respect to some so-called aberrant species or groups of species. Considering how oversplit the whole thing is, the natural course of action would be to include those weird-looking taxa into the larger genera. Never happens. Virtually all remaining large genera have either been split already (Francolinus, Sterna, Columba, and many others) or are about to be split (Larus, Turdus, Zoothera etc.) In many cases, this is a two-step maneuver: first, one species is declared a separate genus, then it is "noticed" that the rest has become paraphyletic. That's what was done with Anas by initially splitting Marmaronetta. In other cases, genetic data which goes contrary to common sense and should be checked and re-checked, is admitted without scrutiny as long as it can be used as an excuse for splitting. Recent suggestion that Parus (one of the few remaining genera that are well-defined and encompass a natural group) is paraphyletic in respect to Tibetan ground-jay was immediately used to split the former. It almost feels like there is some kind of splitting conspiracy.
3. Higher levels
Birds have been considered a separate class since the first Ancient Greek classification systems. However, they are a very uniform group. Even the difference between the most extreme members of Aves (an ostrich and a hummingbird, for example) is much smaller than between a gecko and a legless lizard. Still, lizards are considered a suborder, while birds are divided into a large number of orders, some so similar that their members are difficult to tell apart even at hand. Because the system of dividing birds into orders is so poorly substantiated, it has always been riddled with controversies, and is widely ignored in checklists and field guides. A much more logical, convenient, and scientifically sound approach would be to consider birds an order in class Archosauria. Which, accidentally, is what cladisitics tell us.
The same problems exist within orders, especially in passerines. It is very obvious to any outside observer that this is a very uniform group, best treated as a single family with six subfamilies. However, just the oscine passerines are now split into 80-100 families, and the number keeps growing at an insane rate. Dozens of genera are clearly intermediate between traditional families, cannot be fitted neatly into any such classification, and inevitably end up being proclaimed to be separate families. In absence of any attempts at balanced approach, this runaway frenzy has resulted in levels of oversplitting not seen in systematics for many decades.
Combined, all the problems mentioned above not only make avian systematics oversplit on all levels, they render the whole system increasingly meaningless. Even assuming that at least some of these trends are temporary, and strict scientific standards will again be applied to bird classification, it would take a lot of time and effort to repair the damage.
Let's sum this up. There is always a tendency to oversplit, because researchers love their study subjects and enjoy having long publications lists. People use all kinds of gimmicks to do this, knowingly or unknowingly. I haven't even mentioned less honest approaches, such as splitting species in non-peer-reviewed books (Handbook of the Birds of the World is an infamous example) or in quazi-scientific journals owned by the author of the split or his organization. Unless this tendency is checked by editorial scrutiny and healthy skepticism, the splitfest will continue until the systematics become absolutely meaningless and misleading.