I disagree. In fact, it's very hard to see how the capacity for language -- a digital infinite system -- could have evolved slowly. You don't go from one word to an infinite number of words in gradual steps. The language capacity evolved through some sort of rewiring of the brain, no doubt, but it's more likely this happened in an individual 200,000 or so years ago. — Xtrix
What are you basing this on?
Firstly, our capacity for language is not infinite. The number of possible sequences of sounds we can make is infinite, but we cannot sounds indefinitely. We acquire shared language at a limited rate, and we have a limited capacity to store information pertaining to language (the idea-symbol relationships encoded in the brain).
Secondly, words and language as we know them aren't the only kind of communication. As evolving social animals, our distant ancestors (the tree of hominids we're related to, and beyond) have been refining language capacity for eons. The cortical language centers of the brain don't need auditory information to do language processing, it's just a quicker and easier way to make more language signs than hand movements. This is why deaf people have sign languages that have nothing to do with sound or spoken words whatsoever (their brains repurpose the unused cortical space to do more visual processing (along with other sensory data).
Canines also have language processing centers in their brains; they're less powerful versions of us, and they cannot do vocalizations like us, but if they could you would be surprised how coherent they can be.
We've been eugenically selecting dogs that are able to understand us (at least to some extent), so it's not surprising that they're capable of performing basic feats of language.
You may want to conclude that if we can set dogs down the vocal language road in just a few thousand years of artificial selection, this is evidence of the sudden emergence of communication skills in our ancestral homonids, but we could also interpret this as evidence that the basic language and communication structures are far more ancient (and have been cooking for far longer) than Chomsky wants to reckon.
A sudden "re-wiring of the brain in an individual" is incredibly fantastical. It's entirely possible that a small adaptation which enhanced language capacity snowballed as the mutation spread and refined, but this optimization would be gradual (and is in fact still occurring to this day). Some people are born with cognitive profiles that are better or worse at language processing (for example, Autism has been linked with topological brain differences, and plausibly corresponding differences in cognitive profile, such as the trend of increased spatial reasoning capacity, and reduced social and linguistic capacities).
There is no miraculous infinite language capacity, just a varying spectra of complex
learning structures, all of which take ages to emerge, optimize, and evolve. The emergence of dynamic vocal chords probably played a role in the rapid optimization of preexisting communication faculties (toward,for example, increased vocabulary capacity), but this too would have occurred gradually.