دانلود پایان نامه دکترای زبان انگلیسی AN INVESTIGATION INTO THE USE OF ARGUMENT STRUCTURE AND LEXICAL MAPPING THEORY FOR MACHINE TRANSLATION که شامل 167 صفحه و بشرح زیر میباشد:
نوع فایل : PDF قابل ویرایش
AN INVESTIGATION INTO THE USE OF ARGUMENT STRUCTURE AND LEXICAL MAPPING THEORY FOR MACHINE TRANSLATION
Abstract
In recent work on the Lexical-Functional Grammar (LFG) formalism, argument
structure (a-structure) and lexical mapping theory have been used to explain many
linguistic behaviours across languages. It has been suggested that the combination
of c-structure, f-structure and a-structure might form a suitable architecture
for Universal Grammar. If this suggestion is valid, the LFG formalism would be a
suitable linguistic model for Machine Translation (MT). This thesis reports on the
investigations carried out on using a-structure and lexical mapping theory for aiding
various sub-tasks in MT. The two investigations described in this thesis are the
abilities of a-structure and lexical mapping theory to: (1) aid different kinds of lexical
and structural disambiguations involving verbs and prepositions, and (2) act
as a suitable medium for carrying out source-to-target language transfer. Based
on the results of these investigations, this thesis also gives an evaluation of how
well a-structure and lexical mapping theory can improve the existing models of
linguistic-based MT.
Contents
1 Introduction 1
1.1 Problems of Machine Translation ... 3
1.1.1 Why are problems in MT vital to the application of real-life MT systems? 3
1.1.2 What makes MT so difficult? .... 4
1.1.3 Linguistic Problems ....5
1.1.4 Meaning Representation .... 7
1.2 Motivation and Aims of the Research ...8
1.3 Organisation of this Thesis ... 10
2 Machine Translation 12
2.1 Different Kinds of Ambiguities .... 13
2.1.1 Lexical Ambiguity ... 13
2.1.2 Structural Ambiguity ....14
2.2 Different Kinds of MT Systems .... 15
2.2.1 Direct MT systems ... 15
2.2.2 Indirect MT Systems ....16
2.3 Practical Use of some MT Systems ..... 18
2.3.1 Systran ..... 18
2.3.2 M´et´eo ..... 20
2.3.3 Discussion ...21
2.4 Methods of Transfer ..... 22
2.5 Alternative Approaches to Machine Translation .... 24
2.5.1 Sublanguage Approach .... 25
2.5.2 Statistics-based Approach ... 27
2.5.3 Example-based Approach ... 28
2.6 Conclusion .....29
3 Lexical-Functional Grammar (LFG) 32
3.1 The LFG Formalism ..... 33
3.1.1 Constituent Structure (c-structure) ... 34
3.1.2 Functional Structure (f-structure) ... 35
3.1.3 Semantic Structure (s-structure) ...41
3.2 Lexical-Functional Grammar in Machine Translation ... 43
3.2.1 Kudo and Nomura’s Lexical-Functional Transfer .... 44
3.2.2 Kaplan et al.’s approach to MT ...45
3.2.3 Her et al.’s Lexical and Idiomatic Transfer .... 47
3.3 Conclusion .....51
4 Argument Structure and Lexical Mapping Theory 54
4.1 Thematic Roles .... 55
4.1.1 Agent ..... 56
4.1.2 Beneficiary, Recipient and Experiencer ..... 57
4.1.3 Instrument ...61
4.1.4 Theme and Patient ... 61
4.1.5 Locative .... 64
4.2 Argument Structure ..... 66
4.2.1 How to establish the a-structure(s) for a verb? ..67
4.3 Lexical Mapping Theory .... 69
4.3.1 Thematic Hierarchy ....69
4.3.2 Classification of Syntactic Functions .. 70
4.3.3 Lexical Mapping Principles ..... 71
4.3.4 Well-formedness Conditions .... 76
4.4 Lexical Mapping— A Demonstration ...76
4.4.1 With the Verb ‘give’ ....76
4.4.2 With the Morpholexical Operation ‘passive’ ... 78
4.4.3 With the Morpholexical Operation ‘applicative’ .... 78
4.5 Is A-structure another variant of Case Grammar? .... 79
4.5.1 Case Grammar .... 80
4.5.2 A-structure and Case Grammar — A Comparison ...81
4.6 Conclusion .....83
5 Using A-structure and Lexical Mapping Theory for MT 84
5.1 Parsing Source Language Sentence ..... 84
5.1.1 Differentiating V + PP from Phrasal Verb + NP ..86
5.1.2 Differentiating NP with N and PP from NP + PP .... 92
5.2 Lexical Selection .... 96
5.2.1 Lexical Selection for Ergative Verbs .. 98
5.2.2 Lexical Selection for Verbs ... 101
5.2.3 Lexical Selection for Phrasal Verbs ... 106
5.3 Aiding Sentence Generation ....108
5.3.1 Verb Copying in Chinese ... 109
