Progress up to now
Background
The project builds on previous research done at the IMS as part of the
EUROTRA-D project (1985-1992).
1995
Broad coverage in the areas of NP and VP syntax. Development of
parallel analyses for PPs, adjectives, auxiliaries, and predicative
constructions in German, French, and English.
1996
Parallel Analyses for relative clauses, comparatives and differing
types of clauses (imperative, interrogative, conditional, etc.).
Integration of a morphological analyzer and a large lexicon (40 000
entries) for the German grammar.
1997
Incorporation of a constraint ranking scheme inspired by the
Optimality Theory literature and applied as a means for filtering
ambiguity. Systematic use of means of abstraction like templates and
macros, and lately also parametrized rules. Revision of central parts
of the grammar.
1998
Refinement of NP syntax, including various types of narrow apposition
(titles, measures, etc.). Development of methods for systematic
grammar evaluation by annotating test sentences with target expressions.
1999
Refinement of AP syntax, covering all kinds of arguments as well as
comparison phrases of different types. Development of an automatic
conversion routine mapping TSNLP subcategorization information to XLE
format.
2000
Complete rewriting of the VP syntax using parametrized rules.
Application of the grammar for corpus annotation in the TIGER project. Coverage evaluation with respect to
newspaper texts (Nov. 2001: 50% parsed, 35% parsed with correct
analysis). Experiments with stochastic disambiguation of the grammar
output.
2001
Switch to IMS morphological analyzer (dmor1); use of underspecified
morphological tags. Experiments with integration of derivational
morphological analyzer (DeKo project). Start of detailed documentation of the
grammar.
2002
Implementation of DP analysis; integration of inside-out functional
uncertainty.