If I understand it right, Chomsky and Co say that there just aren't enough inputs into the system (the system here being a child during say the ages of 8 months to 3 years, as she or he goes about acquiring language) too limited to account fully for the richness (and I am guessing accuracy) of the child's language capability.
But it surely can't be the case that all the inputs to the child are well understood.
At least until a few years ago, the argument from ID proponents was that the eye was just way too complex in structure to have just appeared in entirety to fulfil its function. The argument against this (until recently, AFAIK) was not to explicitly demonstrate that each stage of the development of the eye had a particular structure which served as an adaptation. Rather, it was generally offered that just because we do not know the developmental path of the eye does not imply that some unknown mechanism fashioned the eye to serve its purpose.
Similarly, unless Chomsky and Fodor have actually listed all the inputs a child encounters, we can argue that their claim is not necessary simple because they perceive a complexity that is unexplained.
Of course Chomsky is about as far away from an ID theorist as you can get, so my point is not that he believes in mystical magical things. But rather, I think that behavioural scientists may yet produce new insights into how the environment shapes a faculty in an organism, working on nothing beyond a basic, generalised capacity that the organism is endowed with.
Carrol: the meat of the paragraph you quoted was not the "poverty of input" theses, but to list the various arguments that are available to buttress Chomsky's position. Poverty of input is one. The reducibility of the structure of all (most?) human languages to the UG is another. So on.
--ravi