Biblical Hebrew Forum

[Paul Zellmer <zellmer AT cag.pworld.net.ph>]

--- Begin Message ---

• From: "Paul R. Zellmer" <zellmer AT cag.pworld.net.ph>
• To: Biblical Greek <b-greek AT franklin.oit.unc.edu>
• Subject: deriving the accents
• Date: Fri, 21 Aug 1998 10:48:03 +0800

Another old contribution from Vince.

Paul

-----

From: Vincent DeCaen <decaen AT chass.utoronto.ca>
Date: Fri, 17 Jul 1998 10:30:55 -0400 (EDT)
Subject: deriving the accents

dear Kirk et al,

re accents. the independent generative treatment of the accents has
already been looked after by Price 1990, 1996. I think there are
better ways to frame it than 19 seemingly unrelated phrase-structure
rules, but hey.

in the abstract, we can define the disjunctives in terms of a
right-headed tree

N''
/ \
N+1 N'
/ \
N+1 N

the disjunctives may be labelled D0-D3 by depth of recursion. the
basic dynamic is that there are three D1s that take specific D2s which
in turn have specific conjunctives. otherwise, each node is uniquely
and automatically labelled. the conjunctives form a melody of sorts
with indications of rise and fall / \ and rests.

but here's the question: what does the accent parse represent? not
syntax certainly, though it's related. what? why not a prosodic
representation? ie, the natural intonational contour of *spoken*
syntax. if so, we would expect a mapping that follows the syntax in
some ways, and systematically deviates in others. it's called "edge"
mapping: given a syntactic tree, key on certain branches, use edges to
create phonological phrases.

sure enough, looks like you can map syntax-to-prosody with amazing
results. there are some constraints that are well known: a D2 of
certain shape can't support left-recursion on D3 (Price calls the
result "virtual geresh"), etc, etc.

notice this is not what Janis (Ph.d. 1987) did: he used a bottom up
system that I still don't understand. BTW, anyone know what ever
happened to the Harvard grad???????

but the more interesting question is: in what way does the accent
parse deviate from expectation? is there a common denominator? is
there a common "repair strategy"?. the answer appears to be yes. that
's what I'm working on right now. but the basic idea seems to be that
(1) is always good and (2) is usually bad; and this would have a
prosodic explanation.

(1) ( big ) (small)

(2) bad! (small) ( big )

cheers.
-
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Vincent DeCaen, Ph.D.
<decaen AT chass.utoronto.ca>
Faculty of Information Studies, University of Toronto

Hebrew Syntax Encoding Initiative
http://www.chass.utoronto.ca:8080/~decaen/hsei/intro.html
c/o Deparment of Near and Middle Eastern Civilizations
4 Bancroft Ave., 2d floor, University of Toronto, Toronto ON, M5S 1A1

-
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

When I get a little money I buy books; and if any is left, I buy food
and clothes. --Erasmus
--
Paul and Dee Zellmer, Jimmy Guingab, Geoffrey Beltran
Ibanag Translation Project
Cabagan, Philippines

zellmer AT faith.edu.ph


---
B-Greek home page: http://sunsite.unc.edu/bgreek
You are currently subscribed to b-greek as: $subst('PurgeID')
To unsubscribe, forward this message to
leave-b-greek-13725G AT franklin.oit.unc.edu
To subscribe, send a message to subscribe-b-greek AT franklin.oit.unc.edu


--- End Message ---