Noun Phrases, Nouns and Relative Clauses
On Tuesday in Polly’s syntax class, we discussed the possible syntactic phrase structure rule for relative clauses in English. Now, this came after the explanation that proper nouns are not really different from what are regularly categorized as Noun Phrases. They appear to have the same syntactic distribution and therefore proper nouns like Mitka and Gottlob have the same category as some dog and the green man. So with this theory of NPs in mind, we moved on to two possible recursive rules for relative clauses:
NP -> NP RC or N -> N RC
Now, Polly (playing devil’s advocate, I think) pushed the argument for the N -> N RC rule, saying that the semantics really had a strong argument for this structure. And then someone in class made an observation that the N -> N RC rule really does seem to hold up to the distributional facts, while the NP -> NP RC rule didn’t. That is, while the green man who jumped over the moon is well-formed, Gottlob who drank many beers is not (or at least very archaic). I agree with the claim that proper nouns don’t seem to take RCs as well as common nouns, but I’m not convinced that this is an argument for the N -> N RC rule (or an argument against NP -> NP RC).
Because, if you recall, we started this by making an assumption about proper nouns and other NPs belonging to the same category because of supposed similarity in distribution. But if we abandon this assumption and names are names and NPs are NPs and the two are in fact not quite the same, then the fact that NPs take RCs and names don’t is just a fact of their dissimilarity in distribution and evidence that we don’t assume these two as belonging to the same category.
And semantically, this isn’t totally off. Names refer to individuals which stay constant throughout all possible words, whereas the Det-N reference to an individual is world-dependent.
I’m, of course, not completely positive that this is the right way to think about proper nouns and other NPs, but I do think that some of the arguments made in favor of the standard view need to be more carefully evaluated. (And there’s more to be said here about how you would do the semantics for the NP -> NP RC rule, because that certainly still seems to be a problem for this rule. However, there is a solution, and I’ll get to that later.)
emma :: Feb.08.2007 :: misc, syntax, semantics, linguistics :: No Comments »