The most compelling case for language in non-human primates to date has involved a bonobo, Kanzi, raised from a very early age in a language-rich environment and allowed to acquire language naturally through observation (rather than being trained to elicit a reward). The underlying assumptions were that infants acquire language comprehension before language production, and that comprehension doesn’t require positive reinforcement because the listener is naturally motivated by his desire to understand the intentionality of the speaker and so predict what will happen next. Production would then follow as a natural by-product of comprehension.
Exactly what language is will always be up for debate, but it’s certainly not enough for the subject to be able to learn that certain sequences of signs or vocalizations elicit certain results. The challenge for researchers is to design ways of screening out behaviors that could simply be learned by rote or by decoding patterns. There is also the challenge of selectivity bias on the part of the researchers: favoring results that look like language while ignoring those that don’t.
Some of the markers of language researchers look for in NHP studies are: the ability to use words or signs referentially (“this symbol means this thing”) in a variety of contexts and, ultimately without context; the ability to use those symbols syntactically, spontaneously and creatively (including the invention of symbols); arbitrariness of symbols (the ability to use symbols that have no resemblance correlation to the referent); evidence that the subject possesses some theory of mind (understands others as having independent points of view and knowledge); specialization (use of the signs has no other function than to convey meaning); displacement (the ability to “talk” about something that isn’t present); productivity, and the ability to express new concepts by combining previously learned symbols; and traditional transmission (the ability to learn language by from other users).
Various studies of NHPs have demonstrated all of these markers to some extent. The studies involving Kanzi and other very young primates suggest that neuroplasticity is as much of a factor in NHPs as in humans for acquiring language. The same methodology applied to older NHPs produced very different results.
Chomsky rejects out of hand the possibility of language in NHPs because he postulates a language-specific neurological organ (as yet unidentified) that is uniquely human. The affirmation of language in NHPs threatens that theory, as does the observation that language may be a brain-wide phenomenon.