It would seem that the fundamental basis for this allegation is that, according to Stebbins, "Chomsky has consistently denied that UG and other components of the human language capacity can be the result of natural selection acting upon random variation." In Neo-Darwinian theory, organisms evolve as the result of chance mutations which provide some sort of advantage in ensuring an organism's survival, which are then selected and passed on to the organism's descendants. Chomsky, in effect, denies that this how UG emerged. Linguist Ray Jackendoff, as quoted by Stebbins, calls Chomsky's position a "retreat to mysticism".
Moreover, there is another apparent problem with Chomsky's theory vis-a-vis neo-Darwinian evolution, noted by anthropologist John Hawks in a March 11, 2011 blog post at johnhawks.net. Remember that Chomsky argues that UG consists of both principles and parameters (see Part 1 for a more detailed explanation of this). The principles limit what is allowed in any human language, while the parameters provide a limited number of options for the ways in which languages can differ. For example, one parameter restricts languages to either allowing or not allowing null subjects (in some languages, like Japanese, declarative sentences can have no explicit subject, while other languages, like English, disallow this). However, if one (or more) of the options included in a particular parameter is not used in a particular language, it does not became part of an individual's language. Thus, how could knowledge of that option be passed on to one's descendants? It would not be available to be passed on via natural selection.
Nevertheless, despite the claim of "closet creationism," the record would show that Chomsky does hold to neo-Darwinian evolution. In fact, while he denies that UG emerged due to its usefulness in ensuring the survival of the human species, he does, according to Stebbins, allow for "the possibility of an 'exaptationist' scenario, in which UG may have arisen as a byproduct of other evolutionary processes." Thus, he clearly believes UG is the result of some sort of evolutionary mechanism. At the same time, if UG developed by a evolutionary process other than natural selection, then the logical problem involving UG parameters discussd in the paragraph above would seem to disppear.
Why then the notion that Chomsky is a "closet creationist"? I believe there may be (at least) two reasons. First, Chomsky tends to emphasize the idea that language is a unique characteristic of humans, not possesed by (presumptively) other species. This contrasts with the tendeny of advocates of neo-Darwinian evolution to stress the similarities between humans and nonhuman species. The idea that humans are different from animals is more typical of those who believe that humanity's existence is due to the actions of a Creator who created mankind "in His image"--"creationists" in a broad sense. Second, to me at least, a universal linguistic endowment consisting of principles and parameters, as postulated by Chomsky, by which the grammars of thousands of languages can be generated, seems more likely to have been the result of design than chance. In other words, UG would seem to fit very nicely within the theory of Intelligent Design (which advocates that biological organisms show signs of having been designed, without necessarily specifying the designer). To its opponents, though, Intelligent Design is simply "creationism in a tuxedo"--implying that its supporters are (irony alert!) simply more sophisticated versions of those Bible-thumping ignoramuses who believe in a literal six-day creation. In other words, Chomsky's theory may seem to imply a Designer/Creator even if he himself disavows such an idea.
What needs to be kept in mind in this discussion is that Chomsky may be totally wrong among the existence of UG--I think it is safe to say it is still the subject of considerable debate. Nevertheless, if it were to be shown that UG does exist, it could have some interesting implications for the questions of the origins of human language and of humanity itself.