Brought to you by EarthWeb
IT Library Logo

Click Here!
Click Here!

Search the site:
 
EXPERT SEARCH -----
Programming Languages
Databases
Security
Web Services
Network Services
Middleware
Components
Operating Systems
User Interfaces
Groupware & Collaboration
Content Management
Productivity Applications
Hardware
Fun & Games

EarthWeb Direct EarthWeb Direct Fatbrain Auctions Support Source Answers

EarthWeb sites
Crossnodes
Datamation
Developer.com
DICE
EarthWeb.com
EarthWeb Direct
ERP Hub
Gamelan
GoCertify.com
HTMLGoodies
Intranet Journal
IT Knowledge
IT Library
JavaGoodies
JARS
JavaScripts.com
open source IT
RoadCoders
Y2K Info

Previous Table of Contents Next


3.4.4. Linguistic Problems That Affect the Parse Operations

Independent of the grammar and from the type of tool used, all the parsers have to face some well-known linguistic problems, those which hinder a fully unambiguous syntactic analysis and, eventually, render NLP so difficult. We will examine here quickly some of these problems and the proposed solutions.

3.4.4.1. Prepositional Phrases Attachment

A problem similar to that of the "garden-path" sentences evoked in Section 3.4.3, and that is, unfortunately, really ubiquitous, is that of the "prepositional phrases attachment" (PP-attachment problem). This concerns the problem of identifying the constituent in the syntactic analysis to which a modifying phrase (or "adjunct") introduced by a preposition like "of, with, by, in ..." must be attached. We can illustrate the problem by using a famous example like: "I saw the man in the park with a telescope." In this sentence, it is not clear whether the adjunct "with a telescope" modifies "park" (the man has been seen in a particular park that is characterized by the presence of a telescope, may be as a facility for the tourists), "man" (the man had a telescope with him), or "saw" (I made use of a telescope to localize the man); moreover, "in the park" may modify "man" or "saw."

In principle, the best way of dealing with this sort of problem would be that of tackling it at the semantic or pragmatic level, making use of our "context knowledge" about the particularities of the man, the park, etc. in a specific situation. Given, however, the intrinsic uncertainty about the best way of executing semantic analysis, and the difficulties of using semantic criteria on a large scale, the problem is often dealt with at the syntactic parser level, by way of "heuristics" based only on the results of the morphological/syntactic analysis. For example, the "most right association principle" can be used to solve relative clause attachment: this means that these clauses are assumed to modify the rightmost available constituent. In the example: "I know a lady who was attacked by a woman whose husband had deserted her," the relative clause "whose husband had deserted her" is assumed to modify "a woman" and not "a lady." Another popular heuristic rule is the "attach low and parallel" principle, which suggests to attach constituents as low as possible and, if possible, in parallel with other constituents. A specialization of this rule proposes to favor the attachment of postmodifiers to the closest possible site, skipping over proper nouns; a particular case of this rule is used to settle the problems linked with the attachment of prepositional phrases introduced by "of." As J.R. Hobbs remarks, every noun can be subcategorised for "of"; using the postmodifiers rule, "of" prepositional phrases are nearly always attached to the immediate preceding word. According to Hobbs, more than 98% of the ambiguities linked with the presence of "of" can be solved using this very simple heuristic.


Previous Table of Contents Next

footer nav
Use of this site is subject certain Terms & Conditions.
Copyright (c) 1996-1999 EarthWeb, Inc.. All rights reserved. Reproduction in whole or in part in any form or medium without express written permission of EarthWeb is prohibited. Please read our privacy policy for details.