[b-hebrew] Generation grammar and b-hebrew

dwashbur at nyx.net dwashbur at nyx.net
Sat Jun 19 14:59:49 EDT 2010


On 19 Jun 2010 at 14:37, James Christian wrote:

> Hi, by syntax I'm assuming you mean the constituent structure of the
> sentence. e.g. things like: 
> S --> NP VP 
> VP --> V NP 
> NP --> (Det) N 
> In the LFG formalism such rules are referred to as the c-structure
> of a sentence. What are your 
> feelings about functional distinctions? e.g. consider the following
> sentences: 
> 1) Mary hit the dog on the head 
> 2) Mary hit the dog in the park 
> 3) Mary hit the dog on Monday 
> 4) Mary hit the dog on the head in the park on Monday 
> Sentences 1, 2 and 3 could be generated with similar c-structures
> but the PP's in each clearly 
> have different functional meanings. That is to say that the PP in 1
> indicates the part of the dog 
> that was hit, the PP in 2 indicates where the dog was and the PP in
> 3 indicates the time the dog 
> was hit such that we can combine the three different functions and
> produce unambiguous 
> sentence 4. 
> For this reason I find a c-structure definition of PP's to be
> inadequate. Different PP's clearly have 
> different functions. The most common functions being temporal PP's
> and locative PP's. We even 
> see that there is some grammatical structure to the way that these
> PP's can be used. Consider 
> the following sentences: 
> 5) Mary hit the dog on Monday on the head in the park 
> 6) Mary hit the dog in the park on Monday on the head 
> 7) Mary hit the dog on the head on Monday in the park 
> While the common meaning behind sentences 4, 5, 6 and 7 can be
> extracted by a native 
> speaker of English some version of the sentence is more acceptable
> than others as ranking 
> experiments would show. 

This is precisely why I go with a strict separation between syntax and semantics.  The 
internal structure of the PP is the same regardless of what it "means":

PP -> P NP

This phrase structure rule (the same as your c-structure as far as I know, I just use an older 
term because I'm an older term myself ;-)  describes virtually all prepositional phrases in BH.  
Once we start looking at function or meaning, we're out of the realm of syntax.  
Syntactically, 5, 6 and  7 differ in the order of the PP's and nothing else.  The order of these 
PP's is determined by meaning/semantics, i.e. what makes the most "sense" to the speaker 
and hearer, but that is not a syntactic decision.  Syntactically all three are well-formed: NP 
subject, VP predicate with NP direct object followed by a string of properly formed adjuncts 
(in this case, PP's).  

Does that clear it up?  I believe this approach can be vastly useful for discovering a 
generative grammar of BH because it takes us down to the very basic structure of the 
language and, as you suggested in a separate post, works from the bottom up rather than 
the reverse.

Dave Washburn


More information about the b-hebrew mailing list