-
BulTreeBank
-
SFB 441, A5
-
MiLCA,
A4
-
Home page
-
Contact
|
|
For over a decade the linguistics department SfS has been the home to
a number of HPSG research projects. At present, alongside smaller
projects, there are three long term projects with extensive, independent
objectives, and one project which is in its initial phases.
BulTreeBank is a collaborative project of the research group with
Kiril Simov at the Bulgarian Academy of Sciences in Sofia and the
computational linguistics group at the SfS. The objective of
BulTreeBank is to create a large treebank of HPSG-type annotated Bulgarian
sentences, to develop further Bulgarian corpora. Within the framework of
this project, we plan to construct two grammars: an HPSG grammar of
Bulgarian and a grammar suitable for partial parsing. Additionally,
BulTreeBank is developing XML-based software for work with treebanks. There
is a an active exchange between the academic communities of the
SfS and the Bulgarian Academy of sciences.
The project, Grammar Formalisms and Parsing falls within the scope of
federal consortium's Media Intensive Instructive Modules in
Computer Linguistic Education, otherwise known under its anagram
of MiLCA, which is funded by the Federal Ministry for Education and
Research under the umbrella of the Federal Program "Neue Medien in der
Bildung" (New Media in Education). Its origin in this consortium lends
the nickname of this project, MiLCA (sub-project) A4. The goal of the
MilCA project A4 is to design a web-based course on HPSG, bridging the
gap between the topics of linguistic theory formation, mathematical
foundations of feature-based grammar and constraint-based
parsing. This course is being developed in three locations. In
addition to the Linguistics department in Tübingen, both Detmar Meurers
from the Ohio State University in Columbus (USA) and Gerald Penn from
the University of Toronto (Canada) are contributing to the development
of the course.
The Project A5 in Tübingen Special Research Program
(Sonderforschungsbereich) 441 is devoted to the subject of (as it is
called) distributional idiosyncrasies, from the viewpoint of their
corpus linguistic occurrences and grammar theoretical meaning. Simply
said, this research identifies and characterizes linguistic signs
which have specific requirements on the context of their occurrences,
through work with corpora as well as methods of theoretical
linguistics. There is a need to examine the special characteristics of
these signs more closely since they are traditionally neglected
in the context of universal and formally set up grammar theories. Yet
a study of these signs and their characteristics is essential for a
realistic, overall picture of language. On the basis of the empirical
data acquired in the project, grammatical theories will be examined
with regard to their suitability in describing distributional
idiosyncrasies. Our findings on distributional idiosyncrasies will
then be integrated into a corresponding theory in HPSG. In doing this,
we expect important impulses for the architecture of HPSG grammars, as
well as for their computational implementation.
As of July 2002, a project, which Frank Richter and Manfred Sailer in
Tübingen and Gerald Penn from the University of Toronto are participating
in, will be funded by the Strukturfond at the University of
Tübingen for the next year. The project explores the specification and
efficient implementation of a semantic representation language in
constraint-based grammar formalisms, appropriate for empirically and
theoretically interesting questions. The long term goal of this project
is to syntactically integrate a very expressive semantic theory in the
TRALE implementation platform for HPSG grammars, using suited programming
techniques.
|