In 2008 McWhorter published an article  in which he made the argument that it is not just rare that languages evolve to become simpler overall without the language being acquired by adults, but that such simplification is indeed impossible.
He phrases his thesis thus.
In the uninterrupted transmission of a human language, radical loss of complexity throughout the grammar is neither normal, occasional, nor rare, but impossible. The natural state of human language is one saddled with accreted complexity unnecessary to communication. Wherever this complexity is radically abbreviated overall rather than in scattered, local fashion, this is not just sometimes, but always caused by a sociohistorical situation in which non-native acquisition of the language was widespread enough that grammar was transmitted to new generations in a significantly simplified form. This is true not only in the extreme case of creoles, but also to a lesser but robust extent in many languages of the world.Languages become more complex over time when there is no outside influence on them. In McWhorter's words: grammars naturally "maintain a considerable level of complexity over time: simplifications occur, but are counterbalanced by complexifications due to grammaticalization, reanalysis, and new patterns created by phonetic erosion."
The languages of Keo, Rongga, and Ngadh, spoken on the island of Flores in Indonesia, are completely analytic (meaning of a very simplex grammatical form), and yet are surrounded by hundreds of related languages that are much more complex.
Austronesianists treat these languages’ analyticity as a matter of chance, of little inherent interest. But under my analysis, these languages are utterly bizarre. Why did a descendant of Proto-Austronesian wend its way into eschewing all affixation?Indonesia has a lot of spoken languages, and many of them are very much more complicated than they need to be. McWhorter explains that this is how languages often evolve, by picking up with 'complexifications' just because they can. Much to his surprise, he then encountered this small group of languages that are very simple, in contrast with hundreds of related languages. This mystery needed an explanation.
The reason languages like Keo and Ngada are so strangely streamlined on Flores is that an earlier ancestor of these languages, just as complex as its family members tend to be, was used as second language by these other people and simplified. Just as our classroom French and Spanish avoids or streamlines a lot of the "hard stuff," people who learn a language as adults usually do not master it entirely.McWhorter's speculation that it was contact with H. floresiensis that led to the simplification of Keo, Rongga, and Ngadh is to me astonishing. There's currently no evidence for that whatsoever, though, but just to consider these two species communicating is enticing. Imagine what we could learn about ourselves during such an encounter. It would be like speaking with animals (which has been done using sign language, in the case of both gorillas and chimpanzees). He notes that linguists could in that case add interspecies contact to the list of factors that affect how languages evolve.
Specifically, I would hypothesize that the little people were gradually incorporated into modern human society over time—perhaps subordinated in some way—such that modern human children were hearing the little people's rendition of the language as much as a native one.
Whether Homo floresiensis was a different species or not is very contentious at the moment. It really doesn't matter for McWhorter's linguistic explanation of language simplification, though. If H. Floresiensis was a microcephalic human population, the languages could have simplified with contact with them nontheless. However, it does add momentous marvel to the story.
For a great review of the current debate over whether H. floresiensis was a different species of Homo or not, see 'What Is the Hobbit?' by Tabitha Powledge .
Some words that I had to look up to get through this paper: affixation, agglutination, analytic, ellipsis, enclitic, ergativity, inflection, morpheme, periphrastic, typology.
View Larger Map
 John McWhorter (2008). Why does a language undress? Strange cases in Indonesia Miestamo, Matti, Kaius Sinnemäki and Fred Karlsson (eds.), Language Complexity: Typology, contact, change. 2008. xiv, 356 pp., 167-190
 Tabitha M. Powledge (2006). What Is the Hobbit? PLoS Biology, 4 (12) DOI: 10.1371/journal.pbio.0040440