System Description

Figure 3.1 shows the overall flow of the system when parsing a sentence; a summary of each component is presented in Table 3.1. At the heart of the system is a parser for lexicalized TAGs ([#!schabesjoshi88!#,#!schabes90!#]) which produces all legitimate parses for the sentence. The parser has two phases: Tree Selection and Tree Grafting.

  

Figure 3.1: Overview of XTAG system

 

Table 3.1: System Summary

Component	Details
Morphological	Consists of approximately 317,000 inflected items
Analyzer and	derived from over 90000 stems.
Morph Database	Entries are indexed on the inflected form and return
	the root form, POS, and inflectional information.
POS Tagger	Wall Street Journal-trained trigram tagger ([#!kwc88!#])
and Lex Prob	extended to output N-best POS sequences
Database	([#!soong90!#]). Decreases the time to parse
	a sentence by an average of 93%.
Syntactic	More than 30,000 entries.
Database	Each entry consists of: the uninflected form of the word,
	its POS, the list of trees or tree-families associated with
	the word, and a list of feature equations that capture
	lexical idiosyncrasies.
Tree Database	1094 trees, divided into 52 tree families and 218 individual
	trees. Tree families represent subcategorization frames;
	the trees in a tree family would be related to each other
	transformationally in a movement-based approach.
X-Interface	Menu-based facility for creating and modifying tree files.
	User controlled parser parameters: parser's start category,
	enable/disable/retry on failure for POS tagger.
	Storage/retrieval facilities for elementary and parsed trees.
	Graphical displays of tree and feature data structures.
	Hand combination of trees by adjunction or substitution
	for grammar development.
	Ability to manually assign POS tag
	and/or Supertag before parsing