As so many of you have pointed out, splitting/lumping is highly subjective, based on how systematists interpret data. It seems that naming of thought-to-be unique taxa (nomenclature/taxonomy) is separate to how valid those taxa are as taxonomic units (phylogenetics/systematics). It may be important for conservation that xyz population of lemur is threatened with extinction, but is it really important to know whether it's a species, subspecies, or none of the above? If we can differentiate it from its nearest relatives, consistently or otherwise, from physical, genetic, or other characters, then its surely just an academic discussion, right?
I'm working on a project involving working out valid taxonomic names for carnivoran taxa. There have been at least 4000 described carnivorans (some invalidly) since 1758, although the vast majority have been lumped with other taxa. Many of these will be because our understanding has evolved, and we know that a black leopard is just a differently-coloured leopard and not a separate entity, for example, or Tigris regalis is just another name for the Bengal tiger, but one of the "tigris"s was changed due to the thought that you couldn't have a Tigris tigris.
But so many other names have been "lost" because somewhere along the line, one interpretation of data is that b, c, d were synonymised with a. But the person who described b, c, and d might one day be proven right, and the systematist who revised the systematics could be proven wrong.
Case in point: the IUCN Cat specialist group revamped Felidae systematics, lumping a great many subspecies that have been accepted for many years. As new evidence comes in, it's bound to happen that some of their decisions will be reversed. But if I want to know the valid scientific name of the Asiatic lion, it's no longer Panthera leo persica, as Meyer described it more than two centuries ago, but Panthera leo leo. Never mind that it has clear physical differences from other members of that subspecies - Meyer's P. l. persica is valid as it complies with the code for zoological nomenclature (ICZN). One day, perhaps after it's too late, it might come to light that the distinctiveness of Asiatic lions will be reconfirmed.
My point is that perhaps we should not be too centred on whether a population with definable characters is a phylogenetic entity, as that is bound to change.
To bring the topic back to primates, in my opinion, it doesn't really matter that they might be oversplit, so long as we don't lose what might be one day proven to be a valid entity in lumping previous splits. That's the danger in overlumping - if we were to go back to the days where there was only one orangutan, gorilla, and chimpanzee, we might not care so much about the plights of the Tapanuli orangutan, Cross River gorilla, or western chimpanzee as much. There isn't such a danger with oversplitting, so I'd rather oversplit and be wrong than overlump and be very wrong!
