The father of modern philosophy.
Really, the opposition to natural selection by so many orthodox linguists is a scandal. The latest example is in Biolinguistics (here) in which the authors seek to refute Derick Bickerton's paper (PDF here), which I discussed in Biology Without Darwin? Bickerton's point was that even if a process is self-organizing rather than genetic, it becomes fixed in the species only through selection.
I have often thought that if I could just get a grip on the reason Chomskyans have such a distaste for natural selection, I would have a much clearer grasp of what at lies at the root of our disagreement. Chomsky is making an assumption; I hold a counter-assumption. They are so basic that I can stare them in the face and not see their radicalism. So I was relieved to find a paper by Francesco Ferretti and Ines Adornetti titled "Against Linguistic Cartesianism" (abstract here) that at last made the obvious pop out at me.
I first thought Chomsky's main objection was about genetic slowness. Chomsky takes an all or nothing approach to language that does not fit with genetic tweaking. Chomsky does not like the business of a little here, a little there, and ultimately you have language. His point is that if you say—as he does—that a defining trait of language (possibly the defining trait) is that it can generate sentences of unbounded length, then uttering five word sentences gets you no closer to unbounded length than you are at sentences of zero length.
Of course, there is the practical example that children go through a process of meaningless babble, which become one word utterances, two-word utterances, and then phrases that reflect some of the form of the surrounding language. It is many years before we get to the unbounded stage. Chomsky's reply is that for all intents and purposes, children go from not talking to talking perfectly correctly in an instant.
This reply gives the game away. Chomsky is talking about some idealized, abstract world rather than the one we actually inhabit. He takes for his defense the model of Galileo who said heavy things fall at the same speed that lighter things do. This claim drove his contemporaries batty and they did experiments which proved otherwise. Galileo said, no matter, things would fall at the same speed if air was not interfering.
So what might be the frictions that limit a baby to saying mama instead of Oh, how happy I am to see my mother who is the source of all nourishment and whose smile renders me delighted with both the day of my birth and the present hour of my life, and who …? Well, several limits come to mind. One is that the baby's vocabulary is still too small for an oration, its control over its pronunciation still too uncertain, and its working memory too short to remember what it is talking about long enough to construct such a sentence.
The frictions are real. Vocabulary growth might be cultural, but control over pronunciation, and working memory are both biological factors that had to evolve, so language cannot have appeared all at once.
Chomsky's reply is that language began as an instrument of thought and only later was "externalized" as speech or signs. I have always liked the boldness of that claim. It tends to leave one sputtering, "But … but … but … that can't be." Why not? For one thing, there is working memory. That has to function for internal language just as much as external speech and for it to spread throughout the species it would have to be selected. Nor could you argue that working memory was the thing that made unbounded language possible, because working memory still puts a limit to what we can say. We do better than toddlers, but we still have a tough time coping with The boy the man the woman loved saw ran. So that cannot be the mutation that gave us Chomskyan language.
But why this insistance that it all happened in a trice? Couldn't Chomsky's ideas work just as well if language had a long history? I have been assuming that the answer to that one was yes, and that Chomsky was just being loyal to a logical point at the expense of biological credibility. Now that I have examined the paper by Ferretti and Adornetti, however, I realize the answer to that one is no. Language cannot have a natural history similar to, say, humanity's upright walking and still be language as Chomsky defines it.
Ferretti and Adornetti trace the anti-Darwinian efforts of many linguists to Cartesian roots. Descartes done it! "As a good Cartesian, Chomsky has always expressed his complete aversion to the possibility of considering universal grammar as an adaptation of natural selection." [p. 30] I wish the authors had done a little analysis to explain why good Cartesians are anti-Darwin. Surely it is more than that Descartes lived a couple of centuries before Darwin. The authors left it up to me to remember that Descartes saw animals as mindless, soulless machines without sensations or awareness, while humans have a soul that links the human mind to God's mind. Jumping forward to modern, agnostic science we are left with human thought being qualitatively different from animal thought.
