[B-Greek] B-Greek syntax trees and phrase structure rules

James Tauber jtauber at jtauber.com
Sat Aug 12 00:44:03 EDT 2006


On 11/08/2006, at 10:29 PM, Jason Hare wrote:

> Does anyone have any information about Koiné phrase structure rules  
> that
> they could share? I am just beginning to study syntax trees and
> transformational grammar on my own, and I would be interested in  
> any help I
> can get. (Also, if you know, even relating to English, of some online
> resources about syntax trees and relative pronouns, I would be  
> interested in
> URLs.)

A couple of things come to mind:

Palmer, Micheal W, "Levels of Constituent Structure in New Testament  
Greek", Peter Lang.
Black and Marlett (1996), "On Generating the Greek Noun Phrase",  SIL  
UND Work Papers, Volume 40, pp. 89--105.

There's also Daryl Schmidt's dissertation "Hellenistic Greek Grammar  
and Noam Chomsky: Nominalizing Transformations" but I haven't read it.

When you use the term "transformational grammar" are you meaning it  
in the sense of Chomskyan linguistics in the last 25 years (GB, P&P,  
Minimalism) or the 25 years before that (Syntactic Structures,  
Aspects)? One thing to note with the Chomskyan programme is that the  
focus is on what languages tell us about language not what language  
tells us about languages and so the application of Chomskyan models  
to understanding of specific texts is dubious, even to to Chomskyans.  
Of course, there are many linguists who don't even think Chomskyan  
models tell us much about language either (but that's off topic :-)

Also, there is a big difference between studying Koine phrase  
structure rules and studying transformational grammar. If you are  
interested in language-specific phrase structure rules, you generally  
won't find them studied in transformational grammar (except maybe 40  
years ago).You are better off looking at something like Tree- 
Adjoining Grammars. Most linguistic frameworks eschew lists of phrase  
structure rules. Some computationally-focused work still uses them  
(or similar dependency-based rules), although for scalability  
reasons, much practical natural language processing is now  
statistical rather than rule-based.

As a side note, one of my secondary research interests is in the  
inference of syntactic rules from existing analyzed corpora like  
OpenText.org. By using existing bracketing, a lot of ambiguities that  
typically arise in syntactic parsing by machine are eliminated. And  
rule-based approaches still make sense because the corpus is just way  
too small for statistical NLP to really be helpful (although I've  
never formally run the numbers)

James Tauber






More information about the B-Greek mailing list