Colorless green ideas do not sleep furiously—or do they?

Noam Chomsky’s most famous sentence is also one of the strangest ever constructed: “Colorless green ideas sleep furiously.” His point, in contrasting the syntax of this sentence with a variation, “Furiously sleep ideas green colorless,” was that while the former is grammatical and the latter is not, the latter is no less probable to appear in English language than the former.

Because he asserted his point so audaciously—claiming that no statistical model for grammaticalness would show otherwise—he left himself open to Fernando Pereira’s 2000 report on his own analysis. It turned out that a simple statistical model showed the grammatical version to be 200,000 times more likely to appear than the ungrammatical version.

All of this was described by Mark Liberman in a Language Log post which was reprinted in Far from the Madding Gerund.

So it seems Mr. Chomsky was wrong in that particular assertion, but then along came Mikael Parkvall to seal the deal, showing us once and for all that colorless green ideas do, in fact, sleep furiously. His illustration serves as the punch line for this post:

This illustration appeared on the back cover of the original UK edition of Limits of Language, but for some reason was dropped for the US edition. Since I was involved in the decision, I ought to be able to explain, but for some reason I cannot. Take another look at that beautiful rendition and tell me I’m not crazy.