Discussion Questions: 24 April

1. (CP) Kelly & Martin call language "the human duck-bill". I suppose that means that the ability to do probabilistic reasoning is like fur. I can see a correlation between language and probability, but what's the correlation between a duck-bill and fur? In other words.. Why don't rats have language?
2. (CMS) It seems like these statistical learning hypothesis really go against nativist claims. Would there be any way to reconcile statistical learning with a nativist view of the world?
3. (TS) The absence of a shock to a rat seems to have greater (conscious) consequences to the rat than the absence of double-object structures to children. Can we make this analogy more convincing to suggest, like the rat, children would have reason to notice the absence of something.
4. (TS) Is there a difference between rare constructions and idiomatic ones? Linguistic judgements might vary greatly for frequent idiomatic constructions ("I'm liking this!") but not for infrequent or rare constructions.
5. (JN) Could "statistical regularities" account for such domain-specific principles as the whole-object constraint, or novel word to novel object mapping, or must it be supplemented by these?
6. (CAS) Given all the evidence that multiple sources of information which are not in and of themselves perfectly predictive are so prevelently used, does this mean that we should all start data mining on the speach stream to children if we ever want to understand language acqusition?