5.3.2 Positioning PPs within a Chinese Sentence .... 111
5.4 Discussion .....114
5.5 Conclusion .....117
6 Dealing with the Transfer of Passive Sentences 119
6.1 Using F-structure as a medium for Transfer .. 119
6.2 Passive in English ...122
6.3 Passive in Chinese ..... 126
6.4 Differences between Passive Sentences in English and in Chinese ... 129
6.5 The Transfer from English passive sentences to Chinese .... 133
6.6 Discussion .....136
6.7 Conclusion .....140
7 Conclusion and FutureWork 141
7.1 Problems in Using A-structure and Lexical Mapping Theory inMT .... 141
7.1.1 No Matching Source-and-Target Language A-structures ...142
7.1.2 Difficulty in Establishing Appropriate A-structures ... 144
7.2 What makes this investigation successful? ... 147
7.3 FutureWork ..... 148
7.3.1 Disambiguating nouns .... 149
7.3.2 Automatic extraction of a-structures from a corpus ... 150
7.3.3 Reducing the processing time .... 150
7.4 Conclusion .....151
List of Figures
1.1 A Word-for-Word Translation .... 4
2.1 Typical building blocks of a transfer-based MT system ..17
2.2 Building blocks of an interlingual MT system ..... 17
2.3 Building blocks of a multilingual MT system using the interlingual approach .. 18
2.4 A dictionary entry for transferring ‘bug’ suggested by Her et al. (1994) .... 26
3.1 C-structure for the sentence “John played Mary a tune on the violin.” ... 34
3.2 F-structure for the sentence “John tried to play the guitar.” .... 36
3.3 F-structure for the sentence “John played Mary a tune on the violin.” ... 38
3.4 C-structure and F-structure for the sentence “John died.” .... 41
3.5 C-structure & F-structure correspondence of the sentence “John died.” .. 42
3.6 S-structure for the sentence “The baby fell.” ... 42
3.7 C-structure, F-structure and S-structure correspondence of the sentence “John died.” 44
3.8 The correspondences between different structures for source and target languages inLFG ....... 46
3.9 A minimal f-structure for transferring the idiom “to kick the bucket” suggested by Her et al. (1994) ..... 50
5.1 Two potential c-structures for the word sequence “John played on words” ... 85
5.2 F-structure for “John played on words.” ... 89
5.3 F-structure for “John played on the table.” ... 89
5.4 The lexical mapping between a-structure arguments and their corresponding syntactic functions for the sentences in Table 5.1 .... 92
5.5 A possible c-structure for “John bought a book in a bookshop in Prague.” produced
by a syntax-based parser.... 93
5.6 Another possible c-structure for “John bought a book in a bookshop in Prague.” produced by a parser.... 94
5.7 The c-structure for “John saw a girl with a dog with a telescope.” ... 96
5.8 Examples of English ergative verbs with matching Chinese counterpart ... 99
5.9 Examples of English ergative verbs with different Chinese translation in transitive and intransitive cases ... 100
5.10 A-structures and sample sentences for the English verb ‘tell’ and its Chinese counterparts ..... 102
5.11 The use of a-structures for lexical selection .. 103
5.12 Some examples on lexical selection for verbs by using a-structures ... 105
6.1 English and Chinese F-structures for “Mary was killed by John.”...130
6.2 English and Chinese F-structures for “Mary was killed.”.... 131
6.3 The English and Chinese equivalents of the sentence “Mary was given a book by John” .... 132
6.4 Skeleton of Chinese F-structure for “Mary was given a book by John.”.. 135
6.5 The final Chinese F-structure for “Mary was given a book by John.”... 136
6.6 Transferring English passive sentence into Chinese using a-structure and lexical
mapping theory .... 137
List of Tables
1.1 Different meanings of some nouns ..... 7
3.1 Different cases for the Czech proper noun ‘Jan’ ....40
5.1 Some examples of different combinations of verbs and prepositions ... 88
5.2 Different Meanings of ‘look up’ .... 107
5.3 The a-structure arguments for ‘look up’ and its Chinese equivalents ... 108
دانلود پایان نامه دکترای زبان انگلیسی AN INVESTIGATION INTO THE USE OF ARGUMENT STRUCTURE AND LEXICAL MAPPING THEORY FOR MACHINE TRANSLATION
