Grammatical formalisms

Directorate-General For Translation European Commission

Description

Issues in migration
Information technology and telecommunications
Information - Documentation

Sujets

• Information processing
• Linguistics
• Machine translation
• Programming language
• Terminology

Informations

• Publié par: EUROPEAN-COMMISSION
• Nombre de lectures: 15
• Langue: English
• Poids de l'ouvrage: 4 Mo

Extrait

STUDIES IN MACHINE TRANSLATION
AND NATURAL LANGUAGE PROCESSING
Volume 4
GRAMMATICAL FORMALISMS:
ISSUES IN MIGRATION
Edited by
Stella MARKANTONATOU and Louisa SADLER
Commission of the European Communities Studies in machine translation
and natural language processing
Published by:
Office for Official Publications
of the European Communities Managing editor
Erwin Valentini (CEC), Luxembourg
Editorial board
Jacques Durand (University of Sa If ord, United Kingdom)
Frank van Eynde
(Nationaal Fonds voor Wetenschappelijk Onderzoek, België)
Tom C. Gerhardt (CRP-CU i CRETA, Luxembourg)
Steven Krauwer (Rijksuniversiteit Utrecht, Nederland)
Bente Maegaard (Center for Sprogteknologi, Danmark)
Karsten Strørup, (CEC, Luxembourg)
Luxembourg: Office for Official Publications of lhe European Communities, 1994
ISSN 1017-6568
© ECSC-EEC-EAEC. Brussels · Luxembourg. 1994
Primed in Germany Volume 4
Grammatical Formalisms:
Issues in Migration and Expressivity
Edited by
Stella Markantonatou and Louisa Sadler
Commission of the European Communities Volume 4
Grammatical Formalisms:
Issues in Migration and Expressivity
Editors
Stella Markantonatou
Louisa Sadler
Contents
Introduction 7
DOUG ARNOLD
Reusability - General Considerations 11
STEVE PULMAN
Expressivity of Lean Formalisms 35
JOSEF VAN GENAB1TH, STEFAN MOMMA
The ALEP_0 Formalism 61
DIETER KOHL. STEFAN MOMMA
Importation of an LFG Grammar into the ALEP_0 Formalism 7
JOSEF VAN GENABITH, STELLA MARKANTONATOU,
LOUISA SADLER, MARC VERHAGEN
English HPSG in ALEP_0 9
JOSEF VAN GENABITH. PAUL SCHMIDT
The Reuse of ET Resources 117
TONI BADIA
The Spanish Grammar 139
PAUL SCHMIDT
The German Grammar 167 STELLA MARKANTONATOU, LOUISA SADLER
Techniques and Devices 189
JOSEF VAN GENABITH, LOUISA SADLER
Reflections on Bitstrings, Generalisation and Negation 201
JOSEF VAN GENABITH
Template-Driven Phrase Structure Grammar (TPSG) 21
BIBLIOGRAPHY 215 Introduction
The discussion in this volume is mainly concerned with two related topics:1
• The reusability of grammatical resources
• The 'natural' and efficient encoding of grammatical descriptions
1 Reusability
Reusability of grammatical resources is an important idea. Practically, it has obvious
economic benefits in allowing grammars to be developed cheaply; for theoreticians it is
important in allowing new formalisms to be tested out, quickly and in depth, by providing
large-scale grammars. It is timely since substantial computational grammatical resources
exist in various N LP systems, and large scale descriptions must be quickly produced if ap­
plications are to succeed. Meanwhile, in the CL community, there is a perceptible paradigm
shift towards typed feature structure and constraint based systems and, if successful, mi­
gration allows such systems to be equipped with large bodies of linguistic description drawn
from existing resources.
In principle, there are two approaches to achieving the reuse of grammatical and lexical
resources. The first involves storing or developing resources in some theory neutral rep­
resentation language, and is impossible in the current state of knowledge. Our discussion
here focuses on reusability through migration— the transfer of linguistic resources (gram­
matical and lexical descriptions) from one computational formalism (a source formalism)
into another (a target formalism). Migration can be completely manual (as when a linguist
attempts to encode the analyses of a particular linguistic theory in some computationally
interpreted formalism), semi-automatic or automatic. The starting resource can be a paper
description or an implemented, runnable grammar.
The literature on migration is thin, and practical experience is episodic at best. Shieber's
work (e.g. [Shieber (1988)]) is relevant, but this was concerned with relations between
formalisms, rather than on migrating grammars per se. The work discussed in this volume
differs substantially from Shieber's work: the goal here was to provide grammars in the
target formalism(s) that can be directly used for further work by linguists, e.g. extending
the coverage or restructuring the description to express new insights, etc.
The idea of migration raises general questions some of which are:
• What counts as successful migration? (e.g. what properties must the output/target
description have?).
• How conceptually close must source and target be for migration to be successful?
• How far is it possible to migrate descriptions expressed in a richer formalism (e.g.
one that uses many expressive devices) into a poorer formalism?
Much of the work reported in this volume was supported by the CEC as part of the project ET10/52. Studies in NLP and MT
• How important is it that the source formalism have a well-defined semantics? How
far can difficulties in this area be off-set if the grammars/descriptions are well-
documented?
• How does the existence of non-monotonic devices within a source formalism effect
migratability?
The first chapter of this volume discusses general issues about migration between formalisms
and tries to establish criteria for the evaluation of such an enterprise. Issues concerning
differences in philosophy and expressivity between source and target formalisms are mainly
discussed in the chapters about the implemented grammars. Here, the discussion is based on
a series of experiments which constitute attempts at automatic, semi-automatic and manual
migration of implemented grammatical and lexical resources and of textbook specifications,
written in various 'styles', to the ALEP_0 formalism.
A separate chapter gives a description of ALEP.O for the convenience of the reader. The
choice of ALEP was motivated by the assumption the study would be most interesting if
the target formalism is relatively mainstream. (Of course, for practical purposes one might
want to migrate resources to a non-standard formalism, provided it is relatively easy to
understand).
As regards the 'style' and expressivity of source formalisms, the migrations discussed are
from (i) HPSG, which uses fully-typed feature structures and a variety of richly expressive
devices, (ii) ETS grammars and lexicons2; ETS is an untyped stratificational formalism
essentially using rewrite rules for feature structures, and (iii) an LFG grammar3; LFG is a
fairly standard untyped AVS formalism, with a separate CFG component.
2 The Encoding of Grammatical Descriptions
Directly related to issues of reusability but also an important issue in the issue of the
'natural' and 'efficient' encoding of grammatical descriptions. This raises questions of the
following kind:
• Which higher level expressive devices can be directly expressed in a 'lean' formalism,
which ones might be compiled down into a lean formalism, and which ones are truly
problematic? Are there any general hints that might be given for any particular class
of higher level expressive devices? When should effort be put into finding encod­
ings for richer devices, and when should effort go into simply extending the target
formalism?
• With respect to non-monotonic devices, is it possible to identify, for a given source
grammar, uses of these mechanisms that are not truly non-monotonic in nature and
could thus still be modelled inside a monotonie description?
• To what extend are macroe, preprocessors, 'refinement' and 'transfer' facilities a useful
tool in a step-wise migration from source to target?
The chapter called 'Expressivity of Lean Formalisms' discusses ways which are in fact avail­
able for encoding a number of expressive devices (booleans, multiple inheritance, sets and
'Developed at Saarbrücken, Eaaex and UMIST during the EUROTRA project. d at Stuttgart as part of the EUROTRA accompanying re»earch, »ee [Meier (1992)],