Chomsky has been very strong on this point, arguing repeatedly that human thought is qualitatively different from the rest of the animal world. But that cannot be the place where Chomsky and I differ. I agree with him that human language is qualitatively different from animal communications, but I believe we got here through an evolutionary process governed by natural selection. So why do so many Chomskyans insist otherwise?
Ferretti and Adornetti point out another feature of Chomsky's thought that echoes Descartes: "Chomsky's difficulties with evolutionary theory are tied to the fact that U[niversal] G[rammar] is a device inside the mind completely detached from the surrounding environment." There you go! Descartes' most famous proposition—Cogito ergo sum—divorces his identity from his body. Our thoughts are not constrained by reality, so why must it constrain our grammar?
That meditation prepares us for a further link to Descartes: "Despite Chomsky radically changing his conception of UG…, the nub of his thinking is still that language is a device that makes possible the combination of symbols whose functioning is completely independent of the relationship they establish with the reality they represent."*  There is another Descartes-Chomsky link. Descartes was a mathematician, and this description of language applies very well to mathematics.
*Note: Chomsky would object to that final part about 'the reality they represent." He argues that words don't represent reality, but the definition is close enough to pass in this discussion.
Descartes gets dismissed a lot these days because of his mind-body dualism, usually taken to mean that the world is composed of both matter and mind-stuff. I don't think anybody of serious consequence today believes in mind-stuff, which is too atheistic for believers and too spooky for atheists. But Ferretti and Adornetti have forced me to think a bit more deeply about what kinds of dualism exist.
There is qualitative dualism, the belief that humans are qualitatively different from other animals. Chomsky has taken plenty of criticism for standing by this principle, but I am with him on this one, and there is nothing anti-Darwinian in this proposition. Ants are qualitatively different from crickets. Both are natural; both are insects; both evolved, but ants cooperate in a way that is unmatched by anything crickets can do, even when they turn into great locust swarms.
Then there is the anti-Darwinian dualism that sees two sorts of biological reality: that which evolved through a process of natural selection and that which emerged from some other process. I am definitely on the Darwinian side, but many biolinguists are on the other side.
There is also mathematical dualism, the belief that the mathematical world and physical world both exist and follow separate laws. Chomsky is definitely this kind of a dualist, and many other smart people agree with him. Mathematical dualism asserts the independence of its subject from the other reality. If I add a gallon of pure alcohol to a gallon of distilled water, and get slightly less than two gallons of mix, that is nothing against the mathematical proposition that 1+1=2. And Chomsky has always insisted on the independence of his subject. In his first work, Syntactic Structures, the chapter immediately following the introduction is titled, "The Independence of Grammar." That chapter ends with words I underlined decades ago, "I think that we are forced to conclude that grammar is autonomous and independent of meaning…" 
Voila! The radical assumption that has been so apparent I have not noticed its premise. By autonomous, Chomsky doesn't mean merely self-governing or even self-developing. He means natural language occupies its own reality, just as the natural numbers do. Furthermore, they are independent of any meaning assigned to them culturally. Chomsky believes language (or at least its Universal Grammar) is the same way. So, of course, he is not going to expect much help from biological questions in trying to understand the nature of language. Who would turn to biology to understand mathematics?
So at last I get it, and see what is fundamentally different between my assumptions and Chomsky's. I see language as autonomous in the same way, say, the digestive system is autonomous. It follows its own rules and has its own elements, but ultimately must serve adaptive needs. I also see that if Chomsky is right, Darwin must be wrong… and, of course, if Darwin was right, it is Chomsky who is wrong.
I can imagine some readers immediately reaching for the comment button and demanding why language had to evolve if natural numbers did not. But we are not all born with a mathematical module lodged in the brain. Neither mathematics nor even counting to ten is universal among cultures. Meanwhile, says Chomsky, a Universal Grammar module comes with every newborn babe.