I'm told that Etienne Gilson's _From Aristotle to Darwin and Back Again_ quite admirably covers the philosophical difficulties. There are some notes of Gilson's points at this link. I have also found one philosophy of science professor who is engaging the problems an overly metaphysical kind of Darwinism poses for the essence or nature of a creature. See his weblog.
He touches a bit on nominalism, as well. One of the problems is that "species" is now an equivocal term having both an ontological and a biological sense. Curiosity seems to be using "species" in a biological sense, so I think trying to make it fit into its philosphical conception is a big bad category mistake.
Nominalism as a hard theory is definitely incoherent, but it seems to me that science has to proceed using some kind of pragmatically anti-realist or nominalist philosophy. Natural science is the attempt to build a model of nature, and that model must be both open to correction and amenable to scientific consensus. Practical nominalism, instead of "hard" nominalism, is a necessary practice to fit these requirements, so long as we don't get in the habit of being nominalistic everywhere else. (An old thread on Methodical realism might be relevant, or perhaps it contradicts me completely, I forget. Worth a read though.)
Biologically speaking, neo-Darwinism tends to be blind to the organism, generally treating it as an epiphenomenon generated as a byproduct of gene dispersal and duplication. This strikes me as an area with far more potential to attack, or simply to complement, Darwinian reductionism on biological grounds.
"What is it then that the modern biologist wishes to say by declaring that it is scientific to exclude final causality from the explanation of organized living beings?"
Curiously, few Darwinians retain their agnosticism regarding the final purposes of various body systems or organs when their bodily organs are malfunctioning. Most suddenly become aware of the body's... dare I say it... proper functioning.
I suggested a probabalistic framework in which to analize it. Take the red-orange problem. The probability that a given wave of light is classified as red is a function of its length. Thus for wavelength a, the probability that it's red may be 1. Then as you move along the spectrum, that probability dereases and eventually reaches zero. Next, you move on to orange, then yellow, and so on.
You could estimate this function by showing different wavelengths of light to a large sample of people and asking whether it is red or orange.
What do you gentlemen think?
I think I understand your point after having read the links, but it seems to me that scientists must necessarily, at bottom, assume a realist philosophy in all discussions regarding creatures. It seems to me that, in fact, a biologist, when he is writing of "dolphins" in an essay on "dolphin evolution," assumes that the reader will undertand by the term "dolphin" the same thing that the author intends.
Can you give me an example of a case where "pragmatic nominalism" would serve a useful purpose?