Casey Luskin has an interesting article on the Evolution News and Views website today. He first notes the oft-cited principle that the strongest evidences for evolution are not shared instances of excellent design, but shared instances of kludged-up, jerry-rigged, or bizarre design: the inverted retina in vertebrate eyes, the panda's thumb, or (to take today's example) the infamous recurrent largyngeal nerve, which loops around the heart as it connects the brain to the larynx; in giraffes, this takes 20 feet of nerve to accomplish what one would think should take no more than one foot.
Then, citing a paper by "pro-ID biologist" Wolf-Ekkehard Lönnig, "The Laryngeal Nerve of the Giraffe: Does it Prove Evolution?", he notes that the laryngeal nerve is not, in fact, recurrent in all human beings. In about half a percent of the population, the nerve in fact takes a direct and short route to the larynx. From this, and based on the principle that if a mutation can occur once, it has probably recurred many times in humanity's time on Earth (and that of other species), Lönnig reasons that on Darwinian grounds, the circuitous path of the nerve must be more beneficial than a more direct course. He notes that this is sensible since the nerve sends out fibers that in fact innervate several organs before it loops around to get to the larynx, so a direct route might well impair some of the other functions of the nerve.
This, in turn argues against the recurrent laryngeal nerve as evidence for common ancestry: if it is good design, then a competent Designer might well have used it.
I should point out, in passing, that Lönnig offers no actual evidence that variations in the path of the recurrent laryngeal nerve are mutations rather than developmental aberrations: the papers he cites make no mention of hereditability. If these variations were beneficial but offspring were unlikely to inherit them, they could not spread through the population through natural selection. But that's just a nit-pick; it seems very likely that some mutations could affect the path of the nerve.
A more telling response, I think, is this: why, given the multiplicity of nerves in the typical vertebrate body, should a single nerve serve the heart and larynx in all tetrapods? It seems rather bizarre that, given the immense variety of tetrapod anatomies and ecological niches, that the recurrent path of the laryngeal nerve should be optimal in all of them. Why should there not be some species, even if humans are not one of them (or giraffes another), in which a direct path did not prove best?
Common descent can explain this: the nerve, in an ancient vertebrate ancestor, did indeed follow a direct path to its ultimate destination (with, presumably, other stops along the way), as it still does in sharks today. Given that many features of anatomy are affected by many different genes each (polygeny), and that a single gene may have multiple functions (pleiotropy), a change to one trait may affect many others, and "editing" a developmental path may change more than, say, the pathway taken by one nerve -- and of course may not change features that have evolved to take advantage of, e.g. the path of that nerve. There may be no way to reach a particular improvement that does not pass through intermediates that work worse than either the present structure or the possible improved structure (this has been suggested as a reason why we still have the vermiform appendix: reductions in it would make it even more vulnerable to blockage or infection).
But how does design account for such features? As David Hume noted over two centuries ago, if we found a house many of whose features were awkward and inconvenient, we should hardly accept as an adequate defense of the architect the fact that any single change might leave the house worse off; we would wonder why the architect didn't just scrap the whole plan and come up with a better one whose function didn't depend on such awkward features. Evolution can't do that because it can't look ahead to see how mutation and selection will change a given structure in the distant future, so new designs are jerry-rigged out of and within the constraints imposed by older ones.
Meanwhile, in yet a further article on his own blog complaining about evolutionist testimony in Kitzmiller v. Dover, Cornelius Gordon addresses a similar point from a different angle. Shared ERVs and pseudogenes, after all, are evolutionist arguments of precisely the same type (except at the molecular rather than the comparative-anatomical level) as arguments from shared recurrent laryngeal nerves. Dr. Gordon has been reminded, repeatedly, that such comparisons of functionally-unrequired genetic homologies are, after all, the basis of paternity tests whose results we routinely accept in court cases affecting property and other rights. Why should we accept the results of such tests within our own species but reject them between species?
Dr. Gordon, of course, focuses on that "between species" aspect. We know that human beings produce human offspring; the only question is which human being produced this particular human offspring. He then spends a couple of paragraphs pointing out that we've never seen a cat give birth to a puppy, and reminds us that a common ancestor can give rise to both humans and gorillas is the very point at dispute. I don't find this dispositive. We know that speciation occurs. We know that both significant changes in bodily proportions and behavior, and in interfertility between populations, can occur in the course of descent. By this point, it becomes reasonable to demand what is supposed to prevent such changes from accumulating until one achieves descendants as different from one another as humans from rhesus macaques, or indeed humans from sharks?
One might as well point out that in fact not all men can sire children, but enough shared distinctive alleles between a man and his putative child put a strong burden of proof on the apparent father to show that he is one of those men. By the same token, pile up the shared pseudogenes, endogenous retroviruses, and anatomical quirks like the recurrent laryngeal nerve, and the presumption ought to be that those ancient tetrapods really were capable, given time and opportunity, of evolving into ourselves.