20.4197, Diss: Syntax: Haugereid: 'Phrasal Subconstructions: A...'

[linguist at LINGUISTLIST.ORG]

LINGUIST List: Vol-20-4197. Tue Dec 08 2009. ISSN: 1068 - 4875.

Subject: 20.4197, Diss: Syntax: Haugereid: 'Phrasal Subconstructions: A...'

Moderators: Anthony Aristar, Eastern Michigan U <aristar at linguistlist.org>
            Helen Aristar-Dry, Eastern Michigan U <hdry at linguistlist.org>
 
Reviews: Monica Macaulay, U of Wisconsin-Madison  
Eric Raimy, U of Wisconsin-Madison  
Joseph Salmons, U of Wisconsin-Madison  
Anja Wanner, U of Wisconsin-Madison  
       <reviews at linguistlist.org> 

Homepage: http://linguistlist.org/

The LINGUIST List is funded by Eastern Michigan University, 
and donations from subscribers and publishers.

Editor for this issue: Di Wdzenczny <di at linguistlist.org>
================================================================  

To post to LINGUIST, use our convenient web form at
http://linguistlist.org/LL/posttolinguist.html.

===========================Directory==============================  

1)
Date: 08-Dec-2009
From: Petter Haugereid < petter.haugereid at ntnu.no >
Subject: Phrasal Subconstructions: A constructionalist grammar design, exemplified with Norwegian and English
 

	
-------------------------Message 1 ---------------------------------- 
Date: Tue, 08 Dec 2009 14:06:08
From: Petter Haugereid [petter.haugereid at ntnu.no]
Subject: Phrasal Subconstructions: A constructionalist grammar design, exemplified with Norwegian and English

E-mail this message to a friend:
http://linguistlist.org/issues/emailmessage/verification.cfm?iss=20-4197.html&submissionid=2230614&topicid=14&msgnumber=1
  


Institution: Norwegian University of Science and Technology 
Program: Department of Linguistics 
Dissertation Status: Completed 
Degree Date: 2009 

Author: Petter Haugereid

Dissertation Title: Phrasal Subconstructions: A constructionalist grammar
design, exemplified with Norwegian and English 

Dissertation URL:  http://urn.kb.se/resolve?urn=urn:nbn:no:ntnu:diva-5755

Linguistic Field(s): Syntax

Subject Language(s): English (eng)
                     Norwegian, Bokmål (nob)


Dissertation Director(s):
Lars Hellan

Dissertation Abstract:

Every grammatical theory has to take a stand with regard to how the
argument structure of a verb is accounted for. There are two main
approaches to argument structure in the literature; the lexicalist approach
and the constructionalist approach. The lexicalist approach typically
accounts for argument structure by means of specifications on lexical
items, and is normally associated with theories such as Government and
Binding (GB), Head-Driven Phrase Structure Grammar (HPSG), and Lexical
Functional Grammar (LFG). The constructionalist approach on the other hand
typically accounts for argument structure by means of constructions that
lexical items fit into. This approach is associated with Construction
Grammar and some versions of GB/Minimalism.

Two problems with a constructionalist approach, seen from an
implementational point of view, have been the large number of syntactic
rules associated with the grammar, and the lack of specificity in the
lexicon. Given the lack of specifications on the lexical items, the number
of possible structures in an analysis becomes difficult to handle, and
given a more or less one-to-one relation between surface structures and
constructions, the number of constructions may become unmanageable. This
has lead to a common conception of constructionalist grammars as not being
suited for implementation, and grammar implementations are almost without
exception based on a lexicalist approach to argument structure.

However, grammar implementations based on lexicalist theories are also not
without problems. One problem, which I have addressed in my thesis, is the
problem of how to account for valence alternations. Many verbs (or open
class items) are compatible with more than one argument frame. For example
the Norwegian verb presse ('press') may appear with as many as 8 different
argument frames. In principle, this means that 8 lexical entries have to be
assumed for one form presse. This is neither theoretically nor
computationally a good solution.

In my work I have addressed the valence alternation problem of grammar
implementations by assuming a constructionalist grammar design. I have met
the challenges to constructionalist grammar implementations (i. an
unmanageable number of constructions and ii. unconstrained lexical entries)
by making two proposals. The first proposal is that a construction can be
decomposed into subconstructions. Subconstructions are assumed to be
realized by functional signs (valence rules, function words, and
inflections). There are five types of subconstructions, and it is assumed
that different combinations of these subconstructions exhaust the number of
possible constructions. The argument structure of a verb is computed from
the subconstructions introduced by the functional signs in the clause. The
decomposition of constructions into subconstructions allows for a small
inventory of (binary) syntactic rules, comparable to a lexicalist grammar.
The second proposal is to introduce a mechanism which makes it possible to
constrain verbs with regard to what argument frames they are compatible
with. The two proposals make it possible to let a verb like presse have
just one lexical entry, even though it is compatible with a range of
argument frames, and at the same time let the verb be constrained so that
it does not appear in constructions that it is not compatible with. The
proposals have been implemented in a comprehensive computational grammar
for Norwegian, called Norsyg. 




-----------------------------------------------------------
LINGUIST List: Vol-20-4197