Beyond Grammar

by Bod, Rens

• ISBN13: 9781575861500
• ISBN10: 157586150X
• Format: Paperback
• Copyright: 1998-06-01
• Publisher: Stanford Univ Center for the Study

Summary

During the last few years, a new approach to linguistic analysis has started to emerge. This approach, which has come to be known under various labels such as 'data-oriented parsing', 'corpus-based interpretation' and 'treebank grammar', assumes that human language comprehension and production works with representations of concrete past language experiences rather than with abstract grammatical rules. It operates by decomposing the given representations into fragments and recomposing those pieces to analyze (infinitely many) new utterances. This book shows how this general approach can apply to various kinds of linguistic representations. Experiments with this approach suggest that the productive units of natural language cannot be defined by a minimal set of rules or principles, but need to be defined by a large, redundant set of previously experienced structures. Bod argues that this outcome has important consequences for linguistic theory, leading to an entirely new view of the nature of linguistic competence.

Table of Contents

Preface	xi
Introduction: what are the productive units of natural language?	1	(11)
A probabilistic approach to language	2	(2)
Stochastic grammars and the problem of productive unit size	4	(1)
The Data-Oriented Parsing framework: productivity from examples	5	(2)
Language comprehension vs. language production	7	(2)
Evaluation of DOP models	9	(1)
Overview of this book	10	(2)
An experience-based model for phrase-structure representations	12	(12)
Representations	12	(1)
Fragments	13	(1)
Composition operations	14	(2)
Probability calculation	16	(8)
Formal Stochastic Language Theory	24	(16)
A formal language theory of stochastic grammars	24	(2)
DOPI as a Stochastic Tree-Substitution Grammar	26	(1)
A comparison between Stochastic Tree-Substitution Grammar and Stochastic Context-Free Grammar	27	(6)
Other stochastic grammars	33	(5)
Stochastic History-Based Grammar (SHBG)	33	(3)
Stochastic Lexicalized Tree-Adjoining Grammar (SLTAG)	36	(1)
Other stochastic lexicalized grammars	37	(1)
Open questions	38	(2)
Parsing and disambiguation	40	(11)
Parsing	40	(3)
Disambiguation	43	(8)
Viterbi optimization is not applicable to finding the most probable parse	43	(2)
Monte Carlo disambiguation: estimating the most probable parse by sampling random derivations	45	(4)
Cognitive aspects of Monte Carlo disambiguation	49	(2)
Testing the model: can we restrict the productive units?	51	(18)
The test environment	51	(2)
The base line	53	(1)
The impact of overlapping fragments	53	(1)
The impact of fragment size	54	(1)
The impact of fragment lexicalization	55	(2)
The impact of fragment frequency	57	(1)
The impact of non-head words	58	(6)
Overview of the derived properties and discussion	64	(5)
Learning new words	69	(12)
The model DOP2	69	(1)
Experiments with DOP2	70	(2)
Evaluation: what goes wrong?	72	(7)
The problem of unknown-category words	79	(2)
Learning new structures	81	(14)
The problem of unknown structures	81	(3)
Good-Turing: estimating the population frequencies of (un)seen types	84	(1)
Using Good-Turing to adjust the frequencies of subtrees	85	(3)
The model DOP3	88	(1)
Cognitive aspects of DOP3	88	(2)
Experiments with DOP3	90	(5)
An experience-based model for compositional semantic representations	95	(17)
Incorporating semantic interpretation	95	(14)
Assuming surface compositionality	96	(8)
Not assuming surface compositionality: partial annotations	104	(3)
The probability model of semantic DOP	107	(2)
Extending DOP to discourse and recency	109	(3)
Speech understanding and dialogue processing	112	(14)
The OVIS corpus: trees enriched with compositional frame semantics	112	(4)
Using the OVIS corpus for data-oriented semantic analysis	116	(2)
Extending DOP to dialogue context: context-dependent subcorpora	118	(1)
Interfacing DOP with speech	119	(3)
Experiments	122	(4)
Experience-based models for non-context-free representations	126	(18)
A DOP model for Lexical-Functional representations	126	(12)
Representations	126	(2)
Fragments	128	(3)
The composition operation	131	(3)
Probability models	134	(4)
Illustration and properties of LFG-DOP	138	(6)
Conclusion: linguistics revisited	144	(2)
References	146	(17)
Index	163